cond-mat0012164/O3.tex
1: \section{The three-dimensional Heisenberg universality class}
2: \label{O3d3}
3: 
4: The three-dimensional Heisenberg universality class is
5: characterized by a three-component order parameter, ${\rm O}(3)$
6: symmetry, and short-range interactions.
7: It describes the critical 
8: behavior of isotropic
9: magnets, for instance the Curie transition in isotropic ferromagnets 
10: such as Ni  and EuO, and of antiferromagnets such as 
11: RbMnF$_3$ at the N\'eel transition point. 
12: Moreover, it describes isotropic 
13: magnets with quenched disorder, see also Sec.~\ref{lsec-random}.
14: Indeed, since $\alpha < 0$, the Harris
15: criterion \cite{Harris-74} states that disorder is an irrelevant 
16: perturbation. The only effect is to introduce 
17: an additional correction-to-scaling term $|t|^{\Delta_{\rm dis}}$ with 
18: $\Delta_{\rm dis} = - \alpha$. 
19: 
20: Note that
21: the isotropic Heisenberg Hamiltonian is a simplified model for magnets.
22: It neglects several interactions that are present in real materials.
23: Among them, we should mention the presence of interactions with
24: cubic anisotropy due to the lattice 
25: structure and the dipolar interactions. 
26: Even if, in the RG language, 
27: these effects are relevant perturbations
28: of the Heisenberg fixed point 
29: \cite{Fisher-74,Aharony-76,AF-73,CPV-00},
30: the new critical exponents are so close to those of the 
31: Heisenberg universality class that the difference is experimentally
32: very difficult to observe, see, e.g., 
33: Refs.~\cite{BA-74,CPV-00,SKS-00,CPV-02-3} and reference therein.
34: See also Sec.~\ref{lsec-cubic} for a discussion of the 
35: cubic anisotropy.
36: 
37: 
38: \subsection{The critical exponents}
39: \label{expO3}
40: 
41: \subsubsection{Theoretical results}
42: \label{expO3th}
43: 
44: In Table \ref{O3exponents} we report the 
45: theoretical estimates of the critical exponents obtained
46: by various approaches.
47: 
48: 
49: 
50: \begin{table*}[tb]
51: \caption{Estimates of the critical exponents for the Heisenberg universality
52: class.
53: We indicate with an asterisk (${}^*$) the estimates that have been
54: obtained using the scaling relations $\gamma =(2 - \eta)\nu $,
55: $2 - \alpha = 3 \nu $, $\beta = \nu(1 + \eta)/2$, and 
56: $\beta\delta= \beta+ \gamma$. 
57: }
58: \label{O3exponents}
59: \footnotesize
60: \begin{center}
61: \begin{tabular}{rllllll}
62: \hline
63: \multicolumn{1}{c}{Ref.}& 
64: \multicolumn{1}{c}{info}& 
65: \multicolumn{1}{c}{$\gamma$}& 
66: \multicolumn{1}{c}{$\nu$}& 
67: \multicolumn{1}{c}{$\eta$}&
68: \multicolumn{1}{c}{$\beta$}&
69: \multicolumn{1}{c}{$\delta$} \\   
70: \hline  
71: \cite{CHPRV-02} $_{2002}$ & MC+IHT $\phi^4$ & 1.3960(9)
72: &0.7112(5)   & 0.0375(5)  &  0.3689(3)$^*$ &4.783(3)$^*$ \\
73: 
74: \cite{BC-97-2} $_{1997}$ & HT sc & 1.406(3)    & 0.716(2)     &
75: 0.036(7)$^*$ & 0.3710(13)$^*$ &4.79(4)$^*$    \\
76: \cite{BC-97-2} $_{1997}$ & HT bcc & 1.402(3)    & 0.714(2)     &  
77: 0.036(7)$^*$& 0.3700(13)$^*$ & 4.79(4)$^*$ \\
78: 
79: \cite{AHJ-93} $_{1993}$ & HT sc & 1.40(1)    & 0.712(10)     &  
80: 0.03(3)$^*$&  0.368(6)$^*$   &  4.80(17)$^*$\\
81: 
82: \cite{FH-86} $_{1986}$ & HT fcc   & 1.40(3)    & 0.72(1)    &  
83:   0.06(5)$^*$ & 0.38(2)$^*$  & 4.68(27)$^*$ \\
84: \cite{MDH-82} $_{1982}$ & HT sc   & 1.395(5)    &     &  &  &  \\
85: \cite{RF-72} $_{1972}$ & HT sc, bcc, fcc   & 1.375$^{+0.02}_{-0.01}$
86: & 0.7025$^{+0.010}_{-0.005}$  &  0.043(14) & 0.366(14)$^*$  & 4.75(16)$^*$ \\
87: 
88: 
89: \cite{CHPRV-02} $_{2002}$ & MC FSS $\phi^4$ &1.3957(22)$^*$ &0.7113(11)
90: & 0.0378(6)  &  0.3691(6)$^*$ &4.781(3)$^*$ \\
91: 
92: \cite{Hasenbusch-00} $_{2000}$  &  MC FSS $\phi^4$ &  1.393(4)$^*$  &
93: 0.710(2)&  0.0380(10)  &  0.3685(11)$^*$ & 4.780(6)$^*$\\
94: 
95: \cite{CBL-00} $_{2000}$  &  MC FSS double-exchange  & 1.3909(30) &
96: 0.6949(38)  & & 0.3535(30) & \\
97: 
98: \cite{BFMM-96} $_{1996}$  & MC FSS & 1.396(3)$^*$ & 0.7128(14)  &
99: 0.0413(16)  &  0.3711(9)$^*$ & 4.762(9)$^*$\\
100: 
101: \cite{BrCi-96} $_{1996}$ & MC FSS & 1.270(1)$^*$  & 0.642(2)  &
102: 0.020(1)  &  &  \\
103: \cite{HJ-94} $_{1994}$  &  MC FSS  &   &  0.706(8)$^*$  &  &  &  \\
104: \cite{HJ-93} $_{1993}$  &  MC FSS  & 1.389(14)$^*$  &  0.704(6)  &
105: 0.027(2)  & 0.362(3)$^*$  & 4.842(11)$^*$   \\
106: \cite{CFL-93} $_{1993}$  &  MC FSS sc, bcc  & 1.3812(6)$^*$  &
107: 0.7048(30)  &  0.0250(35)   & 0.361(2)$^*$ & 4.85(20)$^*$\\
108: \cite{PFL-91} $_{1991}$  &  MC FSS  & 1.390(18)$^*$  &  0.706(9)  &
109: 0.031(7)   & 0.364(5)$^*$ & 4.82(4)$^*$ \\
110: \cite{DHNN-91} $_{1991}$  &  MC FSS  & &  0.73(4)  & &  &  \\
111: \cite{NB-88}  $_{1988}$  & MC FSS &  & 0.716(40)  & &  & \\
112: 
113: \cite{JK-00} $_{2001}$ & FT $d=3$ exp & 1.3882(10)& 0.7062(7)  & 
114: 0.0350(8) & 0.3655(5)$^*$  & 4.797(5)$^*$ \\
115: 
116: \cite{GZ-98} $_{1998}$ & FT $d=3$ exp & 1.3895(50)& 0.7073(35) &
117: 0.0355(25) & 0.3662(25) & 4.794(14) \\
118: 
119: \cite{MN-91}   $_{1991}$ & 
120: FT $d=3$ exp & 1.3926(13)\{39\} & 0.7096(8)\{22\} &0.0374(4) & & \\
121: 
122: \cite{LZ-77}   $_{1977}$  & 
123: FT $d=3$ exp & 1.386(4)& 0.705(3) &0.033(4) & 0.3645(25) & 4.808(22) \\
124: 
125: 
126: \cite{GZ-98} $_{1998}$ & FT $\epsilon$ exp & 1.382(9) & 0.7045(55) &
127: 0.0375(45) & 0.3655(5)$^*$  & 4.797(5)$^*$ \\
128: 
129: \cite{YG-98} $_{1998}$ & FT $\epsilon$ exp & 1.39$^*$ & 0.708 & 0.037 & 0.367$^*$ & 4.786$^*$ \\
130: 
131: \cite{Kleinert-00} $_{2000}$ & FT $(d-2)$ exp & & 0.695(10) & & & \\
132: 
133: \cite{NR-84}  $_{1984}$ & 
134: SFM  & 1.40(3)   &  0.715(20)   & 0.044(7) & 0.373(11) & 4.75(4)$^*$\\
135: \cite{BSW-01} $_{2001}$ & CRG & 1.45 & 0.74 & 0.038 & 0.37 & 4.78 \\
136: \cite{GW-01} $_{2001}$  & CRG (1st DE) & 1.374 & 0.704  & 0.049  &  0.369 & 4.720 \\
137: \cite{MT-98} $_{1998}$  & CRG (1st DE) & 1.464 & 0.745  & 0.035  &  0.386 & 4.797 \\
138: \cite{BTW-96} $_{1996}$  & CRG ILPA & 1.465 & 0.747  & 0.038  &  0.388 & 4.78 \\
139: \hline
140: \end{tabular}
141: \end{center}
142: \end{table*}
143: 
144: 
145: Accurate results for the critical exponents
146: have been obtained by combining 
147: MC simulations and HT expansions for the  improved $\phi^4$ Hamiltonian 
148: (\ref{latticephi4}) with $\lambda^*=4.6(4)$
149: \cite{CHPRV-02,Hasenbusch-00}, cf. Sec.~\ref{sec-2.3.2}.
150: Using the linked-cluster expansion technique, the HT expansions of 
151: $\chi$ and $\mu_2 \equiv \sum_x |x|^2 G(x)$
152: were computed to 20th order. 
153: The analyses were performed  using the estimates of $\lambda^*$
154: and $\beta_c$ obtained from the MC simulations. The 
155: results 
156: are denoted by MC+IHT in Table~\ref{O3exponents}. 
157: The other results reported in the table were obtained 
158: from the analysis of the HT series
159: for the standard Heisenberg model (HT),  by MC simulations (MC),
160: or by  FT methods (FT).  
161: The HT results of Ref.\ \cite{BC-97-2}
162: were obtained by analyzing 21st-order HT expansions for the standard
163: O(3)-vector model on the simple cubic (sc) and on the 
164: body-centered cubic (bcc) lattice.
165: Most MC results concern the standard Heisenberg model
166: and were obtained using FSS techniques 
167: \cite{BFMM-96,BrCi-96,HJ-94,HJ-93,CFL-93,PFL-91,DHNN-91,NB-88}.
168: The results of Refs.~\cite{CHPRV-02,Hasenbusch-00}  were obtained by simulating
169: the improved $\phi^4$ model. Ref.~\cite{CBL-00} considers 
170: an isotropic ferromagnet with double-exchange interactions,\footnote{
171: Recently, a model with competing superexchange and 
172: double-exchange interactions has been studied 
173: \cite{TL-01}. A 
174: preliminary analysis for the paramagnetic-ferromagnetic
175: transition gives $\nu = 0.720(2)$ and $\gamma = 1.438(8)$.
176: While $\nu$ is in reasonable agreement with the Heisenberg
177: value, $\gamma$ is significantly higher, so that the 
178: identification of this transition as a Heisenberg one is in doubt.}
179: whose Hamiltonian is given by \cite{AH-55}
180: \begin{equation}
181: {\cal H} = - \beta \sum_{\langle ij \rangle} \sqrt{ 1 + s_i\cdot
182: s_j}.
183: \end{equation}
184: The FT results of Refs.~\cite{JK-00,GZ-98,MN-91,LZ-77,YG-98,Kleinert-00} 
185: were derived 
186: by analyzing perturbative expansions in different frameworks:
187: fixed-dimension expansion (6th- and 7th-order series, see 
188: Refs. \cite{BNGM-77,MN-91}),
189: $\epsilon$ expansion (to $O(\epsilon^5)$, see Refs. \cite{CGLT-83,KNSCL-93}),
190: and $(d-2)$-expansion (to $O[(d-2)^4]$, 
191: see Refs. \cite{HB-78,Hikami-83,BW-86}). 
192: We quote two errors for the results of Ref.~\cite{MN-91}: 
193: the first one (in parentheses) 
194: is the resummation error, the second one (in braces) 
195: takes into account the uncertainty of the fixed-point value $g^*$,
196: which was estimated to be approximately 1\% in Ref.~\cite{MN-91}. 
197: %To estimate the second error we use the results of
198: %Ref.~\cite{GZ-98} where the dependence of the exponents on $g^*$ is given.
199: The results of Ref.~\cite{NR-84} were obtained  using the so-called 
200: scaling-field method (SFM). 
201: Refs.~\cite{BSW-01,GW-01,BTW-96,BTW-99} present results
202: obtained by approximately solving continuous renormalization-group
203: (CRG) equations for the average action.
204: We also mention the HT results of Ref. \cite{BC-99}:
205: they performed a direct determination of the exponent $\alpha$ obtaining 
206: $\alpha = -0.11(2)$, $-0.13(2)$ on the sc and bcc
207: lattice. Ref. \cite{MOF-01} computes the critical exponents for a Heisenberg
208: fluid by a canonical-ensemble simulation. Depending on the analysis method, 
209: they find $1/\nu = 1.40(1)$, $1.31(1)$, $\beta/\nu= 0.54(2)$, $0.52(1)$, 
210: and  $\gamma/\nu =1.90(3)$, $1.87(3)$.
211: Overall, all estimates are in substantial agreement. 
212: We only note the quite anomalous
213: result of Ref.~\cite{BrCi-96}, which is further discussed in
214: Refs. \cite{HJ-97,BrCi-97}, and 
215: the apparent discrepancies of the MC+IHT results with   
216: the MC estimates of $\eta$ of Refs.~\cite{BFMM-96,HJ-93},
217: and with the FT results of Ref. \cite{JK-00}. 
218: %However, the reliability of the
219: %error bars reported in Ref.~\cite{JK-00} is unclear: indeed, 
220: %Ref. \cite{GZ-98} analyzes the 
221: %same perturbative series and reports much more cautious error estimates.
222: 
223: Concerning the leading scaling-correction exponent $\omega$,
224: we mention the estimates $\omega=0.782(13)$ obtained from
225: the analysis of the six-loop fixed-dimension expansion \cite{GZ-98},
226: $\omega=0.794(18)$ from the five-loop $\epsilon$ expansion
227: \cite{GZ-98}, $\omega\approx 0.773$ from MC simulations \cite{Hasenbusch-00}.
228: Correspondingly, using \cite{CHPRV-02} $\nu = 0.7112(5)$, we have 
229: $\Delta = \omega\nu = 0.556(9)$, 0.565(13), 0.550.
230: 
231: We finally report some results for the 
232: crossover exponent $\phi_2$ associated with 
233: the spin-2 traceless tensor field 
234: $O^{ab}(x) = \phi^a(x) \phi^b(x) - \case{1}{3}\delta^{ab} \phi(x)^2$,
235: see Sec.~\ref{sec-1.5.8}, which
236: describes the instability of the O(3)-symmetric 
237: theory against anisotropy \cite{FP-72,Wegner-72,FN-74,Aharony-76}. 
238: The crossover exponent $\phi_2$ has been determined 
239: using various approaches, obtaining
240: $\phi_2 = 1.271(21)$ by the analysis of the six-loop expansion
241: in the framework of the fixed-dimension FT expansion
242: \cite{CPV-02}; 
243: $\phi_2\approx 1.22$ 
244: by setting $\epsilon=1$ in the corresponding $O(\epsilon^3)$ 
245: series \cite{Yamazaki-74};
246: $\phi_2=1.250(15)$ by HT expansion
247: techniques \cite{PJF-74}.
248: The exponent $\phi_4$ can be derived from the results of Ref. \cite{CPV-00}
249: for the $O(N)$ model with a cubic-symmetric perturbation, see 
250: Sec. \ref{lsec-cubic}. One finds \cite{CPV-02-4} 
251: $\phi_4 = 0.009(4)$.  Since $\phi_2>0$ and $\phi_4 > 0$, 
252: the spin-2 and the spin-4 (we also expect the 
253: spin-3) operators are relevant perturbations. Higher-order spin operators
254: are expected to be RG irrelevant. 
255: 
256: 
257: 
258: 
259: \subsubsection{Experimental results}
260: \label{expO3ex}
261: 
262: In Table \ref{table_exponents_exp_O3} we report 
263: some recent experimental results for ferromagnets and
264: antiferromagnets. It is not 
265: a complete review of published results, but it is
266: useful to get an overview of the experimental state of the art.  
267: In the table we have also included results for the well-studied 
268: doped manganese perovskites La$_{1-x}$A$_{x}$MnO$_3$, although the 
269: nature of the ferromagnetic transition in these compounds is still 
270: unclear.\footnote{
271: For some dopings and some divalent cation A a first-order
272: transition has been observed. Moreover, in systems in which the transition 
273: appears to be of second order, mean-field critical exponents have been
274: measured. For instance, for La$_{1-x}$Sr$_x$MnO$_3$, a mean-field 
275: value for $\beta$ was observed in 
276: Refs.~\cite{LRKBMT-97,MSKMM-98,SSA-00},
277: while an estimate compatible with the Heisenberg value was found in
278: Refs. \cite{MSEHMT-96,HLHNLKNUMC-96,LBGGKSAM-97,GLGKSAM-98}.
279: For $x=1/3$ there also exists \cite{Ramos-etal-01} an estimate of the exponent
280: $\alpha$, $\alpha = - 0.14\pm 0.10$, in agreement with the 
281: Heisenberg value. See also the recent review \cite{SJ-01}.}
282: 
283: The Heisenberg universality class also describes isotropic 
284: magnets with quenched disorder. 
285: The experimental results confirm this theoretical prediction,\footnote{
286: In order to observe the correct exponents, it is essential to consider 
287: corrections to scaling in the analysis of the experimental data \cite{Kaul-85}.
288: All results reported in Table \ref{table_exponents_exp_O3_2}, except 
289: those of Ref. \cite{SHAM-99,Tsurkan-etal-99}, have been obtained by assuming 
290: scaling corrections of the form $(1 + a |t|^{\Delta_1} + b 
291: |t|^{\Delta_2})$, with $\Delta_1 = 0.11$ and 
292: $\Delta_2 = 0.55$. Note that the value of $\Delta_1$ is slightly lower
293: than the precise theoretical estimate of Ref. \cite{CHPRV-02}, 
294: $\Delta_1 = 0.1336(15)$, and that RG
295: predicts additional corrections 
296: of order $|t|^{2\Delta_1}$, $|t|^{3\Delta_1}$, etc., which are more 
297: relevant than the term $|t|^{\Delta_2}$ and should therefore be taken into 
298: account in the analysis of the data.}
299: as it can be seen 
300: from Table \ref{table_exponents_exp_O3_2} (older experimental results with 
301: a critical discussion are reported in Ref. \cite{Kaul-85}). 
302: Finally, we mention the experiment 
303: reported in Ref.~\cite{CH-88}
304: on Fe$_{1-x}$V$_x$ in the presence of annealed
305: disorder; as predicted by theory, they obtain $\beta=0.362(8)$, in agreement
306: with the corresponding Heisenberg exponent.
307: 
308: Beside the exponents $\gamma$, $\beta$, and $\delta$ there are
309: also a few estimates of the specific-heat exponent $\alpha$, in most 
310: of the cases obtained from resistivity measurements: 
311: $\alpha \approx -0.10$ in Fe and Ni \cite{KHM-81};
312: $\alpha = -0.12(2)$ in EuO \cite{SWSLPK-92}; 
313: $\alpha = -0.11(1)$ in Fe$_x$Ni$_{80-x}$B$_{19}$Si \cite{KR-94};
314: $\alpha = -0.11(1)$ in RbMnF$_3$ \cite{MMFB-96};
315: $\alpha \approx -0.12$ in Sr$_2$FeMoO$_6$ \cite{YCSXM-02}.	
316: 
317: Some experimental estimates of crossover exponent
318: $\phi_2$ are
319: reported in Ref.~\cite{PHA-91}. We mention the experimental
320: result $\phi_2=1.279(31)$ for the bicritical point in MnF$_2$ \cite{KR-79}.
321: 
322: 
323: \begin{table*}
324: \caption{
325: Experimental estimates of the critical exponents for Heisenberg systems.}
326: \label{table_exponents_exp_O3}
327: \footnotesize
328: \begin{center}
329: \begin{tabular}{rllll}
330: \hline
331: \multicolumn{1}{c}{Ref.}& 
332: \multicolumn{1}{c}{Material}& 
333: \multicolumn{1}{c}{$\gamma$}& 
334: \multicolumn{1}{c}{$\beta$}&
335: \multicolumn{1}{c}{$\delta$} \\   
336: \hline  
337: \cite{SBEW-80} $_{1980}$        & Ni     &           & 0.354(14) &       \\
338: \cite{Kobeissi-81} $_{1981}$    & Fe     &           & 0.367(5)  &       \\
339: \cite{Seeger-etal-95} $_{1995}$ & Ni     & 1.345(10) & 0.395(10) & 4.35(6) \\
340: \cite{RKS-95} $_{1995}$     & Gd$_2$BrC  & 1.392(8)  & 0.365(5)  &4.80(25) \\
341: \cite{RKS-95} $_{1995}$     & Gd$_2$IC   & 1.370(8)  & 0.375(8)  &4.68(25) \\
342: \cite{Zhao-etal-99} $_{1999}$ & Tl$_2$Mn$_2$O$_7$ 
343:                                            & 1.31(5)   & 0.44(6)   &4.65(15) \\
344: \cite{Barsov-etal-00} $_{2000}$ & La$_{0.82}$Ca$_{0.18}$MnO$_3$ 
345:                                            &           & 0.383(9)  &         \\
346: \cite{Zhao-etal-00} $_{2000}$ &  La$_{0.95}$Ca$_{0.05}$MnO$_3$  
347:                                            & 1.39(5)   & 0.36(7)   &4.75(15) \\
348: \cite{ArPa-00} $_{2000}$    & Gd(0001)   &           & 0.376(15) &         \\
349: \cite{MRBSGV-00} $_{2000}$  & Gd$_2$CuO$_4$ & 1.32(2)& 0.34(1)   &         \\
350: \cite{Buhrer-etal-00} $_{2000}$ & C$_{80}$Pd$_{20}$ (liq) & 1.42(5) & &    \\
351: \cite{Buhrer-etal-00} $_{2000}$ & C$_{80}$Pd$_{20}$ (sol) & 1.40(8) & &    \\
352: \cite{Brueckel-etal-01} $_{2001}$ & GdS  &           & 0.38(2)   &         \\
353: \cite{Yang-etal-01} $_{2001}$ & CrO$_2$  & 1.43(1)   & 0.371(5)  &         \\
354: \cite{HKH-01} $_{2001}$  &  La$_{0.8}$Ca$_{0.2}$MnO$_3$   
355:                     & 1.45    & 0.36     & \\
356: \cite{YCSXM-02} $_{2002}$ & Sr$_2$FeMoO$_6$ & 1.30 & 0.388 & 4.35 \\
357: \hline
358: \end{tabular}
359: \end{center}
360: \end{table*}
361: 
362: \begin{table*}
363: \caption{
364: Experimental estimates of the critical exponents for Heisenberg
365: systems with quenched disorder.}
366: \label{table_exponents_exp_O3_2}
367: \footnotesize
368: \begin{center}
369: \begin{tabular}{rllll}
370: \hline
371: \multicolumn{1}{c}{Ref.}& 
372: \multicolumn{1}{c}{Material}& 
373: \multicolumn{1}{c}{$\gamma$}& 
374: \multicolumn{1}{c}{$\beta$}&
375: \multicolumn{1}{c}{$\delta$} \\   
376: \hline  
377: \cite{KR-94} $_{1994}$ & Fe$_{10}$Ni$_{70}$Bi$_{19}$Si 
378:                   &    1.387(12)   & 0.378(15)   & 4.50(5)   \\
379: \cite{KR-94} $_{1994}$ & Fe$_{13}$Ni$_{67}$Bi$_{19}$Si 
380:                   &    1.386(12)   & 0.367(15)   & 4.50(5)   \\
381: \cite{KR-94} $_{1994}$ & Fe$_{16}$Ni$_{64}$Bi$_{19}$Si 
382:                   &    1.386(14)   & 0.360(15)   & 4.86(4)   \\
383: \cite{RK-95,RK-95b} $_{1995}$ & Fe$_{20}$Ni$_{60}$P$_{14}$B$_6$
384:                   &    1.386(10)   & 0.367(10)   & 4.77(5)   \\
385: \cite{RK-95,RK-95b} $_{1995}$ & Fe$_{40}$Ni$_{40}$P$_{14}$B$_6$
386:                   &    1.385(10)   & 0.364(5)    & 4.79(5)   \\
387: \cite{BK-97} $_{1997}$ & Fe$_{91}$Zr$_9$ 
388:                   &    1.383(4)    & 0.366(4)    & 4.75(5)   \\ 
389: \cite{BK-97} $_{1997}$ & Fe$_{89}$CoZr$_{10}$ 
390:                   &    1.385(5)    & 0.368(6)    & 4.80(4)   \\
391: \cite{BK-97} $_{1997}$ & Fe$_{88}$Co$_2$Zr$_{10}$ 
392:                   &    1.389(6)    & 0.363(5)    & 4.81(5)   \\
393: \cite{BK-97} $_{1997}$ & Fe$_{84}$Co$_6$Zr$_{10}$ 
394:                   &    1.386(6)    & 0.370(5)    & 4.84(5)   \\
395: \cite{SHAM-99} $_{1999}$ & Fe$_{1.85}$Mn$_{1.15}$Si 
396:                   &    1.543(20)   & 0.408(60)   & 4.74(7)   \\
397: \cite{SHAM-99} $_{1999}$ & Fe$_{1.50}$Mn$_{1.50}$Si 
398:                   & 1.274(60)      & 0.383(10)   & 4.45(19)  \\
399: 
400: \cite{Tsurkan-etal-99} $_{1999}$ & MnCr$_{1.9}$In$_{0.1}$S$_4$ 
401:                   &    1.39(1)   & 0.36(1)   & 4.814(14)   \\
402: 
403: \cite{Tsurkan-etal-99} $_{1999}$ & MnCr$_{1.8}$In$_{0.2}$S$_4$ 
404:                   &    1.39(1)   & 0.36(1)  & 4.795(10)   \\
405: 
406: \cite{Perumal-etal-00} $_{2000}$ & Fe$_{86}$Mn$_4$Zr$_{10}$
407:                   & 1.381          & 0.361       & \\
408: \cite{Perumal-etal-00} $_{2000}$ & Fe$_{82}$Mn$_8$Zr$_{10}$
409:                   & 1.367          & 0.363       & \\
410: \cite{PSKYRD-01} $_{2001}$ & Fe$_{84}$Mn$_6$Zr$_{10}$
411:                   & 1.37(3)        & 0.359       & 4.81(4)    \\
412: \cite{PSKYRD-01} $_{2001}$ & Fe$_{74}$Mn$_{16}$Zr$_{10}$
413:                   & 1.39(5)        & 0.361       & 4.86(3)    \\
414: \hline
415: \end{tabular}
416: \end{center}
417: \end{table*}
418: 
419: 
420: \subsection{The critical equation of state}
421: \label{eqstO3}
422: 
423: \subsubsection{Approximate representations}
424: 
425: The critical equation of state can be determined  
426: using the method described in Sec.~\ref{pareqstXY}
427: in the context of the $XY$ universality class, 
428: i.e. using the small-magnetization expansion of the free energy
429: to construct approximate parametric
430: representations following the schemes A and B,
431: cf. Eqs.~(\ref{scheme1}) and (\ref{scheme2}).
432: 
433: \begin{table*}
434: \caption{
435: Estimates of $g_4^+$, $r_{6}$, and $r_8$ for the Heisenberg universality
436: class. We also mention the estimate 
437: $r_{10}=-6(3)$ obtained 
438: by studying the equation of state \protect\cite{CHPRV-02}.
439: }
440: \label{summarygjO3}
441: \footnotesize
442: \tabcolsep 4pt        % Less than the usual 6pt
443: \begin{center}
444: \begin{tabular}{cllll}
445: \hline
446: \multicolumn{1}{c}{}& 
447: \multicolumn{1}{c}{HT}& 
448: \multicolumn{1}{c}{$d=3$ exp}&
449: \multicolumn{1}{c}{$\epsilon$ exp} &
450: \multicolumn{1}{c}{CRG} \\
451: \hline  
452: $g_4^+$   & 19.13(10) \cite{CHPRV-02} & 19.06(5) \cite{GZ-98} &
453:           19.55(12) \cite{PV-00,PV-gr-98}  & 22.35 \cite{BTW-96,BTW-99}\\
454:  
455:         & 19.31(14),$\;$ 19.27(11) \cite{BC-98}  &  19.06 \cite{MN-91} & & \\ 
456:         & 19.34(16) \cite{PV-gr-98}  & &  &  \\ 
457: 
458: $r_6$   & 1.86(4) \cite{CHPRV-02}  & 1.880 \cite{SOUK-99} &
459:           1.867(9) \cite{PV-00,PV-ef-98} & 1.74 \cite{TW-94} \\
460:         & 2.1(6) \cite{Reisz-95} & 1.884(32) \cite{PV-00} & & \\ 
461: 
462: $r_8$   & 0.6(2)  & 0.975 \cite{SOUK-99} &
463:           1.0(6) \cite{PV-00,PV-ef-98} & 0.84 \cite{TW-94}\\
464: 
465: $r_{10}$   & $-$15(10) & & & \\
466: \hline
467: \end{tabular}
468: \end{center}
469: \end{table*}
470: 
471: 
472: In Table~\ref{summarygjO3} we report a summary of the available results for
473: the zero-momentum four-point coupling $g_4^+$, cf. Eq. (\ref{grdef}), 
474: and for the coefficients $r_{6}$, $r_8$, and $r_{10}$ that  
475: parametrize the small-magnetization expansion of the 
476: Helmholtz free energy, cf. Eq. (\ref{Fzdef}).
477: 
478: Figure \ref{figFzO3}
479: shows the scaling functions $F(z)$, $f(x)$, and $D(y)$, as obtained
480: in Ref. \cite{CHPRV-02}. They used 
481: schemes $A$ and $B$ with $n=0,1$, and
482: the (MC+IHT) estimates of $\gamma$, $\nu$, $r_6$, and $r_8$.
483: The three approximations of $F(z)$ are practically
484: indistinguishable, and differ at most by approximately 2\%
485: (the difference between the two $n=1$ curves is much smaller).
486: The large-$z$ behavior of $F(z)$ is well determined, indeed
487: $F_0^\infty = 0.0266(5)$.
488: The three curves for $f(x)$ 
489: are in substantial agreement, especially 
490: those with $n=1$. Indeed, the difference between them 
491: is within the uncertainty due to the errors on the input parameters.
492: These approximate parametric representations are not 
493: precise at the coexistence curve, providing only a rough estimate of 
494: the coefficient $c_f$, cf. Eq. (\ref{fxcc}),
495: i.e. $c_f=5(3)$.
496: We also report
497: the estimates of  the coefficients $f_n^0$, 
498: cf. Eq.~(\ref{expansionfx-xeq0}) obtained in Ref.~\cite{CHPRV-02}:
499: $f_1^0 = 1.34(5)$, $f_2^0 = 0.20(2)$, $f_3^0 = -0.10(1)$.
500: 
501: The scaling function $f(x)$ was also determined in Ref.~\cite{BTW-96} 
502: by CRG methods using the lowest order of the derivative expansion.
503: In Fig.~\ref{figFzO3}, together with the results of Ref.~\cite{CHPRV-02} for 
504: $f(x)$, we also show the approximate scaling function $f(x)$
505: obtained in Ref.~\cite{BTW-96}.
506: We note sizeable differences between the results of the two approaches.
507: 
508: \begin{figure}[tb]
509: \hspace{0cm}
510: \begin{tabular}{cc}
511: \hskip -0.6truecm
512: \psfig{width=7.5truecm,angle=0,file=fxO3.eps} &
513: \hskip 0.6truecm
514: \psfig{width=7.5truecm,angle=-0,file=FzO3.eps} \\[12mm]
515: \multicolumn{2}{c}{\psfig{width=8truecm,angle=-0,file=Dy.eps}}
516: \end{tabular}
517: \vspace{0cm}
518: \caption{
519: The scaling functions $f(x)$, $F(z)$, and 
520: $D(y)$ 
521: for the Heisenberg universality class. 
522: All results have been obtained in  Ref. \cite{CHPRV-02},
523: except those labelled by CRG (Ref. \cite{BTW-96}).
524: }
525: \label{figFzO3}
526: \end{figure}
527: 
528: 
529: 
530: %% \begin{figure}[tb]
531: %% \hspace{0cm}
532: %% \centerline{\psfig{width=7.5truecm,angle=0,file=fxO3.eps}}
533: %% \vspace{0cm}
534: %% \caption{
535: %% The scaling function $f(x)$ for the Heisenberg universality class. 
536: %% Results from  Refs. \cite{CHPRV-02,BTW-96}. 
537: %% }
538: %% \label{figfxO3}
539: %% \end{figure}
540: %% 
541: %% 
542: %% \begin{figure}[tb]
543: %% \hspace{0cm}
544: %% \centerline{\psfig{width=7.5truecm,angle=-0,file=duO3.eps}}
545: %% \vspace{0cm}
546: %% \caption{
547: %% The scaling function $D_R(y_R)\equiv D(y)/D(y_{\rm  max})$ versus
548: %% $y_R\equiv y/y_{\rm max}$, cf. Eq.~(\ref{defDw}). 
549: %% Results from  Ref. \cite{CHPRV-02}.
550: %% }
551: %% \label{figdyO3}
552: %% \end{figure}
553: 
554: 
555: \subsubsection{Universal amplitude ratios}
556: \label{unraO3}
557: 
558: In Table~\ref{univratiosO3} we report the estimates of  several
559: universal amplitude ratios. 
560: The results denoted by IHT--PR were obtained in
561: Ref.~\cite{CHPRV-02}  using approximate parametric representations
562: of the equation of state.
563: The FT estimates of $U_0$ were obtained from the analysis
564: of the fixed-dimension expansion in the 
565: minimal-renormalization scheme without $\epsilon$ expansion 
566: \cite{LMSD-98,KV-00}
567: and from the standard $\epsilon$ expansion to $O(\epsilon^2)$ 
568: \cite{Bervillier-86}.
569: The CRG estimate of $U_0$ and $R_\alpha$ were  obtained in
570: Ref. \cite{CHPRV-02}
571:  using the expression for $f(x)$ reported in
572: Refs.~\cite{BTW-96,BTW-99}; they significantly differ from the 
573: estimates obtained using other methods.
574: See, e.g., Ref.~\cite{PHA-91} for a more complete review of 
575: theoretical and experimental estimates.
576: 
577: 
578: \begin{table*}
579: \caption{
580: Estimates of universal amplitude ratios for the Heisenberg universality class.
581: The numbers marked by an asterisk have been obtained in Ref. \cite{CHPRV-02}
582:  using the results reported in the quoted references.
583: }
584: \label{univratiosO3}
585: \footnotesize
586: \begin{center}
587: \begin{tabular}{ccccccc}
588: \hline
589:  & IHT--PR \cite{CHPRV-02} & $d$=3 exp & $\epsilon$ exp & CRG & HT & experiments \\
590: \hline
591: $U_0$ & 1.56(4) & 1.51(4) \cite{LMSD-98} & 1.521(22) \cite{Bervillier-86} & 
592:             $^*$1.823 \cite{BTW-96,BTW-99} & & 1.50(5) \cite{KR-94}\\
593:       &         & 1.544 \cite{KV-00} & & & &   1.27(9) \cite{MMFB-96} \\
594:       & & & & & & 1.4(4) \cite{Ramos-etal-01} \\ 
595: $R_\alpha$ & 4.3(3)  &  $^*$4.4(4) \cite{LMSD-98}  & 4.56(9)
596:            \cite{Bervillier-86} & $^*$3.41 \cite{BTW-96,BTW-99} &&  \\
597:            &         &  $^*$4.46 \cite{KV-00}  &   &&&  \\
598: 
599: $R_\chi$ &  1.31(7) &  & 1.33 \cite{AM-78} & 1.11 \cite{BTW-96,BTW-99} & & \\
600: 
601: $R_C$    & 0.185(10) & 0.189(9) \cite{SLD-99} & 0.17 \cite{AH-76} & & & \\
602:          & & 0.194 \cite{KV-00} & & & & \\
603: 
604: 
605: $R_4$ & 7.8(3) & &&& & \\
606: 
607: $R_\xi^+$ & 0.424(3) & 0.4347(20) \cite{BB-85} & 0.42 \cite{Bervillier-76} &&
608: 0.431(5) \cite{BC-99} & \\
609: & & 0.4319(17) \cite{BG-80} & & & 0.433(5) \cite{BC-99}& \\ 
610: 
611: $P_m$ & 1.18(2) & &&& & \\
612: 
613: %$P_c$ & 0.357(5) & &&& & \\
614: 
615: $R_p$ & 2.020(6) & &&& & \\
616: \hline
617: \end{tabular}
618: \end{center}
619: \end{table*}
620: 
621: 
622: \subsubsection{Comparison with the experiments} \label{CES.E}
623: 
624: 
625: In spite of the large number of experiments, at present there is no 
626: accurate quantitative study of the equation of state in the critical 
627: regime. We shall discuss here three different representations that 
628: are widely used in experimental work.
629: 
630: A first possibility \cite{KR-67} consists in studying the behavior of 
631: $h/m\equiv H |t|^{-\gamma}/M$ versus $m^2 \equiv M^2 |t|^{-2\beta}$. 
632: Such a function 
633: can be easily obtained from approximations of $f(x)$, since 
634: $m^2 = B^2 |x|^{-2\beta}$ and 
635: \begin{equation} 
636:    {h\over m} = k |x|^{-\gamma} f(x)
637: \end{equation}
638: where $k = \left(B_c \right)^{-\delta} B^{\gamma/\beta} = {R_\chi/C^+}$.
639: A plot of $m^2/B^2$ versus $C^+ h/m$ is reported in Fig. \ref{figm2ohmO3}.
640: It agrees qualitatively with the analogous experimental ones 
641: reported, e.g., in Refs. \cite{KR-94,BK-97,FKK-01}.
642: 
643: Often, for small $h/m$ one approximates the equation of state by
644: writing 
645: \begin{equation}
646:    {h\over m} = a_{\pm} + b_{\pm} m^2,
647: \label{Kouvell-Rodbell}
648: \end{equation}
649: where $a_{\pm}$ and $b_\pm$ are numerical coefficients depending on the phase.
650: Such an approximation has a limited range of validity. In the HT
651: phase, one obtains for $m^2 \to 0$ \cite{CHPRV-02}
652: \begin{eqnarray}
653: {h\over m} &=& {1\over C^+} \left[ 1 + {R_4\over 6} {m^2\over B^2} + 
654:    \sum_{n=2}^\infty {R^n_4 \ r_{2n+2}\over (2n+1)!} 
655:    \left({m^2\over B^2}\right)^n \right]
656: \nonumber \\
657: &\approx& {1\over C^+} \Bigl[ 1 + 1.30(5)\ {m^2\over B^2} + 
658:               0.94(8)\ \left({m^2\over B^2}\right)^2        
659: + 0.06(2)\ \left({m^2\over B^2}\right)^3 + \cdots\Bigr].
660: \label{hsum-expansion}
661: \end{eqnarray}
662: From Eq. (\ref{hsum-expansion}), one sees that the 
663: approximation (\ref{Kouvell-Rodbell}) is valid only for very small $m^2$, 
664: i.e. at the 1\% level only for $m^2 \lesssim 0.01 B^2$. 
665: The quadratic approximation---i.e. the approximation with an 
666: additional $(m^2)^2$ term---has a much wider range of validity 
667: because of the smallness of the coefficient of $m^6$.
668: 
669: \begin{figure}[tb]
670: \hspace{0cm}
671: \centerline{\psfig{width=7.5truecm,angle=-0,file=m2homO3.eps}}
672: \vspace{0cm}
673: \caption{
674: Plot of $m^2/B^2$ versus $C^+ h/m$. From Ref. \cite{CHPRV-02}.
675: }
676: \label{figm2ohmO3}
677: \end{figure}
678: 
679: In the low-temperature phase, Eq. (\ref{Kouvell-Rodbell}) is theoretically
680: incorrect, since it does not take into account the presence of 
681: Goldstone modes. Indeed, for $m^2/B^2\to 1$, we have 
682: \begin{equation}
683: {h\over m} \approx {k c_f\over 4 \beta^2} \left(1 - {m^2\over B^2}\right)^2,
684: \label{hsum-coex}
685: \end{equation}
686: where $c_f$ is defined in Eq. (\ref{fxcc}).
687: Eq. (\ref{hsum-coex}) is inconsistent with the approximation
688: (\ref{Kouvell-Rodbell}) near the coexistence curve.
689: 
690: Finally, note that for $m^2$ large 
691: we have
692: \begin{equation}
693: {h\over m} \approx k \left({m\over B}\right)^{\delta-1}.
694: \end{equation}
695: A second form that is widely used to analyze the experimental data 
696: is the Arrott-Noakes \cite{AN-67} scaling equation
697: \begin{equation}
698: \left({H\over M}\right)^{1/\gamma} = a t + b M^{1/\beta},
699: \end{equation}
700: where $a$ and $b$ are numerical constants. This approximation 
701: is good in a neighborhood of the critical isotherm $t=0$. 
702: Since
703: \begin{equation}
704: \left({H\over M}\right)^{1/\gamma} k^{-1/\gamma} = 
705: \left({M\over B}\right)^{1/\beta} f(x)^{1/\gamma},
706: \end{equation}
707: one obtains \cite{CHPRV-02}
708: \begin{eqnarray}
709: \left({H\over M}\right)^{1/\gamma} k^{-1/\gamma} = 
710: \left({M\over B}\right)^{1/\beta} + 
711:  0.96(4)\ t 
712: - 0.04(2)\, t^2 \left({M\over B}\right)^{-1/\beta}
713:             - 0.02(2)\, t^3 \left({M\over B}\right)^{-2/\beta}
714:         \cdots 
715: \end{eqnarray}
716: Thus, at a 1\% level of precision the Arrott-Noakes formula is valid approximately 
717: for $t (M B^{-1})^{-1/\beta} \lesssim  25$ which is quite a large interval.
718:   
719: Finally, Ref.~\cite{Zhao-etal-99} reports an experimental study of
720: the behavior of Tl$_2$Mn$_2$O$_7$ along the crossover line,
721: and determines the scaling function $D(y)$, although with
722: different normalizations. The comparison of their results
723: with the theoretical curve $D(y)$ obtained 
724: in Ref. \cite{CHPRV-02}, see Sec.~\ref{eqstO3},
725: shows a very nice  quantitative agreement.
726: 
727:  
728: 
729: 
730: 
731: 
732: 
733: 
734: