Title

This web page documents my photometric sequence observations for the exoplanet IL Aqr (GL 876, GJ 876).

The AAVSO is interested in exoplanet observing projects, and the upcoming exoplanet IL Aqr transit windows of 2004.10.21 and 22 will be the first observing project that the AAVSO is coordinating with the TransitSearch group. Web pages for TransitSearch start at http://www.transitsearch.org/links.htm and the AAVSO web page for the IL Aqr project is at http://www.aavso.org/news/ilaqr.shtml

The exoplanet system IL Aqr is located at 22:53:17.03, -14:15:52.6 (equinox 2000.0, according to my UCAC 2.0 astrometry, with a RMS residual of 0.2 "arc). The following image shows the stars in the IL Aqr region. Here are my suggested "referecne star" magnitudes.

Figure 2. V-magnitudes. FOV = 35.8 x 38.3 'arc, north up, east left. SE uncertainty is 0.05 magnitude.

Aaron Price (AAVSO) has suggested using reference star with a V-magnitude of 11.93 since it is very red, like IL Aqr. This star, hereafter referred to as *119, has a problem, however. It's a double, with the fainter star ~17 "arc to the NNE (pa = 327 degrees), as shown in a low resolution image below. The faint star is 3.0 magnitudes fainter than *119 using a V-band filter.

My photometry of *119 excluded the nearby faint star by employing a technique that would horrify most observers; I used a pixel editing feature to "remove" the faint star from the image (using a copy of the original for the horrific deed), then performed the photometry using that edited image. (I'm practised in this from my faint asteroid light curve work, where an asteroid is always going near faint stars during the course of an evenning's observing.) I didn't have the option of changing the "aperture radius/gap annulus width/sky reference annulus width" settings because doing that would render my equation predicted magnitude (described in the next paragraph) unusable for those changed settings. The way I recommend handling this faint interfering star for exoplanet monitoring is to set the signal aperture so that the faint star is entirely within or entirely outside the signal aperture (and not withni the sky reference annulus). The effect of completetly including the faint star within the signal aperture is to merely introduce a slight offset in the exoplanet magnitudes, and provided all images are reduced with the same aperture settings this will not cause changes in the exoplanet light curve.

Figure 5. This is the preferred way to set aperture radius, gap width and sky reference annulus so that the neaerby interfering star does not affect the measured intensity of reference star 119. FOV = 6.0 x 5.8 'arc, north up, east left (crop of a 72 x 48 'arc image).

Two nights were used to establish the photometric sequence presented above. I will describe obserations for the second night since it consisted of many more Landolt stars, a greater air mass range and an equation for converting intensity to magnitude that included an additional term.

I employ a novel scheme for converting measured star intensity to magnitude (described at http://brucegary.net/AllSky/x.htm). It is based on the idea that the intensity of a star in a CCD image is proportional to the star's brightness, to first order. In addition, it is assumed that allowance must be made for extinction effects, an instrumental response to stars of various colors, and a term that allows for the fact that the broad-band spectrum of a star is affected by total extinction (i.e., zenith extinction times air mass). It is just as intuitively obvious that a reverse calculation should be possible, namely that brightness is proportional to star intensity and all factors that affect apparent brightness. Linear dependencies are assumed, so the desired equation for deriving brightness from measured intensity, air mass, extinction, an instruemntal response term and a product term should look something like the following:

Magnitude = constant + 2.5 * LOG (exposure time / intensity) - magnitude extinction per air mass * air mass + constant * star color + constant * air mass * star color

The procedure I've adopted is to use a large number of Landolt standard stars (with accurate B, V, Rc and Ic magnitudes), observed at a wide range of air mass under clear sky conditions, and perform a least-squares solution for the various constants. Then, with these constants, it is possible to convert any unknown star's measured intensity to magnitude. For star color, such as B-V, an iterative procedure is employed, and this is straight-forward when using a spreadsheet.

When the Landolt stars are subjected to the same procedure as proposed for sue with unknown stars it is possible to assess the accuracy of the equations with their solved-for constants. I shall now present scatter plot graphs of the Landolt stars equation magnitudes versus true magnitudes.

Figure 6. Equation versus true B-magnitude for 56 Landolt stars observed under 4 air mass conditions on 2004.10.15. As noted in the graph the residuals from a perfect fit have an RMS of 0.043 magnitude (for stars brighter than 14.7)

Figure 7. Equation versus true V-magnitude for 56 Landolt stars observed under 4 air mass conditions on 2004.10.15. As noted in the graph the residuals from a perfect fit have an RMS of 0.057 magnitude.

Figure 8. Equation versus true R-magnitude for 62 Landolt stars observed under 4 air mass conditions on 2004.10.15. As noted in the graph the residuals from a perfect fit have an RMS of 0.033 magnitude.

B-mag = 19.31 + LOG (G / Iblu) - 0.228 * m + 0.10 * ((B-V)-0.70) + 0.06 * m * ((B-V-0.70)

R-mag = 19.83 + LOG (G / Ired) - 0.100 * m - 0.110 * ((B-V)-0.70) + 0.02 * m * ((B-V-0.70)

I-mag = 18.723 + LOG (G / Iinf) - 0.09 * m - 0.066 * ((B-V)-0.64) + 0.02 * m * ((B-V-0.64)

     where G is exposure time [seconds] (G stands for "gate time")
               Iblu = intensity using B-band filter (ADU counts), Ivis = V-band intensity, and Ired = R-band intensity, and Iinf = I-band intensity
               m = air mass, and
             B-V is star color (derived iteratively for unknown stars).

In my opinion these equations appear to be simpler and superior to the standard "CCD Transformation Equation" (derived and described in all its gory detail at http://reductionism.net.seanic.net/CCD_TE/cte.html). My all-sky procedure involves noting each reference star's "intensity" (having units of "ADU counts") and entering them in a spreadsheet. When the user completes an iterative solution for two key coefficient values (the first constant and the extinction coefficient) it is possible to convert any star's intensity to an "equation predicted magnitude." When this is done for the Landolt reference stars (of known true magnitude) it is possible to compute an RMS difference between equation predicted and true magnitude for the Landolt stars. This method of assessing RMS accuracy can be used to predict the accuracy of equation predicted magnitudes of any other star for which an intensity has been measured. Based on the RMS scatter of the Landolt stars magnitudes used for deriving the IL Aqr reference star magnitudes I conclude that the B-, V- and R-magnitudes noted in the star field images (Fig's 1 to 3) have an SE uncertainties of 0.043, 0.057 and 0.033 magnitude. The B-V entries are therefore subject to an SE uncertainty of 0.071 magnitude.

Light Curve for 2004.10.23

On October 23 UT I observed IL Aqr for about 6 hours and found no deviation from constant brightness. Here's my V-magnitude light curve.

ILAqr LC V-mag

Figure 9. V-magnitude light curve from a 6-hour observing run on October 23 UT. The gap at 4.1 hours is due to having to perform a "meridian flip" (MF) and re-calibrate the autoguider. (No adjustment was made to the post-MF data to bring it into agreement with the pre-MF data.) Each datum is a 30-second exposure, and their RMS departures from the average line is 0.021 magnitude (corresponding to a 20-minute average SE of 0.0041 magnitude, allowing for download overhead). Three reference stars were used, having V-magnitudes of 11.02, 11.48 and 10.42.

The pre-transit data are noisier than the post-transit data due to differences in autoguider performance. The lack of nearby bright reference stars means that some of the IL Aqr measurement uncertainty is due to the stochastics of using a limited number of relatively faint reference stars. Another source of scatter for these obserations was a high level of sky brrightness caused by the moon, which was 10 degrees away.

Miscellaneous Links

XO-1 observations (serves as an exoplanet observing tutorial) http://brucegary.net/XO1/x.htm
Exoplanet observing tutorial at http://brucegary.net/tutorial_exoplanet/x.htm
AAVSO home web page is at http://www.aavso.org/
TransitSearch home web page is at http://www.transitsearch.org/links.htm
TransitSearch tranist ephemeris for all exoplanets candidates is at http://www.ucolick.org/%7Elaugh
TransitSearch Results for TrES-1 is at http://www.ucolick.org/%7Elaugh/TrES-1b.results.html
AAVSO CCD observer's manual is at http://www.aavso.org/observing/programs/ccd/manual/
My TrES-1 observations of three transits is at http://brucegary.net/TrES-1/x.htm
My HD209458 observations of one transit (in 2002) is at http://reductionism.net.seanic.net/HD209458/ExoPlanet.html
My general interest AstroPhotos web page with many links is at http://reductionism.net.seanic.net/brucelgary/AstroPhotos/x.htm
All-sky procedure using equations for converting intensity to magnitude http://brucegary.net/AllSky/x.htm
My observatory in Hereford, AZ is shown at http://reductionism.net.seanic.net/brucelgary/AstroPhotos/m_hardware.htm
and you can reach me at the following e-mail address: b l g a r y @ u m i c h . e d u

____________________________________________________________________