Authors: Lena Worbs, Tej Varma Yenupuri, Tong You, Filipe R. N. C. Maia
Categories: Article, Imaging and sensing, Optical physics, Biophysics
Source: Communications Physics
Authors: Lena Worbs, Tej Varma Yenupuri, Tong You, Filipe R. N. C. Maia
The study of ultrafine particle aerosols, those with particle diameters of 100 nm or less, is important due to their impact on our health and environment. However, given their small sizes, such particles can be difficult to measure and trace. Most common optical methods are unable to reach this size range. Other methods exist but incur other limitations, such as the need for electrically charged particles. Here we show how light scattering can be used to detect and measure the size and location of single viruses and protein complexes forming an aerosol beam, as well as trace their path. We were able to detect individual particles down to 16 nm in diameter. The primary purpose of our instrument is to monitor the delivery of single bioparticles to the focus of an X-ray laser to image those particles, but it has the potential to study any other aerosols such as those resulting from ultrafine sea spray, with important consequences for cloud formation and climate modeling, or from combustion, responsible for most air pollution and resulting health impacts.
Small particles have a big impact on our life. The range of particles with a diameter below 100 nm, also known as ultrafine particles or PM0.1, includes many important biological entities such as protein complexes and most viruses. The recent COVID-19 pandemic made it evident that understanding the behavior of aerosols containing these minuscule particles holds great importance^1,2^. Furthermore, ultrafine particle aerosols are an important contributor to air pollution, mainly derived from combustion processes, such as in road vehicles. This air pollution inflicts substantial societal and economic burdens on the global environment^3^. Ultrafine particles directly impact our health^4^, first by attacking the respiratory system^5,6^, like other larger particles, but then spreading to other parts of the body^7^. They are involved in cardiovascular diseases^8,9^, damage to the central nervous system^10,11^, increased diabetes^12^ and cancer rates^13^. Naturally generated ultrafine particles, such as the one from sea spray, also have an important impact on our climate through their role in cloud formation^14,15^.
Ultrafine particle aerosols are difficult to measure and visualize due to their small size. Differential mobility particle sizers (DMPS) and scanning mobility particle sizers (SMPS), a combination of a differential mobility analyzer^16^ (DMA) with a condensation particle counter (CPC), are probably the most common way to study such aerosols, measuring their electric mobility. However they have limited temporal resolution and only work with charged aerosols. Optical methods, such as optical particle counters^17^ and aerodynamic particle sizers^18^, while cheap, are typically limited to particles above 100 nanometers in diameter^19^. Microresonator-based systems have shown promise in measuring nanoparticles^20^, but they require the deposition of the aerosol on a substrate and are unable to resolve the spatial distribution of the particles. Recently the ability of mass photometry^21^ to measure the mass of individual, unlabeled biomolecules in dilute solution to high accuracy, using light scattering, has opened a host of exciting applications^22,23^. This inspired us to see how far we could push the direct imaging of PM0.1 aerosols by laser scattering.
Aerosols are also used to carry biological particles to be studied in native mass spectrometry^24^ and single-particle X-ray diffractive imaging (SPI)^25^ on various samples ranging in size from viruses down to single proteins^26–28^. In SPI, 2D X-ray diffraction patterns are recorded from randomly oriented non-crystalline isolated (bio-)nanoparticles to reconstruct their 3D electron density resulting in their structure. Using short X-ray pulses, a diffraction pattern of an intact particle can be recorded before its destruction^29^.
In SPI an aerosol is generated from the sample solution using electrospray ionisation (ESI)^30,31^. The charged aerosol is then neutralized and aerodynamically focused into a particle beam that is intersected by an X-ray pulse in a vacuum chamber^32–34^.
To monitor the quality of the particle beam delivered to the X-rays, the aerosol droplet distribution is typically measured after the ESI source using a differential mobility analyser^16^ in combination with a condensation particle counter (DMA-CPC). However, this distribution does not necessarily reflect the size distribution in the interaction region e.g., due to transmission discrepancies of different sizes in a heterogeneous sample and differences in droplet evaporation.
Previous nanoparticle sizing methods in particle beams were performed using optical scattering^32,35,36^, but were restricted to sizes larger than 40 nm.
In this work, we extend the use of optical scattering microscopy to size small viruses and biological macromolecular complexes in vacuum, down to 16 nm in diameter, which was previously out of reach due to their small size.
By improving the focus of the incident laser and optimizing the experiment geometry we have greatly increased the range of observable particle sizes at the cost of a decreased field of view. This makes it possible to measure the particle size distribution of biologically relevant samples directly in the interaction region. Our method has the potential to make sample delivery of isolated protein complexes in SPI experiments^28^ simpler and more robust opening the way for the study of single-particle ultrafast dynamics. This method can also be used as a general aerosol sizing technique when standard approaches are impractical or impossible.
To size the nanoparticles and characterize the particle-beam properties, we modified our previously used Rayleigh-microscopy setup^32^. In the new design, the laser beam, the particle beam and the CMOS camera are perpendicular to each other (see Fig. 1). The green pulsed Nd:YAG laser (Quantel Evergreen 25100, λ = 532 nm, pulse duration 10 ns, pulse energy 93.3 mJ, repetition rate 15 Hz, single pulse mode) beam is focused using a 50 mm plano-convex lens onto the particle beam, creating a focus size of 35 μm (FWHM) that illuminates the nanoparticles which are imaged through a custom microscope system. The camera and tube lens are set up outside the experimental chamber, whereas the objective is placed inside. The light scattered by the illuminated particles passes through an infinity-corrected 10x (NA = 0.45, 2.7 μm depth of field) objective lens, a 200 mm tube lens and is captured by a CMOS camera (Hamamatsu Orca Flash 4.0 V2).Fig. 1Experimental Rayleigh-scattering microscopy setup.The aerodynamic lens, the objective lens and the optical beam path are perpendicular to each other.
As a comparison measurement we used an electrostatic classifier (TSI model 3080), a differential-mobility analyzer (TSI model 3081) and a condensation particle counter (TSI model 3786) (DMA-CPC) to measure the aerodynamic diameters of the nanoparticles.
The aerosol was generated by an ESI source as previously described^31^. The aerosol then goes through two skimmer pairs to compensate for the high gas input from an ESI source and is focused onto the interaction region by an aerodynamic lens stack (TSI AFL-100). In the skimmer pairs the excess gas is skimmed away using one scroll pump per skimmer box. The particles enter the aerodynamic lens with an entrance pressure of 0.56 mbar and exit through a 1.5 mm aperture, 2 mm above the interaction region in the experimental chamber which is kept at 6 × 10^−5^ mbar. The schematic setup is shown in the Supplementary Fig. 3.
We studied three biological samples with diameters ranging from 13 to 27 nm and for calibration polystyrene spheres (PS) with diameters ranging from 18 to 59 nm. All samples were suspended in 20 mM ammonium acetate (AmAc) solution. We used Bacteriophage MS2, ribosomes (70S ribosomes isolated from Escherichia coli) and apo-ferritin (from equine spleen), with measured aerodynamic diameters of 25.9 nm, 19.5 nm, and 13.6 nm, respectively. From here on, we refer to the biosamples as MS2, ribosomes, and ferritin, respectively. The sample details including manufacturer’s size, DMA-CPC size distributions and particle solution concentrations are given in Supplementary Note 1.
We recorded frames to measure the scattering intensity of the different particles with the laser in single-pulse mode. We collected multiple datasets with a minimum of 1000 frames, where each frame corresponds to a single laser shot. The number of frames was increased to 2000 and 3000 frames for particles with smaller diameters due to decreased hit rates. The camera frames were analyzed with our open-source software package (https://github.com/Toonggg/spts) to determine the particle position and scattered intensity. An example of a single raw frame containing a particle hit is shown in Fig. 2a. The raw data frames were corrected for dead pixels on the camera (see center of frame in Fig. 2a) and the mean laser background, which was recorded without sample being delivered, was subtracted as shown in Fig. 2b. The images were then denoised by convolving them with a Gaussian kernel with a standard deviation of 1.6 pixels. We then threshold the denoised image to determine the pixels with photons with a fixed threshold of 20. Particle hits were determined from the thresholded data and the center was defined as the pixel with the maximum intensity. The total scattered signal per hit was determined by integrating all pixels within a circular window large enough to account for the point spread function of the system. We limited the analysis to focused particles in the laser illumination region. To select the focused particles, we ran the data analysis first with a large circular window of 25 px diameter and a second time with a 5 px circular window. We defined focused particles as those where more than 90% of the scattering intensity determined in the large window fell inside the small 5 px window. Examples of a focused and an unfocused particle hit are shown in Fig. 2c, d, respectively. For each particle size, we determined the peak scattering intensity of all the selected particles.Fig. 2Camera frame examples.a Raw camera image. b Background subtracted camera image in the region of interest with a particle hit. c The particle hit from b.) with the two circular windows used for analysis. A focused particle is shown. Focused particle examples for all particle sized used are shown in Supplementary Fig. 4. d An example of an unfocused particle hit.
To trace individual particles and determine the particle’s speed, we recorded frames with the laser in double-pulse mode. The delay of the pulses was set to 0.3 μs for the PS sample and to 0.2 μs for the biosamples. Each frame corresponds to a double laser shot. The data analysis was performed similar to the data analysis of the single pulse data with the extension of finding frames with a double hit and determining the difference in distance between the hits to determine the particle’s speed.
We calibrated the apparatus by measuring the scattering intensity from mono-disperse aerosols, each containing PS in the size range 18 to 59 nm (peak diameter according to DMA-CPC measurements, see Supplementary Fig. 1a). For particles much smaller than the wavelength of the illumination (dp ≪ λlaser), the scattering follows Rayleigh’s law so the scattered intensity is proportional to the sixth power of their radius. The measured peak sixth root intensity is shown in Fig. 3 as a function of the measured particle peak diameter. We used scipy.optimize.curve_fit to fit a linear function through the \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$
\root 6 \of{I({d}_{{{{\rm{p}}}}})}