Exchange Bias and Spin-Glass-Like Ordering in "-Fe 3 N–CrN Nanocomposites N. S. GAJBHIYE 1;2  and Sayan BHATTACHARYYA 1 1 Department of Chemistry, Indian Institute of Technology, Kanpur 208 016, U.P., India 2 Institute of Nanotechnology, Forschungszentrum Karlsruhe, P.O. Box 3640, 76021 Karlsruhe, Germany (Received July 31, 2006; accepted November 28, 2006; published online March 8, 2007) "-Fe 3 N–CrN nanocomposite system is synthesized by wet chemical technique. Exchange bias coupling exists at the interface of the ferromagnetic (FM) "-Fe 3 N and antiferromagnetic (AF) CrN phases and the FM "-Fe 3 N and AF Fe–O–N/Cr–O–N surface layers. However, the exchange bias is not found to be dependent on the CrN concentration and the coupling of the spins at the surface Fe–O–N/Cr–O–N–"-Fe 3 N has a dominating influence. The maximum shift in hysteresis loop is observed to be 78 Oe and the low magnitude of exchange bias is attributed to the roughness of AF–FM interface that induces spin- disorder and random exchange anisotropy. The relaxation dynamics study indicates the presence of antiferromagnetic ordering in this system. Below 65 K, the broad peak in zero-field cooled (ZFC) magnetization curves (T E ¼  T f ) indicates the presence of unidirectional anisotropy and spin-glass-like ordering. The spin-glass-like ordering, which arises from the freezing of the clustered randomly oriented spins, is confirmed from the frequency dependence in AC susceptibility measurements and dynamic scaling analysis. [DOI: 10.1143/JJAP.46.980] KEYWORDS: nanocomposite, nitride, exchange bias, spin glass 1. Introduction The interactions at the antiferromagnetic–ferromagnetic (AF–FM) interface in atomic proximity have been a subject of extensive research over the last few decades. 1,2) The hysteresis loop 'ðHÞ of FM is generally centered at the origin with 'ðHÞ¼'ðHÞ, where ' is the magnetization and H is the applied field, satisfying time reversal symme- try. 3) However, in an AF–FM system, when the system is cooled through the Ne ´el temperature (T N ) of the AF, in the presence of an applied field, the hysteresis loop is offset from zero applied field, i.e., it exhibits unidirectional anisotropy. This phenomena known as exchange bias, depends on the orientation of the spins at the AF–FM interface. The exchange bias phenomenon has received an increased attention in the recent years, both from the theoretical and the experimental point of view. 1,2) One of the main reasons for the continued interest is the fact that different experimental techniques may yield different aniso- tropy values. 4) Notwithstanding the large number of inves- tigations, complete understanding of the magnetic coupling at the AF–FM interfaces is still unavailable. 5) Exchange bias is mainly studied in thin films having AF–FM bilayers. Some of the different bilayers studied, are NiO/NiFe, 6) Co– FeMn, 5) Co/CoO, 7,8) oxidized Ni–Fe, 9) Fe/FeF 2 , 10) Fe/Cr double superlattices, 11) and few on the nanoparticles. 12–17) The exchange bias systems that are studied for nano- particles/polymer composites are Cr/Cr 2 O 3 , Co/CoO, Fe/ Fe–O, Cu/CuO, and MnFeCo. In the nanoparticles, the AF– FM interface is rough 18) and results in spin flipping due to disorder in the spin orientation and reduction in the overall unidirectional anisotropy. The randomness in the orientation of spins gives rise to spin freezing at lower temperature and the spin-glass state is observed. 19) However, the exchange bias phenomenon and spin-glass studies are very rare in nitrides, and more so in Cr containing nitride materials. The exchange bias is reported in Fe–Fe 2 N 20) and Co–CoN 21) nitrides, while the spin-glass studies are reported for "-Fe 3 N nanoparticles and frozen fluids. 22) Since the "-Fe 3 N–CrN nanocomposites (with AF CrN and FM "-Fe 3 N phases) exhibit such phenomena, it represents an interesting system for study. 2. Methodology and Structural Characterization The "-Fe 3 N–CrN system was prepared by mixing aqueous solutions of 0.22 (N) Fe(NO 3 ) 3 . 9H 2 O, 0.2 (N) Cr(NO 3 ) 3 . 9H 2 O and citric acid in desired proportions according to the stoichiometric Fe/Cr atomic percentage ratios of 93/07, 86/ 14, 79/21, and 72/28. The citrate gel was evaporated to give the Fe–Cr-oxide precursors, which were nitrided in flowing NH 3 (g) at a flow rate of 240 cm 3 /min at 983 –1103 K for 24 h. The XRD-Rietveld and chemical analyses indicate that "-Fe 3 N/CrN, 81/19, 68/32, 58/42, and 53/47-nanocompo- site systems were obtained for the Fe/Cr atomic percentage ratios of 93/07, 86/14, 79/21, and 72/28. For "-Fe 3 N/ CrN, 81/19 system, the solid solution with composition, "-Fe 2:8 Cr 0:2 N was formed crystallizing in the space group P6 3 =mmc. The lattice parameter of CrN (space group: Fm  3m) was the same as the reported value of a ¼ 4:14 A ˚ . 23) For the "-Fe 3 N phase (space group: P6 3 =mmc), lattice parameters varied slightly from the reported values of a ¼ 2:695 and c ¼ 4:362 A ˚ . 24) The X-ray diffraction (XRD) patterns are shown in Fig. 1. The average particle size for the nanocomposites is 25 nm as observed from transmission electron microscope (TEM) studies (Fig. 2). X-ray photo- electron spectroscopy (XPS) [carried out with Al K radiation (1486.6 eV) for the representative "-Fe 3 N/CrN, 53/47-nanocomposites] studies showed the presence of "- Fe 3 N, CrN, oxynitride (Fe–O–N/Cr–O–N) and Fe–O phases at the surface. The XPS depth-profiling (carried out with 4 keV Ar þ ion sputtering) showed the presence of elemental Cr, which remains distributed within the CrN matrix, although the Cr peaks are not observed in the XRD patterns. Figure 3 shows the XPS spectra for the 5 nm sputtered sample. The peak fitting was performed with different reasonable parameters, keeping the standard full-width-at- half-maximum (FWHM) and the binding energy values for the different states into consideration and Fig. 3 represents the best fit to the data. For the 5 nm sputtered sample, the Cr 2 p 3=2 spectrum is deconvoluted into three peaks. The  E-mail address: nsg@iitk.ac.in Japanese Journal of Applied Physics Vol. 46, No. 3A, 2007, pp. 980–987 #2007 The Japan Society of Applied Physics 980