Smog chamber experiments using ambient air as a starting point can improve our understanding of the evolution of atmospheric particulate matter at timescales longer than those achieved by traditional laboratory experiments. These types of studies can take place under more realistic environmental conditions addressing the interactions among multiple pollutants. The use of two identical smog chambers, with the first serving as the baseline chamber and the second as the perturbation chamber (in which addition or removal of pollutants, addition of oxidants, change in the relative humidity, etc.), can facilitate the interpretation of the results in such inherently complex experiments. The differences of the measurements in the two chambers can be used as the basis for the analysis of the corresponding chemical or physical processes of ambient air.
A portable dual-smog-chamber system was developed using two identical
pillow-shaped smog chambers (1.5 m
Teflon reactors, known as smog or atmospheric simulation chambers, have been valuable research tools for the study of the complex chemical interactions that take place in the atmosphere. Studies using such reactors date back to the 1950s (Finlayson and Pitts, 1976). The use of these chambers eliminates many of the uncertainties resulting from the analysis of ambient observations where several variables, such as weather conditions, pollutant emission rates, dilution, and transport, all contribute to the observed changes (Kim et al., 2009). Typically, these reactors are made of Teflon, though there are some chambers that are made of metal or glass (Cocker et al., 2001a; Paulsen et al., 2005; Kim et al., 2009). The volume of these chambers varies from a few hundred liters up to hundreds of cubic meters, with the larger configurations having lower surface-to-volume ratio, thus minimizing the wall effects (Cocker et al., 2001a).
Chambers are placed either indoors or outdoors with the former having the
advantage of a well-controlled environment with constant temperature, light
intensity, etc. and the latter being able to use natural sunlight (Laity,
1971; Jeffries et al., 1976; Leone et al., 1985; Carter et al., 2005). For
the indoor chambers, a variety of UV light sources can be used including
black light lamps (Laity, 1971), xenon, and argon arc lamps (Warren et al.,
2008). The
Different groups around the world have conducted thousands of smog chamber
experiments in order to simulate the behavior of pollutants in ambient air.
These smog chambers have been used to study, for example, secondary organic
aerosol and its dependence on temperature, relative humidity, UV intensity,
There have been a number of studies that used ambient air as the starting
point of the experiment. Roberts and Friedlander (1976) added
There have been a few efforts to use portable smog chamber facilities for
different applications. For example, Shibuya and Nagashima (1981) used a
portable 4.5 m
The interactions of the walls of the chamber with the pollutants inside it
represent a major experimental challenge and have been the topic of several
studies. Gas-phase pollutants (e.g., ozone) are lost to the walls and
increased relative humidity tends to increase the decay rates measured
(Akimoto et al., 1979). The walls can also serve as a source of OH mainly
outgassing nitrous acid (HONO) (Jeffries et al., 1976; Akimoto et al., 1979;
Carter et al., 1982; Sakamaki et al., 1983; Pitts et al., 1984; Jenkin et
al., 1988; Glasson and Dunker, 1989; Killus and Whitten, 1990;
Finlayson-Pitts et al., 2003). In most cases OH production increases with
temperature, humidity, and
Typically, smog chamber experiments isolate a pollutant or a mixture of pollutants emitted by a source and focus on its chemistry. In most cases, clean air is used as the starting point of the experiment. While the corresponding results are clearly valuable, these experiments might miss the potentially important interactions of the examined chemical system with other pollutants existing in ambient air. To close this major gap between the laboratory studies and the ambient atmosphere, a portable dual-smog-chamber system with UV lights is designed and tested in this study. The chamber has been developed to use ambient air rather than clean air as its starting point. Having the advantage of being portable enhances the opportunities to study several environmental scenarios and simulate the processes occurring in previously out-of-reach chemical regimes (e.g., very aged air masses). The preliminary tests of the operation of this system are presented.
Relatively small Teflon reactors were selected for this system so that they
can be filled in a matter of minutes, while having a volume adequate to
support a 4 h batch experiment, losing less than a third of their volume
based on the standard instrumentation sampling flow rates. A set of two
identical smog chambers was constructed from Teflon (PTFE) film (DuPont PTFE
2 mil, 0.5 m wide; 1 mil
Pictures of the portable dual-chamber system:
We constructed the chambers in our laboratory in Patras. The reactors were
thoroughly cleaned and conditioned before the first characterization
experiment. The cleaning procedure involved first introduction of high
Sampling is alternated between the two chambers every 3 min by an automated three-way valve synchronized with the operation of the corresponding instrumentation. This allows a total duration of the experiments of more than 4 h without the addition of makeup air. In order to eliminate interferences and memory effects due to this periodical alteration of the sampling lines, adequate time (30 s) is allowed within the 3 min sampling cycle for the lines to be flushed with the sample air from the next chamber. This is achieved by synchronizing the line flushing with the measuring instrumentation and discarding the data collected during this 30 s period.
A hemispheric design was selected with sixty 36 W UV light lamps (Osram,
L36W/73) in a hexagonal arrangement. The lamps were mounted on five metal
frames (12 per frame) creating five substructures (Fig. 1b) that can be
easily disassembled and transported. Once assembled the UV light support
structure had a footprint of
A dual-head metal bellows pump (model MB-602) is used to fill the chambers
with ambient air delivering 80 L min
If required clean particle-free air can be introduced in the chambers. Dry
air is generated by an oilless compressor (Bambi VT200D) and further
purified by activated carbon (carbon cap, Whatman), HEPA filters (HEPA
capsule, Pall), and silica gel (silica gel rubin, Sigma-Aldrich). The
compressor and air cleaning system are not used for the actual experiments.
In these experiments the chambers are filled with ambient air without the
use of cleaning devices. The compressor and cleaning system are used for the
cleaning of the chambers between experiments and for blank or other chamber
characterization experiments. Typically, the concentrations of ozone,
A subunit including the above systems (except for the filling pump and the
compressor) was added to one of the metal frames of the system. This subunit
also includes a syringe pump, an atomizer (TSI model 3076), and a silica gel
diffusion drier (silica gel rubin, Sigma-Aldrich) for seed generation.
Additionally, a bubbler subsystem for HONO introduction and an ozone
generator (Azcozon, model HTU-500) were used. The concentration of OH when
HONO was added was estimated in all experiments by the decay of
The set of instruments selected for the use with the chamber system include a
high-resolution time-of-flight mass spectrometer (HR-ToF-AMS, Aerodyne Research Inc.), a proton-transfer-reaction mass spectrometer (PTR-MS, Ionicon Analytik), a
scanning mobility particle sizer (SMPS, classifier model 3080, DMA model
3081, CPC model 3787, TSI), an ozone monitor (API Teledyne, model 400E), and
a
The instrumentation is first used to characterize the ambient conditions for at least a couple of hours. After filling of the chambers is completed, sampling is switched from ambient measurements to the chambers and an initial characterization of the sampled air inside the chambers takes place. Then a perturbation (addition of oxidant or pollutant) is implemented in one of the chambers, while the other is used as a reference. Following the completion of the experiment, ammonium sulfate seeds are introduced into both chambers to measure their loss rate on the walls over time. In this step, the chambers may be refilled with particle-free air. This last stage is used to quantify the particle-size-dependent wall loss rate constants in order to make corrections to the rest of the measurements. Finally, the instrumentation is switched back to ambient observations and the chambers are flushed with either ambient air and/or clean air in preparation for the next experiment.
The system was developed and evaluated in Patras, Greece, and also during the
Finokalia Aerosol Measurement Experiment (FAME 16) campaign. Finokalia is a
remote site in Crete, Greece (Kouvarakis et al., 2000). The field campaign
took place during May–June 2016. Additional tests aimed at improving the
performance of the setup were performed indoors at Carnegie Mellon
University in Pittsburgh, United States. The present work is based on the
results of 51 chamber characterization experiments and seven field test
experiments. The characterization experiments include 15 blank or
contamination-related experiments, 14 experiments characterizing wall
losses, six experiments quantifying the ambient air sampling efficiency,
three
experiments for the measurement of
Tests were conducted in the field in order to assess the potential
contamination of the chambers by ambient air. The chambers were filled with
clean (particle free) air and the particle concentration inside the chambers
was monitored by an SMPS. Figure 2 shows the total number concentrations in
the two chambers. The particle number concentrations remained below 10 cm
Total particle number concentrations as a function of time when the chambers were filled with clear air in the field for leak check of the chambers.
Comparison of the measurements between the two chambers and between
ambient measurements:
Similar results should be obtained when identical experiments take place in
the two chambers in order to safely use one of them as reference. To
establish this, ambient air was introduced in both chambers and the evolution
of the concentrations and composition of the particulate matter and gas
pollutants was measured. An SMPS measured the size distribution and an AMS
the particulate composition. The measured chamber and ambient mass
concentrations (Fig. 3a) and the AMS spectra (Fig. 3c and d) were in good
agreement between the two chambers and the ambient air. The particle mass
concentration in the chambers was approximately 85 % of the ambient
levels. The theta angle (Kostenidou et al., 2009) between the organic aerosol
spectra in the two chambers and the ambient air was in the range of
2.5–6
Pump and tubing losses during the filling procedure were evaluated in order to establish the difference between ambient concentrations and the ones obtained in the chambers. The same number distributions are achieved in both chambers after filling them with ambient air. The penetration efficiency through the tubing and the pump for particles with diameter larger than 80 nm is close to 100 %, while for smaller particles due to higher diffusional deposition the penetration efficiency is 45 %. The length of the tubing is approximately 1 m (with a 0.5 in diameter). The estimated losses for this tube for the flow rates used and for particles in the 20–80 nm size range are 1 %–3 %. Therefore, most of the losses occur in the pump.
Loss of particles to the walls is one of the processes that complicates the analysis of smog chamber experiments. The use of smaller reactors with lower surface-to-volume ratios can accelerate these losses. Disturbances of the Teflon reactors tend to increase the wall loss rates due to the buildup of static charges on the chamber walls. Transporting and installing the reactors also results in higher wall loss rate constants (Wang et al., 2018).
Coagulation-corrected particle wall loss rate constant as a function
of particle size for the two chambers
In order to assess the wall loss behavior of the system, experiments were
conducted in both the laboratory and the field. In all cases, ammonium
sulfate seeds were added to the chambers and their decay with time was
measured. Typically, the chambers were first filled with clean (particle-free) air and then ammonium sulfate seeds were introduced. A solution of 5 g L
In order to assess if it is possible to minimize such charges, a test was conducted in the Teflon reactors in the lab. The chambers were moved in a different location inside the building where the lab is located. The two reactors were handled in exactly the same way simulating the handling during a field deployment. One of the chambers was inflated with air to almost half full while the other was empty. The particle wall losses were measured before and after the movement. Figure 5 represents the loss rate constants in the two chambers because of the movement. The loss rate did not change in the partially inflated chamber. The other chamber, though, experienced an increase in the loss rate constants, almost doubling for the particles in the range 50–200 nm, due to stronger friction of the Teflon walls with each other and thus building static charge. No significant change was noticed for particles larger than 250 nm.
Coagulation-corrected particle wall loss rate constant as a function
of particle size for the two chambers after the movement
The high particle wall losses introduce uncertainty in the results because the wall loss corrections dominate the corrected concentration values. If the losses are very high, the maximum duration of such experiments may be limited. We have been working on developing methods to minimize these effects. Moving the chambers to the field site either fully or at least partially inflated inside our mobile laboratory clearly helps.
Concentrations of the VOCs measured by the PTR-MS were within a few percent
of their ambient levels. In most cases, no noticeable differences were seen.
Tests indicated that there was no detectable contamination due to the metal
bellows pump during the filling process of the two chambers. Vapor loss
experiments were also performed for a few selected VOCs. The measured wall
loss rates were quite low (less than 1 % h
A series of experiments were performed to quantify the loss rates of
The performance of the system was tested in experiments that took place
indoors at Carnegie Mellon University (Center for Atmospheric Particle
Studies). The potential aging of urban background air masses in Pittsburgh,
PA, by OH radicals was used as a pilot study for the system evaluation.
Fewer UV lights were used in this test, resulting in a
Prior to the experiment, both chambers were flushed with particle-free air
overnight under UV illumination to remove any residual particles and
gas-phase organics. Both chambers were filled with ambient air using the
metal bellows pump. During the filling procedure, the instruments measured
ambient conditions. After the addition of ambient air in the chambers,
The wall-loss-corrected SMPS-measured aerosol number concentration.
HONO was added only in the perturbed chamber at
Plots of the evolution of particle number distributions during the
HONO perturbation experiment in Pittsburgh.
The particle wall-loss-corrected concentrations of the major
PM
The wall-loss-corrected total particle number concentration as measured by
the SMPS is shown in Fig. 6. The instruments measured ambient
conditions during the filling process. The average ambient number
concentration was around 2500 cm
Based on the AMS measurements, the ambient air used to fill the chambers
contained on average 3.6
To quantify the secondary aerosol formation, data were corrected for both the
collection efficiency and for particle wall losses. Figure 8 shows the
concentrations of the major PM
The organic mass spectra after filling and after 2 h of UV illumination in the perturbed chamber.
The evolution of the oxygen-to-carbon ratio of the organic aerosol in the two chambers is shown in Fig. 10. The O : C of the ambient organic aerosol and of the initial OA in the two chambers was 0.44. After the OH introduction and the SOA production in the perturbed chamber, the O : C decreased slightly to 0.40. The O : C in the control chamber remained approximately the same. The decrease in the O : C in the perturbed chamber indicates that the additional formed SOA had smaller O : C than the ambient air. The average O : C ratio in other ambient experiments conducted in Pittsburgh was around 0.5, indicating an already moderately oxidized aerosol population. For comparison, the average O : C ratio in the FAME 2008 and FAME 2009 campaigns was 0.8 and 0.5, respectively (Hildebrandt et al., 2010). The results of such ambient air experiments with detailed analysis of the formed aerosol and comparison with ambient and laboratory measurements are included in a forthcoming publication.
The O : C ratio evolution for the control and the perturbed chamber. The shaded area indicates that the chambers were in the dark.
A portable dual-chamber system has been developed for field studies using ambient air as a starting point. The system has been evaluated and no contamination was observed during a typical experiment. The concentrations in the two chambers when filled with ambient air are within a few percent of each other. Particle losses during filling were less than 20 %. No noticeable losses or cross-contamination was observed for the measured VOC species.
Higher wall loss rates were observed when the chambers were deployed in the
field, compared to the lower and stable rates observed when the chambers were
inside the laboratory, due to higher electrostatic charges induced during
their movement. A reduction in the wall loss rates with time was observed when the chambers were deployed in the field,
suggesting that they should be measured after each experiment. The losses can
be reduced if the chambers are transported partially inflated. Initial
laboratory experiments show promising results with respect to potential aging
properties of urban background air in Pittsburgh. An additional
1.5
The data in the study are available from the authors upon request (spyros@chemeng.upatras.gr).
The supplement related to this article is available online at:
CK constructed the facility, participated in the experiments, and wrote the paper. SDJ conducted and analyzed the wall loss and test experiments and contributed to the writing of the paper. EL and KF helped in the construction of the facility and assisted in the experiments. SNP was responsible for the design of the study and the synthesis of the results and contributed to the writing of the paper.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Simulation chambers as tools in atmospheric research (AMT/ACP/GMD inter-journal SI)”. It is not associated with a conference.
This research was supported by the European Research Council Project ATMOPACS (Atmospheric Organic Particulate Matter, Air Quality and Climate Change Studies) (grant agreement 267099). This work has also received funding from the European Union's Horizon 2020 research and innovation program through the EUROCHAMP-2020 Infrastructure Activity under grant agreement no. 730997.
This paper was edited by Hartmut Herrmann and reviewed by two anonymous referees.