Context. This paper discusses the sensitivity of water lines to chemical processes and radiative transfer for the protoplanetary disk around TW Hya. The study focuses on the Herschel spectral range in the context of new line detections with the PACS
instrument from the Gas in Protoplanetary Systems project (GASPS). Aims. The paper presents an overview of the chemistry in the main water reservoirs in the disk around TW Hya. It discusses the limitations in the interpretation of observed water line fluxes. Methods. ... (abbreviated) Results. We report new line detections of p-H2O (3_22-2_11) at 89.99 micron and CO J=18-17 at 144.78 micron for the disk around TW Hya. Disk modeling shows that the far-IR fine structure lines ([OI], [CII]) and molecular submm lines are very robust to uncertainties in the chemistry, while the water line fluxes can change by factors of a few. The water lines are optically thick, sub-thermally excited and can couple to the background continuum radiation field. The low-excitation water lines are also sensitive to uncertainties in the collision rates, e.g. with neutral hydrogen. The gas temperature plays an important role for the [OI] fine structure line fluxes, the water line fluxes originating from the inner disk as well as the high excitation CO, CH+ and OH lines. Conclusions. Due to their sensitivity on chemical input data and radiative transfer, water lines have to be used cautiously for understanding details of the disk structure. Water lines covering a wide range of excitation energies provide access to the various gas phase water reservoirs (inside and outside the snow line) in protoplanetary disks and thus provide important information on where gas-phase water is potentially located. Experimental and/or theoretical collision rates for H2O with atomic hydrogen are needed to diminish uncertainties from water line radiative transfer.
Aims. We want to understand the chemistry and physics of disks on the basis of a large unbiased and statistically relevant grid of disk models. One of the main goals is to explore the diagnostic power of various gas emission lines and line ratios for
deriving main disk parameters such as the gas mass. Methods. We explore the results of the DENT grid (Disk Evolution with Neat Theory) that consists of 300 000 disk models with 11 free parameters. Through a statistical analysis, we search for correlations and trends in an effort to find tools for disk diagnostic. Results. All calculated quantities like species masses, temperatures, continuum and line fluxes differ by several orders of magnitude across the entire parameter space. The broad distribution of these quantities as a function of input parameters shows the limitation of using a prototype T Tauri or Herbig Ae/Be disk model. The statistical analysis of the DENT grid shows that CO gas is rarely the dominant carbon reservoir in disks. Models with large inner radii (10 times the dust condensation radius) and/or shallow surface density gradients lack massive gas phase water reservoirs. Also, 60% of the disks have gas temperatures averaged over the oxygen mass in the range between 15 and 70 K; the average gas temperatures for CO and O differ by less than a factor two. Studying the observational diagnostics, the [CII] 158 mum fine structure line flux is very sensitive to the stellar UV flux and presence of a UV excess and it traces the outer disk radius (Rout). In the submm, the CO low J rotational lines also trace Rout. Low [OI] 63/145 line ratios (< a few) can be explained with cool atomic O gas in the uppermost surface layers leading to self-absorption in the 63 mum line; this occurs mostly for massive non-flaring, settled disk models without UV excess. ... abbreviated
Methods. We use the recently developed disk code ProDiMo to calculate the physico-chemical structure of protoplanetary disks and apply the Monte-Carlo line radiative transfer code RATRAN to predict observable line profiles and fluxes. We consider a s
eries of Herbig Ae type disk models ranging from 10^-6 M_Sun to 2.2 10^-2 M_Sun (between 0.5 and 700 AU) to discuss the dependency of the line fluxes and ratios on disk mass for otherwise fixed disk parameters. Results. We find the [CII] 157.7 mum line to originate in LTE from the surface layers of the disk, where Tg > Td . The total emission is dominated by surface area and hence depends strongly on disk outer radius. The [OI] lines can be very bright (> 10^-16 W/m^2) and form in slightly deeper and closer regions under non-LTE conditions. The high-excitation [OI] 145.5 mum line, which has a larger critical density, decreases more rapidly with disk mass than the 63.2 mum line. Therefore, the [OI] 63.2 mum/145.5 mum ratio is a promising disk mass indicator, especially as it is independent of disk outer radius for Rout > 200 AU. CO is abundant only in deeper layers A_V >~ 0.05. For too low disk masses (M_disk <~10^-4 M_Sun) the dust starts to become transparent, and CO is almost completely photo-dissociated. For masses larger than that the lines are an excellent independent tracer of disk outer radius and can break the outer radius degeneracy in the [OI] 63.2 mum/[CII]157.7 mum line ratio. Conclusions. The far-IR fine-structure lines of [CII] and [OI] observable with Herschel provide a promising tool to measure the disk gas mass, although they are mainly generated in the atomic surface layers. In spatially unresolved observations, none of these lines carry much information about the inner, possibly hot regions < 30 AU.
