▎ 摘　　要

Many-body electron-electron interaction effects are theoretically considered in monolayer graphene from a continuum effective field-theoretic perspective by going beyond the standard leading-order single-loop perturbative renormalization group (RG) analysis. Given that the effective (bare) coupling constant (i.e., the fine structure constant) in graphene is of order unity, which is neither small to justify a perturbative expansion nor large enough for strong-coupling theories to be applicable, the problem is a difficult one, with some similarity to (2+1)-dimensional strong-coupling quantum electrodynamics (QED). In this work, we take a systematic and comprehensive analytical approach in theoretically studying graphene many-body effects, primarily at the Dirac point (i.e., in undoped, intrinsic graphene), by going up to three loops in the diagrammatic expansion to both ascertain the validity of a perturbative expansion in the coupling constant and to develop a RG theory that can be used to estimate the actual quantitative renormalization effect to higher-order accuracy. Electron-electron interactions are expected to play an important role in intrinsic graphene due to the absence of screening at the Dirac (charge neutrality) point, potentially leading to strong deviations from the Fermi-liquid description around the charge neutrality point where the graphene Fermi velocity should manifest an ultraviolet logarithmic divergence because of the linear band dispersion. While no direct signatures for non-Fermi-liquid behavior at the Dirac point have yet been observed experimentally, there is ample evidence for the interaction-induced renormalization of the graphene velocity as the Dirac point is approached by lowering the carrier density. We provide a critical comparison between theory and experiment, using both higher-order diagrammatic and random phase approximation (i.e., infinite-order bubble diagrams) calculations, emphasizing future directions for a deeper understanding of the graphene effective field theory. We find that while the one-loop RG analysis gives reasonable quantitative agreement with the experimental data, both for graphene in vacuum and graphene on substrates, particularly when dynamical screening effects and finite carrier density effects are incorporated in the theory through the random phase approximation, the two-loop analysis reveals an interacting strong-coupling critical point in graphene suspended in vacuum signifying either a quantum phase transition or a breakdown of the weak-coupling renormalization group approach. By adapting a version of Dyson's argument for the breakdown of the QED perturbative expansion to the case of graphene, we show that in contrast to QED where the asymptotic perturbative series in the coupling constant converges to at least 137 orders (and possibly to much higher order) before diverging in higher orders, the graphene perturbative series in the coupling constant may manifest asymptotic divergence already in the first or second order in the coupling constant, favoring the conclusion that perturbation theory may be inadequate, particularly for graphene suspended in vacuum. We propose future experiments and theoretical directions to make further progress on this important and difficult problem. The question of convergence of the asymptotic perturbative expansion for graphene many-body effects is discussed critically in the context of the available experimental results and our theoretical calculations.