PHYSICAL REVIEW B 1 JUNE 2000-IVOLUME 61, NUMBER 21 Field and temperature dependence of magnetization in FeCu-based amorphous alloys P. Crespo, M. Multigner, F. J. Castan˜o, R. Casero, and A. Hernando Instituto de Magnetismo Aplicado, P.O. Box 155, Las Rozas, 28230 Madrid, Spain A. Garcı́a Escorial CENIM-CSIC, Avenida Gregorio del Amo 40, 28040 Madrid, Spain L. Schultz Institut für Metallische Werkstoffe, Helmholtzstraße 20, D-01069 Dresden, Germany S. N. Kaul School of Physics, University of Hyderabad, Central University PO, Hyderabad 500134, India ~Received 22 February 2000! In this paper, the production of FeCu-based FeCuZr amorphous alloys by ball milling is reported. The thermal dependence of magnetization for the~Fe0.5Cu0.5!85Zr15 (at. %) amorphous alloy has been found to show a dramatic field dependence of the kink point of the magnetization. This kink corresponds to a tempera- ture different from the Curie temperature, above 400 K, of the ferromagnetic phase, which, according to spin waves fitting, can be induced by applying external fields. Just above 235 K, the thermoremanence increases sharply, and this feature strongly suggests an increase of the ferromagnetic ordering under zero field heating. Neutron diffraction experiments seem to confirm the enhancement of spin alignment. The thermal expansion above the compensation temperature is proposed to be the origin of the thermoremanence enhancement through the anti-Invar effect as might be explained within the framework of recentab initio calculations@M. van Schilfgaardeet al., Nature~London! 400, 46 ~1999!#. iv a- um ng n- m sp fo no d s el ior de bit n n ll als d e o a ys m ly- he see a e s is ion a ded i- ow - of d -Fe per- vice IB ea- re- ure C less ing The spin res - om The FeCu system is a typical one with a high and posit enthalpy of mixing.1 However, extended regions of met stable solubility can be obtained via several nonequilibri processing techniques, such as vapor-quenchi2 sputtering,3 and mechanical alloying.4,5 The crystalline struc- ture of the FexCu12x solid solutions depends on the Fe co tent, being fcc forx,60 at. %, bcc forx.70 at. %, whereas a mixture of fcc and bcc phases is observed for the inter diate composition range. The fcc alloys have attracted a cial interest because they exhibit ferromagnetic behavior a wide range of compositions~above 20 at. % in Fe!, al- though fcc-Cu or fcc-Fe at their ground state, are ferromagnetic.3,6–8In particular, EXAFS experiments carrie out in fcc-Fe0.5Cu0.5 ~Ref. 9! have shown that the lattice i expanded with respect to the fcc Cu lattice, therefore is lik that Invar effects10 are related to the ferromagnetic behav of fcc-FeCu alloys. To induce amorphicity in the FeCu system we have ad 15 at. % of Zr. Notice that FeZr amorphous alloy exhi typical Invar properties as high volume magnetostrictio11 and a remarkable Curie temperature dependence on the drostatic pressure.12 Amorphous powders of nominal compositio ~Fe0.5Cu0.5!85Zr15(at. %) were obtained by high energy ba milling in a planetary mill with hardened stainless-steel vi and 10 mm diameter stainless-steel balls in a ball-to-pow weight ratio of 15:1. In order to avoid the oxidation of th powder during the milling, the vials were sealed under arg atmosphere prior to the milling. The composition of the fin powder was checked by energy dispersive x-ray anal ~EDX!. No traces of others elements, such as Cr or Ni, co ing from the milling vial or balls were detected in the ana sis of the EDX results. X-ray-diffraction pattern shows t PRB 610163-1829/2000/61~21!/14346~4!/$15.00 e , e- e- r t y d hy- er n l is - typical broad halo characteristic of amorphous systems; Fig. 1~a!. The XRD pattern of a nanocrystalline of fcc-~Fe0.5Cu0.5!93Zr7 (at. %) solid solution, with an averag grain size of 6 nm, prepared under the same condition also shown for comparison. The specimens for transmiss electron microscopy~TEM! studies were prepared mixing small amount of the material with acetone and the suspen particles collected on a grid, which was left to dry and d rectly observed in the microscope. TEM observations sh featureless microstructures, Fig. 1~b!, whereas the corre sponding diffraction pattern shows a ring characteristic the amorphous state, Fig. 1~c!. DSC measurements showe that the alloy starts to decompose around 500 K into bcc and fcc-Cu rich phases. Magnetic characterization was formed using a superconducting quantum interference de ~SQUID! and a vibrating sample magnetometer~VSM!. Neutron diffraction experiments were carried out at the D diffractometer of the ILL in Grenoble~France!. The temperature dependence of the magnetization m sured under different applied fields is shown in Fig. 2~for low applied fields! and Fig. 3~for high applied fields!. The ZFC/FC curves shown in Fig. 2 exhibit the onset of the ir versibility between the ZFC and FC curves for a temperat around 225 K. Below this bifurcation temperature, the ZF curve decreases while the FC curve remains more or constant. For applied fields higher than 100 mT, the splitt between the ZFC and FC curves disappears; see Fig. 3. observed behavior points out the existence of several configurations degenerated in energy at low temperatu and weak applied fields. At low applied fields~Fig. 2! the apparent Curie tempera ture, Tco , of the amorphous phase, roughly estimated fr 14 346 ©2000 The American Physical Society ld t b an d e w ro ag- be he- field ot be lied h d d s is the n- Fe - ag- der hat - s ple es se ave ence he ag- nd le to er- it until ence an tic of PRB 61 14 347BRIEF REPORTS the minimum of conventional kink point plots (dM/dT vs T! using the FC measurements, is 235 K. For high applied fie ~Fig. 3! the Curie temperature is above 400 K and can no reached without crystallising the sample. The more remarkable characteristics of the thermal field dependence of the magnetization can be summarize follows: ~i! the enormous and anomalous field dependenc the apparent Curie temperature—notice that it increases an applied field of 1 T more than one hundred degrees;~ii ! at FIG. 1. ~a! XRD pattern of the amorphous (Fe0.5Cu0.5!85Zr15 alloy and of the fcc-~Fe0.5Cu0.5!93Zr7 solid solution;~b! TEM mi- crograph and~c! the corresponding electron diffraction pattern the amorphous (Fe0.5Cu0.5!85Zr15 alloy. FIG. 2. ZFC curves~open circles! and FC curves~solid circles! with low applied fields of 5 and 10 mT. s e d as of ith low fields ~Fig. 2! the magnetization does not drop to ze even for temperatures well above the kink point of the m netization. Magnetic impurities as Fe nanocrystals could reasonably invoked to account for such behavior. Nevert less, the Curie temperature dependence on the applied as well as the thermoremanence data shown below can n attributed to impurity effects. The thermomagnetization data obtained for an app field of 5 T ~see Fig. 3! can be fitted according to a Bloc law, M (T)5M (0)@12BT3/2#. The results of this fitting are M (0)564.07 emu/g andB59831026 K3/2 and the value obtained forB is in good agreement with those reporte earlier13 for highly disordered systems. Thus it is inferre that the low temperature demagnetization under high field dominated by spin-wave excitations thereby indicating ferromagnetic character of the spin configuration. Additio ally, the experimental value of the magnetic moment per atom obtained at 5 K is mFe51.74mB , where we have as sumed no contribution of the Cu and Zr atoms to the m netization. It turns out that the ferromagnetic order is achieved un the action of high applied fields. Therefore,Tco is not a true Curie temperature. Furthermore, it is important to note t the observedTco is in good agreement with those Curie tem peratures reported14–18 for Fe-rich FeZr amorphous alloy with Fe content close to the FeCu content of the sam presently investigated. It is likely that these reported valu also might correspond to compensation temperatures.19 In order to inquire into the nature of the magnetic pha transition observed for temperatures above 235 K, we h measured the temperature dependence of thermoreman ~TRM! for a temperature range between 5 K and 800 K~see Fig. 4!. This experimental data was obtained by cooling t sample from room temperature to 5 K under an applied m netic field of 10 mT. The field was then removed at 5 K, a the remanence was measured while heating up the samp 800 K. The residual magnetic field in the SQUID was det mined to be as low as 0.2 mT. An unexpected behavior of the TRM is found since increases spontaneously for temperatures above 235 K it reaches a plateau. The observed increase of the reman in the absence of an applied field can only be ascribed to increase of the net magnetic moment. According to data shown in Fig. 4, the zero field magne FIG. 3. ZFC~open circles! and FC~solid circles! processes for high applied fields of 100 mT, 500 mT, 1 T, and 5 T. te ld m a et e se e a on e pe ac ro n ro io tu o bl C e g- u ts , ing ing ing ie to ic ic e of to a ture rop he the ol- der th that In- lied ust to rmal - er ear tur to ols line 14 348 PRB 61BRIEF REPORTS structure below 235 K is not ferromagnetic and is charac ized by a negligible average moment. At low applied fie some average magnetic moment is induced. Since the ment observed at low temperature is induced by weak plied fields, there is not contradiction between the magn zation increase at zero field~Fig. 4! and the magnetization decrease at 5 mT~Fig. 2! observed at 235 K. Notice that th magnetization just above 235 K is 0.25 emu/g in both ca Further stronger evidence of the increase of net magn moment associated with the TRM increase has been tempted by neutron diffraction experiments. The diffracti patterns, measured at temperatures aroundTco using steps of 5 K, show a broadS(Q) halo owed to the structure of th sample. Two sets of spectra were measured at each tem ture ~solid and open symbols in Fig. 5!, to insure that the observed behavior was not due to any experimental artif As depicted in Fig. 5 the magnitude of the integrated neut diffracted intensity surprisingly increases aroundTco . This anomalous behavior suggests a reinforcement of the magnetic moment or spin alignment and thus of the fer magnetic order achieved just aboveTco when heating at zero field. The overall behavior can be summarized as follows:~i! by applying high magnetic fields a ferromagnetic configurat over the whole temperature range, with Curie tempera above the crystallization temperature, is achieved;~ii ! at zero field the magnetic structure below 235 K corresponds t noncollinear or disordered spin arrangement with negligi average moment;~iii ! according to differences between ZF and FC curves, the low field spin configurations are deg erated belowTco ; ~iv! weak fields induce an average ma netic moment below 235 K which suddenly drops, witho disappearing, at this temperature; and~v! zero field heating ~TRM! as well as zero field neutron diffraction experimen FIG. 4. Thermoremanence after cooling from room tempera to 5 K under a magnetic field of 10 mT. . r- s o- p- i- s. tic t- ra- t. n et - n re a e n- t point out that just aboveTco , in absence of any applied field a reinforcement of the spin alignment occurs. In summary, it has been shown that mechanical alloy enables the production of amorphous FeCu~Zr! alloys. The magnetic properties of this alloy can be understood by tak into account two magnetic structures:~a! a low field non- collinear ground state with different degrees of spin cant and compensation temperature of 235 K, and~b! a high field induced collinear ferromagnetic configuration, with Cur temperature above 400 K. These results might be tentatively explained according a new idea derived from recent ‘‘ab initio’’ calculations car- ried out by van Schilfgaardeet al.20 related to Invar effect at 0 K in crystalline samples. According to that the interatom distance is related not only to the amplitude of the atom magnetic moment but also to the spin canting. In the cas the sample reported here the ground state corresponds noncollinear spin arrangement. The increase of tempera could give rise to an increase of canting producing the d of magnetization around 235 K. At high applied fields t ferromagnetic configuration is induced as evidenced by accomplishment of the Bloch law. Even though magnetov ume measurements cannot be performed in our pow sample the similarity of its overall magnetic behavior wi the Fe rich FeZr amorphous alloys allows us to suggest the ferromagnetic phase corresponds to the high volume var configuration. It has also been shown that by heating at zero app field a reinforcement of spin alignment could be induced j aboveTco . We propose that this anti-Invar effect is due volume expansion enhanced by the increase of the the expansion coefficient atTco . Probably the controversial fea tures of Fe rich FeZr amorphous alloys13–17 could be ex- plained by a similar continuous transition, induced by eith temperature or applied field changes, between noncollin to collinear spin configurations. e FIG. 5. Integrated neutron diffraction intensity corresponding a (Fe0.5Cu0.5!85Zr15 amorphous sample. Open and solid symb correspond to different sets of diffraction patterns. The dashed is shown as a guide to the eye. a, - 1M. Hansen,Constitution of Binary Alloys~McGraw-Hill, New York, 1958!, p. 580. 2K. Sumiyama, Y. Yoshitake, and Y. Nakamura, Acta Metall.33, 1785 ~1985!. 3C. L. Chien, S. H. Liou, D. Kofalt, W. Yu, T. Egami, and T. R McGuire, Phys. Rev. B33, 3247~1986!. 4K. Uenishi, K. K. Kobayashi, S. Nasu, H. Hatano, K. N. Ishiar and P. H. Shingu, Z. Metallkd.83, 132 ~1992!. 5A. R. Yavari, P. J. Desre´, and T. Benameur, Phys. Rev. Lett.68, 2235 ~1992!. 6A. Hernando, P. Crespo, A. Garcı´a-Escorial, and J. M. Barandia rán, Phys. Rev. Lett.70, 3521~1993!. ı . o . ys. a, a- PRB 61 14 349BRIEF REPORTS 7P. Crespo, A. Hernando, R. Yavari, O. Drbohlav, A. Garc´a- Escorial, J. M. Barandiara´n, and I. Oru´e, Phys. Rev. B48, 7134 ~1993!; P. Crespo, A. Hernando, and A. Garcı´a-Escorial,ibid. 49, 13227~1994!. 8J. Z. Jiang, Q. A. Pankhurst, C. E. Johnson, C. Gente, and Bormann, J. Phys.: Condens. Matter6, 227 ~1994!. 9V. G. Harris, K. M. Kemner, B. N. Das, N. C. Koon, A. E Ehrlich, J. P. Kirkland, J. C. Woicik, P. Crespo, A. Hernand and A. Garcı´a-Escorial, Phys. Rev. B54, 6929~1996!. 10R. J. Weiss, Proc. R. Soc. London, Ser. A82, 281 ~1963!. 11J. Arcas, A. Hernando, J. M. Barandiara´n, M. Schwetz, and R. Grössinger, Appl. Phys. Lett.73, 2509~1998!. 12J. M. Barandiara´n, P. Gorria, I. Oru´e, M. L. Fdez-Gubieda, F Plazaola, and A. Hernando, Phys. Rev. B54, 3026~1996!. R. , 13S. N. Kaul, J. Phys.: Condens. Matter3, 4027~1991!. 14S. N. Kaul, Phys. Rev. B27, 6923~1983!. 15D. H. Ryan and J. M. D. Coey, Phys. Rev. B35, 8630~1987!. 16N. Saito, H. Hiroyoshi, K. Fukamichi, and Y. Nakagawa, J. Ph F: Met. Phys.16, 911 ~1986!. 17S. N. Kaul, J. Phys. F: Met. Phys.18, 2089~1988!. 18J. J. Rhyne, J. H. Schelleng, and N. C. Koon, Phys. Rev. B10, 4672 ~1974!. 19A. Slawska-Waniewska, P. Nowicki, H. K. Lachowicz, P. Gorri J. M. Barandiara´n, and A. Hernando, Phys. Rev. B50, 6465 ~1994!. 20Mark van Schilfgaarde, I. A. Abrokosov, and B. Johansson, N ture ~London! 400, 46 ~1999!.