Photo_ch_15.doc
3-24-05
CHAPTER 15
PHOTOCHEMICAL REACTIONS
1.
Introduction
Photochemical reactions involve reactions induced by action of light. These types of
reactions find many applications in environmental problems and purification of waste
streams. The objective of this chapter is to show some important application, discuss the
theory and show design examples.
1.2
Application Examples
1.2.1
1.2.2
Photo catalytic-oxidation of organic contaminants industrial wastewater
In many cases, a solid catalyst such as TiO2 is also used. The catalyst
enhances the light absorption.
Photo-ozonation of contaminated water
This process has been applied, for example, in the oxidation of cyanides in
wastewater and phenolic compounds. In this process, ozone is photolyzed by
the action of UV radiation in aqueous solution.
This generates hydroxyl radicals that are the major oxidant which then
oxidizes the toxic wastes to innocuous products such as CO2, carboxylic acids,
etc. The schematic of the light absorption processes leading to the formation
of the hydroxyl radicals can be represented as:
hv
O3 + H 2 O ⎯⎯→
H 2 O2 + O2
hv
H 2 O2 ⎯⎯→
2OH *
The hydroxyl radical is one of the most powerful oxidizing species. The
kinetics of oxidation with hydroxyl radical is faster by several order of
magnitude than by use of pure O3 or H2O2. The rate constants are of the order
of 108 to 1010 L/mol.s. Hence photo-oxidation is much faster conventional
oxidation. This is one of the reasons for the growing application of this
process in waste treatment.
1.2.3 Photo-oxidation with H2O2
H2O2 can be activated by initiation reactions induced by UV light. The
maximum absorbance for H2O2 occurs at a wavelength of about 220 nm. The
resulting OH* radicals then participate in oxidizing reactions.
The reaction rate constant of organic molecules with OH radical is an
indicator of effectiveness of a photochemical oxidation. The rate constants of
various organic compounds can be predicted by group contribution methods
as discussed earlier. In general, alkenes have a higher rate constant than
alkanes and hence a higher rate of photo-degradation.
1.2.4 Photo-Fenton Process
The Fenton reagent is highly suitable for oxidation via generation of OH*
radical. The mechanism can be simplified as:
Fe +2 + H 2 O2 → Fe +3 + OH − + OH *
+
−H
Fe +3 + H 2 O2 ←⎯
⎯→ Fe − O2 H + 2 ⇔ Fe + + HO2*
Fe +3 + HO2* ⇔ Fe + + + H + + O2
The combination of Fenton reaction in UV is called the photo-Fenton process
and has been shown to enhance the efficiency of the Fenton process. The
enhancement due to action of light has been demonstrated in practice but the
exact mechanism is not well established.
2
Review of Photon Absorption
In this section, we review the fundamental laws in photochemistry that are of
importance in understanding the kinetics and design of photochemical reactors.
Grottus-Draper Law: Only light absorbed by a system can cause chemical
transformations.
Starke-Einstein Law: Only one quantum of light is absorbed per molecule of absorbing
and reacting species that disappear
Lamberts Law:
Fraction of incident radiation absorbed by a transparent medium is
independent of the intensity of the incident light. Thus, successive layers in the medium
absorb the same fraction of the incident light.
Beer Law:
Absorption of incident light is directly proportional to the
concentration of the absorbing species.
By combining the above two laws, we can express the intensity of radiation as a function
of distance as:
dI
= −α v CI
dz
where α v = proportionality constant, the subscript v indicating the frequency of the
incident light and C = concentration of absorbing species.
Assuming an intensity I0 at z = 0, (for example at the source or lamp) the integrated form
of the Beer-Lambert law can be expressed as
I = I 0 exp(− α v Cz )
Amount absorbed by a medium of thickness L is
I 0 − I = I 0 [1 − exp(− α v CL )]
If several components are absorbing simultaneously then the equation should be modified
as:
[
]
Amount absorbed = I 0 1 − exp(− ∑α v , j C j L )
The units are W/m2 here.
The efficiency of a photochemical reaction was defined by Einstein in terms of the
quantum efficiency, φ.
φ=
Number of molecules decomposed
Number of quanta absorbed
The concept of quantum efficiency follows from the Stark-Einstein law. Many species
may absorb light and get “excited” by the photons. Not all of them participate in the
chemical reaction. Some of them may lose energy by non-chemical pathways. The
quantum efficiency keeps a track of such processes. To see the implication of this,
consider a simple reaction:
A→B
The mechanistic sequence for photo-catalytic reaction can be proposed as follows:
hv
A ⎯⎯→
A*
A* → Pr oducts
k1
k2
where A* is the photon activated complex. But not all A* formed goes to products
yielding quantum efficiencies less than one. For example, the following processes may
also take place:
Fluorescence:
Collision with neutral species
A* → A + hv
A* + M → A + M
k3
k4
Rate of formation of products = k 2 A*
Net rate of formation of A* = k1 A* I − k 2 A* − k 3 A* − k 4 A* M
The species A* is at quasi-steady state and its rate of formation is zero. Hence,
A* =
k1 AI
k 2 + k3 + k 4 M
Equation for the rate of photoreaction can then be written as:
Rate =
k 2 k1 AI
k 2 + k3 + k 4 M
(1)
This is expressed in terms of a photon yield as:
Rate = k1φIA
where φ is the quantum efficiency.
φ=
k2
k 2 + k3 + k 4 M
for the postulated set of mechanism.
The rate of reaction is then simplified as:
rate = k1 AIφ
Hence φ is a parameter which has to be known to design and scale-up the reactor.
Example: Mechanism of ozone photolysis
Reactions leading to ozone photolysis can be represented as follows:
Note that mass transfer of ozone needs to be included separately in the reactor model
along the kinetics of decomposition of the pollutants by the OH radical.
Reactor for UV/H2O2 system
A schematic flow diagram of a typical Calgon UV/H2O2 system is shown below. It has
an oxidation unit with six photo reactors in series with 15-kilowatt UV lamp in each
reactor and holds a total volume of 55 liters. Each UV lamp is mounted inside a quartz
tube in the center of each reactor such that water flows around the quartz tube. The
system is shown below.
Note the split flow arrangement for H2O2 and a flow through pattern for the waste water.
Reactor for UV/H2O2 /O3 system
A reactor diagram for UV/H2O2 /O3 is shown in the next diagram. The UV reactor is
divided by five vertical baffles into six chambers and contains 24 low-pressure mercury
vapor lamps in quartz sleeves. The UV lamps are installed vertically and are evenly
distributed throughout the reactor. Each chamber also has one stainless steel sparger that
extends along the width of the reactor. The sparger uniformly diffuses ozone gas from
base of the reactor into contaminated water. Hydrogen peroxide is introduced in the
reactor from the feed tank. An in-line static mixer is used to disperse the hydrogen
peroxide into the contaminated water in the influent feed. Typical reactor arrangement is
shown below.
Reactor for UV/O3/H2O2 system
Photocatalytic Systems: Kinetic Models
Photocatalytic oxidation (PCO) has been demonstrated to be a highly effective technique
for treating water contaminants with trace amounts of refractory organic pollutants.
These are difficult to remove by conventional methods.
Photocatalytic oxidation occurs as a result of the combined action of three components, a
semi-conductor photocatalyst such as TiO2, sufficiently energetic photon sources (near
UV radiation) and an oxidizing agent (O2). TiO2 is a semi-conductor which has a valance
band and a conduction band. There is a band-gap that separates the two and no electrons
can be found here. Upon the action of UV or visible light, an electron is moved from the
valence band to the conduction band as shown in the figure. A hole is left behind in the
valence band. This hole can be scanvenged by organic molecules that are thereby
oxidized.
Figure 1. Valence and conduction bands in a semiconductor
Conduction band
Eg
Ec
e-
band gap
Ev
h+
Valence band
The mechanistic description of PCO can then be postulated as follows:
Electron-hole pair generation:
(
hv
TiO2 ⎯⎯→
TiO2 e − + h +
)
Electron removal from adsorbed organic:
x
x
h + + Rads
→ Rads
+
Electron removal from solvent (H2O)
h + + H 2 O → H + + OH *
leading to the hydroxide radical which can then participate in the liquid phase oxidation.
Electron Acceptor Reaction:
Electron transfer to other species (inherently present or deliberately added such as H2O2
to the solution) also takes place.
H 2 O2 + e − → OH − + OH *
O2 + e − → O2−
A more detailed mechanistic model has been proposed by Ollis and Turchi.
Some promising applications of PCO are shown in the following table.
Table 1. Photodegradation of organics by TiO2/UV Processes (Legrini et al)
1. Phenol, TOC reduced by 35% in 90 minutes.
2. Acetic, benzoic, formic acids, ethanol, methanol, TOC reduced by >96% in 10
minutes.
3. Degradation of aniline, salicyclic acid and ethanol.
4. 1,2-dimethyl-3-nitrobenzene and nitro-o-xylene from industrial wastewater, >95%
TOC removed in 50 minutes.
5. Degradation of pentachlorophenol.
6. Degradation of other organic compounds: Trichloroethylene, chlorobenzenes,
nitrobenzene, chlorophenols, phenols, benzene and chloroform.
7. Degradation of a pesticide, atrazine
Some advantages of PCO are as follows:
(a) There is complete mineralization of many organic pollutants to
environmentally benign effluents such as CO2, H2O and mineral acids.
(b) No need for expansive oxidizing agents such as O3 and H2O2. Air is often
sufficient.
(c) TiO2 is inexpensive, nonhazardous, stable and reusable.
(d) Light required to activate the catalyst is low energy UV.
(e) It is possible in some cases to use solar energy. However, solar radiation
may not be the best choice for UV/TiO2 because only a small portion of
the total solar spectrum is in the 300-450 nm range.
Large scale applications is lacking due to the fact that the scale-up of these reactors is a
difficult problem. Hence, modeling the system in a detailed manner is an important
research area. In the following sections, we discuss the basic model for photocatalytic
systems.
It may be noted that only those photons which have energy equal to greater than the band
gap will be absorbed and only those will contribute to the electron hole pair formation.
Example: Find the wavelength range of light absorption for rutile which has a band gap
of 3.0 eV.
Solution: hv is the energy for light absorption where v is the frequency. Hence
hv ≥ 3.0 eV
h = Planck’s constant = 6.6256 x 10-34 Js
Also 1eV is the energy required to accelerate 1 electron of charge by 1 volt.
1 electron has a charge of 1.6 x 10-19 C
Hence 1eV = 1.6 x 10-19 J
Solving for v, we have v = 3 x 1.6 x 10-19 / (6.6256 x 10-34) = 7.2543 x 1014 s-1
To find the wavelength we use the relationship, C = vλ where C is velocity of light =
3x108 m/s.
Hence, λ ≤
3 × 10 8
= 4.14 x 10-7 m = 414 nm
14
7.25 × 10
λ < 414 nm for light absorption. Visible light is 414 to 700 nm range and the light of
interest is in the ultra violet range.
Parametric Effects
Qualitative effect of various parameters on the rate of photocatalytic reaction are
discussed below.
(i)
(ii)
(iii)
(iv)
(v)
(vi)
Solute concentration
Reaction rate depends on a power law manner in solute concentration. At low
concentrations, it is first order while at high concentrations, a zero order effect
is observed.
Catalyst loading
Reaction rate increases with loading, then reaches a maximum at some
optimum loading and then decreases. The decrease is often attributed to
obstruction of light transmission (the shielding effect)
Light intensity
Rate is proportional to I β where β is between 0.5 and 1.0.
pH
pH has a direct effect and the best range of pH is between 5.6 and 6.4. The
effect is attributed to change in surface properties of TiO2 with pH.
Partial pressure of oxygen
Rate increases with increase in oxygen partial pressure.
Temperature
Temperature does not have a significant effect. This may be due to decrease
in adsorption coefficient of the solute on the catalyst with an increase in
temperature.
Model Equations for PCO
A general kinetic equation can be used to describe PCO reactions:
R = k e c np
(1)
where k e is an effective rate constant and cp is the pollutant concentration. k e turns out
to be a function of the catalyst concentration (ccat) and the local volumetric rate of
absorption ( e&λm1 ) of those components of the incident radiation whose wavelength ( λ1 ) is
effective in activating the reaction.
These local volumetric absorption rates (LVAR) are related to radiative transfer within
the medium and are key parameters in the derivation of a reliable kinetic equation.
For any effective wavelength λ, the local volumetric rate of radiant energy absorption can
be evaluated at each point within an isotropic medium thus:
e&λm1 (x ) = k λ (x ) ∫ I λw (x )dw
(2)
4π
Eq(2) is schematically illustrated in the following diagram. Also, kλ is the local value of
absorption coefficient for the medium.
Diagram to represent LVAR
θ
x
φ
Radiation leaving
I λw ( x )dw
Spherical coordinate system
solid angle
dw =
=
sin θ dθ dφ
Intensity is defined as energy per unit time per solid angle and unit area. The light
intensity is described by a “radiation transport equation” (RTE) which turns out to be a
complex integro-differential equation. The equation can be solved analytically only for
idealized one dimensional situation. The basic equation is as follows:
dI λw
σ
= −(k λ + σ λ )I λw + λ
ds
4π
∫π I λ
w
(
)
p λ w ' → w dw '
(7)
4
where k λ and σ λ are the adsorption and single-scattering coefficients respectively (the
absorption and scattering cross section per unit volume) and p λ (w ' → w) is the phase
function (which accounts for elastic scattering from the direction w’ to the direction w).
The first term on the RHS represents the loss of photons due to absorption, the loss due to
out-scattering is accounted for by the second term. The last term (integrated term)
accounts for the gain in the radiation due to in-scattering.
The incident intensity at any point from all directions is given as
4π
Gλδ =
∫ Iλ
w
dw
0
and which is the basis Eq. 2 for LVAR. The above expression for G multipled by k
gives LVAR.
The relevance of k λ and σ λ to radiative transfer within a given system may be more
readily understood if they are compared with the relevant characteristic geometric length
of the system, I0. Since k λ and σ λ can be interpreted as the reciprocal of a λ-photon
mean free path before an absorption or scattering event, the following dimensionless
parameters result, respectively:
mλ = (k λ + σ λ )I 0
Cλ =
σλ
kλ + σ λ
the optical thickness
the single scattering albedo
The former compares the dimension of the system with the mean free path of a λ-photon
before an interaction event (irrespective of the type of event), while the latter provides the
probability for an interaction event to result in scattering. No recasting is required for the
phase function pλ since it is already dimensionless: it determines the probability of a λphoton being scattered from the direction w’ to the direction w.
Absorption and scattering coefficients
Values of κ and σ for titania catalyst depend on the wavelength of the light. Therefore,
the RTE should be solved for each of the individual wavelength ranges. However, this
will make the contributions memory-intensive. Consequently, wavelength-averaged
values of the absorption and scattering coefficients (Romero et al 1997) are used:
σ = 3.598 x Wcat
κ = 0.2758 x Wcat
where σ and κ are the wavelength-averaged scattering and absorption coefficients (m-1)
and Wcat is the catalyst loading (gm-3).
Phase Function Parameter
For the isotropic scattering, the phase function parameter has a value of unity. However,
for anisotropic scattering, a number of expressions have been reported and each one of
them is suitable for specific systems. The widely used phase function is of linear
anisotropic form (Fiveland 1984)
p(θ ) = 1 + a 0 cos(θ )
with a 0 =1,0, -1 for forward, isotropic and backward scattering respectively. In terms of
direction cosines, the phase function may be expressed as:
(
)
(
p w m → w l = 1 + a 0 µ l µ m + ξ l ξ m + η lη m
)
where µl,ξl, ηl are the direction cosines of lth direction and µm,ξm, ηm are the direction
cosines of mth direction.
Assumptions involved in Eq(7) are as follows:
1) Homogeneous medium
Since TiO2 concentration is small and particle diameter is much smaller than the
reactor length this is a reasonable assumption.
2) No emission by the medium
Since the PCO process is carried out at near room temperature, this is a valid
assumption.
3) Elastic and independent scattering
i.e., scattering changes only the direction of the light and not its energy and each
particle scatters radiation as though it is isolated.
4) Diffuse radiation
This means I ≠ I(θ). No preferred directionality for the radiation.
Model equations can be solved by Monte-Carlo techniques. Here the fate of statistically
meaningful number of photon bundles emitted by the lamp is followed. The results are
averaged to find LVEA.
Some criterion for reactor design are as follows:
i)
Order of magnitude of the optical thickness should be close to unity.
ii)
For a given value of the absorption coefficient, the catalyst with the lowest
albedo should be selected. This parameter is close to 0.689 for rutile TiO2.
iii)
Other parameters such as mixing and O2 mass transfer play their usual role.
Type of reactors for PCO
Fixed bed (packed) reactors
Catalyst coated tubes
Catalyst coated extremely narrow diameter immersion type lamps
Catalyst coated rotating tube bundles
Taylor vortex photo-reactor (two co-axial cylinders with co- or counter-rotation)
This reactor has been investigated by Butta and Ray and provides a well
characterized flow field. Hence it is useful for kinetic studies.
Matrix UV/TiO2 system
The Matrix system uses UV light with its predominant emission at a wavelength
of 254 nm, TiO2 semiconductor (anatase form), and oxidants to generate OH*. A typical
system contains many photo catalytic reactor cells; the exact number of cells varies
depending on the application. A 75W, 254 nm UV light source is located coaxially
within long quartz sleeve (1.6 m), which is in turn is surrounded by eight layers of
fiberglass mesh bonded with TiO2 and is enclosed in a stainless steel jacket. Each cell is
rated for a maximum flow rate of about 0.8 L/min.
A typical Matrix system consists of two units positioned side by side in a mobile
trailer. Each unit consists of 12 wafers, and each wafer consists of six photo catalytic
reactor cells joined by manifolds. Contaminated water is sent into three reactor cells at a
time. Each set of three cells along the path where the contaminated water flows is
defined as a path length. Therefore, a wafer has two path lengths and each unit has 24
path lengths, resulting in a total of 48 path lengths for the two units. Hydrogen peroxide
and ozone are injected at multiple path lengths throughout the system. The exact number
of injection points varies depending on the application.
The contaminated water enters first wafer, path length 1 (set of three reactor cells
in unit 1) and then path length 2 (the second set of three reactor cell in unit 1). After the
treatment is completed in the first wafer, contaminated water flows to the second wafer
and enters path length 3 (set of three reactor cell in unit 2) and the path length 4 (second
set of three reactor cell in unit 2). This process continues until the contaminated water
has passed through all 24 wafers (48 path lengths). The treated water exiting path length
42 is disposed of appropriately.
References:
D.F. Ollis and C. Turchi, “Heterogeneous photocatalysis for water-purificationcontaminant mineralization kinetics and elementary reactor analysis”, Env. Progress 9 (4)
229-234 1990
O. Legrini, E. Oliveros, AM. Braun, “Photochemical Processes for Water Treatment”,
Chemical Reviews, 93 (2) 671-698 1993
P.K. Dutta and A.K. Ray, “Experimental investigation of Taylor vortex photocatalytic
reactor for water purification”, Chem. Eng. Sci. 59, 5249-5259 2004
H. Feng, L. Lechang, “Degradation kinetics and mechanisms of phenol in photo-Fenton
process”, J. of Shejtang University, 5(2), 198-205, 2004