Documentation for the preliminary Reduction of the

NGC7184 Spectroscopy

Author:

Frank R. Larsen, MSc.

The Lidemarkgroup

Date: 15^th October 2006

Abstract

August 15^th-16^th, 3 amateur astronomers from The Lidemark group used the Nordical Optical Telescope for the acquisition of a highresolution spectre of the Sb:r spiral galaxy NGC7184 using an 1.8” vertical slit and Grism #17 with 2400r/mm.

The goal was to obtain a spectrum that shows how the variation in redshift of emission lines of ionized molecular clouds in starforming HII regions of a galaxy can be used to measure its rotation. This rotation curve can then be used to give a lower limit to the mass of the galaxy.

This article documents the preliminary work with reducing the data and doing the first crude measurements.

Through very simple methods we show that we are able to produce results that are very close to those made earlier by professionals.

NGC7184 unfiltered

Reducing and preparing data

First the calibration and object fits-files were reduced in IRAF using function ccdproc.

Only bias, trim and overscan was applied, as we did not have a reference flat for the spectroscopy.

Spectrum was not corrected for cosmic rays – all though many are present and this step would make it nicer looking and the automated tools might work better.

Also there was made two spectres and stacking could perhaps lower the noise level.

This is a preliminary run where measurement are done manually, so the the cosmics are not a problem.

The spectrum was only 1000seconds under somewhat poor conditions and should probably have been at least 2-3000seconds.

However, we have sufficient SNR to do some measurement on the brighter parts of the galaxy.

As we have not learned all functionality of IRAF yet, further processing was done in IRIS.

Because of strong curved distortion, the calibration and object spectres was transformed with the smile function in IRIS.

Manually the Radius of curvature was calculated from solving the 2.order polynomial

(dx² + dy²)(1/R)² – 2dx(1/R) = 0  

Mean radius was calculated from several points along different lines in calibrations spectres, close to the emission lines found in the object spectra.

The calculations was also done with the ”sky” lines in the object frame and the results concurred.

(lines are however not straightened exactly)

R was estimated to be 10733pixels, and it was estimated that center lies on y=997, however not calculated.

Next step is to find a wavelength solution from the calibration spectres.

We have calibration spectres from both the Ne and the ThAr lamps.

For this preliminary investigation, only the Ne lamps was used and wavelength scale assumed for simplicity to

be linear in the interval from 656nm and through the wavelength where our galaxy spectrum lies.

The 1D Ne Calibration spectre was extracted in IRIS, and in Sao/DS9 numeric data was saved.

From the NOT homepage the wavelength definition for the Ne lamp was downloaded and by manual inspection the x-value for the strongest lines in the transformed calibration spectre was estimated giving this table:

Column #	Wavelength [Å]
185	6383.0
264	6402.2
685	6506.5
788	6532.9
1044	6599.0
1345	6678.3
1489	6717.0

Making a linear fit for column 685,788,1044 and 1345 we get the wavelength solution:

(1)     Wavelength [Å] = 6327.85795Å  + X * 0.260297Å/pixel     : (correlation=0.999983)

We see that the fitted dispersion of 0.2603Å/pixel matches the value supplied by NOT (0.26Å/pix)

Because of the width of the slit, the emission lines are wide and the outlined method is somewhat imprecise.

The slit is 1.8” and 9.5pixels. This gives a resolution of 9.5 x dispersion = 2.47Å.

That is – features in the spectre less than 2.47Å apart cannot be resolved.

In terms of redshift at 6563Å this translates to a resolution in redshift of:

(2)   z = (λ_obs-λ_em) / λ_em

(3)   v_rad = cz

v=c(λ_obs-λ_em) / λ_em = c * 2.47Å / 6563Å = 112.8km/s.

However, using midpoint of measured lines mean values can be measured at least to nearest 10km/s.

Here is the enhanced and filtered object spectrum of galaxy ngc7184. Dispersion is horizontal.

Sky lines are very dominant – evidence of moon, haze and clouds.

Here is 1D spectres from “towards” at line 1406, “center” at line 1022 and “away” at line 652:

It can be seen that spectre must have been slanted a bit, as the “towards” spectre is shiftet 3 pixel relative to center, and away is shiftet 1 pixel the other way.

This will have some impact on the calculations done later.

The high peak in “away” is a cosmicray.

Doing the measurements

By inspection the 2D spectre it can be seen that the “toward” and “away” region shows almost straight and vertical lines.

We use this fact for a fast approximation of the mean radial velocity and rotation speed.

This image shows red lines layered on top of the two feature emmisionlines which we for a start assume is the hydrogene alfa line (6563Å) and the forbidden N[II] line (6583Å):

Inspection of column values of the red lines and converting to wavelength using (1) and calculating redshifts using (2),(3) gives the following table

	column	Wavelength	Redshift z	Redshift vrad	mean	Differentialspeed
6563Å away	1146	6626.2Å	0.00963	2886km/s	2642km/s	244km/s
6563Å toward	1105	6615.5Å	0.00800	2398km/s
6583Å away	1225	6646.7Å	0.00968	2901km/s	2653km/s	249km/s
6583Å toward	1183	6635.8Å	0.00802	2404km/s

The N[II] line values differ by 0.25%-0.5% - which is less than 1 column value – and can thus be explained by the crude method of measurement.

The differential speed can be converted to a rotation speed in the remote galaxy by calculating the inclination angle.

Simbad gives majoraxis=6.025” and minoraxis=1.348”.

Assuming the galaxy is circular we calculate inclination to α=arcsin(1.348/6.025)=12.93˚.

True rotationspeed can then be calculated: v_rot = 247km/s / cos(α) = 253km/s.

Preliminary results:

The redshift of the galaxy ngc7184 has (through manual and somewhat crude methods) been measured to 2648km/s or z=0.00883.

In Simbad there is

6 R_vel measurements: 2617 : 2635 : 2617 : 2626 : 2635 : 2617  and 1 ze measurement :  0.008729

These values are not far from our values – we have made an overestimate of only 1.15%.

The mean rotationspeed of the outer parts of the galaxy has been estimated to be 253km/s which is normal for this type of galaxy.

We have shown (in image) that the outskirts are almost showing a constant speed – this is also normal, but also shows that there must be a lot of mass that cannot be seen.

What next?

Next we will learn spectre analysis in IRAF, and use the extensive functionality of doslit, making the measurements more accurate.

Subpixel tools for grabbing the coordinates and wavelengths of the emmisionlines will further increase the precision.

Also we will make measurements along the emission line in order to produce the rotationcurve.

Visually from the images, interesting details as higher velocities close to the center suggest that an effort should be made here. Unfortunately this part lies in the underexposed parts, so it will just be an estimated curve.

An estimate of possible errors and their sizes should be worked out so we can characterize the validity of the measurement.