Molecular Hydrogen Microstructures in Planetary Nebulae

: Molecular hydrogen (H 2 ) emission is commonly detected in planetary nebulae (PNe), specially in objects with bipolar morphologies. New studies showed that H 2 gas is also packed in microstructures embedded in PNe of any morphological type. Despite the presence of H 2 in cometary knots being known for years, only in the last ﬁve years, much deeper imagery of PNe have revealed that H 2 also exists in other types of low-ionisation microstructures (LISs). Signiﬁcant differences are found between the host PNe of cometary knots and other types of LISs, such as nebula age, central star temperature (evolutionary stage) and the absolute sizes of the microstructure itself.


Introduction
Planetary nebulae (PNe) represent the final evolutionary stage of low-and intermediate-mass stars (1-8 M ), and they are formed by the interaction of two or more mass-loss events during the previous stages. The result of these interactions in conjunction with the evolution of the central star (CS) is the development of structures like shells, rims and haloes. Many PNe, however, have been found to posses microstructures (≤2-3 arcsec) that are noticeable in low-ionisation emission-lines such as [N II], [S II], and [O I] (e.g., [1,2]). Based on their kinematic characteristics, these microstructures are labeled as: (i) fast low-ionisation emission regions (FLIERs; [2]), (ii) slow moving low ionisation emitting regions (SLOWERs; [3]), and (iii) bipolar, rotating, episodic jets (BRETs; [4]).
A comprehensive review on the morphological and kinematic properties of these microstructures was given by [5], also discussing potential links to various formation models. Interestingly, none of the available formation models are able to sufficiently explain all their properties. Due to their substantial different kinematic properties, all the microstructures were included under one class called low-ionisation structures (LISs, [5]). The intriguing enhancement of the low-ionisation emission-lines in these microstructures relative to the surrounding nebular gas has been attributed to either UV radiation from the CS (e.g., [6]) or a combination with shocks (e.g., [7]). Still, little is known about these features.
A particular sub-class of LISs has been found in nearby PNe, namely the cometary knots (CKs) with a long radial tail extended away from the CS [8]. CKs in the Helix, Ring, and Dumbbell nebulae are known to be composed of molecular hydrogen (H 2 ) (e.g., [9][10][11][12]). Gonçalves and co-workers claimed that LISs should also be dense structures composed of molecular gas, similar to CKs [13]. This could explain their systematic lower electron densities compared to the surrounding nebula values (e.g., [6,7,13]).
In this paper, we present H 2 1-0 S(1) narrow-band images of PNe with low-ionisation structures to discuss the different patterns of the host PNe.

Molecular Hydrogen in Microstructures
Besides the presence of H 2 gas in the cometary knots of nearby PNe, new observations over the last five years have unveiled that more LISs are composed of H 2 gas. Fang et al. reported the detection of H 2 emission in microstructures embedded in four PNe. This includes the northwestern knot in Hu 1-2 [14], two pairs of knots in Hb 12 [15], six knots in NGC 7009 [15], and several knots in the halo of NGC 6543 [15] (Figure 1).  [14]), NGC 7009 (top-right panel, [15]), Hu 1-2 (bottom-left panel, [15]) and NGC 6543 (bottom-right panel, [15]). The insets on the top panels show a zoom in the H 2 knots.
Deep and high sensitive H 2 imagery of PNe with LISs obtained with the Near InfraRed Imager and Spectrograph (NIRI) at Gemini North have also detected H 2 emission-lines from LISs (K 4-47 and NGC 7662, [16]; NGC 6543 and NGC 7009, [17]) ( Figure 2). Note that, the higher spatial resolution of the NIRI images ( [17]), revealed H 2 emission from the inner LISs of NGC 6543.
All these H 2 detections associated with LISs (more than 50) have strongly supported the scenario that LISs are composed of molecular gas. A comparison between the CKs and other types of LISs shows a number of differences; so hereafter, we refer to the PNe with CKs as the CKs-PN group and to the PNe with the rest (mostly knots) as the Ks-PN group. Scrutinizing their H 2 images, we find that all knots but those in the elliptical NGC 7662 and in the halo of NGC 6543, lie in the polar direction. This is a perceptible difference between the two groups and it is very likely associated with their formation mechanisms (e.g., thermal instabilities, bullet/jet ejections or AGB fossils).
The two PN groups are characterized by different nebular ages and CS temperatures (T e f f ) (i.e., evolutionary stages). In particular, the CSs in the CKs-PN group have T e f f > 100 kK, whilst those in the Ks-PN group cover a wider range, 40 < T e f f < 120 kK (Table 1).
Another pronounced difference between the CKs and knots is that the former are found in younger PNe (≤2000 yrs) and the latter in older PNe (>7000 yrs) for which the CSs have already entered the cooling track (Table 1). The CKs in NGC 6720 are formed after the PN has entered the recombination phase and its CS has entered the cooling track [18]. This mechanism seems not to be applicable for the Ks-PN group, for which the CSs are in earlier evolutionary stage. What is the link (if any) between CKs and knots? Is the formation mechanism the same?
It is known that the intensities of H 2 lines around 2 µm and Brγ line as well as the R(H 2 )=)H 2 1-0 S(1)/H 2 1-2 S(1) and R(Brγ)=H 2 1-0 S(1)/Brγ ratios, are strongly dependent on time (evolutionary phase) and the dominant excitation mechanism of H 2 gas [19]. In proto-PN early phase, H 2 excitation is dominated by strong UV radiation. Due to high densities, collisional de-excitation becomes important and results in high R(H 2 ) ratios (∼10) and a thermal H 2 emission. For more evolved PNe with T e f f > 100 kK (e.g., NGC 6270), H 2 emission has again a thermal origin because of the more important contribution of X-rays. In intermediate phases, the rapid expansion of PNe results in significant decrease of the density and the strength of UV radiation field. Therefore, the excitation of H 2 is again dominated by the UV radiation field, but the de-excitation of gas by collisions is negligible due to the low densities resulting in a R(H 2 ) ratio close to 3 [20,21].  [16]), NGC 7662 (bottom-left panel, [16]), NGC 7009 (middle panel, [16]) and NGC 6543 (right panel [16]). The circles indicate LIS knots with H 2 emission detected.
NGC 7009 is characterized by low (western knot) and high (eastern knots) R(H 2 ) values [17]. This implies a different origin for the H 2 emission of each knot. The age of NGC 7009 is similar to those of NGC 7662 and NGC 6543 (Table 1) and its H 2 emission should be dominated by fluorescence. However, the high R(H 2 ) found in the eastern knots indicates a thermal origin. A density difference between the knots could explain this discrepancy.
Last but not least, the absolute size of the CKs and knots is also different. A rough estimate of the knot's size is given in Table 1, which corresponds to lower values. Knots are at least thrice as big than the CKs. This suggests that the amount of molecular material is likely most relevant in the former.

Future Work
Little is known about the formation of LISs in PNe and the excitation mechanisms of their H 2 gas. It is suggested from the above discussion that the host PNe of CKs and other types of LISs are in different evolutionary phases. It is, therefore, necessary to observe a large sample of PNe with LISs, covering different morphological types, CS's temperatures and types of LISs to determine the contribution of UV radiation, shocks and X-ray emission to the excitation of their H 2 gas.