PSI Toolkit for the Single Particle Soot Photometer (SP2 ... · PSI Toolkit for the Single Particle...

C O P Y R I G H T © 2 0 1 3 D R O P L E T M E A S U R E M E N T T E C H N O L O G I E S ,

I N C .

PSI Toolkit for the

Single Particle Soot

Photometer (SP2)

Manual

DOC-0359 Revision A

2545 Central Avenue

Boulder, CO 80301-5727 USA

PSI SP2 Toolkit Manual

DOC-0359 Rev A 2

© 2013 DROPLET MEASUREMENT TECHNOLOGIES, INC.

Copyright © 2013 Droplet Measurement Technologies, Inc.

2545 CENTRAL AVENUE

BOULDER, COLORADO, USA 80301-5727

TEL: +1 (303) 440-5576

FAX: +1 (303) 440-1965

WWW.DROPLETMEASUREMENT.COM

All rights reserved. DMT licenses 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

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.

Trademark Information

All Droplet Measurement Technologies, Inc. product names and the Droplet Measurement

Technologies, Inc. logo are trademarks of Droplet Measurement Technologies, Inc.

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

owners.

Risks of Installing Additional Software

Instrument computers from DMT are configured to acquire data in a reliable, robust manner.

Typically, such instruments are either not connected to a network or are connected to a small,

local network that is isolated from the internet, reducing the risk of viruses. Since anti-virus

programs can cause erratic behavior when run in the background on data acquisition computers,

DMT does not install anti-virus, anti-spam, or anti-malware programs. If you choose to install

these programs, you accept the risk associated with them in terms of potential performance

degradation of the software installed by DMT.

For similar reasons, DMT recommends that you do not install or run other software on the

dedicated instrument computer. Although the installation of some software may be unavoidable,

it is particularly important not to run other software while the computer is acquiring data.


DOC-0359 Rev A 3


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

1.0 Introduction ........................................................................ 7

About this Manual ............................................................................... 7

Additional Resources ............................................................................ 7

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 Data Analysis Approach and Data Structure .............................. 11

3.1 Raw Data Files ..........................................................................11

3.2 Program Architecture ..................................................................11

3.3 Schematics of Data Analysis Approach ..............................................12

3.4 Data Structure after loading into Igor Pro ..........................................18

4.0 Loading Data ..................................................................... 20

4.1 Loading Housekeeping Data ...........................................................20

4.2 Loading Particle Trace Data ..........................................................22

4.2.1 SP2 Configuration and Properties ...............................................22

4.2.2 Setting Data Load and Analysis Parameters ....................................22

4.2.3 Loading and Analyzing the Particle-Trace Data ...............................24

4.2.4 Resetting the Sample Flow Rate Data ..........................................26

4.2.5 Resetting the YAG power data ...................................................27

5.0 Trace Analysis ................................................................... 28

5.1 Preparing Settings for Trace Analysis ................................................28

5.1.1 Incandescence Channel Settings .................................................28

5.1.2 Scattering Channel Settings ......................................................30

5.1.3 Split Detector (Position Sensitive) Channel ....................................30

5.1.4 Detector and A-to-D Converter Saturation .....................................31

5.2 Analyzing Raw Traces ..................................................................32

6.0 Post Processing Particle-by-particle Data ................................. 34

6.1 Basic Post Processing ..................................................................34

6.2 Fitting Log-normal Functions to Size Distribution .................................38

6.3 Extracting Time Series of Concentration Data .....................................40

6.4 Extracting Time Series of Size Distribution Data...................................42

6.5 Further post processing ................................................................44

6.5.1 Statistics of Bandratio Data ......................................................44

6.5.2 Delay time data ....................................................................46


DOC-0359 Rev A 4


6.6 Troubleshooting ........................................................................50

7.0 LEO-fit ............................................................................. 50

8.0 Analyzing Calibration Data ................................................... 51

8.1 BC calibration with Separate SP2-files for Each DMA Size ........................51

8.1.1 Preparing Waves for Calibration Coefficients .................................51

8.1.2 Load SP2 Calibration Data ........................................................53

8.1.3 Preparing the Calibration Info Waves ...........................................56

8.1.4 Filling in the “CalibInfo” Table ..................................................57

8.1.5 Processing the Monodisperse Calibration Data ................................59

8.1.6 Determining the Peak Height for Each Measured BC Mass ...................62

8.1.7 Fitting the Calibration Curve ....................................................64

8.1.8 Storing the Calibration Coefficients and Verification of the Calibration .70

8.1.9 Counting Efficiency and Bandratio Calibration ................................71

8.2 Scattering Detector Calibration ......................................................71

8.2.1 Loading SP2 Scattering Calibration Data .......................................71

8.2.2 Entering Calibration Information ................................................72

8.2.3 Preparing Histograms of Scattering Peak Heights .............................72

8.2.4 Inspecting Histograms and Determining Scattering Calibration Coefficients

76

9.0 Analysis of Ice Core Data ...................................................... 81

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

Figure 1: Unzipped Archive .............................................................. 8

Figure 2: Main Procedure Window for Toolkit ....................................... 9

Figure 3: SP2 Toolkit Panel .............................................................. 9

Figure 4: Procedure Window Displaying Version Number ....................... 10

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

Figure 6: Data Browser Displaying SP2 Wave Data ................................ 19

Figure 7: Loading Housekeeping Data – Step 1 ..................................... 21

Figure 8: Loading Housekeeping Data – Step 2 ..................................... 21

Figure 9: Specifying Data Channels to Be Loaded ................................. 22

Figure 10: Specifying Data Loading Parameters ................................... 23

Figure 11: Selecting Beam-Shape Data .............................................. 24

Figure 12: Selecting a Single File to Be Loaded ................................... 25


DOC-0359 Rev A 5


Figure 13: Entering List of Files to be Loaded ..................................... 25

Figure 14: Entering Folder from which to Upload Files ......................... 25

Figure 15: Selecting Housekeeping Files to Be Uploaded ....................... 26

Figure 16: Selecting Source for Sample Flow Rate Data ......................... 27

Figure 17: Resetting YAG Power Data ............................................... 28

Figure 18: Incandescence Channels .................................................. 29

Figure 19: Scattering Channel Settings .............................................. 30

Figure 20: Split Detector Settings .................................................... 31

Figure 21: Detector and A-to-D Converter Settings .............................. 31

Figure 22: Trace Analysis Settings .................................................... 32

Figure 23: Selecting Raw Trace Data ................................................. 33

Figure 24: MaskBadData Wave Shown In File Browser ........................... 34

Figure 25: Settings for Calibration Constants and BC Property ................ 35

Figure 26: Additional Post-Processing Settings .................................... 36

Figure 27: Specifying Trace-Analysis Data Folders ............................... 37

Figure 28: Log-normal Fit Settings ................................................... 38

Figure 29: Specifying Trace-Analysis Data Folder(s) ............................. 39

Figure 30: Settings for Extracting Concentration Time Series ................. 40

Figure 31: Specifying Folders for Extracting Time Series Data ................ 41

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

........................................................................................... 42

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

Distribution Data ..................................................................... 43

Figure 34: Selecting Folder for Log-Normal Fitting of Size Distribution Time

Series Data ............................................................................ 43

Figure 35: add bandratio statistic Button ........................................... 44

Figure 36: Selecting Data Folder for Calculating Bandratio Statistics ........ 45

Figure 37: Setting Bandratio Calculation Options ................................. 45

Figure 38: Generating Histogram of Delay-Time Data ............................ 46

Figure 39: Selecting Data Folder for Delay-Time Data ........................... 47

Figure 40: Setting Options for Delay-Time Histogram ............................ 47

Figure 41: Selecting Bin Numbers for Delay-Time Histogram .................. 47

Figure 42: add delay time number fractions Button ............................. 48

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

Particles ............................................................................... 49

Figure 44: Specifying Settings for Calculating Number Fraction of Thickly

Coated Particles ..................................................................... 49

Figure 45: create waves for calib. coef. Button ................................... 51

Figure 46: Specifying Folder Name for Calibration Coefficients ............... 52

Figure 47: Selecting Calibration Coefficients ...................................... 52


DOC-0359 Rev A 6


Figure 48: Selecting the Number of Points to be Averaged for Peak Height 53

Figure 49: Setting the Channel Configuration and Channels to be Loaded.. 54

Figure 50: Load Options and load raw data Button ............................... 55

Figure 51: Selecting the Range of Raw Data Files to Load ...................... 55

Figure 52: Specifying the Directory for Raw Data ................................. 56

Figure 53: create empty BC calib info waves Button ............................. 56

Figure 54: Selecting the Number of Calibration Points .......................... 57

Figure 55: The CalibInfo Table ........................................................ 57

Figure 56: BC Calibration Material Setting and analyze monodisperse

calibration Button ................................................................... 59

Figure 57: Selecting Calibration Info Waves ........................................ 60

Figure 58: Setting Additional Calibration Parameters ............................ 60

Figure 59: Selecting Raw Data Folder with Calibration Measurements ...... 61

Figure 60: Peak Height Histogram Measured at Calibration Points............ 62

Figure 61: CalibFitPkHt Table ......................................................... 63

Figure 62: Graph of Calibration Curve ............................................... 64

Figure 63: fit calibration curve (spline) Button ................................... 65

Figure 64: Selecting the Calibration Summary Folder ........................... 66

Figure 65: Selecting the Detector Channel ......................................... 66

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

(below) ................................................................................. 67

Figure 67: Deficient Fit at Small Masses/Peak Heights ........................... 68

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

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

Peak Heights .......................................................................... 69

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

Figure 71: Entering PSI Scattering Calibration Information .................... 72

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

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

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

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

Figure 76: Scattering Calibration Optional Parameters .......................... 75

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

Signal ................................................................................... 77

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

Fit ....................................................................................... 77

Figure 79: Eliminating Invalid Calibration Points .................................. 79

Figure 80: Wave Stats Dialog Box ..................................................... 80

Figure 81: Updating Calib_Coef_table ............................................... 80


DOC-0359 Rev A 7


1.0 Introduction

About this Manual

This manual is closely based on an original manual put together by Martin Gysel and

Marie Laborde. If you find ways to improve it, please email your suggestions to

[email protected].

Additional Resources

DMT maintains a YouTube channel that has a video on installing the Toolkit. Visit

http://www.youtube.com/user/dropletmeasurement to view this video. Other videos

on the Toolkit will be available on this channel soon.

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.

mailto:[email protected]

http://www.youtube.com/user/dropletmeasurement?feature=watch


DOC-0359 Rev A 8


Figure 1: Unzipped 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”. 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).


DOC-0359 Rev A 9


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:


DOC-0359 Rev A 1 0


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 analyze 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.

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.


DOC-0359 Rev A 1 1


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

3.0 Data Analysis Approach and Data Structure

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 Error! Reference source not found..

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.)

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.


DOC-0359 Rev A 1 2


3.3 Schematics of Data Analysis Approach

The following pages contain schematics for the data analysis approach.


DOC-0359 Rev A 1 3



DOC-0359 Rev A 1 4



DOC-0359 Rev A 1 5



DOC-0359 Rev A 1 6



DOC-0359 Rev A 1 7



DOC-0359 Rev A 1 8


3.4 Data Structure after loading 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”.

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).


DOC-0359 Rev A 1 9


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.


DOC-0359 Rev A 2 0


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.


DOC-0359 Rev A 2 1


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.


DOC-0359 Rev A 2 2


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, select the type and order of the data channels to be loaded (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 specify the channels to be

loaded in this experiment (box #1 in Figure 9).

Figure 9: Specifying Data Channels to Be Loaded

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.

1

2


DOC-0359 Rev A 2 3


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: if checked, the toolkit adjusts the fraction of loaded particles automatically from file to file such that the number of loaded particles per hour does not exceed the particle rate set with “max. rate [1000/h]” (limit applies to particle rate averaged over the whole file). For example, if the sample concentration increased such that the number of loaded particles doubled, the software would load a lower fraction of 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 analyzing: See Section 5.0 for details.


DOC-0359 Rev A 2 4


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 loading data 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 Analyzing 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

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


DOC-0359 Rev A 2 5


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


DOC-0359 Rev A 2 6


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 analyzed. See 3.4 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]”


DOC-0359 Rev A 2 7


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.


DOC-0359 Rev A 2 8


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.

5.0 Trace Analysis

5.1 Preparing Settings for Trace Analysis

5.1.1 Incandescence Channel Settings

The incandescence channels are highlighted below.


DOC-0359 Rev A 2 9


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 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 that 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.


DOC-0359 Rev A 3 0


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.


DOC-0359 Rev A 3 1


Figure 20: Split Detector Settings

See Section 5.1.1 for a description of all possible settings.

5.1.4 Detector and A-to-D Converter Saturation

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


DOC-0359 Rev A 3 2


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”.


DOC-0359 Rev A 3 3


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:

Figure 23: Selecting Raw Trace Data

Trace analysis results will be written into the folder

“root:YYYYMMDDxnnn_SP2_PBP”.


DOC-0359 Rev A 3 4


6.0 Post Processing Particle-by-particle Data

6.1 Basic Post Processing

Single outliers or a connected part of the raw data can be filtered by writing 1 to the

corresponding point (or range of points) in the “Mask_BadDataMain” wave in the raw

data folder ( “root:'YYYYMMDDxnnn_SP2':”) to 1:

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


DOC-0359 Rev A 3 5


Please check the note of the wave “Mask_BadDataMain” for a description of the exact

consequence of filtering particles in this way.

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.


DOC-0359 Rev A 3 6


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


DOC-0359 Rev A 3 7


BBHG channel to the 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.


DOC-0359 Rev A 3 8


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.


DOC-0359 Rev A 3 9


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…”.


DOC-0359 Rev A 4 0


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.


DOC-0359 Rev A 4 1


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.


DOC-0359 Rev A 4 2


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.


DOC-0359 Rev A 4 3


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



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: Selecting Folder for Log-Normal Fitting of Size Distribution Time Series Data


DOC-0359 Rev A 4 4


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


DOC-0359 Rev A 4 5


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

Figure 36: Selecting Data Folder for Calculating Bandratio Statistics



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.


DOC-0359 Rev A 4 6


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.


DOC-0359 Rev A 4 7


Figure 39: Selecting Data Folder for Delay-Time Data



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


DOC-0359 Rev A 4 8


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.


DOC-0359 Rev A 4 9


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



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.


DOC-0359 Rev A 5 0


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.


DOC-0359 Rev A 5 1


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


DOC-0359 Rev A 5 2


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


DOC-0359 Rev A 5 3


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


DOC-0359 Rev A 5 4


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

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


DOC-0359 Rev A 5 5


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


DOC-0359 Rev A 5 6


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


DOC-0359 Rev A 5 7


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.


DOC-0359 Rev A 5 8


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

time of SP2-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.


DOC-0359 Rev A 5 9


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


DOC-0359 Rev A 6 0


Select the folder containing the calibration info waves:

Figure 57: Selecting Calibration Info Waves

The window below displays various settings, which can be altered later on (except for

the target folder):

Figure 58: Setting Additional Calibration Parameters


DOC-0359 Rev A 6 1


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).


DOC-0359 Rev A 6 2


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


DOC-0359 Rev A 6 3


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:


DOC-0359 Rev A 6 4


Figure 62: Graph of Calibration Curve

Note that the crosses in the above graph appear only if the calibrated peak heights are

also determined for the high gain narrowband channel and if the calibration function is

repeated once more (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:

40

35

30

25

20

15

10

5

0

Pa

rtic

le m

ass [

fg]

3200280024002000160012008004000

Incendescence peak height [d.u.]

high gain broadband: single charge

inferred from high gain narrowband:

single charge

high gain broadband


DOC-0359 Rev A 6 5


Figure 63: fit calibration curve (spline) Button


DOC-0359 Rev A 6 6


Select the calibration summary folder.

Figure 64: Selecting the Calibration Summary Folder

Select the detector channel:

Figure 65: Selecting the Detector Channel


DOC-0359 Rev A 6 7


=> 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


DOC-0359 Rev A 6 8


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


DOC-0359 Rev A 6 9


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:


DOC-0359 Rev A 7 0


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.


DOC-0359 Rev A 7 1


8.1.9 Counting Efficiency and Bandratio Calibration

This section will be available soon.

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.


DOC-0359 Rev A 7 2


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 rangethat can be entered into the first FileNumFirst and

FileNumLastlast file num lists 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).


DOC-0359 Rev A 7 3


Figure 73: Analyse monodisperse calibration button

[Added red box]

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.


DOC-0359 Rev A 7 4


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.


DOC-0359 Rev A 7 5


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


DOC-0359 Rev A 7 6


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

adjust as needed. [This last sentence didn’t make sense to me.]

After user presses Continue, the toolkit will prompt 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.


DOC-0359 Rev A 7 7


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 better fit.


DOC-0359 Rev A 7 8


[Is this supposed to be the better one? I wasn’t sure how it was improved, probably

b/c I don’t have the scientific background…]

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. In this exampleConsider the graph

below.


DOC-0359 Rev A 7 9


Figure 79: 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:


DOC-0359 Rev A 8 0


Figure 80: 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 81: 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.


DOC-0359 Rev A 8 1


9.0 Analysis of Ice Core Data

This section of the toolkit is in development. Contact DMT for more information.

Date post:	12-Jul-2018
Category:	Documents
Upload:	vuquynh
View:	226 times
Download:	0 times

PSI Toolkit for the Single Particle Soot Photometer (SP2 ... · PSI Toolkit for the Single Particle...

Documents