observations
The interacting spiral galaxy M51(NGC5194) was observed in Hsignificant at 2
and L(H
NGC 5194, or Whirlpool galaxy, is an Sbc type grand design spiral galaxy seen nearly face on and is at a distance of 7.3 Mpc (Waller et. al.1997). It is interacting with its companion NGC 5195. Its proximity and face on appearence make it an ideal candidate to study star formation rate(SFR) in interacting galaxies. M51 has been observed in different wavelength bands to study SFR. (Bushouse 1987; Bell & Kennicutt 2001; Scoville et. al. 2001). The consistency of these observations can be checked by comparing star formation rates.
The goal of this project was to determine the star formation rate in M51 from H observations and compare it with previously published results of star
formation rate in various wavelength bands. UV photons with energies
in excess of 13.6 eV ionize the ground state HI gas surrounding the massive stars.
The ionized photons cascade back to the ground state producing recombination lines.
A significant portion of this recombination radiation is an H (n =3 to n=2) transition. H luminosity is directly proportional to the number of
ionizing photons and therefore can be used to determine the star formation rate.
This goal could not be achieved because of extreme non photometric conditions. The
data was used instead to find an upper limit on the flux and the SFR in individual HII regions.
=
13h29m53s.4, 
= +47°11'48" ) was observed in H and H continuum filters on
May 3-4, 2002 with the 0.25 m Great Ohio Telescope(GOT) and ST-8 detector with 1530 x 1020 pixels.
The pixel size is 9 X 9 
m with a pixel
scale of 0.7 arcsec/pixel. The angular size of M51 is 11' x 7', which fits in the 17.5'
x 12.8' field of view of the detector. Conditions were non-photometric and most of
the observations were taken through cirrus clouds. The CCD operating temperature was ~-20°C.
A guiding star was not available; therefore, the exposure time for a single exposure was
restricted to 30s. 22 x 30s exposures were obtained in each filter. The filters were alternated
after each exposure to observe the galaxy in each filter at same airmasses. The ring nebula (M57)
was used as a standard and was observed at two different airmasses. Calibration frames included 11 x
0.11s zeros, 6 x 20s and 5 x 30s darks, and 8 x 20s evening and morning flats in each H and H continuum. Morning flats were taken
after sunrise; the sky was not uniform because of thin cirrus. There may have been a light
leak due to light from the laptop during observations.
and H continuum. The galaxy could not be seen in the final
images and there was a sharp flux gradient in each image. The images were block averaged in 4 x 4
pixels. After block averaging only 8 images in H filter showed very faint nuclei of the two galaxies. These eight 30s exposures
are equivalent to one 163s exposure. A sample 30s image in H is shown in figure 1.
|
Figure 1.Continuum subtracted H
image of M51 in a 30s exposure. Nuclei of two galaxies are faintly visible (indicated by arrows)
|
images were added using the 'imexpr' task to get the final
H image and the corrosponding H
continuum images were added to form the final H continuum image. Since the H continuum filter is
about 4 times wider than the H
filter, one forth of the H continuum image was subtracted from the H image to get the H emission line flux. The image after subtraction was blank(except for statistical noise)
and no part of the galaxy was visible.
This could have been the effect of the flux gradient mentioned earlier. To remove it, the following
steps were performed on the set of eight good images in each filter. The raw images were
dark corrected. They were then averaged using the 'combine' task. The combined image was smoothed
using the 'mkskycor' task and it showed a sharp gradient(figure 2). This smoothed average was subtracted
from the dark corrected object frames. The images were then flat fielded and added to get
combined image in each filter. Though the gradient was not completely removed, the images
looked much flatter than before. They had to be block-averaged to see if the galaxy was
visible. A sample 30s image in H filter
is shown in figure3.
Even after these reductions the final continuum subtracted image was blank. Since no HII
regions were visible this image could not be used to measure the flux. The noise in
the final image was calculated and was used to find an upper limit on flux from star forming
regions at 2 and 3 detection levels. Considering the effect
of block-averaging, the noise in the final image was found to be 16.92/pix.
|
Figure 2. The smoothed average of eight combined H images. The raw images in each filter were
dark corrected and combined into a single image. The resultant is smoothed to obtain the gradient.
Blank sky images were not available, so the images used for combining and finding the smoothed
average were the eight good images.
|
|
Figure 3. A 30s exposure in H
after subtracting the smoothed average from the dark corrected image. The image is block averaged into
4 x 4 pixels.
|
observations, and the analysis of the data
to find an upper limit for the star formation rate.
HI gas surrounding the hot, young stars is photoionized by the UV continuum shortward of the Lyman limit (912 A°). All the HI gas is ionized if the nebula is optically thick. These ionizations are balanced by the total number of recombinations per unit time (Osterbrock 1989).
. . . . . . .(1)
Q(H) is the total number of ionizing photons and L
is the luminosity of star per
unit frequency interval. The luminosity in a particular frequency interval (H) for case B recombination is (In case B recombination any photon emitted in n2P to
12S transition is immediately absorbed by neighbouring Hydrogen atoms, and the downward radiative
transitions to 12S are not considered)
. . . . .(2)
where B
and effH are the recombination coefficients, with 
(Seaton 1959).
Thus the luminosity in H is
proportional to the number of ionizing photons and can be used to calculate the star formation rate.
Assuming a Salpeter initial mass function (IMF) of 0.1-100
, the SFR is given by
. . . . (3)
The significant contribution to H
emission comes from high mass stars (M > 10) with lifetimes < 20Myr. Therefore the SFR measured from
H is independent of the early history
of the galaxy.
|
Figure 4. Combined image in H.
|
| Figure 5. Combined image in H continuum. |
|
Figure 6. Continuum subtracted image of M51.
No HII regions are visible. |
and H continuum images;
and figure 6 shows the continuum subtracted image.
I could not measure the H luminosity,
as the galaxy was not visible in the image.The data was instead used to estimate an upper limit on the
flux at detection levels of 2 and 3.
Assuming that the HII regions appear as point sources with PSF FWHM of 8 pixels or 5.6", the noise
in an individual HII region image is, N/pix from blank sky multiplied by the square root of numer of pixels
in the point source image. This was calculated to be 239.91 counts. For a detection limit of 2 emission line is
redshifted by 10 Ao, where the transmitivity of the filter is 68%. Therefore the rest frame H
flux is 1.0797 x 10-12erg s-1
cm-2, corrosponding to a luminosity of ~1039.8erg s-1. Thus
equation(3) gives an upper limit of 0.0098
for the star formation rate in a single HII region.
A similar calculation at 3. Figure 7 shows the HII region
luminosity function of M51(Rand 1992). The upper limits on H the luminosity are consistent with the published results.
|
Figure 7 HII region luminosity function of M51. (* :Rand 1992; o:Kennicutt et. al.1989) (Rand 1992) |
observations of M51 were performed.
Though the data were inadequate to measure the luminosity from individual HII regions the upper limit on
luminosity was obtained. This upper limit is consistent with the published values of
individual HII region luminosities (Kennicutt et. al.1989, Rand 1992) to a detection limit of 2 and 3. The upper limits on single HII region luminosities and
the corrosponding star formation rates are found to be L(H) ~ 1039.8erg s-1, SFR = 0.0098 significant at 2 and L(H)= 1.035 x 1040erg s-1, SFR = 0.0147 significant at 3. I could not find published values
of star formation rates in single HII regions.
From figure 7 the luminosity function of HII regions peaks at 2.511 x 1037erg s-1.
The S/N needed to detect a single HII region at this luminosity is zero. So it is not possible to
detect single HII regions at this luminosity with our equipment. If conditions had been photometric for the
eight good images it would have been possible to detect HII regions with L(H) = 1039erg s-1, S/N = 1.97 to L(H) = 1039.6erg s-1, S/N = 7.6. This
corrosponds to a total of 18 HII regions. Also, 25 X 30s exposures in photometric conditions would have allowed
detection of 72 HII regions down to a luminosity of 1038.6erg s-1 and S/N = 2.
The noise in a single HII region image would be 239.91 counts as explained in the earlier section. The estimated counts in the brightest HII region with a luminosity of 1039.6erg s-1 are 267. This gives a S/N of 1.1, which is very marginal to detect. This is most likely reason for not being able to detect any HII regions. If conditions were photometric the S/N would have been higher giving a possibility of the detection. To have some detection (say S/N = 3) in these conditions, the required exposure time would have been 7 x 163S exposures.