Luminescence Overview

Lesˇeticky´ L (1985) Isotope separation by chromatographic

methods [in Czech]. Radioisotopy 26: 113–126.

Lockley WJS (1989) Regiochemical differences in the

isotopic

fractionation

of deuterated

benzoic

acid

isotopomers by reversed phase HPLC. Journal of Chrom-
atography 483: 413–418.

Simon H and Palm (1966) Isotope effects in organic chem-

istry and biochemistry [in German]. Angewandte Chemie
78: 993–1007.

Tanaka N and Thornton (1977) Structural and isotopic

effects in hydrophobic binding measured by HPLC. A
stable and highly precise model for hydrophobic inter-
action in biomembranes. Journal of the American Chem-
ical Society 99: 7300–7306.

Wade D (1999) Deuterium isotope effects on noncovalent

interaction

between

molecules.

Chemico-Biological

Interactions 117: 191–217.

LOW ENERGY ELECTRON DIFFRACTION

See

SURFACE ANALYSIS: Low Energy Electron Diffraction

LUMINESCENCE

Contents

Overview

Solid Phase

Overview

N W Barnett and P S Francis

, Deakin University,

Geelong, VIC, Australia

Introduction

The observation and investigation of luminescent
phenomena has a long and delightful history and
some of the more important milestones are briefly
described here. Naturally occurring luminescence,
such as ﬁreﬂies, St Elmo’s ﬁre, and shining ﬂesh has
fascinated humans since the dawn of time, with an
early record appearing in Chinese literature between
1500 and 1000

. A millennium or so later, Aristotle

(384–322

) astutely noted that these emanations

were produced without heat. Subsequently, Caius
Plinius Secundus (Pliny the Elder,

23–79) de-

scribed, in detail, a number of luminous organisms.
Notwithstanding the revolutionary nature of these
observations of bioluminescence, made by two such
great philosophers, a rigorous scientific approach to
the subject was not taken until the mid-sixteenth
century. This study culminated in the publication, in
1555, of a book by Conrad Gesner (1516–1565)
concerned solely with luminescence. However, Sir

Robert Boyle (1627–1691), son of the Earl of Cork,
who came to be known as the father of analytical
chemistry, ﬁrst categorized the essential differences
between incandescence and luminescence in 1668.
The following year saw the ﬁrst artiﬁcial lumines-
cence, which accompanied Hennig Brandt’s dis-
covery and isolation of elemental phosphorus.
More than three centuries passed before the emana-
tions from white phosphorus were correctly charac-
terized as chemiluminescence. Cambridge professor
and president of the Royal Society, Sir George Gab-
riel Stokes (1819–1903), was the ﬁrst to characterize
the bichromatic nature of crystal ﬂuospar as a true
emission (actually phosphorescence) in 1845. He
also coined the term ﬂuorescence as an analogy to
opalescence and noted that in the production of ﬂu-
orescence the absorption of shorter wavelengths re-
sulted in emission at longer wavelengths (Stokes’
law). In 1877, Raziszewski was probably the ﬁrst to
observe chemiluminescence from synthetically pro-
duced organic compounds during the preparation of
lophine (2,4,5-triphenylimidazole) from hydrobenz-
amide.

The earliest usage of the word luminescence is

credited to Eilhardt Wiedemann (1852–1928), who
in 1888 used the term to describe the emission of
light that was not the result of a rise in temperature
(‘cold light’). He deﬁned six classes of luminescence
by the source of energy that stimulated the emission.

LUMINESCENCE

/ Overview

305

The present number of classes of luminescent phe-
nomena is somewhere around 20; however, the
boundaries between some of these classiﬁcations is
more than a little blurred. In stark contrast to in-
candescence, luminescent phenomena are not only
cold but are also of relatively low intensity, for
example:

the mysterious St Elmo’s Fire and aurora australis
(electro- or radioluminescence);

emission from rubbing or shattering crystals
(triboluminescence);

the ephemeral blue haze from a glass of gin and
tonic in the sunlight (ﬂuorescence);

the enchanting emanations from ‘lightening bugs’
(ﬁreﬂies) on a summer evening (chemilumines-
cence).

Luminescence is generally less intense than incan-
descence, but it often emanates from extremely small
amounts of matter, which has beneﬁcial implications
for analytical science. Nevertheless, the utilization of
luminescence for analysis is quite a recent innovat-
ion. The following commentary describes the fun-
damental spectroscopic and chemical principles un-
derlying luminescence in relation to its application in
analytical science. As other articles will deal with
atomic spectroscopy, this discussion will be restricted
to analytical molecular luminescence spectroscopy
including ﬂuorescence, phosphorescence, and chemi-
luminescence (bioluminescence being a special case
of chemiluminescence).

By monitoring the relative intensity of lumines-

cence as a function of concentration it is possible to
quantitatively determine (often at trace levels) a
range of inorganic and organic analytes. Contem-
porary analytical texts reveal that ﬂuorimetric
methodology is far more commonplace than either
phosphorescence or chemiluminescence. The re-
search literature, however, indicates an increased ac-
ceptance of the latter two spectroscopies particularly
when combined with either organized media or ﬂow
analysis, respectively. Luminescence methodology
generally offers superior selectivity, detectability,
and linear calibration range compared to that at-
tainable with absorption spectrometry. Unfortunate-
ly, there are a relatively limited number of molecules
that will exhibit luminescence thus restricting its ap-
plicability compared to absorption techniques. Al-
though all forms of luminescence have a common
quantum mechanical basis for emission, the route to
the excited state deﬁnes their respective class. Given
that the theory of photoluminescence is better un-
derstood than that of the various related phenomena,
a basic synopsis of ﬂuorescence and phosphorescence

is presented prior to an introduction to chemically
induced luminescence.

Photoluminescence

General Principles

Photoluminescence occurs as the result of electronic
excitation within a molecule brought about by
the absorption of a photon. The question therefore
arises, as the vast majority of molecules do not lu-
minescence, what are the quantum mechanisms that
give rise to ﬂuorescence and phosphorescence and
how does molecular structure aid or inhibit these
processes? Figure 1 is a Jabłon´ski diagram, named in
honor of the Ukrainian born physicist, Professor
Alexander Jabłon´ski (1898–1980), who is considered
by many to be the father of ﬂuorescence spectros-
copy. These schematic energy level diagrams are
useful tools for the understanding of molecular elec-
tronic excitation and deexcitation leading to photon
emission. The ﬁrst observation to be made from
Figure 1 is that ﬂuorescence emission (F) results from
an electron falling from the lowest vibrational level
of an electronically excited state (S

) to any of the

vibrational excited levels of the ground state (S

Phosphorescence emission (P) is a similar process
except that the upper electronic state is a triplet (T

The terms singlet and triplet refer to the relative spins
of the electrons in the ground and excited states.
When the spins are paired (antiparallel) the upper
level is termed a singlet, and when the spins are
unpaired (parallel) a triplet state exists. The nomen-
clature arises from the observed multiplicity in spec-
tra measured under the inﬂuence of a magnetic ﬁeld.
It is noteworthy that T

states are less energetic than

; this is a direct consequence of electrons with par-

allel spins being further apart thus exhibiting less
mutual repulsion (spin correlation, Hund’s rule).
Hence, for a given compound the phosphorescence
emission will occur at longer wavelengths than the
ﬂuorescence. The photo-induced promotion of an
electron from S

directly to T

has not been shown in

Figure 1 as the simultaneous change of molecular
orbital and electronic spin has a very low probability
of occurrence. In fact, such transitions are often re-
ferred to as ‘forbidden’ by the spin selection rule.

As most compounds are not luminescent, this im-

plies that there must be more efﬁcient alternative
deexcitation mechanisms available to molecules in S

or T

states to return to S

other than ejection of a

photon. To understand the nature of the spec-
troscopic processes shown in Figure 1, their relative
rates of occurrence and the inﬂuences on the various
pathways must be considered.

306

LUMINESCENCE

/ Overview

At room temperature in solution, we may assume

that all molecules will be in the lowest vibrational
level (v

¼ 0) of the ground state (S

), thus absorption

will originate solely from v

¼ 0. According to the

Franck–Condon principle, the initial step (absorption
of a photon) is extremely rapid, requiring only
10

–10

s. Immediately after absorption; either

of the excited singlets (S

or S

) will undergo vibra-

tional relaxation to their lowest vibrational level
(v

¼ 0). This process is extremely efﬁcient with all the

excess vibrational energy being transferred to the
solvent molecules as heat in around 10

–10

In the case of excitation to S

, the process of internal

conversion to the upper vibrational levels of S

(IC in

Figure 1) is rapid and effective. Thus luminescence in
solution will usually result from the lowest vibrational
level of the S

excited state. This is not always the case

in gas-phase luminescence where collisional relaxa-
tion is far less likely. The ﬂuorescence emission life-
time (t) is of similar magnitude to the mean lifetime of
the excited singlet state (10

–10

s), the latter

being inversely proportional to the molar absorptivity.

As the probability of ﬁnding an excited state molecule
at time t after removal of the excitation source is
e

t/t

, then the relationship between luminescence in-

tensity and the mean excited state lifetime is:

¼ I

t=t

½1

where I is the luminescence intensity at time t and
I

is the maximum luminescence intensity during

excitation. In practical terms the difference in t val-
ues for various analytes can provide extra selectivity
when using time-resolved luminescence spectroscopy.
Figure 1 also illustrates that the ﬂuorescence emis-
sion spectrum is shifted to longer wavelengths com-
pared to that of the absorption (excitation) spectrum.
This is the result of vibrational relaxation both after
excitation and after ﬂuorescence and is known as
Stokes’ shift. From the zeroth vibrational level of the
lowest excited singlet (n

¼ 0, S

) it is also possible for

radiationless deexcitation to convert all the energy to
heat in preference to light. This is termed internal
conversion, the quantum mechanical basis for which
is somewhat vague. The internal conversion from S

″

′

ISC

EC/IC

Singlet states

Triplet states

Figure 1

A Jabłon´ski diagram for a hypothetical luminescent molecule, where A is absorption, F is ﬂuorescence, P is phospho-

rescence, VR is vibrational relaxation, IC is internal conversion, EC is external conversion, ISC is intersystem crossing, v

and v

are

vibrational levels associated with each electronic state, S

is the electronic ground state, S

and S

are excited singlet states and T

and T

are excited triplet states.

LUMINESCENCE

/ Overview

307

to S

for aliphatic compounds can be rationalized by

the overlapping of the upper vibrational levels of S

with those of S

. In this case relaxation is rapid and

efﬁcient and explains why aliphatic molecules rarely
luminesce. Internal conversion also occurs between
excited electronic states (as shown in Figure 1) due to
the overlapping vibrational levels of S

and S

. At the

crossover point the potential energies of the two
excited states are equal and as the efﬁciency of vibra-
tional relaxation is far greater than emission of a
photon, internal conversion to the zeroth vibrational
level of S

occurs rather than ﬂuorescence from S

Therefore, ﬂuorescence from anything but the S

state is rare; the blue hydrocarbon azulene (isomeric
with naphthalene) is the most well-known exception.
In large molecules internal conversion may also re-
sult in predissociation (bond cleavage), when the
electron moves from an excited state to a high-
vibrational level of a lower electronic state. In this
situation the vibrational energy is sufﬁcient to cause
cleavage of bonds in the unstable excited state.
Together with internal conversions, electronically
excited states can be deactivated by virtue of their
interactions with solvent molecules of other concom-
itant species present. Such pathways are termed ex-
ternal conversions. As a general rule, those environ-
mental parameters that lower the probability of
collisional deexcitation (such as lower temperature,
increased viscosity, and organized media) tend also to
enhance luminescence.

Whilst the probability of the direct population of

from S

by absorption is virtually zero, a kinet-

ically efﬁcient pathway exists from the S

state in a

number of molecules. The mechanism is known as
intersystem crossing and it can be considered as a
spin-dependent internal conversion. It should be
borne in mind that singlet–triplet transitions are ap-
proximately a million times less likely to occur than
singlet–singlet or triplet–triplet processes and also
that radiationless vibrational deexcitation occurs in
around 10

s. As the mechanism of intersystem

crossing relies upon vibrational coupling between S

and T

, the time required for this spin forbidden

process can be estimated to be around 10

–10

This is of the same order as the lifetime of the ra-
diative transition and therefore intersystem crossing
competes with ﬂuorescence. The probability of in-
tersystem crossing is enhanced when the energy dif-
ference between S

and T

is small and when the

lifetime of S

is relatively long. Intersystem crossing

is also aided by the presence of heavy atoms (e.g.,
iodine and bromine) as substituents on either the
solute or the solvent molecules. This so-called
heavy-atom effect arises from increased spin/orbital
interactions and as such spin reversal becomes

more probable. After intersystem crossing, the
molecule rapidly undergoes internal conversion
(10

–10

s) to the lowest vibrational level of T

In a similar manner to ﬂuorescence, phosphores-

cence can only occur with a radiative deexcitation
from T

to S

and two factors limit the likelihood of

this event. Firstly, because the energy difference be-
tween S

and T

is smaller than that between S

and

, the vibrational coupling between S

and T

may

be enhanced resulting in more efﬁcient internal
conversion. Of more consequence, however, is the
relatively long intrinsic lifetime of the triplet excited-
state (10

–10

s), which provides more than ample

opportunity for collisional deactivation. In fact, this
mechanism is predominant and explains why phos-
phorescence is rarely observed at room temperature
in simple solutions (2,3-butanedione and tris(2,2

bipyridyl)ruthenium(II) are examples). The relatively
long lifetime of T

arises from the spin forbidden

nature of the triplet–singlet state transition; as a
consequence phosphorescence is often characterized
by an afterglow. This is not seen in ﬂuorescence.
Clearly, radiationless processes are much more likely
to deexcite a triplet state compared to the ejection of
a photon, therefore, phosphorescence is most com-
monly observed from molecules that are either at
very low temperatures (often 77 K), in organized
media (micelles), or adsorbed onto solid surfaces.

Structural and Environmental Inﬂuences
on Photoluminescence

From the preceding section, it is evident that the
molecule’s structure and its chemical environment
determine whether or not it will luminesce, and to
what extent. In order to discuss these variables it is
useful to introduce quantum yield (f) or quantum
efﬁciency. For a luminescent process, f represents the
ratio of the number of molecules that emit to the
total number excited. For highly luminescent sub-
stances f will approach unity and for species that do
not luminesce appreciably, f will tend toward zero.
Thus the luminescence quantum yield for a particular
molecule in a speciﬁed environment will be related to
the relative rate constants of those pathways which
can deexcite the lowest excited state (S

or T

). The

ﬂuorescence quantum yield can therefore be ex-
pressed as follows:

ðk

þ k

½2

in which k represents the various rate constants
as designated by the subscripts, where f is ﬂuores-
cence, i is intersystem crossing, and r is radiationless
energy loss, which includes internal and external

308

LUMINESCENCE

/ Overview

conversions plus predissociation or dissociation.
Thus, if k

cðk

þ k

Þ then f

-1 and if k

{ðk

Þ then f

-0. The phosphorescence quantum yield

) is dependent upon competition between the ra-

diationless routes from T

to S

and phosphorescence

emission. But f

also relies upon the rate of inter-

system crossing which competes with both ﬂuores-
cence and radiationless deactivation from S

and S

Thus f

is given by:

þ k

½3

where k

represents the combined rate constant for

all radiationless pathways from T

to S

. In the ma-

jority of situations k

and k

are related to molecular

structure with only minor dependence upon environ-
mental variables. The magnitude of k

can be affected

by both these parameters. Both k

and k

have only

slight dependence upon molecular structure but are
markedly affected by the molecular environment.

Photoluminescence resulting from absorption of

wavelengths below 200 nm is not common since the
subsequent transitions from interaction with such
energetic photons are likely to result in deactivation
of the excited states via predissociation or dissocia-
tion. It is therefore not surprising that luminescence
due to s

-s transitions are virtually nonexistent, in

fact most ﬂuorescent or phosphorescent emissions
from organic molecules arise from p

-p and some-

times p

-n transitions depending upon which is less

energetic. Fluorescence is more commonly observed
from compounds where the S

-S

transition corre-

sponds to a p

-p emission rather than p

-n, which

implies that the quantum efﬁciency resulting from
p

; p

states is far greater than that from n

states.

This situation can be explained by consideration of
some spectroscopic parameters for each state (see
Table 1).

Because the lower molar absorptivity of the n

-p

transitions translates to lower values of k

, the re-

lative importance (efﬁciency) of intersystem crossing
is in turn enhanced. The smaller energy differences
between S

and T

for n

states also increases the

rate of intersystem crossing. Therefore, if S

is an

state, k

in eqn [2] becomes large with respect to

and as such f

tends toward zero. On the other

hand, when S

is a p

; p

state, lifetimes are much

shorter (k

). There is also less vibrational level

overlap between S

and T

and consequently ﬂuo-

rescence becomes more probable. In general, mole-
cules with low lying n

states do not ﬂuoresce but

may phosphoresce given a suitable environment. In
such a case emission would result from a p

to n

transition, but phosphorescence is also observed
from p

to p transitions.

Analytically usable ﬂuorescence is most often

observed from compounds with aromatic function-
ality. Relatively few nonaromatic molecules exhibit
ﬂuorescence and those that do often contain carbonyl
groups or are highly conjugated. Consistent with the
previous discussion, all these species possess low
energy p to p

to S

) transitions. The majority of

unsubstituted aromatic hydrocarbons exhibit ﬂuo-
rescence in liquid solution and sometimes phospho-
rescence under specific conditions. Increasing the
number of fused rings generally results in emission at
longer wavelengths; thus benzene and naphthalene
ﬂuoresce in the UV whereas anthracene and naph-
thacene exhibit blue and green ﬂuorescence, re-
spectively. Substitution on to the aromatic rings
causes shifts in the absorption wavelengths, which in
turn changes the emission spectra. More importantly,
substitution of particular species can drastically af-
fect the quantum efﬁciency. For example, large
(heavy) atoms increase the probability of intersys-
tem crossing to the triplet state and carboxylic acids
or carbonyl groups reduce the ﬂuorescence quantum
yield since these moieties often possess low level n

transitions. Along with substitution, molecular
geometry also affects luminescence. This can be
illustrated by considering two structurally similar
compounds: ﬂuorescein(1), which is intensely ﬂuo-
rescent; and phenolphthalein(2), which is not ﬂuo-
rescent. The only difference between the two
structures is the oxygen bridge present in ﬂuoresce-
in, which imparts rigidity to the molecule. In phe-
nolphthalein electronic excitation can be lost
internally via vibration and rotation rather than
photon emission. In structurally rigid molecules,
energy dissipation via vibration and internal rota-
tion is far less efﬁcient. Enhanced phosphorescence
can result from adsorption of the emitting species on
to a solid surface to provide added rigidity. Such
molecular inﬂexibility is often used to rationalize the
observed increase in ﬂuorescence intensity of organic
complexing agents when chelated to a metal cation.
Thus, the complexes of Ca

, Mg

, Zn

, and

with 8-quinolinol-5-sulfonic acid are much

more ﬂuorescent than the free ligand molecule itself.

Table 1

A comparison of n

and p

; p

singlet states

; p

Molar absorptivity (1 mol

)

10–10

–10

Lifetime(s)

–10

Energy difference between S

and T

Small

Often large

Rate constant for intersystem

crossing (k

)

LUMINESCENCE

/ Overview

309

It is worth mentioning that while many metal ions
form rigid complexes with 8-quinolinol-5-sulfonic
acid, relatively few exhibit analytically useful ﬂuo-
rescence. This is due to the internal heavy atom effect
and/or paramagnetism causing intersystem crossing,
which is manifested (in some instances) by increased
intensity of phosphorescence.

Together with molecular structure, environmental

parameters such as temperature, solvent type, visco-
sity, pH, and dissolved oxygen content can also affect
luminescence. As the solution temperature rises, it
follows that the number of collisions between the
excited state molecule and the solvent molecules will
increase, thus greatly improving the likelihood of
radiationless deexcitation to the ground state. There-
fore, f

for most compounds decreases with increa-

sing temperature. As mentioned earlier, the effect of
temperature upon f

is even more dramatic due to

the vastly greater lifetime of triplet states.

COO

−

COO

−

(1)

(2)

By similar reasoning an increase in viscosity (via

organized media) will serve to limit collisional energy
transfer and thus enhance luminescence. The heavy
atom effect discussed earlier in relation to molecular
structure can be used to explain the depression of
ﬂuorescence with solvents or solutes containing such
species. Carbon tetrabromide and ethyl iodide can be
used to limit ﬂuorescence by promoting intersystem
crossing with a resultant enhancement of phospho-
rescence. The absorption of a photon causes a change
in electronic distribution and, therefore, molecular
geometry. This can often lead to a considerable dif-
ference in polarity between the ground and excited
states. If the excited state has increased polarity then
in polar solvents luminescence will be shifted to
longer wavelengths compared to that observed in a
nonpolar medium. This results from the increased
stabilization of the excited state relative to the
ground state. Solvent polarity can also affect the
type of luminescent phenomena observed. For ex-
ample, in cyclohexane, isoquinoline exhibits intense
phosphorescence (at 77 K), yet the same compound
in either water or ethanol gives only ﬂuorescence. In
protic solvents the lone pair on the nitrogen (n-or-
bital) is solvated which lowers its energy relative to
the p

orbital; as a consequence the p

; p

becomes

the lowest lying state and, as discussed earlier,
ﬂuorescence is preferred. The reverse is the case in

cyclohexane, with the low lying n

being prone to

intersystem crossing and then phosphorescence.

The control of pH in a solution containing a lu-

minescent analyte is also of great importance for
sensitive and reliable analysis. For aromatic com-
pounds with acidic or basic substituents, excitation
and emission wavelengths of the ionized and free
forms are likely to differ. In the case of ﬂuorescence
from metal chelates, the pH must be controlled to
ensure that the conditional stability constant for
the complex is optimal for the particular analytical
situation.

Dissolved molecular oxygen can quench both ﬂu-

orescence and phosphorescence albeit with different
efﬁciencies and mechanisms. The extent of oxygen
quenching on ﬂuorescence is strongly dependent on
the ﬂuorophore, whereas for phosphorescence the
intensity is always adversely affected. This can be
rationalized using the Stern–Volmer equation [4]
assuming no other quenching species are present.

ð1 þ k

½QÞ

½4

where f

and f

are the luminescence quantum

yields in the presence and absence of the quencher,
respectively, [Q] is the molar concentration of the
quenching species and k

is the rate constant for the

quenching interaction. Given that quenching is dif-
fusion limited (k

E10

mol

l s

) and that the

concentration of oxygen in water at atmospheric
pressure is

B10

mol l

then eqn [4] becomes:

ð1 þ 10

½5

Since phosphorescence lifetimes in the absence of

oxygen are in the range 10

–10

s, the ratio of the

quantum yields (from eqn [5]) will be extremely
small for most phosphorescent species. However, as
ﬂuorescent lifetimes in the absence of oxygen vary
from 10

to 10

s, the diffusion controlled

oxygen quenching of singlet states will be most ef-
ﬁcient for the longer-lived species. Numerous mech-
anisms have been proposed for the quenching of
ﬂuorescence by molecular oxygen. The most likely
pathway involves enhancement of intersystem cross-
ing by the triplet oxygen, which can be summarized
as follows:

-ðX

þ O

Þ-

The mechanism of intersystem crossing (

)

is thought to occur via a transient charge-transfer
species between the excited ﬂuorophore (X

) and

molecular oxygen. Although molecular oxygen
promotes intersystem crossing, it does not enhance

310

LUMINESCENCE

/ Overview

phosphorescence. As discussed earlier, the long life-
time of the radiative transition precludes phospho-
rescence being observed in simple solution due to
quenching by collisional deactivation. The mecha-
nism for the quenching of triplet states in solution by
molecular oxygen (or other triplets) is at best spec-
ulative, possibly involving triplet–triplet annihila-
tion. The ability of molecular oxygen to efﬁciently
quench photoluminescence can, however, be exploit-
ed for the sensitive determination of oxygen in liq-
uids and solids.

Paramagnetic transition metal ions may also

quench solution luminescence by increasing the rate
of intersystem crossing. However, there are numer-
ous exceptions to this simple mechanism including
manganese(II) ions, which are generally poor quenc-
hers of luminescence but are highly paramagnetic.
Diamagnetic main group metal ions are usually in-
efﬁcient quenching agents. It is therefore clear that
other processes may be involved in the interaction of
metal ions with luminescent species. Excited state
complex formation and energy transfer are both
likely in some instances. From a practical analytical
point of view, the quenching of luminescence by any
concomitant species in real samples must be inves-
tigated during method development and either
removed or compensated for.

The Relationship between Photoluminescence
Intensity and Analyte Concentration

The relationship between emission intensity and
analyte concentration can be derived using the Lam-
bert–Beer law in conjunction with the schematic of a
spectroluminometer in Figure 2.

If the incident radiation (l

) has the power P

and

a portion of this is absorbed over b

then the radiant

power striking the central region of the sample (P

)

is given by:

¼ P

½6

where e

is the molar absorptivity at l

. It follows

that the radiant power of the exciting beam after
traversing b

) is:

¼ P

½7

The luminescence intensity (I

) is proportional to

both the amount of light absorbed and the quantum
efﬁciency (f) such that:

¼ fðP

½8

A fraction of the emitted radiation will be ab-

sorbed over the pathlength b

and this can be

expressed in terms of the observed luminescence in-
tensity (I):

¼ I

½9

where e

is the molar absorptivity at l

. By rear-

ranging eqns [6]–[9] the following relationship
between P

and I can be obtained:

¼ fP

ð1 10

Þ10

½10

At low analyte concentrations I is proportional to

C as absorption is small. Since the absorbance in-
creases faster than emission at high analyte concen-
trations the resultant I value reaches a maximum and
then decreases as seen in Figure 3. This is termed self-
absorption and it occurs when l

overlaps the

absorption band. Self-quenching also decreases the
luminescence intensity at high concentrations as it
results from a radiationless energy transfer between
two excited molecules and the solvent in a similar
fashion to external conversion. Under certain condi-
tions self-absorption and self-quenching may result
in a maximum in the calibration function. Equation
[10] can be simpliﬁed at low analyte concentrations;
absorbance will be negligible and the three exponents
will become very small. Therefore, we can replace
the terms 10

and 10

by unity. Clearly,

the term 10

cannot be treated in this way or

eqn [10] would become zero:

¼ fP

ð1 10

½11

The term in eqn [11] can subsequently be expanded
as a MacLaurin series to give:

¼ fP

C ln 10

ðe

C ln 10

ðe

C ln 10

? þ

ðe

C ln 10

½12

When the solution absorbance is small (e

o0.05) it is a good approximation to neglect all

′

Entrance

slit

Excitation

radiation

(

)

Sample cell

Exit

slit

Photoluminescent

emission (

)

′

Figure 2

A schematic representation of the sample compart-

ment of a photoluminescent spectrometer.

LUMINESCENCE

/ Overview

311

but the ﬁrst term in the series:

¼ fP

C ln 10

½13

Therefore, when the absorbance is small (at constant
P

and under a deﬁned chemical environment) the

observed emission intensity is directly proportional
to analyte concentration:

¼ KC

½14

In most analytical applications the above condi-

tions are met; and although the linear dependence of
I upon P

is not inﬁnite, it provides a signiﬁcant

advantage for photoluminescence over absorption
spectrophotometry.

Excitation and Emission Spectra

Figure 4 shows both the emission (A) and excitation
(B) spectra of anthracene in ethanol (c. 1 mg ml

The former was recorded with a ﬁxed excitation
wavelength of 340 nm and the latter was monitored
at an emission wavelength set to 379 nm. The four
peaks in Figure 4A are transitions from the zeroth
vibrational level of S

to the excited vibrational levels

of S

. Conversely, the four peaks in Figure 4B cor-

respond to transitions between the zeroth vibrational
level of S

and the various vibrational levels of S

The approximate ‘mirror image’ appearance of the
two spectra occurs due to the similarities in the
energies of the vibrational levels in S

and S

The difference in wavelength between the excita-

tion and emission spectra is a characteristic of photo-
luminescent molecules and is known as the Stokes’
shift, which can be quantiﬁed into wavenumbers by

the following relationship:

Stokes

shift

¼ 10

½15

where l

and l

are the corrected wavelengths of

the maximum excitation and emission, respectively
(in nanometers). In order to record corrected excita-
tion and emission spectra the wavelength dependence
of the source, monochromators, and photomultiplier
response must be identiﬁed and accounted for. The
presence of the extraneous peak in the emission
spectrum can be attributed to Rayleigh scattering of
the excitation radiation. This is most commonly
observed at either the excitation wavelength or twice
this value due to second order diffraction from the
grating in the emission spectrometer. Anthracene also
exhibits phosphorescence in ethanol at 77 K with the
wavelength of maximum emission being 462 nm.
This is consistent with the lower energy of the lowest
excited triplet state.

Chemiluminescence

All chemical reactions are accompanied by energy
changes. Any excess energy is usually dissipated by
collision in around 10

s. Chemiluminescence is

commonly observed at wavelengths from the near
ultraviolet to the near infrared (which correspond to

Fluorescence intensity (mV)

200

150

100

0.0

0.1

0.2

0.3

0.4

0.5

Ru(bipy)

concentration (mmol l

−

)

Figure 3

Photoluminescence calibration for an aqueous solu-

tion of tris(2,2

-bipyridyl)ruthenium(II) chloride hexahydrate,

¼ 457 nm, l

¼ 609 nm.

(B)

(A)

Relative fluorescence intensity

300

350

400

450

500

Wavelength (nm)

Figure 4

Fluorescence spectra of anthracene (1 mg ml

) in

ethanol. The

emission spectrum (A)

was

obtained with

¼ 340 nm and the excitation spectrum (B) was obtained with

¼ 379 nm.

312

LUMINESCENCE

/ Overview

excess chemical energy in the range from 340 to
130 kJ mol

, respectively). Whilst these amounts of

excess energy are achievable with certain reactions
(often redox), the generation of chemiluminescence
depends upon the suitable molecular structure of in-
termediates or products that will facilitate the
conversion of chemical potential to electronic exci-
tation.

After chemical generation of the excited state has

occurred, all of the environmental factors that inﬂu-
ence photoluminescence are equally valid for chemi-
luminescence. It is generally accepted that most
chemiluminescence reactions can be thought of as
chemically induced ﬂuorescence based upon the rar-
ity of phosphorescence occurring in simple solution.
Two noteworthy exceptions are the reduction of (1)
tris(2,2

-bipyridyl)ruthenium(III)

and

(2)

acidic

potassium permanganate in the presence of certain
polyphosphates. The emission in these cases appears
to originate from very short-lived triplets. Many
chemiluminescent reactions can produce light for
several minutes or even longer. It should be empha-
sized that this is indicative only of the reaction ki-
netics, which produce the emitting species, and not
the lifetime of the excited state.

Chemiluminescence can be rather simply classiﬁed

as either direct or indirect. The former can be
thought of as follows:

þ B-C

þ D

-C þ hn

where A and B are reactants employed to produce
either a product or intermediate in an electronically
excited state (C

) that returns to its ground state by

ejection of a photon (hn).

Indirect chemiluminescence has been used ex-

tensively for analysis and in the so-called chemical
light sources; the most well known of these reactions
involve oxidation of certain diaryl oxalates and ox-
amides. Instead of C

returning to the ground state

by photon ejection (as in the above reaction scheme),
it can undergo energy transfer with a suitable ﬂuor-
ophore, which in turn may then exhibit its charac-
teristic ﬂuorescence emission:

þ fluorophore-½fluorophore

þProducts

½fluorophore

-fluorophore þ hn

The analytical utility of indirect chemilumines-

cence depends upon little or no emission resulting
from C

in the region of maximum emission from the

electronically excited ﬂuorophore together with efﬁ-
cient transfer of excitation energy.

Bioluminescence

The emission of light from ﬁreﬂies has enchanted
observers for millennia. However, it was not until
comparatively recently (1947) that adenosine tripho-
sphate (ATP) was identiﬁed as being the key compo-
nent in the enzymatically controlled luciferin
luminescence. The analytical utility of ﬁreﬂy biolu-
minescence was ﬁrst demonstrated in 1952. Since
that time the luciferin–luciferase reaction has been
extensively employed for analysis in a wide variety of
scientific disciplines.

While there are numerous species of ﬁreﬂy having

similar biochemical pathways for the generation of
light, the proposed mechanism for the luciferin–
luciferase reaction is based upon the system of Photi-
nus pyralis. The oxidation of

-luciferin (3) by

oxygen, which is catalyzed by the enzyme luciferase,
is thought to proceed by the sequence outlined
below:

þ LH

þ ATP "

E-LH

-AMP

þ PP

E-LH

-AMP

þ O

-E þ L ¼ O þ CO

þ AMP

where E is the enzyme luciferase, L is luciferin
(3), AMP is adenosine monophosphate, PP is
pyrophosphate and L

¼ O is oxyluciferin (4).

COOH

(3)

(4)

In the initial step, the

-luciferin is activated via

interaction with Mg-ATP to give the enzyme bond

luciferyl adenylate and pyrophosphate. As shown,
this reaction is reversible and therefore pyrophos-
phate can impede the forward reaction. Generally,
only low levels of pyrophosphate are formed from
the concentrations of ATP and luciferase present un-
der normal reaction conditions. The second step is
an oxidative decarboxylation which produces ox-
yluciferin (4) in an electronically excited state. The
excited oxyluciferin returns to the ground state by
the ejection of a photon (l

¼ 562 nm) with an im-

pressive overall quantum yield of

B0.9 (at 251C and

pH 7.8). The light intensity is directly proportional
to the ATP concentration provided that the level of
luciferin remains constant.

There are numerous other bioluminescent systems

of analytical importance, which have been isolated
from various natural sources. A more detailed

LUMINESCENCE

/ Overview

313

discussion of these systems is given elsewhere in this
encyclopedia.

Other Types of Luminescence

Other classes of luminescence have been used in an-
alytical applications; some examples are provided in
Table 2. Thermally stimulated luminescence (or
thermoluminescence) is the emission arising during
mild heating of a solid material, often after it has
been subjected to ionizing radiation. Thermolumi-
nescence has been extensively employed for dating
and dosimetry, including (for example) the detection
of irradiated foods. Some incandescent solids emit at
much shorter wavelengths than expected due to
‘candoluminescence’, which has been exploited for
the detection of metals such as manganese, antimony,
cerium, europium, terbium, lead, and bismuth.

Radioluminescence – induced by gamma- or X-

rays – has also been exploited for dating. Radiolu-
minescent light sources have been designed for spec-
trometric studies; the radioisotope and scintillation
medium allows independent selection of the spectral
and temporal characteristics. Sonoluminescence, the
emission observed when particular solutions are ex-
posed to ultrasonic waves, has the potential for an-
alytical applications involving the detection of
medium-to-high concentrations (over 10 g dm

) of

elements that have an ionization energy below
7.65 eV and a boiling point below 2700

1C.

Triboluminescence (arising from solids during

structural rearrangements such as crushing) has lim-
ited applicability in analytical chemistry, but has
been examined as a tool for clinical diagnosis with
blood samples. The use of triboluminescent materials
to detect damage in composite structures during im-
pact has been suggested.

See

also:

Bioluminescence.

Chemiluminescence:

Overview. Fluorescence: Overview. Phosphorescence:
Principles and Instrumentation.