Recalcitrant pharmaceuticals such as carbamazepine are known to be removed by
sorption from the water phase in CWs [50]. Higher retention in gravel beds was
observed for carbamazepine (97 ng g
À1 ; 77%) in relation to clofibric acid (14 ng g
À1 ;
11%) and ibuprofen (15 ng g
À1 ; 12%), which was explained by carbamazepine
higher hydrophobicity and electrostatic interactions between the acidic compounds
charged negatively and negatively charged biofilm covering the gravel bed. In
contrast, sorption of polar compounds such as caffeine was considered of minor
importance in CW systems, because of their high water solubility and low hydrophobicity [83]. For ionizable pharmaceuticals, pH-dependent octanol/water partition
coefficient (Log D ow ) has been considered more useful than Log K ow [27]. However,
it has also been observed that pharmaceuticals that are moderately hydrophilic (with
Log D ow ranging from À2.3 to 3) tend not to bind significantly to organic
matter [63].
The majority of the HSSF and VF systems in the different studies revised by
Verlicchi and Zambello [27] have been filled with gravel (generally with a particle
size of 8–15 mm and porosity 30–40%) and in some cases LECA (light expanded
clay aggregate). Gravel and sand are naturally more inert; thus, sorption will be very
much dependent on the organic matter that builds up in time within the systems.
However, optimization may be attempted by selecting materials with high sorption
capacity. Dordio and Carvalho [84] provide an in-depth analysis of the role of the
CW substrate on the removal of xenobiotics, including pharmaceuticals. Less
common media such as LECA or activated carbons and more recent innovations
like kaolinite, diatomite, cork, perlite and zeolites have been tested. However, the
majority of those experiments were carried in the lab with synthetic wastewater (tap
water spiked with a few compounds, not always at realistic conditions), thus raising
concerns with the sustainability and applicability of these new approaches in the
long term [27]. It should be stressed that LECA has been consistently shown in
different studies to provide extensive sorption of carbamazepine and atenolol [62].
One should bear in mind that sorption can be seen not only as an end mechanism
in the form of sedimentation but also a factor affecting the bioavailability of
compounds for plant uptake and for biodegradation processes. Sorption provides
different reaction times for the plants, biofilm and respective communities to interact
with the pollutants. Hijosa-Valsero et al. [85] observed very limited adsorption of the
musk fragrances galaxolide and tonalide (present in personal care products) despite
their high hydrophobicity and strong affinity to media adsorption (Log K ow 6.26 and
6.35, respectively). In fact, it was due to the easy uptake by plant rhizosphere, rather
than being adsorbed by media. Moreover, the substrate itself not only conditions
sorption but also affects the biofilm microbial community. Vesuvianite (or idocrase,
a silicate mineral) had better removal efficiency than gravel and zeolite for treating,
sulfonamides in spite of the limited sorption of these compounds. Yet, vesuvianite
conditioned the growth of the microbial consortium and lengthened HRT via its
larger porosity, thus contributing for an overall better removal efficiency [86]. A
similar behaviour was observed for a study with six commonly used materials in
CWs (sand, zeolite, blast iron slag, petroleum coke, polonite (natural calcium
silicate) and crushed autoclaved aerated concrete). Laboratorial batch tests with
400
P. N. Carvalho
sorption from the water phase in CWs [50]. Higher retention in gravel beds was
observed for carbamazepine (97 ng g
À1 ; 77%) in relation to clofibric acid (14 ng g
À1 ;
11%) and ibuprofen (15 ng g
À1 ; 12%), which was explained by carbamazepine
higher hydrophobicity and electrostatic interactions between the acidic compounds
charged negatively and negatively charged biofilm covering the gravel bed. In
contrast, sorption of polar compounds such as caffeine was considered of minor
importance in CW systems, because of their high water solubility and low hydrophobicity [83]. For ionizable pharmaceuticals, pH-dependent octanol/water partition
coefficient (Log D ow ) has been considered more useful than Log K ow [27]. However,
it has also been observed that pharmaceuticals that are moderately hydrophilic (with
Log D ow ranging from À2.3 to 3) tend not to bind significantly to organic
matter [63].
The majority of the HSSF and VF systems in the different studies revised by
Verlicchi and Zambello [27] have been filled with gravel (generally with a particle
size of 8–15 mm and porosity 30–40%) and in some cases LECA (light expanded
clay aggregate). Gravel and sand are naturally more inert; thus, sorption will be very
much dependent on the organic matter that builds up in time within the systems.
However, optimization may be attempted by selecting materials with high sorption
capacity. Dordio and Carvalho [84] provide an in-depth analysis of the role of the
CW substrate on the removal of xenobiotics, including pharmaceuticals. Less
common media such as LECA or activated carbons and more recent innovations
like kaolinite, diatomite, cork, perlite and zeolites have been tested. However, the
majority of those experiments were carried in the lab with synthetic wastewater (tap
water spiked with a few compounds, not always at realistic conditions), thus raising
concerns with the sustainability and applicability of these new approaches in the
long term [27]. It should be stressed that LECA has been consistently shown in
different studies to provide extensive sorption of carbamazepine and atenolol [62].
One should bear in mind that sorption can be seen not only as an end mechanism
in the form of sedimentation but also a factor affecting the bioavailability of
compounds for plant uptake and for biodegradation processes. Sorption provides
different reaction times for the plants, biofilm and respective communities to interact
with the pollutants. Hijosa-Valsero et al. [85] observed very limited adsorption of the
musk fragrances galaxolide and tonalide (present in personal care products) despite
their high hydrophobicity and strong affinity to media adsorption (Log K ow 6.26 and
6.35, respectively). In fact, it was due to the easy uptake by plant rhizosphere, rather
than being adsorbed by media. Moreover, the substrate itself not only conditions
sorption but also affects the biofilm microbial community. Vesuvianite (or idocrase,
a silicate mineral) had better removal efficiency than gravel and zeolite for treating,
sulfonamides in spite of the limited sorption of these compounds. Yet, vesuvianite
conditioned the growth of the microbial consortium and lengthened HRT via its
larger porosity, thus contributing for an overall better removal efficiency [86]. A
similar behaviour was observed for a study with six commonly used materials in
CWs (sand, zeolite, blast iron slag, petroleum coke, polonite (natural calcium
silicate) and crushed autoclaved aerated concrete). Laboratorial batch tests with
400
P. N. Carvalho