Remarks on the Calibration of the Alpha-Particle-X-Ray Spectrometers on board 
the MER Rovers

R. Gellert, APXS Science Team
3 February 2006

The APXS on both MER rovers don't carry a useful calibration target
for the x-ray channel. The dedicated calibration target on the inner
side of the dust protective doors consists mainly of a thin gold layer
glued on top of the Copper doors. It was specially tailored to fulfill
the needs for the calibration of the alpha particle channels. But the
lowest x-ray lines for gold at around 2.2 keV are too high in energy
to be sensitive to reveal unambiguously subtle changes in detector
properties (either contamination or other changes in the detector
properties). Furthermore it was decided early in the mission to leave
the doors open due to a cumulative unreliable mechanism on the APXS
sensor head.

Therefore the soils at both landing sites were used to monitor the
performance and the calibration of each APXS instrument. This was
triggered by the discovery that the soils on all five known Mars
landing sites were very similar (due to global mixing by Aeolian
processes) and the observation that the composition didn't change much
along the traverse of both MER rovers (except for obvious soil
anomalies mostly in disturbed subsurface soils).
 
Recently it was discovered and proven that an obvious incident
happened to the MERB APXS instrument sometime between sol 250 and sol
368. These were the dates when the APXS measured soils on MERB. The
ratio between the low-z elements Na, Mg and Al and the Si peak areas
are plotted in Figure 1.  This ratio allows comparison despite the
admixture of various amounts of Fe rich spherules in the soils of the
MERB landing site.

It is obvious that in this timeframe a change in the properties of the
detector occurred.  Figure 1 shows that the light elements Na, Mg and
Al are clearly enhanced after sol 250 versus Silicon. The enhancement
shows a certain pattern. The enhancement is bigger for Na than for
Mg. The Mg enhancement is in turn higher than Al. A pattern like this
indicates a change in the detector properties, i.e., by removing an
absorbing element in the path of the detected low energy x-rays. The
fact that the ratios are nearly constant before and after sol 250 also
points to a one time change of the detector properties. The constancy
of the results before and after made it necessary to wait till enough
spots were measured. Recent calculations revealed that the changes can
be simply explained by the removal of 1 bar N2 atmosphere within the
detector assembly. The x-ray detector chip is sealed by glass
feed-throughs and a Be window from the external environment. This
compartment is filled, as a standard practice, with 1 bar of N2 by the
vendor after detector assembly. It seems that either an electrical
feed-through, the housing, or the Beryllium window developed a gas
leakage after sol 250. This incident removed the 1 bar of N2 and
filled the interior of the detector with martian atmosphere. The
derived elemental ratios before and after the incident lead to a
thickness of 1.9 mm of the N2 column. This is in good agreement with
the internal mechanical setup of the x-ray detector. A complete
removal of the Be window can be excluded as it would have had a
greater influence on the results with a different energy pattern.

Derived correction factors of the measured peak areas before and after
the incident are given in Figure 2.

Taking these correction factors into account, the results of the
outcrop analyses show that the composition of the outcrops did not
change significantly after Opportunity left Endurance Crater, which
was about the time when the incident happened.  This conclusion is in
good agreement with all other instruments on the MER rover. Without
the correction a jump of approximately 80 % in Na and a significantly
increased Si/Fe ratio would have been observed in the APXS results.

A final correction table of the peak areas is pending until all
results of the science investigations on board the Opportunity Rover
are taken into account.

The incident bears no challenge for the x-ray detector itself. The
FWHM as well as the peak to background ratio of the APXS before and
after the incident is constant. Because the martian atmosphere is very
dry, no oxidation inside the detector is expected. The fact that
measurements during the morning hours do not show excessive noise due
to light indicates a very small gas leakage only in the x-ray detector
housing.