Chemosphere xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation Pankaj Chowdhury, Hassan Gomaa, Ajay K. Ray ⇑ Department of Chemical and Biochemical Engineering, University of Western Ontario, London, Ontario N6A 5B9, Canada

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Sacrificial H2 generation in 1 sun is

studied with Pt loaded EY-sensitized TiO2.  H2 generation reaction mechanism in UV, visible and solar light is established.  In both solar and visible light HCHO is detected as an intermediate.  H2 generation kinetics follows a Langmuir-type isotherm.  At lower intensity, QYs were much higher due to insignificant charge recombination.

a r t i c l e

i n f o

Article history: Received 15 May 2014 Received in revised form 28 October 2014 Accepted 30 October 2014 Available online xxxx Handling Editor: E. Brillas Keywords: Hydrogen Photocatalysis Dye-sensitization Solar Triethanolamine Eosin Y

a b s t r a c t In this paper, we have studied Eosin Y-sensitized sacrificial hydrogen generation with triethanolamine as electron donor in UV, visible, and solar light irradiation. Aeroxide TiO2 was loaded with platinum metal via solar photo-deposition method to reduce the electron hole recombination process. Photocatalytic sacrificial hydrogen generation was influenced by several factors such as platinum loading (wt%) on TiO2, solution pH, Eosin Y to Pt/TiO2 mass ratio, triethanolamine concentration, and light (UV, visible and solar) intensities. Detailed reaction mechanisms in visible and solar light irradiation were established. Oxidation of triethanolamine and formaldehyde formation was correlated with hydrogen generation in both visible and solar lights. Hydrogen generation kinetics followed a Langmuir-type isotherm with reaction rate constant and adsorption constant of 6.77  106 mol min1 and 14.45 M1, respectively. Sacrificial hydrogen generation and charge recombination processes were studied as a function of light intensities. Apparent quantum yields (QYs) were compared for UV, visible, and solar light at four different light intensities. Highest QYs were attained at lower light intensity because of trivial charge recombination. At 30 mW cm2 we achieved QYs of 10.82%, 12.23% and 11.33% in UV, visible and solar light respectively. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Photocatalytic water splitting for hydrogen production has drawn attention due to its potential to generate green fuel (with ⇑ Corresponding author. Fax: +1 519 661 3498. E-mail address: [email protected] (A.K. Ray).

no CO2 emission) from water. Fujishima and Honda (1972) reported the photocatalytic water splitting over a TiO2 single crystal, following which a remarkable progress was witnessed in the last several decades under ultraviolet (UV) light (Domen et al., 1986; Kudo and Kato, 1997). In different studies, Pt/TiO2, RuO2/ TiO2, reduced SrTiO3 electrode with a platinum counter electrode, platinized SrTiO3, SrTiO3 powder modified by rhodium oxide, and

http://dx.doi.org/10.1016/j.chemosphere.2014.10.076 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

2

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

nickel-loaded SrTiO3 were also investigated for improvement in their photocatalytic activities (Bulatov and Khidekel, 1976; Wrighton et al., 1976; Sato and White, 1980; Wagner and Somorjai, 1980; Lehn et al., 1982; Domen et al., 1986). The primary problem with UV photocatalysis was its limited solar spectrum of 4% compared to 46% for visible light. For visible light water splitting, the photocatalytic materials should have proper band positions and suitable band gap energy (1.23 eV < Eg < 3.0 eV) (Lee, 2005). Oxides such as TiO2, ZnO, and SnO2 have large band gaps (3–3.8 eV) and can absorb only UV part of the solar radiation, and therefore, have low conversion efficiencies. Only few chalcogenides (CdS, CdSe etc.) have band gaps within 1.23 eV and 3.0 eV, which can be activated with visible light. These catalysts, however, cannot be used because of severe photo corrosion problems (Lee, 2005). Semiconductor photocatalyst can be modified to expand the photo-response to visible region in several ways such as doping with cation/anion, spectral sensitization using dye or polymer, coupling with another small band gap semiconductor and implantation of metal ions (Ni et al., 2007). To harvest solar visible light, spectral sensitization of broad band-gap semiconductors such as TiO2, and ZnO by adsorbed dye molecule have been studied for sacrificial hydrogen generation with several electron donors (Abe et al., 2005). Organic dyes, inorganic sensitizers, and coordination metal complexes were mostly studied for sacrificial hydrogen generation. Ruthenium-based dyes which are very expensive and toxic would not be economical in large scale applications. On the contrary, organic dyes are less toxic, less expensive and can be used for dye sensitization process. Several organic dyes such as Eosin Y, Rose Bengal, Merocyanine, Cresyl violet and Riboflavin have been utilized for spectral sensitization of semiconductors (Donghua and Jingfei, 2012). Among these dyes, we have chosen Eosin Y (EY) which shows a good absorption in solar visible spectrum and can be used in solar light assisted hydrogen generation. In dye sensitization process, sacrificial electron donors play a vital role, as the dye regeneration and electron injection to semiconductor are assisted by the electron donor. By choosing an industrial organic effluent as electron donor, one can make the hydrogen generation process self-sustained. Literature shows plenty of electron donors such as ethylenediaminetetraacetic acid (EDTA), acetonitrile, methanol, isopropanol, IO3/I, diethanolamine, triethanolamine, chloroacetic acid, and oxalic acid for such applications. Among these we have chosen triethanolamine (TEOA) as the electron donor which could be achieved easily from industrial effluents related to dry cleaning, cosmetics, shampoo, detergents, surfactant, textile and water repellents (Chan et al., 1992; West and Gonsior, 1996). EY fulfills the thermodynamics for both electron injection and EY regeneration as the LUMO (lowest unoccupied molecular orbital) of EY is 0.92 V compared to the ECB (conduction band energy) of TiO2 of 0.5 V and the HOMO (highest occupied molecular orbital) of EY is +1.15 V compared to the E(TEOA+/TEOA) (reduction potential) of +0.82 V (Kalyanasundaram et al., 1978; Wang et al., 2005). Moreover, EY and TEOA were proven to be an effective dye-electron donor couple for sacrificial hydrogen generation (Abe et al., 2000). EY–TEOA couple was used for visible-light-driven dye-sensitized hydrogen generation with a wide variety of materials such as TS-1 zeolite (Zhang et al., 2008), silica gel (Zhang et al., 2007), multiwalled carbon nanotubes (Li et al., 2007; Kang et al., 2012), nanotube Na2Ti2O4(OH)2 (Li and Lu, 2007), sol–gel TiO2 (Sreethawong et al., 2009), N-doped sol–gel TiO2 (Li et al., 2008), and silane coupled-TiO2 (Abe et al., 2000). Although these studies have provided plenty of information on visible-light-driven dye-sensitized hydrogen generation in presence of sacrificial electron donor, none of them have investigated the dye-sensitized process in complete solar spectrum which is essential for using the photocatalyst in real

applications. Furthermore, there is also insufficient information about the characteristics of the electron donor and dye in presence of solar UV radiation as well as lack of intermediate analysis to assist in establishing the reaction mechanism. In this work, we aim to explore the sacrificial hydrogen generation in complete solar spectrum with EY sensitized platinum loaded TiO2 in aqueous TEOA solution. The photocatalytic behavior has been systematically studied in solar–UV (300–388 nm), solar– visible (420–650 nm) and full solar spectrum (300–650 nm) by varying reaction conditions including (i) light intensity, (ii) solution pH, (iii) platinum content (wt%) on TiO2, (iv) concentration of TEOA, and (v) mass ratio of EY to Pt/TiO2. Photocatalytic reaction mechanisms and charge recombination processes were explored as functions of light intensity in UV, visible and solar light. We also proposed hypothetical reaction schemes for solar and visible light driven hydrogen generation considering both UV and visible light driven processes on the basis of intermediate analysis. To the best of our knowledge, studies of the above mentioned factors are limited, thus, information obtained from this investigation would be helpful in developing proper design and scaling-up methodologies for solar hydrogen generation photo-reactors. 2. Experimental 2.1. Chemicals All chemicals were of analytical grade and were used without further treatment. Aeroxide TiO2 P25 (80–20% anatase to rutile) (TiO2) from Evonik Degussa Corporation was used as a photocatalyst. Eosin Y dye (99.0%, Sigma–Aldrich Canada Ltd.) was used as a sensitizer for TiO2. Triehanolamine (98.0%) and hydrogen hexachloroplatinate (IV) solution (8 wt%) were also purchased from Sigma–Aldrich Canada Ltd. Ultra-pure water (18.2 M O) was prepared from an in-house EASYPureÒ RODI system (Thermo Scientific, Canada) (Chowdhury et al., 2011, 2012). 2.2. Preparation of dye-sensitized photocatalyst Platinum was loaded on TiO2 catalyst by a solar photodeposition method. TiO2 powder was stirred in an aqueous ethanol solution (ethanol/water = 10/90 by volume) with H2PtCl6 (hexachloroplatinate (IV)) solution, the amount of which corresponded to nominal platinum loadings of 0.05, 0.25, 0.5, 0.75, 1.0, and 1.5 wt%. Then, the solution was irradiated under the solar simulator (at 1 sun) for 3 h. Photoreduction of H2PtCl6 (PtIV) occurred, and highly dispersed platinum particles were deposited on the TiO2 surface. After being filtered and washed with water, the powder was dried at 150 °C for 2 h and milled in a mortar resulting in platinum loaded TiO2 (Pt/TiO2). Eosin Y (EY) sensitization was achieved by mixing EY with Pt/TiO2 in aqueous triethanolamine (TEOA) solution, sonicated and then purged with nitrogen gas. 2.3. Reactor, instruments and analytical procedures Photocatalytic reactions were carried out in a gas-tight Pyrex glass made batch reactor (530 mL) with a flat window at the top for illumination. The photocatalyst was suspended in 100 mL TEOA solution (0.05–0.5 M) after pH adjustment with 1:1 HCl. The photocatalyst suspension was dispersed for 5 min in an ultrasonic bath and then the system was degassed by bubbling ultra-pure nitrogen gas for 40 min. Reaction mixture was stirred (500 rpm) using a magnetic stirrer. The photocatalyst was irradiated with a solar simulator from the top. The light source was equipped with AM 1.5 G as well as a 420 nm cut-off filter (Omega optical, USA) to remove all the UV light. A band pass filter was used to eliminate the visible

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

and NIR light for UV operations only. The lamp and reactor were attached with cooling fans to avoid heating up. The gas sampling port in the reactor was sealed with a silicone rubber septum, and the sampling was made intermittently through the septum during the experiments. Hydrogen was analyzed by Shimazu GC 2014 (TCD, ultra pure N2 as carrier gas and HeyeSep D column). The photocatalysts were characterized for their visible light activity with diffuse reflectance spectra (DRS), and for platinum with transmission electron microscopy (TEM) as well as energy dispersive X-ray (EDX). Diffuse reflection spectra (DRS) were collected using UV-3600 (Shimadzu, Japan) equipped with DR integrated sphere coated with barium sulfate. Barium sulfate is also considered to be of suitable standards by the CIE. The spectra were recorded at room temperature in the range of 200–700 nm. The Merck DIN 5033 barium sulfate powder standard (EM Industries Inc., Hawthorne, NY) has an absolute reflectance of 0.973–0.988 in the 380–750 nm wavelength ranges, and >0.95 in the 750– 1500 nm wavelength range. Energy Dispersive X-ray (EDX) was performed with Hitachi S-4500 field emission SEM with a Quartz XOne EDX system. A transmission electron microscopic (TEM) study was performed with a Hitachi H-7000 electron microscope operating at 100 kV. Charge recombination at EY-sensitized Pt/ TiO2 surface was studied by using photoluminecense (PL) spectroscopy. PL spectra were recorded on a Photon Technology International QuantaMaster™ 50 spectrofluorometer. The excitation wavelength of the light source was 490 nm and emission peaks were observed in the range of 525–540 nm. 3. Results and discussion 3.1. DRS, TEM and EDX characterization of the photocatalyst Diffuse reflectance spectra of Aeroxide TiO2 P25 (TiO2) and EYsensitized Pt/TiO2 are shown in Fig. SM-1 (SM: supplementary material). One can clearly observe that the absorption band of TiO2 was in the UV light range of 200–400 nm, and after platinum loading and EY sensitization the absorption edges shifted towards longer wavelength. A broad spectrum of 450–600 nm can be seen in Fig. SM-1. EY dye mainly absorbs visible light with a maximum absorption at 520 nm and thus can provide visible light activity to the resulting photocatalyst. EDX and TEM analysis revealed the presence of platinum metal in the EY-sensitized Pt/TiO2 catalyst (Fig. SM-2). The TEM image clearly showed the dispersed platinum particles on the TiO2 matrix. This was also confirmed by a visual color change of the catalyst. Pure TiO2 had a milky white color, whereas Pt/TiO2 was gray colored and became darker with increasing platinum content (wt%). Finally, with EY-sensitization, the photocatalyst color turned pink.

3

visible light, EY accepts photon and is excited to higher energy level from where it can inject electron to the conduction band of TiO2. Photogenerated conduction band electron then transfers to H+ ion forming gaseous hydrogen. In solar light, the photocatalytic reaction initiates with the formation of electron/hole (e/h+) pairs. With platinum on EY-sensitized TiO2 photocatalyst, platinum can confine electrons, and hydrogen can be produced on platinum particles (Schiavello, 1997). The amount of hydrogen generation increased with irradiation time during 3 h reaction period. However, the rate of hydrogen generation was highest in the first hour, and then decreased gradually with time due to reverse reaction of gaseous products in the system (Sreethawong and Yoshikawa, 2006). The rate of hydrogen generation is presented in Fig. 1 as a function of platinum loading in both visible and solar light irradiation. The rate of hydrogen generation initially increased as the platinum content increased from 0.05% to 0.25% (wt%), but further increase of platinum did not show any positive effect. Loading of platinum metal beyond the optimum value resulted in lower hydrogen generation rate. The possible reasons are as follows: (i) greater light scattering and lower light absorption due to excessive platinum nano-clusters on TiO2, and (ii) at very high metal loading they act as charge recombination centers (Mills and Hunte, 1997). 3.2.2. Dependence of hydrogen generation on solution pH in visible and solar light Solution pH showed notable effect on hydrogen generation over EY-sensitized Pt/TiO2 as discussed in our earlier study (Chowdhury et al., 2011). Acidic pH was not very helpful in hydrogen generation compared to alkaline pH. According to Li and Lu (2007), EY molecule has carboxylic acid group that can interact with the hydroxyl groups on Pt/TiO2 surface and thereby form an ester like linkage in TEOA solution. Again from zeta potential point of view, TiO2 has zero surface charge within pH 5.6–6.8 (Lu et al., 1993; Chowdhury et al., 2012). Hence, at alkaline pH (pH 10.0), the TiO2 surface is negatively charged, and at acidic pH (pH 4.0), the surface is positively charged. Therefore, the pH value has a significant effect on the adsorption-desorption properties at the EY-sensitized Pt/TiO2 surface. Moreover, the ionization of oxidized form of triethanolamine (TEOA+) in aqueous solution depends on solution pH. At alkaline pH (pH = 9), TEOA+ deprotonates to yield a neutral radical with the unpaired electron in alpha-position to either the amino or alcohol group. Such species are expected to exhibit reduction instead of

3.2. Parametric study EY–TEOA system showed significant performance for dye-sensitized hydrogen generation in our studies. TiO2 and Pt/TiO2 showed no activity for hydrogen generation in visible light irradiation. This was because of the higher band-gap energy of TiO2 and Pt/TiO2. On the contrary, EY-sensitized TiO2 and EY-sensitized Pt/TiO2 showed hydrogen generation activity in visible light irradiation. In solar light irradiation all of them showed hydrogen generation activities. 3.2.1. Dependence of photocatalytic hydrogen evolution on platinum loading over EY-sensitized TiO2 in visible and solar light Photocatalytic hydrogen generation from EY-sensitized Pt/TiO2 with various platinum loadings (0.05–1.5 wt%) was studied in aqueous solution of TEOA, in solar and visible light irradiation. In

Fig. 1. Effect of platinum wt% on photocatalytic hydrogen generation over EYsensitized TiO2 in solar (4) and visible (s) light irradiation. (Experimental conditions: [TEOA] = 0.25 M, EY:Pt/TiO2 = 1:83, I = 100 mW cm2, pH = 7, N2 saturated, pre-sonicated, visible 420 nm cut-off filter).

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

4

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

oxidation properties. In acidic pH, the acid base equilibrium of TEOA+ seems to favor the protonated form of the radical. In neutral solution, both protonated and unprotonated forms coexists. The protonated form is less favorable as an electron donor compared to the unprotonated form, thus unprotonated form of the amine is effective as a reducing agent (Kalyanasundaram et al., 1978). Comparable initial rate of hydrogen generation was found in our experiment for both neutral (pH = 7) and alkaline (pH = 10) pH, although after a while the rate of hydrogen generation dropped considerably for alkaline pH. Lower rate of hydrogen generation at alkaline pH was also reported by Zhang et al (2007). As described by Li and Lu (2007), at strong basic solution, a part of hydroxyl group on photocatalyst surface first reacted with H+ and then left a basic group with negative charge. Because of the electrostatic repulsion force, carboxylic acid groups of EY were difficult to adsorb on the Pt/TiO2 surface, which prevented the electron transfer from excited dye (EY⁄) to the TiO2 conduction band. In this paper, we have performed the entire study at neutral (pH 7.0) solutions in both visible and solar light irradiation. 3.2.3. Dependence of hydrogen generation on EY to Pt/TiO2 mass ratio The concentration of EY has a lead role in improving the number of excited electrons and thereby increasing the photocatalytic activity. Again, the dye adsorption and photocatalytic activities are both related to the active sites (or surface area) of the semiconductor photocatalyst. So it was necessary to investigate the effect of the dye and photocatalyst together. In earlier studies, authors have reported different optimum EY to photocatalyst (EY:T) mass ratio for hydrogen generation in visible light. Li and Lu (2007) reported EY:T mass ratio of 1:1 as optimum for Na2TiO4(OH)2/Pt nanotubes in visible light. Zhang et al. (2008) reported an optimum mass ratio (EY:T) of 1:8 for TS-1 zeolite. At low EY concentration, hydrogen generation rate was much higher in solar light compared to visible light because of UV-assisted hydrogen generation from TEOA. This phenomenon is discussed in detail in Sections 3.3.2 and 3.4. We considered optimum mass ratios (EY:Pt/TiO2) of 1:10 and 1:13.3 in visible and solar light respectively (Fig. 2). 3.2.4. Dependence of hydrogen generation on initial TEOA concentration TEOA played a vital role in dye-sensitized process for hydrogen generation. In absence of TEOA, EY-sensitized Pt/TiO2 was unable to produce any hydrogen. TEOA has three different roles: (i) acts

as electron donor, (ii) extends dye stability, and (iii) acts as a buffer. As an electron donor, it can reduce the EY+ species to return back to its ground state (i.e., EY) and also enhance the stability of EY through the Van der Waals interaction with oxidized dye radicals (Zhang et al., 2007). Moreover, aqueous solution of TEOA had a natural pH of 10.4 and it actually acted as a buffer solution throughout the reaction. Initial concentration of TEOA correlated the hydrogen generation rate by Langmuir-type isotherm as discussed in Section 3.5. 3.3. Discussion of reaction mechanism in UV, visible and solar light irradiation In the photosensitization system, a photochemically excited molecule may donate or accept electron depending on the presence of electron acceptor or electron donor respectively (Larson et al., 1992). In case of dye sensitization, dye serves a dual role: (i) sensitizer for semiconductor photocatalyst, and (ii) molecular bridge between semiconductor photocatalyst and electron donor (Donghua and Jingfei, 2012). So far, most of the studies have discussed applications of dye-sensitized photocatalyst in visible light. We have used EY-sensitized Pt/TiO2 photocatalyst in solar light which includes both UV and visible spectrum. Results showed that hydrogen production increased drastically in solar light with TEOA as electron donor. A probable explanation is that, in solar UV light, band-gap excitation produced e/h+ pairs on TiO2. Positive h+ reacted with water to produce hydroxyl radical (HO) which oxidized TEOA to formaldehyde (HCHO) and ammonia (NH3). HCHO was further oxidized to produce hydrogen (H2) and carbon dioxide (CO2). Blank experiments were performed with only Pt/TiO2 (without EY) separately in visible light and solar light in presence of aqueous solution of TEOA to verify the reaction mechanisms. 3.3.1. Reaction mechanism in visible light irradiation In visible light (k P 420 nm), the energy of photon is lower than the band-gap of Pt/TiO2 but higher than the band-gap of dye molecule (EY). So the EY molecules adsorbed on the surface of photocatalyst were excited with visible photons and then injected electron to the conduction band of TiO2. After the electron injection, EY was converted to its oxidized form (EY+) and the electron reduced H+ ion to H2 on the platinum site over the TiO2. The proposed mechanism is shown below: Scheme 1 (Fig. SM-3) (Chatterjee, 2010; Chowdhury et al., 2013a): Visible photon: hm;k6520 nm

TiO2  ðEYÞs ! TiO2  ðEY Þs

ð1Þ

TiO2  ðEY Þs ! TiO2  ðEYþ j þ eCB Þs

ð2Þ

TiO2  ðEYþ þ eCB Þs þ Pt ! TiO2  ðEYþ Þs þ PtðeCB Þ

ð3Þ

TiO2  ðEYþ Þs þ TEOA ! TiO2  ðEYÞs þ TEOAþ

ð4Þ

1 PtðeCB Þ þ H2 O ! Pt þ H2 þ HO 2

ð5Þ

(Considering a maximum of 520 nm wavelength that can excite EY).

Fig. 2. Effect of EY to Pt/TiO2 mass ratio on hydrogen generation rate in solar (4) and visible (s) light irradiation. (Experimental conditions: [TEOA] = 0.25 M, Pt/TiO2 (0.25%) = 1 g L1, I = 100 mW cm2, pH = 7, N2 saturated, pre-sonicated, visible 420 nm cut-off filter).

3.3.2. Reaction mechanism in solar light irradiation Solar light irradiation includes both UV and visible spectrum. UV light can excite TiO2 and produce e/h+ pair by conventional band-gap excitation process. At the same time EY dye is also sensitized in solar light. TEOA oxidation occurred in presence of h+ and/or HO as evident from the formation of formaldehyde as an intermediate. Oxidation of TEOA can be achieved in presence of

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

5

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

several oxidants such as periodate, hypochlorous acid, chlorine dioxide, and hexacyanoferrate (III). But in those cases further oxidation of formaldehyde was not reported or the oxidation of the formaldehyde was extremely slow and, therefore, formaldehyde was reported as the end product (Shukla et al., 1973). Scheme 2 (Fig. 3) (Chowdhury et al., 2013a; Chowdhury et al., 2013b): Visible photon: hm;k6520 nm

TiO2  ðEYÞs ! TiO2  ðEY Þs

ð6Þ

TiO2  ðEY Þs ! TiO2  ðEYþ þ eCB Þs

ð7Þ

TiO2  ðEYþ þ eCB Þs þ Pt ! TiO2  ðEYþ Þs þ PtðeCB Þ

ð8Þ

TiO2  ðEYþ Þs þ TEOA ! TiO2  ðEYÞs þ TEOAþ

ð9Þ

1 PtðeCB Þ þ H2 O ! Pt þ H2 þ HO 2

ð10Þ

(Considering a maximum of 520 nm wavelength that can excite EY). UV photon: hm;k6388 nm

þ

TiO2  ðEYÞs ! TiO2 ðe þ h Þ  ðEYÞs

ð11Þ

þ

ð12Þ

TiO2 ðh Þ þ H2 O ! HO þ Hþ

þ

ð13Þ

TiO2  ðEYþ þ eCB Þs þ TEOA ! TiO2  ðEY þ eCB Þs þ TEOAþ

ð14Þ

TEOA þ HO ! HCHO þ NH3

ð15Þ

HCHO þ HO !! H2

ð16Þ

TiO2 ðh Þ þ HO ! HO

In our case we noticed further oxidation of formaldehyde to hydrogen in solar light. This fact was confirmed by increased hydrogen generation and decreased formaldehyde concentration compared to visible light. This also confirmed much higher oxidizing capability of HO compared to the above mentioned oxidants.

and triethanolamine) in water as a sweetening mixture of hydrogen sulfide. Trace amounts of ethylene glycol was formed in the system, but further oxidation of ethylene glycol was not reported (Eq. (18)). photocatalyst=hm

NðCH2  CH2  OHÞ3 þ 3H2 O ! 3HO  CH2  CH2  OH þ NH3

ð18Þ

In another study, the photodegradtion of ethanolamine did not report any ethylene glycol formation, but formaldehyde formation was mentioned as a major intermediate (Yin et al., 2007). In our case, formaldehyde was detected and quantified as an intermediate in both visible and solar light driven reactions using both semiquantitative and quantitative methods. In case of visible light, only dye sensitization mechanism was possible. So TEOA was oxidized by EY+ species. However, in presence of solar light, TEOA oxidation was also initiated by the valence band hole (h+). This was evident from the fact that, under similar reaction conditions visible light accumulated more formaldehyde than solar light. Fig. SM-4 shows the plot of EY to Pt/TiO2 mass ratio versus intermediate (formaldehyde) accumulation in solar and visible lights. With a higher dye to photocatalyst ratio, the photocatalytic hydrogen generation rate increased. At the same time, formaldehyde accumulation also increased. In solar light, formaldehyde molecule was further oxidized to hydrogen which was not possible in visible light. Figs. 4 and SM-5 show H2 generation and formaldehyde accumulation at different EY to Pt/TiO2 ratio in visible and solar lights, respectively. In visible light, hydrogen was produced via dye-sensitization pathway and the rate of hydrogen generation increased gradually with increase in EY concentration and reached a plateau after EY:Pt/TiO2 mass ratio of 1:10 (Fig. 2). Conversely, in solar light, hydrogen was produced via both dye-sensitization and band gap excitation pathways. TEOA was oxidized by valence band hole and finally produced hydrogen with the help of solar UV radiation. So TEOA solution was consumed at a much faster rate in solar light. As the EY concentration increased, there was less TEOA available for EY+ regeneration and thus dye-sensitized hydrogen generation rate dropped (in solar light). From Fig. 2 it can be observed that at EY:Pt/TiO2 mass ratio of 1:10, the hydrogen generation rate in solar light was lower than that of visible light.

3.4. Photocatalytic oxidation of trietanolamine 3.5. Hydrogen generation kinetics Shukla et al. (1973) reported the oxidation kinetics of TEOA with alkaline hexacyanoferrate (III) in aqueous media. The oxidation reaction is shown below (Eq. (17)):

TEOA reacted with the photogenerated h+ and/or HO and degraded to different compounds. Thus, the concentration of TEOA

NðCH2  CH2  OHÞ3 þ 6HO þ 6FeðCNÞ3 6 ¼ 6HCHO þ NH3 þ 3H2 O þ 6FeðCNÞ4 6

ð17Þ

Naman and Gratzel (1994) studied the colloidal suspension of vanadium sulfide with different percentage of ethanolamine (mono-, di-,

Fig. 3. Hydrogen generation path in solar light.

Fig. 4. Hydrogen generation () and formaldehyde accumulation (r) in visible light irradiation. (Experimental conditions: Pt/TiO2 (0.25%) = 1 g L1, [TEOA] = 0.25 M, Ivis = 100 mW cm2, pH = 7, N2 saturated, pre-sonicated, visible 420 nm cut-off filter).

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

6

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

continuously changed throughout the reaction. In order to determine the effect of initial concentration of TEOA on hydrogen generation, we selected 60 min time interval, as the change of reactant or product can be well determined during this time interval. Fig. 5 describes the effect of TEOA concentration on the rate of hydrogen generation. A significant improvement in hydrogen generation was observed as the concentration of TEOA increased from 0.05 to 0.25 M. However, the hydrogen generation rate was found to be almost independent after 0.25 M of TEOA concentration. Hence, the variation of hydrogen generation rate as a function of TEOA concentration could be best represented according to Langmuir-type expression as described below (Li et al., 2006):

r¼

dCH2 kKCo ¼ dt 1 þ KCo

ð19Þ

where r is the initial rate of hydrogen generation, k, the reaction rate constant, and K the adsorption constant of triethanolamine on to EY-sensitized Pt/TiO2 photocatalyst. According to the Langmuir-type plot, k = 6.77  106 mol min1 and K = 14.45 M1 was obtained.

Attempts were made to model the solar hydrogen generation using a linear function combining UV and visible rate constant data of the form:

kðIsolar Þ ¼ bkðIUV Þ þ ð1  bÞkðIv is Þ

ð20Þ

h i h i kðIsolar Þ ¼ b 9:31  104 I1:66 þ ð1  bÞ 9:79  102 I0:8 v is UV

ð21Þ

where b is the fraction of UV light contributed to solar hydrogen generation. This however was not successful predicting the hydrogen generation in solar light. Instead, a better prediction of the latter was obtained using a simple power law model of the form:

kðIsolar Þ ¼ 0:3017I0:6 solar

ð22Þ +

+

In both cases TEOA oxidation occurred with either EY or h . Therefore in solar light there was always a competition for TEOA. Again at higher light intensities, the process undergoes plenty of charge recombination which was confirmed through photoluminescence (PL) study. The main reasons behind the deviation from Eq. (21) were different excitation mechanisms in UV and visible lights and charge recombination at higher light intensities.

3.6. Dependence of hydrogen generation on light intensity

3.7. Photoluminescence (PL) study under UV, visible and solar light

The incident light intensity is expected to be one of the rate controlling parameters. In order to illustrate this effect, experiments were performed under four levels of light intensity and the hydrogen generation rates were compared. To achieve pure UV light from solar simulator, we used a band pass filter placed above the reactor with an external attachment. To receive visible light we placed an UV cut-off filter inside the solar simulator. We have studied the effect of UV, visible, and solar light at different intensities such as 30 mW cm2, 50 mW cm2, 70 mW cm-2, and 100 mW cm2 (Fig. 6). The reaction rate constant k, typically followed power-law dependence on light intensities (Ray and Beenackers, 1997). The hydrogen generation rate constants were evaluated as a function of UV, visible, and solar light intensities (IUV, Ivis, and Isolar), keeping all other parameters fixed. In UV and visible radiation, the data fitted well with the power law model (k(I) = aIb) compared to that of solar light. The constants and R2 values are presented in Table SM-1. Solar light contains both UV and visible spectrum, so it may be assumed that the hydrogen generation rate in solar light could be predicted from UV and visible light assisted hydrogen generation.

Photoluminescence (PL) technique is an effective methodology to assess the electron transfer performance in photocatalyst. We have studied the PL of EY-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation (Fig. SM-6). The emission peaks could be attributed to the e/EY+ or e/h+ recombination processes. UV and solar lights showed lower PL intensities (6  105–7  105 s1) compared to visible light (13  105 s1). Therefore, higher PL value for visible light indicated a higher e/EY+ recombination process. Again, the charge recombination process also increased with increasing light intensity for all three kinds of light sources.

Fig. 5. Rate of hydrogen generation as a function of initial concentration of TEOA. Plot shows experimental (N) and Langmuir type ( ) model. (Experimental conditions: EY:Pt/TiO2 = 1:10, Ivis = 100 mW cm2, pH = 7, N2 saturated, presonicated, visible 420 nm cut-off filter).

3.8. Apparent quantum yield in UV, visible, and solar light Fig. SM-1 shows the DRS of pure TiO2 and EY-sensitized Pt/TiO2 photocatalyst. Compared to pure TiO2, the EY-sensitized photocatalyst showed good visible light absorption which was also confirmed by its hydrogen generation capability in visible light. Apparent quantum yields were evaluated for hydrogen generation over EY-sensitized Pt/TiO2 according to the formulae shown below

Fig. 6. Dependence of hydrogen generation on light intensity of UV (j), visible ( ) and solar ( ) light. (Experimental conditions: [TEOA] = 0.25 M, I = 100 mW cm2, EY:Pt/TiO2 = 1:20, pH = 7, N2 saturated, pre-sonicated, visible 420 nm cut-off filter, UV – band pass filter).

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

7

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx Table 1 Apparent quantum yield in UV, visible and solar light. Light source

Apparent quantum yield (%)

Wavelength (nm) UV Visible Visible Solar Solar

a

300–388 420–650b 420–520c 300–650d 300–520e

30 mW cm2

50 mW cm2

70 mW cm2

100 mW cm2

10.82 ± 0.271 5.16 ± 0.189 12.23 ± 0.201 5.81 ± 0.234 11.33 ± 0.213

10.65 ± 0.348 3.88 ± 0.208 9.38 ± 0.120 5.23 ± 0.331 10.39 ± 0.407

10.42 ± 0.287 3.26 ± 0.131 8.23 ± 0.201 3.32 ± 0.200 6.69 ± 0.321

9.58 ± 0.226 2.39 ± 0.095 5.73 ± 0.462 2.15 ± 0.196 4.19 ± 0.197

As EY could utilize photons up to a maximum wavelength of 520 nm, we used 300–520 nm wavelength range (a, c, e) for QY calculation. However, we have reported QY values for 300–650 nm wavelength range (b, d) for comparison only.

 Recombination of e/h+ or e/EY+ were determined through PL study, which illustrated the higher recombination rates in case of visible light compared to that of solar and UV light at 1 sun.

(Eq. (23)) (Shimidzu et al., 1985). The quantum yield (/) would certainly be higher than the apparent quantum yield, as the adsorbed photons were a certain fraction of the incident photons. Incident light intensities were measured with StelerNET instrument. Table 1 shows the apparent quantum yield values at different light intensities for UV, visible, and solar lights. At lower intensity, the apparent quantum yields were much higher due to a reduced e/h+ recombination rate. Again at lower light intensity (30 mW cm2), we achieved higher apparent quantum yields in solar and visible light (11.33–12.23%) than that of in UV light (10.82%), by considering 300–520 nm range of wavelengths. Our apparent quantum yield values were either comparable (10% at 520 nm (Abe et al., 2000), 10.4%, k P 420 nm (Zhang et al., 2007), and 12.14%, k P 420 nm (Li et al., 2007)) or better (5.1%, k P 420 nm (Jin et al., 2007), 7.1%, k P 420 nm (Jin et al., 2006), and 9.4%, k P 420 nm (Zhang et al., 2008)) than previously reported values from EY-sensitized hydrogen generation studies.

The authors would like to thank Natural Science and Engineering Research Council of Canada and Western Engineering for their financial support.

/ > apparent quamtum yield; /H2 ð%Þ

References

Acknowledgement

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.10.076.

þ

ðNumber of reacted e or h Þ  100 ¼ ðNumber of incident photonsÞ ¼

ðNumber of H2 molecule evolvedÞ  2  100 Number of incident photons

ð23Þ

4. Conclusions From the results presented in this article, the following could be inferred:  DRS supported the visible light activity of EY-sensitized Pt/TiO2. TEM and EDX results reported the presence of platinum metal on TiO2 surface.  Different parameters showed significant effects on hydrogen generation. Neutral pH, 0.25 wt% of Pt on TiO2 was found to be optimum levels for sacrificial hydrogen generation in solar and visible light.  EY to Pt/TiO2 mass ratio was shown to be a crucial parameter for hydrogen generation. We obtained an optimum mass ratio (EY:TiO2/Pt) of 1:10 and 1:13.3 in visible and solar light respectively.  Concentration of TEOA also played a major role during sacrificial hydrogen generation. Hydrogen generation rate varied as a function of TEOA concentration and it followed Langmuir-type isotherm.  The reaction mechanisms in solar and visible lights were different, although in both cases formaldehyde was detected as an intermediate product. However, in solar light, formaldehyde was oxidized by h+/HO to produce hydrogen.  Light intensity was also found to be an important parameter in photocatalytic hydrogen generation. At lower light intensities, the system had higher quantum yields compared to higher light intensities. At 30 mW cm2 QYs for UV, visible and solar light were 10.82%, 12.23%, and 11.33% respectively.

Abe, R., Hara, K., Sayama, K., Domen, K., Arakawa, H., 2000. Steady hydrogen evolution from water on eosin Y-fixed TiO2 photocatalyst using a silanecoupling reagent under visible light irradiation. J. Photochem. Photobiol. A 137, 63–69. Abe, R., Sayama, K., Sugihara, H., 2005. Effect of water/acetonitrile ratio on dyesensitized photocatalytic H2 evolution under visible light irradiation. J. Sol. Energy Eng. 127, 413–416. Bulatov, A.V., Khidekel, M.L., 1976. Decomposition of water under the effect of UV irradiation and in the presence of platinized titanium dioxide. Izv. Akad. Nauk. SSSR. Ser. Khim., 1902–1903. Chan, J.R., Pan, B.Y., Yen, T.F., 1992. Treatment of wastewater containing citric acid and triethanolamine. U.S. Patent No. 5,139,679. Chatterjee, D., 2010. Effect of excited state redox properties of dye sensitizers on hydrogen production through photo-splitting of water over TiO2 photocatalyst. Catal. Commun. 11, 336–339. Chowdhury, P., Gomaa, H., Ray, A.K., 2011. Factorial design analysis for dyesensitized hydrogen generation from water. Int. J. Hydrogen Energy 36, 13442– 13451. Chowdhury, P., Moreira, J., Gomaa, H., Ray, A.K., 2012. Visible solar light driven photocatalytic degradation of phenol with dye-sensitized TiO2: parametric and kinetic study. Ind. Eng. Chem. Res. 51, 4523–4532. Chowdhury, P., Gomaa, H., Ray, A.K., 2013a. Dye-sensitized photocatalyst: a breakthrough in green energy and environmental detoxification. Sustainable Nanotechnology and the Environment: Advances and Achievements, vol. 1124. ACS Symposium Series, Washington, DC, pp. 231–266. Chowdhury, P., Malekshoar, G., Ray, M.B., Zhu, J., Ray, A.K., 2013b. Sacrificial hydrogen generation from formaldehyde with Pt/TiO2 photocatalyst in solar radiation. Ind. Eng. Chem. Res. 52, 5023–5029. Domen, K., Kudo, A., Onishi, T., 1986. Mechanism of photocatalytic decomposition of water into H2 and O2 over NiO–SrTiO3. J. Catal. 102, 92–98. Donghua, P., Jingfei, L., 2012. Development of visible light-responsive sensitized photocatalysts. Int. J. Photoenergy 2012, 1–13. http://dx.doi.org/10.1155/2012/ 262831. Article ID 262831. Fujishima, A., Honda, K., 1972. Electrochemical photolysis of water at a semiconductor electrode. Nature 238, 37–38. Jin, Z., Zhang, X., Li, Y., Li, S., Lu, G., 2007. 5.1% Apparent quantum efficiency for stable hydrogen generation over eosin – sensitized CuO/TiO2 photocatalyst under visible light irradiation. Catal. Commun. 8, 1267–1273. Jin, Z., Zhang, X., Lu, G., Li, S., 2006. Improved quantum yield for photocatalytic hydrogen generation under visible light irradiation over eosin sensitized TiO2 – investigation of different noble metal loading. J. Mol. Catal. A: Chem. 259, 275– 280. Kudo, A., Kato, H., 1997. Photocatalytic decomposition of water into H2 and O2 over novel photocatalyst K3Ta3Si2O13 with pillared structure consisting of three TaO6 chains. Chem. Lett. 26, 867–868.

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076

8

P. Chowdhury et al. / Chemosphere xxx (2014) xxx–xxx

Kang, S.Z., Chen, L., Li, X., Mu, J., 2012. Composite photocatalyst containing eosin Y and multi-walled carbon nanotubes loaded with CuO/NiO: mixed metal oxide as an active center of H2 evolution from water. Appl. Surf. Sci. 258, 6029–6033. Kalyanasundaram, K., Kiwi, J., Gratzel, M., 1978. Hydrogen evolution from water by visible light, a homogeneous three component test system for redox catalysis. Helv. Chim. Acta 61, 2720–2730. Larson, R.A., Stackhouse, P.L., Crowley, T.O., 1992. Riboflavin tetra-acetate: a potentially useful photosensitizing agent for the treatment of contaminated waters. Environ. Sci. Technol. 26, 1792–1798. Lee, J.S., 2005. Photocatalytic water splitting under visible light with particulate semiconductor catalysts. Catal. Surv. Asia 9, 217–227. Lehn, J.M., Sauvage, J.P., Ziessel, R., Hilaire, L., 1982. Water photocatalysis by UV irradiation of rhodium loaded strontium-titanate catalysts-relation between catalytic activity and nature of the deposit from combined photolysis and ESCA studies. Isr. J. Chem. 22, 168–172. Li, Y., Xie, Y., Peng, S., Lu, G., Li, S., 2006. Photocatalytic hydrogen generation in the presence of chloroacetic acids over Pt/TiO2. Chemosphere 63, 1312–1318. Li, Q., Chen, L., Lu, G., 2007. Visible-light-induced photocatalytic hydrogen generation on dye-sensitized multi-walled carbon nanotube/Pt catalyst. J. Phys. Chem. C 111, 11494–11499. Li, Q., Lu, G., 2007. Visible-light driven photocatalytic hydrogen generation on eosin Y-sensitized Pt-loaded nanotube Na2Ti2O4(OH)2. J. Mol. Catal. A: Chem. 266, 75–79. Li, Y., Xie, C., Peng, S., Lu, G., Li, S., 2008. Eosin Y-sensitized nitrogen-doped TiO2 for efficient visible light photocatalytic hydrogen evolution. J. Mol. Catal. A: Chem. 282, 117–123. Lu, M.C., Roam, G.D., Chen, J.N., Huang, C.P., 1993. Factors affecting the photocatalytic degradation of dichlorvos over titanium dioxide supported on glass. J. Photochem. Photobiol. A 76, 103–110. Mills, A., Hunte, S.L., 1997. An overview of semiconductor photocatalysis. J. Photochem. Photobiol. A 108, 1–35. Naman, S.A., Gratzel, M., 1994. Visible-light generation of hydrogen from hydrogen sulphide in aqueous solutions of ethanolamines containing vanadium sulphide dispersions. J. Photochem. Photobiol. A 77, 249–253. Ni, M., Leung, M.K.H., Leung, D.Y.C., Sumathy, K., 2007. A review and recent developments in photocatalytic water-splitting using TiO2 for hydrogen production. Renew. Sust. Energy Rev. 11, 401–425.

Ray, A.K., Beenackers, A.A.C.M., 1997. Novel swirl-flow reactor for kinetic studies of semiconductor photocatalysis. AIChE J. 43, 2571–2578. Sato, S., White, J.M., 1980. Photodecomposition of water over Pt/TiO2 catalysts. Chem. Phys. Lett. 72, 83–86. Schiavello, M., 1997. Heterogeneous Photocatalysis. Wiley Series in Photoscience and Photoengineering, vol. 3. John Wiley & Sons Ltd., Chichester, UK. pp 87–107. Sreethawong, T., Junbua, C., Chavadej, S., 2009. Photocatalytic H2 production from water splitting under visible light irradiation using eosin Y-sensitized mesoporous-assembled Pt/TiO2 nanocrystal photocatalyst. J. Power Sources 190, 513–524. Sreethawong, T., Yoshikawa, S., 2006. Enhanced photocatalytic hydrogen evolution over Pt supported on mesoporous TiO2 prepared by single-step sol–gel process with surfactant template. Int. J. Hydrogen Energy 31, 786–796. Shimidzu, T., Iyoda, T., Koide, Y., 1985. An advanced visible-light-induced water reduction with dye-sensitized semiconductor powder catalyst. J. Am. Chem. Soc. 107, 35–41. Shukla, K.S., Mathur, P.C., Bansal, O.P., 1973. Oxidation kinetics of triethanolamine by alkaline hexacyanoferrate (III). J. Inorg. Nucl. Chem. 35, 1301–1307. Wagner, F.T., Somorjai, G.A., 1980. Photocatalytic hydrogen production from water on Pt-free SrTiO3 in alkali hydroxide solutions. Nature (London, United Kingdom) 285, 559–560. Wang, Z.S., Sayama, K., Sugihara, H., 2005. Efficient eosin Y dye-sensitized solar cell containing Br/Br3 electrolyte. J. Phys. Chem. B 109, 22449–22455. West, R.J., Gonsior, S.J., 1996. Biodegradation of triethanolamine. Environ. Toxicol. Chem. 15, 472–480. Wrighton, M.S., Ellis, A.B., Wolczanski, P.T., Morse, D.L., Abrahamson, H.B., Ginley, D.S., 1976. Strontium titanate photoelectrodes. Efficient photoassisted electrolysis of water at zero applied potential. J. Am. Chem. Soc. 98, 2774–2779. Yin, Z.H., Li, Y.X., Peng, S.Q., Lu, G.X., Li, S.B., 2007. Photocatalytic hydrogen generation in the presence of ethanolamines over Pt/TiO2. J. Mol. Catal. (China) 21, 154–161. Zhang, X., Jin, Z., Li, Y., Li, S., Lu, G., 2007. Visible-light-induced hydrogen production over Pt-eosin Y catalysts with high surface area silica gel as matrix. J. Power Sources 166, 74–79. Zhang, X., Jin, Z., Li, Y., Li, S., Lu, G., 2008. Photocatalytic hydrogen generation over eosin Y-sensitized TS-1 zeolite. Appl. Surf. Sci. 254, 4452–4456.

Please cite this article in press as: Chowdhury, P., et al. Sacrificial hydrogen generation from aqueous triethanolamine with Eosin Y-sensitized Pt/TiO2 photocatalyst in UV, visible and solar light irradiation. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.076