were obtained with a feed temperature of 50 °C. A flux of
0.392 kg m
−2 h
−1 and a separation factor of 3.98 in favor to
the methanol permeation were measured. By increasing the
methanol concentration in the methanol/MTBE feed composition, the separation factor decreases. Besides, it was
assumed that MTBE could act as a plasticizer in the membrane and would consequently improve the permeability of
both compounds. Later, in another article, the authors went
further in the evaluation of the pervaporation performances
(Villegas et al. 2016). This time, the authors chemically
modified the P(3HB) dense membranes by plasma polymerization with acrylic acid. After membrane modification,
due to the additional layer formed and additional flux
resistance, the flux was decreased. Nevertheless, the separation factor was significantly improved up to 18.6. The
latter is explained by a higher affinity between the plasma
polymerized acrylic acid layer and methanol.
Related to pervaporation applications, Galego et al.
studied the swelling capacity, vapor permeability and
selectivity of PHAs films with different HV content (Galego
et al. 2002). The swelling percentage to water–ethanol
mixtures and the water vapor permeability of P(3HB) was
higher than P(3HB3HV) (22 mol% HV). It is discussed as a
consequence of the rougher surface of P(3HB) films and the
better ability of P(3HB) to make hydrogen bonding interactions with water molecules.
3.2.2 Air Filtration
Nicosia et al. have made PLA/P(3HB) filters intended for air
filtration applications (Nicosia et al. 2015). The filters were
made by electrospinning of a PLA/P(3HB) mixture in
chloroform. The performances were quantified by measuring
the penetration of sodium chloride aerosol particles ranging
from 20 to 600 nm. The performances could be improved by
superposing several thin layers to reduce the overall packing
density of the structure. The results highlight a good
potential for air filtration applications. The best collection
efficiency was measured at 98.5% for 0.3 µm particles.
Additionally, these membranes could be modified by loading of didecyldimethylammonium nitrate in order to improve
their antimicrobial activity.
3.2.3 Liquid Filtration
Mas et al. studied PHAs-based membranes intended for
microfiltration applications (Mas et al. 1996). The membranes were made with P(3HB3HV) having different HV
ratio. For the membrane fabrication, they used the EIPS
coupled with the NIPS. For the dope solution, various
compositions of chloroform/THF mixtures were used. THF
acts as a non-solvent in the dope solution. The non-solvent
bath was composed of a mixture of ethanol/water (4/1 v/v).
The water permeability and ethanol permeability were
measured. The pores size was analyzed by image analysis.
Whether for P(3HB) or for P(3HB3HV) (22 mol% HV), the
best permeabilities were obtained for a dope solution containing 8% of THF. The permeability went up to 600 L
m
−2 h
−1 bar
−1 but no rejection test was performed. For THF
concentrations above 8%, the proper solubility of the polymer was hindered, leading to the densification of the membrane. Among the membranes characterized, the pores size
varied from 0.25 to 2 µm.
Keawsupsak et al. have made PLA/P(3HB3HV) blend
membranes to test them on water permeability and rejection
of bovine serum albumin (Keawsupsak et al. 2014). Membranes were produced by NIPS using NMP as a solvent and
water as a non-solvent. At a small amount of P(3HB3HV)
added to PLA, the membrane exhibits a better permeability
value of 65.2 L m
−2 h
−1 bar
−1 , but it remains quite low. In
that case, the BSA rejection was measured at 78.7%. The
membrane showed a finger-like structure.
Composite membranes made of P(3HB), calcium alginate
and carboxyl multi-walled carbon were made by electrospinning for NF applications (Guo et al. 2016). Their performances were measured by water permeability, separation
of emulsified oil–water solutions and rejection of brilliant
blue. One membrane demonstrated an excellent antifouling
property associated with a pure water flux around 35 L
m
−2 h
−1 at 2 bar, a 98.2% rejection of brilliant blue and a
98% rejection of oil.
Finally, P(3HB3HV) (3 mol% HV)-based membranes
with tailored structures were made in our research group at
the ISCR (Institut des Sciences Chimiques de Rennes,
France), in the Chemistry and Process Engineering team.
The membranes were fabricated via two different techniques,
NIPS or EIPS. By NIPS, the NMP was used as solvent and
water was used for the non-solvent bath. By EIPS, chloroform was employed as volatile solvent. The dope solutions
were casted under controlled conditions. The resulting
membranes are described in Table 7.
For each fabrication method, the influence of the dope
solution composition was studied by changing the P
(3HB3HV) concentration and/or by adding an additive. The
membranes properties were highly dependent on the fabrication method. In case of NIPS, the membranes exhibited
asymmetric structures. On the other hand, the membranes
made by EIPS exhibited symmetric structures.
Figure 5 gives the scanning electron microscopy
(SEM) images of the different membranes made by NIPS.
The cross sections and surfaces were analyzed.
Asymmetric structures with a finger-like structure at the
top and a sponge-like sublayer can be observed for each
membrane. Their microstructures properties are detailed in
Table 8.
As observed on Figure 5 and confirmed by the Table 8,
the increase in P(3HB3HV) concentration tends to decrease
the surface porosity but tends to increase the membrane
188
P. Tomietto et al.
0.392 kg m
−2 h
−1 and a separation factor of 3.98 in favor to
the methanol permeation were measured. By increasing the
methanol concentration in the methanol/MTBE feed composition, the separation factor decreases. Besides, it was
assumed that MTBE could act as a plasticizer in the membrane and would consequently improve the permeability of
both compounds. Later, in another article, the authors went
further in the evaluation of the pervaporation performances
(Villegas et al. 2016). This time, the authors chemically
modified the P(3HB) dense membranes by plasma polymerization with acrylic acid. After membrane modification,
due to the additional layer formed and additional flux
resistance, the flux was decreased. Nevertheless, the separation factor was significantly improved up to 18.6. The
latter is explained by a higher affinity between the plasma
polymerized acrylic acid layer and methanol.
Related to pervaporation applications, Galego et al.
studied the swelling capacity, vapor permeability and
selectivity of PHAs films with different HV content (Galego
et al. 2002). The swelling percentage to water–ethanol
mixtures and the water vapor permeability of P(3HB) was
higher than P(3HB3HV) (22 mol% HV). It is discussed as a
consequence of the rougher surface of P(3HB) films and the
better ability of P(3HB) to make hydrogen bonding interactions with water molecules.
3.2.2 Air Filtration
Nicosia et al. have made PLA/P(3HB) filters intended for air
filtration applications (Nicosia et al. 2015). The filters were
made by electrospinning of a PLA/P(3HB) mixture in
chloroform. The performances were quantified by measuring
the penetration of sodium chloride aerosol particles ranging
from 20 to 600 nm. The performances could be improved by
superposing several thin layers to reduce the overall packing
density of the structure. The results highlight a good
potential for air filtration applications. The best collection
efficiency was measured at 98.5% for 0.3 µm particles.
Additionally, these membranes could be modified by loading of didecyldimethylammonium nitrate in order to improve
their antimicrobial activity.
3.2.3 Liquid Filtration
Mas et al. studied PHAs-based membranes intended for
microfiltration applications (Mas et al. 1996). The membranes were made with P(3HB3HV) having different HV
ratio. For the membrane fabrication, they used the EIPS
coupled with the NIPS. For the dope solution, various
compositions of chloroform/THF mixtures were used. THF
acts as a non-solvent in the dope solution. The non-solvent
bath was composed of a mixture of ethanol/water (4/1 v/v).
The water permeability and ethanol permeability were
measured. The pores size was analyzed by image analysis.
Whether for P(3HB) or for P(3HB3HV) (22 mol% HV), the
best permeabilities were obtained for a dope solution containing 8% of THF. The permeability went up to 600 L
m
−2 h
−1 bar
−1 but no rejection test was performed. For THF
concentrations above 8%, the proper solubility of the polymer was hindered, leading to the densification of the membrane. Among the membranes characterized, the pores size
varied from 0.25 to 2 µm.
Keawsupsak et al. have made PLA/P(3HB3HV) blend
membranes to test them on water permeability and rejection
of bovine serum albumin (Keawsupsak et al. 2014). Membranes were produced by NIPS using NMP as a solvent and
water as a non-solvent. At a small amount of P(3HB3HV)
added to PLA, the membrane exhibits a better permeability
value of 65.2 L m
−2 h
−1 bar
−1 , but it remains quite low. In
that case, the BSA rejection was measured at 78.7%. The
membrane showed a finger-like structure.
Composite membranes made of P(3HB), calcium alginate
and carboxyl multi-walled carbon were made by electrospinning for NF applications (Guo et al. 2016). Their performances were measured by water permeability, separation
of emulsified oil–water solutions and rejection of brilliant
blue. One membrane demonstrated an excellent antifouling
property associated with a pure water flux around 35 L
m
−2 h
−1 at 2 bar, a 98.2% rejection of brilliant blue and a
98% rejection of oil.
Finally, P(3HB3HV) (3 mol% HV)-based membranes
with tailored structures were made in our research group at
the ISCR (Institut des Sciences Chimiques de Rennes,
France), in the Chemistry and Process Engineering team.
The membranes were fabricated via two different techniques,
NIPS or EIPS. By NIPS, the NMP was used as solvent and
water was used for the non-solvent bath. By EIPS, chloroform was employed as volatile solvent. The dope solutions
were casted under controlled conditions. The resulting
membranes are described in Table 7.
For each fabrication method, the influence of the dope
solution composition was studied by changing the P
(3HB3HV) concentration and/or by adding an additive. The
membranes properties were highly dependent on the fabrication method. In case of NIPS, the membranes exhibited
asymmetric structures. On the other hand, the membranes
made by EIPS exhibited symmetric structures.
Figure 5 gives the scanning electron microscopy
(SEM) images of the different membranes made by NIPS.
The cross sections and surfaces were analyzed.
Asymmetric structures with a finger-like structure at the
top and a sponge-like sublayer can be observed for each
membrane. Their microstructures properties are detailed in
Table 8.
As observed on Figure 5 and confirmed by the Table 8,
the increase in P(3HB3HV) concentration tends to decrease
the surface porosity but tends to increase the membrane
188
P. Tomietto et al.