Spherically symmetric models of loop quantum gravity have been studied recently by different methods that aim to deal with structure functions in the usual constraint algebra of gravitational systems. As noticed by Gambini and Pullin, a linear redefinition of the constraints (with phase-space dependent coefficients) can be used to eliminate structure functions, even Abelianizing the moredifficult part of the constraint algebra. The Abelianized constraints can then easily be quantized or modified by putative quantum effects. As pointed out here, however, the method does not automatically provide a covariant quantization, defined as an anomaly-free quantum theory with a classical limit in which the usual (off-shell) gauge structure of hypersurface deformations in spacetime appears. The holonomy-modified vacuum theory based on Abelianization is covariant in this sense, but matter theories with local degrees of freedom are not. Detailed demonstrations of these statements show complete agreement with results of canonical effective methods applied earlier to the same systems (including signature change).

Marginal LTB models with corrections from loop quantum gravity have recently been studied with an emphasis on potential singularity resolution. This paper corroborates and extends the analysis in two regards: (i) the whole class of LTB models, including non-marginal ones, is considered, and (ii) an alternative procedure to derive anomaly-free models is presented which first implements anomaly-freedom in spherical symmetry and then the LTB conditions rather than the other way around. While the two methods give slightly different equations of motion, not altogether surprisingly given the ubiquitous sprawl of quantization ambiguities, final conclusions remain unchanged: Compared to quantizations of homogeneous models, bounces seem to appear less easily in inhomogeneous situations, and even the existence of homogeneous solutions as special cases in inhomogeneous models may be precluded by quantum effects. However, compared to marginal models, bouncing solutions seem more likely with non-marginal models.

Canonical quantum theories with discrete space may imply interesting effects. This article presents a general effective description, paying due attention to the role of higher spatial derivatives in a local expansion and differences to higher time derivatives. In a concrete set of models, it is shown that spatial derivatives one order higher than the classical one are strongly restricted in spherically symmetric effective loop quantum gravity. Moreover, radial holonomy corrections cannot be anomaly-free to this order.

Several properties of canonical quantum gravity modify space-time structures, sometimes to the degree that no effective line elements exist to describe the geometry. An analysis of solutions, for instance in the context of black holes, then requires new insights. In this article, standard definitions of horizons in spherical symmetry are first reformulated canonically, and then evaluated for solutions of equations and constraints modified by inverse-triad corrections of loop quantum gravity. When possible, a space-time analysis is performed which reveals a mass threshold for black holes and small changes to Hawking radiation. For more general conclusions, canonical perturbation theory is developed to second order to include back-reaction from matter. The results shed light on the questions of whether renormalization of Newton's constant or other modifications of horizon conditions should be taken into account in computations of black-hole entropy in loop quantum gravity.