Challenges in constraining gravity with cosmic voids

We compare void size and clustering statistics for nDGP and f (R) gravity models and general relativity (GR) using N-body simulations. We show how it is critical to consider the statistics derived from mock galaxy catalogs rather than the dark matter halos alone. Marked differences between the void size functions for GR and f (R) models which present when voids are identified using dark matter halos are removed when voids are identified, more realistically, from mock galaxy tracers of the halos. The void radial velocities and velocity dispersions in the f (R) and nDGP models are enhanced relative to GR in both halos and mock galaxy identified voids. Despite this, we find that the redshift space void quadrupole moments derived from the mock galaxy tracers are strikingly similar across the three gravity models. The Gaussian streaming model (GSM) is shown to accurately reconstruct ₂ in modified gravity models, and we employ the GSM, using a functional derivative approach, to analyze the insensitivity of ₂ to the gravity model. Assuming linear theory, we show the void quadrupole to be an unbiased estimator of the redshift space growth rate parameter =f/b in the modified gravity theories.