Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Fractal dimension of critical curves in the O(n)-symmetric ϕ4 model and crossover exponent at 6-loop order: Loop-erased random walks, self-avoiding walks, Ising, XY, and Heisenberg models

Mikhail Kompaniets and Kay Jörg Wiese
Phys. Rev. E 101, 012104 – Published 3 January 2020

Abstract

We calculate the fractal dimension df of critical curves in the O(n)-symmetric (ϕ2)2 theory in d=4ε dimensions at 6-loop order. This gives the fractal dimension of loop-erased random walks at n=2, self-avoiding walks (n=0), Ising lines (n=1), and XY lines (n=2), in agreement with numerical simulations. It can be compared to the fractal dimension dftot of all lines, i.e., backbone plus the surrounding loops, identical to dftot=1/ν. The combination ϕc=df/dftot=νdf is the crossover exponent, describing a system with mass anisotropy. Introducing a self-consistent resummation procedure and combining it with analytic results in d=2 allows us to give improved estimates in d=3 for all relevant exponents at 6-loop order.

  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
11 More
  • Received 28 September 2019

DOI:https://doi.org/10.1103/PhysRevE.101.012104

©2020 American Physical Society

Physics Subject Headings (PhySH)

Statistical Physics & Thermodynamics

Authors & Affiliations

Mikhail Kompaniets1 and Kay Jörg Wiese2

  • 1Saint Petersburg State University, Saint Petersburg 199034, Russia
  • 2CNRS-Laboratoire de Physique de l'Ecole Normale Supérieure, ENS, Université PSL, Sorbonne Université, Université Paris-Diderot, Sorbonne Paris Cité, 75005 Paris, France

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand
Issue

Vol. 101, Iss. 1 — January 2020

Reuse & Permissions
Access Options
Author publication services for translation and copyediting assistance advertisement

Authorization Required


×

Images

  • Figure 1
    Figure 1

    Example of a loop-erased random walk on the hexagonal lattice with 3000 steps, starting at the black point to the right and arriving at the green point to the left.

    Reuse & Permissions
  • Figure 2
    Figure 2

    Fractal dimensions of lines in dimension d=3. Two expansions are shown: Direct (in red) and expansion for 1/df (blue). The table compares our values to results from the literature.

    Reuse & Permissions
  • Figure 3
    Figure 3

    The fractal dimension of lines in dimension d=2, as extracted from field theory (colored), and compared to exact results (black dashed line). The different curves are from resummation of df (blue), df1 (red), df2 (cyan), and df2 (green). The table compares the result of our different schemes, with the direct expansion of df used for SC. Note that the error given is the error of the expansion in one scheme. Comparing different expansion schemes, we estimate the overall error to be of order 0.05.

    Reuse & Permissions
  • Figure 4
    Figure 4

    The ratios rn as given in Eq. (45) for α=0 and the fit to Eq. (46).

    Reuse & Permissions
  • Figure 5
    Figure 5

    Resummation of df for LERWs (n=2) in d=3 as a function of the series-order n, setting α=0. One sees that the resummed series converges, for all assumed values of the branch cut, with orange zbc=0.3/a to green with zbc=1/a and ending with cyan zbc=1.1/a, which clearly sits inside the supposed branch cut, which oscillates, and for which only the real part is shown.

    Reuse & Permissions
  • Figure 6
    Figure 6

    Minus the exponential decay constant c from Eq. (46).

    Reuse & Permissions
  • Figure 7
    Figure 7

    (a) In blue the fractal dimension df of LERWs as a function of α. The latter yields bounds for df, i.e., df[1.62378,1.6254], and as a best estimate the mean of the obtained values, df1.62426 (blue dashed line). The numerical result is df=1.62400±0.00005 (orange with error bars in dashes) [48]. (b) Same for d=2. We find df[1.238,1.259], with a mean estimate df=1.244, to be compared to the exact result df=5/4. Using only the 5-loop series gives df(d=3)1.621 and df(d=2)=1.11.

    Reuse & Permissions
  • Figure 8
    Figure 8

    The crossover phase diagram as given in Ref. [11], with λ=m22m12. The thick black line is a line of first-order phase transitions.

    Reuse & Permissions
  • Figure 9
    Figure 9

    Slope of the crossover exponent at n=0 for dimensions 0d4. The black cross is the analytic result from Eq. (102) in d=2.

    Reuse & Permissions
  • Figure 10
    Figure 10

    df as a function of d for LERW (n=2). The red dashed line is the bound df56d [67] (bound continuation to all dimensions guessed). The gray dashed lines are the bounds 1dfdRW=2.

    Reuse & Permissions
  • Figure 11
    Figure 11

    The exponent ν for d=2 (a) and its inverse (b). The different colors come from resummations of ν (blue), 1/ν (red), 1/ν2 (green), 1/ν3 (cyan), and α=2νd (dark green). The dashed black line is from CFT as given by Eq. (100). The shaded errors are (minimal) errors estimated from the uncertainty in the extrapolation, see Sec. 3.

    Reuse & Permissions
  • Figure 12
    Figure 12

    The exponent η in d=2. The blue curve is the direct expansion and the green one a resummation of η, which, as η starts at order ε2, has a regular series expansion in ε. The black solid line is η=4h1/2,0 as given by Eq. (101).

    Reuse & Permissions
  • Figure 13
    Figure 13

    The exponent ω in d=2. Dots represent values reported in the literature, mostly based on CFT. The value ω=7/4 for n=1 is consistent with the O(1) term in Ref. [78], while the reanalysis of Ref. [79] concludes on ω=2. In Ref. [79] it is also argued that ω=2 for n>2. The black dashed line is the guess (103) resulting from the operator generating an intersection between two lines.

    Reuse & Permissions
  • Figure 14
    Figure 14

    The exponent ϕc in d=2. The dashed black line is the analytic result from Eq. (102). The colored lines are resummations of ϕc (blue), 1/ϕc (red), 1/ϕc2 (green), 1/ϕc3 (cyan), 1/ϕc13/4 (magenta), and 1/ϕc3 (gray). Resumming 1/ϕc13/4 considerably improves the precision.

    Reuse & Permissions
  • Figure 15
    Figure 15

    The exponent η in d=3. The SC resummation scheme (in blue) seems to be systematically smaller than the values of KP17 (in red). SC resummation of η (in cyan) works slightly better. Black crosses represent the best values from MC and conformal bootstrap, as given in Ref. [10].

    Reuse & Permissions
  • Figure 16
    Figure 16

    The exponent ν in d=3, obtained from a resummation of 1/ν3. In blue the results from SC and in red using KP17. Black crosses are from MC and conformal bootstrap, as given in Ref. [10].

    Reuse & Permissions
  • Figure 17
    Figure 17

    The exponent ω in d=3 via SC (blue, with shaded error bars) and KP17 (in red). Crosses represent the best values from MC, as given in Ref. [10].

    Reuse & Permissions
  • Figure 18
    Figure 18

    The exponent ϕc in d=3. Crosses are from MC and experiments [54, 58]. The value for n=2 is taken as df/2 with df the fractal dimension of LERWs [48].

    Reuse & Permissions
×

Sign up to receive regular email alerts from Physical Review E

Log In

Cancel
×

Search


Article Lookup

Paste a citation or DOI

Enter a citation
×