2.5 Process Aspects and Mixture Separation
For ZSM-5 and silicalite-1, different authors have been looking at dynamic separations using a fixed-bed adsorption column, mimicking real adsorption-desorption
cycles [48, 52, 53, 97, 103, 109, 114, 119, 121, 124, 138, 143, 178]. Holtzapple and
Brown developed a thermal swing process [114, 178] based on a shell-and-tube heat
exchanger design. The silicalite-1 adsorbent was not pelletized but packed as a pure
powder. Throughout the bed, different tubes allowed to heat the adsorbent bed with
steam. By using ten of these adsorbent cartridge units in parallel, each adsorbent bed
could be heated and cooled in different steps, using hot or cool water and steam.
Furthermore, a purging step was employed, where compressed air was used to flush
the packed adsorbent bed, before complete regeneration by heating. By developing a
detailed model of the proposed process, the authors were able to show that ethanol
could be concentrated from 1 wt% to 24 wt%. Unfortunately, the proposed process
was, at the time, only competitive with conventional distillation technologies for
initial concentrations below 4 wt%, while the typical concentrations after ethanol
fermentation lie between 5 and 14 wt% [179].
In a separate paper, the same authors studied a polyethylene/silicalite composite
sintered to a heat exchanger tube [114]. In this work, the authors were able to show
that by adapting the regeneration procedure, the final ethanol purity could be
increased. By regenerating at low temperatures, most products was recovered at a
lower concentration. However, by first performing a low temperature purging step,
removing a lot of the interstitial water, and subsequently heating to high temperatures, the final ethanol concentration could be greatly increased. However, the total
ethanol recovery (e.g., the amount of ethanol recovered at high concentration) was
observed to decrease.
In more recent work, Águeda et al. developed an adsorption-drying-desorption
method to recover biobutanol from an aqueous solution using silicalite [124]. In this
combined experimental and modeling study, the regeneration of a packed silicalite-1
adsorbent bed was performed by first purging the bed with dry air, removing the
non-adsorbed liquid (the drying step). In a subsequent step, the column was purged
with dry air and simultaneously heated, allowing the removal of the adsorbed
n-butanol present in the micropores. The butanol vapors were subsequently condensed after the column. Employing this method, n-butanol with a purity of around
98 wt% could be obtained, with a recovery of 68% in the condenser. Using their
developed model, the authors were able to estimate the energy cost of their proposed
process, which lies around 3.4 MJ/kg n-butanol and is significantly lower than the
energy content of n-butanol (36 MJ/kg) and the amount of energy employed in the
traditional steam-stripping distillation step (50 MJ/kg) [32].
98
B. Claessens et al.
For ZSM-5 and silicalite-1, different authors have been looking at dynamic separations using a fixed-bed adsorption column, mimicking real adsorption-desorption
cycles [48, 52, 53, 97, 103, 109, 114, 119, 121, 124, 138, 143, 178]. Holtzapple and
Brown developed a thermal swing process [114, 178] based on a shell-and-tube heat
exchanger design. The silicalite-1 adsorbent was not pelletized but packed as a pure
powder. Throughout the bed, different tubes allowed to heat the adsorbent bed with
steam. By using ten of these adsorbent cartridge units in parallel, each adsorbent bed
could be heated and cooled in different steps, using hot or cool water and steam.
Furthermore, a purging step was employed, where compressed air was used to flush
the packed adsorbent bed, before complete regeneration by heating. By developing a
detailed model of the proposed process, the authors were able to show that ethanol
could be concentrated from 1 wt% to 24 wt%. Unfortunately, the proposed process
was, at the time, only competitive with conventional distillation technologies for
initial concentrations below 4 wt%, while the typical concentrations after ethanol
fermentation lie between 5 and 14 wt% [179].
In a separate paper, the same authors studied a polyethylene/silicalite composite
sintered to a heat exchanger tube [114]. In this work, the authors were able to show
that by adapting the regeneration procedure, the final ethanol purity could be
increased. By regenerating at low temperatures, most products was recovered at a
lower concentration. However, by first performing a low temperature purging step,
removing a lot of the interstitial water, and subsequently heating to high temperatures, the final ethanol concentration could be greatly increased. However, the total
ethanol recovery (e.g., the amount of ethanol recovered at high concentration) was
observed to decrease.
In more recent work, Águeda et al. developed an adsorption-drying-desorption
method to recover biobutanol from an aqueous solution using silicalite [124]. In this
combined experimental and modeling study, the regeneration of a packed silicalite-1
adsorbent bed was performed by first purging the bed with dry air, removing the
non-adsorbed liquid (the drying step). In a subsequent step, the column was purged
with dry air and simultaneously heated, allowing the removal of the adsorbed
n-butanol present in the micropores. The butanol vapors were subsequently condensed after the column. Employing this method, n-butanol with a purity of around
98 wt% could be obtained, with a recovery of 68% in the condenser. Using their
developed model, the authors were able to estimate the energy cost of their proposed
process, which lies around 3.4 MJ/kg n-butanol and is significantly lower than the
energy content of n-butanol (36 MJ/kg) and the amount of energy employed in the
traditional steam-stripping distillation step (50 MJ/kg) [32].
98
B. Claessens et al.