NGC 3105: A Young Cluster in the Outer Galaxy

TL;DR

This study presents detailed imaging and spectroscopic analysis of the young open cluster NGC 3105, determining its distance, age, mass function, and stellar properties, revealing insights into its composition and stellar evolution.

Contribution

First detailed photometric and spectroscopic analysis of NGC 3105 extending to sub-solar masses and revealing its age, distance, and stellar population characteristics.

Findings

Distance of 6.6+/-0.3 kpc to NGC 3105.

Age of at least 32 Myr for the cluster.

Presence of infrared excess and Be stars indicating dust and stellar activity.

Abstract

Images and spectra of the open cluster NGC 3105 have been obtained with GMOS on Gemini South. The (i', g'-i') color-magnitude diagram (CMD) constructed from these data extends from the brightest cluster members to g'~23. This is 4 - 5 mag fainter than previous CMDs at visible wavelengths and samples cluster members with sub-solar masses. Assuming a half-solar metallicity, comparisons with isochrones yield a distance of 6.6+/-0.3 kpc. An age of at least 32 Myr is found based on the photometric properties of the brightest stars, coupled with the apparent absence of pre-main sequence stars in the lower regions of the CMD. The luminosity function of stars between 50 and 70 arcsec from the cluster center is consistent with a Chabrier lognormal mass function. However, at radii smaller than 50 arcsec there is a higher specific frequency of the most massive main sequence stars than at larger radii. Photometry obtained from archival SPITZER images reveals that some of the brightest stars near NGC 3105 have excess infrared emission, presumably from warm dust envelopes. Halpha emission is detected in a few early-type stars in and around the cluster, building upon previous spectroscopic observations that found Be stars near NGC 3105. The equivalent width of the NaD lines in the spectra of early type stars is consistent with the reddening found from comparisons with isochrones. Stars with i'~18.5 that fall near the cluster main sequence have a spectral-type A5V, and a distance modulus that is consistent with that obtained by comparing isochrones with the CMD is found assuming solar neighborhood intrinsic brightnesses for these stars.

Tables

Table 1. Table 1 Previous age and distance estimates of NGC 3105

Age	Distance	Reference
(Myr)	(kpc)
$25\pm 10$	$9.5\pm 1.5$	Sagar et al. (2001)
$20\pm 5$	$8.5\pm 1.0$	Paunzen et al. (2005)
$25\pm 5$	$6.8\pm 0.3$	Davidge (2014)

Table 2. Table 2 Summary of GMOS images

Filter	Exposures	FWHM
		(arcsec)
$g^{\prime }$	$1\times 1$ sec	0.7
	$1\times 300$ sec	0.7
$i^{\prime }$	$1\times 1$ sec	0.5
	$5\times 60$ sec	0.5

Table 3. Table 3 Spectroscopic targets

ID	RA	Dec	$i^{\prime }$	$g^{\prime }-i^{\prime }$	RegionaaF = Field, S = Shoulder, C = CenterbbMask alignment stars indicated with a ‘$\ast$’
	(2000)	(2000)
1	150.236249	-54.767349	15.823	0.955	F
2	150.232644	-54.800179	18.941	1.578	F
3	150.228896	-54.806980	15.656	0.917	F
4	150.224848	-54.790791	16.473	1.212	F*
5	150.219455	-54.798790	16.305	1.200	F
6	150.215549	-54.787201	18.850	1.443	F
7	150.206852	-54.789371	18.755	1.519	F
8	150.202804	-54.808510	18.731	1.473	F
9	150.199199	-54.790760	18.671	1.517	S
10	150.195293	-54.784309	20.509	2.310	S
11	150.193750	-54.784167	18.576	2.018	S
12	150.191846	-54.792011	20.532	2.058	S
13	150.188556	-54.800850	20.637	2.551	S
14	150.184951	-54.813831	16.347	1.128	F
15	150.181503	-54.808170	18.467	1.452	F
16	150.177312	-54.799061	20.553	2.114	C
17	150.172648	-54.784229	20.553	2.210	C
18	150.166197	-54.797981	20.486	1.974	C
19	150.162449	-54.801311	20.560	2.422	S
20	150.158844	-54.792580	20.499	2.205	C
21	150.153751	-54.789532	20.560	2.113	C
23 ccThere is no entry for Star 22 as it was found to be an artifact of a bright star.	150.148501	-54.798698	20.572	2.165	S
24	150.145354	-54.794891	18.569	1.484	S
25	150.141006	-54.789539	20.462	2.269	S
26	150.136042	-54.795441	20.473	2.138	F
27	150.135417	-54.795417	21.148	2.566	F
28	150.126901	-54.783588	16.024	0.968	F
29	150.117602	-54.784351	15.980	1.171	F*
30	150.107846	-54.801510	15.844	0.911	F
31	150.103812	-54.774899	16.240	1.180	F*
32	150.099167	-54.801750	15.378	2.632	F
33	150.098405	-54.801609	16.113	1.014	F
34	150.094800	-54.812679	18.582	1.333	F
35	150.091352	-54.787239	15.948	1.018	F

Table 4. Table 4 Na D equivalent widths (EWs) in mean spectra constructed from the Bright sample.

ID	EW
#	(Å )
Blue Sub-sample	$1.5\pm 0.1$
Stars 4, 5, & 14	$1.9\pm 0.1$
Bright Sample	$1.6\pm 0.1$
(No Stars 29, 31, & 32)

Full text

1. INTRODUCTION

Simulations of molecular clouds suggest that stars form in deeply embedded filamentary structures that subsequently collapse and merge to form star clusters and a diffusely distributed component (e.g. Bonnell et al. 2011). Assuming that the majority of stars form in such environments then a census of young stars in clusters and in the so-called ‘field’ suggests that most clusters must be short-lived (e.g. Lada & Lada 2003; Silva-Villa & Larsen 2011; Fall & Chandar 2012). That stars do not form in isolation is important not only for interpreting the nature and evolution of star clusters but also for understanding the properties of stars in the field. This is because the time that stars stay in close physical proximity in their natal environments may affect the properties of the objects that ultimately migrate into the field, such as the binary fraction (e.g. Parker & Meyer 2014) and the low mass cut-off of the main sequence (MS; e.g. Johnstone et al. 1998; Adams et al. 2004).

Feedback from massive stars likely plays a key role in cluster evolution. Feedback may purge the gas that is the dominant contributor to the total system mass early in a cluster’s evolution, thereby altering the gravitational potential so that stars are no longer bound to the cluster. Feedback is also a potential regulator of star formation during embedded phases (e.g. Nakamura & Li 2014), and may trigger star formation in surrounding areas (e.g. Bik et al. 2014; Getman et al. 2014). Feedback may also provide re-cycled material to a star-forming region if star formation proceeds over timescales that are longer than a few Myr (e.g. Palous et al. 2014).

If feedback from massive stars is a significant influence on cluster evolution then the age distribution of embedded clusters should plunge at a time that corresponds to the lifetime of massive stars. Studies of clusters with minimum masses of at least a few $\times 10^{3}$ M*⊙* in the nearby spiral galaxy M83 indicate that the oldest embedded clusters have ages $\sim 6$ Myr (Hollyhead et al. 2015), which is roughly consistent with the lifetime of very massive stars. This age also corresponds to the onset of the red supergiant (RSG) evolutionary phase in stellar systems, and so marks a time when the intensity of the local UV radiation field will drop if very massive stars were initially present.

Factors other than feedback almost certainly also affect cluster evolution. The kinematic state of a cluster at the time of gas expulsion likely plays a significant role in its survivability, in the sense that systems that happen to be in a super-virial state at the epoch of gas expulsion will have a greater chance of surviving than those in a sub-virial state (e.g. Farias et al. 2015). The fate of a cluster may also depend on environmental factors such as proximity to massive molecular clouds (e.g. Kruijssen et al. 2012), location in a galaxy (e.g. Silva-Villa et al. 2014), and the morphological characteristics of the host galaxy (e.g. de Grijs et al. 2013).

Having a large initial mass may not be a guarantee that a cluster will survive for more than a few Myr. Rather, the fate of a young cluster may be influenced by local conditions within the natal environment. Stars in massive star-forming regions may form in a range of environments with surface densities that span at least 4 orders of magnitude (e.g. Kuhn et al. 2015), and it is the densest sub-structures that might be expected to have the highest chances of long-term survival. Not all systems contain such sub-structures. For example, the massive star-forming complex W33 lacks compact sub-structures, and will likely evolve into a loose association (e.g. Messineo et al. 2015).

The star-to-star age dispersion in a cluster is one measure of how long star-forming material is retained, and this can be estimated by measuring the age difference between the youngest pre-MS (PMS) stars and the oldest MS stars. Realistic stellar structure models that span a wide range of masses and evolutionary states are required, and the failure to include key physical processes can skew age dispersion estimates. For example, models presented by Somers & Pinsonneault (2015) indicate that the failure to account for star spots may cause isochrones to detect an erroneous age dispersion in the CMDs of systems that are actually coeval.

Observations suggest that age dispersions up to a few Myr are common in massive young clusters (e.g. Román-Zúñiga et al. 2015; Zeidler et al. 2015; Kuryavtseva et al. 2012). Such a relatively short star-forming history is consistent with observations that point to a rapid collapse for massive systems (e.g. Banerjee & Kroupa 2015) and the timescale over which feedback may start to influence cluster evolution. Still, evidence for larger age dispersions has been found in some cases (e.g. De Marchi et al. 2011b; Lim et al. 2013). Environment may also play a role in determining the duration of star formation. In particular, the spread in age in low mass clusters that form in low density environments may be significantly larger than in more massive – and presumably more compact – clusters (Pfalzner et al. 2014).

The Galactic disc contains young clusters in a range of environments, and the investigation of clusters and their immediate surroundings throughout the Galaxy will provide insights into the processes that affect cluster evolution. With R${}_{GC}=11.4\pm 0.6$ kpc (Paunzen et al. 2005), the open cluster NGC 3105 is in the outer regions of the Galactic disc. Previous estimates of the age and distance of NGC 3105 are summarized in Table 1. The age estimates in Table 1 draw on different age diagnostics and wavelengths. The ages measured by Sagar et al. (2001) and Paunzen et al. (2005) are based on the photometric properties of MS turn-off (MSTO) and post-MS stars at visible wavelengths. In contrast, the age estimated by Davidge (2014) is based on the MS cut-off (MSCO) and the properties of the cluster PMS sequence in the $(K,J-K)$ color-magnitude diagram (CMD).

In the present paper, deep photometric and spectroscopic observations of stars in NGC 3105 are discussed. Images obtained with the Gemini Multi-Object Spectrograph (GMOS) on Gemini South are used to construct a CMD that extends from the very brightest cluster members to $g^{\prime}\sim 23$, which is where PMS stars might be expected based on some previous age estimates. If the input physics used in stellar structure models are correct then age diagnostics that cover different wavelengths and a broad range of magnitudes should yield the same age. The range of magnitudes covered with the GMOS data enables an investigation of cluster age using indicators that span a range of masses and evolutionary states.

In addition to investigating the age, distance, and reddening of NGC 3105, the GMOS CMD was also used to select targets for spectroscopic observation at visible and red wavelengths. Spectra were obtained of stars that span a range of brightnesses, and the data are used to investigate the line-of-sight interstellar absorption, search for line emission, and estimate spectral types. This is the first spectroscopic survey of NGC 3105 that includes stars that are fainter than the MSTO. Finally, images of NGC 3105 that were obtained as part of the SPITZER (Werner et al. 2004) GLIMPSE (Benjamin et al. 2003) survey are also examined. These data are used to probe the light profile of the cluster and investigate the mid-IR properties of bright stars in and around NGC 3105.

Details of the observations and the steps used to remove instrumental signatures from the data are discussed in Section 2. The light profile of NGC 3105 obtained from archival SPITZER images is investigated in Section 3. The photometric measurements obtained from the GMOS and SPITZER images are described in Section 4, and the CMDs and LFs obtained from these data are examined in Sections 5 and 6. The stellar spectra are discussed in Section 7, while a discussion and summary of the results follows in Section 8.

2. OBSERVATIONS & REDUCTIONS

2.1. GMOS Imaging

Images and visible/red spectra were recorded with GMOS (Hook et al. 2004) on Gemini South as part of program GS-2014A-Q-84 (PI: Davidge). The GMOS detector when these data were recorded 333The detector in GMOS has since been changed. was a mosaic of three $2048\times 4068$ EEV CCDs. Each 13.5$\mu$m square CCD pixel subtended 0.073 arcsec per side on the sky. All images and spectra for this program were recorded with $2\times 2$ pixel binning.

$g^{\prime}$ and $i^{\prime}$ images of NGC 3105 were recorded on the night of UT December 31, 2013. The sky conditions were photometric when the data were recorded. The exposure times and the mean full width half maximum (FWHM) of stellar profiles at each exposure time are summarized in Table 2. Short and long exposures were recorded to broaden the magnitude range over which photometry could be performed. The long exposure $i^{\prime}$ images were recorded with a five point dither pattern that sampled the corners and center of a $10\times 10$ arcsec square.