Dust and Volatiles in the Disintegrating Comet C/2019 Y4 (ATLAS)

C/2019 Y4 (ATLAS) is an Oort cloud comet with an orbital period of ∼5895 yr. Starting in 2020 March, its nucleus underwent disintegration. In order to investigate the gas and dust properties of C/2019 Y4 (ATLAS) during its disintegration, we obtained long-slit spectra at 3600–8700 Å and BVRI multiband images with the Xinglong 2.16 m Telescope in 2020 April. Our observations revealed that C/2019 Y4 (ATLAS) exhibited strong emission bands of CN, C2, C3, and NH2, which are superimposed on a dust-scattering continuum, typical of cometary spectra in the optical. The production rates of CN, C2, and C3 derived using the Haser model and the corresponding C2/CN and C3/CN ratios suggest that C/2019 Y4 (ATLAS) is a “typical” Oort cloud comet under the A’Hearn classification, although it appears less dusty, as revealed by the Af ρ quantities. Its dust-scattering reflectivity is slightly red, with a gradient of ∼5% per 103Å. We model the reflectivity gradient in terms of porous dust and find that the red color is accounted for by porous dust.


INTRODUCTION
Cometary nuclei are the most primitive objects in the solar system and have preserved pristine materials from the presolar molecular cloud and from the early stages of the protosolar nebula.Therefore, the origin of cometary nuclei is closely linked to the origin of the solar system.The chemical and physical properties of dust and volatile materials in cometary nuclei and comae shed light on the nature of the primordial interstellar materials present during the solar system formation, and provide clues to the processes of incorporation of these materials into cometary nuclei, as well as the chemical and thermodynamic conditions in the outer solar nebula in which comets formed.
In the absence of a direct analysis of cometary nucleus materials, one often resorts to cometary comae formed by volatiles outgassed from the nucleus and dust particles dragged out by the expanding gas.Once formed and stored in the Oort cloud or the Kuiper belt, however, comets are exposed for ∼ 4.5 billion years to the flux of Galactic cosmic rays and may develop a substantial "crust" of nonvolatile materials (e.g., see Strazzulla 1999).For short-period comets, such a crust could also result from the solar radiation during their numerous close approaches to the Sun (e.g., see Li & Greenberg 1998a).In this scenario, only the sub-surface materials still remain pristine.Therefore, disintegrating comets provide us an unique opportunity to gain access to the pristine, sub-surface materials.
Comet C/2019 Y4 (ATLAS) (hereafter, 19Y4), discovered by the Asteroid Terrestrial-impact Last Alert System (ATLAS) at the Mauna Loa Observatory on 2019 December 28, is a long-period comet.According to the JPL orbital solution #17, it takes 19Y4 5895 ± 23 yr to orbit the Sun in an elliptical trajectory of eccentricity e = 0.999, inclination i = 45.4 • , and perihelion distance q = 0.253 AU.1 Its initial light curve exhibited such a steep brightening that, if continued, 19Y4 would have been visible to naked eyes when approaching its perihelion in 2020 late May.2However, the steep brightening terminated in 2020 mid-March, when 19Y4 started to disintegrate, as revealed by the nongravitational effect and a blue color likely caused by the release of a large amount of volatiles (Hui & Ye 2020).The polarization increased dramatically in 2020 late-March as well, indicating a large increase in the amount of carbonaceous material (Zubko et al. 2020).On 2020 April 6, Ye & Zhang (2020) and Steele et al. (2020) successively reported an elongated nucleus.On 2020 April 13, the nucleus of 19Y4 had already split into at least five condensations (Sekanina et al. 2020).Optical spectroscopy on 2020 April 14 and 16 revealed emission bands of CN, C 2 , C 3 , and NH 2 .The C 2 /CN and C 3 /CN productionrate ratios suggested a "typical" Oort cloud comet under the A'Hearn classification (Ivanova et al. 2021).Continuing disintegration was reported by Ye & Hui (2020), and the evolution of the fragment clusters was investigated in detail through high angular resolution images obtained with the Hubble Space Telescope (HST) on 2020 April 20 and 23 (Ye et al. 2021).
To explore the properties of the pristine volatiles and dust beneath the refractory crust, we have performed long-slit spectroscopic and broad-band photometric observations of 19Y4 during its disintegration.This paper reports and analyzes the spectroscopic and photometric data and is organized as follows.The observation and data reduction are described in §2 and the results are reported in §3.The gas and dust productions are discussed in §4.1 and §4.2, respectively.The dust-scattered light is modeled in §4.3 to infer the properties of the dust.We summarize our major conclusions in §5.

Long-slit Spectroscopy
We obtained long-slit optical spectra of 19Y4 with the Beijing Faint Object Spectrograph and Camera (BFOSC) on board the 2.16-Meter Telescope at Xinglong Station.BFOSC has a 2048×2048 pixel 2 CCD installed, subtending a 9.36 ′ × 9.36 ′ field of view (FOV).The pixel scale is thus ∼ 0.274 ′′ and no binning was performed.The slit length of 9.4 ′ was fixed along the north-south orientation.We note that the north-south orientation is parallactic only when the target is on the meridian.So our spectra are somewhat subjected to atmospheric dispersion (see §3).A slit width of 1.8 ′′ or 2.3 ′′ was chosen, depending on the seeing at the time of observation.The G4 grism, covering a wavelength range of 3600-8700 Å, was used.It has a resolving power of ∼265 at [O iii] λ4959 Å and λ5007 Å, if combined with a slit width of 2.3 ′′ (Fan et al. 2016).
We performed spectroscopic observations of 19Y4 in four nights: 2020 April 6. 51,13.53,20.53,and 23.58 (UT).Although the disintegration of the nucleus had already started in mid-March 2020 (Hui & Ye 2020), no fragment was resolved in our observation on April 6.51 (UT), and the slit was centered on the brightest pixels of the elongated nucleus.On April 13.53 (UT), the slit was centered on the fragment 19Y4-A (see Figure 1a).On April 20.53 (UT), fragments 19Y4-A and 19Y4-B were resolved and they roughly aligned along the north-south orientation (see Figure 1b), therefore, the slit was over both of them.On April 23.58 (UT), the slit was centered on the brightest fragment 19Y4-B.Non-sidereal tracking was used.We alternated between short exposures of 600 s and slit view inspections in order to ensure the guiding accuracy.The iron-argon arc was used as the wavelength calibrator.In addition, we have also obtained the long-slit spectra of several flux standards: Feige 34, He 3, HR 4554, HD 109995, HZ 44 and Feige 98.The spectroscopy log is shown in Table 1.Also tabulated in Table 1 are the kinematic parameters of 19Y4, including its phase angle (ϕ), heliocentric (r h ), and geocentric distances (∆) as well as their temporal changing rates ṙh and ∆, obtained from the JPL Horizons On-Line Ephemeris System web-interface. 3ollowing the general data reduction procedures, we reduced the long-slit spectroscopic data with Python.First, we took bias subtraction, corrected for flat fields, and removed cosmic rays.Then, we corrected for the geometric curvature by fitting the curved arc lines with the B-spline functions.Finally, we combined multiple short exposures in each night, and calibrated their wavelengths and fluxes to derive fully reduced 2-D spectra.

Broadband Photometry
BFOSC was also used to perform broadband photometry over 19Y4 on 2020 April 23.55 (UT).The CCD configuration is the same as that used in the long-slit spectroscopy (see §2.1).We made three exposures of 300 s each, with the BVRI filters in the Johnson-Cousins system.Again, non-sidereal tracking was used.The photometry log is also shown in Table 1.
We developed a Python pipeline to reduce the photometric data.First, we corrected all the science frames for bias, flat field and cosmic ray pixels.Then, we smoothed the corrected frames to detect and center the peaks of the trailed field stars, which were flattened under non-sidereal tracking.We performed photometry at each centroid on the corrected frames with rectangular aperture and annulus to obtain instrumental magnitudes.Next, we used the centroids of the star trails to derive plate solutions via Astrometry.net(Lang et al. 2010), and the instrumental magnitudes to derive flux zeropoints by cross-match with the Sloan Digital Sky Survey Data Release 16 catalog (Ahumada et al. 2022).The BVRI fluxes and errors of the field stars were constructed from the Sloan gri fluxes and errors according to the transformation relations given in Chonis & Gaskell (2008).The flux calibrated frames were finally combined according to their filters.

RESULTS
We adopt two apertures, different in width, to extract 1-D spectra: the narrow one, with a single aperture, has a width of ∼ 27.5 ′′ (corresponding to a physical scale of ∼ 2.4 × 10 4 km); the broad one, divided into 40 sub-apertures, has a width ∼ 165 ′′ in total (corresponding to a physical scale of ∼ 1.2 × 10 5 km; see Figure 1).The narrow aperture is chosen to maximize the signal-to-noise ratio in the dust continuum bands, and the extracted spectra are used to determine the reflectivity gradients (see below).We show in Figure 2 the narrow aperture spectra of 19Y4.The spectra are composed of emission bands from gases and sunlight scattered by dust grains.With the help of an emission line catalog (Brown et al. 1996), we identify the ∆ν = 0 bands of CN, C 2 , and C 3 , the ∆ν = +1 band of C 2 , and a series of bands of NH 2 .The sub-aperture spectra are used to both estimate the influence of atmospheric dispersion on the reflectivity gradients (see below) and model the gas and dust production (see §4.1 and §4.2).
Figure 3 shows the BVRI images of 19Y4 obtained with BFOSC on 2020 April 23.55 (UT).As the band goes from B to I, the coma of 19Y4 shows a less-extended morphology, indicating the broadband colors vary from inner to outer coma.To characterize the color variation, we derive (B − V ) and (V − R) profiles of the coma by performing photometry with a slit-like aperture along the north-south orientation.The aperture is divided into a series of sub-apertures, 4.1 ′′ in height and 2.3 ′′ in width (see Figure 3), which are the same as that used to extract sub-aperture spectra.The color profiles are shown in Figure 4.In general, the (B − V ) profile is slightly redder and the (V − R) profile is significantly bluer than the solar colors of (B − V ) ⊙ = 0.653 ± 0.005 and (V − R) ⊙ = 0.352 ± 0.007 (Ramírez et al. 2012).More specifically, the (B − V ) profile has a local minimum of 0.68 near the nucleus and becomes redder outwards.After reaching ∼ 0.9 at ∼ 30 ′′ from the nucleus, it starts to turn blue.The (V − R) profile has a maximum of −0.07 near the nucleus and becomes bluer outwards.As the broad BVR bands cover multiple gas emission lines and dust continuum, such color variation should not be simply attributed to the intrinsic changes in the nature of gas or dust.Instead, it may merely reflects the spread of gases and dust in the coma.To be more quantitative, we estimate from the spectra that the gas emission bands contribute ∼ 50% and ∼ 25% of the VR fluxes, respectively.This suggests that the blue color of (V − R) does not necessarily indicate a blue dust continuum of 19Y4.
To characterize the dust color, we define seven dust continuum bands free from gas emission and telluric absorption: 5225-5255 Å, 5800-5850 Å, 6410-6450 Å, 6800-6850 Å, 7420-7450 Å, 7540-7580 Å, 7795-7821 Å, and 8000-8050 Å, and derive the reflectivity gradient, following Jewitt & Meech (1986).Observationally, the reflectivity is defined as where F sca λ (r h ) is the dust scattered continuum of 19Y4 at a heliocentric distance of r h ; F ⊙ λ (1 AU) is the solar flux at r h = 1 AU.Here we use the solar reference spectrum from CALSPEC (Bohlin et al. 2014).Since the spectra obtained at different epochs have different dust continuum levels, the reflectivity is further normalized to ⟨S obs ⟩, the mean reflectivity in the observed wavelength range.We fit S obs λ /⟨S obs ⟩ with linear functions and determine the reflectivity gradient, defined as the slope of the best-fit.
As shown in Figure 5, the reflectivity gradients are negative at wavelengths 5000-8200 Å for all the four epochs, indicating that 19Y4 is significantly bluer than the Sun for which, by definition, the reflectivity gradient is 0% per 10 3 Å.It is also bluer than typical comets for which the reflectivity gradient is ∼ 5-18% per 10 3 Å at wavelengths 3500-6500 Å (see Jewitt & Meech 1986).We note that Hui & Ye (2020) also found that 19Y4 at one point had a color that was much bluer than the Sun.However, as we mentioned in §2.1, the spectra of 19Y4 were affected by atmospheric dispersion and so was the color.Therefore, correction is needed to recover the true reflectivity gradient.
To estimate the influence of atmospheric dispersion on the reflectivity gradient, we further derive spectroscopic (B − V ) and (V − R) profiles on April 23.58 (UT) by convolving the sub-aperture spectra with the the transmission functions of the BVR filters, and compare them in Figure 4 with the photometric color profiles on the same date.If the spectra were free from atmospheric dispersion, there should not be any difference between the spectroscopic and photometric color profiles, as the apertures used to derive the color profiles are the same.Nevertheless, as shown in Figure 4, the spectroscopic (B − V ) and (V − R) colors are systematically bluer, indicating the influence of the atmospheric dispersion on the colors.To the first-order approximation, we assume that the color deviation due to the atmospheric dispersion varies linearly with the wavelength, and determine a linear correction function by minimizing the systematic differences in the color profiles.We note that such a correction is only applied to the spectrum obtained on 2020 April 23.58 (UT) and, as shown in Figure 5, the recovered reflectivity gradient is ∼ 5% per 10 3 Å at wavelengths 5000-8200 Å.The results are shown in Figure 5 and will be used to constrain the dust properties in §4.3.

Gas Production
The gas production rate and scale length of a given volatile species are derived by modeling the "observed" column density profile, which, for species j at a projected distance ρ, is where I j (ρ) is the continuum-subtracted radiance of the species measured at ρ and integrated over a bandpass, and g j is the fluorescence efficiency of the species averaged over the band (Langland-Shula & Smith 2011).We average N obs j over the two pairs of the sub-apertures symmetric about the nucleus.In this work, we derive the "observed" column density profiles of CN, C 3 and C 2 through their ∆ν = 0 bands.The bandpasses and fluorescence efficiencies are taken from Langland-Shula & Smith (2011) except for CN (∆ν = 0), the fluorescence efficiency of Schleicher ( 2010) is adopted for which the Swings effect (Swings 1941) is taken into account.We show in Figure 6 the "observed" column density profiles of CN, C 3 , and C 2 .
In the literature, cometary gas production rates, scale lengths and their dependencies on r h have been carefully investigated in numerous statistical works (e.g., see A' Hearn et al. 1995;Fink & Hicks 1996;Langland-Shula & Smith 2011).We apply those results as prior knowledge for Bayesian inference studies.To this end, we assume a log-uniform prior for the gas production rate Q and normal priors for both primary (l 0 ) and product (l 1 ) scale lengths, i.e., log Q ∼ U (22, 28), l 0 ∼ N ( l0 , σ2 l 0 ) and l 1 ∼ N ( l1 , σ2 l 1 ), where l0 , σl 0 and l1 , σl 1 are taken from Langland-Shula & Smith (2011).Here U and N are probability functions.With log Q ∼ U (22, 28), we assume that the production rate Q is uniformly distributed between 10 22 s −1 and 10 28 s −1 in the log space.As we do not have any knowledge about Q, we should assume such a non-informative prior.l 0 and l 1 are scale lengths for Zhao et al.
both of which we assume Gaussian priors, as their values have been well investigated in previous statistical studies.The Gaussian parameters l0 , σl 0 and l1 , σl 1 (i.e., mean and standard deviation) are taken from Langland-Shula & Smith (2011).The use of Bayesian inference is justified later in this section.The corresponding prior probability density function is denoted as p (Q, l 0 , l 1 ).From Bayes' theorem, the posterior density is proportional to the prior multiplied by a likelihood function, i.e., where σ N obs is the uncertainty of N obs .If σ N obs is assumed to be Gaussian, the likelihood function can be defined as where N mod i is the modeled column density in the i-th sub-aperture.With a set of (Q, l 0 , l 1 ) given, N mod i can be derived from eqs. ( 11)-( 13) in Langland-Shula & Smith (2011).The joint posterior distributions are then sampled using the Markov chain Monte Carlo (MCMC) method.The 50% (median) and 50 ± 34% percentiles in the marginalized distributions of Q, l 0 and l 1 are taken as the best-fit values and the uncertainties, respectively.Note that, in the fitting, N obs i with ρ i < 5 × 10 3 km are excluded, considering that the gas distribution in the inner coma may not be spherically symmetric due to the disintegration and also that "holes" may be present in the profiles of C 2 (see Langland-Shula & Smith 2011).
The best-fits to the "observed" column density profiles are shown in Figure 6, with the best-fit model parameters and uncertainties listed in Table 2.The fits to C 3 (∆ν = 0) on 2020 April 6.51 and 23.58 (UT) fail due to the low signal-to-noise ratio of the observational data, and are thus not shown.Typically, to simultaneously determine both l 0 and l 1 , the "observed" profiles are required to extend further than l 1 (Fray et al. 2005).However, this is not the case here.If the widely adopted χ 2 -minimizing method is used, several (l 0 , l 1 ) pairs could lead to equally good fits (Fray et al. 2005;Langland-Shula & Smith 2011).Therefore, it is the Bayesian model or, more specifically, the prior that helps to find unique best-fits.
Since G4 does not cover any strong OH bands, we take Q(CN) as an indicator of the gas production rate of 19Y4.We show in Figure 7a the dependency of Q(CN) on r h .In general, the gas production rate in 19Y4 is "typical" in the sense that both the production rate and its dependency on r h agree with that of the 26 comets reported in Langland-Shula & Smith (2011).
We estimate a mean ratio of log ] follow a power-law dependence on r h and take the power-law slopes of Langland-Shula & Smith (2011).The C 2 /CN and C 3 /CN production rate ratios estimated for 19Y4 suggest a "typical" comet instead of a "depleted" one under the classification of A' Hearn et al. (1995), consistent with that found by Ivanova et al. (2021).

Dust Production
Initially introduced in A' Hearn et al. (1984), the so-called Af ρ quantities are often used to describe cometary dust production, where A is the dust reflectivity (i.e., albedo), f is the dust filling factor and ρ is the projected linear radius of the FOV at the comet as discussed in §4.1.The Af ρ quantity within a circular aperture with a projected radius of ρ is determined directly from observation as where F c is the cometary flux in a chosen dust continuum band, F ⊙ (1 AU) is the solar flux at r h = 1 AU in the same band, and Φ HM (ϕ) is the normalized Halley-Marcus phase function (Schleicher & Bair 2011).In most cases, F c is obtained from narrow-band photometry.But it can also be constructed from long-slit spectroscopy by considering a geometric correction function where ρ i is the projection distance from the i-th sub-aperture to the comet nucleus, and G(ρ i ) is the geometric correction function (Langland-Shula & Smith 2011).The summation is over all the sub-apertures within ρ (i.e., ρ i < ρ).In this work, we calculate Af ρ from the constructed F c in the λ4850 Å and λ5240 Å continuum bands refined in Langland-Shula & Smith ( 2011) with an aperture radius of 10000 km.The results are tabulated in Table 3.We should admit that our method of reconstructing Af ρ from long-slit spectra which only sample a strip along certain part of the coma has a drawback of not accounting for the coma's asymmetry, even more so for a disintegrating comet.However, detailed analysis of dust production during disintegration requires high resolution, spatially resolved photometry and detailed modeling, which is beyond the scope of this work.Nevertheless, we note that our Af ρ values closely agree with that from narrow-band photometry measured with TRAPPIST (Y.Moulane, private conversation).
We obtain a mean Af ρ for 19Y4 by averaging over that in the λ4850 Å and λ5240 Å continuum bands and examine the dependency of Af ρ on r h in Figure 7b.For comparison, we also show in Figure 7b the Af ρ quantities of the 26 comets studied in Langland-Shula & Smith (2011).While 19Y4 appears to be less dusty, it exhibits a r h -dependence more or less resembling that of the Langland-Shula & Smith (2011) samples for which a power-law with an index 5.3±0.6 is suggested, even though most comets in the sample of Langland-Shula & Smith (2011) did not disintegrate when observed.

Scattered Sunlight
The dust-scattered light observationally determined in §3 can be used to infer the dust size, compositonal, and structural properties.To reproduce the normalized reflectivity S obs λ /⟨S obs ⟩ (see §3 and Figure 8), we adopt the cometary dust model of Li & Greenberg (1998b), which assumes cometary dust grains to be porous aggregates of small astronomical silicates, amorphous carbon, and vacuum.We consider two cases for the porosity (i.e., the fractional volumn of vacuum): f vac = 0.9 and f vac = 0.5, which are roughly the upper and lower limits for cometary dust grains (see Greenberg & Hage 1990;Greenberg 1998).For porous dust resulting from coagulational growth, one would expect a porosity higher than f vac = 0.5 (see Blum & Wurm 2008).On the other hand, polarimetric measurements suggest that, for most of the time, amorphous silicates in the coma of 19Y4 account for a fractional volumn (f sil ) of 0.17-0.28(Zubko et al. 2020).Therefore, we assume a bulk volume mixing ratio of f carb /f sil = 3 for our dust model.The refractive indices of amorphous silicates and amorphous carbon are taken, respectively, from Draine & Lee (1984) and Rouleau & Martin (1991).
We use Mie theory combined with the Bruggman effective medium theory (Bohren & Huffman 1983) to calculate C sca (λ, a) and g(λ, a), the scattering cross sections and asymmetry factors of spherical porous aggregates of radii a at wavelength λ.Here the aggregate size a refers to the radius of the sphere encompassing the entire aggregate (we assume that all grains are spherical in shape).We assume a power-law dust size distribution, dn/da ∝ a −α , over the size range a min ≤ a ≤ a max .The model reflectivity, S mod (λ), is calculated as where θ, defined as π − ϕ, is the scattering angle, and Φ HG [g(λ, a); θ] is the Henyey-Greenstein phase function (Henyey & Greenstein 1941).Note that the right-hand side differs from the absolute reflectivity by some scaling factor.However, similar to S obs (λ), S mod (λ) is normalized with respect to the averaged value over the defined continuum wavelengths to produce S mod λ /⟨S mod ⟩ during which the scaling factor is cancelled out.It is apparent that the model reflectivity S mod (λ) contains information about the dust size, composition and morphology since the scattering cross section C sca (λ, a) and the asymmetry factor g(λ, a) are dependent on the dust size a, composition (through the refractive index), and morphology (e.g., porosity).
We calculate normalized reflectivities with different size distributions.Four size ranges are considered, different in the lower cutoff, i.e., a min = 0.1, 1, 10, and 100 µm.The upper cutoff (a max ) is fixed to 1000 µm.The resulting reflectivity gradient is not sensitive to the choice of a max since the scattering in the optical by such mm-sized or larger grains is gray (e.g., see Li 2008).Given a size range, we vary α from 2.5 to 4.1, the flattest and the steepest power-law for cometary dust grains from space measurements (McDonnell et al. 1991;Lasue et al. 2009;Price et al. 2010).The parameters are tabulated in Table 4.
We compare the model reflectivities in Figure 8 with the reflectivity derived from observations in §3.Models with same a min but different α are grouped and shown as shadows of different colors.Figure 8 shows that the "observed" reflectivity gradient (i.e., 5% per 10 3 Å) can be explained by different combinations of dust parameters.For highly porous dust with f vac = 0.9, models with a min = 1 µm and a min = 10 µm are able to reproduce the "observed" reflectivity gradient.For dust with a lower porosity of f vac = 0.5, while models with a min = 10 µm result in reflectivity gradients too flat to be comparable with observed, the "observed" reflectivity gradient can be reproduced by dust with a min = 1 µm.Apparently, although the limited observational constraint does not allow us to uniquely determine the dust parameters, the porous dust model is capable of accounting for the "observed" reflectivity gradient with appropriate combination of parameters.In general, the reflectivity gradient S mod (λ) decreases as the dust becomes more porous or a min becomes smaller.The variation of S mod (λ) with α, the power index of the dust size distribution, depends on the choice of a min .For a min < 1 µm, S mod (λ) becomes bluer as α increases, due to the increasing role played by submicron-sized grains which scatter more at shorter wavelengths.By increasing a min to several microns (and larger), S mod (λ) becomes redder as α increases since these micron-sized grains scatter more effectively at longer wavelengths.With a min reaching several tens of microns, S mod (λ) becomes gray and does not vary with α since these grains are in the geometrical optics limit and their scattering does not vary much with wavelength.Apparently, a min plays a more important role than α in determining S mod (λ).

SUMMARY
We presented long-slit optical spectra at ∼ 3600-8700 Å and multi-band BV RI images of the disintegrating comet 19Y4, from which we derived the gas and dust production rates and modeled the dust-scattering reflectivity.Our principal results are as follows: 1.All of the spectra taken during the disintegration showed strong gas emission.The production rates of CN, C 2 , and C 3 and the corresponding C 2 /CN and C 3 /CN production-rate ratios were derived.Both the gas production rates and their dependencies on r h are consistent with that of "typical" Oort cloud comets, not that of carbon-chain "depleted" Kuiper Belt comets.
The reflectivity gradient was modeled in terms of porous dust.The red color is accounted for by porous dust.13.53 (b),20.53 (c),and 23.58 (UT;d).The gradients are determined by fitting those in the dust continuum bands (black open squares with error bars) with linear functions (red lines with light red shadows representing the fitting uncertainties).As the spectra of C/2019 Y4 (ATLAS) are somewhat subjected to atmospheric dispersion, the gradients need correction.To the first-order approximation, the reflectivity gradient on April 23.58 (UT) is corrected to ∼ 5% per 10 3 Å (blue line; see §3 for details).All quantities are normalized to their mean values in the observed wavelength range.Ye, Q., Jewitt, D., Hui, M.-T., et al. 2021, AJ, 162, 70, doi: 10.3847/1538-3881/abfec3 Zubko, E., Zheltobryukhov, M., Chornaya, E., et al. 2020, MNRAS, 497, 1536, doi: 10. cases are considered for the porosity (i.e., the fractional volumn of vacuum): f vac = 0.9 (left panel) and f vac = 0.5 (right panel).The volume mixing ratio of amorphous silicate and amorphous carbon is fixed to f carb /f sil = 3.For the size distribution, four size ranges are considered, different in the lower cutoff, i.e., a min = 0.1 (gray shadow), 1 (blue shadow), 10 (red shadow), and 100 µm (yellow shadow).The upper cutoff (a max ) is fixed to 1000 µm.Given a size range, the power-law index α is varied from 2.5 to 4.1.The recovered reflectivity gradient (∼ 5% per 10 3 Å) is barely explained by both 1 ≤ a ≤ 1000 µm and 10 ≤ a ≤ 1000 µm when f vac = 0.9, while well explained by the size range 1 ≤ a ≤ 1000 µm when f vac = 0.5.For both cases, larger or smaller a min leads to flatter gradient (or "gray" in color).

fFigure 1 .
Figure 1.Slit views of C/2019 Y4 (ATLAS) on 2020 April 13.53 and 20.53 (UT).The ephemeris positions of the two major fragments A and B (from MPEC 2020-L06 and 2023-J29, respectively) are marked.On April 13.53 (UT), the slit was centered on A. On April 20.53 (UT), the slit was over both A and B. The rectangles, 4.1 ′′ in height and 2.3 ′′ in width, along the slits are sub-apertures.They were used to extract sub-aperture spectra (see §3).The arrows in the upper right corner mark the north and the east directions, and those in the lower right corners mark the velocity and comet-Sun vectors.

Figure 5 .
Figure 5.The "observed" reflectivities (S obs λ /⟨S obs ⟩; gray lines) of C/2019 Y4 (ATLAS) on 2020 April 6.51 (a),13.53(b), 20.53 (c), and 23.58 (UT; d).The gradients are determined by fitting those in the dust continuum bands (black open squares with error bars) with linear functions (red lines with light red shadows representing the fitting uncertainties).As the spectra of C/2019 Y4 (ATLAS) are somewhat subjected to atmospheric dispersion, the gradients need correction.To the first-order approximation, the reflectivity gradient on April 23.58 (UT) is corrected to ∼ 5% per 10 3 Å (blue line; see §3 for details).All quantities are normalized to their mean values in the observed wavelength range.

Figure 7 .Figure 8 .
Figure 7. Top panel (a): Variation of the CN production rate Q(CN) of C/2019 Y4 (ATLAS) (yellow circles) with heliocentric distance r h .Also shown are the CN production rates of the Langland-Shula & Smith (2011) sample of 26 comets (black circles).Middle panel (b): Same as (a) but for Af ρ.Also shown is a linear fit to the r h -dependence of Af ρ for the Langland-Shula & Smith (2011) sample: d(log Af ρ)/d(r h ) = 5.3 ± 0.6 (black dashed line).Bottom panel: Same as (a) but for Af ρ/Q(CN).The power-law has an index of 4.6 ± 0.8.

Table 2 .
Gas production rates and scale lengths.a