A Low Upper Threshold for Saltation-Mediated Triboluminescence at Gale Crater, Mars H. M. Sapers1, J. L.
Kloos2, M. Baker3, D. M. Fey4, H. Kalucha5, M. Lemmon6, M. Minitti7, C. Newman8, J. E. Moores1, 1York University Centre for Research in Earth and Space Science (hsapers@yorku.ca), 2University of Maryland, 3Smithsonian
Institution, 4Malin Space Science Systems, San Diego, CA 92191-0148, USA 5California Institute of Technology,
Division of Geological and Planetary Sciences, 6Space Science Institute, 7Planetary Science Institute, 8Aeolis research

Introduction: Laboratory studies suggest that saltation-driven grain-to-grain collisions are capable of
producing triboelectric discharge in the hyperarid,
near-surface atmosphere on Mars 1,2. Saltation simulation experiments induced triboelectric discharge with a
radiance of approximately 1.4 - 3.2 µW/m2/sr under a
simulated Martian-atmospheric composition at 8 mbar
pressure - capable of ionizing methane and providing a
potential mechanism to explain the enigmatic seasonal
variations of CH4 and O2 3. With observational evidence of wind-mediated saltation 4–7, saltation-induced
triboelectric discharge is theorized to be common on
Mars, but has yet to be observed in situ. We successfully carried out the fist in situ triboelectric observations carried out by the Mars Hands Lens Imager
(MAHLI) onboard the Mars Science Laboratory
(MSL) Curiosity Rover setting an upper threshold for
triboelectric discharge that would not have a measurable effect on near-surface atmospheric composition.
The observed seasonal variation of CH4 and O2 in
the Martian atmosphere is not fully explained by currently understood atmospheric or geological processes.
The apparent correlation between CH4 and O2 variation
suggests a potentially shared mechanism accounting
for an increase in volume mixing ratios (VMR) during
the northern spring/summer followed by apparent
destruction during the northern summer and autumn
(figure 1). An empirical study simulating saltating
grains 3 indicated that saltation mediated triboelectric
discharge is capable of ionizing argon, and therefore
also CH4 and O2, providing a testable hypothesis for a
mechanism behind CH4 and O2 depletion during dust
storm season beginning ~Ls 180°–210° and ending
~Ls 330°–360° during the norther autumn and winter
(figure 1). Methane is of particular interest as ~70% of
methane of Earth is biogenic 8, as such, enigmatic
sources and sinks of methane on Mars hinting at as of
yet unknown processes remain a tantalizing biosignature potentially indicative of putative past or extant
biological activity in the subsurface 9. Subsurface
methanogenic microorganisms on Earth produce methane through anaerobic chemoautolithic metabolisms,
consistent with physicochemical conditions in the
putatively habitable Martian subsurface 10.
Curiosity performed a ‘Sands of Forvie’ (a sand
sheet is approximately 400 m x 1000 m in extent)

observing campaign between Sols 2989 and 3213
during the end of the windy season providing an opportunity to perform a night observation optimized to
detect putative triboluminescence provided that it is
occurring at the rate predicted by 3. Further, we quantify an upper threshold if triboluminescence is not
observed.

Figure 1: Normalized volume mixing ratios of Ar, O2,
and CH4 in the near surface Martian atmosphere as
measured by the Sample Analysis at Mars (SAM) instrument onboard Curiosity over Mars years 31-34
indicating the enigmatic seasonal cycle of CH4 and O2
with a decrease during the dust storm season. Data
from 11,12 (top). MARDI change detection observations
from Mars year 31-34 indicating observed grain
movement (red) and no observed grain movement
(black; bottom). The dotted line indicates the Ls of the
triboelectric observation at the end of the windy season.
Observations: The Mars Science Laboratory
(MSL) rover, Curiosity, performed a ‘Sands of Forvie’
observing campaign between Sols 2989 and 3213
during the end of the windy season providing an opportunity to perform a night observation optimized to
detect triboluminescence. The Sands of Forvie sand
sheet is approximately 400 m x 1000 m in extent. Due
to a combination of spectral range, sensitivity, field of
view, and operability, MAHLI 13 was selected for the

observation. Three main observational experiments
were executed:
1) Triboluminescence observation, (Sol 3017):
Seventeen 60-second full-frame MAHLI images were
acquired at night (without Phobos in the sky) to capture a signal of saltation-mediated triboluminescence.
As documented in a corresponding MAHLI image
acquired during daylight hours, the scene included the
sand sheet, foreground rocks, the local horizon, and the
sky (figure 2);
2) Wind assessment (Sols 3022, 3024): A set of
change detection images, view acquired with the fixedpointing Mars Descent Imager (MARDI) camera, were
used to assess wind/saltation conditions present around
the time of the triboluminescence experiment;
3) Phobos-shine observation (Sol 3215); Seventeen
60-second exposures were acquired of the rover’s
remote sensing mast (RSM) with the white RSM angled to reflect Phobos-light. A corresponding daytime
context image was also obtained (figure 3).
Image processing: Images were downlinked as
losslessly compressed raster 8 bit files using first difference Huffman compression. Following decompression, the data were decompanded into 12 bits and
stored as 16 bit integers. Radiometric calibration was
achieved using the standard MAHLI spatial domain
calibration pipeline [9] including dark current compensation, shutter smear migration, and flat field correction. The radiometrically calibrated and color corrected
prepointed daylight images were used for masking.
Triboelectric observation: Four areas on the daylight pre-pointed image were identified and masked
using RGB band ratios: sky, foreground, sand sheet,
and the Greenheugh pediment (figure 3). The brightest
and darkest 1/6th pixels were masked in each of the 17
images in the night sequence, all 17 images were then
stacked and averaged. The mean digital number (DN)
value was then calculated for each of the masked areas
identified in the pre-pointed image.

uted to tolerance in repositing the rover arm precluded
direct masking based on the daylight image. The
brightest and darkest 1/6th pixels were masked in each
of the 17 images in the night sequence, all 17 images
were then stacked and averaged. Using the signal in
the processed Phobos-shine image, the area of Phoboslight reflected in the RSM was masked and the average
DN value for the green channel was calculated after
applying a median filter.

Figure 2: Pre-pointed daylight image (L) and average
Phobos-shine image (R). Notice the faint signal acquired from Phobos-light reflecting off the RSM.
Observational Radiance calculation: The radiometric calibration pipeline outputs a radiance scaling
factor to convert digital number values to units of I/F.
As input flux is not directly relatable to DN value for
MAHLI, A cross calibration method developed to
obtain MAHLI radiance values through calibration
with MastCam M34 was used [10].

Figure 4: Mean digital number (DN) values for all
pixels in each of the ROIs: sky, pediment, sand sheet,
and foreground in the triboelectric observation indicating no significant difference between each area.
Error bars 1σ.

Figure 3: Pre-pointed daylight image (L) and average
triboluminescence image (R ) indicated no increased
signal in the sand sheet attributable to triboluminescence.
Phobos-Shine: An offset between the pre-pointed
daylight image and the 17 frame night sequence attrib-

Theoretical Radiance calculation: Theoretical radiance limits were also derived for comparison to the
MAHLI triboelectric observational data. The following
radiometric considerations that were used to produce
these theoretical limits for the MAHLI G-band (500600 nm; referred to a ‘in-band’). Two quantities are
estimated: (1) the solar radiance emitted directly by
Phobos at the time of the “Phobos sky” observation,
and (2) the radiance emitted by the RSM while illumi-

Empirical triboelectric discharge
1.4-3.2 uW/m2/sr

Theoretical
upper limit

Observational
upper limit

0.23 uW/m2/sr

0.594 uW/m2/sr

Table 1: Empirical 3, and theoretical and observational radiance limits for triboelectric discharge at
Gale Crater.

nated only by Phobos-shine at the time of the “Phobosshine” observation. The elevation of Phobos during the
observation was 36 deg and azimuth of 278 deg.
Phobos-sky: The radiance [W/m2/sr], Lp, emitted by
the surface of Phobos assuming Lambertian scattering
is given by eq 1:
(1)
where F = 185 W/m2 is the in-band solar flux at 1AU,
d = 1.6 AU is the Phobos-Sun distance at the time of
the observation, and Ap = 0.026 is the albedo of the
surface of Phobos at 550 nm 14. To estimate the radiance that would be observed at the surface of Mars,
Lp’, the atmospheric attenuation must be accounted for
using the Beer-Lambert law in eq 2:
(2)
Where = 0.34 is the atmospheric column opacity and
is the cosine of the angle of incidence that accounts for
the extended path length through the atmosphere. The
Phobos-emitted radiance observed at the surface of
Mars is therefore estimated as 0.28 W/m2/sr; normalized by the idealized Lambertian radiance (i.e. the
solar radiance emitted by a normally illuminated ideal
Lambertian surface) yields an I/F of 0.013.
Phobos-shine: the radiance reflected from the RSM
can be estimated using the Phobos-emitted radiance Lp'
calculated above. The flux incident on the RSM, FRSM,
due to Phobos-shine is given as FRSM = Lp'Ω, where Ω
is the solid angle of Phobos’ disk with respect to the
Martian surface. The radiance reflected from the RSM
toward the MAHLI imager can then be found using an
assumption of Lambertian scattering and the albedo of
the RSM, assumed here to be 0.9 15 assuming a negligible change in reflectance due to dust settling. The
incident flux on the RSM is calculated as 1.3 W/m2
and the RSM-emitted radiance as 0.2 W/m2/sr.
It is noted the that observational and theoretical radiance values for Phobos-shine reflected off the RSM
agree with each other, supporting the observationally
derived upper limit for triboelectric discharge of ~0.6
uW/m2/sr.

Establishing an upper threshold: There was no
significant difference in the mean DN values between
the four areas in the triboluminescence observation
indicating that, if occurring, triboluminescence within
the Sands of Forvie sand sheet was occurring below
the radiance detection limit of MAHLI (figure 4). The
absence of detection necessitated using the Phobosshine observations to quantify an upper limit for putative triboelectric discharge assuming that Phobos-light
reflected off of the RSM comprises the lowest signal
capable of being detected by MAHLI. Using the
MastCam cross-calibration pipeline, the radiance of
reflected Phobos light is calculated as ~0.6 W/m2/sr
(table 1). Given the ionization energy of CH4 of 12.61
eV, and assuming isotropic conditions, 100% ionization efficiency, and that CH4 is well mixed in the lower
10-100 m of the atmosphere, then if occurring and
producing a maximum radiance of ~0.6 W/m2/sr,
triboelectric discharge could ionize only 1 part in
1,000,000 to 1 part in 100,000 of the CH4 emitted each
night assuming a subsurface emission rate 5.6E15 CH4
molecules/sol 16.
Implications for the near-surface atmosphere: Using ~0.6 W/m2/sr as an upper limit for triboelectric
discharge in a sand sheet near the end of the windy
season at Gale crater, the number of molecules of CH4
and O2 that could be ionized are insufficient to have an
appreciable effect on bulk methane or O2 variability or
explain the observed seasonal variation indicating that
triboelectric discharge alone cannot account for the
observed seasonal decrease in CH4 and O2 after ~Ls
150°. Several factors may explain the discrepancy
between the empirical predictions and the observations. It is not currently known how often winds in
Gale crater would reach the 21 m/s speed used in the
empirical studies. Although there was some grain
movement detected in the MARDI images, it was
limited compared to the most active part of the windy
season 6 and ripple migration was not observed suggesting low overall saltation fluxes. The time elapsed
between the triboelectric observation and change detection observation precludes knowledge of the wind
conditions during the main observation. Finally, the
MastCam M34 cross-calibration pipeline relies on the
I/F value calculated using the radiance scaling factor
17
. The radiance scaling factor assumes Solar illumination and may not be valid for night observations. To
mitigate this challenge future work will model the
predicted the radiance of reflected Phobos-light to
quantify the signal captured by MAHLI during the

Phobos-shine observations providing an independent
upper threshold on triboluminescence. Future observations of opportunity when sand sheets are in the
MAHLI field of view at serval times during the windy
season would place more robust limits on the potential
occurrence of wind-mediated triboluminescence.
Conclusion: No evidence of saltation-mediated triboelectric discharge occurring above ~0.6 W/m2/sr
was detected within a sand sheet at Gale crater at the
end of the windy season (Ls 356.3°) and triboluminescence is unlikely to be widespread enough, or producing a high enough flux to affect measurable concentrations of methane or oxygen in the near-surface atmosphere. The diurnal and seasonal variation of CH4 remains enigmatic and a potential biosignature indicative
of putative past of extant subsurface biological activity.
Acknowledgments: The authors wish to thank Ken
Edgett for assistance in designing the MAHLI observations and data analysis discussions, Sean McNair and
Chase Million for assistance with MAHLI calibration
and image processing.
References:
Krauss, C.E., Horányi, M, M. & Robertson, S.
New J. Phys. 5, 70–70 (2003).
2. Krauss, C.E., Horányi, M, M. & Robertson, S
Journal of Geophysical Research 11, 8 (2006).
3. Thøgersen, J., Norskov Bak, E., Finster, K.,
Nørnberg, P. & Jakobsen, H.J. Icarus 332, 14–18
(2019).
4. Greeley, R. et al. J. Geophys. Res. 111, E02S09
(2006).
5. Geissler, P.E. et al. Journal of Geophysical Research: Planets 115, (2010).
6. Baker, M.M. et al. Geophys. Res. Lett. 45, 8853–
8863 (2018).
7. Bridges, N.T. et al. J. Geophys. Res. 110, E12004
(2005).
8. Conrad, R. Environmental Microbiology Reports
1, 285–292 (2009).
9. Krasnopolsky, V.A., Maillard, J.P. & Owen, T.
Icarus 172, 537–547 (2004).
10. Harris, R.L. et al. Sci Rep 11, 12336 (2021).
11. Trainer, M.G. et al. Journal of Geophysical Research: Planets 124, 3000–3024 (2019).
12. Webster, C.R. et al. Science 360, 1093–1096
(2018).
13. Edgett, K.S. et al. Space Sci Rev 170, 259–317
(2012).
14. Pajola, M. et al. ApJ 777, 127 (2013).
15. Wiens, R.C. et al. Space Sci Rev 170, 167–227
(2012).
16. Moores, J.E. et al. Nat. Geosci. 12, 321–325
(2019).
1.

17. Liang, W. et al. Icarus 335, 113361 (2020).