The $Herschel$ Space telescope carried out an unprecedented survey of nearby stars for debris disks. The dust present in these debris disks scatters and polarizes stellar light in the visible part of the spectrum. We explore what can be learned with aperture polarimetry and detailed radiative transfer modelling about stellar systems with debris disks. We present a polarimetric survey, with measurements from the literature, of candidate stars observed by DEBRIS and DUNES $Herschel$ surveys. We perform a statistical analysis of the polarimetric data with the detection of far-infrared excess by $Herschel$ and $Spitzer$ with a sample of 223 stars. Monte Carlo simulations were performed to determine the effects of various model parameters on the polarization level and find the mass required for detection with current instruments. Eighteen stars were detected with a polarization $0.01 \le P \lesssim 0.1$ per cent and $\ge3\sigma_P$, but only two of them have a debris disk. No statistically significant difference is found between the different groups of stars, with, without, and unknown status for far-infrared excess, and presence of polarization. The simulations show that the integrated polarization is rather small, usually $< 0.01$ per cent for typical masses detected by their far-infrared excess for hot and most warm disks. Masses observed in cold disks can produce polarization levels above $0.01$ per cent since there is usually more dust in them than in closer disks. We list five factors which can explain the observed low-polarization detection rate. Observations with high-precision polarimeters should lead to additional constraints on models of unresolved debris disks.

[Abridged] Debris disks are extrasolar analogs to the solar system planetesimal belts. The star Fomalhaut harbors a cold debris belt at 140 AU as well as evidence of a warm dust component, which is suspected of being a bright analog to the solar system's zodiacal dust. Interferometric observations obtained with the VLTI and the KIN have identified near- and mid-infrared excesses attributed to hot and warm exozodiacal dust in the inner few AU of the star. We performed parametric modeling of the exozodiacal disk using the GRaTeR radiative transfer code to reproduce the interferometric data, complemented by mid- to far-infrared measurements. A detailed treatment of sublimation temperatures was introduced to explore the hot population at the sublimation rim. We then used an analytical approach to successively testing several source mechanisms. A good fit to the data is found by two distinct dust populations: (1) very small, hence unbound, hot dust grains confined in a narrow region at the sublimation rim of carbonaceous material; (2) bound grains at 2 AU that are protected from sublimation and have a higher mass despite their fainter flux level. We propose that the hot dust is produced by the release of small carbon grains following the disruption of aggregates that originate from the warm component. A mechanism, such as gas braking, is required to further confine the small grains for a long enough time. In situ dust production could hardly be ensured for the age of the star, so the observed amount of dust must be triggered by intense dynamical activity. Fomalhaut may be representative of exozodis that are currently being surveyed worldwide. We propose a framework for reconciling the hot exozodi phenomenon with theoretical constraints: the hot component of Fomalhaut is likely the tip of the iceberg since it could originate from a warm counterpart residing near the ice line.