Near-IR [Fe II] Line Emission in Nearby Star-Forming Galaxies
and the Supernova Rate

Kathleen Labrie

University of Victoria
Electronic-mail: labrie@uvastro.phys.uvic.ca

and

Christopher J. Pritchet

University of Victoria
Electronic-mail: pritchet@uvastro.phys.uvic.ca

Abstract:

Grain destruction behind supernova remnant (SNR) shock waves is believed to be the source of the near-IR [Fe II] line emission observed in star-forming galaxies. The enhancement of the [Fe II]$\lambda 1.644\mu m$/Pa$\beta$ ratio in SNRs compared to HII regions can be used to detect individual remnants in nearby galaxies. Lumsden & Puxley (1995) observed a linear correlation between the expansion velocity of a sample of SNRs in M33 and their [Fe II] flux. Through the Sedov solution, the expansion velocity of a remnant can be used to derive its age (i.e. time since the explosion). High spatial resolution [Fe II] mapping of nearby galaxies allows us to estimate the number of supernova remnants produced over a time interval calculated from the age estimates obtained from the iron flux of the individual SNRs. The supernova rate of the galaxy observed can then be calculated, without the need for multi-epoch observations. We recently obtained, at CFHT, high spatial resolution near-IR narrow-band images of three nearby star-forming galaxies (NGC 1569, NGC 3738 and NGC 5253). We present some preliminary results.

Introduction

  Supernova rates are generally measured using multi-epoch observations of a large sample of galaxies. This method works well but requires a large amount of telescope time. An alternative and potentially more efficient way to measure the supernova rate would be to study supernova remnants (SNRs). A supernova remnant's interaction with the ISM will stay detectable for thousands of years.

The goal of our study is to develop a reliable technique for measuring the supernova rate of a galaxy from its near-infrared [Fe II] line emission. The general idea was first suggested by Greenhouse et al. (1991). Grain destruction behind SNR shock waves is believed to be the source of the enhanced iron line emission at the remnants' locations (e.g. Greenhouse et al. 1991; van der Werf et al. 1993 and references therein). Here, we briefly present our project and the first results.

Scientific Analysis

 Recent observations of six supernova remnants in M82 (Greenhouse et al. 1997) and a few other remnants in M33 (Lumsden & Puxley 1995) clearly show the strong increase of the [Fe II] line emission at the locations of the SNRs. The interaction of the SNR's shock front with the interstellar dust is thought to be the cause of the [Fe II] emission. Iron is one of the most depleted elements in the ISM, and hence most interstellar iron is believed to be locked up in dust grains. A SNR shock wave leads to grain destruction, resulting in the return to the gas phase of the depleted iron. The significant increase of the iron abundance behind the shock explains the observed increase of the forbidden Fe+ emission.

The [Fe II] $\lambda 1.644\micron$/Pa$\beta$ $\lambda 1.2822\micron$ flux ratio can be used to discriminate between SNRs and HII regions. Indeed, for SNRs, this ratio is an at least an order of magnitude higher than that of HII regions (e.g. Greenhouse et al. 1991).

Once the SNR reaches the adiabatic phase of its evolution, the expansion velocity can be described by the Sedov solution, $v\sim\Bigl(\frac{E_\circ}{n_\circ}\Bigr)^{1/5}t^{-3/5}$, (E$_\circ$ is the initial energy of the shock, $n_\circ$ is the initial ambient density, and t corresponds to the age of the remnant). Also, Lumsden & Puxley (1995) observed a linear correlation between the expansion velocity of a sample of SNRs in M33 and their [Fe II] flux. Using the two relations above, [Fe II] $\sim v$ and $v\sim t^{-3/5}$, one can derive an estimate of the supernova rate within each galaxy.

Using the data gathered at CFHT in January 1998 (see Section 3), we wish (1) to detect individual SNRs in nearby galaxies, (2) to obtain an estimate of the supernova rate from the observation of individual supernova remnants, (3) to obtain an estimate of the supernova rate from the integrated [Fe II] luminosity, and (4) to study the possible use of the integrated [Fe II] luminosity to obtain SN rate estimates at moderate redshift.

Observations

 The observations were conducted at CFHT, on January 29-31, 1998, and consisted of direct imaging using REDEYE-W. High spatial resolution narrow-band imaging of three nearby star-forming galaxies, NGC 1569, NGC 3738, and NGC 5253, was done. We observed the Pa$\beta$ $\lambda 1.2822\micron$ and [Fe II] $\lambda 1.644\micron$ emission lines.

The galaxies had been selected according to their starburst activity, distance, angular size and availability at the time of the observations. Starburst galaxies are more likely to contain a large number of SNRs. Because our primary goal is to show that supernova remnants in nearby galaxies can be detected from their [Fe II] emission, it makes sense to start with galaxies having a potentially high number of SNRs. High spatial resolution is necessary in order to detect individual SNRs. Therefore, an upper limit to the distance, D$_{h_\circ 0.75}< 5$ Mpc, was set, so that the detection of $\sim 10^4$ years old SNRs would be possible. To achieve our goals the entire galaxy had to be observed, and that had to be done in a reasonable amount of time. Because the field of view of REDEYE-W is small, $2.1\arcmin \times 2.1\arcmin$, galaxies with a small angular size were favoured.

Detection

 A strong [Fe II] source has been detected (Figure 1) in the galaxy NGC 1569, approximatively $16.2\arcsec$ south and $38.9\arcsec$ east of star cluster A. At a distance of 3.1 Mpc ($h_\circ=0.75$), the FWHM of the feature is $\sim 35$ pc. More relevant is the [Fe II]/Pa$\beta$ flux ratio. Preliminary analysis indicates a flux ratio close to 1., which is consistent with ratios found for the SNRs in M33 and M82 (Lumsden & Puxley 1995; Greenhouse et al. 1991; Greenhouse et al. 1997). Also, the ratio is a few orders of magnitude larger than for typical HII regions, as expected for SNRs. Figure 2 shows contours of the [Fe II] emission superimposed on the Pa$\beta$ image of the same region; there is no strong hydrogen emission peak at the position of the iron emission source. It is also possible to see that the strongest iron emission occurs towards the densest part the galaxy, as expected if the emission is caused by the interaction of a shock front with the interstellar medium.

Conclusion

  Our analysis has led so far to the detection of a SNR candidate in the galaxy NGC 1569. Several other [Fe II] sources will be carefully inspected, along with the (hopefully) many other SNRs we expect to find. The search for SNRs completed, an estimate of the supernova rate in each galaxy will be obtained based on the number of remnants detected and their [Fe II] line flux.

This work was funded by the Natural Sciences and Engineering Research Council of Canada.

 

Figure 1:   Narrow-band image of the [Fe II] $\lambda 1.644\micron$ emission of NGC 1569. The SNR candidate is indicated by the arrow. The five strange looking feature in the upper part of the image are artifacts of the data processing. The plate scale is $0.5\arcsec$/pixel. North is to the left, east is down.

 

Figure 2:   Narrow-band image of the Pa$\beta$ $\lambda 1.2822\micron$ emission of a section of NGC 1569. Superimposed are the contours of the [Fe II] emission. The [Fe II] emission peak is clearly visible at the center of the image. The interval between the contours is 15 counts. For reference, the HII regions Waller 8 and 9 (Waller, 1991) have been labeled (plate scale and orientation as in Figure 1).