Phase, size and shape controlled formation of aerosol generated nickel and nickel oxide nanoparticles D. Ortega a,b,⇑ , M.V. Kuznetsov c , Yu.G. Morozov d , O.V. Belousova d , I.P. Parkin e a Department of Physics and Astronomy, University College London, WC1E 6BT London, UK b London Centre for Nanotechnology, Gordon Street, WC1H 0AH London, UK c N.P. Ogarev Mordovian State University, Saransk, Republic of Mordovia 430005, Russia d Institute of Structural Macrokinetics and Materials Science, Chernogolovka, Moscow Region 142432, Russia e Department of Chemistry, Materials Chemistry Centre, University College London, 20 Gordon Street, WC1H 0AJ London, UK article info Article history: Received 1 April 2013 Received in revised form 3 June 2013 Accepted 20 June 2013 Available online 29 June 2013 Keywords: Nickel nanoparticles Levitation-jet aerosol synthesis Coalescence Magnetic properties abstract Ferromagnetic Ni nanoparticles were formed by a levitation-jet aerosol synthesis under different gas environments and metal precursor feed rates. At a constant background gas inlet temperature, it was found that a higher Ni loading resulted in enhanced particle growth through coalescence. He partial atmosphere favors surface condensation of evaporated Ni atoms over coalescence as the surface area reduction mechanism in the nanoparticles. A flow of 2.5% air in the background gas mixture was enough to oxidize 75% of the initial Ni load, inducing a drastic destabilization of particle size and shape distribu- tion. Regardless of the background inert gas composition, necked nanoparticles were observed in samples prepared with a 1 g/h Ni feed rate, whereas discrete nanoparticles resulted from a higher feed rate of ca. 4 g/h, confirming the key role of Ni loading on the rate of coalescence. The highest saturation magnetiza- tion (51.75 A m 2 kg 1 measured at 300 K) and the lowest coercivity (0.008 T) were obtained under an Ar flow. Zero-field cooled and field-cooled magnetization curves measured under an applied field of 10 2 T revealed that the blocking processes of nanoparticles are dominated by their particle size distributions, with some features attributable to interparticle interactions. Ó 2013 Elsevier B.V. All rights reserved. 1. Introduction The interest in nickel nanoparticles research mainly comes from their use in catalysis [1–3], although other applications such as en- ergy storage [4] ferrofluids [5], electroanalysis [6], and more re- cently biology and medicine [7–10], have also been important. Compared to the other ferromagnetic transition metals, a signifi- cantly lower number of reports dealing with the synthesis and physicochemical characterization of Ni nanoparticles have been published. Microwave-assisted synthesis, combustion, pulsed laser ablation, vacuum evaporation or chemical precipitation are, among others, some of the methods reported for preparing Ni nanoparti- cles [11–15]. Within these methods, some make use of surfactants for different purposes, such as preparing core–shell structures [16], controlling size and shape distributions [17] or surface functional- ization [18]. Despite these achievements, one of the weak points of these methods is the relatively low saturation magnetization (r s ) of the final nanoparticles when compared to that of the bulk nickel [19–28]. This feature is central for some novel applications like magnetically recoverable Ni catalysts for hydrogenation reactions [29], which aims to reduce the costs of metal-based catalysts and eliminate the leaching effect associated with post-reaction filtra- tion procedures by using magnetic filtration. Apart from a high sat- uration magnetization, both a low coercivity and remanence are also convenient in order to minimize the applied field needed to demagnetize the system and prevent particle agglomeration upon removal of the field. Saturation magnetization, as an intrinsic magnetic property, is structure insensitive and only depends on the magnitude of the atomic magnetic moments and the number of atoms per unit vol- ume. The most common reason for the reported low r s values in metallic (Ni, Fe, Co) nanoparticles is the occurrence of uncontrolled oxidation processes during synthesis, typically leading to the pres- ence of antiferromagnetic oxides, such as NiO, CoO or a-Fe 2 O 3 . In the case of Ni, this is a major drawback stemming from its kineti- cally favoured oxidation [30], which seems to be inherently associ- ated with many deposition methods [6]. In wet-chemical preparations, this issue has been addressed through the incorpora- tion of surfactants or coating agents in the reaction medium at some point [31], but this route is mainly suited for applications in liquid media and often entails a certain performance reduction of the as-synthesized nanoparticles. Furthermore, complete 0925-8388/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jallcom.2013.06.128 ⇑ Corresponding author at: Department of Physics and Astronomy, University College London, WC1E 6BT London, UK. E-mail address: [email protected](D. Ortega). Journal of Alloys and Compounds 579 (2013) 495–501 Contents lists available at SciVerse ScienceDirect Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom
Transcript
Journal of Alloys and Compounds 579 (2013) 495–501
Contents lists available at SciVerse ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier .com/locate / ja lcom
Phase, size and shape controlled formation of aerosol generated nickeland nickel oxide nanoparticles
0925-8388/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.jallcom.2013.06.128
⇑ Corresponding author at: Department of Physics and Astronomy, UniversityCollege London, WC1E 6BT London, UK.
D. Ortega a,b,⇑, M.V. Kuznetsov c, Yu.G. Morozov d, O.V. Belousova d, I.P. Parkin e
a Department of Physics and Astronomy, University College London, WC1E 6BT London, UKb London Centre for Nanotechnology, Gordon Street, WC1H 0AH London, UKc N.P. Ogarev Mordovian State University, Saransk, Republic of Mordovia 430005, Russiad Institute of Structural Macrokinetics and Materials Science, Chernogolovka, Moscow Region 142432, Russiae Department of Chemistry, Materials Chemistry Centre, University College London, 20 Gordon Street, WC1H 0AJ London, UK
a r t i c l e i n f o
Article history:Received 1 April 2013Received in revised form 3 June 2013Accepted 20 June 2013Available online 29 June 2013
Ferromagnetic Ni nanoparticles were formed by a levitation-jet aerosol synthesis under different gasenvironments and metal precursor feed rates. At a constant background gas inlet temperature, it wasfound that a higher Ni loading resulted in enhanced particle growth through coalescence. He partialatmosphere favors surface condensation of evaporated Ni atoms over coalescence as the surface areareduction mechanism in the nanoparticles. A flow of 2.5% air in the background gas mixture was enoughto oxidize 75% of the initial Ni load, inducing a drastic destabilization of particle size and shape distribu-tion. Regardless of the background inert gas composition, necked nanoparticles were observed in samplesprepared with a 1 g/h Ni feed rate, whereas discrete nanoparticles resulted from a higher feed rate of ca.4 g/h, confirming the key role of Ni loading on the rate of coalescence. The highest saturation magnetiza-tion (51.75 A m2 kg�1 measured at 300 K) and the lowest coercivity (0.008 T) were obtained under an Arflow. Zero-field cooled and field-cooled magnetization curves measured under an applied field of 10�2 Trevealed that the blocking processes of nanoparticles are dominated by their particle size distributions,with some features attributable to interparticle interactions.
� 2013 Elsevier B.V. All rights reserved.
1. Introduction
The interest in nickel nanoparticles research mainly comes fromtheir use in catalysis [1–3], although other applications such as en-ergy storage [4] ferrofluids [5], electroanalysis [6], and more re-cently biology and medicine [7–10], have also been important.Compared to the other ferromagnetic transition metals, a signifi-cantly lower number of reports dealing with the synthesis andphysicochemical characterization of Ni nanoparticles have beenpublished. Microwave-assisted synthesis, combustion, pulsed laserablation, vacuum evaporation or chemical precipitation are, amongothers, some of the methods reported for preparing Ni nanoparti-cles [11–15]. Within these methods, some make use of surfactantsfor different purposes, such as preparing core–shell structures [16],controlling size and shape distributions [17] or surface functional-ization [18]. Despite these achievements, one of the weak points ofthese methods is the relatively low saturation magnetization (rs)of the final nanoparticles when compared to that of the bulk nickel[19–28]. This feature is central for some novel applications like
magnetically recoverable Ni catalysts for hydrogenation reactions[29], which aims to reduce the costs of metal-based catalysts andeliminate the leaching effect associated with post-reaction filtra-tion procedures by using magnetic filtration. Apart from a high sat-uration magnetization, both a low coercivity and remanence arealso convenient in order to minimize the applied field needed todemagnetize the system and prevent particle agglomeration uponremoval of the field.
Saturation magnetization, as an intrinsic magnetic property, isstructure insensitive and only depends on the magnitude of theatomic magnetic moments and the number of atoms per unit vol-ume. The most common reason for the reported low rs values inmetallic (Ni, Fe, Co) nanoparticles is the occurrence of uncontrolledoxidation processes during synthesis, typically leading to the pres-ence of antiferromagnetic oxides, such as NiO, CoO or a-Fe2O3. Inthe case of Ni, this is a major drawback stemming from its kineti-cally favoured oxidation [30], which seems to be inherently associ-ated with many deposition methods [6]. In wet-chemicalpreparations, this issue has been addressed through the incorpora-tion of surfactants or coating agents in the reaction medium atsome point [31], but this route is mainly suited for applicationsin liquid media and often entails a certain performance reductionof the as-synthesized nanoparticles. Furthermore, complete
496 D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501
removal of surfactants is not always possible and involves addi-tional processing steps [6]. There is then a clear need for a cost-effi-cient and scalable technique suitable for producing Ninanoparticles of controllable size and shape while keeping theirrS as close as possible to the bulk material.
Levitation-jet aerosol synthesis under an inert gas flow allowsfor simultaneous manipulation of the shape and composition ofnanoparticles without the need of further surfactants or cappingagents, minimizing the amount of precursors used and generallyallowing for a large-scale production of high-purity nanopowders.An important advantage of this technique is that the metal precur-sor load is melted into a droplet and levitated by the action of ahigh-frequency (HF) field in an inert gas flow, precluding any con-tact with surfaces and hence minimizing nanoparticle formationover them. This method is particularly attractive for synthesisingstable metallic nanoparticles due to the close control that can beexerted over oxidation processes by adequately manipulating therelevant experimental conditions. Moreover, the reagents involvedare kept to a minimum, hence reducing both the chances of sidereactions and the necessity of further post-processing stages.
In the present work we report on the successful application ofthe levitation-jet aerosol synthesis under an inert gas flow for pre-paring ferromagnetic Ni nanoparticles of various shapes and sizes.The influence of some crucial experimental parameters on the finalstructural and magnetic properties of the nanoparticles is demon-strated. The latter comprise the composition and flow of the inertgas mixture, as well as the Ni precursor feed rate into the system.
2. Materials and methods
A series of samples were prepared at room temperature by a crucibleless levi-tation-jet aerosol method in the presence of a gaseous mixture with variable com-position. The set of synthetic conditions are summarized in Table 1. A metal dropletlevitated inside a thin quartz tube (HF electromagnetic field levitation) and blownup by a gas stream is heated up to melting and vaporization by the action of thefield. The laminar gas stream consisted of either Ar (99.987 at.%), He (99.99 at.%),He/air or Ar/air mixtures each supplied at the flow rates indicated in Table 1. Thecase of an Ar/He mixture is not considered here as it has been previously studiedelsewhere [32]. The formation of aerosol nanoparticles takes place in a regionaround the droplet, where atoms from the evaporated precursors condense andeventually form clusters. In order to sustain the nanoparticle production, the Nidroplet is fed at a constant rate with the metal precursor (Ni, 99.9 at.%) in the formof a 0.2 mm diameter wire.
Crystal structure and phase composition of samples were determined by X-raydiffraction in a DRON-3M diffractometer using Cu Ka radiation. Powder DiffractionFile (PCPDFWIN ver. 2.02) database was used as the input for CrystallographicaSearch-Match and PowderCell for Windows programs to perform the XRD phase
Table 1Gas flow and Ni feeding rate used during the synthesis of the samples.
Table 2Main properties of the Ni nanoparticles prepared under different conditions. hdi is the adiffractograms, VXRD the at.% Ni calculated from XRD diffractograms, rs the saturation magnthe remanent magnetization, HC the coercive field and S the specific surface area.
a The term crystallite corresponds here to the size of coherently scattering domains.
analysis. Rietveld analysis of the profiles allowed for determining the Ni/NiO ratioin the nanoparticle powders. Particle size and morphology were studied by trans-mission electron microscopy (TEM) using a JEOL JEM 1200EX II electron microscopeoperated at 120 kV. Magnetization measurements at 300 K were made in a 7 TQuantum Design hybrid VSM-SQUID magnetometer, where the samples weremounted into polypropylene sample holders. IUPAC recommendations on reportingand interpreting magnetic properties were followed [33]. A SORBI-M META devicewas used to obtain the specific surface area (BET) of the nanoparticles from 4-pointmeasurements of nitrogen physisorption.
3. Results
Table 2 summarizes the main parameters extracted from thestructural and magnetic characterization. Room temperature XRDdiffractograms (Fig. 1) indicate that Ni is present in samples S1to S4, while NiO appears only in samples (S4 and S5) prepared inthe presence of a variable amount of air. No other phases or impu-rities were detected. A peak broadening effect is observed for sam-ples S1 and S5, those with a smaller crystallite size hdXi (Table 2).
As seen in TEM images, sample S1 forms a fractal-like arrange-ment of spheroidal nanoparticles (Fig. 2a; under He and slow Nifeeding), some of which have partially coalesced (Fig. 2b). This par-ticular arrangement is absent in sample S2 (Fig. 2c; He and fast Nifeeding), where the predominant morphology has now changed tohexagonal (Fig. 2d), even more pronounced than that reported forother Ni nanoparticle systems [34]. A higher Ni feed rate increasesthe average particle size and broadens the size distribution (Fig. 3aand b), furthermore it notably increases the specific surface area (S)compared to sample S1, producing a small increase in the metallicNi content per sample (Table 2).
The changes introduced by switching to an Ar atmosphere arequite remarkable; the average particle size in sample S3 is biggerthan those obtained with a He atmosphere and the size distribu-
verage particle size obtained from TEM images, hdXi the crystallitea size from XRDetization, VMAG the at.% Ni calculated from magnetometry measurements (see [36]) rr
Fig. 1. Room temperature XRD patterns of samples series S1 to S5 indicating thetypical peaks of Ni (gray dots, JCPDS card No. 04-0850) and NiO (black dots, JCPDScard No. 73-1519).
Fig. 2. Conventional TEM images of samples S1 (a and b), S2 (c and d), S3 (e and f),S4 (g and h) and S5 (i and j).
Fig. 3. Size distributions of sample S1 (a) and S2 (b) extracted from thecorresponding TEM images, showing the fit to a lognormal distribution function.
Fig. 4. Hysteresis loops at 300 K for the studied sample series. Inset: zoom of thelow field region for sample S5, revealing the presence of a small coercive field.
D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501 497
tion broadens (Fig. 2e and f). Another interesting feature observedin the TEM images of sample S3 is the reappearance of coalescence,evidenced by the occurrence of necks between neighboring parti-cles (arrows in Figs. e and f). If a small amount of air is incorporated
into the initial gas flow keeping the Ni feed rate constant, a sub-stantial change in morphology is further induced, giving rise to amixture of irregular shapes (Fig. 2g and h). A partial oxidation oc-curs in these conditions (Fig. 1, sample S4), reducing the Ni contentalmost by a factor of four. A gas mixture with a much higher pro-portion of air gives rise to the coexistence of two morphologies;cubes and truncated cubes (Fig. 2i and j). The corresponding
Fig. 5. Detail of the low field region of the hysteresis loops for all samples at 300 K.
498 D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501
X-ray diffractogram clearly shows that the resulting nanoparticlesare almost fully oxidized (Fig. 1, sample S5).
Magnetic characterization revealed that all the samples S1–S4show a ferromagnetic-like hysteresis (Fig. 4). Saturation magneti-zation values at 300 K increase with the Ni content in samples S1to S3 (Table 2), approaching that of bulk Ni at 300 K (54.8 A m2 -kg�1) [35]. A different trend is observed for coercivity (HC) andremanence (rr) (Fig. 5); both parameters increase throughout thesequence S3–S1–S2. For those synthesized in an Ar atmosphere, ahigher coercivity is found in sample S4, while a higher remanenceis shown by sample S3. Sample S5 shows a practically anhystereticmagnetization curve at 300 K.
Zero-field cooled and field cooled (ZFC/FC) curves under an ap-plied field of 10�2 T show irreversibility at around 300 K in all cases(Fig. 6a). The shape of the ZFC branches of samples S1 and S2 in thelow temperature range (5–100 K) are quite similar, showing asteep slope at the beginning followed by an increase at a constantrate towards room temperature, resulting in a straight line from 75
to100 K onwards. It has to be noted that no well-resolved temper-ature peak or even a temperature local maximum is observed inany of the ZFC curves. In sample S3 the slope increase is muchsmoother and it extends throughout the whole temperature rangerather than just at the lower end. Finally, sample S4 shows an al-most linear ZFC branch and, opposite to the rest of the series, theFC branch increases beyond the irreversibility point.
Finally, the specific surface area S is similar in samples S1 andS2 regardless of their different phase composition (Table 2), whilelower values are observed in samples S3 and S4, with a much largeraverage particle size. As expected, the sample with the smallest hdi(S5) possesses the highest S.
4. Discussion
Altering the Ni feed rate has a large impact in the structural andmagnetic properties of the nanoparticles prepared by this levita-tion technique. The most noticeable consequences of increasingthe feeding rate are the increase in the average particle diameter[32] and its contribution to a more pronounced hexagonal mor-phology (Fig. 2c and d). In order to explain these observations,we consider the establishment of a nucleation and growth regionor layer located at a short distance from the Ni evaporation region– the condensation zone [37]. The factors influencing the thicknessof this layer will be discussed later. In principle, increasing the Nisupply to the system would entail the rapid formation of a highernumber of new nuclei as the metal aerosol is being cooled by thegas flow. If this simple hypothesis is accepted, a higher Ni feed rateshould then lead to samples composed of a larger number of smallparticles, since the equilibrium between nucleation and growthwould be heavily shifted. Nevertheless, the opposite situation isexperimentally observed here: the number of particles per unitarea is lower and the average size is visibly larger. This picture ofthe system is more consistent with the existence of a certain crit-ical number of nuclei which, once reached, increases the probabil-ity of particle growth processes to the detriment of nucleiformation events, since the collisions of Ni atoms with the gas willbe less likely each time due to the increasing presence of particles
D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501 499
in the nucleation/growth layer [38]. Parallel to metal-gas interac-tions, interparticle interactions also contributes to the growth pro-cess, as exemplified by the remarkable differences in size andshape between samples S1 and S2. A Ni feed rate of 1 g/h producesan uncompleted coalescence of particles (Fig. 2a and b), whereasloose particles are observed for an increased rate of 3.8 g/h(Fig. 2c and d); in other words, a higher Ni volume loading favorscollision–coalescence kinetics. This experimentally confirms theresults obtained from a kinetic Monte Carlo model of coalescencewithout constraints on the particle size distribution [39], whereit was clearly shown that under constant background gas temper-ature the coalescence process can be enhanced by increasing thecollision rate through larger volume loading when the collisiontime is considerably higher than the coalescence time. In our case,the disappearance of coalesced particles is accompanied by anaverage size increase from S1 to S2 for higher feeding rates (Ta-ble 2), and since a higher precursor loading result in enhancedgrowth rates [39], it is reasonable to point out that coalescence isthe main growth mechanism involved in the formation of the hex-agonal Ni particles. If the background gas composition is changed,its temperature kept constant and the Ni feeding rate is set back to1 g/h, coalescing nanoparticles are obtained again, as shown by theTEM pictures of sample S3. This result confirms that coalescencerate can be modified by adjusting the Ni supply to the system.
A higher feed rate has a twofold effect on the shape of the hys-teresis loop: a ‘‘sharpening’’ effect (Fig. 5, samples S1 and S2) interms of the higher rS and smaller rr, and a widening effect interms of the higher HC. This is intimately related to changes in bothinterparticle interactions and particle size. Considering the dimen-sionless hardness parameter j (0.13), exchange length lex (5.1 nm)and anisotropy constant K (5 J m�3) for Ni [35], the resulting valuesfor its theoretical superparamagnetic size limit Dspm and single-do-main equilibrium diameter Dsd are 34 and 48 nm, respectively. Fol-lowing the typical evolution of coercivity with particle size [40,41],in which HC ? 0 for particles with Dsd < hdi, peaks at Dsd � hdi anddecreases again for Dsd > hdi. In the present study, sample S2 isclose to the Ni Dsd, which is in agreement with its coercivity value,the highest within the series. As expected, sample S1 shows a low-er HC as the average diameter of the particles is below both Dspm
and Dsd. This is related to the different shape of the correspondingZFC curves (Fig. 6a and b), where the magnetization inflectionpoint occurs at a lower temperature in sample S1 because the ear-lier unblocking process associated to their smaller particles with anarrower size distribution compared to sample S2. Since hdi forsamples S3 and S4 are over both Dspm and Dsd, it is expected thatmost of the particles in these samples are multidomain and hencealready blocked either before or around room temperature. Theshape of the ZFC branch for both samples provides additional infor-mation on this matter. This is partially true for sample S3 (Fig. 6c);its unblocking process spans almost over the entire measured tem-perature range as a result of the much wider size distribution,which demonstrates that there is a fraction of superparamagneticparticles. The continuously increasing magnetization and the irre-versibility up to room temperature is the obvious result of the frac-tion of bigger particles observed in TEM images (Fig. 2e). Sample S4magnetization shows a linear temperature dependence in the ZFCbranch, whereas the magnetization increase in the FC branch be-yond the irreversibility point is indicative of less effective interpar-ticle interactions [42] in comparison to the other samples. Thehysteresis measurements, in the low field range of sample S5 isreminiscent of non-compensated antiferromagnets showing para-sitic ferromagnetism or spin-canting of antiparallel magnetic mo-ments [43], where a small ferromagnetic signal is superimposedto a dominant paramagnetic one in the low-field region (Fig. S1,Supplementary information). This feature could be also ascribedto the onset of a superparamagnetic regime due to uncompensated
surface spins [44] in particles at the lower end of the experimentalsize distribution, which would agree with the superparamagneticbehavior reported for NiO nanoparticles below 100 nm, with amagnetic moment inversely proportional to the particle size [45].Nonetheless, two facts lead us to point out a different cause: (i)the presence of metallic Ni traces randomly distributed over thesample (Table 2) may mimic the response of canted antiferromag-nets and (ii) much higher magnetization values would be expectedfor superparamagnetic NiO nanoparticles given that the number ofsurface moments contributes to the total magnetization. We there-fore conclude that a ‘‘diluted ferromagnetism’’ corresponding tothe Ni traces is the most likely explanation for the magnetizationcurve of sample S5.
Other competing factors involved in shaping the hysteresisloops are the compensation of defects in the particles by increasingthe available Ni concentration during synthesis and shape anisot-ropy effects. Decreasing the number of defects homogenizes theexchange interactions within the nanoparticles and also contrib-utes to reduce surface effects due to the loss of translational sym-metry of superficial atoms. Also, the slightly higher rs in sample S2would be in agreement with Ni defect compensation, as it leads toa higher number of magnetic moments per unit mass. This mech-anism is possible provided that the early nuclei spend enough timeat the growth layer before migrating to the cooling zone before col-lection. The previously discussed particle growth associated to ahigher Ni feeding into the system reasonably supports it. In rela-tion to shape effects, a more symmetrical spatial arrangement ofparticles would be needed for shape anisotropy to properly showup and hence be clearly distinguished from other contributionsto the total anisotropy energy [46]. Moreover, the random orienta-tion of the nanoparticles in the studied samples averages any pos-sible modification on their magnetization reversal mechanism dueto shape anisotropy changes.
Another factor of major influence on the physicochemical prop-erties of the nanoparticles is the choice of the gas mixture. The mainexplanation for the difference between the physical properties ofthe nanoparticles produced under either an Ar or He atmospherehas to be found in the much higher density and viscosity of Ar com-pared to He, resulting in a more efficient cooling of the evaporatedNi atoms due to the shorter mean free path associated to Ar by vir-tue of its larger molecular mass [32,38]. In addition to the gas com-position, the gas flow is also a factor influencing the properties of thenanoparticles, especially their average size and size distribution[38]. The processes of nucleation and growth of particles take placeclose to the liquid Ni droplet throughout a layer or region of variablethickness determined by either the availability of vaporized metalatoms or the temperature of the metal aerosol itself [47]. Underthe gas pressure conditions used here, this layer is approx. 1 mmthick and it is located 1 mm apart from the droplet [37]. On onehand, the residence time within that layer will dictate the probabil-ity of effective collisions and consecutive deposition of metal atomson the surface of the already formed nuclei and hence the averageparticle size, which is greater for longer residence times. On theother hand, when the growth layer is controlled by the metal aero-sol temperature, which in turn depends on the gas flow, the appear-ance of particle coalescence as the reigning growth mechanism isfavoured [47]. By comparing the pseudo-hexagonal crystal habitof particles in sample S1 and the spherical shape observed in sampleS3, it becomes apparent that in the former case He favors surfacecondensation of evaporated atoms over coalescence as the surfacearea reduction mechanism. This is confirmed by the well-definedfaceting obtained in sample S2 by just changing the feeding ratewhile keeping the same atmosphere. Although the establishmentof a more or less broad statistical size distribution may be caused,in principle, by a non-homogeneous temperature profile acrossthe path travelled by the particles towards the collector, it can
500 D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501
also be rationalized as the result of a set of discrete coalescenceevents influenced by the cooling of the aerosol particles.
As expected, the presence of air in the gas mixture (S4 and S5)leads to the oxidation of Ni to an extent determined by the per-centage of air allowed into the system. The small air flow used insample S4 is enough to induce dramatic changes in the particleshape. If the oxidation were to take place solely at the condensa-tion stage, the resulting product would consist either in a mixtureof well-defined Ni and NiO nanoparticles or core–shell particles (Nicore + NiO shell) due to a selective oxidation of their surface. Sincethis picture of the system does not match the experimental results,the oxidation should then occur at an earlier stage, eventually lead-ing to a non-homogeneous condensation of partially oxidizedcores. As the nickel adsorption and oxygen chemisorption proceedat the surface of the condensed cores, the lattice mismatch be-tween Ni and NiO is thought to be responsible for the randomshapes observed in TEM (Fig. 2g and h).
5. Conclusions
The changes introduced by different gas mixtures and precursorfeeding rate in the preparation of ferromagnetic Ni nanoparticlesthrough a crucibleless levitation-jet aerosol method are reported.The preparation of stable Ni nanoparticles through this method isstraightforward under appropriate conditions without the needof any surfactant or protective coating during or after the synthesisprocess.
In general terms, either a He or Ar atmosphere prevents a pre-mature oxidation of the nanoparticles, but changes in size, shapeand phase composition are observed depending on the choice ofatmosphere. Specifically, larger spherical particles with the highestNi content are produced by using Ar. Under the same experimentalconditions, increasing the Ni feed rate increases the average parti-cle size and a slightly higher saturation magnetization is measuredin the product, likely due to defect compensation. In addition, amorphological change from spheroidal to hexagonal is observedas the primary particles coalescence comes to completion, allowingthe development of a defined crystal habit. It was also observedthat, regardless of the composition of the background gas and ata constant temperature, the coalescence rate of the primary parti-cles is accelerated by increasing the Ni feeding rate. Once the coa-lescence between particles is completed, minimization of surfaceenergy favors the hexagonal form as the equilibrium morphology.The results suggest that surface condensation of evaporated Niatoms is favoured over coalescence as the particle growth mecha-nism by using a pure He background gas. Nanoparticles producedunder an Ar atmosphere showed the highest rs – close to the valuefor bulk nickel – and the lowest Hc over the series, with a lower rr
than those produced under a He flow. This is ascribed to both theaverage particle size – above Dsd – and presumably the highernumber of defects present in the latter samples, which providesa mechanism for Hc and facilitates the pinning of magnetic mo-ments. The presence of air in the gas stream leads to oxidation ofparticles to an extent determined by the amount of air allowed intothe system. Remarkably, by introducing a 2.5 at.% of air in thebackground gas mixture, about 75 at.% of the initial Ni load oxi-dizes during preparation. A higher oxidation degree is accompa-nied by a drastic drop in particle size, and the associated shapechanges are also remarkable, going from hexagonal/spheroidal tocubic/truncated cubic.
Acknowledgement
D.O. thanks the support received from University CollegeLondon under the Provost’s Strategic Fund.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jallcom.2013.06.128.
References
[1] I. Vicente, P. Salagre, Y. Cesteros, Ni nanoparticles supported on microwave-synthesised hectorite for the hydrogenation of styrene oxide, Appl. Catal. A 408(2011) 31–37.
[2] I. Vicente, P. Salagre, Y. Cesteros, Ni nanoparticles supported on microwave-synthesised saponite for the hydrogenation of styrene oxide, Appl. Clay Sci. 53(2011) 212–219.
[3] J. Cornella, E. Gómez-Bengoa, R. Martin, Combined experimental andtheoretical study on the reductive cleavage of inert C–O bonds with silanes:ruling out a classical Ni(0)/Ni(II) catalytic couple and evidence for Ni(I)intermediates, J. Am. Chem. Soc. 135 (2013) 1997–2009.
[4] L. Wang, R.T. Yang, Hydrogen storage properties of carbons doped withruthenium, platinum, and nickel nanoparticles, J. Phys. Chem. C 112 (2008)12486–12494.
[5] S.P. Gubin, Introduction, in: S.P. Gubin (Ed.), Magnetic Nanoparticles, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2009, pp. 1–23.
[6] C.M. Welch, R.G. Compton, The use of nanoparticles in electroanalysis: areview, Anal. Bioanal. Chem. 384 (2006) 601–619.
[7] M.J. Hajipour, K.M. Fromm, A. Akbar Ashkarran, D. Jimenez de Aberasturi, I.R.d.Larramendi, T. Rojo, V. Serpooshan, W.J. Parak, M. Mahmoudi, Antibacterialproperties of nanoparticles, Trends Biotechnol. 30 (2012) 499–511.
[8] R. Magaye, J. Zhao, Recent progress in studies of metallic nickel and nickel-based nanoparticles’ genotoxicity and carcinogenicity, Environ. Toxicol.Pharmacol. 34 (2012) 644–650.
[9] D. Guo, C. Wu, J. Li, A. Guo, Q. Li, H. Jiang, B. Chen, X. Wang, Synergistic effect offunctionalized nickel nanoparticles and quercetin on inhibition of the SMMC-7721 cells proliferation, Nanosc. Res. Lett. 4 (2009) 1395–1402.
[10] W. Zhang, Y. Li, J. Niu, Y. Chen, Photogeneration of reactive oxygen species onuncoated silver, gold, nickel, and silicon nanoparticles and their antibacterialeffects, Langmuir 29 (2013) 4647–4651.
[11] C. Parada, E. Moran, Microwave-assisted synthesis and magnetic study ofnanosized Ni/NiO materials, Chem. Mater. 18 (2006) 2719–2725.
[12] N.M. Deraz, Magnetic behavior and physicochemical properties of Ni and NiOnano-particles, Curr. Appl. Phys. 12 (2012) 928–934.
[13] T. Seto, H. Akinaga, F. Takano, K. Koga, T. Orii, M. Hirasawa, Magneticproperties of monodispersed Ni/NiO core–shell nanoparticles, J. Phys. Chem. B109 (2005) 13403–13405.
[14] S. Sako, K. Ohshima, M. Sakai, S. Bandow, Magnetic property of NiO ultrafineparticles with a small Ni core, J. Vac. Sci. Technol. B 15 (1997) 1338–1342.
[15] A. Anspoks, A. Kalinko, R. Kalendarev, A. Kuzmin, Atomic structure relaxationin nanocrystalline NiO studied by EXAFS spectroscopy: Role of nickelvacancies, Phys. Rev. B 86 (2012) 174114.
[16] I.S. Lee, N. Lee, J. Park, B.H. Kim, Y.-W. Yi, T. Kim, T.K. Kim, I.H. Lee, S.R. Paik, T.Hyeon, Ni/NiO core/shell nanoparticles for selective binding and magneticseparation of histidine-tagged proteins, J. Am. Chem. Soc. 128 (2006) 10658–10659.
[17] N. Cordente, M. Respaud, F. Senocq, M.-J. Casanove, C. Amiens, B. Chaudret,Synthesis and magnetic properties of nickel nanorods, Nano Lett. 1 (2001)565–568.
[18] D. Guo, C. Wu, X. Li, H. Jiang, X. Wang, B. Chen, In vitro cellular uptake andcytotoxic effect of functionalized nickel nanoparticles on leukemia cancercells, J. Nanosci. Nanotechnol. 8 (2008) 2301–2307.
[19] Y.L. Hou, S. Gao, Monodisperse nickel nanoparticles prepared from amonosurfactant system and their magnetic properties, J. Mater. Chem. 13(2003) 1510–1512.
[20] J. Park, E. Kang, S.U. Son, H.M. Park, M.K. Lee, J. Kim, K.W. Kim, H.J. Noh, J.H.Park, C.J. Bae, J.G. Park, T. Hyeon, Monodisperse nanoparticles of Ni and NiO:Synthesis, characterization, self-assembled superlattices, and catalyticapplications in the Suzuki coupling reaction, Adv. Mater. 17 (2005) 429–434.
[21] C. Estournés, T. Lutz, J. Happich, T. Quaranta, P. Wissler, J.L. Guille, Nickelnanoparticles in silica gel: preparation and magnetic properties, J. Magn.Magn. Mater. 173 (1997) 83–92.
[22] S.H. Wu, D.H. Chen, Synthesis and characterization of nickel nanoparticles byhydrazine reduction in ethylene glycol, J. Colloid Interface Sci. 259 (2003) 282–286.
[23] D.H. Chen, C.H. Hsieh, Synthesis of nickel nanoparticles in aqueous cationicsurfactant solutions, J. Mater. Chem. 12 (2002) 2412–2415.
[24] D.H. Chen, S.H. Wu, Synthesis of nickel nanoparticles in water-in-oilmicroemulsions, Chem. Mater. 12 (2000) 1354–1360.
[25] S. Li, M. Liu, H. Bi, L. Lu, W. Zou, Z. Huang, Y. Du, The enhanced magneticproperties of Ni nanocrystallites induced by the coupling between Ni/NiOinterfaces, J. Alloys Comp. 425 (2006) 1–3.
[26] M.P. Fernandez-Garcia, P. Gorria, M. Sevilla, M.P. Proença, R. Boada, J.s. Chaboy,A.B. Fuertes, J.s.A. Blanco, Enhanced protection of carbon-encapsulatedmagnetic nickel nanoparticles through a sucrose-based synthetic strategy, J.Phys. Chem. C 115 (2011) 5294–5300.
D. Ortega et al. / Journal of Alloys and Compounds 579 (2013) 495–501 501
[27] Y.a. Huang, Z. Xu, Y. Yang, T. Tang, R. Huang, J. Shen, Preparation,characterization, and surface modification of carbon-encapsulated nickelnanoparticles, J. Phys. Chem. C 113 (2009) 6533–6538.
[28] D.S. Sidhaye, T. Bala, S. Srinath, H. Srikanth, P. Poddar, M. Sastry, B.L.V. Prasad,Preparation of nearly monodisperse nickel nanoparticles by a facile solutionbased methodology and their ordered assemblies, J. Phys. Chem. C 113 (2009)3426–3429.
[29] V. Polshettiwar, B. Baruwati, R.S. Varma, Nanoparticle-supported andmagnetically recoverable nickel catalyst: a robust and economichydrogenation and transfer hydrogenation protocol, Green Chem. 11 (2009)127–131.
[30] P. Song, D. Wen, Z.X. Guo, T. Korakianitis, Oxidation investigation of nickelnanoparticles, PCCP 10 (2008) 5057–5065.
[32] T.A. Kondrat’eva, Y.G. Morozov, E.A. Chernov, Effect of conditions ofmanufacture on the properties of ultrafine nickel powder, Powder Metall.Metal Ceram. 26 (1987) 793–795.
[33] S. Hatscher, H. Schilder, H. Lueken, W. Urland, Practical guide to measurementand interpretation of magnetic properties: IUPAC technical report, Pure Appl.Chem. 77 (2005) 497–509.
[34] Y.T. Jeon, J.Y. Moon, G.H. Lee, J. Park, Y. Chang, Comparison of the magneticproperties of metastable hexagonal close-packed Ni nanoparticles with thoseof the stable face-centered cubic Ni nanoparticles, J. Phys. Chem. B 110 (2005)1187–1191.
[35] J.M.D. Coey, Magnetism and Magnetic Materials, Cambridge University Press,2010.
[36] Y.G. Morozov, O.V. Belousova, M.V. Kuznetsov, D. Ortega, I.P. Parkin, Electricfield-assisted levitation-jet aerosol synthesis of Ni/NiO nanoparticles, J. Mater.Chem. 22 (2012) 11214–11223.
[37] M.Y. Gen, M.S. Ziskin, Y.I. Petrov, Study of dispersion of Al aerosols dependingon the conditions of their formation, Dokl. Akad. Nauk. SSSR 127 (1959) 366–368.
[38] Y.I. Petrov, E.A. Shafranovsky, Features of ultrafine inorganic particlepreparation by the gas evaporation method, Bull. Russ. Acad. Sci. Phys. 64(2000) 1236–1244.
[39] D. Mukherjee, C.G. Sonwane, M.R. Zachariah, Kinetic Monte Carlo simulation ofthe effect of coalescence energy release on the size and shape evolution ofnanoparticles grown as an aerosol, J. Chem. Phys. 119 (2003) 3391–3404.
[40] B. Cullity, C. Graham, Introduction to Magnetic Materials, Wiley-IEEE Press,Graham, 2008.
[41] A.E. Petrov, V.I. Petinov, V.V. Shevchenko, Magnetic properties of small aerosolnickel particles in the temperature range 4.2–300 K, Sov. Phys. Solid State 14(1972) 2595–2599.
[42] X. Batlle, N. Perez, P. Guardia, O. Iglesias, A. Labarta, F. Bartolome, L.M. Garcia, J.Bartolome, A.G. Roca, M.P. Morales, C.J. Serna, Magnetic nanoparticles withbulklike properties (invited), J. Appl. Phys. 109 (2011) 07B524.
[43] S. Chikazumi, C.D. Graham, Physics of Ferromagnetism, Oxford UniversityPress, Oxford, 2009.
[44] S.A. Makhlouf, F.T. Parker, F.E. Spada, A.E. Berkowitz, Magnetic anomalies inNiO nanoparticles, J. Appl. Phys. 81 (1997) 5561–5563.
[45] J.T. Richardson, D.I. Yiagas, B. Turk, K. Forster, M.V. Twigg, Origin ofsuperparamagnetism in nickel oxide, J. Appl. Phys. 70 (1991) 6977–6982.
[46] R.P. Cowburn, Property variation with shape in magnetic nanoelements, J.Phys. D: Appl. Phys. 33 (2000) R01–R16.
[47] C. Granqvist, R. Buhrman, Ultrafine metal particles, J. Appl. Phys. 47 (1976)2200–2219.
