An ALMA Molecular Inventory of Warm Herbig Ae Disks. II. Abundant Complex Organics and Volatile Sulphur in the IRS 48 Disk

The Atacama Large Millimeter/submillimeter Array (ALMA) can probe the molecular content of planet-forming disks with unprecedented sensitivity. These observations allow us to build up an inventory of the volatiles available for forming planets and comets. Herbig Ae transition disks are fruitful targets due to the thermal sublimation of complex organic molecules (COMs) and likely H2O-rich ices in these disks. The IRS 48 disk shows a particularly rich chemistry that can be directly linked to its asymmetric dust trap. Here, we present ALMA observations of the IRS 48 disk where we detect 16 different molecules and make the first robust detections of H213CO , 34SO, 33SO, and c-H2COCH2 (ethylene oxide) in a protoplanetary disk. All of the molecular emissions, aside from CO, are co-located with the dust trap, and this includes newly detected simple molecules such as HCO+, HCN , and CS. Interestingly, there are spatial offsets between different molecular families, including between the COMs and sulfur-bearing species, with the latter being more azimuthally extended and radially located further from the star. The abundances of the newly detected COMs relative to CH3OH are higher than the expected protostellar ratios, which implies some degree of chemical processing of the inherited ices during the disk lifetime. These data highlight IRS 48 as a unique astrochemical laboratory to unravel the full volatile reservoir at the epoch of planet and comet formation and the role of the disk in (re)setting chemical complexity.


INTRODUCTION
Due to the sensitivity of the Atacama Large Millimeter/submillimeter Array (ALMA) we now have unmatched access to the volatile reservoir in planetforming disks.In recent years, ALMA has enabled the detection of both new disk molecules including SO 2 and CH 3 CN, and rare isotopologues (e.g., 13 C 17 O and HC 18 O + ) ( Öberg et al. 2015;Booth et al. 2019Booth et al. , 2021a;;Furuya et al. 2022).What is particularly exciting is the detection of complex organic molecules (COMs) which are defined as molecules containing at least 6 atoms and of which at least one is carbon (Herbst & van Dishoeck 2009).Although the first detection of the simplest COM CH 3 OH in a Class II T-Tauri disk (TW Hya) traced a very low abundance of cold CH 3 OH (Walsh et al. 2016) subsequent observations of warmer Herbig Ae transition disks have revealed abundant thermally desorbed CH 3 OH and even other COMs of higher complexity (Booth et al. 2021b;van der Marel et al. 2021a;Booth et al. 2023;Brunken et al. 2022).The detection of abundant COMs in warm Herbig Ae disks is clear evidence for the inheritance of ices from the earlier stages of star formation.This is because CH 3 OH only forms efficiently on the surfaces of cold dust grains and primarily via the hydrogenation of CO ice (Watanabe & Kouchi 2002;Fuchs et al. 2009;Santos et al. 2022).
In the warm young F/Herbig Ae disks HD 100546, IRS 48 and HD 169142 there is no evidence for significant CO freeze-out meaning that the observed reservoir of CH 3 OH cannot have formed in-situ.This was shown directly for the HD 100546 disk using astrochemical models (Booth et al. 2021b).Therefore, in order for CH 3 OH to be present in these systems CH 3 OH rich ices must survive the star formation process and be transported to the inner disk where they thermally sublimate.CH 3 OH will come off the grains at a similar temperature as H 2 O (Minissale et al. 2022) and therefore the bulk of the volatile content of the disks should also be in the gas phase in this region of the disk.These sources therefore give us a window into a typically unobservable molecular reservoir in disks.
The disk most rich in COMs and potentially H 2 Oderived volatiles like SO is the disk around the young star IRS 48.The IRS 48 disk has been well studied with ALMA and hosts the most asymmetric dust trap yet discovered at a distance of 60 au from the central star (van der Marel et al. 2013a(van der Marel et al. , 2021b;;Yang et al. 2023).Its gas mass of only 5.5×10 −4 M ⊙ is much lower than other Herbig Ae disks, yet it is very line rich, with detection's of the CO isotopologues 12 CO, 13 CO, C 18 O, C 17 O along with SO, SO 2 , 34 SO 2 , NO, H 2 CO, CH 3 OH, CH 3 OCH 3 and, tentatively CH 3 OCHO (van der Marel et al. 2013aMarel et al. , 2021a;;Booth et al. 2021a;Brunken et al. 2022;Leemker et al. 2023).The significance of the reported non-detections of CS, C 2 H, CN in IRS 48 were quantified by Booth et al. (2021a) and Leemker et al. (2023) and indicate a C/O ratio in the disk gas that is significantly less than 1.This low C/O and lack of C 2 H is consistent with H 2 O being in the gas-phase and a general lack of volatile depletion at least at the location of the dust trap (Leemker et al. 2023).
There are several key simple molecules that have yet to be targeted in the IRS 48 disk, which would allow for a more complete comparison to other Herbig Ae disks.Here we present the results of an ALMA line survey of the IRS 48 disk where we target >20 molecular species.These data provide key constraints on the abundances of HCO + , HCN, CN, C 2 H and CS in this system.Additionally, we further unravel the volatile sul-phur and complex organic reservoir of the disk and discuss the physical/chemical origin of the molecular substructures observed.We particularly make a direct comparison between the molecular inventory of the IRS 48 and HD 100546 disks where the initial results for the latter are presented in Booth et al. (2024) and, contextualise the detections of COMs in these systems with protostellar environments.The data consist of two spectral settings with four spectral windows each at a spectral resolution of 976.6 kHz (0.84 km s −1 at 350 GHz) and a bandwidth of 1.875 GHz.These spectral windows are centered at 338.790824, 340.732413, 348.916936 and 350.775389 GHz for setting A and, 344.240980, 3459.40999, 354.367095 and 356.067114GHz for setting B. Further details on the individual execution blocks are provided in the Appendix in Table 4 and for full details the data reduction, observational set-up and imaging please refer to the companion paper which also presents data on the HD 100546 system Booth et al. (2024).The self-calibration was performed on the IRS 48 continuum data after flagging the strong lines which resulted in a continuum signal-to-noise increase from ≈475 to ≈3220.This process consisted of four rounds of phasecalibration and one round of amplitude calibration and resulted in the detection of the weak millimetre emission in the north of the IRS 48 disk.The data were imaged in CASA using tCLEAN with the multiscale deconvolver with a uniform velocity resolution of 0.9 km s −1 .These ≈0.′′ 3 data have a beam area 2.5× smaller than that presented the series of papers from van der Marel et al. (2021a); Booth et al. (2021a); Brunken et al. (2022); Leemker et al. (2023).Individual lines were cleaned with Keplerian masks down to 4× the rms of the dirty image where the Keplerian masks were constructed using the properties for the IRS 48 disk as listed in Table 1.The properties of the transitions imaged and the resulting beam sizes and rms noise for each line are listed in Tables 5 and 6 in the Appendix.

Molecules detected
We use matched filtering to make an initial line identification (Loomis et al. 2018b).This technique uses the predictable Keplerian rotation of the disk gas to detect molecular lines in the visibility data via cross correlation of the uv-data with a filter.This filter can be a  smooth model, e.g. a Keplerian mask, the FITS output of a line radiative transfer model or, a strong line detection in the disk.In Figure 1 we present the resulting matched filter response over the full data set for the IRS 48 disk with a Keplerian model with an outer radius of 150 au compared to the HD 100546 response (outer radius of 300 au) that is presented in Booth et al. (2024).HD 100546 was observed in the same manner as IRS 48 and we find that the IRS 48 disk is more line rich but there are different molecules detected in each disk.These differences may be attributed to different physical properties of the systems and/or the dominant chemical processes.Both are disks around young A-type stars and the characteristics of these two systems are compared in Table 1.In Section 4.2 we discuss the similarities and differences both physically and chemically between the two disks.The fully annotated version of the IRS 48 filter response is shown in the Appendix in Figure 7. From this we have detected 16 molecular species in IRS 48 disk where a detection is defined as a matched filter response of at least 4 σ.This includes robust detections of the rare isotopologues H 13 2 CO, 34 SO and 33 SO and, the detection of the first heterocycle -ethylene oxide (c−H 2 COCH 2 ) -in protoplanetary disks.We detect two lines of c−H 2 COCH 2 with the fiducial Keplerian model filter at rest frequencies of 338.7720826GHz and 350.3036524GHz.Using alternative image filters does not yield a significant improvement in the detection strength -likely due to the compact nature of the emission.In the channel maps compact emission from c−H 2 COCH 2 is detected at the 4 σ level over the 3 consecutive channels where the CH 3 OH lines are strongest for both lines.Interestingly, although the isomer acetaldehyde (CH 3 CHO) Note-The presence of H 13 CN in IRS 48 is unclear (indicated with "?") due to line blending with SO 2 .
us typically more abundant (Ikeda et al. 2001

Integrated intensity maps
Figure 2 presents the 0.9 mm continuum map and the integrated intensity maps of the representative transitions of each molecule detected in the IRS 48 disk.This galley does not include the isotopologues of SO which will be the focus of a future work.These line maps were generated using the Keplerian masks generated in the CLEANing with no clipping thresholds.All of the molecules aside from 12 CO and C 17 O only show significant emission in the south of the disk -the same region of the disk as the millimetre dust trap.In the north of the disk the 12 CO emission suffers from cloud absorption along the minor axis of the disk but there is weak millimetre dust and C 17 O emission present here (also seen by Bruderer et al. (2014b) in the C 17 O J = 6 − 5).Previous studies have shown asymmetric emission for SO, SO 2 , NO and several of the large organics (van der Marel et al. 2021a;Booth et al. 2021a;Brunken et al. 2022).Here, we present the first detections of the simple molecules HCO + , HCN and CS, and interestingly find that they all show a similar asymmetric emission morphology.However, not all of the molecules have the exact same asymmetric morphology.

Sub-structures in the IRS 48 disk
The only molecule detected in the north of the IRS 48 disk is CO, while all of the other species are located in the south, but there are variations in where the different molecules peak both radially and azimuthally.Figure 3 shows azimuthal profiles taken from the intensity maps in Figure 2 at the radius where each of the molecules peak along with the normalised azimuthal profile of the millimetre dust.From this, it is clear that there are dips in the intensity of most species at the azimuthal peak of the dust emission.This could be due to line suppression from the optically thick dust (e.g., Weaver et al. 2018;De Simone et al. 2020) but interestingly this is not as apparent for the COMs emission.The COMs emission is also significantly narrower in azimuthal extent than the simpler molecules that are detected and are located at the dust peak with a similar width to the millimetre dust.This is highlighted further in Figure 4 which presents a polar deprojection of the intensity maps.It is clear that the SO and SO 2 peak radially further out in the disk than the CH 3 OH and H 2 CO, which was not clear in the lower spatial resolution data presented in van der Marel et al. (2021a) and Booth et al. (2021a).Furthermore, the HCO + emission is peaking closer to the star, in the gas cavity, than the CH 3 OH, and the HCN is approximately co-spatial with the H 2 CO.The possible physical and chemical explanations for these dif- ferent emission morphologies will be discussed further in Section 4.1.

Disk integrated line fluxes
In Figure 5 we show the disk-integrated fluxes for molecules detected/not detected in the IRS 48 disk compared to the HD 100546 disk, which was observed as part of the same ALMA program (Booth et al. 2024).For some molecules we detected multiple transitions but we only report the flux of a representative transition.These representative transitions are based on the strongest lines detected in the HD 100546 disk.In the case of CN, C 2 H and NO the chosen lines are the strongest of the N = 3 − 2, N = 4 − 3 and J = 7/2 − 5/2 hyper-fine groups, respectively.For SO 2 the J = 6 (4,2) −6 (3,3) is the strongest line detected and for SO the J = 7 8 −6 7 transition is the strongest.For CH 3 OH we pick the J = 7 0 −6 0 transition and for CH 3 OCHO and CH 3 OCH 3 we use the J = 3 1 − 3 0 and J = 19 − 18 transitions which are both blends of multiple transitions.These fluxes are extracted from Keplerian masks which are 2. ′′ 0 and 4. ′′ 0 in radius for the IRS 48 and HD 100546 disks, respectively.If a molecule is undetected we give the 3 σ upper limit on the flux where σ is propagated from the rms in the channel maps and the number of pixels included in the mask (e.g., Carney et al. 2019).All of the line fluxes are listed in Table 6 with their associated errors.After accounting for the different distances to the two sources (110 pc v 135 pc) HD 100546 is brighter in all of the lines aside from: SO, 34 SO, SO 2 , NO, H 13 2 CO, CH 3 OH and CH 3 OCHO.Note that the IRS 48 12 CO (J=3-2) flux is a lower-limit due to foreground cloud absorption (e.g., see Figure 2 in Bruderer et al. 2014b).
To take into account the significantly different gas masses of the HD 100546 and IRS 48 disks (with HD 100546 >100× more massive than IRS 48; see Table 1) we normalise the line fluxes with respect to the C 17 O J = 3 − 2 line.The C 17 O line is the most optically thin CO isotopologue detected in both disks and this flux should be a good proxy for the total gas content in each disk (e.g., Zhang et al. 2021).These flux ratios are shown in Figure 5 and from this, there are significant differences in the relative intensities of the different molecular lines between these two disks.There are caveats to this comparison, e.g., if lines are optically thick in one or both of the disks and/or the excitation temperatures are very different.This is, however, a good starting point for comparing the two sources.The observed line strengths of most of the simple molecules are within a factor of 3 for the two disks.The differences in the line ratios become more significant when looking at the molecules that are already brighter in IRS 48.There is a factor 15 difference for NO and CH 3 OCHO, a factor 50 difference in the SO, SO 2 and H 13 2 CO and a factor 80 difference for CH 3 OH between the two disks.The largest difference is in the 34 SO/C 17 O line ratio which is ≈130× higher in IRS 48 than HD 100546.

Column densities
We estimate column densities following the methods outlined in Loomis et al. (2018a) and in the same manner as Booth et al. (2024).For the molecules where multiple transitions are detected, e.g., CH 3 OH and SO 2 , we pick one representative transition.Future work will focus specifically on constraining the excitation conditions of these molecules individually.We compute azimuthal column density profiles for the IRS 48 disk from the profiles presented in Figure 3 and explore a range of excitation temperatures: 50, 100 and 150 K.These temperatures are motivated by the observations and modelling results from van der Marel et al. (2021a) and Leemker et al. (2023).For the non-detected molecules, we calculate an upper limit propagated from the upper limits on the disk-integrated fluxes (listed in Table 6) assuming a conservative emitting area of a 2" aperture.The resulting profiles are shown in Figure 6 and the main results are as follows: • The peak C 17 O column density is ≈ 2.5 × 10 16 cm −2 at 100 K and the line is optically thin.With the assumption of ISM isotope ratios, this is equivalent to a CO column density of ≈ 5 × 10 19 cm −2 .
In Table 3 we list the peak column density ratios of each of the molecules relative to the average CO column density across the IRS 48 disk.
• The line emission from the simple molecules HCO + , HCN and NO are all optically thin.CN and C 2 H are both undetected with the 3 σ diskaveraged upper limit of CN ≤ 2 × 10 12 cm −2 and C 2 H ≤ 5 × 10 13 cm −2 .
• The radical CS is detected with a peak column density of ≈ 10 13 cm −2 which is a factor of a few lower than the upper-limit report by Booth et al. (2021a).H 2 CS is not detected with a column density upper limit of < 2 × 10 13 cm −2 which relative to CS is not constraining when comparing to other disks.
• SO is abundant in the IRS 48 disk and therefore we use the J = 3 3 − 3 2 transition for the column density calculation.This line has the lowest Einstein coefficient of the three SO transitions detected.The other two SO lines have lower column densities due to their higher optical depths.This results in a peak column density of 5×10 15 cm −2 at 100 K.This is ≈ 4× higher than the SO 2 peak column density and results in a N(CS)/N(SO)≈ 10 −3 .When comparing the derived SO column density with the 34 SO column density, the ratio is consistent with 22, the local ISM 32 S to 34 S ratio (Wilson 1999), indicating the J = 3 3 − 3 2 line is indeed optically thin.A detailed analysis on the S isotopes detected in these data will follow in a future work.OCS is not detected with a column density upper limit of < 10 12 cm −2 , less than a few percent of the SO column density.
• Both H 2 CO and H 13 2 CO are robustly detected and we find a H 2 CO/H 13 2 CO column density ratio of ≈ 4.This is significantly lower than the expected 12 C/ 13 C of 69 (Wilson 1999) indicating optically thick H 2 CO emission or a lower isotope ratio.
• The CH 3 OH column density at peaks at ≈ 2 × 10 15 cm −2 .Using the column density derived for the main H 2 CO isotopologue results in a column density ratio of CH 3 OH/H 2 CO of 14±1 and using the H 13 2 CO and a C isotope ratio of 69 results in a ratio of 0.8±0.1.This means that if the H 2 CO is indeed optically thick the ratio of CH 3 OH to H 2 CO is ≈1.The CH 3 OH emission is still compact in these new data therefore, as discussed in Brunken et al. (2022) the emission may be optically thick and beam diluted.This will be investigated further in Temmink et al. (in prep.)along with the constraints from 13CH 3 OH which remains undetected.
• The peak abundance ratios of the COMs CH 3 OCHO, CH 3 OCH 3 and c−H 2 COCH 2 with respect to the peak CH 3 OH column density are 0.28 ± 0.04, 0.25 ± 0.03 and 0.017 ± 0.006, respectively, at a temperature of 100 K. CH 3 CHO is undetected with an upper-limit of ≈ 4 × 10 13 cm −2 .

DISCUSSION
In this section, we discuss the physical and chemical origins of the observed molecular emission in the IRS 48 disk.We place this unique source in context with another chemically well-characterised protoplanetary disk namely, the HD 100546 disk which has been observed in the same frequecny setting within the same ALMA program.Table 3. Ratios of the peak column density of different molecules (X) relative to disk averaged CO in the IRS 48 disk.

The origin of the molecular sub-structures in the IRS 48 disk
The simplest explanation for the molecular complexity and high relative column densities of oxygen-bearing volatiles in the IRS 48 disk is the sublimation of ices.With this new data, there are clear spatial offsets between the different molecules that complicate this picture.As seen in Figures 2 and 4, the COMs have the most compact emission that peaks with the dust and these are also the species with the highest binding energies.The H 2 CO and HCN emissions are roughly cospatial with a depression in both the H 2 CO and HCO + emissions where the COMs (and dust) emission is brightest.Interestingly, the SO and SO 2 emissions, which Booth et al. (2021a) proposed to originate first from the sublimation and photodissociation of H 2 O and H 2 S to OH and S respectively, are peaking radially further out in the disk compared to the CH 3 OH, the latter of which should trace the same region as the H 2 O.This may point to a different chemical origin for the SO and SO 2 .There may also be a link between gas leading and trailing in the Keplerian orbit of the dust trap.The orbital direction is highlighted in Figure 4.The dust trap in the IRS 48 disk has been proposed to be a large anticyclonic vortex (van der Marel et al. 2013b) therefore, it could be expected that there is additional radial and vertical mixing, or turbulence, and this will affect the disk chemistry.Semenov & Wiebe (2011) find that in their turbulent disk chemistry models the abundances of SO and SO 2 can increase by two orders of magnitude relative to the laminar disk due to the enhanced sublimation of ices.The interplay between the dust and line optical depth may also be influencing the observed emission structures.Therefore, a more detailed analysis of the IRS 48 line emission, including mapping the disk temperature structure, will be the focus of future work (Temmink et al. in prep).

In context with other Herbig disks
Both the IRS 48 and HD 100546 disks show rich reservoirs of complex organics and volatile sulphur that are yet to be detected in most other planet-forming disks (aside from HD 169142; Booth et al. 2023).The simplest explanation for the chemical origin of these species is via the sublimation of H 2 O and COM-rich ices.The brightness temperatures of the 12 CO in these disks are 100 K indicating the physical conditions for ice-sublimation are indeed possible (see Wölfer et al. 2023).There are significant differences between the two disks, especially when considering their mass and size.The HD 100546 disk has a gas mass 500× and dust mass 100× higher than the IRS 48 disk (see Table 1) and the HD 100546 CO disk extends to ≈600 au compared to ≈200 au for IRS 48.Given the different mass reservoirs in the disks, one may expect IRS 48 to have uniformly lower line fluxes than HD 100546 but, as shown in Figure 5, this is not the case.On a disk average level the relative fluxes of simple oxygen molecules (NO, SO, SO 2 ) and larger organics (H 2 CO, COMs) are 15 to 130 times brighter in IRS 48 than HD 100546.
The HCO + abundance in the IRS 48 disk is similarly low as found in HD 100546 and HD 142527 which, is 2 orders of magnitude lower than found in the HD 163296 and MWC 480 disks (see Table 4; Aikawa et al. 2021;Temmink et al. 2023;Booth et al. 2024, priv. comm. Temmink).This may be due to the low stellar X-ray flux of IRS 48 and/or the presence of gas-phase H 2 O (not yet detected in IRS 48 but only inferred, Leemker et al. 2023) which effectively destroys HCO + .We do not detect CN in IRS 48 and similar to HD 100546 it has a low CN/HCN ratio when compared to other disks.The low CS/SO ratio and non-detection of C 2 H is consistent with a disk C/O<1 as reported by Booth et al. (2021a).Similar to HD 100546, NO is the most abundant observed nitrogen carrier in the IRS 48 disk when compared to HCN or CN.The sulphur-bearing equivalent of H 2 CO, H 2 CS, was not detected in IRS 48.Given the high abundance of H 2 CO in IRS 48 this may be surprising but in the HD 100546 disk the H 2 CS follows the CS (as also found in MWC 480 and HD 169142; Le Gal et al. 2021;Booth et al. 2023) and not the sublimating SO in the inner disk.This indicates that the H 2 CS in disks is likely forming in the gas-phase at lower temperatures (<100 K) rather than having a significant abundance on the grains.

Contextualising the volatile sulphur reservoir in the IRS 48 disk
In IRS 48 SO, SO 2 and CS are detected but OCS and H 2 CS are not.With this family of molecules, we can compare the relative column density ratios of these species to both protostars and comets.In IRS 48 SO is the most abundant S-bearing volatile detected with peak column density ratios of SO 2 /SO of ≈26%, CS/SO of ≈0.2%, OCS/SO of <0.3% and, H 2 CS/SO of <1.0%.The SO 2 /SO ratio in IRS 48 is similar to that detected in HD 100546 where again the SO column density is higher than the SO 2 column density (Booth et al. 2024), but this is not the same as observed towards both protostars and comets.Drozdovskaya et al. (2018) compare the volatile sulphur reservoirs in the comet 67P and towards the protostar IRAS 16293-2422 B. Comparing these environments to IRS 48: SO 2 , OCS and H 2 CS are all lower in abundance relative to SO in this disk than could be expected from the sublimation of cometary ices, although the SO/SO 2 ratio from 67 P has been shown to vary in time, exceeding 1 at points (Calmonte et al. 2016).Additionally, Drozdovskaya et al. (2018) show that OCS has strong variations between these two environments with OCS/SO ≈60% in 67P and ≈560% in IRAS 16293-2422 B, where in the latter source OCS is proposed to be enhanced due to UV irradiation.For both ratios, OCS would have been detectable in our data of the IRS 48 (and HD 100546) disk.Boogert et al. (2022) find that the column density of OCS in the ices toward massive young stellar objects (MYSOs) correlates with the abundance of CH 3 OH ice.Therefore, with the detection of CH 3 OH in IRS 48 we may expect to also see OCS if these ices are dually inherited by the disk, but the binding energy of OCS (pure ice 2430 K, Ward et al. 2012) is significantly lower than than of CH 3 OH (on water ice 5000 K Ferrero et al. (2020); Minissale et al. (2022)).The median ice abundance of OCS relative to CH 3 OH towards the MYSOs target by Boogert et al. (2022) is ≈1% and in contrast for IRS 48 we find that gas-phase column density ratio of OCS/CH 3 OH<0.1%.One explanation for the lack of OCS could be that during the disk lifetime, the volatile S in the simple inherited ices is converted to more refractory compounds like S allotropes due to processing via UV irradiation (Cazaux et al. 2022).Formation of S-allotropes can also can also act to destroy OCS on the ice, with models showing that OCS + S → S 2 + CO can be an important destruction pathway for OCS ice (Laas & Caselli 2019).If S 2 is desorbed from grains it can also play an important role in gas-phase SO (and SO2) formation, via reactions with atomic O.All in all, these comparisons show that the gas-phase volatile sulphur in IRS 48 is distinct to both the gas and ice detected towards protostars and in comets.

Molecular complexity as evidence for ice processing?
The degree of molecular complexity detected in the IRS 48 disk is unique for protoplanetary disks with three ≥7 atom COMs detected -CH 3 OCHO, CH 3 OCH 3 and c−H 2 COCH 2 .c−H 2 COCH 2 is the first detection of a heterocyclic molecule in a protoplanetary disk.Heterocycles are abundant in comet 67P (Hänni et al. 2023) and more generally these rings of carbon with an oxygen are of biological importance.The peak abundance ratios of these COMs with respect to the peak CH 3 OH column density show that these COMs have abundances ≈30, 25 and 2% of CH 3 OH, respectively.Similarly, in the HD 100546 disk, in addition to CH 3 OH, CH 3 OCHO is also detected with an abundance of 70% relative to CH 3 OH.Interestingly, CH 3 OCH 3 is undetected in HD 100546 with an upper limit of ≲10% relative to CH 3 OH.The slight differences in binding energies of CH 3 OCHO and CH 3 OCH 3 are not sufficient to explain the lack of CH 3 OCH 3 in HD 100546 since they are both lower than the binding energy of CH 3 OH (Minissale et al. 2022;Ligterink & Minissale 2023).
Typically, in the warm gas around low and high-mass protostars these COMs have fractional abundances of a few percent of CH 3 OH (e.g., Manigand et al. 2020;van Gelder et al. 2020;Chen et al. 2023).The higher abundances we see in the Class II disks may simply be due to an underestimated CH 3 OH column density due to optically thick and beam diluted line emission.Deeper observations to target 13 CH 3 OH isotopologues are needed to test this.Otherwise, if these high ratios are confirmed, these results reflect a different chemistry than is traced in observations of protostars.Similarly, the abundance ratio of CH 3 OCHO to CH 3 OCH 3 has been shown to be remarkably constant across different evolutionary stages of star formation (Coletta et al. 2020;Chen et al. 2023).This ratio of ≈1 is also seen in IRS 48 but not for HD 100546 where we find a ratio ≳7.c−H 2 COCH 2 is an isomer of acetaldehyde (CH 3 CHO) and vinyl alcohol (CH 2 CHOH) both of which are undetected in our data.CH 3 CHO is typically the most abundant of these isomers by at least an order of magnitude: for example, observations of IRAS 16293-2422 find that CH 3 CHO is ≈10× more abundant than c−H 2 COCH 2 and c−H 2 COCH 2 relative to CH 3 OH is ≈0.05% (Lykke et al. 2017;Manigand et al. 2020) whereas, in IRS 48, CH 3 CHO/c−H 2 COCH 2 ≲1.
The high abundance ratios of COMs with respect to CH 3 OH that we have observed so far in Class II disks and the variation in CH 3 OCHO and CH 3 OCH 3 ratios between sources could be the result of the energetic processing of ices in disks.Over the millions of years that ices are present in disks they will be exposed to UV photons, X-ray and cosmic rays -especially if vertical mixing is prominent.These energetic processes can break apart CH 3 OH ice resulting in radicals (CH 3 O, HCO, CH 3 ) that can combine to form the more complex species CH 3 OCHO, CH 3 OCH 3 and CH 3 CHO ( Öberg et al. 2009).The specific branching ratios for these radicals will play a key role in setting the new COMs ice abundances (Laas et al. 2011;Walsh et al. 2014a).c−H 2 COCH 2 has been shown to form in the solid state via the reaction of C 2 H 4 and O where Bergner et al. (2019) find a branching ratio of 0.5 for c−H 2 COCH 2 relative to CH 3 CHO.Given the upper limit on CH 3 CHO in the IRS 48 disk this may indicate that the formation of COMs via oxygen insertion reactions is also impor-tant.Finally, there may also be a non-negligible contribution from gas-phase reactions in the inner disk where the gas is warm >100 K, UV irradiated and at a significantly higher density than in protostellar envelopes.This needs to be tested with astrochemical models which we leave to further work.Additionally, a larger sample of disks is needed to understand the spread of COMs abundances in disks and better place IRS 48 in context.Upper limits on other COMs lines covered in these data and deuterated isotopes, e.g.HDCO, in the IRS 48 disk, will be investigated in Kipfer et al. (in prep) where a further, more complete, comparison to proto-stellar environments and comets will be made.

CONCLUSION
This paper is the second in a series presenting an ALMA molecular line survey towards the disks around the Herbig Ae stars HD 100546 and IRS 48.Here we focus on the IRS 48 disk where we detect 16 different molecular species and our main results are as follows: • We report the first robust detections of H 13 2 CO, 34 SO, 33 SO and c−H 2 COCH 3 in protoplanetary disks and confirm the reported tentative detection of CH 3 OCHO from Brunken et al. (2022) and CH 3 OCH 3 is clearly seen.We also detect the simple molecules HCO + , HCN and CS in the IRS 48 disk for the first time.
• The IRS 48 disk hosts an extremely asymmetric dust trap in the south of the disk.We find that all the molecular lines detected aside from CO show emission in the same region of the disk as the dust trap, including the simple molecules HCO + , HCN and CS.
• The asymmetric molecular emissions from the different molecules are not all co-spatial.There are radial and azimuthal offsets in the peak position most clearly seen between the COMs and the SO and SO 2 .This warrants further investigation of the chemistry in turbulent vortices.
• The low relative abundance of HCO + in IRS 48 is similar to the other Herbig disks HD 100546 and HD 142527, which could reflect the star's lower Xray luminosity when compared to other sources.Similar to regions of the HD 100546 disk, the CN/HCN ratio in IRS 48 is low <1, where the lack of CN may also be due to the low C/O ratio in the IRS 48 disk gas (Leemker et al. 2023).This is distinct from the elemental make up in the other Herbig Ae disks HD 163296 and MWC 480.
• CS and HCN are the only molecules detected in the IRS 48 disk without oxygen and the low CS/SO ratio and the non-detection of C 2 H support the bulk of the gas in south of the IRS 48 disk having a C/O<1.In these data there is no evidence of an enhanced C/O>1 in the non-dust trap region of the disk.Further more, the partition of volatile S between SO, SO 2 and CS and, the non-detected OCS and H 2 CS is distinct to that measured for comets and protostars with OCS/SO <0.3%.
• IRS 48 hosts the most chemically complex disk to date and the high abundances of COMs relative to CH 3 OH when compared to protostars as well as the different relative COMs ratios may indicate processing of the inherited ices in protoplanetary disks.The apparently high column density ratios of COMs to CH 3 OH needs to be confirmed via observations of optically thin tracers of CH 3 OH.i.e., the 13 C isotopologues.
Our results solidify the IRS 48 disk as a unique astrochemical laboratory to study the full volatile reservoir available during planet formation and show the benefits of large unbiased surveys of protoplanetary disks.The clear association of the molecular emissions with the dust trap shows a strong coupling between the dust and ice chemistry.Nine different molecules have been detected for the first time in the IRS 48 disk in only two ALMA observing programs (2017ALMA observing programs ( .1.00834.S, 2021.1.00738) .1.00738)with just ≈10 hours of on-source time.The efficiency of these types of observations will improve dramatically with the planned Wideband Sensitivity Upgrade for ALMA which will increase both the simultaneously observable bandwidth and the imaging speed (Carpenter et al. 2023).
B. MOLECULAR DATA Table 5.Molecular data of the transitions presented in this paper.This covers all of the molecules detected in the disk and particular nondetections of interest but not all of the transitions covered/detected.All data are taken from CDMS except for C 17 O, C 2 H, CH 3 OCHO and CH 3 OCH 3 which are from JPL: (Endres et al. 2016;Pickett et al. 1998).

Molecule
Transition Frequency (GHz) Eup (K) log10(A ul ) gu Detection observed in the ALMA program 2021.1.00738.S (PI. A. S. Booth) and the general properties of the IRS 48 system are listed in Table 1.

Figure 1 .
Figure 1.Matched filter responses for the IRS 48 and HD 100546 (taken from Booth et al. 2024) disks showing the full frequency coverage of the observations and highlighting the main molecules detected in each disk.Note that the HD 100546 response has been inverted and the lines reaching the top and bottom of the y-axis have responses >50σ.The molecule labels in the top and bottom of the plot indicate from which disk the line is more strongly detected and the vertical grey lines show the location of particular molecular transitions in more crowded regions of the spectrum.Both matched filter responses were generated using Keplerian models with an outer radius of 150 au for IRS 48 and 300 au for HD 100546.A fully annotated version of the IRS 48 response is shown in Figure 7.

Figure 2 .
Figure 2. Integrated intensity maps of the 0.9 mm dust continuum emission and molecular line emission from the IRS 48 disk.The continuum map is shown on a color log scale to highlight the weak millimetre emission in the north of the disk.The beam is shown in the left-hand corner of each panel.

Figure 3 .
Figure 3. Azimuthal line emission profiles for the IRS 48 disk generated from the maps presented in Figure 2. The dashed lines show the millimeter dust emission normalised to the peak of the line emission in each panel.

Figure 5 .
Figure 5. Top: Disk integrated fluxes for the molecules detected in the IRS 48 (purple) and HD 100546 (orange) disks.Bottom: Disk integrated fluxes relative to the C 17 O J = 3 − 2 line flux for each disk.Vertical lines and numbers show the relative differences in the different line ratios between each disk where this value is >10×.Triangles pointing down are 3 σ upper limits and triangles pointing up are lower limits.For most of the lines, the ±1 σ error bars are smaller than the plot markers.

Figure 6 .
Figure 6.Azimuthal column density profiles for the IRS 48 disk determined at a range of assumed excitation temperatures.

Table 1 .
Properties of the IRS 48 and HD 100546 star and disk systems.

Table 2 .
Booth et al. (2024)(✓) and not detected (-) in the ALMA observations of the IRS 48 and HD 100546 disks presented in this paper andBooth et al. (2024).
Brunken et al. (2022)gain, as reported byBrunken et al. (2022), and their weak detection of methyl formate (CH 3 OCHO) is clearly confirmed in our data.A investigation into other COMs lines covered in these data and upper-limits on other non-detections will follow inKipfer et al. (in prep).A summary of the molecules detected/not-detected in both the IRS 48 and HD 100546 disks are shown in Table2.It is unclear from visual inspection of the data if H 13 CN is detected in IRS 48 or not as this line is blended with a strong SO 2 line.Using matched filtering and the HCN as a mask we find that HC 15 N, CN and C 2 H are all not detected.