DYNAMIC SCALING IN THE ISING SPIN GLASS Fe0.25Zn0.75F2

Abstract

The ac parallel susceptibility (w, T) was measured in Fe,Znl-,F2 with x = 0.25 and 0.10. X' (w, T) exhibits all features of a canonical spin glass: it has a peak in T which reduces in amplitude and shifts to higher T as w increases; as Ho increases the peak is reduced but does not shift in T. A dynamic scaling analysis for the 0.25 sample yields /3 E 1 and zv E 7.5, exponents appropriate for an Ising spin glass. It has been shown recently (I, 21 that the ran- domly diluted antiferromagnet Fe,Znl-,F2 with x < 0.3 is a short range Ising spin glass. This is a most in- teresting finding because it is known that this system with x < 0.4 under an external uniform fields Ho is an excellent realization of a random-field Ising model (RFIM) system (3). Since spin glasses and random field systems have quite different grounds states, nonequi- librium and critical behaviour, it is remarkable that the same system can display the properties of both types of random magnets. The previous studies of the spin glass phase of Fe,Znl-,F2 were based on mea- surements of the dc magnetization. Here we present ac susceptibility x (w) studies which confirm its spin glass nature. We have measured the temperature (T) dependences of the real (w) and imaginary (w) parts of the sus- ceptibility parallel to the c-axis of Fe,Znl-,F2 using the standard ac inductance method. The data were taken mostly in Ho = 0 at several frequencies w / 271 between 17 Hz and 6 970 Hz with a driving field of m 1 Oe RMS. Only at 155 Hz, where the signal to noise ratio of the bridge is maximum, we took data at several values of Ho applied parallel to the ac field. Figure 1 shows the overall behavior of (T) for the x = 0.25 sample at several frequencies in zero field. Similar results were obtained for the x = 0.10 sam- ple. Similarly to other canonical spin glasses (4), X' (T) has a cusp at Tf (w) which shifts to higher temper- atures, broadens and reduces in amplitude as w in- creases. Note that this is in marked contrast to the behaviour of (T, w) in the RFIM phase (x = 0.46) which shows no shift in the cusp temperature with varying frequency (5). The inset in figure 1 shows the field dependence of (T) at 155 Hz. The smearing of the cusp with increasing field with no shift in the peak