Time-resolved photometry of flare stars
Flare stars are usually M-type dwarfs which display prominent and rapid changes in brightness due to the release of energy in the form of flares. The brightness change is on the order of seconds or a few minutes. Most flares diminish rapidly but some can last for hours. Some flare stars have displayed only one flare while some, such as Wolf 424, show flares every few hours. This huge variability in incidence rates makes them fun and worthwhile objects to observe since you never know when the star will flare. It can be incredibly exciting to watch the light curve develop and see it reveal an immense flare.
Observing the flares is easy. You need a telescope and camera and an ability to perform differential photometry. You will also need software to plate-solve each of the many short-exposure images you will take. I use ASTAP. Lastly, you will need AstroImageJ to draw the light curve, although other software exists for this purpose.
The graph above shows two flares from Proxima Centauri one of the most active flare stars, over a period of about 2.5 hours. The data was taken with 5-second exposures and a blue filter. The two flares both show the characteristic FRED curve: Fast Rise Exponential Decay.
Below is a pair of flares from another flare star, YZ Canis majoris. This star produced two flares in rapid succession during the ~75 min of recording. Notice that the brightness of the star appeared elevated after the second flare had decayed.
Choosing the target
Several published lists of flare stars are available. Some lists identify stars which MAY be flare stars based on only one flare event. These objects need further scrutiny to confirm their identity as flare stars. Many stars are true flare stars which produce flares very rarely. A small number of stars are well-known flare stars which, while not regular in their ability to produce flares, nevertheless are good targets for beginners since they produce flares at least once in several hours of recording. The table below is a very short list of flare stars which I have recorded flares from. Proxima centauri is particularly active.
| Star | RA DEC | Approximate B-band Brightness |
| AT Mic | 20 41 51 -32 26 07 | 12 |
| Proxima centauri | 14 29 43 -62 40 46 | 13 |
| Ross 154 | 18 49 49 -23 50 10 | 12 |
| UV Ceti | 01 39 02 -17 57 01 | 13 |
| Wolf 424 | 12 33 17 +09 01 16 | 14 |
| YZ CMi | 07 44 40 +03 33 09 | 13 |
Equipment
Telescope: Flare stars tend to be faint M-type stars necessitating the need for larger telescopes with greater light gathering abilities. An aperture of 15 cm should be enough but more is better. You need to be able to guide for at least several seconds, but perfect guiding isn’t necessary. Focal length isn’t too important either but you should have sufficient image scale to be able to separate the target star from nearby stars if it is in a crowded field. At least 500 mm focal length should be good. Binning can be useful if you have a camera with tiny pixels. 2X or 4X can be useful if you end up recording thousands of images. The camera should be sensitive and preferably monochrome. It should possess stable cooling. You don’t need a large field of view however this can help in finding comparison stars which you will need when doing the differential photometry. A field of view of at least 20 arcminutes is a good start. Since most flares are low amplitude, you should record in 16-bit which most cameras nowadays can achieve. Filters: Most flare stars produce the greatest energy release in the UV, some in the blue and little in the red. Imaging without a filter (i.e. integrated light) will give you the highest signal/noise (SNR) image of the star, but if you can image with a UV or blue filter, you will get a higher SNR of the flare event. If you use a filter wheel, you can take rapid successive imges of the flare star to see how the flare looks in different wavelengths, however this will reduce the cadence. It is also possible to use a spectroscope to acquire time-resolved spectra of flare stars but this is very challenging. Each spectrum will spread the light of the star over many pixels, which means it will be very faint. This is the most informative way to image flare stars, and I have seen it achieved by a few amateurs (including by Robin Leadbeater and Andrew Smith), but you will need a large telescope, a very sensitive camera, and attention to detail during processing. The spectroscope should be a low-dispersion design. A Star Analyser 100 or 200, or an Alpy200 would be the best choice when trying to acquire time-resolved high-cadence images of flare star spectra.
Method
Exposure Time: Once you have acquired the target flare star in the field of view, set the exposure time of the camera to a few seconds. Short exposures will give you better temporal resolution of the shape of the flare event, but will lead to a noisier image. Long exposures will produce better SNR but will lead to a poor assessment of the temporal shape of the flare profile. Since most flares tend to be about 1 min long, I recommend exposure times that are a few seconds long for bright flare stars and about 15 seconds long for very faint ones. Exposure times of a minute or more will reveal the presence of flares once you have produced the light curve, but will be too long to display the temporal profile, unless the flare is many minutes in duration. These events do occur but are rare. In the case of the Proxima centauri flares in the graph above, you can see that the rise time of the flare is about 30 seconds long, and the decay is several minutes long.
Gain: Flares will make the star brightness increase for a short time. But not by much. Use a gain value for your camera which will make the star image appear about 1/3 of the maximum brightness available. For 16-bit camera, aim for a target star image brightness about 5-10k. High gain will produce noisier images but will help extend the brightness range, in the camera, over which the flare is recorded. Although it is possible that using very high gain values will make a flare appear overexposed, this is unlikely to occur. Very low gain values will lead to images with less noise, but the dynamic range of the flare will occupy few brightness values. You will need to experiment when setting a good gain value.
Imaging: The image file name should contain the name of the object, the time and date of acquisition. If possible, the image metadata should contain the RA and Dec of the object as this will greatly aid in registration of the images. I use N.I.N.A. since it is free, stable and flexible. Take images for as long as possible. If you are taking images without a filter, or with a spectroscope, set the camera acquisition program to take as many images as you have time to image. You will need to image for at least several hours. If you want to use colour filters, set the acquisition program to take alternating filter images (eg red-blue-red-blue etc). During the acquisition, make sure the telescope tracking and guiding is as good as you can make it. The occasional passage of thin cloud is acceptable since differential photometry can usually handle the errors this imposes. However high wind or thick cloud will lead to poor photometry.
Plate solving: Once you have acquired several hundred minutes worth of images you need to plate solve the images. However they aren’t averaged into a single image which is what’s usually done to produce an astrophoto. Instead, the RA and DEC for each image is determined. I use ASTAP since it is extremely fast and batch processing hundreds of images is easy. ImageJ can also solve for WCS in an image however I found it to be slower and, worse still, sometimes it failed to find the WCS for an image.
You can download ASTAP from here.
Instructions for setting up ASTAP for bulk plate solving are here.
Differential photometry with AstroImageJ: Open AstroImageJ and load the files using File/Import/Image Sequence. The rest of the process involves making a light curve using AstroImageJ in a way similar to that used to make an exoplanet light curve.
Below is a pdf of the entire process. If you have a series of images and want to quickly check them for a flare, you can skip the dark and flat calibration steps and simply do differential photometry, (go to STEP 5) but if you do find a flare, you should calibrate the images properly before sharing the data.