Visible-Light-Responsive Photocatalysis: Ag-Doped TiO2 ...

46th International Conference on Environmental Systems ICES-2016-169 10-14 July 2016, Vienna, Austria

Visible-Light-Responsive Photocatalysis: Ag-Doped TiO2

Catalyst Development and Reactor Design Testing

Janelle L. Coutts1,

Engineering Services Contract, Kennedy Space Center, FL, 32899

Paul E. Hintze2, Anne Meier3, Malay G. Shah4

NASA – Kennedy Space Center, FL, 32899

Jan M. Surma5, Robert W. DeVor6, Phillip R. Maloney7, Brint M. Bauer8,

Engineering Services Contract, Kennedy Space Center, FL, 32899

and

David W. Mazyck9

University of Florida, Gainesville, FL 32611

In recent years, the alteration of titanium dioxide to become visible-light-responsive (VLR)

has been a major focus in the field of photocatalysis. Currently, bare titanium dioxide requires

ultraviolet light for activation due to its inherent band gap energy of 3.2 eV. Mercury (Hg)-

vapor fluorescent light sources are used in photocatalytic oxidation (PCO) reactors to provide

adequate levels of ultraviolet light for catalyst activation; these Hg-containing lamps, however,

hinder the use of this PCO technology in a spaceflight environment due to concerns over crew

Hg exposure. VLR-TiO2 would allow for use of ambient visible solar radiation or highly

efficient visible wavelength LEDs, both of which would make PCO approaches more efficient,

flexible, economical, and safe. Over the past three years, Kennedy Space Center has developed

a VLR silver (Ag)-doped TiO2 catalyst with a band gap of 2.72 eV and promising

photocatalytic activity. Catalyst immobilization techniques, including incorporation of the

catalyst into a sorbent material, were examined. Extensive modeling of a reactor test bed

mimicking air duct work with throughput similar to that seen on the International Space

Station was completed to determine optimal reactor design. A bench-scale reactor with the

novel catalyst and high-efficiency blue LEDs was challenged with ethanol, a common volatile

organic compound (VOC) found in ISS cabin air, to evaluate the system’s ability to perform

high-throughput trace contaminant removal. The ultimate goal for this testing was to

determine if the unit would be useful in pre-heat exchanger operations to lessen condensed

VOCs in recovered water thus lowering the burden of VOC removal for water purification

systems.

1 Scientist IV, ESC - Chemical & Biological Sciences, Mail Stop ESC-24, Kennedy Space Center, FL 32899. 2 Chemist, NASA Applied Science Branch, UB-R3, Kennedy Space Center, FL 32899. 3 Chemical Engineer, NASA Materials Analysis Branch, NE-L3, Kennedy Space Center, FL 32899. 4 Pathways Intern, NASA-KSC, Kennedy Space Center, FL 32899. 5 Scientist IV, ESC - Chemical & Biological Sciences, Mail Stop ESC-5, Kennedy Space Center, FL 32899. 6 Scientist V, ESC - Chemical & Biological Sciences, Mail Stop ESC-55, Kennedy Space Center, FL 32899. 7 Scientist IV, ESC - Chemical & Biological Sciences, Mail Stop ESC-5, Kennedy Space Center, FL 32899. 8 Mechanical Technician, ESC - Chemical & Biological Sciences, Mail Stop ESC-55, Kennedy Space Center, FL

32899. 9 Professor of Environmental Engineering, University of Florida, Gainesville, FL 32611.

International Conference on Environmental Systems

2

Nomenclature

ACD = acetaldehyde

CCT = Correlated Color Temperature

CFD = Computational Fluid Dynamic

cfm = cubic feet per minute

d.f. = film depth

DRS = diffuse reflectance spectroscopy

eV = electron volts

Eg = band bap energy

GC-FID = gas chromatograph-flame ionization detector

ISS = International Space Station

KSC = Kennedy Space Center

LED = light emitting diode

LPM = liters per minute

MA = mechanical alloying

PCO = photocatalytic oxidation

PD = photodeposition

RANS = Raynolds-Averaged Navier Stokes

RH = relative humidity

rpm = revolutions per minute

STC = silica-titania composite

TCC = trace contaminant control

UCF = University of Central Florida

UF = University of Florida

UV = ultraviolet

VLR = visible-light-responsive

VOC = volatile organic compound

W = Watt

I. Introduction

ITANIUM dioxide has been one of the most utilized photocatalysts dominating the field of photocatalytic

oxidation (PCO) for decades. Applications for this technology encompass a broad range of processes such as

hydrogen production via photocatalytic water splitting, chemical and biological purification of water and air, and

synthesis of organic compounds. A commercial TiO2, Degussa P25, consisting of a mixture of anatase (70-85%),

rutile, and amorphous (minor) titania1, has demonstrated high PCO activity in numerous studies2-7. The anatase phase

is known for its superior photocatalytic activity relative to the rutile phase8 with a band gap energy of 3.2 eV; the

Degussa P25 mixture has also demonstrated higher photocatalytic activity over pure anatase or rutile phase TiO28.

Based on the anatase band gap, photons with a wavelength of 388 nm or less (i.e., UV) are required to activate anatase

TiO29. Traditionally, mercury (Hg)-vapor light sources are used in PCO systems, but the presence of mercury within

the bulbs precludes their use in crewed spaceflight environments due to the possibility of contamination (i.e., bulb

breaks). Ultraviolet LEDs are an emerging technology, and have proven to be a feasible excitation light source

alternative. However, a survey of commercially-available UV-LEDs shows that they possess low lighting efficiency

(13%)6.

The UV-range photonic energy requirements for bare-TiO2 photocatalysis limits the use of traditional indoor

lighting or solar radiation, which only consists of ~4-6% UV radiation at the Earth’s surface10, as conceivable energy

sources. Furthermore, UV-activated photocatalysis using TiO2 provides only a moderate reaction rate and somewhat

low quantum yield due to a low electron transfer rate and high electron-hole recombination rate11. Thus, in its current

state of development, TiO2-assisted photocatalysis is not feasible for high throughput processes. A solution for

overcoming such limitations is to enable the TiO2 photocatalyst to be activated by visible light to allow for the use of

highly efficient blue or white LEDs or visible band solar radiation (~45% of the solar spectrum lies in the visible

region10). The development of Visible-Light-Responsive (VLR) PCO catalysts would facilitate the acceptance and

incorporation of PCO-based technology for numerous ISS applications including use in air trace contaminant control

(TCC) water recovery systems, low-cost hydrogen production using solar energy12, enhanced chemical and microbial

purification of water13, 14, and potentially in the field of artificial photosynthesis. The past decade has witnessed a shift

in catalyst development towards visible light responsiveness. Efforts have included sensitization of TiO2 with

T


3

absorbed dye molecules15, 16, metal cation17-21 or anion22-

24 doping, and narrow band gap semiconductor and TiO2

coupling25-27 to name a few. Despite the vast amount of

research in this field, conclusions regarding the relative

effectiveness of any given method are difficult to draw

due to the lack of consistent or standardized

experimental conditions.

In the past several years, a team from Kennedy Space

Center (KSC) have and the University of Central Florida

(UCF) developed a VLR catalyst library focused on

coupling narrow band gap semiconductors with TiO2 via

two methods: 1) photodeposition, and 2) mechanical

alloying. Further, rapid screening assays for

photocatalytic activity in both the aqueous and gas

phases were developed to allow for direct comparison of

the catalysts. A total of 45 catalysts were synthesized

and tested; a commercially-available VLR catalyst was

also characterized28. Gas phase testing focused on the

PCO of ethanol; the top five performing catalysts from the rapid assay are shown in Figure 128. Constraints on the

rapid assay reactor, a static vial, limited the analytical techniques and operational optimization available; thus,

measurements focused on the oxidation of ethanol to acetaldehyde (not a favored reaction product, as acetaldehyde is

more harmful than ethanol). From these results, 3%-PbS- and 3%-Ag-doped TiO2, both prepared via the

photodeposition method, were considered for future gas-phase PCO work. While the PbS-doped catalyst showed the

highest relative activity, a determination was made to eliminate it as an option due to safety concerns associated with

lead sulfide in air purification systems.

After the catalyst down-selection was completed from the initial rapid assay testing, the KSC team has continued

work that focuses on the development of a bench-scale, high-throughput reactor incorporating the novel VLR catalyst.

Catalyst immobilization techniques, including incorporation of the catalyst into a sorbent material, were examined in

conjunction with the University of Florida (UF). Extensive modeling of a reactor test bed mimicking air duct work

with throughput similar to that seen on the International Space Station was completed to determine optimal reactor

design. High-efficiency blue LEDs were also utilized in the design to eliminate Hg-containing light sources from the

system in order to better accommodate spacecraft needs. The system was challenged with ethanol, a common volatile

organic compound (VOC) found in ISS cabin air, to evaluate

the system’s ability to perform high-throughput trace

contaminant removal. The ultimate goal for this testing was

to determine if the unit would be useful in pre-heat exchanger

operations to lessen condensed VOCs in recovered water thus

lowering the burden of VOC removal for water purification

systems.

II. High-Throughput Reactor Design

Considerations for a photocatalytic reactor concentrated

on a design that could be implemented in an air duct-style

system, focusing on elements allowing for high-throughput of

contaminated air. A rectangular reactor design was adopted

to simulate a typical air duct shape; in order to provide a low

pressure drop within the system, LED and catalyst panels,

parallel to the air flow, were used. By comparing dimensions,

flow rates, and contact times of several commercially-

available photocatalytic units tested at Marshall Space Flight

Center29, it was determined that the constructed bench scale

reactor would be 15.24 x 15.24 x 30.48 cm and consist of two

air channels processing 25 LPM per channel. To accomplish

comparable processing volumes of the commercial units

Figure 1: Gas phase ethanol oxidation to products

results for top-performing catalysts compared to

Degussa P25 and GENS NANOTM catalysts. PD =

photodeposition method; MA = mechanical alloying

method.

Figure 2: Spectroradiometer integrating

sphere during LED profiling (Top) and

irradiance profiles for LED strips (Bottom).

0

0.05

0.1

0.15

375 395 415 435Irria

dia

nce

(mW

/cm

2)

Wavelength (nm)

SuperbrightLEDs

LEDLightinghut


4

(2500 LPM or 88 cfm), the bench-scale unit would need to be approximately 3.5 times larger with multiple panel sets.

LED Light Source Selection

Commercially-available, flexible LED strips

were chosen for the light source because if their

ease in installation. Two brands of LED strips

were purchased for consideration in the reactor

design; the first set from superbrightleds.com

are described as UV/black-light LEDs with a

Correlated Color Temperature (CCT) of 440

nm. According to the vendor, CCT is the

measure of the color appearance of the light

emitted from the light source and not necessarily

the actual wavelength emitted by the source.

The second set of LED strips were purchased

from ledlightinghut.com and claimed to have a

maximum output at 400-405 nm. Irradiance

profiles were collected for both brands of LED

strips in a dark room using a spectroradiometer

(model OL754C, Optronics Laboratories, Figure

2). As seen in Figure 2, the superbightleds.com

LEDs showed a λmax of 404 nm while the

ledlightinghut.com set had a λmax at 406 nm, as

advertised. The irradiance for a single LED was determined through the integration of the irradiation scan with defined

integration limits of 5% irradiance with respect to the value at the λmax; total irradiance for this range of wavelengths

was 0.692 and 1.76 mW/cm2 for the superbrightleds.com and ledlightinghut.com LEDs, respectively. Based on these

results, the ledlightinghut.com strips were selected as the light source to be used in the reactor design since they have

a near 2.5x higher output than the superbrightled.com LEDs. Two 15.24 cm x 30.48 cm light panels were constructed,

each consisting of 234 LEDs, as tightly packed as possible for use within the reactor system. Due to constraints in the

panel size and soldering requirements of the strips, the layout of the LEDs was limited to vertical lines of LEDS rather

than a staggered pattern which would allow for slightly more uniform illumination.

The LED output and layout within the reactor drives the optimal channel height (distance between catalyst and

LED panels); because LEDs are a point source light, the distance between the LED panel and catalyst panel becomes

a compromise between obtaining evenly distributed light and the highest possible light intensity. The closer the

catalyst is to an LED, the higher the incident intensity it will receive, but the less even the light distribution will be

over the catalyst panel. The further from the LED, the less incident irradiation will be available at the catalyst surface,

but higher uniform light distribution across the panel will be achieved. Based on Figure 3, which shows the pattern

of 120° light output from each LED on the panel, it was determined that the channel height must remain between

0.535 and 8.030 cm to achieve optimal intensity and uniform illumination at lower distances.

Computational Fluid Dynamic (CFD) Analysis

A major consideration within the reactor design was ensuring that flow through each channel allowed for sufficient

interaction of contaminants with the catalyst panel (i.e., that there was turbulent flow rather than laminar flow). In

order to encourage turbulent flow, fins were

attached to the catalyst panel; various

configurations and sizes were explored to

achieve optimal flow dynamics via

Computational Fluid Dynamic (CFD) Analysis

using OpenFOAM v2.4. Designs for the fin

geometry iterated from discrete, alternating

angled fins to continuous V-shaped fins

pointing upstream. A design with parallel,

angled fins was also considered but was found

to perform poorly in comparison to the V-

shaped fins. A successful design was chosen

based on visual inspection of the results as well

Figure 4: CFD Domain

Figure 3: LED panel lighting distribution based on 120°

light spread from each LED (blue); the red line indicates the

minimum distance while the gold line indicates the furthest

distance from the top of the LEDs to the catalyst panel.


5

as the estimated pressure drop across the channel. Figure 4 shows the simulation domain for the chosen design. The

channel was extended slightly at the inlet and outlet in order to avoid instabilities in the simulation.

The turbulent air flow was modeled using the SST k-omega turbulence model in a Reynolds-averaged Navier-

Stokes (RANS) setup. A steady-state solver for incompressible, turbulent flows was used; this solver is called

simpleFoam in OpenFOAM. It was assumed that the air could be treated as an incompressible ideal gas due to its low

speed through the channel (Mach number is much less than 0.3). No chemical reactions were modeled and any thermal

effects were neglected. Boundary conditions used in the simulation included a fixed inlet velocity of 0.2691 m/s

(equivalent to 25 L/min volumetric flow rate), fixed outlet pressure of 101325 Pa, no-slip walls with appropriate wall

functions, and a channel height of 1.016 cm.

With a flow rate of 25 LPM through a single channel, the pressure drop was fairly low: within the range of 1-2 Pa.

As expected, several recirculation regions existed within the channel as air flowed through. The V-shaped fins split

the flow and created two large counter-rotating vortices. Air near the bottom wall was forced outward toward the

sidewalls. Along the top surface, the flow was pushed inward towards the center of the channel. Additional smaller

recirculation regions formed along the immediate downstream side of the ribs. These smaller local vortices forced

the air to repeatedly make contact with the surface as it travelled between fins while the large-scale rotations helped

force air near the top surface downwards. The LED lights along the top surface also helped to disturb any boundary

layer growth, and generally added to the turbulence within the system. The local vortices can be seen via streamlines

in Figure 5. Point and line sources were used to generate streamlines through a particular set of coordinates.

Figure 6: Point source on tip of first rib (Right); point source on tip of third rib (Left).

Figure 5: (A) horizontal line source; (B) vertical line source


6

The large-scale rotations can be seen in Figure 6 from the line source streamlines placed at the inlet. For clarity,

the streamlines were clipped to show the how the flow behaves in the top and bottom halves of the channel. In Figure

6, (A) shows streamlines from a horizontal line source along the channel center while (B) shows streamlines from a

vertical line source towards the left side of the channel. Images of the top halves from both show how the flow is

pushed inward towards the center as it travels downstream. Along the bottom half, the flow travels outward towards

the sidewall while circulating between the fins. The flow break-up as a result of the LED obstruction can also be seen

near the top wall.

III. VLR Catalyst Preparation, Immobilization, and VOC Testing Setup

Catalyst Preparation and Immobilization

Two immobilization techniques were used in these

studies, both employing a sorbent material to enhance the

catalyst contact time. First, a 3%-Ag-doped TiO2 powder was

prepared via photodeposition. 0.25 g of Degussa P25

titanium dioxide and 100 mL of 1:20 degassed ethanol:water

were added to a 125-mL Erlenmeyer flask. An appropriate

amount of 0.1 M metal was added to the flask to achieve a

3.0% dopant loading by weight. The flasks were placed on a

stir plate approximately 3-cm from a 15W, T5 UV-A

fluorescent bulb and allowed to react for 30 minutes (Figure

7). The flasks received a total of approximately 100 mW of

UV radiation during the reaction. The catalyst was then

centrifuged at 4500 rpm for 12 minutes to separate the solid

from the liquid. The catalyst pellet was then washed and

centrifuged in deionized water three times before drying at

70°C overnight.

The catalyst powder was then stirred with DI water (71%

weight) and Ludox colloidal silica (25% weight) for 15 minutes. Individual 2.54 x 2.54-cm ceramic tiles were

submerged into the mixture, removed, and placed horizontally on an oven-safe pan to achieve evenly distributed TiO2

dispersion. The tiles were dried for 75 minutes at 100°C, then allowed to cool for 15 minutes before adding additional

layers of catalyst. The process was repeated four times for all tiles. These tiles, containing a thin layer of the

adsorption-enhanced catalyst were then affixed to a metal panel with Torr Seal for testing within the reactor system

(Figure 8); stainless steel bars were used as the V-shaped fins.

For the second immobilization technique tested, a similar method of photodeposition as described above was used

to dope pre-fabricated, adsorption-enhanced silica-titania-composite pellets (STCs)4, 5, 9, containing bare Degussa P25.

The STCs were 4% P25 by mass, thus requiring 556 µL of 0.1 M silver solution to 5 g of STCs to achieve the same

dopant loading level. The centrifugation step described above was eliminated from the post-reaction washing steps.

Through this process, the pellets, seen in Figure 8, garnered a 104-µm average of Ag-dopant on the pellet surface,

indicating that the light did not penetrate completely through the pellet during photodeposition process. Traditionally,

STC pellets have been used in a packed bed reactor4, 9, 30; with the proposed reactor design, light penetration from

LED panels into a large packed-bed would not be conducive for adequate activation of the catalyst; thus, the pellets

were adhered to a panel using an inorganic cement adhesive (Omegabond 400) as densely as possible.

Figure 7: Photodeposition preparation method

for quantum dot formation on TiO2.

Figure 8: Ag-TiO2-coated ceramic tile (left) and VLR-activated pellets (right) prepared via direct

photodeposition.


7

Catalyst Physical Characterization

1. Diffuse Reflectance Analysis

In order to determine if the emission wavelength range of the selected LEDs was suitable for the VLR catalyst

activation, the band gap of the VLR catalyst was determined. A 60-mm diameter integrating sphere was mounted to

a Jasco V-670 UV-visible spectrophotometer for diffuse reflectance spectra analysis of prepared bulk powder samples.

A Spectralon certified reference was used as a standard. Percent reflectance was measured at ambient temperature

between 300 and 800 nm at 2-nm intervals, 1000 nm/min continuous scan, with a 340-nm light source. Band-gap

energy estimation was measured from the diffuse reflectance spectroscopy (DRS). Through a series of mathematical

transformations31-34, the Tauc plot is created and the tangent line at the point of inflection is drawn; the point at which

the tangent line meets the x-axis equates to the band gap, Eg, value. The calculation of the band gap was completed

using a Matlab script.

When photons are absorbed into a semiconductor at energy greater than the gap of the semiconductor, an electron

is transferred from the valence band to the conduction band where an abrupt increase in the absorbency of the material

occurs to the wavelength corresponding to the band gap energy. The relation of the absorption coefficient to the

incidental photon energy depends on the type of electronic transition. When the electron momentum is conserved

during the transition of electrons, the transition is direct. When the electron momentum is not conserved, and the

transition requires assistance by a photon, it is it is considered an indirect transition. To analyze the electronic

properties of the Ag-doped-TiO2 samples in this experiment, the indirect transition (n=2) was used35, 36.

Figure 9 shows Tauc plots for both bare Degussa P25 TiO2 and the VLR Ag-doped TiO2. It is known that the band

gap for anatase titanium dioxide is ~3.2 eV, and that of rutile titanium dioxide is ~3.0 eV, requiring a wavelength of

388 nm and 414 nm for activation, respectively. Degussa P25 is a mixture of anatase (70-85%) and rutile TiO2 and

has shown superior photocatalytic activity compared to either phase individually8. Figure 9A shows a band gap of

3.15 eV for the Degussa mixture (likely due to the high content of the anatase phase). There also exists a slight

shoulder in the curve to the left of the red band gap determination line. This shoulder could be due to an impurity in

the sample or partially due to the small percentage of rutile phase TiO2 in the Degussa P25 mix. Once doped with

silver, the catalyst exhibits absorption of light over a range of energies as signified by the multiple curves in Figure

9B. From the plot, the band gap for the doped TiO2 was determined to be 2.72 eV; this energy correlates to activation

of the catalyst by a wavelength of 457 nm or lower for activation. By using the selected LEDs with a maximum

irradiance at 405 nm, activation of the dopant and the rutile phase, itself, can be accomplished for further increased

photocatalytic activity. While the rutile phase is known to have inferior activity due to its direct band gap structure8,

with the presence of the anatase phase TiO2 and VLR dopant, less electron-hole recombination events are expected

due to the increased ability for charge separation and electron transfer in the material.

VOC Challenge Testing

1. Catalyst Activity Testing

Due to the suboptimal nature of the vial studies completed previously28, experiments were performed using a

custom-made annular reactor (Southern Scientific, Inc., Micanopy, FL) described in previous work9, 30 to assess the

immobilized catalyst activity in a ruggedized system. The reactor was packed with 14.6 g of VLR-modified STC

pellets to give a 63-mm bed height (0.17 s contact time). Continuous illumination by a custom-built, 12-LED light

Figure 9: Tauc Plots showing band gap energies for A) Degussa P25 TiO2 and B) 3% Ag-doped TiO2

A B


8

stick using the LEDs detailed above was implemented. All tests were carried out in a flow-through mode with

uninterrupted 2 LPM CO2-free air (30.0 ± 1.0% RH) containing either 5 or 50 ppmv ethanol at 25°C as the test volatile

organic compound (VOC), provided by a Kintek Trace Gas Generator. The STC pellets were regenerated in-line

between each test by passing a VOC-free (30.0 ± 1.0% RH) sweeping gas at 25°C through the reactor accompanied

by visible light irradiation. Both influent and effluent streams were sampled alternately every 8.45 minutes using a

VICI Valco sampling and switch valve controlled by a Thermo Scientific Trace Ultra GC. Gas streams were analyzed

for the test contaminant and its oxidation intermediates by GC-FID equipped with an HP Plot Q column (30 m x 0.32

mm, 20 µm d.f.). The effluent stream was also directed to a Li-Cor CO2 analyzer for the determination of the rate of

CO2 production. Data logging of parameters including relative humidity, pressure, and flow rates, which were not

recorded by the GC/FID software, was completed using a custom-programmed OPTO 22 system.

2. High-Flow Test Bed for VOC Challenge Testing

All experiments were performed within the developed high-flow, bench-scale reactor containing two 15.24 x

30.48-cm panels of LEDs and two 15.24 x 30.48-cm catalyst-coated panels with V-shaped fins; each channel was

1.016 cm wide. CO2-free air was supplied to a Kintek Trace Gas Generator to produce a calibrated, 2 LPM

contaminant stream. This stream was then diluted with the desired flow of humidified, CO2-free air produced by a

Miller Nelson HCS-401 Flow-Temperature-Humidity Controller. All tests were carried out in a flow-through mode

with a relative humidity of 30.0 ± 3.0% containing 30 ppmv ethanol. Process monitoring followed equivalent

procedures to the catalyst activity testing described above. Figure 10 shows the high-flow PCO test bed setup.

3. Calculations for PCO Oxidation Rate

PCO performance was quantified by contaminant removal: the measure of the removal of the ethanol regardless

of it being adsorbed or oxidized at pseudo-steady state conditions, and oxidation efficiency (XA), the measure of the

oxidation of the ethanol to CO2. These values were calculated using Eqs. 1 and 2, respectively, where [EtOH]IN and

[EtOH]OUT are the influent and effluent ethanol concentrations and [CO2]OUT is the CO2 generated by the PCO at

pseudo-steady state conditions. In Eq. 2, multipliers are used for EtOH because calculations are based on carbon

concentration. The rate of PCO of the test contaminant, r, was determined based on the formation of ACD or CO2

rather than the disappearance of the contaminant to prevent overestimation due to the VOC adsorption to the silica-

rich photocatalyst.

𝐸𝑡𝑂𝐻 𝑅𝑒𝑚𝑜𝑣𝑎𝑙 = (([𝐸𝑡𝑂𝐻]𝐼𝑁− [𝐸𝑡𝑂𝐻]𝑂𝑈𝑇)

[𝐸𝑡𝑂𝐻]𝐼𝑁) ∗ 100 (1)

𝑋𝐴 = ([𝐶𝑂2]𝑂𝑈𝑇

2∗[𝐸𝑡𝑂𝐻]𝐼𝑁) ∗ 100 (2)

Figure 10: PCO test bed setup.


9

IV. Reactor Testing Results

1. Catalyst Activity Testing

As stated, previous work on the gas-phases photocatalytic activity was performed in a suboptimal rapid assay

designed to quickly assess numerous catalysts in a powder form under equivalent settings28. In order to better gauge

the photocatalytic activity of the VLR catalyst down-selected for these experiments in conjunction with the selected

adsorption-based immobilization technique, the Ag-doped STC pellets were tested in a packed-bed, annular reactor

system. Figure 11 shows the influent and effluent species over the course of the studies when 5 and 50 ppm ethanol

concentrations were used. In both experiment set ups, acetaldehyde and ethanol were the only compounds detected

in the effluent stream; a carbon mass balance verified that no other compounds were produced. Ethanol removal,

regardless of whether it was oxidized or adsorbed to the catalyst pellet, was determined to be 57.81% and 20.13% for

5 ppmv and 50 ppmv influent ethanol, respectively. Complete mineralization to CO2, in any capacity was not achieved

in either study; as a measure of the oxidation efficiency, Eq. 2 was altered to reflect the amount of ethanol oxidized to

the main reaction intermediate detected, acetaldehyde. This oxidation efficiency was calculated to be 56.11% and

15.40% for 5 ppmv and 50 ppmv influent ethanol, respectively. Based on these results, the 5 ppmv ethanol study

shows that the ethanol removal was due, almost entirely, to oxidation nearly, while the removal during the 50 ppmv

challenge was ~75% due to oxidation and 25% due to adsorption onto the pellets. It is obvious that VOC contaminant

concentration has a major effect on the oxidation capacity of the catalyst; the 50 ppmv challenge was used to provide

enough reactant to detect if multiple intermediate species would be produced beyond acetaldehyde, while the 5 ppmv

challenge focused on a contaminant level closer to what would be expected in the ISS environment.

The contact time of 0.17 s was greater than the minimum 0.1 s contact time determined by Perry et al.29 to provide

appreciable VOC degradation in PCO reactor systems and is not expected to be the cause of the low performance seen

in the annular system. The reduced activity has several potential causes including low catalyst loading on the pellets

or insufficient light intensity. As discussed above, the VLR-activated portion of the pellets only accounted for only

~23% of the total pellet volume; incorporating the Ag-doped TiO2 powder during the pellet formation procedure

instead of completing VLR activation by direct photodeposition on the pre-formed pellets would allow for the full 4%

catalyst concentration within the silica scaffold to be VLR. This increased catalyst loading would likely increase the

overall photocatalytic activity of the catalyst system to provide sufficient mineralization of ethanol to CO2 rather than

to the acetaldehyde intermediate. Further, it is known that increasing incident irradiation within a reactor system when

the bare STCs are used in conjunction with a UV light source can increase ethanol removal, increase mineralization

of ethanol to CO2, and lower the amount of acetaldehyde in the effluent stream9. While testing this VLR catalyst was

limited to the selected LEDs, increasing the intensity through selection of different LEDs or using a more-densely

packed source with the current LEDs could also provide synergistic effects for the activity of the reactor. Due to time

constraints in the project, these hypotheses could not be tested to determine the source of the low activity. However,

these experiments served as a basis for the expected activity within the high-throughput reactor.

Figure 11: Influent and effluent profiles for A) 5 ppmv and B) 50 ppmv ethanol challenges in annular reactor

system; (●) designates influent ethanol, (○) designates effluent ethanol, and (◊) designates effluent

acetaldehyde.

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

0 10 20 30 40

Co

nce

ntr

ati

on

(p

pm

v)

Time (hr)

0.00

10.00

20.00

30.00

40.00

50.00

60.00

0 10 20 30

Co

nce

ntr

ati

on

(p

pm

v)

Time (hr)

A B


10

2. High-Flow Test Bed for VOC Challenge Testing

VLR-activated STC pellets immobilized onto metal panels were tested in the high-throughput system at various

flow rates to determine a correlation between reactor contact time and photocatalytic activity. Approximately 75% of

each 464.5-cm2 panel was covered with pellets at an average packing density of approximately 4.24 pellets/cm2. Table

1 details the various contact times tested based on the total reactor volume and on the amount of catalyst contained on

the two catalyst panels within the reactor. The catalyst contact time used in this set of experiments was 1.9-4.4 times

higher than that used in the annular reactor; coupled with the use of the V-shaped fins to ensure that the contaminated

air flow would intimately contact the catalyst within the reactor, it was expected that similar or better activity would

be achieved.

Table 1: High-throughput contact times for varied flow rate tests

Flow rate (LPM) Reactor Contact Time (s) Catalyst Contact Time (s)

35 12.14 0.32

25 17.00 0.44

15 28.31 0.74

As the contact time was increased within the reactor, the ethanol removal capacity decreased from 7.55% to 5.37%

to 4.65% for 0.32, 0.44, and 0.74 s, respectively. While the overall removal decreased, the extent of removal due to

oxidation to acetaldehyde rather than due to adsorption increased from 51% to 100%. While it was expected that

removal would have increased with increased contact time, the lowered flow rate could have resulted in less turbulent

flow, leading to less contact of the contaminated air stream with the catalyst panel. Similar to the annular, packed-

bed testing, it is expected that low catalyst loading and possibly low incident irradiation are the leading causes of low

catalytic activity within the system.

Similar testing was completed with coated ceramic tiles affixed to panels for influent ethanol concentrations of

~30 ppmv and ~5 ppm at a 35 LPM rate. In both studies, no photocatalytic activity was achieved. While the catalyst

was mixed with colloidal silica for enhancing activity through addition of a sorbent material, the lack of a xerogel

silica structure (e.g., the pellet structure) greatly reduced the adsorptive capacity of the catalyst system. The silica-

VLR TiO2 coating was also easily knocked off of the ceramic tiles. The experiment clearly demonstrates that

adsorption-enhanced photocatalysis provided better contaminant degradation and removal. Further, results indicate

that catalyst immobilization remains a challenge in the high-throughput system. Redesign of the system to contain a

substrate with better catalyst adhesion and higher catalyst loading are needed to further the technology as an option

for reduction of polar organic compounds in closed-environment air systems. Once a successful design for catalyst

Figure 12: Influent and effluent profiles high-throughput reactor with pellet-covered panels at varied flow

rates; (●) designates influent ethanol, (○) designates effluent ethanol, and (◊) designates effluent

acetaldehyde. Vertical red lines designate when the system flow rate was altered during the run.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

0.00

10.00

20.00

30.00

40.00

50.00

0 10 20 30 40 50 60 70

Ace

tald

ehy

de

(pp

mv

)

Eth

an

ol

(pp

mv

)

Time (hrs)

35 LPM 25 LPM 15 LPM


11

immobilization is proven, rugged testing with multiple contaminants would allow for a clearer picture of how the

technology could be implemented for trace contaminant control in relevant environments.

V. Conclusions

Alteration of TiO2 to become visible-light-responsive continues to be a leading challenge in the field of

photocatalysis. While KSC researchers were able to create several modified-TiO2 catalysts showing photocatalytic

activity, implementation of the novel catalyst into a high-throughput system still requires further refinement to produce

a system capable of treating polar organic compounds in the gas phase. Suboptimal results attained in this study

demonstrate where efforts in improvements should be focused: 1) optimization of catalyst loading within the reactor

system, 2) validation of light intensity needs for increased catalytic activation, and 3) improved catalyst

immobilization techniques. Increased catalyst concentration within the reactor system will allow for sufficient

available active sites for oxidation of contaminants. As mentioned, previous work demonstrated that increased

irradiance levels allow for increased activity and mineralization when bare TiO2 and UV wavelengths are utilized;

similar results are expected in VLR systems. Most importantly, improving the catalyst scaffold is essential to

achieving a useful high-throughput system; examination of other immobilization techniques such as impregnation of

filter materials or varied catalyst panel designs incorporating sorbents for enhanced activity are warranted. It is

expected that resolution of these key issues would lead to the evolution of the VLR-catalyst reactor presented in this

work to be a useful technology to lower the level of organic contaminants entering water purification systems via heat

exchangers in closed environments.

Acknowledgments

This effort was supported by the Kennedy Space Center 2015 Center Innovation Fund (CIF). The authors would

like to thank Ms. Regina Rodriguez at the University of Florida for her support in preparing the catalyst-coated ceramic

tiles and Mr. Robert Athman for construction of the LED light stick used in the low-flow catalyst testing experiments.

References

1. Ohtani, B., Prieto-Mahaney, O. O., Li, D., and Abe, R. "What is Degussa (Evonik) P25? Crystalline composition analysis,

reconstruction from isolated pure particles and photocatalytic activity test," Journal of Photochemistry and Photobiology

A: Chemistry Vol. 216, No. 2–3, 2010, pp. 179-182.doi: http://dx.doi.org/10.1016/j.jphotochem.2010.07.024

2. Carp, O., Huisman, C. L., and Reller, A. "Photoinduced reactivity of titanium dioxide," Progress in Solid State Chemistry

Vol. 32, No. 1–2, 2004, pp. 33-177.doi: http://dx.doi.org/10.1016/j.progsolidstchem.2004.08.001

3. Luo, Y., and Ollis, D. F. "Heterogeneous Photocatalytic Oxidation of Trichloroethylene and Toluene Mixtures in Air:

Kinetic Promotion and Inhibition, Time-Dependent Catalyst Activity," Journal of Catalysis Vol. 163, No. 1, 1996, pp.

1-11.doi: http://dx.doi.org/10.1006/jcat.1996.0299

4. M. Stokke, J., Mazyck, D. W., Wu, C. Y., and Sheahan, R. "Photocatalytic oxidation of methanol using silica-titania

composites in a packed-bed reactor," Environmental Progress Vol. 25, No. 4, 2006, pp. 312-318.doi: 10.1002/ep.10164

5. Stokke, J. M., and Mazyck, D. W. "Photocatalytic Degradation of Methanol Using Silica–Titania Composite Pellets:

Effect of Pore Size on Mass Transfer and Reaction Kinetics," Environmental Science & Technology Vol. 42, No. 10,

2008, pp. 3808-3813.doi: 10.1021/es703221c

6. Levine, L. H., Richards, J. T., Coutts, J. L., Soler, R., Maxik, F., and Wheeler, R. M. "Feasibility of Ultraviolet-Light-

Emitting Diodes as an Alternative Light Source for Photocatalysis," Journal of the Air & Waste Management Association

(Air & Waste Management Association) Vol. 61, No. 9, 2011, pp. 932-940.doi: 10.1080/10473289.2011.596746

7. Kwon, S., Fan, M., and Cooper, A. T. "Photocatalytic Applications of Micro- and Nano-TiO2 in Environmental

Engineering," Critical Reviews in Environmental Science and Technology Vol. 38, No. 3, 2008, pp. 197-226.doi:

10.1080/10643380701628933

8. Agrios, A. G., Gray, K. A., and Weitz, E. "Photocatalytic Transformation of 2,4,5-Trichlorophenol on TiO2 under Sub-

Band-Gap Illumination," Langmuir Vol. 19, No. 4, 2003, pp. 1402-1409.doi: 10.1021/la026397x

9. Coutts, J. L., Levine, L. H., Richards, J. T., and Mazyck, D. W. "The effect of photon source on heterogeneous

photocatalytic oxidation of ethanol by a silica–titania composite," Journal of Photochemistry and Photobiology A:

Chemistry Vol. 225, No. 1, 2011, pp. 58-64.doi: http://dx.doi.org/10.1016/j.jphotochem.2011.09.026

10. Zayat, M., Garcia-Parejo, P., and Levy, D. "Preventing UV-light damage of light sensitive materials using a highly

protective UV-absorbing coating," Chemical Society Reviews Vol. 36, No. 8, 2007, pp. 1270-1281.

11. Lanfang, L., Janelle, C., Jeffrey, R., Paul, H., and Christian, C. "Review on Transforming TiO2 into a Visible-Light-

Responsive Catalyst for Water and Air Purification," 42nd International Conference on Environmental Systems.

American Institute of Aeronautics and Astronautics, 2012.

12. Maeda, K., and Domen, K. "Photocatalytic water splitting: recent progress and future challenges," The Journal of

Physical Chemistry Letters Vol. 1, No. 18, 2010, pp. 2655-2661.

http://dx.doi.org/10.1016/j.jphotochem.2010.07.024

http://dx.doi.org/10.1016/j.progsolidstchem.2004.08.001

http://dx.doi.org/10.1006/jcat.1996.0299

http://dx.doi.org/10.1016/j.jphotochem.2011.09.026


12

13. Byrne, J. A., Fernandez-Ibañez, P. A., Dunlop, P. S., Alrousan, D., and Hamilton, J. W. "Photocatalytic enhancement for

solar disinfection of water: a review," International Journal of Photoenergy Vol. 2011, 2011.

14. Malato, S., Fernández-Ibáñez, P., Maldonado, M., Blanco, J., and Gernjak, W. "Decontamination and disinfection of

water by solar photocatalysis: recent overview and trends," Catalysis Today Vol. 147, No. 1, 2009, pp. 1-59.

15. Grätzel, M. "Photoelectrochemical cells," Nature Vol. 414, No. 6861, 2001, pp. 338-344.

16. Snaith, H. J., and Schmidt‐Mende, L. "Advances in liquid‐electrolyte and solid‐state dye‐sensitized solar cells," Advanced

Materials Vol. 19, No. 20, 2007, pp. 3187-3200.

17. Anpo, M. "Use of visible light. Second-generation titanium oxide photocatalysts prepared by the application of an

advanced metal ion-implantation method," Pure and applied chemistry Vol. 72, No. 9, 2000, pp. 1787-1792.

18. Anpo, M., Ichihashi, Y., Takeuchi, M., and Yamashita, H. "Design of unique titanium oxide photocatalysts by an

advanced metal ion-implantation method and photocatalytic reactions under visible light irradiation," Research on

chemical intermediates Vol. 24, No. 2, 1998, pp. 143-149.

19. Asahi, R., Morikawa, T., Ohwaki, T., Aoki, K., and Taga, Y. "Visible-light photocatalysis in nitrogen-doped titanium

oxides," science Vol. 293, No. 5528, 2001, pp. 269-271.

20. Choi, W., Termin, A., and Hoffmann, M. R. "The role of metal ion dopants in quantum-sized TiO2: correlation between

photoreactivity and charge carrier recombination dynamics," The Journal of Physical Chemistry Vol. 98, No. 51, 1994,

pp. 13669-13679.

21. Li, X., Li, F., Yang, C., and Ge, W. "Photocatalytic activity of WO x-TiO 2 under visible light irradiation," Journal of

Photochemistry and Photobiology A: Chemistry Vol. 141, No. 2, 2001, pp. 209-217.

22. Ohno, T., Akiyoshi, M., Umebayashi, T., Asai, K., Mitsui, T., and Matsumura, M. "Preparation of S-doped TiO 2

photocatalysts and their photocatalytic activities under visible light," Applied Catalysis A: General Vol. 265, No. 1, 2004,

pp. 115-121.

23. Rengifo-Herrera, J., Kiwi, J., and Pulgarin, C. "N, S co-doped and N-doped Degussa P-25 powders with visible light

response prepared by mechanical mixing of thiourea and urea. Reactivity towards E. coli inactivation and phenol

oxidation," Journal of Photochemistry and Photobiology A: Chemistry Vol. 205, No. 2, 2009, pp. 109-115.

24. Rengifo-Herrera, J. A., Pierzchała, K., Sienkiewicz, A., Forró, L., Kiwi, J., Moser, J. E., and Pulgarin, C. "Synthesis,

characterization, and photocatalytic activities of nanoparticulate N, S-codoped TiO2 having different surface-to-volume

ratios," The Journal of Physical Chemistry C Vol. 114, No. 6, 2010, pp. 2717-2723.

25. Gao, B., Kim, Y. J., Chakraborty, A. K., and Lee, W. I. "Efficient decomposition of organic compounds with FeTiO

3/TiO 2 heterojunction under visible light irradiation," Applied Catalysis B: Environmental Vol. 83, No. 3, 2008, pp.

202-207.

26. Kim, Y. J., Gao, B., Han, S. Y., Jung, M. H., Chakraborty, A. K., Ko, T., Lee, C., and Lee, W. I. "Heterojunction of

FeTiO3 nanodisc and TiO2 nanoparticle for a novel visible light photocatalyst," The Journal of Physical Chemistry C

Vol. 113, No. 44, 2009, pp. 19179-19184.

27. Kang, M. G., Han, H.-E., and Kim, K.-J. "Enhanced photodecomposition of 4-chlorophenol in aqueous solution by

deposition of CdS on TiO 2," Journal of Photochemistry and Photobiology A: Chemistry Vol. 125, No. 1, 1999, pp. 119-

125.

28. Coutts, J. L. H., Paul E.; Clausen, Christian A.; Richards, Jeffrey, T. "Visible-Light-Responsive Catalysts Using Quantum

Dot-Modified TiO2 for Air and Water Purification," 44th International Conference on Environmental Systems, Tucson,

AZ, July 2014.

29. Jay, P., Kenneth, F., Joseph, S., and Dana, R. "A Comparison of Photocatalytic Oxidation Reactor Performance for

Spacecraft Cabin Trace Contaminant Control Applications," 41st International Conference on Environmental Systems.

American Institute of Aeronautics and Astronautics, 2011.

30. Coutts, J. L. "TRACE CONTAMINANT CONTROL: AN IN-DEPTH STUDY OF A SILICA-TITANIA COMPOSITE

FOR PHOTOCATALYTIC REMEDIATION OF CLOSED-ENIVRONMENT HABITAT AIR." University of Central

Florida Orlando, Florida, 2013.

31. Murphy, A. "Band-gap determination from diffuse reflectance measurements of semiconductor films, and application to

photoelectrochemical water-splitting," Solar Energy Materials and Solar Cells Vol. 91, No. 14, 2007, pp. 1326-1337.

32. Tauc, J., Grigorovici, R., and Vancu, A. "Optical properties and electronic structure of amorphous germanium," physica

status solidi (b) Vol. 15, No. 2, 1966, pp. 627-637.

33. Tauc, J., and Abeles, F. Optical properties of solids: North-Holland, Amsterdam, 1972.

34. Davis, E., and Mott, N. "Conduction in non-crystalline systems V. Conductivity, optical absorption and

photoconductivity in amorphous semiconductors," Philosophical Magazine Vol. 22, No. 179, 1970, pp. 0903-0922.

35. Willardson, R. K., and Beer, A. C. Optical properties of III-V compounds: Academic Press, 1967.

36. Dressel, M. "Gruner; G. Electrodynamics of Solids: Optical Properties of Electrons in Matter." Cambridge University

Press: Cambridge, 2002.

Visible-Light-Responsive Photocatalysis: Ag-Doped TiO2 ...

Documents

Transcript of Visible-Light-Responsive Photocatalysis: Ag-Doped TiO2 ...