KAppa Functional Group Analysis (KAFGA)

KAFGA is MATLAB/GNU Octave code to predict the hygroscopicity parameter kappa for organic compounds from functional group composition.

Short summary Organic particles suspended in air serve as nucleation seeds for droplets in atmospheric clouds. Over time their chemical composition changes towards more functionalized compounds. This work presents a model that can predict an organic compounds’ ability promote the nucleation of cloud drops from its functional group composition. Hydroxyl, carboxyl, aldehyde, hydroperoxide, carbonyl, and ether moieties promote droplet nucleation. Methylene and nitrate moieties inhibit droplet nucleation.

Documentation

The algorithms are described in the manuscript and associated supplement. Model calculations graphed in Figures 1-3 (main text) and Figure 1 (supplement) are in the example folders. The most straightforward example is fs01.

Model input files

• each compound is represented by a model input file in folder fs01comp
• an input file consists of 4 columns and 9 rows
• rows and columns correspond to main and subgroups, respectively

• Specifically:

                      Sub1        Sub2        Sub3        Sub4
          Alkane      CH3         CH2         CH          C
          Alcohol     OH          ---         ---         ---
          Water       H2O         ---         ---         ---
          Carbonyl    CH3C(=O)    CH2(C=O)    ---         ---
          Aldehyde    H(C=O)      ---         ---         ---
          Ether       CH3(O)      CH2(O)      CH(O)       THF
          Acid        C(=O)OH     HC(=O)OH    ---         ---
          Nitrate     CH2(ONO2)   CH(ONO2)    C(ONO2)     ---
          Peroxide    CH2(OOH)    CH(OOH)     C(OOH)      ---

• Example input file for ethanol CH3-CH2OH

                        1           1           0            0
                        1           0           0            0
                        0           0           0            0
                        0           0           0            0
                        0           0           0            0
                        0           0           0            0
                        0           0           0            0
                        0           0           0            0
                        0           0           0            0

• Input files for all compounds in gmd-2015-072 are provided in examples

• A component input structure is needed. These are generated in the
  script, for example, fs01_compounds.m. The fields are: ‘name’:
  compound name, ‘file’: name of input text file (see above), ‘sol’
  observed solubility’, MS_obs: observed molecular weight, rho_obs:
  observed density, Cx, Hx, Ox, Nx: # of C, H, N, O atoms, D:
  diameter for which CCN activity is computed. Note that observed
  quantities are not needed for the model. To see an example for
  initialization/generation of model compounds without observations
  see the structure returned for f03_compounds().

Model execution and output

• For example model execution see file fs01.m
• Multiple files can be present in the input directory
• The output directory must exist. It is possible to combine them
  into a single directory
• A file with interaction coefficients anm, located in src/txt can be
  specified at runtime. Current options are anm_standard and
  anm_revised. anm standard corresponds to standard UNIFAC,
  anm_revised to the coefficients used in gmd-2015-172
• To execute now compounds, create a copy of for example f03 and
  rename it. If the copy is not in the example folder, the sourcepath
  needs to be changed in the execution file, as well as the src files
  main.m, load_coefficients.m and unifac.m
• A summary output file is written. The file is appended each time
  the model runs. The file lists: filenamame (compound), # of C, # of H, # of O, # of N, molecular weight, predicted molar volume, Flory-Huggins kappa, Raoult kappa, UNIFAC kappa (reported in the paper), water
  activity of saturated solution, molefraction of xw of phase bounary
  1,molefraction of xw of phase bounary 2.
• The complete output is also saved as a .mat file for quick loading
• An ASCII file (.out) is written that contains xw, xs, gamma_w, gamma_s
• Example f02 contains an initialization script for serial and
  parallel execution. Parallel execution is not supported in
  Octave. Serial execution for f02 produces the same result in octave
  and matlab. Postprocessing in fs01 is not supported in
  Octave.

Citation

This work was supported by the United States Department of Energy, Office of Science, Biological and Environment Research, Grant number DE-SC0010470. If you use KAFGA in your research, please cite

Petters, M. D., Kreidenweis, S. M., and Ziemann, P. J. (2016). Prediction of cloud condensation nuclei activity for organic compounds using functional group contribution methods, Geoscientific Model Development, 9, 111-124, https://doi.org/10.5194/gmd-9-111-2016.