In contrast to all other debris disks, where the dust can be seen via an infrared excess over the stellar photosphere, the dust emission of the Edgeworth-Kuiper belt (EKB) eludes remote detection because of the strong foreground emission of the zodiacal cloud. We accessed the expected EKB dust disk properties by modeling. We treated the debiased population of the known trans-Neptunian objects (TNOs) as parent bodies and generated the dust with our collisional code. The resulting dust distributions were modified to take into account the influence of gravitational scattering and resonance trapping by planets on migrating dust grains as well as the effect of sublimation. A difficulty with the modeling is that the amount and distribution of dust are largely determined by sub-kilometer-sized bodies. These are directly unobservable, and their properties cannot be accessed by collisional modeling, because objects larger than (10 . . . 60) m in the present-day EKB are not in a collisional equilibrium. To place additional constraints, we used in-situ measurements of the New Horizons spacecraft within 20 AU. We show that to sustain a dust disk consistent with these measurements, the TNO population has to have a break in the size distribution at s 70 km. However, even this still leaves us with several models that all correctly reproduce nearly constant dust impact rates in the region of giant planet orbits and do not violate the constraints from the non-detection of the EKB dust thermal emission by the COBE spacecraft. The modeled EKB dust disks, which conform to the observational constraints, can either be transport-dominated or intermediate between the transport-dominated and collision-dominated regime. The in-plane optical depth of such disks is τ (r > 10 AU) ∼ 10 −6 and their fractional luminosity is f d ∼ 10 −7 . Planets and sublimation are found to have little effect on dust impact fluxes and dust thermal emission. The spectral energy distribution of an EKB analog as seen from 10 pc distance peaks at wavelengths of (40 . . . 50) µm at F ≈ 0.5 mJy, which is less than 1% of the photospheric flux at those wavelengths. Therefore, EKB analogs cannot be detected with present-day instruments such as Herschel/PACS.

The Edgeworth-Kuiper belt (EKB) and its presumed dusty debris is a natural reference for extrsolar debris disks. We re-analyze the current database of known transneptunian objects (TNOs) and employ a new algorithm to eliminate the inclination and the distance selection effects in the known TNO populations to derive expected parameters of the "true" EKB. Its estimated mass is M EKB = 0.12 M ⊕ , which is by a factor of ∼15 larger than the mass of the EKB objects detected so far. About a half of the total EKB mass is in classical and resonant objects and another half is in scattered ones. Treating the debiased populations of EKB objects as dust parent bodies, we then "generate" their dust disk with our collisional code. Apart from accurate handling of destructive and cratering collisions and direct radiation pressure, we include the Poynting-Robertson (P-R) drag. The latter is known to be unimportant for debris disks around other stars detected so far, but cannot be ignored for the EKB dust disk because of its much lower optical depth. We find the radial profile of the normal optical depth to peak at the inner edge of the classical belt, ≈40 AU. Outside the classical EKB, it approximately follows τ ∝ r −2 which is roughly intermediate between the slope predicted analytically for collision-dominated (r −1.5 ) and transport-dominated (r −2.5 ) disks. The size distribution of dust is less affected by the P-R effect. The cross section-dominating grain size still lies just above the blowout size (∼1 . . . 2 µm), as it would if the P-R effect was ignored. However, if the EKB were by one order of magnitude less massive, its dust disk would have distinctly different properties. The optical depth profile would fall off as τ ∝ r −3 , and the cross section-dominating grain size would shift from ∼1 . . . 2 µm to ∼100 µm. These properties are seen if dust is assumed to be generated only by known TNOs without applying the debiasing algorithm. An upper limit of the in-plane optical depth of the EKB dust set by our model is τ = 2 × 10 −5 outside 30 AU. If the solar system were observed from outside, the thermal emission flux from the EKB dust would be about two orders of magnitude lower than for solar-type stars with the brightest known infrared excesses observed from the same distance. Herschel and other new-generation facilities should reveal extrasolar debris disks nearly as tenuous as the EKB disk. We estimate that the Herschel/PACS instrument should be able to detect disks at a ∼1 . . . 2M EKB level.

Infrared excesses associated with debris disk host stars detected so far, peak at wavelengths around ∼ 100 µm or shorter. However, six out of 31 excess sources studied in the Herschel a Open Time Key Programme, DUNES, have been seen to show significant -and in some cases extended -excess emission at 160 µm, which is larger than the 100 µm excess. This excess emission has been attributed to circumstellar dust and has been suggested to stem from debris disks colder than those known previously. Since the excess emission of the cold disk candidates is extremely weak, challenging even the unrivaled sensitivity of Herschel, it is prudent to carefully consider whether some or even all of them may represent unrelated galactic or extragalactic emission, or even instrumental noise. We readdress these issues using several distinct methods and conclude that it is highly unlikely that none of the candidates represents a true circumstellar disk. For true disks, both the dust temperatures inferred from the spectral energy distributions and the disk radii estimated from the images suggest that the dust is nearly as cold as a blackbody. This requires the grains to be larger than ∼ 100 µm, even if they are rich in ices or are composed of any other material with a low absorption in the visible. The dearth of small grains is puzzling, since collisional models of debris disks predict that grains of Cold debris disks 3 all sizes down to several times the radiation pressure blowout limit should be present. We explore several conceivable scenarios: transport-dominated disks, disks of low dynamical excitation, and disks of unstirred primordial macroscopic grains. Our qualitative analysis and collisional simulations rule out the first two of these scenarios, but show the feasibility of the third one. We show that such disks can indeed survive for gigayears, largely preserving the primordial size distribution. They should be composed of macroscopic solids larger than millimeters, but smaller than a few kilometers in size. If larger planetesimals were present, they would stir the disk, triggering a collisional cascade and thus causing production of small debris, which is not seen. Thus planetesimal formation, at least in the outer regions of the systems, has stopped before "cometary" or "asteroidal" sizes were reached.