Heard D.E. (editor) Analytical Techniques for Atmospheric Measurement

Подождите немного. Документ загружается.

202 Analytical Techniques for Atmospheric Measurement

with the instrument walls. The air flow was slowed from the aircraft velocity of 200 to

100 m/s in a large bypass duct. A sub-sample of air in this duct entered the detection

duct where flow speeds were regulated between 10 and 100 m/s. Observations showed

that turbulence in the flow at speeds below 40 m/s resulted in loss of OH. The gas passed

by three detection cells used for measurement of OH, HO

, and H

O (Section 4.3.4).

NO was injected after the first cell to allow for detection of HO

in the second cell. An

NO injector was also positioned before the first cell as a diagnostic for the HO

-to-OH

conversion efficiency.

The aircraft platform flew dozens of times in the first deployment of this HO

instrument and hundreds since. Diagnostics showed [OH] was linear in laser power,

confirming non-quadratic behaviour, and showing that retrieved [OH] was independent

of laser pulse energy. Also [OH] was observed to be zero at night, further confirming the

absence of laser-generated OH. Finally, titration of ambient OH with perfluoropropene

C

 was also used to confirm the absence of laser-generated OH. Perfluoropropene

is the compound of choice for chemical modulation as it contains no hydrogen atoms

that could possibly re-form OH while titrating ambient OH (Dubey et al., 1996). Intro-

duction of perfluoropropene to the inlet only removes ambient OH. Laser-generated OH

molecules are detected in nanoseconds after their creation by the same laser pulse, and

their reaction with perfluoropropene is too slow for removal prior to detection.

The background of OH LIF instruments is routinely measured by tuning the laser on

and off the chosen rovibronic absorption feature in the OH absorption spectrum. An

OH ‘reference’ chamber, in which high concentrations of OH are produced, is used to

lock the laser frequency. Ideally all components of the background are invariant over

this small shift in excitation wavelength. Most OH LIF instruments continuously employ

spectral modulation during operation and occasionally use chemical modulation to test

for interferences.

In principle, the calibration of the balloon and aircraft-borne instruments could be

calculated from detailed knowledge of the transmission of the optical system and the

detailed temperature-dependent state-to-state quenching parameters (Smith & Crosley,

1990; Wennberg et al., 1994a). In practice, the optical transmission is not known well

enough and a direct calibration is necessary. Such calculations do, however, provide a

useful check on the signal and magnitude of the observed pressure and temperature

dependence. The balloon-borne and aircraft-borne instruments were calibrated using a

known quantity of OH that was produced by reacting an NO

reference standard with

excess H atoms:

H +NO

→ OH +NO (4.15)

These calibrations were performed over a range of temperatures and pressures and using

and N

/Ar mixtures (Anderson, 1976; Stimpfle and Anderson, 1988; Stimpfle

et al., 1990; Wennberg et al., 1994a).

For the ER-2 instrument, the calibration was checked against absorption at very high

mixing ratios of OH to test the assumption that the OH delivered to the detection region

could be calculated from the concentration of the NO

standard. These two methods

agreed to within 10%. The ER-2 instrument utilized the N

Stokes Raman scattering at

302 nm as its transfer standard. This was more reliable than the Rayleigh transfer standard

Fluorescence Methods 203

used in the balloon instrument because the laser scatter from the detection chamber is

much lower at red-shifted wavelengths than at the Rayleigh lines.

The uncertainty in the ER-2 OH measurements (Perkins et al., 2001) was ±13%1,

which is a combination of the uncertainty of the laboratory calibration and the uncertainty

in the inference of the sensitivity from in-flight diagnostics. The precision was 0.03 ppt in a

two-second integration. For HO

, additional uncertainty in the chemical conversion leads

to a slightly greater uncertainty of ±15%, with a precision of 0.15 ppt in a two-second

integration. Typical detection limits were 0.05 ppt.

4.3.2 Instruments for lower tropospheric OH measurements

The central role of OH in tropospheric chemistry was recognized in the late 1960s and

early 1970s (Weinstock, 1969). In the troposphere, OH removes VOCs such as methane

and isoprene by initiating oxidation. In the presence of NO

and sunlight, the oxidation

results in catalytic production of ozone. Secondary aerosols may also be produced. In the

lower troposphere, ozone is an irritant and the principal component of photochemical

smog. In the upper troposphere, ozone is a potent greenhouse gas.

The high water vapour concentrations in the troposphere have required instruments

developed for tropospheric OH measurement to use resonant excitation and detection

at 308 nm. This is superior to excitation at 282 nm because the product of the ozone

absorption cross section and the photolysis quantum yield of O

D is reduced by a factor

of ∼30 at 308 nm relative to 282 nm. Early instruments for tropospheric OH detection

(Davis et al., 1976; Wang and Davis, 1974; Wang et al., 1975, 1976) failed due to laser-

generated OH. Hard and co-workers (1984) reduced laser-generated OH using FAGE,

as described in Section 2.2.5, and their early success combining FAGE with excitation at

308 nm using high repetition–rate lasers has been adopted as a model by most research

groups using LIF to observe tropospheric OH (Creasey et al., 1997; Faloona et al., 2004;

Hard et al., 1995; Holland et al., 1995; Kanaya et al., 2001).

All of the current generation instruments operate at low pressure, pump optically

with 1–20 mW of 1–10 kHz tunable light at 308 nm, and detect resonance fluorescence

at 308 nm. Most use similar methods for calibration and background measurement and

have similar inlet and optical cell designs. Some of the reviews that have described these

instruments and scientific results in detail are Crosley, 1995, and Heard and Pilling,

2003.

Inlets consist of a critical orifice (∼1 mm diameter) centred on either a flat plate or a

cone. Inlet losses are typically minimized by use of unreactive coatings such as halocarbon

wax and by judicious design of the inlet (Faloona et al., 2004; Stevens et al., 1994). NO

is injected immediately downstream of the orifice, with the reaction time determined by

the volume of the inlet and by the instrument’s pumping speed. Most OH instruments

utilize a single-pass optical cell and detect fluorescence with a PMT. The Pennsylvania

State University design (Faloona et al., 2004) is the exception and uses a multi-pass White

Cell with a microchannel plate detector. This instrument has the best sensitivity reported

to date with a detection limit of 14 ×10

molecules/cm

in a 30-second integration, SNR

of 2, and an uncertainty of ±16% 1 for both OH and HO

. Most other instruments

have reported detection limits in the range of 28 ×10

to 10

molecules/cm

in the same

averaging times and SNR values (Heard & Pilling, 2003).

204 Analytical Techniques for Atmospheric Measurement

Tropospheric OH instruments are usually calibrated by photolysis of water in the

presence of O

to produce OH and HO

in a 1:1 ratio. The 184.9 nm line of a mercury

lamp is used to photolyze water:

O +h

=1849nm

−−−−−→

H +OH (4.16)

H +O

+M →HO

+M (4.17)

The OH production rate is given by

dOH

= H

O



F (4.18)

where 

is the absorption cross section of water at 184.9 nm, 

is the photolysis

quantum yield of water to form OH, and F is the photon flux of the mercury lamp.

Integration yields

OH

= H

O



Ft (4.19)

The absolute water vapour concentration is measured by use of chilled hygrometers or

by IR spectroscopy, F is measured with an absolutely calibrated photomultiplier tube,

and the residence time is determined using calibrated flow measurements. Alternately,

the product Ft, which is difficult to determine accurately, can be measured by an ‘O

chemical actinometer’ (Bloss et al., 2004; Holland et al., 1995; Kanaya et al., 2001; Schultz

et al., 1995). The same mercury lamp photolyzes oxygen:

1849nm

−−−−→

O +O (4.20)

2O +O

+M →O

+M (4.21)

creating ozone, which is measured with a commercial ozone analyzer and compared to

the calculated O

 produced:

O

 = O





Ft (4.22)

or Ft =

O



O





where 

is the absorption cross section of molecular oxygen and  is the quantum

yield for ozone production. Actinometry can also be based on photolysis of N

O instead

of ozone (Faloona et al., 2004). The O(

D) photolysis product reacts with N

O to form

2NO molecules, which can be detected by chemiluminescence (see Chapter 7).

Alternate calibration methods include monitoring the decay of a hydrocarbon by gas

chromatography in an irradiated, continuously stirred reactor in which OH is produced

by photolysis of water vapour (Hard et al., 1984), production of a steady state OH

concentration by alkene ozonolysis (Hard et al., 2002), and absorption of super-ambient

OH concentrations (Kanaya et al., 2001). Typically, these calibration strategies are used

Fluorescence Methods 205

at a single value (often close to zero) of the water vapour concentration. Ambient

measurements require accounting for water vapour quenching in the sample because it

quenches OH 22 times faster than nitrogen (Bailey et al., 1999). Additionally, losses of

radicals due to clustering, which is very dependent upon water vapour concentation, have

been observed in some instruments (Heard & Pilling, 2003), and thus calibrations are

ideally performed at ambient water vapour concentrations.

is detected after chemical conversion to OH by reaction with NO. In tropospheric

instruments, NO is added in the low pressure region downstream of the pressure-

reducing orifice. HO

-to-OH conversion efficiencies range from 30 to 95% (Creasey et al.,

2003; Kanaya et al., 2001; Stevens et al., 1994). In addition to the formation of HONO

(Reaction 4.14), a further complication encountered in the troposphere is conversion of

organic peroxy radicals RO

 to OH by the reaction scheme

+NO →RO +NO

(4.23)

RO +O

→ R



CHO +HO

(4.24)

+NO →OH +NO

(4.13)

where R



CHO represents an aldehyde or ketone. Conversion rates of 5% for CH

and

have been measured (Kanaya et al., 2001; Holland et al., 2003). Ren et al. (2004)

report negligible interference from up to 100 ppt of methane-, ethane-, propane-, and

n-butane-derived RO

species.

The accuracy of LIF HO

measurements has also been assessed by comparison

to measurements using different techniques and therefore presumably subject to

different interferences, sampling artefacts, and calibration errors. Intercomparisons of

OH measurements have been reviewed in detail by Heard and Pilling (2003). For

example, OH measurements by LIF and by folded-path differential optical absorption

spectrometry (DOAS) were obtained in rural north-eastern Germany in an exper-

iment during the summer of 1994. A linear fit to the data gives DOAS measurement

of 109 ×LIF +028 ×10

 molecules cm

−3

with an R

correlation coefficient of 0.8

(Hofzumahaus et al., 1998). Most of the scatter is accounted for by the imprecision of

the measurements. A comparison study of Penn State’s GTHOS (Ground-based Tropo-

spheric Hydrogen Oxides Sensor) LIF detector with NCAR’s perCIMS (peroxy radical

chemical ionization mass spectrometer) detector in HO

mode was conducted at a rural

site in Pennsylvania, USA, in 2002 (Ren et al., 2003a). Measurements of ambient HO

and both instruments’ calibration sources were compared. The comparison showed the

calibration sources were identical to within 2%. The accuracy of perCIMS in HO

mode

was estimated to be 41% at the 95% confidence interval, with 15% of the measured

in perCIMS stemming from CH

. The accuracy of GTHOS was estimated to

be 32% at the 95% confidence level. Figure 4.5 shows a time series of ambient HO

measurements. A linear regression of the comparison data, with an integration period

of 140 seconds, was described by PerCIMS HO

= 096×GTHOS +06 ppt, with an R

value of 0.85.

206 Analytical Techniques for Atmospheric Measurement

June 5 June 6 June 7

Date of 2002 (EDT)

(pptv)

June 8

GTHOS HO

PerCIMS HO

Figure 4.5 Time series of ambient HO

measurements by LIF (GTHOS, solid line) and perCIMS (circles).

(From Ren et al., 2003a.)

4.3.3 OH ﬂuorescence as an indirect detection strategy

In addition to HO

, other species can be measured indirectly by OH fluorescence. Water is

measured by the stratospheric hygrometer described by Weinstock et al. (1994). Photolysis

of water vapour by Lyman  radiation at 121.6 nm produces electronically excited OH:

O +h1216nm →OHA



 +H (4.25)

A fraction of the OH produced is highly rotationally excited in the 



= 1 level of the

A



 electronic state. Following vibrational relaxation, emission from highly excited

rotational levels of the







= 0 state is detected at 305–325 nm in an optical cell

similar to that of the stratospheric HO

instrument aboard the ER-2.

Ambient detection of nitrous acid (HONO) by photo-fragmentation–LIF should also be

feasible and prototypes have been described (Liao et al., 2003; Rodgers and Davis, 1989).

4.3.4 Naphthalene

Martinez et al. (2004) discovered that naphthalene could be measured by their HO

detector thanks to a fortuitous overlap between the absorption spectra of naphthalene

and OH. Naphthalene has a sharp absorption peak at 308.0026 nm, slightly to the red of

the Q

2 rotational line of OH at 307.9951 nm. The naphthalene absorption is negligible

to the blue of the Q

2 line which was used as the off-resonance wavelength for spectral

modulation of OH. Since the Penn State group alternated their offline measurement to

high and low frequencies of the OH peak, they were able to produce accurate OH and

naphthalene observations after the interference was identified as a useful signal.

Fluorescence Methods 207

4.3.5 Measurement of the OH lifetime

Most of the techniques described in this chapter are used to measure the concentration

of a particular atmospheric species. However, the concentration is just one of many

parameters that are useful for testing our understanding of atmospheric chemistry. A few

LIF instruments have been constructed which measure the total reactivity of OH k

total



or equivalently its lifetime 

 defined as:

total

= 

−1



= k

CO +k

NO

 +k

CH

 +··· (4.26)

where k

is the product of an individual reactant concentration and its second order

rate constant with OH. The advantage of measuring the total loss rate is that it is not

possible to measure every compound with which OH reacts; indeed the identity of every

compound that reacts with OH is not known. The two methods of measuring the total

reactivity of OH combine LIF detection of OH with standard techniques borrowed from

laboratory kinetics.

Kovacs and Brune (2001) demonstrated measurement of the OH lifetime in 2001. Their

method, TOHLM (Total OH Loss Measurement), also used by Heard and co-workers at

Leeds (Johnson, 2004), is based on the decay of OH concentrations when mixed with

ambient air in a flow tube as depicted in Figure 4.6. Approximately 500 ppt of OH and

in approximately a 1:1 ratio, produced by photolysis of water vapour by a mercury

lamp, is released into a 5 cm diameter, glass flow tube by a moveable injector. Ambient

air flows into the flow tube and the relative OH concentration is measured downstream

by LIF. As the total flow rate and reaction distance are known, the reaction time between

the injected OH and the ambient air is known for any injector tube position. Plotting OH

concentration versus reaction time yields decay profiles that are fit to single exponentials

to extract the first order rate constant. Injection of a known and high concentration of

a compound with a well-quantified reaction rate with OH, such as carbon monoxide, is

used to ‘calibrate’ the flow conditions, and injection of zero air is used to quantify wall

losses of OH which must be subtracted from the total measured loss. The measurement is

simplified by the fact that the OH concentration does not have to be quantified absolutely.

Rather, only the relative decay of the OH concentration over time is measured.

At a few sites, the total OH reactivity measured by TOHLM was compared with the

total OH reactivity calculated by summing the product of the individually measured OH

sinks (e.g. CO, CH

,NO

, VOCs) with their respective reaction rate coefficients with

OH (i.e. by explicit evaluation of the right hand side of Equation 4.26). In Nashville

during the SOS ’99 study, observed OH reactivities were on average 40% higher than the

calculated reactivities (Kovacs et al., 2003), whereas during the summer of 2001 in the

PMTACS-NY field study in New York City, the calculated and observed OH reactivities

agreed to within 10% (Ren et al., 2003b). In a forest in northern Michigan, the difference

between the observed and the calculated reactivities values was found to be exponentially

dependent on temperature and to closely resemble the temperature profile of terpene

emissions, suggesting the existence of significant concentrations of unmeasured biogenic

VOCs (Di Carlo et al., 2004).

In an alternative strategy described by Calpini and co-workers (Calpini et al., 1999;

Jeanneret et al., 2001) and successfully implemented by Sadanaga and co-workers (2004b),

208 Analytical Techniques for Atmospheric Measurement

FLOW TUBE

and H

Bubbler

Diffuser

MERCURY LAMP

SHIELD

Detection

cell

MCP

Laser

Blower

Hot wire

anemometer

FLOW

Ambient

air in

Calibration

gas in

Vacuum pump

Addition for

detection

Detection

inlet

Figure 4.6 The OH lifetime instrument of Kovacs and Brune. Ambient air enters the ﬂow tube from the

left and mixes with high concentrations of OH released at a series of distances from the LIF inlet. (From

Kovacs and Brune, 2001, with kind permission of Kluwer Academic Publishers.)

OH is produced by flash photolysis of ambient ozone using the fourth harmonic of an

Nd:YAG laser at 266 nm (80 mJ/pulse, 10 Hz) followed by reaction of the O(

D) photolysis

product with water vapour. The subsequent decay of the high OH concentration produced

(up to 10

molecules/cm

) is measured by LIF in real time, thus indicating the OH

reactivity.

4.4 The nitrogen oxides

The nitrogen oxide radicals NO

≡ NO +NO

 are central to the production of ozone

in the troposphere and the destruction of ozone in the stratosphere. Accurate and precise

measurements of the atmospheric abundance of NO

and their oxidation products –

peroxy nitrates, alkyl nitrates and nitric acid – are thus important to understanding and

modeling air quality (Liang et al., 1998; Russell & Dennis, 2000; Thornton et al., 2002),

acid rain and nitrogen deposition (Galloway, 2001; Munger et al., 1998), greenhouse

forcing (Mickley et al., 1999), the carbon cycle (Lerdau et al., 2000; Holland et al.,

1997), and stratospheric ozone (Wennberg et al., 1994b). Measurements of NO

are

Fluorescence Methods 209

also important to combustion science (Docquier & Candel, 2002) and are used in some

strategies for detection of explosives (Steinfeld & Wormhoudt, 1998). Detection of NO

is difficult because high concentrations of other NO

species represent interferences for

many techniques. Additionally, concentrations vary by 1–2 orders of magnitude at a

single location and by several orders of magnitude between the combustion sources and

the remote atmosphere, where mixing ratios are as small as 1–2 ppt. Till the time of

writing, no technique has become the singular method of choice. Instruments based

on IR absorption (Chapter 2), visible absorption (Chapter 3), LIF (this chapter) and

chemiluminescence (Chapter 7) are all in active use.

Although NO is routinely measured accurately by chemiluminescence in scientific and

a wide variety of other settings, high quality measurements of NO

by chemiluminescence

are relatively rare and have been limited to scientific instruments. The most sensitive,

in situ instruments for observing NO

are photofragmentation – chemiluminescence

and direct LIF. The latter is emerging as competitive with the former in both cost and

sensitivity. In many monitoring networks, surface catalysis is used to convert NO

to NO

prior to chemiluminescence detection. These measurements virtually always overestimate

because of the conversion of nitric acid, alkyl nitrates and peroxy-nitrates into NO

As a result, the data from such networks is ignored by the scientific community, although

it serves some useful purposes for development of public policy.

4.4.1 Photophysics of NO

Figure 4.7 depicts the absorption spectrum of NO

. It is highly congested due to the

presence of numerous overlapping ro-vibronic transitions involving the X

and the C

electronic states and as a result of large differences in the geometries

of the ground and the excited states. The rotational structure is embedded in a broad

continuum, which peaks at 413 nm. Photodissociation to NO +O

P with near-unity

quantum yield occurs for excitation wavelengths shorter than 398 nm. At short times

Figure 4.7 NO

absorption spectrum at 293 K. (From Burrows et al., 1998.)

210 Analytical Techniques for Atmospheric Measurement

<1 s following excitation, the emission spectrum of NO

exhibits vibrational structure

at pressures below ∼10 m torr (Donnelly et al., 1979). At longer times or higher pressures,

the emission spectrum resembles a broad peak centred in the near-IR due to efficient

stepwise vibrational relaxation of the excited electronic state (Clough & Thrush, 1967,

Mazely et al., 1994, Patten et al., 1990). The excited state lifetime of NO

 ∼100 sis

much longer than expected based on its integrated absorption cross section and is a

strong function of excitation wavelength, pressure, and spectral region observed during

measurement (Donnelly & Kaufman, 1978, Patten et al., 1990, Sackett & Yardley, 1972).

This long radiative lifetime is a result of strong interaction between the excited electronic

states and upper vibrational levels of the ground electronic state as first described by

Douglas (1966). Quenching of electronically excited NO

is rapid; the air-average value

with 532 nm excitation is 6 ×10

−11

molecule

−1

(Donnelly et al., 1979).

4.4.2 Single-photon NO

instrument design

In contrast to the design of tropospheric OH LIF instruments, which have for the most

part converged, the design and operation of NO

LIF instruments are widely varied

and evolving rapidly. There is no standard excitation wavelength or optical cell design.

Most instruments operate at reduced pressure and use pulsed lasers with time-gated

fluorescence detection. Early attempts at detection of tropospheric NO

via LIF were

conducted at atmospheric pressure but suffered from an aerosol interference (Tucker

et al., 1975). Interest was not renewed until work by Fong and Brune (1997), Perkins

et al. (2001), and Thornton et al. (2000) in the late 1990s.

The choice of excitation wavelength is a trade-off between specificity and sensitivity.

Specificity is maximized in the red end of the spectrum because of the high ratio of

on-resonance to off-resonance absorption, while sensitivity is maximized in the blue

end where the absolute absorption cross section is highest. Due to the greater ease in

producing tunable light and the specificity afforded at longer wavelengths, excitation at

all wavelengths is competitive. Current NO

LIF instruments use excitation wavelengths

ranging from the blue to the red region of the spectrum. Emission from excited NO

peaks in the near-IR, dictating the type of PMT to use. Multi-alkali or cooled GaAs

photocathodes, with quantum efficiencies of ∼8–15% in the range of 800–900 nm are

typically used.

Most NO

LIF instruments use a dual-wavelength technique to demonstrate specificity

to NO

. This spectral modulation is similar to the background measurement technique

used by OH LIF instruments except that for NO

there is still appreciable fluorescence

at the off-resonance frequency due to the continuum in the absorption spectrum. This

technique has been applied in both the stratosphere (Perkins et al., 2001) and the

troposphere (Cleary et al., 2002; Fong & Brune, 1997; Matsumoto and Kajii, 2003;

Matsumi et al., 2001; Matsumoto et al., 2001; Thornton et al., 2000). In the instrument

described in detail by Thornton et al. (2000), the output from a narrow linewidth dye laser

pumped by a 3 W, 8 kHz Nd:YAG laser excites a pair of overlapping rotational lines in the

(430) A

←000X

vibronic band of NO

at 170865cm

−1

(585.257 nm). The laser

frequency is repeatedly tuned on- and off-resonance with this peak, spectrally modulating

Fluorescence Methods 211

the fluorescence signal. The NO

concentration is proportional to the difference between

the ‘online’ and ‘offline’ signal.

The detection limit of the instrument described by Thornton et al. is 6 ppt for a

1-minute integration, with SNR = 2. The precision is shot-noise limited and at 1.5 ppb,

equals 16% 1 in 1 second. If combined with the supersonic jet technology

described by Cleary et al. (Section 4.4.3), an enhancement of ∼20 in the signal rate could

be expected.

Figure 4.8 depicts NO

spectral modulation. The bottom trace is the fluorescence

excitation spectrum of ambient NO

. The top trace is the transmission through a 2.5 cm

reference cell containing 20 torr of pure NO

. After scanning the laser frequency in order

to locate the position of the online and offline peaks, the laser frequency alternates

between the two indicated frequencies. Use of spectral modulation conveys a spectroscopic

fingerprint of NO

in the fluorescence signal itself, greatly increasing the specificity of

the instrument. Any potential interference would only increase both the offline and the

online signals but would not affect the difference between the two, as it is unlikely that

any interference would respond to such a small change in excitation frequency. This

leaves ‘false NO

’ as the most likely potential interference, for example that produced

by the reaction of NO and O

or thermal decomposition of HO

or N

in the

sampling lines. Such interferences can be quantified by calculation and by laboratory

experiment, and are usually insignificant.

Dye lasers (Fong & Brune, 1997; Perkins et al., 2001; Thornton et al., 2000), tunable

diode lasers (Cleary et al., 2002), and optical parametric oscillators (Matsumi et al.,

2001) have all been successfully used for NO

detection. The linewidth of dye lasers

450

400

350

300

250

200

150

100

2100 2200

Time (seconds)

Online

Offline

Signal (cps)

Transmittance

2400 2450

Figure 4.8 Spectral modulation of NO

. From time 2050 seconds to 2300 seconds, the laser frequency is

scanned. The bottom trace is ﬂuorescence from ambient NO

and the top trace is the transmission through

a 2.5 cm reference cell containing 20 torr NO

. At time 2400 seconds, the laser frequency alternates

between the online and ofﬂine positions. The increase in ﬂuorescence at the right of the ﬁgure is from

an increase in the ambient NO

mixing ratio.