Methodology of emissionline surveys

In broad terms, two main methodologies are generally adopted for emission-line surveys, namely imaging through filters with a relatively narrow transmittance curve, designed to select the targeted emission line, and spectroscopy, either with a slit or slitless.

2.3.1 Narrow-band imaging Narrow-band imaging surveys are generally carried out either with traditional transmitting filters or with tunable filters, such as Fabry-Perot interferometers. One of the advantages of this technique is that the acquisition and reduction of imaging data, as well as source identification, are generally relatively straightforward. Imaging also has the further advantage of producing spatial maps of the line emission within the sources, as well as the spatial distribution of the sources themselves. However, only the limited range of wavelengths transmitted through the adopted bandpass can be probed at any one time, resulting in the selection of sources within a very narrow range of radial velocity or redshift.

The images can reach relatively deep flux levels, because the background noise from the sky is greatly suppressed by the narrow bandpass of the spectral element compared with broad-band images of similar exposure time (continuum images). However, the final sensitivity of narrow-band imaging to faint emission lines depends on how well the continuum emission at the wavelength of the line can be measured and subtracted. The measuring of the continuum requires additional imaging in spectral regions coincident with or adjacent to that of the targeted line. This means that either broad-band images (in at least one bandpass) or additional narrow-band images need to be obtained together with the primary narrow-band ones. Figure 2.1 shows the transmission curve of a typical narrow-band filter used in surveys for high-redshift galaxies together with that of the accompanying broad-band filter and the line-emission spectrum of the night sky (from Hu et al. 2004). (The inset in Figure 2.4 shows the case of two broad-band filters used to estimate the continuum.)

Figure 2.2 shows the identification of Lya-emitting galaxies at redshift z = 5 ' 66 by narrow-band imaging centered at A = 8160 A by Taniguchi et al. (2003). A galaxy with strong Lya emission is detected with high signal-to-noise ratio in the narrow-band filter and is undetected in the broad-band passbands, except in the I-band filter, where most of the observed flux is also coming from the line. Figure 2.3 shows the two-dimensional spectrum of the galaxy (top), as well as the extracted spectrum (middle), where only the emission line is observed.

Sources with strong emission lines (large equivalent width) are identified, for example, using color-magnitude diagrams in which the (narrow-broad) color is plotted versus the broad-band magnitude to show the line-flux excess over the continuum. Sources with weak emission lines can be detected only if the line contribution to the total flux recorded in the narrow-band image is larger than the uncertainty in the continuum

Figure 2.1. The transmission curve of the narrow-band filter centered at A = 8160 A used by Hu et al. (2004) in their survey for Lya emitters at redshift z = 5.7 together with that of the Cousins I broad-band filter. Also plotted is the spectrum of the night sky obtained with the LRIS spectrograph on the Keck telescope at Mauna Kea. Note how the narrow-band filter has been strategically chosen to limit the sky background by avoiding strong night-sky emission lines. The filter is also very nicely located near the center of the broad bandpass, which minimizes errors arising from the subtraction of the continuum contribution to the narrow-band images.

Figure 2.1. The transmission curve of the narrow-band filter centered at A = 8160 A used by Hu et al. (2004) in their survey for Lya emitters at redshift z = 5.7 together with that of the Cousins I broad-band filter. Also plotted is the spectrum of the night sky obtained with the LRIS spectrograph on the Keck telescope at Mauna Kea. Note how the narrow-band filter has been strategically chosen to limit the sky background by avoiding strong night-sky emission lines. The filter is also very nicely located near the center of the broad bandpass, which minimizes errors arising from the subtraction of the continuum contribution to the narrow-band images.

contribution. This highlights the fact that the sensitivity of narrow-band imaging surveys is limited both in flux and in equivalent width. The limit in flux determines whether the galaxy is detected at all in the various filters used. The limit in equivalent width determines whether the line emission can be detected over the continuum one. This is illustrated in Figure 2.4, which shows the color-magnitude diagram used by Shimasaku et al. (2004) to select Lya emitters at z ~ 4.79. The narrow-band filter is centered at A = 7040 A, and the continuum image is obtained by summing the R-band and i-band images.

Let's consider a survey with one narrow-band filter and a broad-band one, a configuration often used in narrow-band surveys. Let's also assume that the sources are detected in the broad bandpass. The limit in equivalent width comes from the fact that, in order to detect a line emission, the fraction of the total flux detected through the narrow filter due to the line only must be larger than the total photometric relative error (from both the broad and the narrow filter). In other words, the equivalent width Ew of the line must be such that

where (/>Broad and c/>NalTow are the fluxes collected in the broad and narrow filters, respectively, and FWn is the FWHM of the narrow-band filter. This relation assumes that the contribution of the line emission to the broad-band magnitude is small, which is certainly

Was this article helpful?

0 0

Post a comment