Chapter 9- Atomic Absorption and Atomic Fluorescence Spectrometry

A: Sample Atomization Techniques

Atomic Spectroscopy
Atomic spectroscopy is used for the qualitative and quantitative determination of perhaps
70 elements. Sensitivities of atomic methods lie typically in the parts-per-million to parts-
per-billion range. Additional virtues of these methods are speed convenience unusually
high selectivity and moderate costs.
Spectroscopic determination of atomic species can only be performed on a gaseous
medium in !hich the individual atoms are !ell separated from one another.
"onsequently the first step in all atomic spectroscopic procedures is atomization a
process in !hich the sample is volatilized and decomposed in such a !ay as to produce
an atomic gas. The efficiency and reproducibility of the atomization step in a large
measure determine the method#s sensitivity precision and accuracy so atomization is by
far the most critical step in atomic spectroscopy.
$n AAS and A%S there are t!o factors involved. These are:
The intensity of light source
The probability of transition.
$t should be noted that in AAS and A%S the ground state is the starting point but in
A&S some higher levels partially populated to start !ith. 'oth AAS and A%S obey the
selection rules but A&S does not because it involves the use of thermal e(citation. AAS
!as founded by Allen )alsch *Australia+ in the ,-../s !hile 'unsen founded A&S in the
,-00/s. $n terms of simplicity AAS is the simplest spectroscopic instrument to use
follo!ed by A%S.
%lame Atomization
$n a flame atomizer a solution of the sample is nebulized by the flo! of gaseous o(idant
mi(ed !ith a gaseous fuel and carried into a flame !here atomization occurs.
The first step is de-solvation in !hich the solvent is evaporated to produce a finely
divided solid molecular aerosol. As a result de-solvation of these molecules then leads to
an atomic gas. Some of the atoms ionize and give cations and electrons. 1ndoubtedly
other molecules and atoms are also produced in the flame as a result of interactions of the
fuel !ith the o(idant and !ith various species in the sample. As sho!n in the diagram
above !e see that the heat of the flame also e(cites a fraction of the molecules atoms
and ions thus giving atomic ionic and molecular emission spectra.
Types of Flames
The table belo! lists the common fuels and o(idants employed in flame spectroscopy and
the appro(imate range of temperature realized !ith each of these mi(tures. These
temperatures in the range of ,700-0200
o
" !ere obtained !ith the various fuels !hen air
serves as the o(idant.
Max. flame speed (cm/s Max. temp. (
o
C
Air-"oal gas .. ,320
Air-propane 30 ,-0.
Air-hydrogen 400 00.0
Air-.05 o(ygen-acetylene ,60 0400
7(ygen-nitrogen-acetylene 620 03,.
7(ygen-acetylene ,,40 4060
7(ygen-cyanogen ,20 2620
8itrous o(ide-acetylene ,30 0-..
8itric o(ide-acetylene -0 40-.
8itrogen dio(yde-hydrogen ,.0 0660
8itrous o(ide-hydrogen 4-0 06.0
The burning velocities listed in the middle column are of considerable importance because flames are
stable in certain ranges of gas flo! rates only. $f the gas flo! rate does not e(ceed the burning velocity
the flame propagates itself bac9 into the burner giving flashbac9. As the flo! rate increases the flame
rises until it reaches a point above the burner !here the flo! velocity and the burning velocity are equal.
This region is !here the flame is stable.
Flame Structure
$n the figure !e see all the important
regions of a flame. $t includes the primary
combustion zone the interzonal region and
the secondary combustion zone. The
appearance and relative size of these regions
vary considerably !ith the fuel-to-o(idant
ratio as !ell as !ith the type of fuel and
o(idant. The primary combustion zone in a
hydrocarbon flame is recognizable by its
blue luminescence arising from the band
spectra of "0 and ": and other radicals.
Thermal equilibrium is ordinarily not
reached in this region and it is seldom used
for flame spectroscopy. %urthermore the
interzonal area -!hich is narro! in
stoichiometric hydrocarbon flames may
reach several centimeters in height in fuel
rich acetylene;o(ygen acetylene;nitrous
o(ide sources. The zone is rich in free atoms and it is the most !idely used part of the flame.
Temperature !rofiles
The ma(imum temperature is located in the flame about ,cm above the primary combustion zone. $t is
important- particularly for emission methods- to focus the same part of the flame on the entrance slit for
all calibrations and analytical measurements.
The "ffects of Flame Temperature
'oth emission and absorption spectra are affected in a comple( !ay by variations in
flame temperature. :igher temperatures tend to increase the total atom population of the
flame and sensitivity. )ith certain elements such as the al9ali metals ho!ever this ii-
increase in atom population is more than offset by the loss of atoms by ionization
Also flame temperature determines the relative number of e(cited and une(cited atoms
in a flame. $n an air;acetylene flame for e(ample the ratio of e(cited to une(cited
magnesium atoms can be computed to be about ,0&-3 !hereas in o(ygen acetylene
flame !hich is about 700" hotter this ratio is about ,0&-6. "ontrol of temperature is
thus of prime importance in flame emission methods. %or e(ample !ith a 0.00 #" flame
a temperature increase of $7 #" causes the number of sodium atoms in the e(cited 4p state
to increase by about 45. $n contrast the corresponding decreases ii-a the much larger
number of ground state atoms is only about 0.0005. So emission methods based as they
are on the population of e(cited atoms. <equires much closer control of flame
temperature than do absorption procedures in !hich the analytical signal depends upon
the number of une(cited atoms.
The number of une(cited atoms in a typical flame e(ceeds the number of e(cited ones by
a factor of ,0&4 to ,0&,0 or more. And this fact suggests that absorption methods
should be significantly more sensitive than emission methods.
Flame Absorbance !rofiles
A different portion of the flame should be used for the analysis of different elements. The more
sophisticated instruments for flame spectroscopy are equipped !ith monochromators that sample the
radiation from a relatively small region of the flame= ad>ustment of the position of the flame !ith respect
to the entrance slit is thus critical.
Flame Atomi#ers
A flame atomizer of a pneumatic nebulizer !hich converts the sample solution into a
mist or aerosol that is then fed into a burner. A common type of nebulizer is the
concentric tube type. $n !hich the liquid sample is suc9ed through a capillary tube by -a
high-pressure stream of a gas flo!ing around the tip of the tube. This process of liquid
transport is called aspiration. The high velocity gas brea9s the liquid into the fine
droplets of various sizes !hich are then carried into the flame. "ross-flo! nebulizers are
also employed in !hich the high-pressure gas flo!s across a capillary tip at right angles.
7ften in this type of nebulizer the liquid is pumped through the capillary. $n most
atomizers the high-pressure gas is the o(idant !ith the aerosol containing o(idant being
mi(ed subsequently !ith the fuel.
7n this figure !e see a typical commercial laminar flo! burner that employs a concentric
tube nebulizer.
The aerosol is mi(ed !ith fuel and flo!s past a series of baffles that remove all but the
finest droplets. As a result of the baffles the a-ma>ority of the sample collects in the
bottom of the mi(ing chamber !here it is drained to a !aste container. The aerosol
o(idant alit fuel are then directed into a slotted burner !hich provides a flame that is
usually . or ,0 cm in length.
?aminar flo! burners provide a relatively quiet flame and a long path length. These
properties tend to enhance sensitivity and reproducibility. The mi(ing chamber in this
type of burner contains a potentially e(plosive mi(ture !hich can be ignited by
flashbac9 if the flo! rates are not sufficient.
Fuel $ %xidant &e'ulators
An important variable that requires close control in flame spectroscopy is the flo! rate of
both o(idant and fuel. %uel and o(idant are ordinarily combined in appro(imately
stoichiometric amounts. :o!ever a flame that contains an e(cess of fuel may be more
desirable. The most !idely used device for measuring flo! rates is the rotameter !hich
consists of tapered graduated transparent tube that is mounted vertically !ith the smaller
end do!n.
!erformance Characteristics of Flame Atomi#ers
%lame atomization appears to be superior to all other methods that have been thus far
developed for liquid sample introduction for atomic absorption and fluorescence
spectrometry. :o!ever there are t!o reasons that lo!er the sampling efficiency of the
flame. %irst a large portion of the sample flo!s do!n the drain. Second the residence
time of individual atoms in the optical path in the flame is brief.
Sources of Atomic Spectra
)ithout chemical bonding there can be no vibrational or rotational energy states and
transitions. Therefore atomic emission absorption and fluorescence spectra are made up
of a limited number of narro! pea9s or lines.
&lectrothermal Atomization
&lectrothermal atomizers !hich first appeared in ,-70 generally provide enhanced
sensitivity because the entire sample is atomized in a short period and the average
residence time of the atoms in the optical path is a second or more. They generally have
not been applied for direct production of emission spectra. They are ho!ever beginning
to be used for vaporizing samples for sample introduction in inductively coupled plasma
emission spectroscopy.
"lectrothermal Atomi#ers
@raphite furnace atomic absorption spectrometry *@%AAS+ is also 9no!n by various
other acronyms including electrothermal atomic absorption spectrometry *&TAAS+.
'riefly the technique is based on the fact that free atoms !ill absorb light at frequencies
or !avelengths characteristic of the element of interest *hence the name atomic
absorption spectrometry+. )ithin certain limits the amount of light absorbed can be
linearly correlated to the concentration of analyte present. %ree atoms of most elements
can be produced from samples by the application of high temperatures. $n @%AAS
samples are deposited in a small graphite tube !hich can then be heated to vaporize and
atomize the analyte.
An ideal graphite furnace should fulfill the follo!ing requirements:
A constant temperature in time and space during the interval in !hich free atoms
are produced
Auantitative atom formation regardless of the sample composition
Separate control of the volatilization and atomization processes
:igh sensitivity and good detection limits
A minimum of spectral interferences
%utput Si'nal
At a !avelength at !hich absorbance or fluorescence occurs the transducer output rises
to a ma(imum after a fe! seconds of ignition follo!ed by a rapid decay bac9 to zero as
the atomization products escape into the surroundings. The change is rapid enough to
require a high-speed data acquisition system. Auantitative analysis is usually based on
pea9 height although pea9 area also has been used.
!erformance Characteristics of "lectrothermal Atomi#ers
They offer the advantage of unusually high sensitivity for small volumes of sample. The
relative precision of nonflame methods is generally in the range of .5 to ,05 compared
!ith the ,5 or better that can be e(pected for flame or plasma atomization. %urthermore
furnace methods are slo!- typically requiring several minutes per element. A final
disadvantage is that the analytical range is lo! being usually less than t!o orders of
magnitude.
Analysis of Solids (ith "lectrothermal Atomi#ers
A !ay of performing such measurements is to !eigh the finely ground sample into a
graphite boat and insert the boat manually in the furnace. Another !ay is to prepare a
BslurryB of the po!dered sample by ultrasonic agitation in an aqueous medium. The
slurry is then pipetted into the furnace for atomization.
Specialized Atomization Techniques
7ther atomization methods for sample introduction can be used for atomic absorption
analyses. Three are described in this section.
)lo( *ischar'e Atomi#ation
A glo! discharge device produces an atomized vapor that can be s!ept into a cell for
absorption measurements. $n order for this technique to be applicable the sample must
be an electrical conductor such as finely ground graphite or copper. Solution samples
have also been analyzed by deposition on a graphite aluminum or copper cathode.
Cetection limits !ith this type of device are reported to be in the lo! parts per million for
solid samples.
+ydride Atomi#ation
Atomization of the hydrides requires only that they be heated in a quartz tube.
Cold-,apor Atomi#ation
This is applicable only to the determination of mercury because it is the only metallic
element that has an appreciable vapor pressure at ambient temperature. This
determination is of vital importance because of the to(icity of various organic mercury
compounds and their !idespread distribution in the environment. Cetection limits in the
parts-per-billion range are realized. Several manufacturers offer automatic instruments
for performing this analysis for a variety of types of samples.
': Atomic Absorption $nstrumentation
Flame Atomic Absorption Spectroscopy
%lame atomic absorption spectroscopy is currently the most !idely used of all the atomic
methods. $t is so because of its simplicity effectiveness and relatively lo! cost. The
general use of this technique began in the early ,-.0s and gre! e(plosively after that.
The reason that atomic absorption methods !ere not !idely used until that time !as
directly related to problems created by the very narro! !idths of atomic absorption lines.
The natural !idth of an atomic absorption or an atomic emission line is on the order of ,0
-. nm. T!o effects ho!ever cause line !idths to be broadened by a factor of ,00 or
more.
<adiation Sources
Analytical methods based on the atomic absorption are potentially highly specific
because atomic absorption lines are remar9ably narro! and because electronic transition
energies are unique for each element. 7n the other hand the limited line !idths create a
problem not ordinarily encountered in molecular absorption spectroscopy. The problem
created by the limited !idth of atomic absorption pea9s has been solved by the use of line
sources !ith band!idths even narro!er than absorption pea9s.
*oppler -roadenin'
Coppler broadening results from the rapid motion of atoms as they emit or absorb
radiation. Atoms moving to!ard the detector emit !avelengths that are slightly shorter
than the !avelengths by atoms moving at right angels to the detector. This difference is a
manifestation of the !ell-9no!n Coppler shift= the effect is reversed for atoms moving
a!ay from the detector. The net effect is an increase in the !idth of the emission line.
%or precisely the same reason the Coppler effect also causes broadening of absorption
lines. This type of broadening becomes more pronounced as the flame temperature
increases because of the increased rate of motion of the atoms.
!ressure -roadenin'
Dressure broadening arises from collisions among atoms that cause slight variations in
their ground state energies and slight energy differences bet!een ground and e(cited
states. Dressure broadening becomes greater !ith increases in temperature. So as result
broader absorption and emission pea9s are al!ays encountered at elevated temperatures.
+ollo( Cathode .amps
The most common source for atomic absorption measurements is the hollo! cathode
lamp. This type of lamp consists of a tungsten anode and a cylindrical cathode sealed in
a glass tube that is filled !ith neon or argon at a pressure of , to . torr. The cathode is
constructed of the metal !hose spectrum is desired or serves to support a layer of that
metal. The application of a potential of about 400 E across the electrodes causes
ionization of the argon and generation of a current of . to ,0 mA as the argon cations and
electrons migrate to the t!o electrodes. $f the potential is sufficiently large the argon
cations stri9e the cathode !ith sufficient energy to dislodge some of the metal atoms and
thereby produce an atomic cloud- and this process is called sputtering. The sputtered
metal atoms in a lamp eventually diffuse bac9 to the cathode surface of to the !alls of the
lamp and are deposited. The cylindrical configuration of the cathode tends to concentrate
the radiation in a limited region of the metal tube= this design also enhances the
probability that redeposition !ill occur at the cathode rather than on the glass !all.
"lectrodeless *ischar'e .amps
&lectrodeless discharge lamps are useful sources of atomic line spectra and provide
radiant intensities that are usually one to t!o orders of magnitude greater than their
hollo! cathode counterparts. A typical lamp is constructed from a sealed quartz tube
containing an inert gas such as argon at a pressure of a fe! torr and a small quantity of
the analyte metal. The lamp contains no electrode but instead is energized by an intense
field of radio frequency or micro!ave radiation. The argon ionizes in this field and the
ions are accelerated by the high frequency component of the field until they gain
sufficient energy to e(cite the atoms of the metal !hose spectrum is sought.
Source Modulation
$n the typical atomic absorption instrument it is necessary to eliminate interference
caused by emission of radiation by the flame. %urthermore most of it is removed by the
monochromator. 8evertheless emitted radiation corresponding in !avelength to the
monochromator setting is inevitably present in the flame due to e(citation and emission
by analyte atoms. $n order to eliminate the effects of flame emission it is necessary to
modulate the output of the source so that its intensity fluctuates at a constant frequency.
The detector then receives t!o types of signal an alternative one from the source and a
continuous one from the flame. Then these signals converted to the corresponding types
of electrical response.
A simple and entirely satisfactory !ay of modulating the emission from the source is to
interpose a circular metal dis9 or c.2opper in the beam bet!een the source and the flame.
Alternate quadrants of this dis9 are removed to permit passage of light. <otation of the
dis9 at constant 9no!n rate provides a beam that is chopped to the desired frequency.
Spectrophotometers
$nstruments for atomic absorption !or9 are offered by numerous manufacturers= both
single and double-beam designs are available. The range of sophistication and cost is
substantial.
$n general the instrument must contain:
8arro! band!idth to isolate the line chosen for measurement
Sufficient glass filter
$nterchangeable interference filters
@ood-quality ultraviolet;visible monochromators
Dhotomultiplier tubes
Ficrocomputer systems that are used to control instrument parameters and data
Sin'le--eam /nstruments
A typical single-beam instrument consists of several hollo! cathode sources a chopper
or a pulsed po!er supply an atomizer and a simple grating spectrophotometer !ith a
photomultiplier transducer.
*ouble--eam /nstruments
The beam of the hollo! cathode is split by a mirrored chopper one half passing through
the flame and the other half around it. The t!o beams are then recombined by a half-
silvered mirror and passed into a "zerney-Turner grating monochromator= a
photomultiplier tube serves as the transducer. The output from the latter is fed to a loc9-
in amplifier that is synchronized !ith the chopper drive. The ratio bet!een the reference
and the sample signal is then amplified and fed to the readout !hich may be a digital
meter or a signal recorder.
": $nterferences in Atomic Absorption Spectroscopy
Spectral $nterferences
Spectral interferences arise !hen the absorption or emission of an interfering species
either overlaps or lies so close to the analyte absorption or emission that resolution by the
monochromator becomes impossible.
The T(o-.ine Correction Method
$t employs a line from the source as a reference. The line should lie as close as possible
to the analyte line but must not be absorbed by the analyte. $f the conditions are met it is
assumed that any decrease in po!er of the reference line from that observed during
calibration arises from absorption or scattering by the matri( products of the sample.
The Continuum-Source Correction Method
A deuterium lamp provides a source of continuous radiation throughout the ultraviolet
region. The absorbance of the deuterium radiation is then subtracted from that of the
analyte beam. The attenuation of the po!er of the continuum during passage through the
atomized sample reflects only the broadband absorption or scattering by the sample
matri( components. A bac9ground correction is thus achieved.
-ac0'round Correction -ased on the 1eeman "ffect
)hen an atomic vapor is e(posed to a strong magnetic field a splitting of electronic
energy levels of the atoms ta9es place that leads to formation of several absorption lines
for each electronic transition. This phenomenon !hich is called the Zeeman effect is
general for all atomic spectra. Application of the Geeman effect to atomic absorption
instruments is based upon the differing response of the t!o types of absorption pea9s to
polarized radiation. Geeman effect instruments provide a more accurate correction for
bac9ground than earlier methods. These instruments are particularly useful for
electrothermal atomizers and permit the direct determination of elements in samples such
as urine and blood. The decomposition of organic material in these samples leads to
large bac9ground corrections and consequent susceptibility to significant error.
-ac0'round Correction -ased on the Source Self-&e2ersal
The Smith-Hieftje background correction method is based upon the self-reversal or self-
absorption behavior of radiation emitted from hollo! cathode lamps !hen they are
operated at high currents.
"hemical $nterferences
Chemical interferences result from various chemical processes occurring during
atomization that alter the absorption characteristics of the analyte. They are more
common than spectral ones. Their effects can frequently be minimized by a suitable
choice of operating conditions. 'oth theoretical and e(perimental evidence suggest that
many of the processes occurring in the mantle of a flame are in appro(imate equilibrium.
As a consequence it becomes possible to regard the burned gases as a solvent medium to
!hich thermodynamic calculations can be applied.
The equilibria of principle interest include:
%ormation of "ompounds of ?o! Eolatility
Cissociation <eactions
$onization
C: Atomic Absorption Analytical Techniques
Sample Dreparation
$t is required for flame spectroscopy that the sample be introduced into the e(citation
source in the form of a solution most commonly an aqueous one. The decomposition
and solution steps are often more time consuming and introduce more error than the
spectroscopic measurement itself. Some common methods used for decomposition and
dissolving samples for atomic absorption methods include treatment !ith hot mineral
acids= o(idation !ith liquid reagents such as sulfuric nitric or perchloric acids=
combustion in an o(ygen bomb or other closed container to avoid loss of analyte= ashing
at a high temperature= and high-temperature fusion !ith reagents such as boric o(ide
sodium carbonate sodium pero(ide or potassium pyrosulfate.
Table 3. "!A sample processin' method for metallic element analysis
Analysis Tar'et Method "n2ironmental Matrice
4umber
total recoverable
metals
400. ground !ater;surface !ater
dissolved metals 400. ground !ater;surface !ater
suspended metals 400. ground !ater;surface !ater
total metals 40,0
aqueous samples !astes that contain suspended solids
and mobility-procedure e(tracts
total metals 40,.
aqueous samples !astes that contain suspended solids
and mobility-procedure e(tracts
total metals 4000
aqueous samples !astes that contain suspended solids
and mobility-procedure e(tracts
total metals 40.0 sediments sludges and soil samples
total metals 40., sludges sediment soil and oil
7rganic Solvents
$t !as recognized that enhanced absorption pea9s could be obtained from solutions
containing lo!-molecular-!eight alcohols esters or 9etones. The effect of organic
solvents is largely attributable to an increased nebulizer efficiency= the lo!er the surface
tension of such solutions results in smaller drop sizes and a consequent increase in the
amount of sample that reaches the flame. A most important analytical application of
organic solvents to flame spectroscopy is the use of immiscible solvents such as methyl
isobutyl 9etone to e(tract chelates of metallic ions. The resulting e(tract is then
nebulized directly into the flame. The sensitivity is increased not only by the
enhancement of absorption pea9s due to the solvent but also by the fact that for many
systems only small volumes of the organic liquid are required to remove metal ions
quantitatively from relatively large volumes of aqueous solution.
"alibration "urves
$n theory atomic absorption should follo! 'eer/s la! !ith absorbance being directly
proportional to concentration. :o!ever departures from linearity are often encountered.
A calibration curve that covers the range of concentrations found in the sample should be
prepared. Any deviation of the standard from the original calibration curve can then be
used to correct the analytical result.
The idealized calibration or standard curve is stated by 'eer#s la! that the absorbance of
an absorbing analyte is proportional to its concentration.
Standard Addition Fethod
$t is !idely used in atomic absorption spectroscopy in order to partially or !holly
counteract the chemical and spectral interferences introduced by the sample matri(.
Techni5ues of /nternal Standards
.imitations of /nternal Standard (/S
$S !ill be present in the atomization as the analyte is present and so the $S cannot be present in
the sample to start !ith. $t should be added to the sample *$f present in the original sample $S
should have a constant amount.
$n arc and;or spar9 methods the $S should have the same comparable volatilization as the
analyte.
$S H analyte should have similar e(citation energy
$SH analyte should have similar $.D.#s. <esult: both analyte and $S give rise to similar intensities
have similar intensities.
Similar atomic !eights since diffusion depends upon !eight.
These si( items are especially important !hen A&S quantitative !or9 is involved.
Applications of Atomic Absorption Spectrometry
Cetection ?imits
Accuracy
/nstruments for the Flame "mission Spectroscopy
$nstruments for flame emission are ire similar in design to flame absorption instruments
e(cept that in the former the flame no! acts as the radiation source a hollo!-cathode
lamp and chopper a(e therefore tin-necessary.
&: Atomic %luorescence Spectroscopy
$nstrumentation
Sources
A continuum source !ould be desirable for atomic fluorescence measurements.
:o!ever the output po!er of most continuum sources over a region as narro! as an
atomic absorption line is so lo! as to restrict the sensitivity of the method severely.
:ollo! "athode ?amp- only observed the fluorescent signal during pulses
&lectrodeless Cischarge ?amp- produced intensities that e(ceed those of hollo!
cathode lamps= ho!ever it is not available for many elements
?asers- ideal source !ith high intensities and narro! band!idths= ho!ever the
high cost has discouraged their routine application
*ispersi2e /nstruments
They are made up of a modulated source an atomizer a monochromator or an
interference filter system a detector and a signal processor and readout.
4ondispersi2e /nstruments
They ideally are made up of a source an atomizer and a detector.
Advantages:
Simplicity and lo!-cost instrumentation
<eady adaptability to multi-element analysis
:igh-energy throughput and thus high sensitivity
Simultaneous collection of energy from multiple lines enhancing sensitivity
$nterferences
They appear to be of the same type and of about the same magnitude as those found in
atomic absorption spectroscopy.
Applications
Atomic fluorescence methods have been applied to the analysis of metals in such
materials as lubricating oils sea!ater biological substances graphite and agricultural
samples.
&eferences
!!!.anachemumu.se.htm
!!!.aurora-instr.com;right.htm
!!!.anachem.umu.se;>umpstation.htm
!!!.anachem.umu.se;cgi;>umpstation.e(eIAtomicSpectroscopy
!!!.anachem.umu.se;cgi;>umpstation.e(eI7pticalFolecularSpectroscopy
!!!.minyos.its.rmit.edu.au;Jrcmfa;mstheory.html
http:;;science.!idener.edu;sub;ftir;introKit.html
http:;;!!!.s-a-s.org;
http:;;!!!.chems!.com
http:;;!!!.scimedia.com;chem-ed;spec;atomic;aa.htm
http:;;nercdg.org
http:;;!!!.analyticon.com
!!!.lcgmag.com;
!!!.lcms.com;
!!!.dq.fct.unl.pt;A7%;"hroma.html
!!!-ssg.chem.utas.edu.au;
!!!.yahoo.com;science;chemistry;chromatography;
!!!.onlinegc.com
http:;;!!!.scimedia.com;chem-ed;analytic;ac-meth.htm
http:;;!!!.scimedia.com;chem-ed;spec;atomic;aa.htm
http:;;!!!.scimedia.com;chem-ed;spec;atomic;afs.htm
http:;;!!!.cee.vt.edu;programKareas;environmental;teach;smprimer;aa;aa.htmlL%eatures
http:;;!!!.anachem.umu.se;aas;gfaas.htm
http:;;!!!.agsci.ubc.ca;fnh;courses;food400;atomic;aatomic04.htmLelectro
http:;;!!!.scpscience.com;products;AA;hollo!lamps.asp
http:;;las.per9inelmer.com;catalog;Droduct.asp(IDroduct$dM840.06-,
http:;;elchem.9aist.ac.9r;vt;chem-ed;spec;atomic;aa.htm