HD 37605

HD37605 PHOTOMETRIC SEQUENCE

Bruce L. Gary (GBL), Hereford Arizona Observatory (G95)

HD 37605 is known to have a planet with an orbit that might produce transits. The star system is in Orion, with coordiantes RA = 05:40:01.75, Dec = +06:03:36.9. For brevity I will refer to HD 37605 by the name "E0540" ("E" stands for Exoplanet system, and "0540" is the approxiamte RA position).

Links internal to this web page:

    Star Field Overview
    Photometric Sequence Results
    Suggested R_Band Reference Stars
    Expected Precision for E0540 Monitoring
    Other Links (including Details of All-Sky Photometric Sequence Solution)

Star Field Overview

B&W large FOV

Figure 1. Large FOV, 70.2 x 44.4 'arc, of E0540 star field; north up, east left. E0540 is indicated by a square. [CGE-1400, SBIG ST-8XE, V-filter, 400 sec, Hereford, AZ]

Cropped version of previous

Figure 2. Enlarged and cropped version of the previous image, FOV, 23.4 x 17.6 'arc. E0540 is the bright star in the center. (Limiting magnitude ~ 17, SNR=3.)

Color image ofcropped FOV, faint stars shown Same as left panel, but showing bright star colors

Figure 3. RGB color image of cropped star field. Left panel shows faint stars (bright stars appear white because all colors are saturated); right panel shows only the bright stars. FOV = 24.2 x 30.6 'arc.

Photometric Sequence

All-sky observations were conducted December 1 (UT) for the purpose of establishing a photometric sequence for HD 37605 and HD 74156. Three Landolt star fields were observed at several air mass values. Details of the calibration of these data are given in the next section. The results for E0540 (HD 37605)

BVRI image

Figure 4. Star field for E0540 with 24 stars identified with BVRI magnitude information shown in a table. FOV = 71.0 x 44.23 'arc (north up, east left).

Crop of above

Figure 5. Zoom of crop of above image. FOV = 23.4 x 17.6 'arc.

Table BVRI Table V, B-V, etc

Figure 6. Tables showing all-sky photometric sequence for stars in previous two images, based on observations of 2004.12.10 (reduction version 4C23).

I plan on using an R-band filter for the January 1/2 observing window. A filter is required to avoid saturation of E0540 (especially true for a 14-inch aperture, where 2-second exposures are the maximum possible without a filter). I've chosen the R-band filter for two reasons. First, atmospheric effects for R are smaller than for B and V bands; and second, scintillation is slightly smaller for R than for B and V filters. Zenith extinction values at my site (in the winter) ares typically 0.26, 0.15, 0.13 and 0.09 [mag / air mass] for B, V, R and I. It is my subjective impression that my system is more "stable" using the R-filter than the I filter. Here's the star field and my choices for reference stars for the R-filter.

Suggested R-Band Reference Stars

The R-band all-sky magnitudes in the above table have been used to annotate the following image of the E0540 (HD 37605) star field.

Big FOV RED mags

Figure 7. R-band magnitudes for the E0540 star field.

Cropped RED mags

Figure 8. R-band magnitudes for stars suitable for "reference" (bright and having no nearby confusing stars). Small ovals are suggested "check stars" (having no nearby confusing stars). FOV = 44.3 x 40.3 'arc (north up, east left)

Expected Precision for E0540 Monitoring

When the exoplanet star is brighter than nearby stars, it is important to use as many reference stars as possible. This section illustrates that using 8 of 10 reference stars listed in the above section there is sufficient precision in the E0540 exoplanet's brightness measurement.

SE vs R-mag

Figure 9. Measured RMS precision versus star brightness using an R-band filter and exposing for 10-seconds (corresponding to 18-second intervals between exposures). The dashed line is an empirical model that assumes a constant component (such as due to scintillation) and a stochastic component (related to SNR). Eight reference stars were used (the maximum for MaxIm DL 4.0). [Celestron CGE-1400, prime focus using a Starizona HyperStar adapter lens, SBIG CFW-8, SBIG ST-8XE CCD; Hereford, AZ; 2004.12.01]

This graph shows that 10-second exposures with an R-band filter produces E0540 exoplanet brightness measurements with a 6 milli-magnitude SE at intervals of 18 seconds. By averaging 10 data points the expected SE should become 2 milli-magnitude. This level of uncertainty is much smaller than systematic errors. The expected transit fade is 19 milli-magnitude, lasting ~7 hours. It is unlikely that if a transit occurs no single observer will be fortunate enough to capture ingress AND egress. Therefore, it is important that each observer use the same set of reference ("comp") stars, or at least specify which reference stars were used.

There has been much discussion on the Photometry Discussion Group about how to overcome scintillation effects when observing a bright exoplanet star. The AAVSO also offers tips for observing bright stars: Bright Star Photometry. Scintillation can be the dominant source of uncertainty when observing a bright star with a telescope having an aperture large enough that it delivers too many photons to it's CCD and causes the bright star to saturate for exposures long enough to produce good SNR for the reference stars. Suggestions to overcome this have included: 1) reduce the telescope aperture, 2) use a B-filter to reduce the star's brightness (so that exposures can be 10 seconds or longer (which averages out some of the scintillation), 3) defocusing (in order to not saturate at the bright star's brightest PSF pixel location), 4) use a Barlow lens to spread-out the PSF (point-spread-function) over more pixels (in order to avoid saturation at the brightest pixel location), 5) wait for a night of poor seeing (to avoid saturation), 6) hope for thin cirrus (to reduce star intensity and avoid saturation), 7) use of a neutral filter in front of the CCD plate cover so that it covers the main CCD chip while leaving the autoguider chip clear (to avoid saturation of the star field that includes the bright star while still being able to autoguide), and 8) use of a plate with an occulting spot so that the bright star can be positioned coincident with the occulting spot.

Personally, I recommend doing none of the above! Reducing the telescope aperture increases scintillation effect, using a B-filter invites extinction systematic error variations on a slow time scale (hours, which is comparable to the transit duration), defocusing means nearby stars are more likely to appear in the sky background annulus when doing aperture photometry, a Barlow lens decreases the FOV which means you're limited to only nearby stars that are likely to be too faint for use as reference stars, poor seeing nights tend to be variable seeing nights (which means there will be systematic errors that vary throughout the night since the signal aperture will capture a variable amoutn of the PSF), cirrus clouds will introduce a host of other unknown systematic errors (but "be my guest" if you feel brave), a filter that covers the main CCD chip but leaves the autoguider chip in the clear should allow better autoguiding (but autoguiding isn't necessary if your polar axis is well aligned), and using a spot occulting plate that requires extremely accurate tracking to keep the bright star under the exact same location as the spot seems like a weird idea with negligible payoff (just to avoid saturation wile introducing a host of new systematic errors).

The graph, above, shows that none of these suggested strategies are necessary to achieve good precision on a star as bright as R-mag = 8.2. I'm going to observe with a full aperture, an R-filter, 9-second exposures, with sharp focus, and no Barlow lens (in fact, I'll be using a prime focus configuration, f/1.86 with a 1.2 x 0.8 degree FOV), on a night that hopefully has good seeing, no cirrus, no neutral filters to reduce the bright star light, and no spot occulting plate! Let the observing begin!