Photo-catalytic oxidation reaction of gaseous mercury over titanium dioxide nanoparticle surfaces

Article abstract of DOI:10.1016/j.cplett.2010.03.062

Hg⁰(g) is known to undergo photo-catalytic oxidation by UVA-irradiated TiO₂ surfaces. One micrometre layers of TiO₂ on quartz glass were irradiated within the 240-800 nm range. Gaseous mercury was measured by mass spectrom

Full text of DOI:10.1016/j.cplett.2010.03.062

Chemical Physics Letters 491 (2010) 23–28
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Photo-catalytic oxidation reaction of gaseous mercury over titanium dioxide
nanoparticle surfaces
*
Graydon Snider, Parisa Ariya
Department of Chemistry, McGill University, 801 Sherbrooke St. W., Montreal, PQ, Canada H3A 2K6
Department of Atmospheric and Oceanic Sciences, McGill University, 801 Sherbrooke St. W., Montreal, PQ, Canada H3A 2K6
a r t i c l e i n f o
a b s t r a c t
Hg0(g) is known to undergo photo-catalytic oxidation by UVA-irradiated TiO2 surfaces. One micrometre
layers of TiO2 on quartz glass were irradiated within the 240–800 nm range. Gaseous mercury was mea-
sured by mass spectrometry single ion monitoring. The surface configuration and elemental characteriza-
tion of TiO2 layer was evaluated using scanning electron microcopy with energy dispersive spectroscopy.
The LH adsorption constant of was found to be KHg = (5.1 2.4) ꢀ 10ꢁ14 cm3 and an apparent surface
deposition rate of k = (7.4 2.5) ꢀ 1014 minꢁ1 cmꢁ2 under experimental conditions. Water did not affect
the rate constant. We show TiO2 could be employed to reduce mercury concentrations in gas streams,
even at very high Hg0 concentrations.
Article history:
Received 4 March 2010
In final form 22 March 2010
Available online 27 March 2010
Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction
Titanium dioxide (TiO2) is a popular heterogeneous catalyst for
hydrocarbon oxidation whose surface properties [8] and photocat-
Mercury is a neurotoxic heavy metal [1] released anthropogen-
ically from coal combustion and trash incineration [2]. For in-
stance, nearly half of all power in the United States is derived
from coal [3]. The only long-term solutions for these problems
are drastic reductions on our dependence of non-renewable goods
and maximizing recycling. The use of coal as an energy source is
likely to grow in the coming decade despite research into alterna-
tive energy sources [4]. Interim goals are necessary to reduce hea-
vy metal emissions in the atmosphere. In parallel to the numerous
proposals for CO2 reduction through carbon sequestration, remov-
ing toxic heavy metals from combustion is an equally open and ac-
tive area of research [5]. Here we aim to study the capacity of
titanium dioxide for removing gaseous mercury from air with a fo-
cus on the physical chemistry of surface adsorption, and the effects
of water on uptake efficiency.
alytic potential [9] have been extensively studied. TiO2 can crystal-
lize into rutile, anatase, or (non-photolytic) brookite structures
[10]. A mixture of anatase and rutile TiO2 are typically found in
oxide film coatings [11]. Titanium dioxide is a candidate for scav-
enging gaseous mercury. It has long been known that TiO2 powders
and films are photolytically active under UVA (320 6 k 6 400 nm)
light [11]. Hydrocarbons adsorbed to the surface of TiO2 are oxi-
dized heterogeneously. Such a system has been used to decompose
harmful hydrocarbons and bacteria [8]. Elemental mercury can be
similarly oxidized under room temperature conditions in air, to
mercury oxide [12,13], which is a non-volatile solid characterized
by nano-scale zigzag chains of Hg–O.
The threshold energy required to generate electron–hole pairs
(i.e. excite from the valence to conduction band) in titania is
3.2 eV or about 380 nm (UVA light) [14]. The mechanism of
photo-catalytic oxidation depends to some extent on the oxidant
reaction in question. Here it begins with an adsorbed oxygen mol-
ecule that traps the ‘free’ electron and reducing it to superoxide:
Gaseous mercury, Hg0(g), has a long atmospheric lifetime of
0.5–2 years [6], allowing emissions to disperse globally. Gaseous
elemental mercury deposition, both wet and dry, deposits into
aquatic ecosystems transformed into methyl mercury by sulfate-
reducing bacteria [7]. Methylated mercury can be incorporated
and biomagnified through the food chain eventually leading to fish
advisories from the increasingly dangerous levels of methyl mer-
cury found in edible fish [7].
O2 þ eꢁ ! Oꢀ2ꢁ
ð1Þ
The hole may then oxidize water to the hydroxyl radical:
H2O þ hþ ! OHꢀ þ Hþ
ð2Þ
OH was then assumed to oxidize the adsorbed mercury into HgO(s)
[15].
* Corresponding author at: Department of Atmospheric and Oceanic Sciences,
McGill University, 801 Sherbrooke St. W., Montreal, PQ, Canada H3A 2K6.
E-mail addresses: gradyon.snider@mcgill.ca (G. Snider), parisa.ariya@mcgill.ca
(P. Ariya).
It is assumed that HgO(s) will both decompose into Hg0(g)
above 500 °C [16] and is soluble in nitric acid, whereas TiO2 melts
0009-2614/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2010.03.062

24
G. Snider, P. Ariya / Chemical Physics Letters 491 (2010) 23–28
at 1560 °C and is insoluble in most acids [17]. The proper means for
disposing chemisorbed HgO remains to be determined and is a par-
allel subject of research [18].
a constant flow (1.5 mL minꢁ1) of ultra pure helium. During chro-
matographic runs, we typically kept the GC oven isothermal at
45 °C (0 °C = 273.15 K) for 1 min and increased the temperature
at a rate of 25 °C minꢁ1 from 45 to 80 °C.
The quartz disk was inserted into the flask, a volume of 950 mL.
Except for quartz window, the flask was made of borosilicate glass.
Mercury and toluene samples were added to the reaction flask via
vacuum line and gas syringe, respectively, and measured for con-
sistency via GC/MS. The reaction flask was coated with MTO-Halo-
carbon Wax (Supelco) to inactive surface adsorption of reactants,
products, or reaction intermediates to lead undesired side and sec-
ondary reactions due to non-TiO2 surfaces. The effect of wax coat-
ing has been previously studied by our group [30].
Radiation was produced from a 100 W Hg lamp housed in a
casing (Oriel, 6281 and 60076) attached with a rear reflector. The
radiation power was measured with a UVA detector (PMA2110,
Solar Light Company, Inc. 370 nm peak response). At 15 cm, radia-
tion power was approximately 66 5 mW/cm2. Temperature of
reaction was based on ambient conditions, T = 24 2 °C, P =
770 5 torr. Estimation of errors was determined based on daily
temperature, pressure, and UV emission fluctuations.
Scanning electron microscopy (SEM) and electron dispersive
X-ray spectroscopy (EDS) were performed before and after mer-
cury surface deposition on TiO2 film. What appeared as mercury
deposits were visible in SEM as lightly coloured (white) deposits.
The HgO deposits appeared concentrated in particular regions.
Focusing the X-ray beam on these regions showed the presence
of mercury (Fig. 2). No visible or chemical signs of mercury were
found before irradiation and grey areas in Fig. 2 had a minimum
of mercury. Switching to a topographical display (Fig. 3), the
deposits were apparently localized on the summits of TiO2
growths.
A growing body of research is focused on the oxidation of Hg0(g)
by UVA (320–400 nm) irradiated TiO2 (e.g. Wu et al. [13], Lee et al.
[19], Pitoniak et al. [20] and Prairie et al. [21]). For instance, Li and
Wu [22,23] oxidized Hg0(g) using TiO2–SiO2 nanocomposites, and
Rodríguez et al. [15] oxidized mercury over TiO2 coated on quartz
irradiated at 320 6 k 6 400 nm (UVA). The application of TiO2 films
to mercury oxidation is attractive as it can be performed at room
temperature. Several methods require higher temperatures
(T > 150 °C), such as selective catalytic reduction (SCR), iron oxide
coatings (Fe2O3), fly ash surfaces, or aluminum oxide (Al2O3)
[24,25]. At sufficiently high temperatures (>250 °C), TiO2 also pro-
vides a catalytic surface without UV irradiation [26]. TiO2 can scav-
enge mercury passively through oxide-coated windows while
utilizing solar UV radiation [27]. Such a system may even function
under relatively high mercury concentrations.
In this laboratory, we have previously studied various kinetic,
thermochemical and mechanisms of mercury in gaseous and aqua-
tic phases [28–31]. We have also studied heterogeneous oxidative
surfaces for instance, oleic acid oxidation by ozone over thin water
films via attenuated total reflectance Fourier transform infrared
spectroscopy [32]. In the present study, we compared with litera-
ture sources the mechanisms and kinetics of mercury capture over
irradiated titanium dioxide. Previous studies used significantly
lower mercury concentrations (1–10 ppb) than our experiments.
Consequently we have investigated the exhaustibility of TiO2 over
repeated collections of mercury oxide. We demonstrate that TiO2
could be employed as an efficient means to reduce mercury con-
centrations in gas streams, even at very high (1–2 ppm) elemental
mercury concentrations. We also attempted to resolve whether
high and low levels of water vapour inhibit or promote mercury
surface oxidation.
2.2. Materials
An initial concentration of 1–2 ppm Hg (1 ppm = 2.46 ꢀ
1013 molecules cmꢁ3 at T = 25 °C, P = 760 torr) was used in experi-
ments. Titanium (IV) isopropoxide (97%) was obtained from
Aldrich. HPLC-grade isopropyl alcohol (99.7%), HPLC-grade acetone
(99.5%), and 68–70% Nitric acid were all used as delivered from
ACP Chemicals. HPLC-grade toluene (99.8%) was obtained from
Fisher chemicals.
Standard deviations between repeated experimental trails were
performed when available. For individual datum, errors were esti-
mated from equipment uncertainties.
2. Method
2.1. TiO2 coating procedure
TiO2 coating procedure was followed from Fernandez et al. [33].
Thirteen milliliter of Ti(IV) isopropoxide was added to 87 mL of iso-
propyl alcohol and stirred together in a small beaker. Solution de-
picted slight yellow colour likely due to oxidation with water
vapour. A circular disk of quartz (4.8 cm diameter, 0.32 cm thick)
was lowered edge-wise into the solution. The disk was slowly re-
moved from solution over 10–15 s and left to oxidize from mois-
ture in ambient air for ca., 3 min. Disk was dipped in solution
twice more for a total of three coats. Quartz disk was then heated
to 400 °C in an oven for 2 h. Opaque white coating was observed on
disk. TiO2 surface was washed progressively with nitric acid,
3. Results and discussion
It is reported that the heterogeneous rate of reaction is propor-
tional to the available surface area and light intensity [15,34,35],
r ¼ khIa
ð3Þ
18.2 MX Milli-Q water, and HPLC-grade acetone.
TiO2 coating was also removed from one side of disk. Disk was
weighed (TiO2 mass = 1.9 0.1 mg), and density was given as
where k is the deposition rate constant [units: molecule minꢁ1
a
cmꢁ2 (mW/cm2)ꢁ ], I is the UV light intensity, and h the fraction
qTiO ¼ 3:8 g=cm3 [17]. Approximate average thickness, h, of TiO2
2
of available surface. We have assumed a constant intensity of UV
coating was estimated as h = TiO2 mass/(disk area ꢀ TiO2 den-
a
light, so that I has been incorporated implicitly into k, i.e. k0 ꢃ kI
sity) ꢂ 0.3
lm. The calculated thickness compares reasonably with
(prime is omitted henceforth). Fluctuations in light sources are
therefore an experimental source of error. Further assuming mer-
cury adsorption obeyed the Langmuir–Hinshelwood mechanism
Fernandez et al. [33], who obtained a thickness of 0.2
lm (no error
reported).
The apparent rate constant of the catalytic photo-oxidation of
Hg0(g) over TiO2 reaction was determined by measuring the rela-
tive loss of [Hg0(g)] via electron impact (EI) ionization mass spec-
trometry (HP-5973). We performed the separation of Hg0(g) from
other constituents on a gas chromatograph (HP-6890) equipped
with a 30 m ꢀ 0.25 mm i.d. ꢀ 1.0 mm o.d. crossed-linked phenyl-
methyl–siloxane column (HP5-MS). The column was operated at
KHg½Hgꢄ
h ¼
ð4Þ
1 þ KHg½Hgꢄ
where KHg was the Langmuir adsorption constant (units: cm3/mol-
ecule) and [Hg] was the concentration of mercury in the flask, by
combining (1) and (2) to form a predictive rate law, we obtained:

G. Snider, P. Ariya / Chemical Physics Letters 491 (2010) 23–28
25
1
dHg
KHg½Hgꢄ
mercury losses, converting them into mercury losses per unit
volume.
Rate ¼ ꢁ
¼ k
ð5Þ
ATiO dt
1 þ KHg½Hgꢄ
2
Integrating over [0,t] and [[Hg]0, [Hg]t], we obtained:
ꢀ
ꢁ
d½Hgꢄ ATiO2 k KHg½Hgꢄ
1
KHg
½Hgꢄ
k
Vf
½Hgꢄ0t
ꢁ
¼
ð50Þ
½Hgꢄ0 ꢁ ½Hgꢄt þ
ln
¼
ATiO tirrad:
ð6Þ
2
dt
Vf 1 þ KHg½Hgꢄ
where tirrad. is the irradiation time of the TiO2 plate. Re-arranging,
ATiO was the area of the TiO2 disk and Vf was the volume of the
2
flask. The left hand side of the equation, Vf was divided by absolute
lnð½Hgꢄ0=½Hgꢄ Þ ATiO kKHg
tirrad:
½Hgꢄ0 ꢁ ½Hgꢄt
½Hgꢄ0 ꢁ ½Hgꢄtt
Vf
2
¼
ꢀ
ꢁ KHg
ð60Þ
lnð½Hgꢄ0
=½Hgꢄ Þ versus x ¼
yielded an intercept ꢁKHg
½Hgꢄ0ꢁ½Hgꢄt
tirrad:
t
Plotting y ¼
½Hgꢄ0ꢁ½Hgꢄt
and slope AkKHg/Vf. The intercept hence must be negative to have
physical meaning.
The associated uncertainties are
d½Hgꢄt
2
dy ¼
2 fðlnð½Hgꢄ0=½HgꢄtÞ ꢁ ½Hgꢄ0=½Hgꢄt þ 1Þ
ð½Hgꢄ0 ꢁ ½HgꢄtÞ
2
1=2
þ ðlnð½Hgꢄt=½Hgꢄ0Þ ꢁ ½Hgꢄt=½Hgꢄ0 þ 1Þ g
(
)
1=2
ꢀ
ꢁ
2 þ 2
ꢀ
ꢁ
2
tirrad:
½Hgꢄ0 ꢁ ½Hgꢄt
dtirrad:
tirrad:
d½Hgꢄt
½Hgꢄ0 ꢁ ½Hgꢄt
dx ¼
assuming d½Hgꢄt ¼ d½Hgꢄ0 ꢅ 50 ppb; dt ꢅ 0:04 min.
The values k and KHg can be solved knowing the area of the TiO2
plate and the flask volume. Typical plot of mercury loss via Hg
lamp is shown in Fig. 1. The decay was initially proportional to irra-
diation time; afterward the decay logarithmically decreased with
time. Rate constants are shown in Table 1.
Fig. 1. Plotting Eq. (4), monitoring mercury (Hg0(g)) losses from UVA-irradiated
TiO2. While possible concomitant reactions are occurring to explain this graph, the
LH mechanism is the best approximation we currently have.
Fig. 2. (a) SEM image (3000ꢀ magnification) of TiO2 displaying white patches, confirmed to be HgO via EDX. (b) EDX of marked area shows presence of TiO, and Hg.

26
G. Snider, P. Ariya / Chemical Physics Letters 491 (2010) 23–28
linearity of logarithmic plots at lower mercury concentrations
(<1 ppm) versus time indicates the LH mechanism is more valid
for our experiments.
3.2. Comparison of calculated KHg and k to literature
To our knowledge, no direct measurement of KHg for photo-acti-
vated titanium dioxide surface values has been obtained in the lit-
erature, and our current value represents the first estimation. We
compared with the model of Rodriguez et al. [15] by deriving an
interpretive value for KHg. Rodriguez’s rate-loss for mercury pre-
dicted for trace water vapour conditions was given as:
d½Hgꢄ
1
ba½Hgꢄ
ꢁ
¼
ð7Þ
dt
2 ð1 þ c½HgꢄÞ
In our experiments where low water vapour concentrations were
used (RH 6 1%), Eq. (5) seemed analogous to our own Eq. (3). Com-
paratively, adsorption constant KHg in Eq. (3) corresponded to con-
stant ‘c’, hence:
KHg ¼ c
ð8Þ
Rate constant ‘k’ was also equivalent coefficient in Eqs. (3) and (5):
Vf ab
k ¼
ð9Þ
ATiO 2c
2
Converted values were shown in Table 1.
The reported KHg and k values were found to be within the same
range of magnitudes as Rodriguez et al. [15]. Li and Wu [22], how-
ever, clearly show dissimilar values; their calculated KHg was 30
times larger while k was 10ꢁ5 times smaller. The agreement be-
tween Rodriguez’s data is likely explained by both experiments
having used a pure TiO2 surface rather than a SiO2–TiO2 composite.
It is noteworthy that our operating temperature was lower while
UV light intensities were much higher than Rodriguez et al., the
light intensity effect was studied in the following section.
Fig. 3. (a) TiO2 image using back-scattered electron (BSE). BSE indicates there are at
least two distinct compounds. Deposits are circled. (b) Imaging using environmen-
tal secondary electron detector (ESED). ESED topographically shows HgO is located
on peaks of TiO2.
3.3. Sources of uncertainties
The percentage errors for KHg and k were 47% and 34%, respec-
tively. The error ranges are large, likely due to several factors: (A)
UV intensity varies with distance/angle of lamp while TiO2 films
also varied in thickness depending on XY-surface position. (B) Sur-
face TiO2 temperatures may vary from the measured bulk temper-
ature of the flask. (C) Cleanliness of the TiO2 plate. (D) The presence
of any volatile organic compounds (VOCs). TiO2 is known to oxidize
most VOCs, thus competing with Hg deposition. We did not, how-
ever, see any MS signals for VOCs except acetone, whose concen-
tration remained unchanged during experiments.
3.1. Evaluation of Langmuir–Hinshelwood mechanism
The LH mechanism assumes the reactants (mercury and water)
first adsorb onto the surface and then react on the same surface.
Mechanistically, mercury appears that it must adsorb onto the sur-
face of the TiO2 film before reacting; experiments performed with-
out the TiO2 catalyst in the presence of UV light showed no signs of
mercury oxidation. The adsorption of water is quite strong accord-
ing to previous studies [36]. Finally, the HgO(s) deposit clearly
formed onto the surface to such an extent that it became visible
after several hours of continuous UV exposure.
The Eley–Rideal mechanism [5] is an alternative explanation,
whereby water adsorbs to the surface but mercury reacts while
remaining in the gas phase. This mechanism would imply Hg0(g)
concentrations are linear with reaction rates. Experimentally, the
3.4. Light intensity
Light intensity has been shown to affect the Langmuir adsorp-
tion constant [37]. We measured a broad range of UVA intensities
from the lamp, ranging from 44 to 70 mW cmꢁ2 depending on the
Table 1
Langmuir adsorption constant KHg and rate constant k with comparison of affecting parameters.
Ref.
T
(°C)
UV power
(mW/cm2)
[Hg]
(ppb)
Air flow rate
(L/min)
Detection
KHg
k
(cm3/molecule)
(molecule minꢁ1 cmꢁ2
)
A
B
C
23–26
75–80
43
44–70
1.85
4
1000–2000
0.2–0.3
10–80
0
1
2
MS–EI
CVAA
ZAAS–HFM
5.1 2.4 ꢀ 10ꢁ14
7.4 2.5 ꢀ 1014
13 ꢀ 10ꢁ14
3.8 ꢀ 1014
156 ꢀ 10ꢁ14
(2.4 ꢀ 10ꢁ4) ꢀ 1014
(A) This work, UV source: Oriel 100 W Hg lamp (UVA and B emission), UVA detector: SolarLight PMA2110, MS–EI: HP 5973 Mass Spectrometer Electron Ionization (B)
Rodriguez et al. [15], UV source: XX-40 Spectronics 80 W (UVA emission), CVAA: cold vapour atomic adsorption (Shimadzu UV-1201S Spectrometer), UV detector unknown.
(C) Li and Wu [22,23], UV source unknown, UV detector: UVX radiometer with UVX-36 sensor probe (335–380 nm range), ZAAS–HFM: Zeeman atomic absorption spec-
trometry using high frequency modulated light polarization (OhioLumex Co. RA-915+).

G. Snider, P. Ariya / Chemical Physics Letters 491 (2010) 23–28
27
angle and position of the detector. Hence our Oriel lamp was not a
homogenous light source. An ‘average’ intensity near the centre of
the beam is given here as 60 mW cmꢁ2. According to Fujishima et
al., given this high intensity range of light, our experiments may
actually lie in a mass transport-controlled region of space [34].
Hence UV light may have saturated the reaction rate and lamp fluc-
tuations could be of small concern.
compounds are recommended as such species are known to inhibit
mercury adsorption [27,40,41].
4. Conclusions
Titanium dioxide is an attractive method of oxidizing gaseous
mercury using potentially safe, low-cost procedures. TiO2 has a
high Hg0 uptake capacity, is relatively cheap (1.09–1.19 USD/lb
[42]), and environmentally benign (e.g. currently used in tooth-
paste and suntan lotion). The power cost of running continuous
UV lights remains a problem [40]. Ultra Violet LEDs will save on en-
ergy, hence total cost, provided the LEDs themselves are inexpen-
sive and sufficiently [43]. TiO2 doped for a shift in visible light
conversion will may allow for the use of sunlight radiation in mer-
cury capture [27].
We have measured the overall rate constant k and Langmuir
adsorption constant KHg in dry and humid air at room temperature
and pressure. Measured values were comparable to Rodriguez [15]
but clearly distinct from Wu and Li [22,23]. We have addressed the
impact of water vapor on the adsorption-oxidation efficiency of
mercury on TiO2 surfaces and did not observe any major impedi-
ments on Hg oxidation process even at higher relative humidities.
As for the utility of TiO2 nanoparticles for Hg0 removal in a coal
plant, it is known SO2 will inhibit TiO2 surface reactions [19] but
this is true of other methods as well [5]. Rising temperatures, espe-
cially above 100 °C, might inhibit oxidation [19,35], which in turn
emphasizes on the potential of TiO2 nanoparticles for industrial
usage, particularly as the downstream and upstream cooling are
part of the existing industrial pollution industries. Life-cycle anal-
ysis of photo-activated titanium oxides methods for removal of
mercury and the secondary reactions in the environment should
be studied to assure its benign nature in the environment. Addi-
tives such as gold nanoparticles [44–46] to titanium dioxide coat-
ings should be explored for enhanced mercury adsorption
properties.
3.5. TiO2 disk characteristics and surface area
The optimal surface density for a TiO2 coating has been sug-
gested to be 0.23 mg cmꢁ2 [38]. Thick films attenuate UV light be-
fore reaching the surface while too-thin films do not fully absorb
the UV light. Our films were approximately 0.1–0.2 mg/cm2 corre-
sponding well with this ‘optimal’ value. We assumed the surface
area of the quartz disk would represent, at least, the perpendicular
area exposed to the UV light, about 18 cm2. Scanning electron
microscopy (SEM) analysis of TiO2 plates (Fig. 2) indicated uniform
coverage over the surface, with uneven thicknesses throughout the
deposition. The surface area in our sample was unknown. Back-
scattering images show the areas with HgO accumulation, whereas
environmental secondary electron detector (ESED) imaging show a
topographical image of TiO2 with sharp peaks and valleys. Using
energy dispersive X-ray (EDX) spectroscopy, Ti, O, and Hg signals
were observed.
3.6. Saturated HgO deposits
Experiments in which a mercury-saturated humid air stream
passed over the TiO2 film created a dark deposit of HgO. We
reached the saturation point of HgO, whereby reactivity of the
TiO2 film ceased. The deposit was visible to the naked eye, however
the exact thickness is not known.
3.7. Mechanism of Hgads oxidation (with and without presence of
water)
References
In our experiment, hydroxyl radicals was expected to oxidize
the adsorbed mercury, Hgads into HgO (the intermediate HgOH is
unstable) [15] on the TiO2 surface. We considered the possibility
that ozone was generated in our reaction chamber, however sev-
eral trails were performed on gaseous mercury without the pres-
ence of TiO2 and no reaction took place.
[1] F. Bakir et al., Science 181 (1973) 230.
[2] E.G. Pacyna, J.M. Pacyna, Water Air Soil Pollut. 137 (2002) 149.
[3] J. Luna-Camara, Electric Power Monthly Energy Information Administration,
2009.
[4] H. Yi, J. Hao, X. Tang, Energy Policy 35 (2007) 907.
[5] A.A. Presto, E.J. Granite, Environ. Sci. Technol. 40 (2006) 5601.
[6] C.H. Lamborg, W.F. Fitzgerald, J. O’Donnell, T. Torgersen, Geochim. Cosmochim.
Acta 66 (2002) 1105.
There has been a dispute as to whether water vapour promotes
or inhibits mercury oxidation [15,22,23]. Mechanistically, water
molecules are thought to be required in generating OH radicals
that oxidize mercury. The rate-limiting concentration of water
needed for mercury oxidation is unclear. In experiments where
we heated the TiO2 plate (to 120 °C) prior to chamber assembly
and flushing the flask with dry air, the rate-loss of mercury
remained constant. Hence we found it unnecessary to add water
deliberately to incur reactivity. We estimate between 5 and
20 ppm of water might be present in these ‘dry’ air experiments.
Over a rutile TiO2 surface, water adsorbs to a significant degree;
degassing can be detected above 300 °C under vacuum as Hydroxyl
groups are chemisorbed to the surface [36]. In more humid
conditions (up to 100% relative humidity at 25 °C), mercury oxida-
tion again remained the same. Excessive water did not present a
strong influence over reactivity, yet is required in only minute
concentrations.
[7] F.M.M. Morel, A.M.L. Kraepiel, M. Amyot, Ann. Rev. Ecol. Syst. 29 (1998) 543.
[8] U. Diebold, Surf. Sci. Rep. 48 (2003) 53.
[9] A.L. Linsebigler, G. Lu, J.T. Yates, Chem. Rev. 95 (1995) 735.
[10] K. Rajeshwar, J. Ibanez, Environmental Electrochemistry, Academic Press Inc.,
San Diego, 1996.
[11] R.I. Bickley, T. Gonzalez-Carreno, J.S. Lees, R.J.D. Tilley, L. Palmisano, J. Solid
State Chem. 92 (1991) 178.
[12] U. Kaluza, H.P. Boehm, J. Catal. 22 (1971) 347.
[13] C.Y. Wu, T.G. Lee, G. Tyree, E. Arar, P. Biswas, Environ. Eng. Sci. 15 (1998) 137.
[14] T.G. Lee, J.E. Hyun, Chemosphere 62 (2006) 26.
[15] S. Rodríguez, C. Almquist, T.G. Lee, M. Furuuchi, E. Hedrick, P. Biswas, J. Air
Waste Manag. Assoc. 54 (2004) 149.
[16] X. Feng, J.Y. Lu, D.C. Grègoire, Y. Hao, C.M. Banic, W.H. Schroeder, Anal. Bioanal.
Chem. 380 (2004) 683.
[17] R.C. Weast, CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton, FL,
2007.
[18] J.D. Noel, P. Biswas, D.E. Giammar, J. Air Waste Manage. Assoc. 57 (2007) 856.
[19] T.G. Lee, P. Biswas, E. Hedrick, Am. Inst. Chem. Engrs. 47 (2001) 954.
[20] E. Pitoniak, C.-Y. Wu, D.W. Mazyck, K.W. Powers, W. Sigmund, Environ. Sci.
Technol. 39 (2005) 1269.
[21] M.R. Prairie, L.R. Evans, B.M. Stange, S.L. Martinez, Environ. Sci. Technol. 27
(1993) 1776.
Although some studies indicate that the presence of moisture is
not necessarily essential to keep up the photocatalytic process
[39], further studies under controlled dry conditions and over the
larger range of water and other potential co-pollutants concentra-
tions such as NOx (=NO + NO2), SOx (=SO2 + SO3), volatile organic
[22] Y. Li, P. Murphy, C.-Y. Wu, Fuel Process. Technol. 89 (2008) 567.
[23] Y. Li, C.-Y. Wu, Environ. Eng. Sci. 24 (2007) 3.
[24] C.S. Turchi, Novel Process for Removal and Recovery of Vapor-Phase Mercury,
ADA Technologies, Inc., Littleton, CO, 2000, p. 57.
[25] E.J. Granite, H.W. Pennline, Ind. Eng. Chem. Res. 41 (2002) 5470.
[26] K.C. Galbreath, C.J. Zygarlicke, Fuel Process. Technol. 65–66 (2000) 289.

28
G. Snider, P. Ariya / Chemical Physics Letters 491 (2010) 23–28
[27] E.J. Granite, W.P. King, D.C. Stanko, H.W. Pennline, Main Group Chem. 7 (2008)
227.
[37] Y. Xu, C.H. Langford, J. Photochem. Photobiol. A 133 (2000) 67.
[38] V. Keller, P. Bernhardt, F. Garin, J. Catal. 215 (2003) 129.
[39] S. Hager, R. Bauer, G. Kudielka, Chemosphere 41 (2000) 1219.
[40] C.R. McLarnon, E.J. Granite, H.W. Pennline, Fuel Process. Technol. 87 (2005) 85.
[41] A.A. Presto, E.J. Granite, Environ. Sci. Technol. 41 (2007) 6579.
[42] L. Terry, Titanium Dioxide (TiO2) Prices and Pricing Information, ICIS, 2009.
[43] A. Fujishima, X. Zhang, C. R. Chim. 9 (2006) 750.
[44] M. Daté, Y. Ichihashi, T. Yamashita, A. Chiorino, F. Boccuzzi, M. Haruta, Catal.
Today 72 (2002) 89.
[45] I.M. Arabatzis, T. Stergiopoulos, D. Andreeva, S. Kitova, S.G. Neophytides, P.
Falaras, J. Catal. 220 (2003) 127.
[28] B. Pal, P.A. Ariya, Environ. Sci. Technol. 38 (2004) 5555.
[29] B. Pal, P.A. Ariya, Phys. Chem. Chem. Phys. 6 (2004) 572.
[30] G. Snider, F. Raofie, P.A. Ariya, Phys. Chem. Chem. Phys. 10 (2008) 5616.
[31] P.A. Ariya, K. Peterson, G. Snider, M. Amyot, in: N. Pirrone, R.P. Mason (Eds.),
Mercury Fate and Transport in the Global Atmosphere, Springer US, New York,
2009, p. 459.
[32] H.-M. Hung, P. Ariya, J. Phys. Chem. A 111 (2007) 620.
[33] A. Fernandez et al., Appl. Catal. B 7 (1995) 49.
[34] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C 1 (2000) 1.
[35] T.G. Lee, P. Biswas, E. Hedrick, Ind. Eng. Chem. Res. 43 (2004) 1411.
[36] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids,
Academic Press, London, 1999, p. 287.
[46] A.A. Presto, E.J. Granite, Platinum Met. Rev. 52 (2008) 144.