APEXnnnD.BIN / APEXnnnDIONS.CSV

note

OpenWRaw does not read these files and decodes no drift time from them. The layout notes below, including any dt_bin / tof_bin split of a record word, are unverified working notes. The same split was shown not to hold for _FUNCnnn.DAT (see _FUNCnnn.DAT).

Files produced by the Waters Apex3D peak-detection algorithm. Present only in IMS-MS acquisitions on SYNAPT G2-Si instruments where MassLynx post-processing was run (requires Apex3D module installed).

The nnn index matches the _FUNCnnn function number being processed. In all corpus examples, only FUNC001 (the survey IMS function) is processed: Apex3D.bin and Apex3DIons.csv.

APEXnnnDIONS.CSV

Status: Fully Decoded

Human-readable CSV. All columns are self-describing from the header row.

Observed columns (from PXD068881 CtpA):

Column	Type	Description
Function	int	Source function number (always 1 in corpus)
index	int	Sequential ion index (1-based)
m_z	float	Calibrated m/z (Da, 4 decimal places)
m_zNoCal	float	m/z before T1 calibration polynomial
rt	float	Retention time at peak apex (min)
inten	int	Peak apex intensity (TDC counts)
area	int	Integrated peak area
counts	int	Total raw ion counts within 3D peak
errMzPPM	float	m/z error (ppm), if lock mass applied
errRT	float	RT error (min)
errArea	float	Area error
saturation	int	Saturation flag (0 = not saturated)
filterReason	str	Rejection code (empty = accepted)
m_zNoDT	float	m/z without drift time correction
m_zRaw	float	Raw m/z from integer tof_bin (no sub-bin)
fIndex	float	accIndex value (= K * sqrt(m_zRaw))
accIndex	float	Accumulated TOF index: K * sqrt(m_zRaw) where K = A_us^-1 × pusher_clock_rate
chFWHM	float	Chromatographic peak width (min, FWHM)
msFWHM	float	Mass spectral peak width (Da, FWHM)
atStartTime ... atStopTime	float	Peak boundary retention times (min)
atMsStartMass, atMsStopMass	float	Peak boundary m/z limits (Da)
innerElementCount	int	Number of IMS cells contributing to core peak
centerSumResponse	int	Summed intensity of core peak cells
outerElementCount	int	Number of cells in peak shoulder/tail
outerSumResponse	int	Summed intensity of shoulder cells
nonZeroElementCount	int	Count of non-zero cells across full peak boundary
centerSumCounts	int	Raw TDC counts in core peak
outerSumCounts	int	Raw TDC counts in shoulder/tail

accIndex field

accIndex (= fIndex) encodes the centroid TOF flight position in Apex3D-internal units:

accIndex = K * sqrt(m_zRaw)

where K = 1 / t_bin_us (inverse of the per-bin TOF step in microseconds) is derived from the instrument calibration. For the CtpA dataset K = 4582.48 (units: bins/sqrt(Da)).

This is NOT a raw field from _FUNCnnn.DAT. It is a post-processed quantity computed by Apex3D from the decoded tof_bin using the same calibration constants as the m/z formula.

APEXnnnD.BIN

Status: Container structure decoded; per-section data partially characterized

Self-describing multi-section binary container written by the Apex3D executable. Mirrors the ion data in APEXnnnDIONS.CSV but in binary form plus additional internal state.

File Header (64 bytes)

Offset	Type	Value	Description
0x00	u32	1	File format version
0x04	u32	0	Padding
0x08	u32	64	Header size in bytes
0x0C	u32	52	Section descriptor record size
0x10-0x2F	bytes	0	Padding
0x30	u32	1	Unknown (always 1 observed)
0x34	u32	0	Padding
0x38	u32	varies	Byte offset where section data begins
0x3C	u32	0	Padding

Section Descriptor Table

Immediately follows the 64-byte header. Each descriptor is 52 bytes:

struct SectionDesc {
    u32  reserved_0;      // first scalar field (semantics unclear; see notes)
    u32  data_offset;     // byte offset of this section's data, relative to data_start
    u32  reserved_8;      // second scalar field (semantics unclear)
    char name[40];        // null-padded ASCII section name
};

The number of section descriptors = (data_start - 64) / 52.

The file data area begins at the offset stored in header@0x38. Section data starts at header_data_start + section_desc.data_offset.

Section 0 always has data_offset = 0 (its data starts immediately at data_start). Subsequent sections have data_offset equal to the cumulative byte extent of previous sections.

Sections Observed in CtpA (PXD068881)

13 sections:

Index	Name	Data size	Content
0	AcquisitionParameter	7230	Nested parameter table (13 fields; see below)
1	InstrumentInfo	834	Instrument geometry/state (binary; mostly zeros)
2	FunctionParameter	550	Per-function settings (binary)
3	Calibration	62	ASCII: Apex3D command-line args (lock mass, tolerance)
4	PeakDetectResult	114	ASCII: detection parameters used (threshold, function)
5	ProcessingProgram	117	Binary: program version/timestamp
6	PeakDetectParameter	659	Binary f64 array: detection algorithm parameters
7	DeIsotopeParameter	116	Binary f64 array: de-isotoping parameters
8	BinParameter	71	Binary: binning configuration
9	LmCalibrator	44	Binary f64: lock mass calibration coefficients
10	LockMassStick	65	Binary f64: lock mass reference m/z positions
11	Peak	388	Binary: compact peak summary (format not decoded)
12	PeakEx	8,502,560	Binary: extended peak data (format not decoded)

Calibration and PeakDetectResult (ASCII sections)

These sections contain the Apex3D command-line invocation that produced the file. Example (truncated):

-lockMassZ1 556.2771 -lockMassToleranceAMU 0.2500
-le3DThresholdCounts 130 -function 1 -bEnableCuda 1

AcquisitionParameter Section

The data at data_start is a nested Waters-style parameter table with 13 named fields describing the acquisition setup (scan time, pusher count, m/z range, etc.). The inner structure uses a different record format than _FUNCnnn.STS and is not currently fully decoded.

Peak and PeakEx Binary Data

These sections contain the detected 3D ion peaks in binary form. The file size of PeakEx (~8.5 MB for CtpA, 1138 scans over 2h) suggests it stores per-scan or per-push IMS-MS intensity data associated with each detected peak, not just the apex values.

The APEXnnnDIONS.CSV file provides all peak apex data in human-readable form. Binary decoding of Peak/PeakEx is not necessary for reading or writing spectrum data.

Reference Sources

• Empirical analysis: CtpA (PXD068881)
• CSV format: self-describing column headers
• Binary structure: decoded from hex analysis