Agilent ICP-MS MSProfile.bin File Structure

This file format contains Agilent ICP-MS data (for example, from an Agilent 7700 single-quadrupole or an Agilent 8900 triple-quadrupole instrument). It shares its container and several index files with the HRMS format, but the MSProfile.bin payload is laid out differently and is not compressed.

ICP-MS data is distinguished from HRMS data by the presence of a MSScan_XSpecific.bin file in the AcqData subdirectory.

The data is encoded across several files:

File	Information
MSTS.xml	Number of retention times
MSTS_XSpecific.xml	Number of isotope channels (masses)
MSScan.xsd	File structure of MSScan.bin
MSScan.bin	Per-scan retention time, data offset, and point count
MSScan_XSpecific.bin	Presence flags the directory as ICP-MS data
MSProfile.bin	Intensity values (uncompressed)
MSTS_XAddition.xml	Real isotope m/z labels (located one level above AcqData)

MSTS.xml and MSScan.xsd are read exactly as in the HRMS format: the summed NumOfScans gives the number of retention times, and the XSD defines the ScanRecordType records that make up MSScan.bin (beginning at offset 0x58). For ICP-MS, only the scan’s retention time, SpectrumOffset, and PointCount are needed — the data is uncompressed, so the uncompressed byte count is not used.

MSTS_XSpecific.xml lists one Masses element per isotope channel. Counting them gives the number of channels recorded at every retention time.

MSTS_XAddition.xml (one directory above AcqData) maps each channel index to its real isotope m/z through ProductIonMZ. These become the m/z labels. If the file is absent, the per-channel XValue from MSScan_XSpecific.bin is used as a fallback.

<MSTS_XAddition_IndexedMasses>
    <Index>1</Index>
    <PrecursorIonMZ>12</PrecursorIonMZ>
    <ProductIonMZ>12</ProductIonMZ>
</MSTS_XAddition_IndexedMasses>

MSProfile.bin stores, for each retention time, four parallel blocks of PointCount values. Assume little-endianness.

Block	Data Type
Channel index	Float (4 bytes each)
Reported value (the reported intensity)	Double (8 bytes each)
Raw pulse count	Double (8 bytes each)
Analog value	Double (8 bytes each)

rainbow keeps the reported values, which are the intensities that MassHunter writes to its CSV export. The raw pulse and analog blocks (used for secondary-electron-multiplier detector cross-calibration) are skipped.

A single data segment can be visualized as follows. The inner blocks are not drawn to scale, and each block holds PointCount entries.

+---------+---------++----------+----------++-------+-------++--------+--------+
| index 1 | index 2 || reported | reported || pulse | pulse || analog | analog |
+---------+---------++----------+----------++-------+-------++--------+--------+
|                            repeats for every retention time                  |
+------------------------------------------------------------------------------+

Note

ICP-MS parsing is reached through the MassHunter path, so it requires hrms=True:

import rainbow as rb
datadir = rb.read("path/to/data.d", hrms=True)

Unlike the HRMS format, ICP-MS data is uncompressed and does not require the optional python-lzf dependency.

This parser currently supports time-resolved acquisitions with a single tune mode and one measurement per isotope. Files with multiple tune modes (e.g. several collision/reaction gas settings) or multiple measurements per isotope are not yet handled.

The decoding of this format was contributed by Jeremy Hourigan (UC Santa Cruz); see issue #25.