PSI Toolkit Single Particle Soot Photometer (SP2)

PSI Toolkit

Single Particle Soot

Photometer (SP2)

Operator Manual

DOC-0359 Revision A-1

2400 Trade Centre Avenue

Longmont, CO 80503 USA

DOC-0359 Rev A-1 © 2017 DROPLET MEASUREMENT TECHNOLOGIES

2

T a b l e o f C o n t e n t s

1.0 Introduction .................................................................................................. 5 General Information .................................................................................................................... 5

2.0 Installation and Updating .............................................................................. 7 2.1 Installing IGOR Pro Software ........................................................................................... 7 2.2 Installing the SP2 Toolkit ................................................................................................. 7 2.3 Updating existing IGOR experiments to latest Sp2 toolkit release .................................. 9

3.0 Program Architecture and Data Structure .................................................... 10 3.1 Raw Data Files ................................................................................................................ 10 3.2 Program Architecture .................................................................................................... 10 3.3 Data Structure after Loading Files into Igor Pro ............................................................ 11

4.0 Loading Data ............................................................................................... 13 4.1 Loading Housekeeping Data .......................................................................................... 13 4.2 Loading Particle Trace Data ........................................................................................... 15

4.2.1 SP2 Configuration and Properties ............................................................................ 15 4.2.2 Setting Data Load and Analysis Parameters ............................................................ 16 4.2.3 Loading and Analyzing the Particle-Trace Data ...................................................... 17 4.2.4 Resetting the Sample Flow Rate Data ..................................................................... 19 4.2.5 Resetting the YAG power data ................................................................................. 20

5.0 Trace Analysis .............................................................................................. 21 5.1 Preparing Settings for Trace Analysis ............................................................................ 21

5.1.1 Incandescence Channel Settings .............................................................................. 21 5.1.2 Scattering Channel Settings ..................................................................................... 22 5.1.3 Split Detector (Position Sensitive) Channel .............................................................. 23 5.1.4 Detector and A-to-D Converter Saturation .............................................................. 24

5.2 Analyzing Raw Traces ..................................................................................................... 25

6.0 Post Processing Particle-by-particle Data ..................................................... 26 6.1 Basic Post Processing ..................................................................................................... 26 6.2 Fitting Log-normal Functions to Size Distribution ......................................................... 31 6.3 Extracting Time Series of Concentration Data ............................................................... 33 6.4 Extracting Time Series of Size Distribution Data............................................................ 35 6.5 Further post processing ................................................................................................. 37

6.5.1 Statistics of Bandratio Data ..................................................................................... 37 6.5.2 Delay time data ....................................................................................................... 38

6.6 Troubleshooting ............................................................................................................. 42

7.0 LEO-fit ......................................................................................................... 43

8.0 Analyzing Calibration Data ........................................................................... 44 8.1 BC calibration with Separate SP2-files for Each DMA Size ............................................ 44

8.1.1 Preparing Waves for Calibration Coefficients .......................................................... 44 8.1.2 Load SP2 Calibration Data ....................................................................................... 46 8.1.3 Preparing the Calibration Info Waves ...................................................................... 49 8.1.4 Filling in the “CalibInfo” Table ................................................................................. 50 8.1.5 Processing the Monodisperse Calibration Data ....................................................... 51

PSI SP2 Toolkit Manual

DOC-0359 Rev A-1 3 © 2017 DROPLET MEASUREMENT TECHNOLOGIES

8.1.6 Determining the Peak Height for Each Measured BC Mass ..................................... 54 8.1.7 Fitting the Calibration Curve .................................................................................... 56 8.1.8 Storing the Calibration Coefficients and Verification of the Calibration ................. 61 8.1.9 Counting Efficiency and Bandratio Calibration ........................................................ 61

8.2 Scattering Detector Calibration ..................................................................................... 62 8.2.1 Loading SP2 Scattering Calibration Data ................................................................. 62 8.2.2 Entering Calibration Information ............................................................................. 63 8.2.3 Preparing Histograms of Scattering Peak Heights................................................... 63 8.2.4 Inspecting Histograms and Determining Scattering Calibration Coefficients ......... 67

9.0 Analysis of Ice Core Data.............................................................................. 71

Appendix A: Revisions to Manual ............................................................................. 72

Appendix B: Schematics of Data Analysis Approach ................................................... 72

T a b l e o f F i g u r e s

Figure 1: Unzipped Archive .......................................................................................... 7

Figure 2: Main Procedure Window for Toolkit ............................................................. 8

Figure 3: SP2 Toolkit Panel .......................................................................................... 8

Figure 4: Procedure Window Displaying Version Number ............................................ 9

Figure 5: Procedure Window Updated to Load Latest Version Number ...................... 10

Figure 6: Data Browser Displaying SP2 Wave Data ..................................................... 12

Figure 7: Loading Housekeeping Data – Step 1 ........................................................... 14

Figure 8: Loading Housekeeping Data – Step 2 ........................................................... 14

Figure 9: Specifying Data Channels to Be Loaded ....................................................... 15

Figure 10: Specifying Data Loading Parameters.......................................................... 16

Figure 11: Selecting Beam-Shape Data ....................................................................... 17

Figure 12: Selecting a Single File to Be Loaded ........................................................... 18

Figure 13: Entering List of Files to be Loaded ............................................................. 18

Figure 14: Entering Folder from which to Upload Files ............................................... 18

Figure 15: Selecting Housekeeping Files to Be Uploaded ............................................ 19

Figure 16: Selecting Source for Sample Flow Rate Data .............................................. 20

Figure 17: Resetting YAG Power Data ........................................................................ 20

Figure 18: Incandescence Channels ........................................................................... 21

Figure 19: Scattering Channel Settings ....................................................................... 22

Figure 20: Split Detector Settings .............................................................................. 23

Figure 21: Detector and A-to-D Converter Settings .................................................... 24

Figure 22: Trace Analysis Settings .............................................................................. 25

Figure 23: Selecting Raw Trace Data .......................................................................... 26

Figure 24: MaskBadData Wave Shown In File Browser ............................................... 27

Figure 25: Settings for Calibration Constants and BC Property ................................... 28


4

Figure 26: Additional Post-Processing Settings ........................................................... 29

Figure 27: Specifying Trace-Analysis Data Folders ...................................................... 30

Figure 28: Log-normal Fit Settings ............................................................................. 31

Figure 29: Specifying Trace-Analysis Data Folder(s) .................................................... 32

Figure 30: Settings for Extracting Concentration Time Series ...................................... 33

Figure 31: Specifying Folders for Extracting Time Series Data ..................................... 34

Figure 32: Settings Used in Extracting Time Series of Size Distribution Data ............... 35

Figure 33: Select Data Folder(s) for Extracting Time Series of Size Distribution Data ... 36

Figure 34: Select Folder for Log-Normal Fitting of Size Distribution Time Series Data.. 36

Figure 35: add bandratio statistic Button ................................................................... 37

Figure 36: Selecting Data Folder for Calculating Bandratio Statistics .......................... 38

Figure 37: Setting Bandratio Calculation Options ....................................................... 38

Figure 38: Generating Histogram of Delay-Time Data ................................................ 39

Figure 39: Selecting Data Folder for Delay-Time Data ................................................ 40

Figure 40: Setting Options for Delay-Time Histogram ................................................. 40

Figure 41: Selecting Bin Numbers for Delay-Time Histogram ...................................... 40

Figure 42: add delay time number fractions Button ................................................... 41

Figure 43: Selecting Data Folder for Number Fraction of Thickly Coated Particles ....... 42

Figure 44: Specifying Settings for Calculating Number Fraction of Thickly Coated Particles .................................................................................................... 42

Figure 45: create waves for calib. coef. Button .......................................................... 44

Figure 46: Specifying Folder Name for Calibration Coefficients .................................. 45

Figure 47: Selecting Calibration Coefficients .............................................................. 45

Figure 48: Selecting the Number of Points to be Averaged for Peak Height ................ 46

Figure 49: Setting the Channel Configuration and Channels to be Loaded .................. 47

Figure 50: Load Options and load raw data Button .................................................... 48

Figure 51: Selecting the Range of Raw Data Files to Load ........................................... 48

Figure 52: Specifying the Directory for Raw Data ....................................................... 49

Figure 53: create empty BC calib info waves Button ................................................... 49

Figure 54: Selecting the Number of Calibration Points ............................................... 50

Figure 55: The CalibInfo Table ................................................................................... 50

Figure 56: BC Calibration Material Setting and analyze monodisperse calibration Button ...................................................................................................... 51

Figure 57: Selecting Calibration Info Waves ............................................................... 52

Figure 58: Setting Additional Calibration Parameters ................................................. 52

Figure 59: Selecting Raw Data Folder with Calibration Measurements ....................... 53

Figure 60: Peak Height Histogram Measured at Calibration Points ............................. 54

Figure 61: CalibFitPkHt Table ..................................................................................... 55

Figure 62: Graph of Calibration Curve ........................................................................ 55

Figure 63: fit calibration curve (spline) Button ........................................................... 56

Figure 64: Selecting the Calibration Summary Folder ................................................. 57

Figure 65: Selecting the Detector Channel ................................................................. 57

Figure 66: Calibration Points and Fit Shown in Table (above) and Graph (below) ....... 58

Figure 67: Deficient Fit at Small Masses/Peak Heights ............................................... 59



Figure 68: Modifying Waves to Improve Fit in Specific Ranges ................................... 59

Figure 69: Graph (above) and Table (below) Showing Corrected Fit for Small Peak Heights ..................................................................................................... 60

Figure 70: Comparing Calibrated Measurements to DMA Results ............................... 61

Figure 71: Entering PSI Scattering Calibration Information ........................................ 63

Figure 72: Waves Describing Calibration Data ............................................................ 63

Figure 73: Analyse monodisperse calibration button ................................................. 64

Figure 74: Selecting Scattering Calibration Info Waves ............................................... 65

Figure 75: Selecting SP2 Raw Data ............................................................................. 66

Figure 76: Scattering Calibration Optional Parameters ............................................... 66

Figure 77: Issues with Fitting the Histogram—Large Number of Saturated Signal ....... 68

Figure 78: Issues with Fitting the Histogram—Too Few Data Points, Too Wide Fit ...... 68

Figure 79: Corrected Fit ............................................................................................. 69

Figure 80: Eliminating Invalid Calibration Points ........................................................ 70

Figure 81: Wave Stats Dialog Box .............................................................................. 71

Figure 82: Updating Calib_Coef_table ........................................................................ 71

1.0 Introduction

About this Manual

In no event will Droplet Measurement Technologies, LLC (DMT) be liable for direct, indirect, special,

incidental or consequential damages resulting from any defect or omissions in this manual.

DMT reserves the right to make changes to this manual and the products it describes at any time,

without notice or obligation. Revised editions are found on the manufacturer’s website.

All DMT product names and the Droplet Measurement Technologies Logo are trademarks of Droplet

Measurement Technologies, LLC.

All other brand and product names are trademarks, or registered trademarks, of their respective

owners.

Software License

DMT licenses its software only upon the condition that you accept all of the terms contained in this

license agreement.

This software is provided by DMT “as is” and any express or implied warranties, including, but not

limited to, the implied warranties of merchantability and fitness for a particular purpose are

disclaimed. Under no circumstances and under no legal theory, whether in tort, contract, or


6

otherwise, shall DMT or its developers be liable for any direct, indirect, incidental, special, exemplary,

or consequential damages (including damages for work stoppage; computer failure or malfunction;

loss of goodwill; loss of use, data or profits; or for any and all other damages and losses). Some states

do not allow the limitation or exclusion of implied warranties and you may be entitled to additional

rights in those states.

Warranty

The seller warrants that the equipment supplied will be free from defects in material and

workmanship for a period of eighteen months from date of shipment or 12 months from the date of

either installation or first use whichever comes first. When returning the equipment to DMT for

warranty or service procedures, the equipment owner will pay for shipping to DMT, while DMT will

pay the return shipping expense. Consumable components, such as tubing, filters, pump diaphragms,

and Nafion humidifiers and dehumidifiers are not covered by this warranty.



2.0 Installation and Updating

2.1 Installing IGOR Pro Software

Follow the installation instructions provided by WaveMetrics, the software manufacturer.

PSI recommends creating a folder for all your IGOR procedure files on a backed up hard drive,

e.g.: “M:\LabordeOnly\igor\UserProcs.”

Create a shortcut to the above folder and put this shortcut into IGOR’s user procedure folder

(the subfolder of the program folder), e.g.:

“C:\Program Files\WaveMetrics\Igor Pro Folder\User Procedures.”

View the IGOR Pro tutorials in order to get familiar with the features of this software.

2.2 Installing the SP2 Toolkit

1. Contact DMT for instructions on how to get the toolkit .zip archive:

2. Unzip the archive.

Figure 1: Unzipped Archive


8

3. Copy all procedure files (.ipf files) to the folder that has been created above for storing

the IGOR user procedures, e.g.,

“M:\LabordeOnly\igor\UserProcs”. You can also access this folder by opening Igor Pro

and selecting "Show Igor Pro User Files" from the "Help" menu. The procedure files to

be copied include the following:

• Gysels_SharedProcs_vx.ipf (overwrite any older versions of this file)

• SP2_toolkit_Constants_xxxx.ipf

• SP2_toolkit_GraphProcs_xxxx.ipf

• SP2_toolkit_Procs_xxxx.ipf

These procedure files contain the entire program code for the SP2 Toolkit.

4. Store the SP2 toolkit template “SP2_toolkit_xxxx.pxt” wherever you like. Note: the

template contains the instructions to load all procedure files required for running the

SP2 toolkit.

5. Pressing Ctrl-M will open the main procedure window (shown in Figure 2).

Figure 2: Main Procedure Window for Toolkit

6. Open the template, “SP2_toolkit_xxxx.pxt.”

7. On the SP2_MG menu, select SP2_panel_PSI. This will display the SP2 toolkit panel.

Figure 3: SP2 Toolkit Panel

You are now ready to start using the SP2 toolkit.



2.3 Updating existing IGOR experiments to latest Sp2 toolkit release

1. Contact DMT for information on where to download the latest release of the toolkit.

2. Unzip the archive.

3. Copy all procedure files (.ipf files) to the folder that has been created above for storing

the IGOR user procedures, e.g.,

“M:\LabordeOnly\igor\UserProcs”. The procedure files to be copied include the

following:

• Gysels_SharedProcs_vx.ipf (overwrite any older versions of this file)

• SP2_toolkit_Constants_xxxx.ipf

• SP2_toolkit_GraphProcs_xxxx.ipf

• SP2_toolkit_Procs_xxxx.ipf

These procedure files contain the entire program code for the SP2 Toolkit.

4. Open the file containing the previous SP2 data evaluation.

5. Open the main procedure window by pressing Ctrl-M on the keyboard. Check which release you have been using to analyse the data set so far. In Figure 4, the version number is 1000.

Figure 4: Procedure Window Displaying Version Number

6. Adapt the “#include …” statements in order to load the latest SP2toolkit code. In Figure 5, the new version number is 1100.


1 0

Figure 5: Procedure Window Updated to Load Latest Version Number

7. Kill the previous SP2toolkit panel (close it, not just hiding).

8. Select SP2_panel_PSI from the menu SP2_MG to create the SP2 toolkit panel again.

Note: All settings on the panel are automatically reset and must be re-entered.

3.0 Program Architecture and Data Structure

The following sections provide information on the program architecture and data structure. For

schematics of the Toolkit’s data analysis approach, see Appendix B.

3.1 Raw Data Files

The SP2 data acquisition software writes four different types of files:

• .sp2b files: Main raw data files containing the traces of all triggered particles.

• .hk files: Housekeeping files containing various instrument parameters such as YAG-power, flow rates, pressure, etc. at 1 Hz resolution.

• .ini files: Configuration files containing various software settings such as the trace length, number of triggered particles, trigger threshold, number of saved particles, etc.

• .sum or .log files: Logbook which keeps track of main software actions, i.e. starting new data files etc. (File extension depends on SP2 software version).

3.2 Program Architecture

1. Loading raw data (particle traces and housekeeping data)

see Section 4.0.

2. Trace analysis (determining peak heights for each particle)

see Section 5.0.

3. Post processing particle-by-particle data (e.g. calculation of BC mass from incandescence peak height, calculation of number and mass concentrations and size distributions etc.)


DOC-0359 Rev A-1 1 1 © 2017 DROPLET MEASUREMENT TECHNOLOGIES

see Section 6.0.

4. Further post processing such as calculating time series of various quantities or analysis of LEO-fit results. See sections 6.0 and 7.0.

3.3 Data Structure after Loading Files into Igor Pro

Every detector channel has its specific prefix:

- high gain scattering: “SCHG_...” (previously “scatt_...”)

- high gain broadband incandescence: “BBHG_...” (previously “broad_...”)

- high gain narrowband incandescence: “NBHG_...” (previously “narr_...”)

- high gain split detector: “SPHG_...” (previously “split_...”)

- low gain scattering: “SCLG_...”

- low gain broadband incandescence: “BBLG_...”

- low gain narrowband incandescence: “NBLG_...”

- low gain split detector: “SPLG_...”

And similarly, the combined incandescence data:

- combined BBHG+NBHG: “BHNH_...” (previously “fullBN_...”)

- combined BBHG+NBLG: “BHNL_...”

- combined BBHG+BBLG: “BHBL_...”

- combined BBLG+NBLG: “BLNL_...”

The housekeeping data are loaded to a folder with suffix “YYYYMMDDxnnn_HK” or

“YYYYMMDDhhmmss_HK” according to the name of the housekeeping file “.hk”.


1 2

All raw trace data are loaded to a folder named “YYYYMMDDxnnn_SP2” according to the name

of the raw data file “.sp2b”. The raw trace matrices (BBHG_data, NBHG_data, etc.) can be

deleted after the trace analysis to reduce memory usage considerably. Do not delete any other

wave in this folder. The time stamp of each individual particle is stored in the wave “TimeDate”.

The string “root: YYYYMMDDxnnn_SP2':HKfldrFP” contains the full path to the associated

housekeeping folder if the latter has been loaded.

Within the folder YYYYMMDDxnnn_SP2_PBP you will find all data resulting from the trace

analysis and further post processing on a particle by particle basis. The folder

YYYYMMDDxnnn_SP2_PBP is the folder you should look at most of the time. It contains all the

data in an array type of format. These arrays are used in various combinations to display

different graphs (see “Graphs/Tables” tab).

Information about the content and units of each wave can be found in the wave notes. You can

use the data browser to display them, as shown in Figure 6.

Figure 6: Data Browser Displaying SP2 Wave Data

Some examples of single particle data waves:



Particle type classification Classification and ClassificationByMinCut

Incandescence peak height of the

BBHG-channel

BBHG_FitPeakHt

BC mass of the BBHG-channel BBHG_BCmass

Delay time DelayTimeScattMax2Broad

Concentration time series data are put in the subfolder “:ConcTser.” This subfolder can be

renamed if additional time series data with different time resolution or cut-offs are to be

extracted.

Size distribution time series data are put in the subfolder “:SizeDistTser.” This subfolder can be

renamed if additional time series data with different time resolution or cut-offs are to be

extracted.

LEO-fit results are put in the subfolder “:LEO” and several subfolders therein.

4.0 Loading Data

4.1 Loading Housekeeping Data

The SP2 data acquisition software writes two different types of data files: the housekeeping

data files (“.hk”) and particle trace data files (“.sp2b”). The housekeeping files are required for

information such as the SP2’s sample flow rate or the YAG laser power. Housekeeping files can

automatically be loaded along with the particle trace data (see below) by checking “load

housekeeping file(s)” on the load tab. It is also possible to load the housekeeping data

independently of the particle trace data by following the steps below.


1 4

1. Select load type “all files in folder” and press the button “load housekeeping files.” See

Figure 7.

Figure 7: Loading Housekeeping Data – Step 1

Note: The original 1-s data are remapped to a longer interval if remap when loading is

checked.

2. Select the folder containing the “.hk”-files to be loaded and press OK.

Figure 8: Loading Housekeeping Data – Step 2

The housekeeping data are loaded to a folder called “root:YYYYMMDDxnnn_HK” or

“root:YYYYMMDDhhmmss_HK” depending on the name of the original file. The

housekeeping data can be visualized with the raw data graph “housekeeping data” from

the Graphs/Tables tab.



4.2 Loading Particle Trace Data

4.2.1 SP2 Configuration and Properties

On the Config tab, define the hardware wiring between detectors and DAQ-board channels (box

#1 in Figure 9). The default values are for a factory standard SP2 setup. If you have changed the

order of the detectors (e.g., made the incandescence detector Ch0) you will need to update

these values for each experiment. Then select the detector channels to be loaded to the Igor

experiment (box #2 in Figure 9).

Figure 9: Specifying Data Channels to Be Loaded

1

2


1 6

4.2.2 Setting Data Load and Analysis Parameters

Before loading data, you will need to set parameters on the Load tab. See the definitions below

for information on individual parameters.

Figure 10: Specifying Data Loading Parameters

• load type: select options to load a single or multiple files.

• load one of every n particles: if checked, software will only load one out of every n saved particles is loaded, where n is set with the variable “n=”.

• limit maximum data rate: This option aims at reducing the amount of loaded single particle data at times when particle concentration is high, while retaining good statistics at times when particle concentration is low. Specifically, if checked, all particles are loaded from a raw data file if the trigger rate (averaged over the whole file) was below the specified maximum data rate (given in units of 1000 triggers per hour). If the averaged trigger rate in a raw data file is x times larger than the specified limit, then only one out of every n particles is loaded, with n=floor(x). The fraction of loaded particles is adapted on a file-to-file basis. For example, if the sample concentration increased such that the number of triggers doubled, the software would load half the particles from the subsequent data file.

• concatenate raw files: each raw data file is loaded into a separate folder (if unchecked) or all data are loaded to a single folder (if checked).



• load housekeeping file(s): housekeeping data are automatically loaded (if checked) or taken from previously loaded housekeeping data. Note: only required if sample flow rate data are to be taken from housekeeping file.

• Analyse traces and delete raw data after analysing: See Section 5.0 for details.

• post processing of trace analysis, fit log-normal size distributions, extract concentration time series, and extract size distribution time series: See Section 6.0 for details.

• LEO-fit and post processing of LEO fit: See Section 7.0 for details.

• YAG power [V]: Stores the YAG power value to be used for optical sizing. More details and a description how to reset the YAG power value after file loading are provided in Section 4.2.5. In general, DMT recommends holding the YAG power constant (at the same value applied in scattering calibration work) because it is not necessarily an accurate measure of the inner cavity laser intensity that affects light scattering signals.

4.2.3 Loading and Analysing the Particle-Trace Data

1. Press the button load raw data on the Load tab to load the raw data.

2. If LEO-fit is checked, select folder containing the beam shape data:

Figure 11: Selecting Beam-Shape Data


1 8

3. If loading a single file, select the sp2b-file to be loaded:

Figure 12: Selecting a Single File to Be Loaded

If selecting multiple files, do one of the following:

• Enter the list of files to be loaded (loading “files from list”):

Figure 13: Entering List of Files to be Loaded

• Select the folder containing the shortcuts to sp2b data files (loading “all files in folder”):

Figure 14: Entering Folder from which to Upload Files



4. If the window below appears, select the folder containing the housekeeping data and

press OK:

Figure 15: Selecting Housekeeping Files to Be Uploaded

Note: This dialog appears only if “load housekeeping file(s)” is not checked.

The raw data will be loaded and analysed. See 3.3 for structure of loaded data.

4.2.4 Resetting the Sample Flow Rate Data

The sample flow rate data are required to calculate number and mass concentrations. (This

wave has not been written with older toolkit versions.) The sample flow rate can be reset with

values taken from different sources. There are three options for setting this parameter, as

shown in Figure 16:

• HK: sample flow rate data are taken from the housekeeping file; HK data must have

been loaded beforehand

• INI: the nominal value set in the configuration file (this option is not possible for SP2

software release 4.0 onwards)

• CONST: a constant value set on the panel with “flow [cm³/min]”


2 0

Figure 16: Selecting Source for Sample Flow Rate Data

4.2.5 Resetting the YAG power data

To reset the YAG power data, enter the desired value in YAG power [V] and press write YAG

power wave.

Figure 17: Resetting YAG Power Data



Note: Normally the value for YAG power [V] should be the same as the value used to calculate

the calibration coefficients of the scattering channels from a PSL calibration. This value should

usually be a constant. (The YAG power monitor reading varies with factors such as temperature,

so it may not reflect the true intra-cavity laser power variability.) If the YAG power has changed

since the calibration measurement due to different pump power or mirror contamination,

however, then the YAG power could be adapted accordingly. Ideally a second scattering/PSL

calibration should be performed if the laser power is suspected to have changed, and

coefficients for this second calibration applied to the affected data Contact DMT with questions

with questions about correct treatment of data.

5.0 Trace Analysis

5.1 Preparing Settings for Trace Analysis

5.1.1 Incandescence Channel Settings

The incandescence channels are highlighted below.

Figure 18: Incandescence Channels

fit type for incand. data: The baseline can be obtained by either averaging the pretrigger points

or by fitting a Gaussian to the trace. The latter option is normally not recommended.

match narr to broad: If checked, then the peak of the narrowband channel is only searched in

the range where the broadband peak was observed. This is helpful if the narrowband channel


2 2

contains plenty of noisy spikes. This option is inactive if the fit type “fit Gaussian to get baseline”

is selected.

# points to be averaged for peak height: Number of points around maximum of incandescence

signal that are averaged to get the peak height. Normally 1 or 3 (always an odd number). Note

the same setting must be used for analysing measurements and corresponding calibration data.

filter baseline and full amplitude of baseline noise: This option is only used if the fit type

“average pretrigger points to get baseline” is selected. If filter baseline is checked, then any

signal higher than minimum signal + full amplitude of baseline noise is filtered before averaging

the pretrigger points to get the baseline level.

get # pretrigger pts from configuration file: If checked, then the number of pretrigger points

used for determining the baseline is taken from the number of pretrigger points used during

data recording.

fixed # of pretrigger pts and pretrigger pts for baseline: If fixed # of pretrigger pts is checked,

then the number of pretrigger points used for determining the baseline is taken from the

setvariable # pretrigger pts for baseline.

5.1.2 Scattering Channel Settings

Scattering Channel Settings are shown below.

Figure 19: Scattering Channel Settings

fit type for incand. data: The baseline can be obtained by either averaging the pretrigger points

or by fitting a Gaussian to the trace. The former option is normally recommended.

See Section 5.1.1 for a description of the other settings.



5.1.3 Split Detector (Position Sensitive) Channel

The Split Detector Settings are shown in Figure 20.

Figure 20: Split Detector Settings

See Section 5.1.1 for a description of all possible settings.


2 4

5.1.4 Detector and A-to-D Converter Saturation

Figure 21: Detector and A-to-D Converter Settings

Saturation of the signal normally occurs in the analog to digital converter (DAQ-board). This

means that the maximum possible signal value is 2047 (for a 12-bit A/D-board reporting values

symmetrically about zero). However, it is possible the analogue signal saturates already at small

signal values. In the latter case the signal value corresponding to saturation (i.e. maximum

observed signal minus some allowance for noise) should be entered in the corresponding set

variable. Correct identification of saturated signals is only possible if the saturation limit is

provided.



5.2 Analyzing Raw Traces

Select the channels to be analyzed (if data are available) and post processing options:

Figure 22: Trace Analysis Settings

delete raw data after analysing: If checked, then all raw trace data are deleted from memory after trace fitting. This reduces memory usage massively with the drawback that it is not possible to compare the fit result with the raw traces or to repeat the trace fitting (without reloading the raw traces).

See Section 6.0 for detailed description of the options “post processing of trace analysis”, “fit log-normal size distributions”, “extract concentration time series”, and “extract size distribution time series”.

See Section 7.0 for detailed description of the options “LEO-fit” and “post processing of LEO fit”.

Press the button analyse traces if all options are set. Then you will have to select the folder

containing the raw trace data and confirm it by pressing OK:


2 6

Figure 23: Selecting Raw Trace Data

Trace analysis results will be written into the folder “root:YYYYMMDDxnnn_SP2_PBP”.

6.0 Post Processing Particle-by-particle Data

6.1 Basic Post Processing

A connected part of the raw data can be ignored during data analysis by writing 1 to the

corresponding range of points in the “Mask_BadDataMain” wave in the raw data folder. (Note:

“ignoring” means that neither the measured particle signals nor the corresponding sampling

volume/interval are considered, as if the instrument wasn't running during this period.)



Figure 24: MaskBadData Wave Shown In File Browser

This can be done with the following command for a range of particles by their index:

setdatafolder root:'YYYYMMDDxnnn_SP2'

Mask_BadDataMain[11,101]=1 //filters points 11-101

A certain time period (e.g. 16.06.2010 00:11:53 - 16.06.2010 00:11:53) can be filtered with the

following command sequence:

setdatafolder root:'YYYYMMDDxnnn_SP2'

make /o/n=(numpnts(TimeDate)) Mask_ PeriodToBeFiltered =0

MaskTimeIntervalTStr(TimeDate, Mask_PeriodToBeFiltered, "16.06.2010 00:11:53", "16.06.2010

19:37:00", BadVal=1)

Mask_BadDataMain = Mask_BadDataMain || Mask_ PeriodToBeFiltered //combine the two masks

with logical OR

Warning: This mask wave can also be used to filter outliers (e.g. from electronic signal spikes),

but ONLY IF THE OUTLIERS ARE AN INSIGNIFICANT FRACTION of all triggered particles. If a

significant fraction of the particles are outliers, then a separate filter must be written to get rid

of them. Using this mask wave in such a case would result in an overestimation of the

number/mass concentrations of particles.


2 8

Select the instrument calibration and BC density:

Figure 25: Settings for Calibration Constants and BC Property

• select calib: this field sets the subfolder (in “root:SP2toolkit:CalibData”) which contains the calibration coefficients for the incandescence and scattering channels.

• density of BC particles for diam. calc. [kg/m³]: This density is used to convert the measured BC mass to a mass equivalent diameter of the BC cores. Changing this value will e.g. affect the BC number and size distribution data.



Further settings for basic post processing are shown below.

Figure 26: Additional Post-Processing Settings

# size bins: number of bins for size distributions (number and mass size distribution of BC cores, size distribution of purely scattering particles). The default value is rather high. Reduce the number of size bins (or load more raw data files) if the size distributions are noisy.

min. peak height, …: set the minimum peak height threshold below which the data are ignored in the further data analysis for every channel. Make sure that reliable calibration data are available all the way down to the chosen threshold. The threshold should also be slightly above the corresponding threshold used to trigger the data acquisition during the measurement (use e.g. a threshold of 15 if the trigger threshold was set to 10). If the data analysis threshold is chosen below the trigger threshold, then the resulting size distributions will have an artificial drop across the trigger threshold.

combine BBHG with NBHG: if checked, then the BBHG and NBHG BC data are combined to cover the full measurement range from the lower detection limit of the BBHG channel to the


3 0

upper detection limit of the NBHG channel. Specifically, the BBHG data are used if the BBHG channel is not saturated, the NBHG channel data otherwise. This of course only makes sense if saturation of the NBHG channel occurs at higher BC mass than for the BBHG channel. The checkboxes “combine BBHG with NBLG”, “combine BBHG with BBLG” and “combine BBLG with NBLG” are to combine other pairs of incandescence data in a similar way.

• fit log-normal size distributions: see Section 6.2.

• run LEO post processing: see Section 7.0.

• extract concentration time series: see Section 6.3.

• extract size distribution time series: see Section 6.4.

• sizing of pure scatt: select the refractive index to be used for optical sizing of the purely scattering particles.

• graphs: several standard graphs are created if checked.

The basic post processing of the trace analysis data can be executed by pressing the post

processing of trace analysis button once the above options have been set. This brings up a

browser window that allows you to specify the folder (or multiple folders) containing the trace

analysis data (e.g. 20100618x029_SP2_PBP):

Figure 27: Specifying Trace-Analysis Data Folders

When you have selected the folder(s), press OK.

The results of the basic post processing will be written to the selected folder. Use the wave

notes and the standard graphs from the Graphs/Tables tab to determine the meaning of all

waves.



6.2 Fitting Log-normal Functions to Size Distribution

Figure 28: Log-normal Fit Settings

D-range of measurement to be fitted: The setvariables in this group box are used to set the diameter range of the BC core size distribution which is to be considered for fitting a log-normal distribution.

D-range of fitted log-normal: The setvariables in this group box are used to set the resolution (“number of points”) and the diameter range for which the fitted log-normal curves are calculated.

Note: log-normal fits are done for both BC core number and mass size distributions.


3 2

The log-normal fits can be done by pressing the “fit log-normal to size distributions” button once

the above settings are set:

Figure 29: Specifying Trace-Analysis Data Folder(s)

Select the folder containing the trace analysis data (e.g. 20100618x029_SP2_PBP) and press OK.

The fitted log-normal size distributions will be written to the selected folder. The waves

names contain “…LogNorm…”.



6.3 Extracting Time Series of Concentration Data

Figure 30: Settings for Extracting Concentration Time Series

Set the time resolution (interval of time series [s] (0=all data)) and press the extract

concentration time series button.


3 4

Figure 31: Specifying Folders for Extracting Time Series Data

Select the folder containing the basic post processing data (e.g. 20100618x029_SP2_PBP) and

press OK.

The concentration time series data will be written to the subfolder “:ConcTser” in the

selected folder. This subfolder can be renamed in order to be able to calculate time series with

different time resolution.



6.4 Extracting Time Series of Size Distribution Data

Figure 32: Settings Used in Extracting Time Series of Size Distribution Data

Set the time resolution (interval of SD time series [s] (0=all data)) as well as the size resolution

# bins for size distribution and then press the extract size distribution time series button. Log-

normal fits according to the settings in the group box log-normal fit are calculated if fit log-

normal size distributions is checked.


3 6

Figure 33: Select Data Folder(s) for Extracting Time Series of Size Distribution Data


press OK.

The size distribution time series data will be written to the subfolder “:SizeDistTser” in the

selected folder. This subfolder can be renamed in order to be able to calculate time series with

different time or size resolution.

Fitting the log-normal functions to the size distributions can be added/repeated using the

button “fit log-normal to size distributions” in the group box extract size distribution time

series:

Figure 34: Select Folder for Log-Normal Fitting of Size Distribution Time Series Data

Select the folder containing the size distribution time series data (e.g. “SizeDistTser”) and press

OK.



The fitted log-normal size distributions will be added to this folder.

6.5 Further post processing

6.5.1 Statistics of Bandratio Data

The bandratio for every particle is found in the waves

“root:'YYYYMMDDxnnn_SP2_PBP':BandRatioHG” (ratio of peak heights) and

“root:'YYYYMMDDxnnn_SP2_PBP':BandRatioHGavg” (ratio of peak areas).

Figure 35: add bandratio statistic Button


3 8

Press the button add bandratio statistics to calculate statistics of the band ratio.

Figure 36: Selecting Data Folder for Calculating Bandratio Statistics


press OK.

Figure 37: Setting Bandratio Calculation Options

Set all options and press Continue.

Statistics of the bandratio data will be added to the selected folder. The names of the result

waves contain “…bandratio…Perc…”. Their meaning is described in the wave notes.

6.5.2 Delay time data

The delay time between the peaks of the scattering and incandescence particles for every

particle is found in the wave

“root:'YYYYMMDDxnnn_SP2_PBP': DelayTimeScattMax2Broad” (Note, the measured maximum

of the scattering trace before the half decay of the incandescence peak is used for the position



of the scattering peak, not the centre of the fitted Gaussian => strongly negative numbers are

not possible).

A histogram of measured delay times can be obtained as follows:

Figure 38: Generating Histogram of Delay-Time Data

Press the button add delay time histogram to calculate a histogram of all delay times.


4 0

Figure 39: Selecting Data Folder for Delay-Time Data


press OK.

Figure 40: Setting Options for Delay-Time Histogram

Select the range of broadband peak heights and the range of delay times to be included and

press Continue.

Figure 41: Selecting Bin Numbers for Delay-Time Histogram



Set the number of bins for the delay time histogram and press Continue.

The histogram of measured delay times will be added to the selected folder. The names of

the result waves contain “DelayHisto…”. Their meaning is described in the wave notes.

The number fraction of thickly coated particles in dependence of the incandescence peak height

can be obtained as follows:

Figure 42: add delay time number fractions Button

Press the button add delay time number fractions to calculate the number fraction of thickly

coated particles from the delay time data.


4 2

Figure 43: Selecting Data Folder for Number Fraction of Thickly Coated Particles


press OK.

Figure 44: Specifying Settings for Calculating Number Fraction of Thickly Coated Particles

Select the threshold delay time for thickly coated particles according to the minimum between

the two distinct modes in the delay time histogram. Select the resolution and the range of

broadband peak heights as well as the range of triggered particles to be included in this analysis.

Press Continue.

The number fraction of thickly and thinly coated particles according to the delay time data

will be added to the selected folder. The names of the result waves contain

“DelayNumbFract…”. Their meaning is described in the wave notes. Note that “thinly coated” is

a misnomer, as the coating threshold above which a particle shows up as “thickly coated” is

actually rather high.

6.6 Troubleshooting

Problem Possible cause/solution



Optical sizing of purely scattering particles

fails.

YAG power wave still contains the default

value or the calibration coefficient wave for

the scattering channel still contains the

default value.

Check both and set them correctly.

7.0 LEO-fit This section remains to be explained in detail. Very briefly:

• Load a data file which includes purely scattering particles (without deleting raw traces).

• Run trace analysis and post processing.

• Go to LEO-tab and run "get beam and PSD properties".

• Run "LEO trace analysis" up to ~3% of maximal amplitude and select the folder ..._SP2_PBP':LEO:BeamAndCalib for the beam data (the appropriate limit for the LEO-fit needs to be determined carefully). I recommend to select the option “use fast LEO-fit”.

• Ensure that the baseline of the median baseline is at zero (within <~1‰). If the median baseline deviates from zero, then the trace analysis needs to be repeated with improved “full amplitude of baseline noise”.

• Run "verification and optical sizing" (use RI=2.26+1.26i for core and RI=1.50+0i for coating if no other knowledge for the refractive indices is available).

• Check the scatter plot of the standard and reconstructed scattering peak height. If the slope deviates substantially from unity (>~5%), then you will have to adapt ..._SP2_PBP':LEO:BeamAndCalib:FitSlopeFudgeFactor (constant value for all points) and rerun the verification (without resetting the slope fudge factor) until the slope in the scatter plot becomes ok. Likewise for the PSD channel and the wave ..._SP2_PBP':LEO:BeamAndCalib:SPHG_CalFactMeasFit.

• Now you can use the folder "..._SP2_PBP':LEO:BeamAndCalib" for batch loading of further data (can be copied anywhere if you wish)

• Now you can play with "LEO post processing" and other data files. Make sure that you discard the small optical sizes which are in the noise range.


4 4

8.0 Analyzing Calibration Data

8.1 BC calibration with Separate SP2-files for Each DMA Size

8.1.1 Preparing Waves for Calibration Coefficients

Press the create waves for calib. coef. button:

Figure 45: create waves for calib. coef. Button



Enter a useful folder name for your calibration coefficients:

Figure 46: Specifying Folder Name for Calibration Coefficients

=> A table showing the waves containing the calibration coefficients will automatically pop up.

Select the new calibration coefficients for the further processing of the calibration data:

Figure 47: Selecting Calibration Coefficients


4 6

8.1.2 Load SP2 Calibration Data

Specify the value for # points to be averaged for peak height. Use either 1 or 3. The latter

choice is recommended because it tends to improve the signal to noise a little depending on the

performance of your detectors. Changing this value will change the calibration curves slightly.

This means that the same value has to be used for analyzing a data set as has been used for

determining the selected calibration curve.

Select # points to be averaged for peak height:

Figure 48: Selecting the Number of Points to be Averaged for Peak Height



Set the channel configuration and select the channels to be loaded:

Figure 49: Setting the Channel Configuration and Channels to be Loaded


4 8

Select load options and press the load raw data button:

Figure 50: Load Options and load raw data Button

Select the range of raw data files to be loaded:

Figure 51: Selecting the Range of Raw Data Files to Load



Browse for the directory containing the raw data:

Figure 52: Specifying the Directory for Raw Data

8.1.3 Preparing the Calibration Info Waves

Press the create empty BC calib info waves button:

Figure 53: create empty BC calib info waves Button


5 0

Select the number of calibration points (different dry sizes selected with the DMA):

Figure 54: Selecting the Number of Calibration Points

=> the “CalibInfoTable” will automatically pop up.

8.1.4 Filling in the “CalibInfo” Table

Figure 55: The CalibInfo Table

The CalibInfo table has the following columns:

Sample ID: for your own use.

TimeStartList: start times of calibration points. Automatically determined from start time of

SP2-file.

TimeEndList: end times of calibration points. Automatically determined from end time of file.

TimeCentreList: automatically determined from TimeStartList and TimeEndList

FileDateFirst: date of first SP2 raw data file corresponding to a calibration point.



FileNumFirst: file number of first SP2 raw data file corresponding to a calibration point.

FileDateLast: date of last SP2 raw data file corresponding to a calibration point.

FileNumLast: file number of last SP2 raw data file corresponding to a calibration point.

DiamList: diameter of BC particles selected for a calibration point.

ConcList: mean particle number concentration measured between start and end time by a CPC connected to the DMA in parallel to the SP2. Leave empty if not available.

TClist: temperature in the DMA during a calibration point.

PressList: pressure in the DMA during a calibration point.

8.1.5 Processing the Monodisperse Calibration Data

Select the “BC calibration material,” the press the analyse monodisperse calibration button:

Figure 56: BC Calibration Material Setting and analyze monodisperse calibration Button


5 2

Select the folder containing the calibration info waves:

Figure 57: Selecting Calibration Info Waves

The window below displays various settings, which can be altered (except for the target folder):

Figure 58: Setting Additional Calibration Parameters



Select the SP2 raw data folder containing all calibration measurements:

Figure 59: Selecting Raw Data Folder with Calibration Measurements

=> The complete command for analyzing the monodisperse calibration with above settings will

be printed to the history:

SP2_CalibAnalyzerIncandMono("", DistrMode=0, nHistoBins=75, nSizeBins=100, nGauss=…

Copying the above command is an efficient way to repeat the calibration analysis with slightly

different settings. Note, the complete command has to be stitched together by removing the

line break when it extends across two or more lines.

Plenty of graphs and tables will automatically pop up when the data processing is finished.

Below are definitions for the most important optional parameters in the above command line

(see also notes with the function).


5 4

• DistrMode: see Section 8.1.8.

• nHistoBins: number of bins for peak height histograms. Use rather high number if you want to read the maximum of the mode from the cursor position (see Section 8.1.6). Using fewer bins makes fitting the Gaussians to the peak height modes easier in many cases. The value producing best results may vary between different calibration points.

• nSizeBins: number of size bins used for the size distribution slider discussed in Section 8.1.8.

• nGauss: number of Gaussians fitted. Use integer larger than unity to get the doubly charged mode fitted. nGauss must also be increased if the peak from the saturated signals is bigger than the main mode to be fitted or if some fitted Gaussian just follow the noise.

• FitRangeFact: Determines the range around the maximum in the peak height histograms, which is considered when fitting the Gaussian. Typical values are between ~0.3 (very narrow range considered in the fit) and ~2 (wider range considered for the fit). The value producing best results may vary between different calibration points.

8.1.6 Determining the Peak Height for Each Measured BC Mass

The following example illustrates how to determine peak height for the high gain broadband

channel.

Figure 60: Peak Height Histogram Measured at Calibration Points

This slider graph displays the peak height histogram measured at a calibration points. The singly,

doubly and triply charged particles are clearly visible. The mean peak height of the singly

charged particles has to be determined for every calibration point. The cursor “A” can be set to

the singly charged (leftmost) peak before pressing the get 1e peak from cursor A button.



However, this approach may not be very accurate depending on the resolution of the histogram

resolution and/or the noise in the data. The better alternative is manually copying the fitted

mean diameter from column 0 of the matrix “XXX_GaussModeMat” to the wave

“XXXX_FitPkHt_1e”, both shown in the table below. Only copy those rows with a successful fit,

and be careful not to mix up different detector channels.

Figure 61: CalibFitPkHt Table

The calibrated peak heights copied to “XXXX_FitPkHt_1e” will appear in the following graph

representing the calibration curve:

Figure 62: Graph of Calibration Curve

40

35

30

25

20

15

10

5

0

Part

icle

mass [fg

]

3200280024002000160012008004000

Incendescence peak height [d.u.]

high gain broadband: single charge

inferred from high gain narrowband:

single charge

high gain broadband


5 6

Note that the crosses in the graph appear only if the calibrated peak heights are also determined for the high gain narrowband channel and if the calibration function is repeated either using the corresponding panel button or copying the command from the history.

8.1.7 Fitting the Calibration Curve

Press the fit calibration curve (spline) button:

Figure 63: fit calibration curve (spline) Button



Select the calibration summary folder.

Figure 64: Selecting the Calibration Summary Folder

Select the detector channel:

Figure 65: Selecting the Detector Channel


5 8

=> By default a quadratic polynomial is fitted to the calibration points and the fit results will be

shown in a table and a graph:

Figure 66: Calibration Points and Fit Shown in Table (above) and Graph (below)

Zooming in reveals that the fit is often deficient for the small masses/peak heights:

50

45

40

35

30

25

20

15

10

5

0

Part

icle

mass [fg

]

40003500300025002000150010005000


single charge fitted spline

high gain broadband



Figure 67: Deficient Fit at Small Masses/Peak Heights

The fit can be improved by fitting quadratic spline segments instead. The table below shows

how to modify the “XXXX_SplineCoef” and “XXXX_HoldForFit” waves in order to get a calibration

curve consisting of three quadratic spline segments for the fixed peak height ranges 0-600, 600-

1200 and >1200:

Figure 68: Modifying Waves to Improve Fit in Specific Ranges

Then the fit has to be repeated, either by copying the command from the history or by pressing

the panel button again. The result for our example is:

10

9

8

7

6

5

4

3

2

1

0

-1

Pa

rtic

le m

ass [

fg]

10009008007006005004003002001000


single charge fitted spline

high gain broadband


6 0

and

Figure 69: Graph (above) and Table (below) Showing Corrected Fit for Small Peak Heights

Note, getting the curvature of the calibration curve at the low end right is crucial for deriving the

shape of the mass size distribution near the lower instrument cut-off.

10

9

8

7

6

5

4

3

2

1

0

-1

Pa

rtic

le m

ass [

fg]

10009008007006005004003002001000


high gain broadband: fitted spline single charge

inferred from high gain narrowband:



8.1.8 Storing the Calibration Coefficients and Verification of the

Calibration

When you consider the fit sufficiently good, then the “XXXX_SplineCoefComplete” should be

copied to the corresponding calibration coefficient wave “XXXX_CalCoef” that has been

prepared previously. Afterwards the calibration should be verified. To do this, the data

processing is to be repeated with the optional parameter DistrMode set to 1:

SP2_CalibAnalyzerIncandMono("", DistrMode=1, nHistoBins=300, nSizeBins=100, …

note: the optional parameter nSizeBins determines the resolution of the calculated size

distributions

The “calibration size distribution slider” graphs then reveals whether the calibrated

measurement agrees with the sizes selected by the DMA (rarely perfect for the higher charges):

Figure 70: Comparing Calibrated Measurements to DMA Results

You are done with the calibration of a channel if the above slider shows good results. Otherwise

the peak heights were not properly extracted, or the fitted calibration curve doesn’t fit the

calibration sufficiently well.

8.1.9 Counting Efficiency and Bandratio Calibration

This section will be available soon.


6 2

8.2 Scattering Detector Calibration

This section assumes you have already performed the incandescence detector calibration

described in the previous section.

8.2.1 Loading SP2 Scattering Calibration Data

Select the number of points to be averaged when determining peak heights. As with

incandescence, use either 1 or 3, being sure to use the same value for the rest of the data set. A

value of 3 will improve the signal to noise depending on the performance of the individual SP2

scattering detectors. Choose either load all files in folder or list of files, and load desired .sp2b

files corresponding to the scattering calibration run (usually one file for each PSL size sampled

during the calibration). Load required data using the Load data tab and provide required

information for calibration folder and specific files to load. Make sure concatenate raw data and

analyse traces and post processing of trace analysis options are selected. Other loading options

do not need to be selected, but the calibration processing will still work if they are inadvertently

selected.

The toolkit has an option to normalize the scattering peak height values by the measured YAG

power meter voltage. The scattering response depends on the intensity of the incident light, so

in theory normalizing by the laser power corrects for this effect. The power meter tracks the

inner cavity laser power, but also depends on temperature and other factors related to light that

is transmitted through the output coupler to the YAG power meter. Because of this, normalizing

by YAG power is not recommended at this time. The toolkit default is to not apply a power

correction and instead holds the YAG power at a default value of -9.99. This value is used in both

the calibration and calculation steps, and so cancels out. The actual value of the constant is not

important, but the same value must be used for both calibration processing and data analysis. A

value of -9.99 is useful in that it prevents confusion with the real YAG power value, but also

leads to a negative scattering coefficient. An alternative is to apply a value of 1.0 as a constant,

which keeps the scattering coefficient positive.



8.2.2 Entering Calibration Information

Select create empty scattering calib info waves from the Calibration tab. Enter the full path to

the data folder in Igor Pro that will contain the information waves (by default this is

root:PSLcalib:YYYYMMDD) and enter the number of scattering calibration points.

Figure 71: Entering PSI Scattering Calibration Information

After pressing Continue, a table will appear listing waves describing the calibration data. Users

should enter dates of the calibration files, number of the calibration file (e.g., 2 for

YYYYMMDDx002.sp2b), and the diameter (in nm) that corresponds to the specific calibration file

(e.g., 269 for 269 nm PSL). If multiple files were collected for a single size, this range can be

entered into the FileNumFirst and FileNumLast fields. Otherwise for a single file the first and

last file numbers should be identical. The example below shows partially completed calibration

information waves. The TimeStartList, TimeCentreList and TimeEndList waves are filled

automatically by the toolkit. The ConcList wave is optional and is where CPC concentration

measurements taken during the calibration could be entered.

Figure 72: Waves Describing Calibration Data

8.2.3 Preparing Histograms of Scattering Peak Heights

Select the appropriate refractive index in the pull-down menu in the scattering channel box

under the Configuration tab (typically RI_1dot59_PSL for PSL scattering calibration). Press the

analyse monodisperse calibration button (Figure 73).


6 4

Figure 73: Analyse monodisperse calibration button

The data browser will pop up with a prompt to select the data folder containing the scattering

calibration info waves defined in section 8.2.2. Note: This is not the same folder created for the

incandescence calibration.



Figure 74: Selecting Scattering Calibration Info Waves

After selecting the folder, press OK. A second prompt asks the user to select the folder with the

SP2 raw data. This should have a name with the date and first file number and the suffix "_SP2"

and was generated when the scattering calibration raw data were loaded. Do not select the

"_PBP" data folder.


6 6

Figure 75: Selecting SP2 Raw Data

After selecting the folder, press OK. A dialog box will appear requesting the user input values for

fitting the peak height histograms and selecting the Igor data folder where output will be stored.

Figure 76: Scattering Calibration Optional Parameters



Explanations for these parameters are given below.

target folder for calibration: Default is the "root:PSLcalib:'YYYYMMDD'" where the date

corresponds to the date of the calibration.

maximum number of Gaussians to fit: Default is 3. Select 1 for PSL.

range to be fitted about peak of histogram: Default is 3. Determines how much of histogram to

fit to determine peak mode location. Smaller values restrict the fit to a narrower region

surrounding the peak of the histogram. A value of 1 fits between the approximate location of

the peak half-rise and half-decay.

number of bins for peak height: Default is 200. Can adjust to obtain better histograms

depending on instrument and peak height distribution.

number of leading/trailing histogram bins to ignore: Default is 2. Use larger numbers to

prevent toolkit from fitting noise/saturated signals at extremes of detection range. Will depend

on instrument configuration. If histogram fitting noise/saturated signals increase, exclude more

bins at the upper and lower range of the histogram as needed.

After the user presses Continue, the toolkit will prompt the user to select calibrations for either

high-gain scattering or low-gain scattering channels. Choose high-gain scattering if you have a 4-

channel SP2.

The toolkit will create histograms of measured scattering channel peak heights for each PSL size

and calculate the scattering coefficients for each PSL size based on the location of the peak

value in the histograms. The coefficient is the factor needed to convert the measured peak

height to the scattering cross-section calculated from Mie theory for the scattering calibration

particle diameter. In theory, the same coefficient should be obtained for all PSL sizes; however,

in practice coefficients for different sizes can vary due to variations in the fitting of the peak

height distribution histograms, small fluctuations in the inner-cavity laser power, uncertainty in

the PSL diameter, and other factors. Increasing or decreasing the number of bins in the

histograms can help give better agreement for coefficients determined for different PSL sizes.

The coefficient is also normalized by the YAG power constant used in the section 8.2.1 raw data

loading step. If a negative value is used (e.g., -9.99) the resulting coefficients will also be

negative. The examples shown here use a YAG power value of 1.0.

8.2.4 Inspecting Histograms and Determining Scattering Calibration

Coefficients

The fits to the peak height histograms should be manually inspected to verify they represent the

histograms using the calibration slider graph created in section 8.2.3. The toolkit will calculate a

coefficient even when the scattering particles saturate the detector, so these points should be

identified and ignored. For example, the high-gain scattering detector recorded mostly

saturated signals for 269 nm PSL, as shown below.


6 8

Figure 77: Issues with Fitting the Histogram—Large Number of Saturated Signal

The toolkit still tried to fit the histogram to obtain a peak location, but clearly this fit is not

correct. The histograms should also be inspected to verify the accuracy of the fits for particles

sizes below the saturation limit of the detector. The example below shows the histogram of

peak heights (n = 200) for 220 nm PSL:

Figure 78: Issues with Fitting the Histogram—Too Few Data Points, Too Wide Fit

In this case the number of histograms may be too small and the fit range (3) too wide. Repeating

the monodisperse analysis with 2000 histogram bins and a narrower fitting range (1) yields a

slightly better fit.



Figure 79: Corrected Fit

The Scattering_Calibration_Factor graph that is also generated in Section 8.2.3 is used to

determine the coefficient value to use as the scattering calibration coefficient. Some of the

points in this graph will represent saturated and/or poor fits to the raw peak height histograms

and should be ignored. Consider the graph below.


7 0

Figure 80: Eliminating Invalid Calibration Points

In this case, the largest four diameters saturated the detector and should be ignored. The four

smallest points have closer values, but the 240 nm point is lower compared to the three smallest

diameters. At this point, the decision to include or ignore this calibration point is subjective and

depends on the user analysis goals. It is a good idea to see how big of an impact excluding the

questionable point has on the calibration coefficient and resulting scattering size distributions as

part of the analysis set up.

Once the points used in the calibration have been selected, the next step is to determine the

average coefficient value. An easy way to do this is to use the "WaveStats" command in Igor Pro

combined with cursor selection of the point range. First, bring up the graph info panel by

selecting Show Info using the graph menu in Igor (making sure you have made the calibration

factor the active window by selecting it immediately prior to choosing show info). The shortcut

control-I or command-I will also bring up this sub-panel. Drag the cursors to select the range of

points that will be used in the average. Once the cursors are set, select Wave Stats… from the

Analysis menu in Igor. A dialog box will pop up:



Figure 81: Wave Stats Dialog Box

Select from target and chose the Scatt_Calib… wave that contains the calibration coefficients

calculated for each size. Click the cursors button to use the cursors as the range of points in the

wave to us for the calculation. Press Do it. The statistics should appear in the command window

in Igor. Copy the "V_avg" value to the appropriate box in the "Calib_Coef_table".

Figure 82: Updating Calib_Coef_table

In the above example, we copied 419145 to the SCHG_CalCoef space in the coefficient table.

This value will be used to convert high-gain scattering peak heights to particle diameter in

subsequent calculations/analysis.

Repeat steps 8.2.3 and 8.2.4 for low-gain scattering data if you have an 8-channel instrument.

9.0 Analysis of Ice Core Data This section of the toolkit is in development. Contact DMT for more information.


7 2

Appendix A: Revisions to Manual

Rev. Date Rev. No. Summary Section

10/21/13 A-1 Moved schematics of data analysis approach to

Appendix B

3.3,

Appendix B

Appendix B: Schematics of Data Analysis

Approach

The following pages contain schematics of the PSI Toolkit’s data analysis approach.




7 4




7 6




7 8

Documents

PSI Toolkit Single Particle Soot Photometer (SP2)