Ellis A.M., Feher M., Wright T. Electronic and Photoelectron Spectroscopy: Fundamentals and Case Studies

Подождите немного. Документ загружается.

10 Lasers

Laser

output

Pump

laser

BS1

BS2

OCODCBEE

PDC ADC

Laser cavity

Figure 10.5 Optical arrangement of a tunable pulsed dye laser. The dye laser is pumped by pulsed

radiation from another laser. Abbreviations are as follows: BS, beamsplitter; M, mirror; CL, cylindrical

lens; OC, output coupler (end mirror); ODC, oscillator dye cell; PDC, preampliﬁer dye cell; ADC,

ampliﬁer dye cell; BE, beam expander; E, intracavity etalon (optional); DG, diffraction grating. The

preampliﬁer and ampliﬁer dye cells are used to increase the intensity of the dye laser beam produced

in the laser cavity. This ampliﬁcation process can increase the intensity by more than two orders of

magnitude.

The output wavelength can be extended outside of the traditional dye operating ranges

using non-linear optical techniques. The most commonly used is frequency doubling, in

which the dye laser fundamental is passed through a suitable crystal to generate the second

harmonic (v

out

= 2v

). This crystal must possess the correct non-linear optical properties,

as well as being able to withstand very high laser intensities. -barium borate is one of the

best materials currently available, with KDP as a cheaper alternative for some wavelength

ranges. Efﬁcient harmonic generation requires correct phase matching of the fundamental

and higher harmonic beams. Phase matching is the process by which the refractive indices

of the input and output beams are equalized, and this requires a speciﬁc orientation of

the crystal relative to the incoming laser beam. Frequency doubling allows coverage of

the whole of the near-ultraviolet (205–400 nm), and more advanced techniques can extend

the wavelength into the vacuum ultraviolet region (<200 nm). At the long wavelength end,

tunable radiation beyond 1 m can be generated using difference frequency generation [5].

10.7 Titanium:sapphire laser

The Ti:sapphire laser is a tunable solid state laser based on transitions of Ti

ions doped in

a sapphire host. The crystalline lattice broadens the electronic energy levels of Ti

to such

an extent that tunability far exceeding that of a single laser dye is achieved. However, the

Ti:sapphire laser is not really a competitor to the dye laser since their tunability ranges only

partially overlap. One of the strengths of the Ti:sapphire laser is that much of its tunability

86 Experimental techniques

range, 660–1180 nm, is in a difﬁcult region for dye lasers. It also possesses better frequency

stability and a narrower linewidth than dye lasers. Output in the near-ultraviolet and blue

regions is possible by frequency doubling the fundamental output.

10.8 Optical parametric oscillators

These are tunable laser sources that offer the promise of eventually superseding dye lasers.

Tunability in optical parametric oscillators (OPOs) is achieved by non-linear optical pro-

cessing of a single input (pump) beam. It is useful to think of this as the opposite of frequency

doubling in a non-linear crystal. In essence, a single high intensity laser beam is passed

through the non-linear crystal. The input beam can ‘split’ into two output beams, one known

as the signal and the other the idler, such that v

= v

signal

+ v

idler

. The exact reverse of fre-

quency doubling would correspond to equal idler and signal frequencies. However, any

combination of v

signal

and v

idler

is, in principle, achievable providing the sum equals v

, and

a particular combination can be ampliﬁed if the mixing process is carried out in a tunable

laser cavity. By combining the tunability of a diffraction grating in the laser cavity, and the

orientation of the crystal for optimum phase matching, efﬁcient generation of tunable radi-

ation over a wide spectral range is possible. Commercial OPOs are available which operate

over the whole of the visible region and these can be extended into the near-ultraviolet by

frequency doubling.

References

1. Principles of Lasers,O.Svelto, New York, Plenum Publishing Corporation, 1998.

2. Lasers,A.E.Siegman, Mill Valley, California, University Science Books, 1986.

3. Laser Fundamentals,W.T.Silvast, Cambridge, Cambridge University Press, 1996.

4. R. H. Lipson, S. S. Dimov, P. Wang, Y. J. Shi, D. M. Maxo, X. K. Hu, and J. Vanstone,

Instrum. Sci. Technol. 28 (2000) 85.

5. A. S. Pine, J. Opt. Soc. Am. 70 (1980) 1568.

Optical spectroscopy

Consider a beam of light of intensity I

incident on some absorbing sample. Providing only

a small fraction of the light is absorbed,

and assuming that losses caused by light scattering

are negligible, the transmitted light intensity, I,isgoverned by the familiar Beer–Lambert

law,

A = log





= ε(ν)cl (11.1)

where A is known as the absorbance. The absorbance is dependent upon the concentration

of absorbing species, c, the optical path length, l (distance travelled by the light through

the sample), and the molar absorption coefﬁcient, ε. The molar absorption coefﬁcient is a

measure of the intrinsic absorbing power of the sample and is frequency dependent, which

is why it has been written as ε(ν). It is customary to give c in units of mol dm

−3

and l in

cm, and so ε is often quoted in the rather strange mixture of units dm

mol

−1

.Asone

might expect, ε is related to the Einstein B coefﬁcient introduced in Chapter 7.

The absorbance is an important quantity because it is directly proportional to the con-

centration. If monochromatic radiation is passed through a material of known thickness

and known molar absorption coefﬁcient, the concentration of the absorbing species can

be determined from a measurement of the absorbance. This is a widely used feature of

absorption spectroscopy.

11.1 Conventional absorption/emission spectroscopy

A schematic of an absorption spectrometer is shown in Figure 11.1. Ideally, the light source is

continuous over the wavelength region of interest and shows no major variations in intensity.

Resistively heated ﬁlaments are good sources of near-continuum light. One example is a

white-hot tungsten ﬁlament, which will cover the whole of the visible and parts of the

near-ultraviolet and near-infrared. A wavelength selector is central to the spectrometer and

is usually a monochromator built around a diffraction grating, thus allowing tunability.

In order to obtain a spectrum, light intensity transmitted through the monochromator is

If the fraction of light absorbed is large, then the light intensity varies strongly as the sample is traversed and the

Beer–Lambert law no longer holds.

88 Experimental techniques

Sample

Light

source

Entrance

slit

Exit

slit

Monochromator

Lenses

Diffraction

grating

Figure 11.1 Schematic of a conventional grating-based absorption spectrometer. The monochromator

is of the Czerny–Turner type in which the entrance and exit slits are placed at the focal points of curved

mirrors.

Excitation

source

Sample

Mono-

chromator

Detector

+ PC

Figure 11.2 Block diagram of a standard emission spectrometer.

measured as a function of wavelength. The light intensity is measured by a photomultiplier

tube (PMT), a photodiode, or some other light-detecting device.

In an emission spectrometer, the sample must be driven up to excited quantum states in

order for emission to occur. This is normally achieved by an electrical discharge, although

broadband optical excitation is also possible. As indicated in Figure 11.2, the monochroma-

tor is now used to select a speciﬁc emission wavelength from the sample and the intensity at

this wavelength is measured by imaging the light onto a detector such as a PMT. An emission

spectrum is obtained by recording the PMT signal as a function of emission wavelength.

Monochromators such as that shown in Figure 11.1 have both entrance and exit slits.

These are crucial to the wavelength selection process. Narrowing the entrance and exit slits

can improve the spectral resolution, but it does so at the expense of sensitivity because of

the reduced light throughput. Improvements can be made that make more efﬁcient use of the

available light. For example, the exit slits in an emission spectrometer can be dispensed with

if a multichannel detector is available. Examples are photodiode arrays and charge-coupled

devices (CCDs). These measure the light intensity as a function of position on the detector

surface and so are able to record a large portion of the spectrum simultaneously. Another

alternative is Fourier transform spectroscopy, which does away with both the entrance and

exits slits as well as the diffraction grating. Fourier transform spectroscopy is described

later in this chapter.

11 Optical spectroscopy

Photomultiplier

tube

Collection lens

Dye laser beam

Filter

Figure 11.3 Experimental arrangement for laser-induced ﬂuorescence spectroscopy. Fluorescence

radiates in all directions, a portion of which is collected by the lens and transmitted to the detector, a

photomultiplier tube. The ﬁlter, which is used to reduce the amount of scattered laser light reaching

the detector, is optional. The arrangement shown is for laser excitation spectroscopy. For dispersed

ﬂuorescence spectroscopy the ﬁlter is replaced with a scanning monochromator.

11.2 Laser-induced ﬂuorescence (LIF) spectroscopy

This is one of the principal techniques for studying electronic transitions of both neutral

molecules and molecular ions at high sensitivity and at high resolution. In LIF spectroscopy

an electronic transition of the molecule is excited using a tunable laser and any ﬂuorescence

generated is monitored. There are two complementary methods that parallel, respectively,

conventional absorption and emission spectroscopy.

Suppose the wavelength of a tunable laser is scanned through the electronic absorption

band of a molecule. Absorption will occur at resonant wavelengths and could be monitored

by measuring the intensity of the transmitted laser beam. The high intensity of a laser

can greatly increase the probability of absorption compared with low intensity non-laser

light sources and thus it might be thought that laser absorption spectroscopy would be

very sensitive. Unfortunately, this is not the case because the fractional absorption by a

sample will still normally be very low. Thus a small change in intensity is superimposed

on a large background signal. When ﬂuctuations in intensity of the laser beam and noise

from the light detector are factored in, this approach turns out to have a very limited

sensitivity.

However, instead of measuring absorption directly it can be monitored indirectly by

detecting ﬂuorescence from the excited electronic state. The experimental arrangement is

remarkably simple, and is outlined in Figure 11.3.Atunable laser is passed through the

sample and any ﬂuorescence produced is collected off-axis, usually at right angles to the

laser beam, by a collection lens. The light is then detected by a photosensitive device, most

90 Experimental techniques

usually a PMT. PMTs have phenomenal sensitivities and are even capable of detecting single

photons in some cases. When the laser is off-resonance, no ﬂuorescence will be produced,

and therefore the PMT registers no signal. However, at resonant wavelengths ﬂuorescence

is possible and so absorption can be registered by detecting emission from the excited state.

This is the basic idea of laser excitation spectroscopy,inwhich a spectrum is obtained by

measuring the ﬂuorescence intensity as a function of laser wavelength.

There are several important points to note about laser excitation spectroscopy. First, while

there is a clear similarity between laser excitation spectroscopy and absorption spectroscopy,

there is also an important difference. The intensity of peaks in a laser excitation spectrum

depends on both the absorbance of the sample and the ﬂuorescence quantum yield of the

excited state. The ﬂuorescence quantum yield is deﬁned as



rate of photon emission by excited state

rate of photon absorption

(11.2)

A ﬂuorescence quantum yield of unity implies that all molecules excited to the upper

electronic state relax via photon emission. However, competition from other decay routes

(see below) may not only lower 

,but may also cause it to change from one excited state

level to another. As a result absorption and ﬂuorescence excitation spectra may look very

different.

The high sensitivity of LIF spectroscopy arises from the low background signal received

by the PMT at off-resonance laser wavelengths. Even though any ﬂuorescence produced

may be very small, it is easily detected by the PMT and therefore if the off-resonance

signal is much smaller still then an extremely high signal-to-noise ratio can be achieved. In

practice the off-resonance signal is never zero. The principal cause is scattered light from

the laser. This can be minimized by keeping potential scattering sites out of the path of the

laser. Furthermore, scattered laser light can be virtually eliminated if at least a portion of

the ﬂuorescence is at longer wavelengths than the laser. If this condition is satisﬁed, and

it often is for many molecules, then an optical ﬁlter, which will only transmit wavelengths

longer than that of the laser, can be inserted in front of the PMT.

In laser excitation spectroscopy the ﬂuorescence serves only as a means of detecting the

absorption process. However, the ﬂuorescence itself clearly contains spectroscopic infor-

mation since it arises from emission to lower energy levels. If the emission is dispersed in

a monochromator, the spectrum obtained will be the emission spectrum originating from

a speciﬁc (laser-excited) upper state. This type of spectroscopy goes by several names,

including dispersed ﬂuorescence spectroscopy, laser-excited emission spectroscopy, and

single vibronic level ﬂuorescence spectroscopy;wewill use the ﬁrst of these throughout

this text.

In laser excitationspectroscopythe resolution is often limited by the linewidth of the laser.

For pulsed dye lasers, linewidths of ∼0.03 cm

−1

can be obtained relatively straightforwardly.

If narrower linewidth lasers are used, such as CW dye lasers or specialized pulsed dye lasers,

other factors may begin to limit the resolution, such as Doppler broadening. If steps are taken

to minimize Doppler broadening, a resolution of better than 0.001 cm

−1

can be attained.

With such a high resolution, rotationally resolved electronic spectra of quite large molecules

can be tackled.

11 Optical spectroscopy

A + B

A + B*

AB*

Repulsive

state

Figure 11.4 Predissociation caused by the crossing of two potential energy curves. Notice that only

those energy levels above the crossing point can undergo predissociation.

In dispersed ﬂuorescence spectroscopy, the use of a scanning monochromator is normally

the principal factor limiting the resolution. Even a large monochromator may only have a

resolution of about 1 cm

−1

. Thus dispersed ﬂuorescence spectroscopy is normally concerned

with vibrationally resolved emission spectra.

The principal disadvantage of LIF is the need for a ﬂuorescent excited state. Fast non-

radiative decay routes may reduce the ﬂuorescence quantum yield to zero and in these cases

LIF cannot be used. An example of a non-radiative decay process is predissociation, which is

illustrated in Figure 11.4. Predissociation results from a crossing of potential energy surfaces

of two excited electronic states, one of which is repulsive (dissociative). If the molecule

is excited to the bound potential energy curve, it may hop over onto the repulsive curve at

the crossing point and will then undergo dissociation. If the probability of predissociation

is not too high, there may still be sufﬁcient ﬂuorescence for LIF detection. In such cases,

the occurrence of predissociation manifests itself by a broadening of spectral lines, since

the effect of predissociation is to decrease the lifetime of the level and hence increase the

lifetime broadening.

Depopulation mechanisms such as predissociation are particularly troublesome for large

molecules because of their high density of rovibrational energy levels. Usually the coupling

mechanism, the process which actually brings about the interaction between the electronic

states, will be restricted by symmetry in the same way that symmetry restricts electric dipole

transitions. However, the importance of symmetry restrictions decreases as the overall point

group symmetry of a molecule is lowered, and large molecules tend to have low symmetry.

It is for these reasons that LIF is a particularly powerful technique for investigating small

molecules, but is more limited in scope for large molecules.

92 Experimental techniques

11.3 Cavity ringdown (CRD) laser absorption spectroscopy

Direct laser absorption electronic spectroscopy is appealing for several reasons. First, the

narrow linewidths of lasers can be exploited. Second, it does not rely on the occurrence

of a secondary process for detection, as in LIF spectroscopy. Third, the absorbance can

be directly related to the concentration of the absorbing species, thus allowing absolute

concentration measurements to be made.

However, as discussed in the previous section,

when done in the conventional manner laser absorption spectroscopy is a low-sensitivity

technique.

Cavity ringdown spectroscopy is a form of laser absorption spectroscopy in which the

absorbance is determined but in a rather ingenious manner. It is a relatively new technique,

ﬁrst appearing in 1988, but is based on a simple idea.

Suppose a gas is placed between two highly reﬂecting mirrors which act as an optical

cavity. If a pulse of laser light is injected into the cavity, as shown in Figure 11.5, then

laser light will reﬂect backwards and forwards and, if the spacing between the mirrors is

relatively small, interference will occur as a consequence of the coherence of the laser

beam. However, the coherence of a laser beam is restricted to a ﬁnite distance known as

the coherence length.

The ﬁnite coherence length is brought about by uncertainty in the

frequency of the light, which in turn is a result of the non-zero linewidth. The coherence

length, l

coh

,isgivenby

coh

ν

(11.3)

where ν is the linewidth (FWHM) of the laser. For typical pulsed dye lasers without

intracavity etalons, the linewidth is 0.2 cm

−1

and so equation (11.3) yields l

coh

= 5 cm.

Consequently, if the mirror separation is signiﬁcantly larger than 5 cm, interference is not

an issue; this is the starting point for cavity ringdown spectroscopy.

If the mirrors are able to transmit a small proportion of the incident light, then each time

the laser light pulse impinges on a mirror some is lost from the cavity. Gradually, at a rate

determined by the mirror reﬂectivities, the intensity of the light trapped within the cavity

will decay to zero. In fact the decay is exponential, and the time taken for the intensity

to decay to 1/eofits initial value is known as the ringdown time.Itcan be measured by

placing a sensitive light detector, usually a photomultiplier tube, behind one of the mirrors,

as shown in Figure 11.5.

Now suppose that an absorbing sample is placed inside the cavity. Absorption of the

laser light by the sample will accelerate the ringdown process, resulting in a faster ringdown

time. The larger the absorbance, the shorter the ringdown time. Hence it is possible to record

something akin to an absorption spectrum by measuring the change in ringdown time as

a function of laser wavelength. In fact there is a simple and exact relationship linking

It is difﬁcult to deduce absolute concentrations of an absorbing species from LIF spectroscopy, although changes

in relative concentrations can easily be measured.

The coherence length is a measure of the distance over which the phase relationships between the constituent waves

in a light source are maintained. For optical path differences exceeding this difference, the phase relationships are

lost and so interference effects become negligible.

11 Optical spectroscopy

PMT

Pulsed laser

Sample

Oscilloscope

Figure 11.5 Experimental arrangement for pulsed cavity ringdown laser absorption spectroscopy.

The cavity is deﬁned by the two plano-concave mirrors. Concave mirrors are preferred over plane

mirrors because the former can produce a so-called stable optical cavity, making it easier to ‘trap’ the

radiation within the cavity.

ringdown time to the sample absorbance, A,which takes the form

A =



−



(11.4)

where L is the cavity length, c is the speed of light, and τ

and τ are the ringdown times in

the absence and presence of an absorber, respectively. Thus the cavity ringdown spectrum

can easily be converted into a conventional absorption spectrum.

To achieve high absorption sensitivity, high mirror reﬂectivities are required since these

lengthen the ringdown times and therefore make it easier to observe small changes. Mirrors

with reﬂectivities better than 99.995% are available in the visible and near-ultraviolet. Of

course, one problem with such high mirror reﬂectivities is that only a tiny proportion of the

light from the laser is injected into the cavity in the ﬁrst place since the laser beam enters

through one of the end mirrors! However, with sensitive detectors such as PMTs this does

not cause a signiﬁcant problem.

The discussion so far has focussed on cavity ringdown using pulsed lasers. However, it is

also possible to record CRD spectra with continuous lasers (CW-CRD). Typically, a narrow

linewidth tunable diode laser is employed as the light source. It is still necessary to inject a

pulse of light into the cavity. One way this can be achieved is to scan the cavity length by

mounting one of the end mirrors on a piezoelectric transducer. If the cavity length does not

match one of the longitudinal modes of the cavity, no signiﬁcant light can be injected. This

restriction is normally unimportant in pulsed laser CRD because the relatively broad laser

linewidths mean that there is always some radiation that matches longitudinal modes of the

cavity. As the mirror is moved in CW-CRD, at some stage the standing wave condition will

be met and light will be injected into the cavity. An electro-optical switch, known as an

94 Experimental techniques

acousto-optical modulator, is then used to block the laser beam so that a pulse of laser light

remains in the cavity. A ringdown proﬁle is then measured in the normal manner.

CW-CRD is growing in importance. One reason for this is that it is capable of much

higher spectral resolution than pulsed laser CRD. Much higher pulse repetition rates can

also be employed giving improved detection sensitivity [1].

11.4 Resonance-enhanced multiphoton ionization (REMPI) spectroscopy

Highly excited electronic states can be studied by vacuum ultraviolet (VUV) absorption

spectroscopy. One of the problems in working with VUV light sources is the low resolution

achieved in this region. Although tunable laser radiation can be obtained in parts of the

VUV, this is not as routine to generate, nor is it as cheap, as visible and near-ultraviolet

laser sources. Fortunately, many VUV transitions can be accessed by multiphoton transitions

using visible or near-ultraviolet laser light. Resonance-enhanced multiphoton ionization

(REMPI) spectroscopy is a particularly powerful and widely used example of a multiphoton

spectroscopic technique.

REMPI is a two-stage process. In the ﬁrst step, molecules are promoted to an excited

electronic state by the absorption of one or more photons. It may at ﬁrst sight seem strange

to suggest that more than one photon can be absorbed in a spectroscopic transition, since

we normally regard them as single-photon resonant processes. However, there is nothing

intrinsically impossible in using two or more photons of lower energy to achieve the same

task, providing (i) their combined energy satisﬁes the resonance condition, e.g. for two

photons having the same frequency, E

− E

= 2hv, and (ii) all selection rules are satisﬁed

(see later).

The principal reason why multiphoton transitions are not normally considered is that such

processes are extremely improbable at normal light intensities. The photons must arrive at

the molecule at virtually the same instant in time in order to be simultaneously absorbed.

With ordinary light sources, such as lamps or low intensity lasers, this hardly ever happens.

However, if extremely high light intensities are employed, as is the case with powerful

pulsed lasers, then multiphoton transition probabilities need no longer be negligible. Even

so, it is easy to appreciate that the probability will rapidly decrease as the number of photons

to be absorbed increases.

Once the molecule has reached the excited electronic state by absorption of one or more

photons, it may absorb one or more further photons to climb abovethe ionization limit. This is

a REMPI process. Compare this with direct (non-resonant) multiphoton ionization. Clearly

REMPI and direct (non-resonant) multiphoton ionization have the same overall photon

order, i.e. the same total number of photons is absorbed. However, in REMPI the ionization

is achieved by two steps of lower photon order, each with a much higher probability (many

orders of magnitude) than the non-resonant multiphoton ionization process. In other words,

the ionization probability is dramatically increased by breaking the ionization process down

into two separate, sequential steps.

This suggests a means of detecting electronic transitions. If the laser is tuned to a wave-

length that is not resonant with an energy level in the excited electronic state manifold

of the neutral molecule, then ionization is only possible by the non-resonant route, and