DOI: 10.1126/science.1200448, 746 (2011);331 Science, et al.Xiaobo Chen

Hydrogenated Titanium Dioxide NanocrystalsIncreasing Solar Absorption for Photocatalysis with Black

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): April 11, 2011 www.sciencemag.org (this infomation is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/331/6018/746.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2011/01/19/science.1200448.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/331/6018/746.full.html#ref-list-1, 2 of which can be accessed free:cites 22 articlesThis article

http://www.sciencemag.org/cgi/collection/chemistryChemistry

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2011 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

20. R. H. Baughman, S. Stafström, C. Cui, S. O. Dantas,Science 279, 1522 (1998).

21. U. Mitzdorf, D. Helmreich, J. Acoust. Soc. Am. 49, 723 (1971).22. G. V. Yukhnevich, E. G. Tarakanova, J. Mol. Struct. 447,

257 (1998).23. R. Lakes, K. W. Wojciechowski, Physica Stat. Solidi B 245,

545 (2008).24. J. N. Grima, D. Attard, R. Gatt, Physica Stat. Solidi B 245,

2405 (2008).

25. A. L. Goodwin, Z. Krist. 1 (suppl. 30), 1 (2009).26. We thank I. Wood for useful discussions; the Institut

Laue Langevin and the ISIS facility for beam time; andJ.-L. Laborier, C. Payre, and L. Gendrin for invaluabletechnical assistance in carrying out the high-pressuremeasurements at the ILL. This work was funded underfellowships from STFC (ex Particle Physics andAstronomy Research Council, PPARC) grant nos.PPA/P/S/2003/00247 and PP/E006515/1.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6018/742/DC1Materials and MethodsFigs. S1 to S14Tables S1 to S18References

4 October 2010; accepted 5 January 201110.1126/science.1198640

Increasing Solar Absorption forPhotocatalysis with Black HydrogenatedTitanium Dioxide NanocrystalsXiaobo Chen,1,2 Lei Liu,1,3 Peter Y. Yu,1,3 Samuel S. Mao1,2*

When used as a photocatalyst, titanium dioxide (TiO2) absorbs only ultraviolet light, and several approaches,including the use of dopants such as nitrogen, have been taken to narrow the band gap of TiO2. Wedemonstrated a conceptually different approach to enhancing solar absorption by introducing disorder inthe surface layers of nanophase TiO2 through hydrogenation. We showed that disorder-engineeredTiO2 nanocrystals exhibit substantial solar-driven photocatalytic activities, including the photo-oxidationof organic molecules in water and the production of hydrogen with the use of a sacrificial reagent.

The effectiveness of solar-driven photo-catalytic processes underlying hydrogenproduction and water decontamination is

dictated to a great extent by the semiconductor’scapability of absorbing visible and infrared light,as well as its ability to suppress the rapid com-bination of photogenerated electrons and holes.Nanophase titanium dioxide (TiO2), which has alarge surface area that can facilitate a fast rate ofsurface reactions, is a widely used wide–band-gap semiconductor photocatalyst for a variety ofsolar-driven clean energy and environmental tech-nologies (1–4). In order to increase the limited

optical absorption of TiO2 under sunlight, therehave been persistent efforts to vary the chemicalcomposition of TiO2 by adding controlled metal(5, 6) or nonmetal (7–10) impurities that generatedonor or acceptor states in the band gap. Differentfrom impurity incorporation, self-doping that pro-duces Ti3+ species in TiO2 has also been demon-strated (11). Through doping, the solar absorptioncharacteristics of TiO2 have been improved tosome extent. For example, when nonmetalliclight-element dopants are introduced (9), the op-tical absorption of TiO2 can be modified as theresult of electronic transitions from the dopant2p or 3p orbitals to the Ti 3d orbitals. At present,nitrogen-doped TiO2 exhibits the greatest opticalresponse to solar radiation (3), but its absorptionin the visible and infrared remains insufficient.

We developed an alternative approach toimproving visible and infrared optical absorptionby engineering the disorder of nanophase TiO2

with simultaneous dopant incorporation. In itssimplest form, a disorder-engineered nanophaseTiO2 consists of two phases: a crystalline TiO2

quantum dot or nanocrystal as a core, and a high-ly disordered surface layer where dopants are in-troduced (Fig. 1A). Although an ensemble of

1LawrenceBerkeley National Laboratory, University of Californiaat Berkeley, Berkeley, CA 94720, USA. 2Department of Mechan-ical Engineering, University of California at Berkeley, Berkeley,CA 94720, USA. 3Department of Physics, University of Californiaat Berkeley, Berkeley, CA 94720, USA.

*To whom correspondence should be addressed. E-mail:ssmao@lbl.gov

Fig. 1. (A) Schematicillustration of the struc-ture and electronic DOSof a semiconductor inthe form of a disorder-engineered nanocrystalwith dopant incorpo-ration. Dopants are de-picted as black dots,and disorder is repre-sented in the outer lay-er of the nanocrystal.The conduction and va-lence levels of a bulksemiconductor, EC andEV, respectively, are al-so shown, and the bandsof the nanocrystals areshown at the left. Theeffect of disorder, whichcreases broadened tailsof states extending intothe otherwise forbiddenband gap, is shown at

20 30 40 50 60 70 80

(b) black TiO2

XR

D In

ten

sity

/ a.

u.

2 Theta / degree

(a) white TiO2

C

200 400 600 800 1000

938.

3

849.

1

765.

5

690.

1

352.

9

294.

224

6.9

199.

3

639.

9

517.

1

397.

4

146.

2

(b) black TiO2

Ram

an In

ten

sity

/ a.

u.

Wavenumber / cm-1

(a) white TiO2

D

A B

Ec

EV

E

F

the right. (B) A photo comparing unmodified white and disorder-engineered black TiO2 nanocrystals. (Cand D) HRTEM images of TiO2 nanocrystals before and after hydrogenation, respectively. In (D), a short-dashed curve is applied to outline a portion of the interface between the crystalline core and thedisordered outer layer (marked by white arrows) of black TiO2. (E and F) XRD and Raman spectra of thewhite and black TiO2 nanocrystals.

11 FEBRUARY 2011 VOL 331 SCIENCE www.sciencemag.org746

REPORTS

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

nanocrystals retains the benefits of crystallineTiO2 quantum structures for photocatalytic pro-cesses, the introduction of disorder and dopant attheir surface would enhance visible and infraredabsorption, with the additional benefit of carriertrapping. Large amounts of lattice disorder insemiconductors could yield mid-gap states whoseenergy distributions differ from that of a singledefect in a crystal. For example, instead of formingdiscrete donor states near the conduction bandedge, these mid-gap states can form a continuumextending to and overlapping with the conduc-tion band edge; thus they are often also known asband tail states. Similarly large amounts of dis-order can result in band tail states merging withthe valence band (12–15). These extended ener-gy states, in combination with the energy levelsproduced by dopants, can become the dominantcenters for optical excitation and relaxation. Anadditional potential advantage of these engi-neered disorders is that they provide trapping sitesfor photogenerated carriers and prevent them fromrapid recombination, thus promoting electron trans-fer and photocatalytic reactions. The density of states(DOS) of disorder-engineered semiconductornanocrystals, as compared to those of unmodifiednanocrystals, is shown schematically in Fig. 1A.

To introduce disorders into nanophase TiO2

with simultaneous dopant addition, we generateda porous network of TiO2 nanocrystals, a fewnanometers in diameter. Hydrogenation of thismaterial creates a disordered layer on the nano-crystal surface. We observed a shift in the onsetof absorption in such disorder-engineered TiO2

nanocrystals, from the ultraviolet (UV) to near-infrared after hydrogenation, accompanied by adramatic color change and substantial enhance-ment of solar-driven photocatalytic activity. Aphoto of disorder-engineered black TiO2 nano-crystals, as compared to one of unmodified whiteTiO2 nanocrystals, is shown in Fig. 1B.

We prepared TiO2 nanocrystals with a pre-cursor solution consisting of titanium tetraisopro-poxide (TTIP), ethanol, hydrochloric acid (HCl),deionizedwater, and an organic template, PluronicF127, with molar ratios of TTIP/F127/HCl/H2O/ethanol at 1:0.005:0.5:15:40. The solution washeated at 40°C for 24 hours and then evaporatedand dried at 110°C for 24 hours. The dried pow-ders were calcinated at 500°C for 6 hours toremove the organic template and enhance thecrystallization of TiO2. Both the temperatureramp rate and the cooling rate were approximate-ly 0.3°Cmin–1. The resulting white-colored pow-

ders were first maintained in a vacuum for 1 hourafter being placed in the sample chamber of aHy-Energy PCTPro high-pressure hydrogen sys-tem and then hydrogenated in a 20.0-bar H2

atmosphere at about 200°C for 5 days. Becausehydrogen tends to be attracted to dangling bonds,we expected the concentration of hydrogen to bethe highest in the disordered layer, where there aresubstantially more dangling bonds than in thecrystalline core of black TiO2 nanoparticles.

We investigated the structures of the TiO2

nanocrystals before and after hydrogenation withx-ray diffraction (XRD), Raman spectroscopy, andscanning and transmission electron microscopy(SEM and TEM). The pure TiO2 nanocrystalswere highly crystallized, as seen from the well-resolved lattice features shown in the high-resolution TEM (HRTEM) image (Fig. 1C); thesize of individual TiO2 nanocrystals was approx-imately 8 nm in diameter. After hydrogenation,however, the surfaces of TiO2 nanocrystals be-came disordered (Fig. 1D) where the disorderedouter layer surrounding a crystalline corewas ~1 nmin thickness. StrongXRDdiffraction peaks (Fig. 1E)also indicate that the TiO2 nanocrystals werehighly crystallized. The crystalline phase had ananatase structure with an average crystal size ofapproximately 8 nm, in agreement with HRTEMobservation.

We used Raman spectroscopy to examinestructural changes in the TiO2 nanocrystals afterthe introduction of disorder with hydrogenation.The three polymorphs of TiO2 belong to differentspace groups: D4h

19(I41/amd ) for anatase, D2h15

( pbca) for brookite, and D4h14(P42/mnm) for ru-

tile, which have distinctive characteristics in Ramanspectra. For anatase TiO2, there are six Raman-active modes with frequencies at 144, 197, 399,515, 519 (superimposed with the 515 cm–1 band),and 639 cm

–1

, respectively (3). The unmodifiedwhite TiO2 nanocrystals display the typical anataseRaman bands, but new bands at 246.9, 294.2,352.9, 690.1, 765.5, 849.1, and 938.3 cm–1 emergefor the black TiO2 nanocrystals, in addition to thebroadening of the anatase Raman peaks (Fig. 1F).These Raman bands cannot be assigned to any ofthe three polymorphs of TiO2, which indicatesthat structural changes occur after hydrogenation,resulting in disorders that can activate zone-edgeand otherwise Raman-forbidden modes (such asmodes that are infrared-active only) by breakingdown the Raman selection rule (13).

The solar-driven photocatalytic activity of thedisorder-engineered black TiO2 nanocrystals wasmeasured by monitoring the change in optical ab-sorption of a methylene blue solution at ~660 nmduring its photocatalytic decomposition process.Other than being a nitrogenous reference com-pound for evaluating photocatalysts, methyleneblue can be found as a water contaminant fromdyeing processes. In a typical experiment, 0.15mgof black TiO2 nanocrystals was added to a 3.0-mlmethylene blue solution that had an optical den-sity (OD) of approximately 1.0 under aerobic con-ditions; the results were corrected for methylene

1.0

0.8

0.6

0.4

0.2

0.00 10 20 30 40 50

Irradiation Time / min Irradiation Time / min

O.D

. / a

.u.

O.D

. / a

.u.

00.0

0.2

0.4

0.6

0.8

1.0

0 0 0 0 0 00

B

C

AWhite TiO2Black TiO2

Time / day

Time / hour

H2

gen

erat

ion

/ m

mo

l 1.0

0.8

0.6

0.4

0.2

0.00 10 20 30 40 50 60 70 80 90 100

63 9 12 15 18 21

Fig. 2. (A) Comparison of the solar-driven photocatalytic activity of the black TiO2 nanocrystals with that ofthe white TiO2 nanocrystals under the same experimental conditions. The y axis represents the optical densityof the methylene blue solution, whereas the x axis is the solar light irradiation time. a.u., arbitrary units. (B)Cycling tests of solar-driven photocatalytic activity (methylene blue decomposition) of the disorder-engineeredblack TiO2 nanocrystals. Data in the figure represent the first 8 min of measurements in each of the eightconsecutive photodegradation testing cycles. (C) Cycling measurements of hydrogen gas generation throughdirect photocatalytic water splitting with disorder-engineered black TiO2 nanocrystals under simulated solarlight. Experiments were conducted in a 22-day period, with 100 hours of overall solar irradiation time.

www.sciencemag.org SCIENCE VOL 331 11 FEBRUARY 2011 747

REPORTS

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

blue degradation in the absence of any photo-catalyst [see the supporting onlinematerial (SOM)and fig. S1]. Photodegration was complete after8 min for the black TiO2 nanocrystals, whereasfor the unmodified white TiO2 nanocrystals un-der the same testing conditions, it took nearly1 hour (Fig. 2A). Similar improvement was alsoobserved for the photocatalytic decomposition ofphenol (SOM and fig. S2).

The results of cycling tests of the solar-drivenphotocatalytic activity of black TiO2 nanocrystalsin decomposing methylene blue are shown inFig. 2B. Once the photocatalytic reaction of atesting cycle was complete, the subsequent cyclewas started after an amount of concentrated meth-ylene blue compound was added to make the ODof the solution approximately 1.0. The disorder-engineered black TiO2 nanocrystals did not ex-hibit any reduction of their photocatalytic activityunder solar irradiation after eight photocatalysiscycles.

The disorder-engineered black TiO2 nano-crystals exhibit substantial activity and stabilityin the photocatalytic production of hydrogen fromwater under sunlight. Hydrogen gas evolution asa function of time during a 22-day testing periodof solar hydrogen production experiments usingblack TiO2 nanocrystals as the photocatalysts isshown in Fig. 2C. A full-spectrum solar simu-lator was used as the excitation source, whichproduced about 1 Sun power at the sample con-sisting of black TiO2 nanocrystals loaded with0.6weight%Pt, placed in a Pyrex glass containerfilled with 1:1 water-methanol solution (metha-nol is the sacrificial reagent, and the anodic re-action generating O2 from H2O is not occurringin this system). Measurements were conductedinitially for 15 consecutive days; each day thesample was irradiated for 5 hours and then storedin darkness overnight before testing the next day.We found that 1 hour of solar irradiation gen-erated 0.2 T 0.02 mmol of H2 using 0.02 g ofdisorder-engineered black TiO2 nanocrystals(10 mmol hour–1 g–1 of photocatalysts). This H2

production rate is about two orders of magnitudegreater than the yields of most semiconductorphotocatalysts (2, 16). The energy conversionefficiency for solar hydrogen production, definedas the ratio between the energy of solar-producedhydrogen and the energy of the incident sunlight,reached 24% for disorder-engineered black TiO2

nanocrystals.After testing for 13 days, 30 ml of pure water

was added to compensate for the loss, and mea-surements continued for 2 additional days beforethe sample was stored in darkness for 2 days(days 16 and 17) without measurements. Ex-periments were resumed for 5 more days afterthe 2-day storage period, and similar rates of H2

evolution were still observed. Throughout the test-ing cycles, the disorder-engineered black TiO2

nanocrystals exhibited persistent high H2 pro-duction capability. Under the same experimentalconditions, no H2 gas was detected from theunmodified white TiO2 nanocrystals loaded with

540 536 532 528 524

O1s XPS

(a) white TiO2

(b) black TiO2

Binding Energy / eV

Inte

nst

ity

/ a.u

.

500 700 900 1100 1300

750 950 1150

Ab

sorb

ance

/ a.

u.

Wavelength / nm

(a) white TiO2

(b) black TiO2

(b)

(a)

A B

20 15 10 5 0 -5

VB XPS

Inte

nsi

ty /

a.u

.

Binding Energy / eV

(-0.92 eV)(1.26 eV)

---- black TiO2

---- white TiO2

DCconductionband

valenceband

2.18 eV3.3 eV

white TiO2Black TiO2

1.54 eV

Fig. 3. (A) O 1s XPS spectra of the white and black TiO2 nanocrystals. The red and black circles are XPS data.The green curve is the fitting of experimental data for black TiO2 nanocrystals, which can be decomposedinto a superposition of two peaks shown as blue curves. (B) Spectral absorbance of the white and black TiO2nanocrystals. The inset enlarges the absorption spectrum in the range from approximately 750 to 1200 nm.(C) Valence-band XPS spectra of the white and black TiO2 nanocrystals. (D) Schematic illustration of the DOSof disorder-engineered black TiO2 nanocrystals, as compared to that of unmodified TiO2 nanocrystals.

-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8

DO

S

Energy (eV)

bulk

nano

disorderedBA

-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 80

20

40

60

80

Energy (eV)

Par

tial

DO

S

(sta

tes/

eV/u

nit

cel

l)

Total Ti_s Ti_p Ti_d O_s O_p H_s

C D

Fig. 4. (A) Schematic illustration of a supercell for modeling a disorder-engineered TiO2 nanocrystal(red, O atoms; gray, Ti atoms; white, H atoms), as compared to a supercell modeling bulk anatase TiO2. (B)Calculated DOS of TiO2 in the form of a disorder-engineered nanocrystal, an unmodified nanocrystal, anda bulk crystal. The energy of the valence band maximum of the bulk phase is taken to be zero. (C)Decomposition of the total DOS of disorder-engineered black TiO2 nanocrystals into partial DOS of the Ti,O, and H orbitals. (D) Three-dimensional plot of calculated charge density distribution of a mid-gapelectronic state (at about 1.8 eV) of disorder-engineered TiO2 nanocrystals.

11 FEBRUARY 2011 VOL 331 SCIENCE www.sciencemag.org748

REPORTS

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

Pt. We performed experiments to quantify theamount of hydrogen absorbed in black TiO2

photocatalysts (SOM and fig. S3). The 20.0 mgof black TiO2 photocatalysts that were used togenerate 40 mg of H2 in 100 hours containedonly about 0.05 mg of H2. The black TiO2

photocatalysts did not act as a hydrogen reservoirin these experiments.

We also measured photocatalytic H2 produc-tion, using black TiO2 as the photocatalyst withonly visible and infrared light by filtering outincident light with wavelengths shorter thanabout 400 nm. The rate of H2 production droppedto 0.1 T 0.02 mmol hour–1 g–1 of photocatalysts,reflecting the activity of the extended tail or mid-gap states of the thin disordered layer, whichhave a narrower band gap created by disordering.This reduced H2 production rate was about thesame as that measured under the full solarspectrum but without using any sacrificialreagent. We examined the change of surfacechemical bonding of TiO2 nanocrystals inducedby hydrogenation with x-ray photoelectronspectroscopy (XPS) using a Physical ElectronicsPHI 5400 system. The Ti 2p XPS spectra werealmost identical for both the white and blackTiO2 nanocrystals (SOM and figs. S4 and S5),which indicates that Ti atoms have a similarbonding environment after hydrogenation and donot resemble spectra of TiO2 doped with carbonor other impurities (3, 7–10). The O 1s XPSspectra of the white and black TiO2 nanocrystalsshow dramatic differences (Fig. 3A); the single O1s peak at 530.0 eV, typical for white TiO2, canbe resolved into two peaks at about 530.0 and530.9 eV for the black TiO2 nanocrystals. Thebroader peak at 530.9 eV can be attributed toTi-OH species (17). Diffusive reflectance andabsorbance spectroscopy (Fig. 3B) revealed thatthe band gap of the unmodified white TiO2

nanocrystals was approximately 3.30 eV, slightlygreater than that of bulk anatase TiO2. The onsetof optical absorption of the black hydrogenatedTiO2 nanocrystals was lowered to about 1.0 eV(~1200 nm). An abrupt change in both the reflec-tance and absorbance spectra at approximately1.54 eV (806.8 nm) suggests that the optical gapof the black TiO2 nanocrystals was substantiallynarrowed by intraband transitions. No colorchange was observed for the black TiO2 nano-crystals over 1 year after they were synthesized.

The density of states (DOS) of the valenceband of TiO2 nanocrystals was also measured byvalence band XPS (Fig. 3C). The white TiO2

nanocrystals displayed typical valence bandDOScharacteristics of TiO2, with the edge of the max-imum energy at about 1.26 eV. Because the bandgap of the white TiO2 is 3.30 eV from the opticalabsorption spectrum, the conduction band min-imum would occur at about –2.04 eV. For theblack TiO2 nanocrystals, the valence band max-imum energy blue-shifts toward the vacuum lev-el at approximately –0.92 eV. Combined with theresults from optical measurements that suggest amuch narrowed band gap, the conduction band

DOS of the black TiO2 nanocrystals would nothave as substantial a change. Nevertheless, theremay be conduction band tail states arising fromdisorder that extend below the conduction bandminimum. Optical transitions from the blue-shiftedvalence band edge to these band tail states arepresumably responsible for optical absorptiononset around 1.0 eV in black TiO2. A schematicillustration of the DOS of disorder-engineeredblack TiO2 nanocrystals is shown in Fig. 3D.

To understand the origin of the change in theelectronic and optical properties of black TiO2

nanocrystals, we calculated the energy band struc-tures using a first-principles density functionaltheory (DFT) (18–20). Existing models of mod-ified TiO2 are focused on point defects, whichtend to produce shallow or deep energy levelsnear the conduction band minimum with typicalTi3+ state characteristics (21–23). We studiedon-lattice disorders in TiO2 nanocrystals in thepresence of hydrogen and found that, rather thangenerating levels near the conduction band min-imum, disorder-induced mid-gap states can up-shift the valence band edge of TiO2 nanocrystals.

We first constructed a network of TiO2 nano-crystals without disorder but with fully relaxedsurface dangling bonds. From the calculated totaland projected DOS (SOM and fig. S6), we foundthat the primary effect of surface reconstructionin TiO2 nanocrystals is to produce strong bandtailing near the valence band edge. The valenceand conduction states are derived mainly fromthe O 2p orbitals and the Ti 3d orbitals, re-spectively. Without the introduction of disorder,we examined the band structures of TiO2 nano-crystals containing four types of intrinsic defectswith low formation energies: Ti vacancy (VTi), Ovacancy (VO), interstitial titanium (ITi), and in-terstitial oxygen (IO); and three types of hydrogenimpurities: interstitial H atom (IH), interstitial H2

molecule (IH2), and H atom forming surface OHbonds with oxygen (OHsurface). The three nativedefects, VTi, VO, and IO, do not introduce mid-gap states, which agrees with previous calcula-tions (24). Similarly, no mid-gap states wereproduced as the result of the defects associatedwith hydrogen impurities (22, 23). Based on theestablished DFT (21), without disorder, the onlydefect that could yield a gap state in TiO2 nano-crystals, about 0.5 eV below the conduction bandminimum, is the interstitial Ti atom, ITi.

When we introduced lattice disorders in hy-drogenated anatase TiO2 nanocrystals, mid-gapelectronic states were created, accompanied by areduced band gap. The disordered TiO2 nano-crystal model, in which one H atom is bonded toanO atomwhile another H atom is bonded to a Tiatom, yields electronic band structures consistentwith the valence band XPS measurements. Aschematic illustration of a supercell of the dis-ordered TiO2 nanocrystal model as compared tothat of a bulk anatase TiO2 crystal is shown inFig. 4A, and Fig. 4B plots the calculated DOS ofthe disorder-engineered TiO2 nanocrystals alongwith those of the bulk and the unmodified TiO2

nanocrystals. Two groups of mid-gap states (cen-tered at about 1.8 and 3.0 eV) can be observed inthe DOS of the disordered TiO2 nanocrystals, forwhich the Fermi level was found to locateslightly below 2.0 eV. The different nature ofthese two groups of mid-gap states is revealed bythe calculated partial DOS (Fig. 4C).Whereas thehigher-energy mid-gap states (~3.0 eV) are de-rived from the Ti 3d orbitals only, the lower-energy states (~1.8 eV) are hybridized from bothO 2p orbitals and Ti 3d orbitals, and mainly fromthe valence band states as the result of disordersstabilized by hydrogen. The hydrogen 1s orbitalcoupling to the Ti atom does not make a sub-stantial contribution to either state, which suggeststhat lattice disorder accounts for the mid-gapstates; hydrogen may have stabilized the latticedisorders by passivating their dangling bonds.Because the lower-energy mid-gap states lie be-low the Fermi level, they can account for a largeblue shift of the valence band edge.

We also examined the three-dimensional chargedensity distribution of the mid-gap electronicstates (Fig. 4D) of disorder-engineered TiO2

nanocrystals. Charges associated with the lower-energy mid-gap states distributed around every Oor Ti atom are indicative of the overall impact ofthe disorder. Because the lower-energy mid-gapstates are derived from hybridization of the O 2porbital with the Ti 3d orbital, optical transitionbetween these mid-gap states and the conductionband tail would produce charge transfer from theO 2p orbital to the Ti 3d orbital, similar to thetransition from the valence to the conductionband of bulk TiO2. The localization of bothphotoexcited electrons and holes prevents fastrecombination and presumably is the reason whydisorder-engineered black TiO2 can more effi-ciently harvest the infrared photons for photo-catalysis than what bulk anatase can do to theabove–band-gap UV photons.

References and Notes1. M. Grätzel, Nature 414, 338 (2001).2. X. Chen, S. Shen, L. Guo, S. S. Mao, Chem. Rev. 110,

6503 (2010).3. X. Chen, S. S. Mao, Chem. Rev. 107, 2891 (2007).4. A. Fujishima, X. Zhang, A. D. Tryk, Surf. Sci. Rep. 63,

515 (2008).5. M. R. Hoffmann, S. T. Martin, W. Choi, D. W. Bahnemann,

Chem. Rev. 95, 69 (1995).6. W. Choi, A. Termin, M. R. Hoffmann, Angew. Chem. 106,

1148 (1994).7. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga,

Science 293, 269 (2001).8. S. U. M. Khan, M. Al-Shahry, W. B. Ingler Jr., Science

297, 2243 (2002).9. X. Chen, C. Burda, J. Am. Chem. Soc. 130, 5018 (2008).

10. X. Chen, C. Burda, J. Phys. Chem. B 108, 15446 (2004).11. F. Zuo et al., J. Am. Chem. Soc. 132, 11856 (2010).12. J. Tauc, A. Abraham, L. Pajasova, R. Grigorovici, A. Vancu,

Non-Crystalline Solids (North-Holland, Amsterdam, 1965).13. P. Y. Yu, M. Cardona, Fundamentals of Semiconductors:

Physics and Materials Properties (Springer, Heidelberg,ed. 4, 2010).

14. U. Diebold, Surf. Sci. Rep. 48, 53 (2003).15. T. L. Thompson, J. T. Yates Jr., Chem. Rev. 106, 4428 (2006).16. A. Kudo, Y. Miseki, Chem. Soc. Rev. 38, 253 (2009).17. E. McCafferty, J. P. Wightman, Surf. Interface Anal. 26,

549 (1998).

www.sciencemag.org SCIENCE VOL 331 11 FEBRUARY 2011 749

REPORTS

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

18. G. Kresse, J. Hafner, Phys. Rev. B 47, 558 (1993).19. G. Kresse, J. Hafner, Phys. Rev. B 48, 13115 (1993).20. G. Kresse, J. Furthmüller, Comput. Mater. Sci. 6, 15 (1996).21. C. Di Valentin, G. Pacchioni, A. Selloni, J. Phys. Chem. C

113, 20543 (2009).22. C. G. Van de Walle, Phys. Rev. Lett. 85, 1012 (2000).23. Ç. Kiliç, A. Zunger, Appl. Phys. Lett. 81, 73 (2002).24. S. Na-Phattalung et al., Phys. Rev. B 73, 125205 (2006).25. We thank M. S. Dresselhaus for encouragement, M. T. Lee

and S. H. Shen for their assistance, and R. Greif for

discussions and critical reading of the manuscript. Thisresearch has been supported by the Office of EnergyEfficiency and Renewable Energy of the U.S. Departmentof Energy. S.S.M. and X.C. also acknowledge support fromthe King Abdullah University of Science and Technology–University of California Academic Excellence Alliance. TEMwork was performed at the National Center for ElectronMicroscopy, which is supported by the Office of Science,Office of Basic Energy Sciences of the U.S. Departmentof Energy.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/science.1200448/DC1MethodsSOM TextFigs. S1 to S6References

15 November 2010; accepted 11 January 2011Published online 20 January 2011;10.1126/science.1200448

Complete Fourth Metatarsal andArches in the Foot ofAustralopithecus afarensisCarol V. Ward,1* William H. Kimbel,2 Donald C. Johanson2

The transition to full-time terrestrial bipedality is a hallmark of human evolution. A key correlateof human bipedalism is the development of longitudinal and transverse arches of the foot thatprovide a rigid propulsive lever and critical shock absorption during striding bipedal gait. Evidencefor arches in the earliest well-known Australopithecus species, A. afarensis, has long been debated.A complete fourth metatarsal of A. afarensis was recently discovered at Hadar, Ethiopia. It exhibitstorsion of the head relative to the base, a direct correlate of a transverse arch in humans. Theorientation of the proximal and distal ends of the bone reflects a longitudinal arch. Further, thedeep, flat base and tarsal facets imply that its midfoot had no ape-like midtarsal break. Thesefeatures show that the A. afarensis foot was functionally like that of modern humans and supportthe hypothesis that this species was a committed terrestrial biped.

Although Australopithecus afarensis wasprimarily a terrestrial biped, there con-tinues to be debate over the nature of its

bipedality and the extent to which its morphologyrepresents a compromise between terrestrial bi-pedality and arboreal locomotion. One of the keyadaptations to a human-like striding bipedal gaitis the evolution of permanent transverse and lon-gitudinal pedal arches (1, 2). The arches, supportedby bone and soft tissue, provide an importantmechanism for shock absorption during the stancephase of gait (3) and a rigid lever at heel-off, aswell as permit flexibility during locomotion atdifferent speeds and across irregular terrain (3–6).Muscles that in apes adduct the hallux (such asthem. adductor hallucis andm. fibularis longus), inhumans primarily support the pedal arches (1, 5).Permanent plantar arches are a key component ofhuman bipedal walking and running becausethey contribute to the rigidity of the foot and pro-vide an enhanced mechanical advantage duringthe propulsive phase of gait (1, 7–9). Extant apes,in contrast, exhibit pronounced midtarsal dorsi-flexion during heel-off as a result of a mobilemidfoot, which permits flexibility for negotiatingvariably oriented arboreal substrates [(8, 9) and a

recent review in (10)]. This break is greater inmagnitude and is kinematically and anatomicallydistinct from the medial collapse seen in somehumans (7, 11, 12). Therefore, determining theextent to which the foot of A. afarensis hadpermanent longitudinal and transverse pedal archesis key to deciphering the extent of its commit-ment to terrestrial bipedality.

Skeletal evidence for the presence of pedalarches inA. afarensis has been ambiguous, because

key bones from the midfoot have been lacking.The talus (specimens AL 288-1 andAL 333-147)shows a distinct facet for plantar calcaneonavicular(13) and cubonavicular ligaments (9, 13, 14), whichare indicative of a human-like medial longitudi-nal arch. However, unlike in humans, a groove ispresent for the m. fibularis longus tendon on theplantar surface of the ectocuneiform (AL 333-79)in A. afarensis, as seen in apes, perhaps related tothe lack of a transverse arch. Dorsal inclination ofthe tarsal facets has been interpreted to suggestthe lack of longitudinal arches in A. afarensis(11, 12), and the well-developed navicular tu-berosity is argued to be a weight-bearing struc-ture in these hominins (15, 16).

Here we describe AL 333-160, a complete,nearly perfectly preserved fourth metatarsal ofA. afarensis from Hadar, Ethiopia (Fig. 1). Thisspecimen was recovered from the Hadar locali-tyAL333 in 2000 during sieving of erodedDenenDora 2 submember surface deposits of the HadarFormation. Since 1975, these deposits at the 333locality have yielded more than 250 homininfossils that eroded from an in situ horizon datedto ~3.2 million years ago (17). We assign AL333-160 to A. afarensis, the only homininspecies in an assemblage of >370 homininspecimens so far recovered from the Hadar For-mation (18). Other partial metatarsals attributedto A. afarensis are known from Hadar (19), butnone is complete enough to address the question

1Department of Pathology and Anatomical Sciences, M263Medical Sciences Building, University of Missouri, Columbia,MO65212,USA. 2Institute of HumanOrigins, School of HumanEvolution and Social Change, Post Office Box 874101, ArizonaState University, Tempe, AZ 85287-4101, USA.

*To whom correspondence should be addressed. E-mail:wardcv@missouri.edu Fig. 1. AL 333-160 left fourth metatarsal in dorsal, lateral, medial, plantar, and proximal views.

11 FEBRUARY 2011 VOL 331 SCIENCE www.sciencemag.org750

REPORTS

on

Apr

il 11

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from