RLum.OSL_correction {OSLdecomposition}

Check and correct CW-OSL curves in RLum.Analysis data sets

Description

CW-OSL measurements are often affected by background signals or might be measured under inconsistent detection settings. This function provides tools to test and solve some common problems.

Usage

RLum.OSL_correction(
  object,
  record_type = "OSL",
  background_sequence = NULL,
  subtract_offset = 0,
  check_consistency = TRUE,
  remove_light_off = TRUE,
  limit_duration = 20,
  PMT_pulse_pair_resolution = 18,
  verbose = TRUE
)

Arguments

object

RLum.Analysis or list of RLum.Analysis (required): Data set of one or multiple CW-OSL measured aliquots.

record_type

character (with default): Type of records selected from the input object, see object[[]]@records[[]]@recordType. Common are: "OSL","SGOSL" or "IRSL".

background_sequence

numeric vector (optional): Indices of list items with CW-OSL measurements of empty aliquots. The records in these list items are used to calculate one average CW-OSL background curve with sum_OSLcurves. This background curve is subtracted from each CW-OSL record of the data set. The attributes ⁠@recordType⁠ of the background measurements will be renamed to "{record_type}background".

subtract_offset

numeric (optional): Signal offset value in counts per second (cts/s). Value is handled as background level and will be subtracted from each CW-OSL record.

check_consistency

logical (with default): The CW-OSL component identification and separation procedure requires uniform detection parameters throughout the whole data set. If TRUE, all records are compared for their channel width and their number of channels. Those records with the most frequent set of channel parameters keep their ⁠@recordType⁠ attribute, while records with other sets of measurement parameter will be enumerated record_type "{record_type}2", "{record_type}3" and so on.

remove_light_off

logical (with default): Checks if the records contain zero-signal intervals at beginning and/or end of the measurement and removes them. Useful to tailor single-grain measurements.

limit_duration

numeric (with default): Reduce measurement duration to input value in seconds (s). Long measurement duration can lead to over-fitting at the component identification of Step 1 which may induce systematic errors, see Mittelstrass (2019). Thus, limiting the OSL record length ensures sufficient accuracy regarding the Fast and Medium component analysis. If however, slow decaying components are of interest, limit_duration = NA is recommended.

PMT_pulse_pair_resolution

numeric (with default): Time span of the pulse-pair resolution of the PMT in nanoseconds (ns). If a value is given, the signal values will be corrected for time-resolution related non-linearity at height counting rates, see Details. Set PMT_pulse_pair_resolution = NA if algorithm shall be omitted.

verbose

logical (with default): Enables console text output.

Details

This function processes data sets created within the Luminescence-package (Kreutzer et al. 2012). Those data sets must be formatted as RLum.Analysis objects. Output objects will also be RLum.Analysis objects and are meant for further analysis with RLum.OSL_global_fitting.

The data preparation tools are executed in the following order:

1. check_consistency

2. remove_light_off

3. limit_duration

4. PMT_pulse_pair_resolution

5. background_sequence

6. subtract_offset

Currently, not all functions are available.

Details to remove_light_off: The algorithm does the following: (1) Create global reference curve with sum_OSLcurves (2) Search for the maximum in the first half of the reference curve and remove all data points before the maximum . Do this for all curves of the selected 'record_type'. (3) Search for an infliction point with negative curvature (minimum of second differential) in the second half of the reference curve. If the next data point has at least 50% less signal, remove all data points after the infliction point. Do this for all curves of the selected 'record_type'.

Details to PMT_pulse_pair_resolution:

The algorithm corrects non-linearity of signal values due to insufficient pulse-pair resolution of the photo-multiplier tube (PMT). Equation (6-2) of the Hamamatsu Photomultiplier Handbook is used:

I_corrected = I_measured / (1 - I_measured * resolution)

The algorithm does not account for PMT saturation and PMT aging effects. As default pulse-pair resolution 18 ns is pre-defined, the Hamamatsu H7360 series pulse-pair resolution according to the data sheet. The H7360-02 is the default PMT in Freiberg Instruments lexsyg OSL/TL readers. DTU Physics Risoe TL/OSL reader deploy ET Enterprise 9235B series PMTs as default. For these PMTs, the pulse-pair resolutions is not given in the data sheets and relies on the operation voltage. However, due to the pulse properties given in the data sheets, it is unlikely that those PMTs have a better pulse-pair resolution than 18 ns.

Impact of a pulse-pair resolution correction of 18 ns

Measured signal	Corrected signal	Signal underestimation
1000 cts/s	1000 cts/s	0.00 %
10000 cts/s	10002 cts/s	0.02 %
50000 cts/s	50045 cts/s	0.09 %
100000 cts/s	100180 cts/s	0.18 %
500000 cts/s	504541 cts/s	0.91 %
1000000 cts/s	1018330 cts/s	1.83 %

Value

The input object, a list of RLum.Analysis objects, is given back with eventual changes in the elements object[[]]@records[[]]@recordType and object[[]]@records[[]]@data.

The returned data set contains a new list element object[["CORRECTION"]] which provides a list of the input parameters and additional data depending on the applied tools.

Last updates

2022-01-02, DM: Revised PMT_pulse_pair_resolution algorithm.

Author(s)

Dirk Mittelstrass, [e-mail]

Please cite the package the following way:

Mittelstraß, D., Schmidt, C., Beyer, J., Heitmann, J. and Straessner, A.: R package OSLdecomposition: Automated identification and separation of quartz CW-OSL signal components, in preparation.

References

Hamamatsu, 2007. Photomultiplier Tubes: Basics and Applications, Third Edition (Edition 3A). Hamamatsu Photonics K. K., Hamamatsu City.

Kreutzer, S., Schmidt, C., Fuchs, M.C., Dietze, M., Fischer, M., Fuchs, M., 2012. Introducing an R package for luminescence dating analysis. Ancient TL, 30 (1), 1-8.

Mittelstraß, D., 2019. Decomposition of weak optically stimulated luminescence signals and its application in retrospective dosimetry at quartz (Master thesis). TU Dresden, Dresden.

See Also

RLum.OSL_global_fitting, RLum.OSL_decomposition, sum_OSLcurves

Examples

# 'FB_10Gy' is a dose recovery test with the Fontainebleau quartz
# measured with a lexsyg research with green LED stimulation
data_path <- system.file("examples", "FB_10Gy_SAR.bin", package = "OSLdecomposition")
data_set <- Luminescence::read_BIN2R(data_path, fastForward = TRUE)

# To correct for the background signal, subtracted the average curve from the
# OSL curves of an empty aliquot (list item 11) from all other OSL records:
data_set_corrected <- RLum.OSL_correction(data_set, background = 11, remove_light_off = FALSE)


# Plot background corrected global average CW-OSL curve
sum_OSLcurves(data_set_corrected, output.plot = TRUE, record_type = "OSL")

# Plot background curve
sum_OSLcurves(data_set_corrected, output.plot = TRUE, record_type = "OSLbackground")