Nitrogen Doped 3D Titanium

Nitrogen Doped 3D Titanium Dioxide Nanorods Architecture withSignificantly Enhanced Visible Light PhotoactivityZhaodong Li,† Fei Wang,† Alexander Kvit,†,‡ and Xudong Wang*,†

†Department of Materials Science and Engineering, University of WisconsinMadison, Madison, Wisconsin 53706, United States‡Materials Science Center, University of WisconsinMadison, Madison, Wisconsin 53706, United States

*S Supporting Information

ABSTRACT: Surface-reaction-limited pulsed chemical vapor deposition (SPCVD) is able to create 3D TiO2-based hierarchicalnanowire (NW) architecture with superhigh surface area and good electronic transport properties for high-performancephotoelectrode development. However, how to intentionally dope the nanorods (NRs) through the SPCVD process to improvetheir electronic properties and light absorption behavior is still unexplored. In this paper, a comprehensive study of doping TiO2NRs with nitrogen through the SPCVD technique is reported for the first time. The high-density nitrogen doped TiO2 NRbranches with controlled doping concentrations were synthesized on dense Si NW forest by introducing designed number ofTiN cycles to TiO2 growth cycles. Microscopic studies revealed the influence of nitrogen doping on the crystal growth behaviorand NR morphology, as well as the elements distribution inside the lattices. Nitrogen doping lowered the band gap of TiO2 NRsand effectively activated visible light photoactivity. It also largely improved the incident-photon-to-current-conversion efficiencyin the UV range. Successful synthesis of N doped TiO2 NRs by the SPCVD method introduces a strong new capability to thisnovel and powerful 3D NR growth technique. It enables composition and optoelectronic property control of the novel 3D NRstructures, allowing performance enhancement or creating new functionality.

1. INTRODUCTION

Titanium dioxide (TiO2) is a widely used multifunctionalmaterial due to its superior stability and physical−chemicalproperties, as well as its great accessibility and moderate cost.1

Owing to the intriguing discovery of the photocatalyticproperties on TiO2 in the early 1970s,2 TiO2 quickly attractedtremendous research interest in its applications on environ-mental remediation, hydrogen generation, and solar energyharvesting.3−8 One critical limitation of its application in solarenergy harvesting is its large band gap (∼3.0−3.2 eV) whichexcludes it from absorbing the majority of solar energy withinthe visible light range.8−10 Tailoring the band structure bydoping has been broadly investigated as a common strategy toaddress this challenge.9,11−13 Anion doping (e.g., C, S, N, F, B,and P) has been intensively studied as an effective approach tonarrow the band gap by introducing new p-states near thevalence band or hybridizing the O 2p states to lift the valenceband edge.4,14−22 Among all the anion candidates, nitrogen is apromising dopant due to its comparable atomic size with

oxygen, small ionization energy, and good stability.23,24

Nitrogen doped TiO2 has been created via a variety of methodsand demonstrated efficient visible light absorption up to 600nm,16 which largely enhanced photoelectrochemical watersplitting performance,25 and suppressed the reduction of I3

−

on the TiO2 electrode in dye-sensitized solar cells (DSSC).26

Incorporating cations, such as V, Cr, Mn, Fe, and Ni, into theTiO2 lattice has also been found practical for band gapnarrowing.27,28 The transition metal dopants can introduceintraband discrete states near the edges of conduction band(CB) or valence band (VB), leading to visible light absorptionwith sub-band-gap energies.24,29,30 In addition, hydrogen hasalso been found as an effective dopant that enabled visible andinfrared light absorption in TiO2. The dopant induced a highlydisordered layer on the nanocrystals surface, which yields mid-

Received: December 18, 2014Revised: January 21, 2015

Article

pubs.acs.org/JPCC

© XXXX American Chemical Society A DOI: 10.1021/jp512622jJ. Phys. Chem. C XXXX, XXX, XXX−XXX

pubs.acs.org/JPCC

http://dx.doi.org/10.1021/jp512622j

gap states for optical excitation/relaxation and providestrapping sites for photoexcited charges to reduce rapidrecombination.31 In order to circumvent the impurity-relatedcharge separation and transport issue found in single elementdoped TiO2,

27,32 a donor−acceptor co-doping concept hasbeen proposed. For example, density functional theorycalculations suggested that (W,C), (Mo,C), (Ta,N), (Nb,N),(W,2N), and (Mo,2N) donor−acceptor pairs are good co-doping candidates, which can shift both CB and VB edges ofTiO2 and significantly narrow its band gap.9,33,34 Nevertheless,only a few successful developments were reported on co-dopedTiO2 because of the complexity in element and crystal structurecontrol during the synthesis.9 So far, various solution-based,vapor-based, and ex situ growth techniques have been appliedto the synthesis of doped TiO2 nanostructures. However, howto control dopant concentration and distribution and how thedoping elements influence the morphology and crystal structureof the host materials are inadequately understood. Suchunderstandings are particularly critical for engineering theproperties of TiO2 nanostructures and thereby further advancethe performance of TiO2-nanomaterials-based solar energyharvesters.9,10,12,16−20

Among all TiO2 nanostructure growth techniques, surface-reaction-limited pulsed chemical vapor deposition (SPCVD) isa powerful 1D nanostructure growth technique that has beendeveloped in our lab recently. It mimics the process of atomiclayer deposition (ALD) and offers a unique capability forgrowing high-density nanorod (NR) branches inside highlyconfined spaces. This technique realizes a 3D hierarchicalnanowire (NW) architecture with superhigh surface areadensity (total surface area/projected substrate area of >1000)and good electronic transport properties. Our recent researchrevealed several important growth phenomena in this uniqueanisotropic crystal growth process for morphology andcomposition control, including the oriented attachmentprinciple,8 the Ostwald−Lussac law,35 and the Kekendalleffect.36 However, how to intentionally dope the NRs throughthis process is still unexplored, though it is a critical aspect forcontrolling the electronic properties and improving theperformance. In this paper, we report a comprehensive studyof doping TiO2 NRs with nitrogen through the SPCVDtechnique. Nitrogen doping was achieved by introducingdesigned number of TiN cycles to TiO2 growth cycles. Forthe first time, we successfully synthesized high-density nitrogendoped TiO2 NR branches with controlled doping concen-trations on dense Si NW forest. Microscopic studies wereperformed to reveal the influence of nitrogen dopant to thecrystal growth behavior and NR morphology, as well as theelement’s distribution inside the lattices. We also demonstratedthat the nitrogen doping lowered the band gap of TiO2 NRsand effectively activated photoelectrochemical (PEC) reactionsin the visible light range.

2. EXPERIMENTAL SECTIONSynthesis of Nitrogen Doped TiO2 Branched Nano-

rods Architecture. Highly orientated Si NWs arrays werefabricated on n-type Si wafer by the reported metal assistedelectroless chemical etching method.5 The etchant solution wasa 40 mL mixture of 0.02 M AgNO3 and 5 M HF. The Si NWsarrays were cleaned by DI water and dried by N2 gas and thentransferred into our home-built ALD reaction chamber. In theSPCVD process for synthesizing nitrogen doped TiO2 NRsbranches, deposition cycle ratio of the NH3/H2O (X:Y)

increased from 1:80 to 1:1 to achieve for different Nconcentrations (A2−A7). The undoped TiO2 NRs branches(A1) for comparison were prepared by the same SPCVDprocess with only TiCl4/H2O cycles. The total number ofTiCl4/H2O cycles for each sample (A1−A7) was kept at 400.

Compositions and Structures Characterizations. X-rayphotoelectron spectroscopy (Thermo Fisher Scientific Inc.,Waltham, MA) was applied to characterize the chemicalcompositions of all nitrogen doped TiO2 NR samples. Thesurvey binding energy range is from 0 to 1350 eV. The N 1s, O1s, and Ti 2p spectra were measured for characterizing thedopant concentrations and chemical structure variations. Thedensity of states (DOS) of the valence band maximum profilewas also characterized by the XPS spectra. Both TEM andEELS combined with STEM were performed on highly dopedsample A6 and lightly doped sample A2. EDS analysis wasconducted during SEM analysis.

Photoanode Setup and Photoactivity Characteriza-tions. All nitrogen doped and undoped TiO2 NR samples wereimmersed into the dilute HF and HCl to remove the oxidizedSi and contaminations on the surface. Then 200 cycles of ALDanatase TiO2 overcoating were conducted at 300 °C to protectthe entire sample surfaces. These samples were then covered byepoxy, leaving an average exposed active area of ∼0.22 mm2 asthe photoanodes. PEC characterizations were performed in 1mol L−1 KOH (pH = 14) aqueous solution using a three-electrode electrochemical cell configuration. A saturatedcalomel electrode (SCE) was used as the reference electrode,and a Pt wire was used as the counter electrode. All electrodeswere connected to a potentiostat system (Metrohm Inc.,Riverview, FL) for J−V measurements. Light illumination wasprovided by a 150 W Xe arc lamp (Newport Corporation,Irvine, CA), and the intensity at the PEC anode position wasadjusted to be 100 mW cm−2. An AM 1.5G filter and a UV cut-off filter were also utilized with the lamp for PEC character-izations. A monochromator was used on Xe lamp source toobtain the monochromatic illuminations for IPCE character-izations with the wavelength ranging from 300 to 700 nm. UV−vis absorption spectra were measured by a UV/vis/NIRspectrometer (Jasco, Easton, MD) within the wavelengthrange of 300−700 nm.

3. RESULTS AND DISCUSSIONFigure 1a and Figure 1b show a typical morphology of as-synthesized nitrogen doped TiO2 NRs on Si NW hierarchicalstructures measured by scanning electron microscopy (SEM).The top-view image presents dense nitrogen doped TiO2 NRsbranching from Si NW bundles and protruding outward radially(Figure 1a). Higher magnification image reveals that theapproximate dimensions of the as-fabricated NRs were ∼130nm in length and ∼25 nm in diameter (inset of Figure 1a).Same as the regular SPCVD result, the TiO2 NRs wereuniformly distributed along the entire side surfaces of the SiNWs that were 12 μm in length and only ∼100−200 nm apartfrom each other (Figure 1b). The zoom-in view clearly showsTiO2 NRs had the rectangular prism shape, which were allrooted on the Si surfaces (inset of Figure 1b).The average N concentrations were initially quantified by

energy-dispersive X-ray spectroscopy (EDS). The N signal wasobserved from all nitrogen doped samples (Figure S1a inSupporting Information and inset) but absent from theundoped one (Figure S1b and inset). Table S1 summarizesthe amounts of N dopant corresponding to the doping ratio,

The Journal of Physical Chemistry C Article

DOI: 10.1021/jp512622jJ. Phys. Chem. C XXXX, XXX, XXX−XXX

B


which was controlled by the ratio of NH3 cycle to H2O cycle(X:Y). It showed a very good correlation between N traces andexpected level of nitrogen doping. In particular, A6 (X:Y = 1:5)had the highest N doping concentration (∼2.28%) while A2(X:Y = 1:80) only had 0.82% N. Sample A7 (X:Y = 1:1)exhibited a slightly lower N concentration compared to A6,although it had the most NH3 cycles. This is possibly becausetoo high an NH3 precursor ratio might lead to more TiNgrowth cycles, which typically reduces the incorporation rate ofN elements into the lattice.37

The nitrogen doping effect to the TiO2 NR morphology wasinvestigated from samples grown under different doping ratios(Figure S2a−g). Samples A1−A7 represent undoped (A1) anddoped TiO2 NRs (A2−A7) with six X:Y values of 1:80, 1:40,1:20, 1:10, 1:5, and 1:1, respectively. The lengths and widths ofNRs in all the samples were measured via a statistical samplingmethod38 and shown in Figure 1c. In general, the length of NRsslightly decreased from 143 ± 9 nm (the undoped NRs) to 137± 7 nm when the NH3 cycle ratio was increased to 1:20(sample A4). Meanwhile, the average diameter of the NRsmonotonically decreased from 25 ± 4 nm (A1) to 19 ± 3 nm(A4). When the NH3 cycle ratio was further increased from1:20 to 1:1, a large decrease of NR length from 137 ± 7 nm(A4) to 114 ± 9 nm (A7) was observed. The NR diameter alsorapidly increased from 19 ± 3 nm (A4) to 34 ± 5 nm (A7),indicating a significant change of NR morphology from longand slim to short and plump upon the introduction of Ndopants.The statistical size measurement provides a preliminary

understanding of the TiO2 NR growth behavior by SPCVDunder the influence of NH3 precursor. The highest growth ratewas achieved at ∼3.6 Å/cycle when no NH3 precursor wasinduced (A1). Reduced length growth rate was observed withthe rising of the nitrogen doping ratio. The smallest length withthe average length growth rate of 1.4 Å /cycle was identifiedwhen the most TiCl4/NH3 cycles were programed for nitrogendoping (A7). This observation is similar to the ALD of nitrogendoped TiO2 films because TiN has lower growth rate thanTiO2.

10,37 In the high temperature SPCVD process, we have

found that the unique anisotropic crystal growth behavior ofTiO2 is enabled by the formation and migration of metastableamorphous−crystalline mix-phased nanoparticles (NPs) via theprinciple of vapor-phase oriented attachment.35 We thereforereckon that the deposition of TiN had significantly hinderedthe aggregation and migration of the metastable NPs and thussuppressed the NR length growth. The NR width growth rate,however, significantly increased from ∼0.45 to 0.84 Å/cyclewhen the NH3 precursor ratio was raised from1:20 to 1:1. Thisalso indirectly indicates the immobilization of the metastableNPs, which stay at the side surfaces and induce faster widthgrowth. Nevertheless, it appears that lower NH3 ratio (<1:20)can reduce the width growth rate as well. Therefore, the aspectratios of all samples followed a parabola-like curve as shown inFigure 1d. The maximum aspect ratio was 7.4 from sample A4,which had a NH3 precursor ratio of 1:20. In general, lowdoping concentration has minimal effect on the TiO2 NR’sgrowth rate and morphology, while high doping concentrationcould jeopardize the anisotropic crystal growth and form moreparticle-like nanoscrystals.The elemental characteristic of nitrogen doped TiO2 NRs

was further studied by X-ray photoelectron spectroscopy (XPS)as shown in Figure 2. The N 1s peaks could be identified from

all nitrogen doped TiO2 NR samples (A2−A7) but were absentin the undoped one (A1) (Figure 2a). To analyze the chemicalstructures of nitrogen doped and undoped TiO2 NRs, threeregions of the XPS spectrum were examined: N 1s scan near400 eV, Ti 2p scan near 460 eV, and O 1s scan near 530 eV(Figure 2b). For TiO2 NRs with different nitrogen doping ratio,there are observable variations among each scan. Broad N 1speaks expanding from 398.0 to 403.5 eV were observed in allnitrogen doped samples. The peaks were located at binding

Figure 1. Top view (a) and cross-section view (b) of high-densitynitrogen doped TiO2 NRs on Si NW arrays. Insets are highermagnification SEM images showing the morphology of nitrogen dopedTiO2 NRs branches. (c) Plots of NR length (square symbol) anddiameter (diamond symbol) as a function of nitrogen doping ratio. (d)Plots of aspect ratio of NRs as a function of nitrogen doping ratio.

Figure 2. XPS characterization of nitrogen doped TiO2 NRs withdifferent nitrogen doping ratios (A1−A7): (a) full XPS spectra; (b) N1s scan near 400 eV, Ti 2p scan near 460 eV, and O 1s scan near 530eV of A1−A7 samples; (c) deconvoluted spectra for N 1s peak ofheavily doped sample A6 (1) and lightly doped sample A2 (2), O 1speak of heavily doped sample A6 (3) and undoped sample A1 (4), andTi 2p peak of heavily doped sample A6 (5) and undoped sample A1(6).



C


energies higher than the typical value of TiN (397.2 eV).39 Thisis considered as the consequence of higher N 1s binding energyin an O−Ti−N environment, where the formal charge of N inO−Ti−N is more positive than that in TiN.39,40 The variationin peak intensities and positions also reveals the differentconcentrations and chemical environment. The doublet lines inthe spectrum profiles in the Ti 2p and O 1s scanning regions ofdoped and undoped TiO2 NRs exhibit similar differences. Inparticular, sample A6’s (X:Y = 1:5) exhibit the most significantshift of its Ti 2p3/2 peak to lower binding energy (459 eV) andamplitude attenuation of its O 1s peak at 533 eV (red curve inFigure 2b). The shift in the binding energy of Ti 2p peakstoward lower values indicates a successful incorporation of N inthe TiO2 lattice. The change in O 1s spectrum suggested theformation of O−Ti−N bonds and the modification of OHgroups on the NR surfaces.8,39,40

In order to elucidate the chemical binding states of N, O, andTi elements, the deconvoluted spectra for N 1s, O 1s, and Ti 2ppeaks were investigated and presented in Figure 2c. The N 1sprofile can be split into two characteristic peaks and assigned toO−Ti−N or Ti−N−O bonds at 400.2 eV (green peak) andNH4

+ adsorbed on NR surfaces at 402.5 eV (purple peak) asreported in the literature (Figure 2c1 and Figure 2c2).41,42 Asshown in Figure 2c1, the area ratio of the O−Ti−N peak (at∼400.2 eV) to the total N 1s peak in sample A6 (X:Y = 1:5) is∼0.86 while the area ratio of NH4

+ (at ∼402.5 eV) is ∼0.14.When the doping ratio decreased to X:Y = 1:80 (sample A2shown in Figure 2c2), the two ratios became 0.61 and 0.39,respectively, suggesting that a more substitutional doping mighthappen at 1:5 doping ratio. By comparison of the nitrogendoped sample A6 and undoped sample A1, more notablevariations were observed in O 1s (Figure 2c3 and Figure 2c4)

and Ti 2p (Figure 2c5 and Figure 2c6) signals regarding Nincorporation. The O 1s peak of A6 was deconvoluted intothree contributions at 530.6 eV (green), 532.4 eV (purple), and533.4 eV (yellow) which are identified as Ti−O, O−Ti−N, andsurface OH groups, respectively. The peaks of sample A1 wereonly deconvoluted into two components of Ti−O (green) andOH groups (yellow), while the O−Ti−N signal at 532.4 eV wasabsent. This further confirms that the O−Ti−N bonds wereformed in the nitrogen doped TiO2 NRs. The surface OHgroup related peak at 533.4 eV is common in XPS studies ofTiO2. The area ratio of this component was reduced in sampleA6, possibly because of more substitutions of O atoms by Nand the consumption of surface OH groups for NH3protonation (Figure 2c1).42

Earlier research mentioned that the valence state of the Tications in TiO2 should be reduced when N substitutes O bydoping. The binding energy of Ti 2p peaks shifted to lowervalues because of Ti4+ being reduced to Ti3+.40,43 Developmentof the O−Ti−N bonds through partially substituting O with Nin TiO2 lattice can be indicated by the shifts of Ti 2p bindingenergy.40 This effect was illustrated by XPS profiles of Ti 2pscans.41,44 Figure 2c5 is the deconvoluted spectra of the Ti 2ppeaks of sample A6. The extra Ti3+ signals appeared at 459.3 eVin Ti 2p3/2 (green) and 464.5 eV in Ti 2p1/2 (yellow), whichwere located at the lower binding energy side of each Ti−Opeak. The undoped one only showed two peaks of typical Ti−O bonds (Figure 2c6).Transmission electron microscopy (TEM) was implemented

to study the morphology and crystal structure of the nitrogendoped TiO2 NRs. Figure 3a shows high resolution TEM(HRTEM) images of NRs from sample A6 that exhibits thehighest detected N concentration. The images were taken from

Figure 3. Structure and species distribution in TiO2 NRs. (a) TEM image of nitrogen doped TiO2 NRs bundles on a Si NW. (b, c) High resolutionTEM images taken from the tip (b) and body (c) regions. The NR is covered by a ∼2.5−3.5 nm thick percolated layer as marked by the dashedlines. Inset in (c) is the FFT pattern of the lattice image. It confirms the anatase phase. (d) Voids observed in the percolated layer, as marked bywhite dashed lines. (e) EELS spectrum image measurements on a nitrogen doped TiO2 NR showing the distributions of Ti, O, and N elements, aswell as the concentration difference between Ti and O (Ti/O) along the entire NR body. (f) N concentration profile along the NR cross section asmarked by the square dots in the left panel of Figure 4e.



D


the tip and body regions. The truncated tip of this NR istypical, and the facets could be indexed as the (001), (011),(01 ̅1 ̅), and (01 ̅1) planes. The tip angle of NR is 111.4°,representing the angle between the (01 ̅1) and (001) planeswith lattice spaces measured to be 0.35 and 0.48 nm,respectively (Figure 3b). The NR tip and body were completelycovered by a ∼2.5−3.5 nm thick percolated layer, a mixture ofrandomly oriented nanocrystals with amorphous intercalations(Figure 3c). This is different from the single crystalline sharpsurfaces that were observed from the undoped TiO2 NRs.45

Beneath the percolated layer, the NR exhibits a nearly perfectcrystal lattice with perfect {011} facets, which is consist withthe observation for undoped TiO2 NRs.

45 Analysis of the fastFourier transform (FFT) pattern of a cut image showing thecentral part of N-doped NR reveals the anatase structure ofTiO2 (inset of Figure 3c).To analyze the element distribution and understand the

composition of the percolated layer, electron energy lossspectroscopy spectrum image (EELS SI) was taken from asingle nitrogen doped TiO2 NR (center and left panel of Figure3e). Appreciable N signal in EELS spectrum was detected fromthe NR (left panel of Figure 3e). Elemental mapping clearlyreveals the distributions of Ti, O, and N elements, as well as theconcentration difference between Ti and O (Ti/O) along theentire NR body (right panel of Figure 3e). More Ti was foundnear the NR surface, where less O element was identified. Themap of Ti/O reveals the O depleted surface region and thestoichiometric NR body. The N map demonstrates aconcentrated N region along the NR surface profile. Such anonuniform N distribution is further confirmed by measure-ment of the N concentration profile along the NR cross-sectionas shown in Figure 3f, where depth of surface layer withincreased concentration of nitrogen was ∼4 nm.It is known that the impurity formation energy in doped

nanocrystals increases as the crystal size decreases.46 Impurityatoms can diffuse to the crystal surface, referred to as self-

purification.46,47 Meanwhile, point defects such as vacancies andinterstitials can diffuse outward and aggregate to form voids andgrain boundaries.48,49 We believe the formation of thepercolated layer with highly concentrated N atoms manifestsabove processes. In the high-temperature SPCVD process,oxygen vacancies associated with O−Ti−N structures coulddiffuse to the surface while the TiO2 NR body was crystallized,forming voids (Figure 3d) in the percolated layer.50 As a result,the nitrogen-rich percolated layers were formed near the NRsurfaces. Because the NR formation mechanism has been foundto be the migration of crystallized particles on manifold bodyfollowed by orientated attachment on high surface energy facets(the tip part of NRs), a significant amount of the percolatedlayer covering the NRs might be an obstacle for the migrationand attachment of crystal particles to the tip area of NRs.Theoretical and experimental work suggested that N

impurity could introduce localized states above valence bandmaximum (VBM) or VBM lifting, enabling photoactiviy in thevisible light range.4,8,12,51 The surface disorder and oxygenvacancies were also known to be able to shift the VBM andnarrow the band gap of TiO2.

52 Therefore, introducingnitrogen doping to the SPCVD TiO2 NR branch growth isexpected to realize a promising 3D NR architecture for high-performance PEC photoanode across a wide spectrum range.To investigate the doping-related PEC performance, the sixnitrogen doped TiO2 NR samples (A2−A7) and the undopedsample (A1) were used as PEC photoanodes and theirphotocurrent density (Jph) characteristics versus bias potentialwere measured under the same conditions. Corresponding PECefficiencies were estimated using following equation:6

η = − | − |⎡⎣⎢

⎤⎦⎥J E E E

I% ( )

100ph rev

0bias aoc

0 (1)

where Ebias is the bias potential, Erev0 = 1.23 V is the standard

state reversible potential for water-splitting reactions, and Eaoc =Voc is the open circuit voltage (vs SCE). Figure 4 shows

Figure 4. (a−c) J−V curves of nitrogen doped and undoped TiO2 NRs−Si NW hierarchical architectures under illuminations of 100 mW cm−2 Xelamp source, lamp with AM 1.5G filter, and lamp with UV cutoff filter, respectively. (d) XPS demonstrates that density of states (DOS) shift in thevalence band happened to all nitrogen doped samples. The maximum shift is ∼2.4 eV in sample A6. (e, f) IPCE of nitrogen doped and undopedTiO2 NRs−Si NW hierarchical architectures within the wavelength range of 300−425 and 425−700 nm, respectively.



E


representative Jph−V curves collected under the illumination ofa 100 W cm−2 Xe lamp without and with AM 1.5G or UVcutoff filters. The dark current densities of each PEC testremained at fairly low values (<10−4 mA cm−2) within the biaspotential range between −1.1 and 0.5 V (vs SCE), indicatinghigh quality crystal surfaces of the as-synthesized TiO2 NRarchitectures. Among all the samples, A6 exhibited the highestJph under all wavelength conditions. Generally, higher nitrogendoping yielded high photocurrent and higher efficiency. All thephotocurrent and efficiency data are summarized in Table 1.Under direct Xe lamp illumination, the maximum efficiency wascalculated to be 1.35% at approximately −0.75 V from sampleA7 (Figure S4a). This value is close to that of sample A6(1.23%) and 48% higher than that of the undoped TiO2 NRphotoanodes (0.91% at approximately −0.77 V). However,other nitrogen doped TiO2 NR samples (A2−A5) all yieldedlower photocurrent and efficiency compared to the undopedone. This is possibly due to the defective crystal structurearound the NR surfaces (Figure 3), which would jeopardize thecharge transport even though their visible light photoactivitymight be higher than that of the undoped sample.When an AM 1.5G filter was applied, sample A6 yielded the

maximum efficiency of 0.26% at approximately −0.36 Vwhereas the efficiency of A7 rapidly dropped to 0.19%, whichwas only 73% of that of A6. Other than A5, whose Jph was aboutthe same level as A7, all the other nitrogen doped NRsexhibited the same level of performance as the undoped one(∼0.08%). Therefore, under AM 1.5G illumination, themaximum efficiency offered by nitrogen doping was 225%higher than that of the undoped TiO2 NRs.When the samples were only exposed to visible light by

applying a UV cutoff filter (cutoff wavelength of 400 nm), thephotocurrents were further reduced. As shown in Figure 4c,Figure S4c, and Table 1, sample A6 still exhibited the highestperformance with a maximum efficiency of 0.04% atapproximately −0.54 V. Under this situation, the Jph andefficiency of sample A7 further dropped to the same level ofsample A4 (X:Y = 1:20) and were outperformed by sample A5.Similar to the observations under UV-involved illuminations,no significant difference was found between the low-doping-concentration samples (A2 and A3) and the undoped one,where the photocurrents remained at the 0.01 mA cm−2 level.Because of the rapid photocurrent drop when the entire UVspectrum was cut off, 1 order of magnitude enhancement wasidentified when appropriate N was doped into TiO2 NRs.The as-synthesized heavily nitrogen doped TiO2 NRs (A6)

appeared darker in color compared to the undoped TiO2 NRs(A1, inset Figure S3), demonstrating its more effectiveabsorption of visible light. Corresponding UV−vis absorptionspectra revealed a significantly enhanced absorption in the

visible light regime (less than ∼600 nm) as a result of nitrogendoping (Figure S3). XPS was then used to further understandthis phenomenon from the aspect of density of states (DOS)shift in the valence band (Figure 4d). From the XPS spectra,the undoped TiO2 NRs (A1) exhibited a VBM at ∼4.8 eV.While sample A2, which had the lowest N dopingconcentration, only exhibited a shift of 0.5 eV, all the othernitrogen doped samples showed a shift of 2.2−2.4 eV. Thechange of VBM indicates the emergence of absorptioncapability in the visible light range, which could possiblyenable photoactivity in broad-band illumination and thusenhanced PEC performance. These observations indicate thesuccessful activation of visible-light photoactivity induced by Ndoping.Incident-photon-to-current-conversion efficiency (IPCE) was

used to quantitatively understand the N doping effect on thephotoactivity as a function of incident wavelength.53,54 IPCEcharacterizations (Figure 4e and Figure 4f) were conducted onall nitrogen doped and undoped samples at the potentialswhere Jph approaches saturation (e.g., −0.55 V vs SCE for A6).IPCE values were calculated via the following equation:55

λ= J PIPCE (%) (1240 )/( )ph mono (2)

where Jph is the photocurrent density at the incident wavelengthλ. Pmono is the power density of the incident monochromaticlight. All the samples exhibited significant photoactivity over theentire UV region and decreased sharply when the wavelength isgreater than 360 nm. The lowest IPCE (∼35−40%) was foundfrom samples A2 and A3, which was consistent with the lowestJph and efficiency observed from these two samples in allwavelength ranges. We believe this is due to the very low levelof nitrogen doping that was not sufficient to completely altertheir optoelectronic behavior but rather introduced defectsjeopardizing the charge transport properties. It is also surprisingto see a very high IPCE from relatively heavily doped TiO2 NRsamples (A6 and A7). Their IPCE values reached over 95% inthe UV range, which was 27% higher than that of undopedTiO2 NRs. Considering the absorption of heavily doped TiO2in this wavelength range is comparable to that of undoped one,we attribute this enhancement to their more efficient utilizationof UV light where the separation and transportation ofphotoexcited charge carriers are more effective.9,53 Sample A6exhibited the highest high IPCE (∼94%) at ∼360 nm and thelongest cutoff wavelength at 425 nm, which is consistent withits highest photoactivity among all samples. A blue shift on thecutoff wavelength was observed when the dopant concentrationdecreased and the undoped TiO2 NRs exhibited the shortestcutoff wavelength at ∼400 nm, which is consistent with theirUV−vis spectra.

Table 1. Summary of Photocurrent Density and PEC Efficiency of Nitrogen Doped and Undoped TiO2 NR−Si NWArchitectures Measured under Different Illumination Conditions

sample

A1 A2 A3 A4 A5 A6 A7

undoped 1:80 1:40 1:20 1:10 1:5 1:1

J (mA cm−2) lamp source 2.081 1.650 0.668 1.271 2.159 3.920 3.187UV cut-off filter 0.011 0.011 0.015 0.040 0.060 0.095 0.039AM 1.5G filter 0.175 0.187 0.140 0.159 0.426 0.858 0.466

η (%) lamp source 0.905 0.362 0.324 0.531 0.492 1.232 1.357UV cut-off filter 0.004 0.003 0.005 0.014 0.028 0.040 0.015AM 1.5G filter 0.082 0.059 0.074 0.060 0.149 0.257 0.188



F


Figure 4f shows IPCE in the visible light region. Here, we canallocate their visible light performance into three groups. GroupI contains relatively heavily doped TiO2 NRs (A6 and A7).They exhibited the highest IPCE in the 450−700 nm range(0.8−1.0%) with an obvious peak at ∼610 nm. Group IIcontains moderately doped TiO2 NRs (A4 and A5). Theyshowed a broad IPCE peak in the 450−700 nm regime with amaximum value of ∼0.6% across a broad band (500−600 nm).Group III contains lightly doped and undoped TiO2 NRs (A1−A3). Their IPCE values were nearly zero within this wavelengthrange, suggesting that the low concentration of N is not able tointroduce photoactivity in the visible light range. ConsideringTi3+ in nitrogen doped TiO2 is referred to as a color centerwhich induces interbands for visible light absorption,41,56,57 weattribute the conspicuous IPCE around ∼610 nm of group Initrogen doped TiO2 to the Ti3+ related interbands lightabsorption.In addition, the IPCE investigation also helps in under-

standing the strong wavelength dependence of sample A7’sPEC performance. Since both A6 and A7 exhibited comparablyhigh IPCE in the UV range (<350 nm), they showed similarPEC performance under Xe lamp illumination which had asignificant UV portion. IPCE of A7 rapidly dropped when thewavelength was greater than 350 nm. Therefore, its PECperformance was lowered to the level of sample A5 when theAM 1.5G filter was applied. Once the UV illumination wascompletely blocked by a UV cut-off filter, the Jph and efficiencyof A7 further dropped to the same level of A4, despite its higherIPCE in the visible light region.

4. CONCLUSIONIn summary, we successfully introduced nitrogen dopingcapability in the SPCVD process for synthesizing high-densitynitrogen doped TiO2 NR branches on dense Si NW forests.The N elements were introduced by adding NH3 vapor asadditional precursor pulses between typical H2O and TiCl4cycles for growing TiO2 NRs at 600 °C. The dopantconcentration can be rationally controlled by adjusting theratio between the doping cycles and regular TiO2 growthcycles. HRTEM and EELS mapping revealed that the Ndopants were mostly accumulated around the surface of NR,forming a percolated layer. XPS study confirmed the formationof O−Ti−N bonds in nitrogen doped TiO2 NRs. Ti

3+ signalswere also observed, which are associated with the reduction ofTi4+ when N substitutes O by doping. Compared to anundoped one, the valence band maximum of nitrogen dopedTiO2 NRs exhibited a notable lifting up. The largest shift valuereached ∼2.4 eV in the most highly doped sample. The as-synthesized nitrogen doped TiO2 NR branched heterostruc-tures were used as PEC photoanodes, and their performancewas characterized within controlled illumination wavelengthranges. The nitrogen doped TiO2 NRs demonstrated fairlycompetitive photoactivity under 100 W cm−2 Xe lamp andsimulated solar illuminations. Appreciable PEC activity was alsoobserved in the visible light range from 450 to 700 nm. Themaximum efficiencies offered by nitrogen doping were 225%higher than those of the undoped TiO2 NRs under AM 1.5Gillumination. The enhancement of PEC efficiency was morethan 900% when only visible light was illuminated. Successfulsynthesis of nitrogen doped TiO2 NRs by the SPCVD methodintroduces a strong new capability to this novel and powerful3D NR growth technique. It enables composition andoptoelectronic property control of the NR branches, allowing

performance enhancement or creating new functionality.Understanding the doping behavior and controlling the dopingconcentration in SPCVD 3D NR growth hold a great promisefor engineering the electrical properties of functional nanoma-terials for PEC and other energy harvesting and conversionapplications.

■ ASSOCIATED CONTENT*S Supporting InformationEDS spectra, summary of the average concentrations of Ndopant in all doped and undoped samples, SEM images of allsamples, UV−vis absorption spectra of heavily doped (A6) andundoped (A1) TiO2 NRs, PEC efficiencies of all doped andundoped TiO2 NRs−Si NW hierarchical architectures underdifferent illuminations. This material is available free of chargevia the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: [email protected]. Phone: +1-608-890-2667.Fax: +1-608-262-8353.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis research is primarily supported by the U.S. Department ofEnergy (DOE), Office of Science, Basic Energy Sciences (BES),under Award DE-SC0008711. F.W. thanks the support of AirForce Office of Scientific Research (AFOSR) under AwardFA9550-13-1-0168 for TEM characterization.

■ REFERENCES(1) Wang, X.; Li, Z.; Shi, J.; Yu, Y. One-Dimensional TitaniumDioxide Nanomaterials: Nanowires, Nanorods, and Nanobelts. Chem.Rev. 2014, 114, 9346−9384.(2) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at aSemiconductor Electrode. Nature 1972, 238, 37−38.(3) Xiong, Z.; Zhao, X. S. Nitrogen-Doped Titanate-Anatase Core-Shell Nanobelts with Exposed {101} Anatase Facets and EnhancedVisible Light Photocatalytic Activity. J. Am. Chem. Soc. 2012, 134,5754−5757.(4) Wang, J.; Tafen de, N.; Lewis, J. P.; Hong, Z.; Manivannan, A.;Zhi, M.; Li, M.; Wu, N. Origin of Photocatalytic Activity of Nitrogen-Doped TiO2 Nanobelts. J. Am. Chem. Soc. 2009, 131, 12290−12297.(5) Shi, J.; Hara, Y.; Sun, C.; Anderson, M. A.; Wang, X. Three-Dimensional High-Density Hierarchical Nanowire Architecture forHigh-Performance Photoelectrochemical Electrodes. Nano Lett. 2011,11, 3413−3419.(6) Li, Z.; Yao, C.; Yu, Y.; Cai, Z.; Wang, X. Highly Efficient CapillaryPhotoelectrochemical Water Splitting Using Cellulose Nanofiber-Templated TiO2 Photoanodes. Adv. Mater. 2014, 26, 2262−2267.(7) Cho, I. S.; Chen, Z.; Forman, A. J.; Kim, D. R.; Rao, P. M.;Jaramillo, T. F.; Zheng, X. Branched TiO2 Nanorods for Photo-electrochemical Hydrogen Production. Nano Lett. 2011, 11, 4978−4984.(8) Shi, J.; Wang, X. Hierarchical TiO2−Si Nanowire Architecturewith Photoelectrochemical Activity under Visible Light Illumination.Energy Environ. Sci. 2012, 5, 7918−7922.(9) Cho, I. S.; Lee, C. H.; Feng, Y.; Logar, M.; Rao, P. M.; Cai, L.;Kim, D. R.; Sinclair, R.; Zheng, X. Codoping Titanium DioxideNanowires with Tungsten and Carbon for Enhanced Photo-electrochemical Performance. Nat. Commun. 2013, 4, 1723.(10) Pore, V.; Heikkila,̈ M.; Ritala, M.; Leskela,̈ M.; Areva, S. AtomicLayer Deposition of TiO2−XNx Thin Films for PhotocatalyticApplications. J. Photochem. Photobiol., A 2006, 177, 68−75.



G

http://pubs.acs.org

mailto:[email protected]


(11) Ohno, T.; Murakami, N. Development of Visible-Light Active SCation-Doped TiO2 Photocatalyst. Curr. Org. Chem. 2010, 14, 699−708.(12) Park, J. H.; Kim, S.; Bard, A. J. Novel Carbon-Doped TiO2Nanotube Arrays with High Aspect Ratios for Efficient Solar WaterSplitting. Nano Lett. 2006, 6, 24−28.(13) Khan, S. U.; Al-Shahry, M.; Ingler, W. B., Jr. EfficientPhotochemical Water Splitting by a Chemically Modified N-TiO2.Science 2002, 297, 2243−2245.(14) Chen, X.; Burda, C. The Electronic Origin of the Visible-LightAbsorption Properties of C-, N- and S-Doped TiO2 Nanomaterials. J.Am. Chem. Soc. 2008, 130, 5018−5019.(15) Morikawa, T.; Asahi, R.; Ohwaki, T.; Aoki, K.; Taga, Y. Band-Gap Narrowing of Titanium Dioxide by Nitrogen Doping. Jpn. J. Appl.Phys. 2001, 40, L561−L563.(16) Burda, C.; Lou, Y.; Chen, X.; Samia, A. C. S.; Stout, J.; Gole, J. L.Enhanced Nitrogen Doping in TiO2 nanoparticles. Nano Lett. 2003, 3,1049−1051.(17) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Visible-Light Photocatalysis in Nitrogen-Doped Titanium Oxides. Science2001, 293, 269−271.(18) Tang, Y.-C.; Huang, X.-H.; Yu, H.-Q.; Tang, L.-H. Nitrogen-Doped Photocatalyst Prepared by Mechanochemical Method: DopingMechanisms and Visible Photoactivity of Pollutant Degradation. Int. J.Photoenergy 2012, 2012, 1−10.(19) Dunnill, C. W.; Parkin, I. P. N-Doped Titania Thin FilmsPrepared by Atmospheric Pressure CVD Using t-Butylamine as theNitrogen Source: Enhanced Photocatalytic Activity under VisibleLight. Chem. Vap. Deposition 2009, 15, 171−174.(20) Shen, H.; Mi, L.; Xu, P.; Shen, W.; Wang, P.-N. Visible-LightPhotocatalysis of Nitrogen-Doped TiO2 Nanoparticulate FilmsPrepared by Low-Energy Ion Implantation. Appl. Surf. Sci. 2007,253, 7024−7028.(21) Yang, Y.; Zhong, H.; Tian, C. Photocatalytic Mechanisms ofModified Titania under Visible Light. Res. Chem. Intermed. 2010, 37,91−102.(22) Morikawa, T.; Asahi, R.; Ohwaki, T.; Aoki, K.; Suzuki, K.; Taga,Y. Visible-Light Photocatalyst - Nitrogen-Doped Titanium Dioxide.R&D Rev. Toyota CRDL 2005, 40, 45−50.(23) Burda, C.; El-Sayed, M. A. High-Density FemtosecondTransient Absorption Spectroscopy of Semiconductor Nanoparticles.A Tool To Investigate Surface Quality. Pure Appl. Chem. 2000, 72,165−177.(24) Devi, L. G.; Kavitha, R. A Review on Non Metal Ion DopedTitania for the Photocatalytic Degradation of Organic Pollutants underUV/Solar Light: Role of Photogenerated Charge Carrier Dynamics inEnhancing the Activity. Appl. Catal., B 2013, 140−141, 559−587.(25) Hoang, S.; Berglund, S. P.; Hahn, N. T.; Bard, A. J.; Mullins, C.B. Enhancing Visible Light Photo-Oxidation of Water with TiO2

Nanowire Arrays via Cotreatment with H2 and NH3: SynergisticEffects between Ti3+ and N. J. Am. Chem. Soc. 2012, 134, 3659−3662.(26) Tian, H.; Hu, L.; Zhang, C.; Liu, W.; Huang, Y.; Mo, L.; Guo,L.; Sheng, J.; Dai, S. Retarded Charge Recombination in Dye-Sensitized Nitrogen-Doped TiO2 Solar Cells. J. Phys. Chem. C 2010,114, 1627−1632.(27) Choi, W.; Termin, A.; Hoffmann, M. R. The Role of Metal IonDopants in Quantum-Sized TiO2: Correlation between Photo-reactivity and Charge Carrier Recombination Dynamics. J. Phys.Chem. 1994, 98, 13669−13679.(28) Liu, B.; Chen, H. M.; Liu, C.; Andrews, S. C.; Hahn, C.; Yang, P.Large-Scale Synthesis of Transition-Metal-Doped TiO2 Nanowireswith Controllable Overpotential. J. Am. Chem. Soc. 2013, 135, 9995−9998.(29) Devi, L. G.; Kottam, N.; Kumar, S. G. Preparation andCharacterization of Mn-Doped Titanates with a Bicrystalline Frame-work: Correlation of the Crystallite Size with the Synergistic Effect onthe Photocatalytic Activity. J. Phys. Chem. C 2009, 113, 15593−15601.(30) Yu, J.; Xiang, Q.; Zhou, M. Preparation, Characterization andVisible-Light-Driven Photocatalytic Activity of Fe-Doped Titania

Nanorods and First-Principles Study for Electronic Structures. Appl.Catal., B 2009, 90, 595−602.(31) Chen, X.; Liu, L.; Yu, P. Y.; Mao, S. S. Increasing SolarAbsorption for Photocatalysis with Black Hydrogenated TitaniumDioxide Nanocrystals. Science 2011, 331, 746−750.(32) Devi, L. G.; Kumar, S. G. Influence of Physicochemical−Electronic Properties of Transition Metal Ion Doped PolycrystallineTitania on the Photocatalytic Degradation of Indigo Carmine and 4-Nitrophenol under UV/Solar Light. Appl. Surf. Sci. 2011, 257, 2779−2790.(33) Yin, W.-J.; Tang, H.; Wei, S.-H.; Al-Jassim, M. M.; Turner, J.;Yan, Y. Band Structure Engineering of Semiconductors for EnhancedPhotoelectrochemical Water Splitting: The Case of TiO2. Phys. Rev. B2010, 82, 045106.(34) Ma, X.; Wu, Y.; Lu, Y.; Xu, J.; Wang, Y.; Zhu, Y. Effect ofCompensated Codoping on the Photoelectrochemical Properties ofAnatase TiO2 Photocatalyst. J. Phys. Chem. C 2011, 115, 16963−16969.(35) Shi, J.; Li, Z.; Kvit, A.; Krylyuk, S.; Davydov, A. V.; Wang, X.Electron Microscopy Observation of TiO2 Nanocrystal Evolution inHigh-Temperature Atomic Layer Deposition. Nano Lett. 2013, 13,5727−5734.(36) Yu, Y.; Yin, X.; Kvit, A.; Wang, X. Evolution of Hollow TiO2

Nanostructures via the Kirkendall Effect Driven by Cation Exchangewith Enhanced Photoelectrochemical Performance. Nano Lett. 2014,14, 2528−2535.(37) Elers, K.-E.; Winkler, J.; Weeks, K.; Marcus, S. TiCl4 as aPrecursor in the Tin Deposition by Ald and Peald. J. Electrochem. Soc.2005, 152, G589−G593.(38) Wang, F.; Hwang, Y.; Qian, P. Z.; Wang, X. A Statistics-GuidedApproach to Precise Characterization of Nanowire Morphology. ACSNano 2010, 4, 855−862.(39) Saha, N. C.; Tompkins, H. G. Titanium Nitride OxidationChemistry: An X-ray Photoelectron Spectroscopy Study. J. Appl. Phys.1992, 72, 3072−3079.(40) Chen, X.; Burda, C. Photoelectron Spectroscopic Investigationof Nitrogen-Doped Titania Nanoparticles. J. Phys. Chem. B 2004, 108,15446−15449.(41) Jagadale, T. C.; Takale, S. P.; Sonawane, R. S.; Joshi, H. M.;Patil, S. I.; Kale, B. B.; Ogale, S. B. N-Doped TiO2 Nanoparticle BasedVisible Light Photocatalyst by Modified Peroxide Sol−Gel Method. J.Phys. Chem. C 2008, 112, 14595−14602.(42) Sham, T. K.; Lazarus, M. S. X-ray Photoelectron Spectroscopy(XPS) Studies of Clean and Hydrated TiO2 (Rutile) Surfaces. Chem.Phys. Lett. 1979, 68, 426−432.(43) Hashimoto, S.; Murata, A.; Sakurada, T.; Tanaka, A. Alternationof Ti 2p XPS Spectrum for TiO2 by Ar Ion Bombardment. J. Surf.Anal. 2002, 9, 459−462.(44) Cai, J.; Huang, Z. a.; Lv, K.; Sun, J.; Deng, K. Ti Powder-Assisted Synthesis of Ti3+ Self-Doped TiO2 Nanosheets withEnhanced Visible-Light Photoactivity. RSC Adv. 2014, 4, 19588−19593.(45) Shi, J.; Sun, C.; Starr, M. B.; Wang, X. Growth of TitaniumDioxide Nanorods in 3D-Confined Spaces. Nano Lett. 2011, 11, 624−631.(46) Dalpian, G.; Chelikowsky, J. Self-Purification in SemiconductorNanocrystals. Phys. Rev. Lett. 2006, 96, 226802.(47) Norris, D. J.; Efros, A. L.; Erwin, S. C. Doped Nanocrystals.Science 2008, 319, 1776−1779.(48) Wang, F.; Seo, J. H.; Li, Z.; Kvit, A. V.; Ma, Z.; Wang, X. Cl-Doped Zno Nanowires with Metallic Conductivity and TheirApplication for High-Performance Photoelectrochemical Electrodes.ACS Appl. Mater. Interfaces 2014, 6, 1288−1293.(49) Yankovich, A. B.; Puchala, B.; Wang, F.; Seo, J. H.; Morgan, D.;Wang, X.; Ma, Z.; Kvit, A. V.; Voyles, P. M. Stable P-Type Conductionfrom Sb-Decorated Head-to-Head Basal Plane Inversion DomainBoundaries in ZnO Nanowires. Nano Lett. 2012, 12, 1311−1316.(50) Rosenberg, R. Inhibition of Electromigration Damage in ThinFilms. J. Vac. Sci. Technol. 1972, 9, 263−270.



H


(51) Morisaki, H.; Hariya, M.; Yazawa, K. Anomalous Photoresponseof N-TiO2 Electrode in a Photoelectrochemical Cell. Appl. Phys. Lett.1977, 30, 7−9.(52) Liu, L.; Chen, X. Titanium Dioxide Nanomaterials: Self-Structural Modifications. Chem. Rev. 2014, 114, 9890−9918.(53) Wang, G.; Wang, H.; Ling, Y.; Tang, Y.; Yang, X.; Fitzmorris, R.C.; Wang, C.; Zhang, J. Z.; Li, Y. Hydrogen-Treated TiO2 NanowireArrays for Photoelectrochemical Water Splitting. Nano Lett. 2011, 11,3026−3033.(54) Murphy, A.; Barnes, P.; Randeniya, L.; Plumb, I.; Grey, I.;Horne, M.; Glasscock, J. Efficiency of Solar Water Splitting UsingSemiconductor Electrodes. Int. J. Hydrogen Energy 2006, 31, 1999−2017.(55) Rao, P. M.; Cai, L.; Liu, C.; Cho, I. S.; Lee, C. H.; Weisse, J. M.;Yang, P.; Zheng, X. Simultaneously Efficient Light Absorption andCharge Separation in WO3/BiVO4 Core/Shell Nanowire Photoanodefor Photoelectrochemical Water Oxidation. Nano Lett. 2014, 14,1099−1105.(56) Zhang, Z.; Yang, X.; Hedhili, M. N.; Ahmed, E.; Shi, L.; Wang,P. Microwave-Assisted Self-Doping of TiO2 Photonic Crystals forEfficient Photoelectrochemical Water Splitting. ACS Appl. Mater.Interfaces 2014, 6, 691−696.(57) Ventosa, E.; Xia, W.; Klink, S.; La Mantia, F.; Mei, B.; Muhler,M.; Schuhmann, W. Ammonia-Annealed TiO2 as a Negative ElectrodeMaterial in Li-Ion Batteries: N Doping or Oxygen Deficiency?Chemistry 2013, 19, 14194−14199.



I


Documents

Nitrogen Doped 3D Titanium