PIXL instrument chart sample Credit:NASA/JPL-CALTECH

Of course PIXL has been used already in Jezero crater, but as far as I know only to identify mineral compositions and textures, as described in this Lunar and Planetary Science Conference paper.
The article How vanadium could be a beacon for Martian life shows that knowledge about the presence and/or quantity of some trace elements could be important.
According to the page for scientists PIXL could even detect important trace elements such as Rb, Sr, Y and Zr at 10's ppm level.

  • 2
    $\begingroup$ I don't get your question. As you say PIXL has been used to identify mineral compositions, and trace elements are part of the composition. What are you saying hasn't been done? $\endgroup$
    – GdD
    May 16, 2022 at 14:03
  • 3
    $\begingroup$ You can start your own investigation right now, all the data is readily available! pds-geosciences.wustl.edu/missions/mars2020/pixl.htm $\endgroup$
    – asdfex
    May 16, 2022 at 16:56
  • 3
    $\begingroup$ It's a bit too much work for me right now. I suggest checking the documentation "RDR SIS" for the "RPM" data format. That's basically a table for 22 elements for each scanned pixel, including Cr and Ti. First plots can easily be made with any spreadsheet software. The main issue is that the measurement time is too short to see trace elements properly, so you need to average over a large part of each image. $\endgroup$
    – asdfex
    May 16, 2022 at 19:50
  • 1
    $\begingroup$ Quite possibly because identifying minerals gives a bigger picture of the geology of a region. Trace elements are minor elements that may add intrigue. $\endgroup$
    – Fred
    May 16, 2022 at 20:07
  • 3
    $\begingroup$ @OscarLanzi This is not rock, but NIST 612 reference material. Magnesium shows up just right of Sodium - but at a 30x smaller peak, so it vanishes in the gap between Sodium and the unlabeled Aluminum peak. $\endgroup$
    – asdfex
    May 23, 2022 at 9:32

2 Answers 2


As @Cornelis noted, PIXL data are readily available for download. Here I used data taken on Sol 560. At this link you can find a visual image of the rock as well as pixel-resolved spectra. Obviously, individual pixel data have much lower count rates than the integral of all pixels and are more difficult to analyze. To get a quick impression I only plotted the integrated data - x-axis is the energy in a range between 1 keV and 18 keV and the y-axis shows the count rate for each energy bin.

Histogram of one x-ray scan

For reference I added the main spectral lines (K$\alpha$) of elements between Sodium and Molybdenum at the top of the plot and labeled a few secondary lines on the bottom. Elements lighter than Sodium don't have spectral lines that are energetic enough to be seen in the detector. Heavier elements can be detected by their secondary lines (e.g. L$\alpha$) but these are hard to identify without further work. You might notice some peaks that I left unlabeled for this reason.

To really extract information about trace elements one needs to do several additional steps of data processing. The most important one is to subtract the background from the plot in a way similar to that shown in the plot shown in the question. Then one needs to subtract the contributions of all spectral lines of already identified elements as well. This needs information about the relative intensities of lines which are known, but not available from simple x-ray emission tables like the one used to add labels.

  • $\begingroup$ Fine, now I also know what I did wrong to find the MSA files in the Mars 2020 Analyst's Notebook. I've read somewhere that it's just a matter of how much time you're willing to spend in beaming the X-rays on a specific target to be able to distinguish the trace elements. Btw, the top near the Kr line, could that be from thorium (Th) ? $\endgroup$
    – Cornelis
    Apr 23, 2023 at 15:31
  • 2
    $\begingroup$ This peak is at 12.75keV, too low for Thorium. You're right that more time helps: It's just a matter of statistics to get better data. Noise rises only with the square root of counts, so peaks are easier to spot. $\endgroup$
    – asdfex
    Apr 23, 2023 at 16:47

Thanks to @asdfex for his link to PIXL data in the PDS Geosciences Node.
There's a link to the Mars 2020 Analyst's Notebook tool with the PIXL instrument's data archive.
In the archive I found two MSA files each containing so-called histogram bulk summed spectra (RBS). (page 41 for explanation)

Each of the 4096 channels of the PIXL instrument that counts the re-emitted X-rays from a target, has its own unique 8 eV range and they fit together to cover the total 32.768 keV energy range.

The MSA files contain a header followed by 4096 rows, each with 2 values, one for each detector, representing bulk sum intensity at each channel.

So I looked for all the maximum and minimum values in the rows and with the help of the histogram representation in figure 14 and this energy table for EDS analysis I could find the max. X-ray counts for the different elements.

               Sol 140  (275 min.)                Sol 167 (468 min.)        

Element     Top      Left      Right           Top      Left      Right

 Na        5361        -        524          11202        -       3289
 Mg       25328     20492     11723          20302     12390      5925
 Al      114989    101384     57502         161257    146880     74567
 Si     1020232    962744    982363        1682175   1595484   1602593
 P        45481      7612     28329         105844     26262     77713
 S       383118    365966    329125         297494    269363    229758
 Cl      286871    232879    277149         792033    724297    772907
 K        67628     57905     52775         117093     97968     88794
 Ca      580670    565817    548754        1356676   1328377   1279895
 Ca       91180     59264     78017         209764    132983    188783
 Ti       90868     77706     71454         293708    272727    257282
 V        29756     10342     10416          72306     35880     45220
 Cr       38408     19068     20287          37335     10250      9939
 Mn       90854     72733     59558         164931    137535    116169
 Fe     2940961   2909664   2916046        5212902   5164162   5168024
 Fe      441308    416393    428950         777404    732526    759197
 Ni       15684      3326      4956            -         -         -
 Cu       14117      3390      2744          21443      4105      4379
 Zn       14562      3189        -           19656      2568        -


  • Every value in the table is the mean of the 2 detector values.
  • A value in a "Left" column is the difference between the "Top" value and the next minimum value to the left of the "Top" value in a from the table produced histogram.
  • A value in a "Right" column is the difference between the "Top" value and the next minimum value to the right of the "Top" value in a from the table produced histogram.
  • Ca and Fe have 2 max. values next to each other.
    (See the histogram representation in fig.14)
  • $\begingroup$ Could you add a link to the source of this table? $\endgroup$
    – asdfex
    May 31, 2022 at 17:34
  • $\begingroup$ I guess you used ps__0140_0679334390_000rbs__00518120532812811688___j02.msa . So I made: i.sstatic.net/ZrPrE.png (to be refined ...) $\endgroup$
    – asdfex
    May 31, 2022 at 18:18
  • $\begingroup$ @asdfex Yes indeed, that's a great histogram, much more user friendly. The other one was (if you didn't guess already) ps_0167_0681742997_00rbs_...., I think they are the only rbs files so far. Very suitable for an "extra" answer ! $\endgroup$
    – Cornelis
    May 31, 2022 at 19:07
  • 1
    $\begingroup$ @OscarLanzi Didn't see your comment last year. Argon isn't labeled but just on the falling slope of the Chlorine peak. Nothing clearly visible there. The small bump under the "Rh*" label is the K-beta line of Chlorine. $\endgroup$
    – asdfex
    Apr 22, 2023 at 10:43
  • 1
    $\begingroup$ Latest data was added a month ago: pds-geosciences.wustl.edu/m2020/urn-nasa-pds-mars2020_pixl/… Regarding the plot I think I have to add a couple of lines with explanations $\endgroup$
    – asdfex
    Apr 23, 2023 at 11:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.