Up-scaling the synthesis of Cu2O submicron particles with controlledmorphologies for solar H2 evolution from water

Enrique Carbó-Argibay a, Xiao-Qing Bao a, Carlos Rodríguez-Abreu a, M. Fátima Cerqueira b,Dmitri Y. Petrovykh a, Lifeng Liu a, Yury V. Kolen’ko a,⇑a International Iberian Nanotechnology Laboratory, Braga 4715-330, Portugalb Center of Physics, University of Minho, Braga 4710-057, Portugal

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 22 April 2015Revised 8 June 2015Accepted 11 June 2015Available online 17 June 2015

Keywords:Cu2O colloidsScale upReductionMechanism of formationPhotocathodeWater splitting

a b s t r a c t

The synthesis of Cu2O was studied to examine the effects of up-scaling on the size and morphology of theresultant particles. As a result, a successful protocol employing an automated laboratory reactor wasdeveloped for large-scale synthesis of phase-pure Cu2O colloids with specific sizes in the submicron tomicrometer range (0.2–2.6 lm). The as-synthesized products have been studied by means of powderX-ray diffraction, Raman spectroscopy, X-ray photoelectron spectroscopy, UV–Vis–NIR spectroscopy,scanning electron microscopy, and photoelectrochemical measurements. A broad range of morphologies,both equilibrium (stellated octahedrons, cubes, cuboctahedrons, truncated octahedrons, truncatedcuboctahedrons) and metastable (cage-like hierarchical structures, microspheres with flower-like tex-ture), with uniform sizes have been selectively prepared either by careful tuning of synthesis conditions.Recrystallization of primary aggregates through Ostwald ripening is proposed as the formation mecha-nism for these Cu2O structures. As a photocathode for photoelectrochemical H2 evolution, Cu2O submi-cron cubes with exposed {001} facets exhibit a high open-circuit potential of ca. 0.9 V vs. the RHE atpH 1.

! 2015 Elsevier Inc. All rights reserved.

http://dx.doi.org/10.1016/j.jcis.2015.06.0140021-9797/! 2015 Elsevier Inc. All rights reserved.

⇑ Corresponding author.

Journal of Colloid and Interface Science 456 (2015) 219–227

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

journal homepage: www.elsevier .com/locate / jc is

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jcis.2015.06.014&domain=pdfhttp://dx.doi.org/10.1016/j.jcis.2015.06.014http://dx.doi.org/10.1016/j.jcis.2015.06.014http://www.sciencedirect.com/science/journal/00219797http://www.elsevier.com/locate/jcis

1. Introduction

Cu2O belongs to a class of relatively rare p-type semiconductingoxide materials, and holds promise for applications in photovoltaics(PVs) [1], thin-film transistors (TFTs) [2], photocatalytic water split-ting [3], and solar-assisted photoelectrochemical (PEC) H2 evolu-tion from water [4]. These intensive research activities in Cu2Oapplications stem from several advantageous properties of thismaterial. In particular, Cu2O has direct electronic band gap of 1.9–2.1 eV [5–8], high absorption coefficient in the violet-to-greenregion above the gap [9], hole mobility of up to 100 cm2 V!1 s!1

and minority-carrier diffusion lengths of 10 lm [8]. Moreover, cop-per is naturally abundant and a low-cost element.

Cu2O is especially appealing to us given our continuing effortsto implement Cu2O in solar H2 fuel technology as a viable replace-ment for current Pt-containing photocathodes in PEC water split-ting. Furthermore, interested in the possibility of fabricatinglarge-area films using doctor blade technique, our attention wasconcentrated on large-scale synthesis of Cu2O colloids that serveas precursors for ink formulation. Notably, doctor blade is particu-larly attractive for Cu2O film manufacturing compared to the mostcommon thermal oxidation, electrodeposition or physical evapora-tion techniques, because of the convenient large-area depositionusing simple apparatus at ambient conditions. The need for largequantities of Cu2O ink precursors with controlled morphologiesand sizes is, however, a prerequisite for efficient film manufactur-ing by this process. Hence, a reproducible, robust, and controlledsynthesis protocol is desirable for Cu2O colloids to enable the inkpreparation on a large scale, which is critical for the realizationof emerging and future Cu2O applications in clean-energytechnology.

A major advantage of Cu2O material is that it can be producedby cheap and environmentally friendly wet-chemistry methods.In particular, Cu2O particles with various morphologies have beenproduced via different synthesis protocols, including thermaldecomposition [10–12], solvothermal [13], and hydrothermal[14] methods, polyol synthesis [15,16], seed-mediated chemicaldeposition [17], and via the most common liquid-phase reductionof Cu2+ either by hydrazine monohydrate [18–20], sodiumborohydride [21], ascorbic acid [22], or hydroxylamine hydrochlo-ride [23]. These syntheses are often conducted under control of asurfactant, such as oleic acid, poly(vinylpyrrolidone), or sodiumdodecyl sulfate, which is used to generate colloidal dispersionsand to direct the microstructure of the resultant Cu2O materials.

Despite the advances in Cu2O synthesis, the reproducibility ofsynthesis protocols is a challenge. This is mainly due to the highsensitivity of the final Cu2O products to quality of starting reagents,their concentrations, shelf-time, as well as to many reaction condi-tions, such as stirring speed, temperature, tempering time, etc.Moreover, scale-up production of nanomaterials is, in general, anon-trivial task [24], since mass and heat-transfer rates are signif-icantly altered during up-scaling [25]. This alteration stronglyaffects the principal processes of nucleation and crystal growthof the targeted compound, leading to low reproducibility of theresults from smaller scale protocol. Typically, this is reflected bye.g. different particle size, particle size distribution, morphology,and phase composition of the resultant up-scaled products.

There are only few reports focused on the preparation of gramquantities of uniform and shape-controlled Cu2O particles in thenm–lm size regime [26]. Therefore, successfully adapting and fur-ther optimizing convenient and controllable synthesis protocolsup-scaled from the original milligram-scale methods is not com-monly achieved or reported for Cu2O materials. Our recent successin developing large-scale synthesis of iron oxide nanocolloids usingan automated laboratory reactor [27] motivated us to study the

possibility of extending this approach to the synthesis of Cu2Ocolloids. Accordingly, here we report (1) the effect of synthesisup-scaling on the properties of the resultant Cu2O colloids, (2)preparation of large quantities of Cu2O materials having differentuniform shapes, (3) the mechanism of particle formation, and (4)a preliminary investigation of the PEC performance of theas-synthesized Cu2O materials.

2. Materials and methods

2.1. Starting materials

CuCl2"2H2O (99%), sodium dodecyl sulfate (SDS) (98%), andNH2OH"HCl (99%) were obtained from Sigma–Aldrich; NaOH(98.9%) and EtOH (analytical reagent grade) were purchased fromFisher Scientific. All reagents were used as received without furtherpurification. Ultrapure water used in the study was produced usinga Milli-Q Advantage A10 system (Millipore).

2.2. Synthesis

Cu2O samples were prepared by modifying and up-scaling theprocedure outlined by Huang and co-workers [23]. One batch ofCu2O cubes was prepared following exactly the literature method.The 10-fold up-scaling of this synthesis was conducted followingthe same procedure in a single-neck round-bottom flask chargedwith 10 times larger amount of all reagents.

The 100-fold up-scaling was realized in an automated synthesissystem (Atlas Potassium, Syrris), as illustrated by a typical time logof the synthesis in Fig. 1. First, a 2 L reaction glass vessel wascharged with 1 L of water (pH 7.7) and mechanical stirrer wasset to 300 rpm. The reaction system was heated up to 34 "C usinga temperature-control system (LH85 PLUS, Julabo), while adding55 mL of 0.1 M CuCl2 solution at 20 mL/min via a syringe pumpto obtain a clear faint-blue solution (pH 4.8). Next, 9.7 g of SDSwas added, and the mixture was stirred for 10 min to achieve com-plete dissolution of SDS (pH 5.2). Then, 20 mL of 1.0 M NaOH solu-tion was added at 10 mL/min using a second syringe pump toobtain a milky-blue slurry (pH # 11.4). After the addition was com-plete, 45 mL of 0.1 M NH2OH"HCl reducing solution was rapidlyadded into the system while stirring at 600 rpm for 20 s, turning

Fig. 1. Time log of an automated reactor synthesis of Cu2O submicron cubesshowing temperature, turbidity, and pH. (1) Addition of 0.1 M CuCl2 solution; (2)adjusting temperature to 34 "C; (3) addition of SDS; (4) addition of 1.0 M NaOHsolution; (5) addition of 0.1 M NH2OH"HCl solution; (6) aging. The inset shows theautomated reactor glass vessel at the final stage of the synthesis.

220 E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227

the solution to a green slurry (pH 11.1). Then, the stirring wasstopped and the system was aged at 34 "C for 1 h (pH 11.3).During this hour, the color of the slurry was typically observed togradually turn from green to orange as a result of Cu2O formation.

Additional experiments to study the influence of the reducingagent on the morphology of the resultant particles were conductedby varying the amount of added NH2OH"HCl. To neutralize the HClgenerated by hydrolysis of NH2OH"HCl, and thus avoidacid-etching of Cu2O, equimolar amounts of aqueous NaOH wereadded. All reagents were mixed in a manner similar to thatdescribed above for the 100-fold synthesis, but with varyingamounts of the reducing agent: 19, 21, 22, 24, and 28 mL of1.0 M NaOH solution was charged into the system at 10 mL/minfollowed by the addition of 35, 55, 65, 85, and 125 mL 0.1 MNH2OH"HCl reducing agent solution, respectively.

To investigate the influence of different synthesis parameterson the microstructural properties of the resultant products, severalsyntheses were carried out with varying concentrations ofreagents, mixing, and without neutralization of HCl fromNH2OH"HCl.

2.3. Washing

After aging, the reaction product was collected by centrifuga-tion at 9000 rpm for 5 min and thoroughly washed with waterand ethanol to remove excess SDS and byproducts. Specifically,the resultant solid from centrifugation was redispersed in 80 mLof water/ethanol mixture (1:1) using ultrasonication in anElmasonic P bath (Elma) for 10 min with vortex mixing (VWR)in-between. The obtained orange suspension was centrifuged at5000 rpm for 3 min. The supernatant was discarded, while theresulting solid was washed 2 more times with water/ethanol(1:1) and 2 more times with ethanol using copious workup, asdescribed above, to yield an orange product. Finally, the productwas dispersed and stored as a suspension in 5 mL of ethanol. Theas-derived Cu2O is stable under ambient air and does not hydro-lyze in water.

2.4. Characterization

The products were characterized by scanning electron micro-scopy (SEM) (Quanta 650 FEG microscope, FEI Company), energydispersive X-ray spectroscopy (EDX) (INCA 350 spectrometer,Oxford Instruments), powder X-ray diffraction (XRD) (X’Pert PROdiffractometer, PANalytical), Raman spectroscopy (alpha300 R con-focal microscope, WITec), X-ray photoelectron spectroscopy (XPS)(ESCALAB 250 Xi, Thermo Scientific) [27,28], and UV–Vis–NIR spec-troscopy (Lambda 950 spectrophotometer, Perkin Elmer). For moredetails regarding characterization, see the Supplementary Material(SM).

2.5. Film fabrication

For PEC measurements, where specimens are required in theform of thin films, Cu2O samples were deposited using the doctorblade technique onto glass plates coated with fluorine-doped tinoxide (FTO, TEC 15, 12–14 X/h). Before deposition, each glass plate(3.5 $ 3.5 cm2) was sequentially cleaned with soap solution andacetone, and then dried. To obtain a uniform Cu2O film, 0.1 mL ofhomogeneous paste-like Cu2O suspension in ethanol was spreadonto the FTO-coated surface of the glass plate using the doctorblade tool with thickness set at 15 lm. The film was allowed todry under ambient conditions, and was then subjected to rapidthermal annealing in a tube furnace (Carbolite) at 450 "C for10 min in high-purity N2 (99.999%) prior to the PEC measurements.

2.6. PEC measurements

PEC H2 evolution properties of as-fabricated Cu2O films werestudied in a three-electrode configuration with a Cu2O film asthe working electrode (photocathode), a Pt coil as the counter elec-trode, and a Hg/Hg2SO4 (saturated K2SO4) electrode (MSE) as thereference. The electrolyte consisted of 0.5 M K2SO4 solution withpH value adjusted to 1 or 4 using H2SO4. The dynamic photore-sponse of electrodes was recorded at a scan rate of 20 mV/s(toward cathodic direction) under chopped illumination (choppingfrequency: 0.5 Hz) using a Zennium electrochemical workstation(Zahner). The light source was a calibrated tungsten lamp workingat 100 mW/cm2 (Zahner). The temperature of the sample wasmaintained at 23 ± 1 "C during the test using a home-made coolingbox connected with a refrigerated chiller (HAAKE Phoenix II,Thermo Scientific). All potentials are given vs. reversible hydrogenelectrode (RHE) by converting the measured potentials vs. MSEaccording to: URHE = UMSE + 0.654V + 0.059 pH.

3. Results

3.1. Structure and composition

EDX analysis reveals the presence of only Cu and O in theobtained products, exhibiting homogeneous distribution withinthe particles (Fig. S1, SM). According to the XRD (Fig. S2, SM), allas-synthesized samples are phase-pure cuprite, Cu2O, (ICDD No.01-080-7711, cubic, Pn-3m) having a minor admixture of tenorite,CuO, (ICDD No. 04-004-5425, monoclinic, C2/c), which is inevitablypresent at the oxidized surface of copper(I) oxide. The phase purityof the products was also confirmed by Raman spectroscopy [29–33], as detailed in the SM (Figs. S3 and S4, SM).

XPS data in Cu 2p region clearly indicate that the surface of theparticles is dominated by Cu2+, the presence of which is evidencedby the strong satellites in 940–945 eV and 960–965 eV ranges aswell as shoulders at ca. 935 and 955 eV (Fig. 2). The Cu LMMAuger peak (data not shown), which corresponds to a samplingdepth larger by ca. 2 nm than that of the Cu 2p peak, qualitativelyindicates that Cu+ is present below the oxidized surface layer.

3.2. Size and morphology

3.2.1. Effect of up-scalingThe up-scaling of the original synthesis results in procedures

that efficiently produce Cu2O particles of well-defined regular

Fig. 2. High-resolution XPS data in Cu 2p region for Cu2O particles produced by the100-fold up-scaling process.

E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227 221

cubic morphology surrounded by six square {001} facets (Fig. 3).Furthermore, the size of the particles changes with up-scaling.Specifically, synthesis using the original method produces submi-cron cubes exhibiting a uniform size distribution around 0.22 lm(Fig. 3a). Syntheses with 10-fold and 100-fold up-scaling produceuniform 0.65-lm (Fig. 3b) and 0.42-lm (Fig. 3c) particles, respec-tively, with overall cubic shapes, sometimes with slightly concavesides.

3.2.2. Effect of synthesis conditionsTo study the formation mechanism, the influence of the differ-

ent synthesis conditions on the growth of Cu2O products was clar-ified, as described in the SM (Figs. S5–S8, SM). In particular, it was

not possible to obtain Cu2O particles with continued 300-rpm stir-ring during 1 h aging at 34 "C, or in the absence of eitherNH2OH"HCl reducing agent or SDS, highlighting the importanceof these specific synthesis parameters.

The microstructure of Cu2O particles strongly depends on theconcentration of the reagents in the precursor solution. With theamount of water and SDS unchanged, lowering concentration ofthe reactants by a factor of 3 relative to our reference 100-foldsynthesis (Fig. 3c) resulted in the Cu2O product that exhibiteduniform 0.21-lm stellated octahedron shape featuringhalf-spindle-like structures on their surface (Fig. 4a). The forma-tion of such a nonequilibrium crystal shape is energetically unfa-vorable [34], indicating limited diffusion of the reacting speciestoward the interface during the synthesis. Conversely, a 3-foldincrease in the concentration of the reactants enhanced bothnucleation and diffusion, resulting in energetically favorable struc-tures, namely truncated octahedron or cuboctahedron shapes [35]with a non-uniform size distribution in the range from 0.4 to1.3 lm (Fig. 4b).

The morphology of Cu2O is also effectively controlled by thespeed and duration of the stirring just after the addition ofNH2OH"HCl. When the speed and duration were increased to800 rpm and 60 s, respectively, cage-like hierarchical structureswith a relatively uniform size of ca. 0.85 lm were obtained(Fig. 4c). As in the case of the reduced reagent concentrations,the formation of such energetically unfavorable structures sug-gests limited diffusion during the synthesis. The cage-like struc-tures seem to form by the self-assembly of many Cu2Ospindle-like nanoparticles, which are interwoven into a hierarchi-cal structure. This well-organized open porous structure isexpected to facilitate gas/liquid diffusion within the particles,which could be of interest for catalytic applications.

As expected, the pH value was found to strongly affect themicrostructure of Cu2O products. Specifically, the hydrolysis ofNH2OH"HCl reducing agent generates HCl, which can partially etchthe Cu2O particles. Accordingly, our 100-fold synthesis with anincreased quantity of NH2OH"HCl (125 mL) but without subse-quent neutralization of the resultant HCl by an equimolar amountof NaOH (pH 6.3 vs. pH 11.3 under reference conditions) producedtruncated octahedron Cu2O particles with an average size of ca.1.42 lm (Fig. 4d). The >1 lm size of these particles is directlyrelated to the increased quantity of the reducing agent comparedto the reference 100-fold procedure, while their morphology isproduced by the HCl etching.

The hexagonal facets of the truncated octahedrons in Fig. 4dhave clearly been partially dissolved, generating oval-shaped flatfacets. The HCl etching may also produce voids within the struc-tures, as indicated by the round holes on surfaces of somemicroparticles in Fig. 4d. The etching process was further enhancedunder the same pH value of 6.3 if the stirring was continued at300 rpm during aging, resulting in the formation of monodisperse2.6-lm microspheres featuring flower-like surface texture com-posed of smooth nanopetals (Fig. 4e). Our Raman analysis of thesamples clearly indicates that etching gives rise to the intensityinversion of the 219 and the 100 cm!1 modes, which is most likelydue to the development of Cu deficiency in the final product(Figs. S3 and S4, SM).

3.2.3. Amount of reducing agentThe examples in Fig. 4d and 4e clearly demonstrate how the

reduction capacity of the solution can dramatically affect the mor-phology of Cu2O particles. To further explore the possibilities of tun-ing the morphology of submicron particles, the amount of thereducing agent was systematically varied: Fig. 5 presents a series ofsamples that were prepared by adding 35, 45, 55, 65, 85, and125 mL of 0.1 M NH2OH"HCl solution. Etching observed in

Fig. 3. The effect of up-scaling on the morphology and size of the Cu2O particles.Typical low- (main panels) and high- (insets) magnification SEM images of Cu2Osubmicron cubes obtained according to the reported method (a) [23] as well as via10-fold (b) and 100-fold (c) up-scaling of that method.

222 E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227

Fig. 4d and 4e was minimized via neutralization of the HCl fromhydrolyzed NH2OH"HCl by an equimolar amount of NaOH, asdescribed in Section 2. By controlling both the pH and reductioncapacity, a wide range of largely uniform shapes was obtained,including 0.81-lm stellated octahedrons (Fig. 5a, 35 mLNH2OH"HCl), 0.42-lm slightly concave cubes (Figs. 3c and 5b,45 mL NH2OH"HCl), 0.55–lm lamella-covered cubes (Fig. 5c, 55 mLNH2OH"HCl), 0.68-lm distorted cubes (Fig. 5d, 65 mL NH2OH"HCl),0.71-lm cuboctahedrons (Fig. 5e, 85 mL NH2OH"HCl), and 0.52-lm

truncated octahedrons (Fig. 5d, 125 mL NH2OH"HCl). In agreementwith previous reports [23], Fig. 5 illustrates the versatility of tuningthe morphology of submicron Cu2O particles by adjusting theamount of the reducing agent.

3.3. Optical properties

The UV–Vis–NIR absorbance spectra in Fig. 6 show thesize-dependent optical properties of reference Cu2O submicron

Fig. 4. Effects of synthesis parameters on the microstructure of Cu2O particles. Representative low- (left panels) and high- (right panels) magnification SEM images illustratethe following variations of our reference 100-fold synthesis (Fig. 3c): 3 times lower (a) or higher (b) concentrations of the reactants, maintaining water and SDS unchanged;(c) 1 min stirring at 800 rpm just after addition of NH2OH"HCl; (d) adding 125 mL of NH2OH"HCl without neutralizing the resultant HCl by an equimolar amount of NaOH; (e)stirring at 300 rpm during aging for 1 h in addition to the synthesis conditions of (d).

E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227 223

Fig. 5. Comparison of the morphological features of Cu2O products prepared using automated synthesis system. Representative low- (left panel) and high- (right panel)magnification SEM images illustrate Cu2O samples prepared with different amounts of NH2OH"HCl added to the reaction—35 (a), 45 (b), 55 (c), 65 (d), 85 (e), and 125 mL (f)—while neutralizing the resultant HCl by an equimolar amount of NaOH to avoid etching.

224 E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227

cubes imaged in Fig. 3. The spectra show a sharp band at 500–510 nm corresponding to intrinsic absorption of Cu2O. Direct opti-cal bandgap estimates for our materials, based on the Tauc plotanalysis [36], are 1.99–2.05 eV, in agreement with the 1.9–2.1 eVrange of theoretical and experimental literature values [5–7]. Abroad absorption band at ca. 660 nm (for 0.42-lm and 0.65-lmcubes) is most likely associated with the presence of the Cu(OH)2precursor strongly adsorbed on the surface of Cu2O particles(Fig. S9, SM).

Interestingly, all the samples also exhibit a very broad absorp-tion band throughout the visible region and NIR with maxima ataround 950 nm for 0.22-lm cubes and 1500 nm for 0.42-lm and0.65-lm cubes, respectively. We presume that this NIR absorbancearises from excitations of the free carriers in the highly-dopedp-type Cu2O, potentially with plasmonic contributions [7].

3.4. PEC H2 evolution properties

To examine the potential of as-synthesized Cu2O colloids forsolar H2 evolution from water, the PEC properties of the submicronparticles were investigated under simulated solar illumination atdifferent pH values. A concentrated ethanol suspension of0.65-lm Cu2O cubes from 10-fold synthesis (Fig. 3b) was depositedas a thick Cu2O film (photocathode) on a conducting FTO substrateby the doctor blade technique. To prevent peeling off the Cu2Omaterial from FTO substrate and to ensure good electrical contactbetween Cu2O and FTO, the as-fabricated photocathode was sub-jected to quick annealing at 450 "C for 10 min in high-purity N2(inset in Fig. 7).

The PEC performance comparison of the photocathode at pHvalues of 1 and 4 is shown in Fig. 7, highlighting its excellent pho-toresponse. Even without surface protection against corrosionunder acidic conditions, a high open-circuit potential close to0.9 V vs. RHE was observed at pH 1. This potential is higher thanthe difference (ca. 0.54 V) between the Fermi-level of Cu2O(0.48 V vs. SHE) and the RHE potential (!0.06 V vs. SHE). This highphotovoltage of 0.9 V is, apparently, dictated by the equilibrium ofthe Fermi-level of Cu2O with the mixed potential arising from bothH+/H2 and Cu2O/Cu redox couples. When a bare Cu2O electrode isimmersed in acidic media, a fraction of the photocurrent is inducedby the reduction of Cu2O to Cu0 (E0 = !0.36 V vs. SHE) according tothe Eq. (1) [4,37]:

Cu2Oþ 2Hþ þ 2e! ¼ 2CuþH2O ð1Þ

The photoresponse decayed rapidly along the cathodic scandirection and finally disappeared beyond 0.2 V, indicating fastdegradation of the electrode due to the dissolution of Cu2O atpH 1. In contrast, at pH 4 a slight decrease in photoresponse wasobserved, while the stability was significantly improved (Fig. 7).Nevertheless, without a corrosion-protective coating on the photo-cathode surface, Cu2O will degrade in acidic electrolyte with time,depending on the solution pH, wherein lower pH enables a highercorrosion rate. Interestingly, an increase in dark current duringpotential sweep towards cathodic direction was observed. Mostlikely this dark current results from tunnelling breakdown acrossthe shallow space charge region of Cu2O dictated by thehighly-doped nature of p-type Cu2O [37]. The good PEC H2 evolu-tion performance of our Cu2O submicron cube photocathode maystem from enhanced visible light absorption among Cu2O cubesdue to multi-reflection (cf. Fig. 6). Moreover, the capability of{0 0 1} facets of the cubic particles to dissociate adsorbed watermolecules can additionally enhance their PEC performance[38,39]. Further improvements of the electrode stability as wellas shape and size effect of Cu2O particles on the solar H2 evolutionperformance are being investigated.

4. Discussion

4.1. Synthetic aspects of up-scaling

The large-scale synthesis of earth-abundant advanced materialsfor clean energy technology is of particular interest in our labora-tory. Specifically, our interest in Cu2O arises from its good perfor-mance in PEC and photovoltaic applications. As pointed out inthe introduction, the synthesis of Cu2O particles with various mor-phologies can be achieved using a number of wet-chemistry syn-thesis protocols. After screening several selected methods[18,21,23] for convenience and reproducibility of the results, theprotocol developed by Huang and co-workers [23] has been chosenfor our experiments. This protocol is reproducible, convenient, andeasy to modify. Additionally, the protocol is the synthesis of choiceto obtain Cu2O particles with various morphologies, such as cubic,octahedral, and rhombic dodecahedral. These particles are synthe-sized under the control of SDS via a simple reduction of Cu(OH)2 byNH2OH"HCl at 34 "C. Besides the regulatory effect of the SDS sur-factant on the product morphology, the use of SDS also renders dis-persibility of the as-synthesized particles in water or ethanol. The

Fig. 6. Normalized optical absorption spectra for three sizes of submicron Cu2Ocubes dispersed in ethanol.

Fig. 7. Current-potential characteristics of photocathode made from 0.65-lm Cu2Ocubes. PEC performance recorded under chopped-light illumination at pH of 1 and4. The inset shows the respective Cu2O electrode film on FTO obtained by the doctorblade technique.

E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227 225

colloidal stability of the resultant dispersions is, however, not sig-nificant, since the particles are large in size and tend to sedimentwithin a short period of time (

[3] M. Hara, T. Kondo, M. Komoda, S. Ikeda, K. Shinohara, A. Tanaka, J.N. Kondo, K.Domen, Chem. Commun. (1998) 357–358.

[4] A. Paracchino, V. Laporte, K. Sivula, M. Grätzel, E. Thimsen, Nat. Mater. 10(2011) 456–461.

[5] R.J. Elliott, Phys. Rev. 108 (1957) 1384–1389.[6] T. Gershon, K.P. Musselman, A. Marin, R.H. Friend, J.L. MacManus-Driscoll, Sol.

Energ. Mat. Sol. C 96 (2012) 148–154.[7] J.P. Hu, D.J. Payne, R.G. Egdell, P.A. Glans, T. Learmonth, K.E. Smith, J. Guo, N.M.

Harrison, Phys. Rev. B 77 (2008) 155115.[8] L.C. Olsen, F.W. Addis, W. Miller, Sol. Cells 7 (1982) 247–279.[9] C. Malerba, F. Biccari, C.L.A. Ricardo, M. D’Incau, P. Scardi, A. Mittiga, Sol. Energ.

Mat. Sol. C 95 (2011) 2848–2854.[10] L.I. Hung, C.K. Tsung, W.Y. Huang, P.D. Yang, Adv. Mater. 22 (2010) 1910–1914.[11] H.J. Yang, S.Y. He, H.L. Chen, H.Y. Tuan, Chem. Mater. 26 (2014) 1785–1793.[12] M. Yin, C.K. Wu, Y.B. Lou, C. Burda, J.T. Koberstein, Y.M. Zhu, S. O’Brien, J. Am.

Chem. Soc. 127 (2005) 9506–9511.[13] Y. Chang, J.J. Teo, H.C. Zeng, Langmuir 21 (2005) 1074–1079.[14] X.A. Lan, J.Y. Zhang, H. Gao, T.M. Wang, CrystEngComm 13 (2011) 633–636.[15] W. Chen, L.L. Li, Q. Peng, Y.D. Li, Nano Res. 5 (2012) 320–326.[16] M.H. Kim, B. Lim, E.P. Lee, Y.N. Xia, J. Mater. Chem. 18 (2008) 4069–4073.[17] M.D. Susman, Y. Feldman, A. Vaskevich, I. Rubinstein, ACS Nano 8 (2014) 162–

174.[18] M.L. Pang, H.C. Zeng, Langmuir 26 (2010) 5963–5970.[19] H.L. Xu, W.Z. Wang, W. Zhu, J. Phys. Chem. B 110 (2006) 13829–13834.[20] Y. Yu, L.Y. Zhang, J. Wang, Z. Yang, M.C. Long, N.T. Hu, Y.F. Zhang, Nanoscale

Res. Lett. 7 (2012).[21] J.T. Zhang, J.F. Liu, Q. Peng, X. Wang, Y.D. Li, Chem. Mater. 18 (2006) 867–871.[22] C.C. Yec, H.C. Zeng, Chem. Mater. 24 (2012) 1917–1929.[23] W.C. Huang, L.M. Lyu, Y.C. Yang, M.H. Huang, J. Am. Chem. Soc. 134 (2012)

1261–1267.[24] C.N.R. Rao, H.S.S.R. Matte, R. Voggu, A. Govindaraj, Dalton Trans. 41 (2012)

5089–5120.[25] C.J. Tighe, R.Q. Cabrera, R.I. Gruar, J.A. Darr, Ind. Eng. Chem. Res. 52 (2013)

5522–5528.[26] J.C. Park, J. Kim, H. Kwon, H. Song, Adv. Mater. 21 (2009) 803–807.[27] Y.V. Kolen’ko, M. Bañobre-López, C. Rodríguez-Abreu, E. Carbó-Argibay, A.

Sailsman, Y. Piñeiro-Redondo, M.F. Cerqueira, D.Y. Petrovykh, K. Kovnir, O.I.Lebedev, J. Rivas, J. Phys. Chem. C 118 (2014) 8691–8701.

[28] M.P. Seah, L.S. Gilmore, G. Beamson, Surf. Interface Anal. 26 (1998) 642–649.[29] M. Balkansk, M.A. Nusimovi, J. Reydelle, Solid State Commun. 7 (1969) 815–

818.[30] C. Carabato, Phys. Status Solidi 37 (1970) 773–779.

[31] C. Carabato, B. Prevot, Phys. Status Solidi B 44 (1971) 701–712.[32] R.J. Elliott, Phys. Rev. 124 (1961) 340–345.[33] K. Huang, Z. Phys. 171 (1963) 213–225.[34] C. Herring, Phys. Rev. 82 (1951) 87–93.[35] S. Alvarez, Dalton Trans. (2005) 2209–2233.[36] J.M. Essick, R.T. Mather, Am. J. Phys. 61 (1993) 646–649.[37] A. Paracchino, N. Mathews, T. Hisatomi, M. Stefik, S.D. Tilley, M. Gratzel, Energ.

Environ. Sci. 5 (2012) 8673–8681.[38] M. Meng, X.L. Wu, X.B. Zhu, L. Yang, Z.X. Gan, X.S. Zhu, L.Z. Liu, P.K. Chu, J. Phys.

Chem. Lett. 5 (2014) 4298–4304.[39] M. Sun, S.J. Xiong, X.L. Wu, C.Y. He, T.H. Li, P.K. Chu, Adv. Mater. 25 (2013)

2035–2039.[40] J. Chen, B. Sarma, J.M.B. Evans, A.S. Myerson, Cryst. Growth Des. 11 (2011)

887–895.[41] H.R. Zhang, C.M. Shen, S.T. Chen, Z.C. Xu, F.S. Liu, J.Q. Li, H.J. Gao,

Nanotechnology 16 (2005) 267–272.[42] M.J. Siegfried, K.S. Choi, J. Am. Chem. Soc. 128 (2006) 10356–10357.[43] Y.S. Luo, S.Q. Li, Q.F. Ren, J.P. Liu, L.L. Xing, Y. Wang, Y. Yu, Z.J. Jia, J.L. Li, Cryst.

Growth Des. 7 (2007) 87–92.[44] C.H. Kuo, M.H. Huang, J. Phys. Chem. C 112 (2008) 18355–18360.[45] C.H. Kuo, T.E. Hua, M.H. Huang, J. Am. Chem. Soc. 131 (2009) 17871–17878.[46] X.D. Liang, L. Gao, S.W. Yang, J. Sun, Adv. Mater. 21 (2009) 2068–2071.[47] D.F. Zhang, H. Zhang, L. Guo, K. Zheng, X.D. Han, Z. Zhang, J. Mater. Chem. 19

(2009) 5220–5225.[48] Y.B. Cao, J.M. Fan, L.Y. Bai, F.L. Yuan, Y.F. Chen, Cryst. Growth Des. 10 (2010)

232–236.[49] S.D. Sun, C.C. Kong, S.C. Yang, L.Q. Wang, X.P. Song, B.J. Ding, Z.M. Yang,

CrystEngComm 13 (2011) 2217–2221.[50] K.F. Chen, D.F. Xue, CrystEngComm 14 (2012) 8068–8075.[51] P.G. Vekilov, Cryst. Growth Des. 10 (2010) 5007–5019.[52] N.C. Christov, N.D. Denkov, P.A. Kralchevsky, K.P. Ananthapadmanabhan, A.

Lips, Langmuir 20 (2004) 565–571.[53] Y.V. Kolen’ko, V.D. Maximov, A.V. Garshev, P.E. Meskin, N.N. Oleynikov, B.R.

Churagulov, Chem. Phys. Lett. 388 (2004) 411–415.[54] S.G. Kwon, T. Hyeon, Small 7 (2011) 2685–2702.[55] V.K. LaMer, Ind. Eng. Chem. 44 (1952) 1270–1277.[56] V.K. LaMer, R.H. Dinegar, J. Am. Chem. Soc. 72 (1950) 4847–4854.[57] N.T.K. Thanh, N. Maclean, S. Mahiddine, Chem. Rev. 114 (2014) 7610–7630.[58] A. Ruditskiy, S.I. Choi, H.C. Peng, Y.N. Xia, MRS Bull. 39 (2014) 727–737.[59] X.F. Yang, J.X. Fu, C.J. Jin, J.A. Chen, C.L. Liang, M.M. Wu, W.Z. Zhou, J. Am. Chem.

Soc. 132 (2010) 14279–14287.

E. Carbó-Argibay et al. / Journal of Colloid and Interface Science 456 (2015) 219–227 227

http://refhub.elsevier.com/S0021-9797(15)00552-4/h0015http://refhub.elsevier.com/S0021-9797(15)00552-4/h0015http://refhub.elsevier.com/S0021-9797(15)00552-4/h0020http://refhub.elsevier.com/S0021-9797(15)00552-4/h0020http://refhub.elsevier.com/S0021-9797(15)00552-4/h0025http://refhub.elsevier.com/S0021-9797(15)00552-4/h0030http://refhub.elsevier.com/S0021-9797(15)00552-4/h0030http://refhub.elsevier.com/S0021-9797(15)00552-4/h0035http://refhub.elsevier.com/S0021-9797(15)00552-4/h0035http://refhub.elsevier.com/S0021-9797(15)00552-4/h0040http://refhub.elsevier.com/S0021-9797(15)00552-4/h0045http://refhub.elsevier.com/S0021-9797(15)00552-4/h0045http://refhub.elsevier.com/S0021-9797(15)00552-4/h0050http://refhub.elsevier.com/S0021-9797(15)00552-4/h0055http://refhub.elsevier.com/S0021-9797(15)00552-4/h0060http://refhub.elsevier.com/S0021-9797(15)00552-4/h0060http://refhub.elsevier.com/S0021-9797(15)00552-4/h0065http://refhub.elsevier.com/S0021-9797(15)00552-4/h0070http://refhub.elsevier.com/S0021-9797(15)00552-4/h0075http://refhub.elsevier.com/S0021-9797(15)00552-4/h0080http://refhub.elsevier.com/S0021-9797(15)00552-4/h0085http://refhub.elsevier.com/S0021-9797(15)00552-4/h0085http://refhub.elsevier.com/S0021-9797(15)00552-4/h0090http://refhub.elsevier.com/S0021-9797(15)00552-4/h0095http://refhub.elsevier.com/S0021-9797(15)00552-4/h0100http://refhub.elsevier.com/S0021-9797(15)00552-4/h0100http://refhub.elsevier.com/S0021-9797(15)00552-4/h0105http://refhub.elsevier.com/S0021-9797(15)00552-4/h0110http://refhub.elsevier.com/S0021-9797(15)00552-4/h0115http://refhub.elsevier.com/S0021-9797(15)00552-4/h0115http://refhub.elsevier.com/S0021-9797(15)00552-4/h0120http://refhub.elsevier.com/S0021-9797(15)00552-4/h0120http://refhub.elsevier.com/S0021-9797(15)00552-4/h0125http://refhub.elsevier.com/S0021-9797(15)00552-4/h0125http://refhub.elsevier.com/S0021-9797(15)00552-4/h0130http://refhub.elsevier.com/S0021-9797(15)00552-4/h0135http://refhub.elsevier.com/S0021-9797(15)00552-4/h0135http://refhub.elsevier.com/S0021-9797(15)00552-4/h0135http://refhub.elsevier.com/S0021-9797(15)00552-4/h0140http://refhub.elsevier.com/S0021-9797(15)00552-4/h0145http://refhub.elsevier.com/S0021-9797(15)00552-4/h0145http://refhub.elsevier.com/S0021-9797(15)00552-4/h0150http://refhub.elsevier.com/S0021-9797(15)00552-4/h0155http://refhub.elsevier.com/S0021-9797(15)00552-4/h0160http://refhub.elsevier.com/S0021-9797(15)00552-4/h0165http://refhub.elsevier.com/S0021-9797(15)00552-4/h0170http://refhub.elsevier.com/S0021-9797(15)00552-4/h0175http://refhub.elsevier.com/S0021-9797(15)00552-4/h0180http://refhub.elsevier.com/S0021-9797(15)00552-4/h0185http://refhub.elsevier.com/S0021-9797(15)00552-4/h0185http://refhub.elsevier.com/S0021-9797(15)00552-4/h0190http://refhub.elsevier.com/S0021-9797(15)00552-4/h0190http://refhub.elsevier.com/S0021-9797(15)00552-4/h0195http://refhub.elsevier.com/S0021-9797(15)00552-4/h0195http://refhub.elsevier.com/S0021-9797(15)00552-4/h0200http://refhub.elsevier.com/S0021-9797(15)00552-4/h0200http://refhub.elsevier.com/S0021-9797(15)00552-4/h0205http://refhub.elsevier.com/S0021-9797(15)00552-4/h0205http://refhub.elsevier.com/S0021-9797(15)00552-4/h0210http://refhub.elsevier.com/S0021-9797(15)00552-4/h0215http://refhub.elsevier.com/S0021-9797(15)00552-4/h0215http://refhub.elsevier.com/S0021-9797(15)00552-4/h0220http://refhub.elsevier.com/S0021-9797(15)00552-4/h0225http://refhub.elsevier.com/S0021-9797(15)00552-4/h0230http://refhub.elsevier.com/S0021-9797(15)00552-4/h0235http://refhub.elsevier.com/S0021-9797(15)00552-4/h0235http://refhub.elsevier.com/S0021-9797(15)00552-4/h0240http://refhub.elsevier.com/S0021-9797(15)00552-4/h0240http://refhub.elsevier.com/S0021-9797(15)00552-4/h0245http://refhub.elsevier.com/S0021-9797(15)00552-4/h0245http://refhub.elsevier.com/S0021-9797(15)00552-4/h0250http://refhub.elsevier.com/S0021-9797(15)00552-4/h0255http://refhub.elsevier.com/S0021-9797(15)00552-4/h0260http://refhub.elsevier.com/S0021-9797(15)00552-4/h0260http://refhub.elsevier.com/S0021-9797(15)00552-4/h0265http://refhub.elsevier.com/S0021-9797(15)00552-4/h0265http://refhub.elsevier.com/S0021-9797(15)00552-4/h0270http://refhub.elsevier.com/S0021-9797(15)00552-4/h0275http://refhub.elsevier.com/S0021-9797(15)00552-4/h0280http://refhub.elsevier.com/S0021-9797(15)00552-4/h0285http://refhub.elsevier.com/S0021-9797(15)00552-4/h0290http://refhub.elsevier.com/S0021-9797(15)00552-4/h0295http://refhub.elsevier.com/S0021-9797(15)00552-4/h0295

SUPPLEMENTARY MATERIAL FOR:

Up-scaling the Synthesis of Cu2O Submicron Particles with Controlled Morphologies for Solar H2 Evolution from Water Enrique Carbó-Argibay,1 Xiao-Qing Bao,1 Carlos Rodríguez-Abreu,1 M. Fátima Cerqueira,2 Dmitri Y. Petrovykh,1 Lifeng Liu,1 and Yury V. Kolen’ko*,1 1International Iberian Nanotechnology Laboratory, Braga 4715-330, Portugal 2Center of Physics, University of Minho, Braga 4710-057, Portugal *yury.kolenko@inl.int

This SM file includes a detailed description of characterization methods, Raman spectroscopy results, synthesis results, and Figs. S1 to S9.

CHARACTERIZATION

Scanning electron microscopy and energy dispersive X-ray spectroscopy Scanning electron microscopy (SEM) investigation was performed using Quanta 650 FEG microscope (FEI Company), equipped with INCA 350 spectrometer (Oxford Instruments) for energy dispersive X-ray spectroscopy (EDX). The samples for SEM were prepared by dropping diluted suspension of Cu2O in ethanol onto Si wafer (1 x 1 cm2) followed by the evaporation of the solvent.

Powder X-ray diffraction Powder X-ray diffraction (XRD) data were collected on a X’Pert PRO diffractometer (PANalytical) set at 45 kV and 40 mA, and equipped with Cu Kα radiation (λ=1.541874 Å) and a PIXcel detec-tor. Data were collected using Bragg-Brentano geometry in the 15 to 80° 2θ range with a scan speed of 0.006 °/s. Ge was used as an in-ternal standard. The XRD patterns were matched to International Centre for Diffraction Data (ICDD) PDF-4+ database using HighScore software package (PANalytical).

Raman spectroscopy Raman spectroscopy measurements were carried out on alpha300 R confocal Raman microscope (WITec) using a 532 nm Nd:YAG laser for excitation. The system was operated with an output laser power of around 100 µW to avoid sample degradation due to laser-induced heating. The laser beam was focused on the powder by a ×50 lens (Zeiss); and the spectra were collected with an 1800 groove/mm grating using 40 acquisitions with a 2 s acquisition time.

RAMAN SPECTROSCOPY ANALYSIS Cuprite Cu2O crystallizes in a cubic structure of space group !!! (Pn-3m) [Ref. 29−33, MS]. Its unit cell contains two Cu2O units, i.e., six atoms, yielding 15 optical phonon modes in addition to the three acoustic lattice vibrations. The symmetries of the vibrational modes

at k = 0 are A2u ⊕ Eu ⊕ T2g ⊕ 3T1u ⊕ T2u. Phonons with A, E, and T symmetry are one-fold, two-fold and three-fold degenerate, respec-tively. The three acoustic phonons have T1u symmetry. The two re-maining modes of T1u symmetry are infrared active optical lattice vibrations. Phonons with A2u, Eu, and T2u symmetry are silent modes, neither Raman nor infrared active. The only Raman active lattice vibrations in Cu2O belong to the three-fold-degenerate T2g symmetry. Accord-ing to the group theoretical analysis, the Raman spectrum of a per-fect Cu2O crystal should exhibit only one phonon Raman signal, that of the three-fold-degenerate T2g mode. However, a typical Raman spectrum of cuprous oxide, independently of the synthesis method used, is much richer with a multitude of Raman signals that have been assigned to different one-phonon scattering processes in addi-tion to a background due to two-phonon scattering. This fact is ex-plained interpreting the spectra in terms of bi-phonons or multi-phonons processes in light absorption. Fig. S3 shows the Raman spectra of the two selected samples, name-ly, 0.42−µm Cu2O cubes (Fig. 3c, MS) and strongly HCl-etched 1.42−µm Cu2O truncated octahedrons (Fig. 4d, MS). The collected Raman spectra of both samples are different, mainly in the 80−160 cm−1 range, which corresponds to the range of T2u and Eu symmetry modes. The Raman spectrum of 0.42−µm Cu2O cubes and the corresponding fit is presented in Fig. S4a. It is seen that this sample presents the “characteristic” phonon frequencies of the crystalline Cu2O. The 219 cm−1 peak (the strongest one) is associated with the second-order Raman mode of Cu2O crystals (2Eu). The peak at 146 cm−1 most likely attributed to Raman scattering from phonons of symmetry T1u(TO). The mode at 410 cm−1 is assigned to four-phonon mode (3Eu + T1u(TO)). The 630 cm−1 mode is also related with Cu2O crys-tals, and is assigned to infrared-allowed mode (T1u(TO). The mode present at 100 cm−1 is assigned to the inactive Raman mode (Eu) in crystals, while the 192 cm−1 mode is assigned to the T1u mode. We were not able to identify the peak observed at 486 cm−1. The Raman spectrum and correspondent fit of the etched 1.42−µm Cu2O truncated octahedrons is shown in Fig. S4b. As already men-

S2

tioned the Raman spectrum of this sample is different in the lower wavenumber range as compared to the cubes. At the higher wave-number range, besides the aforementioned modes observed for the Cu2O cubes, the sample shows one more Cu2O mode at 515 cm−1, which corresponds to the T2g symmetry mode (the only Raman mode allowed, according with the group theory). In the lower wavenumber ranges, all the modes seen in Cu2O cubes are present also in etched Cu2O truncated octahedrons, although broadened, indicating lower crystallinity as compared to the cubes. Moreover, the relative intensity between the modes is strongly dif-ferent (Fig. S3), namely, the intensity ratio between the 219 and the 100 cm−1 modes is 3.8 for cubes and 0.5 for octahedrons. Taking into account that the 100 cm−1 mode is associated with breathing mode of the Cu tetrahedron and that our SEM study evidenced strong etching of the octahedrons by HCl, it is reasonable to deduced that the observed inversion of intensity is most likely a result of Cu deficiency in the etched octahedrons.

X-ray photoelectron spectroscopy The surface chemistry of Cu2O particles was probed by X-ray photo-electron spectroscopy (XPS) using an ESCALAB 250 Xi system (Thermo Scientific). The specimens of Cu2O were prepared for XPS analysis by drop casting the colloids onto Si wafer with subsequent drying in vacuum at room temperature. ESCALAB 250 Xi is equipped with a monochoromated Al Kα X-ray source, a hemispherical electron energy analyser, an automated sam-ple stage, and a video camera for viewing the analysis position. The analysis spot of about 650×400 µm2 was defined by the micro-focused X-ray source. The energy of the monochromated Al Kα X-ray source was measured to be within < 0.2 eV from 1486.6 eV. The binding energy (BE) scale of the analyser was calibrated to produce < 50 meV deviations of the three standard peaks from their standard values: 83.98 eV for Au 4f7/2, 368.26 eV for Ag 3d5/2, and 932.67 eV for Cu 2p3/2 [Ref. 28, MS]. Spatially uniform charge neutralization was provided by beams of low-energy (≤ 10 eV) electrons guided by a magnetic lens and by Ar+ ions. The aliphatic C 1s peak appeared at 284.9–285.0 eV under these measurement conditions. The measurements were performed at room temperature in an ultra-high vacuum chamber with the base pressure < 5×10-10 mbar; the charge neutralization device produced ca. 2×10-7 mbar partial pressure of Ar during measurements. High-resolution elemental XPS data in C 1s, O 1s, Cu 2p, and Cu LMM regions were acquired with the analyser pass energy set to 20 eV (corresponding to energy resolution of about 0.36 eV) and the step size set to 0.1 eV. All the spectra were acquired in normal emis-sion with an effective analyser collection angle of ca. 30°.

UV−Vis−NIR spectroscopy Room temperature UV−Vis−NIR spectra were collected in the 250-2600 nm range using a Lambda 950 spectrophotometer (Perkin Elmer). The measurements were performed on diluted suspension of Cu2O in ethanol using 1 mm precision cell made of quartz (Hellma Analytics).

EFFECT OF SYNTHESIS CONDITIONS As expected, the presence of NH2OH⋅HCl reducing agent is a critical factor for the preparation of Cu2O particles. Continuous spertiniite Cu(OH)2 (Fig. S5) nanowires with uniform width of around 35 nm and a length in the range of a few µm, instead of Cu2O submicron cubes, were generated in 100-fold synthesis, wherein the NH2OH⋅HCl was absent (Fig. S6a). Further, we probed the influ-ence of SDS on the Cu2O formation. Our attempt mimicking the 100-fold synthesis but without use of SDS does not delivered the desired Cu2O. The as-prepared product was found to be a phase mix-ture of Cu(OH)2, Cu2O, and minor CuO (Figs. S6b and S7).

The agitation during the final aging step of the synthesis has been found to be also important for the Cu2O formation. When we contin-ued the stirring at 300 rpm during 1 h aging at 34 °C, the Cu2O sub-micron cubes did not form. As shown in Fig. S6c, the size and shape of the resultant crystals are very similar to those of Cu(OH)2 seen in Figs. S6a and S6b, which were obtained in the absence of NH2OH⋅HCl or SDS, respectively. Seemingly, the absence of SDS or agitation during aging step some-what prevents the reduction of the Cu(OH)2 to Cu2O. This is clearly confirmed by XRD, which shows that the samples are mostly consist of Cu(OH)2 with admixture of Cu2O and CuO or anatacamite Cu2Cl(OH)3 in the case of SDS absence or continuous agitation, respectively (Figs. S7-S8).

FIGURES

Fig. S1. Representative SEM image and the corresponding EDX elemental and chemical mappings of 0.65−µm Cu2O submicron cubes, revealing the homogeneous distribution of Cu and O within the particles (scale bar: 3 µm).

Fig. S2. Typical XRD patterns from Cu2O submicron particles ob-tained according to a reported method of Ref. 25 in the MS (green line), as well as via 10-fold (blue line) and 100-fold (red line) up-scaling of the method. Tick marks below the pattern correspond to the positions of the Bragg reflections expected for the cuprite (ICDD no. 01-080-7711, cubic, Pn-3m). The primary (11-1) reflection of the surface CuO admixture at 2θ = 35.54 is denoted by * (ICDD no. 04-004-5425, monoclinic, C2/c).

S3

Fig. S3. Room temperature Raman spectroscopy data for the select-ed 0.42−µm Cu2O cubes (Fig. 3c, MS) and HCl-etched 1.42−µm Cu2O truncated octahedrons (Fig. 4d, MS).

Fig. S4. Lorentzian fit using seven Lorentzians of Raman data for selected 0.65−µm Cu2O cubes (a) and HCl-etched 1.42−µm Cu2O truncated octahedrons (b). The insets are enlargements of the Raman shift region from 80 to 300 cm−1, which are provided for clarification of the fitting.

Fig. S5. XRD pattern (red line) from the product (Fig. S6a) synthe-sized in the absence of NH2OH⋅HCl reducing agent. Tick marks below the pattern correspond to the positions of the Bragg reflections expected for the spertiniite Cu(OH)2 (ICDD no. 01-072-0140, orthorhombic, Cmcm).

Fig. S6. Representative low- (left panels) and high- (right panels) magnification SEM images of the prepared products showing the effects of the different synthesis parameters on the microstructure: (a) absence of NH2OH⋅HCl reducing agent; (b) absence of the SDS; (c) stirring at 300 rpm during the final aging for 1 h.

S4

Fig. S7. XRD pattern (red line) from the product (Fig. S6b) synthe-sized in the absence of SDS. Blue, green and grey tick marks below the pattern correspond to the positions of the Bragg reflections ex-pected for the cuprite Cu2O (ICDD no. 01-080-7711, cubic, Pn-3m), the spertiniite Cu(OH)2 (ICDD no. 00-013-0420, orthorhombic, Cmcm) and tenorite CuO (ICDD no. 01-089-2531, monoclinic, C2/c), respectively.

Fig. S8. XRD pattern (red line) from the product (Fig. S6c) synthe-sized using stirring at 300 rpm during the aging for 1 h. Blue and green tick marks below the pattern correspond to the positions of the Bragg reflections expected for the spertiniite Cu(OH)2 (ICDD no. 00-013-0420, orthorhombic, Cmcm) and anatacamite Cu2Cl(OH)3 (ICDD no. 04-016-8161, anorthic, P-1), respectively.

Fig. S9. UV−Vis spectrum of bare Cu(OH)2 (confirmed by XRD) precursor admixture extracted during multiple washings of the re-sultant Cu2O product with water/ethanol (1:1).

Up-scaling the synthesis of Cu2O submicron particles with controlled morphologies for solar H2 evolution from water1 Introduction2 Materials and methods2.1 Starting materials2.2 Synthesis2.3 Washing2.4 Characterization2.5 Film fabrication2.6 PEC measurements

3 Results3.1 Structure and composition3.2 Size and morphology3.2.1 Effect of up-scaling3.2.2 Effect of synthesis conditions3.2.3 Amount of reducing agent

3.3 Optical properties3.4 PEC H2 evolution properties

4 Discussion4.1 Synthetic aspects of up-scaling4.2 Mechanism of formation

5 ConclusionsAcknowledgmentsAppendix A Supplementary materialReferences