Magnetic Field Dependence of the Small-Angle Neutron Scattering (SANS) in Amorphous DyCu and NdFe2

The magnetic correlations and the response of the spin system to an applied field have been studied in sputtered amorphous NdFe~ using high-resolution small angle neutron

This study was undertaken to obtain additional information about them.
Small-angle neutron scattering data-on amorphous NdFe~ were taken in zero field and in applied fields up to 18 kOe~ In zero field, the lineshape of the scattering was Lorentzian above the transition temperature and became increasingly deviated from this· form below this temperature. The spin correlation length showed a broad maximum near to, but loweI than, the transition temperature and then decreased below this temperature. SANS data obtained with the applied field s.howed a sharp decrease in intensity with increasing field. The spiL correlation length also decreased with increasing field.
The response of the SANS from amorphous DyCu to applied fields up to 15 kOe was examined. Application of the field caused a decrease in SANS intensity with .1..L.L increasing field and also caused a decrease in the correlation length which was obtained from the lineshapes which were Lorentzian to 6 A from the zero field value of c 13.5 Vi. vii LIST OF FIGURES 1. Scattered neutron intensity versus temperature in amorphous NdFe 2 for several small Q . . . . . . . . . . . . . 12 2. f h . . .
In particular ferromagnetic spin ordering will not occur at any temperature in a system with finite random anisotropy  [2]. At lower temperatures an enormous increase in the scattering occurs which re£lects the formation of spin clusters as opposed to ferromagnetic order for which the SANS would nearly vanish below T at these Q. Figure 2 plots the inverse intensity versus Q-squared for temperatures both above and below the transition. In a conventional ferromagnet the scattering for T > Tc has a Lorentzian lineshape reflected as straight lines in Figure 2. Note that for T > Tc the lineshape is Lorentzian, but as T is decreased through Tc the data become increasingly curved.
We have analyzed this data using a Lorentzian plus Lorentzian-squared cross section which is appropriate for a random anisotropy system, as proposed by Aharony and Pytte [1 ]. · Its form is + ( 1) where K is the i.n verse of the spin correlation length.. The tem~erature dependences of the parameters A and B are shown in Figure 3-Note that there is a sharp peak in the Lorentzian coefficient at 340 K, which is approximately b the critical temperature as derived from the peak in the scattering intensity. Also observe that, at lower temperatures, the values of both coefficients increase sharply~ Values of the spin correlation length as a function of temperature axe also given in Figure 3. Note that the correlation length has a broad maximum of 170 angstroms near to, but lower than, the critical temperature.
Below this temperature the correlation length decreases to approximately 80 angstroms at 100 K, after which it remains consta.n t.
These results may be understood as follows.
As T is decreased from above T~, the alloy appears to approach a conventional ferromagnetic phase transition. This is reflected in the rise in the correlation length as Tc is approached from above. However, below Tc the presence of the random anisotropy on the rare earth sites precludes the correlation of spins beyond a certain length, resulting in clusters of spins. As the temperature is £urther decreased, the effective local random anisotropy increases, thereby decreasing the correlation length even further.
The low T state of this amorphous alloy may be classed as a spin glass.

FIELD DEPENDENCE
In zero field, constant intensity lines on the SANS area detector appear as circles (isotropic scattering)-As

CONCLUSIONS
These results plus the previous magnet~zation data imply that NdFe~ does not undergo a second order phase transition to a ferromagnetic state, but rather exhibits a spin-frozen cluster state which has a strong sensitivity to applied fields. This is consistent with the prediction of Aharony and Pytte [1] that, in a random anisotropy alloy, no ferro-magnetic phase transition is possible in three dimensions. Appl. Phys. 50, 1958Phys. 50, (1979    .   the fiela caused a decrease of the total SANS intensity and a decrease 0£ the correlati~n length derived from the Lorentzian lineshapes to -6 angstroms from the zero field value of 13.5 angstroms.
Field-cooling and relaxation effects were compared with measurements.
also observed in the SANS intensity and are similar phenomena in magnetization

INTRODUCTION
Recently we reported [1] the observation of small angle neutron scattering
Since magnetization measurements on this compound have revealed unusual field cooling and relaxation ef£ects, we thought it would be interesting to explore the response o:f the SANS to an applied field, and aiso to note whether this response is different from other amorphous compositions (3)(4) with stronger exchange forces.

RESULTS AND DISCUSSION
The most obvious effect of the applied field was a sharp decrease of the total SANS intensity, as depicted in Figure 1, which shows data taken after cooling in zero field. Upon removal of the field and going through a demagnetization cycle, the intensity reverted to its initial value .. Different behavior was observed when the sample was cooled through T in a field of 4 kOe.. As indicated in the figure, the intensity re:sulting from this procedure was smaller, and, after removal of the field, did not regain its original value. However, after a time lapse of about ~O minutes the intensity had relaxed considerably towards its zero field value and eventually reached essentially the same level.
Accompanying this decrease in intensity was a marked decrease in the correlation length, as shown in Figure   correlation length decreases to about 6 angstroms from its zero field value of 13.5 angstroms and then appears to make a sli ght but not necessarily sta.tistically significant Although these results by no means paint an unambiguous picture of the magnetization distributions in this compound, they suggest several qualitative features and mechanisms.
The decrease in scattered intensity observed here is similar to that observed in the amorphous compositions TbFea reported by Rhyne and Glinka [3], and NdFe~ reported by Spano et al [4]. Both these compositions are different in that, because of the high iron concentration, they display higher ordering temperatures and larger magnetic moments.. In those cases the decrease in SANS intensity was tied to formation of an infinite percolating ferromagnetic cluster, which, since its lineshape is a delta function in q-space, removes intensity from the window seen by the detector-Undoubtedly, this is occurring i.n the present instance as well, since we know from magnetization measurements (2] that a large moment is developed by a field of a £ey kOe.
In the previous work on TbFeoz [ 3 ] significant a.nisotropy was found in the scattering contours on application of the external magnetic field. The zero field circular constant intensity contours became elliptical in. a field with the major axis along the field direction.
Counting statistics on this very small sample of Dycu precluded a quantitative deter~ination of any scattering anisotropy, however within experimental error, no scattering a.D.isotropy was observed.
our final comment refers to the relaxation effects observed in the field cooled experiment, which we believe are the first example of such effects observed in SANS~ similar time dependent effects observed by Coey et..al.[2] in the remanent moment of this comp?sition, but the time constant for the situation most closely related to the present experiment is much longer.  Saturation of the magnetization was not achieved even at fields on the order of 150 kDe, which indicates that the ordering was not ferromagnetic. These effects suggested that the alloy had random and anisotropic spin freezing which suggested an asperomagnetic spin glass-like magnetic structure below its transition temperature.
Additional evidence that some sort of phase transition occurs in amorphous DyCu vas found in a small-angle neutron scattering (SANS) experiment performed on this alloy by pickart et al [2]. In this study, a break was observed in the zero field scattering patterns at a temperature roughly that of the transition temperature thus indicating some sort of phase tra~sition.
Previous studies on the other amorphous alloy, NdFe~, have also provided information on magnetic behavior in a Knowledge about the structure ·of amorphous metallic alloys is important to both scientists and engineers.
Engineers are interested in the structure of these •aterials because of the relationship between properties and the structure, scientists because this knowledge will lead to a better understanding 0£ the solid state.
Some classes of metallic glasses [10] are ( the transition metal -zirconium or hafnium (TM-Zr-Hf) alloys, Also, these metallic glasses are particularly interesting because complex magnetic orderings are expected in certain compositions of these materials because of effects such as random anisotropy.

.preparation
preparation techniques for the formation of amorphous aetallic alloys can be grouped into two categories: (1) direct quenching from the melt (process of solidification by rapid cooling from the liguid state ) and (2)  and sulfates which makes them attractive for use in marine and medical environments as cables, chemical filters, and scalpels a~ong other things.
One area of large application for these alloys is as transformer cores. The properties of low power losses and low coercivity as well as the ease of large scale production by melt-spinning make them attractive for this application. Also, high permeability and high yield stress has led to the utilization of these alloys as magnetic shielding_ In particular, the amorphous RE-TM alloys are being considered for use as magnetic bubbles for the storage of data in computer memory because of their low thermal conductivity and lack of grain boundaries, which are requirements for the alloys, the size of storage of data. By using these tne data storage area can be reduced thereby leading to smaller components for computer memory devices, an ongoing concern of many engineers. The large spin-disorder scattering also makes them attractive as aaqneto-resistance sensors. Amorphous speriaagnets can be regarded as canted or noncollinear ferrimagnets. In these materials, one of the magnetic sublattices usually has its spins aligned parallel as in a ferromagnet while the other magnetic sublattice has its spins frozen into random orientations that are not collinear. There are two types of sperimagnets that are distinguishable. For illustrative purposes, the rare-earth transition-metal alloys, Dyco 1 and Ndco 3 , will be examined.
In the one type, Dyco 3  However, some magnetic atoms may still belong to smaller clusters whose magnetic moments within the cluster are correlated but whose net cluster moment is free to fluctuate in many directions. (1) a cusp in both the ac and de susceptibility at the transition or spin-freezing temperature [ 12] (2} sensitivity of the susceptibility to applied magnetic field at and near t~e transition temperature.. The application of a field on the order of a few tens of gauss completely smears out the susceptibility cusp. Also, time dependent and history dependent effects are observed in the low-field magnetization [17]. (3) the absence of an anomaly,which is expected for a phase transition, in the specific heat.
The extra specific heat due to the impurities is a linear function of T, C = ~ T, where ¥ is independent of concentration and much larger than the ordinary term of normal metals (18]. Because of (c) , the absorption of neutrons by matter is often very small, and large sample volumes can be studied by this technique [20]. in addition, for amorphous NdFe~, the applied field dependence is also different from that expected for a ferromagnet.

exhibited by
The behavior o.f this a.lloy is similar to that a system of super-paramagnetic spins or clusters. Thus, an examination of the field dependent data result in t~is alloy being considered a system of spin clusters of varying size, whose larger clusters are driven into a near-infinite ferromagnetic cluster by the application of an external field.
Since this cluster will scatter only infinitesimally near to Q : 0 , only the smaller clusters will contribute to the SANS patterns.
This accounts for the resultant decrease in scatteriHg with the application of a field.
The applied field dependence that is saen in amorphous Dycu is similar to that observed in TbFe~ [23] and NdFe~ [24]in these alloys, this behavior was attributed to the formatioD of a near-infinite ferromagnetic cluster with regions of smaller clusters that do not become part of the near-infinite cluster.
The relaxation effects observed after the field-cooling in Dycu are believed to be the first observance of such effects in SANS.

F.SIJMMARY
The SANS measurements demonstrate that amorphous NdFe~ and DyCu do not have a magnetic phase transition to long-range ferromagnetism. .They indicate that these alloys undergo a transition to a state that has some features of a spin glass. This behavior is consistent with what is predicted for a strong random anisotropy system.