O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) scalar models: including 𝒪 ( ∂ 2 ) \mathcal{O}(\partial^{2}) corrections in the Functional Renormalization Group analysis.

Carlos A. Sánchez-Villalobos, Bertrand Delamotte, Nicolás Wschebor · 2024

Read O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) scalar models: including 𝒪 ( ∂ 2 ) \mathcal{O}(\partial^{2}) corrections in the Functional Renormalization Group analysis. by Carlos A. Sánchez-Villalobos, Bertrand Delamotte, Nicolás Wschebor (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2411-02616.txt.

Description

Abstract The study of phase transitions in frustrated magnetic systems with O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) symmetry has been the subject of controversy for more than twenty years, with theoretical, numerical and experimental results in disagreement.

Catalog record

CREATORCarlos A. Sánchez-Villalobos, Bertrand Delamotte, Nicolás Wschebor
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE87 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2411-02616

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) scalar models: including 𝒪 ( ∂ 2 ) \mathcal{O}(\partial^{2}) corrections in the Functional Renormalization Group analysis. Carlos A. Sánchez-Villalobos, Bertrand Delamotte, Nicolás Wschebor Abstract The study of phase transitions in frustrated magnetic systems with O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) symmetry has been the subject of controversy for more than twenty years, with theoretical, numerical and experimental results in disagreement. Even theoretical studies lead to different results, with some predicting a first-order phase transition while others find it to be second-order. Recently, a series of results from both numerical simulations and theoretical analyses, in particular those based on the Conformal Bootstrap, have rekindled interest in this controversy, especially as they are still not in agreement with each other. Studies based on the functional renormalization group have played a major role in this controversy in the past, and we revisit these studies, taking them a step further by adding non-trivial second order derivative terms to the derivative expansion of the effective action. We confirm the first-order nature of the phase transition for physical values of N N , i.e. for N = 2 N=2 and N = 3 N=3 in agreement with the latest results obtained with the Conformal Bootstrap. We also study an other phase of the O ⁡ ( N ) × O ⁡ ( 2 ) O(N)\times O(2) models, called the sinusoidal phase, qualitatively confirming earlier perturbative results. I Introduction The nature of the phase transition in scalar models with O(N)\times O(2) symmetry has been the subject of a long-standing controversy for more than two decades [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ] . These models are expected to describe two important families of frustrated magnets: the stacked triangular antiferromagnets (STAs) and the helimagnets. The above controversy is therefore not just theoretical but has experimental implications [ 9 , 5 ] . In a Ginzburg-Landau formulation, such models can be described by means of a pair of N- component fields, \vec{\varphi}_{1} and \vec{\varphi}_{2} and a Hamiltonian of the form: \displaystyle H \displaystyle=\int d^{d}x\Big\{\frac{1}{2}(\partial_{\mu}\vec{\varphi}_{1})^{2}+\frac{1}{2}(\partial_{\mu}\vec{\varphi}_{2})^{2}+\frac{r}{2}\big(\vec{\varphi}_{1}^{2}+\vec{\varphi}_{2}^{2}\big) \displaystyle+\frac{u}{4!}\big(\vec{\varphi}_{1}^{2}+\vec{\varphi}_{2}^{2}\big)^{2}+\frac{v}{4!}\Big(\frac{1}{2}(\vec{\varphi}_{1}^{2}-\vec{\varphi}_{2}^{2})^{2}+2(\vec{\varphi}_{1}\cdot\vec{\varphi}_{2})^{2}\Big)\Big\} (1) For v>0 , the ground state corresponds to orthogonal \vec{\varphi}_{1} and \vec{\varphi}_{2} of the same modulus and is called orthogonal. As the temperature is varied, this system undergoes a phase transition to a paramagnetic phase and the controversy concerns the order of the transition for N=2 and 3 , which are the physically interesting cases. To present the problem, let us first recall what happens near the upper critical dimension d_{c}=4 of this system around which it is possible to perform a controlled perturbative expansion in \epsilon=4-d [ 10 , 11 , 9 ] . Even for small \epsilon , this model has a rich structure of fixed points of the renormalization group (RG) which varies according to the value of N , see Fig. 1 : i) For N>N_{c}^{+}(d) , with \displaystyle N_{c}^{+}(d) \displaystyle=4(3+\sqrt{6})-4\Big(3+\frac{7}{\sqrt{6}}\Big)\epsilon+O(\epsilon^{2}) \displaystyle\simeq 21.8-23.4\epsilon+O(\epsilon^{2}), (2) the model has four perturbative fixed points. The first two are the Gaussian fixed point ( G ) and the standard Heisenberg ( H ) or Wilson-Fisher O(2N)- invariant fixed point, corresponding both to multicritical transitions. In addition to these, it has two “anisotropic” fixed points usually referred to as “chiral” ( C_{+} ) and “anti-chiral” ( C_{-} ). The C_{+} fixed point is once unstable and is associated with a phase transition between the ordered orthogonal phase and a paramagnetic phase. The C_{-} fixed point is twice unstable and is associated with the tricritical transition between the same two phases. ii) For N_{c}^{+}(d)>N>N_{c}^{-}(d) , with \displaystyle N_{c}^{-}(d) \displaystyle=4(3-\sqrt{6})-4\Big(3-\frac{7}{\sqrt{6}}\Big)\epsilon+O(\epsilon^{2}) \displaystyle\simeq 2.2-0.57\epsilon+O(\epsilon^{2}), (3) only the Gaussian and Heisenberg fixed points exist. They are both multicritical and the system undergoes generically a first-order phase transition. By decreasing N at fixed d , C_{+} and C_{-} approach each other, “annihilate” for N=N_{c}^{+}(d) and become complex for N_{c}^{+}(d)>N>N_{c}^{-}(d) . iii) For N_{c}^{-}(d)>N>N_{c}^{H}(d) , with N_{c}^{H}(d)=2-\epsilon+O(\epsilon^{2}), (4) the model has again four fixed points but, unlike in case (i), in addition to the Gaussian and Heisenberg fixed points, it has two fixed points called “sinusoidal” ( S_{+} ) and “anti-sinusoidal” ( S_{-} ). These correspond to transitions between a paramagnetic phase and an ordered phase where \vec{\varphi}_{1} and \vec{\varphi}_{2} are parallel in the ground state. The “sinusoidal” fixed point corresponds to a second-order phase transition and the others to multicritical transitions. As before, the line N_{c}^{-}(d) corresponds to the location where S_{+} and S_{-} “annihilate”. iv) For N<N_{c}^{H}(d) , the Heisenberg fixed point exchanges stability with the sinusoidal one that becomes a chiral fixed point C and the second-order transition has an emergent O(2N)- symmetry at large distance. It is important to note that only cases (i) and (ii) are relevant for the phase transition in the STA, but the other cases are of theoretical interest for a better understanding of the model, which is why this article deals with all of them. Figure 1: Flows and perturbative fixed points in d=3.9 in the one-loop approximation for different values of N . Top left: N=32>N_{c}^{+} ; Top right: N_{c}^{-}<N=3<N_{c}^{+} ; Bottom left: N_{c}^{H}<N=2.18<N_{c}^{-} ; Bottom right: N=1<N_{c}^{H} . Figure inspired from [ 12 ] . The represented couplings are the dimensionless \hat{u}=w_{d}\tilde{u} and \hat{v}=w_{d}\tilde{v} where w_{d}^{-1}=2^{d-1}\pi^{d/2}\Gamma(d/2) (see Eq. ( 17 )). The Gaussian fixed point is called G and the Heisenberg (or O(2N) invariant) fixed point H. For N=1 , the chiral fixed point is denoted by C and the sinusoidal by S . The above results become controversial when the curve N_{c}^{+}(d) is prolonged down to d=3 [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ] . First of all, among the methods based on resummed perturbation theory, the \epsilon=4-d expansion at 6-loops presents a first order transition for N=2, and 3 [ 8 ] but the perturbative expansion (in the constants u and v ) in fixed dimension presents a second order transition for those values of N . This expansion is performed at 5-loops in the \overline{MS}- scheme [ 6 ] and at 6-loops in the mass zero-momentum scheme [ 2 , 3 , 4 ] . In fact, as shown in Fig. 2 , the curves corresponding to both these perturbative analyses start almost identically near d=4 but the curve of the \epsilon- expansion is monotonous while the curve given by the perturbative expansion at fixed dimension has a “S”-shape around d=3 that puts the values N=2 and N=3 on the second order side of the N_{c}^{+}(d) curve in d=3 . It should be noted, as discussed in Ref. [ 7 ] , that the shape of the “S” curve is quite unstable and depends substantially on the details of the resummation procedure employed. In Fig. 2 , an example is shown in which the return point occurs at d\approx 3.2 but in some other resummation procedures (see, for example, [ 6 ] ) the return point occurs slightly below d=3 and the upper branch of the “S” curve looks very similar to the \epsilon- expansion even in d=3 . However, as N is further decreased in d=3

Read the complete text