Interpretation of experimental data near lambda-transition point in liquid helium

TL;DR

This paper analyzes experimental specific heat data near the lambda-transition in liquid helium, comparing interpretations from perturbative RG and a new theory, and suggests potential changes in behavior very close to the transition.

Contribution

The paper demonstrates that different theoretical approaches can fit the data similarly, and proposes a new set of critical exponents that better fit certain experimental results.

Findings

Both RG and new theory fit the data well over a range of temperatures.

The effective exponent alpha_eff(t) indicates possible behavior change very close to the transition.

New exponents nu=9/13 and Delta=5/13 fit superfluid fraction data better than RG exponents.

Abstract

The recently published experimental data for specific heat C_p of liquid helium in zero gravity conditions very close to the lambda-transition have been discussed. We have shown that these data allow different interpretations. They can be well interpreted within the perturbative RG approach and within our recently developed theory, as well. Allowing the logarithmic correction, the corresponding fits lie almost on top of each other over the whole range of the reduced temperatures t (for bin averaged data) 6.3 x 10^{-10} < t < 8.8 x 10^{-3}. However, the plot of the effective exponent alpha_eff(t) suggests that the behaviour of C_p, probably, changes very close to the lambda-transition temperature. To clarify this question, we need more accurate data for t<10^{-7}. In addition, we show that the experimental data for superfluid fraction of liquid helium close to the critical point within 3 x 10^{-7} < t < 10^{-4} can be better fit by our exponents nu=9/13, Delta=5/13 than by the RG exponents (nu approximately 0.6705 and Delta about 0.5). The latter ones are preferable to fit the whole measured range 3 x 10^{-7} < t < 10^{-2} where, however, remarkable systematic deviations appear. Our estimated value 0.694 +/- 0.017 of the asymptotic exponent nu well agrees with the theoretical prediction nu=9/13.