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

Подождите немного. Документ загружается.

278 Analytical Techniques for Atmospheric Measurement

properly, so their trajectories tend to follow the gas streamlines, and they are also

the particles most affected by Brownian motion, which adds an inherently random

component to their velocities. Larger particles can overshoot the axis during focusing at

a particular lens stage, in effect becoming defocused, or more simply they can impact

on the interior surfaces of the lens system during the rapid changes of gas velocity.

These factors result in decreased focusing efficiencies for particles outside these diameter

ranges. As aerodynamic lenses are normally best suited for submicron particles, a more

traditional nozzle inlet is often preferable when studying coarse particles. An unavoidable

limitation of aerodynamic inlets is encountered when sampling certain types of non-

spherical particles, as these are subject to lift forces perpendicular to the direction of

acceleration, in effect causing extra divergence.

6.4.3 Particle sizing methods

As well as the chemical nature of aerosol particles, aerosol mass spectrometers are often

capable of providing information on the size of the individual particles sampled, allowing

the direct linkage of chemical compositions with particle size. The ability to do this is

invaluable when deriving mixing states, applying compositions to particle populations

measured by sizing instruments, or indeed applying accurate size and time information to

chemical composition data from offline aerosol analysis techniques. A method favoured

in some of the earlier instruments was to measure the total response in the mass

spectrometer, in effect measuring the total amount of constituents and therefore the

volume of the aerosol (e.g. Allen & Gould, 1981; Myers & Fite, 1975), but as this is only

reliable for aerosols of known composition, it is not suitable for ambient measurements.

For this reason, it is desirable for an aerosol mass spectrometer to employ an independent

sizing method.

6.4.3.1 Sizing by monodisperse sampling

Certain designs of aerosol mass spectrometer are intrinsically incapable of sizing the

particles on their own, due to the handling of material in bulk during analysis (e.g.

TDPBMS, TD-CIMS) or a lack of triggering (e.g. RSMS-II). However, they are still

capable of delivering size-resolved particle data if the sampled aerosol particles are selected

according to size before analysis, which can be performed by coupling the instrument in

series with a DMA. By stepping the DMA and taking repeated measurements with the

aerosol mass spectrometer, size-resolved (in D

space) data can be derived. Corrections

for multiply charged particles can be made in a similar manner as an SMPS. An alternative

method, used by the RSMS II and III, is to exploit the size biases that nozzle inlets

incur. The particles are selected aerodynamically by carefully designing the instrument’s

inlet to have an optimum aerodynamic diameter for delivery to the instrument, with the

particles of other sizes being removed by the differential pumping system. This optimum

can then be controlled by varying the inlet pressure (Mallina et al., 2000; Phares et al.,

2002). The problem with monodisperse sizing is that multiple data captures are needed to

generate a complete distribution, leading to long scan times. While this is not a problem

in laboratory scenarios where there are no shortages of particles, and systems are stable

Mass Spectrometric Methods for Aerosol Composition Measurements 279

or repeatable, this sizing method is not ideal for ambient sampling, especially in relatively

clean environments or during aircraft measurements.

6.4.3.2 Optical sizing

Optical sizing by scattering intensity measurements is in principle compatible with

aerosol mass spectrometry because it is passive, i.e. the particles are unaffected by the

measurement, so an in situ probe can be used to measure the particles prior to desorption.

Also, many designs of laser-based instruments require the particles to be detected in

order to trigger the desorption laser (see Section 6.4.4), so in this manner a detection

laser can serve a dual purpose. This method has been used in the PALMS instrument in

particular. However, it is intrinsically limited by the wavelength and intensity of the light

used and subject to all the extra complications associated with shape and composition,

and variations in scattering intensity with angle that are associated with all optical sizing

methods.

6.4.3.3 Aerodynamic sizing

At the time of writing, most designs of aerosol mass spectrometer use aerodynamic

sizing to deliver size-resolved information of the particles in a similar manner as an

APS. The basic principle is to measure the velocity of the particles after the nozzle

acceleration, which can be linked to their aerodynamic diameter. Salt et al. (1996) showed

this to be a much more reliable sizing method than single-angle optical scattering, due

to the elimination of the complex issues associated with optical sizing mentioned above.

A version of selective, monodisperse aerodynamic sizing was used in the original LAMPAS

instrument. This featured a spatial separation on the particle beam path between the

optical detection point and the focal point of the desorption laser. By adjusting the

time delay between detection and triggering of the desorption laser, only particles of

a particular velocity are detected. A polydisperse aerodynamic method first introduced

by Prather et al. (1994) is used in more recent instruments such as the ATOFMS,

Chapel Hill Instrument and LAMPAS 2. With these, the velocity is measured by using

two optical scattering detectors at different stages of the particle beam and the velocity

calculated accordingly. As well as providing sizing information, fast electronics derive the

appropriate triggering delay for the desorption laser on the fly, so that analysis can be

performed on particles of all sizes with an optimum hit rate. In exactly the same manner

as OPCs and APSs, limitations are imposed on this sizing method by the wavelength and

intensity of the light used in the detectors, so sizes of particles of around 200 nm or less

in optical diameter cannot be delivered.

The AMS adopts a slightly different approach to aerodynamic sizing. During normal

operation, the instrument alternates between two modes (Jimenez et al., 2003a). In the

first mode, particles of all sizes are sampled as the mass spectrometer scans to deliver

complete average mass spectra of the entire particle ensemble sampled. The second mode

uses a chopper wheel (a rotating disc with radial slits cut into it) to deliver pulses of

particles into the instrument (at a rate of around 100–130 Hz) while the mass spectrometer

stays fixed at one of a selection of specific m/z’s. A TOF region separates the particles

280 Analytical Techniques for Atmospheric Measurement

temporally according to their velocities before they are detected. As flash vaporisation

and quadrupole mass spectrometry operate on very short timescales for non-refractory

components (of the order of s), it can measure the particles’ times of flight (of the order

of ms) by comparing the detection times with the slit positions as measured by an optical

sensor. This method can detect individual particles as pulses in the mass spectrometer

signal or, by averaging the measured signals over successive chopper periods, can derive

an ion count rate as a function of particle TOF, which can in turn be transformed into

particle diameter (Allan et al., 2003a). This method has the advantage of not being limited

by the optical size of the particles, therefore the lower sizing limit of the instrument is only

limited by the inlet transmission. However, it is intrinsically limited to continuous ion

source methods, so it is not compatible with laser-based instruments. It also cannot deliver

size information as accurately as the opto-aerodynamic method due to the finite width

of the chopper slits (typically 1–2% of the chopper circumference). Finally, it assumes

that the particle vaporisation and ion generation and detection times are short; particles

that vaporise slowly or those that generate very high mass ions can be sized inaccurately.

Another limitation of aerodynamic sizing in general is the need to introduce a particle

TOF region to the instrument between the inlet and the detection region. The larger the

region, the more accurate the sizing, but this increases the instrument’s susceptibility to

particle beam divergence, as described in the inlet Section 6.4.2 above, thereby reducing

its sensitivity and possibly increasing any relative biases towards particle of certain sizes.

The differences between methods must be taken into consideration when comparing

size-resolved aerosol mass spectrometry data with those of another particle sizing

instrument. In particular, the aerodynamic diameters delivered by aerosol mass spectrom-

eters must be treated with some caution due to the low pressures that these instru-

ments are operated at. Traditional aerodynamic sizing methods, for example impactor or

cyclone inlets, cascade impactors and aerodynamic particle sizers, typically size at near

atmospheric pressure, where the size of the particle is comparable to or much greater than

the mean free path of the gas molecules (known as the transition and continuum regimes

respectively). Providing the Reynolds’ numbers remain low, the particle mechanics are

normally adequately described by Stokes’ formula, where the drag force exerted on a

particle is proportional to its diameter (Fuchs, 1964). This leads to the aerodynamic

diameter D

, as measured by these instruments, to have a square root dependence on the

particle density when relating it to the volume-equivalent diameter D

. The relationship

is given by (DeCarlo et al., 2004; Kelly & McMurry, 1992; McMurry et al., 2002):







= D













(6.1)

where 

is the particle density (in g cm

−3

), 

the unity density 1g cm

−3

 and  the

dynamic shape factor of the particle. The latter quantity is unity for perfect spheres and

increases as the particle becomes more irregular and therefore more subject to drag forces

(Brockmann & Rader, 1990). The ratio of the relative particle density to the dynamic

shape factor is often referred to as the particle ‘effective density’ or 

eff

for convenience.

CD is a slip correction function, applied to take account of the kinetic discontinuity at

the particle’s surface (Cunningham, 1910). For large particles, continuum fluid dynamics

applies and this quantity is close to unity, but as the particle sizes begin to approach the

Mass Spectrometric Methods for Aerosol Composition Measurements 281

mean free path of the gas molecules (67.3 nm for air at 23



C and 730 torr), it increases

and can become significant (Allen & Raabe, 1982).

Because the aerodynamic acceleration in aerosol mass spectrometers usually takes place

in the form of a supersonic expansion into high vacuum, the particles can be sized under

conditions where the mean free path of the gas molecules is much greater than the size

of the particles (known as the free molecular regime). This is especially true of systems

using aerodynamic lenses or those operated at high-altitude, where the pressure upstream

of the nozzle is also low. Under these conditions, the drag force exerted on a particle

is proportional to its cross-sectional area rather than its diameter (Fuchs, 1964) and

therefore the aerodynamic diameter measured is a fundamentally different property to the

‘classical’ aerodynamic diameter measured at higher pressures. Jimenez et al. (2003b,c)

introduced this as the ‘vacuum aerodynamic diameter’ or D

of a particle. The empirical

definition is essentially the same as the classical version, being the diameter of a sphere of

unit density that would behave the same way under the same conditions, but it is related

to the particle’s physical properties as follows (DeCarlo et al., 2004):





(6.2)

Because of this, assumptions about particle shape and density must be made when

comparing size-resolved aerosol mass spectrometry data with those of other instruments,

even when using other aerodynamic methods. For instance, with heavier-than-unity

particles typically found in the atmosphere, a vacuum aerodynamic instrument will report

slightly larger diameters than a classical aerodynamic instrument. These differences can

be used to a scientist’s advantage, as the coupling of two sizing methods in series can be

used to derive properties such as the effective density of the particle (e.g. Jimenez et al.,

2003b; McFiggans et al., 2004; Murphy et al., 2004).

Another limitation of this sizing technique is the effect the sudden change of pressure

has on the particles, which is again especially pronounced for systems that use aerody-

namic lenses, as there is a significant residence time between the critical orifice and

the nozzle. For example, aerosol particles are often sampled as aqueous droplets but

an amount of the water fraction of these may evaporate during aerodynamic focusing

and sizing. As the precise amount of water lost is difficult to quantify or predict, this

often causes problems when trying to report the ‘true’ sizes of particles (Buzorius et al.,

2002). For this reason, it is often pertinent to control the humidity of the aerosols

sampled.

6.4.4 Ablation and ionisation techniques

One of the principle features of any aerosol mass spectrometry technique is the conversion

of some or all of the condensed phase particulate matter into gas-phase ions. The methods

employed by the various instrument designs can be separated into three main categories:

thermal methods, where the particle constituents are vaporised thermally and ionised

using heat, electrons or chemical ions; single-step laser methods, where a single laser pulse

is used to simultaneously vaporise and ionise the particulate matter; and two-step laser

282 Analytical Techniques for Atmospheric Measurement

methods, where a laser pulse performs the ionisation but the vaporisation is performed

thermally or using a second, weaker laser. The technical details of the different methods,

along with their strengths and weaknesses, examples of instruments that use them and

how the data they generate can be used are presented in the following sub-sections.

6.4.4.1 Thermal methods

The earliest designs of aerosol mass spectrometers vaporised and ionised the particulate

matter in a single step by impacting the particle beam upon a (normally metallic) surface

heated to a very high temperature, known as the vaporiser (e.g. Davis, 1973; Myers

& Fite, 1975). This technique, known as surface ionisation, is used in some modern

instruments such as the SI-PBMS. Its advantage is that it yields large numbers of ions

and is very useful for the quantitative study of specific inorganic components of aerosol

particles, in particular the alkali elements (Davidsson et al., 2002; Jäglid et al., 1996). This

makes the instrument very useful for the study of sea-salt based particles, in particular.

However, the strong biases towards specific elements, based on their electronic structure

and the work function of the vaporiser material, makes surface ionisation unsuitable for

general-purpose aerosol instrumentation.

These issues can be overcome by separating the vaporisation and ionisation stages of

the analysis; the impaction surface is heated to a lower temperature so that the particles

are vaporised in the form of a neutral gas and are subsequently ionised independently.

Early examples of instruments that could do this include CAART (Chemical Analysis

of aerosols in Real Time) (Allen & Gould, 1981), and PAMS (Particle Analysis by Mass

Spectrometry) (Sinha et al., 1982). The CAART design is shown in Figure 6.3, as this

employed the main basic principles of decoupled thermal vaporisation. The variations

on these principles that have been employed since will be discussed below.

Particles enter the instrument through a nozzle, where they are collimated into a beam

before passing through two differentially pumped chambers to reduce the gas-phase

pressure to less than 10

−6

torr. Once in the detection region, they enter a vaporiser-ioniser

cell, which serves to vaporise the particles, contain the neutral gas molecules until a

reasonable fraction of them can be ionised, and direct the ions to the mass spectrometer.

The vaporisation surface consists of a resistively heated tantalum wire in a coiled cone

shape, as shown in Figure 6.4, which presents a continuous surface to the particle beam

but is 50% open to the side. The wire is heated to around 1500 K and the particles are flash

vaporised upon impaction. The ionisation region consists of a cylindrical cage 2 cm in

length by 1 cm in diameter. Electrons are produced by a thoriated tungsten filament and

are accelerated into the cage by the potential difference (70 V by default), creating a cloud

of high-energy electrons in the centre. The energy transferred to the vapour particles

through random electron impacts (EI) is enough to cause ionisation, creating positive

ions that are extracted and focused by electric fields and analysed by the quadrupole mass

spectrometer. Normally, 70 eV is sufficient to break covalent bonds within the molecules,

so the ions detected from a single chemical species are frequently taken from a selection

of ionised fragments of the parent molecules. However, providing that the electrons

have the correct energy, these fragmentation patterns are very repeatable, even between

instruments (McLafferty & Turecek, 1993).

Mass Spectrometric Methods for Aerosol Composition Measurements 283

Ion lenses

Window

Collimator

Vic

B-A

Guage

Skimmer

Mechanical

pump

Mechanical

pump

Nozzle skimmer

B-A

Guage

Quadrupole

Ion multiplier

Beam-forming section

Inlet positioner

Inlet nozzle

Analysis

section

10 cms

Figure 6.3 Schematic of the CAART. This instrument employed the basic principles of decoupled thermal

vaporisation and ionisation used by all modern thermal aerosol mass spectrometers. The B-A gauges are

beta attenuation pressure gauges. The ‘VIC’ is the vaporiser-ioniser cell referred to in Figure 6.4. (Allen &

Gould, 1981, reprinted with permission from American Institute of Physics.)

Extractor

Ion lenses

Quadrupole

Aerosol beam

Collimator

Vaporizer

Grid

Ionizer

filament

Ion beam

axis

Figure 6.4 The vaporiser-ioniser cell and ion optics of the CAART. (Allen & Gould, 1981, reprinted

with permission from American Institute of Physics.)

284 Analytical Techniques for Atmospheric Measurement

Most decoupled thermal methods employ variations on this basic principle, whereby

particles are collected on a surface, vaporised using heat and ionised separately using a

continuous ion source before analysis. In instruments such as the AMS, the vaporiser

is contained within the ionisation region itself, to maximise the fraction of gas-phase

molecules that are ionised, thereby increasing the sensitivity. However, care must be taken

when using this configuration to ensure that the presence of the metallic surface does

not interfere with the electrical potential of the ionisation region or the field geometries

of the mass spectrometer. The filament also radiatively heats the vaporisation surface,

imposing a minimum vaporisation temperature and making the configuration unsuitable

for instruments such as the TDPBMS (see below).

There are two general methods of performing the vaporisation itself: one is to keep the

surface at a constant temperature, allowing the particles to vaporise as they hit the surface,

a technique used by the surface ionisation methods, CAART and the AMS amongst others.

As this is normally performed in ultra high vacuum with temperatures of hundreds of



the particles flash vaporise almost instantly. To prevent particles leaving the surface before

vaporisation is complete, in the AMS the conductively heated molybdenum vaporiser

was coated with layers of mesh. This was later replaced with an inverted cone vaporiser

made from porous tungsten. A second method is to collect the particles on a cooler

surface in bulk and subsequently heat it during a separate analysis stage. While this

does interrupt sampling, thereby reducing the time resolution of the instrument, it does

allow a large amount of material to be collected for the analysis runs. Examples of this

technique include the TDPBMS when operated in temperature-programmed mode, which

impacts the particles upon a cryogenically cooled surface (a conductively heated, v-shaped

molybdenum foil) during sampling. By carefully ramping the surface temperature at the

analysis stage and scanning the mass spectrometer continuously, it permits the study of

composition as a function of component volatility. Another example is the TD-CIMS. As

this collects the particles at atmospheric pressure, it cannot form a particle beam and use

impaction. Instead, the particles pass through a bi- or unipolar charger and precipitate

electrostatically onto the collection surface, which is in the form of a nichrome filament.

After collection, the filament is injected into a separate analysis chamber for vaporisation

and ionisation.

An advantage of 70 eV EI is that it is powerful enough to ionise almost all organic and

inorganic molecules, which is useful when trying to achieve mass closure of secondary

aerosol species. As the use of 70 eV electrons for ionisation is a very well established

convention in mass spectrometry, the mass spectral data produced (such as in the AMS

and TDPBMS) are directly comparable with those obtained using other laboratory instru-

ments. The established signatures of fragmentation of many different specific chemicals

are held on databases/atlases, compiled, for example, by NIST (2003), making the inter-

pretation of data much more straightforward. When analysing ambient particles, common

inorganic species such as nitrate and sulphate manifest themselves at specific m/z ratios

in the mass spectrum, making it easy to identify and quantify them (Allan et al. , 2003a).

Organic species tend to generate much more complex patterns and are therefore harder

to identify, especially when multiple species are present.

Thermal vaporisation is not limited to standard quadrupole mass spectrometry and

EI ionisation. For instance, the APCI-MS and TD-CIMS use chemical ionisation, the

former in conjunction with an ion trap and the latter with a triple quadrupole.

Mass Spectrometric Methods for Aerosol Composition Measurements 285

Chemical ionisation (CI) is a much ‘gentler’ method that charges the gas-phase

molecules by reaction with ions generated from a radioactive source or corona discharge

(e.g. N

−

and CO

−

). For more details of CI, see Chapter 5,

Sections 3.2.2 and 3.2.4. The resultant ions produced are chemically different from the

parent analyte molecules, a common result being the addition or removal of a proton,

but as there is no molecular fragmentation, it is highly suited to the study of the organic

fraction in particular. Chemical ionisation is a very selective form of ionisation that

can be tailored to specific chemical fractions by controlling the initial ions produced.

Another advantage of chemical ionisation is that the particle collection, vaporisation and

ionisation stages can be performed at higher pressures. The neutral gas molecules can be

removed by differential pumping as the ions are accelerated into the mass spectrometer,

rather than at the aerosol inlet before vaporisation. This means perturbations to the

particles through exposure to high vacuum can be eliminated, as too can the losses of

smaller particles associated with particle beam formation, allowing the online study of

the chemistry of ambient ultrafine particles as small as 6 nm with the TD-CIMS (Smith

et al., 2004).

The main advantage of the decoupled thermal methods so far discussed is the ability

to be quantitative, as it can be assumed that all (or a defined fraction) of the available

particulate matter is collected and vaporised. As the ionisation occurs in the gas phase,

the components do not interact once vaporised, allowing the measurements of individual

chemical fractions to be made independently of one another (Bley, 1988), in turn meaning

that accurate mass concentrations can be derived for many chemical species in a sample

simultaneously. With the AMS and TDPMS, their absolute quantitative capabilities are

dependent upon the particles reliably hitting the vaporisers, so they are in principle

susceptible to the biases associated with beam divergence. Their use of aerodynamic lenses

mitigates this over large ranges of sizes. Thermal vaporisation is not without its issues and

limitations. The act of heating imparts extra energy to the subject molecules, which can

change the fragmentation behaviour during ionisation when compared with purely gas-

phase mass spectrometry. In the more extreme cases, the heat may fragment the molecules

as they are vaporised, meaning the neutral gas-phase species produced are not chemically

the same as the parent condensed-phase components of the particles. This phenomenon

is most pronounced for complex organic molecules. A specific example is the behaviour

of di- and polycarboxylic acids in the AMS, a common group of chemical species in

atmospheric aerosols. Rather than vaporising in their molecular forms, they tend to

pyrolyse on the vaporiser surface and produce large numbers of simpler molecules such

as CO

, CO and H

O instead. However, these molecules can still be used as quantitative

markers for the parent organic species.

Another intrinsic limitation is that these instruments are blind to the refractory fraction

of the particles. While they can very effectively study components such as ammonium,

nitrate, sulphate and organics, the common components with much higher melting points

such as elemental carbon and crustal material cannot be vaporised and are therefore not

detected. Finally, the systems that use quadrupole or magnetic sector mass spectrometers

are limited to studying one m/z at any given time, which, coupled with the fact that

ions are generated on a continuous basis, means that complete composition data cannot

be derived for individual particles; they can only either generate averaged mass spectra

286 Analytical Techniques for Atmospheric Measurement

for an aerosol sample or information on a single m/z for a given particle. This makes

probing mixing states inherently difficult if not impossible.

Electron impact and CI are generally not well suited for conventional TOF mass

spectrometry, which requires a pulsed ion source. Coggiola et al. (2000) presented

a thermal vaporisation-EI aerosol mass spectrometer that held the ions produced in

a potential well before being ejected by a short ≈100 ns voltage pulse that provided

the start of the TOF in the mass spectrometer (Grix et al., 1989). The voltage pulses

were gated by an optical particle detector located in the particle beam. Unfortunately,

this design suffered from the problem that the number of ions coexisting in the ion

source region was intrinsically space-charge limited to around 10

. This meant that when

a particle’s composition was dominated by one particular component (e.g. water), the

other ion species were detected with drastically reduced sensitivity.

Future thermal instrument developments could benefit from Orthogonal extraction

Reflectron Time of Flight (ORTOF) mass spectrometers (Chen et al., 1999), a new type

of mass spectrometer designed for continuous ion sources. Pulses of ions are extracted

at right angles from a focused ion beam, itself extracted from the ionisation region,

which are analysed using a reflectron TOF configuration. As the spatial and velocity

distributions of the ions in the axis of extraction are intrinsically constrained, very high

m/z resolutions are possible, permitting a short TOF region, which in turn permits a

high pulse rate and near-continuous data collection. At the time of writing, Aerodyne

Research Inc. is offering an ORTOF mass spectrometer as an interchangeable alternative

to the standard quadrupole in the AMS. Initial results are promising, effectively giving

size-resolved data for all m/z channels simultaneously (Drewnick et al., 2005).

6.4.4.2 Single-step laser methods

Laser desorption and ionisation (LDI) is a different approach that uses a high-powered,

focused laser to deliver a large amount of energy to the particle while airborne to

simultaneously vaporise and ionise the components (Johnston, 2000; Noble & Prather,

1998). The earliest examples of this technique being used to study mobile particles include

those presented by Sinha et al. (1982) and McKeown et al. (1991). The original PALMS

instrument, as described by Murphy & Thomson (1995), is shown as an example in

Figure 6.5, as this has all the common elements of most current LDI designs, although

many new innovations have occurred since and will be discussed below. The aerosols

enter the instrument through a capillary tube and pass through differentially pumped

chambers before entering a high vacuum region, where they pass through the focal

point of the desorption laser. A KrF eximer laser was used, which delivers UV light at

a wavelength of 248 nm. As the particles tend to be fast moving after introduction into

the vacuum of the instrument, high amounts of energy must be transferred in very short

spaces of time to hit and vaporise them successfully. The most effective way of providing

enough laser energy is through the Q-switching of the desorption laser, which delivers

a short, efficient and powerful pulse orthogonal to the particle beam. To trigger these

laser pulses, an additional, lower-powered but continuous laser beam (He:Ne laser at

633 nm) in conjunction with an avalanche photodiode was used along the same beam

path to detect the particles as they entered the instrument. The scattering events as seen

by the photodiode were used to trigger the Q-switch of the desorption laser. The pulse

Mass Spectrometric Methods for Aerosol Composition Measurements 287

Excimer laser

He–Ne laser

Trigger

Differentially pumped nozzle

Logarithmic

amplifiers

100

MHz

digitizers

Turbopump

Avalanche photodiode

for scattered light

Two stage ion detector

ZAP!

Gas with particles

Figure 6.5 Basic schematic of the original PALMS instrument, which incorporates most of the funda-

mental principles of laser-based aerosol mass spectrometry. (Murphy & Thomson, 1995, reprinted with

permission from American Association of Aerosol Research, published by Elsevier Science Inc.)

energy delivered by the desorption laser was sufficient to both vaporise and ionise the

components of the particles. The use of a pulsed ion source has the added advantage

of being able to act as the start gate in a TOF mass spectrometer, which is used to

analyse ions of all m/z produced from the individual particles simultaneously. The mass

spectrometer could be operated in either positive or negative ion mode. The PSPF method

was also used, acting on the ions of an m/z between 24 and 350, and gave the instrument

an m/m FWHM resolution of >200.

There have been many variations on this principle over the years, employing many

different alternative technologies. For instance, other designs have used Ar

(514 or

488 nm) (e.g. Salt et al., 1996), He:Cd (442 nm) (e.g. Carson et al., 1995), frequency

doubled Nd:YAG or Nd:YVO

(532 nm) (e.g. Thomson et al., 2000) or diode lasers (e.g.

Woods et al., 2001) in the particle detection systems, and photomultipliers rather than

photodiodes are now used in most implementations to detect the scattered light. In

addition to the power consumption, cost and portability, the choice of laser technology

also affects the detection capabilities, as was discussed in the sizing Section 6.4.3.2,

together with aerodynamic sizing of the particles using two detection lasers. Particles

smaller than the lower limit of the detection and triggering system can be studied

by firing the desorption laser arbitrarily, without triggering. This results in a lack of

direct particle sizing information, but a DMA placed in series or a size-selective inlet

as described in the sizing section can be used to generate size-resolved data (e.g. Ge

et al., 1998; Kane et al., 2001). As there is no triggering, the hit rate is significantly

reduced but this is mitigated in the RSMS II, where the desorption laser beam and the

particle beam are counter-propagating, as opposed to the orthogonal alignment used in

other instruments, thereby increasing the coincident volume of the laser and particle

beams.