Photocatalytic oxidation of cyanide: Kinetic and mechanistic studies

Article abstract of DOI:10.1016/S1381-1169(02)00512-5

The kinetics and mechanism of the photocatalytic oxidation of cyanide in the presence titanium dioxide catalyst were investigated in this study. By displacing the surface hydroxyl groups on the surface of titanium dioxide with fluoride ions, it was deduced that cyanide is oxidised via a pure heterogeneous pathway, i.e. oxidised by the holes trapped at the surface hydroxyl groups ≡TiO^.. The quantum efficiency of the photocatalytic oxidation was found to be low (ca. 0.1-0.2%) and this was mainly due to (1) the low adsorption of cyanide ions onto the titanium dioxide surface, (2) the absence of homogenous reaction between cyanide ions and diffused hydroxyl radicals, and (3) the high electron hole recombination rate within the titanium dioxide photocatalyst. A kinetic model was developed to describe the mechanism involved in the photocatalytic oxidation of cyanide.

Full text of DOI:10.1016/S1381-1169(02)00512-5

Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Photocatalytic oxidation of cyanide:
kinetic and mechanistic studies
K. Chiang∗, R. Amal, T. Tran
Centre for Particle and Catalyst Technologies, School of Chemical Engineering and Industrial Chemistry,
The University of New South Wales, Sydney, NSW 2052, Australia
Received 22 July 2002; accepted 20 August 2002
Abstract
The kinetics and mechanism of the photocatalytic oxidation of cyanide in the presence titanium dioxide catalyst were
investigated in this study. By displacing the surface hydroxyl groups on the surface of titanium dioxide with fluoride ions,
it was deduced that cyanide is oxidised via a pure heterogeneous pathway, i.e. oxidised by the holes trapped at the surface
•
≡
hydroxyl groups TiO . The quantum efficiency of the photocatalytic oxidation was found to be low (ca. 0.1–0.2%) and this
was mainly due to (1) the low adsorption of cyanide ions onto the titanium dioxide surface, (2) the absence of homogenous
reaction between cyanide ions and diffused hydroxyl radicals, and (3) the high electron hole recombination rate within the
titanium dioxide photocatalyst. A kinetic model was developed to describe the mechanism involved in the photocatalytic
oxidation of cyanide.
© 2002 Elsevier Science B.V. All rights reserved.
Keywords: Advanced oxidation technologies; Heterogeneous photocatalysis; Titanium dioxide; Cyanide
1. Introduction
in the literature and many of these are concerned with
the degradation of organic and inorganic compounds
present in the wastewater. The photocatalytic oxida-
tion of cyanide ion is one of the most frequently stud-
ied reactions due to the acute toxicity of the cyanide
ions. One of the earliest studies on heterogeneous
photocatalytic oxidation of cyanide was reported by
Frank and Bard [1]. In most cases, cyanate is found
to be the oxidation product which could be further
oxidised simultaneously on the surface of titanium
dioxide to produce nitrate and carbonate [2–5]. The
hydrolysis of cyanate is also possible which gives rise
to ammonium and carbonate as products as shown in
Eqs. (1) and (2).
In the past 30 years, extensive research has
shown that heterogeneous photocatalysis is a viable
alternative to treat wastewater. The degradation of
contaminants in wastewater using titanium dioxide
photocatalysis is attractive because titanium diox-
ide could be used repeatedly without any noticeable
deterioration of photoactivity and is non-toxic and
economical. The kinetics and mechanism of many
photocatalytic reactions have been described in detail
∗
Corresponding author. Present address: Institute of Environ-
mental Science and Engineering, Nanyang Technological Uni-
versity, Innovation Centre, Block 2, Unit 237, 18 Nanyang
Drive, Singapore 637723, Singapore. Tel.: +65-6794-3864; fax:
+65-6792-1291.
CN− + 2OH− + 2hv+b
→
TiO2+UV OCN− + H2O
(1)
(2)
2−
E-mail address: kchiang@ntu.edu.sg (K. Chiang).
OCN− + 2H2O → NH4+ + CO3
1381-1169/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S1381-1169(02)00512-5

286
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Although many research groups have demonstrated
According to Minero et al. [11,12] the adsorption
of fluoride ions onto titanium dioxide surface occurs
most efficiently at a pH of 3.5. The suspensions were
mixed well using a rotary shaker at 25 ◦C for 12 h
to allow ion exchange to taken place. The titanium
dioxide particles in each suspension was recovered by
vacuum filtration and then re-suspended in water to
remove any un-adsorbed fluoride ions and nitrate ions
on the surface of titanium dioxide. This step was re-
peated five times in order to remove any un-adsorbed
fluoride ions. The ion-exchanged titanium dioxide
samples were finally dried in an oven at 110 ◦C for
12 h and stored under vacuum. The titanium dioxide
samples which had been mixed with pure water or
exchanged with 0.01 M NaF solution were labelled
given the name F0/TiO2, and F1/TiO2, respectively, to
represent the untreated and fluoride-treated titanium
dioxide samples. The fluoride-exchanged titanium
dioxide was used in order to study the effect of surface
hydroxyl group concentration on the photocatalytic
activity of titanium dioxide.
the possibility of degrading cyanide through titanium
dioxide photocatalysis, no conclusion can be made in
most of these studies on the exact reaction mechanism
of cyanide oxidation. The oxidation of cyanide over
titanium dioxide has been reported to occur via both
direct electron transfer to the holes [2,5–7] and homo-
geneous hydroxyl radical attack in the bulk solution
[8–10]. Therefore, the aim of this work is to inves-
tigate the kinetics and mechanism involved to pro-
vide further insight into the photocatalytic oxidation
of cyanide.
2. Experimental procedures
2.1. Materials and apparatus
The titanium dioxide used in the study was De-
gussa P25. The primary crystallite size was found
to be 30 nm as determined from TEM studies. The
material had a surface area of 51 m2/g and contained
predominantly anatase (79% anatase and 21% rutile
as determined from X-ray diffraction). Analytical
grade sodium cyanide (NaCN) and sodium cyanate
(NaOCN) were used as the source of cyanide and
cyanate. Nitric acid and sodium hydroxide were used
to adjust the pH of solution. Ultra pure gas (either air
or N2) were used in all experiments. All solutions were
prepared using freshly prepared Milli-Q water and all
chemicals were used without further purification.
The procedure to prepare fluoride-exchanged ti-
tanium dioxide is as follow. Sodium fluoride (NaF)
was first dissolved into two 500 ml aliquots of water
to give a fluoride concentration of 0.02 M. Two sus-
pensions of titanium dioxide were also prepared by
adding 5 g of Degussa P25 into two separated 500 ml
of water. After sonification for 30 min to break up
loosely aggregated particles, one titanium dioxide sus-
pension was mixed with the prepared 500 ml sodium
fluoride solution for fluoride exchange reaction. An-
other blank titanium dioxide slurry was also prepared
in the same way except it was mixed with 500 ml of
water instead of the sodium fluoride solution. The re-
sultant slurries contained 5 g TiO2/l and final fluoride
concentrations of 0 and 0.01 M, respectively. The pH
of the titanium dioxide suspensions was then adjusted
to 3.5 by the addition of concentrated nitric acid.
2.2. Photoactivity tests
All experiments were carried out using a 1 l flat
bed reactor to assess the photocatalytic degradation of
cyanide. Since the pKa of hydrogen cyanide is equal
to 9.3 at 25 ◦C, it is necessary to work in an alka-
line condition in order to avoid the volatilisation of
toxic hydrogen cyanide gas during the experiment.
Any gaseous hydrogen cyanide formed during the ex-
periment would be captured by a caustic trap (1 M
NaOH).
2.2.1. Series 1
The first series of experiment involves the as-
sessment of the photoactivity of titanium dioxide to
degrade cyanide. One gram of the titanium diox-
ide sample was suspended in 500 ml Milli-Q water
and sonicated for 30 min. A pre-weighted amount of
NaOH was first added to the suspension followed by
the addition of NaCN. The volume was then made
up to 1 l by further addition of water to reach a
pre-determined pH (8.5–12) and cyanide concentra-
tions (0.39–3.85 mM). Unless otherwise stated, air
was bubbled through the suspension at a flowrate of
500 ml/min for 15 min before the reaction started.
The solution was kept aerated while it was irradiated

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
287
for 3–7 h at 30 ◦C After the complete disappearance
of cyanide, the solutions were irradiated further for
2 h more to study the oxidation of cyanate.
concentration of 10 ppm. The pH of the solution was
adjusted by the addition of NaOH and HCl and the
electrophoteric mobility and zeta potential of titanium
dioxide particles were determine in the range of pH
3–12.
2.2.2. Series 2
The pathway of photocatalytic oxidation of cyanide
was studied using fluoride-exchanged titanium diox-
ide. In these experiments, the initial concentration of
cyanide was kept at 3.85 mM with the pH varying
from 9.5 to 12.0, respectively. The concentration of
the photocatalysts (F0/TiO2 and F1/TiO2) was kept at
1 g/l. The concentrations of cyanide and cyanate were
monitored with time.
3. Results and discussion
3.1. Cyanide oxidation by O2/UV/TiO2
It has been reported that the initial rate of photocat-
alytic degradation of many pollutants is a function of
the photocatalyst concentration [13–15]. In fact, the
reactor configuration plays a key role in determining
the photon absorption characteristic of the design and
therefore the optimum catalyst concentration. The ef-
fect of P25 TiO2 concentration (0.1–4.0 g/l) on the rate
of photocatalytic degradation of cyanide was evalu-
ated at pH 12 and the results are shown in Fig. 1.
These results verify that the rate of cyanide degrada-
tion increases with TiO2 loading up to a concentration
of 1.0 g/l (or 0.1 wt.%). The enhancement is due to the
increasing volumetric photon absorption by the pho-
tocatalyst, therefore providing a higher concentration
of the charge carrier per unit volume for cyanide oxi-
dation. However, as the loading was increased beyond
the optimum value, the degradation rate decreased due
to the increased opacity of the suspension and light
scattering. The penetration depth of the UV photon is
lowered and less titanium dioxide particles could be
activated. Therefore, in this study, a TiO2 loading of
1 g/l corresponding to the highest absorption of inci-
dent photons by the photocatalyst was used for subse-
quent experiments.
It was also found that the rate of photocatalytic
degradation of cyanide is dependent on the air
flowrate inside the reactor. The aeration process not
only provides molecular oxygen to slow down the
electron-hole recombination reaction on the titanium
dioxide particles, but also facilitates the mixing of
photocatalyst in the photoreactor. As shown in Fig. 2,
an air flowrate of 500 ml/min provides adequate oxy-
gen concentration and agitation to the system. As
shown in Table 1, the adsorption of free cyanide onto
titanium dioxide particles (10 g/l) at pH 12 in the
absence of light for 24 h is small. This is ascribed to
the negative charge of CN− and a negative surface
2.3. Analysis
The concentration of free cyanide was determined
by potentiometric titration using a standardised silver
nitrate solution. An automated system consisting of
a Metrohm 665 Dosimat and a silver Titrode in con-
junction with a 682 Titroprocessor was used to detect
the end-point of titration. Samples were first filtered
through 0.45 ␮m syringe filters and 1 ml aliquots were
titrated against a standardised AgNO3 solution. This
method is suitable to analyse cyanide concentration
between 1 and 1000 mg/l CN− with an accuracy of
1 mg/l. The concentration of cyanate was determined
spectrophotometrically in a buffered 2-aminobenzoic
acid solution. The concentration of nitrate was de-
termined by using a Shimadzu high-performance ion
chromatography system equipped with a conductivity
detector. A Hamilton PRP-X100 inorganic ion column
(150 mm × 4.1 mm) was employed for the separation
and the mobile phase used was a mixture of 4.0 mM
p-hydroxybenzoic acid and 2.5% methanol at pH 8.9
and temperature of 30 ◦C.
A Philips CM200 Field Emission Gun Electron
Transmission Microscope FEGTEM were used for
micro-structural analysis of the surface of TiO2. The
specific surface areas of all samples were measured
using the single point dynamic surface area analyser,
which allowed the accurate determination of surface
area from 0.5 to 1000 m2/g. A Cary 5 UV-Vis-NIR
spectrophotometer was used to record absorbance
data of the solution samples. The Zeta potential of
the titanium dioxide samples was measured using
the Brookhaven Zetaplus system. Titanium dioxide
samples were suspended in 0.01 M NaCl solution at a

288
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Fig. 1. Effect of titanium dioxide concentration on the rate of photocatalytic cyanide degradation. The solution was at an initial CN−
concentration of 3.85 mM and pH 12.
Fig. 2. Effect of air flowrate on the rate of photocatalytic cyanide degradation at 3.85 mM initial CN− and pH 12.
charge of titanium dioxide particles induced by the
Table 1
adsorption of OH− ions onto the titanium dioxide
Dark adsorption of cyanide ion for 24 h in the absence and presence
surface at high pHs. Fig. 3 shows that the zeta po-
tential of the titanium dioxide particles at pH > 9 in
the presence and absence of cyanide ions (3.85 mM)
are similar. Any change in the point of zero charge
of titanium dioxide at pH lower than 9.3 in the pres-
ence of cyanide ions could not be determined due to
the possibility of cyanide volatilisation which might
affect the accuracy of the measurement. The change
in zeta potential therefore confirms that the titanium
dioxide surface only interacts and adsorbs cyanide
slightly under alkaline condition.
(10 g/l) of TiO2
Initial [CN−]
(mM)
Final [CN−] (mM)
In the absence
of TiO2
In the presence
of TiO2
0.40
0.98
1.89
2.91
3.91
0.39
0.98
1.88
2.91
3.91
0.35
0.87
1.73
2.79
3.71
The initial concentrations of cyanide were 0.40, 1.00, 1.90, 2.89,
and 3.90 mM onto and the pH for adsorption was 12.

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
289
Fig. 3. Zeta potential of P25 TiO2 in the absence (᭹) and in the presence (×) of 3.85 mM CN−. The matrix solution contains 0.01 M
NaCl solution.
Similar to other studies, cyanide ion at various
concentrations (0.39–3.85 mM CN−) can be oxidised
photocatalytically by titanium dioxide [1,2,5,7–10].
Volatilisation of cyanide in the experiment, as de-
termined from the concentration of cyanide from
the caustic trap, was insignificant at the pH and air
flowrate used in the experiments. Fig. 4 shows the
concentration-time profile of cyanide during the pho-
tocatalytic oxidation at pH 12. A constant initial
degradation rate of cyanide was observed at a cyanide
concentration higher than 2.0 mM and a straight line
could best fit the data during the initial course of oxi-
dation. This does not apply for an initial concentration
of cyanides below 2.0 mM. Based on the change in
cyanide concentration and the number of incident
photons impinging on the front window of the reactor
cell, the photonic efficiency (ζr) of a photocatalytic
reaction could be calculated [16,17]. Similarly, the
photonic efficiency of cyanide degradation in this
study was determined using Eq. (3):
ꢀ[CN−] × V × Em
ζr =
× 100%
(3)
a × I0 × ꢀt
where ꢀ[CN−] is the change in concentration of
Fig. 4. Concentration profile of cyanide vs. irradiation time at various initial cyanide concentration at 1 g/l TiO2 and at pH 12. (᭹)
3.85 mM, (᭜) 2.9 mM, (᭡) 2 mM, (᭿) 1 mM, and (×) 0.4 mM.

290
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Table 2
Table 3
Calculated reaction rate and photonic efficiencies (ζ) obtained
from different cyanide concentration at an initial pH of 12
Calculated reaction rate constants (k) and photonic efficiencies (ζ)
obtained from different pH at an initial cyanide concentration of
3.85 mM
Initial cyanide
Parameter
Initial pH of solution
Parameter
concentration (mM)
Rate (M/min)
ζ (%)
r (M/min)
ζ (%)
0.390
0.944
2.036
2.926
3.850
6.18 × 10−6
8.85 × 10−6
1.09 × 10−5
1.18 × 105
0.11
0.15
0.19
0.20
0.20
8.5
9.5
10.5
12.0
1.81 × 10−5
1.37 × 10−5
1.32 × 10−5
1.24 × 10−5
0.29
0.23
0.22
0.21
1.25 × 10−5
further irradiated for 2 h in order to study the degra-
dation of cyanate. The concentration of nitrate was
undetectable throughout the experiments and there-
fore cyanate was confirmed to be the only product of
cyanide degradation. Since there was an insignificant
change in cyanate concentration at the end of the
experiment, it can be concluded that the oxidation of
cyanate was insignificant.
The effect of pH on the rate of cyanide oxidation is
shown in Table 3 and Fig. 6. As the pH of the solution
decreases, the rate of cyanide degradation increases.
At pH 8.5, around 66% of the cyanide disappeared
through volatilisation as hydrogen cyanide. This was
confirmed with the detection of cyanide in the gas trap
and the un-closure mass balance of CN− and OCN−
in the solution at the end of the experiment. Never-
theless, the photocatalytic degradation of cyanide be-
comes more favourable at lower pH as indicated by the
results at pH 9.5, 10.5, and 12. The main cause of the
observed lower cyanide degradation rate at high pHs
is due to the unfavourable adsorption of cyanide ions
onto the more negatively charged surface of titanium
dioxide.
cyanide expressed in mol/l. A summary of the results
for the photocatalytic oxidation of cyanide by P25
TiO2 is tabulated in Table 2. The values displayed in
the table represent the lower limit of the photonic ef-
ficiency. It can be seen that these values are gener-
ally very low and are always around 0.1–0.2%. The
low value of ζr is firstly due to the fact that partic-
ulate TiO2 could never absorb all the incident pho-
tons. The photons emitted from the UV lamp could
be lost through reflection and light scattering due to
the high refractive index of titanium dioxide. It is
well-known that the reflected photon could sometimes
reach 4.5–74% of the incident flux [18] and the theo-
retical maximum quantity of photon absorption could
never exceeds 65% of the incident flux [19]. Secondly,
the charge carriers recombination process is fast, es-
pecially under high photon flux condition, therefore,
consumes the electrons and holes without contributing
to any redox reactions. The results in Table 2 show
that the photonic efficiency increases when the initial
concentration of cyanide in the system increases. It is
clear that the photogenerated holes could be scavenged
more efficiently in the presence of higher cyanide
concentration leading to the increase of the photonic
efficiency.
3.2. Mechanism of cyanide photooxidation
Fig. 5 shows the concentration profiles of cyanide,
cyanate, and nitrate during different experiments at
different initial cyanide concentrations. It can be seen
that as cyanide was being degraded, the concentration
of cyanate increased. The concentration of cyanate
reached maximum when all the cyanide initially
presents in the system was completely degraded. As
shown from the carbon mass balance, the sum of
cyanide and cyanate ion concentrations at any time
during the experiment remained constant. After the
completion of cyanide degradation, the solution was
Many research groups support the direct charge
transfer for the photocatalytic oxidation of cyanide by
titanium dioxide photocatalyst [1,2,5–7,20]. The oxi-
dation of cyanide via indirect pathway mediates by ad-
sorbed hydroxyl or homogeneous pathway by diffused
hydroxyl radical were also reported [8–10]. Reactions
(4)–(20) summarise the major elementary reactions
which contribute to the overall degradation of cyanide
on the surface of titanium dioxide photocatalyst in
present study. Both heterogeneous and homogeneous

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
291
Fig. 5. (a)–(e) Concentration profiles of cyanide and intermediate products plotted against irradiation time at different initial cyanide con-
centration. (᭡) Cyanide concentration, (᭜) cyanate concentration, (᭿) nitrate concentration, and (᭹) total carbon balance (CN− +CNO−).
The concentration of TiO2 is 1 g/l and an initial pH of 12. (a) 3.85 mM, (b) 2.89 mM, (c) 1.93 mM, (d) 0.96 mM, and (e) 0.39 mM.
pathways for cyanide oxidation are also considered in
the reaction scheme. Reaction (9) and (10) represent
the formation of hydroxyl radical HO• from the reac-
tions betw•een adsorb•ed hydroxide ion HO− with ht+r
cyanate ions (OCN−) and formamide (HCONH2) as
shown in reactions (11)–(13). These represent the ho-
mogeneous oxidation of cyanide ions in the solution.
For direct heterogeneous charge transfer pathway,
cyanide ions must be first adsorbed and subseq+uently
react with surface superficial trapped holes htr and
≡
and TiO . The HO radicals could diffuse into the
bulk solution and react with the cyanide ions to form

292
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Fig. 6. The influence of pH on the photocatalytic oxidation of cyanide in the presence of 1 g/l TiO2. (᭹) At initial pH of 12, (×) at initial
pH of 10.5, (᭡) at initial pH of 9.5, and (᭜) at an initial pH of 8.5.
•
≡
TiO as illustrated in reactions (14)–(17).
2H2O2 → 2H2O + O2
(20)
hν
TiO2 → TiO2{ec−b · · · hv+b}
(4)
(5)
(6)
It is well known from the radiation chemistry of
cyanide that the reaction between HO• radical and
cyanide ion leads to the production of formamide radi-
cal [21–24]. Upon disproportionation, two formamide
radicals yield one cyanate ion and one formamide as fi-
nal products. If the photocatalytic oxidation of cyanide
by titanium dioxide follows a pure homogeneous path-
way via HO• radical attack, at least two intermediates
will be formed, namely cyanate ions and formamide.
In addition their concentrations should be close to,
if not equal to, 1:1 stoichiometric ratio. However, as
shown from the carbon balance in Fig. 5, the sum of
cyanide ion and cyanate ion concentration in the sys-
tem was always constant and the mass balance was
closed. There was no evidence of formation of inter-
mediates other than cyanate ions from the present spe-
ciation study. There is no abstractable hydrogen atom
in CN− to react with the HO• radical and there was
no evidence of HO• radical addition product forma-
tion. These observations suggest that the indirect ho-
mogeneous oxidation of cyanide ions by diffuse HO•
radicals is not a correct description of the oxidation
mechanism. Secondly, if the diffused hydroxyl radical
is the active species oxidising cyanide, the oxidation
of cyanide ion should become more favourable at a
higher pH [25]. This is because as the pH increases,
more HO• radicals could be formed and consequently
ec−b + hv+b → heat/hν
O2 + H2O + 2e− → HO− + HO2
−
cb
hv+b
→
shallow trap ht+r∗
(7)
deep trap
→
+
−
•
≡
≡
TiO + hvb
TiO
(8)
(9)
HO−ads + ht+r → HO•
HO−ads
+
TiO → HO + TiO
(10)
(11)
−
•
•
≡
≡
HO• + CN− → HO•CN
HO•CN H→2O C•ONH2
2C•ONH2 → HCONH2 + HOCN
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
−
+
•
CNads + htr → CNads
CN−ads
+
TiO → CNads
+
−
TiO
•
•
≡
≡
•
2CNads → (CN)2
(CN)2 + 2HO− → OCN− + CN− + H2O
−
−
•
≡
≡
TiO + ecb
→
TiO
•
2HOads → H2O2

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
293
Table 4
the oxidation rate of cyanide would be enhanced. How-
ever, the opposite trend was observed which strongly
support the proposed mechanisms that the photocat-
alytic oxidation of cyanide is a surface catalysed re-
action but not via diffused HO• radicals.
Calculated reaction rate and photonic efficiencies (ζ) obtained for
two type of titanium dioxide samples at 1 g/l and an initial cyanide
concentration of 3.85 mM
Parameter
Initial pH of solution
9.5
10.5
12.0
To further consolidate the present postulation, the
surface hydroxyl groups on titanium dioxide surface
were exchanged with fluoride ions and the sample was
used to diagnose the reaction pathway [26–30]. Re-
cently, Minero et al. [11,12] showed that it is possible
to differentiate whether the photocatalytic oxidation
of phenol proceeds via superficial hydroxyl radical or
via a direct interaction with the hole hv+b by the adsorp-
tion of fluoride ion on the surface of titanium dioxide.
Table 4 and Fig. 7 summarise the results of cyanide
oxidation using F0/TiO2 and F1/TiO2 at different pHs
(9.5, 10.5, and 12.0). It is clear that by incorporating
F− ions onto the surface of TiO2, the photoactivity
(a) F0/TiO2
Rate (M/min) 1.21 × 10−5
1.12 × 10−5
1.07 × 10−5
ζ (%)
0.20
0.19
0.19
(b) F1/TiO2
Rate (M/min) 9.98 × 10−6
9.32 × 10−6
8.48 × 10−6
ζ (%)
0.17
0.16
0.14
of titanium dioxide for cyanide degradation decreases
and the photonic efficiency is reduced by 25.1, 17.9,
and 18.1% at pH 9.5, 10.5, and 12.0, respectively.
≡
After fluoride exchange, the concentration of TiOH
Fig. 7. (a)–(c) The rate of cyanide photooxidation by (᭹) F0/TiO2, and (᭺) F1/TiO2 at different pH. (a) At pH 12.0, (b) at pH 10.5, and
(c) at pH 9.5.

294
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Fig. 8. The pH point zero charge of two different samples. (᭜) F0/TiO2 and (᭹) F1/TiO2.
−
≡
and TiO on the titanium dioxide surface decreased
due to the irreversible displacement of the basic hy-
droxyl groups by fluoride ions according to the fol-
lowing reaction.
3.3. Derivation of reaction kinetics for cyanide
oxidation
The essential reactions concluded from the previous
discussion are summarised in the following kinetic
scheme as illustrated by reactions (R1)–(R13). Since
the homogeneous reaction between CN− ions and free
HO• radicals is unlikely to occur in present system,
only the heterogeneous pathway is considered here.
The concentration of any species is represented with
square brackets “[ ]” in the rate expressions.
−
−
≡
≡
TiF + HO
TiOH + F
→
(21)
As a result, the ability of the photocatalyst to trap
•
≡
the hole as TiO decreases. Although the amount of
hydroxyl groups displaced cannot be shown quantita-
tively, the displacement of surface hydroxyl group on
titanium dioxide is evidenced by the shift of pHPZC
from 6.68 to 6.41 as shown in Fig. 8. The lower-
ing of the concentration of surface hydroxyl group
would induce a higher concentration of untrapped hole
or free HO• radical available for the oxidation of
cyanide ion [11,12]. However, no enhancement was
observed in this study which mea•ns that concentra-
hν
TiO2→TiO2{ec−b · · · hv+b}
→
Trapping ec−b + ht+r
(R1)
√
r1 = k1 I0 = I0.5
ec−b + ht+r → heat/hν
(R2)
(R3)
(R4)
(R5)
(R6)
r2 = k2[e− ][ht+r ] = k2∗[ht+r ]
cb
O2 + H2O + 2e− → HO− + HO2
−
≡
tion of surface trapped hole ( TiO ) is a determining
cb
factor for the photocatalytic reaction. As shown pre-
viously, no homogenous oxidation of cyanide by hy-
droxyl radicals was observed and the photocatalytic
oxidation of cyanide can only proceed via a pure het-
erogeneous route, i.e. cyanide can be oxidised by the
surface trap•ped holes. Although the surface trapped
r3 = k3[O2][e− ]
cb
+
−
•
TiO
+
≡
≡
TiO + hvb
→
+
−
∗
≡
r4 = k4[ TiO ][htr ] = k4[htr ]
−
+
•
HOads + htr → HOads
≡
hole ( TiO ) is known to possess a lower oxidation
r5 = k5[HO− ][ht+r ]
potential than the untrapped holes, it has a longer life-
time which make the oxidation of cyanide ions sta-
tistically more favourable compared to the short-life
untrapped holes.
ads
HO−ads
+
TiO → HOads
+
TiO
−
•
•
≡
≡
r6 = k6[HO−ads][ TiO ]
•
≡

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
295
−
+
•
on the concentration of ht+r only [33,34]. Reactions
(R3) and (R4) represent the trapping of photogener-
ated electrons and holes by adsorbed molecular oxy-
gen and surface or adsorbed hydroxyl groups. The rate
constant of reaction (R3), k3, in both air and oxygen
saturated titanium dioxide suspension was determined
to be 7.6 × 107 l/mol s and the total concentration of
hydroxyl groups in Degussa P25 TiO2 has been shown
to be 4.7 × 10−4 mol/g [27,35]. Since these concen-
trations rem−ain constant on the surface of TiO2, the
CNads + htr → CNads
(R7)
(R8)
r7 = k7[CN− ][ht+r ]
CN−ads
r8 = k8[CN−ads][ TiO ]
ads
−
•
•
≡
≡
+
TiO → CNads
+
TiO
•
≡
−
−
•
≡
≡
TiO + ecb
r9 = k9[ TiO ][ecb] = k9[ TiO ]
→
TiO
(R9)
−
•
∗
•
≡
≡
•
2CNads → (CN)2
r10 = k10[CNads
(R10)
(R11)
(R12)
(R13)
≡
• 2
term [ TiO ] can be substituted into k4 and the rate
]
expression can be simplified by substituting with a
second rate constant k4∗.
(CN)2 + 2HO− → CNO− + CN− + H2O
During irradiation, the potential electron donors are
r11 = k11[(CN)2][HO−]2
adsorbed HO−+and CN− ions and they would inter-
•
•
≡
act with the htr and TiO to different extents. As
2HOads → H2O2
r12 = k12[HOads
discussed in last section, reactions (R5)–(R13) are the
main photocatalytic reactions occurring on the surface
of titanium dioxide proposed for this study. In reac-
tion (R5)–(R8), k5, k6, k7, and k8 are rate constants
for the surface oxidation reactions between adsorbed
• 2
]
2H2O2 → 2H2O + O2
r13 = k13[H2O2]2
+
HO− ions with htr , adsorbed HO− ions with TiO ,
•
≡
Reaction (R1) represents the photogeneration
of electron-hole pair and the rate constant for
electron-hole pair generation is k1 (I0 is the incident
photon flux and I is the rate of photon absorption).
Since the UVA lamp (66.7 mW/cm2 as measured
by a UV radiometer) used in the experiments has a
photon output higher than the low intensity region
(25 mW/cm2) as defined in the literature, the genera-
tion rate is assumed to have a square root dependency
of light intensity [31,32]. The photoge+nerated holes
adsorbed CN− ions with ht+r , and adsorbed CN−• ions
•
≡
≡
with TiO , respectively. The un-reacted TiO will
recombine with ec−b and return to the ground state as
shown in reaction (R9). The rate of this reaction de-
•
≡
pends on the concentration of TiO on the surface
of titanium dioxide an•d is always a pseudo-first-order
≡
with respect to TiO as shown in the rate expres-
sion. The reaction of CN• radicals and the hydrolysis
of (CN)2 molecules are shown in reactions (R10) and
(R11). The HO• radicals are also consumed via the
formation of H2O2 which decomposes rapidly in the
presence of TiO2 and UV irradiation, to give H2O and
O2 as shown in re−action (R12) and (R13).
+
•
≡
hvb will be instantaneously trapped as htr and TiO ,
if not recombined. This process occurs so rapidly in
colloidal TiO2 particles that it does not contribute to
the overall kinetics of the charge transfer events and
the term trapped hole ht+r is used in other rate ex-
pressions instead of hv+b to indicate the actual species
involved in cyanide photooxidation process.
≡
Since the TiO groups are the sites for cyanide ox-
−
≡
idation, charge transfer will occur at TiO . The ini-
tial pH of the system and zeta potential measurement
indicate that HO− ions are adsorbed and concentrated
near the surface of the titanium dioxide photocatalyst.
Since CN− ions are not well adsorbed on the surface of
titanium dioxide, the adsorbed HO− ions will compete
with cyanide ions for the trapped and untrapped holes.
This is supported by the findings of Bahnemann et
al. [35] who reported that substrate molecules, which
are weakly adsorbed on the surface of titanium diox-
ide, are generally less efficient to react with the holes
because the life time of the holes are very short (on
Reaction (R2) represents the recombination reaction
of the photogenerated charges with rate constant k2. In
this study, since the solution was continuo−usly sparged
with air, the conduction band electrons ecb were effi-
ciently scavenged by+molecular oxygen. Therefore, the
valence band holes htr would have slightly longer life-
time compared to the ec−b. Under these circumstances,
there would be no significant build-up of electron con-
centration within the TiO2 photocatalyst and the rate
of recombination reaction (R2) would be dependent

296
K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
Fig. 9. A plot of the reciprocal of initial cyanide photooxidation rate (1/Ri) vs. the reciprocal of initial cyanide concentration (1/Ci) at pH 12.
nanosecond range). Due to the low concentration of
CN− ions adsorbed on TiO2 surface, it is reasonable
to assume that the rate of cyanide oxidation by holes
is much lower than the oxidation of HO− by holes.
Since O2 is in surplus, the overall rate of cyanide oxi-
dation would be reaction R8. Summing up, the overall
cyanide photooxidation rate (r) can be written as:
line. Such a plot can be found in Fig. 9 with a cor-
relation coefficient R2 = 0.995 which indicates that
there is a good agreement between the prediction of
equation and present experimental results. Although
Eq. (23) resembles to a conventional reciprocal plot of
a Langmuir–Hinshelwood type reaction, it is believed
that the parameters in equation are the integrals of the
kinetic parameters for all the fundamental reactions
(reactions (R1)–(R13)) in present CN−/TiO2 system.
This behaviour was also reported by Turchi and Ollis
[34] on the photocatalytic oxidation of many organic
compounds.
r = r8 = k8[CN−ads][ TiO ]
(22)
•
≡
By using the relationships of r5 ꢀ r7, [HO−ads] >
[CN−ads], and under steady state concentration of
•
≡
TiO , the following rate expression for the photo-
catalytic oxidation of cyanide ion is obtained:
ꢀ
ꢁ ꢀ
ꢁ
k6[HO−ads] + k9∗
k2∗ + k4∗ + k5[HOa−ds
]
4. Conclusion
1
=
r
k8
k4∗I0.5
The kinetics and oxidation mechanism of cyanide
in the presence of UV and titanium dioxide catalyst
was studied. Cyanide ions were found to be oxi-
dised photocatalytically to cyanate on the surface of
titanium dioxide under UV illumination. The photo-
catalytic oxidation rate of cyanide is dependent on
catalyst concentration, air concentration, and initial
pH. By displacing the surface hydroxyl groups on the
surface of titanium dioxide with fluoride ions, it was
confirmed that cyanide is oxidised via a pure het-
erogeneous pathway at the surface hydroxyl groups.
This was explained in terms of the low adsorption of
cyanide on titanium dioxide and the longer lifetime
of the trapped holes. The photonic efficiency of the
k2∗ + k4∗ + k5[HOa−ds
]
1−
×
+
]
(23)
k4∗I0.5
[CNads
It can be seen from Eq. (23) that the rate of cyanide
oxidation is dependent on the CN− ion adsorption on
the surface of titanium dioxide. As evidenced by the
results from the dark adsorption experiment, cyanide
ions are not well adsorbed by titanium dioxide and
therefore the mass transfer process plays a significant
role in the photocatalytic oxidation of cyanide ions. In
accordance to Eq. (23), a plot of the reciprocal of ini-
tial cyanide photooxidation rate versus the reciprocal
of initial cyanide concentration should be a straight

K. Chiang et al. / Journal of Molecular Catalysis A: Chemical 193 (2003) 285–297
[14] H. Kawaguchi, Environ. Technol. 15 (1994) 183.
297
process remains very low and is generally around
0.1–0.2%. Such a low value of photonic efficiency is
due to of the lack of interaction of cyanide with the
surface of titanium dioxide. In addition, cyanide ions
showed low reactivity towards diffused hydroxyl rad-
icals, the majority of the photogenerated holes were
not be efficiently scavenged and became recombined.
A kinetic expression was developed to model the
photocatalytic oxidation reaction.
[15] U. Stafford, K.A. Gray, P.V. Kamat, J. Catal. 167 (1997)
25.
[16] N. Serpone, R. Terzian, D. Lawless, P. Kennepohl, G. Sauve,
J. Photochem. Photobiol. A. Chem. 73 (1993) 11.
[17] I. Arslan, I.A. Balcioglu, D.W. Bahnemann, Appl. Catal. B.
Environ. 26 (2000) 193.
[18] M. Schiavello, V. Augugliaro, L. Palmisano, J. Catal. 127
(1991) 332.
[19] I. Rosenberg, J.R. Brock, A. Heller, J. Phys. Chem. 96 (1992)
3423.
[20] K. Kogo, H. Yoneyama, H. Tamura, J. Phys. Chem. 84 (1980)
1705.
[21] D. Behar, J. Phys. Chem. 78 (6) (1974) 2660.
[22] B.H.J. Bielski, A.O. Allen, J. Am. Chem. Soc. 99 (18) (1977)
5931.
[23] H. Buchler, R.E. Buhler, R. Copper, J. Phys. Chem. 80 (14)
(1976) 1549.
[24] F. Munoz, M.N. Schuchmann, G. Olbrich, V. von Sonntag,
J. Chem. Soc., Perkin. Trans. 2 (2000) 655.
[25] C. Galindo, P. Jacques, A. Kalt, J. Photochem. Photobiol. A.
Chem. 130 (2000) 35.
References
[1] S.N. Frank, A.J. Bard, J. Am. Chem. Soc. 99 (1) (1977) 303.
[2] H. Hidaka, T. Nakamura, A. Ishizaka, M. Tsuchiya, J. Zhao,
J. Photochem. Photobiol. A. Chem. 66 (1992) 367.
[3] B.V. Mihaylov, J.L. Hendrix, J.H. Nelson, J. Photochem.
Photobiol. A: Chem. 72 (1993) 173–177.
[4] A. Bravo, J. Garcia, X. Domenech, J. Peral, Electrochimica.
Acta 39 (6) (1994) 2461.
[26] H.P. Boehm, Discuss. Fara. Soc. 52 (1971) 264.
[27] J.A. Rob van Veen, F.T.G. Veltmaat, G. Jonkers, J. Chem.
Soc., Chem. Commun. (1985) 1656.
[28] K. Kobayakawa, Y. Nakazawa, M. Ikeda, Y. Sato, A.
Fujishima, Ber. Bunsenges. Phys. Chem. 94 (1990) 1439.
[29] C. Kormann, D.W. Bahnemann, M.R. Hoffmann, Environ.
Sci. Technol. 25 (1991) 494.
[30] A. Hattori, M. Yamamoto, H. Tada, S. Ito, Chem. Lett. (1998)
707.
[31] J.S. Curran, J. Domenech, N. Jaffrezic-Renault, R. Philippe,
J. Phys. Chem. 89 (1985) 957.
[32] J.M. Herrmann, Catal. Today 53 (1999) 115.
[33] C. Minero, Catal. Today 54 (1999) 205.
[34] C.S. Turchi, D.F. Ollis, J. Catal. 122 (1990) 178.
[35] D.W. Bahnemann, M. Higendorff, R. Memming, J. Phys.
Chem. 101 (1997) 4265.
[5] V. Augugliaro, V. Loddo, G. Marci, L. Palmisano, M.J.
Lopez-Munoz, J. Catal. 166 (1997) 272.
[6] T.L. Rose, C. Nanjundiah, J. Phys. Chem. 89 (1985) 3766.
[7] J. Peral, J. Munoz, X. Domenech, J. Photochem. Photobiol.
A. Chem. 55 (1990) 251.
[8] N. Serpone, E. Borgarello, M. Barbeni, E. Pelizzetti, P. Pichat,
J.M. Hermann, M.A. Fox, J. Photochem. 36 (1987) 373.
[9] M.S. Ahmed, Y.A. Aitta, J. Non-Cryst. Solid 186 (1995) 402.
[10] C.A. Young, in: B. Mishra (Ed.), EPD Congress, The
Minerals, Metals and Materials Society, 1998, p. 877.
[11] C. Minero, G. Mariella, V. Maurino, E. Pelizzetti, Langmuir
16 (2000) 2632.
[12] C. Minero, G. Mariella, V. Maurino, D. Vione, E. Pelizzetti,
Langmuir 16 (23) (2000) 8964.
[13] A. Mills, S. Morris, R. Davies, J. Photochem. Photobiol. A.
Chem. 70 (1993) 183.