Atmospheric high-viscosity organic particles (HVOPs) were observed in samples of ambient aerosols collected in April and May 2016 in the Southern Great Plains of the United States. These particles were apportioned as either airborne soil organic particles (ASOPs) or tar balls (TBs) from biomass burning based on spetro-microscopic imaging and assessments of meteorological records of smoke and precipitation data. Regardless of their apportionment, the number fractions of HVOPs were positively correlated (
Regional and global atmospheric transport models are commonly used to predict the impact of aerosols on radiative forcing (Feng et al., 2013). The efficacy of these models relies on estimates of the types, number concentrations, spatial distribution, and emission sources of aerosols. One challenge of continued scientific discussion is how (and to what extent) industry and other anthropogenic activities contribute to climate forcing. Emissions of soot, one of the most well-studied anthropogenic aerosols emitted by fossil fuel combustion, have been shown to have a strong climate warming factor comparable to carbon dioxide (Bond et al., 2013). However, soot is not the only light-absorbing carbon-containing aerosol of
concern. Less absorbing but often more abundant light-absorbing organic
carbon aerosol, known as brown carbon (
A defining characteristic of
To better characterize highly viscous organic particles (HVOPs) appearing as
solid spheres, samples of ambient particles were collected at the atmospheric radiation measurement (ARM) facility in Lamont, Oklahoma, located in the Southern Great Plains (SGP) as part of the Holistic Interactions of Shallow Clouds, Aerosols, and Land-Ecosystems (HI-SCALE) field campaign (Fast et al., 2019). To compare the spectroscopic signatures of atmospheric particles with SOM typical for the area of study, aquatic samples of the surface layer of muddy puddles were also collected around the sampling site. These aquatic samples were then filtered, nebulized, and the resulting particles were impacted onto microscopy substrates. The purpose of collecting these samples was to use them to compare particle morphology and composition when the AAE was high, as indicated by online measurements performed at the site (Springston et al., 2016). The particle samples were
analyzed with both scanning electron microscopy (SEM) and scanning
transmission X-ray microscopy coupled with near-edge X-ray absorption fine
structure (STXM-NEXAFS) spectroscopy. SEM images were taken at a
75
The present work evaluates the appearance of ASOPs following rain events and
determines their particle-specific spectroscopic characteristics that would
enable us to distinguish them from biomass burning TBs, secondary organic aerosols (SOAs), and other anthropogenic sources (Parworth et al., 2015; Sheridan et al., 2001).
Characterizing the properties and emission sources of unaccounted
Samples of atmospheric particles were collected at the ARM SGP field site
located in north central Oklahoma (36
Particles were collected by impaction using a Micro-Orifice Uniform Deposit
Impactor (MOUDI; MSP 100) attached to a rotating motor that rotates the
eight impaction surfaces (known as stages) to facilitate uniform particle
deposition. The impactor was connected to a
Samples for analysis were selected from two stages with the following
particle size cutoff ranges: stage 7 (0.32 to 0.56
In addition to impaction samples of ambient particles, aquatic samples
containing SOM brine were collected in 50–200 mL aliquots via syringes from
mud puddles surrounding the SGP field site. This was performed to evaluate a
connection between organics from terrestrial aquatic samples, the
hypothesized source of ASOPs, and ambient ASOPs. Four SOM aquatic samples were
collected on 17 May 2016 prior to the offline analysis of microscopy samples. The obtained samples were then nebulized in laboratory experiments using a Collison Nebulizer (3 jet MRE, CH Technologies, United States) and collected on stage 8 of a 10-stage impactor (110-R, MSP, Inc.) to produce 300–500 nm diameter particles, in which subsequent tilted SEM imaging revealed that ASOPs comprised up to 80 % of the particles by number. In addition, an aliquot of 30–40 mL of one SOM sample was used to generate particles by bubbling
The STXM instruments (beamline 5.3.2.2 ALS, Berkeley, CA, United States, and the SM1
beamline CLS, Saskatoon, SK, Canada) used in this work are located in the
Advanced Light Source (ALS) at the Lawrence Berkley National Laboratory and
at the Canadian Light Source (CLS) at the University of Saskatchewan (Kilcoyne et al., 2003; Regier et al., 2007). Briefly, monochromatic soft
X-rays are focused down to a spot size ranging from 20 to 40 nm in diameter.
The sample is raster scanned after a region of sufficient particle
concentration is found, and individual images are captured at selected photon
energies. Maps were collected in addition to spectra, which are images
consisting of eight energies around the elemental absorption
The SEM analysis of particle samples was performed at the Environmental
Molecular Sciences Laboratory (EMSL) located at Pacific Northwest National
Laboratory (PNNL). Particles were imaged using a computer-controlled
scanning electron microscope (FEI Quanta 3D FEG, Hillsboro, AL, United States). SEM
images were initially taken orthogonally to the substrate until a particle-laden region on the substrate was identified. The substrate mount was then
tilted by 75
The particle soot absorption photometer (3-
Spherical HVOPs have similar spectral characteristics to
Time series for
Both the absorption coefficient and AAE time series data were collected with minute time resolution; the data have been averaged over 30 min time windows to emphasize longer-term data trends instead of short-term fluctuations. The AAE sometimes shows an increase after rain events, though it is not consistent; because aerosol production from rain is more complex than rainfall amounts (depending on droplet size and impact velocity, as well as soil characteristics), it is difficult to see a direct correlation between rain and AAE or particle concentration. While ASOP emissions are expected during rain events, precipitation scavenging is also occurring (Joung and Buie, 2015). The net effect of these two competing processes likely depends on many environmental conditions and is not yet clear. During a rain event, ASOPs will contribute to an increased AAE but because AAE is a bulk optical property, the presence of other absorbing aerosols like black carbon or the washout of mineral dust (large particles with a high AAE) can dampen the effect that rain events have on AAE (Bergstrom et al., 2007).
From this time record, a few samples stand out: the nights of 28 April and 5 and 14 May as these samples had elevated AAE greater than 1.7. On 5 May and 28 April, measured AAE values were above 2, warranting the analysis of particle samples (Lack and Langridge, 2013). While the 14 May sample does not show particularly high AAE values compared to the entire time series, it was collected after a heavy rain storm passed through the area. Also, the particle concentration immediately following the rain event on 14 May shows a significant level of precipitation scavenging, reducing the number of background particles present during sampling and concentrating any ASOPs produced (Hegg et al., 2011). Of note, the lower the
Bulk optical properties, like an elevated AAE, may suggest the presence of spherical HVOPs; therefore, these measurements were used to select samples for detailed chemical imaging of particles. First, tilted SEM images were taken, and HVOP fractions were observed in individual samples ranging from 5 % to nearly 70 % by number. Figure 2 shows representative microscopy and spectro-microscopy images for 3 d when HVOP fractions were high. The top row shows the tilted SEM images used to identify HVOPs. Magenta arrows point to a few identified HVOPs to highlight how much they stick out above the substrate compared to the others. Of note, the SEM images also show the presence of what looks like fractal soot particles in the 28 April and 5 May samples.
(Top row) Tilted (75
Second, the same samples were later imaged by STXM spectro-microscopy.
The middle and bottom row of images shown in Fig. 2 are STXM chemical
speciation maps and total carbon absorbance (TCA) maps, respectively.
Following the procedure described in Moffet (2010) for chemical speciation
maps, each pixel is assigned as either inorganic dominant, organic dominant,
or as a region with high
Sample collection information from the seven samples for which the fractions of HVOPs
were calculated is presented in Table 1 below. The highest HVOP fraction was
observed with the samples taken on 5 May. The prevalence of these particles can be seen in Fig. 2 in the top row. Elevated fractions of HVOPs were also found for the 28 April night sample taken at 18:30 LT and for the 14 May sample. While the 28 April and 5 May samples showed elevated HVOP fractions for both stage 7 and 8 samples, the 14 May sample is unique in that a higher HVOP fraction was only found for
the smaller stage. In addition, the 28 April and the 5 May samples both have elevated particle concentrations and
Optical properties of individual HVOPs from this same data set have been
investigated in our previous work (Veghte et al., 2017). There, the complex refractive index from 200 to 1200 nm was calculated for HVOPs found in the 28 April sample using electron energy loss spectroscopy (EELS). The imaginary part (
Ambient sampling information. Stage 8 values in parentheses when available.
Even though elevated HVOP numbers were identified on a number of days, Fig. 1 shows that there is no clear relationship between the particle concentration and AAE. One reason for this is the presence of other absorbing or non-absorbing aerosols which will increase the measured particle concentration while affecting the AAE differently than HVOPs are expected to. To address this, a correlation plot was made (using values found in Table 1) between the AAE values and the HVOP fractions, and a strong correlation was found (
The appearance of viscous HVOPs at the SGP site has been reported previously (Wang et al., 2016). There, they showed that the viscous HVOPs at SGP had an elevated total carbon absorption (TCA; defined by the pre-edge optical density, OD, subtracted from the post-edge OD) compared to other carbonaceous particles of similar sizes. High TCA values indicate particles that were not deformed upon impaction suggesting a high viscosity. Figure 3 shows TCA values of individual particles plotted against circular equivalent diameter (CED) for four samples reported in this work. The 28 April and 5 and 14 May samples all have elevated TCA whereas the 2 May sample shows lower carbon absorption, in line with the TCA values characteristic of lab-generated SOA particles (Wang et al., 2016). Note that the 5 May sample, when the highest HVOP fraction was identified, has the smallest particles with TCA values above the ambient organic particle regions. Contrast this with the 2 May sample, which shows very few particles with high TCA values and a correspondingly low HVOP fraction.
Correlation between total carbon absorption (TCA) and size measured by circular equivalent diameter for four sampling dates with colored best-fit lines (anchored at 0) for each sample. Blue and gray shaded regions show regions characteristic of ambient organic particles and lab-generated secondary organic aerosols reported in previous study (Wang et al., 2016). The sharp cutoff at about 0.2
So far, the above analysis applies to a general class of HVOPs, which can include both ASOP and TB particles. Additional considerations are necessary, however, before any conclusions are tied to ASOPs exclusively. The two samples with the highest HVOP fractions and AAE values were those collected more than 39 h after a rain event. Because the emission of ASOPs is associated with the bursting of bubbles at flooded soil surfaces after rain, the HVOPs found in the 28 April and 5 May samples are likely not locally emitted ASOPs, while ASOPs might be present in the sample of 14 May.
To investigate the nature and source of HVOPs and determine which can be confidently classified as ASOPs, smoke and fire from biomass burning sources (NOAA:OSPO, 2019) and precipitation data (NWS, 2019) were used along with the calculations of air backward trajectories using a hybrid single particle Lagrangian integrated trajectory model (HYSPLIT) (Stein et al., 2015; Rolph et al., 2017). These data for the events when the three samples had elevated HVOP fractions are shown in Fig. 4 for 28 April and 5 and 14 May. Additional information about the HYSPLIT trajectory conditions, as well as trajectories calculated from multiple starting altitudes, is available in Fig. S3.
The 28 April sample had a moderate fraction of HVOPs (
Smoke, fire, and precipitation data along with HYSPLIT back trajectories for three sample dates. The red circle represents the sampling site, while the small red triangles represent fires. The gray overlays seen in the top row represent detected smoke particles (overlapping smoke plumes are shown in darker shades of gray). The bottom row shows the 24 h average precipitation amount over the sampling date. The top row maps were obtained using the AirNow-Tech navigator using the Hazard Mapping System Smoke Product from NOAA (NOAA, 2019). Source: US EPA AirNow-Tech (Sonoma Technologies Inc., 2019). HYSPLIT trajectories for 28 April and 5 May are for 24 h (Stein et al., 2015; Rolph et al., 2017). The 14 May back trajectory was truncated at 10 h due to a rain event with significant precipitation scavenging. Precipitation maps were made using the NWS Advanced Hydrologic Prediction Service (AHPS) (NWS, 2019).
The 5 May sample has the highest HVOP fraction and the highest average AAE, but it had also been days (140 h) since the last rain event, making ASOPs unlikely. There were many fires surrounding the sampling site compared to the other sampling periods, and the back trajectories show air masses passing directly over some of these fires, suggesting the presence of associated smoke emissions. Precipitation data show that no rainfall was observed anywhere near the sampling site. Figure 1 also shows that this sampling date coincided with a slight particle concentration enhancement and the highest
The last sample date shown (14 May) has regions of smoke away from the sampling site with backward trajectories heading from just outside the smoke-filled region. However, because a rain event was recorded 10 h prior to sampling when significant precipitation scavenging was observed (see the particle concentration decrease in Fig. 1), no influence from biomass burning was observed in this sample. The precipitation map shows that precipitation was observed over the sampling site, as well as in many of the surrounding areas (Radke et al., 1980). Because the microscopy samples were taken shortly after it had rained, the
With the HVOPs observed in the microscopy images being around 0.6
The influence of smoke shown in Fig. 4 may account for the enhancement of
HVOP fractions without rainfall in the 28 April and 5 May samples likely due to TBs. The carbon STXM-NEXAFS spectra of TBs have been recorded previously, and their characteristic features are shown in Fig. 5 (Tivanski et al., 2007). The same figure compares the STXM-NEXAFS spectra for both ambient particles collected during this study and lab-generated ASOP proxies. Figure 5b also includes three characteristic spectra of HVOP particles from the 28 April and 5 and 14 May samples. Even though 5 May and 28 April had the highest AAE values and the highest HVOP fractions, many hours since the last rain event along with the presence of smoke suggest they might be TBs, which is consistent with their NEXAFS spectral features. Three apparent peaks are common for these spectra: the
Comparison of NEXAFS spectra between laboratory-generated ASOP
proxies
Upon comparison with the 28 April and 5 May samples, NEXAFS spectra of HVOP particles from the 14 May sample (taken 10 h after raining) show a slightly enhanced
The chemical composition of TBs has been reported in the literature, with fresh
TBs being comprised of various biomass tar products with a substantial degree of
aromaticity. The nonpolar products were most strongly associated with
the wavelength dependence in absorption seen in TBs and were found in greater
number in fresh TBs (Li et al., 2019). Photochemical oxidation in the presence of
Also shown is a spectrum of organic particles not associated with HVOPs. This
spectrum is characterized by small
Figure 5a shows STXM-NEXAFS spectra of ASOP proxies generated
from the SOM brine. The top four spectra from the puddle water samples all
show a fairly strong carbonate signal at around 290.1 eV along with two
broad potassium peaks (
For more quantitative comparison, two sets of peak ratios were calculated:
the first between the
Plot of peak ratios for (blue dots) carboxylic acid peak at 288.6 eV and the carbon double bond peak at 285.3 eV and (red dots) carboxylic acid peak and the COH peak at 286.7 eV. ASOP proxies refer to particles generated via bubbling through the aquatic sample of SOM.
Tilted SEM was used to identify HVOP fractions in a number of samples taken
during this study, and the fractions of HVOPs present for each sample were
determined based on the aspect ratios of individual particles. The HVOP
fractions showed a strong correlation with the average AAE over the sampling
periods with an
Chemical imaging showing the differences between ASOP- and TB-laden samples
was performed using STXM-NEXAFS spectro-microscopy. Samples unaffected by
recent rain, collected while smoke plumes were present, showed a higher
HVOPs are a subclass of
The MATLAB code used for the current work is available as a Supplement.
The data set used here is available for download as a .zip file at
The supplement related to this article is available online at:
MF, DJB, SC, DV, MKG, AL, and RCM were involved in the project conceptualization. MF, DJB, and RCM developed the software. MF and DJB performed the formal analysis. MF, DJB, SC, DQP, DV, JW, GK, and RCM were involved with the experimental investigation. KT, MKG, AL, and RCM provided resources. MF, DJB, SC, DV, and RCM assisted with data curation. MF and DJB assisted in writing the original draft. All authors were involved in writing, reviewing, and editing. MF, DJB, and SC created the displayed visualizations. MKG, AL, and RCM assisted with project supervision, project administration, and funding acquisition.
The authors declare that there is no conflict of interest.
PSAP records obtained from the Atmospheric
Radiation Measurement (ARM) user facility of OBER were used in this study.
The chemical imaging of particles was performed at the Advanced Light Source at
Lawrence Berkeley National Laboratory. The beamline staff of 11.0.2 and
5.3.2 helped make the current work possible, specifically David Shapiro,
David Kilcoyne, Matthew Markus, and Hendrik Ohldag. A portion of the
research was performed using EMSL (grid.436923.9), a DOE Office of Science
user facility sponsored by the Office of Biological and Environmental
Research. Part of the research described in this paper was performed at the
Canadian Light Source, a national research facility of the University of
Saskatchewan, which is supported by the Canada Foundation for Innovation
(CFI), the Natural Sciences and Engineering Research Council (NSERC), the
National Research Council (NRC), the Canadian Institutes of Health Research
(CIHR), the Government of Saskatchewan, and the University of Saskatchewan.
The CLS work was done at the SM beamline 10ID-1 with help of its staff: Jian
Wang, Yingshen Lu, and Jan Geilhufe. The NOAA HYSPLIT transport and
dispersion model (
This research has been supported by the US Department of Energy, Biological and Environmental Research (award no. DE-SC0018948).
This paper was edited by Markus Petters and reviewed by two anonymous referees.