22 replies to this topic

[Abdulrahman Toonsi]

hello I was recently imaging the sky looking for exoplanets and one of the stars showed this dip in brightness, I used AstroImageJ to prosses the data I took. I used a 200mm diameter SCT and a CCD camera. I have attached the images of my light curves. the flat lightcurve is the same star the following night showing no signs of transits(I hope this means that the transit of the previous night wasn't a malfunction). I also put an image of lightcurves of HAT-p-30b taken with similar sized equipment(not by me) and they are very similar to what I have detected. I have checked NASA's exoplanet archive and have not found this planet. AstroImageJ said that from the transit depth it has calculated the planet's size to be from 3.22 to 5.22 Jupiter radii (it varies because the type of the host star is not know, I assumed that it is a G type or F type because hot Jupiters are most common around them). the transit duration is 28 minutes.

 

 

could this be a false detection caused by some sort issue with the processing?

Attached Thumbnails

• 
• 
• 
• 

[D_talley]

Great job.  The transit time of 28 minutes is short. I would image that star for a week to see if you can pick up another transit to confirm that it is a transit. 

[rutherfordt]

Yes, it certainly resembles the light curves of the exoplanet transits that I have seen.  D_Talley's advice is really good though-- image the star regularly for a week (or two or more)-- if its a transit then it will repeat at some point.  If you can get a second image like this then you can start to figure out what the period might be and once you have that, you can make predictions of when they will occur again-- if the predictions match what you are able to detect then that would be some strong evidence that you have detected an exoplanet transit of some sort.  Now, it won't be absolute evidence since there could still be a few other things that it could be, but being able to successfully predict when these events will occur will eliminate a lot of the "something elses." 

 

Also, at some point, you will need a positive observation from another observer which would really strengthen your case that this is a transit. 

 

Good job!

[pvdv]

The curve has the generic aspect of a fairly standard exoplanet transit. There are a couple of problems though.

1- at 28 minutes, it might be the fastest transiting exo-planet ever. It may also be on an impossible orbit. It is hard to say for sure since you attempt to hide what you are observing and don't have enough information avout the star, but something like a hot jupiter with 4 Jupiter radius transiting in 28 minutes doesn't feel right.

2. an extremely fast transiter would have a different looking curve (see that one for example) https://arxiv.org/pdf/1909.12424.pdf for a bunch of reasons.

[StarmanDan]

As others have mentioned, more data is needed to determine if you really caught something or not. Comparing your data to HAT-p-30b is not an adequate comparison as HAT-p-30b has about a 2 hour transit duration. Not having accurate stellar spectral data also severely limits your ability to calibrate the data to determine planetary mass. It is not good science to make assumptions on a star's spectral class to determine this.

The first thing you need to do is collect more data over a longer time period to see if the transit repeats itself. You should get at least three confirmed transits which should allow you to make predictions on when the next will occur. If one happens the fourth time around at about the predicted time, theres a good chance it's a real phenomenon.

In addition, we could help you help you out a lot more easily if you were more forthcoming with your data. While it's understandable to hide some data feeling like any potential "discovery" might be stolen from you, I assure you this won't happen. Your data can easily prove you discovered it first should anyone try to do so. Besides, science is a collaborative effort. It is not uncommon for astronomers to see something interesting or odd and contact others if for nothing other than to provide a sanity check and to see if they see the same thing as you. In fact, at some point, if you plan to announce any discovery, you're going to have to provide that data so that others can confirm or deny it. So if you want us to help you as much as we can, we really full disclosure.

[rutherfordt]

The 28-minute transit time is actually not unusual-- TOI 1048.01 transits in 44 minutes, TOI 10474.01 transits in 40 minutes and TOI 2821.01 has a transit time of only 17 minutes.  Most transits are, of course, much longer than this, but there are some that are this short.

 

The NASA Exoplanet Archive (https://exoplanetarc...ac.caltech.edu/) is a good resource for anyone trying to do stuff like this.

[pvdv]

I'd say the jury is still out on 2821.01

https://exofop.ipac....e.php?id=252826

Anyway, initial observation here obviously worth confirmation and full documentation.

[RedLionNJ]

This looks like some really nice work, Abdulrahman. There are a couple things difficult to account for in the data, but they don't preclude the very likely probability the transit is real:

 

1. Those 'high outliers' in the data, right before the transit starts - might there be a probable cause for them? Nothing like that appears either after the transit has completed or the following night in the non-transit data

 

2. Agreed 100% with some of the comments above regarding star identification, comparison stars, precise equipment and methodology - if this is a newly-discovered exoplanet, there is 100% no doubt the provenance lays with you. Nobody can "steal away the discovery" or any other such nonsense. Nobody else is going to have data from Feb 13, 2024 for this star/planet.

 

3. The 28 minute transit time for such a large object (up to 5 Jupiter radii) is a little worrisome.

 

Again, I have to repeat point #2: part of the scientific method is to share your methodology and details of your equipment and see if others can reproduce your results with their own equipment.

[StarmanDan]

Another aspect you need to consider is have you tested your camera for linearity? This is a crucial step to ensure you're using the proper exposure time and not saturating, under exposing, or using an exposure time that causes the target or comparison stars to delve into the realm of your camera's non linearity curve. Also, Dennis Conti, a leading expert in exoplanet transits, now recommends guiding when performing an exoplanet run. This is to keep the stars in the image on the same pixels on the camera throughout the duration of the transit. Otherwise, if the target is allowed to drift slightly, you may encounter pixel variations in sensitivity that could masquerade as a transit. While you performed a second observation the next night and saw no transit, it is unlikely you placed the stars on exactly the same pixels as the night before.

[robin_astro]

If I am reading your graph right the eclipse depth is ~5% (~6.75 to 6.4 flux) which is large for an exoplanet (~3x larger than your Hat 30b comparison and larger than any known I think) Have you checked that your target is not a known binary ?

 

Cheers

Robin

[robin_astro]

First check your images to make sure the "transit" is not caused by a cold pixel wandering in and out of the aperture.

 

If all is well,  so far all you have is a star that may have dropped in brightness once.

 

Next step would be to find two more similar equally spaced transits . This would confirm you have a binary system.

 

To distinguish an exoplanet from a binary star system you would then need some professional help to get a spectrum

 

Cheers

Robin

[robin_astro]

 (~3x larger than your Hat 30b comparison and larger than any known I think)

Excluding planets orbiting white dwarfs like WD1856b which is much bigger than its star so is more like an eclipse than a transit ! 

https://britastro.or...34e10f42c4fc746

 

Cheers

Robin

[StarmanDan]

[quote name="robin_astro" post="12516227" To distinguish an exoplanet from a binary star system you would then need some professional help to get a spectrum

Cheers
Robin[/quote]

Not necessarily. Binary star light curves have a distinguishable difference than exoplanets. Dennis Conti describes how to analyze the data for follow up observations of TESS candidates to determine which is which.

[robin_astro]

Not necessarily. Binary star light curves have a distinguishable difference than exoplanets. Dennis Conti describes how to analyze the data for follow up observations of TESS candidates to determine which is which.

Can this distinguish between a brown dwarf and an exoplanet? (Wouldn't a brown dwarf have an identical transit light curve to a hot Jupiter at visual wavelengths?) My understanding is to confirm an exoplanet you need the mass which can only be determined from radial velocity measurements.

 

Cheers

Robin 

Edited by robin_astro, 23 February 2023 - 06:45 AM.

[Abdulrahman Toonsi]

Yes, it certainly resembles the light curves of the exoplanet transits that I have seen.  D_Talley's advice is really good though-- image the star regularly for a week (or two or more)-- if its a transit then it will repeat at some point.  If you can get a second image like this then you can start to figure out what the period might be and once you have that, you can make predictions of when they will occur again-- if the predictions match what you are able to detect then that would be some strong evidence that you have detected an exoplanet transit of some sort.  Now, it won't be absolute evidence since there could still be a few other things that it could be, but being able to successfully predict when these events will occur will eliminate a lot of the "something elses." 

 

Also, at some point, you will need a positive observation from another observer which would really strengthen your case that this is a transit. 

 

Good job!

after a couple of night of imaging I think I found another transit. 

 

the transit duration is almost exactly the same as the previous one, the first transit was 28 minutes and this transit was 29 minutes. almost everything is the same except that this transit is a bit shallower than the previous one, giving me estimates of Jupiter sized planets instead of super Jupiters like the last one did. the time between the first and second transit is approximately 221 hours, what could the orbital period be?

Attached Thumbnails

• 
• 

[rutherfordt]

There seems to be a lot more scatter in the data for the second possible transit than there was in the first that you posted.  Is the target closer to the horizon now or was it perhaps not as good a night?

 

As far as a period-- if the second detection is real, then the period can be about anything up to 221 hours-- you don't know how many events might have passed between your first image and this one.  Getting a third event would narrow things down a lot-- getting a fourth even more so.  So-- keep monitoring the star and see if you pick more of these up.

 

Also, as I and a couple of other people have mentioned earlier-- at some point, you will need to disclose the identity of the star that you are monitoring so that someone else can verify your results.  There are several people on here who could do that-- its a very important step and really needs to be done to verify your claim.  You should also be prepared for the possibility that an outside observer will disprove your claim (this might turn out to be something else)-- as disappointing as that would be for you, it happens in science all the time.

 

So-- keep monitoring the star and see where things go from here.

[Abdulrahman Toonsi]

There seems to be a lot more scatter in the data for the second possible transit than there was in the first that you posted.  Is the target closer to the horizon now or was it perhaps not as good a night?

 

As far as a period-- if the second detection is real, then the period can be about anything up to 221 hours-- you don't know how many events might have passed between your first image and this one.  Getting a third event would narrow things down a lot-- getting a fourth even more so.  So-- keep monitoring the star and see if you pick more of these up.

 

Also, as I and a couple of other people have mentioned earlier-- at some point, you will need to disclose the identity of the star that you are monitoring so that someone else can verify your results.  There are several people on here who could do that-- its a very important step and really needs to be done to verify your claim.  You should also be prepared for the possibility that an outside observer will disprove your claim (this might turn out to be something else)-- as disappointing as that would be for you, it happens in science all the time.

 

So-- keep monitoring the star and see where things go from here.

yes, the star was closer to the horizon than the first transit and that's very probably why there is more scatter.

[pvdv]

What you need for a decent foundation can be found in this paper from 2002

https://arxiv.org/pd...-ph/0206228.pdf

Yes, the paper is a bit old (relatively speaking) but it remains suitable for amateurs.

Then, ideally, you start worrying about limb darkening, probabilistic distributions of the parameters, etc...

Assuming a star type and ending up with (Rp/R*)^2 of 0.0770 and 0.1317 may suggest that you could benefit from more consistency in your analysis.

If you are observing a TOI, there should already be lightcurves available, those can usually be downloaded as .FIT files but in a format that typically isn't supported by programs amateur use. The pyhton package "lightkurve" 

It is really simple to use.

import lightkurve as lk
import matplotlib.pyplot as plt

# open the TESS FITS file
lc = lk.open('hlsp_qlp_tess_ffi_s0043-0000000429501231_tess_v01_llc.fits')

# plot the light curve
lc.plot()
plt.show()

[RedLionNJ]

On the orbital period (assuming the exoplanet really does exist) - you could narrow this down a bit by analyzing all the negative observations - those time windows where you were recording data which resulted in no transit-like dip.  My hunch is you would have to be really, really lucky to hit such a rapid transit with such a significant dip both times in 221 hours if the period was anywhere close to that.

 

Possibilities include:

 

a) the period is quite rapid, which would account for both the chance of capturing it twice in only 221 hours AND the relatively rapid transit period

b) the period could actually be TWICE the nominal 221 hours and this is a close-contact binary star and you captured one primary and one secondary event, 221 hours apart (or 111 hours apart, or 74 hours apart, or ...)

c) something else altogether causing the observations

 

 

I really don't understand your reluctance to share much in the way of data & methodology, other than a handful of anonymous light-curves. This is not a healthy way to "do science" and contributes greatly to people's doubts about the authenticity.

 

I don't think there is anyone reading this thread who does not want this to be genuine - one more incremental set of data points in our understanding of exoplanets. And what anyone wants isn't really relevant anyway - let the science speak for itself.

[Astrolyn]

I ran into a similar situation and come to find out TESS had data on it using Transit Least Squares I found an orbital period of p ~1.72 days with the highest power for that period specifically and what made this interesting is not only did my observations matched Transit Least Squares and my math combined, but the planet had a high impact parameter and edge on orbital inclination which was the cause of the 39.25 (+/- 4 minutes) minute transit duration. This was now confirmed as a Short Period Hot Jupiter with a even transit ratio of 15:1 meaning there was only one odd transits that were off center by 55 arcseconds from Transit Least Squares Python Package with 15 even transits, meaning that they showed no out of transit events. Impact parameter from TESS was 0.88, our Ground Based Telescopes got (b) ~ 0.859 with an uncertainty of (+/-0.021) from TESS. Orbital Distance was also at 0.0668 au away from its star TESS showed 0.0518au with an uncertainty of 0.015au. Ratio Planet and Ratio Star was estimated to be 0.0140%. Got to say this was an interesting investigation we were not sure what it was at first until we investigated further. At first we thought it was Intrinsic Stellar Variability and or a NEB and or EB but was easily disproved with TESS and Ground-Based confirmation. Only thing that would really worry me is when we do chromaticity study where we take separate exposures with different filters different observations is if we run into a huge variation in flux which would be a prime example of a astrophysical false positive in this case during the study. Limb Darkening parameters showed that this was a planet and did not drastically alter the light curve in any way as it was still a U-shaped curve. Amateur Astronomy and discovering, confirming these exoplanets play a huge role and pave a way for a lot of future studies with the general public that have no knowledge in this at all. I am really excited for the future of astronomy as a lot more people are becoming more interested in Exoplanets and how they are confirmed and detected.

[Xilman]

I thought I may have found an exoplanet transit of 25mmag. Yours seems to be a bit deeper than that (or I mis-read your light curves, which is eminently possible) and rather deeper than is typical.

 

Searching for other minima was successful but, unfortunately, the data folded on the detected period showed another 10mmag minimum and so what was actually found was a hitherto unknown Algol-like eclipsing variable.  I suggest that you should also check for this possibility.  In my case, I found that a supposedly good sequence star wasn't useful for high precision photometry though still fine for visual estimates.

 

Good luck with getting to the bottom of this one and I hope that you have indeed discovered a new exoplanet.

[River Hills]

I am just a novice at photometry, but the two different sets of data have nearly identical lengths of time for the diminished signal observed and this seems suggestive of an issue here on earth. . As Robin pointed out above, the depth is about 5% and I am wondering if this might be due to the time for the image to drift in the field of view over pixels obscured by dust that is not corrected for by a good flats calibration. I struggle with flats, so that came to my mind as a possible explanation for the two lengths of time being so similar. 

[Astrolyn]

Good evening! Going back looking at that transit with what I know now, I think you may have stumbled upon a rare find. I am confident you spotted a Grazing Exoplanet Transit with a short transit duration because it is near the stellar limb of the star on the Y axis! The reason for the Shorter Duration is due to that factor. I also believe you stumbled on a relatively big gas giant given the Rp/Rs and if I calculate the radius based off what you have against our sun I would say confidently it is a Gas Giant Exoplanet with a relatively good Signal To Noise Ratio (SNR), and if you could give me the star that you were monitoring I could take to TESS data and report back to you on what I think it is. Some candidate work I am also doing with our amateur setup. This target i'm working on is 4,514 light years away! Talk about a hop skip and a jump away! If you would like to join our Exoplanet Watch Group via Slack join via https://www.exoplanets.nasa.gov/ I am heavily active on that group and can help you proceed further in your investigation! Also 9.2083 days is your orbital period give or take, considering your time in between transits! All you would need to do is measure against two transit events however, three would also be the best to do Transit Timing Variations TTV evaluations. With that slight difference in one minute I would say there could very well be gravitational influence of the possibility of another planet in the system if you keep closely monitoring the star! Happy Hunting!  

Edited by Astrolyn, 15 February 2025 - 05:44 PM.