YAADA > Description

YAADA Description

Software Toolkit to Analyze Single-Particle Mass Spectral Data

Researchers are now able to measure the size and composition of single aerosol particles using Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) instruments developed by Prof. Kimberly Prather and her research group at the University of California. In 1996 the Prather group and Prof. Glen Cass's group at the California Institute of Technology measured atmospheric aerosols at three sites in Southern California. During this study, three ATOFMS instruments measured the size of approximately 3 x 106 particles and the composition of 3 x 105 particles. The data collected during such multi-day, multi-instrument study has proven too large for ad hoc data analysis techniques. Since 1996, numerous design improvements to the ATOFMS have resulted in increased particle detection and hit rates; these improvements will likely continue. ATOFMS data management and analysis will require more attention as the number of ATOFMS instruments grows and the quality of these instruments also improves.

YAADA was developed as part of the research collaboration between the Prather and Cass groups to manage and analyze the large ATOFMS data sets. YAADA is an object-oriented data framework and software toolkit written for Matlab, a programming environment widely used for scientific and engineering computation. The main components of YAADA import, query, plot, and quantitatively analyze ATOFMS data. The data are organized as objects which uniquely identify each instrument, particle, spectrum, and spectral peak. The database is organized hierarchically for efficient data access and analysis. Some important features of YAADA are:

YAADA is written in the Matlab programming language and is available as open source software. Users can write Matlab functions to extend YAADA in order to develop novel analyses of ATOFMS data.

Example From Manual

In this section we demonstrate YAADA's query functions and find sets of particles based on their composition using the criteria developed by Noble and Prather (1996). Noble and Prather call the most common particle type for Southern California ``Organic/Nitrate''. These are most likely particles originally emitted by combustion sources on which ammonium nitrate has condensed. These particles have peaks at m/z equal to 12 (C+), 18 (NH4+), 24 (C2+), 30 (NO+), and 36 (C3+). To find this type of particle in the demonstration data use:

>> OrganicNitrate1 = run_query('mz = 12 and mz = 24 and mz = 36 and mz = 18 and mz = 30');

YAADA will find 40 particles that match this criterion. You can view the particles and their mass spectra with the msview program. The sets of particles in the Matlab workspace are shown in a pull-down menu in the upper left corner of the YAADA MS Viewer figure. Select OrganicNitrate and the particles will be listed in the left-hand column. The particles are listed in the left-hand column ordered by the three letter instrument code, date, time, and aerodynamic diameter in um. Click on one of these particles to view its mass spectra (see Figure). The mass spectra are shown as lines at integral mass-to-charge ratios. Some particles have both positive and negative mass spectra, others have only one mass spectrum.

Mass Spectral Viewer showing ``Organic/Nitrate'' particles.

Search criteria can also specify the mass spectral response, ``area'', for each peak. A more complex search for Organic particles is

>> OrganicNitrate2 = run_query('Area{12} > 50 and Area{24} > 75 and Area{36} > 50 and Area{18} > 50 and Area{30} > 75');

Here Area{12} > 200 finds those spectra which have peaks in the mass-to-charge range 11.5 to 12.5 Daltons that have a total response greater than 200. Only 12 particles in the demonstration data set match this criterion. To view these with msview select from the menu MSViewer/Refresh PartID List, then select Organic. Another common class of compounds identified by Noble and Prather is the ``Marine'' particle class identified by ion peaks at m/z = 23, 39, 81, and 83 DA, representing Na+, K+, Na235Cl+ and Na237Cl+, respectively. Search for these with the command

>> Marine = run_query('Area{23} > 1000 and Ar>ea{39} > 200 and Area{81} > 50 and Area{83} > 50');

To analyze many mass spectra, it is useful to aggregate the mass spectra and plot the aggregate. The digital mass spectrum aggregates mass spectra by collecting for each integral m/z value the fraction of spectra that have a peak (Liu et al., 2001). Plot digital mass spectrum of the Marine particles to see the aggregate mass spectrum as

>> figure; % open new figure
>> clf; % clear figure
>> digital_ms(Marine,1,200);

The peaks in the selection criterion, m/z = 23, 39, 81, and 83 DA, are present in nearly all the spectra as expected. The digital mass spectrum shows that many of the spectra also contain distinctive peaks, for example Na2NO3+ (m/z = 108) and Na3SO4+ (m/z = 165).

Digital Mass Spectrum of ``Marine'' Particles