1. Introduction

ZnO is a II^b–VI semiconductor with direct wide band gap (3.37 eV) and high exciton binding energy (60 meV) at room temperature^{[1, 2]}. Doping of transition metal (Mn, Fe, Ni, Co and Cu, etc.) in ZnO forms diluted magnetic semiconductor (DMS), exhibiting multifunctional properties such as magnetic, semiconducting, and optical properties coinciding in the single material at RT, which make it prominent in spintronic, photonic and optoelectronic applications^[3–6].

A lot of research has been done on the TM-doped ZnO systems prepared by different methods^[7–10]. However, the reports on magnetic and optical studies of Cr-doped ZnO are very interesting and conflicting. Cr and its various impurity phases such as Cr₂O₃ and Cr₃O₄ are antiferromagnetic by nature, but become ferromagnetic in Cr₂O and Cr₂N. So, it is interesting to observe whether the coupling remains ferromagnetic or not, with Cr doping in ZnO. In 2001, Jin et al.^[11] prepared ZnO thin films doped 3d TMs using molecular beam epitaxy (MBE) technique. They did not find any indication of ferromagnetism in Cr-doped ZnO film down to 3 K. In addition, Udea et al.^[12] used pulsed-laser deposition (PLD) method to fabricate 3d TM doped ZnO films. Contrary to theoretical predictions, they also have not detected ferromagnetism in Cr-doped ZnO films. On the other hand, Wang et al.^[13] investigated magnetic properties of Cr-doped ZnO films by first principle calculations and conclude that Cr–Cr coupling is ferromagnetic in bulk case, while coupling is antiferromagnetic in the surface case, and this antithesis arises from the different morphologies associated with Cr doping. Further, Duanet al.^[14] used a novel auto-combustion technique to synthesize Cr-doped ZnO nanoparticles and observed ferromagnetic behavior at room temperature. Similarly, Pang et al.^[15] also observed ferromagnetism in Cr-doped ZnO thin films prepared using RF magnetron sputtering method. Li et al.^[16] and Naqvi et al.^[17] observed the opposite trend of variation in the band gap with Cr doping, contrary to the results reported by Hassan et al.^[18], Safa et al.^[19] and Sartiman et al.^[20]. Yukai An et al.’s^{[39, 40]} work provides strong evidence that the oxygen vacancies play an important role in activating the ferromagnetic interaction in TM-doped In₂O₃ thin films. Although a number of reports are available, confusion persists on the existence, nature, and origin of room temperature magnetism in Cr-doped ZnO. In this paper, we have used the auto-combustion method for synthesis of the samples which is a very simple, quick, and low-cost method. This technique is least used to synthesize Cr-doped ZnO. The analysis of VSM data with different concentrations, shows significant enhanced room temperature ferromagnetism in Cr-doped ZnO samples. The measured value of magnetization is 12 times larger than the value reported by Daun et al. (2010). Room temperature ferromagnetism of the nanoparticles is very meaningful for spintronics applications.

2. Experiment

2.1 Synthesis

The nanocrystalline powder samples of Zn_1−xCr_xO (x = 0.00, 0.01, 0.03, 0.05, 0.07 and 0.09) were synthesized by auto-combustion method. Fig. 1 shows the schematic diagram of an auto-combustion method to prepare Cr-doped ZnO nanoparticles. The appropriate amounts of zinc nitrate [Zn(NO₃)₂·6H₂O, Aldrich, 99%], chromium nitrate [Cr(NO₃)₃·9H₂O, Aldrich, 99%] and glycine (H₂NCH₂COOH, Panreac, 99.8%) were dissolved in 50 mL distilled water. Then, the solution was heated on a hot plate at 250 °C under constant stirring for 1 h. After that, the solution was transformed into a gel that further swelled into the foam and then the self-propagating combustion reaction was carried out. As a result, off-white color un-doped ZnO and moss green color Cr-doped ZnO powders were obtained, which were ground well in an agate mortar for half an hour. The colors of undoped ZnO and Cr-doped ZnO powders were turned into white and gray, respectively, after the calcination at 600 °C for 1 h.

2.2 Characterization

The structure and the average size of the nanoparticles were determined by XRD using X-ray diffractometer (STOE Powder Diffraction System, Model: Theta-Theta) employing CuK_α radiation. The morphology of the samples was observed in the field emission scanning electron microscope (QUANTA, Model: FEG 250). EDS analysis has been made by using the MIRA3 TESSCAN EDS. Absorption spectra of all samples in the ultraviolet–visible spectral region were recorded using UV–Vis spectrophotometer (SHIMADZU, Model: UV–1800). The magnetization curves of the nanopowders were obtained using VSM (Lake Shore, Model: 7410).

3. Results and discussion

3.1 XRD analysis

XRD data of Zn_1−xCr_xO samples (x = 0.00, 0.03, 0.05, 0.07, and 0.09) were collected on a STOE Powder Diffraction System using CuK_α-radiation (λ = 1.5406 Å) over a range 20° < θ < 80°. Fig. 2 shows the patterns of the pure ZnO and the Cr-doped ZnO. The diffraction peaks corresponding to the (100), (002), (101), (102), (110), (103), (200), (112), (201), (004) and (202) lattice planes signify a hexagonal wurtzite structure of the nanocrystals. From Fig. 2 one can observe that the Bragg angle of the intense (101) diffraction plane slightly shifted towards higher values relative to that of pure ZnO. It is seen that when the content of Cr (x) is 0.03, a very weak secondary phase is to be observed, and at 0.05, 0.07, and 0.09, the phase appearance slightly improved with Cr concentration, which agrees with the results reported by Duan et al.^[14]. Fig. 2 inset shows the peaks position of (101) plane. The inset shows that the intensity of the diffraction peaks decreases gradually with the increase of doping concentration, indicating a loss of crystallinity due to lattice distortion. When Cr ions are incorporated into the system of ZnO, this results in the alternation of the lattice periodicity and decrease in crystal symmetry. As can be seen from the XRD patterns, the diffraction peaks get broadened as the Cr concentration is increased, suggesting a systematic decrease in the grain size. It can be seen from the estimated data of our samples, that the doped Cr nanoparticles deviate considerably from those of the un-doped sample due to the incorporation of Cr ions into the ZnO lattice, which induces the local distortion of the crystal structure^[21].

Compared to pure ZnO, small change in diffraction angle associated with Cr-doped ZnO samples reveals an overall variation of the lattice parameter. So, there is a small change in the lattice constants, which depends on the amount of glycine employed in the sample preparation. The measured values of a and c are in the range of 3.2526–3.2417 and 5.1984–5.204 Å, respectively. The average grain size for all the samples was calculated using Scherrer’s formula;

$$D = \frac{{0.9\lambda }}{{\beta \cos \theta }},$$

(1)

where D is the average crystallite size, λ = 1.5406 Å, the wavelength of X-ray (for CuK_α), β is the FWHM in radian and θ is the diffraction angle. Fig. 3 shows the variation of crystallite size with Cr concentration. The calculated crystallite size of each sample is depicted in Table 1. It can be observed from Table 1 that the crystallite size of ZnO decreased from 30.7 to 9.2 nm when Cr content increased from 0.00 to 0.07 and again increased by increasing the Cr content up to 0.09. One reason in size reduction may be a particle growth prevention, the motion of a grain boundary must be impeded^[22]. We can elucidate the hindrance on a movement of the grain boundaries by Zener pinning. When the moving boundaries attached the zinc interstitial and the substituted Cr ions, they will offer a retarding force on the boundaries. If the retarding force was generated more than the driving force for grain growth, the particles cannot grow any longer^[23]. The XRD spectra have also been used to study the crystallinity of the samples. The doping of Cr in ZnO not only lowers the particle size but also degrades the crystallinity of the nanoparticles. As the Cr content increases, the intensity of XRD peaks decreases and FWHM increases which is due to the degradation of crystallinity. This means that even though the Cr ions occupy the regular lattice site of Zn²⁺, it produces crystal defects around the dopants and these defects change the stoichiometry of the materials. In this context, it is well established that the lattice distortion due to the defects (vacancies, interstitial, substitutional, local structure transformation etc) may cause the shift in XRD peak position depending on the type of strain in the crystal.

Further, when impurities are incorporated into lattice sites, they act as nucleation centers and therefore nucleation rates increase. Moreover, addition of impurity increases interfacial energy; thus, further crystal growth is inhibited. Consequently, it is considerably more difficult for the doped crystal to grow compared to the pure host crystal in the absence of impurities, and particle clustering occurs to reduce the interfacial energy which induced by decreasing of the particle size^[24]. Meanhile, the increase in size for Zn_0.91Cr_0.09O (shown in Fig. 3) may be due to possible residence of some Cr atoms on the octahedral interstitial site, as the strong preference of Cr for an octahedral rather than tetrahedral coordination with oxygen^[18].

3.2 SEM analysis

The morphology of pure ZnO and Cr-doped ZnO (Zn_0.97Cr_0.03O) samples has been analyzed by using field emission scanning electron microscope (FESEM). Fig. 4 shows the FESEM photomicrographs of pure ZnO sample with magnification ×3000. It is observed that the microstructure consists of many rounded shape grains, however, there is the agglomeration in utmost regions, as higher flame temperature of glycine results in the formation of semi-sintered^[25].

3.3 EDS analysis

Energy-dispersive spectra (EDS) measurements were used to determine the chemical composition of the Cr-doped ZnO samples. Fig. 5 shows EDS spectra of the Cr-doped ZnO samples. The content of different elements (Zn, O, and Cr) in the sample can be observed in the spectra, confirming the incorporation of Cr atoms into the nanocrystalline ZnO particles. The EDS results indicate that the amounts of Cr incorporated in the samples were slightly lower than the amounts of Cr introduced in the synthesis. The spectra reveal that only three elements, Zn, O, and Cr exist in Cr-doped ZnO nanoparticles. No traces of other elements were found in the pure ZnO and Cr-doped spectra which confirm the purity of the samples. The atomic ratio (x) of Cr to Zn is determined from the relation x = Cr/([Zn + Cr]) and it is found to be 0.011, 0.040, 0.050, 0.07, and 0.11 doping. EDS analysis gives what amount of Cr atom has really entered into the ZnO nanoparticles. We actually doped 0.01, 0.03, 0.05, 0.07 and 0.09 Cr. But EDS spectra show a slightly increased atomic ratio of Cr at 0.09. It may be associated to Zn atoms that should have been lost during reaction process in weight ratio (or) impurity should have been formed. But, we have not recorded any peak accountable for impurity. Therefore, it is assumed that Zn atoms could have been removed in weight ratio. Atomic percentage (%) is shown in Table 2. Thus, composition analysis confirms the incorporation of Cr element in the samples.

3.4 Optical characterization

UV–visible absorption spectroscopy is a powerful technique to study the optical properties of semiconducting nanoparticles. The optical band gap of the Zn_1−xCr_xO (x = 0.00, 0.01, 0.03, 0.05, 0.07, and 0.09) samples were calculated using UV−visible absorption data. Fig. 6 shows UV−Vis absorption spectra of all samples calcined at 600 °C. The UV–visible spectra shown in Fig. 6 attributes that strong UV absorption is characteristic of all measured samples, which attains a plateau above 378 nm. The absorption peak of pure ZnO is observed at 375 nm and the peaks of Cr-doped ZnO samples are observed in the range 376–378 nm. The peaks below 380 nm shows decreasing trend in absorption as we increased the Cr concentration. For un-doped ZnO nanoparticles the band gap comes out to be 2.83 eV and for doped samples are 2.76 to 2.42 eV. This indicates that the band gap of ZnO material decreases with the increasing doping concentration of Cr ions. The band gap value of un-doped is slightly lower than the reported value in literature and this decreasing trend in band gap may be attributed to fuel selection for auto-combustion reaction. Carrero et al. reported the band gap of undoped ZnO 2.75 eV synthesized by auto combustion method. Moreover, when grain size increases band gap energy is found to be decreased. This is due to quantum confinement. A literature^[42–44] survey also indicates significant variation in band gap of ZnO.

The energy band gap of the samples is calculated by the following relation presented by Tauc^[27]:

$$\alpha hv = C{\left( {hv-{E_{\rm g}}} \right)^{\frac{1}{2}}},$$

(2)

where C is constant, α is the absorption coefficient, hv is the incident photon energy and E_g is the energy band gap. The plot of (αhv)² versus hv yields the value of the energy band gap by extrapolating the linear portion of the plot on to the energy axis as shown in Fig. 7. The observed values of band gaps are shown in Table 1. The band gap decreases from 2.83 to 2.35 eV for 0.00 to 0.07 Cr concentration and grain size also decreases consistently, while, an increase in band gap is observed for 0.09, as its grain size is also increased. This variation of band gap with Cr doping concentration is in good agreement with the results reported by^{[18–20, 28]}.

According to Yilmaz et al.^[29], the decrement in the band gap with increasing Cr content may be due to the formation of various optically active deep sub-levels through the band gap. The Cr⁺³ ions generate an optical sub level due to the difference of their ionic number with Zn⁺². The blue shift can be attributed to the sp-d spin-exchange interactions between delocalized s- or p-type band electrons of ZnO and delocalized d-electrons of Cr⁺³ ions. The exchange interactions cause a decrease in conduction band edge, and an increase in valence band edge, resulting in a decrement in the band gap.

In the present work, the variation in band gap is consistent with the grain size, which can be explained by a simple model introduced by Suryanarayan et al.^[30]. Per this model, trapping of impurities arising out of doping cause the periodic variations in potential within the grain. Such periodic variation of potential, further leads to the band bending or band tailing effect^[31]. Impurity band formation is an obvious consequence of increased doping concentration and the trapping of the dopant atoms at the grain boundary leads to the formation of defect states within the forbidden band. By increasing doping concentration, density of these defect states increases, leading to the decrease of band gap or red shift. Tuning the energy band gap of ZnO nanoparticles by Cr doping make it promising for photocatalytic and photoelectronics applications.

3.5 Magnetic characterization

Magnetic measurements of Cr-doped ZnO samples were performed by using VSM under high magnetic field in the range of ±20 500 Oe at room temperature. Magnetization (M) versus magnetic field (H) loops of Zn_1−xCr_xO (x = 0.00, 0.01, 0.03, 0.05, 0.07, and 0.09) samples is shown in Fig. 8. These M−H loops are found to be hysteretic, signifying room temperature ferromagnetism (FM). Table 1 shows the corresponding values of magnetization of all Zn_1−xCr_xO samples. The value of magnetization of Zn_0.99Cr_0.01O sample is 0.097 emu/g, which shows 12 times increase in magnetization than the value reported by Daun et al.^[14]. A decrease in magnetization up to x = 0.05 and 0.07 Cr content may be due to the competition between ferromagnetic and antiferromagnetic coupling that occurs at short distances between a pair of Cr ions^[20]. However, sample with x = 0.09 Cr content contains a relatively large fraction of Cr atoms to contribute in remarkably enhanced ferromagnetism.

ZnO is diamagnetic type compound and Cr atom is antiferromagnetic. Also, the impurity phases in Zn–Cr–O system, such as Cr₂O₃, Cr₃O₄ and ZnCr₂O₄ are antiferromagnetic, while, CrO₂ is ferromagnetic phase^{[32, 33]}. However, CrO₂ is unstable and decomposes under normal conditions^[34].

According to Duan et al.^[14], increasing super exchange (SE) interactions with doping content cause FM for Cr-doped ZnO. Since the favored ferromagnetic SE interactions between Cr⁺³ take place when a majority p-electron is shifted, the spin S on the TM is maximized, lowering the ground state energy in conformity with the Hund’s rule. Karmaker et al.^[32] concluded that the native defects in ZnO host can originate the FM in TM-doped ZnO system. Jin et al.^[33] proposed that bound magnetic polaron (BMP) model is valid for room temperature FM, in which the long-range ordering could be formed by either direct overlaps between BMPs or indirect BMP-magnetic impurity-BMP interactions. Cr doping content increases the density of magnetic impurities, which can facilitate the FM alignment of BMPs. According to Sundaresan et al.,^[34] when no magnetic impurities are present, the ferromagnetism of metal oxide nanoparticles can be associated with the exchange interactions between localized electron spin moments resulting from oxygen vacancies at the surfaces of nanoparticles. Ghosh et al.^[35] found that the presence of large numbers of V_O⁺ defects are responsible for the observed FM in n-type pristine SnO₂ thin films.

There is also an emerging consensus that ferromagnetic behavior in transition-metal-doped ZnO is correlated with defects such as oxygen or zinc vacancies^[36–39]. Yukai An et al.^{[34, 40–41]} reported that oxygen vacancy around the doped ions is crucial to the observation of RT ferromagnetism in Mn-doped In₂O₃ films. Considering the literature, the existence and improved nature of ferromagnetism is explained as follows. From the EDS spectra of Cr-doped ZnO NPs, the atomic percentage of oxygen is found to be increased with increasing chromium concentration and the atomic percentage of zinc is found to be decreased. In our case it is clearly shown from EDS spectra that content of oxygen and zinc shows significant variation with chromium concentration. This variation of oxygen and zinc contents is very well matched with the M–H loops as shown in Fig. 8. From EDS spectra, the variation of oxygen contents and M–H loops variations are as per the following order as 0.09 > 0.01 > 0.03 > 0.05 > 0.07 > 0.00 of chromium concentration. The values of magnetization (emu/g) from Table 1 for different concentrations also supported this trend as mentioned above. The qualitative literature survey also suggests that the origin of RTFM in these systems is strongly dependent on the synthesis route.

4. Conclusion

Pure and Cr-doped ZnO nanopowders were prepared by an auto-combustion method. The average grain size of nanocrystals was estimated using Debye Scherrer’s formula, which showed a decline in the range of 30.7–9.2 nm with increasing Cr concentration up to 0.07. This may be due to the smaller ionic radius of Cr⁺³. An increase in Zn_0.91Cr_0.09O sample is due to possible residence of some Cr atoms on the octahedral interstitial site. FESEM analysis indicated that the morphology of nanopowder is highly effected by Cr doping. EDS analysis shows host and doping elements without impurities. The optical absorption spectra indicated a red shift in the absorption band edge with an increase in Cr doping content. The band gap decreased from 2.83 to 2.35 eV for 0.00 to 0.07 Cr concentration and increased for 0.09. The decrement in the band gap with increasing Cr content is attributed to sp-d interactions and smaller average grain size of Cr-doped ZnO nanoparticles. The magnetization varied in the similar trend of variation as in the related band gaps with increasing Cr content. The room temperature ferromagnetic behavior is ascribed to the intrinsic property of Cr-doped ZnO nanoparticles. Room temperature ferromagnetism makes Cr-doped ZnO nanopowders attractive to the potential applications of the spin-based electronic devices.