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

Подождите немного. Документ загружается.

Plate 3a The diurnal cycle of NO

at eight sites in the UK, showing the morning rush hour maximum

due to vehicle traffic and a secondary maximum in the evening. (Figure 6.14 of AQEG, “Nitrogen Dioxide

in the United Kingdom”, 2004, printed with permission from Crown.) (See Figure 7.10a.)

Plate 3b The decadal trend in NO

at nine sites in the UK, showing the gradual decline due to the tight-

ening of vehicle emission standards. (Figure 6.5 of AQEG, “Nitrogen Dioxide in the United Kingdom”,

2004, printed with permission from Crown.) (See Figure 7.10b.)

(a)

200

180

160

140

120

100

(µg m

–3

, as NO

)

London Bloomsbury (µg m

–3

, as NO

)

London Bexley (µg m

–3

, as NO

)

West London (µg m

–3

, as NO

)

Manchester Town Hall (µg m

–3

, as NO

)

Glasgow City Chambers (µg m

–3

, as NO

)

Belfast Centre (µg m

–3

, as NO

)

Port Talbot (µg m

–3

, as NO

)

Lullington Heath (µg m

–3

, as NO

)

01234567891011

Hour of Day

12 13 14 15 16 17 18 19 20 21 22 23

225

(b)

200

175

150

125

100

Bridge Place

Annual mean NO

(mg m

–3

, as NO

)

London Bloomsbury

London Bexley

West London

Manchester Town Hall

Glasgow City Chambers

Belfast Centre

Port Talbot

Lullington Heath

1982

1983

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

ALT-Colorplate 1/19/06 12:53 PM Page 2

Plate 3c The geographical distribution of NO

in the UK, derived from monitoring-network data inter-

preted with an empirical model. (Figure 3.1 of Stedman et al., 2002, printed with permission from AEA

Technology.) (See Figure 7.10c.)

ALT-Colorplate 1/19/06 12:54 PM Page 3

50 55 61 66

72 77 82

75 84

400

300

200

100

0 10 20 30 40 50 60 70 80 90 100

51015202530354045505560 7065 75 80

Retention time (min)

1st retention time (min)

2nd column RT / secs

Plate 4 Comparison of single column (upper) and GC × GC separations (lower) of a Leeds urban air sam-

ple. Areas of the full chromatogram are successively extracted at higher gain to illustrate increasing isomeric

complexity at higher boiling points. GC × GC chromatograms are annotated with start of individual C

isomer band (running right to left) where A = C

, B = C

, C = C

, D = C

, E = C

, F = C

, G = C

H = naphthalene. Chemical banding assignments, 1; aliphatics, 2; olefinics, 3; oxygenated, 4; mono

aromatics, 5; polyaromatics. (From Hamilton & Lewis, 2003, reproduced with permission from Elsevier.)

(See Figure 8.10.)

Plate 5 Two-dimensional GC × GC-TOFMS total ion current (tic)-plot of an aerosol sample in 2D-contour

plot: (A) showing the full chromatogram of the analysed aerosol with (B) the extraction of the selected section

for data analysis. (From Welthagen et al., 2003, reproduced with permission from Elsevier.) (See Figure 8.11.)

GC × GC plot of Aerosol sample

(A)

(B)

4.0

3.5

3.0

2.5

2.0

1.5

1.0

4.0

3.5

3.0

2.5

2.0

1.5

1.0

Polarity 2nd dimension (seconds)Polarity 2nd dimension (seconds)

1000 1500 2000

2500 3000 3500 4000

600 1250 2500 3750 5000 6250 7500 8750 10

000

Volatility 1st dimension (seconds)

000

000 40 000

ALT-Colorplate 1/19/06 12:54 PM Page 4

Mass Spectrometric Methods for Aerosol Composition Measurements 271

dislodged after impaction and become re-entrained in the air flow, meaning they may

be collected on the wrong stage or not at all. To mitigate this problem, the substrates

are often pre-treated or coated before sample collection to promote adhesion. However,

these methods are only permitted when the treatment or coating will not perturb

the aspect of particle chemistry of interest or interfere with the analytical procedure

being used.

While these bulk sampling methods can employ many powerful and well-established

analytical techniques to determine the chemical nature of the particles, they carry many

intrinsic limitations. First, a measurable amount of particulate material has to be collected,

and the detection limits of laboratory chemical analysis instrumentation usually require

sampling times from several hours to days, so temporal resolution is generally poor.

Secondly, because of the intervening time between sample collection and analysis, volatile

components of the aerosol may evaporate and be lost or chemically unstable compounds

may react in the interim (Chow, 1995; Zhang & McMurry, 1987, 1992). Semi-volatile

chemicals may also interchange with the gas phase during sampling or before analysis

and there is the potential for contamination during handling. Thirdly, while cascade

impactors provide some size-resolved information, the resolution is generally poor and

particle bounce may result in erroneous size information. Finally, because the aerosol is

handled in bulk, no information on the extent of internal or external chemical mixing is

retained.

A common remedy for the problem of gas-phase interchanges during bulk sampling

is to place a suitable denuder (tubing lined with a reagent) upstream of the collector to

absorb any gas phase chemicals that would otherwise contaminate the substrate. A second

denuder or chemically treated filter can also be placed downstream and subsequently

analysed to measure the amount of particle-phase material that has evaporated from the

substrate during sampling (e.g. Mader et al., 2003).

6.3.2 Online bulk sampling and analysis methods

Some of the problems associated with bulk sampling such as gas-phase partitioning after

collection and contamination during handling can be avoided by performing the analysis

in situ by an automated instrument, for example the Thermo Electron Series 8400N

Ambient Particulate Nitrate Monitor. This collects a sample over time through impaction

and then, through heating, converts the nitrate fraction of a sample to NO

, which it

then measures to calculate the particulate nitrate mass concentration (Stolzenburg &

Hering, 2000). While this offers an increased time resolution (minutes are possible), it

is limited to one aspect of particle chemistry, although instruments employing variations

of this principle can be used in parallel to study other chemical species (e.g. sulphate and

OC/EC).

The particle into liquid sampler (PILS) (Orsini et al., 2003; Weber et al., 2001) and

the steam jet aerosol collector (SJAC) (Khlystov et al., 1995) are examples of automated,

in situ methods that use a more generalised approach. Particles are grown to large sizes

by introducing them into a supersaturated water environment, and then impacted on a

surface, over which water is flowed continuously. The sample water is collected period-

ically and analysed, frequently using IC, to quantify the amounts of various chemicals

272 Analytical Techniques for Atmospheric Measurement

present in the aerosol. This allows high time resolution data (minutes) on many ionic

components of the aerosol to be collected simultaneously. The collection technique is

also not limited to IC; with a suitable detector, it can be used to quantify fractions such

as water-soluble organic carbon as well.

Another novel in situ method is the thermal desorption aerosol GC-MS/FID (TAG)

(Williams et al., 2004). This collects particles in bulk on a cooled surface through

impaction over a period of time. After collection, the surface is isolated and then heated,

vaporising the components. The organic fraction is then transferred to a gas chromato-

graph for analysis, using either a flame ionisation detector for total mass measurements

or a mass spectrometer for compound identification. This method is similar to the online

methods involving thermal decomposition, discussed later. The analysis produced by this

instrument is in many ways analogous to the results obtained by offline GC analysis;

while many specific compounds can be positively identified and quantified, the majority

of the organic matter in particles remains unresolved. These methods are similar to those

employed for the gas-phase speciation of organic compounds, as discussed in Chapter 8.

These instruments are automated and therefore less labour-intensive and so are ideally

suited for continuous monitoring operations. However, while these online instruments

are very powerful and reliable, they have inherent limitations in that size-resolved data

are not delivered, and as is common to all forms of bulk sampling instrumentation,

cannot provide any information regarding mixing states.

6.4 Online aerosol mass spectrometry

The field of online aerosol mass spectrometry has emerged to provide a real-time method

for the chemical analysis of aerosols. These techniques overcome the problems associated

with bulk collection and offline analysis by analysing the composition in situ on a particle-

by-particle basis. The basic principle of an aerosol mass spectrometer is to introduce

airborne particles into the instrument, vaporise and ionise the material and then analyse

the ions produced using mass spectrometry. There have been many different designs of

aerosol mass spectrometer over the years, most developed in-house at research institu-

tions around the world and tailored to specific tasks, usually related to either ambient

sampling or specific laboratory studies. However, one of the more useful features of these

instruments is their versatility; in most cases, the individual instruments have proved

useful in multiple applications.

At the time of writing, two models of aerosol mass spectrometer are available commer-

cially: the TSI model 3800 Aerosol Time Of Flight Mass Spectrometer (ATOFMS), an

instrument based on the pioneering work of Prather et al. (1994) and developed further

by Gard et al. (1997), and the Aerodyne Research Inc. Aerosol Mass Spectrometer (AMS;

Billerica, MA, USA), a design introduced by Jayne et al. (2000). Both of these instruments

are designed to be portable and suitable for both field and laboratory studies, but the data

produced are fundamentally different. The ATOFMS, being a laser desorption/ionisation

and time of flight-based instrument, delivers quantitative size and largely qualitative

composition information on individual particles, while the AMS, being based around

thermal desorption, electron impact ionisation and quadrupole mass spectrometry, gives

Mass Spectrometric Methods for Aerosol Composition Measurements 273

quantitative data on both the size and composition of the entire aerosol ensemble, but

gives only limited data on specific particles and cannot study refractory components.

The technical aspects of these and many other examples of similar instruments will

be discussed in this section and the specific advantages of the various approaches

explored. Instruments developed at the beginning of twenty-first century that will be

referred to by name include the Aerosol Composition Mass Spectrometer (ACMS),

Atmospheric Pressure Chemical Ionisation Mass Spectrometer (APCI-MS), the Chapel Hill

instrument, the Laser Mass Analyzer for Particles in the Airborne State (LAMPAS), the

Particle Analysis by Laser Mass Spectrometry (PALMS), the Particle Blaster, the Rapid

Single particle Mass Spectrometry (RSMS), the Surface Ionisation Particle Beam Mass

Spectrometer (SI-PBMS), the Thermal Desorption Chemical Ionisation Mass Spectrometer

(TD-CIMS) and the Thermal Desorption Particle Beam Mass Spectrometer (TDPBMS).

Table 6.1 lists these instruments, their measurement technologies and selected publica-

tions. The list is not exhaustive, but represents a wide spectrum of the novel techniques

used. Examples of further implementations of these techniques can be found elsewhere

in the literature. (e.g. Hearn & Smith, 2004; Hunt & Petrucci, 2002; Öktem et al., 2004).

6.4.1 Mass spectrometer types

The principle of a mass spectrometer is to separate and count ions according to their

mass-to-charge ratios (m/z). A detailed discussion is given in Chapter 5, and so the

operation of a mass spectrometer is considered only briefly here. Quantitatively, m is

taken as the mass of the ion relative to the standard atomic mass (defined as one twelfth

of the rest mass of a

C atom, or 16606×10

−27

kg) and z is the charge relative to e, the

elementary charge 1602 ×10

−19

C. The charge of most ions detected is normally ±1e,

although multiple charging is possible, depending upon the composition of the ion and

the ionisation technique used. In this chapter, m/z is treated as being dimensionless,

although atomic mass units (amu or u), Daltons (Da) and Thompsons (Th) are used as

units for the same quantity elsewhere in the literature.

The most basic design of a mass spectrometer is the magnetic sector mass spectrometer,

which accelerates and focuses ions using electric fields and then bends their paths with trans-

verse magnetic fields. As the ions are accelerated and deflected by electric and magnetic fields

of specific strengths, their velocities and deflected trajectories are therefore dependent on

their m/z. Therefore, by using a fixed detector (such as an electrometer), the ions are filtered

according to their m/z prior to counting. This type of mass spectrometer can be scanned by

varying the electric or magnetic field strengths. While this method is generally not favoured

in current designs of aerosol mass spectrometers due to its bulk, it is capable of very high

resolutions, so is often used in other laboratory applications where mass measurements

of fractions of amu are needed, such as when identifying specific elements. An exception

is the RCMS, which has used the compact design introduced by Nier & Schlutter (1985).

Quadrupole mass spectrometers work by again accelerating the ions using electric fields

but this time, the ions are selected by passing them between four parallel rods. A voltage is

applied between the two sets of opposing rods, which consists of AC and DC components.

The ions adopt oscillating trajectories as they travel the length of the rods (see for example

Chapter 5, Figure 5.4), the magnitude of the oscillations dependent on their m/z, the AC

274 Analytical Techniques for Atmospheric Measurement

Table 6.1 Instruments to measure aerosol composition using mass spectrometry

Instrument Vaporisation

and ionisation

Mass

spectrometer

Sizing Selected literature

ACMS Thermal-EI Quadrupole or

magnetic sector

None Schreiner et al.,

2002

AMS Thermal-EI Quadrupole Polydisperse

aerodynamic

Jayne et al., 2000;

Jimenez et al.,

2003a

APCI-MS Thermal-CI Ion Trap None Hoffmann et al.,

2002

ATOFMS

(TSI 3800)

Single laser

∗

Bipolar reﬂectron

(originally

monopolar

linear) TOF

Polydisperse

opto-aerodynamic

Gard et al., 1997;

Prather et al.,

1994

Chapel Hill

Instrument

Dual laser or

thermal-laser

Linear TOF Polydisperse

opto-aerodynamic

Sykes et al., 2002;

Woods et al.,

2001

LAMPAS/

LAMPAS 2

Single laser Bipolar

(originally

monopolar)

linear TOF

Polydisperse

(originally

monodisperse)

opto-aerodynamic

Hinz et al., 1994;

Trimborn et al.,

2000

PALMS/

WB-57F

PALMS

Single laser Reﬂectron

(originally PSPF)

TOF

Polydisperse

optical, later with

opto-aerodynamic

Murphy and

Thomson, 1995;

Thomson et al.,

2000

Particle

Blaster

Single laser Reﬂectron TOF None Reents et al.,

1995; Reents and

Ge, 2000

RSMS/RSMS

II/RSMS III

Single laser Linear (originally

reﬂectron) TOF,

later bipolar

Monodisperse

aerodynamic

(originally none)

Carson et al.,

1995; 1997a;

Lake et al., 2003

SI-PBMS Thermal Quadrupole None Svane et al., 2004

TD-CIMS Thermal-CI Triple quadrupole None Voisin et al., 2003

TDPBMS Programmable

thermal-EI

Quadrupole None Tobias and

Ziemann, 1999;

Tobias et al., 2000

∗

Two-step laser desorption and ionisation was ﬁrst demonstrated on an instrument of this type (Morrical et al., 1998),

although it has not been widely used since.

and DC voltages applied and the frequency of the AC voltage. Normally, the frequency is

kept fixed during operation, as it is easier to vary the magnitudes of the voltages. The AC

voltage acts as a high-pass m/z filter, as ions with too low an m/z adopt trajectories with

oscillations larger than the spacing of the rods, causing them to strike the rods, become

neutralised and not be detected. In a similar manner, the DC voltage acts as a low-pass

Mass Spectrometric Methods for Aerosol Composition Measurements 275

m/z filter by further perturbing the oscillations of high m/z ions. Therefore, if the two

voltages are selected correctly, a quadrupole can act as a filter for all ions but those of

the desired m/z.

After the rods, the ions are detected using devices such as electron multipliers or

conversion dynodes, sometimes after extraction using additional electric fields. While

essentially performing the same filtering function as a magnetic sector mass spectrometer,

quadrupoles are much more compact, robust and generally easier to implement, making

them ideal for field instrumentation. Their accuracy is also not affected by the initial

velocities of the ions prior to selection. However, a disadvantage of selecting the ions this

way is that the resolving power of quadrupoles is limited by the amount of time the ions

spend in the rod region, which is typically less than a microsecond. Because of this, the

m/z resolution of quadrupole mass spectrometers is normally around unity.

A third common type of mass spectrometry is the time-of-flight (TOF) method. The

basic principle is to again accelerate the ions over a specific electric potential, but the

time taken to travel a set distance to the detector (e.g. microchannel or microsphere

plates) is measured, from which the velocity and therefore the m/z can be calculated (see

Figure 5.6 in Chapter 5). This technique requires accurate timing electronics, rapid data

collection and the precise gating of the start of the time of flight, meaning the ions must

be delivered in discrete pulses. It does have a major advantage in that ions of many m/z

can be detected simultaneously, which is not possible with the previous two methods,

and does not have a maximum m/z imposed upon it. The TOF mass spectrometer

is capable of sub-unity m/z resolution, but is limited by several factors, including the

duration of the ion pulse, the spatial distribution of the ions before extraction and the

distribution of velocities of the ions prior to acceleration. There are ways of mitigating

these losses in resolution, post source pulse focusing (PSPF) being one such method

(Kinsel & Johnston, 1989). This involves applying an additional voltage pulse to part of

the drift section after the ionisation event, which causes any lagging ions to catch up with

those of the same m/z. While this produces a satisfactory improvement in the resolution,

it is intrinsically limited to a particular portion of the mass spectrum. The wider the

section it is applied to, the less pronounced the improvement. Many implementations

make use of a reflectron instead, as shown in Figure 6.1, which specifically reduces the

broadening due to uncertainties in the ion velocities in the drift region. Electric fields are

used to reverse the direction of travel of the ions halfway through their flight. As ions with

a velocity higher than intended for their m/z penetrate deeper into the reflectron, the

reflection time is increased, so the otherwise shortening of their time of flight is largely

Ion source

Reflectron

Detector

Figure 6.1 The basic principle of a reﬂectron time-of-ﬂight mass spectrometer. Trajectories of ions with

higher and lower velocities than desired are shown with dashed lines.

276 Analytical Techniques for Atmospheric Measurement

cancelled out, providing the electric fields are tuned correctly (Weickhardt et al., 1996).

An advantage of this over PSPF is that providing it is correctly tuned, the increase in

resolution benefits the ions of all m/z.

Ion trap mass spectrometers use a system of two end-cap electrodes either side of a

ring electrode, which is effectively the rod geometry of a quadrupole wrapped around

on itself. By applying the AC and DC voltages between the caps and the ring in the

same manner as in a quadrupole, ions of a range of m/z’s are indefinitely held in

stable three-dimensional orbits between the electrodes. The voltages can be initially set

to capture and hold all ions and then subsequently ramped to selectively release ions

according to their m/z for detection. Like the TOF mass spectrometer, this is capable

of delivering a complete mass spectrum from a single ionisation event, but is not reliant

upon having a pulsed ion source. Also, it is capable of higher resolutions than quadrupole

mass spectrometers due to the increased resolving time. However, because it must be

scanned, it takes much longer than a TOF mass spectrometer to deliver a complete mass

spectrum, so is unsuitable for rapid data collection.

A method sometimes used to further investigate the origin of specific peaks within a

mass spectrum is tandem mass spectrometry or MS-MS. This can be performed using a

triple quadrupole mass spectrometer, which is in effect three quadrupoles in series. The

first quadrupole selects an m/z of interest, while the second does not act as a filter but as

a region where the ions are further fragmented by techniques such as collision induced

dissociation (CID). The third quadrupole then performs a scan to give a spectrum of

the secondary fragments, which is of particular use in identifying organic chemicals.

By disabling the initial selection and secondary fragmentation, this configuration is also

capable of performing the same task as a standard quadrupole, so the peaks of interest in

a sample can be identified. An MS-MS is also possible with ion trap mass spectrometers;

the basic method is to release all ions except for those of the m/z of interest from the

trap before returning it to a state where it will hold all ions. The remaining ions are then

fragmented using CID before the mass spectrometer is scanned as normal.

6.4.2 Inlet technologies

Mass spectrometers must be operated at high or ultra-high vacuum, because random

collisions with gas molecules can prevent ions from being detected. Therefore, a key

feature of aerosol mass spectrometry involves removing as much of the gas-phase material

from the aerosol sample as possible while retaining the particulate fraction for analysis.

The exceptions are the thermal desorption-chemical ionisation methods (e.g. APCI-

MS, TD-CIMS) because while the mass spectrometer is operated at high vacuum, the

collection, desorption and ionisation can take place at atmospheric or near atmospheric

pressure. The most common method of removing gas from the sample is to form a

collimated particle beam from the aerosol at the inlet and skim off the majority of the

gas-phase material using a series of apertures and differential pumping. In most early

designs, the particle beam was formed by accelerating the particles through a capillary

tube or nozzle into the vacuum of the instrument. Due to their inertia, the particulate

fraction of the aerosols forms a less divergent jet than the gas fraction at the point of

expansion. The particles pass through a series of channels or apertures, allowing the

Mass Spectrometric Methods for Aerosol Composition Measurements 277

excess gas to be removed by the pumping system. There is still some divergence in the

particle beams, with the amount of divergence dependent on the aerodynamic size of the

particles. Particles must hit a physical target or pass through laser focal points in order to

be detected and so there is a tolerance in the solid angle in which particles can travel and

be detected. This in turn leads not only to a reduction in sensitivity, but also to a bias

towards particles of a given aerodynamic size. These can both be mitigated by reducing

the distance between the nozzle and the detection point.

Since mid-1990s, the implementation of aerodynamic lenses prior to a nozzle expansion

has sought to address these problems (Liu et al., 1995a,b; Petrucci et al., 2000; Zhang

et al., 2002, 2004), as shown in Figure 6.2. Particles are drawn through a series of

concentric apertures with successively decreasing diameters, which causes the gas stream-

lines to rapidly compress and expand as the aerosol flows through the lens system. These

successive compressions cause the particle streamlines to converge on the axis of the

lens, so that when they are accelerated through the nozzle, the radial components of

their velocities are small and the streamlines take the form of a tightly collimated beam.

As the lens must be operated at a low pressure to reduce defocusing due to Brownian

motion, the aerosol must normally pass through a critical orifice before entering the lens.

The main advantage of aerodynamic lenses is that particles over a wide range of sizes

are focused into a narrow and very parallel beam, thereby helping to eliminate any size-

dependent detection biases an instrument may have over a given size range and improving

quantitative capabilities. The tighter beam also increases the number of particles detected,

which in turn improves the instrument sensitivity and overall data quality, particularly

for the smaller particles (Kane et al., 2001; Su et al., 2004). The TDPBMS, ACMS, AMS,

SI-PBMS and Chapel Hill instrument (see Table 6.1) are among the instruments that use

this technology and TSI have manufactured an aerodynamic lens (model 3801-030 AFL)

as an optional accessory for the ATOFMS. In the case of the ACMS, the lens design was

modified to allow sampling at high altitudes (Schreiner et al., 1998).

Normally, a particular lens implementation can effectively focus a specific range of

particle sizes; for instance, the 3801-030 is specified for particles between 30 and 300 nm

in aerodynamic diameter. Smaller particles do not posses enough inertia to be focused

Radial coordinate (m)

Axial coordinate (m)

0.00

–0.005

0.005

0.000

0.10 0.20 0.30

Figure 6.2 Computed 100 nm particle trajectories within an aerodynamic lens system. Several aerosol

trajectories enter from the left and become focused down the axes of the lenses during the successive gas

expansions and contractions. The right side is where the aerosol enters the ﬁrst vacuum chamber of the

instrument. (Jayne et al., 2000, reprinted with permission from 2000 American Association for Aerosol

Research, published by Taylor and Francis.)