cond-mat0504275/lfm.tex
1: %\documentclass[twocolumn,aps,showkeys,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2: \documentclass[preprint,aps,showkeys,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
3: 
4: %\documentclass[preprint,aps]{revtex4}
5: %\documentclass[preprint,aps,draft]{revtex4}
6: %\documentclass[prb]{revtex4}
7: 
8: \usepackage{graphicx}
9: \usepackage{dcolumn}
10: \usepackage{bm}
11: 
12: %\nofiles
13: 
14: \begin{document}
15: 
16: %\preprint{APS/123-QED}
17: 
18: \title{Role of heterostructures and multiple magnetic phases in 
19: the low-field magnetization of Fe-Cr GMR multilayers}
20: 
21: \author{R. S. Patel}
22: \author{A. K. Majumdar}
23:  \email{akm@iitk.ac.in}
24: \affiliation{Department of Physics, Indian Institute of Technology,
25: Kanpur-208 016, India}
26: 
27: \author{A. K. Nigam}
28: \affiliation{Tata Institute of Fundamental Research, Homi Bhabha Road,
29: Mumbai-400 005, India}
30: 
31: \author{D. Temple}
32: \author{C. Pace}
33: \affiliation{MCNC, Electronic Technologies Division,
34: Research Triangle Park, North Carolina, 27709}
35: 
36: 
37: 
38: %\date{\today}
39:              
40: 
41: \begin{abstract}
42: 
43: Zero-field-cooled (ZFC) and field-cooled (FC) magnetization
44: along with ac-magnetization vs. temperature
45: and m-h loop measurements are reported 
46: for two series of ion-beam sputtered Fe-Cr GMR multilayers
47: where the interface roughness is different. The exchange 
48: coupling between the Fe layers varies from ferromagnetic (FC) 
49: to antiferromagnetic (AF) depending upon the Cr layer
50: thickness. The ZFC and FC magnetization data 
51: follow different curves below an irreversible temperature 
52: ($T_{irr}$). The FC data shows a $T^{3/2}$ thermal demagnetization
53: behavior at lower temperatures with very small spin-wave stiffness constant 
54: (as compared with that of bulk Fe obtained from 
55: Bloch's $T^{3/2}$ law) but it goes as 1/T at higher temperatures (above ($T_{irr}$)). 
56: This behavior is interpreted in terms of the coexistence of spin-glass
57: (SG)/superparamagnetic, FM and AF phases. ac-magnetization 
58: vs. temperature shows a peak at $T_g$. This peak
59: shifts towards higher temperatures and its amplitude decreases
60: with increasing frequency of the driving ac field.
61: \end{abstract}
62: 
63: \pacs{75.70.Cn, 75.10.Nr, 75.60.Nt}
64:                              
65: \keywords{Magnetization, spin glass, Fe-Cr multilayers, GMR}
66: 
67: \maketitle
68: 
69: 
70: \section{INTRODUCTION}
71: The giant magnetoresistive (GMR) property was discovered in 1988 \cite{Baibich:1988}. Since then
72: it has become an established area of promising technological applications. Sensors
73: and read heads in hard disks are major attractions of GMR materials. Tera byte
74: capacity hard disks are not far from coming to the market. In the present 
75: Fe-Cr GMR materials ferromagnetic (FM)
76: Fe layers are stacked antiferromagnetically with non-magnetic Cr spacer layers.
77: This antiferromagnetic arrangement of the Fe layers is engineered by varying the Cr
78: spacer layer thickness. With varying Cr thickness successive Fe layers show
79: oscillatory antiferromagnetic (AF) and FM couplings.
80: In an external magnetic field $H>H_{sat}$, the Fe layers
81: of these Fe-Cr samples align ferromagnetically whereas in zero field they are in 
82: antiferromagnetic configuration.
83: Our study is focused on finding out the magnetic behavior of multilayers in low
84: external magnetic fields (up to a few hundred gauss). The low field response
85: of multilayers is important for technical applications as well as for interesting
86: physics behind it.
87: 
88: 
89: Kravtsov et al. \cite{Kravtsov:2001} studied the interface formation and magnetic
90: ordering in Fe-Cr multilayers with Fe thickness 2, 4, and 13 \AA \ by polarised
91: neutron reflectometry. They found that the samples with Fe layer thickness
92: of 2 \AA \ show pure superparamgnetic behavior. Increasing the Fe thickness
93: to 4 \AA \ led to a crossover from superparamagnetic to a mixed ferromagnetic
94: and spin-glass state. The sample with 13 \AA \ Fe layer displays the usual magnetic
95: properties of a GMR multilayer. Here one should keep in mind that 2 \AA \ is
96: too thin for a layer which will rather be a discontinuous magnetic layer. Fert et al. \cite{Fert:1995}
97: referred such type of multilayers as hybrid nanostructured multilayers because 
98: they consist of both usual layers and layers having nano-scale lateral structure,
99: like clusters. Low-field magnetization study of CoFe-Al$_2$O$_3$ multilayers by
100: Kakazei et al. \cite{Kakazei:2003} showed that depending upon the CoFe thickness
101: and temperature, the system shows superparamgnetic, spin-glass or a FM like state.
102: They also concluded a co-existence of different phases in the mixed state. An 
103: experimental investigation of the magnetic properties of multilayer (Gd/Si/Co/Si)$_n$
104: films in low magnetic fields by Patrin et al. \cite{Patrin:2003} had shown a spin-glass
105: like behavior. They explained the existence of the spin-glass state of multilayer
106: films in terms of the bi-quadratic exchange interaction. de Oliveira et
107: al. \cite{Oliveira:1999} explained the magnetic irreversibility in Fe-Cu multilayers
108: at low temperatures in terms of the presence of some interdiffusion between 
109: Fe and Cu at the interfaces.
110: They also found diffused Fe atoms aggregate in clusters at the interfaces.
111: The magnetization behavior of thin (10 - 50 \AA) epitaxial Fe films on Cr
112: studied by Berger and Hopster \cite{Berger:1994} by the magneto-optical 
113: Kerr effect revealed that the exchange coupling between Cr and Fe overlayers
114: depends very much on the intrinsic antiferromagnetic properties of Cr. At 123 K, Cr
115: shows a phase transition from longitudinal spin density wave (SDW) to
116: transverse SDW.
117: 
118: 
119: \section{EXPERIMENTAL DETAILS}
120: Our samples are grown on Si substrate by ion beam sputter deposition
121: technique using Xe ion at 900 V with a beam current of 20 mA and 1100 V 
122: with a beam current of 30 mA. The typical
123: structures are Si/Cr(50 \AA)/[Fe(20 \AA)/Cr(t \AA)]$\times$ 30/Cr(50 -t \AA).
124: Sample 1 has t = 10  \AA \ and samples 2 and 3 have t = 12  \AA \ but they are
125: deposited under different base pressure at 900 V. Different base pressure results
126: in different surface roughness. Let us call these samples 1 - 3,
127: {\it series A} samples. Samples 4 - 8 have t = 6, 8, 10, 12, and 14 \AA, \ respectively 
128: and they are sputtered at 1100 V. Let us call these samples 4 - 8, {\it series B} samples.
129: These samples 
130: are well characterised and the details have been given elsewhere
131: \cite{Lanon:2002, Majumdar:2002}.
132: All the experiments were done with a Quantum Design superconducting 
133: quantum interference device (SQUID) magnetometer
134: (MPMS). The magnetic field is applied in the plane of the multilayer samples.
135: Samples 1 - 8 have GMR $(=(\rho(H)-\rho(0))/\rho(0) \times 100 \%)$ of $\sim$ 20, 21, 21,
136: 0.4, 31, 33, 32, and 29 \%, respectively at 4.2 K in a longitudinal magnetic field $\sim$ 1 tesla.
137: ac-magnetization was done using a Quantum Design PPMS with some
138: modifications at Forschungszentrum Karlsruhe, Germany.
139: 
140: \section{RESULTS AND DISCUSSION}
141: \subsection{dc-magnetization}
142: When we cool the Fe-Cr multilayer samples from room temperature to a lower temperature
143: of interest in zero magnetic field and then apply a small magnetic field in the plane
144: of the multilayer, moments in iron layers start responding to this external magnetic 
145: field. Those moments of Fe layers, which are not aligned in the direction of external 
146: field try to align in the direction of the applied magnetic field. This gives a finite moment
147: for the whole sample. Good GMR multilayer systems are AF in the sense
148: that the FM Fe layers are coupled antiferromagnetically in zero field. When we increase the 
149: temperature from the lowest temperature, the thermal energy starts disrupting this
150: nearly perfect AF alignment. Thermal fluctuations equally affect both types of Fe layers, i.e.,
151: the Fe layers with magnetic moments aligned parallel to the applied magnetic field and 
152: those with magnetic moments aligned antiparallel to the field. However, the
153: external magnetic field will try to suppress the fluctuations of moments which
154: are in the direction of the field and will effectively increase the fluctuations of moments which are
155: aligned in the opposite direction. In other words, the decrease in the magnetization due to the 
156: thermal energy is less for Fe moments parallel to the applied field. Thus the magnetization 
157: increases with the increasing temperature. But when the temperature is beyond a 
158: critical value, the thermal energy starts disrupting all the ordered moments thus 
159: decreasing the net magnetization with increasing temperature. So for a small 
160: external magnetic field we observe a peak temperature $T_m$ as 
161: shown in Fig. \ref{fig:s1raw} for sample 1. This behavior has been seen prominently
162: in samples 1, 2, 3, 5, and 6 and presented in Table \ref{tab:dct}. This peak temperature is 
163: a function of the applied magnetic field.
164: 
165: 
166: If we cool the sample in the presence of a small magnetic field, we do not find any peak 
167: in M(T), rather M increases at lower temperatures (more ordered state).
168: This feature is similar to the history-dependent effect of spin-glass systems 
169: where zero-field cooled (ZFC) and field-cooled (FC) magnetization curves follow 
170: different paths at low temperatures and low fields. The motivation behind the present 
171: study is to explore such behavior at low fields in these GMR multilayers. $T_m$ (which approaches the 
172: glass transition temperature $T_g$ when $H \rightarrow 0$) decreases with the increasing 
173: external magnetic field because a higher magnetic field is able to align the fluctuating 
174: moments to their maximum value at lower temperatures. In the high-temperature limit ZFC and FC 
175: magnetizations show the same temperature dependence. A characterstic temperature 
176: $T_{irr}$ can be defined below which the sample shows such history dependent effects. 
177: This temperature also decreases with increasing field and $T_{irr}$ is 
178: always greater than $T_m$. These samples show another characterstic temperature, namely, a point of
179: inflection $T_{inf}$ in FC magnetization vs. T curve. Below this temperature, 
180: FC curves are convex upwards
181: and above it they are concave upwards. Thus one more characterstic temperature 
182: $T_{inf}$ can be derived apart from $T_m$ and $T_{irr}$. This $T_{inf}$ is 
183: found to be independent of the external magnetic field as shown in Table \ref{tab:dct}.
184: Similar to this work Durand et al. \cite{Durand:1993} found different 
185: magnetization in FC and ZFC measurements in Fe-Cu multilayers at low temperatures. 
186: They interpreted this phenomena by interdiffusion of Fe atoms in Cu
187: layers. They also found a very low value of the saturation magnetization,
188: nearly one order of magnitude smaller than that of bcc Fe.
189: 
190: 
191: 
192: 
193: Sample 4 did not show any $T_m$  as shown in the inset of Fig.~\ref{fig:ox3raw}.
194: It has a small Cr thickness ($t_{Cr}$) of 6 \AA \ which leads to a FM coupling between the Fe layers
195: and not an AF coupling. This fact is supported by a GMR of less than 1 \% for this sample.
196: For the same external magnetic field, it can be seen from Table~\ref{tab:dct} that 
197: $T_m$ shifts to lower temperatures from 75 to 20 K for sample 5 ($t_{Cr}$ = 8 \AA) to sample 
198: 6 ($t_{Cr}$ = 10 \AA) for an applied field of 50 Oe. This suggests that $T_m$
199: may be at much lower temperatures for samples 7 and 8. Our measurement
200: temperature range is 5 to 300 K. Sample 7 did not show any peak in a magnetic 
201: field of 200 Oe, however, in a magnetic field of 50 Oe there is a signature of a peak
202: at 8 K (notice the first 3-4 points of ZFC curve of sample 7 in Fig.~\ref{fig:ox3raw}). 
203: Sample 8 showed a totally different behavior at higher temperatures 
204: compared to the other samples in this series as shown in Fig.~\ref{fig:ox3raw}.
205: Here the magnetization increases with increasing temperature above 150 K
206: which could not be understood.
207: 
208: 
209: If we take the magnetization value ($m(5 K)_{ZFC}$) of the samples at the lowest temperature
210: from the ZFC magnetization measurements at different applied external magnetic fields, then 
211: we find that ($m(5 K)_{ZFC}$) is roughly linear with the applied field as in a 
212: paramagnet. This has been shown in Fig. \ref{fig:gm0ZFCall}.
213: 
214: 
215: From experimental point of view some major signatures of spin glasses are: (i) below $T_g$ 
216: spin-glass systems show history-dependent behavior, i.e., the magnetization
217: measured in a field-cooled condition is different from that under the zero-field-cooled 
218: condition \cite{Chowdhury:1986} and (ii) low field, low frequency ac susceptibility $(\chi_{ac}(T))$ 
219: exhibits a cusp at a temperature $T_g$. The origin of the spin-glass-like behavior here
220: may be the interface roughness and interdiffused clusters. 
221: These are sputtered samples and the surface roughness
222: may lead to Fe-Fe, Fe-Cr and/or Cr-Cr frustrations \cite{Berger:1994, Pierce:1999}.
223: There may be Fe clusters interdiffused inside Cr layers and vice versa.
224: In spin glasses, $T_g$ is the temperature below which the spins are frozen, i. e.,
225: the FC magnetization curve is flat below $T_g$. However, in multilayers, in low fields
226: AF Cr's role may be quite complex.  Bulk Cr has a N\'{e}el temperature $T_N$ = 311 K. 
227: N\'{e}el temperature of Cr layers in these multilayers can be very much different from
228: that of the bulk. It is known that the Curie temperature of Fe films is less than
229: that of bulk Fe. In Cr, as the thickness decreases, the N\'{e}el temperature also
230: decreases, either due to the decoupled AF state behaving like a thin film, or
231: due to the increasing spin frustration due to closer interfaces. The mechanism of 
232: antiferromagnetism in Cr is well explained by spin density wave 
233: (SDW) \cite{Berger:1994, Overhausser:1960}. Below
234: $T_N$, the SDW is transverse, i.e., the magnetic moments are perpendicular to the 
235: SDW wavevector. This transverse SDW state shows a transition to longitudinal 
236: SDW state at the spin-flip transition temperature $T_{SF}$ = 123 K \cite{Berger:1994}.
237: This adds to the difficulty in understanding the role of Cr in the magnetization
238: study of these multilayers. The magnetization behavior of thin Fe films on Cr
239: studied by Berger et al. \cite{Berger:1994} by magneto-optical Kerr effect gave indication that
240: the magnetization in Fe layers is oriented in the Fe-Cr film plane. In the real case, ion
241: beam sputtered layers may contain atomic steps at surfaces. These atomic steps 
242: cause frustration while aligning the magnetic moments at the interface. Figure 5 of 
243: ref. \onlinecite{Berger:1994} and Fig. 3 of ref. \onlinecite{Pierce:1999} give beautiful illustrations
244: of different atomic interface situations. In summary, we can say that in these 
245: measurements we get contributions from many sub-systems like FM
246: Fe films, AF Cr films which show SDW transition at $\sim$ 123 K (below this 
247: temperature Cr may lead to the formation of small domains in Fe films), and SG-like interdiffused 
248: clusters, spin waves, etc. 
249: 
250: 
251: The antiferromagnetic exchange coupling between two successive Fe layers
252: decreases with increasing temperature. Most of the theories assume that the coupling involves
253: only those few atomic layers in the FM which are nearest to the 
254: interfaces\cite{Heinrich:1994, Cullen:1993}. Experiments 
255: introducing other FM materials at the interfaces have confirmed this point of view to some extent.
256: Each FM layer can be divided into ``bulk" and ``surface" regions consisting of 
257: superparamagnetic and/or spin-glass particles/clusters, etc.. 
258: We find that the FC magnetization data of our multilayers fit well to 
259: the power law of the form
260: 
261: 
262: \begin{equation}
263: \frac{\Delta M}{M(0)} = \frac{M(T)-M(0)}{M(0)}=- A T^{3/2},
264: \label{eq:pwr}
265: \end{equation}
266: \noindent
267: where A is a constant of proportionality. The values of $\chi^2$ are 
268: consistent with the experimental
269: resolution. The values of $\chi^2, R^2$(correlation coefficient),$ M(0),$ and A are presented in Table \ref{tab:pwr}.
270: The typical fits are presented in Fig. \ref{fig:3by2}. The proportionality constant A,
271: which is analogous to the the spin-wave stiffness constant in Bloch's $T^{3/2}$ law,
272: depends on the magnetization. We have earlier done magnetization measurements 
273: (above H$_{sat}$ where all Fe layers are aligned in FM configuration) with temperature
274: in these multilayers \cite{Patel:2005}. We fitted the data to Bloch's $T^{3/2}$ law.
275: We found that the values of A are of the order of $1\times10^{-5}$
276: K$^{-3/2}$. In low fields, on the other hand, A is 30 times more compared to the above 
277: value. To explain this 
278: we have to consider the behavior of Fe clusters in Cr layers. It has been found by Vega et al. 
279: \cite{Vega:1994} that Fe$_N$ ($N$ = number of Fe atoms in Fe clusters) clusters can behave as a FM or an AF. 
280: The magnetic order within the Fe and Cr slabs is qualitatively the same as in the
281: corresponding pure solids and that the coupling at the interfaces is usually AF.
282: They found strong environment dependence of the local magnetic moments.
283: The Fe moments at the interface are reduced with respect to the bulk value,
284: while in the middle of the slab they are sometimes enhanced. The Cr moments
285: are not only modified close to the interface, but they are extremely sensitive to 
286: the compatibility of their spin density wave (SDW) state and the AF coupling
287: at the Fe-Cr interface. They have calculated that small Fe$_N$ clusters in Cr
288: order AF as the Cr matrix for $N\le4$. For larger Fe$_N (N \ge 6)$, the magnetic 
289: order within the Fe cluster is FM-like. So `A' value will depend on the behavior
290: of these Fe clusters. Large values of A will give small spin-wave stiffness constant.
291: Similar results were found by de Oliveira et al. on Fe-Cu multilayers. Sample 4
292: gives good fit to Eq.~\ref{eq:pwr} for the whole temperature range and in both
293: ZFC and FC cases. Here the value of the coefficient A is close to that given in 
294: ref. \onlinecite{Patel:2005}. This means that the magnetization behavior
295: of sample 4 even in low fields is more like the magnetization behavior of 
296: multilayers in the
297: presence of magnetic filed greater than saturation field.
298: 
299: Above $T_{irr}$ the magnetization decreases as 1/T in these multilayers as shown in Fig.
300: \ref{fig:onebyT}. At high temperature the thermal energy disrupts all magnetic alignments. 
301: Similar studies of low-field magnetization in ion-beam deposited
302: metal-insulator $CoFe-Al_2O_3$ multilayers were carried out by Kakazei et al.
303: \cite{Kakazei:2003}. They also reported that the low-field magnetization curve
304: M(H,T) generally displays the Curie-Weiss behavior at high temperatures 
305: and below a certain blocking temperature the FC-ZFC curves split.
306: Figure \ref{fig:onebyT} shows the typical fits to $M \sim 1/T$. 
307: We found that the $R^2 > 0.99$. We 
308: have taken the data only above $T_{irr}$ for the above analysis.
309: 
310: Now we proceed to find out the external field
311: dependence of $T_m$. We find that the de Almeida and Thouless (AT) behavior
312: of the form $H/T \propto (T_g/T-1)^{3/2}$, where $T_g$ is the spin-glass
313: temperature, gives an unique $T_g$ for each sample which is 
314: $\sim (1.2-1.5) \times T_{inf}$ of that sample.
315: Fig.~\ref{fig:at} shows the plots of $T_m$ vs. 
316: $H^{2/3}$ for samples 1, 2, 3, 5, and 6. 
317: $T_g$ is the intercept of the best-fitted straight line with the $T_m$-axis, i.e., 
318: $T_m \rightarrow T_g$ when $H \rightarrow 0$.
319: This $T_g$ is also pronounced in $\chi_{ac}$
320: measurements, i.e., we got a peak in ac-magnetization 
321: measurements at roughly the same
322: temperature as shown in Fig. \ref{fig:s2ac} and 
323: discussed in subsection~\ref{ss:acm}. 
324: 
325: 
326: \subsection{M-H loop}
327: 
328: m-h measurements of these samples are shown in Fig. \ref{fig:mh}. 
329: All the samples except 4 show typical hysteresis of AF coupled multilayers.
330: From these measurements we
331: are able to extract the following important information:
332: 
333: \begin{enumerate}
334: \item {\it B series} samples have better antiferromagnetic coupling (except sample 4) between 
335: Fe layers compared to {\it A series} samples. We have made a comparison between samples 
336: according to their remanent to saturation magnetization ratio $(M_r/M_s)$. 
337: We define a quantity called the antiferromagnetic fraction(AFF) as
338: 
339: \begin{equation}
340: AFF (\%) = \left(1 - \frac{M_r}{M_s} \right) \times 100 \%. 
341: \end{equation}
342: 
343: Samples 1, 2, and 3 have AFF $\sim$ 80 \% at 5 K. Sample 4, which has $<$ 1 \% GMR at 5 K has 
344: FM alignment of Fe layers even in zero applied magnetic field, has the least AFF $\sim$ 50 \%.
345: Samples 5, 6, 7, and 8 have AFF $\sim$ 95 \% at 10 K. This gives us an important information about
346: the coupling of Fe layers. Fe layers in samples 5, 6, 7, and 8 are coupled antiferromagnetically
347: better than samples 1, 2, and 3 in zero field. Samples 1, 2, and 3 have more interface roughness 
348: compared to sample 4, 5, 6, 7, and 8 (concluded from the fact that samples 
349: 5, 6, 7, and 8 have less residual resistivity ($\sim$ 30 $\mu \Omega$cm) compared to 
350: samples 1, 2, and 3 ($\sim$ 45 $\mu \Omega$cm). The
351: saturation field for {\it A series} samples is about 1.3 tesla but the saturation 
352: fields for for {\it B series} samples decreases gradually from 1.2 to 0.5 tesla.
353: This implies that the strength of coupling in {\it A series} samples
354: is stronger than {\it B series} samples.
355: For {\it A series} samples the coercive field is about 200 Oe at 10 K. 
356: The coercive field of samples 5 - 8 is about 70 Oe. This lower value of the
357: coercive field indicates that samples 5 - 8 have ``cleaner'' interfaces,
358: i. e., in the absence of a field or in small fields the Fe layers are aligned ``perfectly'' AF. Samples 1, 
359: 2, and 3 have ``rougher'' interfaces and there are more Fe cluster embedded in Cr layer
360: and vice versa. Probably this is one of the reasons why the history-dependent behavior
361: was more prominently seen in samples 1 - 3. The MBE grown samples,
362: however, have never shown any hysteresis effect due to their much smoother interfaces.
363:  
364:  \item The magnetization data with external applied magnetic field at 
365:  room temperature fit well to the Langevin function of the form
366:  \begin{equation}
367:  \label{eq:lang}
368:  M(H)=N \bar{\mu} \left[ \coth \left( \frac{\bar{\mu}H}{kT} \right) - \frac{kT}{\bar{\mu}H} \right],
369:  \end{equation}
370:  where $\bar{\mu}$ is the average magnetic moment of the clusters, N is the number 
371: of clusters in the sample and $N\bar{\mu}$ is the amplitude of the superparamagnetic 
372: contribution. Figure~\ref{fig:lang} shows the Langevin function fit for samples 2, 3, 5, 6, and 7
373: below saturation field. The average magnetic moment per superparamagnetic cluster
374: $\bar{\mu}$ in Bohr magnetons have been found $\sim 1200 \pm 20, 1280 \pm 70,
375: 1200 \pm 30, 825 \pm 100,$ and $1150 \pm 230 \ \mu_B$ 
376: for samples 2, 3, 5, 6, and 7, respectively. This gives
377: an indication that the average interdiffused particle volume is about the same in both
378: the series and history dependent behaviour is affected prominently by 
379: interface heterostructure. We did not get Langevin function fit for sample 8
380: which has the least saturation field of about half a tesla. Here, the
381: exchange coupling between successive Fe layers are weak but the 
382: interdiffused clusters are strongly coupled to the nearby Fe layers.
383: 
384: \end{enumerate}
385: 
386: 
387: \subsection{ \label{ss:acm} ac-magnetization}
388: We have measured the ac-magnetization vs. temperature in these multilayers at  
389: different frequencies of an ac field of 10 Oe. We have found that the ac-magnetization 
390: has peaks as shown in Fig.~\ref{fig:s2ac} for sample 2. As the signals are quite low (micro-gram of Fe in these samples) we got rather noisy data.
391: Here the peaks in the ac-magnetization show a temperature shift, although by a 
392: very small amount (as in spin glasses), when we change the frequency of the driving signal. 
393: For comparison, a quantitative measure of the frequency shift in terms of $(\Delta T_g/T_g)\times 100 \%$
394: per decade of $\omega$ in canonical spin-glass is $Cu$Mn (0.5 \%), $Au$Mn (0.45 \%), and $Ag$Mn
395: (0.6 \%)~\cite{Mydosh:1993}. For the insulating spin glasses the frequency dependence is larger.
396: It is much more in superparamagnets like $a-CoO.Al_2O_3SiO_2$ (6 \%) 
397: and $a-(Ho_2O_3)(B_2O_3)$ (28 \%).
398: Spin glasses show cusp-like behavior at $T_g$. The cusp gets smeared even at 
399: fields only as high as 50 Oe. In our measurements we found only rounded peaks.
400: The peak shifts to higher temperatures as we increase the frequency.
401: To analyse this data we did peak fitting to each data set. From this peak fit we are able to
402: observe clearly that the peak shifts toward higher temperature as shown in Fig.~\ref{fig:tg}
403: and the magnitude of the magnetization decreases with increasing frequency. 
404: To interpret this result we assume the presence of clusters of different volumes 
405: (giving rise to a distribution of moments) at the interfaces and/or interdiffused in the bulk. 
406: At lower frequency moments of all size will be able to follow the magnetic field while 
407: at larger frequencies only small moments will be following the external field and the 
408: large clusters (large moments) will not be able to respond. This explains the larger 
409: magnitude of magnetization for lower frequency. Any peak is a competition of two 
410: processes. At the lowest temperature the moments are frozen but
411: as we start raising the temperature the thermal energy gives them 
412: some freedom to move and we get higher magnetization with increasing temperature. 
413: Above a certain temperature there is no freezing and we get a maximum. Further 
414: increase in the thermal energy disrupts all the alignments and so the moment
415: starts falling. Clusters of smaller size are unlocked at lower temperatures 
416: compared to those having large moments. The Arrhenius law to
417: explain the frequency shift in superparamagnets 
418: is given by $\omega = \omega_0 \ exp[-E_a/k_BT_g]$, where $\omega$
419: is the driving frequency of $\chi_{ac}$ measurements, $E_a$ is the energy
420: barrier height (=$KV$, the anisotropy constant times the cluster volume), and 
421: $T_g$ is the peak temperature. If the volume of clusters is fixed, then
422: $T_g$ increases with increasing frequency. However, if there is a
423: cluster volume distribution, then 
424: at higher frequency only smaller clusters respond and so the peak temperature 
425: shifts towards lower temperatures with increasing frequency. 
426: In multilayers, $T_g$ is shifting towards higher temperatures with
427: increasing frequency. This suggests that the change in $T_g$ 
428: due to cluster volume change is small in the present investigation.
429: 
430: 
431: 
432: 
433: 
434: To conclude, the low-field magnetization behavior has been studied in ion-beam sputtered Fe-Cr GMR
435: multilayers. Magnetic heterostructure due to the interdiffused particles/cluster and the interfacial
436: imperfactions play an important role in the low-field magnetization behavior. In low fields and at low 
437: temperatures the presence of spin-glass-like phase is established. At higher temperatures these 
438: multilayers behave more like a superparamagnet.
439: 
440: \begin{acknowledgments}
441: One of us (R.S.P.) acknowledges CSIR, Government of India for financial support. We thank Dr.
442: Ashna Bajpai for help and encouragement during some measurements. We sincerely thank
443: Dr. R. N. Viswanath of Forschungszentrum, Karlsruhe for doing the ac-magnetization measurements.
444: \end{acknowledgments}
445: 
446: %\references
447: 
448: %\bibliography{lfm}
449: \begin{thebibliography}{6}
450: 
451: \bibitem{Baibich:1988} M. N. Baibich, J. M. Broto, A. Fert, F. Nguyen Van Dau, F. Petroff,
452: P. Etienne, G. Greuzet, A. Friederich, and J. Chazelas, Phy. Rev. Lett., {\bf 61}, 2472 (1988).
453: 
454: \bibitem{Kravtsov:2001} E. Kravtsov, V. Lauter-Pasyuk, H. J. Lauter, B. Toperverg,
455: O. Nikonov, A. Petrenko, M. Milyaev, L. Romashev, and V. Ustinov, Physica B {\bf 297} 118 (2001).
456: 
457: \bibitem{Fert:1995} A. Fert, A. Barthelemy, P. Galtier, P. Holody, R. Loloee, R. Morel,
458: F. Petroff, P. Schroeder, L. B. Steren, and T. Valet, Mater. Sci. Eng., B {\bf 31}, 1 (1995).
459: 
460: \bibitem{Kakazei:2003} G. N. Kakazei, Yu. G. Pogorelov, J. A. M. Santos, J. B. Sousa, 
461: P. P. Freitas, S. Cardoso, N. A. Lesnik, and P. E. Wign, J. Magn. Magn. Mater. {\bf 266},
462: 57 (2003).
463: 
464: \bibitem{Patrin:2003} G. S. Patrin, V. O. Vas'kovskii, D. A. Velikanov, A. V. Svalov,
465: and M. A. Panova, Phys. Lett. A {\bf 309}, 155 (2003).
466: 
467: \bibitem{Oliveira:1999} A. J. A. de Oliveira, W. A. Ortiz, D. H. Mosca, N. Mattoso, 
468: I. Mazzaro, and W. H. Schreiner, J. Phys.: Condens. Matter., {\bf 11}, 47 (1999).
469: 
470: \bibitem{Berger:1994} A. Berger and H. Hopster, Phys. Rev. Lett. {\bf 73}, 193 (1994) and 
471: references therein.
472: 
473: \bibitem{Lanon:2002} J. M. Lannon Jr., C. C. Pace, D. Temple G. E. McGuire,
474: A. F. Hebard, M. Ray, Materials Research Society Symposium - Proceedings, 
475: {\bf 690}, 157 (2002).
476: 
477: \bibitem{Majumdar:2002} A. K. Majumdar, A. F. Hebard, Avinash Singh, and D. Temple,
478: Phys. Rev. B {\bf 65}, 054408 (2002).
479: 
480: \bibitem{Durand:1993} O. Durand, J. M. George, J. R. Childress, S. Lequien. A. Schuhl, and A. Fert
481: J. Magn. Magn. Mater., {\bf 121}, 140 (1993).
482: 
483: \bibitem{Chowdhury:1986} D. Chowdhury, {\it Spin glasses and other frustrated systems},
484: Princeton series in Physics, Princeton University Press, Princeton, New Jersey (1986).
485: 
486: \bibitem{Pierce:1999} D. T. Pierce, J. Unguris, R. J. Celotta, and M. D. Stiles,
487: J. Magn. Magn. Mater., {\bf 200}, 290 (1999).
488: 
489: \bibitem{Overhausser:1960} A. W. Overhausser and A. Arrott, 
490: Phys. Rev. Lett. {\bf 4}, 226 (1960).
491: 
492: \bibitem{Heinrich:1994} B. Heinrich and J. A. C. Bland (Eds.), {\it Ultrathin Magnetic Structures II},
493: Springer-Verlag, Berlin, Heidelberg (1994).
494: 
495: \bibitem{Cullen:1993} J. R. Cullen, and K. B. Hathaway, Phys. Rev. B {\bf 47}, 14998 (1993).
496: 
497: \bibitem{Patel:2005} R. S. Patel, A. K. Majumdar, A. F. Hebard, and D. Temple, 
498: J. Appl. Phys., {\bf 97}, 033309 (2005). 
499: 
500: \bibitem{Vega:1994} A. Vega, L. C. Balb\'{a}s, J. Dorantes-D\'{a}vila, and G. M. Pastor,
501: Phys. Rev. B, {\bf 50}, 3899 (1994).
502: 
503: \bibitem{Mydosh:1993} J. A. Mydosh, {\it Spin glasses: an experimental introduction}, 
504: Taylor \& Francis, London (1993).
505: 
506: \vspace{15cm}
507: \end{thebibliography}
508: 
509: \begin{table*}
510: \caption{\label{tab:dct} Three characterstic temperatures $T_m$, $T_{irr}$, and
511: $T_{inf}$ as functions of the applied magnetic field.}
512: %\begin{ruledtabular}
513: \begin{tabular}{|c|c|c|c|c|c|c|}
514: \hline
515: \emph{Characterstic} & \multicolumn{6}{c|}{\emph{Applied magnetic field}} \\ \cline{2-7}
516: \emph{Temperatures} & 10 Oe & 50 Oe & 100 Oe & 150 Oe & 200 Oe & Sample \\ \hline
517:    & 135 & 150 & 123 & 105 & 100 & 1   \\ \cline{2-7}
518: $T_m$ (K) & 41 & 41 & 35 & 29 & 24 & 2  \\ \cline{2-7}
519: ($\pm$ 2 K) & 95 & 97 & 83 & 74 & 60 & 3  \\ \cline{2-7}
520: & NP\footnotemark[1] &  & NP &  & NP  & 4  \\ \cline{2-7}
521: &  & 75 & 62 & 58 & 58 & 5  \\ \cline{2-7}
522: & 30 & 20 & 14 &  & 8 & 6  \\ \cline{2-7}
523: &  & 8 &  &  & NP & 7  \\ \cline{2-7}
524:    &  & NP & NP &  & NP & 8  \\ \hline
525:  &  & 140 & 140 & 135 & 140 & 1  \\ \cline{2-7}
526: $T_{inf}$ (K) & 43 & 45 & 45 & 43 & 47 & 2 \\ \cline{2-7}
527: ($\pm$ 5 K) & 91 & 97 & 97 & 98 & 98 &  3 \\ \cline{2-7}
528:    & NP &  & NP &  & NP & 4  \\ \cline{2-7}
529: &  & 75 & 75 & 75 & 85 & 5  \\ \cline{2-7}
530: & 20 & 22 & 20 &  & 22 & 6  \\ \cline{2-7}
531: &  & NP &  &  & NP & 7  \\ \cline{2-7}
532:    &  & NP & NP &  & NP & 8  \\ \hline
533:  &  & 290 & 245 & 225 & 220 & 1  \\ \cline{2-7}
534: $T_{irr} (K) $ & 191 & 175 & 149 & 139 & 210  & 2  \\ \cline{2-7}
535: ($\pm$ 5 K) & 265 & 230 & 205 & 200 & 195 & 3  \\ \cline{2-7}
536:    & NP &  & NP &  & NP & 4  \\ \cline{2-7}
537: &  & 280 & 200 & 200 & 195 & 5  \\ \cline{2-7}
538: & 245 & 210 & 200 &  & 145 & 6  \\ \cline{2-7}
539: &  & 100 &  &  & 200 & 7  \\ \cline{2-7}
540:    &  & NP & NP &  & NP & 8  \\ \hline
541: \end{tabular}
542: \footnotetext[1]{NP: Feature Not Present}
543: %\end{ruledtabular}
544: \end{table*}
545: 
546: 
547: \begin{table}
548: \caption{Values of $\chi^2$, correlation coefficient 
549: R$^2$, the parameters $M(0)$, and $A$ of Eq. (~\ref{eq:pwr}).}
550: \begin{tabular}{ccccccc} \hline
551: H(Oe)  &  $\chi^2(10^{-6})\footnotemark[2]$  &  $R^2$  &  $M(0)(emu/cc)$      &  A$(10^{-4}K^{-3/2})$
552:      &  Range of T(K)   &  \# of data points \\ \hline \hline
553: {\it Sample 1}  &    &    &    &    &    &   \\  
554: 50  &  3.984  &  0.9998  &  138.0$\pm$0.1  &  2.422$\pm$0.003  &  5 - 140  &  68 \\  
555: 100  &  2.242  &  0.9999  &  227.9$\pm$0.1  &  2.467$\pm$0.003  &  5 - 140  &  68 \\  
556: 150  &  10.807  &  0.9995  &  353.9$\pm$0.2  & 2.570$\pm$0.008  &  10 - 140  &  41 \\  
557: 200  &  11.037  &  0.9991  &  390.7$\pm$0.2  &  2.351$\pm$0.007  &  5 - 140  &  60 \\
558: 13000\footnotemark[3]  &  0.005  &  0.9901  &  1284.3$\pm$0.1  &  0.101$\pm$0.001  &  5 - 100  &  95 \\ \hline 
559: 
560: {\it Sample 3}  &    &    &    &    &        &  \\  
561: 10  &  8.345  &  0.9990  &  45.63$\pm$0.03  &  2.63$\pm$0.01  &  5 - 100  &  47 \\  
562: 50  &  19.325  &  0.9985  &  130.7$\pm$0.1  &  3.28$\pm$0.02  &  5 - 100  &  47 \\  
563: 100  &  32.110  &  0.9978  &  199.4$\pm$0.2  &  3.40$\pm$0.02  &  5 - 100  &  47 \\  
564: 150  &  6.232  &  0.9995  &  269.3$\pm$0.2  &  3.27$\pm$0.01  &  5 - 100  &  46 \\  
565: 200  &  7.480  &  0.9994  &  303.2$\pm$0.2  &  3.19$\pm$0.01  &  5 - 100  &  46 \\ 
566: 13000\footnotemark[3]  &  0.006  &  0.9963  &  1463.0$\pm$0.1  &  0.100$\pm$0.001  &  5 - 100  &  95 \\ \hline 
567: 
568: {\it Sample 4}  &    &    &    &    &       &  \\  
569: 10 (FC) &  1.066  &  0.9992  &  1105.4$\pm$0.2  &  0.218$\pm$0.001  &  5 - 300  &  114 \\
570: 10 (ZFC) &  1.359  &  0.9990  &  1106.4$\pm$0.2  &  0.221$\pm$0.001  &  5 - 300  &  114 \\
571: 100 (FC) &  1.065  &  0.9977  &  1459.0$\pm$0.3  &  0.130$\pm$0.001  &  5 - 300  &  60 \\  
572: 100 (ZFC) &  1.448  &  0.9972  &  1463.1$\pm$0.4  &  0.137$\pm$0.001  &  5 - 300  &  60 \\ \hline
573: 
574: {\it Sample 5}  &    &    &    &    &     &  \\  
575: 50  & 0.608   & 0.9996 & 138.6$\pm$0.1 & 3.42$\pm$0.01 & 5 - 80 & 23  \\
576: 100 & 65.962  & 0.9976 & 190.4$\pm$0.4 & 4.48$\pm$0.03 & 5 - 80 & 26  \\  
577: 150 & 37.940  & 0.9983 & 243.5$\pm$0.3 & 2.27$\pm$0.02 & 5 - 80 & 23  \\
578: 200 & 17.731  & 0.9979 & 356.4$\pm$0.2 & 3.82$\pm$0.02 & 5 - 80 & 76 \\ \hline \hline
579: 
580: 
581: \end{tabular}
582: \label{tab:pwr}
583: \footnotetext[2]{ $\chi^2$ is defined as 
584: \begin{equation}
585: \chi^2 = \frac{1}{N} \sum_{i=1}^{N} 
586: \frac{(M_{i-measured}-M_{i-fitted})^2}{M_{i-mean}^2} \nonumber
587: \end{equation}
588: }
589: \footnotetext[3]{Ref.~\onlinecite{Patel:2005}}
590: \end{table}
591: 
592: 
593: 
594: 
595: \begin{figure}
596: \includegraphics[bb=0.4cm 0.5cm 8.7cm 7cm,width=8.5cm]{Fig-1}
597: \caption{\label{fig:s1raw} (Color online) Magnetization {\it vs.} temperature (raw data without any
598: corrections for the diamagnetism of Si substrate and paramagnetism of the packing materials) for sample 1.
599: Three characterstic temperatures $T_m$(marked with $\uparrow$), $T_{irr}$(marked with
600: $\downarrow$), and $T_{inf}$ (dashed line) are also presented. $T_m$ and $T_{irr}$ decrease
601: to lower temperatures with increasing applied external magnetic field. $T_{irr}$ is found to be 
602: independent of the applied external magnetic field. Samples 2, 3, 5, and 6 also show this type of
603: history-dependent magnetization behaviour.}
604: \end{figure}
605: 
606: \begin{figure}
607: \includegraphics[bb=14pt 16pt 271pt 220pt,width=246pt]{Fig-2}
608: \caption{\label{fig:ox3raw} (Color online) Magnetization {\it vs.} temperature (raw data without any
609: corrections for the diamagnetic Si substrate and the paramagnetic packing materials) for sample 
610: 7, 8 and 4 (in the inset).}
611: \end{figure}
612: 
613: \begin{figure}
614: \includegraphics[bb=0.4cm 0.6cm 9.5cm 7.6cm,width=8.5cm]{Fig-3}
615: \caption{\label{fig:gm0ZFCall} (Color online) $m(5 K)_{ZFC}$ {\it vs.} applied external magnetic field for samples 1, 2,
616: 3, 5 and 6. $m(5 K)_{ZFC}$ increases almost linearly with the applied magnetic field. Data and
617: fit for sample 5 and 6 are shifted along y-axis by adding 1.5 $\times 10^{-4}$ emu and 2 $\times 10^{-4}$
618: emu, respectively for clarity.}
619: \end{figure}
620: 
621: \begin{figure}
622: \includegraphics[bb=0.4cm 0.5cm 10.5cm 7.8cm,width=8.5cm]{Fig-4}
623: \caption{\label{fig:3by2} (Color online) Magnetization {\it vs.} temperature for samples 1 and 3.
624: The solid lines are the fits to Eq. (\ref{eq:pwr}). Data are fitted below $T_{inf}$.}
625: \end{figure}
626: 
627: \begin{figure}
628: \includegraphics[bb=0.4cm 0.5cm 10.5cm 7.8cm,width=8.5cm]{Fig-5}
629: \caption{\label{fig:onebyT} (Color online) Magnetization {\it vs.} temperature for samples 1, 2, 3
630: 5, and 6 above $T_{irr}$. The solid lines are the fits to an equation of the form $M \sim 1/T$.}
631: \end{figure}
632: 
633: \begin{figure}
634: \includegraphics[bb=0.4cm 0.5cm 9.5cm 7.2cm,width=8.5cm]{Fig-6}
635: \caption{\label{fig:at} (Color online) The plot of $T_m$ vs. $H^{2/3}$.
636: $T_m$-axis intersection gives the $T_g$ for that sample (written in bracket).}
637: \end{figure}
638: 
639: \begin{figure}
640: \includegraphics[bb=0.4cm 0.5cm 10.5cm 7.8cm,width=8.5cm]{Fig-7}
641: \caption{\label{fig:mh} (Color online) Magnetization {\it vs.} external magnetic field
642: for samples 1 at 5 K, 4 at 10 K, and 6 at 10 K. Insets are the same M(H) 
643: plots for samples 1 and 6 but on different expanded scales.}
644: \end{figure}
645: 
646: \begin{figure}
647: \includegraphics[bb=0.4cm 0.5cm 10.5cm 7.8cm,width=8.5cm]{Fig-8}
648: \caption{\label{fig:lang} The Langevin function fits for samples 2, 3,
649: 5, and 6 below $H_{sat}$. Data and fit for samples 3, 5, and 6 are 
650: shifted along y-axis by adding -0.2, 0.2, and 0.4, respectively for clarity.}
651: \end{figure}
652: 
653: \begin{figure}
654: \includegraphics[bb=0.4cm 0.5cm 9.5cm 7.2cm,width=8.5cm]{Fig-9}
655: \caption{\label{fig:s2ac} (Color online) ac-magnetization {\it vs.} temperature for
656: sample 2 at different frequencies. The points represent the data 
657: whereas the solid lines are the fits to find $T_g$.}
658: \end{figure}
659: 
660: \begin{figure}
661: \includegraphics[bb=0.4cm 0.5cm 9.5cm 7.2cm,width=8.5cm]{Fig-10}
662: \caption{\label{fig:tg} (Color online) Variation of peak temperature ($T_g$) with frequency
663: of the applied magnetic field. The solid lines are just guides to the eye.}
664: \end{figure}
665: 
666: \end{document}
667: