This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Materials Chemistry and Physics 122 (2010) 602–608

Contents lists available at ScienceDirect

Materials Chemistry and Physics

journa l homepage: www.e lsev ier .com/ locate /matchemphys

Synthesis, characterisation and stability of Cu2O nanoparticles produced viareverse micelles microemulsion

D. Dodoo-Arhina, M. Leonia,∗, P. Scardia, E. Garnierb, A. Mittigac

a University of Trento, Department of Materials Engineering and Industrial Technologies, via Mesiano 77, 38100 Trento, Italyb Laboratoire de Catalyse en Chimie Organique, UMR CNRS 6503, Université de Poitiers, 40 Av. du Recteur Pineau, 86022 Poitiers, Francec ENEA C.R. Casaccia, via Anguillarese 301, 00123 Roma, Italy

a r t i c l e i n f o

Article history:Received 27 August 2009Received in revised form 3 February 2010Accepted 17 March 2010

Keywords:Powder diffractionChemical synthesisSemiconductors

a b s t r a c t

Cuprite (Cu2O) nanoparticles were synthesized at room temperature via reduction of CuCl2·2H2O byNaBH4 in water/n-heptane microemulsion stabilised by the non-ionic Brij30 surfactant. Whole PowderPattern Modelling of the X-ray diffraction patterns shows the presence of a bimodal size distribution inthe nanopowders, with a fraction of domains in the 10–40 nm range and a smaller one below 10 nm.Linear and planar defects are absent.

A relationship between the average size of the larger particles and the quantity of water in the systemwas obtained. The stability of cuprite under visible light irradiation both during the synthesis and afterthe preparation was investigated, showing that a self-catalytic conversion of Cu2O into CuO takes placein water.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Nanoparticles of metals, semiconductors and oxides keepattracting the attention of the scientific community because of theirexceptional and in some cases, unexpected physical and chemicalproperties coming from the quantum confinement at the nano-scale.

Synthesis of inorganic nanoparticles still remains a challengingtask owing to intrinsic difficulties in the control of compositionand morphology [1–3]. Among the proposed synthesis routes,water-in-oil (W/O) microemulsions are promising as they providenanoreactors where size and morphology of nanoparticles can bewell controlled [4–8].

Copper oxides are useful reference systems for the study ofcomplex cuprates, most of which show high-Tc superconductiv-ity. Of particular interest is cuprous oxide (Cu2O cuprite), a p-typesemiconductor due to the presence of Cu vacancies which form anacceptor level 0.4 eV above the valence band [9].

As such, Cu2O is attracting the current interest owing to the widerange of potential applications. For instance, cuprite is a promisingsolar cell material (band gap of 2.0–2.2 eV) [10], but it can also beused as anode material for lithium ion batteries [11], as a photocat-alyst for water splitting under visible light irradiation [12,13] andas a sensing material in gas detectors [9].

∗ Corresponding author. Tel.: +39 0461282416; fax: +39 0461281977.E-mail address: Matteo.Leoni@unitn.it (M. Leoni).

For a wider application of the material, however, it would benecessary to have powders characterised by a nanometric size andfree of lattice defects, often required to maximise the efficiency ofthe corresponding devices. Unfortunately, it is not easy to obtainnanopowders of controlled size and shape, with an intermediatevalence (Cu+ vs. Cu0 or Cu2+) and free of defects. Cu2O particleswith a controlled shape (e.g. cuboids [14], octahedra [15], or thick-shell hollow spheres [16]) have been synthesized, but they are fartoo large (few hundred nm) and imperfect to be useful. Yin etal. [17] were able to produce 5 nm cuprite nanoparticles (coatedwith a thin CuO layer), but their process leads to the presenceof stacking defects. More recently, 10–15 nm nanoparticles wereproduced on multiwall carbon nanotubes [18], but in this casethe particles are produced at high temperature and come alreadywith a support, that in most cases is not necessary. The synthesisof cuprite nanoparticles in microemulsions seems promising, butagain a large quantity of defects seems always being present [19].

Furthermore, the analysis of the size of those nanoparticlesas well as their structure is, in some cases, quite naive andcan lead to severe errors (see e.g. [20] for the description of acommon mistake): most authors use simplified diffraction-basedtechniques (e.g. the Scherrer formula [21]) or transmission electronmicroscopy for the size analysis. Quoting an average “crystallitesize” or obtaining a size distribution from the analysis of a fewdozen well visible grains has a limited statistical validity. Simpleyet advanced techniques nowadays exists for a complete structuraland microstructural characterisation of nanocrystalline powders,based on the analysis of the whole X-ray diffraction (XRD) pattern,either by using the Whole Powder Pattern Modelling (WPPM) tech-

0254-0584/$ – see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.matchemphys.2010.03.053

Author's personal copy

D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608 603

Table 1Experimental parameters for the SA and SB sets.

Sample Brij30 (g) n-Heptane (g) 0.2 M CuCl2 (ml) NaBH4 (mg) H2O (%) ω

SA1 6.287 22.430 0.4 31.0 1.0 1.3SA2 6.289 22.158 0.8 61.0 2.0 2.6SA3 6.285 21.891 1.2 92.0 3.0 3.8SA4 6.286 21.613 1.6 122.0 4.0 5.1SA5 6.287 21.345 2.0 152.0 5.0 6.4

SB1 6.284 22.434 0.4 31.0 1.0 1.3SB2 6.286 22.153 0.8 61.0 2.0 2.6SB3 6.286 21.886 1.2 91.0 3.0 3.8SB4 6.284 21.614 1.6 122.0 4.0 5.1SB5 6.286 21.332 2.0 152.0 5.0 6.4

nique [22], or by the Debye equation [23]. In WPPM, microstructuralparameters such as domain size, shape and distribution, as wellas type and quantity of linear and planar defects, are employedto build a computer-generated diffraction pattern of the materialunder study. The parameters are then varied through a non-linearleast squares routine until the best fit is reached between modeland experimental data. This guarantees a self-consistent extrac-tion of microstructural information from the measured XRD pattern[22,24].

In this paper we report preliminary results on room-temperature synthesis and characterisation of lattice defect-free Cu2O nanoparticles. Powders produced via water-in-oilmicroemulsion and analysed both from the morphological, struc-tural, microstructural and optical point of view, show a clearphotoactivity.

2. Experimental

2.1. Materials and specimens

All experiments were conducted at room temperature (24 ± 2 ◦C) in a poorly illu-minated environment: just in one case (cf. Table 1), the preparation was conductedunder light, to test for photocatalytic effects.

Two sets of five batches (40 ml each) of microemulsion were prepared (identifiedas SA and SB, plus the batch number). The two batches differ in the quantity of water(see below and cf. Table 1), i.e. in the parameter ω = nH2O/nBrij30, defined as the ratiobetween the number of molecules of water and surfactant in the system.

A stable inverse-micelle microemulsion was obtained by mixing oil and surfac-tant, and subsequently adding water. The oil-surfactant dispersion was created bymixing n-heptane (Sigma–Aldrich, 99% purity) with the non-ionic surfactant Brij30®

(polyoxyethylene 4 lauryl ether (C2H4O)nC12H26O, n ≈ 4, mean weight M = 362.6,Sigma–Aldrich) in a 80.46/16.54 vol/vol ratio. The two components were weightedand mixed in a 200 ml high-density polyethylene graduated bottle, closed with apolypropylene cap (to limit the volatilization of the hydrocarbon) after each mixingphase of the synthesis. The dispersion was sonicated for 3 min to favour the mixingof the two phases. Sonication was always done in a thermostatic bath (25 ◦C) at afrequency of 59 kHz and 125 W.

A 0.2 M solution of CuCl2·2H2O (Sigma–Aldrich, 6174-250GF) was formed byadding the salt to deionised water (<1.8 �S cm−1). Variable aliquots (1–5 vol.%,according to Table 1) of this solution were added to the dispersion and sonicated for6 min in order to create a homogeneous microemulsion. At the end of the processa homogeneous and transparent sky-blue microemulsion free from any precipitatewas obtained.

Sodium borohydride (Sigma–Aldrich, 99% purity, 213462-25G) was used asreducing agent: a 1 wt% (0.400 mg) was added to the emulsion and sonicated for1s to start the reaction. To limit unwanted products (e.g. CuO), the microemulsionwas continuously stirred for 10 min. The addition of the reducing agent turned theemulsion first to yellow and then to deep brown, with evident evolution of gas andabsence of any precipitation.

To break the microemulsion, 20 ml of acetone (purity 99%, Sigma–Aldrich) wereadded. Breaking the micelles causes the nanoparticles to slowly precipitate, with apowder clearly becoming visible on the bottom of the bottle after 10 min. To removethe organic phase residuals, 40 ml of a 1:1 mixture of acetone and ethanol (purity98%, Sigma–Aldrich) was added to the gelatinous precipitate, mildly sonicated andcentrifuged at 6000 rpm for 10 min. Subsequently 20 ml ethanol and 20 ml deionisedwater were added to the dispersion that was sonicated and centrifuged again at6000 rpm for 10 min.

The particles were further washed several times with deionised water. Whenneeded for the analysis, the particles were laid on an h00 silicon wafer and driedunder Argon flux.

A commercial Cu2O powder (Sigma–Aldrich, 99% purity), named SC, wasemployed as a reference for the analyses. The SC powder was heat treated at 850 ◦Cand 9.5 mbar for 40 min to remove any residual Cu.

2.2. Characterisation techniques

X-ray powder diffraction (XRD) patterns were collected on a Rigaku PMG-VHBragg-Brentano diffractometer operating a copper tube at 40 kV and 30 mA. Thegoniometer is equipped with a high resolution set up (1◦ divergence slit, 2◦ inci-dent and diffracted beam Soller slits, 0.15 mm receiving slit) and a curved-crystalgraphite analyser, providing a narrow and symmetrical instrumental profile overthe investigated angular range.

The instrumental resolution function was characterised with the NIST SRM 660a(LaB6) standard [25]: all peak profiles were simultaneously fitted with symmetricalpseudo-Voigt functions whose width and shape were constrained according to theCaglioti et al. formulae [26]. The XRD patterns of all specimens were recorded in the25–85◦ 2� range with a step size of 0.05◦ and a counting time of 60 s per step.

Microstructural analysis was performed using the Whole Powder Pattern Mod-elling (WPPM) method [22], using the PM2K software [27].

A Jeol JSM 5500 LV microscope operated at 20 kV was employed for the mor-phological characterisation: prior to investigation, specimens were sputtered withAu/Pd.

Spectrophotometry was performed on an UV–vis-NIR V570 Jasco (Japan Spec-troscopic Co. Ltd., Tokyo, Japan) spectrophotometer in air in the 200–1200 nm range.Transmission FTIR spectra were recorded on an Avatar 550 (Thermo Optics) instru-ment in the 4000–400 cm−1 range with 2 cm−1 resolution; the Spectrum v5.3.0software was employed for the FTIR analysis.

3. Results and discussion

3.1. Synthesis

The redox potential for the Cu2+/Cu+ couple(Cu2+(aq) + e−(aq) → Cu+(aq)) is + 0.153 V, i.e. less than half ofthe +0.34 V needed to fully reduce Cu2+ to Cu0 [28].

The water pools in reverse micellar solutions provide a largenumber of reaction sites which are separated from each other bythe non-aqueous medium (oil/heptane) [29]. In aqueous solution,CuCl2 dissociates into [Cu(H2O)6]2+ ions (responsible of the sky-blue colouring) and Cl− anions, that can partially coordinate withthe copper ions. In [Cu(H2O)6]2+, the six water molecules com-pletely surround the Cu2+ ion, shielding it. The addition of NaBH4to the microemulsion causes the following reaction (occurring inthe water pools):

NaBH4(s) + 2H2O → NaBO2(aq) + 4H2(g) + heat

Hydrogen, interacting with the hydroxyl ions, causes the produc-tion of solvated electrons:

H2(g) + 2OH− → 2H2O + 2e−(aq)

that can penetrate the hexaaquacopper(II) ion reducing it toCu+(H2O)x (x = 1–4) ions. The accessibility of the copper ion in thosecomplexes is highly favoured, as the coordination with water is lessstrong [30].

Author's personal copy

604 D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608

Table 2Results of WPPM on the SA set of specimens. Data are reported with significant figures only. The column report the water/surfactant ratio ω, mean domain size 〈D〉 (calculatedas first moment of the size distribution), variance of the distribution �, cell parameter a0, scale parameter k and Goodness of Fit (GoF).

Sample ω (A) One Phase (B) Two phases

〈D〉 (nm) � (nm) GoF k a0 (Å) 〈D〉 (nm) � (nm) a0 (Å) 〈D〉 (nm) � (nm) GoF

SA1 1.3 2.6 2.9 1.22 0.459 4.2644 17.6 8.6 4.2961 1.2 0.69 1.25SA2 2.6 15 7.8 1.71 0.667 4.2623 25.6 11.4 4.2867 1.4 1.28 1.18SA3 3.8 19 8.8 1.56 0.748 4.2637 34.4 5.9 4.2815 1.5 1.56 1.43SA4 5.1 4.8 4.6 1.21 0.384 4.2602 37.5 11.7 4.2691 1.5 1.81 1.10SA5 6.4 4.8 4.5 1.22 0.358 4.2589 41.9 4.2 4.2685 1.6 1.91 1.17

The reduced Cu+ ion can lead to the formation of cuprite via thefollowing reaction chain (see e.g. [31]):

Cu+(aq) + H2O → CuOH(aq) + H+(aq) (1)

2CuOH(aq) → Cu2O(s) + H2O (2)

The need to control the reaction, and in particular to control theH+/OH− ratio and the homogeneity of the OH− distribution, is arequirement for the production of pure Cu2O. In fact, in excess ofOH− and at temperature higher than room temperature, CuO canalso form through the following scheme [31]:

Cu2+(aq) + 2OH−(aq) → Cu(OH)2(aq) (3)

Cu(OH)2(aq) + 2OH−(aq) → Cu(OH)42−(aq)

↔ CuO(s) + 2OH−(aq) + H2O (4)

The disproportionation of Cu+ can also lead to the full reduction ofthe copper ions to metallic copper as

2Cu+(aq) → Cu2+(aq) + Cu0(aq) (5)

Nanoparticles formed inside reverse micelles may undergo furthergrowth or aggregation yielding particles larger than their initialnanodroplets, which may result in a bimodal size distribution.

3.2. X-ray diffraction analysis and transmission electronmicroscopy

The XRD patterns show a predominant presence of Cu2O withtraces of CuO and Cu appearing for increasing ω.

The broadening of the line profiles decreases with the increaseof ω: this may indicate an increase in the size of the particles, asdefects contribution to line profiles is expected to be negligible in aslow bottom-up synthesis such as the one investigated here. Quan-titative microstructural information was obtained from XRD databy means of the recently developed WPPM approach [22], a physi-cally sound alternative to traditional line profile analysis based onthe Scherrer formula, Williamson–Hall or Warren–Averbach meth-ods [21,32,33]. WPPM directly connects a physical model for themicrostructure with the diffraction pattern, allowing an extractionof microstructure parameters without recurring to arbitrary peakshapes to fit diffraction peak profiles.

The WPPM results obtained assuming the presence of a singleCu2O phase with a lognormal distribution of spherical domains (cf.Table 2), lead to the conclusions that: (i) there is no evidence ofadditional contributions due to line defects, and in general no con-tributions to line broadening other than those from the small sizeof the crystalline domains; (ii) mean domain sizes range betweena few nanometres (SA1) to nearly 20 nm (SA3), although stan-dard deviations change quite significantly. Apparently, there is noevident correlation between mean domain size and ω when weassume a single mode distribution.

Apart from SA1, the modelling result is not entirely satisfactory.Residuals (difference between model and data) are not completely

flat and featureless, as it would be expected in case of good match.As an example, Fig. 1 shows some peak-like structures in the resid-ual for specimen SA3. The problems in modelling may be due tothe non-ideal conditions of the XRD measurements, for which onewould need larger quantities of crystals, whereas the yield of eachsample preparation was definitely too low to allow for a propermount. However, also in consideration of the large standard devi-ation, the unsatisfactory modelling may originate form a morecomplex size distribution. WPPM was thus tested by including asecond independent domain size distribution (see Table 2), con-sidering a different unit cell parameter (as if in a second cupritephase) to account for possible surface relaxation effects frequentlyobserved in nanocrystalline powders [34,35]. To reduce parame-ter correlations and to allow for a more robust modelling, relativeintensities were constrained to be the same in the two phases, anda scale parameter was refined to determine the volume fractions ofthe two phases.

Although micellar systems are usually thought to provide asingle and narrow distribution of domains [36,37], bimodal distri-butions (sharp XRD peaks and quite broad tails) have been observedtheoretically [38] and experimentally [39]. From this point of view,the use of a method such as WPPM, which considers fine fea-tures of the diffraction pattern, is important: Scherrer formula orWilliamson–Hall method, so far extensively used by the scientificcommunity and based on the estimation of the sole line profilebreadth, easily miss the presence of small domains, related to fea-tures hidden in the tails of the diffraction peaks.

As shown by the Goodness of Fit (GoF) in Table 2, the qual-ity of the modelling improves significantly when adding a secondphase. As an example, Fig. 1 shows a comparison of the residuals forSA3 obtained by using single and bimodal distributions in WPPM:

Fig. 1. WPPM modelling of the SA3 data using two size distributions (dots – rawdata, line – model). For comparison purposes, the difference between data andmodel (residual) is proposed both for the case of 2 distributions and 1 distribution.Residuals are drawn with a 3× magnification factor to highlight differences.

Author's personal copy

D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608 605

Fig. 2. (a) Bimodal particle size distribution for three representative specimens: SA1(square), SA3 (circle) and SA5 (dot). In (b) the trends of average diameter versus ω.Diffraction data (larger domains, square; smaller domains, circle) are shown withresults of the simulation (as of Eq. (3), diamond).

the improvement when using two phases is quite evident. Fig. 2(a)shows the size distributions for specimen SA1, SA3 and SA5. Distri-butions are multiplied by crystallite volume, so that the area undereach distribution is directly equal to the volume fraction.

According to the literature (see e.g. [40,41]), a monomodaldistribution is mostly expected from syntheses done in micellarsystems. In most practical cases, however, it is possible that amultimodal distribution is obtained, but the minority fraction isnot taken into account or not thoroughly studied. For instance, abimodal distribution has been observed in the micellar productionof silver nanodisks by Maillard et al. [42]: the presence of largedomains with a definite shape (discoidal in this case) plus a frac-tion of spherical nanometric domains is clear in the cited literature.For a further validation and comparison with the cited literature,the morphology of the nanocrystals was observed using transmis-sion electron microscopy (TEM). Fig. 3 shows the TEM micrographof a cluster of Cu2O nanocrystals (SA3), statistically well represent-ing the specimen. A multimodal size distribution is evident, whosefeatures are very similar to those observed in the cited paper [42].

The growth above the spherical micelle size limit can beattributed to an interplay between micelles in the microemul-sion. In fact, in a dynamical condition like that of a microemulsion,micelles will not be static isolated objects, but will move and even-tually collide [41,43]. Collisions lead to intermicellar exchange and

Fig. 3. TEM micrograph of sample SA3 showing the presence of both large and smalldomains.

therefore growth of the nanoparticles through mechanisms involv-ing autocatalysis and ripening [41,43]. When an empty micelleimpinges a filled one, the reactant can be transferred and supportthe growth of the existing particle (autocatalysis). The larger theparticle, the larger the area available for further growth. On theother hand, collisions between filled micelles can lead to growthof the larger particle due to phenomena such as Ostwald ripening,that would again favour larger particles with respect to smallerones [41,43]. The incidence of those phenomena is related to therate of micelle encounter and to the rigidity of the surfactant protec-tive layer. The more rigid the layer, the lower the exchange rate and,consequently, the lower the particle growth. Conversely, a less rigidlayer would allow the micelles to be easily deformed: the exchangecan then occur in completely filled micelles, allowing growth abovethe micelle size limit.

Ultrasonication employed to evenly distribute the reducingagent in the micelles may be responsible for the presence of thesmall particles: the excess energy introduced in the system by ultra-sonication may promote an initial formation of an excess of nucleithat remained till the point of emulsion breaking [44]. The absenceof ultrasonication, on the other hand, would not allow a uniformnucleation in the system, whereas an excess ultrasonication mayaid in weakening the surfactant film rigidity, thus promoting anexcessive growth.

It is to exclude that growth occurred after the microemulsionwas broken: breaking the surface tension using acetone just allowsfor precipitation of the particles under the influence of gravity. Thesurfactant is still protecting the surfaces at this stage: the subse-quent removal of the organic phase using ethanol and water, alsoremoves the possible reactants remained in the system, thus effec-tively limiting the possibility of further growth [45]. At this stage wewould expect to see the formation of the nanoparticles aggregatesobserved under the TEM, more than the coalescence of existingdomains.

The fraction of large crystal versus ω is shown in Table 2: SA3 hasa larger fraction of large crystals (74.8%) with respect to the othersamples. Fig. 2(b) illustrates the trend of the average domain sizeof the two size distributions: the average size of the larger fractionclearly increases with increasing ω. An increase in the water con-tent reduces the interaction between the surfactant head groups inthe micelles and further weakens the surfactant layer rigidity. Thisshifts the nucleation mechanism from intramicellar to intermicel-lar, with a tendency to lower the nucleation rate, which favours anoverall growth of larger particles [46].

Just for completeness, we report that the unit cell parameterof the two distributions is slightly decreasing with ω (cf. Table 2):

Author's personal copy

606 D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608

this decrease is compatible with a surface relaxation effect, albeit acertain degree of correlation between cell parameter and specimendisplacement (due to the non-ideal specimen mount in the powderdiffractometer) exists.

3.3. Modelling of the micellar system

There is a rich literature dealing with the correlation betweenparameters of the reverse micelles (such as size or ω) and the size ofthe final nanoparticles. The results, however, are discordant: in cer-tain cases a linear relationship between grain size and ω is observed[47], but in a number of technologically important cases (e.g. Ptor Ag), the correlation is not easily spotted out, and a levelling isobserved starting from a certain diameter [48].

Reverse micelles formation is determined largely by the molec-ular packing of the surfactant Pc = v/(a0lc) [49], where v is thevolume of the hydrocarbon (hydrophobic) chain, a0 is the opti-mal head group area per surfactant molecule and lc is the criticallength of the hydrophobic tail. For Brij30, v = (27.4 + 26.9nc)Å3 andlc = (1.5 + 1.265nc) Å where nc is the number of carbons, 27.4 Å3 isthe volume of CH3-group, 26.9 Å3 is the volume of a CH2 group, 1.5 Åis Van der Waals radius of the terminal CH3 and 1.265 Å is C–C bondlength. Spherical reverse micelles are formed when Pc > 2 [50].

It is possible, using simple thermodynamic arguments, to find arelationship between micelle size and ω [49]. The specific interfacearea between surfactant and water can be defined on the basis ofa0 as ˙ = nBrij.a0. This area can be correlated with ω, noting that:

˙ = nBrij.a0 = Va0

v0.˛

ω(6)

where V is the total volume of the system, ˛ = VH2O/V is the volumefraction of water and v0 = 29.2 Å3 mol−1 is the effective volume of awater molecule.

Critical parameter is a0, determined by the balance of the attrac-tive forces between the surfactant heads caused by the interfacefree energy per unit area � (of the order of ca. 20–50 mJ m−2) andthe repulsive forces due to steric hinderance. The latter can be con-sidered, in a first order approximation, as inversely proportional tothe surface area a occupied by the head group. Thus, the chemicalpotential can be written as

(�0N)surf = �a + B

a(7)

where B is a proportionality constant. The minimum of the chemicalpotential, leading to the optimal head group area, is a0 =

√B/� .

From the surface to volume ratio of a sphere the average micellediameter can be expressed as

D = 6v0

a0ω from which ˙ = 6

V

D˛ (8)

The total number of micelles in the solution can then be found asthe ratio between the available interfacial area and the area of asingle micelle of radius R:

Nm = ˙

�D2= ˙a2

0

36�v20

.1

ω2(9)

whereas the number of surfactant aggregates is obtained as

K = �D2

a0= 36�v2

0

a30

.ω2 (10)

Eq. (3) predicts a linear relationship between the diameter of theparticles and ω.

Fig. 2(b) shows the trend of the predicted diameter superim-posed with the results of XRD–WPPM investigation: calculations

were done considering a0 = 50 ´A for Brij30 [36]. The slope of the

Fig. 4. FTIR spectra of the SA set of specimens. Data have been shifted for clarity.

curve agrees with the large fraction of domains, but the actual val-ues differ. This can be a further indication that the synthesis was notentirely conducted inside the micelles or that, in the solution, themicelles dynamically aggregate to form larger ones. In the litera-ture [36] it is reported that micelles of C12E4/hydrocarbons systemsmay tend to assume non-spherical shapes beyond ω = 4; it is there-fore possible that for Brij30 this threshold is lower. This is furthersupported by the fact that the size of the small fraction is in goodagreement with the literature data for the employed microemul-sion [36,51].

3.4. FTIR spectroscopy

A check for the presence of the organic surfactant on thesurface of the as-obtained nanoparticles was done via infrared spec-troscopy.

Fig. 4 shows the FTIR spectra of the SA set of specimens (theSB set is analogous). All spectra show a prominent band at around1300-1000 cm−1 due to C–O stretching from the surfactant, pos-sibly adhered to the surface of the Cu2O particles. Two bands arerelative to Cu–O modes: the first one, around 625 cm−1, is the sig-nature for Cu2O as it corresponds to Cu(I)–O stretching. The secondone, around 512 cm−1, results from Cu(II)–O stretching and is themarker for the presence of CuO in the system. Traces of CuO arepossibly present in all patterns, as the band can be hidden underthe tail of the stronger Cu2O signal. In the SA5 specimen, the quan-tity of CuO seems higher. The peak at ca. 1632 cm−1 is due to thebending vibrations of physically adsorbed water [52] while OH in-plane bending is at 1260 cm−1. All other peaks are due to organicphases. The peak at 1470 cm−1 is attributed to CH2 in-plane bend-ing while that at 725–720 cm−1 is due to CH2 in-plane rocking. Thebroad band between 1200 and 1050 cm−1 is due to C–O stretch-ing (polyoxyethylene bond, present in the head group of Brij30)and the peak around 668 cm−1 is the bending of CO2. All those sig-nals are compatible with the presence of the Brij30 surfactant. Thehigh-wavenumber region of the spectra is not shown for clarity, butis characterised by the sole presence of the strong O–H stretchingband around 3500–3200 cm−1, mainly due to water and ethanolresidues.

It is worth noting that the FTIR features attributed to the organ-ics phase increase with ω: this is compatible with the observedincrease in the domain/particle size.

Author's personal copy

D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608 607

Fig. 5. UV–vis-NIR spectra of the SA set.

3.5. UV–vis-NIR spectroscopy

Fig. 5 shows the absorption spectra of the as-prepared Cu2Onanoparticles of set SA: absorption features are present in both thevisible range (between 484 and 615 nm) and in the UV range (inthe 264–364 nm region).

The blue shift observed for both peaks with increasing ω isattributed to quantum confinement effects. The general broadeningand lack of sharpness in the absorption spectra may be the resultof increasing light scattering caused by the decrease in grain sizeand widening of the nanocrystal size distribution. A similar effectin the visible range for Cu2O has been observed by other authors[53,54]. All observations give a further indication of the nanomet-ric nature of the particles. The effects seem related to the quantityof nanocrystalline domains, as the sharpest feature is observedfor SA3, the specimen showing the maximum fraction of largerdomains.

3.6. Stability of the Cu2O nanoparticles dispersed in water

As Cu2O is known to have photocatalytic activity, it was arguedwhether the cleaned Cu2O particles dispersed in water would bestable under light irradiation. Aliquots of the same dispersion (SB3specimen) were thus kept under daylight illumination (SB3I) andin the dark (SB3D) for a maximum time of 3 months. Every month,a couple of specimens (one from the SB3I and the other fromthe SB3D sets) were analysed by X-ray diffraction. The specimenskept in the dark did not undergo any visible transformation. Onthe contrary (cf. Fig. 6), a large fraction of Cu2O transformed intoCuO after 2 months illumination. This evidence suggests the exis-tence on an incubation time for the transformation to start: as itis a self-catalytic reaction, presumably CuO forming on the sur-face of Cu2O particles slows the transformation down, acting as ashield for the light. Once a critical size of the nucleus is reached,the transformation starts again with a much higher thermody-namic driving force. The result of a quantification of CuO done viathe Rietveld method is shown in Table 3. It cannot be excludedthat an analogous mechanism is involved in the generation of CuOduring the synthesis, even if the time scale is certainly shorter.However this is a further proof of the self-catalytic transformationof Cu2O into CuO activated by light irradiation in the presence ofwater.

Fig. 6. Phase transformation under visible light. As produced specimen (a) and spec-imen kept under daylight illumination for 1 month (b), 3 months (c) and 5 months(d). Data are shifted and rescaled for clarity.

Table 3Quantitative phase analysis (%wt) in aged cuprite nanostructured powders keptunder illumination.

Cu2O CuO

As-prepared 100% 0%1 month 98.1(5) % 1.9(5) %2 months 9.6(5) % 90.4(5) %5 months 2.0(3) % 98.0(3) %

4. Conclusions

Nanocrystalline Cu2O was synthesized in a reverse micellemicroemulsion. The domains are defect free and show a tendencyto develop a bimodal size distribution, as evidenced by an advancedXRD line profile analysis carried out by Whole Powder Pattern Mod-elling and validated by electron microscopy observation.

The effect of light on the stability of Cuprite in a water-basedenvironment was also studied: Cu2O particles left under daylightillumination start transforming into CuO after about 1 month, withthe whole conversion being completed in a few months time.

Acknowledgements

We acknowledge Profs. Riccardo Ceccato and Alberto Quarantafor assistance in spectroscopy investigations and useful discus-sions. Dr. Fabrizio Girardi, Dr. Valeria Tagliazucca and Mrs. WilmaVaona are also acknowledged for their help in the FTIR and scatter-ing measurements.

References

[1] C. Burda, X.B. Chen, R. Narayanan, M.A. El-Sayed, Chemistry and properties ofnanocrystals of different shapes, Chem. Rev. 105 (2005) 1025.

[2] A. Roucoux, J. Schulz, H. Patin, Reduced transition metal colloids: a novel familyof reusable catalysts, Chem. Rev. 102 (2002) 3757.

[3] A.P. Alivisatos, Semiconductor clusters, nanocrystals and quantum dots, Sci-ence 271 (1996) 933.

[4] V. Pillai, P. Kumar, M.J. Hou, P. Ayyub, D.O. Shah, Preparation of nanoparticlesof silver halides, superconductors and magnetic materials using water-in-oilmicroemulsions as nano-reactors, Adv. Colloid Interface Sci. 55 (1995) 241.

[5] J. Eastoe, B. Warne, Nanoparticle and polymer synthesis in microemulsions,Curr. Opin. Colloid Interface Sci. 1 (1996) 800.

[6] L.M. Qi, J.M. Ma, H.M. Cheng, Z.G. Zhao, Reversed micelle based formation ofBaCO3 nanowires, J. Phys. Chem. B 101 (1997) 3460.

Author's personal copy

608 D. Dodoo-Arhin et al. / Materials Chemistry and Physics 122 (2010) 602–608

[7] M.P. Pileni, in: J.H. Fendler (Ed.), Nanoparticles and Nanostructured Films:Preparation, Characterization and Applications, Wiley–VCH, Weinheim, 1998,p. 71.

[8] L.M. Qi, J.M. Ma, H.M. Cheng, Z.G. Zhao, Preparation of BaSO4 nanoparticles innon-ionic w/o microemulsions, Colloids Surf. A 108 (1996) 117.

[9] S.T. Shishiyanu, T.S. Shishiyanu, O.I. Lupan, Novel NO2 gas sensor based oncuprous oxide thin films, Sens. Actuators B 113 (2006) 468.

[10] A. Mittiga, E. Salza, F. Sarto, M. Tucci, R. Vasanthi, Heterojunction solar cell with2% efficiency based on a Cu2O substrate, Appl. Phys. Lett. 88 (2006) 163502.

[11] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Taracon, Nano-sized transition-metal oxides as negative-electrode materials for lithium-ion batteries, Nature407 (2000) 496.

[12] M. Hara, T. Kondo, M. Komoda, S. Ikeda, K. Shinohara, A. Tanaka, J.N. Kondo, K.Domen, Cu2O as a photocatalyst for overall water splitting under visible lightirradiation, Chem. Commun. 3 (1998) 357.

[13] P.E. de Jongh, D. Vanmaekelbergh, J.J. Kelly, Cu2O: a catalyst for the photochem-ical decomposition of water? Chem. Commun. 12 (1999) 1069.

[14] L. Gou, C.J. Murphy, Solution-phase synthesis of Cu2O nanocubes, Nano Lett. 3(2) (2003) 231.

[15] P. He, X. Shen, H. Gao, Size-controlled preparation of Cu2O octahedronnanocrystals and studies on their optical absorption, J. Colloid Interface Sci.284 (2005) 510.

[16] J. Gao, Q. Li, H. Zhao, L. Li, C. Liu, Q. Gong, L. Qi, One-pot synthesis of uniform Cu2Oand CuS hollow spheres and their optical limiting properties, Chem. Mater. 20(2008) 6263.

[17] M. Yin, C.-K. Wu, Y. Lou, C. Burda, J.T. Koberstein, Y. Zhu, S. O’Brien, Copperoxide nanocrystals, J. Am. Chem. Soc. 127 (26) (2005) 9506.

[18] A. Martínez-Ruiz, G. Alonso-Nunez, New synthesis of Cu2O and Cu nanopar-ticles on multi-wall carbon nanotubes, Mater. Res. Bull. 43 (6) (2008)1492.

[19] C. Salzemann, I. Lisiecki, A. Brioude, J. Urban, M.-P. Pileni, Collections of coppernanocrystals characterized by different sizes and shapes: optical response ofthese nanoobjects, J. Phys. Chem. B 108 (2004) 13242.

[20] P. Scardi, M. Leoni, On the crystal structure on nanocrystalline Cu2O, Mater. Sci.Eng. A 393 (2005) 396.

[21] P. Scherrer, Nachr. Ges. Wiss. Göttingen, 26 September (1918) 98.[22] P. Scardi, M. Leoni, Whole powder pattern modelling, Acta Cryst. A 58 (2002)

190.[23] A. Cervellino, C. Giannini, A. Guagliardi, M. Ladisa, Nanoparticle size distribution

estimation by a full-pattern powder diffraction analysis, Phys. Rev. B 72 (2005)035412.

[24] P. Scardi, M. Leoni, Line profile analysis: pattern modelling versus profile fitting,J. Appl. Cryst. 39 (2006) 24–31.

[25] J.P. Cline, R.D. Deslattes, J.-L. Staudenmann, E.G. Kessler, L.T. Hudson, A. Hennins,R.W. Cheary, NIST Certificate, SRM 660a Line Position and Line Shape Standardfor Power Diffraction, Gaithersburg, USA, NIST Standard Reference MaterialsProgram, NIST, USA, 2000.

[26] G. Caglioti, A. Paoletti, F.P. Ricci, Choice of collimators for a crystal spectrometerfor neutron diffraction, Nucl. Instrum. Methods 3 (1958) 223.

[27] M. Leoni, T. Confente, P. Scardi, PM2K: a flexible program implementing wholepowder pattern modelling, Z. Kristallogr. Suppl. 23 (2006) 249.

[28] Y. Gao-Qing, H.-F. Jiang, C. Lin, S.-J. Liao, Shape- and size-controlled electro-chemical synthesis of cupric oxide nanocrystals, J. Cryst. Growth 303 (2007)400.

[29] M.P. Pileni (Ed.), Reactivity in Reverse Micelles, Elsevier, New York,1989.

[30] T. Iino, K. Ohashi, Y. Mune, Y. Inokuchi, K. Judai, N. Nishi, H. Sekiya, Infrared pho-todissociation spectra and solvation structures of Cu+(H2O)n (n = 1–4), Chem.Phys. Lett. 427 (1–3) (2006) 24.

[31] C. Qingde, X. Shen, H. Gao, Formation of nanoparticles in water-in-oilmicroemulsions controlled by the yield of hydrated electron: the controlledreduction of Cu2+, J. Colloid Interface Sci. 308 (2007) 491–499.

[32] B.E. Warren, X-ray Diffraction, Addison-Wesley, Reading, MA, 1969.[33] E.J. Mittemeijer, P. Scardi (Eds.), Diffraction Analysis of the Microstructure of

Materials, Springer Verlag, Berlin, 2004.[34] A.C. Nunes, D. Lin, Effects of surface relaxation on powder diffraction patterns

of very fine particles, J. Appl. Cryst. 28 (1995) 274.[35] M. Leoni, P. Scardi, Grain surface relaxation effects in nanocrystalline powders,

in: E.J. Mittemeijer, P. Scardi (Eds.), Diffraction Analysis of the Microstructureof Materials, Springer Verlag, Berlin, 2004, pp. 413–454.

[36] M. Vasilescu, A. Caragheorgheopol, M. Almgren, W. Brown, J. Alsins, R. Johanns-son, Structure and dynamics of nonionic polyoxyethylenic reverse micelles bytime-resolved fluorescence quenching, Langmuir 11 (1995) 2893.

[37] P. Angshuman, S. Shah, S. Devi, Preparation of silver, gold and silver–goldbimetallic nanoparticles in w/o microemulsion containing TritonX-100, Col-loids Surf. A: Physicochem. Eng. Aspects 302 (2007) 483.

[38] C. Tojo, M.C. Blanco, M.A. López-Quintela, Preparation of nanoparticles inmicroemulsions: a Monte Carlo study of the influence of the synthesis variables,Langmuir 13 (17) (1997) 4527.

[39] M.P. Pileni, Reverse micelles as microreactors, J. Phys. Chem. 97 (27) (1993)6961.

[40] M. Fanun (Ed.), Microemulsions. Properties and Applications, Taylor & FrancisGroup, Boca Raton, FL, 2009.

[41] L. Qi, Synthesis of inorganic nanostructures in reverse micelles, in: P. Somasun-daran, A. Hubbard (Eds.), Encyclopedia of Surface and Colloid Science, Taylorand Francis, 2006, pp. 6183–6207.

[42] M. Maillard, S. Giorgio, M.P. Pileni, Silver nanodisks, Adv. Mater. 14 (2002) 15.[43] M. de Dios, F. Barroso, C. Tojo, Nanoparticle formation in microemulsions:

mechanism and Monte Carlo simulations, in: M. Fanun (Ed.), MicroemulsionProperties and Applications, vol. 14, CRC Press, Taylor & Francis Group, BocaRaton, FL, 2009, pp. 455–463.

[44] P. Barnickel, A. Wokaun, Synthesis of metal colloids in inverse microemulsions,Mol. Phys. 69 (1) (1990) 1.

[45] A. Manna, T. Imae, T. Yogo, K. Aoi, M. Okazaki, Synthesis of gold nanoparticles ina Winsor II type microemulsion and their characterization, J. Colloid InterfaceSci. 256 (2002) 297.

[46] M.M. Husein, E. Rodil, J.H. Vera, A novel approach for the preparation of AgBrnanoparticles from their bulk solid precursor using CTAB microemulsions,Langmuir 22 (5) (2006) 2264.

[47] M.P. Pileni, The role of soft colloidal templates in controlling the size and shapeof inorganic nanocrystals, Nat. Mater. 2 (2003) 145.

[48] J. Tothova, M. Richterova, V. Lisy, On two direct methods for measurementof interfacial tension at microdroplet surfaces, Institute of Physics, P.J. SafarikUniversity, Slovakia, Paper deposited on arXiv:cond-mat/0411341v1 (2004).

[49] J.N. Israelachvili, Intermolecular and Surface Forces, 2nd ed., Academic Press,1994.

[50] D.J. Mitchell, B.W. Ninham, Micelles, vesicles and microemulsions, J. Chem. Soc.Faraday Trans. 2 (77) (1981) 601.

[51] F. Luo, D. Wu, L. Gao, S. Lian, E. Wang, Z. Kang, Y. Lan, L. Xu, Shape-controlledsynthesis of Cu2O nanocrystals assisted by Triton X-100, J. Cryst. Growth 285(2005) 534.

[52] A.S. Arico, V. Baglio, A. Di Blasi, V. Antonucci, FTIR spectroscopic investigationof inorganic fillers for composite DMFC membranes, Electrochem. Commun. 5(2003) 862.

[53] M. Yang, J. Zhu, Spherical hollow assembly composed of Cu2O nanoparticles, J.Cryst. Growth 256 (2003) 134.

[54] K. Borgohain, N. Murase, S. Mahamuni, Synthesis and properties of Cu2O quan-tum particles, J. Appl. Phys. 92 (3) (2002) 1292.