by charge repulsion [79]. However, the rejection rate of some uncharged and small
molecular weight organic contaminants by NF/RO membrane can be low [80].
The low rejection of some small molecular weight and uncharged NSAIDs by
NF/RO membranes, as mentioned above, has been widely reported in the literature
[8, 80–84]. For example, at extended stages of filtration, there was poor rejection of
chloroform and bromoform by RO (e.g., TFC-HR and XLE) and NF membranes
(e.g., NF-90 and TFC-SR2) [84]. Chloroform and bromoform are both neutral and
have a molecular weight of 119.4 and 252.7 g/mol, respectively. The charge of the
trace organic contaminants and that on the membrane can play a significant role in
the rejection of TrOCs. For example, rejection of a charged compound by NF/RO
membranes is usually higher than for a neutral compound with the same molecular
weight or size [84]. Since most pharmaceuticals are negatively charged particularly
at neutral pH, a considerable number of these compounds may be completely
rejected by charge repulsion between the compound and membrane charges
[79]. Xu et al. [84] reported that highly negative surface charge membranes such
as the loose NF-200 membrane, with a molecular weight cutoff (MWCO) of 300 g/
mol, could reject more than 89% of low molecular weight negatively charged
compounds such as ibuprofen. A high rejection of other pharmaceuticals such as
dichloroacetic acid (91%) and trichloroacetic acid (94%) was also achieved using the
ESNA (NF) and RO-XLE (RO) membranes [8].
Membranes with a high degree of desalting showed the highest rejection of most
NSAIDs [81]. A UTC60 aromatic polyamide membrane (an NF membrane) which
has a low NaCl rejection (55%) demonstrated a poor rejection of several trace
organics such as 47% for bisphenol A and 5% for chloroform [71]. Moreover, higher
membrane roughness has been highly correlated with a lower rejection of large
organic contaminants [81].
The hydrophobicity and charge of an active layer of the membrane can also affect
the rejection of various NSAIDs [84]. The surface hydrophobicity of a membrane
can be determined by measuring the contact angle. The rejection of some organics
could be improved by increasing the hydrophobicity of the membrane because it
reduces the affinity between the neutral organic solute and the surface of the
membrane [85]. Furthermore, the amount of charge in the surface of the membrane
affects the degree of electrostatic repulsion and rejection of negatively charged
solutes that are subjected to dynamic property changes during the membrane process
[84]. For example, Bellona and Drewes [86] studied the rejection of negatively
charged organic acids (2-naphthalenesulfonic acid and 1,4-dinaphthalenesulfonic
acid) by negatively charged NF membranes (e.g., NF-90 and NF-200). According to
their findings, the rejection was larger than expected based on steric exclusion and
was mainly driven by the surface charge of the membrane and correlated with the
degree of ionization of these compounds [86].
Operational parameters such as feed solution pH, salinity, temperature, pressure,
and cross-flow velocity can influence the rejection of NSAIDs by NF/RO membranes. The feed solution pH can govern the speciation of ionizable NSAIDs (and to
a lesser extent, the membrane surface charge) and thus their rejection. For instance,
Bellona et al. [81] claimed that when using NF/RO at pH values between 3 and
228
L. N. Nguyen et al.
molecular weight organic contaminants by NF/RO membrane can be low [80].
The low rejection of some small molecular weight and uncharged NSAIDs by
NF/RO membranes, as mentioned above, has been widely reported in the literature
[8, 80–84]. For example, at extended stages of filtration, there was poor rejection of
chloroform and bromoform by RO (e.g., TFC-HR and XLE) and NF membranes
(e.g., NF-90 and TFC-SR2) [84]. Chloroform and bromoform are both neutral and
have a molecular weight of 119.4 and 252.7 g/mol, respectively. The charge of the
trace organic contaminants and that on the membrane can play a significant role in
the rejection of TrOCs. For example, rejection of a charged compound by NF/RO
membranes is usually higher than for a neutral compound with the same molecular
weight or size [84]. Since most pharmaceuticals are negatively charged particularly
at neutral pH, a considerable number of these compounds may be completely
rejected by charge repulsion between the compound and membrane charges
[79]. Xu et al. [84] reported that highly negative surface charge membranes such
as the loose NF-200 membrane, with a molecular weight cutoff (MWCO) of 300 g/
mol, could reject more than 89% of low molecular weight negatively charged
compounds such as ibuprofen. A high rejection of other pharmaceuticals such as
dichloroacetic acid (91%) and trichloroacetic acid (94%) was also achieved using the
ESNA (NF) and RO-XLE (RO) membranes [8].
Membranes with a high degree of desalting showed the highest rejection of most
NSAIDs [81]. A UTC60 aromatic polyamide membrane (an NF membrane) which
has a low NaCl rejection (55%) demonstrated a poor rejection of several trace
organics such as 47% for bisphenol A and 5% for chloroform [71]. Moreover, higher
membrane roughness has been highly correlated with a lower rejection of large
organic contaminants [81].
The hydrophobicity and charge of an active layer of the membrane can also affect
the rejection of various NSAIDs [84]. The surface hydrophobicity of a membrane
can be determined by measuring the contact angle. The rejection of some organics
could be improved by increasing the hydrophobicity of the membrane because it
reduces the affinity between the neutral organic solute and the surface of the
membrane [85]. Furthermore, the amount of charge in the surface of the membrane
affects the degree of electrostatic repulsion and rejection of negatively charged
solutes that are subjected to dynamic property changes during the membrane process
[84]. For example, Bellona and Drewes [86] studied the rejection of negatively
charged organic acids (2-naphthalenesulfonic acid and 1,4-dinaphthalenesulfonic
acid) by negatively charged NF membranes (e.g., NF-90 and NF-200). According to
their findings, the rejection was larger than expected based on steric exclusion and
was mainly driven by the surface charge of the membrane and correlated with the
degree of ionization of these compounds [86].
Operational parameters such as feed solution pH, salinity, temperature, pressure,
and cross-flow velocity can influence the rejection of NSAIDs by NF/RO membranes. The feed solution pH can govern the speciation of ionizable NSAIDs (and to
a lesser extent, the membrane surface charge) and thus their rejection. For instance,
Bellona et al. [81] claimed that when using NF/RO at pH values between 3 and
228
L. N. Nguyen et al.