1: \documentclass[12pt]{article}
2: %\documentstyle[12pt,epsf]{article}%
3: \textwidth16cm
4: \textheight23.5cm
5: \oddsidemargin 0cm
6: \topmargin-25mm
7: \renewcommand{\baselinestretch}{1.1}
8: \usepackage{amsmath,amssymb,epsfig}
9: \begin{document}
10: \parindent=0pt
11: \parskip=6pt
12: \rm
13:
14:
15:
16: \begin{center}
17: {\bf \large Magnetic effects on the phase transitions in\\ unconventional
18: superconductors}
19:
20:
21:
22:
23: {\bf Dimo I. Uzunov}
24:
25: Max-Plank-Institut f\"{u}r Physik komplexer
26: Systeme, N\"{o}tnitzer Str. 38, 01187 Dresden, Germany,
27:
28: [Permanent address: CP Laboratory, Institute of Solid
29: State Physics, Bulgarian Academy of Sciences, BG-1784 Sofia,
30: Bulgaria.]
31:
32:
33: \end{center}
34:
35: \begin{abstract}
36: The effect of magnetic fluctuations on the critical behaviour of unconventional
37: ferromagnetic superconductors (UGe$_2$, URhGe, etc.) and
38: superfluids is investigated by the renormalization-group method.
39: For the case of isotropic ferromagnetic order a new unusual
40: critical behaviour is predicted. It is also shown that the
41: uniaxial and bi-axial magnetic symmetries produce fluctuation
42: driven phase transitions of first order. The results can be used
43: in interpretations of experimental data and for a further
44: development of the theory of critical phenomena in complex
45: systems.
46: \end{abstract}
47:
48: Experiments~\cite{Saxena:2000, Huxley:2001} at low temperatures
49: ($T \sim 1$ K) and high pressure ($P\sim 1$ GPa) demonstrated the
50: existence of spin triplet superconducting states in the metallic
51: compound UGe$_2$. The superconductivity is triggered by the
52: spontaneous magnetization of the ferromagnetic phase that occurs
53: at much higher temperatures~\cite{ShopovaPL:2003,Shopova:2005,
54: Shopova:2006}. The ferromagnetic order coexists with the
55: superconducting phase in the whole domain of its existence below
56: $T \sim 1.2$. The same phenomenon of existence of
57: superconductivity at low temperatures and high pressure in the
58: domain of the $(T,P)$ phase diagram where the ferromagnetic order
59: is present was observed in URhGe~\cite{Aoki:2001}). These
60: remarkable phenomena occur through phase transitions of first and
61: second order and multi-critical points which present a
62: considerable experimental and theoretical interest. A fragment of
63: ($P,T$) phase diagrams of itinerant ferromagnetic
64: compounds~\cite{Saxena:2000} is sketched in fig.~1, where the
65: lines $T_F(P)$ and $T_c(P)$ of the
66: paramagnetic(P)-to-ferromagnetic(F) and
67: ferromagnetic-to-coexistence phase(C) transitions are very close
68: to each other and intersect at very low temperature or terminate
69: at the absolute zero ($P_0,0$). At low temperature, where the
70: phase transition lines are close enough to each other, the
71: interaction between the real magnetization vector
72: $\mbox{\boldmath$M$}(\mbox{\boldmath$r$)} =
73: \{M_j(\mbox{\boldmath$r$}); j=1,...,m \}$ and the complex order
74: parameter vector of the spin-triplet Cooper
75: pairing~\cite{Sigrist:1991}, $\psi(\mbox{\boldmath$r$}) =
76: \{\psi_{\alpha}(\mbox{\boldmath$r$}) = (\psi^{\prime}_{\alpha} +
77: i\psi_{\alpha}^{\prime \prime}); \alpha =1,.... n/2\}$ ($n=6$)
78: cannot be neglected~\cite{Uzunov:1993} and, as shown here, this
79: interaction produces new fluctuation phenomena.
80:
81:
82: In this letter a new critical behavior for this type of systems is
83: established and described. The new critical behaviour occurs in
84: real systems with isotropic magnetic order but does not belong to
85: any known universality class~\cite{Uzunov:1993}. Thus it could be
86: of considerable experimental and theoretical interest. Due to
87: crystal and magnetic anisotropy a new type of fluctuation-driven
88: first order phase transitions occur, as shown in the present
89: investigation. The quantum effects~\cite{Uzunov:1993, Hertz:1976,
90: Shopova:2003} on the properties of these novel phase transitions
91: are briefly discussed.
92:
93:
94: Both thermal fluctuations at finite temperatures ($T>0$) and
95: quantum fluctuations (correlations) near the $P$--driven quantum
96: phase transition at $T=0$ should be considered but at a first
97: stage the quantum effects~\cite{Shopova:2003} can be neglected as
98: irrelevant to finite temperature phase transitions ($T_F \sim T_c
99: >0$). The present treatment of a recently derived free energy
100: functional~\cite{Machida:2001, Walker:2002} by the standard
101: Wilson-Fisher renormalization group (RG)~\cite{Uzunov:1993} shows
102: that unconventional ferromagnetic superconductors with an
103: isotropic magnetic order ($m=3$) exhibit a very special
104: multi-critical behavior for any $T> 0$, whereas the magnetic
105: anisotropy ($m =1,2$) generates fluctuation-driven first order
106: transitions~\cite{Uzunov:1993}. Thus the phase transition
107: properties of spin-triplet ferromagnetic superconductors are
108: completely different from those predicted by mean field
109: theories~\cite{ShopovaPL:2003, Shopova:2005, Shopova:2006,
110: Machida:2001, Walker:2002}. The results can be used in the
111: interpretation of experimental data for phase transitions in
112: itinerant ferromagnetic compounds~\cite{Pfleiderer:2002}.
113:
114: The study presents for the first time an example of complex
115: quantum criticality characterized by a double-rate quantum
116: critical dynamics. In the quantum limit ($T\rightarrow 0$) the
117: fields $\mbox{\boldmath$M$}$ and $\psi$ have different dynamical
118: exponents, $z_M$ and $z_{\psi}$, and this leads to two different
119: upper critical dimensions: $d_U^M = 6-z_M$ and $d_{\psi}^U =
120: 6-z_{\psi}$. The complete consideration of the quantum
121: fluctuations of both fields $\mbox{\boldmath$M$}$ and $\psi$
122: requires a new RG approach in which one should either consider the
123: difference $(z_M-z_{\psi})$ as an auxiliary small parameter or
124: create a completely new theoretical paradigm of description. The
125: considered problem is quite general and presents a challenge to
126: the theory of quantum phase transitions~\cite{Shopova:2003}. The
127: obtained results can be applied to any natural system with the
128: same class of symmetry although this letter is based on the
129: example of itinerant ferromagnetic compounds.
130:
131: The relevant part of the fluctuation Hamiltonian of unconventional
132: ferromagnetic superconductors~\cite{Shopova:2005, Shopova:2006,
133: Machida:2001, Walker:2002} can be written in the form
134: \begin{equation}
135: \label{eq1} {\cal{H}}= \sum_{\mbox{\boldmath$k$}}\left[ \left(r +
136: k^2 \right)|\psi(\mbox{\boldmath$k$})|^2 + \frac{1}{2}\left(t +
137: k^2\right)|\mbox{\boldmath$M$}(\mbox{\boldmath$k$})|^2 \right]
138: +\frac{ig}{\sqrt{V}}\sum_{\mbox{\boldmath$k$}_1,{\mbox{\boldmath$k$}_2}}{\mbox{\boldmath$M$}}
139: \left(
140: {\mbox{\boldmath$k$}}_1\right).\left[\psi\left({\mbox{\boldmath$k$}}_2
141: \right)\times \psi^{\ast}\left({\mbox{\boldmath$k$}}_1 +
142: {\mbox{\boldmath$k$}}_2 \right) \right]
143: \end{equation}
144: where $V \sim L^d$ is the volume of the $d-$dimensional system,
145: the length unit is chosen so that the wave vector
146: ${\mbox{\boldmath$k$}}$ is confined below unity ($ 0 \leq k =
147: |{\mbox{\boldmath$k$}}| \leq 1)$, $g \geq 0$ is a coupling
148: constant, describing the effect the scalar product of
149: ${\mbox{\boldmath$M$}}$ and the vector product
150: $(\psi\times\psi^{\ast})$ for symmetry indices $m = (n/2)=3$, and
151: the parameters $t \sim (T-T_f)$ and $r \sim (T-T_s)$ are expressed
152: by the critical temperatures of the generic ($g\equiv 0$)
153: ferromagnetic and superconducting transitions. As mean field
154: studies indicate~\cite{Machida:2001,Shopova:2003}, $T_s(P)$ is
155: much lower than $T_c(T)$ and $T_F(P) \neq T_f(P)$.
156:
157: The fourth order terms ($M^4, |\psi|^4, M^2|\psi|^2$) in the total
158: free energy (effective Hamiltonian)~\cite{Shopova:2005,
159: Shopova:2006, Machida:2001, Walker:2002} have not been included in
160: eq.~(1) as they are irrelevant to the present investigation. The
161: simple dimensional analysis shows that the $g-$term in eq.~(1)
162: corresponds to a scaling factor $b^{3-d/2}$ and, hence, becomes
163: relevant below the upper borderline dimension $d_U=6$, while
164: fourth order terms are scaled by a factor $b^{4-d}$ as in the
165: usual $\phi^4-$theory and are relevant below $d<4$ ($b > 1$ is a
166: scaling number)~\cite{Uzunov:1993}. Therefore we should perform
167: the RG investigation in spatial dimensions $d = 6-\epsilon$ where
168: the $g$--term in eq.~(1) describes the only relevant fluctuation
169: interaction. Moreover, the total fluctuation
170: Hamiltonian~\cite{Shopova:2005, Shopova:2006, Machida:2001}
171: contains off-diagonal terms of the form
172: $k_ik_j\psi_{\alpha}\psi^{\ast}_{\beta}$; $i\neq j$ and/or $\alpha
173: \neq \beta$. Using a convenient loop expansion these terms can be
174: completely integrated out from the partition function to show that
175: they modify the parameters ($r,t,g$) of the theory but they do not
176: affect the structure of the model (1). Such terms change auxiliary
177: quantities, for example, the coordinates of the RG fixed points
178: (FPs) but they do not affect the main RG results for the stability
179: of the FPs and the values of the critical exponents. Here we
180: ignore these off-diagonal terms.
181:
182: One may consider several cases: (i) uniaxial magnetic symmetry,
183: ${\mbox{\boldmath$M$}} = (0,0,M_3)$, (ii) tetragonal crystal
184: symmetry when $\psi = (\psi_1,\psi_2,0)$, (iii) {\em XY} magnetic
185: order $(M_1,M_2,0)$, and (iv) the general case of cubic crystal
186: symmetry and isotropic magnetic order ($m=3$) when all components
187: of the three dimensional vectors ${\mbox{\boldmath$M$}}$ and
188: $\psi$ may have nonzero equilibrium and fluctuation components.
189: The latter case is of major interest to real systems where
190: fluctuations of all components of the fields are possible despite
191: the presence of spatial crystal and magnetic anisotropy that
192: nullifies some of the equilibrium field components. In one-loop
193: approximation, the RG analysis reveals different pictures for
194: anisotropic (i)-(iii) and isotropic (iv) systems. As usual, a
195: Gaussian (``trivial'') FP ($g^{\ast} =0$) exists for all $d>0$
196: and, as usual~\cite{Uzunov:1993}, this FP is stable for $d>6$
197: where the fluctuations are irrelevant. In the reminder of this
198: letter the attention will be focussed on spatial dimensions $d <
199: 6$, where the critical behavior is usually governed by nontrivial
200: FPs ($g^{\ast} \neq 0$). In the cases (i)-(iii) only negative
201: (``unphysical''~\cite{Lawrie:1987}) FP values of $g^2$ have been
202: obtained for $d<6$. For example, in the case (i) the RG relation
203: for $g$ takes the form
204: \begin{equation}
205: \label{eq2} g^{\prime}= b^{3-d/2-\eta}g\left(1 +
206: g^2K_d{\mbox{ln}}b\right),
207: \end{equation}
208: where $g^{\prime}$ is the renormalized value of $g$, $\eta =
209: (K_{d-1}/8)g^2$ is the anomalous dimension (Fisher's
210: exponent)~\cite{Uzunov:1993} of the field $M_3$; $K_d =
211: 2^{1-d}\pi^{-d/2}/\Gamma(d/2)$. Using eq.~(2) one obtains the FP
212: coordinate $(g^2)^{\ast} = -96\pi^3\epsilon$. For $d < 6$ this FP
213: is unphysical and does not describe any critical behavior. For $d
214: > 6$ the same FP is physical but unstable towards the parameter
215: $g$ as one may see from the positive value $y_g = -11\epsilon /2
216: >0$ of the respective stability exponent
217: $y_g$ defined by $\delta g^{\prime} = b^{y_g}\delta g$. Therefore,
218: a change of the order of the phase transition from second order in
219: mean-field approximation to a fluctuation-driven first order
220: transition when the fluctuation $g$--interaction is taken into
221: account, takes place. This conclusion is supported by general
222: concepts of RG theory~\cite{Uzunov:1993} and by the particular
223: property of these systems to exhibit first order phase
224: transitions~\cite{Shopova:2003} in mean field approximation for
225: broad variations of $T$ and $P$.
226:
227: In the case (iv) of isotropic systems the RG equation for $g$ is
228: degenerate and the $\epsilon$-expansion breaks down. A similar
229: situation is known from the theory of disordered
230: systems~\cite{Lawrie:1987} but here the physical mechanism and
231: details of description are different. Namely for this degeneration
232: one should consider the RG equations up to the two-loop order. The
233: derivation of the two-loop terms in the RG equations is quite
234: nontrivial because of the special symmetry properties of the
235: interaction $g$-term in eq.~(1). For example, some diagrams with
236: opposite arrows of internal lines, as the couple shown in
237: fig.~(2), have opposite signs and compensate each other. The terms
238: bringing contributions to the $g$--vertex are shown
239: diagrammatically in fig.~3. The RG analysis is carried out by a
240: completely new $\epsilon^{1/4}$-expansion for the FP values and
241: $\epsilon^{1/2}$-expansion for the critical exponents; again
242: $\epsilon = (6-d)$. RG equations are quite lengthy and here only
243: the equation for $g$ is discussed. It has the form
244: \begin{equation}
245: \label{eq3} g^{\prime}=
246: b^{(\epsilon-2\eta_{\psi}-\eta_M)/2}g\left[1 + Ag^2 +
247: 3(2B+C)g^4\right],
248: \end{equation}
249: where
250: \begin{equation}
251: \label{eq4} A=\frac{K_d}{2}\left[2{\mbox{ln}}b + \epsilon
252: ({\mbox{ln}}b)^2 + (1-b^2)(2r+t)\right],
253: \end{equation}
254: \begin{equation}
255: \label{eq5} B= \frac{K_{d-1}K_d}{192}\left[9(b^2-1) -
256: 11{\mbox{ln}}b - 6\left({\mbox{ln}}b\right)^2\right],\;\;\;\; C=
257: \frac{3K_{d-1}K_d}{64}\left[{\mbox{ln}}b +
258: 2\left({\mbox{ln}}b\right)^2\right],
259: \end{equation}
260: $\eta_M$ and $\eta_{\psi}$ are the anomalous dimensions of the
261: fields ${\mbox{\boldmath$M$}}$ and $\psi$, respectively. The
262: one-loop approximation gives correct results to order
263: $\epsilon^{1/2}$ and the two-loop approximation brings such
264: results up to order $\epsilon$. In eq.~(4), $r$ and $t$ are small
265: expansion quantities with equal FP values $t^{\ast} = r^{\ast} =
266: K_d g^2$. Using the condition for invariance of the two
267: $k^2$-terms in eq.~(1) one obtains $\eta_M =\eta_{\psi} \equiv
268: \eta$, where
269: \begin{equation}
270: \label{eq6} \eta =\frac{K_{d-1}}{8}g^2\left(1-
271: \frac{13}{96}K_{d-1}g^2\right).
272: \end{equation}
273: Eq.~(3) yields a new FP
274: \begin{equation} \label{eq7}
275: g^{\ast}=8\left(3\pi^3\right)^{1/2}\left(2\epsilon/13\right)^{1/4},
276: \end{equation}
277: which corresponds to the critical exponent $\eta =
278: 2(2\epsilon/13)^{1/2} - 2\epsilon/3$ (for $d=3$, $\eta \approx
279: -0.64 $).
280:
281:
282: The eigenvalue problem for the RG stability matrix
283: $\hat{{\cal{M}}} = \left[(\partial\mu_i/\partial
284: \mu_j);(\mu_1,\mu_2,\mu_3) = (r,t,g)\right]$ can be solved by the
285: expansion of the matrix elements up to order $\epsilon^{3/2}$.
286: When the eigenvalues $\lambda_j= A_j(b)b^{y_j}$ of
287: $\hat{{\cal{M}}}$ are calculated dangerous large terms of type
288: $b^2$ and $b^2(\mbox{ln}b)$, ($b \gg 1$)~\cite{Aharony:1974} in
289: the off-diagonal elements of the matrix $\hat{{\cal{M}}}$ ensure
290: the compensation of redundant large terms of the same type in the
291: diagonal elements $\hat{{\cal{M}}}_{ii}$. This compensation is
292: crucial for the validity of scaling for this type of critical
293: behavior. Such a problem does not appear in standard cases of RG
294: analysis~\cite{Uzunov:1993, Aharony:1974}. As in the usual
295: $\phi^4$--theory~\cite{Aharony:1974} the amplitudes $A_j$ depend
296: on the scaling factor $b$: $A_1=A_2=1+(27/13)b^2\epsilon$,
297: $A_3=1-(81/52)\epsilon(\mbox{ln}b)^2$. The critical exponents
298: $y_t=y_1$, $y_r=y_2$ and $y_g$ = $y_3$ are $b$--invariant:
299: \begin{equation}
300: \label{eq8} y_r = 2 + 10\sqrt{\frac{2\epsilon}{13}} +
301: \frac{197}{39}\epsilon,
302: \end{equation}
303: $y_t = y_r-18(2\epsilon/13)^{1/2}$, and $y_g = -\epsilon > 0$ for
304: $d < 6$. The correlation length critical exponents $\nu_{\psi} =
305: 1/y_r$ and $\nu_M = 1/y_t$ corresponding to the fields $\psi$ and
306: ${\mbox{\boldmath$M$}}$ are
307: \begin{equation}
308: \label{eq9} \nu_{\psi}= \frac{1}{2} -
309: \frac{5}{2}\sqrt{\frac{2\epsilon}{13}} +
310: \frac{103}{156}\epsilon,\;\;\;\;\; \nu_M = \frac{1}{2} +
311: 2\sqrt{\frac{2\epsilon}{13}} - \frac{5\epsilon}{156}.
312: \end{equation}
313: These exponents describe a quite particular multi-critical
314: behavior which differs from the numerous examples known so far.
315: For $d = 3$, $\nu_{\psi}$ = 0.78 which is somewhat above the usual
316: value $\nu \sim 0.6 \div 0.7$ near a standard phase transition of
317: second order~\cite{Uzunov:1993},
318: but $\nu_M = 1.76$ at the same dimension ($d=3$) is unusually large.
319: The fact that the Fisher's
320: exponent~\cite{Uzunov:1993} $\eta$ is negative for $d=3$ does not
321: create troubles because such cases are known in complex systems,
322: for example, in conventional superconductors~\cite{Halperin:1974}.
323: The present $\epsilon$-expansion is valid under the conditions
324: $\epsilon^{1/2}b < 1$, $\epsilon^{1/2}(\mbox{ln}b) \ll 1$ provided
325: $b > 1$. These conditions are stronger than those corresponding to
326: the usual $\phi^4$-theory~\cite{Uzunov:1993,Aharony:1974}. This
327: means that the present expansion in non-integer powers of
328: $\epsilon$ has a more restricted domain of validity than the
329: standard $\epsilon$-expansion. Using the known
330: relation~\cite{Uzunov:1993} $\gamma = (2-\eta)\nu$, the
331: susceptibility exponents for $d=3$ take the values $\gamma_{\psi}
332: = 2.06$ and $\gamma_M = 4.65$. These values exceed even those
333: corresponding to the Hartree approximation~\cite{Uzunov:1993}
334: ($\gamma = 2\nu = 2$ for $d=3$) and can be easily distinguished in
335: experiments.
336:
337:
338: The critical behavior discussed so far may occur in a close
339: vicinity of finite temperature multi-critical points ($T_c=T_f>0$)
340: in systems possessing the symmetry of the model (1). In certain
341: systems, as shown in Fig.~1, this multi-critical points may occur
342: at $T=0$. In the quantum limit ($T\rightarrow 0$), or, more
343: generally, in the low-temperature limit [$T \ll \mu; \mu\equiv
344: (t,r);k_B=1$] the thermal wavelengths of the fields
345: $\mbox{\boldmath$M$}$ and $\psi$ exceed the inter-particle
346: interaction radius and the quantum correlations fluctuations
347: become essential for the critical
348: behavior~\cite{Shopova:2003,Hertz:1976}. The quantum effects can
349: be considered by RG analysis of a comprehensively generalized
350: version of the model~(1), namely, the action ${\cal{S}}$ of the
351: referent quantum system. The generalized action is constructed
352: with the help of the substitution $(-{\cal{H}}/T) \rightarrow
353: S[{\mbox{\boldmath$M$}}(q),\psi(q)]$. Now the description is given
354: in terms of the (Bose) quantum fields $\mbox{\boldmath$M$}(q)$ and
355: $\psi(q)$ which depend on the $(d+1)$-dimensional vector $q =
356: (\omega_l, {\mbox{\boldmath$k$}})$; $\omega_l = 2\pi lT$ is the
357: Matsubara frequency ($\hbar=1;l = 0, \pm1,\dots$). The
358: ${\mbox{\boldmath$k$}}$-sums in eq.~(1) should be substituted by
359: respective $q$-sums and the inverse bare correlation functions ($r
360: + k^2$) and ($t + k^2$) in eq.~(1) contain additional
361: $\omega_l-$dependent terms, for example\cite{Hertz:1976,
362: Shopova:2003}
363: \begin{equation} \label{eq10}
364: \langle|\psi_{\alpha}(q)|^2\rangle^{-1} = |\omega_l|+ k^2 + r.
365: \end{equation}
366: The bare correlation function $\langle|M_j(q)|\rangle^2$ contains
367: a term of type $|\omega_l|/k^{\theta}$, where $\theta = 1$ and
368: $\theta =2$ for clean and dirty itinerant ferromagnets,
369: respectively~\cite{Hertz:1976}. The quantum dynamics of the field
370: $\psi$ is described by the bare value $z=2$ of the dynamical
371: critical exponent $z=z_{\psi}$ whereas the quantum dynamics of the
372: magnetization corresponds to $z_M = 3$ (for $\theta =1$), or, to
373: $z_M = 4$ (for $\theta = 2$). This means that the
374: classical-to-quantum dimensional crossover at $T\rightarrow 0$ is
375: given by $d \rightarrow (d + 2)$ and, hence, the system exhibits a
376: simple mean field behavior for $d \geq 4$. Just below the upper
377: quantum critical dimension $d_U^{(0)} =4$ the relevant quantum
378: effects at $T=0$ are represented by the field $\psi$ whereas the
379: quantum $(\omega_l$--) fluctuations of the magnetization are
380: relevant for $d < 3$ (clean systems), or, for even for $d < 2$
381: (dirty limit)~\cite{Hertz:1976}. This picture is confirmed by the
382: analysis of singularities of the relevant perturbation integrals.
383: Therefore, the quantum fluctuations of the field $\psi$ have a
384: dominating role below spatial dimensions $d <4$.
385:
386: Taking into account the quantum fluctuations of the field $\psi$
387: and completely neglecting the $\omega_l$--dependence of the
388: magnetization ${\mbox{\boldmath$M$}}$, $\epsilon_0 =
389: (4-d)$--analysis of the generalized action ${\cal{S}}$ has been
390: performed within the one-loop approximation (order
391: $\epsilon_0^1$). In the classical limit ($r/T \ll 1$) one
392: re-derives the results already reported above together with an
393: essentially new result, namely, the value of the dynamical
394: exponent $z_{\psi}= 2 - (2\epsilon/13)^{1/2}$ which describes the
395: quantum dynamics of the field $\psi$. In the quantum limit ($r/T
396: \gg 1$, $T\rightarrow 0$) the static phase transition properties
397: are affected by the quantum fluctuations, in particular, in
398: isotropic systems ($n/2=m=3$). For this case, the one-loop RG
399: equations corresponding to $T = 0$ are not degenerate and give
400: definite results. The RG equation for $g$,
401: \begin{equation} \label{eq11} g^{\prime}=
402: b^{\epsilon_0/2}g\left(1 + \frac{g^2}{24\pi^3}\mbox{ln}b\right),
403: \end{equation}
404: yields two FPs: ({\em a}) a Gaussian FP ($g^{\ast} =0$), which is
405: unstable for $d<4$, and ({\em b}) a FP $(g^2)^{\ast} =
406: -12\pi^3\epsilon_0$ which is unphysical [$(g^2)^{\ast}<0$] for
407: $d<4$ and unstable for $d \geq 4$. Thus the new stable critical
408: behavior corresponding to $T>0$ and $d<6$ disappears in the
409: quantum limit $T\rightarrow 0$. At the absolute zero and any
410: dimension $d > 0$ the $P-$driven phase transition (Fig.~1) is of
411: first order. This can be explained as a mere result of the limit
412: $T \rightarrow 0$. The only role of the quantum effects is the
413: creation of the new unphysical FP ({\em b}). In fact, the referent
414: classical system described by ${\cal{H}}$ from eq.~(1) also looses
415: its stable FP (7) in the zero-temperature ({\em classical}) limit
416: $T\rightarrow 0$ but does not generate any new FP because of the
417: lack of $g^3$--term in the equation for $g^{\prime}$; see
418: eq.~(11). At $T=0$ the classical system has a purely mean field
419: behavior~\cite{Shopova:2003} which is characterized by a Gaussian
420: FP ($g^{\ast} = 0$) and is unstable towards $T$--perturbations for
421: $0<d<6$. This is a usual classical zero temperature behavior where
422: the quantum correlations are ignored. For the standard $\phi^4-$
423: theory this picture holds for $d<4$. One may suppose that the
424: quantum fluctuations of the field $\psi$ are not enough to ensure
425: a stable quantum multi-critical behavior at $T_c=T_F=0$ and that
426: the lack of such behavior is in result of neglecting the quantum
427: fluctuations of $\mbox{\boldmath$M$}$. One may try to take into
428: account these quantum fluctuations by the special approaches from
429: the theory of disordered systems, where additional expansion
430: parameters are used to ensure the marginality of the fluctuating
431: modes at the same borderline dimension $d_U$ (see, e.g.,
432: Ref.~\cite{Shopova:2003}). It may be conjectured that the
433: techniques known from the theory of disordered systems with
434: extended impurities cannot be straightforwardly applied to the
435: present problem and, perhaps, a completely new supposition should
436: be introduced.
437:
438: In conclusion, the present results may be of use in
439: interpretations of recent experiments~\cite{Pfleiderer:2002} in
440: UGe$_2$, where the magnetic order is uniaxial (Ising symmetry) and
441: the experimental data, in accord with the present consideration,
442: indicate that the C-P phase transition is of first order. Systems
443: with isotropic magnetic order are needed for an experimental test
444: of the new multi-critical behavior.
445:
446:
447:
448:
449:
450:
451:
452:
453:
454:
455:
456: {\bf Acknowledgments.} The author thanks the hospitality of MPI-PKS
457: (Dresden) and ICTP (Triest) where a part of this research has been
458: performed. Financial support by grants No. P1507 (NFSR, Sofia) and
459: No. G5RT-CT-2002-05077 (EC, SCENET-2, Parma) is also acknowledged.
460:
461: \begin{thebibliography}{ll}
462: \bibitem{Saxena:2000} Saxena S. S., Agarwal P., Ahilan K., Grosche F. M.,
463: Haselwimmer R. K. W., Steiner M. J., Pugh E., Walker I. R., Julian
464: S. R., Monthoux P., Lonzarich G. G., Huxley A., Sheikin I.,
465: Braithwaite D. and Flouquet J., Nature {\bf 406} {2000}
466: 587.
467: \bibitem{Huxley:2001}
468: Huxley A., Sheikin I., Ressouche E., Kernavanois N.,
469: Braithwaite D., Calemczuk R., and Flouquet J., Phys. Rev.
470: B {\bf 63} (2001) 144519.
471: \bibitem {ShopovaPL:2003} Shopova D. V. and D. I. Uzunov, Phys. Lett. A {\bf
472: 313} (2003) 139.
473: \bibitem {Shopova:2005} Shopova D. V. and D. I. Uzunov, Phys. Rev. B {\bf 72}
474: (2005) 024531.
475: \bibitem {Shopova:2006}
476: Shopova D. V. and Uzunov D. V., in: Progress in Ferromagnetism Research,
477: ed. bg V. N. Murrey
478: (Nova Science Publishers, New York, 2006) p.223.
479: \bibitem{Aoki:2001} Aoki D., Huxley A., Ressouche E., Braithwaite D.,
480: Flouquet J.,
481: Brison J-P., E. Lhotel and Paulsen C., Nature {\bf 413} (2001) 613.
482: \bibitem {Uzunov:1993}
483: Uzunov D. I., Theory of Critical Phenomena,
484: (World Scientific, Singapore, 1993).
485: \bibitem{Hertz:1976} Hertz J. A., Phys. Rev. B {\bf 14} (1976) 1165.
486: \bibitem {Shopova:2003} Shopova D. V. and Uzunov D. I., Phys. Rep. C {\bf 379}
487: (2003) 1.
488: \bibitem {Sigrist:1991} M. Sigrist M. and Ueda K., Rev. Mod.
489: Phys. {\bf 63} 1991) 239.
490: \bibitem{Machida:2001} Machida K. and Ohmi T., Phys. Rev. Lett. {\bf 86}
491: (2001) 850.
492: \bibitem{Walker:2002} M. B. Walker M. B.
493: and K. V. Samokhin K. V., Phys. Rev.
494: Lett. {\bf 88} 2002) 207001.
495: \bibitem{Pfleiderer:2002} Pfleiderer C. and Huxley A. D., Phys. Rev. Lett.{\bf
496: 89} (2002) 147005.
497: \bibitem{Aharony:1974} Bruce A. D., Droz M. and Aharony A., J. Phys. C: Solid
498: State Physics {\bf 7} (1974) 3673.
499: \bibitem {Lawrie:1987} Lawrie I. D., Millev Y. T. and D. I. Uzunov,
500: J. Phys. A: Math. Gen. {\bf 20} (1987) 1599.
501: \bibitem{Halperin:1974} Halperin B. I., Lubensky T. C. and
502: Ma S. K., Phys. Rev. Lett. {\bf 32} (1974) 292.
503: \end{thebibliography}
504:
505:
506: \begin{figure}
507: \epsfig{file=UzunovFF1}
508: \caption{($P,T$) diagram with a
509: zero-temperature multicritical point $(P_0,0)$. Para- (P),
510: ferromagnetic (F), and coexistence (C) phases, separated by the
511: lines $T_f(P)$ and $T_c(P)$ of P-F and F-C phase transitions,
512: respectively.}
513: \end{figure}
514:
515:
516:
517:
518:
519:
520:
521:
522:
523:
524: \begin{figure}
525: \epsfig{file=UzunovFF2}
526: \caption{A sum of $g^5$--diagrams equal to
527: zero. The thick and thin lines correspond to correlation functions
528: $\langle |\psi_{\alpha}|^2\rangle$ and $\langle |M_j|^2\rangle$,
529: respectively; vertices ($\bullet$) represent $g$--term in
530: eq.~(1).}
531: \end{figure}
532:
533: \begin{figure}
534: \epsfig{file=UzunovFF3}
535: \caption{Diagrams for $g^{\prime}$ of third
536: and fifth order in $g$. The arrows of the thick lines have been
537: omitted.}
538: \end{figure}
539:
540:
541:
542:
543:
544:
545:
546: \end{document}
547:
548:
549:
550:
551:
552:
553:
554:
555:
556:
557:
558:
559:
560:
561:
562:
563:
564:
565:
566:
567:
568:
569:
570:
571:
572:
573:
574:
575:
576:
577:
578: