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

Подождите немного. Документ загружается.

428 Analytical Techniques for Atmospheric Measurement

When driven by solar radiation, this mechanism produces O

P atoms about five to ten

times faster than O

D (Table 9.1). In the troposphere, however, only the formation

of O

D is relevant. The reaction of O

D with water vapour produces the important

atmospheric OH radicals (see Section 9.1.1), whereas tropospheric O

P atoms are

converted back completely to ozone via reaction R 9.2, resulting in a do-nothing cycle.

The photolysis frequency for O

→ O

D is generally called jO

D. It has been

measured in various tropospheric field experiments by chemical actinometers, which

are listed in Table 9.3. The measurement of jO

D by chemical actinometry involves

the difficult task of detecting extremely reactive O

D atoms. A suitable scheme is the

scavenging of O

D by N

O and measurement of the resulting reaction products (Bahe

and Schurath, 1978; Dickerson et al., 1979):

O

D +N

O −→ N

(R 9.11a)

−→ NO+NO (R 9.11b)

The resulting NO, however, is not stable and is converted into higher oxides by reactions

with excess ozone:

NO +O

−→ NO

(R 9.12)

−→ NO

(R 9.13)

+NO

+M  N

+M (R 9.14)

The O

D production rate can be quantified from measurements either of N

or of the

nitrogen oxides, NO

=NO, NO

,NO

(Table 9.3). In any case, the branching ratio

must be accurately known, with k

and k

being the bimolecular rate constants

for reactions R 9.11a and R 9.11b, respectively. Furthermore, if NO

is measured, its

chemical yield  per initially formed NO molecule must be known. With this information

the photolysis frequency can be determined as

jO

D =



1 +



O



N



t

(9.25)

jO

D =



1 +





O



NO



t

(9.26)

Published measurements of k

/k

 have been critically evaluated by Cantrell et al.

(1994), who recommend a value of 061 ±008 (95% confidence interval). This value

was adopted by IUPAC (Atkinson et al., 2004) and agrees within the uncertainty also

with the current recommendation of 0.58 by NASA-JPL (Sander et al., 2003). The value

by Cantrell et al. (1994) corresponds to k

= 064, which may be compared with the

data used by different research groups for their chemical actinometers (Table 9.3). The

error of the ratio is usually the dominating uncertainty in the measurement of j(O

D) by

chemical actinometry, which has a total uncertainty of about ±(10–15)%.

Measurement of Photolysis Frequencies in the Atmosphere 429

When jO

D is determined from NO

measurements, an additional error arises from

the determination of , which has to be calibrated individually for each chemical

actinometer. The calibration can be performed by adding a known amount of NO to the

actinometer gas flow at the entrance of the photolysis reactor and measurement of the

corresponding increase of the NO

signal.

The photolysis of ozone to O

D atoms is a process with a pronounced temperature

dependence. In order to measure representative jO

D values, it is therefore necessary

to operate the chemical actinometer at ambient temperature. If the actinometer gas is

not in thermal equilibrium with the surrounding atmosphere, temperature corrections

have to be applied to the jO

D data. See Section 9.5.4 for a discussion of the expected

temperature dependence.

Static ozone actinometers use the measurement of N

as a proxy for O

D. This concept

has the advantage that nitrogen is a stable compound which undergoes no further

chemical reactions. A disadvantage is the limitation of the N

detection by spurious

background signals which are caused by possible air leaks in the instrumental setup,

or by N

contaminations in the bottled gases from which the actinometer mixture is

prepared. Typical background levels of 1–3 ppmv N

have been reported for the chemical

actinometers listed in Table 9.3. In order to achieve sufficiently high N

concentra-

tions that exceed the background significantly by photolysis, relatively high initial ozone

concentrations and long irradiance times have to be applied. For example, given a

photolysis frequency of jO

D = 3×10

−5

−1

(a typical clear-sky value at overhead sun

and 300 DU of total ozone) and an exposure time of 30 minutes, 2 hPa O

in an excess

of N

O will produce approximately 40 ppmv N

at 1 atm of total pressure.

An example for a static actinometer is shown in Figure 9.10 (Bahe & Schurath, 1978;

Bahe et al., 1979). A quartz bulb (∼7cm dia.) was filled with a mixture of about 13 hPa

and 1.5–3 atm N

O. After exposure to solar radiation, gas samples were drawn from

Quartz actinometer

(pressurized)

He carrier gas

Needle

valve

Sampling

loop

He purge

gas

Flow meter

GC column

GC detector

Figure 9.10 Schematic of a chemical actinometer system for measurement of the photolysis frequency

of O

. In the quartz reactor a gas mixture of N

O and O

is exposed to solar radiation. After exposure the

amount of photolytically formed N

is analysed by gas chromatography (GC) (from Bahe and Schurath,

1978, used with permission of Birkhäuser/Springer).

430 Analytical Techniques for Atmospheric Measurement

the reactor for N

analysis with a gas chromatographic system. The data were evaluated

following Equation 9.25, but corrections ∼50% had to be applied since the actinometer

was not optically thin and the ozone concentration was not stable. About 50% O

loss

within 1 hour of exposure was caused by photolysis and O

consumption by the reaction

with O

P atoms:

O

P +O

−→ O

(R 9.15)

The concept by Bahe et al. (1979) was improved by reducing the initial O

concentrations

∼2 hPa to ensure optically thin conditions (Junkermann et al., 1989; Müller et al., 1995).

Furthermore, the O

losses during sun exposure were reduced by addition of 100 hPa

(Müller et al., 1995). While the O

removes a few per cent of O

D by collisional

deactivation, it has the advantage to recycle all O

P atoms back to O

(via R 9.2),

reducing the effective loss of O

to about 10% per hour. This technique has been applied

successfully for calibration of jO

D filter radiometers (see Section 9.5.4), but owing to

its poor time resolution it is less useful for extensive atmospheric field measurements.

Flowing ozone actinometers use online methods with high sensitivity for NO

detection,

which has several advantages. First, the high sensitivity allows to measure much

smaller product yields than has been possible with N

detection, reducing the necessary

measurement time for jO

D to less than a minute (Table 9.3). Second, the online mode

enables continuous measurements, which makes them ideal for atmospheric field exper-

iments over extended time periods. Third, small flow tube diameters (typically 1–2 cm)

and relative low ozone concentrations provide optically thin conditions which require no

corrections in the derived j-values.

An example of a chemical actinometer developed for airborne operation is shown

schematically in Figure 9.11 (Dickerson et al., 1982). Two flow reactors, one for O

photolysis and one for NO

photolysis, are mounted atop the aircraft with free view of

the sky. The reactors are supplied with gas mixtures prepared online from stored gases

(NO, N

O, O

). For the determination of jO

D, a flow of oxygen is ozonized and mixed

with a larger flow of N

O, providing about 2000 ppmv of ozone, which is passed through

the photolysis reactor. The resulting NO

is reduced catalytically to NO by passing the

gas over heated Palladium at 650



C, which also destroys the remaining ozone. The NO is

then measured by a chemiluminescence detector. Care had to be exercised in the choice

of the catalyst to avoid conversion of N

O into NO (Dickerson et al., 1979). The efficiency

of the NO retrieval from NO

is calibrated by injection of a known flow of a standard

NO gas mixture upstream of the reactor inlet.

Following the concept by Dickerson et al. (1982), other O

actinometers were developed,

but with major modifications in the NO

detection. Blackburn et al. (1992) passed the

gas mixture after its exposure through liquid methanol, where NO

and N

react with

traces of water to form nitric acid. The resulting change in electrical conductivity was

calibrated and used to quantify the NO formation in the actinometer. Bairai and Stedman

(1992) operated their chemical actinometer with much less ozone ∼5 ppmv, resulting

in conversion of NO into NO

without significant production of higher oxides. The NO

was then measured by detection of the chemiluminescence produced during the reaction

of NO

with a luminol solution. In the actinometer developed by Shetter et al. (1996),

the NO is converted mostly into N

, which is then transferred through a permeable

Measurement of Photolysis Frequencies in the Atmosphere 431

FM FM FM

FC FC FC

RXN

chamber

Ozonizer

Quartz flow reactors (2x)

→ O(

→ NO

Pressure

Bulkhead

PMT

conv.

650°

conv.

Vacuum

pump

Ozone

detector

tank

killer

tank

50 ppm

tank

Analyzer

permeation

oven

Solenoid

Manual shutoff

3-Way solenoid

Capillary

Frit

Figure 9.11 Schematic of an airborne chemical actinometer for measurement of photolysis frequencies.

Two photolysis ﬂow reactors, one for O

and one for NO

, are mounted atop the aircraft with free view

of the sky. NO is photolytically formed in both actinometers and is measured by an NO

analyser that

takes alternatingly samples from one or the other reactor. Abbreviations used: FC, ﬂow controller; FM,

ﬂow meter; Conv., conversion of nitrogen oxides into NO; RXN, reaction chamber; PMT, photomultiplier

tube (adapted from Dickerson et al., 1982, printed with permission from American Geophysical Union).

Nafion® membrane into liquid water. The increase in electrical water conductivity due

to formation of nitrate ions is then measured to quantify the initially formed NO.

Measurement examples are shown in Figures 9.12 and 9.13. The first figure displays a

typical diurnal profile, which was measured by a chemical flow actinometer during a field

experiment on Mauna Loa in 1992 (Shetter et al., 1996). Smooth jO

D variations under

almost cloudless sky can be seen in the morning and late afternoon, demonstrating a high

measurement precision. The fluctuations occuring in the early afternoon were caused by

clouds that influenced the solar actinic flux. The general shape of the diurnal profile is

determined mainly by the jO

D dependence on the slant ozone column t

/ cos ,

which effectively absorbs solar radiation at wavelengths below 320 nm (see Section 9.1.6

and Figure 9.3). Examples of the functional dependence of jO

D on atmospheric ozone,

measured at different ground-based locations, are shown in Figure 9.13. Good agreement

of the clear-sky data is found within the errors associated with each measurement

technique. For further comparisons of published jO

D data, see Hofzumahaus et al.

(1992) and Kraus and Hofzumahaus (1998).

432 Analytical Techniques for Atmospheric Measurement

Time (HST)

j (O

D) (10

–5

–1

)

6 8 10 12 14 16 18

Figure 9.12 Diurnal variation of jO

D measured by a chemical actinometer at the NOAA Mauna Loa

Observatory on the island of Hawai, on 3 May, 1992 (adapted from Shetter et al., 1996, printed with

permission from American Geophysical Union).

200 400 600 800

Slant ozone column (DU)

j (O

D) (10

–5

–1

)

Figure 9.13 Dependence of jO

D on the slant ozone column. Data are from: • Mauna Loa on 6 May,

1992, measured by a chemical actinometer (Shetter et al., 1996);  Atlantic Ocean 28



N 30



W on

24 September, 1988, measured by a calibrated ﬁlter radiometer (Hofzumahaus et al., 1992);  Denver,

Colorado, measured by chemical actinometer on 18 May, 1991 (Bairai and Stedman, 1992) (adapted

from Shetter et al., 1996, printed with permission from American Geophysical Union).

9.3.5 Nitrogen-dioxide chemical actinometers

The photodissociation of NO

has one energetically allowed pathway in the troposphere,

which exhibits only a weak temperature dependence (Shetter et al., 1988).

+h < 420 nm −→ NO +O

P (R 9.3)

Since the photolytical loss of NO

is directly related to the primary NO formation, the

photolysis frequency jNO

 can be determined either from the measured loss rate of NO

Measurement of Photolysis Frequencies in the Atmosphere 433

or from the rate of NO production. However, secondary reactions of the O

P atoms

must be taken into account in the evaluation of the jNO

 data.

The O-atom chemistry is a function of the operating conditions of the chemical

actinometer and must be analysed for each experimental setup. In a system of pure NO

(or NO

mixed in N

), O

P will react at low pressure primarily with NO

and produce

another NO molecule:

O

P +NO

−→ NO+O

(R 9.16)

Thus, the effective NO yield per absorbed photon is two. This condition has been realised

in a chemical actinometer, for example, by Parrish et al. (1983), who photolysed 30 ppmv

in 60 hPa ultrapure nitrogen.

When the N

pressure is increased, side reactions start to compete with R 9.16 and

reduce the effective yield of NO:

O

P +NO

−→ NO

(R 9.17)

+NO −→ NO

+NO

(R 9.18)

O

P +NO +N

−→ NO

(R 9.19)

At a total pressure of 1 atm N

, the NO yield reaches a value of 161 ±007 for a range

of initial concentrations of 1–4 ppmv NO

and 0–0.3 ppmv NO (Zafonte et al., 1977).

The yield is additionally influenced by trace amounts of O

, which may be present as a

contaminant in the nitrogen supply. A modified value of 169±008 was estimated for

an impurity of 0.5 ppmv O

in 1 atm N

(Dickerson and Stedman, 1980).

Another approach to perform NO

actinometry involves the addition of a large amount

of oxygen or air to the actinometer. At sufficiently high pressure, for example 1 atm,

nearly all O

P atoms are converted into O

by R9.2. When the actinometer conditions

are chosen properly, for example when the time between formation and analysis of NO

is short, the influence of consecutive O

reactions like R9.12 can be neglected and the

effective NO yield becomes unity (Harvey et al., 1977; Shetter et al., 1992).

Static nitrogen-dioxide actinometers generally use pure NO

(1–5 hPa) as actinometer

gas (Table 9.4). During photolysis the NO

decay over time is either monitored directly

by photometric measurements (Bahe et al., 1980; Schultz et al., 1995) or derived indirectly

from the pressure increase associated with the NO

photodecomposition in the sealed

actinometer cell (Madronich et al., 1983, 1985). The primary photodissociation R9.3

followed by R9.16 results in the net reaction

2NO

+h −→ 2NO +O

(R 9.20)

Considering the stoichiometry of this reaction, Equation 9.24 can be modified to calculate

jNO

 from the NO

concentration measured before t = 0 and after t = t light

exposure:

jNO

 =−

2t

NO



t

NO



t=0

(9.27)

434 Analytical Techniques for Atmospheric Measurement

Equation 9.27 is only accurate when low NO

pressures are used. At high concentra-

tions, NO

produces significant amounts of the dimer N

in a temperature-dependent

equilibrium:

+NO

+M  N

+M (R 9.21)

When NO

is photolysed, the fast equilibrium will replenish some NO

by thermal

decomposition of N

and apparently slows down the NO

decay. The N

concen-

tration that is present in the actinometer can be calculated from the equilibrium constant

K = p

(147 hPa at 298 K; Sander et al., 2003). At room temperature and 1 hPa

, the fraction of N

is 0.7% which can be neglected. At higher NO

concentrations

or much lower temperatures, N

becomes relevant and the equilibrium R9.21 must be

explicitly taken into account in the jNO

 evaluation (for details, see Madronich et al.,

1983).

An example for a static jNO

 actinometer, developed for measurements on a strato-

spheric balloon payload (Madronich et al., 1983, 1985), is shown in Figure 9.14. The

inner quartz cylinder serves as a photolysis cell which can be filled with NO

from a

reservoir, or can be pumped by a 7 litre pre-evacuated tank. The cell is equipped with

sensors for pressure and temperature and the NO

decay is determined from the total

pressure increase associated with R9.20. The photolysis cell is thermally isolated from the

ambient air by a second quartz cylinder, with the annulus evacuated. A cylindrical shutter

which encloses the photolysis assembly can be driven up and down for exposure of the

photolysis cell. This system is the only chemical actinometer that was deployed in the

stratosphere. The measurements, which had uncertainties of ±6% to ±15%, were used

for tests of radiative-transfer model predictions for this atmospheric regime (Madronich

et al., 1985). Other applications of static jNO

 actinometers have involved tests and

calibrations of actinic-flux filter radiometers for tropospheric use (Schultz et al., 1995;

Volz-Thomas et al., 1996; Kraus et al., 1998). In these cases the actinometer measurements

had typical uncertainties of about ±5%.

Flowing nitrogen-dioxide actinometers use highly sensitive chemiluminescence instru-

ments for online detection of NO and NO

, which allows continuous measurements of

jNO

 with high time resolution (Table 9.4). An example is the airborne actinometer

shown schematically in Figure 9.11 (Dickerson et al., 1982). For the determination of

jNO

, a flow of pure oxygen is doped with about 20 ppmv NO

from a permeation

device. The mixture is passed through a photolysis reactor mounted atop the aircraft and

is analysed after short irradiation t ∼ 1s by a chemiluminescence instrument. The

jNO

 is then derived from the measured increase NO and the initial concentration

NO



jNO

 =

t

NO

NO



(9.28)

Other chemical flow actinometers listed in Table 9.4 follow essentially the same principle.

Their measurement uncertainties are typically in the range between ±5% and ±12%

(e.g., Kelley et al., 1995; Lantz et al., 1996; Shetter et al., 2003).

Measurement examples are shown in Figures 9.15 and 9.16. The first figure displays

a typical diurnal profile of jNO

, which was measured by a chemical actinometer at

Measurement of Photolysis Frequencies in the Atmosphere 435

M1 M2

75 CC

Figure 9.14 Schematic of a balloon-borne chemical actinometer for measurement of the photolysis

frequency of NO

. In the quartz cell, pure NO

is photolysed by solar radiation, leading to a measurable

increase of total cell pressure. Abbreviations used: P, absolute pressure transducer; T, thermistors; B1,

B2, B3 bellows valves (closed during ﬂight); M1, M2, latching solenoid valves; 75 cc, reservoir ﬁlled with

; 7 L, volume evacuated before ﬂight (from Madronich et al., 1985, with kind permission of Springer

Science and Business Media).

an observatory on Mauna Loa in 1988, and is compared with corresponding theoretical

model calculations (Shetter et al., 1992). The maximum jNO

 value at noon is around

9 ×10

−3

−1

and corresponds to a photolytical NO

lifetime of about 1.9 minutes. The

diurnal profile of jNO

 may be compared to the one of jO

D measured at the same

place and season in a different year (Figure 9.12). The jO

D profile is narrower, because

436 Analytical Techniques for Atmospheric Measurement

5791113151719

Time (HST)

j (NO

) (10

–3

–1

)

Figure 9.15 Diurnal variation of jNO

 measured by chemical actinometry at the Mauna Loa Obser-

vatory on the island of Hawai, at an elevation of 3410 m, on 21 May, 1988. Experimental data are

shown as points; the curves represent theoretical model calculations assuming slightly different amounts

of albedo enhancement by clouds lying in a valley below the observation point (adapted from Shetter

et al., 1992, printed with permission from American Geophysical Union).

Solar zenith angle (degree)

40 60 80

j (NO

) (10

–3

–1

)

50 70

Figure 9.16 Dependence of jNO

 on the solar zenith angle under cloud-free conditions. Data obtained

by chemical actinometry: — — Dickerson et al. (1982), × Madronich et al. (1983), - - Parrish et al. (1983),

— Müller and Schurath (1986),  Shetter et al. (1992),  Lantz et al. (1996); spectroradiometry: ♦

Kraus and Hofzumahaus (1998); ﬁlter radiometry:  Kraus and Hofzumahaus (1998),  Brauers and

Hofzumahaus (1992). Solid and dashed lines are parametrisations from original data (adapted from Kraus

and Hofzumahaus, 1998, with kind permission of Springer Science and Business Media).

it depends on radiation that is strongly absorbed by stratospheric ozone at low sun,

whereas jNO

 is essentially independent on ozone absorption. Examples of the functional

dependence of jNO

 on the solar zenith angle are given in Figure 9.16 displaying

observations at different ground-based locations for clear-sky conditions.

Measurement of Photolysis Frequencies in the Atmosphere 437

9.4 Actinic-ﬂux spectroradiometry

Spectroradiometry is the measurement of absolute photon fluxes in narrow spectral

intervals at selected wavelengths (see e.g. McCluney, 1994). This task requires an optical

instrument (spectrometer) that contains three basic elements:

•

A receiver optic for radiation incident from outside.

•

A device for the spectral dispersion of radiation into different wavelengths.

•

A detection system for quantification of the dispersed light.

If a spectrometer is primarily designed and calibrated for the absolute measurement of

photon-flux densities, it is called a spectroradiometer.

A scanning spectroradiometer uses a monochromator which selects radiation in a small

spectral interval that can be scanned through the wavelength range of interest over time.

In this operational mode the radiation spectrum is measured sequentially.

A spectrograph exposes a dispersed spectrum of a whole range of wavelengths at once

to a many-element detector array, such that radiation of certain wavelengths can be

assigned to specific detector elements. In this mode of operation the radiation spectrum

is measured simultaneously.

Spectroradiometers can measure different radiometric quantities, usually irradiances

or radiances, depending on the design of the reception optics. In atmospheric physics,

spectroradiometers have been used for many decades for measuring solar UV and visible

irradiances (see Section 9.6). Only since the second half of 1990s, spectroradiometers

have been developed and used for the direct measurement of solar actinic fluxes, in

order to determine atmospheric photolysis frequencies (e.g. Müller et al., 1995; Kraus &

Hofzumahaus, 1998; Shetter & Müller, 1999; Webb et al., 2002b; Eckstein et al., 2003;

Edwards & Monks, 2003; Kanaya et al., 2003; Jäkel et al., 2005). Instruments that have

been used for field measurements are listed in Table 9.5.

9.4.1 Principle of actinic-ﬂux spectroradiometry

Actinic-flux spectroradiometers collect incident solar radiation with isotropically sensitive

receiver optics and record the spectrum of the radiation by either a scanning double-

monochromator system or a spectrograph. A photolysis frequency can then be calculated

from the measured spectrum (cf. Equation 9.12), when three conditions are fulfilled:

1. The intensities and wavelengths of the measured radiation are absolutely calibrated.

2. The absorption spectrum  and quantum yield  of the relevant photolysis reaction

are accurately known.

3. The measured radiation spectrum covers the full range of wavelengths that contributes

to the photodissociation.

This concept has the important advantage that photolysis frequencies of a variety of

atmospheric gases can be obtained from the same measurement of radiation obtained by

a single instrument.