PHANGS–JWST First Results: Massive Young Star Clusters and New Insights from JWST Observations of NGC 1365

In order to build a complete census of massive, young clusters in NGC 1365, we start by identifying all optically selected clusters that have estimated ages younger than 10 Myr and masses higher than 10⁶ M_⊙, from the cluster catalog and age-dating analysis based on multiband HST optical images presented in Turner et al. (2021) and Whitmore et al. (2023). Three clusters are removed from the original list since they have very faint F2100W magnitudes indicative of old populations, unlike the rest of the young clusters in our final sample that are all very bright in F2100W. We find 19 clusters that satisfy our criteria (these are the 19 clusters in Table 1 that have Cluster IDs in column (2)).

Table 1. Census of Massive Young Star Clusters in NGC 1365

ID	Cl ID a	Other ID	R.A.	Decl.	Log Age b	Log Mass c	Av d	CI200 e	Class f	Detected g	Detected h
			(J2000)	(J2000)	(H, H-JW)	(H, JW)	(H, JW)			B, V, I	PAHs
			(deg)	(deg)	(Myr)	(M⊙)	(mag)	(mag)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
1	⋯	⋯	53.400281	−36.142577	⋯, 6.5	⋯, 6.2	⋯, 12.5	2.05	3	⋯, ⋯, I	3.3, 7.7, 11.3
2	996	⋯	53.399891	−36.142575	6.0, 6.0	7.1, 6.6	4.6, 4.4	1.83	2	B, V, I	3.3, 7.7, 11.3
3	1051	J i	53.400691	−36.142127	6.0, 6.0	7.2, 6.9	4.6, 5.4	1.70	2	B, V, I	3.3, 7.7, 11.3
4	1064	⋯	53.399386	−36.142085	6.6, 6.6	6.1 6.2	3.1, 4.3	1.69	2	B, V, I	3.3, 7.7, 11.3
5	⋯	⋯	53.400338	−36.142072	⋯, 6.5	⋯, 6.00	⋯, 4.9	1.64	2	B, V, I	3.3, 7.7, 11.3
6	⋯	⋯	53.399056	−36.141991	⋯, 6.5	⋯, 5.8	⋯,⋯	1.53	3	⋯, ⋯,⋯	3.3, 7.7, 11.3
7	1116	M2 j , SSC3 k	53.399870	−36.141833	6.5, 6.5	6.9, 6.6	1.7, 1.0	1.45	1	B, V, I	3.3, 7.7, 11.3
8	1141	⋯	53.401402	−36.141753	6.7, 6.7	6.6, 6.5	5.4, 4.6	2.05	2	B, V, I	3.3, 7.7, 11.3
9	1251	⋯	53.400184	−36.141315	6.0, 6.0	6.9, 6.6	4.6, 6.3	1.88	2	B, V, I	3.3, 7.7, 11.3
10	1259	⋯	53.402834	−36.141274	6.7, 6.7	6.0, 6.2	4.9, 5.8	1.59	2	B, V, I	3.3, ⋯,⋯
11	⋯	⋯	53.399771	−36.141127	⋯, 6.5	⋯. 6.5	⋯, 4.5	1.66	2	B, V, I	3.3, 7.7, 11.3
12	1296	M3 j , SSC6 k	53.399552	−36.141121	6.5, 6.5	6.6, 6.6	1.2, 2.0	1.62	1	B, V, I	3.3, 7.7, 11.3
13	⋯	⋯	53.402935	−36.140825	⋯, 6.5	⋯, 5.9	⋯, ⋯	1.39	3	⋯, ⋯,⋯	3.3, 7.7,⋯
14	⋯	⋯	53.402318	−36.140610	⋯, 6.5	⋯, 5.8	⋯, 3.8	2.69	1	B, V, I	⋯, 7.7, 11.3
15	⋯	⋯	53.403212	−36.140610	⋯, 6.5	⋯, 6.3	⋯, 12.2	1.52	3	⋯, ⋯, I	3.3, 7.7,⋯
16	⋯	⋯	53.402911	−36.140455	⋯, 6.5	⋯, 5.8	⋯, 8.0	1.73	2	B, V, I	3.3, 7.7, 11.3
17	⋯	⋯	53.402548	−36.140381	⋯, 6.5	⋯, 6.0	⋯,⋯	1.79	3	⋯, ⋯,⋯	3.3, 7.7, 11.3
18	1506	⋯	53.403394	−36.140211	6.0, 6.0	6.4 6.2	4.6, 6.5	1.76	2	B, V, I	3.3, ⋯,⋯
19	⋯	⋯	53.399524	−36.140090	⋯, 6.5	⋯, 6.5	⋯, 8.4	1.89	3	⋯, V, I	3.3, 7.7, 11.3
20	1528	⋯	53.403379	−36.140083	6.5, 6.5	6.2, 6.1	4.6, 6.1	1.74	2	B, V, I	3.3, 7.7,⋯
21	⋯	⋯	53.403331	−36.139860	⋯, 6.5	⋯, 6.2	⋯, 6.9	1.52	2	B, V, I	3.3, 7.7, 11.3
22	⋯	⋯	53.400461	−36.139849	⋯, 6.5	⋯, 6.7	⋯, 7.4	1.72	2	B, V, I	3.3, 7.7, 11.3
23	⋯-	⋯	53.403374	−36.139568	⋯, 6.5	⋯, 6.3	⋯, 5.6	1.74	2	B, V, I	3.3, 7.7,⋯
24	1640	⋯	53.399837	−36.139432	6.0, 6.0	6.3, 5.8	4.6, 2.4	2.03	1	B, V, I	3.3, ⋯,⋯
25	⋯	⋯	53.400928	−36.139100	⋯, 6.5	⋯, 6.7	⋯,⋯	1.84	3	⋯, ⋯, I	3.3, ⋯, 11.3
26	1700	C1 i	53.404026	−36.139017	6.7, 6.7	6.9, 6.7	4.6, 2.9	1.75	1	B, V, I	3.3, 7.7, 11.3
27	⋯	⋯	53.402918	−36.138860	⋯, 6.5	⋯, 6.6	⋯, 9.5	1.79	3	⋯, ⋯, I	3.3, ⋯,⋯
28	⋯	M6 j , G i	53.403181	−36.138475	⋯, 6.5	⋯, 6.6	⋯, 19.3	1.85	3	⋯, ⋯, I	3.3, SAT, 11.3
29	⋯	M4 j , D i	53.401702	−36.138433	⋯, 6.5	⋯, 6.0	⋯, ⋯	1.62	3	⋯, ⋯,⋯	3.3, SAT, 11.3
30	1782	⋯	53.400982	−36.138416	6.5, 6.5	6.2, 6.3	4.9, 7.5	1.55	2	B, V, I	3.3, 7.7, 11.3
31	1794	⋯	53.401669	−36.138367	6.6, 6.6	6.1, 6.0	4.6, 5.8	1.68	2	B, V, I	3.3, 7.7, 11.3
32	1810	⋯	53.401369	−36.138271	6.9, 6.9	6.2, 6.8	4.6, 7.7	1.97	2	⋯, V, I	3.3, ⋯, ⋯
33	1920	M5 j , E i	53.402514	−36.137680	⋯, 6.5	⋯, 6.6	⋯, 4.3	1.55	2	B, V, I	3.3, SAT , 11.3
34	1932	⋯	53.402534	−36.137596	6.7, 6.7	6.4, 5.8	5.0, 7.4	1.45	2	B, V, I	3.3, 7.7, 11.3
35	1948	⋯	53.403066	−36.137494	6.5 6.5	6.5, 5.8	6.9, 6.6	1.57	2	B, V, I	3.3, 7.7, 11.3
36	⋯	⋯	53.403115	−36.137151	⋯, 6.5	⋯, 6.3	⋯, 6.4	1.42	2	B, V, I	3.3, 7.7, 11.3
37	2242	⋯	53.407333	−36.135815	6.6, 6.5	6.2, 6.3	2.0, 2.6	1.71	1	B, V, I	3.3, 7.7, 11.3

Notes.

^a Cluster ID from the NGC 1365 compact cluster catalog, as found at https://archive.stsci.edu/hlsp/phangs-hst. ^b Log age—First value: Hubble value from Whitmore et al. (2023). Typical errors are 0.4 dex. Second value: adopted value from first value if available (Hubble), or 6.5 otherwise (JWST—typical errors are 0.6 dex). See Section 3.2.4 for discussion. ^c Log mass—First value: Hubble value from Whitmore et al. (2023). Second value: value from mean of values based on F200W formula (i.e., $\mathrm{log}\,(M/{M}_{\odot })\,=$ −0.4068 × m_F200W + 13.83) and F300M formula (i.e., $\mathrm{log}\,(M/{M}_{\odot })\,=$ −0.3639 × m_F300M + 13.25). Typical errors are 0.35 dex. See Section 3.2.3. ^d Extinction (Av)—First value: Hubble value from Whitmore et al. (2023). Second value: value from formula A_V (JWST) = [(m_I − m_F360M ) × 0.810 + 0.99] × 3.1. Typical errors are 1.0 mag. See Section 3.2.2. ^e Concentration Index (CI) using the difference in aperture magnitudes in the F200W filter using radii of 1 and 4 pixels. See Section 3.2.1. ^f Visual estimate of how embedded the cluster is. Class 1 = largely unobscured (A_V ≲ 3 mag); Class 2 = partly embedded (3 ≲ A_V ≲ 10 mag); Class 3 = embedded (A_V ≳ 10 mag). ^g Brighter than 26th mag in relevant band. ^h Brighter than 20th mag in relevant band. ⁱ Radio source ID (Sandqvist et al. 1995). ^j MID IR ID (Galliano et al. 2008). ^k SSC = super star cluster ID (Kristen et al. 1997).

Download table as:  ASCIITypeset image

To this list, we add two "historical" clusters (M4 and M6) that were previously discovered in radio continuum images (e.g., Sandqvist et al. 1982, 1995), with estimated ages less than 10 Myr and masses in the range 3 × 10⁶ to 10⁷ M_⊙ (Galliano et al. 2008, 2012). These are comparable to the most massive, recently formed clusters in the merging Antennae (Whitmore et al. 2010) and NGC 3256 (Zepf et al. 1999; Mulia et al. 2016; Adamo et al. 2020b) galaxies.

Finally, we add 16 new, partly or heavily embedded massive young clusters identified in the new JWST images. These have been selected to be brighter than 20.0 mag in at least one of the three filters that contain polycyclic aromatic hydrocarbon (PAH) emission (F335M, F770W and F1130W). As shown in Table 1, many sources meet this criterion in all three filters, but a few only qualify in one or two.

We determined the magnitude limit of 20.0 mag to be included in our sample by requiring that the 19 clusters detected from HST (with one or two exceptions) be included. As will be discussed in Section 3.2.1, we also require the F200W magnitude to be brighter than 20.0 mag to ensure there is a bright stellar continuum source present, rather than just a dense cloud of gas/dust energized by nearby sources. We also require that the concentration index (CI; defined as the magnitude difference measured in 1 and 4 pixel aperture radii) in the F200W band to be larger than that of a point source: CI_F200W > 1.4 mag.

Different cluster selection criteria are possible, and in fact several other works within the PHANGS consortium have selected clusters using a single filter: the F335M filter (Rodriguez et al. 2023), the F1000W filter (Schinnerer et al. 2023), and the F2100W filter (Hassani et al. 2023). In a future paper we will examine the impact of using different selection strategies in more detail.

Our final census of 37 massive (mass ≳10⁶ M_⊙), young (age ≲10 Myr) candidate clusters in NGC 1365 is compiled in Table 1 and shown in Figure 2. Most of the clusters are in an inner star-forming ring, as discussed by Schinnerer et al. (2023), while a few more are in the bar just to the NE of this ring (i.e., objects 33 through 37).

We find that NGC 1365, at a distance of about 20 Mpc, has the richest population of massive young clusters of any known galaxy within 30 Mpc, with about 30 clusters in this mass and age range compared with roughly a dozen in the merging Antennae galaxies (Whitmore et al. 2010), located at ∼22 Mpc. More distant galaxies with comparable populations are NGC 3256 (65 clusters, distance of 38 Mpc) and NGC 3690 (27 clusters, distance of 42 Mpc), according to Adamo et al. (2020a). Both of these are ULIRGs and ongoing merging systems, with estimated star formation rate = SFR ≈ 40 M_⊙ yr⁻¹, roughly twice as large as in NGC 1365 (Lee et al. 2022).

We are now ready to address the question: "how many massive young clusters are so enshrouded in dust that they are missing at optical wavelengths?" From column (11) of Table 1 we find only 4 out of 37 (11%) of our cluster catalog are so dusty that they are not clearly visible in the I-band image (longest wavelength observed from HST). However, age and mass estimates are quite challenging without photometry in at least three optical bands, which means clusters that are not also detected in the B and V filters are unlikely to be included in optically selected cluster catalogs, and hence might be considered missing in some cases. For NGC 1365, we find that 9 of 37 (∼24%) clusters are "missing" in the V band and 11 of 37 (∼30%) are "missing" in the B band. We therefore find that between ∼11%–30% of the clusters are "missing" from optical catalogs because they are embedded at some level. This is similar to the situation in the Antennae, where only 2/13 = 15% of bright thermal radio continuum sources were "missing" at optical wavelengths, and 38% of fainter clusters (with S > 70 μJy) were missing (Whitmore & Zhang 2002).

Figure 3 reveals a number of interesting features in Region 1. The top left panel shows the strong dust lane cutting through an optical IVB image taken with HST. Three different JWST+HST filter combinations (as indicated along the upper right corner of each panel) reveal embedded young clusters, some of which are very bright and saturated (e.g., M4, M5, and M6 in the F2100W and F770W images), as well as some blue, slightly older clusters outside the dust lane. All three of these saturated clusters are also strong radio continuum sources (corresponding to sources D, E, and G, respectively, in Sandqvist et al. 1995).

Zoom In Zoom Out Reset image size

There are also very thin, dark filaments visible in the F335M/F200W/I-band image; a few are labeled in Figure 3. These are typically found along the inside edge of some of the strongest F335M PAH ridges, and have widths of a few tens of parsecs and lengths of a few hundred parsecs, similar to the massive filaments found in the Milky Way (Hacar et al. 2022), normal spiral galaxies (e.g., Thilker et al. 2023; Meidt et al. 2023), as well as other star bursting galaxies such as the Antennae (Whitmore et al. 2014).

Figure 4 presents four diagnostic diagrams that will be used in this section to estimate several basic properties of the clusters. These include a color–magnitude diagram, a color–color diagram, and two flux versus concentration index diagrams using different bands. A visual estimate of the level of embeddedness is included in Table 1: Class 1 = largely unobscured (A_V ≲ 3), Class 2 = partly embedded (3 ≲ A_V ≲ 10), and Class 3 = embedded (A_V ≳ 10). The location of data points for three clusters is shown in each panel to help demonstrate the properties of a largely unobscured cluster (M3, shown in blue), a partially embedded cluster (obj 34, shown in orange), and an embedded cluster (M6, shown in red). Arrows show how properties track in the diagram, including PAH emission strength (youngest clusters), extinction (dustiest clusters), and size (to help separate stars from barely resolved clusters).

Zoom In Zoom Out Reset image size

We create several different training sets to illustrate a number of features. These include the 37 massive young clusters discussed in Section 3.1, now broken into three subsamples: unobscured, partly embedded, and embedded. Other training sets include older clusters (with ages around 300 Myr), a set of largely embedded F200W sources in the dust lane around M4 (see Figure 3), and objects that are likely to be red super giants (RSGs) in a stellar association near the eastern end of the bar. The color coding for the different samples is shown in the top left panel of Figure 4.

In the next subsections, we show how we separate individual stars from clusters and then describe methods to estimate the extinction (Section 3.2.2), mass (Section 3.2.3), and age (Section 3.2.4) of embedded clusters. We adopt the estimated properties of relatively unobscured clusters determined in Whitmore et al. (2023) using Hubble data, as included in Table 1, to establish correlations that allow us to estimate properties for the more deeply embedded clusters observed primarily in the JWST bands.

3.2.1. Separation of Clusters and Stars

3.2.2. Extinction Estimates for Embedded Clusters

There are two approaches we can take to estimate extinction. The first is to use the Galactic extinction law from (Fitzpatrick 1999). The I–F360M versus F300M–F335M diagram shown in the upper right panel of Figure 4 demonstrates the range of extinction (A_V ) values found in our sample. A reddening vector following a Galactic extinction law is shown for A_v = 10 mag. We find that this predicts 4.75 mag of reddening in I–F360M, hence the coefficient connecting A_v and I–F360M is 10/4.75 = 2.11. Very small deviations from exactly vertical are ignored since the horizontal dimension is primarily an indication of PAH emission strength rather than reddening. For example, both M6 and obj 34 are found in regions with strong F335M emission, explaining their positions on the right side of both the color–magnitude and color–color diagrams. We find that the majority of the clusters have A_V values less than 10 mag. The underlying assumption needed to estimate extinction using the I–F360M color index is that all the (young) clusters in our sample have similar intrinsic colors; hence, the redder color is caused by extinction rather than due to age. The fact that unobscured clusters (blue points) all have I–F360M values around 0 supports this assumption.

A second approach is to use the empirical correlation between the E(B − V) values from the spectral energy distribution (SED) -fitting based on Hubble data versus I–F360M, as shown in the bottom panel of Figure 5. We use the sample of Class 1 and 2 (symmetric and asymmetric clusters, human classified sample; see Whitmore et al. 2021) in NGC 1365 to make this determination. Turner et al. (2021) describe the basic SED age-dating method used in the PHANGS–HST pipeline. More recently, Whitmore et al. (2023) described modifications that identify and correct the age estimates for old globular clusters in the pipeline; many old globular clusters have incorrect age estimates in the default pipeline due to challenges in disentangling age/reddening/metallicity. This helps remove interlopers that might otherwise pollute our young clusters sample. The fraction of clusters with incorrect age estimates in NGC 1365 are reduced from ≈20% to ≈10% according to Whitmore et al. (2023).

Zoom In Zoom Out Reset image size

Only the log age < 6.7 clusters are included to make the sample similar to our program sample of young massive clusters in NGC 1365. The rms scatter (0.95 mag) is relatively large and the correlation coefficient (0.49) is only fair; hence, our estimates of extinction will only be approximate. The resulting formula is A_V (JWST) = [(m_I − m_F360M) × 0.810 + 0.99] × 3.1, which provides estimates of extinction for our partially embedded clusters. Hence, the coefficient connecting A_v and I–F360M is 0.810 × 3.1 = 2.51, relatively similar to the estimate from the Galactic extinction law discussed above (i.e., 2.11). This is the formula we will use in Table 1.

In principle it would be advantageous to use only JWST observations to determine the correlation, since four of our clusters do not have I-band photometry. However, we found that the smaller baseline resulted in considerably more scatter, and hence was not practical. We note that the relationship we derive is specific to the analysis presented here since aperture corrections have not been made to our data yet; other researchers will need to determine the constants from their own photometry.

Table 1 shows that the highest estimated extinction value is for the intrinsically bright object 28 (=M6/G, e.g., saturated in the F770W and F2100W filter observations, as shown in Figure 3), which has an estimate of A_V = 19.3 mag. The value would be about A_V = 15 mag if the Galactic extinction formula was used instead. This extinction is even larger than WS80, a deeply embedded star cluster in the Antennae galaxies (Whitmore & Zhang 2002) that is the strongest CO and 15 μm source in that system, and has an estimated age of 2 Myr, A_V ≈ 8 mag, and log mass ≈ 6.6. This indicates that the very different environments of a chaotic merger and the central region of a barred spiral galaxy can both produce very massive, heavily extincted, young clusters.

3.2.3. Mass Estimates for Embedded Clusters

3.2.4. Age Estimates for Embedded Clusters

In the future, we will combine HST and JWST bands to improve age estimates, especially for partially embedded clusters. For now, we will use the same basic approach as in the previous two subsections to estimate the ages of the embedded and partly embedded clusters based on HST estimates.

As shown in Figure 6, the age estimates are somewhat more complicated than extinction and mass estimates, with more scatter and a nonlinear dependence. This figure plots HST-based ages versus source brightness in the F335M filter, which includes the 3.3 μm PAH feature and also continuum emission. While there appears to be a possible correlation for ages between log age = 6 and 8.3 (e.g., correlation coefficient = 0.44), there is essentially no correlation at older ages.

Zoom In Zoom Out Reset image size

However, at the bright end (i.e., with F335M < 20.0 mag) we find that most sources are young, with log age less than 7. A visual examination of the clusters in the upper right part of the diagram show that five of the six (circled) are in regions with strong Hα emission and are likely to be young; hence, the HST-based age estimates are likely to be incorrect. An example is M5 (object 28), which is a very strong radio source that has been age dated using spectral observations indicating that it is clearly younger than 10 Myr (Galliano et al. 2012). We note that the plots for other PAH bands, notably F770W and F1130W, look very similar to that for F335M.

Based on this evidence, we believe at least 90% of the objects with F335M brighter than 20 mag are younger than 10 Myr. We note that this would be compatible with the results from Whitmore et al. (2023), who find that approximately 10% of the HST-based age estimates of clusters in the inner region of NGC 1365 are overestimated, due to the large amount of dust that makes the clusters appear red and potentially old. While low-resolution (08) Hα imaging from MUSE (Emsellem et al. 2022) has been used in Whitmore et al. (2023) to improve the age-dating to some extent, this is not as good as having high-resolution Hubble Hα imaging.

Figure 7 shows there are physical reasons why we expect most of the brightest PAH emitters to be young. This figure shows a normalized (at 1 Myr) plot of predicted flux values in various MIRI bands from the Draine models of interstellar dust emission (Draine et al. 2021). There are similar profiles for the NUV and FUV emission that are the main driver of the PAH emission. In all these cases the peak flux occurs for ages around 2 to 4 Myr, with relatively rapid declines within just 10 Myr.

Zoom In Zoom Out Reset image size

As expected, the nebular emission (purple line) shows an even more rapid drop-off, falling to just a few percent by 6 or 7 Myr. This is why nebular lines, such as Hα, are so valuable for age-dating; if you see strong emission associated with a cluster, the source is young (e.g., Whitmore & Zhang 2002; Anders & Fritze-v. Alvensleben 2003; Chandar et al. 2010; Fouesneau et al. 2012). The PAH band strengths have a similar predicted time evolution, although they typically drop to only ∼20% of their initial flux by 10 Myr. This is why the PAH emission is correlated with young cluster ages in Figure 6. We note that this drop-off in PAH strength as a function of time also appears to be compatible with the results of Lin et al. (2020) for LEGUS galaxies.

Since the scatter in the relationship between F335M magnitudes and log age based on HST is fairly large (∼0.6 dex), we have chosen to simply assign log age values of 6.5 (3 Myr) for the clusters without HST age estimates. This is consistent with the results from Figure 6, the flux profiles in Figure 7, and the comparison with Hubble age estimates in Table 1. Hence, our final age estimates are included as the second entry in column (6) in Table 1 and are either the HST-based SED-fitting value if it exists or log age = 6.5 (3 Myr) for the clusters without estimates from Hubble.

A complete census of massive clusters from the optical and infrared allows us to estimate the time that clusters remain in the deeply embedded phase, specifically the length of time they are not detectable at optical wavelengths. While early estimates (e.g., Kawamura et al. 2009) were longer than 10 Myr, more recent values have generally been in the range 1–5 Myr (e.g., Lada & Lada 2003; Whitmore et al. 2014; Hollyhead et al. 2015; Grasha et al. 2018; Chevance et al. 2020; Kim et al. 2021; Hannon et al. 2022; He et al. 2022; Kim et al. 2023). NGC 1365 provides a particularly interesting case because of the large amount of gas and dust that might, in principle, result in longer timescales. It also provides a unique view of the embedded phase of extremely massive (≥10⁶ M_⊙) clusters, which can eventually be compared with results from lower-mass clusters to assess whether or not the duration of this phase has any dependence on cluster mass.

We calculate the duration of the deeply embedded phase for massive clusters in NGC 1365 as No. missing / total × age (i.e., sample interval) = 4/30 × 10 Myr = 1.3 ± 0.7 Myr, a very short timescale. If we instead consider the timescale for massive clusters to be detectable in the B, V, and I bands (a minimum needed for most optically based age-dating methods), we find a somewhat longer timescale No. missing/ total × age = 11/30 × 10 Myr = 3.7 ± 1.1 Myr.

Our estimates for how long very massive clusters in NGC 1365 remain embedded in their natal interstellar medium (ISM) clouds are consistent with those found by previous works, without being obviously longer or shorter relative to estimates made for lower-mass clusters in other galaxies. It is important to establish if the length of the embedded phase depends on cluster mass or not, because this provides direct constraints on early feedback processes that may be important for limiting the masses of clusters.

While the strongest star formation in NGC 1365 is occurring in the inner region (i.e., 36 of the 37 youngest most massive clusters as we saw in Section 3.1), a moderate amount is also taking place along the bar, an example of which is shown in Figure 8. In particular, Bubble 1 is object 37 from Table 1, with a log mass = 6.6. Much of the star formation in Region 2 appears to be triggered by dust and molecular gas from dark filamentary "streamers" that reach the bar, colliding and merging with the gas and dust as illustrated in the bottom panel of Figure 8, and as discussed in Elmegreen et al. (2009). Another excellent candidate for this type of triggered cluster formation is Bubble 2 (3 Myr, A_V = 2.1 mag; age and extinction estimates from Whitmore et al. 2023), as shown in Figure 8. A careful look at Figure 1 shows several other examples of regions of star formation near the intersection of streamers and the bar.

Zoom In Zoom Out Reset image size

Previous kinematic work has shown strong streaming motions due to the bar in NGC 1365 (e.g., Sakamoto et al. 2007; Lindblad et al. 1996; Schinnerer et al. 2023). Simulations provide a more visual way to understand the dynamics of streamers of gas as they approach, interact with, and sometimes overshoot the bar lane. We encourage interested readers to watch this video. ²⁸ Figure 11 in Sormani et al. (2020) also provides insight to understand the dynamics and the role of streamers as they interact with the bar, potentially triggering star formation. One key new result from JWST observations is how ubiquitous these streamers are in spiral galaxies (see Meidt et al. 2023; Thilker et al. 2023).

Figure 9 shows a larger field of view around Region 2 (i.e., "Region 2 extended" from Figure 1) and contains additional kinematic information that can be used to help test the triggering hypothesis. A full kinematic analysis is beyond the scope of the current Letter, however. The top panel shows an IVB Hubble image that highlights the dust lanes and regions of young stars. Besides the streamer we show in Figure 8, which is labeled Streamer 1, we find several other candidate streamers. The second panel uses the F770W PAH emission filter (shown in the red) to highlight the youngest star-forming regions. The third panel shows CO (2–1) moment-1 velocity from the PHANGS–ALMA observations (Leroy et al. 2021), color-coded from −220 km s⁻¹ (dark blue) to +220 km s⁻¹ (dark red and white). CO clouds are marked using ellipses and box-circle symbols. Ellipses are sized to represent the deconvolved major/minor cloud axes (FWHM), whereas box-circles indicate unresolved clouds (E. Rosolowsky et al. 2023, in preparation).

Zoom In Zoom Out Reset image size

We first note the large number of CO clouds in Streamer 1, indicating that much of the gas and dust feeding into the bar, and eventually into the inner region of the galaxy, is coming from the streamers rather than from the bar itself. This figure shows that the gas velocities along Streamer 1 are similar (i.e., this feature appears as a uniform, green color, indicative of velocities around 0 km s⁻¹), as expected for a coherent physical component.

Gas velocities along the bar (the upper blue arrows) are more mixed, with values ≈+100 to +200 km s⁻¹ along most of the bar, but with velocities ≈+50 km s⁻¹ in regions where the streamers intersect the bar. This appears to be what is happening in the region of Bubble 2, with the velocities shown by a light blue-green color indicating a velocity that is intermediate between that of the bar and Streamer 1, as we would expect if this formed from the collision of the two components. The situation is less clear for Bubble 1, but the color appears to be light blue (≈+100 km s⁻¹), as expected if it is a combination of dark blue (just to the north and south) and green from Streamer 1.

We also find a less CO-populated streamer farther out in the bar, which we have labeled Streamer 2. It appears to be associated with three or four small star formation regions rather than a large event like the formation of Bubbles 1 and 2. However, we find that it has a similar morphology with uniform green colors until it mixes with the bar and then shows a combination of green and blue velocities beyond that point.

The bottom panel of Figure 9 shows the CO velocity residuals based on a model of the inner rotating disk (Schinnerer et al. 2023). This supports the same basic interpretation, but provides an even stronger connection between Streamer 1 and the feeding of the inner region, with a relatively uniform, narrow red component all the way into the central region. We also note that the association of Bubble 1 with the streamer is clearer in the residual velocity image than the original velocity map just above it.

An important capability of JWST is the ability to penetrate through the dust and identify structures that are not observable at optical wavelengths. An example is the NE bridge, which is identified in the top panel of Figure 8, but is much less obvious in the HST image in Figure 1. Features like this bridge are often seen in hydrodynamical simulations of bars (e.g., Sormani et al. 2020; Tress et al. 2020; Henshaw et al. 2022; Schinnerer et al. 2023).

Gas in a barred potential travels almost radially along the bar dust lanes (in the frame corotating with the bar) toward the inner parts of the galaxy. However, not all the gas transitioning from the dust lanes to the inner regions does so smoothly—some overshoots and collides with gas traveling down the opposite dust lane.

We believe the NE bridge feature is a clear example of material that has overshot the inner ring after falling toward the center along the southern bar dust lane (bottom right of Figure 1). When the material reaches the inner ring, gas and stars behave differently because they obey different equations of motion. Gas is affected by cloud collisions and thus dissipates some of its energy and eventually ends up accreting onto the inner ring. Stars, on the other hand, do not feel pressure forces and are unaffected by cloud collisions, so they fly undisturbed through the gas clouds and continue on their ballistic orbits toward the bar lane on the opposite (northern) bar lane, as appears to be the case with the NE bridge.

These, and several other features obvious in the new JWST observations (e.g., the overshoot regions marked in Figures 3 and 2), provide important insights into how the orbits of stars and gas vary depending on where they form and how this results in the population of older clusters outside the inner star formation ring. A related phenomenon is discussed in Whitmore et al. (2023; i.e., the "300 My" overshoot region, as identified in Figure 2).