Journal of Hazardous Materials 243 (2012) 340–349

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Adsorption performance of titanium dioxide (TiO2 ) coated air filters for volatile
organic compounds
Lexuan Zhong, Chang-Seo Lee, Fariborz Haghighat ∗
Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Quebec H3G 1M8, Canada

h i g h l i g h t s

 Systematic approach for evaluation of adsorption performance of UV-PCO is developed.

 Adsorption behaviors of TiO2 coated air filters for various VOCs were examined.
 Adsorptive performance is the sum of interactions between substrates and VOCs.
 The presence of water vapor plays an important role on the adsorption behavior.
 Photocatalytic activity is closely related to the adsorption process.

a r t i c l e i n f o a b s t r a c t

Article history: The photocatalytic oxidation (PCO) technology as an alternative method for air purification has been
Received 9 August 2012 studied for decades and a variety of PCO models indicate that the adsorption of reactants on the catalyst
Received in revised form 19 October 2012 surface is one of the major physical and chemical processes occurring at a heterogeneous photocatalytic
Accepted 21 October 2012
reaction. However, limited study explored the adsorption effect of a photocatalyst. This study carried
Available online 29 October 2012
out a systematic evaluation of adsorption performance of titanium dioxide (TiO2 ) coated fiberglass fibers
(FGFs), TiO2 coated carbon cloth fibers (CCFs), and original CCFs air filters at various relative humidity
Keywords:
conditions for nine volatile organic compounds. TiO2 /FGFs, TiO2 /CCFs, and CCFs were characterized by
Filters
Titanium dioxide (TiO2 )
SEM for morphology and N2 adsorption isotherm for BET surface area and pore structure. A bench-scale
Adsorption adsorption test setup was constructed and adsorption tests were performed at various relative humidity
Relative humidity (RH) conditions and four different injected concentrations for each compound. The isothermal adsorption
Photocatalytic reaction curves at low concentration levels were obtained and they were well described by Langmuir isotherm
model. It was noticed that there were significant differences between the adsorption behaviors and
photocatalytic activities of TiO2 /FGFs and TiO2 /CCFs. It was concluded that adsorption performance is
closely related to the characteristics of substrates and therefore, the development of a substrate with
high adsorption ability is a promising trend for improving the performance of the UV-PCO technology.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction Purification mechanism of PCO technology is that electron–hole

The indoor air quality (IAQ) has received enormous attention in
the area of industrial, educational and medical care sectors world-
wide. It is anticipated that this concern will continue in the near
future, and it will be driven by a multitude of strategies to miti-
gate IAQ problems. Due to the increased demand of energy savings,
air filtration and purification can be an alternative approach to
improve IAQ, rather than solely relying on the dilution ventilation.
Photocatalytic oxidation (PCO), an innovative and developing air
purification technology, gradually becomes a promising approach
for such application [1,2].
∗ Corresponding author. Tel.: +1 514 848 2424x3192; fax: +1 514 848 7965.
E-mail address: haghi@bcee.concordia.ca (F. Haghighat).
pairs generated from the surface of a semiconductor under UV light
irradiation break down adsorbed organic pollutants, theoretically,
into water and carbon dioxide. In general, it is assumed that the
organic pollutant adsorbed on the substrate undergoes oxidation
reactions through a surface-bound hydroxyl radical formed from
electrons transfer, rather than direct holes transfer, although in
some cases there is some evidence that it may occur [3]. Although it
is disputable that the PCO reaction may take place at fluid stream,
most of researchers propose that only those pollutant molecules
in direct contact with catalyst surface undergo PCO reaction [4–6].
In other words, adsorption process is an important prerequisite for
PCO reaction; it also affects the efficiency of PCO air cleaners.
As an alternative method for purification of air stream, the
PCO technology has been studied for decades, and numerous
papers focusing on various aspects of PCO technology have been

0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jhazmat.2012.10.042
 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349
reported. They include the improvement of photocatalysts, the
evaluation of PCO performance of different bench-scale reactors,
potential intermediates, kinetic study and PCO modeling develop-
ment [7–11]. Although a variety of PCO models [11,12] indicate
adsorption of reactants on the catalyst surface is one of physical and
chemical processes occurring at a heterogeneous photocatalytic
reaction, literatures especially aiming to explore the adsorption
effect of a photocatalyst are limited. In addition, different forms
of Langmuir–Hinshelwood (L–H) models have been developed; a
Langmuir adsorption constant is a critical parameter in L–H mod-
els. Usually, the values of adsorption coefficient were determined
by fitting the kinetic PCO model with experimental data [9,13,14].
In fact the values derived from a kinetic study are usually much
higher than those obtained from the Langmuir adsorption isotherm
in the absence of light [3]. In few articles relevant to the adsorp-
tion isotherms [15–18], the adsorption performance was examined
with a limited number of compounds of interest, such as toluene,
acetone, and trichloroethylene. Moreover, the effect of relative
humidity on adsorption isotherms has not been fully examined.
To the best of our knowledge, no systematic studies have been
conducted with respect to an adsorption phenomenon of a photo-
catalyst for a wide range of VOCs. Also research on the adsorption
properties of a photocatalyst placed on various substrates is limited.
Hence, this paper reports the outcomes of a systematic approach for
evaluation of the adsorption system performance at various relative
humidity conditions for various VOCs. It also provides explanations
to the differences of adsorption features observed between three
air filters based on a fundamental analysis of the mass transports,
and how these differences can influence the photocatalytic activity.
Experimental determination of adsorption coefficient is a basic
research for the purpose of widespread commercial utilization
of UV-PCO technology since the adsorption coefficient is a crit-
ical parameter that influences the surface coverage of adsorbed
compounds, thereby affecting the photocatalytic oxidation rate. In
this study, the adsorption performance of titanium dioxide (TiO2 )
loaded on two different substrates, fiberglass fibers (FGFs) and
carbon cloth fibers (CCFs) has been experimentally investigated.
For the FGFs, this paper demonstrates a systematic evaluation of
adsorption performance at various relative humidity conditions
(9.6 ± 0.6%–70.2 ± 2.7%) and at room temperature of 22.8 ± 0.5 ◦ C
for nine compounds: toluene, p-xylene, ethanol, 1-butanol, methyl
ethyl ketone (MEK), acetone, hexane, octane, and limonene. The
challenge air concentrations of each selected compound are from
0.5 ppm to 5 ppm. For the CCFs with and without TiO2 coating,
ethanol and hexane have been employed to examine the adsorption
performance, as a typical polar and a non-polar VOC, respec-
tively, and at various relative humidity conditions. This paper
provides a profound insight into the basic knowledge of TiO2
adsorption mechanism, helps to determine the values of adsorption
coefficients for a PCO modeling, and further reveals that the adsorp-
tion behavior is closely related to the characteristics of substrates.
2. Methodology
2.1. Materials
Two commercially available PCO filters, TiO2 coated on fiber-
glass fibers and TiO2 coated on carbon cloth fibers, were examined
in this study. The exact properties of filters and catalyst prepa-
ration cannot be given due to the proprietary of these TiO2
filters. The morphology of these materials was evaluated by
scanning electron microscope (SEM, Hitachi S-4700 Model).
Brunauer–Emmett–Tele (BET) surface area and pore parameters
were determined by nitrogen adsorption–desorption isotherm
measurements (Micromeritics ASAP 2000).
341
Nine reagent grade chemicals were selected as representative
of indoor air contaminants [19], and they cover major chemical
categories and have a wide range of different physical properties
like molecule weight and polarity, which are shown in Table 1. They
included toluene (99.9%), p-xylene (99.9%), 1-butanol (99.9%), n-
hexane (96%), octane (95%), MEK (99.9%), acetone (99.5%) and d-
limonene (97%) from Fisher Scientific Inc. (Canada), and ethanol
(99%) from SAQ (Société des alcools du Québec – Québec Alcohol
Board).
2.2. Adsorption experimental setup
To quantitatively study equilibrium adsorption and to examine
the impact of relative humidity, the bench-scale adsorption test
setup complying with ASHRAE Standard 145.1 [21] was used (see
Fig. 1). The laboratory compressed air was used as the carrier gas,
its flow rate was controlled by a mass flow meter, and it was puri-
fied by passing it through a granular activated carbon (GAC) filter to
remove potential contaminants. The relative humidity of the mixed
stream was varied from (9.6% ± 0.6%) to (70.2% ± 2.7%) by adjusting
the flow rate of the compressed air into the distilled water bot-
tle. In addition, the temperature of the distilled water bottle was
maintained constant through a water bath so that it could provide
water vapor with a steady concentration. The selected chemicals
were in liquid state at room temperature; they were automatically
injected through a syringe pump (KD Scientific). The injected con-
centrations of selected chemicals were calculated on the basis of
airflow rate, chemical injection rate and the chemical properties.
Before connecting to a media column, a PTFE tube (Fig. 1(1)) was
first connected to the test system and was used to transport gaseous
pollutant to the calibrated multi-gas analyzer (Innova AirTech
Instruments 1312). After the inlet concentration reached to a steady
state, the PTFE tube was replaced by a media column filled with
TiO2 /FGFs, TiO2 /CCFs, or CCFs (Fig. 1(2)). The concentration of chal-
lenge VOCs after the media column was measured continuously
till it was equal to the previous stabilized concentration. For each
compound, the adsorption test was performed at T = 22.8 ± 0.5 ◦ C
and four different injected concentrations (0.496 ± 0.013 ppm,
0.990 ± 0.010 ppm, 1.971 ± 0.023 ppm, and 4.955 ± 0.063 ppm).
Before performing each adsorption experiment, the air filter was
conditioned overnight under the corresponding humid conditions
to be tested; after each adsorption test, the air filter was conditioned
again for regeneration by passing through humidified compressed
air.
2.3. Adsorption isotherm
Many different expressions that describe dynamic equilibrium
of sorbed-phase and gaseous phase have been proposed, among
which the Langmuir isotherm is the most widely applied in the
field of surface kinetics. The Langmuir isotherm model describes
adsorbate–adsorbent systems in which the extent of adsorbate cov-
erage is limited to monolayer coverage of the surface, especially at
low loadings. The sorbed-phase concentration of the VOC at the
catalyst fibers surface, CS , can also be estimated by this model.
CS0 KC
CS = f (C) = (1)
1 + KC
where CS0 is the maximum sorbed-phase concentration corre-
sponding to monolayer complete coverage, C is the gaseous phase
concentration, K is an equilibrium constant, which is the adsorption
constant divided by the desorption constant. When the concentra-
tion of challenge gases is very low, that is KC  1, Eq. (1) can be
simplified to;
CS = f (C) = CS0 KC = KL C (2)
342 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349

Table 1
Physical properties of the selected VOCs [20].

Chemical VOCs class Molecular formula MWa (g/mol) BPb (◦ C) VPc (mmHg) DCd

Ethanol Alcohols C2 H6 O 46.1 78.4 44 24.3

1-Butanol C4 H10 O 74.1 117.2 6 16.68
Acetone Ketones C3 H6 O 58.1 56.1 180 20.7
MEK C4 H8 O 72.1 79.4 78 18.51
Toluene Aromatic C7 H8 92.1 111.1 21 2.38
p-Xylene C8 H10 106.2 138.3 9 2.2
n-Hexane Alkane C6 H14 86.2 68.9 124 1.88
Octane C8 H18 114.2 125.6 10 1.94
d-Limonene Terpenes C10 H16 136.2 176 2 2.3

Note: Water (very polar) has a dielectric constant of 80.10 at 20 ◦ C.

a
MW stands for molecular weight.
b
BP stands for boiling point.
c
VP stands for vapor pressure.
d
DC stands for dielectric constant at 20–25 ◦ C.

where KL is the synthetic Langmuir parameter that embeds both
the saturation capacity, CS0 , and the equilibrium constant, K. Eq. (2)
can be changed into Eq. (3) by converting the sorbed-phase concen-
tration, CS , to the mass of VOCs adsorbed in the surface of catalyst,
ms , and introducing the mass of air filters, mf , to compensate its
impact to the adsorption behavior.
ms
= KL C
mf
where ms is the mass of adsorbed VOCs, mf is the mass of air filters,
and KL is the adsorption coefficient which is defined as the ratio
between the mass of adsorbed VOCs in solid phase and the con-
centration of VOCs in air phase at adsorption equilibrium per unit
weight of air filters.
2.4. Adsorption analysis method
Based on the experimental data obtained from both upstream
and downstream air concentrations measurements, adsorption
capacity of the fiberglass PCO filter for a certain VOC gas can be
evaluated. This capacity is expressed as the ratio of the adsorbed
mass of contaminant gas over the removal media weight [22]:
 Tads (t) (t)
0
Q (Cup − Cdown )dt
RC =
MPCO
where RC is the filter capacity when it reaches equilibrium; Tads
is the elapsed time of adsorption test (min); Q is the airflow rate
(t) (t)
(m3 /min); Cup and Cdown are the upstream and downstream con-
centration (g/m3 ) as a function of elapsed time, respectively; MPCO
is the mass of removal media (g): for TiO2 /FGFs, it was 2.715 g; for
TiO2 /CCFs with TiO2 , it was 0.890 g; and for original CCFs, it was
(t)
0.386 g. It should be noted that in this study Cup is computed as the
average upstream concentration (before adsorption occurs).
2.5. Photocatalysis experimental setup
A single pass mode test rig was built to evaluate the PCO effec-
tiveness for treating ethanol and hexane with inlet concentrations
(3) of 0.25 ppm, 0.5 ppm, and 1 ppm. The test rig was composed of four
parallel test ducts with 0.3 m by 0.3 m of cross-sectional area each.
UV-PCO section was designed to be versatile so that different UV-
PCO systems with various geometries can be installed. Each duct
has a fan with a variable speed control so that the airflow rate can
be controlled regardless of the flow resistances of different UV-
PCO systems. Since the length of the duct before the fan is long
enough, it is reasonable to assume the airflow as ideal plug flow.
This was verified through measurement at various locations both
at the upstream and downstream. The geometry dimensions of test
rig on the elevation view are presented in Fig. 2, and detailed setup
can be found in [23]. In this study, the UV-PCO reactor contained
three 0.3 m by 0.3 m TiO2 /FGFs air filters or TiO2 /CCFs air filters
arranged in two banks, and in each bank two UV lamps (2 × 18.4 W
UVC lamps or 2 × 18.4 W VUV lamps) were employed (total four)
for providing irradiation. The nominal 254 nm UV output power at
100 h and 26.7 ◦ C was 5.8 W, and 254 nm UV output at 1 m from a
lamp was 59 mW/cm2 . The vertical distance between the surfaces
(4)
of UV lamps and the PCO filter was approximate 5 cm.
The target concentration of the selected challenge gases was
at sub-ppm level (0.25–1 ppm) which represented indoor air pol-
lution. All PCO experiments were carried out at room conditions.
Table 2 summarizes the detailed experimental conditions on which
the experiments were performed. The experimental data col-
lected from upstream and downstream measurement ports were
employed to calculate the effectiveness of a UV-PCO air cleaner.

Fig. 1. Adsorption test setup.

 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349 343

355.60 cm
(140.00’’)
76.20 cm 60.96 cm 76.20 cm 60.96 cm 35.56 cm 45.72 cm
(30.00’’) (24.00’’) (30.00’’) (24.00’’) (14.00’’) (18.00’’)

30.99 cm 38.61 cm
PCO reactor (12.20’’) (15.20’’)

128.90 cm 121.28 cm 59.28 cm

(50.75’’) (47.75’’) (23.34’’)

30.99 cm 38.61 cm
PCO reactor (12.20’’) (15.20’’)

60.96 cm 76.20 cm 60.96 cm 35.56 cm 45.72 cm

(24.00’’) (30.00’’) (24.00’’) (14.00’’) (18.00’’)

Fig. 2. Dimensions of the test rig on elevation view.

Table 2 branch of nitrogen isotherm by the BJH (Barrett–Joyner–Halenda)

Photocatalysis experimental conditions.
method [24].
Parameter Experimental conditions The adsorption isotherms (plotted in Fig. 3) indicate the order
Single VOC Ethanol and hexane of adsorption capacity is CCFs > TiO2 /CCFs > TiO2 /FGFs, which is
Inlet concentration (ppm) 0.25–1 also further verified by the measured BET surface areas (Table 3).
Volumetric flow rate (m3 /h) 170 (100 cfm) The BET surface area of TiO2 /FGFs was much smaller than that of
Face velocity (m/s) 0.5 TiO2 /CCFs. It is worth mentioning that BET surface area decreased
Relative humidity (%) 15–30
from 1490.8605 m2 /g to 887.6638 m2 /g when TiO2 was loaded on
Light intensity (W/m2 ) 29–36
Temperature (◦ C) 23 the surface of CCFs. Additionally, the pore size distribution curve
manifests that the pore volume of TiO2 /CCFs decreased to some
extent in the range of mesopores and formation of macropores
Single-pass removal efficiency, which was determined by the resulting from TiO2 particle agglomeration were observed com-
amount of the removed pollutant from the air stream after it went pared with the original CCFs (see Fig. 5). This observation indicates
through the air cleaner, is widely used to evaluate the performance the process of TiO2 coating did not destroy the pore structure, and
of an air cleaner. the presence of the agglomeration of catalyst was at the external
Cup,t − Cdown,t surface of fibers. This is consistent with the observation reported
t % = × 100 (5) by Guo et al. [25] when they examined similar materials.
Cup,t
Fig. 5 also shows that pore diameter distributions were nar-
3. Results row for CCFs and TiO2 /FGFs, and pores of three materials (CCFS ,
TiO2 /FGFs, TiO2 /CCFS ) were mainly mesopores. It can also be found
Fig. 3 presents the equilibrium adsorption isotherms for the from Table 3 that the pore volume and pore size of TiO2 /CCFs are
selected challenge VOCs under different RH conditions. The dimen- larger than those of TiO2 /FGFs.
sionless mass ratio (g/g) was used to facilitate the comparison. This
figure shows, at four RH levels, all isotherms are linear with respect 4.2. Adsorption coefficients verification
to the equilibrium concentration. This indicates that the adsorption
behaviors of TiO2 /FGFs, TiO2 /CCFs, and CCFs follow ideal mono- Demeestere et al. [15] presented the linear adsorption curves for
layer adsorption at low ppm concentrations. It should be noted trichloroethylene, toluene, and chlorobezene on the photocatalyst
that the slope of each regression line is the adsorption coefficient TiO2 Degussa P25 at 25.0 ◦ C. They found that the adsorption coef-
of individual VOC for various RH. The results clearly indicate that ficient of toluene was 0.00253 m3 /g and 0.000103 m3 /g for RH = 0%
the adsorption performance of the TiO2 catalyst varies for different and 57.8%, respectively. Coronado et al. [26] reported the adsorp-
compounds and substrates. And adsorption coefficient is signifi- tion constant of acetone for the weak sites of TiO2 thin film was
cantly affected by the presence of water vapor. 0.2 m3 /g and 0.046 m3 /g for RH = 0% and 25%, respectively. Tomida
et al. [16] investigated the adsorption isotherms of acetone on the
4. Discussions photocatalyst of TiO2 coated on silica beads and obtained similar
results. All the reported adsorption coefficient values are at the
4.1. Characterization same order of magnitude as those obtained in this study.

Fig. 4 shows the scanning electron microscopy (SEM) images of
TiO2 /FGFs and TiO2 /CCFs which shows fibers are randomly oriented
for both media and TiO2 /CCFs displays fracture surfaces. Nano-TiO2
particles were coated on the substrates, and the magnified SEM
images in Fig. 4b and d demonstrate nano-TiO2 powders were more
uniformly distributed on the FGFs’ surface than those loaded on
the CCFs due to the CCFs’ fiber roughness. Fig. 5 presents nitro-
gen adsorption isotherms for CCFs, TiO2 /CCFs and TiO2 /FGFs and
their pore size distributions which were obtained from desorption
4.3. Correlation between adsorption coefficient and properties of
VOC
Results of TiO2 /FGFs indicate that for the polar VOCs, such as
ethanol, isobutanol, acetone, and MEK, the adsorption coefficients
are roughly one order of magnitude higher than those of non-
polar VOCs for a given RH level. Also, the results further indicate
the TiO2 /FGFs filter, a polar substance, has a higher adsorp-
tion preference for polar VOCs. The order of adsorption capacity
344 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349

Fig. 3. Adsorption isotherms of the selected challenge gases at various RH conditions (9.6 ± 0.6%–70.2 ± 2.7%) and at 22.8 ± 0.5 ◦ C for (a) TiO2 /FGFs (b) CCFs (c) TiO2 /CCFs.
 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349 345

Fig. 3. (Continued).

for the selected chemical classes follows the sequence of alco-
hols > ketones and terpenes > aromatics > alkanes. This feature is
attributed to the strength of the corresponding intermolecular
forces between adsorbed VOCs and the catalyst surface. Disper-
sion forces are the main intermolecular forces holding non-polar
alkanes in the solid phase, which are weaker than van der Waals
interactions for aromatic hydrocarbons. Due to the high dipole
moment of the carbonyl group, dipole–dipole interactions for
ketones are stronger than van der Waals attractions for hydro-
carbons. In addition to van der Waals interactions, hydrogen

Fig. 4. (a) SEM images of TiO2 /FGFs and (b) magnified image of TiO2 /FGFs (c) SEM images of TiO2 /CCFs and (d) magnified image of TiO2 /CCFs.
346
Isotherm Plot
700
CCFs
Volume adsorbed (cm3/g)
600
TiO2/CCFs
500
TiO2/FGFs
400
300
200
100
0 0
0 0.2 0.4
Relative pressure (P/P0)
Fig. 5. (a) N2 adsorption isotherms and (b) pore size distributions.
bonding plays a greater role for the attraction between alcohols
and hydrated catalyst surface. This order agrees with the photocat-
alytic oxidation rates reported by Hodgson et al. [27] and Obee and
Hay [28]. Thus, it is inferred that the PCO reaction rate is closely
related to the adsorption process.
Another obvious trend has been observed from the results of
TiO2 /FGFs. That is, for the compounds in the same chemical class,
the adsorption coefficient increases with the increase of molecular
weight and boiling point. This agrees with the adsorption char-
acteristics for sorption-based media. For the structurally similar
compounds, the higher the boiling point is the greater the inter-
molecular forces. Hence, the van der Waals forces of the heavier
compounds make them more likely to be adsorbed to the TiO2
catalyst.
Adsorption results of CCFs and TiO2 /CCFs demonstrate that CCFs
belongs to a non-polar material due to the higher adsorption capac-
ity of hexane compared with that of ethanol (see Fig. 3). Moreover,
adsorption capacity of original CCFs is higher than that of TiO2 /CCFs
for both hexane and ethanol since the BET surface area of CCFs is
larger than that of TiO2 /CCFs, which results in the fact that adsorp-
tion sites on CCFs are obviously more than that those on TiO2 /CCFs.
4.4. Effect of relative humidity
The effect of RH on the adsorptive performance of three
substrates was investigated in the range of 9% (2300 ppm)–70%
(19,600 ppm) at room temperature of 22.8 ± 0.5 ◦ C. From Fig. 6, it
can be noticed that the adsorption coefficient of each tested com-
pound decreases dramatically as RH rises from 9% to 70% for three
substrates. Obviously, the presence of water vapor plays an impor-
tant role on the adsorption behavior of VOCs on the catalyst.
Fig. 7 shows the variation of KL as a function of RH. This fig-
ure indicates that KL varies exponentially: an increasing amount of
physically adsorbed water vapor significantly results to reducing
the adsorbed organic molecules at the surface of catalyst. This phe-
nomenon can be attributed to the competition of water molecules
and VOCs molecules at the surface sites. Hydrogen bonding for
water is stronger than van der Waals interactions for most tested
VOCs so that water is more easily adsorbed on the surface of photo-
catalysts. According to Maudhuit et al. [18], the formation of one
Table 3
BET surface areas and pore structure parameters for three PCO filters.
Parameter
BET surface area (m2 /g)
BJH cumulative desorption pore volume (cm3 /g)
BJH desorption average pore diameter (nm)
TiO2 loading (wt%)
L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349
CCFs TiO2/CCFs TiO2/FGFs
1.2
Pore volume (cm 3/g)
1
0.8
0.6
0.4
0.2
0.6 0.8 1 1 10 100
Pore diameter (nm)
(a) (b)
or several layers of water clusters at the surface of TiO2 in humid
air conditions decreases accessibility of the pollutants to active
sites. In addition, decreasing trends of adsorption capacity with
RH also indicates the tested three PCO air filters in this study are
hydrophilic. Fig. 7 shows there is a linear relationship between
ln(KL ) and RH.
4.5. Comparison adsorption performance of TiO2 /FGFs and
TiO2 /CCFs
Table 3 shows that the BET surface area of TiO2 /CCFs is higher
than that of TiO2 /FGFs results in TiO2 /CCFs having a greater adsorp-
tion capacity. This is ascribed to the fact that a large number of
adsorption sites on the TiO2 /CCFs surface are beneficial to rapid
molecular diffusion of VOCs, thereby increasing the adsorption
capacity. Moreover, CCFs as a supporting substrate have high
adsorption ability. When the challenge VOCs come in contact with
the surface of PCO filter, part of the molecules are directly adsorbed
by the TiO2 , while the rest may be physically captured by CCFs. Test
results of hexane shown in Fig. 6(b) are an excellent interpretation
of the significant impact of the substrate on the adsorption per-
formance. For TiO2 /FGFs, the adsorption capacity of hexane was
very low. Even when RH was greater than 9.6 ± 0.6%, the adsorp-
tion phenomena were not clearly observed (Fig. 3(a)). While for
TiO2/ CCFs, the adsorption of hexane was obviously observed at
various RH values (Fig. 3(c)). The adsorption capacity was found
as 1.1098 m3 /g (9.6 ± 0.6%), 0.3358 m3 /g (32.9 ± 2.5%), 0.0981 m3 /g
(48.3 ± 1.8%) and 0.0555 m3 /g (70.2 ± 2.7%), which was greater than
the adsorption capacity of ethanol under the same RH conditions,
respectively (Fig. 3(c)). This is consistent with results reported in
previous studies [29] that removal performance of the granular
activated carbon filters is positively correlated to the VOCs molec-
ular weight.
Fig. 7 shows that the presence of water vapor has less influence
on the adsorption behavior of TiO2 /CCFs than that of TiO2 /FGFs.
The strength of hydrophilicity of a substrate determines the extent
of interactions of adsorbed water film on the TiO2 surface. Hence,
FGFs is more hydrophilic compared with CCFs so that affinity of
water for FGFs is stronger than that of CCFs. This can be interpreted
with the fact that adsorption energy of water vapor for fiberglass is
CCFs TiO2 /CCFs TiO2 /FGFs
1490.8605 887.6638 105.7063
0.260631 0.638990 0.103594
2.75819 9.58866 3.59489
– 14.32 4.63
 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349 347

2
0.9 TiO2/FGFs
1.8
0.8 Hexane(CCFs)

Adsorption coefficienct (m 3/g)

Adsorption coefficienct (m3/g)

1-Butanol 1.6
0.7 Hexane(TiO2/CCFs)
MEK 1.4 Ethanol(CCFs)
0.6 d-Limonene Ethanol(TiO2/CCFs)
1.2
0.5 Ethanol
1
Acetone
0.4 0.8
p-Xylene
0.3 Toluene 0.6
0.2 0.4
0.1 0.2
0 0
0% 20% 40% 60% 80% 0% 20% 40% 60% 80%
RH RH

(a) TiO2/FGFs (b) TiO2/ CCFs and CCFs

Fig. 6. Relation between adsorption coefficient and RH for different VOCs at 22.8 ± 0.5 ◦ C: (a) TiO2 /FGFs, and (b) TiO2/ CCFs and CCFs.

RH (%)
0 20 40 60 80
0
Acetone lnKL = -6.8496RH - 1.3540 (R² = 0.9983)
-1
MEK lnKL = -7.5066RH - 0.4227 (R² = 0.9836)
-2
Ethanol lnKL = -5.7982RH - 1.3543 (R² = 0.9815)
lnKL(m3/g)

-3 lnKL = -7.0135RH + 0.5550 (R² = 0.9977)

1-Butanol

-4 Toluene lnKL = -5.8133RH - 3.4396 (R² = 0.9541)

p-Xylene lnKL = -7.8335RH - 2.0199 (R² = 0.9790)

-5

d-Limonene lnKL = -9.8801RH - 0.4312 (R² = 0.9978)

-6

-7

-8

(a) TiO2/FGFs

RH (%)
0 20 40 60 80
1

-1
lnKL(m3/g)

-2

-3
Hexane(CCFs) lnKL = -2.9326RH + 0.7969 (R² = 0.9619)
-4
Hexane(TiO2/CCFs) lnKL = -5.0844RH + 0.4676 (R² = 0.9416)

-5 Ethanol(CCFs) lnKL = -1.4064RH - 0.4912 (R² = 0.9937)

Ethanol(TiO2/CCFs) lnKL = -3.5504RH - 0.7666 (R² = 0.9983)

-6
(b) TiO2/ CCFs and CCFs

Fig. 7. Adsorption profiles of various compounds with relative humidity range from 9.6% to 70.2% at 22.8 ± 0.5 ◦ C for (a) TiO2 /FGFs, and (b) TiO2 /CCFs and CCFs.
348 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349
Fig. 8. Single-pass UV-PCO removal efficiency of ethanol and hexane at three
initial concentration levels: 0.25 ppm, 0.5 ppm, and 1 ppm (flow rate = 170 m3 /h,
RH = 15–30%, irradiance = 29–36 W/m2 ).
0.57 J/mol [30], which is much smaller than that of 27.2 kJ/mol for
activated carbon [31]. Cao et al. [32] reported that TiO2 by itself is a
kind of strongly hydrophilic substance. Therefore, RH parameter, to
some extent, influences the adsorption property of air filters with
surfaces coated with TiO2 .
4.6. Photocatalytic activity of TiO2 /FGFs and TiO2 /CCFs
Since adsorption process is one of important steps involved
in the UV-PCO technology, the study of photocatalytic activity of
TiO2 /FGFs and TiO2 /CCFs is necessary for better understanding of
the influence of adsorption process on the UV-PCO. Fig. 8 shows
that single-pass UV-PCO removal efficiency of TiO2 /CCFs is dis-
tinctly higher than that of TiO2 /FGFs for both compounds at various
concentrations. Generally, the larger specific surface area helps
to increase photocatalytic activity, which can be interpreted from
three aspects. First, more active sites are provided through coat-
ing of TiO2 nano-particles on larger BET surface (887.6638 m2 /g) of
TiO2 /CCFs (Table 3). Second, under the same level of illumination,
extremely dispersed TiO2 particles on the surface of TiO2 /CCFs gen-
erate more negative electron–positive hole pairs which may reacts
with water vapor for the formation of hydroxyl radicals, thus reduc-
ing the recombination of electron–hole pairs. Third, large specific
surface area is conducive to the rapid molecular diffusion of VOCs.
Matos et al. [33] also observed the enhanced photoactivity through
combination of photocatalysis and activated carbon for treatment
of toluene. They demonstrated that the contact interface between
TiO2 and activated carbon permits adsorbed pollutants diffuse from
activated carbon to TiO2 for photooxidation. Then activated carbon
was able to adsorb more pollutants from gas stream, and passed
them to TiO2 again. It is this pollutant transfer cycle, as a driven
force, that promotes the photocatalytic conversion of pollutants.
For TiO2 /FGFs air filter, the ordering of photooxidation activity
is consistent with the rank of adsorption characteristics discussed
above, that is, adsorbed ethanol at the surface of TiO2 /FGFs reacts
more readily with hydroxyl radicals than adsorbed hexane does.
This conclusion is in agreement with the results reported by Hodg-
son et al. [27]. For TiO2 /CCFs air filter, although the adsorption
capacity of ethanol is lower than that of hexane, the PCO removal
efficiency of ethanol is still higher than that of hexane, indicat-
ing in addition to adsorption process, UV-PCO performance is also
affected by the photochemical reactions of VOCs. When heteroge-
neous mass transfer of pollutants from gas phase to solid phase does
not limit the PCO process, surface PCO reaction is the controlling
step during UV-PCO technology.
Therefore, the fundamental mechanisms involved in whole PCO
process are divided into several elemental mass transfer processes
occurring in series, namely (1) convection, diffusion and bound-
ary transfer of contaminants in the air-phase; (2) inter-phase mass
transfer of reactant species; (3) adsorption and UV-PCO reaction in
the solid-phase [34]. The above mentioned UV-PCO test results fur-
ther verify that the adsorption process is an important prerequisite
for PCO reaction. And the large specific surface area could increase
the photocatalytic reaction rate as a local high pollutant concentra-
tion can be formed on the surface of TiO2 by the adsorption of CCFs.
It is worthwhile to mention PCO reaction occurs on the surface of
catalyst fibers illuminated with UV lights. That is, only adsorbed
VOCs on the solid surface that is exposed to UV irradiation effec-
tively participate in UV-PCO reactions. As for the synergistic effect
of adsorption and UV-PCO on the removal of indoor VOCs, it is nec-
essary to conduct more tests in the future to further the knowledge
of two processes affecting the conversion efficiency.
5. Conclusions
In this study, the adsorption performance of TiO2 /FGFs,
TiO2 /CCFs, and CCFs air filters for several compounds at vari-
ous RH conditions was tested using adsorption isotherm method.
This method can be employed to separately examine the adsorp-
tion behaviors of different air filters when the PCO technology
works in the absence of illumination. The isothermal adsorption
curves at low concentration levels were well described by Langmuir
isotherm model. TiO2 /FGFs, TiO2 /CCFs, and CCFs were character-
ized by SEM for morphology and N2 adsorption isotherm for BET
surface area and pore structure. The following conclusions can be
drawn from this study:
(1) Adsorption capacity can be ranked as:
CCFs > TiO2 /CCFs > TiO2 /FGFs, which are consistent with
the order of measured BET surface areas of three air filters. This
shows the fact that the adsorption performance is affected by
the properties of substrates.
(2) The test results of TiO2 /FGFs demonstrate that TiO2 /FGFs air fil-
ter presents hydrophilic property, and compounds with high
polarity show higher affinity to the surface of TiO2 /FGFs air
filter due to the strong intermolecular forces. The adsorption
capacity of the selected chemical classes ranks as follows: alco-
hols > ketones and terpenes > aromatics > alkanes. Moreover, a
larger compound in the same chemical class has a greater
adsorption coefficient because of van der Waals interactions.
The test results of TiO2 /CCFs show CCFs belongs to a non-
polar substrate which prefers to adsorb non-polar compounds.
Therefore, the adsorptive performance is the sum of interac-
tions between surface constitutes and a specific VOC.
(3) An increase of RH decreases the adsorption capacity for three
tested media attributable to the strong hydrogen bonding for
water. In addition, the influence of RH on the adsorption behav-
ior of TiO2 /CCFs is less important than that on TiO2 /FGFs.
(4) Due to the differences in adsorption performance, photocat-
alytic activity of the TiO2 /CCFs air filter is obviously higher
than that of TiO2 /FGFs. High performance of TiO2 /FGFs provides
a promising direction to explore other supporting substrates
with high adsorption ability.
Acknowledgements
The authors would like to express their gratitude to the Natu-
ral Science and Engineering Research Council Canada (NSERC) for
the financial support through a CRD grant and Circul-Aire, Inc. for
the support in design and construction of the experimental facility
at Concordia. Electron microscopy research group of McGill Uni-
versity is acknowledged for the help of SEM analysis. The authors
thank Dr. Raymond Le Van Mao and Mr. Hai Tao Yan for technical
assistance of BET analysis.
 L. Zhong et al. / Journal of Hazardous Materials 243 (2012) 340–349
References
[1] B. Wang, R. Mortazavi, F. Haghighat, Evaluation of modeling and measurement
techniques of ultraviolet germicidal irradiation effectiveness – towards design
of immune building, Indoor Built Environ. 18 (2009) 101–113.
[2] L. Zhong, F. Haghighat, P. Blondeau, J. Kozinski, Modeling and physical inter-
pretation of photocatalytic oxidation efficiency in indoor air applications, Build
Environ. 45 (2010) 2689–2697.
[3] A. Mills, S.L. Hunte, An overview of semiconductor photocatalysis, J. Photochem.
Photobiol. A Chem. 108 (1997) 1–35.
[4] T.L. Thompson, J.T. Yates, Surface science studies of the photoactivation of TiO2 –
new photochemical process, Chem. Rev. 106 (2006) 4428–4453.
[5] V.I. Gaya, A.H. Abdullah, Heterogeneous photocatalytic degradation of organic
contaminants over titanium dioxide: a review of fundamentals, progress and
problems, J. Photochem. Photobiol. C 9 (2008) 1–12.
[6] O. Debono, F. Thevenet, P. Gravejat, V. Hequet, C. Raillard, L. Lecoq, N. Locoge,
Toluene photocatalytic oxidation at ppb levels: kinetic investigation and carbon
balance determination, Appl. Catal. B Environ. 106 (2011) 600–608.
[7] J. Krysa, M. Keppert, J. Jirkovsky, V. Stengl, J. Subrt, The effect of thermal treat-
ment on the properties of TiO2 photocatalyst, Mater. Chem. Phys. 86 (2004)
333–339.
[8] H. Liu, X. Ye, Z. Lian, Y. Wen, W. Shangguan, Experimental study on photocat-
alytic oxidation of formaldehyde and its by-products, Res. Chem. Intermediat.
32 (2006) 9–16.
[9] M. Sleiman, P. Conchon, C. Ferronato, J.M. Chovelon, Photocatalytic oxidation of
toluene at indoor air levels (ppbv): towards a better assessment of conversion,
reaction intermediates and mineralization, Appl. Catal. B Environ. 86 (2009)
159–165.
[10] Q.L. Yu, M.M. Ballari, H.J.H. Brouwers, Indoor air purification using heteroge-
neous photocatalytic oxidation. Part II: kinetic study, Appl. Catal. B Environ. 99
(2010) 58–65.
[11] V. Tomasic, F. Jovic, Z. Gomzi, Photocatalytic oxidation of toluene in the gas
phase: modeling an annular photocatalytic reactor, Catal. Today 137 (2008)
350–356.
[12] R.G. Changrani, G.B. Raupp, Two-dimensional heterogeneous model for a
reticulated-foam photocatalytic reactor, AIChE J. 46 (2000) 829–842.
[13] A. Bouzaza, C. Vallet, A. Laplanche, Photocatalytic degradation of some VOCs in
the gas phase using an annular flow reactor—determination of the contribution
of mass transfer and chemical reaction steps in the photo-degradation process,
J. Photochem. Photobiol. A Chem. 177 (2006) 212–217.
[14] F. Shiraishi, T. Ishimatsu, Toluene removal from indoor air using a miniaturized
photocatalytic air purifier including a preceding adsorption/desorption unit,
Chem. Eng. Sci. 64 (2009) 2466–2472.
[15] K. Demeestere, J. Dewulf, H.V. Langenhove, B. Sercu, Gas–solid adsorption of
selected volatile organic compounds on titanium dioxide Degussa P25, Chem.
Eng. Sci. 58 (2003) 2255–2267.
[16] T. Tomida, N. Okada, M. Katoh, Adsorption and photocatalytic decomposition of
volatile organic compounds on photocatalyst of TiO2 - silica beads, Adsorption
11 (2005) 865–869.
[17] J. Chen, G. Li, Z. He, T. An, Adsorption and degradation of model volatile organic
compounds by a combined titania-montmorillonite-silica photocatalyst, J. Haz-
ard. Mater. 190 (2011) 416–423.
349
[18] A. Maudhuit, C. Raillard, V. Hequet, L.L. Coq, J. Sablayrolles, L. Molins, Adsorp-
tion phenomena in photocatalytic reactions: the case of toluene, acetone and
heptane, Chem. Eng. J. 170 (2011) 464–470.
[19] D.W. VanOsdell, Evaluation of test methods for determining the effectiveness
and capability of gas-phase air filtration equipment for indoor air applications-
phase I: literature review and test recommendations, ASHRAE Trans. 100
(1994) 511–523.
[20] NIOSH Pocket Guide to Hazardous Chemicals, National Institute of Occupa-
tional Safety and Health, 2010, http://www.cdc.gov/niosh/npg/ (last accessed
05.03.12).
[21] ASHRAE, ANSI/ASHRAE Standard 145.1: Laboratory Test Method for Assessing
the Performance of Gas-Phase Air-Cleaning Systems: Loose Granular Media,
American Society of Heating, Refrigerating and Air-Conditioning Engineers,
Inc., Atlanta, Georgia, 2008.
[22] F. Haghighat, C.S. Lee, B. Pant, G. Bolourani, N. Lakdawala, A. Bastani, Evaluation
of various activated carbons for air cleaning: towards the design of immune and
sustainable buildings, Atmos. Environ. 42 (2008) 8176–8184.
[23] C.S. Lee, L. Zhong, D. Farhanian, C. Flaherty, F. Haghighat, Development of a
parallel test system for the evaluation of UV–PCO systems, in: Cold Climate
Conference, Calgary, Alberta, Canada, 2012.
[24] E.P. Barrett, L.G. Joyner, P.P. Halenda, The determination of pore volume
and area distributions in porous substances. I. Computations from nitrogen
isotherms, J. Am. Chem. Soc. 73 (1951) 373–380.
[25] T. Guo, Z. Bai, C. Wu, T. Zhu, Influence of relative humidity on the photocatalytic
oxidation (PCO) of toluene by TiO2 loaded on activated carbon fibers: PCO rate
and intermediates accumulation, Appl. Catal. B Environ. 79 (2008) 171–178.
[26] J.M. Coronado, M.E. Zorn, I.T. Tejedor, M.A. Anderson, Photocatalytic oxidation
of ketones in the gas phase over TiO2 thin films: a kinetic study on the influence
of water vapor, Appl. Catal. B Environ. 43 (2003) 329–344.
[27] A.T. Hodgson, H. Destaillats, D.P. Sullivan, W.J. Fisk, Performance of ultravio-
let photocatalytic oxidation for indoor air cleaning applications, Indoor Air 17
(2007) 305–316.
[28] T.N. Obee, S.O. Hay, The estimation of photocatalytic rate constants based
on molecular structure: extending to multi-component systems, J. Adv. Oxid.
Technol. 4 (1999) 1–6.
[29] H. Kholafaei, C.S. Lee, F. Haghighat, Full-scale test method for the performance
evaluation of in-duct filters with mixture of VOC, in: Clima Conference, Turkey,
2010.
[30] R.J. Huang, T. Demirel, T.D. Mcgee, Calculation and interpretation of surface free
energy of wetting of E-glass by vapors, J. Am. Ceram. Soc. 56 (1973) 87–91.
[31] B.F. Delage, P. Pre, P.L. Cloirec, Effects of moisture on warning of acti-
vated carbon bed during VOC adsorption, J. Environ. Eng. 125 (1999)
1160–1167.
[32] L. Cao, Z. Gao, S.L. Suib, T.N. Obee, S.O. Hay, J.D. Freihaut, Photocatalytic oxi-
dation of toluene on nanoscale TiO2 catalysts: studies of deactivation and
regeneration, J. Catal. 196 (2000) 253–261.
[33] J. Matos, A. Garcia, J.M. Chovelon, C. Ferronato, Combination of adsorption
on activated carbon and oxidative photocatalysis on TiO2 for gaseous toluene
remediation, Open Mater. Sci. J. 4 (2010) 23–25.
[34] L. Zhong, F. Haghighat, Modeling and validation of a photocatalytic oxidation
reactor for indoor environment applications, Int. J. Chem. Eng. Sci. 66 (2011)
5945–5954.