Matlab v r2017b

All incinerated virgin and NECs used in this study were provided by Dr. P. Demokritou’s lab at Harvard University following extraction from in-line filters. Generation of incinerated thermoplastic samples using the Harvard INEX system and particle extraction are described elsewhere [94 (link)–96 (link)]. Previous work using these materials fully characterized their physicochemical properties, including CNT loading and signature impurities [29 (link), 42 (link)].
Particle deposition for thermoplastics was modeled using the Harvard Distorted Grid (DG) model as described in GM DeLoid, JM Cohen, G Pyrgiotakis and P Demokritou [97 (link)] to estimate average deposited dose up to 72 h post-treatment. Deposited modeling was conducted using MatLab v. R2017b (MathWorks, Inc., Natick, MA). Except for data derived from dynamic light scattering and deposited modeling, methodological approaches for ascertaining medium density, viscosity, and refractive index may be found within Table S1 and Supplemental Methods.

NMR spectra were imported into MatLab v.R2017b (Mathworks, Natick, WA, USA) using the ProMetab v.1.1 script [53 (link)]. Each raw spectrum consisted of 31,313 data points between −0.60 and 10.00 ppm.
Statistical analyses were performed in MatLab utilizing the PLS Toolbox v. 8.5.2 (Eigenvector Research Inc., Manson, WA, USA). The spectral region containing the residual water peak (δ 4.68–5.00) was removed. Spectra were aligned using the correlation optimized warping algorithm [54 (link)] to account for chemical shift drift, then normalized to total signal area to account for inherent concentration differences between samples. After normalization, spectral regions containing methanol (δ 3.32–3.36) and DSS-d6 (δ 0.4–−0.60) peaks, and the non-informative region beyond 9.00 ppm were removed. Finally, spectra were baseline corrected using automatic weighted least squares, and scaled by mean centering. After editing, a total of 24,349 chemical shift datapoints were included in subsequent statistical analyses.
For multivariate analyses, unsupervised principal component analysis (PCA) was used. Peaks of interest were identified using the Chenomx NMR suite software v.8.4 (Chenomx Inc., Edmonton, AB, Canada), comparison to the literature, and 2D NMR analysis.