(748e) Continuous Butanol Extraction Using Supercritical Carbon Dioxide

Butanol has promise as a
drop-in biofuel or as a fuel additive that can be blended with gasoline at much
higher proportions than ethanol.  To date, economical production of butanol has
been hampered largely by its cytotoxicity, which becomes limiting at levels as
low as several weight percent.  In this work, we demonstrate n-butanol
extraction from dilute aqueous solutions used to simulate fermentation broths
using supercritical carbon dioxide (scCO₂).  In this talk, we will
focus on process engineering of the n-butanol extraction.  Our process
engineering work is complemented by microbiological and metabolic engineering
work that has identified a scCO₂-tolerant bacterial strain and genetically
modified it for biofuel production.  Compared to other butanol recovery
approaches, scCO₂-extraction has sterilization and potential energy
balance advantages; because scCO₂ selectively extracts butanol
instead of water, a highly concentrated butanol stream can be recovered which
requires minimal post-processing purification.

Specifically in this work,
we studied n-butanol extraction performance and material/energy balance
analysis.  For butanol extraction performance, we used batch-wise extraction to
study the effects of initial n-butanol concentration, extraction vessel
pressure, and scCO₂ volumetric flow rate on n-butanol extraction
rate.  Additionally, we modeled the data using a standard liquid-liquid mass
transfer model to determine values for the mass transfer coefficient, k_la. 
Figure 1 shows representative extraction data compared to the model predictions
using fit mass transfer coefficients.  In all cases, the mass transfer model
adequately described the experimental data.  Best-fit values of k_la
did not vary within our estimated limits of uncertainty for variation in
extraction pressure (from 10.3 to 13.7 MPa – the range over which the scCO₂-tolerant
bacterial strain has exhibited growth); therefore, we conclude that operation
at lower pressures should be favored to achieve better process economy.  Similar
analysis was performed to interpret the mass transfer coefficient from
correlations developed for gas-liquid and liquid-liquid extraction and compared
to interfacial area results obtained from scCO₂-droplet size
analysis.  In addition to n-butanol, we investigated extraction of other
alcohol products (e.g., iso-butanol, pentanol, and hexanol) as these are
potential metabolic targets for a modified scCO₂-tolerant organism. 
Finally, we studied the extraction of the aldehyde intermediate, butanal, to
set limits on its accumulation in the fermentation products.

Figure 1. Representative
n-butanol extraction data obtained at 10.3 MPa and 40 °C.

In addition to extraction
experiments and modeling, we performed preliminary material and energy balance
calculations of a conceptual continuous process.  Thermodynamic and extraction
data were used to model n-butanol removal from the fermenter and fermentation
kinetics were modeled based on activity of similar organisms.  The
Peng-Robinson equation of state was used for modeling thermodynamic
properties.  To improve energy efficiency, the scCO₂ extraction
stream was only partially de-pressurized from 10 to 5 MPa, resulting recovery
of  n-butanol at 79 wt% purity.  Energy was recovered during the
de-pressurization, which partly offset the energy required for CO₂
compression.  Analysis was performed assuming the availability of either 10 or
0.1 MPa CO₂.  Overall, we found that n-butanol with 95% purity could
be produced at an energy cost of 3.9 MJ kg^-1 (assuming that a 1 bar
CO₂ source is available) or 3 MJ kg^-1 (assuming a 10 MPa
CO₂ source).  These figures compare well with existing methods of
butanol recovery, which range from 9 to more than 20 MJ kg^-1, and
the lower heating value of the butanol fuel itself (33.1 MJ kg^-1). 
Further analysis is needed to account for heat loss and heat integration and to
optimize process efficiency.