Journal of Nanoparticle Research

, Volume 13, Issue 1, pp 375–384

Synthesis of nickel nanoparticles in silica by alcogel electrolysis


  • Muhammad Z. Rana
    • Department of Chemical and Materials Engineering (DCME), National Centre for Nanotechnology (NCN)Pakistan Institute of Engineering and Applied Sciences (PIEAS)
    • Department of Chemical and Materials Engineering (DCME), National Centre for Nanotechnology (NCN)Pakistan Institute of Engineering and Applied Sciences (PIEAS)
  • Jamil Ahmad
    • Department of Chemical and Materials Engineering (DCME), National Centre for Nanotechnology (NCN)Pakistan Institute of Engineering and Applied Sciences (PIEAS)
  • Muhammad Aslam
    • Department of Chemical and Materials Engineering (DCME), National Centre for Nanotechnology (NCN)Pakistan Institute of Engineering and Applied Sciences (PIEAS)
  • Syed K. Hasanain
    • Department of PhysicsQuaid-i-Azam University
  • Sohail Hameed
    • National Institute for Biotechnology and Genetic Engineering
Research Paper

DOI: 10.1007/s11051-010-0040-1

Cite this article as:
Rana, M.Z., Mehmood, M., Ahmad, J. et al. J Nanopart Res (2011) 13: 375. doi:10.1007/s11051-010-0040-1


We report a novel technique for the formation of metal nanoparticles, based on electrolysis of the alcogels containing metal chlorides. The alcogel was formed from TEOS, water, ethanol, and nickel chloride, and subjected to galvanostatic electrolysis. This resulted in successful formation of Ni nanoparticles inside the silica gel. Average particle size of FCC Ni lies between 18 and 20 nm. The formation of tetragonal nickel (a sub-oxide of nickel) as well as NiO were also detected by XRD and SAED. The resistivity measurements showed that the nickel nanoparticles were separated from each other by Ni(O) present between them. Magnetic studies based on ZFC and FC measurements below room temperature (up to 5 K) and above room temperature (up to 700 K) were conducted using SQUID and Magnetic TGA, respectively, which showed strong magnetic irreversibility as attributable to exchange interaction between metallic and oxide phases and mutual interactions among metallic particles in the network structure. The blocking temperature (~600 K) of the samples was above room temperature. M–H studies based on VSM showed an increase in magnetic coercivity with the formation of NiO. A magnetic transition associated with tetragonal nickel was seen at 10 K.


Sol–gelAlcogel electrolysisNanoparticlesTetragonal nickelExchange interactionSynthesis at room temperatureMagnetism


Nanomaterials exhibit interesting electrical (Moriarty 2001), optical (Moriarty 2001; Cai and Zhang 1997) and magnetic properties (Moriarty 2001; Dorman and Fiorani 1992; Ennas et al. 2004; Cintora-Gonzalez et al. 2001; Skumryev et al. 2003; Tang et al. 2006), in addition to high catalytic activity (Polshettiwar and Molnár 2007).

Nanoparticles of catalytic metals and alloys dispersed in mesoporous (oxide) matrices provide with extra-ordinary catalytic activity in comparison with their other forms (Polshettiwar and Molnár 2007). This is due to large surface area of the nanoparticles (Brinker and Scherer 1989) available for reaction and/or synergistic effects offered by the nanoparticles and the matrix phases at the interphase boundaries. The latter emerge from the fact that the two phases may adsorb mutually different species facilitating the reactions at their interfaces. These nanocomposites may also be extremely useful for electrochemical catalysts and sensors provided that the nanoparticles are interconnected among each other.

Nanocomposites composed of metallic nanoparticles dispersed in oxide phase exhibit interesting magnetic properties. For instance, magnetic nanoparticles isolated by an oxide (or non-magnetic materials) may exhibit superparamagnetism depending on the size and distribution of nanoparticles. Although superparamagnetism is an extremely interesting physical phenomenon, investigations to explore the magnetic nanoparticles for memory devices have compelled the researchers to enhance the magnetic coercivity of small nanoparticles caused by exchange interaction between the ferromagnetic core and the shell of an antiferromagnetic oxide phase (Skumryev et al. 2003). Both the above phenomena have been extensively investigated. A new approach has been suggested in which the interparticle interaction of magnetic nanoparticles lying at a distance shorter than exchange length produces high saturation magnetization, high initial permeability, and low eddy current losses (Peng et al. 2008; Ma et al. 2009). This approach may produce a second generation of core materials for high frequency applications, replacing the conventionally employed ferrites. However, these materials with network structure of the metallic nanoparticles covered by insulating phase have been scarcely investigated for their magnetic properties and synthesis.

The insulating phases which have generally been employed include ceramic as well as polymeric materials, such as SiO2 (Polshettiwar and Molnár 2007; Fidalgo and Ilharco 2005), Al2O3 (Liu et al. 2005, 2006), ZrO2 (Tom et al. 2003), TiO2 (Tom et al. 2003; Boiadjieva et al. 2003), polyurethane (Jamal et al. 2009), and block copolymers (Chipara et al. 2004), etc.

In general, it is difficult to prepare metallic nanoparticles directly through chemical methods employed for the preparation of nanocomposites and a subsequent pyrolysis and/or hydrogen reduction treatment becomes almost essential (Cai and Zhang 1997; Cintora-Gonzalez et al. 2001; Kan et al. 2003; Wu et al. 2001; Estournes et al. 1997; Ennas et al. 1998). This often results in undue growth of particles due to high temperature.

In this study, we have suggested a combination of sol–gel technique and alcogel (wet-gel) electrolysis to prepare composite of networked metallic (nickel) nanoparticles in silica matrix at room temperature. Emphasis has been laid on magnetic properties including ZFC and FC responses at cryogenic and elevated temperatures up to Curie temperature, as well as ACsusceptibility, coercivity, saturation magnetization, and conductivity.



Typically, a mixture containing TEOS (ACROS 98%), ethanol (94% by vol.), water (double distilled), and metallic salt NiCl2·6H2O (Riedel-de-Hauen 97% min.) was prepared in a conical flask. Then, 0.012 N HCl (Merck) was added drop-wise with continuous stirring in about 20 min. Stirring was continued for about 2 h, keeping a lid in place to avoid drying. The molar ratio of TEOS:Ethanol:Water:Ni:HCl was maintained around 1:4:11:0.33:0.005. The resulting sol was then transferred into an open beaker and heated slowly from room temperature (25 °C) to reach 50 °C in about 30 min. Within a few minutes afterward, gel-point was approached resulting in sudden rise in viscosity to form alcogel. Before its complete solidification, alcogel was transferred onto a copper plate, used as a cathode. Another plate of nickel to be used as anode was placed over alcogel. A load of 1.5 kg was applied against electrodes to maintain contact between the electrodes and the alcogel. The exposed area of one electrode was of the order of 12 cm2, and separation between them was about 3–4 mm. The electrolysis was performed galvanostatically using power supply G-W INSTEK PSP-405. After several hours, the gel was removed for characterization. To study the effect of current density, Ni nanoparticles were synthesized as above at various current densities ranging from 25 to 416 mA/cm2. To study the effect of concentration, alcogels with varying concentrations of Ni (i.e., mole fraction of Ni from xNi = 0.00232 to xNi = 0.021440) were subjected to electrolysis at a prescribed current density.


The crystal structure of nanoparticles was characterized by X-ray powder diffraction (XRD) using Bruker D8 Discover (Cu Kα radiation). The XRD patterns were analyzed using evaluation softwares EVA 11 and TOPAS-P. Transmission Electron Microscopy (TEM JEOL-1010) was performed to reveal the microstructure. Brunauer–Emmet–Teller (BET) surface area measurement was performed on micrometric adsorption instrument (Quantachrome NOVA 2000) using N2 as analysis gas. Data analysis was performed on adsorption isotherm using NovaWin 2.1 software.

Electrical resistance/conductivity of pressed samples (pellets) were measured using Autoranging Multimeter (Kiethly 175). Samples were pressed in a SANS electromechanical universal machine (CMT 4304), using a Teflon cylindrical die of 4-mm diameter, provided with removable metallic base and piston, which also functioned as electrodes for measuring the electrical resistance of the powder sample under a prescribed load.

M–H plots were obtained on Vibrating Sample Magnetometer (VSM). Ac-susceptibility was measured by the mutual inductance method using SR530 lock-in amplifier at a lock-in frequency of 270 Hz, and AC field of 0.8 Oe. Low temperature M-T (ZFC–FC) measurements were recorded using a SQUID device (Quantum Design MPMS XL). Magnetic Thermogravimetric Analysis (M-TGA/MDTG) was performed on a Thermo Gravimetric Analyzer (TGA/SDTA851e Mettler Toledo).

Results and discussion

Figure 1 shows typical XRD patterns of the powder samples prepared by alcogel electrolysis. FCC nickel forms predominantly, irrespective of the concentration of nickel in the gel (Fig. 1a) or applied current density (Fig. 1b). It shows that the metal–solution interfaces have moved much deeper inside the porous gel structure to allow for the formation of a high phase fraction of nickel. This is also in agreement with visual inspection, as the color of the gel changes up to a depth of about 2 mm away from the cathode. The nature of the process dictates the formation of networked particulate structures keeping connectivity with the cathode. At xNi = 0.00232, X-ray reflections of NiO are also seen in the XRD patterns (Fig. 1a), which disappear with increasing concentration of nickel in the precursor solution. However, when the concentration of nickel in the precursor solution increases, two other peaks appear at 35° and 59°. Similar peaks were also observed for samples prepared at different current densities from gel with xNi = 0.02114 (inset in Fig. 1b). The intensity of these X-ray reflections increases with current. Identical X-ray reflections have been seen by Nayak et al. (2005), who have assigned them to amorphous oxide (Ni–O) formed by adsorbed oxygen. Relatively sharper X-ray reflections at the same positions have been observed by Roy et al. (2005) who have clearly assigned them to the so-called tetragonal nickel, proposed as a solid solution of interstitial oxygen in nickel.
Fig. 1

Typical XRD patterns of samples, after electrolysis, prepared at various nickel concentrations in the gel (a) and applied current densities (b) along with plots of estimated particle size vs. nickel concentration (c) and vs. current density (d)

The above groups have formed sub-oxide of nickel by borohydride reduction of nickel chloride in aqueous solution. In this study, it seems to be formed by the surface adsorption of oxygen on small nickel particles or their partial oxidation.

At lower concentrations of nickel (xNi = 0.00232) and moderate current densities (50 mA/cm2), solution phase of gel, in front of the advancing metal–solution interface, may become depleted in nickel ions due to limited diffusivity of nickel, facilitating side reaction of hydrogen evolution. The hydrogen evolution reaction may cause a localized increase in the OH ion concentration or pH, leading to the enhanced oxidation of small nickel particles, which results in the formation of NiO. At higher concentrations of nickel, pH does not increase much, and oxidation of nickel is diminished. Therefore, as the nickel ion concentration of the precursor solution is increased, limited adsorption of oxygen is observed resulting in the formation of oxygen-stabilized tetragonal nickel. Increased current density also facilitates the side reaction of hydrogen evolution resulting in higher pH. Accordingly, the phase fraction of tetragonal Ni(O) increases, although no NiO is formed due to fast kinetics.

The average particle size of FCC Ni, as estimated from X-ray peak broadening (using Scherrer’s formula), lies in the range of 17–20 nm, irrespective of the nickel concentration in the gel and applied current density, as shown in Fig. 1c and d, respectively. The relatively broader X-ray reflections of tetragonal Ni(O) suggest an apparent particle size of about 2–3 nm.

Figure 2a shows a typical TEM image of the sample prepared at a current density of 50 mA/cm2. Nanoparticles dispersed in the gel are visible. The selected area electron diffraction (SAED) pattern, as shown on Fig. 2b, confirms the formation of FCC nanoparticles in amorphous silica matrix. Additional reflections of NiO are also seen in SAED, although XRD pattern of the sample (Fig. 1b) does not reveal the formation of NiO. The absence of tetragonal nickel in the SAED pattern suggests that the small particles of tetragonal nickel oxidize to form NiO during sample preparation.
Fig. 2

Bright field TEM image of sample prepared at 50 mA/cm2 along with corresponding SAED pattern (a and b, respectively), a dark field image from the same sample (c), and bright field images of samples prepared at 25, 75, and 100 mA/cm2 (d–f, respectively) along with a bright field image of sample synthesized by hydrogen reduction of alcogel (g)

The particle size (Fig. 2a), lying mostly smaller than 20 nm, is in agreement with the estimates made by X-ray peak broadening. Small size of nanocrystallites is also confirmed by a typical dark field image, as shown in Fig. 2c. Figures 2d–f present, respectively, the TEM images of the samples prepared at 25, 75, and 100 mA/cm2, for comparison. In agreement with Fig. 1d, the particle size remains almost unchanged with the applied current density.

Figure 2g shows TEM image of a sample prepared by reduction in hydrogen at 600 °C. Comparing with Fig. 2a–f, it can be noticed that the particle size obtained by alcogel electrolysis at room temperature is smaller than that obtained by elevated temperature hydro-reduction.

It appears that the alcogel primarily constitutes of silica-rich solid network, encompassing nickel chloride-rich liquid. Therefore, limited diffusivity of the electro-active species and constrictive action of the solid part of gel seems responsible for the formation of small nanoparticles of nickel dispersed in silica matrix. It also explains why the particle size is independent of nickel ion concentration and current density.

For applications at high frequency, resistivity of the magnetic materials should be high. This aspect was focused upon by measuring the electrical resistance of the samples. About 30 mg of sample was acquired in a cylindrical die (4-mm ID) of Teflon, and uni-axial load was applied using metallic pistons, which were also used as electrodes to measure the electrical resistance of the samples as a function of the applied load. Figure 3 is a plot of the electrical resistance as a function of the applied load. It may be noticed that about 100-kg load is required for complete electric contact (short circuit). It appears that particles are mostly separated from each other by non-conducting oxide—silica or Ni–O. The oxidation of small nanoparticles and inter-connecting necks may partly be responsible for the separation of the metallic nanoparticles in addition to breaking of the necks caused by stresses developed in the gel due to shrinkage during subsequent aging. Complete short circuit is possible when nanoparticles are pushed together or extruded through the oxide particles at a load of 100 kg. When the sample prepared by alcogel electrolysis is subsequently reduced in hydrogen atmosphere at 600 °C, electric contact is easily established through the powder at a small load of only about 10 kg, which supports our presumption that the nickel oxide (or sub-oxide) isolates the nanoparticles of FCC nickel.
Fig. 3

A plot of electrical resistance as a function of applied load for a typical sample after electrolysis and after subsequent hydrogen reduction

Porosity of samples is often desirable for catalytic activity. The specific surface area of a typical (50 mA/cm2) as-synthesized Ni-SiO2 nanocomposite was determined from N2 adsorption experiments. The BET area was found to be about 250 m2/g suggesting that the present technique can be used for preparing metal–ceramic catalysts.

In order to determine the magnetic properties of the samples, hysteresis loops were obtained using vibrating sample magnetometer. Figure 4a shows typical M–H curves of the composite samples as a function of solution chemistry. The saturation magnetization values as obtained from these curves have been plotted in Fig. 4c. It may be noticed that the saturation magnetization increases with an increase in the nickel ion concentration in the precursor solution up to xNi = 0.01, which seems to be due to an increase in the amount of metallic nickel in the sample. A further increase in concentration of nickel in precursor solution does not increase saturation magnetization and, consequently, the amount of metallic nickel. Applying higher current densities (at xNi = 0.02) does not enhance the nickel yield as saturation magnetization does not vary with current density (Fig. 4b, c, d).
Fig. 4

M–H curves as a function of nickel concentration and current density (a and b, respectively), magnetization as a function of nickel concentration and current density (c and d, respectively), and coercivity as a function of nickel concentration and current density (e and f, respectively)

The maximum saturation magnetization, which is obtained for composite samples prepared at xNi = 0.02, is about 30 emu/g. It is known that saturation magnetization of pure nickel is about 55 emu/g, suggesting that the maximum metallic nickel content of our samples may be about 55 wt%, which decreases with nickel in the precursor solution below xNi = 0.01.

Figure 4e shows magnetic coercivity of the samples with nickel concentration in the precursor solution. Coercivity of all the samples is much higher as compared to bulk nickel (about 0.7 Oe.). It decreases substantially with increase in the nickel concentration in the precursor solution. This change in coercivity may be related with the reduction in the amount of NiO (Fig. 1a) and exchange bias existing at the interface between the ferromagnetic nickel and antiferromagnetic Ni(O) (Ohldag et al. 2001; Sun and Dong 2002). The coercivity as measured from M–H curves does not vary with current density (Fig. 4f).

AC susceptibility of the samples was measured in the range of 298–725 K, typically as shown in Fig. 5. The magnetic susceptibility diminishes above the Curie temperature of FCC nickel (about 623 K). A broad peak appears at about 523 K, which is close to Neel temperature of NiO. A relatively sharp transition around this temperature has been observed by van Lierop et al. (2002) as related with the existence of exchange bias. In our case, the broad transition may be due to non-stoichiometric nature of oxide. In addition, the magnetic coupling between nickel nanoparticles due to their close proximity to each other may also be operative. An edge is observed in the susceptibility curve around 575 K (inset in Fig. 5). It may be related with the termination of Ni–Ni(O) coupling, after which only interparticle interaction exists.
Fig. 5

AC susceptibility curves of samples prepared at different current densities and nickel concentration of alcogel along with that of bulk nickel for comparison

A SQUID device was used for studying magnetization, M, as a function of temperature, T, under applied field strength of 100 Oe, as shown in Fig. 6. The sample was first zero-field cooled to about 4 K. ZFC measurements were then performed during heating up to a prescribed temperature of 300 or 400 K that was followed by cooling to 4 K for FC measurements. It can be noticed that ZFC and FC curves are widely separated from each other. The irreversibility in the magnetic response may be arising from exchange bias between ferromagnetic nickel and its sub-oxide along with magnetic interactions among the nickel nanoparticles lying at close proximity to each other.
Fig. 6

ZFC–FC magnetization curves for samples prepared at current densities of 50, 75, and 100 mA/cm2 (ac, respectively): applied field for FC being 100 Oe

It may be noticed that a peak appears in the ZFC curve at about 10 K. Such a transition has been found by Roy et al. (2006) at 15–20 K and by Nayak et al. (2005) at 10–25 K. As proposed by Roy et al., the presence of interstitial oxygen does not allow parallel alignment of magnetic spins of nickel atoms, which is responsible for very weak magnetization of this phase at room temperature. However, a transition to ferromagnetic nature possibly occurs at about 10–25 K causing enhanced magnetic susceptibility. The peak intensity in the ZFC and FC curves at 10 K increases with the increase in current density along with an associated decrease in susceptibility at higher temperatures. This suggests that higher amounts of phase fractions of tetragonal nickel being formed at the increased current density leads to enhanced coercivity (lower susceptibility) of FCC nickel at low applied field (100 Oe).

In order to compare FC and ZFC magnetic response at elevated temperatures, Magnetic Thermogravimetric Analysis (M-TGA) was performed as shown in Fig. 7a. Temperature was first raised above the Curie temperature of nickel (curve 1) followed by cooling to room temperature (curve 2) without magnetic field. Weight changes during heating (curve 1) are due to drying of the sample. After cooling, a magnet was placed below the sample, which caused a deflection in the balance indicating a pseudo-rise in weight proportional to magnetic force between the magnet and the sample. The temperature was then raised to a prescribed temperature for ZFC measurements (e.g., curve 3) followed by cooling to record FC response (curve 4). It may be noticed from ZFC curve that susceptibility starts decreasing beyond about 520 K. However, FC and ZFC curves do not meet each other up to about 600 K. This aspect has been further confirmed in typical measurements performed below Curie temperature (Fig. 7b). After obtaining ZFC curve up to a prescribed temperature, the temperature was intermittently decreased and then again increased to obtain several FC curves. When compared with the behavior of bulk nickel (Fig. 7c), it can be seen that the samples prepared by alcogel electrolysis exhibit irreversibility/magnetic blocking at elevated temperatures, beyond the Neel temperature of NiO, which may be due to non-stoichiometry of the oxide as well as inter-particle magnetic interaction due to close proximity of nickel nanoparticles.
Fig. 7

M-TGA curves of a typical sample after electrolysis (a and b) and that of bulk nickel for comparison (c)

The Curie temperature of the nickel nanoparticles as obtained from Fig. 7a is 622 K, which is comparable to the Curie temperature of bulk nickel (625 K): the variation lying within the experimental limitations. A relatively sharp transition in susceptibility (ZFC curve) is observed at about 575 K. A transition at the same temperature was also observed in the AC susceptibility curve (inset in Fig. 5) and appears to be associated with termination of bias coupling between nickel and its oxide.


A new technique based on electrolysis of alcogels has been employed for the synthesis of nickel nanoparticles. Nickel chloride as nickel precursor was introduced in the solution, which after gelation resulted in uniform distribution of nickel precursor throughout the pores of alcogel. Electrolysis of as-generated alcogels (i.e., without any subsequent treatment) resulted in the formation of nickel nanoparticles. Estimations based on comparison between the saturation magnetization of samples with bulk nickel indicated a nickel content of 27–55 wt% in samples depending on the concentration of nickel chloride precursor in solution. Average size of nickel nanoparticles lied between 17 and 20 nm. The presence of both tetragonal nickel (a solid solution of nickel containing oxygen) and NiO has been confirmed by diffraction experiments. Resistivity measurements show that nickel nanoparticles are separated from each other due to the presence of Ni(O) between them. A ferromagnetic–antiferromagnetic transition corresponding to tetragonal nickel was observed at about 10 K. The samples remained in the blocked state up to about 600 K seemingly due to exchange bias coupling between nickel and its oxide and interparticle interactions between nickel nanoparticles lying in close proximity to each other. A relatively sharp transition in the susceptibility behavior is observed around 575 K, which seems to be related to termination of nickel–nickel oxide exchange bias effects.

Copyright information

© Springer Science+Business Media B.V. 2010