Introduction

Diluted magnetic semiconductors (DMS) (Ohno 1998) produced by doping transition metal or rare earth metal ions into non-magnetic semiconductors have been of great interest to realize spintronics devices such as spin LED (Fiederling et al. 1999, p 787), spin field effect transistor (Datta and Das 1990), etc., in the near future. Some authors report ferromagnetic behavior at room temperature in pure TiO2, HfO2, and ZnO nanoparticles (Coey et al. 2005; Hong et al. 2006; Dietl et al. 2000) due to defects and or oxygen vacancies, hence it is crucial to differentiate the source of magnetism intrinsic either from the dopant or from impurity phases. Lee et al. (2003) reported magnetization of Fe-doped TiO2 samples to decrease with Fe composition. Apart from oxides, transition metal doped semiconductor systems such as (GaMn)As (Ohno et al. 1996), (GaMn)N (Reed et al. 2001), (InMn)As (Ohno et al. 1992), and Cr-doped CuZnSe2 (Paul Joseph and Venkateswaran 2010) DMS also exhibit ferromagnetism. Recently, a wide variety of magnetic behavior was observed in SnO2 doped with V, Cr, Mn, Fe, Co, and Ni (Hong and Sakai 2005; Fitzgerald et al. 2004; Wang et al. 2006; Hong et al. 2005; Ogale et al. 2003; Coey et al. 2004).

Tin dioxide is an n-type wide band gap semiconductor that has variety of applications such as solid-state gas sensors, surge arresters, transparent conductors and oxidation catalysts (Fukano et al. 2005; Paul Joseph et al. 2009; Korotcenkov 2005). ‘Er’ has been doped into SnO2 and studied mainly for its optical properties (Brovelli et al. 2006; Wu and Coffer 2007; Morais and Luis Scalvi 2007) and a very few from DMS point of view (Mohan Kant et al. 2005). SnO2-based powders are obtained by means of a variety of synthesis techniques including the mixed oxides route, co-precipitation and sol–gel methods (Wang et al. 1994; Tarey and Raju 1985; Minami et al. 1988). In this study, we prepared Er3+doped SnO2 nanoparticles by sol–gel technique, a method employed quite frequently to synthesize nanoparticles because of its low cost and good stoichiometric control. The structural, optical, morphological, and magnetic properties of the prepared nanoparticles are investigated.

Experimental details

Synthesis of Er-doped SnO2 nanoparticles

Undoped and Er-doped SnO2 DMS nanoparticles were prepared by a sol–gel technique. To achieve the required composition of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1), appropriate amount of tin chloride (SnCl4·5H2O) and erbium chloride (ErCl3·6H2O) was dissolved in 75 mL of distilled water at 80 °C along with 6 mL of polyglycol and citric acid (to attain pH = 1.5) was continually stirred for 10 min until a sol is formed. Ammonia solution (NH3·H2O (28%)) was added drop wise to the above mixture until pH = 8. The formed hydroxide product was stirred for 3 h to form a gel, and finally dried at 120 °C/12 h and calcined at 400 °C/2 h in air.

Measurements

The X-ray diffraction (XRD) patterns were obtained using X’PERT PRO X-ray diffractometer with CuKα = 1.5406 Å radiation. Transmission Electron Micrographs (TEM) were recorded in (JEOL-TEM 2010) with an accelerating voltage of 200 kV. The optical absorption measurements were performed in a JASCO-V-670 spectrophotometer. Electron spin resonance (ESR) spectra of powder samples were recorded at room temperature using X-Band JEOL, JES PX 2300 spectrometer in the frequency range of 8.8–9.6 GHz. The magnetic measurements were carried out using a superconducting quantum interference device (SQUID, Quantum Design MPMS-XL7).

Results and discussion

The XRD patterns of Sn1−x Er x O2 (with x = 0.0, 0.02, 0.04, and 0.1) DMS nanoparticles (Fig. 1) reveal that all the samples have a rutile-type cassiterite (tetragonal) phase of SnO2, and the doping does not change the tetragonal structure (JCPDS # 41-1445) of SnO2. Furthermore, we could not find any diffraction peak corresponding to any impurity phase within the limit of instrumental sensitivity. The peak positions do not show any measurable change, while the intensities of the peaks increase with increasing Er content. The diffraction peaks of undoped SnO2 are broadened and the average crystallite size was estimated to be 16.1 nm. As Er content increases, the XRD peaks appear to be sharper with decreased full width at half the maximum (FWHM), indicating possible increase in crystallite size. The average crystallite size of the Er-doped samples was found to be in the range 16–19 nm using the Scherrer equation (Table 1). The TEM measurements were performed to confirm the nanocrystalline nature and to study the morphology of the particles. Typical TEM micrograph of Sn1−x Er x O2 (x = 0.02) sample (Fig. 2) shows well isolated and nearly spherical shaped crystallites. The distribution plot fitted with a Gaussian profile (Inset of Fig. 2) shows narrow distribution in size with an average crystallite size of 12 nm. The TEM micrograph and SAED pattern of sample with x = 0.04 also show well isolated nanoparticles (Fig. 3). The high resolution TEM image of sample with x = 0.04 shows highly crystallized spherical and few elongated particles with clear lattice fringes and with almost no grain boundaries (Fig. 4). The calculated d spacing value of 1.18 Å for the lengthy elongated rod correspond to the (400) plane (JCPDS # 41-1445). The elongated rod was found to have high aspect ratio of 5.33. The estimated average crystallite size from size distribution plot (Inset of Fig. 4) is 4 nm. The average crystallite size estimated from XRD and TEM have contradicting trend. In this case, due to the method of preparation and subsequent annealing, the particles are well crystallized and hence the FWHM value decreases indicating increasing crystallite size with Er concentration. However, the direct observation by TEM and analysis of the data confirms decrease in size with Er concentration.

Fig. 1
figure 1

XRD patterns of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) nanoparticles

Full size image
Table 1 Average crystallite size, band gap and g values of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) nanoparticles
Full size table
Fig. 2
figure 2

TEM micrograph of Sn1−x Er x O2 (x = 0.02) nanoparticles with the inset showing the size distribution plot

Full size image
Fig. 3
figure 3

TEM micrograph of Sn1−x Er x O2 (x = 0.04) sample showing smaller sized nanoparticles. Inset is the SAED pattern

Full size image
Fig. 4
figure 4

High resolution TEM micrograph of Sn1−x Er x O2 (x = 0.04) nanoparticles with the inset showing the size distribution plot

Full size image

The optical properties of semiconductor nanoparticles exhibiting interesting behavior have been studied extensively in recent years. Optical absorption spectra of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) nanoparticles shown in Fig. 5 indicate absorption edge to shift to shorter wavelengths implying a blue shift in band gap with respect to bulk SnO2 (3.6 eV at 300 K). Similar blue shift in band gap has been observed in the case of pure SnO2 due to size effect by Das et al. (2006). This blue shift in band gap is due to size effect, and observed when the particle size of a semiconductor becomes comparable to the Bohr radius of the exciton leading to variations in the properties of the material due to quantum confinement. The increase in band gap value with decreasing crystallite size induced by Er content is listed in Table 1.

Fig. 5
figure 5

Optical absorption spectra of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) nanoparticles

Full size image

The ESR technique has been employed to study rare earth ions in a variety of host lattices (Abragam and Bleany 1970) in which the study of ground state of the rare earth impurity reveals the symmetry of the occupied state. The ESR spectra of Er3+ in SnO2 have been recorded at room temperature in the field range 320–350 mT, at 100 kHz field modulation to obtain the first derivative spectrum. No resonance signal was detected in the ESR spectra of undoped SnO2 nanoparticles. Figure 6 shows the single resonance peak observed for Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) samples confirming Er3+ ion substituting the Sn4+ sites. This resonance signal could be attributed to Er3+ ions in SnO2 nanoparticles due to ground state of free Er3+ ion in 4f11 electronic configuration (Yang et al. 2009). The g values of Er-doped samples are listed in Table 1. Ting et al. (2001) reported g value of 2.0 for a relatively small peak of Er3+ doped TiO2. The intensity of ESR signal is initially high for x = 0.02 and then it decreases with increasing Er content which may be because of the possibility of antiferromagnetic interactions induced in the sample due to higher Er concentration.

Fig. 6
figure 6

The ESR spectra of Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) samples

Full size image

The temperature dependent (M(T)) (at 500 Oe) and field dependent (M(H)) (at 300 K) magnetization measurement of undoped nanocrystalline SnO2 sample exhibits diamagnetic nature (Fig. 7) confirming that there is no positive susceptibility contribution from defects and oxygen vacancies of SnO2 which has been cautioned to be a universal feature of non-magnetic oxide nanoparticles (Sundaresan et al. 2006). The magnetic hysteresis loops measured at 5 and 300 K of Sn1−x Er x O2 with x = 0.04 and 0.1 nanoparticles are shown in Figs. 8 and 9, respectively. The magnetization at 5 K, though very small due to the doped Er3+ ions, it is comparatively higher than that at 300 K. It can be seen that hysteresis with a small coercivity (34 Oe) is observed at room temperature for x = 0.04 (bottom inset of Fig. 8) which is not the case at 5 K (top inset of Fig. 8) indicating loss of the observed weak magnetic behavior at low temperatures. For x = 0.1, we do not observe any hysteresis behavior both at 5 and 300 K (Top and bottom inset of Fig. 9) indicating loss of weak ferromagnetic behavior with increasing Er concentration. This behavior can be justified considering antiferromagnetic interactions to increase with Er content thereby decreasing the observed weak ferromagnetic interactions. Mohan Kant et al. (2005) reported intrinsic ferromagnetism in rare earth (Gd, Dy, and Er) doped SnO2 thin films by pulsed laser deposition. However, antiferromagnetic behavior has also been reported in the case of Co-doped ZnO and Fe-doped ZnO, TiO2, and SnO2 (Bouloudenine et al. 2005; Soumahoro et al. 2010; Lee et al. 2003; Sambasivam et al. 2011). Lawes et al. (2005) reported antiferromagnetic interactions in Mn- and Co-doped ZnO to depend on concentration of the magnetic dopant.

Fig. 7
figure 7

The diamagnetic behavior of nanocrystalline SnO2 sample as a function of temperature. Inset shows the field dependent diamagnetic signal at 300 K

Full size image
Fig. 8
figure 8

The M(H) curves of Sn1−x Er x O2 (x = 0.04) nanoparticles at 5 and 300 K. Insets show the hysteresis behavior at lower fields

Full size image
Fig. 9
figure 9

The M(H) curves of Sn1−x Er x O2 (x = 0.1) nanoparticles at 5 and 300 K. Insets show the paramagnetic behavior without any hysteresis behavior at lower fields

Full size image

The molar susceptibility (χ(T)) plot from the temperature-dependent magnetization (M(T)) of the Sn1−x Er x O2 (x = 0.04 and 0.1) nanoparticles under field-cooled (FC) (at 500 Oe) mode are shown in Fig. 10. Significant decrease in the positive susceptibility is observed with Er content increasing from 0.04 to 0.1 in the temperature range 300–50 K, below which the susceptibility merge and start increasing. The M(T) curves of x = 0.04 and 0.1 show a steep rise in magnetization value below 50 K without any distinct magnetic phase transition. There is no bifurcation observed between the ZFC–FC curves of x = 0.04 and 0.1 (Inset of Fig. 10), thereby indicating absence of intrinsic ferromagnetic behavior in the samples.

Fig. 10
figure 10

The molar susceptibility curves calculated from field-cooled (at 500 Oe) curves of Sn1−x Er x O2 (x = 0.04 and 0.1) nanoparticles. The inset shows the ZFC–FC behavior of the samples

Full size image

The χ(T) of Sn1−x Er x O2 (x = 0.04 and 0.1) (Figs. 11 and 12) are fitted to Curie–Weiss law

$$ \chi = \frac{C}{T + \theta } $$
(1)

where C is the Curie constant, θ is the Curie–Weiss temperature, and T is the temperature.

Fig. 11
figure 11

The molar susceptibility curves calculated from field-cooled (at 500 Oe) curves of Sn1−x Er x O2 (x = 0.04) nanoparticles. The inset shows the variation of inverse susceptibility with temperature

Full size image
Fig. 12
figure 12

The molar susceptibility curves calculated from field-cooled (at 500 Oe) curves of Sn1−x Er x O2 (x = 0.1) nanoparticles. The inset shows the variation of inverse susceptibility with temperature

Full size image

The inverse susceptibility (1/χ(T)) of x = 0.04 sample (Inset of Fig. 11) has a nearly flat region starting from 300 to around 175 K, below which there is a distinct curvature up to 5 K. In order to obtain θ, which reflects the strength and nature of the magnetic interaction, we have extrapolated the straight line fitting of the linear region at low temperatures of 1/χ(T). The intercept of the extrapolated line yielded a negative value of −5 K indicating presence of antiferromagnetic interactions. For x = 0.1, the behavior of χ(T) (Fig. 12) is similar to that of x = 0.04, however, the 1/χ(T) of x = 0.1 (Inset of Fig. 12) is distinctly different with less curvature and more or less similar to a straight line. The linear fit intercept at −12 K indicates increased antiferromagnetic interactions on increasing Er content to x = 0.1. The value of θ being negative for Er-doped SnO2 samples, −5 K for x = 0.04 and −12 K for x = 0.1 confirms the presence of antiferromagnetic interactions.

According to RKKY theory (Ruderman and Kittel 1954; Yosida 1957), the observed magnetic properties are due to the exchange interaction between local spin-polarized electrons (such as the electrons of Er3+ ions) and conduction electrons. The conduction electrons are regarded as a media to interact among the Er3+ ions. This interaction leads to the spin polarization of conduction electrons. Subsequently, the spin-polarized conduction electrons perform an exchange interaction with local spin-polarized electrons of other Er3+ ions. However, the exchange interaction is short ranged and oscillating nature based on the concentration of Er and its nearest neighbor distance. This may be the plausible explanation of the observed antiferromagnetic interactions.

Conclusion

Sn1−x Er x O2 (x = 0.0, 0.02, 0.04, and 0.1) DMS nanoparticles prepared by sol–gel technique had rutile structure without any impurities as confirmed by XRD and TEM measurements. High resolution TEM results indicate the nanoparticles to be highly crystalline and the crystallite size to decrease with increasing Er content. The optical absorption spectra showed blue shift in band gap with increasing Er content in SnO2 nanoparticles due to size effect. ESR analysis shows a single resonance peak due to Er in 3+ state in SnO2 and also decrease in intensity above x = 0.02 Er content. The inverse susceptibility data from the field-cooled and zero-field-cooled magnetization measurements reveal that Er-doped SnO2 nanoparticles tend toward antiferromagnetic behavior with Er content.