Feedback and discussions on the .ort specs
In the following you find a rather unsorted collection of feedback and ideas about the .ort
specs and the related orsopy
package.
last modified: 2024-03-11
confusion of physical terms
(Jochen)
We have a confusion of what we use as key words. Since the german terms are different I had probplems figuring out the correct English definitions…
official definitions
What can be measured or calculated is a physical quantity.
E.g. the incident angle
This has a dimension = dim(physical quantity) relating it to a set of base quantities like length, time, charge, temperature etc. The dimension is no unit, nor can it be used to unambigiously describe a physical quantity (plane angle does not tell between scattering angle, incident angle, total reflection angle, …).
dim( incident angle ) = plane angle
The physical quantity is often refered to by using a symbol.
one possible symbol for incident angle is $\alpha_i$ (or alpha_i in the orso header)
The physical quantity is composed of a numerical magnitude times unit. Depending on the chosen unit, the numerical magnitude changes.
$\alpha_i = 2.3 \cdot \mathrm{deg}$
guidelines for writing and reading
- avoid contradicting information (e.g. single incident angle in the header for angle-disperse measurement)
- The
.ort
format contains the reduced data. Its content should thus be limited to information useful to the scientist and to some analysis software. The rest is still available in the raw files…..
open issues for lab x-ray reflectometers
Which of the keys discussed below should be included in the specs to (better) incorporate lab x-ray data files?
When attempting to convert the ASCII output files of various commercial lab x-ray reflectometers (diffractometers) it became obvious that the present dictionary misses several entries.
-
It is not exactely clear where to put the brand, model and probably configuration information.
experiment: title: ... instrument: type: x-ray lab source (neutron reflectometer, synchrotron diffractometer, ....) brand: Brucker model: Discovery hardware_indicator: 65519
- The wavelength is often defiend via the anode material, the line(s) and probably the presence of a monochromator.
- The slit sizes are reported to enable resolution calculation.
- Often a long list of hardware settings is supplied, e.g. tube current, temperature, configuration, etc.
These things do not really belong to a reduced data file, but we should at least recommend a place for these entries. In the example below I put it as a multy-line string in
instrument_settings.details
.
measurement:
instrument_settings:
incident_angle:
min: 0.1
max: 6.0
unit: deg
wavelength:
magnitude: 1.54184
unit: angstrom
anode: Cu
lines:
- name: K_alpha1
magnitude: 1.5405980
weight: 2/3
- name: K_alpha2
magnitude: 1.5444260
weight: 1/3
scan_type: continuous
details: |
"Configuration=Reflection-Transmission Spinner 3.0, Owner=user, Creation date=3/5/2021 8:12:09 AM"
"Goniometer=Theta/Theta; Minimum step size 2Theta:0.0001; Minimum step size Omega:0.0001"
"Sample stage=Reflection-transmission spinner 3.0; Minimum step size Phi:0.1"
- The
.ort
specs clearly separate data origin and data reduction. For lab reflectometers it often the same software for instrument control and reduction. - Information about the facility, the owner and the sample is often missing.
new column type: flag
suggestd by Artur, draft by Jochen
# columns:
...
# - flag_is:
# 0: electric field off
# 1: electric field on, positive
# -1: electric field on, negative
or
# columns:
...
# - flag_is:
# 0: ignored for fitting
# 1: used for fitting
The flag key is to be treated as an integer (by cutting off after the point).
date
vs. datetime
for experiment date
The header entry data_source.experiment.start_date
presently shows the date and time. The intention here was to report on the date only. The time is listed anyway for all individual measurements below.
Can we change this to date only?
slit info format
Should we define a format for slits? I.e. distance to sample, horizontal or vertical openings, radius, shape ….
siggestion:
# data_source:
# measurement:
# instrument_settings:
# apertures:
# - name: D1
# shape: rectangular
# horizontal_opening: 20
# vertical_opening: 0.5
# distance_to_sample: -983
# unit: mm
# - name: D2
# shape: circular
# radius: 2
# distance_to_sample: 30
# unit: mm
with
shape
: circular, rectangular, slit- dimensions: vertical_opening, horizontal_opening, radius
distance_to_sample
with negative sign befre the sample (strictly speaking, it is no longer a disctance then). Maybeposition_wrt_sample
?
If this format is opened to other devices, e.g. a collimator tube, the distance_to_sample
would need to allow for a range.