Non-metal doping of transition metal oxides for visible-light photocatalysis

Please cite(2013), htt

ARTICLE IN PRESSG ModelCATTOD-8718; No. of Pages 25Catalysis Today xxx (2013) xxx xxx

Contents lists available at ScienceDirect

Catalysis Today

jou rn al hom epage: www.elsev ier .com

Non-m or photoc

Roland Ma Institute of Phb ARC Centre of ity of QAustralia

a r t i c l

Article history:Received 29 JuReceived in re10 SeptemberAccepted 8 OcAvailable onlin

Keywords:PhotocatalystsMetal oxidesNon-metal dopingBand gap engineeringVisible light photocatalysis

ides re kn. Howr suring op of sompaous nd to s

of an open crystal structure on the efciency of the doping process. We then present the highlights andbreakthroughs of the last ten years in the research eld and point out major improvements importantfor future applications, covering all the available non-metal doped transition metal oxides concerningphotocatalytic reactions.

2013 Elsevier B.V. All rights reserved.

1. Introdu

In timeslation growis one of thconversion the major cfrom waterto fossil fueactual hydrcheap and rcombusting

While pfrom solar semiconducinto clean hfeatures, seysis. By usthe lled v

Correspon Correspon

E-mail [email protected]

0920-5861/$ http://dx.doi.o this article in press as: R. Marschall, L. Wang, Non-metal doping of transition metal oxides for visible-light photocatalysis, Catal. Todayp://dx.doi.org/10.1016/j.cattod.2013.10.088

ction

of an increasing demand for clean energy, global popu-th and global warming, a sustainable energy productione priority issues for humanity. Therefore, solar energyto generate electricity or produce solar fuels is one ofhallenges scientists are facing nowadays. Hydrogen e.g.

splitting is often discussed as future fuel as alternativels, especially with respect to the environmental issues ofogen production [1]. Both sunlight and water representenewable energy sources, and the only waste product

hydrogen as fuel would be water [2].hotovoltaic devices can be used to generate electricitylight, photoelectrochemistry and photocatalysis usingtor materials are direct processes converting solar lightydrogen without CO2 emission. Due to their electronicmiconductor materials are mostly used in photocatal-ing light irradiation, electrons can be excited fromalence band (VB) of a semiconductor into its empty

ding author. Tel.: +61 7 33654218.ding author. Tel.: +49 641 9934585.resses: [email protected] (R. Marschall),u.au (L. Wang).

conduction band (CB) if the energy of the incident photon equals orexceeds the energy of the band gap (Eg) of the semiconductor (Fig. 1)[36]. After photoexcitation, the charge carriers can be separatedand diffuse onto the particles surface, where they may performredox catalytic reactions. The photoexcited negative charge carri-ers are strong reductants and can be used to generate e.g. hydrogenfrom protons, while the positive charge carriers in the VB, strongoxidants, can be used to oxidize adsorbed molecules, e.g. water togenerate oxygen.

For an energetically-feasible overall water splitting reaction, theCB minimum of the semiconductor has to be more negative thanthe reduction potential of H+/H2, while the VB maximum needsto be more positive than the oxidation potential of O2/H2O. Thismeans that the Eg of the photocatalyst must be larger than 1.23 eVwith appropriate band positions.

Already in 1972, Honda and Fujishima were able to performoverall water splitting using a photoelectrochemical setup [7].However, since they used TiO2 as light absorber, they were not ableto make use of 95% of the solar spectrum: due to its Eg of 3.0 (rutile)to 3.2 eV (anatase), only UV light can be used to photogeneratecharge carriers in TiO2.

In the last decade, many different materials for photocatalyticwater splitting have been prepared, and the main challenges,including broad absorption capability and reduced recombinationprobability, have been identied to synthesize a good material for

see front matter 2013 Elsevier B.V. All rights reserved.rg/10.1016/j.cattod.2013.10.088etal doping of transition metal oxides fatalysis

arschall a,, Lianzhou Wangb,

ysical Chemistry, Justus-Liebig-University Giessen, D-35392 Giessen, Germany Excellence for Functional Nanomaterials, School of Chemical Engineering, The Univers

e i n f o

ne 2013vised form

2013tober 2013e xxx

a b s t r a c t

Transition metal oxides and mixed oxtions. Many highly active compounds adegradation and solar fuel generationform photocatalytic reactions at theithe recent progress in non-metal dopstrategies to reduce the large band galine the advantages of this strategy cstress the effect of efcient homogeneproperties of photocatalysts, compare/ locate /ca t tod

visible-light

ueensland, St. Lucia, Brisbane, Qld. 4072,

are the largest group of materials for photocatalytic applica-own from literature for environmental remediation, pollutantever, most of these oxides can only absorb UV light to per-

face due to their large band gap. In this review, we presentf transition metal oxides and mixed oxides, one of the majoremiconductor materials into the visible light range. We out-red to other band gap engineering methods, and especiallyon-metal doping on the optical, electronic and photocatalyticurface doping and surface modication, including the effects


ARTICLE IN PRESSG ModelCATTOD-8718; No. of Pages 252 R. Marschall, L. Wang / Catalysis Today xxx (2013) xxx xxx

ials (D

photocataly[817]. Butphotocataly

The maintion, chargsurface, andmolecules/aare inuenctocatalyst. Tmaterials hincluding csitization ometal/non-decoration

Controlltance for pastrongly affciency of phcan lead to facets are eductor is alsto amorphocharge carris thereforefor charge the surfacemolecular oto increase[20,21], exccupied mol(mostly TiOconductor ssolar cells a redox elethe formatiing direct e[22].

If no suiface, the penergy putally, photochange in Gthat from ato water is

les oo-cawithenerte th

formther d tuement abg vi

c redide oly difion, 24]. Troveen a

can ing les

inual st

effen-mFig. 1. Basic principle of photocatalysis with semiconductor mater

tic water splitting by research groups all over the world many of these materials still utilize only UV light forsis.

steps for a photocatalytic reaction are the light absorp-e carrier separation and diffusion onto the particle

the subsequent surface redox reactions with adsorbedtoms at the active sites of the photocatalyst. These stepsed by bulk, surface and electronic structure of the pho-herefore, several strategies to optimize photocatalyticave been performed with regards to those three steps,rystal growth and shape control [18,19], surface sen-r modication [2023], heterostructuring [21,24], andmetal doping [2527] or plasmonic metallic particle[2830] for visible light absorption.ing the crystal growth and shape is of special impor-rticles size, crystal phase and crystallinity, all of themecting the separation of charge carriers and the ef-otocatalysts. Tuning the crystal shape of a photocatalysta more efcient material when the most active crystalxposed. A highly crystalline structure of the semicon-o advantageous for photocatalytic reactions comparedus structures due to the higher mobility of excitediers within the defect-free solid. One of the key tasks

to synthesize materials with short diffusion lengthcarriers and abundant reaction sites by increasing

area. Surface sensitization and modication combine

exampwith cences photogseparation.

Theis anotion ana Z-schdifferesplittintrolytitwo oxstrongexcitatbility [to impthat evphases

Dopmoleculyst bychemicing theand no this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

r polymeric absorbers with semiconductor materials the absorption range. Using a molecular sensitizerited electrons are transferred from the lowest unoc-ecular orbital (LUMO) into the CB of the semiconductor2), undergoing further reduction reactions on the semi-urface. These systems are comparable to dye-sensitized[31], in which the sensitizer (dye) is regenerated byctrolyte. With polymeric surface modication, evenon of charge transfer complexes is possible, exhibit-xcitation from the polymeric VB into the CB of TiO2

table electron or hole scavenger is present on the sur-hotogenerated charges can also recombine, and the

into the systems is lost unused (Fig. 1). Addition-catalytic water splitting comes with a large positiveibbs free energy (RG = 237 kJ mol1) [10], meaning

thermodynamically point of view, the back reaction favored and has to be inhibited. Therefore, many

years. Especarbon has band gap se

This revof transitiotems, summDifferent dent possibito other revinuence othe distributronic structo other revbut cover alconductorsthe preferamixed oxidcussed and: electron donor; A: electron acceptor).

f chemical modication of semiconductor surfacestalysts [3236] or carbon scaffolds [37,38, and refer-in] (e.g. graphene) were reported, in order to retractated charge carriers form the VB or CB to spatiallyem from the semiconductor, and to inhibit recombina-

ation of photocatalyst heterostructures or compositesvery prominent example for charge carrier separa-ning the absorption properties of a photocatalyst. In

photocatalyst systems [39], two semiconductors withsorption range are combined to drive overall watera multiphoton process, and are connected via elec-ox shuttle. Heterostructured photocatalysts exhibitingr chalcogenide materials in close contact, both havingfering band positions for charge carrier separation aftercan readily reduce charge carrier recombination proba-hus, the lifetime of charge carriers is increased, leadingd photocatalytic performance. It was recently shown

composite of two closely related Ba-tantalate crystalimprove the charge carrier separation dramatically [40].with metal cations, non-metal anions or non-metalcan strongly enhance the absorption of a photocata-encing the electronic structure of a semiconductor. Theate of the dopant and the location are strongly inuenc-ctiveness of the doping procedure. Both options, metaletal doping, have attracted much attention in the last metal oxides for visible-light photocatalysis, Catal. Today

cially non-metal doping with e.g. nitrogen, sulphur andbeen investigated thoroughly, to decrease the Eg of widemiconductors into the visible light range.iew focuses on the recent progress of non-metal dopingn metal oxides, including ternary and quaternary sys-arizing the progress in the eld in the last ten years.

oping pathways will be presented, resulting in differ-lities for the origin of visible-light activity. In contrastiew articles, particular focus will also be given to thef an open crystal structure of the semiconductors ontion of dopants, and the resulting effects on the elec-ture of the doped photocatalyst. However, in contrastiew articles, we will not focus on few special examples,l types of non-metal doped transition metal oxide semi-. Homogeneous non-metal doping will be discussed asble strategy for Eg tuning. Especially non-metal dopede photocatalysts with open crystal structure will be dis-

reviewed as highly visible light active photocatalysts.

Please cite ition(2013), htt

ARTICLE IN PRESSG ModelCATTOD-8718; No. of Pages 25R. Marschall, L. Wang / Catalysis Today xxx (2013) xxx xxx 3

2. Non-metal doping of oxide photocatalysts

To dope or not to dope. This question was raised by Kamatin 2011 [41]. Yet, a conclusive answer has not been found yet. Thequestion itsdoping, anddoped phothighly activthe electrontion of the of the dopeedge of the pcan be absoity decreasecan be the can act as rdopant in trecombine photocatalythe dopantsor novel CBtion potentundoped m

More anities after between ththe doped pas we will sfacts aboutvisible-lighmetal dopinthe non-me(chapter 3)giving rate some typicanon-metal als exhibitiincluding pperformed in the given

2.1. Possibl

In 2001,doping confull-potentiism [27]. Thefciently otransfer of within theiof N with 2VB edge updoping of Sradius of S wlattice of TiOintroduced riers were dA nitrogen in an N2 (4N2 for abouing TiO2 in reported toand gaseou0.42% at 43metal dopinband gap n

materials. Several review articles have already discussed non-metaldoping of TiO2 [6,11,12,42], including TiO2 nanotube arrays [43]and density functional theory (DFT) [44]. Nitrogen is by far the mostinvestigated dopant for narrowing the Eg of semiconductor photo-

ts, th light

yeafor thtates

wasEg ofmbinorpot anoistener couch n ceansfeationt absitest froitrogies fr

areonr

thentri006of ventecancible l

inble odi

ighelso crea wnd 4ine w

2010quesn dn, reutione ce+ N3

nitrenertic Ns upo. Harge-s

s upoN-doe speed sted oradiated baraco fou

ng th this article in press as: R. Marschall, L. Wang, Non-metal doping of transp://dx.doi.org/10.1016/j.cattod.2013.10.088

elf can be asked either for metal doping or non-metal addresses two complementary effects in the eld ofocatalysts: semiconductors with large Eg are usuallye under UV light irradiation. But after doping, whenic structure has been modied to extend the absorp-

photocatalyst into the visible-light region, the activityd photocatalyst deteriorates: Although the absorptionhotocatalyst is red-shifted, meaning that more photonsrbed by the semiconductor, the photocatalytic activ-s under the same irradiation conditions. Several effectsreason for this behavior. Firstly, introduced dopantsecombination sites. Depending on the location of thehe semiconductor crystal, diffusing charge carriers canat dopant sites, and the input energy is lost. Thus, thetic activity is decreased. Secondly, without considering

as recombination sites, the discrete interstitial states and/or VB edges can reduce the reduction or oxida-ial of the modied semiconductor in comparison to theaterial, also resulting in decreased activity.d more examples are known showing increased activ-non-metal doping. Interestingly, a direct connectione homogeneity of the doping, the crystal structure ofhotocatalyst, and the resulting activity can be drawn,how in this review. We will rst summarize the known

non-metal doping in general, including the origin oft activity, and the importance of homogeneous non-g. We will then give a comprehensive overview abouttal doped semiconductor materials known in literature, including some highlights on important ndings, andconstants or quantum efciencies (were available) ofl photocatalytic reactions for comparison. The knowndoped transition metal oxide semiconductor materi-ng photocatalytic activity are summarized in Table 1reparation techniques and the photocatalytic reactionsunder visible light ( > 400 nm) irradiation as reported

references.

e origin for visible-light absorption

Asahi et al. calculated the effects of the substitutionaltents of C, N, F, P and S for O in anatase TiO2 using aal linearized augmented plane wave (FLAPW) formal-ey noted that the formed states upon doping shouldverlap with the TiO2 band states to ensure an efcientphotoexcited charge carriers to reactive surface sitesr lifetime. Thus, it was found that mixing of 2p statesp states of O results in forming a new VB with shiftedwards to narrow down the band gap of TiO2. Although

had resulted in a similar band gap narrowing, the ionicas reported to be too large to be incorporated into the2. Dopants C and P were found to be less effective as thestates were so deep that photo-generated charge car-ifcult to be transferred to the surface of the catalyst.doped TiO2 thin lm was prepared by sputtering TiO20%)/Ar gas mixture, followed by annealing at 550 C int 4 h. N-doped TiO2 powder was also prepared by treat-NH3 (67%)/Ar at 600 C for 3 h. The N-doped TiO2 was

be effective for methylene blue (MB) decompositions acetaldehyde decomposition (quantum yield, Q.E.:6 nm) under visible light ( > 400 nm). Since then, non-g with nitrogen is by far the most popular procedure forarrowing of semiconductors, especially for TiO2-based

catalysvisible

Twodence N 2p s[45]. Itin the as recothe inca slighThe exwas lat[49]. Snitrogetron trexplanble lighdefect and nolater, nvacancdefectslater crmingboth co

In 2origin Color cgen vaUV/vis

Alsoof visiment/mwith h[59], a[60]. U300 amelamTiO2.

In techninitrogetill thesubstittoactiv[O Ti4

and a photogmagnespecieof Ti3+

and ranspecietional the sollocalizenhanclight irpromotion chwas alsreduci metal oxides for visible-light photocatalysis, Catal. Today

us we will use this example to discuss the origins of activity.rs after Asahi, other research groups found new evi-e electronic structure of nitrogen doped of TiO2. Isolated

above the VB of TiO2 were determined experimentally also shown for the rst time that these nitrogen states

TiO2 thin lms prepared via magnetron sputtering actation sites [46], and that upon nitrogen doping, besidesrated isolated nitrogen states at a TiO2 electrode surface,dic shift of the CB edge can also be observed [47,48].ce of localized N 2p states above the VB in anatase TiO2nrmed by density functional theory (DFT) calculationslocalized states can also be generated by single-atomnters, playing an important role in photoinduced elec-r to reducible adsorbates [50]. In the same year, a third

was introduced by Ihara et al., concluding that the visi-sorption of nitrogen doped TiO2 originates from oxygen

in grain boundaries generated upon nitrogen doping,m the nitrogen states itself [51]. As it was calculateden doping decreases the formation energy of oxygenom 4.2 to 0.6 eV [52], making it very likely that oxygen

generated upon nitrogen doping. These results weremed by combined experimental and DFT results, con-

formation of oxygen defects upon nitrogen doping, andbuting to visible light photoactivity [53]., Serpone et al. provided another explanation for theisible light absorption upon nitrogen doping [54,55].rs are formed in the Eg by dopants associated with oxy-ies, which can be either formed via heat treatment oright irradiation [56,57].

2006, Mitoraj and Kisch determined the originlight activity of nitrogen-doped TiO2 after treat-cation with urea was related to surface sensitizationr melamine condensation products [58] like melonalled polyheptazine [22] or polymeric carbon nitrideas converted to ammonia and isocyanic acid between20 C in the presence of TiO2, and the resultingas further condensated to melon on the surface of

, Zhang et al. determined by several spectroscopic the photoactive centers for visible light activity inoped TiO2 [61]. In contrast to literature reports upporting Ti O N [62] or O Ti N [52,63] species foral or interstitial N-doping, they found that the pho-

nters of substitutional N-doped TiO2 are diamagnetic Ti4+ VO] clusters containing an oxygen vacancyogen anion. These clusters act as a source for theated electron transfer to reducible adsorbates. The dia-3 species are transformed into paramagnetic N2

n visible light irradiation, followed by the appearanceb et al. concluded in 2011, using spin-polarized DFTeparated hybrid HSE06 functional to calculate differentn N doping, that diamagnetic species in the substitu-ped TiO2 including oxygen vacancies (TiO(23x)N2x) arecies which induce an optical band-gap narrowing via

ates above the VB [64], and the only species explainingptical absorption in N-doped TiO2 material upon visibletion. The formation of these species is also chemicallyy using doping agents combining reducing and nitrida-ter, like NH3, N2H4, or (N2 + H2) mixture. Moreover, itnd that hydrogen can passivate the effects of nitrogen,e effective nitrogen concentration involved in visible

Please cite this article in press as: R. Marschall, L. Wang, Non-metal doping of transition metal oxides for visible-light photocatalysis, Catal. Today(2013), http://dx.doi.org/10.1016/j.cattod.2013.10.088


Table 1Visible light-active non-metal doped photocatalysts overview.

Pristinephotocatalyst

Non-metaldopant

Doping procedures Visible light photocatalyticreaction for

References

TiO2 N For synthesis details of Ndoping please

Dye degradation MB[27,68,92,96,101,102,104,108,110,117,119,123,129,132,147,279,335]

Compare chapter 3.1 RhB [103,120,130,131,134,268,272,331]Rhodamine G (RhG) [113], MO [116,267,337]Reactive Red 198 (RR198) [132]; sulforhodamine B (SRB) [152]

Terephthalic acid (TA)hydroxylation

[132]

Liquid phaseoxidation/degradation

2-Propanol [69,128]; KI [138]; 4-CP [47,48,112,145,330]; bisphenol A[283]; 2,4-dichlorophenol [134]; hydroquinone [112]; formic acid[112,146,151]; stearic acid [275,284]; trichloroethylene [112];benzamide [86]; nicotine [118]; microcystin-LR [127]; E. colisterilization [347]; 13C-acetone, 13C-ethanol, 13C-trichloroethylene[62]; pure MeOH dehydrogenation [151]; pure EtOH oxidation [241]

Hydroamination ofphenylacetylene

[152]

Gas phase oxidation Trichloroethylene [306]; NOx [121,124]; 2-propanol [45,124];acetaldehyde [27,47,97,112,133,306]; benzene [47,112]; CO [47,112];acetone [51,61]; ethylene [104]; propylene [150]; sweat acid [118]

H2 production From H2O/EtOH (Pt) [125,153]; from H2O/MeOH (Pt) [135]; fromH2O/EDTA-2Na (Pt) [269]; from H2O/Na2SO3 [139]; plasmonicenhanced [148]

Photoelectrochemicalreactions

H2O splitting [69,87]; CO2 reduction [93]; MO degradation [93]

Improved photocurrents [89,91,100,101,123,140,143,144,148,149,269]B Solvation/evap. Dye degradation Acid naphthol red (ANR) [266]

solgel Dye degradation MO [267]; RhB [268,272]Liquid phaseoxidation/degradation

Trichlorophenol [157]; 2,4-dichlorophenol [157]; sodium benzoate[157]; methyl tertiary butyl ether [159]; pentachlorophenol [162]

Hydrothermal Dye degradation RhB [270]TA hydroxylation [270]H2 production From H2O/MeOH (Pt) [163]O2 production From aq. AgNO3 [163]

SSR Liquid phaseoxidation/degradation

Phenol [160]

C Hydrolysis Dye degradation Remazol red [166]; MB [189]; RhB [182]Liquid phaseoxidation/degradation

Salicylic acid [166]; 4-CP [166]; stearic acid [284]; 4-CP [175,179,185]

Gas phase oxidation Acetaldehyde [166]; benzene [166]; CO [166]; NOx [178]Ultrasonication Dye degradation Reactive black 5 (Rb5) [193]


E. coli sterilization [193]; S. aureus sterilization [193]

TiC annealing Dye degradation MB [167]Liquid phaseoxidation/degradation

2-Propanol [165]; trichloroacetic acid [177]

Gas phase oxidation Toluene [195]Improved photocurrents [184]

Solgel Dye degradation MB [188,279]; RhB [193]Liquid phaseoxidation/degradation

4-CP [190]; Iodide [190]

Gas phase oxidation NOx [188]Improved photocurrents [190]

Microwave Dye degradation MB [194]Liquid phaseoxidation/degradation

S. aureus sterilization [194]

Hydrothermal Liquid phaseoxidation/degradation

bisphenol A [283]

CVD Gas phase oxidation NOx [183]Anodization Photoelectrochemical

reactionsH2O splitting [186]

Gas treatment Improved photocurrents [174,180,181]F Hydrothermal Dye degradation MB [328]


4-CP [199]

Gas phase oxidation Styrene [277]Solvation/evap. Dye degradation ANR [266]Solgel Liquid phase

oxidation/degradationFormic acid [245]

Spray pyrolysis Gas phase oxidation Acetaldehyde [197,198,306]; trichloroethylene [198,306]Cl Hydrothermal Dye degradation RhB [255]

Plasma oxidation Dye degradation MO [256]



Table 1 (Continued)


Non-metaldopant


References

I Hydrothermal Dye degradation MB [205]; RhB [209]TA hydroxylation [208,209]Liquid phaseoxidation/degradation

Phenol [211]; 4-CP [207]

Gas phase oxidation Acetone [208]CO2 reduction [216]

Solgel Dye degradation RhB [214]Liquid phaseoxidation/degradation

Phenol [217]; E. coli sterilization [217]; 2,4-dichlorophenol (2,4-CP)[214]

Hydrolysis Liquid phaseoxidation/degradation

Phenol [204,210]; 2-CP [213]

I2 Solgel Dye degradation MB [206]P Solgel Dye degradation MB [222,224,226,229]; MO [227]


4-CP [219,222,330]

Hydrothermal Dye degradation MB [228]; RhB [331]Hydrolysis Liquid phase

oxidation/degradationPhenol [230]

S TiS2 annealing Dye degradation MB [232,239]Thiourea SSR Dye degradation MB [234,235]


2-Propanol [235,236];

Adamantane hydroxylation [235]Hydrothermal Dye degradation MB [328,335]


4-CP [238]; phenol [248]; pure EtOH [241]

Solvothemral Liquid phaseoxidation/degradation

Nitrobenzene [243]

Solgel Dye degradation RhB [193,242]; Rb5 [193]Liquid phaseoxidation/degradation

Microcystin-LR [246]; E. coli sterilization [193]; S. aureus sterilization[193]; Micrococcus lylae sterilization [237]

Gas phase oxidation Acetaldehyde [299]SSR Dye degradation Bromocresol green (BCG) [247]

Peroxidation ofphosphatydilethanolAmine [106]

B-C Solvothermal Dye degradation Acid orange 7 (AO7) [258]TA hydroxylation [258]

Hydrothermal dye degradation RhB [260]liquid phaseoxidation/degradation

2,4-CP [259]

B-F Solgel Dye degradation MB [264]Solvation/evap. ANR [266]CVD Improved photocurrents [263]


MO degradation [263]

B-F-S Solvation/evap. Dye degradation ANR [266]B-N Solgel Dye degradation MB [276]; MO [267]; RhB [272]


Methyl tertiary butyl ether [159];

H2 production from H2O/EDTA-2Na (Pt) [269]Improved photocurrents [269]

Hydrothermal/NH3 Dye degradation RhB [268,270]TA hydroxylation [270]

Hydrolysis Liquid phaseoxidation/degradation

Stearic acid [275]

Hydrolysis/NH3 Improved photocurrents [274]Photoelectrochemicalreactions

H2O splitting [274]

C-F Hydrothermal Gas phase oxidation Styrene [277]C-N Solgel Dye degradation MB [279]; MO [281]


Microcystin-LR [288]

Hydrothermal Dye degradation MB [287]Liquid phaseoxidation/degradation

stearic acid [284]

Solvothermal Liquid phaseoxidation/degradation

Bisphenol A [283]

Ti2CN calcination Dye degradation MB [282]Mechanochemistry Gas phase oxidation NOx [280]

C-N-S Solgel Dye degradation Reactive red 2 (X-3B) [292]Liquid phaseoxidation/degradation

Tetracycline [294];

Hydrothermal Liquid phaseoxidation/degradation

Potassium ethyl xanthate [295,296]



Table 1 (Continued)


Non-metaldopant


References

Gas phase oxidation NOx [293]Solvation/evap. Gas phase oxidation Toluene [290]Thiourea SSR Gas phase oxidation Formaldehyde [291]

C-S Thiourea SSR Dye degradation MB [297]Liquid phaseoxidation/degradation

2-methylpyridine [297]

Hydrolysis Dye degradation MB [300]Liquid phaseoxidation/degradation

4-CP [298]; acetaldehyde [301]; E. coli sterilization [301]; B. subtilissterilization [301]

Gas phase oxidation Acetaldehyde [299]F-N Hydrothermal Dye degradation MB [315]; MO [322]


4-CP [307]

Supercritical cond. Dye degradation MB [316]Pyrolysis Dye degradation MB [317]Impregnation Dye degradation MB [321]Hydrolysis Dye degradation MO [308]Solgel Dye degradation MO [311]


4-CP [320]; microcystin-LR [314,318]

Spray pyrolysis Gas phase oxidation Acetaldehyde [305,306]; trichloroethylene [306]; NOx [324]NH4TiOF3 nitridation H2 production From H2O/MeOH (Pt) [309]

O2 production From aq. AgNO3, La2O3 [309,313]TiOF2 nitridation O2 production From aq. AgNO3, La2O3 [319]

F-P Anodization Improved photocurrents [326]F-S Solvothermal Dye degradation MB [328]N-P Hydrothermal/urea Dye degradation RhB [331]

Solgel Liquid phaseoxidation/degradation

4-CP [330]

N-S Hydrothermal/NH3 Dye degradation MB [335]Liquid phaseoxidation/degradation

Potassium ethyl xanthate [350]; pure EtOH [241]

Hydrolysis Dye degradation RhG [343]Liquid phaseoxidation/degradation

Phenol [338]

Gas phase oxidation Formaldehyde [336]Solgel Dye degradation MO [337](NH4)2TiO(SO4)2 dec. Dye degradation MO [346]

H2 production From H2O/MeOH (Pt) [346]Thiourea SSR TA hydroxylation [349]


Phenol [342]; 4-CP [349]; E. coli sterilization [341,342,347]

SSR H2 production From H2O/MeOH (Pt) [351]

SrTiO3 F Mechanochemistry Gas phase oxidation NOx [356,357]S-C SSR thiourea Liquid phase

oxidation/degradation2-Propanol [358]

N Mechanochemistry Gas phase oxidation NOx [361,362]Solvothermal Gas phase oxidation NOx [366]Solgel Dye degradation MB [367]; MO [367]; RhB [367]

N-S Mechanochemistry Gas phase oxidation NOx [373]

H2Ti4O9 N SSR urea Dye degradation RhB [378]

K2Ti4O9 N Urea solvation/evap. Dye degradation MB [379]

K2La2Ti3O10 N Annealing in NH3 Dye degradation RhB [380]HNO3/NH3/calc. Dye degradation MO [381]; cibacron yellows [381], erionyl red [381]

H2 production From H2O/MeOH [381]

La2Ti2O7 N Annealing in NH3 Dye degradation MO [382]H2 production From H2O (with and without Pt) [383]

Cs0.68Ti1.83O4 N Annealing in NH3 Dye degradation RhB [70]TA hydroxylation [70,71]

Ti0.91O2 N Annealing in NH3 Improved photocurrent [386]I2 Solution

impreg-nation/occulationDye degradation RhB [390]

Ta2O5 N Annealing in NH3 Gas phase oxidation 2-Propanol [393]; toluene [402]CO2 reduction with Ru-complex sensitizer [400]

Reactive sputtering Improved photocurrents [399]

NaTaO3 N Hydrothermal Dye degradation MB [405]; MO [406]Gas phase oxidation Formaldehyde [403]

Annealing in NH3 H2 production From H2O/MeOH, La co-dopant [408]I Hydrothermal Dye degradation MB [409]

K2Ta2O6 N Hydrothermal Gas phase oxidation Formaldehyde [404]



Table 1 (Continued)


Non-metaldopant


References

KTa0.92Zr0.08O3 N Annealing in NH3 Overall water splitting From H2O (Pt) [410]

InTaO4 N Annealing in NH3 Gas phase oxidation 2-Propanol [411]

Sr2Ta2O7 N Annealing in NH3 H2 production, AM 1.5 From H2O/MeOH (Pt + graphene) [415]H2 production From H2O/MeOH (Pt) [419]O2 production From aq. AgNO3, La2O3 (CoOx) [419]

CsCaTa3O10-nanosheets

N Annealing in NH3 O2 production From aq. AgNO3, La2O3 [417]

N Annealing in NH3 H2 production From H2O/MeOH (Pt/Ru/Rh) [418]O2 production From aq. AgNO3 [418]

Ba5Ta4O15 N Annealing in NH3 H2 production, AM 1.5 From H2O/EtOH (Pt) [72]H2 production From H2O/MeOH (Pt) [419]O2 production From aq. AgNO3, La2O3 (CoOx) [419]

Sr5Ta4O15 N Annealing in NH3 H2 production From H2O/MeOH (Pt/Rh/Ir) [419]O2 production From aq. AgNO3, La2O3 (IrO2/CoOx) [419]

Nb2O5 N Annealing in NH3 Gas phase oxidation 2-Propanol [393]SSR urea Dye degradation RhB [425]

HNb3O8 N SSR urea Dye degradation RhB [378,425]

HNb3O8-SiO2 N SSR urea Dye degradation RhB [426]

Sr2Nb2O7 N Annealing in NH3 H2 production From H2O/MeOH (Pt) [75]O2 production From aq. AgNO3, La2O3 [75]

NaNbO3 N Annealing in NH3 Gas phase oxidation 2-Propanol [424]

Nb6O174 N Photo-N-doping H2 production From H2O/MeOH (Pt) [427]

HTiNbO5-TiO2 N SSR urea Dye degradation [428]

WO3 N2 Thermolysis Photoelectrochemicalreactions

HCl oxidation [439]; O2 production [439]

N Thermal decomp. Gas phase oxidation MeOH oxidation [440]Annealing in NH3 Dye degradation MO [442]


MO oxidation [442]

Improved photocurrents [441,442]C Spray pyrolysis Improved photocurrents [443]S SSR thiourea Improved photocurrents [447]

O2 production From aq. Fe3+ [447]

Zr2W2O8 S Thiourea/EtOH O2 production From aq. AgNO3 [457,75]

Bi2WO6 F Hydrothermal Dye degradation (sim. sunlight) RhB [458]HF Dye degradation MB [459]

N Hydrothermal Dye degradation RhB [460,461]Gas phase oxidation Acetaldehyde [460]Improved photocurrents [460]

I Hydrothermal Dye degradation RhB [464]

CsTaWO6 N Annealing in NH3 TA hydroxylation [73]H2 production From H2O/EtOH, no co-catalyst [73]

N-S Annealing S/NH3 TA hydroxylation [74]H2 production From H2O/EtOH, no co-catalyst [74]

ZnO N Decomposition Dye degradation MO [470]liquid phase reduction Cr2O72 [470]

Annealing in NH3 Improved photocurrents [474,478]ZnOHF nitridation O2 production From aq. AgNO3, La2O3 (with/without IrO2) [482]

Improved photocurrents [482]

BiVO4 PO4 oxoanion Precipitation Improved photocurrents [493]O2 production From aq. AgNO3 [493]

F Hydrothermal Dye degradation RhB [494]Liquid phaseoxidation/degradation

Phenol [495]

S aq. thiourea Dye degradation MB (CoOx) [496]

Co3O4 F PE-CVD H2 production (sim. sunlight) From H2O/EtOH, no co-catalyst [497]

LaCoO3 C Solgel CO2 reduction From aq. Na2CO3 [498]

Nb2Zr6O17 N Annealing in NH3 H2S decomposition From KOH (Pt, RuO2) [500]

In2Ga2ZnO7 N SSR H2 production From H2O/MeOH (Pt) [501]

-Fe2O3 N Sputtering Improved photocurrents [502]



Fig. 2. 6 type tor fosemiconducto he Eg;diamagnetic in

gen cReprinted with

light photocomplex wa

The diffeare shown s

All thesenon-metal dbeen identistates in theing O2p anthese studidefects in tto some exthe dopantwith the amin photocatbetter visibity (with andifferent fo

In 2012, tion to unveTiO2 [66]. Sto a titaniudefect NOTiband thermmined. Howyet to be pr

Very reccalculationssubstitutionThey founding correlaboron givinlocalized sting lled sTi3+ centersof uorine agle nitrogenreducing thing to redu

s (seoratits in e poug

havtentiaterifacton th

sible

n thoped T of schemes depicting the electronic situation after modication of a semiconducr with narrowed Eg; (C) localized states below CB; (D) color centers formed in tterstitial N species plus oxygen vacancies.

Fig. 3. Synergistic effects of uorine and nitro permission from reference [67]. Copyright 2013 Elsevier.

catalytic activity [65]. However, this passivating N Hs reported to be dissociable at modest temperatures.rent concepts for doping/modifying TiO2 with nitrogenchematically in Fig. 2.

mechanism can explain the visible light absorption ofoped photocatalysts, however a general origin has noted yet. Especially evidences for both localized N 2p

Eg above the VB or the formation of a new VB by mix-d N 2p states is found in literature until today. Whates also show is that non-metal doping always createshe oxide semiconductor, e.g. oxygen vacancies, whichtent can also act as recombination sites in addition to

itself. The amount of defects increases quantitativelyount of dopant, which can lead to successive decreasealytic activity. Thus, an optimum doping amount forle light absorption and enhanced photocatalytic activ-

procesincorpdopanthe sam

Althdopingless athost mthese effect oalysts.

2.2. Vi

Evegen do this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

acceptable amount of defects) has usually to be foundr a specic photocatalyst system.Umezawa and Ye used DFT plus onsite Coulomb interac-il the most stable nitrogen defects in N-doped anataseubstitutional nitrogen on an oxygen site strongly bindsm atom on an interstitial site. A stable and complexi was identied, and the formation of a defect-impurityally connected with the host VB maximum was deter-ever, experimental evidence for this explanation has

ovided.ently, Di Valentin and Pacchioni used periodic DFT

to investigate the electronic structure upon oxygen in TiO2 by non-metal dopants B, C, N and F [67].

that the position of the localized states upon dop-tes with the effective nuclear charge of the element,g states high in the Eg, nitrogen doping resulting inates just above the VB, and uorine doping generat-tated below the O2p VB leading to the formation of

due to charge compensation. Additionally, co-dopingnd nitrogen was proposed to be advantageous over sin-

doping, since both dopants act as donor acceptor paire number of intrinsic defects in TiO2 after doping, lead-ced electronhole recombination in the photocatalytic

[27], the chnext to theBurda et al.610 nm [6dramatic, shthe visible l

These rcan be obseafter non-mwith steep excitation, reason for t

The nitrpared by trammonia gBurda et alnitrogen prhomogeneoof nitrogencharacterisprepared inmagnetronIn accordanr Eg narrowing: (A) localized states above VB; (B) non-metal doped (E) surface modication with nitrogen-containing compounds; (F)

o-doping in TiO2.

e Fig. 3). Also, uorine doping facilitates the subsequenton of nitrogen, leading to increased amounts of nitrogenthe lattice compared to pure nitrogen doped TiO2 underreparation conditions.h the changes in the electronic structure after non-metale been thoroughly investigated in the last decade, muchon has been paid to the crystal structure effects of theal and the homogeneity of the dopants. We will addressrs in the following chapter, since both have a lastinge photocatalytic activity of non-metal doped photocat-

light absorption bands and homogeneous doping

ugh Asahi et al. presented the rst example on nitro-iO2 with increased light absorption in the visible region metal oxides for visible-light photocatalysis, Catal. Today

anges were quite minor, and only a small shoulder absorption edge of TiO2 was visible (Fig. 4). In 2003,

prepared nitrogen doped TiO2 nanoparticles of around8]. Here, the changes in the absorption spectra were veryowing a complete red-shift of the absorption edge intoight region.st two examples from literature showed already whatrved in reports today, that the resulting absorption bandetal doping can either have a shifted absorption edge

and parallel characteristics, also called a band-to-bandor only a small shoulder and tail. But what is the mainhe different behaviors?ogen doped TiO2 lms from Asahi et al. had been pre-eating the pure TiO2 thin lms post-synthetically inas. In contrast, the nitrogen doped nanoparticles by. had been prepared via solgel synthesis, adding theecursor (triethylamine) to the synthesis gel ensuring aus distribution of the dopant. Hence, the distribution

in the nal material might be crucial for the absorptiontics of non-metal doped semiconductors. Kitano et al.

2006 nitrogen doped TiO2 thin lms via radio frequency sputtering using a N2/Ar mixture as sputtering gas [69].ce with Asahi et al., the introduction of nitrogen was



2; leftReprinted with s showform Ref. [68],

Fig. 5. Homog represin this gure l

performed of the shiftein contrast preparationtant, since isputtering r

A third vdoping wasTo prove thmain inueLiu et al. inva layered titalso treatednitridation layered strufavors the homogeneothe semicondopants waping. The abexcitation oregion, thetrast, nitrogto doping ostructure. Anarrowing dO 2p states.

In summexcitation cthe dopant a small absFig. 4. Different absorption behaviors of nitrogen doped TiO permission from Ref. [27] Copyright 2001 AAAS; right: nitrogen doped nanoparticle

Copyright 2003 American Chemical Society.

eneous doping in layered material vs. surface doping in bulk materials (yellow dots egend, the reader is referred to the web version of the article.) this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

post-synthetically, however a band-to-band excitationd absorption edge with nitrogen content was observedto the results of Asahi. This result indicates that the

technique for the nitrogen doping is equally impor-n both cases the TiO2 substrate was identical. Physicalealized homogeneous nitrogen doping in this case.ery important factor for the effectiveness of nitrogen

found to be the crystal structure of the host material.e theory that homogeneity of nitrogen doping is thence on the resulting electronic band structure of TiO2,estigated nitrogen doping by treatment in ammonia onanate material, Cs0.68Ti1.83O4 [70]. In comparison, they

commercial TiO2 nanoparticles (P25) under the sameconditions, and they found that the lepidocrocite-typecture of Cs0.68Ti1.83O4 with unique lamellar structures

diffusion of ammonia into the materials, leading to aus distribution of substitutional nitrogen throughoutductor particles. The homogeneous distribution of thes conrmed by XPS depth proles and elemental map-sorption spectrum therefore exhibited a band-to-bandf the absorption deeply shifted into the visible light

resulting Eg was determined to be 2.73 eV. In con-en doped P25 only showed an absorption shoulder duenly on the particle surface, and not throughout thedditional spin-polarized DFT calculation conrmed Egue to a newly formed VB, consisting of mixed N 2p and

ary, a shift of the absorption edge with band-to-bandan only be achieved by homogeneous doping. Whendistribution is inhomogeneous e.g. surface doping, onlyorption shoulder or tail will be observed. Furthermore,

the nature othe dopantis sensitive doping, sev

(i) In the cceduredeposit

(ii) In the cment ineeds tstitutiostructuThese nitride

(iii) If the hstructuprimardopingdoping

(iv) In situ to a homassure rities twith cational c

Fig. 5 cobulk photolayered struof shorter : nitrogen doped TiO2 thin lm. complete red-shift of the absorption edge, reprinted with permission

ent non-metal dopants). (For interpretation of the references to color metal oxides for visible-light photocatalysis, Catal. Today

f extended band-to-band absorption is independent on amount, but the threshold of the extended absorptionto dopant concentration [71]. To achieve homogeneouseral pathways are possible:

ase of post-synthetically doping, the experimental pro- needs to be of high energy e.g. physical sputteringion, so that a homogeneous doping can be achieved.ase of post-synthetically nitrogen doping via heat treat-n ammonia, the crystal structure of the host materialo be penetrable for the gas, so that a homogeneous sub-n of lattice oxygen can occur. Materials with layeredre or otherwise open crystal structure are favorable.conditions are also applicable for the oxidation ofs or sulphides to get nitrogen or sulphur doped oxides.ost material for nitrogen doping has no open crystalre, homogeneous doping can only be achieved when they particle size is usually less than 10 nm, when surface

of such small particles becomes equal to homogeneous throughout the particle [68].doping during the material preparation can also leadogeneous distribution of dopants, however one has to

that pure nitrogen doping without any additional impu-akes place. For example, in the case of in situ dopingrbon-containing nitrogen precursors like urea, addi-arbon impurities can remain in the material.

mpares the different doping situations of layered andcatalysts. In contrast to bulk photocatalysts with non-cture, layered photocatalysts also have the advantage

diffusion pathways for charge carriers and improved



WO6Reprinted with ructu

charge carrThe oxidation the surfaholes may elight excitaother speci[11]. Note ain terms ofthe main adpossibility owhere seve

The ideaalready beealysts. Muk(1 1 1)-layegas stream band-to-badoping, redlayered matgen generatunder sunliundoped la0.1 wt.% Ptrmed the states, shiftunchanged

More exbe given into a non-laCsTaWO6 wheat treatmstructure ooctahedra, section (Figrandomly inthe crystall

The chantageous forA distinct swith band-the resultin2.4 eV. The improved pylation and(20 vol.%, 173 mol/h/

ol/h witthe dto oter, tAsahey aarroe cr

. Thu to ootocandu

strur in ow chano berschulph

absducer co-n/su

ligh) eve300 W

amo crysicoFig. 6. (left) UVvis absorption spectra of (A) CsTa permission from Ref. [73] Copyright 2011 Wiley-VCH, Weinheim; right: crystal st

ier separation of photogenerated electronhole pairs.on and reduction sites are isolated, being present eitherce and the edges of the thin unit sheets. Electron and/orasily diffuse to the surface of the unit sheets from the

tion cites, where they can react with interlayer water ores (if present), separating the holes from the electrons

layered structure might not guarantee an advantage light penetration compared to non-layered materials,vantages lie in the guest species penetration and thef subsequent delamination (please compare chapter 3ral layered photocatalysts will be discussed).

of optimized and homogeneous non-metal doping hasn proven to be applicable to other layered photocat-herji et al. prepared nitrogen doped Ba5Ta4O15 withred perovskite structure via heat treatment in ammonia[72]. An extraordinary red shift of the absorption withnd excitation character was observed after nitrogenucing the Eg down to 1.78 eV. After nitrogen doping, theerials showed considerably increased activity in hydro-ion from aqueous methanol solution (20 vol.%, 100 mL)ght irradiation (300 W AM1.5G, 1 sun) compared to theyered tantalate (495 mol/h/0.1 g vs. 305 mol/h/0.1 g,; 42 mol/h/0.1 g at > 420 nm). DFT calculation con-nature of Eg narrowing upon mixing of N 2p and O 2ping the VB maximum to lower potential, leaving the CB.amples for non-metal doped layered photocatalysts will

chapter 3. This strategy was however already appliedyered material exhibiting an open crystal structure.ith defect-pyrochlore structure was nitrogen-doped by

(26 mformed

All cable HowevWhen [27], thlar Eg ninto thradiusradiusthe phsemicocrystalples fowill shslight can alsing. Mawith sing thewas resulphunitrogevisible100 mLtively,

Theon thethe sem this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

ent in ammonia gas stream [73]. The defect-pyrochloref CsTaWO6 consists of corner-sharing TaO6 and WO6forming parallel tunnels each with a hexagonal cross. 6). Both Ta and W ions are considered to be distributed

the framework, and the cesium atoms are located onographic 8b sites.nel structure of CsTaWO6 turned out to be as advan-

gas phase nitrogen doping as for the layered materials.hift of the absorption edge into the visible light rangeto-band excitation was observed after nitrogen doping,g Eg was determined to be reduced from 3.8 eV down toresulting photocatalyst CsTaWO6xNx showed stronglyhotocatalytic activities for terephthalic acid hydrox-

hydrogen generation from aqueous ethanol solution00 mL) under simulated sunlight (147 mol/h/0.1 g vs.0.1 g, 300 W AM1.5G, 1 sun) and pure visible light

ture shouldsince the cent crystal the crystal the photocet al. investvia ammonand 1273 Kical variablthe layeredtemperaturnon-layeredphase compence of non

By usinof ammonand (B) CsTaWO6xNx .re of CsTaWO6 with pathways for nitrogen doping.

/0.1 g, > 420 nm). Hydrogen generation was even per-hout any noble metal co-catalyst.iscussed trends for nitrogen doping are equally appli-her non-metal dopants and to non-metal co-doping.he size of the dopant atom needs to be considered.i et al. performed their theoretical calculation in 2001lready predicted that nitrogen and sulphur show simi-wing, but sulphur would be difcult to be incorporatedystal structure of TiO2 because of the larger ionics, the non-metal dopant should have comparable ionicxygen to ensure a homogeneous distribution withintalyst. This assumption is of course valid for all oxide

ctors, not only for TiO2; however the rigidity of thecture also needs to be taken into account. Some exam-situ sulphur doping of TiO2 are already known, as wein chapter 3. When the crystal structure can acceptges in the lattice parameters, a homogeneous doping

achieved using e.g. sulphur via post-synthetical dop-all et al. doped the mixed-oxide photocatalyst CsTaWO6ur, and reported band-to-band excitation after shift-orption edge into the visible light region [74]. The Egd to 2.7 eV, and could be further reduced by nitrogen-doping down to 2.06 eV. Both materials, sulphur andlphur co-doped CsTaWO6, generated hydrogen undert ( > 420 nm) from aqueous ethanol solution (20 vol.%,n without co-catalyst (17 and 21 mol/h/0.1 g, respec-

Xe lamp AM1.5G, > 420 nm).unt of non-metal dopant can have a profound effecttal structure. To still call the chemical modication ofnductor oxide a doping procedure, the crystal struc- metal oxides for visible-light photocatalysis, Catal. Today

remain unharmed. This is of inevitable importance,omparison of two photocatalysts exhibiting a differ-structure after doping would be pointless. Only whenstructure remains, the effect of non-metal doping onatalytic activity can be unambiguously determined. Jiigated nitrogen doping of the layered niobate Sr2Nb2O7ia treatment at different temperatures between 973

[75]. They found that the crystal structure was a crit-e for their photocatalytic activities, since they changed

niobate material gradually with increasing reactione into an oxynitride (SrNbO2N) material with cubic,

crystal structure. As mentioned before, a change inosition impedes a thorough investigation of the inu--metal doping on photocatalytic activity.g higher reaction temperatures or higher ow ratesia, even pure nitrides can be formed from oxide



semiconductors. A famous example would be the formation ofTaON and Ta3N5 from Ta2O5 [8]. Although the Eg can be steadilynarrowed by extended nitridation, according to a shift of the VBmaximum of the materials to less positive potentials, the crys-tal structurmaterials h

In the nematerials aof non-metmaterials wexisting rev

3. Non-me

3.1. TiO2

TiO2 is mconductorsonly simplepowders [8[43,85]. Andnon-metal d

TiO2 cananatase andhedra with connected brookite viaEg of 3.2 ean inuenc

As alreato report nof effective[27]. They cthe electronpared nitrogInspired byfor non-meconductor o

N-dopedods: by spuby pulsed deposition solgel memal methodby mechan[138,139] oies, also pu[44,49,64,6

Shi et alanatase TiOtroscopy, inwas describsion of oxygions in thefor the re-owas estimafrom the Ar((1.96 0.1release hasof the Eg updiffusion m600 C, at a ature. The rtheir nitrog

Cao et al. used PLD to prepare selectively local nitrogen dopedTiO2 thin lm photoelectrodes [95]. By changing the reactiongas during the preparation, they prepared TiO2 thin lms withnitrogen incorporated close to the back electrode (120 nm thick

but thic TiOurre(SHEeneodes t-phUV rnsity

ascrn-in

can, andresupingts in ntialomptheirent i-bon

nanles, odamles.

of throge

et titalenets didn do/disc

inte that ogenari

tase ic an doce the e

sepataly86 mthe e

veryion oO15omp

ides ing b03], 8,99o re).on d

TiO2ifted this article in press as: R. Marschall, L. Wang, Non-metal doping of transp://dx.doi.org/10.1016/j.cattod.2013.10.088

es change as well [76,77]. Thus, the activity of theseas to be interpreted individually.xt chapters, we will mainly focus on non-metal dopednd the inuence of crystal structure on the efciencyal doping. Oxynitrides, oxysulphides and comparableill not be considered, we refer the readers to otheriew articles [8,14,78,79].

tal doped transition metal oxides

aybe the most investigated photocatalyst of all semi-. It can be prepared in many different morphologies, not

powders, but also dense thin lms [80], mesoporous1] and mesoporous thin lms [8284], and nanotubes

all of these examples have been already modied withopants to reduce their Eg.

crystallize in three different structure types; rutile, brookite. The crystal structures consist of TiO2 octa-different interconnection: in rutile, the octahedral arevia two common edges, in anatase via four, and in

three. That results in different Eg, anatase exhibits anV, and rutile of 3.0 eV. The initial Eg has of course alsoe on the Eg narrowing upon non-metal doping.dy mentioned in chapter 2, Asahi et al. were the rston-metal doping of TiO2 underlining the importance

overlap between dopant states and TiO2 band statesalculated the effect of different non-metal dopants onic structure of TiO2, but experimentally they only pre-en-doped mixed phase (rutile/anatase) TiO2 thin lms.

the work of Asahi, nitrogen is the most popular dopanttal doping, not only for TiO2, but for all kinds of semi-xide photocatalyst.

TiO2 can be prepared by many different meth-ttering [27,46,69,86,87], by ion implantation [8891],laser deposition PLD [53,9295], by physical vaporPVD [96], from TiN [97103], by hydrolysis or

thods [47,48,5052,62,63,68,104127], by solvother-s [128133], by microemulsion techniques [134,135],ochemistry [136,137], or by heat treatment in urear NH3 [45,61,140153]. Besides experimental stud-re theoretical studies were reported on N-doped TiO27,154156].. investigated the oxidation kinetics of nitrogen doped2 thin lms prepared via PLD by in situ optical spec-

pure oxygen or nitrogen [94]. The oxidation kineticsed by a parabolic rate law, either controlled by the diffu-en vacancies or by the interstitial diffusion of titanium

oxidized part of the thin lm. The activation energyxidation process of nitrogen doped anatase thin lmsted from the rate constants at different wavelengths,rhenius plot an activation energy of (189 12) kJ mol12) eV) was calculated. The reaction kinetics of nitrogen

also been determined. The time dependant blue shifton nitrogen release via oxidation was interpreted by aodel. The nitrogen release was complete after 30 min atsimilar time found for the re-oxidation at this temper-e-oxidation of nitrogen doped anatase thin lms anden release are therefore strongly correlated.

layer),(80 nmdopedphotoctrode homogelectroincidenin the rent deusuallynitrogedopingtrodesThese gen dodopandiffere

In city of treatmsurfaceof themolecusulforhmolecusurfacethe nittion.

Kimtypicaldiethyauthornitrogechargedue toshowsfor nitr

Etchof anatraacetnitrogeto reduwhile tcarrierphotoc(k = 0.0about with aevolutBa5Ta4lamp, c[40].

Besinclud[1962phur [9ways tTable 1

Bordopedwas sh metal oxides for visible-light photocatalysis, Catal. Today

without any nitrogen doping at the electrode surfacek), conrmed by EDS mapping. The inner nitrogen2 photoelectrodes exhibited signicantly increasednt (400 A/cm2 at 1.23 V vs. standard hydrogen elec-) under 1 sun) compared to undoped (150 A/cm2),usly doped (100 A/cm2) or outer doped photo-(50 A/cm2), and also demonstrated much higheroton-to-current efciencies (IPCE, up to 95% at 320 nm)egion. The authors showed that, since poor photocur-

of a homogeneously nitrogen doped TiO2 electrodes isibed to the electronhole recombination caused by theduced state below the conduction band, local nitrogen

lead to completely different properties of photoelec- to improved photocurrents related to water splitting.lt also indicated that poor photoactivities after nitro-

cannot purely attributed to overall nitrogen doping, butdifferent photoelectrode locations need to be discussedly.arison, Zheng at al. improved the photocatalytic activ-

nitrogen doped anatase TiO2 nanobers (from heatn ammonia gas) via post-treatment in air to removeded N species [152]. By that procedure, the surfaceobers became more effective for adsorbing organicleading to improved photoinduced degradation ofine B, and also to improved activation for oxygen

These properties were attributed to a restored anatasee nanobers after the second calcination in air, whereas

n remains inside the nanobers for visible light absorp-

al. prepared nitrogen doped TiO2 nanobers from anium isopropoxide sol including small amounts ofriamine via electrospinning [126]. Unfortunately, the

not investigate the photocatalytic properties, but theirped TiO2 nanobers showed improved conductivity andharge characteristics for lithium ion batteries, but hererstitial nitrogen near the surface of the nanobers. Itdepending on the application, different types of location

doping are favorable.et al. prepared nitrogen doped TiO2 heterojunctionsand rutile via solgel route using ethylenediaminete-cid (EDTA) as nitrogen source [119]. Substitutionalping into the lattice of the TiO2 phases was determinedhe Eg of anatase and rutile for visible light absorption,xistence to the heterojunction lead to improved chargearation upon light irradiation, leading to improvedtic activities in the photoinduced degradation of MBin1 compared to 0.004 min1 for P25). This resultffect of a photocatalyst heterojunction is in agreement

recent report about improved photocatalytic hydrogenf a Ba5Ta4O15/Ba3Ta5O15 composite compred to pure(1885 mol/h/0.5 g, in 600 mL of MeOH/H2O, 350 W Hgared to 800 mol/h/0.5 g of pure Ba Ta O compound)

nitrogen doping, doping TiO2 with other non-metalsoron [115,157163], carbon [98,99,164195], uorineiodine [201,204217], phosphorus [218230] and sul-,170,193,223,231249] have been reported as effectiveduce the Eg of TiO2 into the visible light range (see

oping was rst reported by Zhao et al., preparing B-by solgel technique using boric acid [157]. The Eg

to 2.93 eV via boron doping, and was further reduced



Fig. 7. Schemair, 2 h, heatin

Reprinted with

by performrmed the was succesdoped and activity thato Cl afterof boron-dosubstitutionor as intersare possibledonor withcoexist, whet al. contrmicrospheroxygen evopared with doping. Upboron conteoped as a reshell (Fig. 7

The boris consistendoping is ledoping alsoward shift by band beration (5TiO2 microshydrogen (Pt).

Carbon dwith tetrabprepared Tition shouldof differentchlorophento initial 4-Xe lamp, ddoped TiO2with subseqTiO2 nanotuand more e(>420 nm) tprepared aanatase anmicrowave

identied upon carbon doping reducing the Eg of the TiO2 com-posite, and the increased photocatalytic activity of the compositein visible light (MB degradation, 0.008 min1, 0.004 min1 for P25,

catalyst in 50 mL MB solution [105 mol/L]) was attributedffectrookides ith caup oith hly sh

actieadinorineed bya mixered9,252n conin Tirystan [25

mod fromia hySubssiblehe sag phationn [2 ed a

multe latt

of Tion fible l > 42% miolved to remety baes, be

TiO22], dansficaenti

impatic evolution of boron distribution of upon heat treatment (600 C ing rate 5 C/min).

permission from Ref. [163] Copyright 2012 Wiley-VCH, Weinheim.

ing Ni B co-doping to 2.85 eV. XPS measurements con-existence of Bi O and Ti B bonds, and trichlorophenolsfully decomposed under visible light illumination. B-Ni-B co-doped TiO2 showed both higher photocatalyticn the undoped TiO2 (80% of total chloride converted

4 h). Finazzi et al. calculated the electronic structureped anatase TiO2 [161]. Boron can be incorporated asal boron for oxygen which gives rise to midgap states,

titial boron where coordinations of boron with oxygen. Boron on interstitial sites behaves as three-electron

formation of B3+. Both types of doping can howeverich could also result in internal charge transfer. Liuolled the spatial dopant distribution of boron in TiO2es to modulate photocatalytic hydrogen evolution andlution preferences [163]. TiO2 microspheres were pre-a boron-containing core exhibiting substitutional boronon calcination in air, a boron gradient with maximumnt at the outer surface of the microspheres was devel-sult of thermal diffusion of boron from the core to the).on-doped shell consisted of interstitial boron, whicht with results from Finazzi that substitutional boronss stable than interstitial boron doping [161]. Boron

lead to no reduction of the overall Eg, but a down-of the electronic band edges of the boron-doped shellending effects resulted in preferential oxygen gen-0 mol/h/0.1 g, 300 mL AgNO3 [16.7 mmol/L]), whilepheres with boron-doped core preferentially generated75 mol/h/0.1 g, 300 mL MeOH/H2O [10 vol.%], 1 wt.%

0.06 g to the eCB of b

Besed wthe grotion wrecentcan beTiO2, l

Flureportide in consid[67,19TiO2. Iphase high cductio

Theknownet al. v[204]. and viTiO2. Tshowindegradsolutioimprovpared into ththe CBtransitthe vislamp, 5 h, 83are invreducemeasuimpuri2p statdopedVB [21their trbe signet al. idcrucial this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

oping was reported by Kisch et al. by hydrolysis of TiCl4utylammonim hydroxide [166]. After calcination, theO2 contained carbon and exhibited a signicant absorp-er in the visible light range. Photocatalytic oxidation

gaseous and liquid model compounds including 4-ol (4-CP) was shown under visible light irradiation (upCP disappeareance rates of 7.65 108 mol/L s, 150 Wifferent cut-off lters). Park et al. prepared carbon-nanotube array photoanodes by anodizing titanium foiluent annealing in CO gas ow [174]. The carbon-dopedbe arrays showed much higher photocurrent densities

fcient water oxidation under visible-light illuminationhan pure TiO2 nanotube arrays. Etchari et al. recently

composite photocatalyst system of carbon-dopedd carbon-doped brookite through non-hydrothermal-assisted synthesis [194]. Interband C 2p states were

anatase TiOI O Ti statbelow the CWhile the imodicatiofor photocabeyond 600adducts (I2)ing powderabsorption sensitizer aCB of TiO2, odation, 0.05irradiation)

Besides reported fo metal oxides for visible-light photocatalysis, Catal. Today

ive electron transfer of photogenerated electrons in theite to the CB of anatase.carbon doping, the surface of TiO2 can also be modi-rbonaceous species acting as sensitizers, as shown byf Kisch [250,251], comparable to the surface sensitiza-igher melamine condensation products [58]. It was alsoown that under visible light irradiation even benzenevated to form carbonaceous polymeric deposits ontog to visible light photocatalytic activity [23].-doped TiO2 with anatase and brookite phase was rst

Ye et al. via hydrolysis of titanium tetraisopropox-ed NH4F-water solution [196]. Fluorine doping is now

to not cause any shift in the absorption edge of TiO2,253], but is responsible for Ti3+ defect generation intrast, NH4F doping stabilizes the most active anataseO2 up to high calcinations temperatures, and ensuresllinity for enhanced photoinduced charge carrier pro-4].ication of TiO2 with other halogen elements is also

literature. Iodine doping of TiO2 was reported by Hongdrolysis of tetrabuthyl titanate in the presence of HIO3titutional I5+ ions were identied as dopant species,

light absorption was achieved upon iodine doping ofmples were used for photocatalytic phenol degradation,otocatalytic activity under visible light irradiation (59%

after 2 h, 47 % mineralized, 0.1 g catalyst in 100 mL MB 105 mol/L], 18 W lamp, > 420 nm), and also stronglyctivities under UVvis light irradiation. Su et al. pre-ivalent iodine doped TiO2, incorporation both I7+ and I

ice of anatase TiO2 [208]. Iodine interband states belowiO2 were identied via DFT calculations, and electronrom the TiO2 VB to the interband states contribute toight activity of the iodine-doped TiO2 (500 W tungsten0 nm, gaseous acetone oxidation, 94% degradation in

neralized to CO2). However, since surface-IO4-speciesd in the photocatalytic reaction of this materials andI during reaction, the material deactivates in long termnts. By rst-principle calculations, He et al. identied annd above the VB of rutile TiO2, consisting of I 5s and Oing responsible for the visible light absorption of iodine

. This impurity band was fully overlapping with the TiO2elocalizing photoexcited charge carriers and favoringer. As a result the Eg of iodine doped rutile TiO2 was canntly reduced, absorbing visible light up to 480 nm. Liued near-surface I O I and I O Ti structures to be ofortance for efcient photocatalysis with iodine-doped2 [209]. Photonic excitation occurred from occupiedes just above the VB of TiO2 to unoccupied I O I statesB of TiO2, resulting in light absorption up to 800 nm.

ntrinsic Eg still was measured to be 3.14 eV, the surfacen with the two different iodine species was responsibletalytic activity in Rhodamine B (RhB) degradation even

nm. Usseglio et al. encapsulated iodine molecules andn in nanocavities of mixed phase TiO2 [206]. The result-

exhibited a red/brown color, and a strongly red-shiftedshoulder. In this case, the (I2)n adducts acted as a dyebsorbing light and injecting excited electrons into then whose surface the photocatalytic reaction (MB degra-

g in 50 mL MB solution [1 104 mol/L], 24 h sunlight occurred.uorine and iodine, chlorine and bromine have also beenr doping TiO2 [201,252,255257]. Luo et al. prepared



ClBr-co-doped TiO2 nanoparticles with different phase composi-tions in a solvothermal process using TiCl4, hydrobromic acid andethanol [257]. The Eg of TiO2 was narrowed after ClBr-co-dopingdown to 2.85 eV for co-doped rutile TiO2. The highest activity foroverall watNa2CO3 aqulamp) was a30% rutile, bH2, 0.3 wt.%another exalyst compocarbon-dop

Phosphoing non-ordapproach [2mixture, P5

structures. phosphor dwas observanatase TiOacid [219]. Aafter phosplight irradiaanions usinabsorption surface of Tcesses durinusing phospfrom H3PO4tigations prPhotocatalydegradation(Q.E.: 0.66%

Sulphur et al. [231],be as efciedation of Tiafter sulphureported. Brmed beinand co-worcations into(1 mmol/m(5 mmol/h/signicant ble light raTiO2 with S[244]. Withanion-dopedoping wasabsorption.

Besides doping andin literaturenation of pB/F [26326[278289], [326], F/S [doping and

Especialtion, since Lmaterial bygen dopantthe materiacore. IntersTi O bonds

by nitrogen from ammonia gas treatment. Additionally, no Ti3+

defects due to nitrogen doping were formed in this red anatase, dueto charge compensation by boron. Photocurrent measurements ofprepared photoanodes were performed under visible light irradi-

givinnghtas rescopse an

also TiO2tionsopinerefo

is coining

tanat

en Tinthaomend cromi

struts for

uon [3UV isibl

00 nmlengtditiof SrF2ivityeparh sut actibon aheyith

y undpositgatedope

O 2Pd S

C,S d the ubstbly sites tion.ogensed e lathe asserv

Sinceen 837orouore gals wl ora this article in press as: R. Marschall, L. Wang, Non-metal doping of transp://dx.doi.org/10.1016/j.cattod.2013.10.088

er splitting (stoichiometric ratio of H2:O2 2:1) in aeous solution (93.2 g in 400 mL) in UV light (450 W Hgchieved by co-doped TiO2 containing 70% anatase andeing higher than that of Evonik P25 (140 mol/h/0.3 g

Pt, compared to 65 mol/h/0.3 g H2 for P25). This ismple for the favored electron transfer in photocata-

sites, comparable to the already shown examples fored TiO2 [194] and Ba5Ta4O15 [40].r doped TiO2 was rst reported by Yu et al., prepar-ered mesoporous TiO2 in a triblock-copolymer assisted18]. By adding phosphoric acid (H3PO4) to the reaction+ was introduced into the TiO2 lattice, forming Ti O PAlthough the mesoporous structure was stabilized byoping, no extended absorption in the visible light rangeed. Lin et al. reported yellow-colored phosphor doped2 prepared by solgel method with hypophosphorousn absorption tail in the visible light range was observedhor doping, and 4-CP was decomposed under visibletion. Iwase et al. prepared anatase TiO2 doped with P3

g phosphide precursors [230]. A clear red-shift in thewas observed after P3 doping, but PO4-species on theiO2 were still detected arising from the oxidation pro-g synthesis. However, the absorption features of P-TiO2hides are quite different to those of P-TiO2 prepared. Moreover, rst-principle calculations and XPS inves-oved the substitutional P3 doping for lattice oxygen.tic activity in pure visible light was found for phenol

(up to k 0.005 h1) and acetaldehyde degradation at 410440 nm wavelength).doping of TiO2 was rst performed by Umebayashi

after Asahi had predicted that sulphur doping wouldnt in Eg narrowing of TiO2 as nitrogen [27]. After oxi-S2 in air, they showed that the Eg of TiO2 was narrowedr doping, a clear shift in the absorption spectrum wasy ab initio band calculations, Eg narrowing was con-g caused by mixing of S 3p states with the TiO2 VB. Ohnokers rst incorporated sulphur as S6+ [234] or S4+ [235]

TiO2, gaining visible light activity for MB degradationin at 440 nm wavelength) and 2-propanol oxidation

0.1 g). The latter was prepared using thiourea, and ashift of the absorption was observed far into the visi-nge. Zheng et al. compared substitutional S doping ofe and Te doping using rst-principle DFT calculations

increasing atomic number, it becomes harder to formd TiO2, but easier to achieve cationic doping. Anionic

furthermore found to be more effective for visible light

single non-metal doping, a variety of non-metal co- metal/non-metal co-doping strategies were reported

to modify the electronic structure of TiO2. The combi-ossible non-metal co-dopants includes B/C [258262],5], B/F/S [266], B/N [115,159,267276], C/F [277], C/NC/N/S [290296], C/S [297302], F/N [303325], F/P327,328], I/N [329], N/P [330334], N/S [335353] co-

OCN doping [354].ly B/N co-doping has recently attracted much atten-iu et al. were able to prepare red-colored anatase TiO2

controlling the spatial distribution of boron and nitro-s inside anatase TiO2 microspheres [274]. The Eg ofls varied from 1.94 eV on its surface to 3.22 eV in thetitial boron doping of TiO2 microspheres weakened the

for additional and easy substitution of lattice oxygen

ation, wavele

It wspectrothis cait wasdopedcalculaafter ding. Thsystemand ga

3.2. Ti

Wh(Ln = lastoichitions amost povskitedopanformedreactiounder ity in vthan 4wavenrial. Adtime othe actples pr1 h withighes

Car[358]. Tously wactivitdecominvestication-sists ofC 2p angap of rmedthree spreferaon Sr sabsorp

NitrThey uinto thwith twas obof NO.have bies [36mesopmesopmaterimethy metal oxides for visible-light photocatalysis, Catal. Today

g IPCE values of 0.8% to 0.15% from 400 nm to 500 nm, and 0.1% up to 600 nm.cently reported that electron paramagnetic resonancey (EPR) is a powerful tool to investigate non-metal (inion) doped TiO2 and co-doped TiO2 [355]. However,

stressed that only careful investigation of non-metalby multiple characterization techniques and theoretical

can give detailed insight into the materials propertiesg, and the changes in photocatalytic activity after dop-re, a wider perspective would be needed in which thensidered as whole, abandoning the view on a local site

insights especially in charge delocalization.

es

O2 reacts with other metal oxides (e.g. SrO, BaO, Ln2O3nide)) usually by sintering at high temperatures intric mixtures, metal titanates with different composi-rystal structures can be obtained. Being one of thenent titanates for photocatalysis, SrTiO3 exhibiting per-cture has been already doped with different non-metal

visible-light absorption (see Table 1). Wang et al. per-rine doping of SrTiO3 in 2003 by mechanochemical

56]. The photocatalytic NO oxidation was investigatedand visible light, and a signicantly increased activ-e light up to three time higher at wavelengths longer

for F doped SrTiO3 was obtained, two times higher aths longer than 510 nm compared to the undoped mate-nally, the same group investigated the effect of grindingwith SrTiO3, calcination time and uorine content on

of uorine doped SrTiO3 for NO oxidation [357]. Sam-ed by grinding mixture of 5% SrF2 and 95% SrTiO3 forbsequent heat treatment at 150 C for 2 h showed thevities (40% NO oxidation).nd sulphur doping of SrTiO3 was reported by Ohno et al.

used thiourea as precursor and doped SrTiO3 simultane-sulphur cations and tetravalent carbon. Photocatalyticer visible light irradiation was reported for 2-propanolion, with up to 1.2 mmol/h/0.1 g consumed. Liu et al.d the electronic structure and optical properties of C, Sd SrTiO3 by DFT [359]. The VB of C,S doped SrTiO3 con-

orbitals mixing with C 2s and S 3s orbitals, while Ti 3d,3p contribute to the CB. Narrowing of the optical bandoped SrTiO3 was also conrmed by DFT. Li and Yao con-inuence of carbon doping of SrTiO3 depending on theitutional positions [360]. Carbon doping was found toubstitute Ti rather than Sr or O. However, also carbon

and on oxygen reduces the Eg, resulting in visible light

doped SrTiO3 was rst reported by Wang et al. [361].a mechanochemical approach to incorporate nitrogentice of SrTiO3 by mixing different nitrogen precursors-prepared oxide in a ball mill. Photocatalytic activityed for NO oxidation in visible light, oxidizing up to 47.8%e then, several studies about nitrogen doped SrTiO3published [362367], including pure theoretical stud-2]. For example, Zou et al. prepared nitrogen-dopeds SrTiO3 using glycine as both nitrogen source andenerator in a template-free synthesis [367]. The nalere able to decompose model dye compounds like RhB,nge (MO) (40% degradation after 180 min) and MB

Please cite ition metal oxides for visible-light photocatalysis, Catal. Today(2013), htt


(80% degradation after 80 min) under pure visible light irradiation( > 420 nm). Incorporation of nitrogen additionally led to oxygenvacancies and Ti3+ centers in SrTiO3 conrmed by EPR measure-ments. Eg narrowing of about 0.3 eV was therefore concluded tooccur due tand oxygen

SrTiO3[372374] aLa N co-doto generatelight irradiaproduced d

As descrgen dopingspaces to acgen in thestructure hlayered titaachieve vislar titanatenitrogen vinitrogen doH2Ti4O9 fro2.5 eV, resphelped to exchanged acidbase ithe heatingmaterials sdiation for more visibldegradationXe lamp, ity of the smolecules tpair separatpared to thewhich showtion after 5acidity and doping.

Xiong eand TiO2 aanatase nancompletelyfacets idenanatase [18visible-lighelectronhoactivity in vdegradation

The advaby Kumar ewith Ruddleused urea K2La2Ti3O1as precursointo the latwere obserby Huang eammonia sthe carbonshift in absand enhanclight in comK2La2Ti3O1

ayer-gold c

d wit

ng eO7 wO7 pting

in thnitridg rean w

to shotoation

lamtocaol/h

gle-laep smon-do

otoninatio

prepi0.91sed with(PEI)yer ht irradiation compared to multilayer lms prepared withed titania nanosheets.ng this sequential layer-by-layer deposition technique, mul-

hybrid lms of nitrogen-doped Ti0.91O2xNx nanosheetsing in situ reduced Au nanoparticles (Fig. 8) [387] or

ted hematite nanocubes [388] could be prepared. The Au-ing hybrid thin lms showed plasmon resonance due to theoparticles. Such Au nanoparticle hybrid lms might be veryin photocatalytic reactions, efciently separating electronsles under illumination due to gold acting as electron trap,as demonstrated for mesoporous TiO2 thin lms containinganostructures [389].91O2 nanosheet suspensions prepared from undopedi1.83O4 can also be mixed with Iodine solutions, resulting ineneously I2-modied Ti0.91O2 nanosheets [390]. I2-Ti0.91O2eets can be occulated with protons, and exhibit visiblebsorption with an Eg of 2.38 eV. Due to the homogeneousution of I2 dopants within the whole bulk of occulated this article in press as: R. Marschall, L. Wang, Non-metal doping of transp://dx.doi.org/10.1016/j.cattod.2013.10.088

o nitrogen-related energy states slightly above the VB, vacancy-related state below the CB minimum.has also been co-doped with multiple non-metalsnd with non-metal/metal cations [374,375], includingping [376] and Cr N co-doping [377]. The latter is able

hydrogen from H2O/MeOH solution under pure visibletion with a Q.E. of 3.1% at 420 nm (Q.E.: number of H2ivided by the number of absorbed photons).ibed in chapter 2.2, in 2009 Liu et al. reported nitro-

of a layered titanate, Cs0.68Ti1.83O4, using its interlayerhieve homogeneous distribution of substitutional nitro-

crystal lattice [70]. The advantages of this layeredave been described earlier. Since then, several othernates have been modied with non-metal dopants toible light photocatalytic activity. Li et al. doped lamel-

acid H2Ti4O9 and the layered titanate K2Ti4O9 witha solid state reaction with urea [378]. The Eg of theped samples were reduced after nitrogen doping form 3.1 eV to 2.1 eV, and for K2Ti4O9 down from 3.4 eV toectively. The layered crystal structure of the materialsintercalate urea, the acidic character of the proton-solid acid further improved the intercalation due tonteraction, helping to stabilize the crystal structure in/doping process. Both nitrogen-doped layered titanatehowed photocatalytic activity under visible light irra-RhB photodegradation, the protonated sample absorbse light and shows superior photoactivity (90% of RhB

after 100 min, 0.3 g, 100 mL solution [12 mg/L], 300 W > 420 nm), which was attributed to the high acid-olid acid, leading to facilitated intercalation of waterhrough hydrogen bonding for enhanced electronholeion at the interlayer surface. The results were also com-

nitrogen-doped layered niobate solid acid HNb3O8-N,ed even higher photocatalytic activity (100% degrada-0 min, 0.3 g, 100 mL solution [12 mg/L]) due to higherbetter retained layered crystal structure upon nitrogen

t al. doped composites of layered K2Ti4O9 nanobeltsnatase nanorods with nitrogen from urea [379]. Theorods formed a shell around the titanate nanobelts

covering its surface, and with dominantly exposedtied as {0 0 1} planes, the highly reactive facets of]. The nitrogen-doped composite showed enhancedt absorption, excellent charge carrier mobility and lowle recombination, leading to improved photocatalyticisible-light MB (k = 1.4 102 min1) and benzoic acid

compared to the undoped composite.ntage of a layered titanate crystal structure were usedt al. to dope the (1 0 0) layered perovskite K2La2Ti3O10sden-Popper structure with nitrogen [380]. When theyas nitrogen precursor, they modied the surface of0 with carbon nitride polymers, but using ammonia gasr led to incorporation of small amounts of nitrogentice at 400 C. Only minor red shifts in the absorptionved after ammonia treatment, which was also reportedt al. after immersion of K2La2Ti3O10 in nitric acid andolutions with subsequent calcination [381]. However,

nitride modied samples showed a remarkable redorption, the Eg was narrowed from 3.63 eV to 2.92 eV,ed photocatalytic activity for RhB degradation in visibleparison to the ammonia treated sample and unmodied0 was reported.

Fig. 8. Lanionic

Reprinte

MeLa2Ti2La2Ti2exhibichangewhen and lonnitrogeleadingered pdegradvisiblefor pho(25 m

Sintwo-stalkylamnitrogeing prdelamtion toSince Twere usition imine multilaible ligundop

UsitilayercontaindeposicontainAu nanactive and hoas is wgold n

Ti0.Cs0.68Thomognanoshlight adistribby-layer self assembly of PEI with N-doped titania nanosheets andomplex ions.

h permission from Ref. [387] Copyright 2012 Elsevier.

t al. doped the (1 1 0) layered perovskite materialith nitrogen using ammonia gas [382,383]. The barehotocatalyst was prepared via hydrothermal methodnanosheet morphology. Upon ammonia treatment, noe crystal structure was observed, which can be the caseation is performed using high gas ows (1 dm3 min1)ction times (>15 h) forming LaTiO2N [384]. 23.3 at.% of

ere detected via XPS after nitrogen doping of La2Ti2O7,ignicant red shifts of the absorption edge of the lay-catalyst. Photocatalytic activities were reported for MO

in pure visible light (90% degradation after 10 h, 148 Wp, 10 mg in 10 mL MO solution [5 mg/mL] [382]), andtalytic hydrogen generation in pure visible light, with/g) and without (20 mol/h/g) Pt co-catalyst [383].yer titanate nanosheets can also be prepared e.g. byynthesis, protonation and delamination with bulkynium salts [385]. Liu et al. applied this strategy onped layered Cs-titanate (Cs0.68Ti1.83O4xNx), perform-

exchange of cesium cations using HCl solution, andn with tetrabutylammonium hydroxide (TBAOH) solu-are individual exfoliated Ti0.91O2xNx nanosheets [386].O2xNx nanosheets have a negative surface charge, theyas building blocks for sequential layer-by-layer depo-

the positively charged polymers such as polyethylene via dip-coating on conducting glass substrates. Theselms showed enhancements in photocurrent upon vis-



Fig. 9. [Ru(bpvisible light irr

Reprinted with

Ti0.91O2 nanelevated, alble light absactivity of of RhB (80%[4 105 mlight irradia

3.3. Ta2O5 a

Ta2O5 hneeded for water in UV1 wt.% RuONiOx (1154[391,392]. TTa2O5 was din ammoniaof Ta2O5xNgen (x = 0.10temperaturatures 80crystal strupropanol in700 h), and(Q.E. from 1irradiated li

Several malytic appliet al. foundnia treatmsensitized ble light-insemiconducfrom acetonnitrogen doactivity. TEOgen doped Tselectivity f405 nm was

The chaalso been defer from shRu complexfor the charin Ta2O5xNthe determcomplexes.Ta2O5xNx

Ullah et aNb2O5 nan

decomposition (300 mg of catalyst, 100 ppm gas ow, 300 W Xelamp) [402]. Ta2O5xNx showed much better performance in thisreaction than Nb2O5xNx, either in articial sunlight (70% vs 15% ofdecomposition) or pure visible light (30% decomposition), which

tribuies.

rsthermg maried

diffed inS, and ve. Nander2 h fsed lattising 6 naowimpo

lightp, >ntentivityhermheremogaO3sis usorp-lighals e

in 5n [2

lamp lighto et t ph

UV lia doptivitunatd aft-dopnera

+ 220hermationAs a mouith t

waty)2(CO)2]2+ sensitized nitrogen doped Ta2O5 for CO2 reduction underadiation.

permission from Ref. [400] Copyright 2010 Wiley-VCH.

osheets, the VB maximum and width are respectivelyso conrmed by DFT calculations, increasing both visi-orption and hole mobility, giving rise to photocatalyticocculated I2-Ti0.91O2 nanosheets in the decomposition

removal after 60 min, 50 mg in 100 mL RhB solutionol/L], 300 W Xe lamp, > 420 nm) under pure visibletion.

nd tantalates

as an Eg of 3.94.0 eV, thus UV light irradiation isphotocatalytic activity. It can photocatalytically split

light into stoichiometric amounts of H2 and O2 using2 (32/17 mol/h/g, 350 mL, 400 W Hg lamp) or 1 wt.%/529 mol/h/g, 350 mL, 400 W Hg lamp) as co-catalystso expand its adsorption into the visible light range,oped with different amounts of nitrogen via annealing

gas atmosphere [393]. The red shift of the absorptionx was dependent on the amount of incorporated nitro-, 0.17, 0.24, 0.35), which was controlled via annealinge (600, 620, 650, and 700 C, respectively). At temper-0 C TaON was formed accompanied with a change incture. Nitrogen doped Ta2O5 was able to decompose 2-

pure visible light in long term measurements (up to the activity also depended on the amount of nitrogen.3% for x = 0.1 to 0% for x = 0.35) and the wavelength ofght.ore studies about nitrogen doped Ta2O5 for photocat-

cations have been reported [394398], after Morikawa p-type conduction and Eg narrowing after ammo-

ent [399]. Sato et al. used nitrogen doped Ta2O5with Ru-(2,2-bipyridine)-based complexes for visi-duced CO2 reduction to formic acid [400]. Only thetor-complex hybrids showed activity for CO2 reductionitrile/triethanolamine (TEOA) solution (Fig. 9), neitherped Ta2O5 nor the bare Ru-complexes gave rise to anyA was found to be acting as electron donor for nitro-a2O5, and as proton source for the Ru-complexes. Theor formic acid was more than 75%, and a Q.E. of 1.9% at

reported.

was atvacanc

Thehydrotstartinwas vania forresulteby XPshowe550 nmhyde uafter 1decreaber of way uK2Ta2Ocess shwith covisibleXe lamgen cothe achydrotatmospto inho

NaTsyntheThe abvisiblemateridationsolutio250 Wvisible

Katefcienunder bine Lhigh acUnfortreachethe cogen geMeOHhydrotdegrad[409]. small aalyst woverall this article in press as: R. Marschall, L. Wang, Non-metal doping of transitionp://dx.doi.org/10.1016/j.cattod.2013.10.088

rge carrier dynamics and electron transfer rates havetermined for such systems [401]. For the electron trans-allow traps in nitrogen doped Ta2O5 to the adsorbedes, a time constant of 12 1 ps was determined, whilege trapping process from shallow to deep defect sitesx a time constant of 24 1 ps was found, explaining

ined reaction pathways of CO2 reduction at the Ru- No electron transfer was found from Ru complex towhen the complex was directly excited.l. compared nitrogen doped Ta2O5 with nitrogen doped

oparticles in terms of photocatalytic gaseous toluene

300 W Xe, Nitrogen

also preparphase chantion shouldwere obserdoped InTaoxidative 2-diation (Q.Eof InTaO4 wIt was fou metal oxides for visible-light photocatalysis, Catal. Today

ted to a larger number of defect states and oxygen

example of nitrogen doped NaTaO3 was prepared viaal synthesis [403]. Ta2O5xNx and NaOH were used asterials, and the amount of nitrogen doping in NaTaO3

by using nitrogen doped Ta2O5 annealed in ammo-rent temperatures. Hydrothermal treatment at 200 C

pure NaTaO3 phase, nitrogen content was conrmedd the optical absorption of nitrogen doped NaTaO3ry broad absorption shoulders with absorption up toTaO3xNx was able to decompose gaseous formalde-

visible light irradiation (maximum 80% degradationor x = 0.047, 500 W Xe lamp, &g

Documents

Non-metal doping of transition metal oxides for visible-light photocatalysis