J0139 (ZTF J013906.17+524536.89) Photometry Monitoring by Amateur Bruce Gary
Bruce L. Gary, Last Updated 2021.07.12, 15 UT

This web site reports photometry measurements of a white dwarf with a suspected planetesimal that is shedding dust debris when its 106-day elliptical orbit brings it close to the star's Roche distance. The presence of fades means that the orbiting debris crosses in front of the star. My intent is to observe this star at intervals of a few days to record future fade events in order to see if periodicity is regular or chaotic. So far the dust clouds appear to recur at intervals between 104 and 110 days, and each dust cloud has a different structure. The structure within a dust cloud appears to consist of regularly-spaced clumps. The 2019 October clumps were 5 hours apart whereas the 2020 November clumps are 2.4 hours apart. For both events the amplitude of variation on the several hour timescale was ~ 80 mmag.   

Status

Fade events are now known to consist of many independent dips with durations typically 1 or 2 hours and depths ~ 50 % deeper than the depth of the long timescale (daily averages) plot. During a typical observing session when long-timescale fade is deep there are brief periods when brightness is close to the OOT level (between brief dips). Presumably, the previously observed week-long fade events really consist of a multitude of short (1.5-hour) dips. This new information has important constraints on modeling the dust cloud spatial structure, and dust production mechanism. In other words, instead of a week-long fade event being produced by one broad dust cloud, which was a default assumption, a week-long fade consists of 100 to 200 small dust clouds produced by the same number of sources. The hundreds of sources can be thought of as fragments from an earlier (tidal) break-up, or collision. Because an observing session can consist of a half dozen independent dips, with brightness ranging from close to OOT to 40 % depth, for example, it is possible during such an observing session to measure depth vs. wavelength and determine the PSD (particle size distribution) of each dust cloud. The published periodicity of 107.2 days is somewhat supported by the recent fade event.

We are currently observing a brightening above the OOT level, which might be caused by forward scattering of a big dust cloud that failed to produce a fade recently (might have just missed our LOS).

Introduction

J0139 is a white dwarf (DA type, hydrogen atmosphere) that appears to undergo 30 - 45 % fades at ~ 107-day intervals. Five fades have been measured so far (Vanderbosch et al., 2019). I have adopted the following tentative ephemeris for the middle of the fades:  JD = 2458661 + E 107. The fade expected in late 2020 January did not occur! The discovery paper, Vanderbosch et al. (2019), suggests that the fades are due to a planetesimal in a 110-day orbit that is very eccentric (e > 0.97), and during periastron it is close to the Roche radius of the WD and fragments are dislodged that form a dust cloud in the same orbit. RA/DE = 01:39:06.2 +52:45:37. The observing season is centered on October 22.

List of internal links

    Results to date  
    AAVSO observations
    Physical model suggestion 
    Observing sessions 
    Finder image 
    Observing and Analysis Tips 
    References 

Results to date



Figure 1a. Are we going to see a 2-fade event this month? If the brightening now underway is real then it could be due to forward scattering from a big dust cloud that might produce a fade in early December (if inclination is favorable).


Figure 1b. The two months before this fade activity was "quiet."  


Figure 1c. A 1.5-year LC showing presene of odd-numbered dips and absnece of even-numbered dips.


Figure 2.  Detail of above graph. The variations within an observing session are real (as the next two figures show).

The following figures show a few AHS fits to each observing session, starting with the latest.














Figure 3. Normalized flux detail for the last few of observing sessions, showing possible dust cloud dips (using AHS functions).

    _______________________________________________________  PREVIOUS FADE EVENTS    ________________________________________________________________


Figure 4. This is the 2020 January "no show" (no dip when expected). There's a brightening instead! Could the current "dip" be a "forward scattering" brightening from a dust cloud that is just off the line-of-sight? Or are we using a wrong P?

Comment: There is unusual variability that I didn't see during the 2019 October observations, so it might be real. The only way to assess this is to compare near-simultaneous observations. Until such a possibility is established the apparent variations will have to remain simple speculation.


Figure 5. There is an approximate repeat of the two dip structures, 428 days apart. The depth of the current one is ~ 70 % of the earlier one.


Figure 6. P = 107.1 days provides good agreement between this season's big dip and the one 4 periods ago.


AAVSO Observations

AAVSO observers submitted 249 observations of J0139 for the date interval 2019 Aug 28 to 2020 Mar 28. The best observations were made by observers with codes TRT, SPP and GBL. The other observations (observer codes VMT, DFS and LPAC) exhibited a scatter larger than the dip depths that are known to occur. The three best AAVSO data sets are shown in the next two LC graphs.


Figure 7.


Figure 8.
 
Notice that the transit due in late 2020 January did not occur. This was the 5th transit starting with the first one, the discovery transit using ZTF data (2018 Jul 30).
 

Physical Model Suggestion

The 2019 Sep/Oct dip appears to have a two-component shape, that somewhat resembles the 428-day shifted first dip (dotted trace in above LCs). The dip structure consists of an abrupt onset, or ingress, and has a U-shape that is symmetric until near the end of egress. Before egress is complete the second component becomes important and, based on the first dip (shifted 428 days), this second component exhibits a long and slow decline.

This might be understood as a spherical dust coma that produces the 6-day wide first component. The second component "dust tail" lasts ~ 16 days. The dust tail could be produced by small particles pushed outward (and at trailing azimuths) by radiation pressure. If the small particles have a specific lifetime function (due to sublimation?) then the tail should gradually end at some distance which we can associate with the 20 days that it is seen to exist (referenced to the center of the U shape).

There's a way to test this suggestion. If the main dip consists of large particles (e.g., radius > 1 micron) then this dip will have the same depth at all optical wavelengths, such as those observed by Zachery Vanderbosch (g'- & r'-band). If the second component "tail" consists of small particles (e.g., radius < 1 micron) then this dip should be shallower at the longer optical wavelengths.

Another noteworthy feature in the observations is a sinusoidal variation on the two observing sessions at the minimum brightness, Oct 02 and Oct 03. The period of these variations is 4.9 hours and the semi-amplitude is 11 %. If true, I have no idea what could be producing it.

Observing sessions

2020 Fall 2020 Dip (due 2020 Aug 20 and 2020 Dec 05)


2020.12.20  
2020.12.07  
2020.12.04  
2020.12.02  
2020.12.01  
2020.11.30  
2020.11.29  
2020.11.28  
2020.11.27  
2020.11.26  
2020.11.25  
2020.11.23  
2020.11.20  
2020.11.17  
2020.11.16  
2020.11.15  
2020.11.14  
2020.11.13  
2020.11.12  
2020.11.11  
2020.11.10  
2020.11.05  
2020.11.01  
2020.10.28  
2020.10.24  
2020.10.22  
2020.10.19  
2020.10.15  
2020.10.12  
2020.10.06  
2020.10.05  
2020.09.27  
2020.09.24  
2020.09.20  
2020.09.17  
2020.09.16  
2020.09.15  
2020.09.11  
2020.09.10  
2020.09.08  
2020.09.06  
2020.09.05  
2020.08.25  

2020 Winter Dip (due 2020 Jan 18)

2020.02.16  
2020.02.15  
2020.02.14  
2020.02.09  
2020.02.08  
2020.02.07  
2020.02.06  
2020.02.05  
2020.02.03  
2020.02.02
2020.01.31  
2020.01.29  
2020.01.28  
2020.01.27  
2020.01.24  
2020.01.23  
2020.01.20  
2020.01.19  
2020.01.18  
2020.01.14  
2020.01.12  
2020.01.11  

2019 Fall Dip

2019.10.16  
2019.10.15 
2019.10.14  
2019.10.13  
2019.10.12 
2019.10.11  
2019.10.09  
2019.10.08  
2019.10.07  
2019.10.06  
2019.10.05  
2019.10.03  
2019.10.02  
2019.10.01  
2019.09.30  
2019.09.29  
2019.09.28  
2019.09.27  
2019.09.23  
2019.09.20  
2019.09.18  
2019.09.16
2019.09.13
 
2019.09.12  
2019.09.11  
2019.09.10  
2019.09.09  
2019.09.04  
2019.09.02  
2019.09.01 
2019.08.28

______________________________   Observing Session Details   ___________________________


2020.12.20 

2020.12.07 

 

2020.12.04  



2020.12.02 

 

2020.12.01  

Full moon degraded SNR so I averaged groups of 10 images & processed them as usual.

2020.11.30  


2020.11.29  

2020.11.28  

I processed this long observing session as two segments to see if there was a trend within a 10-hour interval. There wasn't..


2020.11.27 

Image set processed two ways, and both give similar answer.


2020.11.26  

2020.11.25  


2020.11.23  


2020.11.20  

2020.11.17 



2020.11.16  



2020.11.15 



2020.11.14  




2020.11.13  



2020.11.12 




2020.11.11  


The 2.4-hr sinusoidal variation and slope are statistically significant.  A recovery appears to be in progress.


2020.11.10  

It appears that a dip has begun with a priod of 104.6 days.

2020.11.05 

 

2020.11.01 


Disregard the wrong date in the title.

2020.10.28 

2020.10.24
 
 

2020.10.22  


2020.10.19  



2020.10.15 

2020.10.12 

 
Disregard the date in the title.

2020.10.06  

2020.10.05  


2020.09.27 


2020.09.24  

2020.09.20 


2020.09.17  


2020.09.16  

2020.09.15 


2020.09.11  



2020.09.10  

2020.09.08  

2020.09.06  


2020.09.05 



2020.08.25  

Weighted average = 18.385 0.020

Starting now, 2020.08.25, I will use a different image procedure procedure. It emphasizes the small photometry aperture measurements.


_________________________________________________________   end of "Winter 2020"   ______________________________________________________________________

2020.02.16 


2020.02.15  


2020.02.14  

Finally, some good weather!


2020.02.09  

2020.02.08  

2020.02.07  

2020.02.06  

Very windy, so systematic error is likely.

2020.02.05  

Very windy, so systematic error is likely.

2020.02.03  

Very windy, so systematic error is likely.

2020.02.02  

2020.01.31  

2020.01.29  

2020.01.28 


2020.01.27 

2020.01.24  

2020.01.23 


2020.01.20 


2020.01.19  

2020.01.18  

2020.01.14 



2020.01.12  

2020.01.11  


    ___________________________________________________      PREVIOUS DIP    ____________________________________________________

2019.10.16  



2019.10.15 


2019.10.14  


Very noisy due to full moon.

2019.10.13  




2019.10.12 



2019.10.11  



2019.10.09 



2019.10.08  



2019.10.07  

2019.10.06  

2019.10.05  


2019.10.03  


Need someone else's LC to know if this sinusoidal variation is real.

2019.10.02  


I think these variations are real.


2019.10.01  



2019.09.30  

2019.09.29  



2019.09.28 


 

2019.09.27  




Still no evidence for target variability during observing session.

2019.09.23  


2019.09.20  


Still no evidence for target variability during observing session.

2019.09.18  



Still no evidence for target variability during observing session.

2019.09.16  


Still no evidence for variability during observing session.

2019.09.13  




Still no evidence for variability during observing session.

2019.09.12  


2019.09.11  

2019.09.10  



Still no evidence for variability during observing session.

2019.09.09  




No evidence for J0139 variability during observing session.

2019.09.04  



No evidence for J0139 variability during observing session.

2019.09.02  




No evidence for J0139 variability during observing session.

2019.09.01  




No evidence for J0139 variability during observing session.




Could this be a pulsation?


The sine fit is statistically significant (at 4-sigma), but it might just be a fluke "curiosity."  Could it be related to star rotation?

2019.08.28



v

Finder image

J0139 is located at RA/DE = 01:39:06.2 +52:45:37.


Figure F1. SDSS image, northeast upper-left.



Figure F3. SDSS image.



Figure F3. Image taken with my AstroTech 16" telescope. FOV = 12.5 x 9.0 'arc, north up, east left.


Figure F4. Some of the stars I use for reference. See below for r'-mags that I've adopted for these stars.

Star #
r'-mag
B-V-0.64
2
15.882
0.13
4
13.854
0.78
5
15.181
0.25
7
14.446
0.21
8
13.338
0.14
9
14.291
0.02
10
15.562
0.25
11
15.014
0.20
13
14.692
0.41
14
15.770
0.25
15
15.704
0.25
16
13.560
0.65
17
14.167
-0.02
18
15.088
0.58
19
12.828
0.22
20
15.852
0.25
21
15.887
0.19
22
13.495
-0.17


Observing and Analysis Tips for this Target

The target star, J0139, is faint (V = 18.4) so unfiltered is the best option. However, consider that the target is bluer than all nearby reference star candidates. The combination of "unfiltered with very blue target" means that systematic offsets will differ for each observer (because each observer's unfiltered effective wavelength will be different). Therefore we should expect that comparing measurements from different observers will require determination of an empirical offset for each observer. This is easy to do. To maximize the usefulness of this process it will be important that each observer adopt a FOV placement that is the same every night (so that flat field systematics, which everyone has, are the same for each observing session). Also, use the same reference stars every night.

 The goal is "day timescale" variations, not hourly or shorter timescale variations. With a 14" telescope, unfiltered, no full moon, 100-second exposures will yield SNR per image ~ 10 (i.e., 10 % SE). This SE per image is large compared with all other systematics. For example, scintillation is typically 5 to 10 mmag per image, or 0.5 to 1.0 %. It is also not necessary to keep the star field fixed with respect to the pixel field throughout the night to minimize flat field variations during an observing session.

The most important tip is to choose photometry setting carefully and stick to them for processing every observing session. This is because there's a brighter star close to the target star (9 "arc away), as shown in the next image. The danger we want to avoid is for atmospheric seeing changes to cause some of the flux of the brighter star from entering the signal circle and making the target star appear brighter than it is. If one observing session has bad seeing compared to another, there's a risk of the target reading for that night to be influenced by the nearby star. My rule of thumb is to choose a signal aperture radius that places the circle less than halfway to the interfering star.


Figure T01. Subset of a single image (3.5 x 3.0 'arc, north up, east left) showing photometry circles centered on the target. The brighter star within the gap annulus is 9 "arc away. All of its flux must be kept within the gap annulus in order to avoid some of its flux from entering the signal circle.

As an aside, you'll note three "hot pixels" in the above image. One of them is in the background annulus. This is OK because it will be ignored (by MaxIm DL, and presumably any other good quality photometry program). Hot pixels inside the signal circle cannot be tolerated, so placement of the FOV should take this into account.

My Collaboration Policy

Please don't ask me to co-author a paper! At my age of 80 I'm entitled to have fun and avoid work. Observing and figuring things out is fun; writing papers is work. If my data is essential to any publication just mention this in the Acknowledgement section.


References

Veras, Dimitri, Catriona H. McDonald and Valeri V. Makarov, "Constraining the Origin of the Planetary Debris Surrounding ZTF J0139+5245 Through Rotational Fission of a Triaxial Asteroid,"   https://arxiv.org/abs/2001.08223

Vanderbosch, Z., J. J. Hermes, E. Dennihy and 8 others, arXiv: https://arxiv.org/abs/1908.09839


Related Links

Wikipedia description of J0139 
WD1145 photometry monitoring 
Resume

____________________________________________________________________

This site opened 2019.08.30.