Current mm/submm interferometers provide images of the dust and gas outer disks at an angular resolution of $\sim 1^{″}$ or 150  AU at the distance of the nearest star forming regions. Molecular hydrogen is, by far, the major constituent of the disk (with a dust-to-gas ratio taken as $\sim 100$) but it remains difficult to observe essentially because it does not possess a dipolar rotational spectrum. Its quadrupolar rotational lines, in the MID-IR domain, mostly trace material at hot/warm temperatures in the disk surface close to the star and do not sample the bulk of the cold gas located along the mid-plane. After H₂, carbon monoxide is the most abundant molecule (with abundance CO/H₂ of order $\sim 10^{-4}-10^{-6}$). Its first rotation lines are observable with current mm arrays allowing to trace the outer gas disk properties. Since the density in disks is very high ($n({\text{H}}_2)>10^6$  cm⁻³), the $J=1$-0 and $J=2$-1 CO lines are thermalized by collision with H₂ in almost the whole disk. Hence, a simple model of a Keplerian disk assuming LTE conditions is sufficient to retrieve the CO disk properties [9].

CO maps not only reveal that disks are in Keplerian rotation around TTauri [18] and Herbig Ae stars [20], they also allow us to retrieve the physical conditions in outer disks. Comparing resolved CO maps to a disk model by performing a ${\chi }^2$ minimization of the disk parameters gives quantitative information on the density and temperature radial and vertical distributions, provided the analysis properly takes into account the transfer function of the interferometer [11,14].

The radial profiles of the temperature deduced from¹²CO images appear to be consistent with models of stellar heating in flared disks, the outer disk being colder than the inner one. Moreover, since the ${}^{12}\text{CO}$ and ${}^{13}\text{CO}$ $J=1$-0 and $J=2$-1 lines have different opacities, they sample different disk layers and allow us to probe the existence of a vertical temperature gradient. Dartois et al. [8] have shown that the ‘CO disk surface’, traced by ${}^{12}\text{CO}$, is significantly hotter than the mid-plane traced by the optically thinner ${}^{13}\text{CO}$ $J=1$-0 transition, in the DM Tau disk. This effect appears in the region of the disk where the dust is still optically thick in absorption to the stellar radiation while it is already optically thin to its own emission, around $r\sim 50$–200 AU in the DM Tau case. Beyond $r\ge 200$ AU where the dust becomes optically thin to both processes, the temperature profile appears vertically isothermal. These observations were the first direct confirmation of models of the vertical thermal structure in disks (e.g., [5–7]). More recently, Piétu et al. [22] have extended this type of analysis to several objects, confirming the existence of vertical and radial temperature gradients. They also find that the turbulence appears to be subsonic in the outer part of disks [22].

It is also important to mention that CO in the gas phase has been observed in all TTauri disks at a temperature which appears to be below the CO freeze out point (17 K). This suggests that the vertical turbulence should play an important role by providing chemical mixing between the vertical molecular layers.

Searching for molecules other than CO remains sensitivity limited to the most abundant molecules found in molecular clouds. So far, there are only a few attempts to survey a large set of molecules in protoplanetary disks [10,17,27,11]. In addition to ${}^{13}\text{CO}$ and C¹⁸O, only HCO⁺, H ${{}^{13}\text{CO}}^{+}$, DCO⁺, CS, HCN, HNC, DCN, CN, H₂CO, N₂H⁺and C₂H have been detected. This domain will also benefit from ALMA, the large submm interferometer currently in construction in Chili.

Continuum observations of the dust, both at mm and NIR wavelengths, imply that the dust in such objects has evolved compared to that found in the interstellar medium. In particular, the particle size has started to increase from sub-micron size up to a few centimeters [25]. One key problem in planetary formation is the understanding of the dust evolution in term of grain size, vertical and radial distribution, composition and chemical nature from the inner disk to the outer disk. For example, in the interstellar medium (ISM), the silicate grains are amorphous while in our Solar system, comets and meteorites present crystalline silicates. The new generation of optical, Near-IR and MID-IR interferometers begin to provide direct insights on the inner disks where planetary formation is thought to occur. The observations discussed below present a qualitative but major improvement in the knowledge of the dust properties in inner disks.

Van Boekel et al. [26] recently observed three disks orbiting young stars of $\sim 2$ M_⊙ and a few million year old with MIDI, the 10 $\mu$m interferometer operated on the Very Large Telescope Interferometer (VLTI) by ESO. They compare the spectra of the silicate bands at 10 $\mu$m of the very inner disks (radius $R<2$–4 AU) observed with the VLTI with the spectra of the disks outside $R>2$–4 AU. The comparison reveals that in the inner disks the silicate grains appear more crystallized (1) than any other dust observed around young stellar objects and (2) than in the outer part of the disks. Fig. 3 shows the spectra. In the very inner disks, the spectra present the three peaks which are characteristics of crystalline silicates; the same peaks are also observed in the spectra of comets such as Halley or Hale-Bopp. The spectra in the outer part of the disks are more similar to those observed in the ISM, characteristics of amorphous silicates. The presence of crystalline silicates in inner disks of a few million years old suggests that silicates crystallize very early in the evolution of the disks, likely before the formation of terrestrial planets.

So far, these recent data bring the first observational constraints on the properties of dust grains at the scale of a few AU in disks. These kinds of observations will likely provide in the next years the first detailed analysis of the dust mineralogy in inner disks, leading to a better understanding of the dust evolution in the earliest phases of the planetary formation.

More generally, recent progress in observational astrophysics begins to provide the sensitivity and the angular resolution needed to study the regions of proto-planetary disks where planetary formation should occur.

Observations of numerous TTauri stars located in star forming regions of various ages have been made by several groups. Fig. 4 is a compilation by [15] of the IR excess (traced by the H–K excess sampling the very inner disk) versus the age of the stars. There are many uncertainties associated to this kind of analysis, the most important being (1) the errors on the estimate of the IR excess and (2) the derivation of the stellar age (which is taken as the median apparent age of the individual objects belonging to the same cluster). However, there is a clear trend which can be seen on this plot: at age around 1 million years, 80–90% of the TTauri stars are surrounded by an innermost dust disk while the fraction of inner disks drops to 30–40% at 2–3 million years and beyond 5 million years, almost all disks have disappeared.

Comparison of the IR excess with the observations of the dust observed in the mm/submm domain [3] and CO data [12] suggest that the dissipation of the inner and outer disks are roughly simultaneous. However, since the statistics in the mm/submm domain are poor, it is difficult to derive quantitative numbers, yet.

Nevertheless, taking into account the various types of surveys and their results, it seems reasonable to conclude that the dissipation is globally a fast process since more than half of the inner disks disappear after 2–3 Myr.

There are many physical reasons which should explain the apparent dispersal of disks:

• • the formation of a planetary system;
• • the formation of a binary or a multiple stellar system;
• • the tidal truncation by stellar encounters;
• • the photo-evaporation of the disk.

Large optical/NIR surveys of star forming clusters have shown that many stars form in binary and/or multiple systems (e.g., [19]). The formation of a binary or a multiple stellar system strongly affects the original distribution of material which is tidally disrupted. For example, observations of the TTauri binary system GG Tau [9] have shown that the gas and the dust are located inside a circumbinary disk orbiting the binary while very small individual disks exist around each stellar component, inside the Roche lobe. Hence, even if the innermost disks can exist, statistics provided by studies such as those presented above, may be partially biased by the presence in the samples of unresolved multiple stellar systems which are not yet known.

In some clusters, such as $\rho$ Ophiuchus or Orion, the stellar density can be as high as 2–$5\times 10^4$  pc⁻³ (eg [16] and external tidal truncations of disks by stellar encounters should be a relatively common phenomenon. For a matter of sensitivity and angular resolution, we do not have yet enough resolved images of disks in these clusters to estimate the importance of this effect. It is, however, reasonable to take it seriously into account, at least as a phenomenon which should be responsible for the existence of smaller disks in dense stellar clusters.

The photo-evaporation of disks is invoked as a possible phenomenon responsible for the disruption of inner disks [2]. UV photons coming from the young stars reach the disk surface where they can create a ionization front, heating the material behind. The ionized gas becomes gravitationally unbound and flows in a wind away from the disk surface. This phenomenon is observed for the proplyds orbiting TTauri stars located in the Orion Trapezium region [4], close to the OB association. The strong ionizing UV flux from the OB stars photo-evaporates the outer disks in their surroundings. Alexander and collaborators have shown that in some cases photo-evaporation should clear the inner disk very efficiently, creating a central hole, and disrupt the whole disk on a time-scale of about $5\times 10^5$ years after a disk lifetime of 2–3 Myr. This physical process has then to be taken into account because the time-scales (both lifetime and dispersal) are those which are observed for TTauri disks, as illustrated in the Section 3.1. However, there is yet no clear observational evidences of the importance of this process for the disk clearing. Dedicated observations, such as sensitive and resolved mm/submm images of inner disks, have to be performed in the next years in order to disentangle between the various scenarii which can be responsible for the existence of an inner hole in disks: photo-evaporation or planetary formation.

Moreover, whatever the frequencies of the various physical processes described in this section are, all of them may affect planet formation since they have a direct impact on the dust and gas distribution close to the stars with the same time-scales.