1:
2: \documentstyle[amssymb,preprint,aps]{revtex}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: %TCIDATA{OutputFilter=LATEX.DLL}
5: %TCIDATA{Created=Thu Sep 07 18:33:03 2000}
6: %TCIDATA{LastRevised=Tue Jun 12 17:26:16 2001}
7: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
8: %TCIDATA{Language=American English}
9: %TCIDATA{CSTFile=revtex.cst}
10:
11: \draft
12:
13: \begin{document}
14: \title{Evidence for Internal Field in Graphite: a Conduction Electron Spin
15: Resonance Study}
16: \author{M.S. Sercheli, Y. Kopelevich, R. Ricardo da Silva*, J.H.S. Torres, and C.
17: Rettori,}
18: \address{Instituto de F\'{i}sica ``Gleb Wataghin'', UNICAMP,13083-970, Campinas-SP,\\
19: Brazil.}
20: \date{\today }
21: \maketitle
22:
23: \begin{abstract}
24: We report conduction electron spin resonance measurements performed on
25: highly oriented pyrolitic graphite samples between $10$ K and $300$ K using
26: S ($\nu $ $=4$ GHz), X ($\nu $ $=9.4$ GHz), and Q ($\nu $ $=34$.4 GHz)
27: microwave bands for the external $dc$-magnetic field applied parallel ($%
28: H\parallel c$) and perpendicular ($H\perp c$) to the sample hexagonal $%
29: c-axis $. The results obtained in the $H\parallel c$ geometry are
30: interpreted in terms of the presence of an effective internal
31: ferromagnetic-like field, $H_{int}^{eff}$($T,H$), that increases as the
32: temperature decreases and the applied $dc$-magnetic field increases. We
33: associate the occurrence of the $H_{int}^{eff}$($T,H$) with the
34: field-induced metal-insulator transition in graphite and discuss its origin
35: in the light of relevant theoretical models.
36: \end{abstract}
37:
38: \pacs{76.30.Pk,71.30.+h,71.27.+a}
39:
40: Recent experiments revealed a number of unexpected phenomena in graphite
41: such as a linear energy dependence of the quasiparticle damping, \cite{Xu}
42: high-$T$ superconducting and ferromagnetic (FM) correlations,\cite
43: {Kopelevich,Y. Kopelevich} and magnetic-field-induced metal-insulator
44: transition (MIT).\cite{Kopelevich,Kempa} These properties could be
45: understood taking into account electron-electron interactions within the
46: graphite layers.\cite{Gonzales} A possible proximity of graphite monolayer
47: to the excitonic instability was an alternative explanation for both, the
48: linear energy dependence of the quasiparticle damping and the weak FM\cite
49: {Khveshchenko}. These observations bring graphite to the class of materials
50: which attract a broad scientific interest, including high-$T_{c}$
51: superconducting (HTS) cuprates, in which a strong Coulomb repulsion between
52: carriers is the key issue,\cite{Anderson} hexaborides demonstrating a high-$%
53: T $ weak FM,\cite{Young} and most recently discovered superconducting MgB$%
54: _{2}$, \cite{Nagamatsu} a material isoelectronic to graphite. \cite{Baskaran}
55:
56: It has been proved long ago \cite{Dresselhaus} that the electron spin
57: resonance (ESR) is a powerful experimental technique that allows to probe
58: directly the properties of conduction electrons (holes) in graphite.
59: However, the observed $g$-factor anisotropy and its temperature dependence
60: remains unexplained by existing theories which consider{\bf \ }the
61: spin-orbit coupling mechanism{\bf ,} \cite{Dresselhaus} indicating the
62: necessity of further experimental and theoretical work.
63:
64: In this letter we report the experimental results which provide a fresh
65: insight on the conduction ESR (CESR) in highly oriented pyrolitic graphite
66: (HOPG) samples. In particular, we show that the results can be interpreted
67: in terms of an effective internal FM-like field, $H_{int}^{eff}(T,H)$,
68: induced by the external $dc$-magnetic field, $H_{ext}=H$, applied along the
69: sample hexagonal $c-axis$. The results obtained in this work indicate that
70: the occurrence of $H_{int}^{eff}(T,H)$ intimately couples to the
71: magnetic-field-induced MIT.
72:
73: CESR measurements performed on three HOPG samples, synthesized at the
74: Research Institute ''Graphite'' (Moscow) \cite{Kopelevich,Y.
75: Kopelevich,Kempa} and the Union Carbide Co. were carried out in a
76: temperature interval from $10$ K to $300$ K by means of a Bruker ELEXSYS-CW
77: spectrometer operating at three microwave frequency bands, S ($\nu =4$ GHz),
78: X ($\nu $ $=9.4$ GHz), and Q ($\nu $ $=34.4$ GHz). The measurements with
79: mutually perpendicular microwave and $dc$-magnetic fields were performed in
80: both $H\parallel c$ and $H\perp c$ configurations. In order to warrant a
81: maximal microwave penetration and to avoid a line-shape change caused by
82: anisotropic skin depth effects, \cite{Walmsley} the microwave field was
83: always kept perpendicular to the sample $c-axis$. Similar results were
84: obtained for all graphite samples studied in this work. Complementary
85: low-frequency ($\nu =1$ Hz) standard four-probe basal-plane resistance $%
86: R(T,H)$ and $dc$-Hall effect measurements were performed for $H\parallel c$
87: using a Quantum Design PPMS-$9$ T and Janis-$9$ T magnet systems.
88:
89: The obtained low field in-plane carrier mobility $\sim 7\cdot 10^{5}-2\cdot
90: 10^{6}$ cm$^{2}$/V$\cdot $s ($T=4.2$ K) indicates the high quality of our
91: samples. \cite{Dresselhaus} The high sample quality has also been confirmed
92: by means of Shubnikov-de Hass oscillations measurements, where, e.g., two
93: resistance minima at $H=6.7$ T and $H=7.5$ T, corresponding to the spin
94: splitting of the $n=1$ Landau level for electrons,\cite{Iye,Timp,Y.Iye}
95: could clearly be resolved.
96:
97: In all our CESR measurements a Dysonian line-shape with $A/B$ asymmetry
98: ratio between $4$ and $9$ was observed.\cite{Dyson} The CESR parameters
99: (resonance field and line-width) were obtained assuming an admixture of
100: absorption/dispersion of a Lorentzian line. \cite{Pake} The integrated
101: absorption part of the resonance was found to be $T$-independent within the
102: accuracy of the measurements, confirming that the ESR originates from
103: itinerant carriers. \cite{Abragam,Taylor}
104:
105: Figure 1 (a, b, c) presents the $T$-dependencies of the resonance field $%
106: H_{R}(T)$ for one of our studied samples (HOPG-3, see Ref. [4]) in $%
107: H\parallel c$ and $H\perp c$ geometry using S, X, and Q bands, respectively.
108: As can be seen from Fig. 1, $H_{R}$ obtained for $H\perp c$ is $T$%
109: -independent within the experimental resolution, whereas $H_{R}(T)$ measured
110: for $H\parallel c$ is a decreasing function of $T$. Note, that $%
111: H_{R\parallel }-H_{R\perp }<0$ which can be re-written in terms of the Lande
112: $g$-factor shift $\Delta g=g_{\parallel }-g_{\perp }>0$. This is a well
113: known experimental fact being, however, in conflict with the theoretical
114: predictions of $\Delta g<0$ based on the spin-orbit interaction. \cite
115: {Dresselhaus} Figure 2 shows $T$-dependencies of the difference $\Delta
116: H_{R}=H_{R\perp }-H_{R\parallel }$ extracted from the data of Fig. 1. These
117: data demonstrate that $\Delta H_{R}(T,H)$ increases as the temperature
118: decreases and the measuring field (frequency) increases. The inset in Fig. 2
119: depicts the behavior of $g_{\parallel }$ also extracted from the data of
120: Fig. 1. As can be seen from this plot, the $g_{\parallel }(T,H)$ is both $T$%
121: - and $H$-dependent. Note a rather complex and non-systematic behavior of $%
122: g_{\parallel }(T,H)$: while $g_{\parallel }(T,H)$ obtained from the S band
123: data coincides with that obtained from Q band for $T\gtrsim 70$\ K, it
124: coincides with the $g_{\parallel }(T,H)$ measured with X band for $T<$\ $70$%
125: \ K, and all three bands give the same value of the $g$-factor at $T=300$\ K
126: (see inset in Fig. 2){\bf . }On the other hand, a comparative analysis of
127: the CESR and transport measurements revealed an intriguing correlation
128: between $\Delta H_{R}(T,H)$ and the basal-plane resistance, $R(T,H)$,
129: suggesting that the description of the results presented in Fig. 1 in terms
130: of $\Delta H_{R}(T,H)$ is more appropriate than that in terms of $%
131: g_{\parallel }(T,H)$. It is shown in Fig. 2 that $\Delta H_{R}(T,H)$ can be
132: well approximated by the empirical equation:
133:
134: \begin{equation}
135: \Delta H_{R}(T,H)=H_{0}ln[R(T,H)/R_{0}(H)]
136: \end{equation}
137: where $H_{0}$ and $R_{0}(H)$ are fitting parameters. Both, $\Delta
138: H_{R}(T,H) $ and $R(T,H)$ decrease approaching the transition from
139: insulating-like ($dR/dT<0$) to the metallic-like ($dR/dT>0$) resistance
140: behavior,\cite{Kempa} implying that the appearance of the $\Delta
141: H_{R}(T,H)>0$ is essentially coupled to the field-induced MIT.
142:
143: It is already established that the CESR results of Fig. 1 are characteristic
144: of both HOPG and single crystalline graphite.\cite{Dresselhaus} Because this
145: and in order to be more specific on our claim regarding the relationship
146: between $\Delta H_{R}(T,H)$ and MIT, we present in Fig. 3 (a, b) the $R(T,H)$
147: data obtained for HOPG (Union Carbide Co.) and Kish single crystalline
148: graphite samples. The data of Fig. 3 (a, b) provide an unambiguous evidence
149: that the field-induced MIT is an intrinsic property of graphite. We note
150: that the low temperature leveling-off of the insulating-like $R(T,H)$ is
151: associated with the reentrant transition to the metallic state, \cite
152: {Kopelevich} which will be not discussed in the present work. A careful
153: analysis of the results given in Fig. 3 (a, b) revealed that the
154: insulating-like resistance $R(T,H)$ develops below a well defined
155: field-dependent temperature $T_{min}(H)$ which is an increasing function of $%
156: H$. This is exemplified for the Kish graphite in Fig. 4 where the reduced
157: resistance $R(T)/R(T_{min})$ vs$.$ $T$ is plotted. The inset in Fig. 4 shows
158: $T_{min}(H)$ which can be best described (dashed curve) by the formula:
159:
160: \begin{equation}
161: T_{min}(H)=A(H-H_{c})^{1/2} \label{eq. 1}
162: \end{equation}
163: with $A$ and $H_{c}$ being the fitting parameters. For $H>1$ kOe, no minimum
164: in $R(T)$ has been found in the studied temperature interval.
165:
166: In attempts to understand the origin of $\Delta H_{R}(T,H)>0$, which
167: suggests the existence of an effective internal FM-like field $%
168: H_{int}^{eff}(T,H)$, as well as its coupling to the MIT we note that the
169: characteristic feature of the band structure of a single graphene layer is
170: that there are two isolated points in the first Brillouin zone where the
171: band dispersion is linear $E$({\bf k}) = $\hbar $v$\mid ${\bf k}$\mid $(v = v%
172: $_{F}$ $\sim 10^{6}$ m/s is the Fermi velocity), so that the electronic
173: states can be described in terms of the Dirac equations in two dimensions.
174: \cite{Gonzales,Khveshchenko,Balents} Besides, an opening of the gap in the
175: spectrum of Dirac fermions by an applied magnetic field (the so-called
176: ''magnetic catalysis'' phenomenon), accompanied by the time-reversal
177: symmetry breaking, has been discussed in a number of recent theoretical
178: works (see e.g., \cite{Laughlin,Ferrer}, and references therein). The
179: equation for the field-induced transition temperature $T_{c}(H)$ to the
180: gapped state is given by:
181:
182: \begin{equation}
183: k_{B}T_{c}(H)=C\hbar \text{v}(2e/\hbar )^{1/2}(H-H_{c})^{1/2},
184: \end{equation}
185: which is similar to the Eq. (2) derived from the experiment. Here $C$ is the
186: constant, $\Delta _{0}$ = $\hbar $v($2e/\hbar $)$^{1/2}$($H-H_{c}$)$^{1/2}$
187: is the induced zero-temperature gap, and $H_{c}$ is the critical field.
188: According to the theory, the development of a magnetic moment proportional
189: to the $\Delta (T,H)$ is also expected. In the context of the present work
190: this would imply $H_{int}^{eff}(T,H)\sim \Delta (T,H)$ which can account for
191: the correlation between $H_{int}^{eff}(T,H)$ and $R(T,H)$, see Fig. 2.
192: Whereas in Refs. [18, 19] the gap occurrence is related to the development
193: of an additional superconducting order parameter in HTS, the gap in the case
194: of graphite is associated with the insulating-like state. However, as
195: emphasized by Laughlin \cite{Laughlin} the theory [Eq. (3)] considers energy
196: scales only. In the light of theoretical results of Ref. [6], it is possible
197: that the here observed phenomenon is related to the field-driven
198: metal-excitonic insulator transition.
199:
200: While the analysis of the experimental results in terms of an effective
201: internal FM-like field occurrence appears to be self-consistent, we cannot
202: exclude a complex dependence of the $g$-factor on the applied magnetic field
203: and temperature (see inset in Fig.2) as an alternative approach. However a
204: lack of a proper theory for both magnetic field and temperature dependence
205: \cite{Dresselhaus} of the $g$-factor prevents us to speculate on this matter.
206:
207: Finally, we present in Fig. 5 the CESR half line-width, $\Delta H_{1/2}(T)$,
208: measured at various frequencies ($\nu $ $=4$ GHz, $\nu $ $=9.4$ GHz, and $%
209: \nu =34.4$ GHz). Due to an intrinsic $mosaic$ effect in HOPG the line-width
210: for $H\parallel c$ is always broader than for $H\perp c$. As can be observed
211: in Fig. 5, $\Delta H_{1/2}(T)$ increases with temperature decreasing, in
212: agreement with previous studies.\cite{Dresselhaus} The novel result,
213: however, is that $\Delta H_{1/2}(T)$ is essentially field
214: (frequency)-independent. Then, the induced effective internal FM-like field
215: should be highly uniform. In principle, various physical processes such as a
216: distribution of $g$-factor values, motional narrowing, many-body exchange,
217: and relaxation effects can contribute to $\Delta H_{1/2}(T)$,\cite
218: {Dresselhaus,Lubzens} so that a further work is needed to provide a
219: plausible explanation for $\Delta H_{1/2}(T)$. For example, pulsed CESR
220: experiments where the spin-spin and spin-lattice relaxation times can be
221: measured, are highly desirable.
222:
223: In summary, CESR measurements performed on HOPG samples using S ($\nu =4$
224: GHz), X ($\nu $ $=9.4$ GHz), and Q ($\nu =34.4$ GHz) microwave bands
225: revealed a complex behavior of the $g$-factor as a function of the applied
226: magnetic field which cannot be understood within a framework of the existing
227: theories. Alternatively, we show that the results can be self-consistently
228: described in terms of the occurrence of an effective internal FM-like field
229: induced by the applied $dc$-magnetic field ($H\parallel c-axis$). It is also
230: found that the occurrence of the internal field is intimately coupled to\
231: the magnetic field-induced MIT. We argue that the magnetic field-induced gap
232: in the energy spectrum of Dirac fermions in graphite can be responsible for
233: the here observed phenomena.
234:
235: {\it \bigskip }
236:
237: This work is supported by FAPESP, CNPq, and CAPES Brazilian science
238: agencies. We thank V.V. Lemanov and P. Scharff for supplying the graphite
239: samples used in the present study, and D.V. Khveshchenko for discussions.
240:
241: * Also at FAENQUIL, 12600-000, Lorena, SP, Brazil.
242:
243: \begin{references}
244: \bibitem{Xu} S. Xu, J. Cao, C.C. Miller, D.A. Mantell et al., Phys. Rev.
245: Lett. {\bf 76}, 483 (1996).
246:
247: \bibitem{Kopelevich} Y. Kopelevich, V.V. Lemanov, S. Moehlecke, and J.H.S.
248: Torres, Fizika Tverd. Tela (St. Petersburg) {\bf 41}, 2135 (1999) [Phys.
249: Solid State {\bf 41}, 1959 (1999)].
250:
251: \bibitem{Y. Kopelevich} Y. Kopelevich, P. Esquinazi, J.H.S. Torres, and S.
252: Moehlecke, J. Low Temp. Phys. {\bf 119}, 691 (2000).
253:
254: \bibitem{Kempa} H. Kempa, Y. Kopelevich, F. Mrowka, A. Setzer, et al.,
255: Solid State Commun. {\bf 115}, 539 (2000).
256:
257: \bibitem{Gonzales} J. Gonzalez, F. Guinea, and M.A.H. Vozmediano, Phys.
258: Rev. Lett. {\bf 77}, 3589 (1996); Phys. Rev. B {\bf 63}, 13442 (2001).
259:
260: \bibitem{Khveshchenko} D.V. Khveshchenko, cond-mat/0101306.
261:
262: \bibitem{Anderson} P.W. Anderson, Science {\bf 235}, 1196 (1987); Physica C
263: {\bf 341}, 1575 (2000).
264:
265: \bibitem{Young} D.P. Young, D. Hall, M.E. Torelli, Z. Fisk et al., Nature
266: {\bf 397}, 412 (1999).
267:
268: \bibitem{Nagamatsu} J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani,
269: and J. Akimitsu, Nature {\bf 410}, 63 (2001).
270:
271: \bibitem{Baskaran} G. Baskaran, cond-mat/0103308.
272:
273: \bibitem{Dresselhaus} M.S. Dresselhaus, G. Dresselhaus, K. Sugihara, I.L.
274: Spain, and H.A. Goldberg in: ''Graphite Fibers and Filaments'', Springer
275: Series in Material Science Vol. {\bf 5}, pp. 179-188 (Springer-Verlag,
276: Berlin, N. Y., London, Paris, Tokyo), and references therein; M.S.
277: Dresselhaus and G. Dresselhaus, Adv. Phys. {\bf 30}, 139 (1981).
278:
279: \bibitem{Walmsley} L. Walmsley, G. Ceotto, J.H. Castilho, and C. Rettori,
280: Synthetic Metals {\bf 30}, 97 (1989).
281:
282: \bibitem{Iye} Y. Iye, P.M. Tedrow, G. Timp, M. Shayegan et. al., Phys. Rev.
283: B {\bf 25}, 5478 (1982).
284:
285: \bibitem{Timp} G. Timp, P.D. Dresselhaus, T.C. Chieu, G. Dresselhaus, and
286: Y. Iye, Phys. Rev. B {\bf 28}, 7393 (1983).
287:
288: \bibitem{Y.Iye} Y. Iye, L.E. McNeil, and G. Dresselhaus, Phys. Rev. B {\bf %
289: 30}, 7009 (1984).
290:
291: \bibitem{Dyson} F.J. Dyson, Phys. Rev. 98, 349 (1955); G. Feher and A.F.
292: Kip, Phys. Rev. {\bf 98}, 337 (1955).
293:
294: \bibitem{Pake} G.E. Pake and E.M. Purcell, Phys. Rev. {\bf 74}, 1184 (1948).
295:
296: \bibitem{Abragam} A. Abragam and B. Bleaney, EPR of Transition Ions
297: (Clarendon Press, Oxford, 1970).
298:
299: \bibitem{Taylor} R.H. Taylor, Adv. in Physics {\bf 24}, 681 (1975); S. E.
300: Barnes, ibid. {\bf 30}, 801 (1981).
301:
302: \bibitem{Balents} L. Balents and M.P.A. Fisher, Phys. Rev. B {\bf 55},
303: 11973 (1997).
304:
305: \bibitem{Laughlin} R.B. Laughlin, Phys. Rev. Lett. 80, 5188 (1998).
306:
307: \bibitem{Ferrer} E.J. Ferrer, V.P. Gusynin, and V. de la Incera,
308: cond-mat/0101308.
309:
310: \bibitem{Lubzens} D. Lubzens, M.R. Shanabarger, and S. Schultz, Phys. Rev.
311: Lett. {\bf 29}, 1387 (1972); D. Lubzens and S. Schultz, Phys. Rev. Lett.
312: {\bf 36}, 1104 (1976).
313:
314: \begin{figure}[tbp]
315: \caption{{}$T$-dependencies of the resonance field, $H_{R}(T)$, measured in
316: the HOPG-3 sample for $H$ $\Vert $ $c$ (squares) and $H$ $\bot $ $c$
317: (triangles) geometries using S ($\protect\nu =4$ GHz), X ($\protect\nu =9.4$
318: GHz), and Q($\protect\nu =34.4$ GHz) microwave bands. The indicated $g$%
319: -factors correspond to the values at room-$T$. The continuous lines are
320: guide for the eye.}
321: \label{Figure 1}
322: \end{figure}
323:
324: \begin{figure}[tbp]
325: \caption{$\Delta H_{R}(T)$ obtained from the data of Fig. 1 (solid symbols).
326: Open symbols correspond to $\Delta H_{R}(T)$ calculated from Eq. (1) using
327: basal-plane resistance data ($H\parallel c$) obtained for the HOPG-3 sample
328: with $H=1.34$ kOe, ($\bigtriangleup $), $H=3.14$ kOe ($\Box $), and $H=11.5$
329: kOe ($\circ $) and fitting parameters $H_{0}=150$ Oe, $R_{0}=0.014$ $\Omega $%
330: ; $H_{0}=150$ Oe, $R_{0}=0.01$ $\Omega $; $H_{0}=150$ Oe, $R_{0}=0.004$ $%
331: \Omega $, respectively. Inset depicts $g_{\parallel }(T,H$) extracted from
332: the $H_{R}(T,H)$ data given in Fig. 1. The continuous lines are guide for
333: the eye.}
334: \end{figure}
335: \begin{figure}[tbp]
336: \caption{Basal-plane resistance, $R(T,H)$, measured at various magnetic
337: fields (in kOe) applied parallel to the sample $c-axis$ in (a) HOPG (Union
338: Carbide Co.) and (b) single crystalline Kish graphite. }
339: \end{figure}
340: \begin{figure}[tbp]
341: \caption{ Reduced basal-plane resistance $R(T)/R(T_{min})$ measured in Kish
342: graphite single crystal with $H=0.04$ T ($\Box $), $H=0.05$ T ($\circ $), $%
343: H=0.075$ T ($\triangle $), and $H=0.1$ T ($\triangledown $) applied parallel
344: to the $c-axis$, where $T_{min}(H)$, denoted by arrows, is the
345: field-dependent temperature separating insulating-like ($T<T_{min}$) and
346: metallic-like ($T>T_{min}$) resistance behavior. The continuous lines are
347: guide for the eye. Inset shows $T_{min}(H)$; the dashed line is obtained
348: from Eq. (2) with the fitting parameters $A=350$ K/T$^{1/2}$ and $H_{c}=0.038
349: $ T. }
350: \end{figure}
351:
352: \begin{figure}[tbp]
353: \caption{ The CESR half line-width, $\Delta H_{1/2}(T)$, measured in the
354: HOPG-3 sample for both $H\parallel c$ and $H\perp c$ geometries using S ($%
355: \protect\nu =4$ GHz), X ($\protect\nu =9.4$ GHz), and Q($\protect\nu =34.4$
356: GHz) microwave bands. The dotted lines are guide for the eye.}
357: \end{figure}
358: \end{references}
359:
360: \end{document}
361: