A similar study was performed by Faisal et al. [48], who also measured single
solute isotherms of acetone, ethanol, butyric acid, and acetic acid in water on an
H-ZSM-5 (Si/Al 230). As in the study of Oudshoorn et al., the single solute isotherms of butyric acid and n-butanol were observed to be completely coinciding.
These authors also studied the pH dependency of the adsorption equilibria of butyric
acid, showing a decrease in adsorption with increasing pH, with the adsorption
capacity of butyric acid dropping from 100 mg/g at pH 4 to 20 mg/g at pH 6 (butyric
acid concentration of 0.4 wt%). The affinity for acetic acid was observed to be much
lower than for butyric acid. Again, batch measurements showed the capacity of
n-butanol to be largely unaffected by the presence of other fermentation
components.
Bowen and Vane studied the effect of acetic acid on the adsorption of bioethanol
on ZSM-5 (Si/Al 137) and silicalte-1 via liquid batch measurements [118]. These
authors mentioned a decrease in ethanol capacity from 110 mg/g to 60 mg/g when
increasing the acetic acid concentration to 3.9 wt%, leading to a pH between 2.9 and
3.6. However, at such pH values, dealumination of the zeolite framework is a risk
[141], so it is unclear whether the decrease in ethanol capacity is caused by
competitive adsorption or destruction of the zeolite framework.
2.2 Which Channel to Choose? The Adsorption Mechanism
of Alcohols on MFI Zeolites
The first studies on the adsorption mechanism of alcohols in MFI zeolites were
limited to the measurement of liquid and vapor phase isotherms, combined with
calorimetric studies [104, 106–108, 110–112, 116, 117, 138, 142–145]. With the
advent of molecular simulations, the microscopic study of the adsorption mechanism
of alcohols on MFI zeolites became possible [98, 125, 146–154]. The MFI zeolite
topology possesses three main regions for adsorption of alcohol molecules: the
zigzag and straight channels (visible in Fig. 3) and the intersections between those
channels. In the case of silicalite-1, the mechanism seems to be dominated by the
Van der Waals interactions of the hydrophobic alcohol tail with the framework
[120, 125, 142, 155, 156]. The affinity of the material for linear alcohols increases
with increasing chain-length, with a typical increase in heat of adsorption of 12 kJ/
mol per methyl group for adsorption from the vapor phase [142, 155, 156]. GC-MC
simulations point out that at low concentrations in the material pores (i.e., low
amounts adsorbed), the zigzag channels are the preferred adsorption sites for methanol and ethanol in silicalite-1 [120]. However, due to its bigger molecular size,
n-butanol molecules appear to fill both channels evenly at low equilibrium capacities, with the straight channels being the last to be completely filled, due to their
larger volume [125].
When considering ZSM-5, compared to silicalite-1, an extra factor of complexity
arises, due to the presence of Al sites and corresponding extra-framework cations.
92
B. Claessens et al.
solute isotherms of acetone, ethanol, butyric acid, and acetic acid in water on an
H-ZSM-5 (Si/Al 230). As in the study of Oudshoorn et al., the single solute isotherms of butyric acid and n-butanol were observed to be completely coinciding.
These authors also studied the pH dependency of the adsorption equilibria of butyric
acid, showing a decrease in adsorption with increasing pH, with the adsorption
capacity of butyric acid dropping from 100 mg/g at pH 4 to 20 mg/g at pH 6 (butyric
acid concentration of 0.4 wt%). The affinity for acetic acid was observed to be much
lower than for butyric acid. Again, batch measurements showed the capacity of
n-butanol to be largely unaffected by the presence of other fermentation
components.
Bowen and Vane studied the effect of acetic acid on the adsorption of bioethanol
on ZSM-5 (Si/Al 137) and silicalte-1 via liquid batch measurements [118]. These
authors mentioned a decrease in ethanol capacity from 110 mg/g to 60 mg/g when
increasing the acetic acid concentration to 3.9 wt%, leading to a pH between 2.9 and
3.6. However, at such pH values, dealumination of the zeolite framework is a risk
[141], so it is unclear whether the decrease in ethanol capacity is caused by
competitive adsorption or destruction of the zeolite framework.
2.2 Which Channel to Choose? The Adsorption Mechanism
of Alcohols on MFI Zeolites
The first studies on the adsorption mechanism of alcohols in MFI zeolites were
limited to the measurement of liquid and vapor phase isotherms, combined with
calorimetric studies [104, 106–108, 110–112, 116, 117, 138, 142–145]. With the
advent of molecular simulations, the microscopic study of the adsorption mechanism
of alcohols on MFI zeolites became possible [98, 125, 146–154]. The MFI zeolite
topology possesses three main regions for adsorption of alcohol molecules: the
zigzag and straight channels (visible in Fig. 3) and the intersections between those
channels. In the case of silicalite-1, the mechanism seems to be dominated by the
Van der Waals interactions of the hydrophobic alcohol tail with the framework
[120, 125, 142, 155, 156]. The affinity of the material for linear alcohols increases
with increasing chain-length, with a typical increase in heat of adsorption of 12 kJ/
mol per methyl group for adsorption from the vapor phase [142, 155, 156]. GC-MC
simulations point out that at low concentrations in the material pores (i.e., low
amounts adsorbed), the zigzag channels are the preferred adsorption sites for methanol and ethanol in silicalite-1 [120]. However, due to its bigger molecular size,
n-butanol molecules appear to fill both channels evenly at low equilibrium capacities, with the straight channels being the last to be completely filled, due to their
larger volume [125].
When considering ZSM-5, compared to silicalite-1, an extra factor of complexity
arises, due to the presence of Al sites and corresponding extra-framework cations.
92
B. Claessens et al.