Article

pubs.acs.org/est

A Measurement of Total Reactive Nitrogen, NOy, together with NO2,

NO, and O3 via Cavity Ring-down Spectroscopy
Robert J. Wild,†,‡ Peter M. Edwards,†,‡,⊥ William P. Dubé,†,‡ Karsten Baumann,§ Eric S. Edgerton,§
Patricia K. Quinn,∥ James M. Roberts,‡ Andrew W. Rollins,†,‡ Patrick R. Veres,†,‡ Carsten Warneke,†,‡
Eric J. Williams,‡ Bin Yuan,†,‡ and Steven S. Brown*,‡
†
Cooperative Institute for Research in the Environmental Sciences, University of Colorado, Boulder, Colorado 80309, United States
‡
Chemical Sciences Division, Earth System Research Laboratory, National Oceanic and Atmospheric Administration, Boulder,
Colorado 80305, United States
§
Atmospheric Research and Analysis, Inc., Cary, North Carolina 27513, United States
∥
Pacific Marine Environmental Laboratory, National Oceanic and Atmospheric Administration, Seattle, Washington 98115, United
States

ABSTRACT: We present a sensitive, compact detector that measures

total reactive nitrogen (NOy), as well as NO2, NO, and O3. In all
channels, NO2 is directly detected by laser diode based cavity ring-
down spectroscopy (CRDS) at 405 nm. Ambient O3 is converted to
NO2 in excess NO for the O3 measurement channel. Likewise, ambient
NO is converted to NO2 in excess O3. Ambient NOy is thermally
dissociated at ∼700 °C to form NO2 or NO in a heated quartz inlet.
Any NO present in ambient air or formed from thermal dissociation of other reactive nitrogen compounds is converted to NO2
in excess O3 after the thermal converter. We measured thermal dissociation profiles for six of the major NOy components and
compared ambient measurements with other instruments during field campaigns in Utah and Alabama. Alabama measurements
were made in a rural location with high biogenic emissions, and Utah measurements were made in the wintertime in unusual
conditions that form high ozone levels from emissions related to oil and gas production. The NOy comparison in Alabama, to an
accepted standard measurement method (a molybdenum catalytic converter/chemiluminescence instrument), agreed to within
12%, which we define as an upper limit to the accuracy of the NOy channel. The 1σ precision is <30 pptv at 1 s and <4 pptv at 1
min time resolution for all measurement channels. The accuracy is 3% for the NO2 and O3 channels and 5% for the NO channel.
The precision and accuracy of this instrument make it a versatile alternative to standard chemiluminescence-based NOy
instruments.

■ INTRODUCTION
Reactive nitrogen compounds play a central role in atmospheric
they require calibrations at least every few days and must be
reconditioned, or cleaned, every few months of continuous
chemistry. Nitrogen oxides (NOx ≡ NO + NO2) strongly affect operation, although the latter depends on the history of
the oxidative capacity of the atmosphere through the catalytic exposure.1−3 Additionally, the chemical processes involved in
cycle that produces ozone (O3) in the lower atmosphere. Total the catalytic conversion are not fully understood.4 Inlet design
reactive nitrogen (NOy) includes NOx and all its reservoirs: can also play a major role, as some NOy species, notably HNO3,
can suffer significant losses on nonheated inlet surfaces.2,5
NOy ≡ NO + NO2 + NO3 + 2N2O5 + HNO3 + HONO In this paper, we report a new method for detection of NOy
+ HO2 NO2 + PAN (peroxy acetyl nitrates) as part of a compact system that measures NO, NO2, NOy, and
O3 based on cavity ring-down spectroscopy (CRDS). Similar to
+ aerosol nitrates + organic nitrates + ... direct absorption spectroscopy, CRDS is an absolute measure-
ment of trace gas concentration, with an accuracy inherently
Knowledge of the abundance of this chemical family, as well
limited only by knowledge of the absorption cross section and
as NO, NO2, and the related compound O3, is a useful indicator
of total nitrogen emissions, air mass age, competition between potential interfering absorbers. For the instrument described
different chemical processes, and the efficiency of ozone here, small species-dependent corrections must be made only
production associated with particular emission sources. Stand- for sampling losses, conversion efficiencies, and dilution factors.
ard measurements of NOy rely on catalytic decomposition of
NOy to NO, followed by NO detection using chemilumines- Received: April 17, 2014
cence.1,2 The most commonly used materials for conversion are Revised: July 2, 2014
gold and molybdenum. However, catalytic converters are prone Accepted: July 14, 2014
to deterioration, affecting conversion efficiencies. As a result, Published: July 14, 2014

© 2014 American Chemical Society 9609 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615
Environmental Science & Technology Article

Furthermore, the instrument has lower power, size, weight, and mirror degradation in the NOy channel under conditions of
vacuum requirements than a chemiluminescence-based instru- extremely large hydrocarbon concentration in ambient air, to be
ment while approaching its sensitivity, precision, and time discussed below.
response. In the NOy CRDS instrument we describe here, NOy The optical system is mounted in a custom designed cage
and its components are converted into NO2 by thermal system with 1/2 in. carbon fiber rods providing stability against
decomposition (TD) in a fused silica inlet (henceforth referred mechanical and thermal stress. A schematic of the cage system
to as quartz following convention), followed by the addition of can be seen in Figure 1. To increase stability and compactness
ozone to convert NO to NO2. The successful use of TD has
been demonstrated for various individual NOy compounds
(N2O5,6 peroxy nitrates,7,8 HONO,9 ClNO2,10 alkyl nitrates,
and HNO38,11), but the combination of thermal dissociation
plus ozone has not, to our knowledge, been demonstrated
previously for measurement of total NOy. It represents an
analytical simplification for NOy measurements over methods
that only detect NO2 through thermal dissociation, since it does
not require the summation of separate measurements. We
combine this with existing techniques for measuring NO, NO2,
and O3, allowing all four of these important and related species
to be measured simultaneously with precision, accuracy, and
time response sufficient for their measurement in ambient air
across a range of environments and measurement platforms, Figure 1. Schematic of the custom cage system that houses all the
including measurements from moving vehicles. optical components. A large section has been cut out for ease of

■ EXPERIMENT
The instrument is based on in situ detection of NO2 via cavity
illustration. The laser is shown in blue, and the optics have been left in
place. The combination of aluminum plates and carbon fiber rods give
the system high mechanical rigidity, helping to achieve a measurement
precision of a few pptv on a minute time scale.
ring-down spectroscopy.12−16 The working principle of CRDS,
as well as the conversion of O3 and NO to NO2, has been
described in detail in previous papers16,17 and will only be over previous designs, the distance between cavity mirrors has
summarized here with emphasis on the design improvements. been reduced from 1 m to 50 cm. We reuse the cavity mirrors
Sampling is done by pulling ambient air through Teflon inlet with 1 m radius of curvature and have found a minimal loss of
lines or our NOy converter, and data are acquired in 1 s sensitivity. The corresponding ring down time constant (τ0),
intervals. An 80 mW laser diode centered at 405 nm and determined by a combination of mirror reflectivity and Rayleigh
modulated with a 2 kHz square wave provides light for the four scattering losses at ambient pressure, is approximately 30 μs,
measurement channels. A Faraday optical isolator provides with a precision (1σ, 1s) of 6 ns.
protection from optical feedback into the laser. When the light The instrument as a whole is compact and requires low
is modulated off, the intensity inside the optical cavities power and maintenance. It measures approximately 110 cm
decreases exponentially as measured by four photomultiplier high with a 50 by 70 cm footprint and consumes 300 W of
tubes that detect the light intensity transmitted through the rear power at its peak. It weighs 95 kg, which includes the sample
mirrors. The exponential decays are coadded and fitted once pump, a zero air generator, and the data acquisition system.
per second to extract the time constant with (τ) and without The data acquisition system, however, has not been designed
(τ0) the absorber present. The number density of the absorber for low weight and currently contributes 15 kg, a figure which
is then given by could be substantially reduced. The instrument also requires a
cylinder of O2 and a cylinder of NO (2000 ppm in nitrogen).
1 ⎛1 1⎞ These additions can be contained in 1.2 L cylinders mounted
[NO2 ] = ⎜ − ⎟
cσNO2 ⎝ τ τ0 ⎠ (1) directly in the instrument rack and last for 4 days of continuous
operation. Standard 30 L cylinders would provide >120 days of
where c is the speed of light and σNO2 is the NO2 absorption continuous operation.
cross section. Previous 405 nm CRDS instruments from this NO2 is directly measured in all channels via eq 1, whereas
group have included purge volumes adjacent to the mirrors to NO, O3, and NOy are quantitatively converted to NO2 prior to
maintain their cleanliness. Purge volumes require a multi- measurement. Atmospheric NO is converted to NO2 via
plicative factor, RL = d/l in eq 1, where d is the mirror reaction with excess O3, created by flowing O2 over a Hg Pen-
separation and l is the length over which the sample is present. Ray lamp (185 nm). As described in detail by Fuchs et al.,16 the
This factor may have had a dependence on pressure that added O3 results in a small percentage (<1%) of NO2
introduces uncertainties into the measurement.18 The instru- conversion to NO3, which subsequently reacts with NO2 to
ment described here does not currently include mirror purge form N2O5. Similarly, O3 is converted to NO2 via the same
volumes, which simplifies the calibrations and eliminates a reaction by the addition of excess NO as described by
possible source of error. However, the elimination of the purge Washenfelder et al.,17 but without the N2O5 interference
volumes can lead to degradation in mirror reflectivity under because NO + NO3 → 2NO2.
some circumstances. Since the sampled air is in direct contact Conversion of the NOy species is performed via gas-phase
with the mirrors, some compounds can condense onto the thermal dissociation in a quartz inlet, chosen for its high
mirrors, causing a decrease of the time constants. This has been melting point and because it is relatively unreactive.11 Figure 2
observed in an environment with high relative humidity and depicts a conceptual drawing of the NOy converter. During
required drying of the sampled air. We have also encountered normal operations, the front half of a quartz tube is heated such
9610 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615
Environmental Science & Technology Article

The overflow during the zeroing process creates a pressure

change of <0.2 hPa (0.03%) for the annular inlet and a change
of 4 hPa (0.5%) in the NOy inlet. The changes in pressure for
the NOy zeroing changes the Rayleigh scattering of the air
sample, requiring a correction of approximately 60 pptv
equivalent NO2, which is based on well-known Rayleigh cross
sections and is accurate to well within 3 pptv. We zero the
channels every 7 min during mobile platform operations but
increase this time to 15 min for stationary settings (a zero takes
between 30 and 90 s, depending on the length of the inlet).

Figure 2. Diagram of the NOy converter. The heated section consists

of nichrome wire wrapped around a quartz tube. It is thermally
■ RESULTS AND DISCUSSION
Laboratory Tests. To illustrate the conversion of the
insulated with fiberglass insulation and encased in a metal shell ending various NOy components, we show measured temperature
in a cone, protecting the quartz and conducting heat to the tip. A profiles of the signal from several NOy species in Figure 3.
profile of air temperature as a function of position inside the heater
with a flow of 1.5 standard liters per minute at atmospheric pressure is
shown (with set point higher than normal operation). The total
residence time in the heated section is about 48 ms. The zero air
addition is used for a periodic measurement of τ0. Just before the
optical cavity, excess O3 is added to convert any NO to NO2.

that the sample air reaches a temperature of approximately 700

°C. The inset shows a typical temperature profile of the gas in
the heater (set to heat the gas to 750 °C), measured by
insertion of a thermocouple probe in the gas flow during
ambient air sampling. For a 1.5 slpm flow, the plug flow
residence time in the heated section is approximately 48 ms.
The Reynolds number is <1000 for the given range of
temperatures, well in the laminar flow regime. The metal shell
ends in a cone-shaped nozzle to ensure that the quartz is heated Figure 3. (Upper plot) Temperature profiles of several of the most
to the inlet end in order to minimize losses of HNO3, which abundant components of NOy. HNO3 and alkyl nitrates were
can be significant on colder surfaces, especially quartz.5 After compared to a known standard, and the other compounds are scaled
the heated section, NOy components have been converted to to unity at high temperatures. (Lower plot) Temperature profiles from
NO2 (and possibly some NO), which can be transported to the known quantities of possible interference compounds. NH3 in dry air
CRDS measurement cell through an arbitrary length of Teflon resulted in the only measurable interference of <1% at 700 °C. The
tubing without significant losses. Since the converter itself interference was less than 0.1% for air with a relative humidity of 10%
functions as the inlet, we anticipate particulate sampling with or greater.
high efficiency and include particulate nitrate in our NOy
budget.19 Particulate nitrate entering the converter should Many of these conversions have been previously demonstrated
evaporate rapidly, eliminating inertial loss normally associated with heated quartz,7,9−11 and we have repeated tests for the
with particulate sampling. However, we have not attempted to compounds that were readily available. For most samples, the
determine a characteristic cut point (a size at which particles are output concentration was uncalibrated and scaled to unity in
no longer transmitted effectively) for the converter as a whole. Figure 3, but a leveling off at high temperature strongly suggests
The thermal conversion process likely produces mainly NO2, unit conversion. We did not explicitly test conversion of the
although some compounds, such as HONO, may dissociate to nitrate radical, NO3, but expect full conversion at the operating
NO.9 A small fraction of the NO2 may also be reduced to NO temperature.20 We directly compared our measurements of
by reaction with atomic oxygen in the converter.11 Finally, the HNO3 and PAN to that of a heated Mo catalytic converter and
thermal converter is not designed to convert NO to NO2. Any saw full conversion at our operating temperature of 700 °C.
NO in the air sample after the heating and cooling process is The alkyl nitrates (a mix of methyl, ethyl, i-propyl, n-propyl, i-
converted to NO2 via addition of excess O3 just prior to the pentyl, and i-butyl nitrate from a calibration cylinder) show a
CRDS measurement cell, as in the NOx channel. The O3 is rapid conversion to NO2 up to 300 °C, then a slower increase
generated from the same Pen-Ray ultraviolet lamp source and up to 800 °C, where we see full conversion to within the
split evenly between the two channels using a pair of critical cylinder specifications (±10%). The slower conversion at
orifices to divide the flow. higher temperature differs from previously reported temper-
To zero the instrument (i.e., measure τ0 in eq 1), we flush the ature profiles for organic nitrates11 and may be due to the
sample tube by overflowing the inlets with air from which all temperature profile specific to our inlet. Since our design goal is
NOx and O3 has been removed (zero air) that is generated in total rather than speciated NOy, the behavior of organic nitrates
situ (or supplied by a cylinder when necessary). We overflow at intermediate temperatures does not significantly affect the
using an annular inlet (i.e., with zero air added through a large performance of the total NOy measurement.
diameter tube that surrounds and extends just beyond the Some nitrogen-containing gases that are not products of NOx
sample inlet) for the NO, NO2, and O3 channels, and a simple oxidation and therefore not components of NOy as traditionally
tee fitting, shown in Figure 2, is used for the separate NOy inlet. defined could conceivably be converted to NOx at high
9611 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615
Environmental Science & Technology Article

temperatures. If they are present at mixing ratios comparable to

or larger than NOy in the atmosphere, as can be the case for
NH3, N2O, and nitriles, these compounds could represent a
significant interference for an NOy measurement based on
thermal conversion to NO2. We sampled known standards of
NH3, N2O, and acetonitrile with the thermal converter, as
shown in the lower plot of Figure 3. Ammonia resulted in the
only nonzero interference, reaching about 1% conversion at our
working sample temperature of 700 °C. However, this
interference was only present in dry air from a cylinder (<1
ppm water vapor mixing ratio). A relative humidity of 10% was
enough to suppress the NH3 conversion to NO2 to below 0.1%,
such that it can be considered a negligible interference in nearly
all field situations. Like Day et al., we do not expect to be
Figure 4. Dependence of 1σ precision on integration time (Allan
sensitive to aromatic nitro compounds.11 Furthermore, they are deviation plots) for the four channels measuring NO2 in zero air under
not typically expected to be a large fraction of NOy. Reports of laboratory conditions. The dashed lines show the expected trend for
particle bound nitro-PAHs in Los Angeles, for example, have an statistically random noise.
equivalent gas phase concentration <0.1 pptv.21
Other interferences to NO2 detection using 405 nm CRDS
have been described before. A water vapor interference results The uncertainty of the NOy measurement must include the
from the change in Rayleigh scattering, which was measured by conversion efficiency uncertainties for the different NOy
Fuchs et al.16 and results in a small correction to our data. compounds. The comparisons to known concentrations
Absorbing gases other than NO2, such as α-dicarbonyls, are a (HNO3, PAN, and alkyl nitrates) showed full conversion to
within the uncertainty of the standards (approximately 10%,
direct interference but constitute a small percentage of NOx in
from the uncertainty in conversion efficiency of the Mo
most scenarios.16 However, since these compounds require a
converter and the alkyl nitrate cylinder mixing ratio). But for
higher dissociation temperature than our NOy converter the other compounds, quantitative conversion was implied by
provides, they only represent an error in the NO2 baseline. the temperature profile, and uncertainties are unavailable. We
The subtractive measurements of NO, O3, and NOz (≡ NOy − derive a total uncertainty empirically by comparison of the
NOx) are therefore unaffected by α-dicarbonyls. Direct thermal dissociation CRDS instrument to a standard NOy
absorption by ambient O3 in the NO and NO2 channels instrument under field conditions. The results of this
represents a negligible interference because the absorption comparison, discussed below, suggest a limiting uncertainty in
cross section of O3 at 405 nm is 1.5 × 10−23cm2 (∼4 × 104 the NOy measurement of 12%.
times smaller than that of NO2).16,22 For 50 ppbv of O3, the Field Comparisons. The new NOy detection scheme was
optical extinction is 1.9 × 10−11cm−1, or equivalent to quantitatively tested during two field comparisons in 2013. One
approximately 1 pptv of NO2. The optical extinction due to was the Uintah Basin Winter Ozone Study (UBWOS) 2013,
the added 30 ppmv O3 in the NOx and NOy channels is during which our measurement was compared to the sum of
measurable, but this signal is constant across instrument the separately measured NOy components. The other
zeroing and thus does not contribute to the measurement. comparison occurred during the Southeast Oxidant and
However, it provides a convenient means to measure the added Aerosol Study (SOAS), where our total NOy measurement
O3 by switching the O3 addition on and off during periods of was directly compared to total NOy measured with a standard
zero air sampling. Mo catalytic converter.
The NO2 calibrations of the four channels have also been The UBWOS 2013 campaign was a four-week study in the
described in previous papers.16,17 O3 is measured by a Uintah Basin of Utah in February and March. The area is host
commercial UV absorption O3 instrument and then quantita- to extensive oil and gas operations and regularly experiences
tively converted to NO2 in excess NO. The NO2 is then strong temperature inversions. This results in very high ozone
measured by CRDS in the four channels. This provides four events as well as elevated levels of NOz (≡ NOy − NOx,
measurements of the effective absorption cross-section, σNO2, averaging about 12 ppbv over the campaign), making it an ideal
field test for the NO y converter. In addition to our
which differ by less than 2% between the channels. The day-to- measurement of total NOy and NOx, two chemical ionization
day variability of these measurements is less than 1%. Figure 4 mass spectrometers measured concentrations of HNO3,
shows a typical plot showing the relation between integration ClNO2, HONO, and PAN.23,24 A particle-into-liquid sampler
time and 1σ precision (Allan deviation plot) for the four measured particle phase inorganic nitrate, and a separate cavity
channels during sampling of zero air. The dashed lines indicate ring-down system measured NO3 and N2O5.6
the expected square root relationship for statistically random During this campaign we observed degradation of the mirror
noise. All channels follow a nearly statistical noise distribution reflectivity to unworkable values over the course of 5 min time
out to 100 s integration time, resulting in a minimum detection scales when sampling behind the heated inlet and without
limit of a few parts per trillion volume. The uncertainties of the purge volumes in place. The cause of this degradation was not
NO2 measurement in the four channels are due to calibration clear, but it may have been the result of production of
uncertainties and result in a 3% uncertainty in the base condensable gases from hydrocarbon degradation occurring in
measurement.16 Conversion of NO2 to N2O5 in the two the NOy converter. Measured hydrocarbon mixing ratios were
channels that add excess O3 (NOx and NOy) increases the anomalously large, in excess of 0.5 ppmv on average and
uncertainty of NOx measurements up to 5%. frequently several parts per million, due to the location of the
9612 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615
Environmental Science & Technology Article

Figure 5. (a) A representative time series showing several days of measurement comparing the total NOy as measured by our instrument with the
sum of NOy components during the February/March UBWOS field campaign, during which NOx contributed 28% to NOy. (b) Correlation plot of
NOy to the sum of components for the entire campaign. (c) A time series comparing the total NOy measured by TD to total NOy measured by a Mo
catalytic converter used by Atmospheric Research and Analysis, Inc. (ARA) during the June/July SOAS field campaign, during which NOx
contributed 43% to NOy. (d) Correlation plot for the two methods for the entire campaign. All data shown are 5 min averages.

sampling site in a natural gas extraction field.25 Installation of a inlets were separated by about 110 m horizontally and 15 m
hydrocarbon scrubber tube downstream of the NOy converter vertically, with the ARA inlet positioned a few meters above
prevented mirror reflectivity degradation without apparent ground. Although mirror purge volumes were also not used for
influence on measured NOy, increasing response time by a this campaign, a hydrocarbon scrubber as in the UBWOS
mere 150 ms.26 We observed the scrubber to be required for campaign was not required. However, the high humidity of air
measurements during this campaign, though subsequent trials, sampled during summertime in Alabama required drying the
including another campaign in the Uintah Basin in 2014, did sample air with a cold trap for all the channels (after the
not require use of the scrubber. In the upper part of Figure 5, converter for the NOy channel) in order to avoid signal
we show the comparison between our total NOy measurement degradation due to water vapor condensation on the mirrors.
and the sum of the NOy components as measured by different This modification is similar to that used on commercial CRDS
instruments (not including organic nitrates), for which NOx instruments that measure NO2 alone.15 The cold trap reduced
represents a 28% contribution to NOy. Figure 5a shows a the relative humidity to <15% but did not measurably influence
representative time series over a few days. In Figure 5b, we plot the transmission of NO2.
the correlation between the NOy measurement and the sum of The lower part of Figure 5 shows a comparison of the two
the components over the whole campaign. The data presented measurements. In this case, the average NOx contribution to
are 5 min averages because inlets had small differences in
NOy was 43% during the SOAS campaign. High concentration
location (several meters separation) at the field site, and
spikes tended to show poorer agreement, most likely due to the
because of frequent transient high-NOx spikes that were
separation between the inlets. In Figure 5c, we plot a time
inhomogeneously mixed. On average, the new NOy instrument
measured 0.9 ppbv (6%) higher NOy than the sum of series of the two instruments spanning 2 days with low
components. This difference may be due to organic nitrates, occurrences of large concentration spikes. In Figure 5d, we
for which data are not available in 2013, or simply from the show a correlation plot of all the data from the campaign. The
combined uncertainties of the individual measurements. slope is equal to unity to within our 5% measurement
The SOAS campaign was a six-week study in central Alabama uncertainty, but there is an average offset of about 140 pptv.
in June and July of 2013. The site is colocated with a SEARCH This offset may be caused by variation in the sensitivity to
network monitoring station maintained by Atmospheric specific components of NOy, such as coarse aerosol nitrate,
Research and Analysis, Inc. (ARA), which collects long-term between the two instruments, a real difference in ambient NOy
measurements of a suite of gases, including total NOy. The due to the separation of the inlets, a systematic error between
ARA NOy measurement is performed using a standard Mo the two instruments, or a combination of these. Nevertheless,
catalytic converter that converts NOy to NO, followed by the observed offset provides an upper limit for our NOy
chemiluminescent detection of NO. The SOAS campaign thus measurement uncertainty relative to an accepted standard
provided an opportunity for direct comparison between the measurement method. Taking the difference between the two
TD-CRDS NOy and a more conventional NOy instrument. The measurements, dividing it by our measurement, and averaging
9613 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615
Environmental Science & Technology Article

the result over the whole campaign results in a 12% upper limit J.; Newman, L.; Olszyna, K.; Meagher, J.; Hartsell, B.; Edgerton, E.;
uncertainty estimate for our NOy measurement. Pearson, J. R.; Rodgers, M. O. Intercomparison of ground-based NOy
In summary, we have described a compact CRDS instrument measurement techniques. J. Geophys. Res.: Atmos. 1998, 103, 22261−
measuring NO, NO2, O3, and total NOy. Although we 22280.
encountered potential artifacts due to response of the mirrors (3) Crosley, D. R. NOy Blue Ribbon panel. J. Geophys. Res.: Atmos.
1996, 101, 2049−2052.
to relative humidity and extreme VOC concentrations, the
(4) Kliner, D. A. V.; Daube, B. C.; Burley, J. D.; Wofsy, S. C.
instrument in its current form is robust, able to operate Laboratory investigation of the catalytic reduction technique for
essentially unattended for days to weeks at a time. We have measurement of atmospheric NOy. J. Geophys. Res.: Atmos. 1997, 102,
further demonstrated a new technique for measuring NOy using 10759−10776.
thermal decomposition to NO2 in a heated quartz inlet, (5) Neuman, J. A.; Huey, L. G.; Ryerson, T. B.; Fahey, D. W. Study
followed by conversion of NO to NO2 in excess O3. of Inlet Materials for Sampling Atmospheric Nitric Acid. Environ. Sci.
Temperature profiles of individual components are consistent Technol. 1999, 33, 1133−1136.
with unit conversion efficiency for every NOy component (6) Dubé, W. P.; Brown, S. S.; Osthoff, H. D.; Nunley, M. R.; Ciciora,
measured, and there is no significant interference from other S. J.; Paris, M. W.; McLaughlin, R. J.; Ravishankara, A. R. Aircraft
reduced nitrogen compounds. Instrument performance has instrument for simultaneous, in situ measurement of NO3 and N2O5
been demonstrated with comparisons at two ground sites. via pulsed cavity ring-down spectroscopy. Rev. Sci. Instrum. 2006, 77.
These field tests show agreement both with the sum of NOy (7) Wooldridge, P. J.; Perring, A. E.; Bertram, T. H.; Flocke, F. M.;
components measured by separate, independent instruments Roberts, J. M.; Singh, H. B.; Huey, L. G.; Thornton, J. A.; Wolfe, G.
M.; Murphy, J. G.; Fry, J. L.; Rollins, A. W.; LaFranchi, B. W.; Cohen,
and with a measurement of total NOy by a standard
R. C. Total Peroxy Nitrates (Σ PNs) in the atmosphere: the Thermal
molybdenum catalytic converter to within <1% correlation Dissociation-Laser Induced Fluorescence (TD-LIF) technique and
and an absolute offset of 140 pptv. The latter comparison comparisons to speciated PAN measurements. Atmos. Meas. Tech.
allows us to empirically define an upper limit on the NOy 2010, 3, 593−607.
measurement uncertainty of 12%, although the uncertainty may (8) Di Carlo, P.; Aruffo, E.; Busilacchio, M.; Giammaria, F.; Dari-
depend on the relative concentrations of NOy components. Salisburgo, C.; Biancofiore, F.; Visconti, G.; Lee, J.; Moller, S.; Reeves,
The instrument performance was equivalent while sampling C. E.; Bauguitte, S.; Forster, G.; Jones, R. L.; Ouyang, B. Aircraft based
from within a moving vehicle (instrumented van) and is likely four-channel thermal dissociation laser induced fluorescence instru-
to be appropriate for aircraft sampling based on our previous ment for simultaneous measurements of NO2, total peroxy nitrate,
experience.18 total alkyl nitrate, and HNO3. Atmos. Meas. Tech. 2013, 6, 971−980.
Future work will involve investigation of methods to prevent (9) Pérez, I. M.; Wooldridge, P. J.; Cohen, R. C. Laboratory
mirror degradation without the need for drying the sample air evaluation of a novel thermal dissociation chemiluminescence method
or scrubbing it of hydrocarbons. Designs are currently being for in situ detection of nitrous acid. Atmos. Environ. 2007, 41, 3993−
tested for smaller purge volumes that require smaller path 4001.
(10) Thaler, R. D.; Mielke, L. H.; Osthoff, H. D. Quantification of
length corrections than our previously reported instruments, or
Nitryl Chloride at Part Per Trillion Mixing Ratios by Thermal
for testing other methods to make the mirrors less susceptible Dissociation Cavity Ring-Down Spectroscopy. Anal. Chem. 2011, 83,
to condensation effects. The current instrument, while smaller 2761−2766.
than previously reported versions, can be further reduced in (11) Day, D. A.; Wooldridge, P. J.; Dillon, M. B.; Thornton, J. A.;
size and weight to allow for more versatile deployments on Cohen, R. C. A thermal dissociation laser-induced fluorescence
mobile platforms.

■
instrument for in situ detection of NO2, peroxy nitrates, alkyl nitrates,
and HNO3. J. Geophys. Res.: Atmos. 2002, 107, ACH 4-1−ACH 4-14.
AUTHOR INFORMATION (12) Mazurenka, M. I.; Fawcett, B. L.; Elks, J. M.; Shallcross, D. E.;
Corresponding Author Orr-Ewing, A. J. 410-nm diode laser cavity ring-down spectroscopy for
*E-mail: [email protected]. trace detection of NO2. Chem. Phys. Lett. 2003, 367, 1−9.
(13) Wada, R.; Orr-Ewing, A. J. Continuous wave cavity ring-down
Present Address spectroscopy measurement of NO2 mixing ratios in ambient air.
⊥
Department of Chemistry, University of York, York, YO10 Analyst 2005, 130, 1595−1600.
5DD, United Kingdom (14) Hargrove, J.; Wang, L.; Muyskens, K.; Muyskens, M.; Medina,
Notes D.; Zaide, S.; Zhang, J. Cavity Ring-Down Spectroscopy of Ambient
The authors declare no competing financial interest. NO2 with Quantification and Elimination of Interferences. Environ. Sci.

■ ACKNOWLEDGMENTS
The authors thank Joost de Gouw for help with calibrations and
Technol. 2006, 40, 7868−7873.
(15) Castellanos, P.; Luke, W. T.; Kelley, P.; Stehr, J. W.; Ehrman, S.
H.; Dickerson, R. R. Modification of a commercial cavity ring-down
spectroscopy NO2 detector for enhanced sensitivity. Rev. Sci. Instrum.
data analysis. This work was funded in part by an Innovative 2009, 80.
Research Proposal Grant from the Cooperative Institute for (16) Fuchs, H.; Dubé, W. P.; Lerner, B. M.; Wagner, N. L.; Williams,
Research in the Environmental Sciences at the University of E. J.; Brown, S. S. A Sensitive and Versatile Detector for Atmospheric
Colorado and in part by the NOAA Atmospheric Chemistry, NO2 and NOx Based on Blue Diode Laser Cavity Ring-Down
Carbon Cycle and Climate (AC4) Program.

■
Spectroscopy. Environ. Sci. Technol. 2009, 43, 7831−7836, PMID:
19921901.
REFERENCES (17) Washenfelder, R. A.; Wagner, N. L.; Dubé, W. P.; Brown, S. S.
(1) Fahey, D.; Eubank, C.; Hübler, G.; Fehsenfeld, F. Evaluation of a Measurement of Atmospheric Ozone by Cavity Ring-down Spectros-
catalytic reduction technique for the measurement of total reactive copy. Environ. Sci. Technol. 2011, 45, 2938−2944.
odd-nitrogen NOy in the atmosphere. J. Atmos. Chem. 1985, 3, 435− (18) Wagner, N. L.; Dubé, W. P.; Washenfelder, R. A.; Young, C. J.;
468. Pollack, I. B.; Ryerson, T. B.; Brown, S. S. Diode laser-based cavity
(2) Williams, E. J.; Baumann, K.; Roberts, J. M.; Bertman, S. B.; ring-down instrument for NO3, N2O5, NO, NO2 and O3 from aircraft.
Norton, R. B.; Fehsenfeld, F. C.; Springston, S. R.; Nunnermacker, L. Atmos. Meas. Tech. 2011, 4, 1227−1240.

9614 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615

Environmental Science & Technology Article

(19) Rollins, A. W.; Smith, J. D.; Wilson, K. R.; Cohen, R. C. Real

Time In Situ Detection of Organic Nitrates in Atmospheric Aerosols.
Environ. Sci. Technol. 2010, 44, 5540−5545.
(20) Johnston, H. S.; Cantrell, C. A.; Calvert, J. G. Unimolecular
decomposition of NO3 to form NO and O2 and a review of N2O5/
NO3 kinetics. J. Geophys. Res.: Atmos. 1986, 91, 5159−5172.
(21) Reisen, F.; Arey, J. Atmospheric Reactions Influence Seasonal
PAH and Nitro-PAH Concentrations in the Los Angeles Basin.
Environ. Sci. Technol. 2005, 39, 64−73.
(22) Axson, J. L.; Washenfelder, R. A.; Kahan, T. F.; Young, C. J.;
Vaida, V.; Brown, S. S. Absolute ozone absorption cross section in the
Huggins Chappuis minimum (350 - 470 nm) at 296 K. Atmos. Chem.
Phy. 2011, 11, 11581−11590.
(23) Slusher, D. L.; Huey, L. G.; Tanner, D. J.; Flocke, F. M.;
Roberts, J. M. A thermal dissociation-chemical ionization mass
spectrometry (TD-CIMS) technique for the simultaneous measure-
ment of peroxyacyl nitrates and dinitrogen pentoxide. J. Geophys. Res.:
Atmos. 2004, 109.
(24) Roberts, J. M.; Veres, P.; Warneke, C.; Neuman, J. A.;
Washenfelder, R. A.; Brown, S. S.; Baasandorj, M.; Burkholder, J. B.;
Burling, I. R.; Johnson, T. J.; Yokelson, R. J.; de Gouw, J. Measurement
of HONO, HNCO, and other inorganic acids by negative-ion proton-
transfer chemical-ionization mass spectrometry (NI-PT-CIMS):
application to biomass burning emissions. Atmos. Meas. Tech. 2010,
3, 981−990.
(25) Helmig, D.; Thompson, C. R.; Evans, J.; Boylan, P.; Hueber, J.;
Park, J.-H. Highly Elevated Atmospheric Levels of Volatile Organic
Compounds in the Uintah Basin, Utah. Environ. Sci. Technol. 2014, 48,
4707−4715.
(26) This hydrocarbon scrubber is the same as the one used in the
Thermo Electron Corporation’s Model 43C-TLE SO2 instrument.
Documentation can be found at www.thermo.com/eThermo/CMA/
PDFs/Product/productPDF_20982.pdf.

9615 dx.doi.org/10.1021/es501896w | Environ. Sci. Technol. 2014, 48, 9609−9615