Observations of Total Peroxy Nitrates during TOPSE

1
Observations of Total Peroxy Nitrates during TOPSE
J. A. Thornton1, P. J. Wooldridge1, D. A. Day1,
R. S. Rosen1, R. C. Cohen1, F. Flocke2, A.
Weinheimer2, B. A. Ridley2 Email
cohen_at_cchem.berkeley.edu 1 Department of
Chemistry University of California, Berkeley
Berkeley, CA 94720 2 Atmospheric Chemistry
Division, NCAR Boulder CO

• Introduction
• We present in-situ high time resolution
  measurements of NO2 and total peroxynitrate
  mixing ratios made aboard the C-130 aircraft as
  part of the Tropospheric Ozone Production about
  the Spring Equinox (TOPSE) campaign from January
  May 2000. A two channel Thermal Dissociation
  Laser Induced Fluorescence (TD-LIF) instrument
  was used to measure ambient NO2 and total
  peroxynitrate (?PAN ? PAN PPN MPAN HNO4
  N2O5 RC(O)OONO2) concentrations. Air was
  drawn from a single point 18 cm from the belly
  of the aircraft and then split between the two
  channels. One channel of the instrument measured
  ambient NO2 by LIF. The other channel had an
  oven (30 cm long x 6 mm OD pyrex tube) set at a
  temperature of 450 K to effect the quantitative
  conversion of peroxynitrates to NO2 and a
  companion radical. In this channel, the sum of
  the resulting NO2 and ambient NO2 was measured by
  LIF. The difference between the two channels is
  the total peroxynitrate concentration. Total
  peroxynitrate concentrations ranged from near
  zero to greater than 1500 ppt. We discuss the
  instruments performance, an informal
  intercomparison between the TDLIF ?PAN
  measurements and the NCAR GC-ECD measurements of
  PAN and PPN, and the importance of other
  peroxynitrates such as pernitric acid (HNO4) in
  the arctic winter troposphere.

IV. ?PAN Measurements
V. ?PAN and the NOy Budget
II. Thermal-Dissociation Laser-Induced
Fluorescence
The above figures demonstrate the TD-LIF
techniques capabilities. On the left, 12-second
average ?PAN (triangles) and NO2 (open squares)
concentrations measured by the TD-LIF instrument
are plotted versus time for a local research
flight based out of Churchill, Canada April 11,
2000. Arrows indicate when the thermal
dissociation oven was either turned on or off.
When the oven is off, the two channels measure
the same NO2 concentrations, and thus the
difference in signal between the two channels is
zero on average. When the oven is on, the
difference in signal between the two channels is
proportional to the sum total peroxynitrate
concentration which ranged from near zero to 750
ppt during this flight. To the right, we
demonstrate the high time resolution by showing
the the 12-second average data from a narrow time
window (56500 sec 62500 sec) together with the
NCAR (PANPPN) data (solid black squares) which
was measured on a slower time interval by Gas
Chromatography ECD.
XNO2 heat ? X NO2
The two plots above show the ratio of ?NOyi/NOy
versus ambient pressure. On the left, ?NOyi NO
NO2 ?PAN HNO3, where the NO and NO2
measurements were made by NCAR, HNO3 by the UNH
mist chamber instrument, and ?PAN by the TD-LIF
instrument. The ratio of the sum of these NOy
species to the total is scattered about the value
of 1 indicating closure in the NOy budget. On
the right, ?NOyi NO NO2 GC_(PANPPN)
HNO3, where the NCAR GC measurements of PAN and
PPN are used in place of the TD-LIF ?PAN
measurements. The ratio of the sum of these NOy
species to the total exhibits a similar trend
versus pressure as that observed in the
comparison to the TD-LIF instrument. The
apparent calibration offset between the three
instruments (TD-LIF, GC, total NOy) affects our
ability to assess the abundance of other peroxy
nitrates such as HNO4.
The NO2 yield vs. temperature curve for a
combination of different classes of nitrate
compounds can be predicted and measured. Tlt 325
K NO2 T 500 K PAN PPN HNO4 N2O5
RO2NO2 NO2 T 650 K RONO2 RO2NO2 NO2
VI. O3 and ?PAN Correlations
Inlet/Calibration
Inlet/Calibration
PAN source
Zero Gas
PFA Teflon Manifold
Oven
1/4diameter, 25 cm long pyrex tube wrapped in
nichrome
wire for heating.
Sampling
Cooling Section
Cooling Section
Calibrating
Calibrating
3m of tubing
To LIF
To LIF
Zeros
Zeros
Cell 2
Cell 2
Pinholes to reduce pressure
(increased flow rate)
Premixed NO2 Calibration Gas
The two figures above show ?PAN (black squares)
and GC_(PANPPN) (red circles) measurements
plotted versus time for two consecutive flights.
Up to three 12-second ?PAN measurements were
averaged to each GC_(PANPPN) measurement. The
flights are the same as those shown in section
III NO2 Measurements. Flight 40, on the left,
shows periods where the TD-LIF ?PAN measurements
are higher than the GC_(PANPPN) measurements as
well as periods when they are lower. During
Flight 41, on the right, the measurements agree
well during periods when concentrations were
changing rapidly.
To LIF
-
Cell 1

• The TD-LIF instrument fit into a standard
  research aircraft rack (see photo), weighed 200
  kg, and consumed 1.1 kW of aircraft power (4 Amps
  of 110V 400 Hz and 23 Amps of 28V DC).
• The inlet (see schematic) was located
  approximately 3 m from the detection axis at the
  belly of the aircraft. It consists of a small (
  9 cm3) PFA teflon manifold. Air is sampled from
  the manifold through two separate PFA lines by a
  combination of a rotary vane vacuum pump and a
  roots blower.
• One line is directed straight to an LIF detection
  cell (Cell 1) to measure ambient NO2
  (NO2signal_1 NO2ambient), and the other
  passes through a 25 cm long, 6 mm OD pyrex oven
  to affect the dissociation of organic
  peroxynitrates (PAN, PPN, HNO4, etc) and N2O5 to
  a sister radical and NO2. After the oven, the gas
  is directed to a second LIF detection cell (Cell
  2) to measure ambient NO2 plus NO2 from the
  dissociation (NO2signal_2 NO2ambient
  NO2ROONO2Heat).
• Calibrations to NO2 and zero artifact tests were
  done routinely during flights by flooding the
  inlet manifold with a gas mixture or zero air.
  Tests of the dissociation oven were done by
  adding a PAN standard mixture to the inlet
  manifold between missions.

• The TOPSE experiment was designed to improve our
  knowledge of the mechanisms responsible for the
  springtime increase in tropospheric O3 over the
  Arctic. The figure above, left, shows O3 versus
  the LIF ?PAN concentration binned by altitude for
  data collected north of 50?. The importance of
  separating the role of mixing from chemistry is
  apparent in this figure
• Above 6km (blue circles), we observe recent
  mixing of stratospheric (high ozone, low ?PAN)
  and tropospheric air (low ozone, high ?PAN)
• Between 2 and 5 km (black triangles), we observe
  layers with 1600ppt ?PAN and 75 ppb O3 indicating
  mixing of mid latitude air.
• Below 2km (red squares), ?PAN is mostly below
  400ppt, and there is little evidence for mixing.

In the figure above, right, we plot O3 and ?PAN
mixing ratios averaged over an entire flight
versus day of year. When TD-LIF ?PAN
measurements were not available we used the NCAR
GC_(PANPPN) measurements in the average. The
correlated increase of both O3 and ?PAN in the
Arctic provide a strong constraint for models
aimed at describing the roles of both transport
and chemistry. While the increase in
tropospheric O3 as observed in the figure could
be explained solely from an increase in mixing of
stratospheric air masses, the correlated increase
in ?PAN mixing ratios (R2 0.63) can not be
since ?PAN mixing ratios are essentially zero in
the stratosphere.
III. NO2 Measurements
VII.Conclusions
The results of a point-by-point comparison of all
TD-LIF ?PAN and NCAR GC_(PANPPN) measurements
are shown above. 12-second average TD-LIF data
that fell within 36 seconds of a GC_(PANPPN)
measurement were averaged for the comparison. A
linear fit to the comparison data suggests the
two data sets agree to with 4 on average with a
1.6 ppt offset and good correlation (R2 0.77).
However, at the lower concentrations the TD-LIF
measurements are biased low relative to the GC
measurements. This bias is also evident in the
plot to the left. The ratio of LIF_
?PAN/GC_(PANPPN) is plotted versus ambient
pressure (mbar). At the lowest pressures
(highest altitudes), the TD-LIF instrument
measures slightly higher (10-15) total
peroxynitrates than the GC_(PANPPN) measurements
on average, and at the highest pressures (lowest
altitudes) the TD-LIF measurements are 30 less
than the GC measurements. Because ?PAN
concentrations are correlated with ambient
pressure (altitude), it is difficult to determine
the cause of this trend. The trend maybe due to
a relative calibration offset between the two
instruments and to the presence of peroxynitrates
other than PAN and PPN such as HNO4.
The TD-LIF measurements of ?PAN are in good
general agreement with the GC_(PANPPN)
measurements. These two data sets provide
evidence that the major NOy species in the Arctic
troposphere is PAN. Calibration offsets between
the two data sets (10-30) make estimates of the
abundance of other peroxynitrates difficult.
However, future analysis will be aimed at
constraining the abundance of HNO4 in the Arctic
troposphere using the TD-LIF ?PAN
measurements. The altitude dependence to the
correlation of ?PAN with O3 demonstrates the
importance of separating the roles of mixing and
chemistry. The observed seasonal increase of
both ?PAN and O3 concentrations provides a strong
constraint for models aimed at describing the
roles of both transport and chemistry.
Examples of the LIF NO2 measurements (red
circles) made on the C-130 are shown above for
two consecutive flights along with the NO2
measurements made by NCAR using Photo-Fragment
Chemiluminescence. The measurements shown here
are 1-minute averages. The LIF and NCAR NO2 are
in reasonable agreement above 10 ppt, and the
comparison is noisy below that. Both channels in
the TD-LIF instrument measured the same NO2
concentrations.