CheMin data sets

Data Set Overview

The CheMin instrument captures diffracted and fluoresced X-rays on a CCD detector. Up to 2730 individual images can be stored in the instrument, but the downlink capability of the rover does not support the transmission of these raw frames in their entirety. Processing onboard the rover is used to reduce the downlink data volume while preserving the scientific content of the data.

Processing

CheMin EDR data products are considered CODMAC Level 2 or Edited Data (equivalent to NASA level 0) products. The EDR data files are generated from Level 1 or Raw Data, which are the telemetry packets within the project specific Standard Formatted Data Unit (SFDU) record.

The CheMin EDR data products will be generated by the MIPL (Multimission Image Processing Laboratory) at JPL under the Operations Product Generation System (OPGS). The EDR data products will be raw uncalibrated data reconstructed from telemetry data products in the format described by the CheMin EDR SIS. Meta-data collected when the rover prepared the science data for downlink will be provided in PDS label.

There will be multiple versions of a CheMin EDR. If telemetry data are missing during the initial downlink from the rover memory, partial data sets will be created and the missing data will be filled with zeroes. The data will be reprocessed after all data are received and the original version will be overwritten. There can be up to 2 EDRs created, one with a status of partial and the other with a complete status in the filename.

Data

The data types available from CheMin are described in the EDR SIS and are summarized as follows:

1. Diffraction - All: Multiple (e.g., 120) raw images are combined into a single array with the same dimensions as the original images where only pixel values within a specified window of DN values are counted. This window will typically be selected to represent diffracted Co-Kalpha X-rays. This will be the primary data product used for diffraction studies.
2. Diffraction - Single: Same as 'Diffraction - All' except pixel values within the specified window are counted only if they are surrounded entirely by pixels below a selected background threshold. This data set minimizes the contribution of diffracted Co-Kbeta X-rays and other extraneous events, but excludes many valid X-rays.
3. Diffraction - Split: Same as 'Diffraction-All' except pixels are counted only if two adjacent values sum to produce a value that is within the specified window. This data set attempts to recover split events (X-rays depositing charge over multiple pixels) and can be used when the count-rate is low.
4. Film: Multiple (e.g., 120) raw images are summed into a single array. This product will have a higher background noise level given that all fluoresced X-rays are added in, but it will be the best product to use for diffraction studies if the grain motion is poor and there is a large occurrence of pixels capturing multiple diffracted X-rays within a single image.
5. Energy - All: The individual pixel values from multiple (e.g., 120) raw images are combined into a single energy histogram representing counts versus instrumental digital number (DN) energy bins.
6. Energy - Single: Same as 'Energy-All' except pixel values are added to the histogram only if they are surrounded entirely by pixels below a selected background threshold. This will be the primary product used for selection of the diffraction window and for any chemical assessments of the data.
7. Energy - Split: Same as 'Energy-All' except the only values added to the histogram are from the sum of two adjacent pixels which are surrounded entirely by pixels below the background threshold. This data set can be used to add counts to 'Energy- single' when the overall count rate is low.
8. Raw frames: Unprocessed, single images are periodically retained for downlink to allow the science team to assess the performance of the onboard algorithms. One dark frame (no x-ray source) is collected before and after each analysis to track the health of the detector. CheMin can also be commanded to output 'thresholded' raw data where a selected range of raw frames are extracted from the instrument but all pixels with values below a specified threshold and all pixels with values above a second specified threshold are set to zero. Even though data in this format will compress well, some valuable information will be discarded and the downlink requirements will still be relatively large. Nonetheless, this data format could be used if issues are identified in the onboard processing.
9. Housekeeping: A full set of parameters, temperatures, voltages, and other health/status data is captured periodically for downlink.
10. CCD Frame: Individual CCD frames will be downlinked periodically to analyze CCD dark current, hot pixels, and overall CCD health. In the event that an off-nominal result is obtained such that K-alpha thresholds are incorrect or some other unforeseen circumstance renders the data reduction algorithms ineffectual, complete individual frames can be downlinked and processed using algorithms that can be written on the ground and uplinked to the rover.

Data File Names

Data file names in the CheMin EDR and RDR archive are named according to the following scheme:

<instr><config><sclk><prod><sol><site><drive><seq><venue><ver>.<ext>

where

<instr>  = CM (CheMin)

<config> = A_, B_, or __ (instrument configuration)

<sclk>   = 9-character spacecraft clock count

<prod>   = 3-character product type

<sol>    = 4-digit sol number

<site>   = 3-character site location

<drive>  = 4-character drive position

<seq>    = 7-character CheMin sequence identifier

<venue>  = 1-character producer identifier

<ver>    = 1-character version identifier

<ext>    = IMG (image), DAT (binary table), CSV (ASCII delimited table)

See the CheMin EDR Data Product Software Interface Specification in the Document directory of the archive for details about the file naming scheme.

Confidence Level Overview

The data presented in the CheMin EDR is intended to be the least processed, most primitive data set released for CheMin. Data presented here are an accurate representation of CheMin data as received from the rover.

Review

The CheMin EDR products are reviewed internally by the CheMin team prior to release to the PDS. PDS also performs an external peer review of the CheMin EDR.

Data Coverage and Quality

This section will be updated with the first release of CheMin EDR data to the PDS.

Limitations

The major limitation of this data set is that this is level 0, minimally processed data. The data are received from the spacecraft telemetry and ingested into a database. If gaps exist in the telemetry, data are lost.

Data Compression

No data compression is used.

Data Set Overview

The CheMin instrument determines the mineralogy and elemental composition of powdered samples through the combined application of X-ray diffraction (XRD, producing mineral identification and quantification) and X-ray fluorescence (chemical analysis based on Energy Dispersive Histograms, EDH). This CheMin RDR data set contains tables of energy and X-ray diffraction data derived from raw CheMin observations.

A higher level CheMin RDR data set is derived from these products, the CheMin Mineral Abundance data (MSL-M-CHEMIN-4-RDR-V1.0), which is archived separately.

Processing

CheMin RDR data products are considered CODMAC Level 4 or Resampled Data, equivalent to NASA Level 1B products. The RDR data files are generated from CheMin Level 2 (NASA Level 0) EDR Data (MSL-M-CHEMIN-2-EDR-V1.0), archived separately.

Processing of XRD data consists of conversion of the 2D CCD images of Debye diffraction rings into 1D files of intensity versus 2-theta. Counts within arcs of the Debye rings are adjusted for variable arc lengths intersected by the CCD. Processing of EDH data consists of conversion of digital number (DN) to X-ray energy in keV. Data

The data types available in this data set are described in the RDR SIS and are summarized as follows:

1. Co K-alpha or K-beta XRD single-pixel data summed along 2 theta radii, adjusted for arc length and presented as 2-theta versus intensity tables (product type RD1).
2. Co K-alpha or K-beta split pixel XRD data summed along 2-theta radii, adjusted for arc length and presented as 2-theta versus intensity tables (product type RDS).
3. Co K-alpha or K-beta all-hits XRD data summed along 2-theta radii, adjusted for arc length and presented as 2-theta versus intensity tables (product type RDA).
4. Raw frames processed for XRD analysis, summed along 2-theta radii, adjusted for arc length and presented as 2-theta versus intensity tables (product type RTR).
5. Film images processed for XRD analysis, summed along 2-theta radii, adjusted for arc length and presented as 2-theta versus intensity tables (product type RDF).
6. Energy histograms of all hits converted from DN to energy (product type REA).
7. Energy histograms of single-pixel hits converted from DN to energy (product type RE1).
8. Energy histograms of split-pixel hits converted from DN to energy (product type RES).

Confidence Level Overview

X-ray diffraction reliability and accuracy of mineral identifications is comparable to that of commercial portable XRD instruments such as the Terra system manufactured by inXitu (http://www.inxitu.com/html/Terra.html). This level of performance produces accurate identification and detection for virtually all crystalline phases at abundance greater than ~1 wt%. EDH capability is limited to elements of Z=14 (Si) and higher, due to poor CCD quantum efficiency at low X-ray energies (below 1.7 keV).

Review

The CheMin RDR products are reviewed internally by the CheMin team prior to release to the PDS. PDS also performs an external peer review of the CheMin RDR.

Data Coverage and Quality

This section will be updated with the first release of CheMin RDR data to the PDS.

Limitations

X-ray amorphous materials may be detected as a broad XRD amorphous background but amorphous materials have higher detection limits with poorer accuracy in quantification. Quantification of EDH data (chemical abundances from X-ray diffraction) is not currently possible but is a research effort.

Data Compression

No data compression is used.

Data Set Overview

The CheMin instrument determines the mineralogy and elemental composition of powdered samples through the combined application of X-ray diffraction (XRD, producing mineral identification and quantification) and X-ray fluorescence (chemical analysis based on Energy Dispersive Histograms, EDH). This CheMin RDR data set contains mineral identification, abundance, and estimated errors derived from CheMin observations (RDR product type MIN).

Processing

These CheMin RDR data products are considered CODMAC Level 5 or Derived Data, equivalent to NASA Level 2 products. The RDR data files are generated from CheMin Level 4 (NASA Level 1B) RDR Data (MSL-M-CHEMIN-4-RDR-V1.0), archived separately.

Processing requires use of one of several available programs for mineral identification and XRD phase quantification applied to any of the Level 1 CheMin 1D diffraction patterns (D1A, D1B, DSA, DSB, DAR, DMR). Mineral identification may be based on various diffraction data libraries (ICDD PDF database, AMCSD, etc.); pattern processing and phase quantification may be based on any of several available software packages (Jade, TOPAS, FULLPAT, etc.). Because of the multiple options for deriving mineral identification and quantification there may be several reprocessed level 2 products based on a single XRD analysis.

Data

The data types available in this data set are described in the RDR SIS. A list of minerals present in a CheMin analysis is compiled from the NASA Level 1B processed XRD data using one or more mineral identification programs; the identification and quantification programs(s) used are specified in the file metadata. Identifications are scrutinized for reliability against detection limits and potential complications (e.g., peak overlaps). Mineral identifications are further evaluated against information from EDH element identifications and other instrument datasets (particularly ChemCam, APXS, and SAM) to evaluate whether minor phases near detection limits or phases problematic in XRD analysis (e.g., poorly crystalline or amorphous) need to be considered. Rietveld or other pattern-fitting methods are used to generate a table of mineral abundances in weight percent along with estimated errors.

Confidence Level Overview

X-ray diffraction reliability and accuracy of mineral identifications is comparable to that of commercial portable XRD instruments such as the Terra system manufactured by inXitu (http://www.inxitu.com/html/Terra.html). This level of performance produces accurate identification and detection for virtually all crystalline phases at abundance greater than ~1 wt%. Quantification of mineral abundances is generally best when based on Rietveld pattern fitting; however, noncrystalline or poorly crystalline materials (amorphous samples and many clay minerals) require other approaches (e.g., FULLPAT pattern fitting methods) where confidence is related to use of appropriate library samples for pattern fitting.

Review

The CheMin mineral abundance products are reviewed internally by the CheMin team prior to release to the PDS. PDS also performs an external peer review of these products.

Data Coverage and Quality

This section will be updated with the first release of CheMin RDR data to the PDS.

Limitations

X-ray amorphous materials may be detected as a broad XRD amorphous background but amorphous materials have higher detection limits with poorer accuracy in quantification. Quantification of some clay minerals, particularly smectites, requires use of appropriate library patterns or structural modeling methods.

Data Compression

No data compression is used.