Quantifying functional group abundances of ultrafine aerosol particles with a NEMS-FTIR system for source attribution studies

The high abundance of ultrafine particles (PM0.1) in the atmosphere not only implies significant interaction with the already critical climate system, but is also associated with severe health risks. The chemical composition of ultrafine particles is not only decisive for their specific health effects but also carries information on the dominant sources of these particles. However, chemical composition analysis techniques of sub-100 nm particles are currently either complex or high-cost, limiting their applicability to large-scale environmental source attribution.
Nanoelectromechanical sensors coupled to a Fourier transform infrared spectrometer, short NEMS-FTIR, is a promising tool for large-scale chemical characterization of ultrafine aerosol particles. It provides a high sensitivity down to few picograms of sample coupled with easy-to-use sampling directly onto the sensor, and subsequent high throughput analysis at a centralized facility equipped with the IR-spectrometer. 
However, a fundamental step towards the quantification of the chemical composition of ultrafine aerosol samples and related source attribution is the calibration of the sampling system with known functional-group abundance such that IR signals can be translated into quantitative chemical composition data. Here we show the characterization of NEMS for the usage in the sub-100 nm range, enabling the quantification of functional group abundances in ultrafine aerosol samples. 
Using particle number measurements in a simple transmission experiment we show that the size-dependent particle collection efficiency of the NEMS-chips is in the order of around 50% in the sub-100 nm range. The collected ultrafine mass on the filters is verified through ion chromatography and then used to obtain functional group-specific calibration coefficients translating infrared absorbance units into abundance of functional groups. We find detection limits e.g., well below 1 ng of collected ammonium sulfate. 
The knowledge of the resulting calibration curves of individual organic and inorganic compounds will enable chemical composition analysis, which we showcase here with selected ambient air measurements from two very different environments: Vienna, Austria and the highly-polluted station in Sonipat, India, close to New Delhi. The long-term goal focuses on the application of functional group analysis on a bigger amount of ambient air samples covering a broad temporal range, which ultimately enables source apportionment through Positive Matrix Factorization.