Sn

SMa

b

a

ARRAA

KCASAP

1

demaoeoracChbd

i

Lf

0h

Catalysis Today 199 (2013) 27– 35

Contents lists available at SciVerse ScienceDirect

Catalysis Today

jou rn al h om epage: www.elsev ier .com/ locate /ca t tod

calable synthesis and photoelectrochemical properties of copper oxideanowire arrays and films

wathi Sunkaraa,1, Venkat Kalyan Vendraa,1, Jeong Hoon Kimb, Thad Druffelb,ahendra K. Sunkaraa,b,∗

Department of Chemical Engineering, University of Louisville, Louisville, KY 40292, United StatesConn Center for Renewable Energy Research, University of Louisville, Louisville, KY 40292, United States

r t i c l e i n f o

rticle history:eceived 28 November 2011eceived in revised form 8 February 2012ccepted 11 March 2012vailable online 10 April 2012

a b s t r a c t

Here, we present a scalable method based on both atmospheric plasma and wet chemical oxidation meth-ods to synthesize thin films and nanowire arrays of both cupric and cuprous oxides. In terms of nanowirearrays, the wet chemical oxidation is shown to produce copper hydroxide nanowire arrays on copper foilsusing hydrogen peroxide as the oxidant similar to ammonium persulfate. Experiments using differentconcentrations of hydrogen peroxide at constant pH resulted in higher nucleation density and smaller

eywords:opper oxide nanowirestmospheric plasmaolar water splittingtomic layer depositionhotoelectrochemical characterization

diameter of copper hydroxide nanowires. A scheme involving a short period of wet oxidation followedby plasma annealing resulted in high number density of Cu2O nanowire arrays. The overall process israpid on the order of 1 min reaction time scale. Photoelectrochemical characterization of titania coatedcopper oxide nanowire and thin film electrodes showed that the nanowire array electrodes exhibitedsignificantly higher photoactivity than the thin film electrodes. The performance of resulting Cu2O NWarray electrodes can be optimized further with other protective coatings.

. Introduction

Photoelectrochemical water splitting for solar hydrogen pro-uction represents one of the grand challenges towards carbon freenergy generation. However, there are no known semiconductoraterials with the appropriate band gap, band edge positions and

queous stability required for solar water splitting [1,2]. Cuprousxide (Cu2O) is an attractive material because of the non-toxicity,arth abundance of its constituent elements and has a band gapf approximately 2.2 eV, with its band edges straddling the watereduction and oxidation potential [3–5]. As-synthesized, Cu2O is

p-type semiconductor due to the acceptor states resulting fromopper vacancies located at 0.4 eV above the valence band [6].uprous oxide can potentially act as efficient photocathode forydrogen evolution. However, the valence band position is justelow the oxygen evolution potential and this would imply less

riving force for oxygen evolution.

One of the main challenges with cuprous oxide is its inherentnstability in aqueous solutions [7]. The photostability challenge

∗ Corresponding author at: Department of Chemical Engineering, University ofouisville, Louisville, KY 40292, United States. Tel.: +1 5024395896;ax: +1 5028526355.

E-mail address: mahendra@louisville.edu (M.K. Sunkara).1 These authors contributed equally to this work.

920-5861/$ – see front matter. Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.cattod.2012.03.014

Published by Elsevier B.V.

posed by the Cu2O electrodes has been addressed by depositing thinfilms of titania on cuprous oxide by atomic layer deposition (ALD)[8]. The resulting hetero-junction is not only expected to improvestability but also reduces recombination by enhancing the electrontransport due to the inbuilt electric-field at the hetero-junction.Gratzel and co-workers modified the cuprous oxide photocathodeby coating it with a layer of titania followed by aluminum dopedzinc oxide and then electrodepositing platinum nanoparticles toimprove the stability and performance of the Cu2O electrodes [9]. Itwas found that the titania coated cuprous oxide photocathodes hadpoor stability in aqueous solutions and exhibited no photoactivityafter 20 min. Pin holes formed during the ALD of titania have beenfound to result in instability in aqueous solutions. Additional layersof zinc oxide and alumina improved the photostability (78% of theshort-circuit current density retained after 20 min) and also play animportant role in enhancing the charge separation and mitigatingthe electron–hole recombination in Cu2O.

The poor electrical conductivity of Cu2O has been a limiting fac-tor in achieving high efficiency for Cu2O based solar cells and is alsoa major concern for solar water splitting cells [10]. The absorptiondepth corresponding to the band gap of Cu2O has been reportedto be 10 �m [11]. However, the diffusion length of minority carri-

ers in electrodeposited copper oxide (Cu2O) has been found to beon the order of 20–100 nm [12]. In view of this, 1D single crystalarchitectures in the form of nanowire arrays could offer the thick-ness necessary for optical absorption and short length scales for

2 ysis To

dfowat

atboaSmettwomCuwCTdadtinIaccp[

hliatfsaadohopc

2

udutiftn

8 S. Sunkara et al. / Catal

iffusion of minority carriers to semiconductor–electrolyte inter-ace to drive the water splitting reaction. The single crystal nature ofne-dimensional structures should allow direct conduction path-ays for faster charge transport [13,14]. However, the performance

nd stability of single crystal copper oxide nanowires coated withitania for solar hydrogen production has not been investigated.

The synthesis of polycrystalline cuprous oxide films has beenccomplished with a number of techniques such as electrodeposi-ion, photochemical deposition, RF magnetron sputtering, chemicalath deposition, anodic oxidation and thermal oxidation meth-ds [9,15–19]. However, the synthesis of cuprous oxide nanowirerrays has been achieved using a limited number of techniques.ynthesis using template and surfactant assisted methods hasostly yielded polycrystalline cuprous oxide nanowires with the

xception of a few reports describing the growth of single crys-al cuprous oxide nanowires [20–23]. Wu and co-workers reportedhe synthesis of linearly aligned complex chains of metal cationshich could be reduced by glucose to form single crystal cuprous

xide nanowires with a diameter of 20 nm and lengths of severalicrons [24]. Li et al. demonstrated the liquid phase synthesis of

u2O NW (diameter ∼ 20 nm) by the reduction of cupric acetatesing reducing agents with methoxy group [25]. Ajayan and co-orkers reported a simple electrochemical method utilizing onlyu foil and DI water to synthesize cuprous oxide nanowires [26].o the best of our knowledge, there have not been any reportsescribing the growth of single crystalline cuprous oxide nanowiresrrays directly on conducting substrates by thermal or plasma oxi-ation. Several reports suggest thermal oxidation of copper foils atemperatures between 400 and 700 ◦C over several hours resultedn bicrystalline CuO nanowire arrays [27–31]. To synthesize Cu2Oanowire arrays, a two-step process has been primarily developed:

n the first step, copper is subjected to wet chemical oxidation usingmmonium persulfate in sodium hydroxide solutions to produceopper hydroxide nanowire arrays; and in the second step, theopper hydroxide nanowire arrays are thermally annealed at tem-eratures around 450 ◦C to convert them to Cu2O nanowire arrays32].

All the synthesis techniques discussed above require severalours of oxidation or annealing time scales to produce few microns

ong Cu2O nanowire arrays on copper foils. In addition, the result-ng nanowires had diameters in excess of 100 nm and it is not clears to what factors control the diameters and the number density ofhe resulting nanowire arrays. Here, we present a scalable approachor producing Cu2O nanowire arrays on copper foils and on otherubstrates such as quartz, fluorinated tin oxide (FTO) substratesnd other metallic foils. Specifically, we investigated the use oftmospheric plasma combined with wet chemical oxidation to pro-uce cuprous oxide nanowire arrays. Also, we performed a seriesf experiments to understand nucleation and growth of copperydroxide and copper oxide nanowires using both wet chemicalxidation and plasma oxidation processes. Most importantly, thehotoelectrochemical performance of films and nanowire arraysoated with thin layers of titania are investigated and compared.

. Experimental

Atmospheric plasma exposure experiments were conductedsing a microwave plasma discharge whose details have beenescribed elsewhere [33]. Various experiments were conductedsing atmospheric plasma discharge at powers ranging from 500o 900 W and air flow rates ranging from 5 to 10 lpm. Wet chem-

cal oxidation experiments were conducted by immersing copperoils in sodium hydroxide solutions at pH values ranging from 7o 14 containing different amounts of oxidizers such as ammo-ium persulfate ((NH4)2S2O8) and hydrogen peroxide. Atomic layer

day 199 (2013) 27– 35

deposition of titania films on the copper oxide thin films andnanowire arrays was carried out using a Savannah 100 ALD sys-tem (Cambridge Nanotech). The deposition was carried out at apressure of 600 mTorr and a temperature of 250 ◦C using water andtitanium isopropoxide as the precursors. The deposition processwas carried out for 650 cycles to form a 20 nm titania coating onthe electrodes. The synthesized samples were analyzed using X-raydiffraction (Bruker Instruments) and Scanning electron microscopy(Nova FEI SEM). The optical band gap of the films was determinedusing diffuse reflectance obtained with a UV–vis spectrophotome-ter (PerkinElmer, Lambda 900).

For photoelectrochemical characterization, the resulting copperoxide film and nanowire arrays samples are made into electrodesusing the following procedure. The copper foils are connected toan electrical wire at the back using conductive silver epoxy andthen the entire non-active area was covered with Hysol epoxy (Loc-tite 9464). The electrodes were cured for 110 ◦C for 1 h. A HysolE120HP, which has excellent chemical resistance, was applied overthe previously cured epoxy coating and was allowed to dry at roomtemperature for 12 h.

All the photoelectrochemical measurements were performedin photoelectrochemical cell equipped with a quartz window. Athree electrode configuration comprising of copper oxide work-ing electrode, Ag/AgCl reference electrode and a platinum meshcounter electrode was used for the characterization. A solutionof sodium sulfate (1 M), buffered to pH 4.9 using a potassiumphosphate solution (0.1 M) was used as the electrolyte. For ana-lyzing current–potential (J–E) curves, the potentials measuredw.r.t. Ag/AgCl electrode was converted to potential w.r.t. referencehydrogen electrode (RHE). The data were not corrected for anyother extrinsic losses such as the iR losses. The photoresponse ofelectrodes was measured under chopped AM 1.5 illumination froma 200 W Xe lamp (Newport Instruments). The light intensity wasadjusted to 100 mW/cm2 (1 Sun illumination) using a calibrated Siphotodiode. The typical active area of the electrodes was 0.68 cm2.For the linear sweep voltammetry experiments (J–E curves), ascan rate of 5 mV/s was used. All the electrochemical measure-ments were recorded using an EG&G Princeton Applied Research273 A potentiostat. The Mott–Schottky analysis was done under1 Sun illumination and a frequency of 10 Hz was used. Platinumloading onto the nanowire arrays was performed using a three elec-trode setup using copper oxide as the working electrode, Ag/AgClas the reference and a Pt mesh as the counter electrode. An aque-ous solution of 1 mM chloroplatinic acid was used as the electrolyteand a potential of −0.1 V vs. Ag/AgCl was applied for 10 min. Cop-per electroplating on the gold sputtered fluorinated tin oxide (FTO)glass slides was performed using a two electrode configurationwhere FTO glass was used as the counter electrode, a copper foilwas used as the working electrode. An aqueous solution of 1 M cop-per sulfate and 0.5 M sulfuric acid was used as the electrolyte anda current density of −10 mA/cm2 was applied for 10 min.

3. Results and discussion

Several researchers have performed wet chemical oxidationusing ammonium persulfate and sodium hydroxide solutions as theprecursors. The experiments performed at pH values >10 resultedin copper hydroxide (Cu(OH)2) nanowire arrays on copper foil [34].Similarly, such wet chemical oxidation method has been shownto work with several other metals in producing their respectivehydroxide nanowire arrays [35–37]. However, it is not clear about

the underlying reasons for producing one-dimensional structuresand the role of ammonium or sodium persulfate. Here, a set ofexperiments are conducted to gain insight in to wet chemical oxi-dation procedure. Firstly, experiments are performed at different

S. Sunkara et al. / Catalysis To

F ◦

sg

praituelair

C

uf

C

C

C

tfacnp

onddwiewti

ig. 1. Pourbaix diagram of copper–water system at 25 C generated using Medusaoftware. [Cu2+] concentration of 1 �M was used. The dotted lines show the hydro-en and oxygen evolution potential.

H values using sodium hydroxide and potassium hydroxide. Theesults showed that copper hydroxide nanowire arrays resultedt pH >10 and no differences were seen between experimentsnvolving sodium hydroxide and potassium hydroxide. In ordero understand the role of oxidizer, experiments are performedsing hydrogen peroxide instead of ammonium persulfate. Thexperiments resulted in copper hydroxide nanowire arrays simi-ar to those obtained using ammonium persulfate. In the case ofmmonium persulfate as oxidizer, the growth of copper hydrox-de nanowires proceeds by an oxidation process according to theeaction [34].

u + 4NaOH + (NH4)2S2O8 → Cu(OH)2 + 2Na2SO4 + 2NH3 + H2O

The above overall reaction can be understood better with these of hydrogen peroxide as the oxidizer. The oxidation of copperoils with hydrogen peroxide is expected to follow two steps:

u + H2O2 → CuO + H2O (Step 1)

uO + 2OH− → Cu(OH)2 + (1/2)O2 (Step 2)

u + H2O2 + 2OH− → Cu(OH)2 + (1/2)O2 + H2O (Overall reaction)

The above proposed two-step reaction suggests that the oxida-ion proceeds in forming copper oxide first and then hydrolyzing toorm copper hydroxide. The copper hydroxide, copper oxide phasesre not stable when the pH of the solution is less than 10 as indi-ated by the Pourbaix diagram (Fig. 1) and this explains why noanowire formation was observed in the experiments carried withH < 10.

Experiments using different concentrations of hydrogen per-xide at constant pH of 12 showed that the number density ofanowire arrays increased and diameter of resulting nanowiresecreased with increasing concentration of hydrogen peroxide asepicted in Fig. 2. It is important to note that pH of the solutionould change with time as sodium hydroxide is being consumed

n the reaction and hence it is essential to maintain constant pH to

nsure proper growth of the nanowires. As the reactions proceeds itas observed color of the solution changed to dark blue suggesting

he presence of copper ions in the solution phase. These copper ionsn the solution could cause additional growth via copper hydroxide

day 199 (2013) 27– 35 29

precipitation from solution resulting in deviations from nanowiregrowth when pH is not maintained.

The length of the copper hydroxide nanowires increased withincreased oxidation times. As wet oxidation time increased, thelength of the nanowires was found to increase from 500 nm (30 simmersion time) to 3 �m (1–2 min immersion time) as shown inFigure S2 (refer Supplementary Information). There was no signif-icant change in the nanowire length at longer immersion times of30 min. These observations support the argument that nanowireformation mechanism is through basal growth. To further vali-date the growth mechanism, short copper hydroxide nanowires(∼500 nm in length) were grown by wet chemical oxidation andthen exposed to plasma oxidation for 20 s or thermal oxidation for10 min. The increase in length of the nanowires after the oxidationstep shows that nanowires nucleated by the wet oxidation methodgrow longer by preferential oxidation at the nanowire/oxide filminterface. Hence, the nucleation and growth mechanism of cop-per hydroxide nanowires can be thought to proceed similar to theplasma oxidation of metal foils to form metal oxide nanowires. Seeschematic in Fig. 3. The only difference is that high pH conditionsresult in the formation of metal hydroxide nanowires, whereas inplasma oxidation metal oxide nanowires are formed.

Direct plasma oxidation of copper foils resulted in thin films ofCu2O and CuO with different exposure conditions. Cu2O thin filmswere formed when Cu foils were exposed to atmospheric plasmaflame that resulted in higher temperatures <500 ◦C. CuO filmsresulted with exposures that resulted in temperatures > 500 ◦C forthe copper foils. The resulting films were about 2–3 �m thicknesswith only about exposure over 2–3 min exposure unlike thermaloxidation or other methods that require processing over severaltens of minutes. Figure S3 shows the SEM images of the copperoxide thin films produced by plasma oxidation. A few of the thinfilm samples prepared by the plasma oxidation technique alsoshowed a low density of nanowires as shown in Figure S2 in sup-plementary information document. Prior work on the growth ofhematite nanowire arrays on iron foil showed that the plasmaoverheating or the initial temperature rise before the substratereaches an equilibrium temperature is a critical factor controllingthe nucleation density [38,39]. Also, earlier reports have shownthat the thermal oxidation of copper foils over 4 h at atmosphericpressure results in cupric oxide (CuO) nanowire arrays in temper-ature window of 400–700 ◦C for 4 h [27]. These findings suggestthat temperature control of the copper foil will play a key role indetermining the number density of the nanowires (Fig. 4).

In order to avoid long durations involved with thermal oxi-dation/annealing step, the synthesis of cuprous oxide nanowireswas carried out using a two-step approach: in the first step, cop-per hydroxide nanowires were formed by a wet oxidation usinghydrogen peroxide and in the second step, the copper hydrox-ide nanowires were annealed in the atmospheric air plasma forfew seconds to convert the copper hydroxide nanowires to cop-per oxide nanowires. Fig. 3 illustrates the synthesis scheme formaking copper oxide nanowires. The X-ray diffractogram of thecopper oxide thin films and nanowires are shown in Fig. 5. Anal-ysis of the peak positions showed that a small fraction of CuOphase was present in the Cu2O nanowire samples. In addition, thesynthesis of nanowire arrays was also carried out on glass slidescoated with fluorinated tin oxide (FTO). Copper was electrode-posited on gold sputtered FTO slides which were then immersedinto a solution of hydrogen peroxide and sodium hydroxide fora very short period of time ∼2 min (sometimes less than 20 sdepending up on thickness). Unlike Cu foils, the electrodeposited Cu

formed a film of copper hydroxide nanoparticles. Interestingly, cop-per oxide nanowires were observed when these copper hydroxidenanoparticles were annealed using plasma oxidation. Even thoughcopper hydroxide nanowires are not formed in the wet oxidation

30 S. Sunkara et al. / Catalysis Today 199 (2013) 27– 35

ire de

pgm

asftalvtptsomNdc

tptptaa

Fig. 2. Scanning electron microscopy images showing increase in nanow

rocedure, the results suggest that high densities of nucleation androwth of copper oxide nanowires from copper hydroxide films isuch more easier and faster than that directly from copper.The optical band gap of the as-synthesized samples (without

ny titania coating) is determined using diffuse reflectance data ashown in Fig. 6. The band gap values are found to be about 1.5 eVor CuO and 2.1 for Cu2O NW respectively. The band gap for the CuOhin films was found to be 1.5 eV and 2.4 eV for Cu2O. These valuesre slightly higher than the range of band gap values reported initerature for cuprous oxide (2.0–2.3 eV [8,22]) and the range ofalues reported for cupric oxide (1.2–1.4 eV) [28,29]. Analysis ofhe Tauc plot for Cu2O NW samples shows evidence for some CuOresent in the Cu2O NW samples and this is also in agreement withhe XRD results. The phase purity of Cu2O NW arrays has also beenhown to be a challenge with the nanowires annealed by thermalxidation [31,32]. The CuO in the Cu2O nanowire array sample isost likely to be present as a polycrystalline layer beneath the Cu2OW arrays. The background signal from the CuO would make itifficult to draw an accurate baseline in the Tauc plots which canause variations in the estimated values for band gaps [2].

The Mott–Schottky plots of the Cu2O NW and thin film elec-rodes are represented in Fig. 7. The intercept of the Mott–Schottkylot gives the value Vfb + kT/q, where Vfb is the flat band poten-ial, k is Boltzmann constant, q is the electron charge. The flat band

otential for the Cu2O NW arrays and the thin films was estimatedo be 0.012 V vs. Ag/AgCl or −4.05 eV vs. vacuum. The CuO NW had

flat band potential of 0.01 V vs. Ag/AgCl while CuO thin films had flat band potential of 0.06 V vs. Ag/AgCl. The negative slope in the

nsity with increase in H2O2 concentration: (a) 0.5 M; (b) 1 M; (c) 1.5 M.

Mott–Schottky plots indicates that Cu2O and CuO electrodes showp-type behavior. The hole concentration can be estimated using thefollowing equation:

N = 2

qε0εrd

dV

(1

C2

)

where N is the carrier density, ε0 = (8.854 × 10−12 F/m) is permit-tivity of free space, εr is the dielectric constant, and (d/dV) (1/C2) isthe slope of the Mott–Schottky curve. A dielectric constant of 6.3 forcuprous oxide and 10.26 for cupric oxide was used for the calcula-tions [40,41]. Basing on the slopes obtained from the Mott–Schottkyplots the carrier densities were calculated as 5.6 × 1022 cm−3 forthe Cu2O thin films and 1.9 × 1023 cm−3 for the Cu2O nanowires.The carrier density was estimated to be 5 × 1021 cm−3 for CuO thinfilms and 1.37 × 1022 cm−3 for CuO NW respectively. The carrierconcentration in copper oxide nanowires is an order of magnitudehigher than thin films suggesting higher carrier generation andlower recombination losses. The Mott–Schottky analysis was per-formed under 1 Sun illumination (100 mW/cm2), resulting in highcarrier concentration due to the generation of excess carriers uponillumination. The open-circuit potential data as shown in Fig. 8indicate slightly negative values for nanowire arrays compared tothin films. This difference is similar to that observed with flat band

potential values using Mott–Schottky technique. The flat potentialindicates the position of Fermi level in the semiconductor. Higherhole density for nanowires samples (obtained from Mott–Schottkyplots) implies that the Fermi level is more closer to the valance

S. Sunkara et al. / Catalysis Today 199 (2013) 27– 35 31

F e nann ges ofn

bllnlaamccnwctctfclstparpiTnTt

ig. 3. Schematic explaining the mechanism for the formation of copper hydroxiducleation of CuO; (b) copper oxide is converted to copper hydroxide and early staanowires by basal growth.

and for the nanowires when compared to thin films. The Fermievel of n-type titania is closer to its conduction band. Upon Fermievel equilibration, a higher shift in the Fermi level is expected foranowires when compared to thin films. If the change in the Fermi

evel is significant after equilibration, this effect would be furthermplified when the flat band potential of thin films and nanowiresre compared. However, this is not reflected in the experimentallyeasured flat band potentials. The flat potentials for the titania

oated copper oxide thin films and nanowire samples appear to belose. This would imply that a thin film of titania (∼20 nm) doesot deplete the charge carriers at the copper oxide surface andould result in no significant change in the flat band potential after

oating the copper oxide with titania. Another possible explana-ion is that the Fermi level could be pinned at the acceptor statesausing no variation in the flat band potential after titania deposi-ion. The J–E curves measured under chopped AM 1.5 illuminationor Cu2O, CuO thin films and nanowires are shown in Fig. 9. Thehopped illumination method has the advantage of monitoring theight and dark current simultaneously. Cu2O NW array electrodeshowed significantly higher photocurrent than the thin film elec-rodes. At a potential of 0 V w.r.t. reference hydrogen electrode, thehotocurrent density (Jlight − Jdark) was found to be 0.06 mA/cm2

nd 0.26 mA/cm2 for the thin film and Cu2O NW array electrodesespectively. CuO thin films and nanowire arrays showed very lowhotocurrents. Phase purity of copper oxide and removal of the

nterfacial CuO layer is important to achieve higher photocurrents.

he presence of pin holes in the titania coating and film delami-ation could also result in variations and lower the photocurrent.he CuO samples exhibited more delamination as they were syn-hesized at higher temperatures.

owires: (a) Cu foil reacts with the oxidizer to form copper oxide and subsequent nucleation of copper hydroxide nanowire growth; (c) increase in the length of the

The work function of copper (4.65 eV) [42] is less than that ofcopper oxide (4.8–5.2 eV) [43], causing the bands bending down-wards at the copper oxide–copper interface after the Fermi levelsequilibrate. This results in the formation of a Schottky barrier forthe holes at the copper oxide–copper interface [44]. The experi-mentally measured barrier heights range from 0.74 to 0.84 eV [45].For a p-type semiconductor, the holes have to move through thebulk of the semiconductor to the counter electrode to participatein water oxidation reaction, while the electrons are transferredfrom the surface of the semiconductor to electrolyte for the hydro-gen evolution reaction. The Schottky-barrier for the holes at thecopper-oxide/copper interface would result in slower transportof holes across the interface. This would increase the recombina-tion of holes with electrons and result in lower current density.To eliminate this Schottky barrier, the synthesis was carried outwith FTO slides, sputtered with gold and subsequently coated withcopper by electrodeposition. The SEM images of cuprous oxidenanowires grown on FTO are shown in Fig. 10. Interestingly, thenanowires grown on FTO also showed similar performance as thatof nanowires grown on Cu foil, suggesting the minimal role of theSchottky barrier in affecting the performance (Figure S6, Supple-mentary information). The low photocurrents seen with nanowireelectrodes could be due to recombination arising at the polycrys-talline copper oxide film present beneath the nanowire arrays(Figure S7, Supplementary information). Also, the presence ofTi3+ trap states in titania could hinder the transport of elec-

trons at the semiconductor–electrolyte interface and lower theobserved photocurrents [46]. Another possibility is the sluggishkinetics of H2 reduction at the titania interface and a suitable cat-alyst like platinum could help overcome this problem. Preliminary

32 S. Sunkara et al. / Catalysis Today 199 (2013) 27– 35

F d theiS

enis(olicit

c

ig. 4. (a) SEM image showing the nucleation of NW by wet chemical oxidation anchematic illustrating growth of wires through plasma oxidation.

xperiments involving loading platinum particles onto the tita-ia coated nanowire arrays showed no significant improvement

n photoactivity. Also, it was found that the photocurrent den-ity saturates when the light intensity is increased beyond 1 SunFigure S8, Supplementary Information). These results clearly pointut the detrimental role of the interfacial cupric oxide layer whichimits the performance of the copper oxide nanowire arrays. Thenterfacial oxide layer not only acts as transport barrier for theharge carriers but also increases the recombination losses. The

nterfacial oxide layer has to be eliminated to achieve higher pho-ocurrents from the nanowire arrays.

The stability of the electrodes was tested by monitoring the openircuit voltage under chopped AM 1.5 conditions (100 mW/cm2).

r growth by (b) thermal oxidation for 10 min and (c) plasma oxidation for 20 s. (d)

The open circuit voltage vs. time plots are shown in Fig. 8. and itcan be seen that both the nanowire and thin film electrodes arestable for atleast 1000 s. A significant loss of photoactivity wasseen after long exposure times (>2000 s). The loss in photoactiv-ity was found to be more rapid when a potential was applied.The loss in photoactivity could be due to the pin holes in thetitania coated Cu2O electrodes. After 300 s no photoactivity wasobserved with cuprous oxide nanowire electrodes when a poten-tial of −0.55 V vs. Ag/AgCl was applied. Similar observations on the

stability of titania coated Cu2O electrodes, were made by Gratzeland co-workers who found that a black layer, corresponding to for-mation of copper, was found on the illuminated region of the Cu2Ophotocathodes. The CuO thin film and NW electrodes were only

S. Sunkara et al. / Catalysis Today 199 (2013) 27– 35 33

Fig. 5. (a) XRD diffractograms of the Cu2O and CuO thin films and (b) NW arrayssynthesized on copper foils.

Fig. 6. Tauc plots for CuO and Cu2O NW for determining the band gap. The insetshows the Tauc plots for CuO and Cu2O thin films. Both Cu2O thin films and nanowirearrays samples show the presence of CuO, which is most likely to be present as anoxide layer beneath the nanowires.

Fig. 7. Mott–Schottky plot for Cu2O NW and thin films obtained at1 Sun illumination (100 mW/cm2) and 10 Hz frequency.

Fig. 8. Photovoltage transients for the Cu2O and CuO thin films and nanowires underopen circuit condition. The electrodes were illuminated with a chopped AM 1.5 light(100 mW/cm2).

Fig. 9. Current density vs. potential characteristics for the Cu2O, CuO thin films andnanowire arrays measured under chopped AM 1.5 illumination (100 mW/cm2).

34 S. Sunkara et al. / Catalysis Today 199 (2013) 27– 35

F on tha

s5

ealnsfdita

4

otstccopattrCsctor

A

tam–t

[[

[

[

[

[

[

[17] E. Fortin, D. Masson, Photovoltaic effects in Cu2O/Cu solar cells grown by anodic

ig. 10. Scanning electron microscopy images of (a) copper hydroxide film formedfter plasma annealing for 2 min.

table for few minutes and showed no photocatalytic activity after min.

The fabrication of a thick mesoporous cuprous oxide nanowirelectrode without any interfacial oxide layer would be the idealrchitecture. This would provide the necessary thickness foright absorption, provide more intimate contact between theanowires and electrolyte for better charge transport at theemiconductor–electrolyte interface, solve the issue of phase purityor interfacial layers and would eliminate recombination lossesue to the polycrystalline oxide layer. Our ongoing research efforts

nclude the fabrication of these mesoporous electrodes and inves-igation of alternative protective layers for improving the stabilitynd performance of the cuprous oxide nanowires.

. Conclusions

A generic and scalable synthesis technique for growing cuprousxide nanowire arrays on different substrates is presented. The syn-hesis of nanowire arrays was carried out on copper foils, glasslides coated with fluorinated tin oxide, and could be extendedo other substrates. Experiments using hydrogen peroxide for wethemical oxidation suggest that the nucleation and growth ofopper hydroxide nanowires proceed in similar fashion to thatf plasma oxidation. Also, the combination of wet chemical andlasma oxidation can be interesting for reducing the time scales for

variety of other materials systems. Photoelectrochemical charac-erization of the copper oxide films and nanowire arrays showedhat the nanowires show a significant enhancement in photocur-ents due to increased carrier generation and lower recombination.uprous oxide nanowires grown on gold sputtered FTO slideshowed a performance similar to that of the nanowires grown onopper foils suggesting that the presence of Schottky barrier is nothe major factor limiting the performance. The presence of a cupricxide layer beneath the nanowire arrays limits the high photocur-ents that can be achieved with nanowire arrays.

cknowledgements

The authors acknowledge support from DOE toward infras-ructure support to Conn Center for Renewable Energy Researcht University of Louisville (DE-EE0003206), Kentucky Depart-

ent of Energy Development and Independence (KY DEDI

RFP–1271000000448) and personnel, infrastructure support fromhe Conn Center.

[

e FTO slides after the wet chemical oxidation and (b) growth of Cu2O NW arrays

Appendix A. Supplementary data

Supplementary data associated with this arti-cle can be found, in the online version, athttp://dx.doi.org/10.1016/j.cattod.2012.03.014.

References

[1] M.G. Walter, E.L. Warren, J.R. McKone, S.W. Boettcher, Q.X. Mi, E.A. Santori, N.S.Lewis, Solar water splitting cells, Chemical Reviews 110 (2010) 6446–6473.

[2] Z.B. Chen, T.F. Jaramillo, T.G. Deutsch, A. Kleiman-Shwarsctein, A.J. Forman,N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E.W. McFarland,K. Domen, E.L. Miller, J.A. Turner, H.N. Dinh, Accelerating materials develop-ment for photoelectrochemical hydrogen production: standards for methods,definitions, and reporting protocols, Journal of Materials Research 25 (2010)3–16.

[3] A.O. Musa, T. Akomolafe, M.J. Carter, Production of cuprous oxide, a solar cellmaterial, by thermal oxidation and a study of its physical and electrical prop-erties, Solar Energy Materials and Solar Cells 51 (1998) 305–316.

[4] A.E. Rakhshani, Preparation, characteristics and photovoltaic properties ofcuprous-oxide – a review, Solid-State Electronics 29 (1986) 7–17.

[5] A.P. Chatterjee, A.K. Mukhopadhyay, A.K. Chakraborty, R.N. Sasmal, S.K. Lahiri,Electrodeposition and characterization of cuprous-oxide films, Materials Let-ters 11 (1991) 358–362.

[6] G.P. Pollack, D. Trivich, Photoelectric properties of cuprous-oxide, Journal ofApplied Physics 46 (1975) 163–172.

[7] H. Gerischer, Stability of semiconductor electrodes against photodecomposi-tion, Journal of Electroanalytical Chemistry 82 (1977) 133–143.

[8] W. Siripala, A. Ivanovskaya, T.F. Jaramillo, S.H. Baeck, E.W. McFarland, ACu2O/TiO2 heterojunction thin film cathode for photoelectrocatalysis, SolarEnergy Materials and Solar Cells 77 (2003) 229–237.

[9] A. Paracchino, V. Laporte, K. Sivula, M. Gratzel, E. Thimsen, Highly active oxidephotocathode for photoelectrochemical water reduction, Nature Materials 10(2011) 456–461.

10] B.P. Rai, Cu2O solar-cells – a review, Solar Cells 25 (1988) 265–272.11] C.J. Engel, T.A. Polson, J.R. Spado, J.M. Bell, A. Fillinger, Photoelectrochemistry

of porous p-Cu2O films, Journal of the Electrochemical Society 155 (2008)F37–F42.

12] P.E. de Jongh, D. Vanmaekelbergh, J.J. Kelly, Photoelectrochemistry of electrode-posited Cu2O, Journal of the Electrochemical Society 147 (2000) 486–489.

13] M.K. Sunkara, C. Pendyala, D. Cummins, P. Meduri, J. Jasinski, V. Kumar, H.B.Russell, E.L. Clark, J.H. Kim, Inorganic nanowires: a perspective about their rolein energy conversion and storage applications, Journal of Physics D: AppliedPhysics 44 (2011) 174032–174041.

14] V. Chakrapani, J. Thangala, M.K. Sunkara, WO(3) and W(2)N nanowire arrays forphotoelectrochemical hydrogen production, International Journal of HydrogenEnergy 34 (2009) 9050–9059.

15] M. Izaki, K.-t. Mizuno, T. Shinagawa, M. Inaba, A. Tasaka, Photochemical con-struction of photovoltaic device composed of p-copper(I) oxide and n-zincoxide, Journal of the Electrochemical Society 153 (2006) C668–C670.

16] Z.G. Yin, H.T. Zhang, D.M. Goodner, M.J. Bedzyk, R.P.H. Chang, Y. Sun, J.B. Ket-terson, Two-dimensional growth of continuous Cu2O thin films by magnetronsputtering, Applied Physics Letters 86 (2005) 061901–061904.

oxidation, Solid-State Electronics 24 (1982) 281–283.18] C.X. Xiang, G.M. Kimball, R.L. Grimm, B.S. Brunschwig, H.A. Atwater, N.S. Lewis,

820 mV open-circuit voltages from Cu(2)O/CH(3)CN junctions, Energy & Envi-ronmental Science 4 (2011) 1311–1318.

ysis To

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[solar-cells, Solar Cells 7 (1982) 247–279.

[45] L.C. Olsen, R.C. Bohara, M.W. Urie, Explanation for low-efficiency Cu2O

S. Sunkara et al. / Catal

19] M. Ristov, G. Sinadinovski, I. Grozdanov, Chemical-deposition of Cu2O thin-films, Thin Solid Films 123 (1985) 63–67.

20] W.Z. Wang, G.H. Wang, X.S. Wang, Y.J. Zhan, Y.K. Liu, C.L. Zheng, Synthesisand characterization of Cu2O nanowires by a novel reduction route, AdvancedMaterials 14 (2002) 67–69.

21] Z.C. Orel, A. Anzlovar, G. Drazic, M. Zigon, Cuprous oxide nanowires preparedby an additive-free polyol process, Crystal Growth & Design 7 (2007) 453–458.

22] A.S. Walton, M.L. Gorzny, J.P. Bramble, S.D. Evans, Photoelectric properties ofelectrodeposited copper(I) oxide nanowires, Journal of the ElectrochemicalSociety 156 (2009) K191–K200.

23] H.S. Shin, J.Y. Song, J. Yu, Template-assisted electrochemical synthesis ofcuprous oxide nanowires, Materials Letters 63 (2009) 397–399.

24] Y.J. Xiong, Z.Q. Li, R. Zhang, Y. Xie, J. Yang, C.Z. Wu, From complex chains to 1Dmetal oxides: a novel strategy to Cu2O nanowires, Journal of Physical ChemistryB 107 (2003) 3697–3702.

25] Y.W. Tan, X.Y. Xue, Q. Peng, H. Zhao, T.H. Wang, Y.D. Li, Controllable fabrica-tion and electrical performance of single crystalline Cu2O nanowires with highaspect ratios, Nano Letters 7 (2007) 3723–3728.

26] S. Sahoo, S. Husale, B. Colwill, T.M. Lu, S. Nayak, P.M. Ajayan, Electric fielddirected self-assembly of cuprous oxide nanostructures for photon sensing,ACS Nano 3 (2009) 3935–3944.

27] X.C. Jiang, T. Herricks, Y.N. Xia, CuO nanowires can be synthesized by heatingcopper substrates in air, Nano Letters 2 (2002) 1333–1338.

28] C.H. Xu, C.H. Woo, S.Q. Shi, Formation of CuO nanowires on Cu foil, ChemicalPhysics Letters 399 (2004) 62–66.

29] Y. Wang, R.Q. Shen, X.Y. Jin, P. Zhu, Y.H. Ye, Y. Hu, Formation of CuO nanowiresby thermal annealing copper film deposited on Ti/Si substrate, Applied SurfaceScience 258 (2011) 201–206.

30] M. Kaur, K.P. Muthe, S.K. Despande, S. Choudhury, J.B. Singh, N. Verma, S.K.Gupta, J.V. Yakhmi, Growth and branching of CuO nanowires by thermal oxi-dation of copper, Journal of Crystal Growth 289 (2006) 670–675.

31] A.L. Jin-Woo Han, P. Nobuhiko, M. Kobayashi, 1. Meyyappan:, Evolutional trans-formation of copper oxide nanowires to copper nanowires by a reductiontechnique, Materials Express 1 (2011) 176–180.

32] F. Qian, G.M. Wang, Y. Li, Solar-driven microbial photoelectrochemical cellswith a nanowire photocathode, Nano Letters 10 (2010) 4686–4691.

33] V. Kumar, J.H. Kim, J.B. Jasinski, E.L. Clark, M.K. Sunkara, Alkali-assisted, atmo-spheric plasma production of titania nanowire powders and arrays, CrystalGrowth & Design 11 (2011) 2913–2919.

[

day 199 (2013) 27– 35 35

34] W.X. Zhang, X.G. Wen, S.H. Yang, Y. Berta, Z.L. Wang, Single-crystalline scroll-type nanotube arrays of copper hydroxide synthesized at room temperature,Advanced Materials 15 (2003) 822–825.

35] N.K. Chaudhari, B. Fang, T.S. Bae, J.S. Yu, Low temperature synthesis of singlecrystalline iron hydroxide and oxide nanorods in aqueous media, Journal ofNanoscience and Nanotechnology 11 (2011) 4457–4462.

36] C.J. Jia, L.D. Sun, Z.G. Yan, L.P. You, F. Luo, X.D. Han, Y.C. Pang, Z. Zhang, C.H. Yan,Iron oxide nanotubes – single-crystalline iron oxide nanotubes, AngewandteChemie International Edition 44 (2005) 4328–4333.

37] H. Hou, Y. Xie, Q. Yang, Q. Guo, C. Tan, Preparation and characterization of�-AlOOH nanotubes and nanorods, Nanotechnology 6 (2005).

38] U. Cvelbar, Z.Q. Chen, M.K. Sunkara, M. Mozetic, Spontaneous growth ofsuperstructure alpha-Fe2O3 nanowire and nanobelt arrays in reactive oxygenplasma, Small 4 (2008) 1610–1614.

39] U. Cvelbar, K. Ostrikov, I. Levchenko, M. Mozetic, M.K. Sunkara, Control ofmorphology and nucleation density of iron oxide nanostructures by elec-tric conditions on iron surfaces exposed to reactive oxygen plasmas, AppliedPhysics Letters 94 (2009) 741–745.

40] L.C. Wang, M. Tao, Fabrication and characterization of p–n homojunctions incuprous oxide by electrochemical deposition, Electrochemical and Solid StateLetters 10 (2007) H248–H250.

41] K. Nakaoka, J. Ueyama, K. Ogura, Photoelectrochemical behavior of electrode-posited CuO and Cu2O thin films on conducting substrates, Journal of theElectrochemical Society 151 (2004) C661–C670.

42] Z. Mao, W. Song, L. Chen, W. Ji, X.X. Xue, W.D. Ruan, Z.S. Li, H.J. Mao, S. Ma, J.R.Lombardi, B. Zhao, Metal-semiconductor contacts induce the charge-transfermechanism of surface-enhanced raman scattering, Journal of Physical Chem-istry C 115 (2011) 18378–18380.

43] M. Kevin, W.L. Ong, G.H. Lee, G.W. Ho, Formation of hybrid structures: copperoxide nanocrystals templated on ultralong copper nanowires for open networksensing at room temperature, Nanotechnology 22 (2011) 235701–2357011.

44] L.C. Olsen, F.W. Addis, W. Miller, Experimental and theoretical-studies of Cu2O

schottky-barrier solar-cells, Applied Physics Letters 34 (1979) 47–49.46] C. Richter, C.A. Schmuttenmaer, Exciton-like trap states limit electron mobility

in TiO(2) nanotubes, Nature Nanotechnology 5 (2010) 769–772.