Journal of Colloid and Interface Science 312 (2007) 68–75www.elsevier.com/locate/jcis

Generation of metal oxide nanoparticles in optimised microemulsions

Ali Bumajdad ∗, Julian Eastoe 1, Mohamed I. Zaki 2, Richard K. Heenan 3, Lata Pasupulety

Kuwait University, Chemistry Department, Faculty of Science, P.O. Box 5969, Safat 13060, Kuwait

Received 11 May 2006; accepted 6 September 2006

Available online 9 September 2006

Abstract

The phase behavior and structure of aqueous-in-n-heptane microemulsions, stabilized by surfactant mixtures of di-n-didodecyldimethylam-monium bromide, DDAB, and Brij®35 were studied by small angle (neutron or X-ray) scattering techniques. The aqueous nanodroplets containeither a precursor reactive salt or a precipitating agent, so that simple mixing induces nanoparticle formation. These formulated microemulsionsdisplay good phase stability against added polar additives such as monovalent, divalent, trivalent metal ions, ammonia solution, tetrabutylammo-nium hydroxide, and their mixtures. Nanoparticle formation was demonstrated via precipitation of metal oxides inside the water nanodroplets,affording control over the resulting particle size. Nanoparticle characteristic size (XRD- and HR-TEM derived sizes) and specific surface areas(SBET (m2 g−1)) for iron oxide and CeO2 prepared in these mixed microemulsions, are compared with those stabilized by single surfactantsDDAB and Pure AOT.© 2006 Elsevier Inc. All rights reserved.

Keywords: Surfactant mixtures; Nanoparticles; Water-in-oil microemulsions; DDAB; Brij; Metal oxide; SANS; SAXS

1. Introduction

In recent years, nanosized particles have been the focus ofsustained research and industrial interest [1], which is expectedto continue for years to come. This is due to various potentialapplications in fields like electronics and magnetism [1], bio-medicine [1], automation and environmental protection [2–4].Important futures are the high surface energies and chemicalreactivity of nanoparticles. Therefore, a variety of techniquessuch as precipitation, sol–gel processing, hydrothermal synthe-sis, sonochemical treatment, microwave decomposition, flamespray, amongst others (see for example [4–8]) have been suc-cessfully used to prepare finely dispersed nanoparticles. Bothprecipitation and sol–gel processing can be carried out in aque-ous and self-assembly systems (e.g., liquid crystals [9], micelles

* Corresponding author. Fax: +965 4816482.E-mail address: bumajdad@kuc01.kuniv.edu.kw (A. Bumajdad).

1 Permanent address: School of Chemistry, University of Bristol, Bristol BS81TS, UK.

2 Permanent address: Chemistry Department, Faculty of Science, Minia Uni-versity, El-Minia 61519, Egypt.

3 Permanent address: ISIS Facility, Rutherford Appleton Laboratory, Chilton,Oxon OX11 0QX, UK.

0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2006.09.007

[10] and water-in-oil microemulsions [11–14]). Structural anddynamical characterization of self-assembled templates is use-ful for understanding the mechanism and growth control ofnanoparticles. Nevertheless, only a few studies have fully char-acterized the parent self-assembled systems used in the parti-cle preparation; for example, the work reported by Caragheo-rgheopol et al. for micellar systems [10].

Microemulsions have been successfully employed to obtainrelatively small particles (i.e. high surface area), with well con-trolled properties (see for example [11–13]). Such dynamiccolloidal templates are known to produce particles of smallersize than those obtained via normal precipitation in aqueoussystems. In fact water-in-oil microemulsions have successfullybeen used to produce a variety of nanoparticle shapes and sizes[12,14]. Nevertheless, it is not known how such templates con-trol the size and shape of the resulting nanoparticles, and there-fore, this issue still requires further examination. It is knownthat type and concentration of surfactant [15–17], both oil[13,15], and alcohol types [18–22], droplet core size [11], speedof microemulsion mixing [23] and particle aging [24], influenceparticle shape, size and/or growth rate. For example, it has beenfound that the particle size of CdS, ZnS or AgCl, can be con-trolled by the microemulsion water pool size (in terms of the

A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75 69

water solubility parameter W = [water]/[surfactant]) [10,12].In contrast, the synthesis of some pure metal particles, such asCu, shows no such correlation [25]. This difference in behaviormay be attributed to weaker surfactant-head group interactionwith the compound particles, hindering the formation of largerparticle sizes. Although microemulsion templating is known toproduce particles of desirable properties, there it does appear[26] that the crucial size-controlling parameter(s) of nanopar-ticles is not only the template itself, and there must be otherimportant factors. Some researchers [11] believe that water-in-oil microemulsion aggregates can be employed to control thesize of spherical nanoparticles, especially at low water con-tents [W ]. It is thought that the water structure in these lowW reversed micelle compartments plays an important role [27].Control over nanoparticle shape, however, is more complexand anisotropic microemulsions do not naturally lead to non-spherical templated particles [11]. Others [28] think that themicroemulsion droplet size only affects the size of the nuclei,and not the final particle size. The nature of the adsorbed sur-factant film is also an important factor for generating a narrowparticle size distribution.

Due to its exceptional performance, Aerosol-OT (AOT,sodium bis(2-ethyl-1-hexyl) sulfosuccinate) is one of the mostwidely used microemulsion surfactants. Excellent reviews cov-ering behavior and properties of AOT are available [29,30].From the technical viewpoint, AOT microemulsions suffer fromtwo major disadvantages for nanoparticle generation: (i) lowaqueous phase solubility when salts (precursors) are present,and (ii) formation of the single phase region (L2) is shifted tohigher (less practical) temperatures.

As it has been shown previously [31,32], mixing cationicdouble-tailed surfactant with non-ionic surfactant of the typeC12EJ (with J � 5) enhances the water solubility limits, byadjusting the preferred interfacial curvature. In this paper it isshown that such mixed systems can overcome the two main dis-advantages noted above for AOT systems. In other words, thepresent work shows that cationic–non-ionic surfactant mixturesnot only enhances aqueous phase solubility, but also lead to ex-cellent thermal stability.

Interestingly, mixing di-n-didodecyldimethylammoniumbromide (DDAB) with small amounts of the non-ionic Brij®35(on average polyoxyethylene 23 dodecyl ether, C12E23), givesrise to microemulsions offering an excellent environment fornanoparticle synthesis [33]. These optimised DDAB–Brij(®35microemulsions have the advantages of both low thermal sen-sitivity and stability toward added inorganic salt type and con-centration. The positive head-group charge of the cationic sur-factant ensures bulk solubility of the inorganic metal cation,whereas the long polyethene oxide chain of the non-ionicsurfactant facilitates increased steric stability. Furthermore,mixed systems have not been investigated before for thispurpose. A recent review article [34] reports many exam-ples of nanoparticle preparation with non-ionic-, anionic-, andcationic-surfactant microemulsions, but does not mention anywork using double-tailed DDAB as a microemulsion stabilizerto generate nanoparticles. This may be due to the narrow single-phase water-in-oil (w/o) microemulsion region with DDAB. In

this work it is shown that replacing 10 mol% of the DDAB withBrij®35 results in a much broader single-phase w/o microemul-sion region, making the optimised microemulsion systems suit-able for metal oxide synthesis.

Phase behavior (water solubility limits), small-angle neutronscattering (SANS), and small-angle X-ray scattering (SAXS),were used to characterize the structure of the parent microemul-sions, whereas, X-ray powder diffractometry (XRD), and nitro-gen sorptiometry were applied to characterize the final metaloxide nanoparticles.

2. Experimental

The di-chain cationic di-n-dodecyldimethylammonium bro-mide (DDAB) surfactant was prepared and purified as describedpreviously [35]. Elemental analysis (CHNS-LECO 932), melt-ing point (Gallenkamp), IR (Perkin–Elmer system 2000) and1H NMR (Bruker DPX 400) were all consistent with the de-sired product. In terms of microemulsion phase behavior at0.10 mol dm−3 surfactant in n-heptane the maximum water sol-ubility, Wmax, also agreed well with literature (Wmax = 25 ± 1at 25 ◦C).

The non-ionic polyoxyethylene 23 lauryl ether (Brij®35)and the branched di-chain anionic sodium bis(2-ethylhexyl) sul-fosuccinate (AOT) were Sigma products and used as received.

The metal salts (Ce(NO3)3·6H2O, Fe(NO3)3·9H2O, Cr-(NO3)3·9H2O, Ni(NO3)2·6H2O, KNO3, KCl, LiNO3, andLiCl), were all Aldrich products of purity �99%. The ammo-nium hydroxide (NH4OH, Aldrich, 28% NH3 in water), 40 wt%tetramethylammonium hydroxide (25 wt% (CH3)4N+OH−,Aldrich), tetraethylammonium hydroxide (40 wt% (C2H5)4N+-OH−; Aldrich), and tetrabutylammonium hydroxide (20 wt%(C4H9)4N+OH−; Aldrich) were diluted with an appropriateamount of distilled water to give 10 wt% base solution.

Deuterated water, D2O (99.9% D-atom D2O; Fluorochem),was used as received, whereas liquid H2O was from a RO100HP Purite Water system. Proteated n-heptane (HPLC-grade; Aldrich) was passed through a fresh alumina columnbefore use, to remove any traces of semi-polar impurities.

Details of microemulsion and nanoparticle preparation andcharacterization techniques can be found in supplementary ma-terials. This section also outlines the small-angle neutron andsmall-angle X-ray scattering methods (SANS and SAXS), em-ployed to analyze the microemulsions as well as the X-raydiffraction (XRD), Nitrogen sorptiometry and high-resolutionTEM (HRTEM) used to characterize the nanoparticles.

3. Results and discussion

3.1. Microemulsion phase behavior

Unsurprisingly, the phase behavior of AOT-microemulsionswas found to be sensitive to the metal salt addition. At roomtemperature, the water solubility limit of AOT-stabilized water-in-n-heptane microemulsion is known to be around W ≈ 64[36]. When 0.05 M solution of a metal salt (e.g., Ce(NO3)3or Fe(NO3)3) was used instead of pure water, the single-phase

70 A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75

Table 1Microemulsion designations and values derived from fits to SANS data (corecontrast, D/H/H) and from Porod plots of pure and additive-containing mi-croemulsionsa

Microemulsionb W c T LCµ (Å) p

DDAB + Brij®35-based Model Porod

DB pure 1 50 25 135.6 154.3 0.23DB pure 2 70 25 178.4 192.9 0.22DB pure 3 60 25 146.5 174.2 0.22DB pure 1 + 2 60∗ 25 145.8 174.2 0.22DB NH3aq 70 25 187.4 192.8 0.24DB TBAH 70 25 173.1 192.8 0.22DB Fe 50 25 143.9 154.2 0.23DB Fe + DB NH3aq 60∗ 25 157.0 166.1 0.22DB Fe + DB TBAH 60∗ 25 150.4 174.2 0.23DB Ce 50 25 143.4 151.7 0.21DB Ce + DB NH3aq 60∗ 25 157.9 166.2 0.22DB Ce + DB TBAH 60∗ 25 151.7 174.2 0.21DB Ni 50 25 132.2 154.2 0.22DB Cr 50 25 127.5 148.4 0.20DB Cu 50 25 127.0 150.2 0.22DB K 50 25 133.6 156.4 0.22DB K 0.2 47 25 125.1 151.6 0.23DB K 0.5 41 25 115.0 136.8 0.23DB CeKd 50 25 134.8 156.8 0.22

a LCµ is the droplet average core diameter and p is the model-fitted polydis-

persity. Unless otherwise stated next to the salt symbol, the salt concentrationwas 0.05 M. The overall surfactant concentration = 0.10 M, and ammonia andTBAH solution was 10 wt%. Uncertainties: W ± 1, LC

µ (Å) (Model) ±2 Å,LC

µ (Å) (Porod) ±10 Å, p ± 0.02.b The microemulsion designation consists of the surfactant(s) used, fol-

lowed by the aqueous additive and the additive concentration. The sign (+)is inserted when two microemulsions are mixed. The symbol DB stands forDDAB + Brij® (90%:10% molar ratio) stabilized microemulsion. Pure is formicroemulsions containing no additives.

c ∗ mark is for average W of a mixture of two microemulsions.d This sample contains 0.042 M Ce(NO3)3 and 0.008 M KNO3, with an over-

all salt concentration of 0.05 M.

(L2) microemulsion did not form at room temperature, but itwas found necessary to increase the temperature up to 70 ◦C.The water solubility limit at this temperature, was found to beW ≈ 30. In other words, the phase boundaries were shiftedtoward higher temperatures. This behavior is likely to be dueto the interaction between the mono-negative head-group andthe multi-positive metal ions. Hence, it was necessary to carryout the metal oxide nanoparticle preparation at temperatures�70 ◦C. In this work, the lowest possible temperature (70 ◦C)was employed and, hence, some solvent evaporation is in-evitable. It is known that AOT-microemulsions are temperature-sensitive [37]. Hence, the different droplet sizes observed forpure-D2O microemulsions and the salt-containing ones, hav-ing the same W (Table 1), could be due either to the differ-ent temperatures, the salt screening effect, or a combination ofboth factors. However, these differences are not of major sig-nificance when considering nanoparticle formation describedbelow.

On the other hand, the DDAB–Brij®35 system behavioris much different from AOT. It was tested in the presence ofCe(NO3)3, Fe(NO3)3, Ni(NO3)2, Cr(NO3)3, Cu(NO3)2, KNO3,KCl, Ce(NO3)3:KNO3 (92:8 mole fraction %); and LiNO3 ad-

ditives, all at 0.05 M salt concentration, and found to formsingle phase (1φ) microemulsions in the range W ≈ 12–50 atroom temperature. Hence, the advantage of this mixed systemis a lower sensitivity to the metal salt type. Moreover, the watersolubility limits were studied as a function of KNO3 concentra-tion (0.05, 0.2, and 0.5 M) and it was found to be less sensitivetoward salt concentration than the single component AOT sys-tems. For 0.05, 0.2, and 0.5 M KNO3, the single-phase region(1φ) was found to commence at W = 12 for all the three con-centrations, whereas the upper solubilization limit was W = 50,47, and 41, respectively, for these additive concentrations. Thislater feature is useful for enhancing the yield of nanoparticlesdescribed below. It is worth mentioning, that at 0.10 M over-all surfactant concentration and W = 60 (the water solubilityafter mixing the two microemulsions), and assuming spheri-cal droplets for these systems [31,32] the droplet concentrationis estimated as 5.3 × 1019 droplet L−1. Now, for an equivalent0.05 M salt concentration, the metal ion concentration in theaqueous dispersed phase works out to be 1.63 × 1021 ion L−1,making about 30 metal ions per droplet. Of course, increas-ing the salt concentration, while maintaining constant surfactantconcentration and W , would further increase the metal ion oc-cupancy number inside the droplets.

The single-phase region limits in the presence of 10 wt%ammonia (aqueous NH3), 10 wt% TBAH, and 10 wt% TMAHwere found to be at W ≈ 8 to ≈75, W ≈ 15 to >100, or W ≈ 16to ≈35, respectively. Aqueous NH3 is soluble in the bulk wa-ter rather than residing on the interface (this is known fromthe SANS data analyses for the D2O/NH3 mixture describedbelow). Due to its weak ionization tendency (i.e., low chargestrength and, hence, weak surfactant head-group screening ef-fect), the aqueous phase solubility limits are expected to bebetter than when salt is present. Note that the 10 wt% is equiv-alent to 6.5 M, 1.2 M, and 0.42 M for NH4OH, TMAH, andTBAH, respectively. This would result in 2.12 × 1023 NH4OHmoieties L−1 and around 4000 NH4OH per droplet when sur-factant concentration is 0.10 M and W is 60. Thus the effec-tive concentration of NH4OH is 130 times that of metal salt,and for M3+ salts about 43 times the nitrate concentration.Because of its lower concentration, the TBAH concentrationwould lead to 1.37 × 1022 TBAH units L−1 and 260 TBAHunits per droplet.

The high water solubility in the presence of TBAH (evenbetter than for the pure microemulsion system) is due to its sur-face activity. For TMAH-containing microemulsions, however,the aqueous phase solubility limits are lower than those of metalsalt-only containing microemulsions. This behavior may be as-cribed to the high ionic nature and a low interfacial activity ofTMAH (methyl groups are insufficiently hydrophobic to be sur-face active).

3.2. Microemulsion structure

To explore the general effects of additive reagents (metalions and precipitating agents) on a standard microemulsion,trial studies were performed with AOT systems. These resultsare documented in supporting material, but the findings are:

A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75 71

Fig. 1. (a) Core contrast SANS data and fits for D2O-in-n-heptane,Ce(NO3)3-containing D2O-in-n-heptane, and Fe(NO3)3-containing D2O-in-n-heptane microemulsions stabilized by DDAB–Brij®35 mixtures at W = 50.Salt concentration = 0.05 M in the aqueous phase. The data and fit of DB Ceand DB pure 1 samples multiplied by a factor of 2 and 4, respectively, for clar-ity. Inset is for the same data but without multiplication factor. (b) Porod plotof the same data without the multiplication factor.

(a) the presence of Ce(III) and Fe(III) nitrates, at typical levelsmentioned above, appears to have negligible effects on thedroplet size (Fig. S1, supplementary material);

(b) added NH3 alone leads to a decrease in aqueous nan-odroplet size (Fig. S2, supplementary material);

(c) heating the NH3-containing AOT-stabilized nanodropletsfrom 25 to 70 ◦C had no measurable effect on water coreradius (Table S1, supplementary material);

(d) the structure of a microemulsion containing a mixture ofmetal ion and ammonia reagents is very similar to a controlsystem with water only (Fig. S3, supplementary material).

Having established these effects for standard AOT mi-croemulsions similar, but more extensive, experiments werecarried out with the new mixed DDAB–Brij®-stabilized sam-ples.

3.2.1. Effect of additives on DDAB–Brij® microemulsions:SANS study

Since the DDAB–Brij®-stabilized systems show minimalsensitivity towards the metal salt additives and high thermal sta-bility, it was possible to conduct the study at room temperature.Fig. 1 compares the SANS data of Fe(NO3)3-containing andCe(NO3)3-containing microemulsions with the data obtainedfor pure D2O at constant W = 50. The polydisperse spheremodel (supplementary material) was found to fit the data well,

Fig. 2. (a) Core contrast SANS data and fits for two samples prepared by mixingequal volumes of two DDAB–Brij®35 microemulsions, one containing metaladditives and the other containing ammonia. The average aqueous solubilityafter mixing is W = 60. The sample compositions and nomenclature are definedin Table 1. The DB Ce + DB NH3 data and fit multiplied by a factor of 4 forclarity. Inset is for the same data but without this multiplication factor. (b) Porodplot of the same data without the multiplication factor.

and parameters can be found in Table 1. Interestingly, there isonly a minimal effect of the different for salt additives on themicroemulsion structure, as can be seen in the inset of Fig. 1aand the Porod plot shown in Fig. 1b.

As shown in supplementary material (Fig. S4) there is verylittle difference in structure between systems with NH3 orTBAH.

The structure of nanoparticle-containing microemulsionsstabilized by DDAB–Brij® was also examined. Fig. 2 showsthe SANS data and model fits for a 1:1 mixture of Fe(NO3)3-containing and ammonia-containing microemulsions (DB Fe +DB NH3aq) and 1:1 mixture of Ce(NO3)3-containing andammonia-containing microemulsions (DB Ce + DB NH3aq).The average water solubility, Wav, for both samples was 60. It isclear by inspection of Fig. 2 and also from the plot Porod Qmax

(Table 1), that the two different samples possess essentiallythe same underlying structure and droplet size. Examination ofFig. 3 reveals similar behavior, for 1:1 mixtures of DB Fe + DBTBAH and DB Ce + DB TBAH (fit parameters found in Ta-ble 1).

Fig. 4 compares four different microemulsions having thesame level of aqueous components. Two of these microemul-sions (DB pure 1 + DB pure 2 and DB pure 3) contain nometal salt, whereas the other two (DB Fe + DB TBAH andDB Ce + DB TBAH) contain trivalent metal ions and TBAH.

72 A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75

Fig. 3. (a) Core contrast SANS data and fits for two samples prepared by mixingequal volumes of two DDAB–Brij®35 microemulsions one containing metaland the other containing TBAH. The average aqueous solubility after mixingwas W = 60. The sample compositions and nomenclature are defined in Ta-ble 1. TBAH concentration is 10 vol% in the aqueous phase. The DB Ce + DBTBAH data and fit multiplied by a factor of 4 for clarity. Inset is for the samedata without multiplication factor. (b) Porod plot of the same data without mul-tiplication factor.

Fig. 4. SANS Porod plot for DDAB + Brij-stabilized microemulsions compar-ing the effect of mixing equal volumes of the indicated microemulsions. Themetal nitrate concentration was 0.05 M in the aqueous phase. The average watersolubility W = 60 and the sample identities are indicated in Table 1. Tempera-ture is 25 ◦C. The data of DB Fe + DB TBAH, DB pure 3 and DB pure 1 + DBpure 2 samples multiplied by a factor of 2, 2 and 4, respectively, for clarity.

From the first maximum in the Porod plots one can clearly seethat mixing equal volumes of the salt-containing microemul-sions results in essentially the same droplet size, compared tothat found following mixing equal volumes of pure (additive

Table 2Porod-derived microemulsion core size (LC

µ (±2 Å)) and droplet size (LDµ

(±5 Å)) of some parent microemulsions

Sample W LCµ (Å)

(SANS data)LD

µ (Å)(SAXS data)

DB pure 1 50 154.3 163.6DB NH3aq 70 192.8 217.0DB TBAH 70 192.8 217.0DB Fe(NO3)3 0.05 50 154.2 168.8DB Ce(NO3)3 0.05 50 151.7 163.6DB Ni(NO3)2 0.05 50 154.2 177.1DB Cr(NO3)3 0.05 50 148.4 163.6DB Cu(NO3)2 0.05 50 150.2 163.6DB KNO3 0.05 50 156.4 163.6DB KNO3 0.20 47 151.6 168.8DB KNO3 0.50 41 136.8 150.0DB Ce(NO3)3KNO3 0.05 50 156.8 163.6

free) microemulsion. Both samples have the same droplet sizeas regular additive free microemulsions, prepared at the sameW of 60.

3.2.2. Effect of additives on DDAB–Brij® microemulsions:SAXS study

The microemulsion samples were also studied by SAXS[38]. Based on electron density the expected order of scatteringintensity is Ce- > Cu- > Ni- > Fe- > Cr- > K-salt and this wasobserved. Fig. 5 shows an example of the comparison betweenSANS and SAXS Porod plots for DDAB–Brij®-stabilized sys-tems containing water only, and the aqueous additives Ce(III),Fe(III) and NH3. Data for other conditions, compositions andadditive types are shown in supplementary material (Figs. S5and S6). Particle dimensions, estimated from Porod plot Qmax

peaks for all the conditions studied are given in Table 2. Thenomenclature employed is Lc

µ to represent the SANS-deriveddiameter, and LD

µ to identify the value reported by SAXS mea-surements. Owing to a subtle shift in the plane of contrast (theouter extent of D2O for SANS, and the electron-dense head-group/counterion region for SAXS) the SAXS-derived diam-eters are slightly larger (Table 2). It is obvious in Fig. 5 thatthe first maximum occurs at essentially the same Q value, i.e.a very similar droplet size. The decrease in SANS intensityobserved for the DB NH3aq sample may be attributed to theH2O/D2O mixing effect described above. Another observation,is that the scattering intensity of DB pure 1, DB Ce, and DBFe is essentially the same in the SANS curves, whereas a cleardifference between the SAXS scattering curves is manifestedfor the same set of samples. This is because the SANS scatter-ing intensity depends essentially on D2O content, whereas theSAXS scattering power depends on the electron density. Hence,in SAXS the scattering intensity is expected to be in the orderDB pure 1 < DB Fe < DB Ce.

All the obtained results for DDAB–Brij® microemulsionsystems show minimal effects on both core (SANS results) anddroplet (SAXS) size. Thus, there is broad agreement betweenSAXS and SANS methods for characterising these nanoparti-cle reactor systems.

A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75 73

Fig. 5. SANS- and SAXS-based Porod plot for pure D2O, Ce(NO3)3-containing, Fe(NO3)3-containing and NH3-containing D2O-in-n-heptane microemulsionsstabilized by a DDAB + Brij®35 surfactant mixture, at W = 50. The metal nitrate concentration was 0.05 M in the aqueous phase. The sample identities areindicated in Table 1. Temperature is 25 ◦C.

Fig. 6. X-ray powder diffractograms observed for iron oxide particles (calcinedand uncalcined as indicated). Peaks of 400-calcined sample are indicated bythe corresponding reflecting crystal facet. Peaks labelled with an asterisk areoriginally from the sample holder (Rh/Pd).

3.3. Nanoparticle characteristics

The DDAB + Brij®35-stabilized aqueous-in-n-heptane mi-croemulsions were successfully used to synthesize iron oxide(α-Fe2O3) and cerium oxide (CeO2) nanoparticles of size inthe range 30–60 Å, using ammonia as a precipitant. The crys-talline structure and average particle size were determined bymeans of XRD and HRTEM, whereas the surface area was mea-sured by BET-analysis of N2 adsorption data. Figs. 6 and 7show XRD patterns for the as-precipitated and calcined iron

Fig. 7. X-ray powder diffractograms observed for cerium oxide particles (cal-cined and uncalcined as indicated). Peaks of 800-calcined sample are indicatedby the corresponding reflecting crystal facet. Peaks labelled with an asteriskoriginate from the filament.

oxide and cerium oxide, respectively. The labelled peaks inFig. 6 are for hematite (α-Fe2O3), according to JCPDS CardNo. 39-1346 [39]. Hence, the crystalline phase emerges moreclearly upon calcination of the parent material at �400 ◦C. Itis worth mentioning, that in the literature some microemulsion-mediated iron oxide syntheses yield γ -Fe2O3 [40], and othersare thought to result in α-Fe2O3 [41] depending on the pre-cursor. Our study of the as-precipitated iron oxide/hydroxidematerials using slower step time (8 s) shows that at least some ofthe as-precipitated materials are of maghemite/magnetite struc-

74 A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75

Table 3Parent microemulsion core size (Lc

µ ±2 Å), XRD-determined average particlesize (Lp

XRD ±1 Å), TEM-determined average particle size (LpTEM ±2 Å) and

N2-determined surface area (SBET ±2 m2 g−1) of the test α-Fe2O3 and CeO2

Sample Lcµ

(Å)Lp

XRD

(Å)Lp

TEM

(Å)SBET

(m2 g−1)

α-Fe2O3(DB)NH3a 157.0 – 40.7 247.2

α-Fe2O3(DB)TBAHb 150.4 – – –

CeO2(AOT)NH3a 70.6 89.9 – 144.3

CeO2(DB)NH3a 157.9 58.9 42.9 239.0

CeO2(DB)TBAHb 151.7 100.4 − 173.5

a The particles were prepared by mixing two DDAB + Brij®-stabilized mi-croemulsions one contains the metal precursor and the other contains 10%ammonia solution.

b The particles were prepared by mixing two DDAB + Brij®-stabilized mi-croemulsions one contains the metal precursor and the other contains 10%TBAH solution.

ture. This feature is not discernable in Fig. 6 because the shorterstep time of 0.2 s.

The XRD results show, moreover, that the crystallizationinto α-Fe2O3 is enhanced upon increasing the calcination tem-perature. Similar results were observed for iron oxide precipi-tated by TBAH. The uncalcined iron oxide that had been pre-cipitated by TBAH showed no crystallinity, whereas the systemcalcined above 400 ◦C was crystalline (data not shown). In con-trast, cubic-CeO2 was found to be an immediate product of thereaction between Ce(NO3)3 and NH4OH solution. Notably, theXRD peaks shown on Fig. 7 are almost identical to those filedfor cubic-CeO2 in the standard JCPPS Card No. 04-0593 [39]).In other words, crystallite CeO2 appears to form, even withoutcalcination.

HRTEM images obtained for α-Fe2O3 and CeO2 nanopar-ticles generated in these microemulsions are shown in Fig. 8.The images reveal clearly nanosized particles in the range of30–60 Å. Table 3 compares the particle dimensions as deter-

Fig. 8. HRTEM image of some nanosize metal oxides prepared by mixing equalvolumes of two DDAB+Brij microemulsions (one contains the metal precursorand the other contains 10% NH3 solution).

mined by XRD (by adopting the line-broadening technique andSherrer formula [42]) and TEM. For α-Fe2O3, XRD could notbe used to determine the particle size at room temperature be-cause the samples were only weakly crystalline.

4. Summary

This work describes novel water-in-oil mixed surfactantmicroemulsion systems, which can be used for the synthe-sis of metal oxide nanoparticles. The system probed is com-posed of n-heptane/90 mol% fraction DDAB/10 mol% fractionBrij®35/water. The metal salt and the precipitant can be ac-commodated at relatively high concentration in the water cores,with only minimal effects on the structure and phase behav-ior of the microemulsion. Carefully designed SANS and SAXSexperiments showed that the system is only weakly sensitivetoward mono-, di- and trivalent metal salts and high ammonia

Scheme 1. A flow chart outlining the experimental procedure.

A. Bumajdad et al. / Journal of Colloid and Interface Science 312 (2007) 68–75 75

and TBAH levels that are required for nanoparticle formation.The experiments also show that increasing the metal salt con-centration results in a drop in water solubility, however the finallimits are still useable. Unlike AOT-stabilized microemulsions,the single-phase microemulsion regions of the new systems areavailable at around room temperature. Incorporating the Brij®non-ionic offers an important advantage over the pure DDABmicroemulsions in that the single-phase regions are consid-erably enhanced in the mixed surfactant systems: this widerstability window is beneficial for formulating samples contain-ing reagent additives needed to synthesize the nanoparticles.Hence, the proposed microemulsion system is better for thisapplication than both AOT- and DDAB-stabilized microemul-sions. The SANS data show that the reactive microemulsionsystems are stable and not perturbed by the soluble (beforeprecipitation) or insoluble (after precipitation) additives. Usingthese microemulsions as nanostructured reactions media, oxidenanoparticles as small as 50 Å were prepared.

Acknowledgments

The financial support of Kuwait University Research Ad-ministration Grant No. SC02/01 is greatly appreciated. The ex-cellent technical assistance found at the XRD (GS03/01), EMU(GS03/01) and other analytical units of SAF (GS01/01) is pro-foundly acknowledged. The Institute Laue Langevin (France)and CCLRC-ISIS (UK) are thanked for provision of beamtime,travel and consumables grants.

Supplementary material

The online version of this article contains additional supple-mentary material.

Please visit DOI: 10.1016/j.jcis.2006.09.007.

References

[1] K.J. Klabunde (Ed.), Nanoscale Materials in Chemistry, Wiley, New York,2001.

[2] A. Ahniyaz, T. Fujiwara, T. Fujino, T. Fujino, M. Yoshimura, J. Nanosci.Nanotechnol. 4 (2004) 233.

[3] S. Matsumoto, Catal. Today 90 (2004) 183.[4] A. Trovarelli (Ed.), Catalysis by Ceria and Related Materials, Imperial

College Press, London, 2002.[5] Y. He, B. Yang, G. Cheng, J. Mater. Sci. 57 (2003) 1180.[6] S. Rossignol, F. Gérard, D. Duprez, J. Mater. Chem. 9 (1999) 1615.[7] R. Si, Y.-W. Zhang, C.-X. Xiao, S.-J. Li, B.-X. Lin, Y. Kou, C.-H. Yan,

Phys. Chem. Chem. Phys. 6 (2004) 1056.[8] A. Berkó, F. Solymosi, J. Phys. Chem. B 104 (2000) 10215.[9] T.M. Dellinger, P.V. Brawn, Chem. Mater. 16 (2004) 2201.

[10] A. Caragheorgheopol, H. Caldararu, M. Vasilescu, A. Khan, D. Ange-lescu, N. Zilkova, J. Cejka, J. Phys. Chem. B 108 (2004) 7735.

[11] M.P. Pileni, Nature Mater. 2 (2003) 145.[12] I. Capek, Adv. Colloid Interface Sci. 110 (2004) 49.[13] M. Summers, J. Eastoe, S. Davis, Langmuir 18 (2002) 5023.[14] K. Holmberg, J. Colloid Interface Sci. 274 (2004) 355.[15] I. Lisieeki, M.P. Pileni, J. Am. Chem. Soc. 115 (1993) 3887.[16] P.S. Shah, J.D. Holmes, K.P. Johnston, B.A. Korgel, J. Phys. Chem. B 106

(2002) 2545.[17] C. Kitchens, M.C. McLeod, B. Roberts, J. Phys. Chem. B 107 (2003)

11331.[18] K.E. Marchand, M. Tarret, J.P. Lechaire, L. Normand, S. Kasztelan,

T. Cseri, Colloids Surf. A 214 (2003) 239.[19] F. Porta, L. Parti, M. Rossi, G. Scari, Colloids Surf. A 211 (2002)

43.[20] M.L. Curri, A. Agostiano, L. Manna, D.M. Monica, M. Catalano, L. Chi-

avarone, V. Spagnolo, M. Lugarà, J. Phys. Chem. B 104 (2000) 8391.[21] M. Husein, E. Rodil, J. Vera, Langmuir 19 (2003) 8467.[22] T. Charinpanitkul, A. Chanagul, J. Dutta, U. Rungsardthong, W. Tantha-

panichakoon, Sci. Technol. Adv. Mater. 6 (2005) 266.[23] R. Singh, S. Kumar, Chem. Eng. Sci. 61 (2006) 192.[24] R. Loukanov, C.D. Dushkin, K.I. Papazova, A.V. Kirov, M.V. Abrashev,

E. Adachi, Colloids Surf. A 245 (2004) 9.[25] J. Cason, M.E. Miller, J.B. Thompson, C.B. Roberts, J. Phys. Chem. B 105

(2001) 2297.[26] A. Khan-Lodhi, B.H. Robinson, T. Towey, C. Hermann, W. Knoche,

U. Thesing, in: D.M. Bloor (Ed.), The Structure, Dynamics and Equilib-rium Properties of Colloidal Systems, in: NATO ASI Series C, vol. 324,Kluwer, Dordrecht, 1990, p. 373.

[27] M.P. Pileni, Langmuir 13 (1997) 3266.[28] L. Yongcheng, C.-W. Park, Langmuir 19 (1999) 952.[29] J. Eastoe, B.H. Robinson, D.C. Steytler, D. Thorn-Leeson, Adv. Colloid

Interface Sci. 36 (1991) 1.[30] T.K. De, A. Maitra, Adv. Colloid Interface Sci. 59 (1995) 95.[31] A. Bumajdad, J. Eastoe, P. Griffiths, D.C. Steytler, R.K. Heenan, J.R. Lu,

P. Timmins, Langmuir 15 (1999) 5271.[32] A. Bumajdad, J. Eastoe, D.C. Steytler, R.K. Heenan, P. Timmins, Lang-

muir 19 (2003) 2560.[33] A. Bumajdad, M.I. Zaki, J. Eastoe, L. Pasupulety, Langmuir 20 (2004)

11223.[34] S. Eriksson, U. Nylén, S. Rojas, M. Boutonnet, Appl. Catal. A 265 (2004)

207.[35] A. Bumajdad, J. Eastoe, J. Colloid Interface Sci. 1 (2004) 268.[36] S. Nave, J. Eastoe, R.K. Heenan, D. Steytler, I. Grillo, Langmuir 16 (2000)

8741.[37] H. Kunieda, K. Shinoda, J. Colloid Interface Sci. 75 (1980) 601;

M. Kolarchyk, Phys. Rev. A 29 (1984) 2054;S.H. Chen, J. Chem. Phys. 93 (1990) 1907;S.H. Chen, S.-L. Chang, R. Stray, J. Chem. Phys. 93 (1990) 1907.

[38] See for example: A. Gunier, G. Fournet, in: O. Glatter, O. Kratky (Eds.),Small Angle Scattering of X-Ray, Wiley, New York, 1955;O. Glatter, O. Kratky (Eds.), Small Angle X-Ray Scattering, AcademicPress, New York, 1982.

[39] International Center for Diffraction Data, 12 Campus Boulevard, NewtonSquare, PA 19073-3273, USA.

[40] V. Chhabra, P. Ayyub, S. Chattopadhyay, A.N. Maitra, Mater. Lett. 26(1996) 21.

[41] W. Zhonghua, G. Lin, L. Qianshu, Z. Hesun, J. Phys. Condens. Matter 11(1999) 4961.

[42] R.J. Matyi, L.H. Schwartz, J.B. Butt, Catal. Rev. Sci. Eng. 29 (1987)41.