Physical Properties of H ii Regions in M51 from Spectroscopic Observations

The H ii regions were selected from the continuum-subtracted Hα image from NASA/IPAC Extragalactic Database. The sources are detected in this Hα image by SExtractor (Bertin & Arnouts 1996). The spectroscopic targets of H ii regions are selected as those sources with at least 25 pixels (about 75 pc in radius) whose Hα flux is larger than a critical value. Possible contaminations by foreground bright stars selected from Two Micron All Sky Survey Point Source Catalog (Skrutskie et al. 2006) were eliminated. The optical spectra of M51 were taken by the Optomechanics Research Inc. (OMR) long-slit spectrograph of the 2.16 m telescope (Fan et al. 2016) at the Xinglong Station of NAOC and the Hectospec multi-fiber positioner and spectrograph of the 6.5 m MMT telescope (Fabricant et al. 2005). The OMR spectrograph is deployed at the Cassegrain focus. It provides a dispersion of 4.8 Å pixel⁻¹ and a resolution of about 10 Å. The wavelength coverage is about 3600–8000 Å. The Hectospec is a moderate-resolution, multi-object optical spectrograph at the Cassegrain focus of the 6.5 m MMT telescope. The Hectospec fiber positioners allow the users to reconfigure 300 fibers to targets over the 1° focal surface in 300 s. Each fiber has a diameter of 15, corresponding to about 57 pc at the distance of M51. The Hectospec 270 gpm grating with blaze wavelength at 5200 Å was used. The spectral coverage ranges from 3650 to 9200 Å. The dispersion is 1.21 Å pixel⁻¹ and the resolution is ∼5 Å.

The M51 observations with the OMR spectrograph of the NAOC 2.16 m telescope started in 2008 March 9 and ended in 2014 March 23. A total of 19 nights were used. Bias and dome flat frames were taken at the beginning and end of every night. The He–Ar arc lamp was used for wavelength calibrations. Normally, we took two exposures for each object and each was about 1800 s. The left panel of Figure 1 presents the slit positions on the sky. The length of the spectrographic slit was about 4' and the width was about 25, corresponding to about 96 pc at the distance of M51. A total of 30 slit positions were selected. Some of them were placed along the major-axis and minor-axis directions of M51 and NGC 5195, which was aimed to obtain high signal-to-noise ratio (S/N) continua of the bulges. For other slits, we manually placed them to cover as many H ii regions as possible. The sky spectra with exposure time of 1200 s were taken between two exposures of each slit position. Proper flux standard stars were selected from the catalog of International Reference Stars (Corbin et al. 1991). Their slit spectra were used for the flux calibration.

Zoom In Zoom Out Reset image size

We were awarded half a night on 2012 February 10 by the Telescope Access Program (TAP⁹ ) to use the multifiber spectrograph of the MMT telescope. The H ii regions of M101 were preferentially targeted due to its larger size. M51 was considered as a secondary target to supplement the long-slit observations. It was unfortunate that the weather got worse after the M101 observation. Only a few spectra with good quality were obtained. The exposure time was 3600 s. The observations were taken at airmass of about 1.06 and the seeing was about 08. The right panel of Figure 1 shows the fiber positions.

The raw data taken by the NAOC 2.16 m telescope observations are processed using the IRAF software.¹⁰ We perform the bias subtraction, flat-field correction, cosmic-ray removal, and the sky-background subtraction. The spectrum of each H ii region is extracted based on the dispersion trace of the flux standard star and then wavelength-calibrated by using the extracted spectrum of the He–Ar lamp at the same CCD position. Flux calibration is performed based on the observation of the flux standard star and the mean atmospheric extinction coefficients at the Xinglong Station.

The MMT data are reduced with the publicly available HSRED v2.0 software.¹¹ After bias subtraction and flat-field correction, the fiber spectra are extracted and wavelength calibrated. The background sky emission is estimated by taking the average spectra with "blank sky" fibers in the same exposure and is subtracted from the individual spectrum. There is no observation for the flux standard star. The spectral energy distributions from the 15 intermediate-band photometric images of the Beijing–Arizona–Taipei–Connecticut Color Survey of the Sky are utilized to calibrate the spectra (Lin et al. 2017).

There is a total of 250 extracted spectra, 113 of which are visually checked to have either good stellar continua or emission lines. Among the 113 spectra, 99 spectra are from the slits and the others are from the fibers. The furthest location of the spectra was ∼10.5 kpc in terms of galactocentric distance of M51. All these spectra are resampled (keeping energy conservation) with a wavelength step of 1 Å in the range of 3700–7500 Å using IRAFs dispcor procedure. The 1 Å step is selected according to the suggestion of STARLIGHT¹² spectral synthesis code (Cid Fernandes et al. 2005), as we want to get the stellar population age. These spectra are corrected for the Galactic extinction using the extinction law of Cardelli et al. (1989) and the reddening value of $E(B-V)=0.031$ from the Galactic dust map of Schlegel et al. (1998).

The underlying stellar continuum of each spectrum is modeled with the STARLIGHT spectral synthesis code. A grid of 150 simple stellar population (SSP) templates from Bruzual & Charlot (2003) ¹³ with the Chabrier (2003) initial mass function (IMF) are used. These SSP templates cover 25 different ages (1 Myr–18 Gyr) and six metallicities (0.005–2.5 Z_☉). The extinction law of Cardelli et al. (1989) is adopted to add dust extinction to the models. The observed spectrum is considered as a linear combination of the SSP templates. During the fitting of STARLIGHT, the wavelength regions covered by nebular emission lines and atmospheric absorption lines are masked. Figure 2 gives an example of the spectral fitting for two H ii regions observed by the NAOC 2.16 telescope and MMT. The best-fit model spectra are also overlapped in this figure.

Zoom In Zoom Out Reset image size

For each spectrum, we derive a stellar population age, which is a light-weighted average of the SSP compositions. In order to get a relatively reliable age estimate, the continuum S/N at ∼5500 Å is required to be higher than 10. A total of 86 spectra have the age measurements. The uncertainty in logarithmic stellar age estimated by Cid Fernandes et al. (2005) is about 0.08 dex for S/N > 10.

After the subtraction of the stellar continuum as modeled in Section 3.1, the emission-line fluxes of Hβ, [O iii]λ5007, Hα, and [N ii]λ6583, which are used in this paper, are measured through the Gaussian function fitting. Only a few spectra (e.g., the ones at the galactic cores as shown in Figure 2) have the detections of other emission lines, such as [O ii] and [S ii], so their flux measurements are not provided. Due to relatively low spectral resolution and low S/N of the spectra, the flux ratio between [N ii]λ6583 and [N ii]λ6548 is tied to theoretical values of 0.348, respectively. The profiles of Hα, [N ii]λ6548, and [N ii]λ6583 are fitted simultaneously. Following Hu et al. (2018), the flux uncertainty of each emission line is estimated with the standard deviation within a wavelength region of 200 Å in width at the central wavelength of the emission line and errors induced by the Gaussian fitting.

The gas-phase extinction can be estimated with the Balmer line ratios, such as Hα/Hβ, Hα/Hγ, and Hβ/Hγ. The extinction estimation based on the Balmer lines is regarded to be independent on the physical conditions of the gas, such as the volume density and temperature (Domínguez et al. 2013). The Hα and Hβ lines are detected in most of our spectra, so we use Hα/Hβ to estimate the dust extinction.

Based on the assumptions of the reddening law of Cardelli et al. (1989), R_V = 3.1, and the intrinsic Hα/Hβ of 2.86 under the Case B recombination of T_e = 10,000 K and n_e = 100 cm⁻³ (Osterbrock & Ferland 2006), the reddening value of $E(B-V)$ can be calculated as

$$\begin{eqnarray}&&E(B-V)=\displaystyle \frac{2.5}{k({\rm{H}}\beta )-k({\rm{H}}\alpha )}\mathrm{log}\displaystyle \frac{{\left({\rm{H}}\alpha /{\rm{H}}\beta \right)}_{\mathrm{obs}}}{{\left({\rm{H}}\alpha /{\rm{H}}\beta \right)}_{\mathrm{int}}},\end{eqnarray} \tag{ 1 }$$

where k(Hα) and k(Hβ) represent the values of the reddening law at the wavelengths of the Hα and Hβ lines, respectively. (Hα/Hβ)_obs and (Hα/Hβ)_int are the observed and intrinsic flux ratios, respectively. As described in Catalán-Torrecilla et al. (2015), the extinction curves and dust-to-stars geometries would have little effect on the attenuation calculation. We also check the effect of adding the dust curve of stellar content in the SSP fitting on the gas-phase extinction and find that it makes the gas-phase extinction difference within about 0.03 mag, which is quite smaller than the average measurement error of about 0.1 mag. In this paper, if the observed Hα/Hβ ≤ 2.86, $E(B-V)$ is set to zero. All the fluxes of the emission lines used in this paper are corrected for this gas-phase extinction.

The SFR surface density is calculated as Σ_SFR = SFR/area. The area for a long-slit spectrum is calculated as the one of a rectangle, whose length is the aperture size used for extracting the spectrum and width is the slit width. The area for a fiber spectrum is calculated as the one of an circular, whose diameter is the fiber size. The SFR in ${M}_{\odot }\ {\mathrm{yr}}^{-1}$ is estimated with the extinction-corrected Hα luminosity (L(Hα)) by using the relation from Kennicutt (1998):

$$\begin{eqnarray}&&\mathrm{SFR}=7.9\times {10}^{-42}L({\rm{H}}\alpha )(\mathrm{erg}\ {{\rm{s}}}^{-1}).\end{eqnarray} \tag{ 2 }$$

This SFR calculator is derived based on the Salpeter IMF with stellar masses in the range of 0.1–100 M_☉. Note that the Hα luminosity is calibrated by using the aperture flux of 3'' in diameter in the Sloan Digital Sky Surveys r-band image.

The so-called direct T_e method is the most reliable way to measure gas-phase oxygen abundance, which is based on the ratio of auroral line intensities, such as [O iii]λ4363/λ5007 (Osterbrock & Ferland 2006). However, it is difficulty to detect these auroral lines in our spectra, which are ∼100–1000 times fainter than Hβ. There are multiple strong-line methods to estimate the gas-phase oxygen abundance of galaxies, including R23 (Kobulnicky et al. 1999; Pilyugin & Thuan 2005), N2 (Pettini & Pagel 2004; Marino et al. 2013), O3N2 (Pettini & Pagel 2004; Marino et al. 2013), N2O2 (Kewley & Dopita 2002; Bresolin 2007), etc. Since only a few spectra have the detection of [O ii]λ3727, those metallicity estimators related to this line is not applicable. Although [N ii]λ6583 and Hα lines are detected for most of our spectra, the metallicity based on the N2 index (N2 ≡ log([N ii]λ6583/Hα)) reaches saturation in the solar and super-solar metallicity regime with the N2 calibration (Pettini & Pagel 2004). Thus, we use the O3N2 index to estimate the gas-phase oxygen abundance, where

$$\begin{eqnarray}&&{\rm{O}}3{\rm{N}}2=\mathrm{log}\left(\displaystyle \frac{[{\rm{O}}\,{\rm{III}}]\lambda 5007/{\rm{H}}\beta }{[{\rm{N}}\,{\rm{II}}]\lambda 6583/{\rm{H}}\alpha }\right).\end{eqnarray} \tag{ 3 }$$

We adopt the O3N2 metallicity calibration of Marino et al. (2013), which is expressed as

$$\begin{eqnarray}&&12+\mathrm{log}({\rm{O}}/{\rm{H}})=8.533-0.214\times {\rm{O}}3{\rm{N}}2.\end{eqnarray} \tag{ 4 }$$

This calibration is valid in the range of −1.1 < O3N2 < 1.7 and the calibration uncertainty is about 0.18 dex (Marino et al. 2013). The random error of the metallicity, which is provided in the rest of this paper, comes from the emission-line measurements.

Based on the Balmer decrement and assumed reddening law of Cardelli et al. (1989), we derive the extinction in A_V and map the extinction distribution in Figure 4. In this figure, we also present the extinctions derived by Croxall et al. (2015), who obtained the spectra of 59 H ii regions in M51 from the CHemical Abundances Of Spirals project (Berg et al. 2015). In the paper of Croxall et al. (2015), the extinction was given at Hβ. The Hβ extinction c(Hβ) is converted to A_V following c(Hβ) = $1.43E(B-V)$ and ${A}_{V}=3.1E(B-V)$ (Berg et al. 2015). The overall average gas-phase extinction is about 1.18 mag. The core region of M51 presents a larger gas-phase extinction than the outside.

Zoom In Zoom Out Reset image size

The radial extinction distribution is shown in Figure 5(a). We can see that there is a mild extinction gradient, although the radial extinction presents a relatively large dispersion. From the extinction map in Figure 4, we can also see that the extinction in the northern spiral arms is generally smaller than the southern ones. Therefore, we try to divide the spectral samples into north and south regions and present their radial extinction distributions in Figures 5(b) and (c), respectively. The line of the minor axis of M51 is considered as the boundary, which is plotted in dashed line in Figure 4. From these radial distributions, we find that the dispersions get reduced compared to the whole distribution and the gas-phase extinction of the north region, which is close to NGC 5195, is generally smaller than that of the south one. The M51–NGC 5195 system has undergone more than once close encounters and the last one possibly occurred 300–500 Myr ago, which was inferred from kinematic and hydrodynamic simulations (Salo & Laurikainen 2000; Dobbs et al. 2010). The close encounter might disperse the gas distribution in the north parts more seriously and lead to the low gas-phase extinction, although more observational evidence is needed.

Zoom In Zoom Out Reset image size

The left panel of Figure 6 shows the distribution of the stellar age. The middle panel of Figure 6 shows the spatial distribution of the equivalent width of the Hα line (EW(Hα)). The EW(Hα) is sensitive to the ratio of present to past SFRs (see Kong et al. 2004, and references therein) and is related to the strength of ionization of the ionization source (Lacerda et al. 2018; Sánchez 2020). We have eliminated the regions influenced by AGN. The EW(Hα) of the majority of our samples are larger than 14 Å, so they are related to the ionization due to young stellar population. The maps of the stellar age and EW(Hα) are generally consistent as shown in Figure 6. The bulge of M51 is older than the outskirt and the inner arms are older than the outer arms. This kind of decreasing radial age profile supports the "inside-out" galaxy growth scenario, which is suitable for a majority of disk galaxies (Sánchez-Blázquez et al. 2014; Sánchez 2020). There are considerable H ii regions in the outer arms with age of 50–500 Myr. The recent close encounter between M51 and NGC 5195 can trigger substantial star formation and form young stellar populations. We also obtain the light-weighted mean stellar age of three regions in NGC 5195, including one in the galactic core and two in the disk. All of them are very old stellar population and the average age is about 10 Gyr. Almost no recent star formation is found in this galaxy, since there is lack of any Hα radiation.

Zoom In Zoom Out Reset image size

Calzetti et al. (2005) reported that M51 is a typical star-forming galaxy with the total SFR ∼ 3.4 M_☉ yr⁻¹ and Σ_SFR = 0.015 M_☉ yr⁻¹ kpc⁻². The right panel of Figure 6 shows Σ_SFR distribution of the H ii regions in M51. We can also see that there is on-going star formation in the bulge region of M51. There is no obvious difference of Σ_SFR between the northern and southern spiral arms. Although the close encounters can trigger star formation through the perturbation, they might disperse the gas more widely in the north, such that the north part presents similar Σ_SFR to the south. It can be also seen from the Hα map that the Hα emission looks more diffuse in the northern arms. Nevertheless, we need more observational evidence to support it.

In Figure 7, we present the radial distribution of Σ_SFR. From this figure, we can see there is a monotonic decrease of the radial Σ_SFR in M51, although the radial profile presents a relatively large dispersion. The gradient is about −0.55 dex ${R}_{e}^{-1}$. González Delgado et al. (2016) characterized the radial structure of Σ_SFR of the CALIFA galaxies, and they found that Σ_SFR of all spiral galaxies decreases with radial distance. The typical gradient in the central 1 × R_e is about −0.78 dex/R_e. Sánchez (2020) presented the radial distributions of Σ_SFR for low-redshift galaxies of different stellar mass and morphology based on the recent IFS surveys. M51 is a Sbc-type galaxy and its stellar mass is about 4.7 × 10¹⁰ M_☉ (Mentuch Cooper et al. 2012). We overlay the radial Σ_SFR profile from Sánchez (2020) for galaxies with the same morphological type and similar stellar mass in Figure 7. It shows that the radial Σ_SFR gradient of M51 is similar to the galaxies of similar mass and morphology in the nearby universe.

Zoom In Zoom Out Reset image size

4.3.1. The Spatial Distribution of Oxygen Abundances

There are a total of 67 H ii regions in M51 whose gas-phase oxygen abundances are estimated using the O3N2 calibration. The two-dimensional distribution of the oxygen abundance is shown in the left panel of Figure 8. The data obtained by Croxall et al. (2015) are also overplotted in this figure. Although Croxall et al. (2015) derived the oxygen abundance through the direct T_e method, we recalculate the metallicity using the same O3N2 calibration as used in this paper. Note that different strong-line calibrators may give distinct absolute values of metallicity.

Zoom In Zoom Out Reset image size

The right panel of Figure 8 shows the radial profile of the oxygen abundance. The data points from Croxall et al. (2015), recalculated by with the same O3N2 calibration, are also presented in this figure. These two data sets are complementary. It is interesting that there are two different gradients in the inner and outer regions. The inner region presents a positive gradient and the outer disk region shows a negative one. We perform two independent linear fits to the radial profile. The cutoff point is manually set at R/R_e = 0.4, which is around the turn point of the radial profile as seen in Figure 8. It is corresponding to the angular distance of about 50'' and the galactocentric distance of 1.9 kpc. The region of R/R_e < 0.4 includes the bulge (size of 11'' × 16'', Lamers et al. 2002) and surrounding star-forming ring. The linear fits are expressed as

$$\begin{eqnarray}&&\begin{array}{l}12+\mathrm{log}({\rm{O}}/{\rm{H}})[{\rm{O}}3{\rm{N}}2]\\ \quad =\,(8.70\pm 0.02)-(0.08\pm 0.015)\times R/{R}_{e}\\ \quad =\,(8.70\pm 0.02)-(0.22\pm 0.04)\times R/{R}_{25},\end{array}\end{eqnarray} \tag{ 5 }$$

for R/R_e > 0.4 or R/R₂₅ > 0.15, and

$$\begin{eqnarray}&&\begin{array}{l}12+\mathrm{log}({\rm{O}}/{\rm{H}})[{\rm{O}}3{\rm{N}}2]\\ \quad =\,(8.53\pm 0.02)+(0.26\pm 0.09)\times R/{R}_{e}\\ \quad =\,(8.53\pm 0.02)+(0.71\pm 0.25)\times R/{R}_{25},\end{array}\end{eqnarray} \tag{ 6 }$$

for 0 < R/R_e < 0.4 or 0 < R/R₂₅ < 0.15. In Figure 8, we also present the linear fits with only our data and only the data of Croxall et al. (2015). The corresponding gradients in the outer region are −0.05 ± 0.025 and −0.10 ± 0.02 dex ${R}_{e}^{-1}$, respectively. These gradients are consistent with that of the combined data within the uncertainty. Combining two data sets should provide more reliable gradient measurements, because we use the same metallicity calibrator. The metallicity gradients as shown in Equations (5) and (6) are adopted in the following analyses.

4.3.2. Comparison with Previous Work