Caballero B. (ed.) Encyclopaedia of Food Science, Food Technology and Nutrition. Ten-Volume Set

Подождите немного. Документ загружается.

0020 The problem of calibrating and handling such com-

plex fluorescence information will be discussed

below. Likewise, autofluorescence of animal tissue

components and parasites, detected in the fluores-

cence microscope and defined in the spectrofluorom-

eter, can be exploited in digital video techniques

coupled to image analyzers and robots, e.g., to facili-

tate automatic butchering operations and cutting out

bones from cod fillets (Figure 7:3A–D). In the labora-

tory, naturally fluorescent compounds such as vita-

min E components can be separated and detected in

an HPLC system coupled to a fluorometer. In com-

mercial-scale fermentation processes, the fluorescing

reduced forms (NADH and NADPH) of the respira-

tory coenzymes nicotinamide adenine dinucleotide

(NAD) and nicotinamide adenine dinucleotide phos-

phate (NADP) can be used online to follow the rate of

the fermentation process. (See Tocopherols: Proper-

ties and Determination.)

0021Similarly, a food chemist in need of a certain assay,

e.g., to monitor the breakdown of mixed-linkage

b-glucan in the malting and brewing industry or to

Ellipsoid

M (E) 5

Source

Mirror

M (S) 3

Mirror

M (T) 4

Mirror

M (F) 2

Mirror

M (F) 2

Mirror

M (T) 1

Grazing

G (1200)

Mirror

M (S) 3

Sample

photomultiplier

Mirror

M (T) 1

Grating

G (1440)

Reference

photomultiplier

Sample

Beam

splitter

Slit

excitation

entry

Slit

excitation

exit

Slit

emission

entry

Slit

emission

exit

Zero-order

mirror

accessory

Excitation

polarizer

Emission

polarizer

Emission

filter

fig0006 Figure 6 Fluorescence spectrophotometer (Perkin Elmer LS-50). The source (S) is a pulse xenon flashtube (less than 002 s for a

complete pulse). Light from S is reflected by a mirror through a slit to the diffraction grating (G1) which has 1440 lines per millimeter.

The resulting narrow-wavelength band is transferred via another system of mirrors and slit to the sample (A). The effective wavelength

of the excitation beam is determined by the setting of the grating, the angle of which is controlled by means of a stepping motor. A very

small proportion of the excitation light is reflected by a beam splitter (B) to a reference photomultiplier. To correct for the response of

the reference multiplier, a rhodamine correction curve is stored within the software in the instrument computer. Light emitted by the

sample is transferred via slits and mirrors to a second diffraction grating (G2), which has 1200 lines per millimeter and is controlled by

another stepping motor, and on to the sample photo multiplier (C). A microprocessor controls the xenon tube, grating stepping motors,

slit apertures, and photo multipliers. The digitized signal is accumulated for about 008 s, i.e., four flashes of the source. The sample

accumulation is divided by the corrected reference accumulation to obtain a ratio that is unaffected by variation in source intensity.

Output is given by the microprocessor on a data screen and a recorder. The computer is programmed so as to allow calibration

spectra to be subtracted from sample spectra to generate difference spectra. The wavelengths of the instrument can be calibrated by

the characteristic fluorescence spectrum of benzene vapors, for example, or from fluorescent standards cast into plastic blocks

furnished by the instrument manufacturers. It is important in some applications to be able to record accurately several of the internal

settings of the instrument, including spectral correction and different scanning modes, and this can be done with a high-specification

spectrofluorometer. The practicable wavelength ranges for the standard version of the instrument are about 230–600 nm for excitation

and about 250–650 nm for emission. Reproduced from Spectroscopy: Fluorescence, Encyclopaedia of Food Science, Food Technology

and Nutrition, Macrae R, Robinson RK and Sadler MJ (eds), 1993, Academic Press.

SPECTROSCOPY/Fluorescence 5437

ABCD

Days

275 µm 1600 µm 100 µm

Video Cutting robot

Conveyor

Separator

fig0007 Figure 7 (see color plate 129) Examples of fluorescence microscopy and imaging reproduced from articles referred to below in

Munck L. (ed.) (1989) Fluorescence Analysis in Foods, Harlow: Longman. pp. 289. 7:1. Cryostat sections of 0, 1-, 3- and 5-day malted

barley viewed through a 4 microscope objective. The sections were stained by (A) Calcofluor/Fast green photographed at 400 nm

excitation/418 nm emission in order to show the patterns of cell wall (b-D-glucan) breakdown and by (B) FITC-labelled antibodies

against a-amylase viewed with BP 455-490 nm exciter/LWP515 nm emission filter. Sections are identical pairwise – first stained with

method B followed by A. It is seen that a-amylase is spread from the germ (the structure on the left of each section) and from the

aleurone layer enclosing the starchy endosperm. The enzyme invades only the endosperm areas where cell walls are absent.

(Compare B with A on each day of malting.) From Munck L. Practical experiences in the development of fluorescence analyses in an

applied food research laboratory, ibid. p. 1-32. 7:2:1-3. represent a wheat and barley material with natural and added fluorochromes.

7:2:1. A view of ground whole wheat flattened slightly with cover glass, excited at 365 nm with a barrier filter from 420 nm. Bran

fragmentsarereadilydetectedbythenaturalbrightbluefluorescencecharacteristicofthealeuronelayer(arrows).Seealso Figure

2b.7:2:2.Alow-powerviewofalongitudinallyhalvedbarleykerneltreatedwithamultiplefluorescentstain(Safranin,BasicFuchsin,

Alcian Blue) to show several grain components simultaneously. The starchy endosperm (central region of the kernel) fluoresces bright

yellow due to starch-safranin interaction under 450-490 nm excitation, while the outer bran layers are deep red as a result of Basic

Fuchsin interaction with lignified and other phenolic-enriched structures such as cell walls (arrows). In this case, safranin is used to

detect starch, while the other stains are added as counterstains to surpress non-starch specific fluorescence. 7:2:3. As in 7:2:2, but

showing the bran tissues at higher magnification and 365 nm excitation. In this case, the procedure clearly differentiates between I.

ferulic acid-enriched aleurone cell walls (lower left) detected due to their natural fluorescence (as in 7:2:1), II. surface trichomes

(arrows) which do not bind significant amounts of the red fuchsin stain and III. the pericarp cell walls (right diagonal) which are

intensely stained deep red. Starch does not fluoresce significantly at these wavelengths. From Fulcher RG, Irwing DW and de

Francisco A. Fluorescence microscopy: Applications in food analysis, ibid. 59-109. 7:3:A-D. Operation of an automatic Fish Fillet

Deboning Line based on fluorescent technology. Skinned cod fillets are fed onto a steel conveyor while moving at one m/s. The video

detector (see drawing above the photographs) detects a fillet (A). The detector consists of two cameras and two light sources for

visible and UV light for fluorescence, respectively. The visual light is projected in stripes (A) in order to detect (I) the boundaries of the

filet and (II) the topography of the fillet for later partition (D) into desired weight portions. The video UV fluorescence system detects the

naturally fluorescent bones (ex. max 340/em. max 390) seen on a TV monitor (B). The information from the two cameras is directed to a

computer which directs a water jet nozzle (diameter 0.1 mm, pressure 300 bar) robot with an operating space of 1 m. The cutting of the

fillet bones (C) and further partition of the fillet (D) is performed while the steel conveyor is moving and completed within the time span

of one second. Bones and fillet pieces are separated and graded in the separator unit drawn above. This principle may also be used to

remove bones from chicken breast fillets. From Jensen SAa, Reenberg S and Munck L. Fluorescence analysis in fish and meat

technology, ibid. 171–180.

5438 SPECTROSCOPY/Fluorescence

(a)

Raw data: 268 samples

from a sugar campaign

Raw data

(b)

PARAFAC

1. Initialize B and C

2. A =

k=1

−1

(B'B)*(C'C)

3. B =

k=1

(A' A)*(C' C)

−1

4. D

5. Go to 2 until convergence

−1

(B' B)*(A' A) diag (A' X

(c)

40 50

24 26 28 30 32 34

30 40 50

30 40

Emission (nm  10) Excitation (nm  10)

Comp 2 Tryp Try Comp

Deconvoluted spectra

(d)

036 9

12345

Time/day

Score

Concentrations

Tyrosine

Comp. 2

Tryptopha

Comp. 1

6789

(e)

Predicting color

Day

Predicting CaO

(f)

0123456789

Predicted

Reference

Day

fig0008 Figure 8 An example of mathematical chromatography is multiway modeling of fluorescence landscapes of sugar into the pure

excitation and emission spectra of the underlying fluorophores. A total of 268 samples of sugar have been dissolved in water and

measured spectrofluorometrically, spanning the sugar campaign with three samples representing three 8-h periods per 24 h. Each

such measurement gives a characteristic landscape of a sample (a). By using the so-called parallel factor analysis (PARAFAC) model

(b), it is possible to describe these landscapes by the variation in four estimated fluorophores (c). Thus, at any time the sugar

landscape of any sugar sample can be described by having different amounts expressed in scores of these four fluorophores, defined

by their estimated excitation and emission spectra. The former is described less accurately than the latter due to sampling (20 nm).

The amounts of the four fluorophores in the 268 samples thoughout the season are shown in (d). Furthermore, it is possible to qualify

which chemical analytes these four fluorophores represent, just as in ordinary chemical chromatography. In this case, two of the

analytes have been identified as the uncolored amino acids tyrosine and tryptophan which have the potential for color formation

through reaction with reducing sugars. (Estimated and true emission spectra are shown.) The other two fluorophores represent

colored fluorescent high-molecular Maillard products. These interpretations have been checked with high-performance liquid

chromatography. Thus, a chemical understanding of the fluctuations during the campaign is provided. In (e) and (f) it is demonstrated

that this spectrofluorometric chemically based variation can be used for creating cheap online screening analyses for prediction of,

e.g., important process variables such as CaO in sugar juice or quality parameters such as color in sugar by using a multiple linear

regression model based on scores of the four fluorophores.

SPECTROSCOPY/Fluorescence 5439

screen for lipases in oat products, may select from

potentially useful fluorochomes (Table 1) to solve the

analytical problems by employing secondary fluores-

cence techniques: for example, calcofluor white Mr2

new and fluorescein dibutyrate can be used to deter-

mine mixed-linkage b-glucan and esterase/lipase con-

tents, respectively, of cereal grains. In the case of the

calcofluor/b-glucan assay, two official European

brewing convention methods, one visual – malt

modification analysis (Figure 7:1A) – and one chem-

ical – FIA for mixed-linkage b-glucan in wort and

beer – have been developed. These methods have

helped to strengthen the insight of brewing chemists

into the strong relationship between the physical cell

wall structure of barley and malt and the chemistry of

wort and beer related to viscosity and filterability due

to the influence of b-glucans. Mg-ANS (1-aniline-8-

naphthalene sulfonic acid) can be used to detect dead

cells in pitching brewer’s yeast, and fluorescein iso-

thiocyanate (FITC)-immunofluorescent visual and

chemical methods can also be applied successfully,

for example, to monitor the distribution of a-amylase

activity in malting barley seeds (Figure 7:1B). A range

of components such as aleurone, starch, b-glucans,

protein bodies, fat globules, protein, and crystalline

inclusions can be convincingly demonstrated by

employing a range of fluorochromes and treatments

one at a time or in combination as demonstrated in

an example with wheat in Figure 7:2:2 and 3. Several

of these methods can also be used quantitatively to

determine components in homogenized samples by

spectrofluorometry.

Fluorescence Spectrofluorometry for Prediction of

Food Quality Calibrated with Chemometrics

0022 Due to its specificity and sensitivity, spectrofluoro-

metric analysis allows a unique characterization of a

fluorescent food sample, thus complementing near

infrared spectrophotometry. Fluorescence measure-

ment with fluorochromes in vitro, e.g., to detect

lipase (esterase) activity with fluorescein dibutyrate

at 495 nm excitation and 525 nm emission (Figure 1)

is a rather straightforward analysis using the equa-

tions given above. However, spectra from complex

fluorescence systems, such as that from wheat flour

(Figure 2), need advanced data analysis techniques

(chemometrics) similar to those developed for protein

analysis of seeds by near infrared spectroscopy. These

use complete spectra, and involve principal compon-

ent analysis for discrimination between normal and

deviating samples and partial least-squares analy-

sis (PLS) for calibration to chemical analyses for

prediction. Precise calibration of the spectrofluorom-

eter that is stable from day to day is obligatory

here, if data from whole spectra are to be utilized.

Spectrofluorometric data in the form of excitation/

emission landscapes are ideally resolved by multiway

chemometric algorithms such as parallel factor analy-

sis (PARAFAC: Figure 8) which in the fluorescence

example from process control in a beet sugar factory

covers three dimensions: excitation, emission, and

sample. In a PLS evaluation of a complex fluores-

cence landscape the latent factors, or loadings, are

difficult to interpret due to orthogonality. In contrast,

PARAFAC loadings are interpretable and expressed

in the form of estimates of the underlying excitation

and emission spectra of the corresponding fluoro-

phores. If stable computerized instruments with en-

larged measurement area for fluorescence monitoring

(mathematical chromatography) of inhomogeneous

samples in the food industry can be developed, the

chemometric software already developed will be able

to create a measurement system with unique potential

with regard to sensitivity and specificity to measure

those quality characteristics which depend on fluor-

ophores.

See also: Carotenoids: Occurrence, Properties, and

Determination; Chromatography: Thin-layer

Chromatography; High-performance Liquid

Chromatography; Polycyclic Aromatic Hydrocarbons;

Retinol: Properties and Determination; Spectroscopy:

Near-infrared; Tocopherols: Properties and

Determination

Further Reading

Bro R (1999) Exploratory study of sugar production using

fluorescence spectroscopy and multi-way analysis. Jour-

nal of the Chemometrics and Intelligent Laboratory

Systems 46: 133–147.

Engelsen SB (1997) Explorative spectrometric evaluations

of frying oil deterioration. Journal of the American Oil

Chemist Society 74: 1495–1508.

Engelsen SB, Mikkelsen E and Munck L (1998) New ap-

proaches to rapid spectroscopic evaluation of properties

in pectic polymers. Progress in Colloid Polymer Science

108: 166–174.

Guilbault GG (ed.) (1990) Practical Fluorescence – Theory,

Methods and Techniques. New York: Marcel Dekker.

Hurtubuise RJ (1984) Solid Surface Luminescence Analysis.

New York: Marcel Dekker.

Lakowicz JR (1999) Principles of Fluorescence Spectros-

copy. New York: Kluwer Academic/Plenum Publishers.

Morgan LC, Gormally J, Hubbard AR, d’Lacey C and

Ockleford CD (1999) Confocal microscopy: theory and

applications. Methods in Molecular Biology 114: 51–74.

Munck L (ed.) (1989) Fluorescence Analysis in Foods.

Harlow: Longman.

Munck L, Nørgaard L, Engelsen SB, Bro R and Andersson

CA (1998) Chemometrics in food science – a demonstra-

tion of the feasibility of a highly exploratory, inductive

5440 SPECTROSCOPY/Fluorescence

evaluation strategy of fundamental scientific signifi-

cance. Journal of the Chemometrics and Intelligent

Laboratory Systems 44: 31–60.

Rost FWD (1992) Fluorescence Microscopy, vols 1 and 2.

Cambridge: Cambridge University Press.

Von Sengbusch G and Thaer AA (1973) In: Thaer AA and

Sernetz M (eds) Fluorescence Techniques in Cell Biology,

p. 31. New York: Springer.

Wang XF and Herman B (eds) (1996) Fluorescence Imaging

Spectroscopy and Microscopy. New York: Wiley.

Wolfbeis OS (ed.) (1993) Fluorescence Spectroscopy: New

Methods and Applications. Berlin: Springer.

Atomic Emission and

Absorption

N Ulrich, University of Hannover, Hannover, Germany

Summary

0001 Atomic spectrometry is suitable for the determination

of trace concentrations of most elements. The

principle is the absorption or emission of light of

certain wavelengths after the atomization and excita-

tion of the sample. The wavelength gives qualitative

information about the element, whereas the intensity

of the emission or the quotient of the intensity before

and after the absorption is proportional to the con-

centration of the analyte. Several spectral and non-

spectral interferences must be considered.

Background

0002 Atomic emission and absorption spectrometry are

used for the qualitative and quantitative determin-

ation of chemical elements, mainly metals and semi-

metals in a huge variety of different matrices. Both

techniques are based on the interaction of atoms

and electromagnetic energy. Kirchhoff and Bunsen

grounded the atomic spectrometry in the middle of

the nineteenth century. They observed that free, gas-

eous atoms of an element can absorb or emit radi-

ation of discrete and specific wavelengths. In atomic

spectrometry, wavelengths between 170 and 800 nm

are investigated.

The Atomizing Process – Important for

Both Techniques

0003 Prior to the excitation or absorption step, free atoms

of the elements must be formed. In classic techniques,

the samples are introduced into the system as aerosols,

e.g., solutions of sodium chloride. These aerosols are

desolvated and evaporated rapidly by adding thermal

energy to form microparticles. By interactions with

the environment in the flame or the plasma, free

atoms are formed that are capable of absorbing light

of specific wavelengths (atomic absorption) or, if

excited by thermal energy, can emit light of specific

wavelengths (atomic emission). The wavelength is

equal for both processes. Figure 1 shows the atomiza-

tion process for aerosols using sodium chloride

(NaCl) as an example.

Excitation and Relaxation of Valence

Electrons

0004After the atomization, the excitation of the electrons

(absorption) and relaxation (emission) take place.

Only the valence electrons are capable of the

radiation used in atomic spectrometry. In their basic

form, these electrons are in certain energy levels,

representing the ground state for an atom of a given

element. The electrons can absorb energy and form

energy-enriched, excited states. Two forms of excita-

tion are of importance for atomic spectrometry:

0005Thermal excitation: The electrons can be excited

by thermal energy. At temperatures of 2000–

3000



C, most of the atoms are in the ground

state, following the Boltzmann distribution.

More states can be populated at higher tempera-

tures, e.g., in the Ar-plasma of an inductively

coupled plasma (ICP). The relaxation takes place

as electromagnetic radiation, which is measured in

atomic emission spectrometry (AES).

0006Electromagnetic excitation: The electrons can

absorb light of a certain wavelength according to

Einstein’s law of absorption and emission, which

is measured in atomic absorption spectrometry

(AAS). The wavelength depends on the energy dif-

ference between the ground and excited electronic

state of the atom. The relaxation may take place as

thermal or electromagnetic radiation.

There are many emission lines in different series

starting from the ground state or an excited state, but

Drying:

+ CI

−

+∆E

NaCI

Evaporation:

NaCI

+∆E

Ionization:

∗

+∆E

Excitation:

•

+CI

•

+∆E

AES:

∗

•

+hn

Dissociation:

NaCI

•

+CI

•

+∆E

AAS:

•

+hn

∗

fig0001Figure 1 Processes in the atomizer of an atomic spectrometer.

SPECTROSCOPY/Atomic Emission and Absorption 5441

not every emission line is also a detectable absorption

line in AAS. The intensity of the absorbance line

depends on the population of that state (,) which

absorbs the light. In most cases, only the ground

state is populated to such an extent that intensive

radiation is emitted.

Atomic Emission Spectrometry

0007 In AES, the atoms are excited by thermal energy and

emit light of specific wavelengths. The frequency of

the light is proportional to the energy difference of

both states. For optical emission spectrometry, the

wavelengths are in the ultraviolet/visible region. This

light is measured and can be correlated by the wave-

length to the element (qualitative information) and

by the intensity to the concentration (quantitative

information).

0008 The intensity of the emitted light depends on the

number of atoms in the excited state. This number

will be larger for higher temperatures given by the

Boltzmann distribution. At temperatures below

2500



C, only transitions between the ground state

and one of the first excited states are possible for most

of the elements. The alkali metals can be excited in

sufficient rates at temperatures of c. 2000



C. The

concentration of the analyte in the sample is directly

proportional to the intensity of the emitted light.

Principal Components of an Atomic

Emission Spectrometer

0009 The main components of an atomic emission spec-

trometer are the sample introduction device, which is

often a pneumatic nebulizer system, the atomization

and excitation source, one or more monochromators,

and a detector, mostly a photomultiplier. The princi-

pal set-up of an AES is shown in Figure 2. AES can be

used as single element method or as a simultaneous

measurement device, as detailed below.

Atomizing and Excitation Techniques

0010 There are many techniques for atomizing and

exciting samples, like bow and spark excitation or

glow discharge excitation. The two most important

methods are flame photometry, which is suitable pri-

marily for alkaline and some earth alkaline elements,

and ICP, which is used widely as an energy source in

atomic emission spectrometry.

Flame Photometry (Atomic Emission Spectrometry)

0011 Flame photometry was established by Bunsen. It is

called photometry, although it is an emission tech-

nique. The correlation between the intensity and

concentration is linear, and the Lambert–Beer law

cannot be applied. The atoms are excited in a flame,

and the intensity of the emitted light is measured.

0012The apparatus of FAES contains a nebulizer system,

a burner, suitable for different gases like nitrous oxide

or methane, a monochromator, which is important

for the quality of the measurement, and a detector.

Often, flame atomization absorption spectrometry

systems with the lamps turned off are used as the

FAES device.

0013Applications and analytical performance In the

early years, FAES was used for all metals. However,

the exciting efficiency of the flames was poor, and

the detection limits were quite high. Today, FAES is

mainly used for the determination of alkali metals in

solutions. The sensitivity for these elements, espe-

cially Na and K, can be better than for the FAAS.

Typical detection limits are 0.01 mg l

1

(Na) and

0.1 mg l

1

(K) in liquid biological samples. For other

elements, AES with plasma excitation, atomic

absorption spectrometry, or mass spectrometry is

used nowadays.

0014Problems and interferences Spectrometric determin-

ations are relative methods, so calibration is neces-

sary. Differences between the samples and the

standard solutions can lead to systematic errors,

called interferences. In spectrometry, there are two

types of interferences. The spectral interferences are

problems caused by unspecific radiation or absorp-

tion, whereas nonspectral interferences affect the con-

centration of the analyte in the atomizer and exciting

zone.

0015Spectral interferences in flames include the spectral

lines of matrix compounds and the high continuous

background of the flame, emitting light of all wave-

lengths. In relatively cold flames (2000



C), emission

Atomizer

(excitation)

Monochromator

Detector

Nebulizer

Sample

fig0002Figure 2 Principal set-up of an AES apparatus.

5442 SPECTROSCOPY/Atomic Emission and Absorption

spectra with few lines are observed. However, excesses

of one element, e.g., 1000-fold excess of Na for Li

determination can lead to an overlapping of the sig-

nals. In addition, Ca and Sr can lead to interferences.

0016 Nonspectral interferences can be classified as

chemical interferences, transport interferences, ion-

ization interferences, and self-absorption. If stable

compounds of the analyte and matrix components

are formed, these will hinder or even completely

suppress the atomization and excitation. Organic

substances can alter the flame temperature, which is

important for the FAES, and can emit continuous

background radiation. In addition, the analytes can

react with the flame gases, e.g., Ca can form CaO

with oxygen.

0017 Transport interferences occur when the amount of

sample introduced into the atomizer changes over

time. The main reason is a difference in viscosity

between water solution and organic solvents. In

addition, larger concentrations of salts and matrix

compounds can disturb the nebulizing process.

0018 Ionization interferences occur, when part of the

analyte is ionized within the flame. It only takes

place to a small extent, but causes a reduction of the

population in the excited state. One problem with a

high concentration of the analyte is self-absorption.

Because emission and absorption take place at the

same wavelength, analyte atoms can absorb light,

which is emitted by other analytes. This leads to a

significant curving of the calibration graph.

Inductively Coupled Plasma–Atomic Emission

Spectrometry (ICP-AES)

0019 The inductively coupled plasma was introduced into

atomic emission spectrometry by Greenfield in 1964.

ICP-AES is today one of the most widely used tech-

niques for trace metal and semimetal determination

in a huge variety of different samples.

0020 An argon plasma with a gas consumption of c.17l

min

1

is used as atomizer. The gas is delivered

through three concentric quartz tubes (the torch),

which are surrounded by an induction coil. This coil

is operated at a frequency of 27.1 MHz and a power

of 1000–2000 W. Gas temperatures of 6000–8000



are observed in the plasma, which, in combination

with the relative long residence time of the analytes in

the plasma, lead to an effective energy transfer on to

the analytes. Even refractive metals and oxides can be

atomized to a great extent. The spectra obtained are

often complicated and contain many lines, especially

for the transition metals. The plasma can be observed

either axially or radially.

0021 One of the main advantages with respect to AAS is

the ability of multielement determinations, because

sequential and simultaneous set-ups can be used.

Analytical wavelengths range from 170 to 800 nm.

For sequential investigations, the monochromator,

often of the Czerny–Turner type, is tuned for

each element. Several monochromators of un-

changeable wavelengths are fixed in the detector,

mostly forming a Rowland circle, for simultaneous

measurements.

0022Applications and analytical performance The ICP-

AES is suitable for a wide range of elements, because

the high temperatures provide very good atomizing

conditions. All metals and semimetals and even some

of the nonmetals (e.g., sulfur, phosphorus, and iodine)

can be detected. The simultaneous spectrometers

are limited to six to 30 elements, depending on the

number of monochromators in the detector. Quanti-

tative results for these elements can be obtained

in c. 1 min. Typical detection limits range from

0.01 mgl

1

for Ca, Mn, Mg, and Mo to 20 mgl

1

for

As, Bi, and Sn. The dynamic range reaches usually

over five to six decades, so that matrix compounds

and trace elements often can be detected in one

analysis.

0023Gaseous, liquid, and solid samples can be detected

using ICP-AES. Liquids can be nebulized with any of

the standard nebulizers. Flow injection and on-line

coupling to chromatographic devices are other

options. Gases, e.g., metal hydrides, can be brought

directly into the plasma. The dry plasma and the

nearly complete sample introduction lead to a signifi-

cant enhancement of the detection limits. For solid

samples, mostly slurry techniques with particles less

than 2 mm in diameter are used.

0024Problems and interferences The drawbacks of ICP-

AES are the high costs of the apparatus and mainten-

ance, and the high spectroscopic background of the

plasma.

0025In addition, there are numerous interferences ob-

served in the spectra. Spectral interferences arise from

the plasma itself, which is an intensive source of

unspecific radiation and argon emission lines. Both

depend on the temperature and the composition of

the plasma gas and vary, e.g., with the introduction of

water from the sample. Further, there are emission

lines of hydrogen, nitrogen, and oxygen. Different

molecular species are formed in the plasma, e.g.,

, OH, NH, or NO, which show a large number

of rotation and vibration bands.

0026Nearly all elements in the sample emit characteris-

tic spectral lines not only from the ground state but

from several excited states, because the high tempera-

tures in the plasma lead to a remarkable population

of numerous states. These interferences can be re-

solved to some extent by the use of high-efficiency

SPECTROSCOPY/Atomic Emission and Absorption 5443

monochromators, but many lines are too close

together for separation.

0027 Chemical interference is not important in most

cases. Transport interference can occur, as described

for FAES.

Other Plasma Sources for Atomization

0028 Microwave-induced plasmas (MIP) are formed by

microwaves (frequency 2.45 GHz, power 50–200 W).

This plasma is mainly used as a detector for gas

chromatography. Mostly, He is used as the plasma

gas, providing very high electron temperatures and

excitation energies, e.g., for the nonmetals.

0029 Several other plasma sources have been applied for

AES. Glow discharges, sparks, and bows are mostly

used for metallic samples. In addition, capacity

coupled plasmas can be utilized to excite metal atoms.

Atomic Absorption Spectrometry

0030 This technique is based on the selective absorption of

light by the gaseous free atoms. The absorbance is the

logarithm of the quotient of the radiation intensity

before the absorption, I

, and after the absorption, I.

According to the Lambert–Beer law, it is directly

proportional to the concentration of the absorbing

atoms in the atomizer, c, the length of the atomizer,

d, and the absorption coefficient, a.

A ¼ 1



¼ adc: ð1Þ

The coefficient, a, depends on the element and the

wavelength, g, and can be related to the Einstein

transition probability of an absorption process, B

and the Planck constant, h.

a ¼ B

: ð2Þ

Principal Components of an Atomic

Absorption Spectrometer

0031 The main component of an analytical absorption

spectrometer is the light source, providing monochro-

matic light for the absorption process. Two types of

light sources are mostly used. Hollow-cathode lamps

contain a cathode of the analyte element and an

anode, and are filled with a noble gas. There is a

glow discharge between the cathode and the anode,

in which positive gas ions are formed, which sputter

element atoms of the cathode at relatively low

temperatures. These lamps emit mostly lines excited

from the ground state and show only slight line

broadening. The other type of lamps are electrodeless

discharge lamps that contain the element in a small

quartz tube filled with a noble gas. A high-frequency

field (c. 27 MHz) leads to a plasma within the tube, in

which the element is excited and emits specific light.

The advantages are the better stability of the lamp,

especially for elements like antimony, mercury and

tin, and the higher intensity. A problem is the forma-

tion of ions, which reduce the effective concentration

of the excited atoms. Both lamp types generate

narrow emission lines and are available for all elem-

ents that can be determined by AAS. AAS is mostly a

single element or sequential analytical technique, be-

cause a specific light source is used for each analyte.

0032The light beam is sent through the atomized sample

and then via a monochromator to the detection

system similar to that of AES. Pneumatic nebulizers

are used as well for AAS. In addition, slurry and solid

samples have been investigated. The atomizer in the

light path can be of different types. The principal set-

up of an AAS is shown in Figure 3.

Atomizing Techniques for Atomic

Absorption Spectrometry

0033The purpose of the atomizer is the reproducible for-

mation of gaseous atoms of the analyte in the ground

state. There are three main atomizing techniques in

atomic absorption spectrometry. The classical method

is flame atomization. Electrothermal atomization pro-

vides better detection limits for most elements, and

chemical vapor atomization is well suited for gaseous

samples like the metal and semimetal hydrides and

gaseous mercury.

Flame Atomic Absorption Spectrometry

0034In FAAS, the liquid sample is introduced as an aerosol

in a flame with temperatures between 2000 and

3000



C. The flame should have a low self-absorption

at the analyte wavelength and be a low source of

Atomizer

Monochromator

Detector

Nebulizer

Sample

Lamp

fig0003Figure 3 Principal set-up of an AAS apparatus.

5444 SPECTROSCOPY/Atomic Emission and Absorption

white light to provide adequate signal-to-noise ratios.

In addition, the residence time of the analyte atoms in

the flame should be sufficient for the absorbing

process, and therefore, lower flow rates for the

flame gases should be used.

0035 The most important is the acetylene/air flame,

which is rather transparent over a broad region of

wavelengths and, at temperatures of c. 2500



C, is

suitable for most analytes. Higher temperatures are

needed for some elements like Be, Ca, Ti, V, Mo, and

the rare earths, so acetylene/nitrous oxide flames with

temperatures of c. 3000



C are used. Other flame

types include hydrogen/air and methane/air.

Applications and Analytical Performance

0036 FAAS is suitable for samples with analyte concentra-

tions of mg l

1

to the higher mgl

1

. The sensitivity

is often given as a characteristic concentration, c

which depends on the parameter of the instrument

(e.g., the geometry of the atomizer) and the element.

Linear calibration functions are obtained up to 100

, and the precision is frequently better than 0.5%.

0037 FAAS can be used to determine the macronutrient

elements in human body fluids and other biological

matrices. Whereas sodium and potassium are mainly

investigated by flame photometry, calcium and mag-

nesium are analyzed by FAAS by direct measurement

after dilution.

Problems and Interferences

0038 The matrix in the samples can influence the drying,

crystal formation, atomization, and distribution of

elements in the flame. Mostly, the kinetic of the atom-

ization is changed by analyte compounds with high

boiling points. The energy of the flame can lead to an

ionization of certain elements like the alkali or earth

alkali metals and to curved calibration functions. The

addition of electron donators like Cs shifts the equi-

librium to the atoms and provides a better linearity

of the calibration. Chemical reactions and spectral

effects are less important for the FAAS, because of

the dilution of the analytes and the matrix by the

flame gases. Problems can arise from organic solvents

and high salt concentrations, which alter the viscosity

of the solution.

Electrothermal Atomic Absorption

Spectrometry (ETAAS)

0039 One major problem of the FAAS is the residence time

in the atomizer. In order to improve this time, the

ETAAS has been invented by L’vov in the middle of

the twentieth century and is now one of the most

widely applied techniques for trace and ultratrace

metal determinations.

0040In the ETAAS, the atomizer is a small quartz tube

mounted between two electrodes. Samples containing

a few microliters of liquid or micrograms of solid are

injected through a small hole on top of the tube. The

oven is heated electrically following a temperature

program:

0041The drying step, in which the water or the solvent

is evaporated.

0042The sample preparation step, in which matrix

components are evaporated or altered.

0043The atomization step, in which the analytes

are atomized at constant temperatures, 1500–

2500



C, depending on the analyte.

0044The cleaning step, in which the temperature is

brought to a maximum, and the graphite furnace

is cleaned.

In order to improve the sensitivity of the ETAAS and

to inhibit the formation of carbides by elements like

Mo, V, and Ti, the furnace can be coated by different

materials including Os and Pd. In contrast to the

continuous signal of the FAAS, in ETAAS, a transient

signal, mostly a heavy tailing peak, is obtained.

Analytical Performance

0045The ETAAS has several advantages in respect to the

FAAS. Slurry and solid samples can be used without

dissolution, and the analytes can be separated from

the matrix using a program of different temperatures.

The method provides a high sensitivity and very

low sample consumption, because nearly the whole

sample is introduced into the atomizer. Drawbacks

are the complicated apparatus and handling, the

low sample throughput, and the large number of

interferences. The ETAAS is now widely used for

the determination of trace elements in biological

samples.

0046The detection limits of the ETAAS are from

0.5 mgl

1

for elements like V and Li to 0.005 mgl

1

for elements like K, Cd, and Zn (10–1000-fold better

than for the FAAS). The absolute sensitivity depends

on the dimensions and geometry of the quartz tube

and further experimental parameters. It is given as

characteristic mass, m

. The precision is c. 1% stand-

ard deviation at 10 m

, and the linear range ends by

c. 100 m

, similar to the FAAS.

Problems and Interferences

0047In the ETAAS, numerous spectral interferences have

been determined. These interferences are based on

the insufficient separation of the analytes and matrix.

The radiation from the light source can be absorbed

unspecifically by atoms and molecules, leading to

overlapping atomic and molecular bands or can be

scattered on particles like undissociated metal salts

SPECTROSCOPY/Atomic Emission and Absorption 5445

in the light beam. At high atomizing temperatures,

sublimating graphite particles can lead to problems.

0048 The background depends strongly on the matrix,

so that background correction is much more import-

ant than for FAAS. Modern apparatus use either a

deuterium lamp or the Zeeman effect for correction

of the data. The standard addition method should be

used for complex matrices.

Chemical Vapor Atomic Absorption

Spectrometry (CVAAS)

0049 CVAAS is used for the determination of gaseous metal

and semimetal compounds. It is an important detector

for the speciation of hydride-forming elements. There

are two sectors of the CVAAS: the cold vapor tech-

nique, which is only suitable for mercury, and the

hydride generation process, which can be applied to

Se, Te, As, Sb, Bi, Sn, and Pb. In the latter case, the

elements are reduced to the corresponding hydrides

mostly with sodium tetrahydridoborate in acidic solu-

tion. A quartz furnace is used as the atomization unit,

which is typically heated electrically to 900–1000



The samples are introduced as gases to nearly 100% by

an inert carrier gas stream. The residence time in the

atomizer depends on this gas stream.

0050 Mercury compounds can be reduced by sodium

tetrahydridoborate to elemental mercury, which can

be introduced by inert gas into the atomizer nearly

without heating. The quartz furnace is used at 120



in order to inhibit the condensation of mercury and to

reduce the scattering on water particles in the oven.

Analytical Performance

0051 The absolute performance of the hydride generation

is c. 100-fold less than the ETAAS, but the larger

sample volumes (up to 100 ml in one batch reactor)

lead to an increase in the detection ability usually at a

factor of 10. The determination of 0.01 mg per liter of

arsenic or selenium is possible. In addition, the oxi-

dation states and organic species of the hydride

formers can be detected separately, so that more infor-

mation about the species distribution of these elem-

ents in different matrices can be gathered. In addition,

the matrix is nearly completely separated from the

analytes. The drawbacks are the complicated experi-

mental set-up, relatively long analysis times, the in-

stability of the formed hydrides, and the high risk of

contaminations by the addition of high excesses of

solid and difficult-to-clean reagents (e.g., sodium

tetrahydridoborate, sodium hydroxide).

0052 The CVAAS for the mercury determination pro-

vides very good detection limits (0.001 mgl

1

). Even

enrichment is possible by using gold foils for

amalgam formation.

Problems and Interferences

0053In the CVAAS, the separation of the analytes of the

matrix is nearly complete for nonhydride formers.

However, the hydride-forming elements can strongly

interfere with the reduction of one another (e.g., As

interfering with the reduction of Sb). In addition, the

reduction process can be inhibited by oxidants or com-

plexing agents in the matrix or even by some transition

metals. The platinum group elements adsorb large

amounts of hydrogen, so that the effective reduction

ability of the solution is lowered. Standard addition

for calibration is therefore strongly recommended.

0054The volume of the sample must be constant. Foam

formation can influence the degassing of the analytes,

especially in biological matrices. This should be

suppressed.

Conclusion

0055Atomic spectrometry is suitable for trace concentra-

tions of most elements. The principle is the absorp-

tion and emission of light of a certain wavelength

after atomization of the sample. The wavelength

provides qualitative information about the element,

whereas the intensity of the emission (AES) or the

quotient of the intensity before and after the absorp-

tion (AAS) is proportional to the concentration of the

analyte. Several spectral and nonspectral interfer-

ences must be considered.

See also: Aluminum (Aluminium): Properties and

Determination; Copper: Properties and Determination;

Electrolytes: Analysis; Iron: Properties and

Determination; Lead: Properties and Determination;

Lithium; Mass Spectrometry: Principles and

Instrumentation; Mercury: Properties and Determination;

Mineral Water: Types of Mineral Water; Potassium:

Properties and Determination; Selenium: Properties and

Determination; Sodium: Properties and Determination;

Spectroscopy: Overview; Tin; Trace Elements; Water

Supplies: Chemical Analysis