Traffic-Emitted Amines Promote New Particle Formation at Roadsides

Abstract

New particle formation (NPF) is a major source of atmospheric aerosol particles, significantly influencing particle number concentrations in urban environments. High condensation and coagulation sinks at highly trafficked roadside sites should suppress NPF due to the low survival probability of clusters and new particles, however, observations show that roadside NPF is frequent and intense. Here, we investigate NPF at an urban background and roadside site in Central Europe using simultaneous measurements of sulfuric acid, amines, highly oxygenated organic molecules (HOMs), and particle number size distributions. We demonstrate that sulfuric acid and amines, particularly traffic-derived C₂-amines, are the primary participants in particle formation. C₂-amine concentrations at the roadside are enhanced by over a factor of 4 relative to the background, overcoming the effect of enhanced coagulation and condensation sinks. Using machine learning we identify a further but uncertain enhancing role of HOMs. These findings reveal the critical role of traffic emissions in urban NPF.

Keywords: NPF; aerosol; nucleation; pollution; traffic.

1

Diel cycles of particles and gases at the background and roadside sites: (a) the particle number size distribution; (b) concentrations of sulfuric acid dimer and monomer, as well as extremely low volatility organic compounds (ELVOC) and ultralow volatility organic compounds (ULVOC) as measured by NO₃ ^– CIMS, alongside the ACDC-modeled sulfuric acid dimer-DMA cluster concentration; (c) amines as measured by NO₃ ^– CIMS; (d) black carbon, measured by MAAP and condensation sink (CS) as calculated from the particle number size distribution; (e) modeled and measured J_1.5 from ACDC and the particle number size distribution, respectively.

2

Mechanism of particle formation at the roadside and background: (a) relationship between H₂SO₄ dimer and C₂-amines as measured by NO₃ ^– CIMS; (b) H₂SO₄ dimer plotted against H₂SO₄ monomer from our measurements, alongside CLOUD chamber data from ref (c) J _1.5 and J _1.7 versus H₂SO₄ from our measurements, as well as CLOUD chamber studies. Purple points from ref pink and light green points from ref (d) J _1.5 plotted versus (H₂SO₄)₂ ²·C₂-amine^0.5·ULVOC for both measurement sites.

3

Quantitative performance of the machine learning models. (a) Learning curve for the mean absolute error (MAE) of the J ₃. For each training size the mean value and variance are obtained by training the model five times by randomly reshuffling the data set; (b) correlation between the measured and predicted J₃ (cm^–3 s^–1) of the best performing models of the five data set random reshuffles for the background site and roadside site.

4

Top 14 most impactful features for the prediction of J ₃ with the RF models. (a) SHapley Additive exPlanations (SHAP) values for the background model; (b) SHAP values of the roadside model. Higher SHAP values indicate that the feature increased the model’s prediction of J ₃, with color showing whether high or low values of the feature caused that effect. The features are ranked by their highest mean |SHAP| values, with only the top 14 shown. Midday photochemistry, morning photochemistry, and nighttime chemistry refer to the HOM PMF factors of ref A SHAP value of 0 represents the average J ₃ value. Positive SHAP values indicate an increase in J ₃, while negative values indicate a decrease. The spread of SHAP values and the color gradient show the strength and direction of the correlation, with a clear gradient suggesting a strong positive or negative relationship.

5

Simulated particle formation rates under varying scenarios. Hourly daytime maxima in J _1.5 are presented. J _1.5 was modeled using the mean cycle of sulfuric acid, C₂-amine (assumed to be dimethylamine), temperature, relative humidity, and condensation sink on NPF days at each site. The leftmost bar represents background conditions, the rightmost bar depicts roadside conditions, and the middle bar shows roadside conditions adjusted to have the same amine concentrations as the background.