LISs in Planetary Nebulae

• Low-ionization structures (LISs) are compact nebular features in planetary nebulae, identified by enhanced emission in lines like [N II] and [S II] and exhibiting diverse morphologies such as knots, filaments, and jets.
• Studies reveal that LISs have systematically lower electron densities in their ionized skins compared to surrounding rims, with hidden molecular and neutral gas layers uncovered through IR and optical diagnostics.
• Diagnostic frameworks using ratios of molecular and atomic lines indicate that excitation in LISs spans from UV photoionization to shock-dominated regimes, offering insights into their formation and evolution.

Searching arXiv for recent and foundational papers on low-ionization structures in planetary nebulae. Low-ionization structures (LISs) are small-scale nebular features in planetary nebulae (PNe) that are prominent in emission from low-ionization species such as [N II], [S II], [O II], and [O I]. They include knots, filaments, jets, clumps, ansae, and related condensations embedded within rims, shells, and haloes, and they are found in essentially all morphological classes of PNe. Their importance follows from the fact that they are sites where fast collimated outflows, shocks, photoevaporation, and strong UV irradiation interact with dense nebular material, so they preserve information on nebular shaping, kinematic history, and the balance between radiative and mechanical excitation (Akras et al., 2024, Mari et al., 27 Aug 2025, Mari et al., 2023).

1. Definition, taxonomy, and observational identity

LISs are defined observationally by their enhanced low-ionization emission relative to the surrounding nebula. Across the literature synthesized in recent statistical and object-by-object studies, they appear as compact knots, filaments, jets, jet-like systems, ansae, arcs, and isolated condensations, and are often identified in [N II] $\lambda\lambda 6548,6584$, [S II] $\lambda\lambda 6716,6731$, [O II] $\lambda 3727$, and sometimes [O I] $\lambda 6300$ (Mari et al., 2023, Akras et al., 2015).

They span a wide kinematic range. One review-oriented observational study describes LISs as compact structures with velocities from about $30\ \mathrm{km\ s^{-1}}$ to $350\ \mathrm{km\ s^{-1}}$, and later work continues to treat them as a heterogeneous family that includes both slowly moving peripheral knots and highly collimated fast outflows (Akras et al., 2016, Akras et al., 2016). Terminology such as FLIER-like features and fast-moving low-ionization emission regions is used in parts of the literature, but current work generally treats these morphologies collectively as LISs when their defining property is enhanced low-ionization emission (Akras et al., 2020, Ali et al., 2017).

This heterogeneous morphology is not a classification nuisance but a physical clue. A plausible implication is that LISs do not arise from a single formation channel; rather, they sample multiple dynamical regimes that later become processed by the same PN radiation field.

2. Plasma properties and the density problem

The most robust global result is that LISs have lower electron densities than the rims or shells of their host nebulae. In the statistical analysis of 33 PNe containing 104 LISs and 88 rims/shells, LISs have a median electron density of about $\sim 1700~\mathrm{cm^{-3}}$, while rims/shells have $\sim 2700~\mathrm{cm^{-3}}$; the interquartile ranges are $800\!-\!2700~\mathrm{cm^{-3}}$ for LISs and $1800\!-\!5000~\mathrm{cm^{-3}}$ for rims/shells (Mari et al., 2023, Mari et al., 27 Aug 2025). Long-slit and IFU studies of individual nebulae reinforce the same pattern: LISs are characterized by lower or at most equal $\lambda\lambda 6716,6731$0 than associated rims and shells, and in NGC 3242 the lower density is localized to the LISs while an increase is observed at the rim (Mari et al., 2022, Konstantinou et al., 20 Mar 2025).

By contrast, temperatures are much less distinctive. The broad consensus is that $\lambda\lambda 6716,6731$1 does not show significant variations between LISs and surrounding nebular components, while $\lambda\lambda 6716,6731$2 may be slightly higher in LISs but with uncertainties large enough that object-by-object significance is weak (Mari et al., 2023, Mari et al., 2022). In NGC 3242, the ordering derived from multiple diagnostics is

$\lambda\lambda 6716,6731$3

with temperatures generally higher in the inner nebular structures rather than specifically enhanced in the LISs (Konstantinou et al., 20 Mar 2025).

Chemical abundances are likewise not a primary discriminator. Statistical and spatially resolved studies consistently report no significant LIS/rim difference in total abundances of He, N, O, Ne, Ar, Cl, and S, with only limited suggestions of slightly higher nitrogen abundances in some knots of NGC 5307 (Mari et al., 2023, Ali et al., 2017, Akras et al., 2015). The low-ionization brightness of LISs therefore does not require a chemically distinct ejecta component.

The central tension is that many formation models require total densities well above $\lambda\lambda 6716,6731$4, and in some scenarios as high as $\lambda\lambda 6716,6731$5, whereas the ionized gas traced by optical diagnostics is systematically less dense than the surrounding nebula (Mari et al., 27 Aug 2025). Current work increasingly interprets this not as a contradiction, but as evidence that optical diagnostics sample only the ionized skin while a neutral or molecular component remains hidden.

3. Molecular and neutral counterparts

Near-infrared imaging has established that LISs can contain molecular gas. Deep NIRI@Gemini observations detected H$\lambda\lambda 6716,6731$6 emission in the LISs of K 4-47 and NGC 7662, showing that both highly moving and slowly moving LISs can host molecular material (Akras et al., 2016, Akras et al., 2016). Subsequent very deep narrow-band H$\lambda\lambda 6716,6731$7 images of NGC 7009 and NGC 6543 detected H$\lambda\lambda 6716,6731$8 emission in LISs in both nebulae, with H$\lambda\lambda 6716,6731$9 1-0 S(1) surface brightnesses of $\lambda 3727$0 in NGC 7009 and $\lambda 3727$1 in NGC 6543 (Akras et al., 2020).

A larger H$\lambda 3727$2 imaging survey later nearly doubled the number of host PNe with confirmed LIS/H$\lambda 3727$3 associations. In 19 LISs across five Galactic PNe, the warm H$\lambda 3727$4 masses range from $\lambda 3727$5 to $\lambda 3727$6, with an average of $\lambda 3727$7, and the excited H$\lambda 3727$8 molecular mass is between 200 and 5000 times lower than the corresponding ionized gas mass (Mari et al., 27 Aug 2025). This was used to argue that the systematically lower electron density in LISs is linked to the presence of H$\lambda 3727$9 molecular gas.

Neutral atomic gas has now been added to this picture. In NGC 7009, the first spatially resolved detection of [C I] $\lambda 6300$0 from the outer pair of LISs showed intense [C I] emission where He I $\lambda 6300$1 is absent, while the host nebula shows the opposite behavior (Akras et al., 2024). In NGC 3242, MUSE data revealed [C I] $\lambda 6300$2 primarily emitted from the LISs, spatially associated with [O I] $\lambda 6300$3 and [N I] $\lambda 6300$4 (Konstantinou et al., 20 Mar 2025). Because [C I] $\lambda 6300$5 traces the transition zone between ionized and neutral gas, these detections support a layered structure in which LISs are not purely ionized condensations but contain neutral and plausibly molecular material (Akras et al., 2024, Konstantinou et al., 20 Mar 2025).

Taken together, the H$\lambda 6300$6 and [C I] results have shifted the interpretation of LISs from purely optical low-ionization features toward partially molecular, stratified condensations with ionized exteriors and neutral or molecular interiors.

4. Excitation mechanisms and diagnostic frameworks

The central interpretive problem is whether LIS emission is powered mainly by stellar photoionization, by shocks, or by both. Early spectroscopic work on five PNe argued that LISs are mainly shock-excited, introducing diagnostic diagrams based on $\lambda 6300$7 and proposing the empirical regimes

• $\lambda 6300$8 for shock-excited structures,
• $\lambda 6300$9 for photo-ionized structures,
• $30\ \mathrm{km\ s^{-1}}$0 for a transition zone (Akras et al., 2015).

Later analyses became more cautious. A spatially resolved study of six PNe concluded that shocks are not negligible but that their inferred importance depends strongly on orientation (Mari et al., 2022). A larger statistical study comparing LISs with photoionization and shock grids from Cloudy and MAPPINGS V found substantial overlap between the mechanisms in classical low-ionization diagrams and concluded that shocks cannot be the main source of excitation for most LISs of PNe (Mari et al., 2023). In that work, one of the more useful shock indicators is

$30\ \mathrm{km\ s^{-1}}$1

provided the density is below the regime where [O III] $30\ \mathrm{km\ s^{-1}}$2 is collisionally de-excited (Mari et al., 2023).

Near-infrared diagnostics have added new leverage but not complete uniqueness. For molecular gas, the framework based on

$30\ \mathrm{km\ s^{-1}}$3

was applied to LISs in NGC 7009 and NGC 6543. High $30\ \mathrm{km\ s^{-1}}$4 often suggests thermal or shock excitation, and low values often suggest UV fluorescence, but the observations and model grids overlap too strongly to identify a unique dominant mechanism in general (Akras et al., 2020).

For atomic gas, a pilot [Fe II] survey targeted the [Fe II] $30\ \mathrm{km\ s^{-1}}$5m line as a shock tracer. Because the narrowband filter includes both [Fe II] $30\ \mathrm{km\ s^{-1}}$6m and H I 12-4 at $30\ \mathrm{km\ s^{-1}}$7m, the theoretical H I 12-4 flux was computed from either Br$30\ \mathrm{km\ s^{-1}}$8 or H$30\ \mathrm{km\ s^{-1}}$9 using Case B recombination for $350\ \mathrm{km\ s^{-1}}$0 and $350\ \mathrm{km\ s^{-1}}$1, with

$350\ \mathrm{km\ s^{-1}}$2

and then subtracted from the observed narrowband fluxes (Akras et al., 2024). The resulting diagnostic,

$350\ \mathrm{km\ s^{-1}}$3

is interpreted such that low values favor UV photoionization, high values favor shock excitation, and intermediate values indicate a mixture of both (Akras et al., 2024).

The net result is not a single universal excitation law. Instead, current diagnostics show that LISs occupy a continuum from predominantly photoionized to clearly shock-dominated, with many objects in an intermediate regime.

5. Empirical archetypes in well-studied nebulae

A small set of nebulae has become archetypal because their LISs occupy different parts of the excitation sequence.

Object	Diagnostic result	Interpretation
K 4-47	H$350\ \mathrm{km\ s^{-1}}$4 1-0 S(1)/2-1 S(1) $350\ \mathrm{km\ s^{-1}}$5 7–10	shock interaction
NGC 7662	H$350\ \mathrm{km\ s^{-1}}$6 1-0 S(1)/2-1 S(1) = 2–3	photo-ionization by the central star
NGC 6543	$350\ \mathrm{km\ s^{-1}}$7	UV radiation from the central star is likely dominant
NGC 7009	outer LISs $350\ \mathrm{km\ s^{-1}}$8	combination of UV photoionization and shock contribution
IC 4634	$350\ \mathrm{km\ s^{-1}}$9	shock excitation
NGC 6571	$\sim 1700~\mathrm{cm^{-3}}$0	shock excitation as the main driver

These empirical cases come from direct imaging of molecular and atomic tracers associated with LISs (Akras et al., 2016, Akras et al., 2024). K 4-47 represents the fast, highly collimated, shock-dominated end of the class: H$\sim 1700~\mathrm{cm^{-3}}$1 emission comes from the walls of bipolar outflows and from a pair of knots at their tips, and the system has been interpreted as a jet or bullet interacting with surrounding asymptotic giant branch material (Akras et al., 2016). NGC 7662 represents the slower, predominantly photoionized end: H$\sim 1700~\mathrm{cm^{-3}}$2 is found in peripheral LISs but not in the main shell, and only four knots are detected in H$\sim 1700~\mathrm{cm^{-3}}$3 $\sim 1700~\mathrm{cm^{-3}}$4 S(1), giving line ratios between 2 and 3.5 (Akras et al., 2016).

NGC 7009 and NGC 6543 are especially important because they contain both molecular and atomic tracers. H$\sim 1700~\mathrm{cm^{-3}}$5 was detected in LISs in both nebulae, but the 2020 analysis concluded that the diagnostic diagram could not decide uniquely between UV or shocks as the mechanism behind the molecular emission (Akras et al., 2020). In NGC 7009, the later [C I] $\sim 1700~\mathrm{cm^{-3}}$6 detection in the outer LIS pair provided direct evidence that these LISs are photodominated regions, while the [Fe II] results for the same nebula indicate a moderate regime in which both mechanisms probably contribute (Akras et al., 2024, Akras et al., 2024). This juxtaposition does not represent a contradiction; it indicates that different tracers sample different layers and processes within the same LIS.

6. Internal structure and formation scenarios

Several recent studies converge on a layered structural model. In NGC 7009, the radial stratification places moderate-ionization lines closer to the central star, low-ionization lines farther out, and H$\sim 1700~\mathrm{cm^{-3}}$7 emission still farther out at the molecular edge (Akras et al., 2024). In NGC 3242, low-ionization and neutral lines such as [N II], [S II], [Fe II], [O I], and [C I] peak at the LIS positions, while high-ionization lines peak closer to the center or around the rim (Konstantinou et al., 20 Mar 2025). This is consistent with a mini-PDR-like or transition-layer geometry rather than a uniform ionized clump.

The preferred physical picture in several papers is that LISs are photoevaporated dense molecular clumps. In that scenario, the clump is dense enough to shield interior H$\sim 1700~\mathrm{cm^{-3}}$8 from UV destruction, the central-star radiation drives a photoevaporative flow from the surface, low-ionization emission arises in the skin of this flow, and the molecular core remains farther out and is traced by H$\sim 1700~\mathrm{cm^{-3}}$9 (Akras et al., 2024). The [C I] $\sim 2700~\mathrm{cm^{-3}}$0 line is particularly valuable because it probes the transition zone between fully ionized and neutral gas and, in the wording of the NGC 7009 study, “goes deeper into the molecular gas” than classical optical low-ionization lines (Akras et al., 2024).

The [Fe II] imagery adds a dynamical refinement to this model. Spatial offsets among [Fe II], H$\sim 2700~\mathrm{cm^{-3}}$1, and Br$\sim 2700~\mathrm{cm^{-3}}$2 in some nebulae were interpreted as especially suggestive of photoevaporating clumps with embedded shocks rather than simple direct shock interaction alone (Akras et al., 2024). In that picture, UV photons ionize the clump surface, drive a photoevaporative flow, and generate a slow reverse shock that can liberate Fe from dust and produce [Fe II] emission (Akras et al., 2024). This suggests that radiative and mechanical processes can coexist within a single LIS but dominate in different spatial zones.

Formation remains open. The literature summarized in recent surveys allows density enhancements produced by instabilities, stagnation points, asymptotic giant branch fossil condensations, jet-driven structures, and binary-related shaping, while emphasizing that present-day observables are further modified by shielding, photoevaporation, and the evolving UV field (Mari et al., 27 Aug 2025, Akras et al., 2015).

7. Open questions and current synthesis

Despite nearly 30 years of detailed study, the origin and nature of LISs remain unresolved (Mari et al., 2023). Three questions dominate the current literature.

The first is the density problem: LISs are observed to have lower electron densities than the surrounding rims and shells, yet model expectations often require dense condensations. The detection of H$\sim 2700~\mathrm{cm^{-3}}$3 in an increasing number of LISs strongly supports the interpretation that optical $\sim 2700~\mathrm{cm^{-3}}$4 measures only the ionized fraction and misses a larger neutral or molecular reservoir (Mari et al., 27 Aug 2025).

The second is excitation ambiguity. H$\sim 2700~\mathrm{cm^{-3}}$5 ratios, classical optical diagnostic diagrams, and even some [C I] constraints do not uniquely separate UV and shocks across the full class, because photoionization and shock models overlap substantially in the relevant line-ratio space (Akras et al., 2020, Mari et al., 2023). The [Fe II] $\sim 2700~\mathrm{cm^{-3}}$6m survey has sharpened the issue by providing the first direct detections of [Fe II] associated with LISs in four out of five surveyed PNe and by showing object-to-object variation from UV-dominated to shock-dominated regimes (Akras et al., 2024). Still, no single tracer appears sufficient on its own.

The third is internal structure. Recent [C I], H$\sim 2700~\mathrm{cm^{-3}}$7, and [Fe II] detections favor a stratified model in which LISs are dense, partially molecular, externally illuminated, and in some cases photoevaporating (Akras et al., 2024, Akras et al., 2024). This suggests that the term “low-ionization structure” is observationally accurate but physically incomplete: the low-ionization optical skin is only one layer of a more complex condensation.

The current synthesis is therefore conservative but specific. LISs are not chemically distinct sub-populations; they are small-scale nebular condensations with systematically low electron densities in their ionized skins, frequent molecular and neutral counterparts, and excitation that ranges from photoionization-dominated to shock-dominated depending on local kinematics, UV field, density structure, and geometry (Ali et al., 2017, Mari et al., 2022, Mari et al., 2023). A plausible implication is that progress will continue to come from combining spatially resolved optical IFU data with narrow-band near-infrared imaging and spectroscopy, because the decisive information lies in how ionized, partially ionized, neutral, and molecular layers are arranged within individual LISs.