Although nitrogen-doping has been demonstrated to provide visible-light photocatalytic activity to TiO 2 by introducing isolated anion dopant states
Figure 2.6 (a) Optical absorbance (in term of Kubelka–Munk equivalent absorbance units) of nitrogen-doped titanium oxide (TiON) powders obtained by calcinating xerogels in air at 500 o C for 3 hours with various initial tetramethylammonium hydroxide (TMA)/titanium tetraisopropoxide (TTIP) ratios at 1:3 (blue line), 1:5 (red line), and 1:10 (black line), compared with the optical absorbance of Degussa P25 powder (brown line). (b) Tauc Plot constructed from (a). Band-gap values are determined from the extrapolation of the linear Tauc Region line to the photon energy abscissa. Adapted from [ 54 ].
(a) (b)
Savage_Ch02.indd 24
Savage_Ch02.indd 24 11/11/2008 2:14:23 PM11/11/2008 2:14:23 PM
within the band gap of TiO 2 , the intraband states also create the serious problem of massive charge carrier recombination, thus largely limiting their photoactivity. It is generally believed that metal ion dopants may infl uence the light absorption and photoreactivity of TiO 2 most signifi cantly by acting as electron (or hole) traps [ 18– 21 , 57 ]. The trapping of charge carriers can decrease the e – /h + pair recombination rate and subsequently increase the lifetime of charge carriers, which is benefi cial to improving the photoreactivity of TiO 2 . Thus, metal ions were added to TiON to reduce the charge carrier recombination and the photoactivity of TiON [ 58 ].
Palladium was added to TiON to form palladium-modifi ed N-doped TiO 2 (TiON/PdO) nanoparticle photocatalysts. The concentration of palladium was controlled by the precursor ratio in a sol–gel process. The sol–gel process uses the optimized processing parameters for TiON synthesis [ 54 ], only adding palladium acetylacetonate (using CH 2 Cl 2 as the solvent) as the palladium source. Figure 2.7(a) shows a typical XRD pattern of TiON/PdO powders.
Major XRD peaks belong to anatase with no rutile phase observed. Besides
Figure 2.7 (a) Typical X-ray diff raction patterns of nitrogen-doped titanium oxide (TiON)/palladium-modifi ed (PdO) powders. (b) High-resolution X-ray photoelectron spectroscopy scan over Pd 3d peaks of TiON/PdO nanoparticle powders. (c) TEM image of TiON/PdO nanoparticle powders. Adapted from [ 58 ].
(c)
(a) (b)
Savage_Ch02.indd 25
Savage_Ch02.indd 25 11/11/2008 2:14:24 PM11/11/2008 2:14:24 PM
the anatase-type structure, a peak assigned to PdO (101) can be clearly identifi ed, suggesting that palladium additive exists as PdO and is not incorporated into the anatase structure. High-resolution XPS scan over Pd 3d ( Fig. 2.7(b) ) peaks shows that the binding energy of Pd 3d 5/2 is approximately 336.20 eV, which can be attributed to PdO species. This observation is in accordance with our XRD experiment result. The Brunauer-Emmett-Teller (BET) surface specifi c areas of these powders are approximately 50 m 2 /g, corresponding to an average particle diameter approximately 30 nm. Figure 2.7(c) shows the TEM image of a TiON/PdO nanoparticle sample. The sample was composed of nanosized particles with nonuniform shapes. TEM revealed a particle size very similar to those obtained from the BET measurements.
The optical properties of these TiON/PdO nanoparticle powders were investigated by measuring the diff use refl ectance spectra. From the refl ectance data, optical absorbance can be approximated by the Kubelka–Munk function [ 55 ]. Figure 2.8(a) shows the light absorbance of TiON/PdO nanoparticle powders, compared with TiON nanoparticle powders with similar nitrogen dopant concentration (N/Ti atomic ratio at approximately 0.06).
TiON powders show a clear shift into the visible-light range (> 400 nm) as expected, which can be attributed to the nitrogen doping. TiON/PdO powders show much higher visible-light absorption than TiON powders. The visible- light absorbance increased with the increase of the Pd ion concentration, which suggests that Pd additive promotes visible-light absorption in TiON nanoparticle photocatalysts. Figure 2.8(b) shows the Tauc Plot [ 55 ] constructed from Fig. 2.8(a) to determine the semiconductor band gap. TiON has a band gap at approximately 2.87 eV, which is consistent with its visible-light absorbance
Figure 2.8 (a) Optical absorbance (in term of Kubelka–Munk equivalent absorbance units) of nitrogen-doped titanium oxide (TiON)/palladium-modifi ed (PdO) powders, compared with the optical absorbance of TiON powder. (b) Tauc Plot constructed from (a). Band-gap values are determined from the extrapolation of the linear Tauc Region line to the photon energy abscissa. [Note: black line for TiON, red line for TiON/
PdO (0.1 percent), blue line for TiON/PdO (0.5 percent), brown line for TiON/PdO (1.0 percent), and green line for TiON/PdO (2.0 percent)] Adapted from [ 58 ].
(a) (b)
Savage_Ch02.indd 26
Savage_Ch02.indd 26 11/11/2008 2:14:25 PM11/11/2008 2:14:25 PM
ability. TiON/PdO powders show a much smaller band gap, which explains their better visible-light absorbance ability. With the increase of the Pd content from 0.1 to 2.0 percent, the band gap of TiON/PdO powders decreases from approximately 1.76 to 1.54 eV.
To understand the mechanism of metal ion modifi cation in this novel material system, TiON/PdO thin fi lm samples with similar structure, composition, and light absorption properties to TiON/PdO nanoparticle were synthesized by ion-beam-assisted deposition (IBAD) [ 59 ]. The fi lm samples were used in the mechanism study because they are easier to be rid of surface contamination and to be analyzed by various material characterization tools than nanoparticle powders. In the TiON/PdO nanoparticle photocatalyst, palladium additions exist as PdO species and are not incorporated into the anatase structure. In situ XPS study revealed changes in the valence states of the palladium species when illuminated. In the in situ experiment, the sample was fi rst placed in the dark for 3 hours right before the XPS high-resolution scan over Pd 3d peak in the dark. Then, XPS scans were performed with a visible-light illumination on the sample simultaneously after 0.5 hour, 1 hour, and 2 hours. Diff erences in the Pd 3d peak shape and position were observed between scans in the dark and under visible-light illumination. For comparison purpose, the same in situ XPS study was also conducted on the TiO 2 /PdO sample. Figure 2.9 shows the comparison of XPS high-resolution scans over Pd 3d peaks in the dark and after 1-hour visible-light illumination, respectively. In the dark, the binding energy of Pd 3d 5/2 is approximately 336.20 eV, indicating that Pd dopant exists as PdO. Under visible-light illumination, however, Pd peaks are broadened and the binding energy of Pd 3d 5/2 shifts to approximately 335.30 eV.
The broad Pd 3d 5/2 is best fi t as a combination of Pd 2+ 3d 5/2 (peak at 336.2 eV) and Pd 0 3d 5/2 (peak at 335.2 eV), which clearly demonstrates that a portion of PdO semiconductor nanoparticles in TiON/PdO thin fi lm is reduced to metallic
Figure 2.9 In situ X-ray photoelectron spectroscopy high-resolution scan over Pd 3d peaks on nitrogen-doped titanium oxide (TiON)/TiON thin fi lm in dark (green line) and after 1-hour visible-light illumination (black line). Note that the red dashed curve fi ts Pd 0 3d peaks, whereas the blue dashed curve fi ts Pd 2+ 3d peaks after 1-h visible-light illumination. Adapted from [ 59 ].
Savage_Ch02.indd 27
Savage_Ch02.indd 27 11/11/2008 2:14:26 PM11/11/2008 2:14:26 PM
Pd nanoparticles under visible-light illumination. The change in the valence state of Pd is consistent with the notion that metal-ion dopants in TiO 2 can act as electron traps to alter the electron–hole pair recombination rate [ 18– 21 , 57 ].
The metallic nanoparticles can also create the surface plasma resonance originated from collective oscillation of free electrons to enhance the visible- light absorption of TiON/PdO as observed in Fig. 2.8 . In contrast, TiO 2 /PdO sample shows no change in Pd 3d peak shape and position under the same visible-light illumination condition because TiO 2 thin fi lm absorbs mainly UV light. Thus, no signifi cant interaction should be expected. The binding energy of Pd 3d 5/2 in TiO 2 /PdO sample is approximately 336.20 eV, indicating that Pd dopant exists as PdO in the dark or under visible-light illumination.
The photocatalytic activities of TiON/PdO nanoparticle photocatalysts were characterized by photocatalytic degradation of humic acid (HA) under visible-light illumination, with the TiON nanoparticle photocatalyst as a comparison basis. Photocatalytic degradation of HA was conducted by exposing the HA solution to various photocatalysts under visible light ( l > 400 nm) for varying time intervals. After the centrifugation to recover photocatalysts, the light absorption of the clear solution was measured and the remaining percentage of HA in the solution was calculated by the ratio between the light absorptions of photocatalyst-treated and untreated HA solutions. Figure 2.10(a) shows representative light absorption spectra of HA solutions with diff erent photocatalyst treatment times. As the illumination time increased, light absorption decreased steadily for HA solutions treated by TiON/PdO photocatalysts at their l max (approximately 280 nm), indicating more and more HA was degraded under the photocatalyst treatment. Figure 2.10(b) summarizes
Figure 2.10 (a) Typical absorption spectra of humic acid (HA) solution during photodegradation under visible-light illumination. (Note that the black line stands for 0 hours, red line for 2 hours, blue line for 5 hours, and green line for 10 hours.) (b) HA residue percentage versus visible-light (> 400 nm) illumination time with nitrogen-doped titanium oxide (TiON) (), TiON/0.1 percent PdO (), TiON/0.5 percent PdO (), TiON/1.0 percent PdO (), and TiON/2.0 percent PdO (), respectively. (Note that the lines are fi tted into the fi rst-order exponential formula.) Adapted from [ 58 ].
(a) (b)
Savage_Ch02.indd 28
Savage_Ch02.indd 28 11/11/2008 2:14:27 PM11/11/2008 2:14:27 PM
the residual percentage of HA as a function of visible-light illumination time for TiON/PdO or TiON photocatalysts in a semi-logarithmic plot format. A fi rst-order exponential decay of HA was observed, which can be fi tted into Equation 2.2:
RP = ae–bt, (2.2)
where RP is the residual percentage of HA, t is the visible-light illumination time, a and b are the fi rst-order exponential fi tting constants. Constant b , the decay rate, can be used as a parameter to compare the photodegradation effi ciency of diff erent photocatalysts. An enhanced photodegradation effi ciency was observed within a certain range of palladium additive concentration. With the addition of palladium ion, the decay rate constant b increases from 0.012 for TiON to 0.018 for TiON/PdO containing 0.5 percent PdO, representing an approximately 50 percent increase. When the palladium dopant concentration was further increased, an inverse eff ect was observed that the decay rate constant b decreased to 0.009 for TiON/PdO with 2.0 percent PdO, an approximately 25 percent decrease when compared to the TiON with no PdO additives.
It is widely reported that transition-metal-ion additives play a complex role in aff ecting the photocatalytic activity of TiO 2 [ 18– 21 , 57 ]. Besides acting as charge trapping center, they can also serve as charge recombination center and aff ect the charge release and migration. In the work by Choi et al. [ 19 ], the ability of a metal-ion additive to function as an eff ective trap in TiO 2 was dependent on its concentration, as found here for the metal-ion-modifi ed TiON.
Thus, although stronger visible-light absorbance may be achieved by increasing the Pd additive concentration, there is no simple relationship between visible- light absorbance and photocatalytic effi ciency. Other metal-ion-modifi cations of TiON, such as Ag + , Nd 3+ , Cu 2+ , Y 3+ , Ce 4+ , and Fe 3+ , had also been explored.
They demonstrated similar enhancements on the photoactivity of TiON under visible-light illumination, whereas diff erent performances were observed between diff erent species. To achieve high photocatalytic effi ciency in this transition-metal-ion-modifi ed TiON photocatalysts, a careful optimization of the transition-metal-ion-additive concentration is needed.