AAS-GTA-reading-material.pdf

Concepts, Instrumentation and
Techniques in Atomic
Absorption Spectrophotometry
Richard D. Beaty
and
Jack D. Kerber
Second Edition
THE PERKIN-ELMER CORPORATION

Copyright © 1993 by The Perkin-Elmer Corporation, Norwalk, CT, U.S.A. All
rights reserved. Printed in the United States of America. No part of this publication
may be reproduced, stored in a retrieval system, or transmitted, in any form or by
any means, electronic, mechanical, photocopying, recording, or otherwise, with-
out the prior written permission of the publisher.
ii

ABOUT THE AUTHORS
Richard D. Beaty
Since receiving his Ph.D. degree in chemistry from the University of Missouri-
Rolla, Richard Beaty has maintained an increasing involvement in the field of
laboratory instrumentation and computerization. In 1972, he joined Perkin-Elmer,
where he held a variety of technical support and marketing positions in atomic
spectroscopy. In 1986, he founded Telecation Associates, a consulting company
whose mission was to provide formalized training and problem solving for the
analytical laboratory. He later became President and Chief Executive Officer of
Telecation, Inc., a company providing PC-based software for laboratory automat-
ion and computerization.
Jack D. Kerber
Jack Kerber is a graduate of the Massachusetts Institute of Technology. He has
been actively involved with atomic spectrometry since 1963. In 1965 he became
Perkin-Elmer’s first field Product Specialist in atomic absorption, supporting ana-
lysts in the western United States and Canada. Since relocating to Perkin-Elmer’s
corporate headquarters in 1969, he has held a variety of marketing support and
sales and product management positions. He is currently Director of Atomic Ab-
sorption Marketing for North and Latin America.
iii

ACKNOWLEDGEMENT
The authors gratefully acknowledge the contributions and assistance they have re-
ceived from their colleagues in preparing this book. We are particularly indebted
to Glen Carnrick, Frank Fernandez, John McCaffrey, Susan McIntosh, Charles
Schneider and Jane Sebestyen of The Perkin-Elmer Corporation for the hours they
spent proofreading the several revisions and to Jorn Baasner, Horst Schulze, Ger-
hard Schlemmer, Werner Schrader and Ian Shuttler of Bodenseewerk
Perkin-Elmer GmbH for their invaluable input on Zeeman-effect background cor-
rection and graphite furnace atomic absorption techniques.
iv

TABLE OF CONTENTS
1 Theoretical Concepts and Definitions
The Atom and Atomic Spectroscopy . . . . . . . . . . . . . . . 1-1
Atomic Absorption Process . . . . . . . . . . . . . . . . . . . . 1-3
Quantitative Analysis by Atomic Absorption . . . . . . . . . . . 1-4
Characteristic Concentration and Detection Limits . . . . . . . . 1-6
Characteristic Concentration . . . . . . . . . . . . . . . . . 1-7
Detection Limits . . . . . . . . . . . . . . . . . . . . . . . . 1-7
2 Atomic Absorption Instrumentation
The Basic Components . . . . . . . . . . . . . . . . . . . . . . . 2-1
AA Light Sources . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
The Hollow Cathode Lamp . . . . . . . . . . . . . . . . . . 2-3
The Electrodeless Discharge Lamp . . . . . . . . . . . . . . 2-6
Optical Considerations
Photometers . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Single-beam Photometers . . . . . . . . . . . . . . . . . . . 2-7
Double-beam Photometers . . . . . . . . . . . . . . . . . . 2-8
Alternative Photometer Designs . . . . . . . . . . . . . . . 2-9
Optics and the Monochromator System . . . . . . . . . . . 2-10
The Atomic Absorption Atomizer
Pre-mix Burner System . . . . . . . . . . . . . . . . . . . . 2-14
Impact Devices . . . . . . . . . . . . . . . . . . . . . . . . 2-15
Nebulizers, Burner Heads and Mounting Systems . . . . . . 2-16
Electronics
Precision in Atomic Absorption Measurements . . . . . . . 2-17
Calibration of the Spectrometer . . . . . . . . . . . . . . . 2-18
Automation of Atomic Absorption
Automated Instruments and Sample Changers . . . . . . . . 2-19
Automated Sample Preparation . . . . . . . . . . . . . . . . 2-20
The Stand-alone Computer and Atomic Absorption . . . . . 2-20
3 Control of Analytical Interferences
The Flame Process . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Nonspectral Interferences . . . . . . . . . . . . . . . . . . . . . 3-3
Matrix Interference . . . . . . . . . . . . . . . . . . . . . . 3-3
Method of Standard Additions . . . . . . . . . . . . . . . . 3-4
Chemical Interference . . . . . . . . . . . . . . . . . . . . . 3-5
v

3 Control of Analytical Interferences (continued)
Ionization Interference . . . . . . . . . . . . . . . . . . . . 3-6
Spectral Interferences
Background Absorption . . . . . . . . . . . . . . . . . . . . 3-7
Continuum Source Background Correction . . . . . . . . . 3-8
Introduction to Zeeman Background Correction . . . . . . . 3-11
Other Spectral Interferences . . . . . . . . . . . . . . . . . 3-14
Interference Summary . . . . . . . . . . . . . . . . . . . . . . . 3-14
4 High Sensitivity Sampling Systems
Limitations to Flame AA Sensitivity . . . . . . . . . . . . . . . 4-1
The Cold Vapor Mercury Technique
Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Advantages of the Cold Vapor Technique . . . . . . . . . . 4-2
Limitations of the Cold Vapor Technique . . . . . . . . . . 4-3
The Hydride Generation Technique
Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Advantages of the Hydride Technique . . . . . . . . . . . . 4-4
Disadvantages of the Hydride Technique . . . . . . . . . . 4-4
Graphite Furnace Atomic Absorption
Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
Advantages of the Graphite Furnace Technique . . . . . . . 4-5
5 Introduction to Graphite Furnace Atomic Absorption
Considerations in Ultra Trace Analysis
Performance Criteria . . . . . . . . . . . . . . . . . . . . . 5-1
Graphite Furnace Applications . . . . . . . . . . . . . . . . 5-2
Components of the Graphite Furnace System
The Graphite Furnace Atomizer . . . . . . . . . . . . . . . 5-2
The Graphite Furnace Power Supply and Programmer . . . 5-5
Summary of a Graphite Furnace Analysis . . . . . . . . . . . . . 5-5
Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
The Drying Step . . . . . . . . . . . . . . . . . . . . . . . . 5-7
The Pyrolysis Step . . . . . . . . . . . . . . . . . . . . . . 5-8
The Pre-atomization Cool Down Step . . . . . . . . . . . . 5-8
The Atomization Step . . . . . . . . . . . . . . . . . . . . . 5-8
The Clean Out and Cool Down Step . . . . . . . . . . . . . 5-9
Fast Furnace Analysis . . . . . . . . . . . . . . . . . . . . . . . 5-9
vi

5 Introduction to Graphite Furnace Atomic Absorption (continued)
Measuring the Graphite Furnace AA Signal
Nature of the Graphite Furnace Signal . . . . . . . . . . . . 5-10
Peak Height Measurement . . . . . . . . . . . . . . . . . . 5-10
Peak Area Measurement . . . . . . . . . . . . . . . . . . . 5-11
Solid Sampling With the Graphite Furnace . . . . . . . . . . . . 5-12
6 Control of Graphite Furnace Interferences
Interferences and the Graphite Furnace . . . . . . . . . . . . . . 6-1
Spectral Interferences
Emission Interference . . . . . . . . . . . . . . . . . . . . . 6-2
Background Absorption . . . . . . . . . . . . . . . . . . . . 6-3
Background Reduction Techniques . . . . . . . . . . . . . . 6-3
Automated Instrumental Background Correction . . . . . . 6-6
Interpolated Background Correction . . . . . . . . . . . . . 6-6
Nonspectral Interferences
Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Method of Standard Additions . . . . . . . . . . . . . . . . 6-8
The Graphite Tube Surface . . . . . . . . . . . . . . . . . . 6-9
The L’vov Platform . . . . . . . . . . . . . . . . . . . . . . 6-10
Matrix Modification . . . . . . . . . . . . . . . . . . . . . . 6-11
Maximum Power Atomization . . . . . . . . . . . . . . . . 6-12
Peak Area Measurement . . . . . . . . . . . . . . . . . . . 6-13
Fast Electronics and Baseline Offset Correction . . . . . . . 6-14
Stabilized Temperature Platform Furnace
The Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-15
The STPF System . . . . . . . . . . . . . . . . . . . . . . . 6-15
7 Alternate Analytical Techniques
Direct Current Plasma (DCP) Emission . . . . . . . . . . . . . . 7-1
Inductively Coupled Plasma (ICP) Emission . . . . . . . . . . . 7-2
Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) . . . 7-3
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-4
vii

1THEORETICAL CONCEPTS
AND DEFINITIONS
THE ATOM AND ATOMIC SPECTROSCOPY
The science of atomic spectroscopy has yielded three techniques for analytical
use: atomic emission, atomic absorption, and atomic fluorescence. In order to un-
derstand the relationship of these techniques to each other, it is necessary to have
an understanding of the atom itself and of the atomic process involved in each
technique.
The atom is made up of a nucleus surrounded by electrons. Every element has a
specific number of electrons which are associated with the atomic nucleus in an
orbital structure which is unique to each element. The electrons occupy orbital po-
sitions in an orderly and predictable way. The lowest energy, most stable electronic
configuration of an atom, known as the ‘‘ground state’’, is the normal orbital con-
figuration for an atom. If energy of the right magnitude is applied to an atom, the
energy will be absorbed by the atom, and an outer electron will be promoted to a
less stable configuration or ‘‘excited state’’. As this state is unstable, the atom will
immediately and spontaneously return to its ground state configuration. The elec-
tron will return to its initial, stable orbital position, and radiant energy equivalent
to the amount of energy initially absorbed in the excitation process will be emitted.
The process is illustrated in Figure 1-1. Note that in Step 1 of the process, the ex-
citation is forced by supplying energy. The decay process in Step 2, involving the
emission of light, occurs spontaneously.
Figure 1-1. Excitation and decay processes.

The wavelength of the emitted radiant energy is directly related to the electronic
transition which has occurred. Since every element has a unique electronic struc-
ture, the wavelength of light emitted is a unique property of each individual ele-
ment. As the orbital configuration of a large atom may be complex, there are many
electronic transitions which can occur, each transition resulting in the emission of
a characteristic wavelength of light, as illustrated in Figure 1-2.
The process of excitation and decay to the ground state is involved in all three
fields of atomic spectroscopy. Either the energy absorbed in the excitation process
or the energy emitted in the decay process is measured and used for analytical pur-
poses. In atomic emission, a sample is subjected to a high energy, thermal envi-
ronment in order to produce excited state atoms, capable of emitting light. The
energy source can be an electrical arc, a flame, or more recently, a plasma. The
emission spectrum of an element exposed to such an energy source consists of a
collection of the allowable emission wavelengths, commonly called emission
lines, because of the discrete nature of the emitted wavelengths. This emission
spectrum can be used as a unique characteristic for qualitative identification of the
element. Atomic emission using electrical arcs has been widely used in qualitative
analysis.
Emission techniques can also be used to determine how much of an element is pre-
sent in a sample. For a ‘‘quantitative’’analysis, the intensity of light emitted at the
wavelength of the element to be determined is measured. The emission intensity
at this wavelength will be greater as the number of atoms of the analyte element
increases. The technique of flame photometry is an application of atomic emission
for quantitative analysis.
If light of just the right wavelength impinges on a free, ground state atom, the atom
may absorb the light as it enters an excited state in a process known as atomic ab-
Figure 1-2. Energy transitions.
1-2 Concepts, Instrumentation and Techniques

sorption. This process is illustrated in Figure 1-3. Note the similarity between this
illustration and the one in Step 1 of Figure 1-1. The light which is the source of
atom excitation in Figure 1-3 is simply a specific form of energy. The capability
of an atom to absorb very specific wavelengths of light is utilized in atomic ab-
sorption spectrophotometry.
ATOMIC ABSORPTION PROCESS
The quantity of interest in atomic absorption measurements is the amount of light
at the resonant wavelength which is absorbed as the light passes through a cloud
of atoms. As the number of atoms in the light path increases, the amount of light
absorbed increases in a predictable way. By measuring the amount of light ab-
sorbed, a quantitative determination of the amount of analyte element present can
be made. The use of special light sources and careful selection of wavelength al-
low the specific quantitative determination of individual elements in the presence
of others.
The atom cloud required for atomic absorption measurements is produced by sup-
plying enough thermal energy to the sample to dissociate the chemical compounds
into free atoms. Aspirating a solution of the sample into a flame aligned in the light
beam serves this purpose. Under the proper flame conditions, most of the atoms
will remain in the ground state form and are capable of absorbing light at the ana-
lytical wavelength from a source lamp. The ease and speed at which precise and
accurate determinations can be made with this technique have made atomic ab-
sorption one of the most popular methods for the determination of metals.
A third field in atomic spectroscopy is atomic fluorescence. This technique incor-
porates aspects of both atomic absorption and atomic emission. Like atomic ab-
sorption, ground state atoms created in a flame are excited by focusing a beam of
light into the atomic vapor. Instead of looking at the amount of light absorbed in
the process, however, the emission resulting from the decay of the atoms excited
by the source light is measured. The intensity of this ‘‘fluorescence’’ increases
with increasing atom concentration, providing the basis for quantitative determi-
nation.
Figure 1-3. The atomic absorption process.
Theoretical Concepts and Definitions 1-3

The source lamp for atomic fluorescence is mounted at an angle to the rest of the
optical system, so that the light detector sees only the fluorescence in the flame
and not the light from the lamp itself. It is advantageous to maximize lamp inten-
sity with atomic fluorescence since sensitivity is directly related to the number of
excited atoms which is a function of the intensity of the exciting radiation.
Figure 1-4 illustrates how the three techniques just described are implemented.
While atomic absorption is the most widely applied of the three techniques and
usually offers several advantages over the other two, particular benefits may be
gained with either emission or fluorescence in special analytical situations. This
is especially true of emission, which will be discussed in more detail in a later
chapter.
QUANTITATIVE ANALYSIS BY ATOMIC ABSORPTION
The atomic absorption process is illustrated in Figure 1-5. Light at the resonance
wavelength of initial intensity, Io, is focused on the flame cell containing ground
state atoms. The initial light intensity is decreased by an amount determined by
the atom concentration in the flame cell. The light is then directed onto the detector
where the reduced intensity, I, is measured. The amount of light absorbed is de-
termined by comparing I to Io.
Figure 1-4. Atomic spectroscopy systems.

Several related terms are used to define the amount of light absorption which has
taken place. The ‘‘transmittance’’ is defined as the ratio of the final intensity to
the initial intensity.
T = I/Io
Transmittance is an indication of the fraction of initial light which passes through
the flame cell to fall on the detector. The ‘‘percent transmission’’ is simply the
transmittance expressed in percentage terms.
%T = 100 x I/Io
The ‘‘percent absorption’’ is the complement of percent transmission defining the
percentage of the initial light intensity which is absorbed in the flame.
%A = 100 - %T
These terms are easy to visualize on a physical basis. The fourth term, ‘‘absor-
bance’’, is purely a mathematical quantity.
A = log (Io/I)
Absorbance is the most convenient term for characterizing light absorption in ab-
sorption spectrophotometry, as this quantity follows a linear relationship with con-
centration. Beer’s Law defines this relationship:
A = abc
Figure 1-5. The atomic absorption process.

where ‘‘A’’ is the absorbance; ‘‘a’’ is the absorption coefficient, a constant which
is characteristic of the absorbing species at a specific wavelength; ‘‘b’’is the length
of the light path intercepted by the absorption species in the absorption cell; and
‘‘c’’ is the concentration of the absorbing species. This equation simply states that
the absorbance is directly proportional to the concentration of the absorbing spe-
cies for a given set of instrumental conditions.
This directly proportional behavior be-
tween absorbance and concentration is ob-
served in atomic absorption. When the ab-
sorbances of standard solutions containing
known concentrations of analyte are meas-
ured and the absorbance data are plotted
against concentration, a calibration rela-
tionship similar to that in Figure 1-6 is es-
tablished. Over the region where the
Beer’s Law relationship is observed, the
calibration yields a straight line. As the
concentration and absorbance increase,
nonideal behavior in the absorption proc-
ess can cause a deviation from linearity, as
shown.
After such a calibration is established, the absorbance of solutions of unknown
concentrations may be measured and the concentration determined from the cali-
bration curve. In modern instrumentation, the calibration can be made within the
instrument to provide a direct readout of unknown concentrations. Since the ad-
vent of microcomputers, accurate calibration, even in the nonlinear region, is sim-
ple.
CHARACTERISTIC CONCENTRATION AND DETECTION LIMITS
Characteristic concentration and detection limit are terms which describe instru-
ment performance characteristics for an analyte element. While both parameters
depend on the absorbance observed for the element, each defines a different per-
formance specification, and the information to be gained from each term is dif-
ferent.
Figure 1-6. Concentration versus
absorbance.

Characteristic Concentration
The ‘‘characteristic concentration’’ (sometimes called ‘‘sensitivity’’) is a conven-
tion for defining the magnitude of the absorbance signal which will be produced
by a given concentration of analyte. For flame atomic absorption, this term is ex-
pressed as the concentration of an element in milligrams per liter (mg/L) required
to produce a 1% absorption (0.0044 absorbance) signal.
Char Conc. (mg/L) = Conc. of Std. (mg/L) x 0.0044
measured absorbance
As long as measurements are made in the linear working region, the characteristic
concentration of an element can be determined by reading the absorbance pro-
duced by a known concentration of the element and using the above equation.
There are several practical reasons for wanting to know the value of the charac-
teristic concentration for an element. Knowing the expected characteristic concen-
tration allows an operator to determine if all instrumental conditions are optimized
and if the instrument is performing up to specifications by simply measuring the
absorbance of a known concentration and comparing the results to the expected
value. A known characteristic concentration value also allows one to predict the
absorbance range which will be observed from a known concentration range or
to determine the concentration range which would produce optimum absorbance
levels.
Detection Limits
It should be noted that, while the magnitude of the absorbance signal can be pre-
dicted from the value given for characteristic concentration, no information is
given on how small of an absorbance signal can be measured. Therefore, it is not
possible to predict the minimum measurable concentration from a known charac-
teristic concentration value. To determine this quantity, more information on the
nature of the measured absorbance signal must be considered.
The smallest measurable concentration of an element will be determined by the
magnitude of absorbance observed for the element (characteristic concentration)
and the stability of the absorbance signal. An unstable or ‘‘noisy’’ signal makes it
more difficult to distinguish small changes in observed absorbance which are due
to small concentration differences, from those random variations due to ‘‘baseline
noise.’’ Figure 1-7 illustrates the concept of the effect of noise on the quantitation
of small absorbance signals. Signal "A" and signal "B" have the same magnitude.

However, the much lower
variability ("noise") of sig-
nal "B" permits even
smaller signals to be de-
tected. The sensitivity of the
two signals is the same, but
there is a real difference in
detection limits.
The term ‘‘detection limit’’
incorporates a considera-
tion of both signal size and baseline noise to give an indication of the lowest con-
centration of an element which can be measured. The detection limit is defined
by the IUPAC as the concentration which will give an absorbance signal three
times the magnitude of the baseline noise. The baseline noise may be statistically
quantitated typically by making 10 or more replicate measurements of the baseline
absorbance signal observed for an analytical blank, and determining the standard
deviation of the measurements. The detection limit is then defined as the concen-
tration which will produce an absorbance signal three times the standard deviation
of the blank.
Routine analytical measurements at the detection limit are difficult, due to the fact
that, by definition, noise makes up a significant percentage of the total measurable
signal. By definition, the precision obtained at detection limit levels is ± 33% RSD
(relative standard deviation) when a three standard deviation criterion is used.
Therefore, while it is possible to distinguish analyte concentrations at the detection
limit from zero, for good precision it is necessary to limit routine analytical work
to concentrations higher than the detection limit.
Figure 1-7. AA measurements near detection
limits.

2ATOMIC ABSORPTION
INSTRUMENTATION
THE BASIC COMPONENTS
To understand the workings of the atomic absorption spectrometer, let us build
one, piece by piece. Every absorption spectrometer must have components which
fulfill the three basic requirements shown in Figure 2-1. There must be: (1) a light
source; (2) a sample cell; and (3) a means of specific light measurement.
In atomic absorption, these functional areas are implemented by the components
illustrated in Figure 2-2. A light source which emits the sharp atomic lines of the
element to be determined is required. The most widely used source is the hollow
cathode lamp. These lamps are designed to emit the atomic spectrum of a particu-
lar element, and specific lamps are selected for use depending on the element to
be determined.
Figure 2-1. Requirements for a spectrometer.
Figure 2-2. Basic AA spectrometer.

It is also required that the source radiation be modulated (switched on and off rap-
idly) to provide a means of selectively amplifying light emitted from the source
lamp and ignoring emission from the sample cell. Source modulation can be ac-
complished with a rotating chopper located between the source and the sample
cell, or by pulsing the power to the source.
Special considerations are also required for a sample cell for atomic absorption.
An atomic vapor must be generated in the light beam from the source. This is gen-
erally accomplished by introducing the sample into a burner system or electrically
heated furnace aligned in the optical path of the spectrophotometer.
Several components are required for specific light measurement. A monochroma-
tor is used to disperse the various wavelengths of light which are emitted from the
source and to isolate the particular line of interest. The selection of a specific
source and a particular wavelength in that source is what allows the determination
of a selected element to be made in the presence of others.
The wavelength of light which is isolated by the monochromator is directed onto
the detector, which serves as the ‘‘eye’’ of the instrument. This is normally a
photomultiplier tube, which produces an electrical current dependent on the light
intensity. The electrical current from the photomultiplier is then amplified and
processed by the instrument electronics to produce a signal which is a measure of
the light attenuation occurring in the sample cell. This signal can be further proc-
essed to produce an instrument readout directly in concentration units.
LIGHT SOURCES
An atom absorbs light at discrete wavelengths. In order to measure this narrow
light absorption with maximum sensitivity, it is necessary to use a line source,
which emits the specific wavelengths which can be absorbed by the atom. Narrow
line sources not only provide high sensitivity, but also make atomic absorption a
very specific analytical technique with few spectral interferences. The two most
common line sources used in atomic absorption are the ‘‘hollow cathode lamp’’
and the ‘‘electrodeless discharge lamp.’’
The Hollow Cathode Lamp
The hollow cathode lamp is an excellent, bright line source for most of the ele-
ments determinable by atomic absorption. Figure 2-3 shows how a hollow cathode
lamp is constructed. The cathode of the lamp frequently is a hollowed-out cylinder

of the metal whose spectrum is to be produced. The anode and cathode are sealed
in a glass cylinder normally filled with either neon or argon at low pressure. At
the end of the glass cylinder is a window transparent to the emitted radiation.
The emission process is illustrated in Figure 2-4. When an electrical potential is
applied between the anode and cathode, some of the fill gas atoms are ionized. The
positively charged fill gas ions accelerate through the electrical field to collide
with the negatively charged cathode and dislodge individual metal atoms in a proc-
ess called ‘‘sputtering’’. Sputtered metal atoms are then excited to an emission
state through a kinetic energy transfer by impact with fill gas ions.
Hollow cathode lamps have a finite lifetime. Adsorption of fill gas atoms onto the
inner surfaces of the lamp is the primary cause for lamp failure. As fill gas pressure
decreases, the efficiency of sputtering and the excitation of sputtered metal atoms
also decreases, reducing the intensity of the lamp emission. To prolong hollow
cathode lamp life, some manufacturers produce lamps with larger internal vol-
umes so that a greater supply of fill gas at optimum pressure is available.
Figure 2-3. Hollow cathode lamp.
Figure 2-4. Hollow cathode lamp process, where Ar+
is a positively-charged ar-
gon ion, Mo
is a sputtered, ground-state metal atom, M* is an excited-state metal
atom, and l is emitted radiation at a wavelength characteristic for the sputtered
metal.
Atomic Absorption Instrumentation 2-3

The sputtering process may remove some of the metal atoms from the vicinity of
the cathode to be deposited elsewhere. Lamps for volatile metals such as arsenic,
selenium, and cadmium are more prone to rapid vaporization of the cathode during
use. While the loss of metal from the cathode at normal operating currents (typi-
cally 5-25 milliamperes) usually does not affect lamp performance, fill gas atoms
can be entrapped during the metal deposition process which does affect lamp life.
Lamps which are operated at highly elevated currents may suffer reduced lamp
life due to depletion of the analyte element from the cathode.
Some cathode materials can slowly evolve hydrogen when heated. As the concen-
tration of hydrogen in the fill gas increases, a background continuum emission
contaminates the purity of the line spectrum of the element, resulting in a reduction
of atomic absorption sensitivity and poor calibration linearity. To eliminate such
problems, most modern hollow cathode lamps have a tantalum ‘‘getter’’on the an-
ode which irreversibly adsorbs evolved hydrogen as the lamp is operated.
The cathode of the hollow cathode lamp is usually constructed from a highly pure
metal resulting in a very pure emission spectrum of the cathode material. It is
sometimes possible, however, to construct a cathode or cathode insert from several
metals. The resulting ‘‘multi-element’’ lamp may provide superior performance
for a single element or, with some combinations, may be used as a source for all
of the elements contained in the cathode alloy. However, not all metals may be
used in combination with others because of metallurgical or spectral limitations.
Special consideration should be given before using a multi-element lamp as ana-
lytical complications may result. Often the intensity of emission for an element
in a multi-element lamp is not as great as that which is observed for the element
in a single-element lamp. This loss of intensity could be a disadvantage in appli-
cations where high precision or low detection limits are required. The increased
spectral complexity of multi-element lamps may require that alternate wave-
lengths or narrower slits be used, which may also adversely affect sensitivity or
baseline noise.
Each hollow cathode lamp will have a particular current for optimum perform-
ance. In general, higher currents will produce brighter emission and less baseline
noise. As the current continues to increase, however, lamp life may shorten and
spectral line broadening may occur, resulting in a reduction in sensitivity and lin-
ear working range. The recommended current specified for each lamp will usually
provide the best combination of lamp life and performance. For demanding analy-

ses requiring the best possible signal-to-noise characteristics, higher currents can
be used for the lamp, up to the maximum rated value. Lower lamp currents can
be used with less demanding analyses to prolong lamp life.
Confusion over exactly what current is being used for a hollow cathode lamp may
occur due to the method used for lamp modulation. As explained earlier, the source
for atomic absorption must be modulated in order to accomplish selective ampli-
fication of the lamp emission signal. This can be accomplished mechanically by
using a rotating chopper or electronically by pulsing the current supplied to the
lamp, as illustrated in Figure 2-5. Both methods produce similar results; however,
in some instruments the use of electronic modulation may create the impression
that a lower lamp current is being applied than is actually the case.
The cause for the apparent difference in measured currents with mechanically and
electronically modulated systems is also shown in Figure 2-5. For mechanical
modulation, the lamp is run at a constant current. Under these conditions, an am-
meter reading of lamp current will indicate the actual current flow. For electronic
modulation, the current is switched on and off at a rapid rate. An ammeter nor-
mally will indicate the time-averaged current rather than the actual peak current
which is being applied.
While some instruments are designed to apply a correction factor automatically
to electronically modulated lamp current readings to provide true peak current val-
ues, many do not. For electronically modulated systems without such correction,
the actual peak current can be approximated from the measured current by divid-
Figure 2-5. Mechanical vs. electrical modulation.

ing it by the ‘‘duty cycle’’, the fraction of time that current is applied to the lamp.
For example, for a duty cycle of 40% and a measured lamp current of 10 milliam-
peres, the actual peak operating current for an electronically modulated system is:
10 milliamperes/0.4 = 25 milliamperes
Specified lamp current settings may appear to be lower for atomic absorption in-
struments which modulate the source electronically and do not apply correction.
The only valid basis of comparison between the current settings used by two dif-
ferent systems is one which includes compensation for the duty cycle, as shown
above.
The Electrodeless Discharge Lamp
For most elements, the hollow cathode lamp is a completely satisfactory source
for atomic absorption. In a few cases, however, the quality of the analysis is im-
paired by limitations of the hollow cathode lamp. The primary cases involve the
more volatile elements where low intensity and short lamp life are a problem. The
atomic absorption determination of these elements can often be dramatically im-
proved with the use of brighter, more stable sources such as the ‘‘electrodeless dis-
charge lamp’’.
Figure 2-6 shows the design of the Perkin-Elmer System 2 electrodeless discharge
lamp (EDL). A small amount of the metal or salt of the element for which the
source is to be used is sealed inside a quartz bulb. This bulb is placed inside a small,
self-contained RF generator or ‘‘driver’’. When power is applied to the driver, an
RF field is created. The coupled energy will vaporize and excite the atoms inside
the bulb, causing them to emit their characteristic spectrum. An accessory power
supply is required to operate an EDL.
Figure 2-6. Electrodeless discharge lamp.

Electrodeless discharge lamps are typically much more intense and, in some cases,
more sensitive than comparable hollow cathode lamps. They therefore offer the
analytical advantages of better precision and lower detection limits where an
analysis is intensity limited. In addition to providing superior performance, the
useful lifetime of an EDL is typically much greater than that of a hollow cathode
lamp for the same element. It should be noted, however, that the optical image for
the EDL is considerably larger than that in a hollow cathode lamp. As a result, the
performance benefits of the EDL can only be observed in instruments with optical
systems designed to be compatible with the larger image.
Electrodeless discharge lamps are available for a wide variety of elements, includ-
ing antimony, arsenic, bismuth, cadmium, cesium, germanium, lead, mercury,
phosphorus, potassium, rubidium, selenium, tellurium, thallium, tin and zinc.
OPTICAL CONSIDERATIONS
Photometers
The portion of an atomic absorption spectrometer’s optical system which conveys
the light from the source to the monochromator is referred to as the photometer.
Three types of photometers are typically used in atomic absorption instruments:
single-beam, double-beam and what might be called compensated single-beam or
pseudo double-beam.
Single-Beam Photometers
The instrument diagrammed in Figure 2-7 represents a fully functional ‘‘single-
beam’’ atomic absorption spectrometer. It is called ‘‘single-beam’’ because all
measurements are based on the varying intensity of a single beam of light in a sin-
gle optical path.
Figure 2-7. A single-beam AA spectrometer.

The primary advantage of a single-beam configuration is that it has fewer com-
ponents and is less complicated than alternative designs. It is therefore easier to
construct and less expensive than other types of photometers. With a single light
path and a minimum number of optical components, single-beam systems typi-
cally provide very high light throughput. The primary limitation of the single-
beam photometer is that it provides no means to compensate for instrumental
variations during an analysis, such as changes in source intensity. The resulting
signal variability can limit the performance capabilities of a single-beam system.
Double-Beam Photometers
An alternate photometer configuration, known as ‘‘double-beam’’ (Figure 2-8)
uses additional optics to divide the light from the lamp into a ‘‘sample beam’’ (di-
rected through the sample cell) and a ‘‘reference beam’’(directed around the sam-
ple cell). In the double-beam system, the reference beam serves as a monitor of
lamp intensity and the response characteristics of common electronic circuitry.
Therefore, the observed absorbance, determined from a ratio of sample beam and
reference beam readings, is more free of effects due to drifting lamp intensities
and other electronic anomalies which similarly affect both sample and reference
beams.
Modern atomic absorption spectrometers are frequently highly automated. They
can automatically change lamps, reset instrument parameters, and introduce sam-
ples for high throughput multielement analysis. Double-beam technology, which
automatically compensates for source and common electronics drift, allows these
instruments to change lamps and begin an analysis immediately with little or no
Figure 2-8. A double-beam AA spectrometer.

lamp warm-up for most elements. This not only reduces analysis time but also pro-
longs lamp life, since lamp warm-up time is eliminated. Even with manual analy-
ses, the ability to install a lamp or turn on the instrument and start an analysis
almost immediately is a decided advantage for double-beam systems.
Double-beam photometers do divert some source energy from the sample beam
to create the reference beam. Since it is the signal:noise ratio of the sample beam
which determines analytical performance, modern double-beam instruments typi-
cally devote a much higher percentage of the source emission to the sample beam
than to the reference beam. For example, a modern double-beam system which
uses a beam splitter to generate sample and reference beams may use 75% of the
source emission for the sample measurement and only 25% for the reference meas-
urement. Using such techniques, modern double-beam instruments offer virtually
the same signal-to-noise ratio as single-beam systems while enjoying the high-
speed automation benefits and operational simplicity of double-beam operation.
Alternative Photometer Designs
There are several alternative system designs which provide advantages similar to
those of double-beam optical systems and the light throughput characteristic of
single-beam systems. Such systems can be described as compensated single-beam
or pseudo double-beam systems. One such design uses two mechanically-adjusted
mirrors to alternately direct the entire output of the source through either the sam-
ple path (during sample measurements) or through a reference path (Figures 2-9
and 2-10).
These alternative photometer designs provide light throughput comparable to that
provided by single-beam photometer systems. They also compensate for system
variations in a manner similar to that of double-beam photometers----similar, but
not the same. This type of photometer performs compensation for drift much less
frequently than do double-beam systems, typically only once per analytical read-
ing. Double-beam systems typically provide drift compensation at rates in excess
of 50 times per second. The lower compensation frequency limits the ability of
alternative photometer systems to compensate for large, quickly changing vari-
ations in source intensity such as those that frequently occur when a source is first
lighted.

Optics and the Monochromator System
As previously discussed, an important factor in determining the baseline noise in
an atomic absorption instrument is the amount of light energy reaching the
photomultiplier (PMT). Lamp intensity is optimized to be as bright as possible
while avoiding line broadening problems. The impact of single-beam and double-
beam photometer systems has been discussed above. But the impact of other com-
ponents must also be considered to determine the capabilities of the complete
optical system.
Figure 2-9. A compensated single-beam system with source light directed
through the sample path.
Figure 2-10. A compensated single-beam system with source light directed
through the reference path.

Light from the source must be focused on the sample cell and directed to the mono-
chromator, where the wavelengths of light are dispersed and the analytical line of
interest is focused onto the detector. Some energy is lost at each optical surface
along the way. Front-surfaced, highly reflective, mirrors can be used to control the
focus of the source lamp and the field of view of the light detector precisely and
with minimal light loss. Alternately, focusing can be accomplished by refraction
instead of reflection by using a lens system. Since the focal length of a lens varies
with wavelength, additional optics (which may further reduce energy throughput)
or complex optical adjustments must be used to obtain proper focus over the full
spectral range for atomic absorption.
Particular care must be taken in the monochromator to avoid excessive light loss.
A typical monochromator is diagrammed in Figure 2-11. Wavelength dispersion
is accomplished with a grating, a reflective surface ruled with many fine parallel
lines very close together. Reflection from this ruled surface generates an interfer-
ence phenomenon known as diffraction, in which different wavelengths of light
diverge from the grating at different angles. Light from the source enters the mono-
chromator at the entrance slit and is directed to the grating where dispersion takes
place. The diverging wavelengths of light are directed toward the exit slit. By ad-
justing the angle of the grating, a selected emission line from the source can be
allowed to pass through the exit slit and fall onto the detector. All other lines are
blocked from exiting.
Figure 2-11. A monochromator.

The angle of dispersion at the grating can be controlled by the density of lines on
the grating. Higher dispersion will result from greater line density, i.e., more
lines/mm. High dispersion is important to good energy efficiency of the mono-
chromator, as illustrated in Figure 2-12.
The image of the source focused on the entrance slit and dispersed emission lines
at the exit slit are shown for both a low-dispersion and a high-dispersion grating.
In order to isolate a desired line from nearby lines, it is necessary to use a narrower
exit slit in the low-dispersion example than is required in the high-dispersion case.
Good optical design practices dictate that the entrance and exit slits be similarly
sized. The use of a larger entrance slit will overfill the grating with the source im-
age, while the use of a smaller entrance slit restricts the amount of light entering
the monochromator. Both reduce the amount of energy available at the exit slit.
For a low dispersion grating, this means that the size of the monochromator en-
trance slit is limited to the narrow size demanded of the exit slit to exclude nearby
lines. Thus, much of the available light energy is prevented from ever entering the
monochromator. In contrast, the greater wavelength separation provided by a
high-dispersion grating allows the use of wider slits, which make use of more of
the available light without any sacrifice in resolution.
To a first approximation, the energy throughput of a monochromator is propor-
tional to the illuminated ruled grating area and inversely proportional to the re-
ciprocal linear dispersion. To obtain the full energy benefit of high dispersion, it
is necessary to use a grating with a ruled surface area large enough to capture all
of the light from the magnified slit image. Large, quality gratings of high disper-
sion are difficult and expensive to make. Therefore, the incentive is great to accept
smaller gratings with lesser line densities and poorer dispersion for atomic absorp-
tion instrumentation. However, better instruments take advantage of the superior
energy throughput afforded by larger gratings.
Figure 2-12. Advantages of high dispersion.

Another factor affecting the optical efficiency of the monochromator is the blaze
angle of the grating, whether it is mechanically ruled or holographically generated.
An illustration of a mechanically-ruled blaze angle appears in Figure 2-13.
Mechanical grating rulings are in the form of V-shaped grooves carved into the
surface of the grating. As discussed earlier, an interference phenomenon causes
light of different wavelengths to diverge from the grating at different angles. The
particular wavelength which diverges from the blazed surface at an angle corre-
sponding to specular reflectance (i.e., angle of reflection equals angle of inci-
dence) will suffer the least loss in intensity as a result of the diffraction process.
A grating can be constructed for a blaze at any desired wavelength by controlling
the angle of cut during ruling. The farther removed a given wavelength of light is
from the wavelength for which a grating is blazed, the greater will be the extent
of monochromator light loss at that wavelength.
The useful atomic absorption wavelength range runs from 189 to 851 nanometers.
With one grating blazed somewhere in the middle of this range, significant energy
fall-off occurs at the wavelength extremities due to energy inefficiencies in the dif-
fraction process. One technique used to overcome this problem and to provide en-
hanced energy throughput at the wavelength extremities is to equip the instrument
with two gratings, one blazed in the ultraviolet and the other blazed in the visible
region of the spectrum. Then by choosing the grating blazed nearest the working
Figure 2-13. Grating blaze angle.

wavelength, the optimum energy throughput can be achieved. Alternately, a single
‘‘dual-blazed’’ grating can be used, with two regions blazed for the two spectral
regions. As the dual blazed grating rotates from one wavelength extreme to an-
other, the region blazed for the current working wavelength is brought into align-
ment with the optical beam, thereby offering improved efficiency compared with
a single grating blazed at one wavelength.
THE ATOMIC ABSORPTION ATOMIZER
Pre-Mix Burner System
The sample cell, or atomizer, of the spectrometer must produce the ground state
atoms necessary for atomic absorption to occur. This involves the application of
thermal energy to break the bonds that hold atoms together as molecules. While
there are several alternatives, the most routine and widely applied sample atomizer
is the flame.
Figure 2-14 shows an exploded view of an atomic absorption burner system. In
this ‘‘premix’’design, sample solution is aspirated through a nebulizer and sprayed
as a fine aerosol into the mixing chamber. Here the sample aerosol is mixed with
fuel and oxidant gases and carried to the burner head, where combustion and sam-
ple atomization occur.
Figure 2-14. Premix burner system.

Fuel gas is introduced into the mixing chamber through the fuel inlet, and oxidant
enters through the nebulizer sidearm. Mixing of the fuel and oxidant in the burner
chamber eliminates the need to have combustible fuel/oxidant in the gas lines, a
potential safety hazard. In addition to the separate fuel and oxidant lines, it is ad-
vantageous to have an auxiliary oxidant inlet directly into the mixing chamber.
This allows the oxidant flow adjustments to be made through the auxiliary line
while the flow through the nebulizer remains constant. Thus, for a burner system
with an auxiliary oxidant line, the sample uptake rate is independent of flame con-
dition, and the need to readjust the nebulizer after every oxidant flow adjustment
is eliminated.
Only a portion of the sample solution introduced into the burner chamber by the
nebulizer is used for analysis. The finest droplets of sample mist, or aerosol, are
carried with the combustion gases to the burner head, where atomization takes
place. The excess sample is removed from the premix chamber through a drain.
The drain uses a liquid trap to prevent combustion gases from escaping through
the drain line. The inside of the burner chamber is coated with a wettable inert plas-
tic material to provide free drainage of excess sample and prevent burner chamber
‘‘memory.’’ A free draining burner chamber rapidly reaches equilibrium, usually
requiring less than two seconds for the absorbance to respond fully to sample
changes.
Impact Devices
The sample aerosol is composed of variously sized droplets as it is sprayed into
the mixing chamber. Upon entering the flame, the water in these droplets is va-
porized. The remaining solid material must likewise be vaporized, and chemical
bonds must be broken to create free ground state atoms. Where the initial droplet
size is large, the sample vaporization and atomization process is more difficult to
complete in the short time in which the sample is exposed to the flame. Incomplete
sample vaporization and atomization will lead to increased susceptibility to ana-
lytical interferences.
Impact devices are used to reduce droplet size further and to cause remaining
larger droplets to be deflected from the gas stream and removed from the burner
through the drain. Two types of impact device are used typically, impact beads and
flow spoilers.

Impact bead systems are normally used to improve nebulization efficiency, the
percentage of sample solution converted to smaller droplets. The impact bead is
normally a spherical bead made of glass, silica or ceramic. Glass or quartz impact
beads may be less corrosion resistant and may cause more contamination problems
than more chemically inert ceramic beads.
The impact bead is positioned directly in the nebulizer spray as it exits the nebu-
lizer. The sample spray contacts the impact bead at high speed, causing some of
the larger droplets to be broken up into smaller droplets. The design and position-
ing of the impact bead are critical in determining how well it will work. Properly
designed impact bead systems will improve nebulization efficiency and remove
many of the remaining large droplets from the spray. However, poorly designed
or positioned impact beads may have little or no effect on nebulization efficiency
and may be very inefficient at removing larger droplets from the spray. The in-
creased population of large droplets in the aerosol may create undesirable effects,
including poorer precision and increased interferences. Additionally, burner sys-
tems using an impact bead may exhibit memory problems with high concentration
solutions or solutions with high dissolved solids content.
The quality of an impact bead system can frequently be determined by the increase
in sensitivity it provides for selected elements. A poorly designed system will pro-
vide improved sensitivity for easily atomized elements simply because more sam-
ple is transported to the flame and less to the drain. However, there normally will
be little or no improvement in sensitivity for the less easily atomized elements. A
properly designed impact bead system will provide improved nebulization effi-
ciency and improved sensitivity for all elements.
Flow spoilers normally do not improve nebulization efficiency. The primary use
of a flow spoiler is to remove the remaining large droplets from the sample aerosol.
The flow spoilers used in atomic absorption burner systems normally are placed
between the nebulizer and the burner head. They typically have three or more large
vanes constructed from or coated with a corrosion resistant material. Smaller drop-
lets are transported through the open areas between the vanes while larger droplets
contact the vanes and are removed from the aerosol.
For routine atomic absorption analyses where maximum sensitivity is not re-
quired, use of an efficient flow spoiler alone will provide the required analytical
stability and freedom from interference. A burner system optimized for maximum
sensitivity and performance should include both a high nebulization efficiency ce-
ramic impact bead and an efficient flow spoiler.

Nebulizers, Burner Heads and Mounting Systems
Several important factors enter into the nebulizer portion of the burner system. In
order to provide efficient nebulization for all types of sample solution, the nebu-
lizer should be adjustable. Stainless steel has been the most common material used
for construction of the nebulizer. Stainless steel has the advantage of durability
and low cost but has the disadvantage of being susceptible to corrosion from sam-
ples with a high content of acid or other corrosive reagents. For such cases, nebu-
lizers constructed of a corrosion resistant material, such as an inert plastic,
platinum alloys or tantalum should be used.
Burner heads typically are constructed of stainless steel or titanium. All-titanium
heads are preferred as they provide extreme resistance to heat and corrosion.
Different burner head geometries are required for various flame or sample condi-
tions. A ten-centimeter single-slot burner head is recommended for air-acetylene
flames. A special five-centimeter burner head with a narrower slot is required
when a nitrous oxide-acetylene flame is to be used. Burner heads also are available
for special purposes, such as use with solutions that have exceptionally high dis-
solved solids contents.
In addition to the flame, there are several options for atomic absorption atomizers.
These options are discussed in detail in Chapter 4. Most of these options require
removal of the premix burner system and replacement by an alternate atomizer in
the spectrometer sample compartment. Since these alternate atomizers offer com-
plementary and extended analytical capabilities, it is likely that the analyst will
want to alternate between the use of flame AA and one or more of the other sys-
tems. A ‘‘quick change’’atomizer mount is an important item to facilitate conven-
ient changeover from one device to another without the use of tools. In addition
to convenience, a ‘‘quick change’’ mount may reduce or eliminate entirely the
need for realignment of the atomizer when it is replaced in the sample compart-
ment.
ELECTRONICS
Precision in Atomic Absorption Measurements
We have already discussed the effects of light energy on the precision of an atomic
absorption measurement. The analyst will have little control over these optical fac-
tors, as they are an inherent part of the instrument design. However, the analyst
can exercise some degree of control over precision by proper selection of integra-
tion time with flame atomic absorption.

Observed precision will improve with the period of time over which each sample
is read. Where analyte concentrations are not approaching detection limits, inte-
gration times of one to three seconds will usually provide acceptable precision.
When approaching instrument detection limits where repeatability is poor, preci-
sion can be improved by using even longer integration times, up to 10 seconds.
In most instances; however, there is little advantage to using integration times
longer than 10 seconds. (To a first approximation, improvement in signal:noise ra-
tio is proportional to the square root of the ratio of integration times.)
Since the detection limit is defined based on the observed precision, the detection
limit also can be improved by increasing the integration time. The analyst, there-
fore, has control over the priorities for a particular analysis, maximum speed or
optimum precision and detection limits.
Current instruments offer statistical functions for averaging and calculating stand-
ard deviation and relative standard deviation or coefficient of variation of replicate
measurements. These functions can be used to determine the precision under vari-
ous experimental conditions, allowing the analyst to optimize method parameters
for each individual requirement.
Calibration of the Spectrometer
Most modern atomic absorption instruments include microcomputer-based elec-
tronics. The microcomputer provides atomic absorption instruments with ad-
vanced calculation capabilities, including the ability to calibrate and compute
concentrations from absorbance data conveniently and accurately, even for non-
linear calibration curves. In the linear region, data on as little as one standard and
a blank may be sufficient for defining the relationship between concentration and
absorbance. However, additional standards are usually used to verify calibration
accuracy. Where the relationship becomes nonlinear, however, more standards are
required. The accuracy of a calibration computed for a nonlinear relationship de-
pends on the number of standards and the equations used for calibration.
For the equation format which optimally fits atomic absorption data, it has been
experimentally shown that accurate calibration can be achieved with a minimum
of three standards plus a blank, even in cases of severe curvature. Figure 2-15 il-
lustrates the accuracy of microcomputer-calculated results for cobalt with single
standard ‘‘linear’’ and three-standard ‘‘nonlinear’’ calibrations. After the instru-
ment was calibrated using the specified procedure, a series of standards were ana-
lyzed. Figure 2-15 shows plots of the actual concentrations for those standards
versus the measured values for both calibration procedures. The results obtained

with ‘‘linear’’ calibration are accurate only where the absorbance:concentration
relationship is linear, up to about 5 m
g/mL. The results obtained with three-stand-
ard ‘‘nonlinear’’ calibration are still accurate at 30 m
g/mL, significantly extending
the useful working range. For versatility, current instruments allow fitting of more
than three standards to these same basic equations.
AUTOMATION OF ATOMIC ABSORPTION
Automated Instruments and Sample Changers
One of the greatest contributions to the efficiency of the analytical laboratory is
the automated atomic absorption spectrometer. Automatic samplers were the first
step in freeing the analyst from the monotonous task of manually introducing each
and every sample.
Figure 2-15. Cobalt Calibration Accuracy

However, the real advancement in analysis automation came in the late 1970’s,
when automated multielement atomic absorption was introduced. In addition to
automatic sample introduction, these instruments offer automatic setup of instru-
ment parameters to preprogrammed values. These instrument ‘‘programs’’ can be
accessed sequentially, making it possible to analyze a tray full of samples for mul-
tiple elements, without any operator intervention.
Automated Sample Preparation
While automated instrumentation has meant considerable time savings to the ana-
lyst, analytical throughput (i.e., the number of samples which can be analyzed in
a given time) frequently is limited by the time required to prepare the sample. Even
when the sample is available in a suitable solution form, there often are pretreat-
ment steps which must be performed prior to analysis. The introduction of com-
mercial systems based on techniques such as flow injection have directly
addressed the need for automated sample preparation capabilities. Flow injection
techniques can be used to automate relatively simple procedures such as dilution
or reagent addition. They can also be used to automate complex chemical pretreat-
ments, including analyte preconcentration and cold vapor mercury and hydride
generation procedures.
The Stand-alone Computer and Atomic Absorption
Stand-alone and personal computers have extended the automation and data han-
dling capabilities of atomic absorption even further. These computers can inter-
face directly to instrument output ports to receive, manipulate, and store data and
print reports in user selectable formats. Also, data files stored in personal com-
puters can be read into supplemental software supplied with the system or third
party software such as word processor, spreadsheet and database programs for
open-ended customization of data treatment and reporting.

3CONTROL OF ANALYTICAL
INTERFERENCES
THE FLAME PROCESS
Atomic absorption is known as a very specific technique with few interferences.
The ultimate analytical method which is absolutely free of any interferences from
the nature of the sample will probably never exist. The next best thing to not having
interferences is to know what the interferences are and how to eliminate them or
compensate for them. The interferences in atomic absorption are well-defined, as
are the means for dealing with them. In order to understand these interferences
thoroughly, we will examine what goes on in the flame atomization process of
atomic absorption.
In order to get the atomic absorption process to occur, we must produce individual
atoms from our sample which starts out as a solution of ions. This process is dia-
grammed in Figure 3-1. First, by the process of nebulization, we aspirate the sam-
ple into the burner chamber,
where it mixes as a fine aero-
sol with the fuel and oxidant
gases. At this point, the metals
are still in solution in the fine
aerosol droplets. As these tiny
droplets pass into the heat of
the flame, the process of
evaporation or desolvation re-
moves the solvent and leaves
tiny solid particles of sample
material. As more heat is ap-
plied, liquefaction will take
place, and additional heat will
vaporize the sample. At this
point the metal of interest,
called the analyte, is still
M+
+ A-
(Solution)
1) Nebulization ↓
M+
+ A-
(Aerosol)
2) Desolvation ↓
MA (Solid)
3) Liquefaction ↓
MA (Liquid)
4) Vaporization ↓
MA (Gas)
5) Atomization ↓
Mo
+ Ao
(Gas)
6) Excitation ↓
M* (Gas)
7) Ionization ↓
M+
+ e-
(Gas)
Figure 3-1. The flame process. "M+
" is a metal
cation and "A-
" is the associated anion.

bound up with some anion to form a molecule which does not exhibit the atomic
absorption phenomenon we wish to measure. By applying still more heat energy,
this molecule is dissociated into the individual atoms which make it up.
Since the thermal energy from the flame is responsible for producing the absorbing
species, flame temperature is an important parameter governing the flame process.
Temperatures for some flames that have been used in atomic absorption are listed
in Table 3-1. Cooler flames are subject to more interference problems resulting
from insufficient energy for complete atomization. The two premix flames now
used almost exclusively for atomic absorption are air-acetylene and nitrous ox-
ide-acetylene. While the air-acetylene flame is satisfactory for the majority of ele-
ments determined by atomic absorption, the hotter nitrous oxide-acetylene flame
is required for many refractory-forming elements. The nitrous oxide-acetylene
flame is also effective in the control of some types of interference.
Table 3-1
Temperatures of Premix Flames
Oxidant-Fuel Temp., °C
Air-Methane 1850-1900
Air-Natural Gas 1700-1900
Air-Hydrogen 2000-2050
Air-Acetylene 2125-2400
N2O-Acetylene 2600-2800
The number of ground state metal atoms formed in step 5 of the flame process
(Figure 3-1) will determine the amount of light absorbed. Concentration is deter-
mined by comparing the absorbance of the sample to that of a known standard con-
centration. The relationship between the number of atoms in the flame and the
concentration of analyte in solution is governed by the flame process. If any con-
stituent of the sample alters one or more steps of this process from the performance
observed for a standard, an interference will exist, and an erroneous concentration
measurement will result if the interference is not recognized and corrected or com-
pensated.
NONSPECTRAL INTERFERENCES
Interferences in atomic absorption can be divided into two general categories,
spectral and nonspectral. Nonspectral interferences are those which affect the for-
mation of analyte atoms.

Matrix Interference
The first place in the flame atomization process subject to interference is the very
first step, the nebulization. If the sample is more viscous or has considerably dif-
ferent surface tension characteristics than the standard, the sample uptake rate or
nebulization efficiency may be different between sample and standard. If samples
and standards are not introduced into the process at the same rate, it is obvious
that the number of atoms in the light beam and, therefore, the absorbance, will not
correlate between the two. Thus, a matrix interference will exist.
An example of this type of interference is the effect of acid concentration on ab-
sorbance. From Figure 3-2, it can be seen that as phosphoric acid concentration
increases (and the sample viscosity increases), the sample introduction rate and
the sample absorbance decrease. Increased acid or dissolved solids concentration
normally will lead to a negative error if not recognized and corrected. Matrix in-
terferences can also cause positive error. The presence of an organic solvent in a
sample will produce an enhanced nebulization efficiency, resulting in an increased
absorption. One way
of compensating for
this type of interfer-
ence is to match as
closely as possible the
major matrix compo-
nents of the standard
to those of the sample.
Any acid or other re-
agent added to the
sample during prepa-
ration should also be
added to the standards
and blank in similar
concentrations.
Method of Standard Additions
There is a useful technique which may make it possible to work in the presence
of a matrix interference without eliminating the interference itself, and still make
an accurate determination of analyte concentration. The technique is called the
method of standard additions. Accurate determinations are made without elimi-
nating interferences by making the concentration calibration in the presence of the
Figure 3-2. Matrix interference from viscosity effects.
Control of Analytical Interferences 3-3

matrix interference. Aliquots of a standard are added to portions of the sample,
thereby allowing any interferent present in the sample to also affect the standard
similarly.
The standard additions technique is illustrated in Figure 3-3. The solid line passing
through the origin represents a typical calibration line for a set of aqueous stand-
ards. Zero absorbance is defined with a water blank, and, as the concentration of
analyte increases, a linear increase in absorbance is observed.
Let us now take equal aliquots of the sample. Nothing is added to the first aliquot;
a measured amount of standard is added to the second; and a larger measured
amount is added to the third. The first volume of added standard is usually selected
to approximate the analyte concentration in the sample, and the second volume is
normally twice the first
volume. However, for
the method of standard
additions to be used ac-
curately, the absor-
bances for all of the
solutions must fall
within the linear portion
of the working curve.
Finally, all portions are
diluted to the same vol-
ume so that the final
concentrations of the
original sample con-
stituents are the same in
each case. Only the
amount of added analyte
differs, and then by a
known amount.
If no interference were present in this sample, a plot of measured absorbance ver-
sus the concentration of added standard would be parallel to the aqueous standard
calibration, and offset by an absorbance value resulting from the analyte present
in the unspiked sample. If some material is present in the sample which causes a
matrix interference, the number of ground state atoms producing atomic absorp-
tion will be affected, as will be the absorbance from the analyte in the unspiked
sample. However, the absorbance increase from added standard will also be
Figure 3-3. The method of standard additions.

changed by the same proportional amount since the concentration of interferent
is the same in each solution. Therefore, a straight line will still result, but because
of the interference, its slope will be different from that observed for the aqueous
standards.
In this situation, if the absorbance of the unspiked sample were to be compared
directly to the aqueous calibration, an error would result. If, however, the slope
determined by the standard additions to our sample is used as the calibration slope,
an accurate determination of the sample concentration can still be made. By con-
tinuing the concentration calibration on the abscissa backward from zero and ex-
trapolating the calibration line backward until it intercepts the concentration axis,
the concentration responsible for the absorbance of the unspiked sample is indi-
cated. An accurate determination has been made by calibrating in the presence of
the interference.
Properly used, the method of standard additions is a valuable tool in atomic ab-
sorption. The presence of an interference can be confirmed by observing the slope
of the spiked sample calibration and determining whether or not it is parallel to
the aqueous standard line. If it is not, an interference is present. If an interference
is present, the method of standard additions may allow an accurate determination
of the unknown concentration by using the standard additions slope for the cali-
bration. Caution should be used with the technique, however, as it can fail to give
correct answers with other types of interference. The method of standard additions
will not compensate for background absorption or other types of spectral inter-
ference, and normally will not compensate for chemical or ionization types of in-
terference.
Chemical Interference
A second place where interference can enter into the flame process is in step num-
ber 5 of Figure 3-1, the atomization process. In this step, sufficient energy must
be available to dissociate the molecular form of the analyte to create free atoms.
If the sample contains a component which forms a thermally stable compound
with the analyte that is not completely decomposed by the energy available in the
flame, a chemical interference will exist.
The effect of phosphate on calcium, illustrated in Figure 3-4, is an example of a
chemical interference. Calcium phosphate does not totally dissociate in an air-
acetylene flame. Therefore, as phosphate concentration is increased, the absor-
bance due to calcium atoms decreases.

There are two means of deal-
ing with this problem. One is
to eliminate the interference
by adding an excess of another
element or compound which
will also form a thermally sta-
ble compound with the inter-
ferent. In the case of calcium,
lanthanum is added to tie up
the phosphate and allow the
calcium to be atomized, mak-
ing the calcium absorbance in-
dependent of the amount of
phosphate.
There is a second approach to
solving the chemical interfer-
ence problem. Since the prob-
lem arises because of insufficient energy to decompose a thermally stable analyte
compound, the problem can be eliminated by increasing the amount of energy; that
is, by using a hotter flame. The nitrous oxide-acetylene flame is considerably hot-
ter than air-acetylene and can often be used to minimize chemical interferences
for elements generally determined with air-acetylene. The phosphate interference
on calcium, for instance, is not observed with a nitrous oxide-acetylene flame,
eliminating the need for the addition of lanthanum.
Ionization Interference
There is a third major interference, however, which is often encountered in hot
flames. As illustrated in Figure 3-1, the dissociation process does not necessarily
stop at the ground state atom. If additional energy is applied, the ground state atom
can be thermally raised to the excited state or an electron may be totally removed
from the atom, creating an ion. As these electronic rearrangements deplete the
number of ground state atoms available for light absorption, atomic absorption at
the resonance wavelength is reduced. When an excess of energy reduces the popu-
lation of ground state atoms, an ionization interference exists.
Ionization interferences are most common with the hotter nitrous oxide-acetylene
flame. In an air-acetylene flame, ionization interferences are normally encoun-
Figure 3-4. Interference of phosphate on
calcium.

tered only with the more easily ionized elements, notably the alkali metals and al-
kaline earths.
Ionization interference can be eliminated by adding an excess of an element which
is very easily ionized, creating a large number of free electrons in the flame and
suppressing the ionization of the analyte. Potassium, rubidium, and cesium salts
are commonly used as ionization suppressants. Figure 3-5 shows ionization sup-
pression for the determination
of barium in a nitrous oxide-
acetylene flame. The increase
in absorption at the barium
resonance line, and the corre-
sponding decrease in absorp-
tion at the barium ion line as a
function of added potassium,
illustrate the enhancement of
the ground state species as the
ion form is suppressed. By
adding 1000 mg/L to 5000
mg/L potassium to all blanks,
standards and samples, the ef-
fects of ionization can usually
be eliminated.
SPECTRAL INTERFERENCES
Spectral interferences are those in which the measured light absorption is errone-
ously high due to absorption by a species other than the analyte element. The most
common type of spectral interference in atomic absorption is ‘‘background ab-
sorption.’’
Background Absorption
Background absorption arises from the fact that not all of the matrix materials in
a sample are necessarily 100% atomized. Since atoms have extremely narrow ab-
sorption lines, there are few problems involving interferences where one element
absorbs at the wavelength of another. Even when an absorbing wavelength of an-
other element falls within the spectral bandwidth used, no absorption can occur
unless the light source produces light at that wavelength, i.e., that element is also
present in the light source. However, undissociated molecular forms of matrix ma-
Figure 3-5. Effect of added potassium on
ionization.

terials may have broadband absorption spectra, and tiny solid particles in the flame
may scatter light over a wide wavelength region. When this type of nonspecific
absorption overlaps the atomic absorption wavelength of the analyte, background
absorption occurs. To compensate for this problem, the background absorption
must be measured and subtracted from the total measured absorption to determine
the true atomic absorption component.
While now virtually obsolete, an early method of manual background correction
illustrates clearly the nature of the problem. With the ‘‘two line method’’, back-
ground absorption, which usually varies gradually with wavelength, was inde-
pendently measured by using a nonabsorbing emission line very close to the
atomic line for the analyte element, but far enough away so that atomic absorption
was not observed, as illus-
trated in Figure 3-6. By sub-
tracting the absorbance meas-
ured at the nonabsorbing line
from the absorbance at the
atomic line, the net atomic ab-
sorption was calculated.
Nearby, nonabsorbing lines
are not always readily avail-
able, however, and inaccura-
cies in background correction
will result if the wavelength
for background measurement
is not extremely close to the
resonance line. Therefore, for
accuracy, as well as conven-
ience, a different method was
needed.
Continuum Source Background Correction
Continuum source background correction is a technique for automatically meas-
uring and compensating for any background component which might be present
in an atomic absorption measurement. This method incorporates a continuum light
source in a modified optical system, illustrated in Figure 3-7.
The broad band continuum (‘‘white’’ light) source differs from the primary
(atomic line) source in that it emits light over a broad spectrum of wavelengths
Figure 3-6. Two-line background correction.

Figure 3-7. Continuum Source Background Corrector.
Figure 3-8. Atomic and background absorption with a primary (line) source and
a continuum (broadband) source.

instead of at specific lines. From Figure 3-8, it can be seen that atomic absorption,
which occurs only at very discrete wavelengths, will not measurably attenuate the
emission from the continuum source. However, background absorption which has
very broad absorption spectra will absorb the continuum emission as well as the
line emission.
As shown in Figure 3-7, light
from both the primary and
continuum lamps is combined
and follows a coincident path
through the sample, through
the monochromator, and to the
detector. The two lamps are
observed by the detector alter-
nately in time, and as illus-
trated in Figure 3-9,
instrument electronics sepa-
rate the signals and compare
the absorbance from both
sources. An absorbance will
be displayed only where the
absorbance of the two lamps
differs. Since background ab-
sorption absorbs both sources
equally, it is ignored. True
atomic absorption, which ab-
sorbs the primary source
emission and negligibly ab-
sorbs the broad band contin-
uum source emission, is still
measured and displayed as
usual.
Figure 3-10 shows how back-
ground absorption can be
automatically eliminated
from the measured signal us-
ing continuum source back-
ground correction. In the ex-
Figure 3-9. Simplified timing diagram.
Figure 3-10. Automatic background correction.

ample, a lead determination is shown without background correction (A) and with
background correction (B). Both determinations were performed at the Pb 283.3
nm wavelength with 15x scale expansion and a 10-second integration time.
Continuum source background correction is widely applied, and except in some
very unusual circumstances, is fully adequate for all flame AA applications. There
are some limitations to continuum source background correction, however, which
especially impact graphite furnace atomic absorption, to be discussed in later
chapters. These limitations are summarized in Table 3-2.
Table 3-2
Limitations of Continuum Source Background Correction
1. Requires additional continuum light source(s) and electronics.
2. Requires the intensities of the primary and continuum sources
to be similar.
3. Two continuum sources are required to cover the full wavelength
range.
4. Requires critical alignment of the continuum and primary sources
for accurate correction.
5. May be inaccurate for structured background absorption.
The fact that continuum source background correction requires two sources (pri-
mary and continuum) imposes convenience, economic, and performance consid-
erations on the use of the technique. The convenience and economic factors come
from the fact that the continuum source has a finite lifetime and must be replaced
on a periodic basis. The performance factor originates from the fact that the back-
ground component of the absorption signal is measured from one source, while
the total uncorrected signal is measured with another. This leads to the possibility
of inaccurate compensation if the two sources do not view exactly the same portion
of the atom cloud, especially at higher background absorption levels. Finally, since
the two sources are spectrally different, background absorption exhibiting fine
spectral structure may attenuate the two source lamps to different degrees, leading
to inaccuracies in background correction for such cases.
Introduction to Zeeman Background Correction
For those applications where the limitations of the continuum source approach are
significant to the analysis, the Zeeman background correction system may be pref-
erable. Zeeman background correction uses the principle that the electronic energy
levels of an atom placed in a strong magnetic field are changed, thereby changing

the atomic spectra which are a
measure of these energy lev-
els. When an atom is placed in
a magnetic field and its atomic
absorption profile observed
with polarized light, the nor-
mal single-line atomic ab-
sorption profile is split into
two or more components sym-
metrically displaced about the
normal position, as illustrated
in Figure 3-11. The spectral
nature of background absorp-
tion, on the other hand, is usu-
ally unaffected by a magnetic
field.
By placing the poles of an
electromagnet around the at-
omizer and making alternat-
ing absorption measurements
with the magnet off and then
on, the uncorrected total ab-
sorbance (magnet off) and
‘‘background only’’ absor-
bance (magnet on) can be
made, as in Figure 3-12.
The automatic comparison
made by the instrument to
compensate for background
correction is similar to that for
the continuum source technique, except that only the one atomic line source is
used. As a result, there are no potential problems with matching source intensities
or coincident alignment of optical paths. Also, background correction is made at
the analyte wavelength rather than across the entire spectral bandwidth, as occurs
with continuum source background correction.
With Zeeman background correction, the emission profile of the line source is
identical for both AA and background measurements. As a result, most complex
Figure 3-11. The Zeeman effect.
Figure 3-12. Zeeman effect background
correction.

structured background situations can be accurately corrected with Zeeman back-
ground correction. This can be seen in Figure 3-13, where background absorption
due to the presence of aluminum in a graphite furnace determination of arsenic is
completely compensated using Zeeman correction but produces erroneously high
results with continuum source background correction. Table 3-3 summarizes the
advantages of Zeeman effect background correction.
Table 3-3
Advantages of Zeeman Effect Background Correction
1. Corrects for high levels of background absorption.
2. Provides accurate correction in the presence of structured
background.
3. Provides true double-beam operation.
4. Requires only a single, standard light source.
5. Does not require intensity matching or coincident alignment
of multiple sources.
The examples used above to illustrate Zeeman effect background correction are
based on the use of a transverse AC Zeeman system, the type most commonly used
Figure 3-13. Zeeman vs. Continuum Background Correction. From
Letourneau, Joshi and Butler, At. Spectrosc. 8, 146 (1987).

with commercial AA instrumentation. However, there are three types of Zeeman
effect background correction systems available on commercial atomic absorption
instruments: DC Zeeman, transverse AC Zeeman and longitudinal AC Zeeman.
These systems differ in the way the magnetic field is applied and by the means
used to measure the combined (atomic absorption plus background absorption)
and background absorption only signals. DC Zeeman systems use a permanent
magnet and a rotating or vibrating polarizer to separate the combined and back-
ground only signals. AC systems use an electromagnet, and measure the combined
and background only signals by turning the magnetic field on and off. The dif-
ference between transverse (magnetic field applied across the light path) and lon-
gitudinal (magnetic field applied along the light path) AC Zeeman systems is that
transverse systems uses a fixed polarizer, while the longitudinal system does not
require a polarizer. The advantages and limitations of each type of Zeeman system
are summarized in Tables 3-4 and 3-5 on the following page.
Other Spectral Interferences
If the atomic absorption profile for an element overlaps the emission line of an-
other, a spectral interference is said to exist. As has already been mentioned, this
is an infrequent occurrence, because of the very wavelength-specific nature of
atomic absorption. Even if an absorption line for an element other than the analyte
but also present in the sample falls within the spectral bandpass, an interference
will occur only if an emission line of precisely the same wavelength is present in
the source. As the typical emission line width may be only 0.002 nanometers, ac-
tual overlap is extremely rare. The chances for spectral interference increase when
multi-element lamps are used, where the source may contain close emission lines
for several elements. Particular care should be exercised when secondary analyti-
cal wavelengths are being used in a multi-element lamp. Procedures for circum-
venting spectral interference include narrowing the monochromator slit width or
using an alternate wavelength.
INTERFERENCE SUMMARY
The major interferences in atomic absorption include: (1) matrix interference, (2)
chemical interference, (3) ionization interference, and (4) background absorption.
For the first type, special considerations in sample preparation or the use of the
method of standard additions may compensate for the problems generated. For the
second and third, addition of an appropriate releasing agent or ionization buffer
or changing the flame type used will normally remove the interference. For the
fourth, background absorption, an instrumental correction technique will automat-

ically compensate for the biasing effects. Application of the techniques described
here will make possible accurate atomic absorption determinations in very com-
plex samples.
Table 3-4
DC Zeeman Systems
Advantages:
Less expensive to operate (lower power consumption)
Disadvantages:
Has poorer sensitivity and analytical working range relative
to AC Zeeman systems.
The polarizer reduces light throughput by as much as 50%,
affecting analytical performance.
A mechanical assembly is required to rotate or vibrate the
polarizer.
Table 3-5
AC Zeeman Systems
Advantages:
Offers better sensitivity and expanded analytical working
ranges relative to DC Zeeman systems.
No polarizer is required, so it provides higher light throughput
and improved analytical performance. (Longitudinal AC
Zeeman systems only)
Requires no additional mechanical devices.
Disadvantages:
Requires more electrical power than DC Zeeman systems, so has
higher operating expenses.
The polarizer causes reduced light throughput by as much as 50%,
affecting analytical performance. (Transverse AC Zeeman
systems only.)

4HIGH SENSITIVITY
SAMPLING SYSTEMS
LIMITATIONS TO FLAME AA SENSITIVITY
Flame atomic absorption is a rapid and precise method of analysis. Determinations
of analyte concentrations in the mg/L concentration region are routine for most
elements. However, the need for trace metal analyses at µg/L and even sub µg/L
levels calls for a more sensitive technique.
For atomic absorption to occur, free ground state atoms must be placed in a beam
of light of a wavelength corresponding to an appropriate electron transition of the
analyte. Any sampling process conceived must therefore address the process of
creating ground state atoms and directing them to the spectrometer light path.
In examining the flame AA process, we can find a number of areas limiting the
sensitivity of the technique. The absorbance depends on the number of atoms in
the optical path of the spectrometer at a given instant. The nebulization process,
which draws sample solution into the burner chamber at approximately 3-8 mil-
liliters per minute, limits the sample introduction rate, and, therefore, the amount
of sample available for transport to the flame. Further, the premix burner design,
which has been universally adopted due to its many desirable characteristics, has
the undesirable characteristic of being very wasteful of sample. Only a small frac-
tion of the sample nebulized ever reaches the flame, with the remainder being di-
rected to the drain. Finally, that sample which is introduced into the flame is
resident in the light path for only a fleeting moment as it is propelled upwards
through the flame.
The sensitivity of atomic absorption can be improved by addressing the limitations
of flame sampling. By improving the sampling efficiency and/or constraining ana-
lyte atoms to the light path for a longer period of time, a greater absorption for
the same analyte concentration can be achieved.

THE COLD VAPOR MERCURY TECHNIQUE
Principle
Since atoms for most AA elements cannot exist in the free, ground state at room
temperature, heat must be applied to the sample to break the bonds combining at-
oms into molecules. The only notable exception to this is mercury. Free mercury
atoms can exist at room temperature and, therefore, mercury can be measured by
atomic absorption without a heated sample cell.
In the cold vapor mercury technique, mercury is chemically reduced to the free
atomic state by reacting the sample with a strong reducing agent like stannous
chloride or sodium borohydride in a closed reaction system. The volatile free mer-
cury is then driven from the reaction flask by bubbling air or argon through the
solution. Mercury atoms are carried in the gas stream through tubing connected
to an absorption cell, which is placed in the light path of the AA spectrometer.
Sometimes the cell is heated slightly to avoid water condensation but otherwise
the cell is completely unheated.
As the mercury atoms pass into the sampling cell, measured absorbance rises in-
dicating the increasing concentration of mercury atoms in the light path. Some sys-
tems allow the mercury vapor to pass from the absorption tube to waste, in which
case the absorbance peaks and then falls as the mercury is depleted. The highest
absorbance observed during the measurement will be taken as the analytical sig-
nal. In other systems, the mercury vapor is rerouted back through the solution and
the sample cell in a closed loop. The absorbance will rise until an equilibrium con-
centration of mercury is attained in the system. The absorbance will then level off,
and the equilibrium absorbance is used for quantitation.
The entire cold vapor mercury process can be automated using flow injection tech-
niques. Samples can be analyzed in duplicate at the rate of about 1 sample per min-
ute with no operator intervention. Detection limits are comparable to those
obtained using manual batch processes. The use of flow injection systems also
minimizes the quantity of reagents required for the determination, further reduc-
ing analysis costs.

Advantages of the Cold Vapor Technique
The sensitivity of the cold vapor technique is far greater than can be achieved by
conventional flame AA. This improved sensitivity is achieved, first of all, through
a 100% sampling efficiency. All of the mercury in the sample solution placed in
the reaction flask is chemically atomized and transported to the sample cell for
measurement.
The sensitivity can be further increased by using very large sample volumes.
Since all of the mercury contained in the sample is released for measurement, in-
creasing the sample volume means that more mercury atoms are available to be
transported to the sample cell and measured. The detection limit for mercury by
this cold vapor technique is approximately 0.02 µg/L. Although flow injection
techniques use much smaller sample sizes, they provide similar performance ca-
pabilities, as the entire mercury signal generated is condensed into a much smaller
time period relative to manual batch-type procedures.
Where the need exists to measure even lower mercury concentrations, some sys-
tems offer an amalgamation option. Mercury vapor liberated from one or more
sample aliquots in the reduction step is trapped on a gold or gold alloy gauze. The
gauze is then heated to drive off the trapped mercury, and the vapor is directed
into the sample cell. The only theoretical limit to this technique would be that im-
posed by background or contamination levels of mercury in the reagents or system
hardware.
Limitations to the Cold Vapor Technique
Of all of the options available, the cold vapor system is still the most sensitive and
reliable technique for determining very low concentrations of mercury by atomic
absorption. The concept is limited to mercury, however, since no other element
offers the possibility of chemical reduction to a volatile free atomic state at room
temperature.
HYDRIDE GENERATION TECHNIQUE
Principle
Hydride generation sampling systems for atomic absorption bear some resem-
blances to cold vapor mercury systems. Samples are reacted in an external system
with a reducing agent, usually sodium borohydride. Gaseous reaction products are
then carried to a sampling cell in the light path of the AA spectrometer. Unlike
the mercury technique, the gaseous reaction products are not free analyte atoms
High Sensitivity Sampling Systems 4-3

but the volatile hydrides. These molecular species are not capable of causing
atomic absorption. To dissociate the hydride gas into free atoms, the sample cell
must be heated.
In some hydride systems, the absorption cell is mounted over the burner head of
the AA spectrometer, and the cell is heated by an air-acetylene flame. In other sys-
tems, the cell is heated electrically. In either case, the hydride gas is dissociated
in the heated cell into free atoms, and the atomic absorption rises and falls as the
atoms are created and then escape from the absorption cell. The maximum absorp-
tion reading, or peak height, or the integrated peak area is taken as the analytical
signal.
Advantages of the Hydride Technique
The elements determinable by hydride generation are listed in Table 4-1. For these
elements, detection limits well below the µg/L range are achievable. Like cold va-
por mercury, the extremely low detection limits result from a much higher sam-
pling efficiency. In addition, separation of the analyte element from the sample
matrix by hydride generation is commonly used to eliminate matrix-related inter-
ferences.
Table 4-1
Hydride Generation Elements
As Bi Ge
Pb Sb Se
Sn Te
The equipment for hydride generation can vary from simple to sophisticated. Less
expensive systems use manual operation and a flame-heated cell. The most ad-
vanced systems combine automation of the sample chemistries and hydride sepa-
ration using flow injection techniques with decomposition of the hydride in an
electrically-heated, temperature-controlled quartz cell.
Disadvantages to the Hydride Technique
The major limitation to the hydride generation technique is that it is restricted pri-
marily to the elements listed in Table 4-1. Results depend heavily on a variety of
parameters, including the valence state of the analyte, reaction time, gas pressures,
acid concentration, and cell temperature. Therefore, the success of the hydride
generation technique will vary with the care taken by the operator in attending to

the required detail. The formation of the analyte hydrides is also suppressed by a
number of common matrix components, leaving the technique subject to chemical
interference.
GRAPHITE FURNACE ATOMIC ABSORPTION
Principle
By far the most advanced and widely used high sensitivity sampling technique for
atomic absorption is the graphite furnace. In this technique, a tube of graphite is
located in the sample compartment of the AA spectrometer, with the light path
passing through it. A small volume of sample solution is quantitatively placed into
the tube, normally through a sample injection hole located in the center of the tube
wall. The tube is heated through a programmed temperature sequence until finally
the analyte present in the sample is dissociated into atoms and atomic absorption
occurs.
As atoms are created and diffuse out of the tube, the absorbance rises and falls in
a peak-shaped signal. The peak height or integrated peak area is used as the ana-
lytical signal for quantitation.
Advantages of the Graphite Furnace Technique
Detection limits for the graphite furnace fall in the ng/L range for most elements.
The sample is atomized in a very short period of time, concentrating the available
atoms in the heated cell and resulting in the observed increased sensitivity. Even
though this technique uses only microliter sample volumes, the small sample size
is compensated by long atom residence times in the light path. This provides de-
tection limits similar to the techniques discussed above which use much larger
samples.
The graphite furnace is much more automated than the other techniques. Even
though heating programs can be very sophisticated, the entire process is automated
once the sample has been introduced and the furnace program initiated. Automat-
ic samplers make completely unattended operation for graphite furnace AA pos-
sible.
Early experiences with the graphite furnace were plagued with interference prob-
lems, requiring detailed optimization procedures for every sample to obtain accu-
rate results. However, extensive studies into the theory of the furnace technique
combined with the development of improved instrumentation have changed fur-
nace AA into a highly reliable, routine technique for trace metal analysis.
High Sensitivity Sampling Systems 4-5

The final and most obvious advantage of the graphite furnace is its wide applica-
bility. The graphite furnace can determine most elements measurable by AA in a
wide variety of matrices. The importance of this technique requires a more de-
tailed discussion in the following chapters.

AAS-GTA-reading-material.pdf

More Related Content

AAS-GTA-reading-material.pdf