An Extension of the Pro/Am White Dwarf Monitoring Project

This web page describes a project inspired by an unexpected result of the PAWM project, during which we observed 46 white dwarfs (WDs) in a search for exoplanet transits. No transit events were discovered, but we found that 3 WDs were variable. One of them led to a publication (2013 SAS Conference) where it is argued that the 5 mmag amplitude, 2.7-hour variation could be produced by either a "hot Jupiter" exoplanet that reflects WD light or a starspot (we now favor the starspot explanation). The high incidence of variability among WDs was unexpected. A continuation of PAWM, called PAWM2, is justified since professional astronomers seem to have overlooked the merits of conducting photometric monitoring surveys of WDs and this task is easy for amateurs.
At the present time observers are expected to find their own WD targets for a night's observing. Please check this web page and the PAWM web page to be sure that a tentative target has not already been observed. Observers must process their own data and create a light curve graphical plot for submission to the webmaster. The principal goal for these observations is to identify WD exoplanet candidates that produce small amplitude variability (<8 mmag) with periods that could be produced by exoplanets in the WD's habitable zone (2 to 10 hrs). So far, 6 out of 66 WDs have been shown by PAWM1 and PAWM2 to be variable (9.1 3.7 %), so the variability search has a relatively high probability of success. A secondary goal is to observe an exoplanet transit! This goal is secondary because it is unlikely to be achieved.

You can find thousands of WDs at the Villanova White Dwarf Catalog web site. If no variability is seen on a single 6-hour or longer observing session I will consider moving that WD to the Completed Target List. If variability is present, or suspected, I will put it on the Current Target List for confirmation.
Links on this web page

    What's new  
    List of PAWM1 & PAWM2 WDs that have been observed  
    Link to PAWM1  
    Completed target list  
    PAWM1 & 2 variability lists & statistics  
    Related observing project  (M. Wilson, Univ. Iowa, 2003)
    PAWM2 light curves  
    Instructions for Choosing a WD to Observe and Preparing a Light Curve Graph for Submission
    PAWM1 Completed Observing List
    Related links  

What's New
Oct 29: Added list of PAWM1 & PAWM2 WDs

PAWM1 & PAWM2 WDs that have been observed

This is a complete list of PAWM1 & PAWM2 WD that have been observed.

User, select from Villanova White Dwarf Catalog

Completed Target List (PAWM2 observations, no more observations needed)

  0250+250, 13.2, 4927-Benni; smooth
  0232+525, 13.9, 4926-Benni; smooth
  1936+327, 13.6, 4901-Benni; smooth
  2122+282, 14.?, 4829-Ogmen; 9-hr variation is comparable to obs'g session length, so AMC could account
  1935+276, 13.0, 4825-Ogmen; known variable (ZZ Ceti type)
  2014+568, 13.6, 4809 & 4817 - Benni; smooth
  2032+248, 11.5, 4711 & 4712- Benni
  1713+695, 13.3, 4711-Benni

   2028+390, 13.4, 4624, 4629-Benni
  2136+828, 13.0, 4630-Benni
  1634+396, 14.1, 4622, 4623-Benni.
Smooth. No more data for this one needed.
  1632+177, 13.1, 4619, 4620, 4622 by Benni. Smooth. No more data for this one needed.

     1626+368, 13.8: 4531-Benni, 4601-Benni, 4606-Benni, 4607-Benni
     1232+479, 14.5: 4528-Benni
     1655+215, 14.1: 4524-Ogmen
     1756+827,  13.9: 4522-Benni, 4523-Garlitz, 4525-Garlitz. No more LCs needed.
     1900+705,  13.2:  4517-Bennni, clean (no variability or transit)

  0943+441, 13.3: 4416-Benni, "dip" near end associated with clouds.
  1748+708, 14.1: 4416-Benni, no variation.
  1104+602, 13.9: 4405-Benni (variation amp = 3.8 0.9 mmag, 3.6-sigma)
, Benni 4406 & 4408 too short to verify, but long 4409 showed no var'n.
  2006+481, 15.3: 3723-Garlitz, 3818-Garlitz; 3901-Garlitz, 3908-Ogmen, 3908-Salas, 3909-Garlitz
  1821+643, 15.0: 3707-Garlitz, 3708-Garlitz, 3709-Garlitz, 3710-Garlitz, 3713-Garlitz, 3714-Starkey, 3723-Starkey, 3819-Jensen, 3910-Starkey, 3914-Starkey, 3923-Starkey
  1735+610, 16.0: 3706-Garlitz, 3718-Garlitz
  1739+419, 15.1: 3704-Ogmen
  1647+375, 14.9: 3704-Garlitz

  1707+475, 15.2,
  1554+322, 16.0: 3628-Garlitz

    1829+547, 15.5: 3620-Starkey, 3626-Gary, 3627-Gary
1729+371, 16.0: 3615-Garlitz, (DAZd, dust disk, IR excess)
  1639+537, 15.1: 3528-Mendez, 3601 & 3605 & 3608 & 3610-Garlitz, 3605-Ogmen, 3611-Gary, 3611-Ogmen, 3612-Gary. No more observations needed.
  1430+136, 16.1: 3502-Garlitz, 3531-Wiggins
1302+597, 14.5:
3509-Garlitz, 3510-Garlitz, 3512-Gary
  1337+705, 12.8: 3504-Ogmen, 3507-Garlitz
  1211+332, 14.9: 3427-Gary
  1254+223, 13.3: 3418-Mendez
  1333+487, 14.0: 3419-Gary
  1408+322, 14.2: 3418-Gary
  1314+293, 12.7: 3427-Mendez, 3429-Gary 
(very hot DA (100,000 K) with M-dwarf binary)
  1213+528, 12.6: 3403, 3404, 3406, 3407, 3408, 3414-Gary
  1134+300, 12.7: 3401-Gary
  0117+383, 19.1: 1930-Gary
  2359-434, 13.0: 1a02-Curtis
  1036+433, 11.5: 3326-Gary

Note: I abbreviate dates using hexadecimal notation. Thus, 1a02 means 2011 October 2 (and 3401 = 2013 April 01).

PAWM1&2 Variability Results: Links Describing PAWM2 WDs Found to be Variable or Other Interesting Stars Found Nearby (external to this web page)

Among the 36 PAWM2 WDs targeted there are 4 variables (1213+528, 1302+597, 1639+537, 2006+481), 1 EB, 1 RR Lyr variable.
Among the 48 PAWM1 WDs there were 3 variables (2359-434, 2229-311,2326+049).
PAWM1&2 stat's: 85 stars & 7 variable, or 8.3 3.1 % of WDs are variable! Variability is most likely due to star spots (e.g., WD2359-434).

Note: At web site I show the following empirical relationship for "spurious sinusoidal variations", which I will use in assessing credibility of sinusoidal fits to the PAWM2 light curves on this web page:

            Variation is Possibly Real if Amplitude > 1.45 (RMS10 - 0.2 mmag).
PAWM2 Variables:
    WD1756+827 a spurious "transit feature" has been explained (bad master dark frame).
    WD2006+481 shows variations that come and go. Even the period and amplitude vary. This star is a puzzle. Time-out to think!
WD 1213+528 (EG UMa)
is a known variable binary (white & red dwarfs). It's not what we're looking for (since it's a known binary).
    NSV 19335 is an irregular variable of a very red star It's not what we're looking for (since it's a red dwarf).
    WD1302+597, variable, 6 mmag, 3.5 hrs - BINGO! This is the kind of variable WD that PAWM2 is looking for! However, WD expert Bergeron reviewed our web page for this object and wrote "I will let you know if I ever go back to study this star in the future." He also said he'd keep in mind others who would be interested in it.
    WD1639+537, another variable: 8.2 mmag, 1.9278 hours. This WD variable can't be an exoplanet candidate because it has spectrae that vary with the same 1.93-hr period as its brightness. No more observations needed. WD expert Bergeron reviewed our web page for this object and wrote "...this magnetic WD is DA & therefore probably H rich; because of dipole field the variation probably due to a spot." (my condensed version).

Related Observing Project

Michael Wilson, a student at the University of Iowa's Department of Physics and Astronomy, observed 75 white dwarfs with a robotic telescope for his master's thesis (submitted in 2003). No transit features were found. Among the total of 75 WDs, only 49 were observed for > 5 hrs, and 33 were observed for > 7 hrs. A few of these targets overlap with the PAWM1 & 2 observed list (nr = 14). As of 2014 August 17, the total number of WDs from both lists that have been observed for > 5 hours is 133, and
the total number of WDs from both lists that have been observed for > 7 hours is 117.

According to the calculations in M. Wilson's thesis, if all WDs have a planet in orbit between the Roche limit (23 planetary radii) and 0.01 a.u. (33 planetary radii), where orbital period ranges from 4.5 hrs and 8 hours (and transit length ranges from 3.6 min and 4.2 min), geometry considerations lead to the probability for transit of ~ 1.2% per WD. Therefore, with 117 stars being observed without transit for > 7 hrs the probability for encountering one transit is ~ 75%.

Variability at the level expected for a starspot (a few mmag) is not possible with the Wilson data because they are too noisy. Only one WD was identified as possibly variable, but the variability was not sinusoidal (and it was too large for a starspot explanation). Therefore, only the PAWM1&2 WDs can be used for placing limits on starspot variability, and as of this date this list shows that 7 out of 84 stars are variable (
8.3 3.1 %) at the ~3 mmag level.

Paul Benni sent me a copy of M. Wilson's thesis (thanks Paul!). Here's a link to a brief description of this project: 

PAWM2 Light Curves
(most recent at top)




No variations.


No variations.


Air mass curvature (AMC) is large, and might account for a residual variation with P ~ observing session length.


This is a known variable, a ZZ Ceti type (non-radial pulsations) that occur for all (?) WDs as they cool through the 11,000 to 13,000 K region. McGraw et al (1981) report a complex period structure, with frequencies ranging from 0.5 to 14 mHz (80 sec < P < 500 sec).  This isn't the type of variable WD that PAWM is looking for (either starspot or exoplanet related).


Same WD, better observing conditions, confirming lack of sinusoidal variation.

This sinusoidal fit is probably "spurious," in spite of the 2.0-sigma, based on my empirical relation (described above). 


My tentative conclusion for this WD is: I don't know if WD2032+248 undergo sinusoidal variations in the period interval that we can sample. The case for there being no variations is based on one night (Jul 17) when two telescopes were used by the same observer (same clear filter), showing statistically significant variability with the 8-inch telescope (5.1
0.8 mmag) but no apparent variability with the 11-inch telescope (0.45 0.43 mmag). On the other hand, a similar variability was observed on back-to-back nights (Jul 21 & 22) using the 11-inch telescope (but with different filters, Cb and B). Both had an amplitude of ~ 1.6 mmag and P = 2.1 hrts, but only one was statistically significant. 

B-band, so is noisier. Variability is comparable to that for the previous night (Cb-band), but is not statistically significant.

Statistically significant sinusoidal variations, but having longer period than the previously significant variations.

No variability.

Same night, same observer, different telescope, same unfiltered: There's no statistically significant sinusoidal variation (0.45 0.43 mmag), yet the noise level is better (notice that the magnitude scale is magnified compared to the 8-inch telescope light curve (below this one). This must mean that the 8-inch variations were systematic, and not produced by the star.

This sinusoidal variation is statistically significant; however, look at the next plot (above this one).

Same data as in plot two below this one, but using 2 ref stars with same blue color as the WD. Sinusoidal variation is same, but is not significant.

Same data as in plot two below this one, but using 2 ref stars with same blue color as the WD. Sinusoidal variation is same, but is only marginally significant (2.1-sigma).

This follow-up observation had to use I-band filter due to full moon nearby. No new info about the short period variability. Used 3 ref stars, one was red the other two were blue (like the WD).

The 0.7-hr variation has marginal statistical significance (2.1-sigma).
Used 3 ref stars, one was red the other two were blue (like the WD).


No variations or transit feature.



Smooth. No evidence for short term variations.


Smooth, no need for more observations.


We can consider this WD to be a non-variable.

Smaller aperture means more noise.

Smooth (all data shown).


The first two LCs (Ma 31 & Jun 01) were suspected (by observer Paul Benni) of exhibiting brief fades at intervals of 1.2489 hours with depths that were slightly greater than expected on the basis of stochastic noise. The two follow-up LCs (Jun 05 & 06) didn't show any similar features. This WD is metal-rich and suspected of having a debris field, so it was reasonable for observer Benni to search for short and shallow fades caused by small, asteroid-size objects. This object is tentatively assigned "completed" status, but I suspect Benni will observe it on occasion in the future. Good work, Paul!


No transits, but only 5 hrs; no sinusoidal variation. More data needed.


No transits & no sinusoidal variation.


No repeat of transit.

No events for this 6.5-hr observing session. This is a high priority target (explanation later).


Good quality data with no features. Add one more to non-variable category.


Paul Benni notes that clouds were probably present when the "dip" occurred at 6.0 UT, based on sky background level, and that the final fade could be related to a brightening dawn sky. Because of these associations I'm going to declare that this LC is "normal."


No sinusoidal variation present, and no transit features (spectral class DXP).


Good LC; shows no evidence for variation. I think the Apr 5 variation was statistical noise.

Obs'g session length too short for verifying 3.8-hr periodicity.

Obs'g session length too short for verifying 3.8-hr periodicity.

This sinusoidal variation has marginal statistical significance, and deserves follow-up observations (Benni has 3 more LCs submitted, awaiting processing).
I note that my empirical threshold for variability being real is for:  Amp > 1.45 x (RMS10 - 0.2 mmag). According to this the variation is probably not real.


Two LC versions. The "pro" version (below) makes it easier to see the 1.4 mmag semi-amplitude variation, which is not statistically significant (1.3-sigma).

Two LC versions. The "pro" version (below) makes it easier to see the 2.0 mmag semi-amplitude variation, which is almost statistically significant (1.9-sigma).


 No variations (of the 2 or 3-hour type). What going on with this star?

 No variation (but noisy, due to weather).

  No variation!


 data  The 2.2-hr variation is becoming believable! Two out of 2 LCs show it.

  data  This variation is statistically not significant, but I'd like to see another LC.


 Sinusoidal variation not statistically significant.

 data   Sinusoidal variation not statistically significant.

Re-analysis of original images.

   data   Variation is statistically significant (2.4-sigma). Amplitude is larger at this shorter wavelength.

   data   This variation is statistically significant, and resembles the variation seen in the Jul 7 LC - but not the same as seen in the previous two LCs.


   data   Insufficient SNR for confirming previous LC's variation. Let's try again.

   data   Good! This variation might be real. Let's observe it some more!


OK, I'm now convinced that there is no variability at the 3 mmag level.

No transit features & variation is not statistically significant. Need at least one more LC.


The variation is not statistically significant.


No variations


No variation!


This is a pretty faint WD for a 12-inch telescope. It appears to have no variability.


Different "variation" each of 3 observing sessions; no evidence for a consistent variation - so abandon this WD. [Ignore the title; this is really an observation of WD1829+547.]

   data    Sinusoidal amplitude has SNR = 2.0, so it's on the verge of being statistically significant. Worth more observations.

Sinusoidal amplitude has SNR ~ 1.7, so it's technically not statistically significant. Worth more observations.





Definitely insufficient SNR (small aperture & use of filter).





   data   Another winner! Probably a starspot but worth more observations!

First obs'n of this target; Manuel apologized for poor quality and vowed to improve with clearer skies.




There appears to be a strong "air mass curvature" present, due to the WD being bluer than the reference stars. However, this shape differs somewhat from the previous night, so maybe there's a real variation lurking.

There appears to be a strong "air mass curvature" present, due to the WD being bluer than the reference stars.


Supports Ogmen's non-variability observation.

Shorter LC than desired, but good data showing no variability. Need one more LC.


This "air mass curvature" is consistent with the target being much bluer than the reference stars.



   1333+487: 13:36:02 +48:28:44, V=14.0, binary (DB+dM), 2.9" sep'n

As an aside, a nearby star was found to be an eclipsing binary. These are "vermin in the sky" for PAWM2:

Nearby EB. Any depth > ~ 40 mmag for a non-dwarf star is almost certainly an EB eclipse (not an exoplanet transit).



Air mass curvature suggests star is bluer than ref stars, which is expected given WD's 100,000 K Teff.

Dew on corrector plate, middle of obs'g session; then wind changed & dried plate. Perhaps light pollution present. (Curvature correlated with air mass, and can probably be attributed to star color issues.) This blue star should have opposite air mass curvature (using V filter).

Observations discontinued because I learned that this is a well-studied binary using another name (EG UMa)








[Will be updated soon]

[Will be updated soon]






Instructions for Choosing a WD to Observe and Preparing a Light Curve Graph for Submission

If the "Current Targets" list does not include a WD that is suitable for you to observe (e.g., too faint), then a WD candidate for observation can be chosen from the Villanova White Dwarf Catalog. Choose the Villanova listing corresponding to your midnight LST (local sidereal time) and search for WDs with a declination and brightness that's observable. Then check the above two listings of already observed WDs to be sure you haven't chosen a target that has already been observed sufficiently.

After processing your image set, and performing photometric measurements, create a graphical light curve plot for submission to this web site's webmaster. Any spreadsheet can be used, or you may use any program that creates a graph from data (e.g., Gnuplot). It's not necessary to "fit" the data, or even to overlay group averages, to see whether or not variability or a deep transit is present. 

I'd like the LC plot that's submitted to have an x-axis range of at least 7 hours. A longer one will be necessary when the observing session is longer, of course.
The x-axis can have units of hours, or JD (HJD is OK). If JD is used, then make the range at least 0.25 days (if it's possible).

I'd like the y-axis range to be 0.1 magnitude, unless the data is so good that a smaller range seems appropriate. If a y-axis range of more than 0.1 magnitude is needed, then use the larger range (or consider not submitting the LC if a large range is needed because the data is noisy).

A title would be very useful, and it should include a UT date corresponding to the UT start of observations (preferred format YYYYMMDD). The title should also include the WD target name (e.g., WD1234+567), and the observer's name. All other things are optional: observer country, JD range, length of observing session, size of telescope, filter used, exposure times, etc.

Observers who have to perform meridian flips should indicate that one was performed on the LC plot (if that's possible). This can be done using an information box; if information boxes are not possible then use the title for stating when the flip occurred. It would be nice to include how big a flip correction was applied.

If your graphical program does not permit information boxes or titles, then in the e-mail submission please include a caption for your LC that includes all of the above required information.

If you're unsure about the suitability of a LC plot, just submit it and I'll give you feedback.

Submit graphical LC plots as an e-mail attachment (or embed) to: 

PAWM1 Completed Observing List (no more observations needed)

PAWM Targets already observed sufficiently.

Related Links 
PAWM (also referred to as PAWM1) Project similar to PAWM based in Victoria, Canada using a professional 1.85-m telescope.

Thanks to Manuel Mendez for creating this animation for PAWM.

References:  More support for WD debris disks (& therefore planets)

B. Gary.  This site opened:  2011.10.04. BGary Web Sites