Investigation of computational upscaling of adsorption of SO₂ and CO₂ in fixed bed columns

Abstract

Fixed bed adsorption is an economical method of removing harmful gases, such as SO₂ and CO₂, from industrial flue gas. It is possible to reduce the cost and environmental impact of fixed bed adsorption by repurposing waste materials to be used as adsorbents, such as semi-coke derived from oil shale, a possible alternative to fossil fuels. Fixed bed adsorption systems are difficult and time consuming to characterize experimentally, especially on large scales. Computational fluid dynamics (CFD) can expand researchers understaning of how these systems are affected by material selection and operating conditions. This study uses CFD to characterize fixed bed adsorption of SO₂ and CO₂. The research primarily focuses on SO₂ adsorption on semi-coke, with an extension to CO₂ adsorption on commercial carbon. The CFD modeling was able to describe the amount of the pollutants each material was able to adsorb over time based on a variety of inputs on a larger scale than experimental research. The model was further able to give more detailed comparisons of materials and operating conditions than the experiments, particularly the SO₂ and semi-coke system.

Appendix

1.1 Semi-coke experiment details

1.1.1 Semi-coke preparation

The oil shale was ground in a planetary ball mill and sieved to a particle size between 0.5 and 1 mm. To prepare the “semi-coke,” the fractionated oil shale was placed in a porcelain boat and pyrolyzed in a Thermo-Scientific Lindberg Blue 1” tube furnace under high-purity nitrogen gas to mimic ex situ retorting conditions. Samples were heated at a rate of 10 °C and held at 110 °C for 30 min to drive off moisture, then heated to 600 °C at the same rate and held at this temperature for 1 h before cooling under $\text {N}_{2}$. To activate the semi-coke, 2.5 g was mixed in a 2:1 (weight) ratio of concentrated $\text {H}_{2}\text {SO}_{4}$:semi-coke and an additional 50 mL of deionized water was added, then stirred at 80 °C for 2 h. The mixture was filtered, then mixed in a 1:1 ratio with NaOH and 50 mL of deionized water, again stirred at 80 °C for 2 h. The resulting solid was filtered, dried in a laboratory oven at 80 °C overnight, and then pyrolyzed in the tube furnace at 600 °C using the same program as before.

The surface area was analyzed on a Quantachrome Autosorb iQ gas sorption analyzer. The sample was outgassed for a minimum of 6 h at 180 °C and weighed pre- and post-degas on a Sartorius semi-microbalance to the 0.1 mg. Data were fitted to the BET isotherm to find total surface area over a partial pressure range of 0–0.30 P/P0, a range suitable for microporous samples (Gregg and Sing 1982). Proximate analysis was conducted on a Mettler–Toledo Thermogravimetric Analyzer-Differential Scanning Calorimeter (TGA-DSC). Approximately 10 mg of sample was loaded into a 70 $\upmu$L alumina crucible. The sample was heated under 50 mL/min of high-purity $\text {N}_{2}$ to 110 °C and held for 30 min to establish a dry baseline. Because carbonate mineral matter begins to decompose around 620 °C, the dry sample was heated at 10 °C to 610 °C and held for 60 min; mass loss in this period was attributed to volatile matter (Williams and Ahmad 1999). The sample was further heated to 900 °C under air (50 mL/min) and held for 60 min; loss over this regime is considered to be fixed carbon. Remaining mass is termed “ash” and consists of non-oxidizable mineral matter. Scanning electron microscopy (SEM; Zeiss Supra 55VP) provided qualitative insight into the effect of the acid-base activation.

Fig. 10

SEM images of pyrolyzed White Rive Mine oil shale “semi-coke” and acid + base activated semi-coke at a magnification of × 20,000 (scale bars indicate 500 nm)

1.1.2 Experimental characterization

During pyrolysis, the kerogen contained in the oil shale devolatilizes, leaving behind a semi-carbonaceous rock with moderately enhanced porosity but lower carbon content. The White River Mine oil shale had an initial volatile matter (VM) content of $41.3 \pm 1.5$ wt% and fixed carbon (FC) of $28.0 \pm 1.4$ wt% (balance ash). Upon pyrolysis, the VM of the semi-coke drops to $9.56 \pm 0.7$ wt%, FC to $8.4 \pm 0.4$ wt%. After activation, the VM content increases to $14.22 \pm 0.8$ wt% and FC drops to $7.5 \pm 0.6$ wt%, suggesting that the acid + base treatment removes both some carbonate minerals and some of the silica preset in the inorganic ash (Gregg and Sing 1982; 2007). This is supported by the SEM images of Fig. 10, where we see upon activation a considerable reduction of the smooth amorphous regions and loss of the white semi-crystalline silica (identified via EDX) as the porous nature of the material expands. The particles, originally 0.5–1.0 mm in diameter, experienced shrinkage upon pyrolysis and activation, and after treatment ranged from approximately 0.05 to 0.4 mm in diameter as measured by optical microscopy (and were subsequently modeled as having a diameter of 0.2 mm).

Across the literature we see semi-coke samples with surface areas up to $100\text { m}^{2}/\text {g}$; however, these are usual for finely ground ($\text {d}_{p}<150\,\upmu \text { m}$) samples (An et al. 2010; Luan et al. 2010; Williams and Ahmad 1999). The as-produced semi-coke sample, given its larger, more commercially viable particle size, had a surface area of only $18.420 \pm 0.553\text { m}^{2}/\text {g}$. The combined acid/base activation treatment yielded an activated semi-coke sample with a surface area of $59.196 \pm 1.776\text { m}^{2}/\text {g}$. This is considerably higher than a typical class F coal fly ash at $5\text { m}^{2}/\text {g}$, which has long been used as a sorbent material for flue gas control (Külaots et al. 2010).