Roberto A. Rodríguez Torres

REU program-Summer 2011
Univ. of Wisconsin - Madison
Madison, WI 53706

roberto dot rodriguez17 at upr dot edu

Research projects of other REU students

Pine Bluff Observatory

Spectral Response of the [OII] Spatial Heterodyne Spectrometer


The temperature and ionization conditions within the WIM appear to differ significantly from conditions within the classical HII regions immediately surrounding O stars. The evidence is provided by observed spatial variations in the intensity ratios of various nebular forbidden lines. For example, the anomalously strong [NII] 6584/Hα and [SII] 6716/Hα, and weak [OIII] 5007/Hα emission line ratios (compared to classical HII regions) indicate a low state of ionization with few ions present that require ionization energies greater than 23 eV (Madsen et al. 2006; Haffner et al. 1999; Rand 1997).

While photoionization models incorporating a low ionization parameter U (the ratio of photon density to gas density) are generally successful in accounting for the low ionization state of the ions (see, e.g., Domgörgen & Mathis 1994), they fail to explain the observations in detail, in particular the large variations in some of the line ratios. Haffner et al. (1999) have shown that these emission line observations can be readily explained if the large variations in [NII]/Hα and [SII]/Hα are due primarily to variations in the electron temperature Te rather than due to variations in the ionization parameter. For the Milky Way, they found that variations in Te between 7000 and 10,000 K would produce the observed variations in the line ratios. Therefore, an important question is whether the line ratio variations are indeed due primarily to variations in temperature, and if so, what is the source of this extra heat.

The interstellar medium (ISM) plays a vital role in the ongoing cycle of stellar birth and death, and galactic evolution. However the role of interstellar matter, from how its properties are influenced by stars to how, in turn, its properties influence star formation is poorly understood. Within the past decade substantial strides have been made towards unraveling the mysteries of a major ISM component, the widespread warm ionized medium (WIM). The advances were enabled by innovative spectroscopic techniques to detect and study extremely faint interstellar emission lines in the visible spectral region. With such observations it is possible to explore the connection between the Galactic disk and halo as energy and gas are transferred away from massive star-forming regions to large distances from the midplane.

An especially exciting development in this area is the evidence for temperature variations and the existence of a previously unrecognized source of heating within the WIM. The emission line of ionized oxygen in the near ultraviolet spectral region (3727 Å ) is key to exploring variations in temperature and ionization state within the gas, and for investigating the role of this additional heating. However, conventional Fabry Perot spectrometers are not well suited for such wavelengths. For this reason, a technique known as Spatial Heterodyne Spectroscopy (SHS) was used this summer in order to study this emission line.

This instrument is a type of Fourier transform spectroscopy (FTS) based on a modified Michelson interferometer in which the mirrors have been replaced by diffraction gratings. These gratings produce Fizeau fringes of wavenumber-dependent spatial frequency. They also have optical tolerances much more relaxed than conventional Fabry Perots.


Previous results from Mierkiewicz et al. and other investigators have shown the viability of using the [OII] emission as a potential tracer of temperature variations to confirm the indications made by the [NII] and [SII] line ratios (see citations above). However, an absolute intensity calibration has not been performed for this doublet. To accomplish this is the ultimate goal of this research project. This is intended to be done by imaging planetary nebulae with well known line ratios and calibrating the [OII] line emission.

However, a problem here arises. The instrument's field of view (FOV) is two degrees, whereas the planetary nebulae calibration sources can be just a few arcminutes in size. It was thus important to investigate how the spectral response of the SHS changes with variations in input area. That is the focus of this summer research project.


The data reduction started with the cropping of the interferogram to avoid the sharp transitions to the light to dark areas, which could cause problems with the Fourier transforms.

Raw Inteferogram

The lines show where the cropping was applied

After the crop, a Hanning function was applied to smooth the transitions down to zero intensity at the edges (once again to avoid problems with the Fourier transform).

Hanning function applied

The Fourier transform was then applied, which produces a "power spectrum". The horizontal white lines running across the power spectrun are the signal from the observation. The rows of bins that contain the signal are selected and their relative intensities are added to form the spectrum. The data was also "zero padded" (null valued pixels were added to aid in the fitting). Various zero paddings were investigated, partly to determine if any significant difference between them, and the preliminary results pointed towards that changes between one padding and another were minimal outside of simple scaling (which was expected). Ultimately, a three times oversampling was chosen for all data out of consistency considerations.

Power Spectrum

Data was periodically collected from cerium-neon (CeNe), rubidium-neon (RbNe), cesium-argon (CsAr) and cerium-argon (CeAr) lamps sources for calibration purposes.

RbNe calibration spectra

In order to generate a wavelenght per bin ratio, peaks from the calibration sources were fitted using a Voigt fitting program called VoigtFit.

Pixel vs Wavelength (Ångstroms) dispersion (RbNe, CeAr, CsAr)

The previous dispersion was created by taking emission data from each lamp separately and plotting the centroids (determined in VoigtFit) versus a list of known wavelengths, taken from the NIST atomic spectra database. The centroids of the spectral features that were thought to correspond to a given line emission were then plotted against the wavelength values supplied by the database.

Pixel vs Wavelength (Ångstroms) dispersion (RbNe and CeAr)

This dispersion was created from taking rubidium-neon (RbNe) and cerium-argon (CeAr) calibration lamps and collecting data from them simultaneously. The centroids were once again plotted against their known wavelengths.

Dispersion table

This is a table of every spectral line we were able to confidently identify from the calibration sources. It can be seen that an Ångstrom per bin ratio of approximately 0.0245 Ångstroms per bin was succesful for our configuration. Different configurations (i.e. different zero paddings) produced different ratios, but they were reasonably consistent with the expected scaling factors.

Aperture vs Relative Intensity

This plot shows how the relative intensities of the emission lines fall off with the square of the aperture diameter of the instrument's iris


Useful links

The following links are very useful for looking up info on UNIX, web page making, and astrophysical data and journals.

SIMBAD (Stellar/Galactic database)

NED (Extragalactic database)

UNIX tutorial

Web page basics

NASA Astrophysics Data Service