1. Atomic Absorption Spectroscopy (AAS)
See also: Fundamental reviews in Analytical Chemistry
e.g. Bings, N. H.; Bogaerts, A.; Broekaert, J. A. C. Anal Chem. 2002, 74, (“Atomic Spectroscopy”)
1802 Wollaston observes absorption lines in solar spectrum
1914 Hollow cathode lamp
1955 Walsh describes analytical AAS
1959 1st Commercial Flame AAS
1960s L’vov and Massman describe graphite furnace
(commercial in 1970s)
Recall: A = -log10(T)

2. Hollow Cathode Lamp
Typical primary source of radiation: Hollow cathode lamp
* Typically one lamp per element
* Different intensities for different elements
* Multielement lamps for multielement analysis
Continuum sources (e.g. Xe arc lamp) only for multielement analysis
Kellner et al.,
Analytical Chemistry

3. Ideal Atomizer Provide complete atomization of the element of interest
Atomic vapor should not be highly diluted by the atomizer gas
Excitation of the analyte (and other species) should be minimal

4. Flame AAS
At <5000 K most atoms are predominantly in their electronic ground state.
Slot burners with 5-10 cm path lengths.
Kellner et al.,
Analytical Chemistry
Ingle and Crouch, Spectrochemical Analysis

5. Electrothermal Atomization
Heating current of several hundred A
Heating rates of up to 1000 °C/s
LOD 100 times lower than flame AAS
Heated in three stages:
Ingle and Crouch, Spectrochemical Analysis

6. Electrothermal Atomization
Typical furnace material: Graphite
 Graphite Furnace AAS
Graphite tube mm
Samples uL
200 to 1000 cycles
Temperature up to 3000 °C to avoid graphite decomposition
Carbon may be reducing agent for metal ions
Argon flow avoids oxidation
Other furnace materials: Ta, W, Pt
High melting point required
Should not emit brightly at high temperature (disadvantage for W
and Ta)

7. Are you getting the concept?
Is ICP a good source for AAS?

8. Double-Beam AAS Single-Beam I0 Double-Beam Isample
Skoog Principles of Instrumental Analysis
What is the problem with just measuring Isample/I0?

9. Background Sources of Background: scattering or molecular emission
Background Correction:
With blank sample
Ac = At - Ablank
Deuterium lamp (arc in deuterium atmosphere;
continuum nm)
 absorption of deuterium lamp represents Abackground
 absorption of HCL radiation represents At
Advantage over blank sample: observe fluctuations in flame
Culver et al, Anal. Chem., 47, 920, 1975.

10. Background Correction with Zeeman Effect
Lines differ by ~0.01 nm
Ingle and Crouch,
Spectrochemical Analysis

11. Background Correction with Zeeman Effect
Unpolarized light from HCL (A) passes through the rotating polarizer (B)
Light is separated into perpendicular and parallel components (C)
The light enters the furnace with an applied magnetic field, producing
3 absorption peaks (D)
Either analyte or analyte + matrix absorb light (E)
A cyclical absorbance pattern results (F)
Subtract absorbance during perpendicular half of cycle from absorbance
during parallel half of cycle to get the background corrected value
Skoog, Principles of Instrumental Analysis

12. Background Correction with Zeeman Effect
DC on atomizer
AC on atomizer
DC on source
Ingle and Crouch, Spectrochemical Analysis

13. AAS: Figures of Merit
Linearity over 2 to 3 concentration decades (can be problem for
multielement analysis)
Probability for line overlap is small  Resolution not as critical
as for AES
Precision: Typically a few % for graphite furnace
0.3 to 1.0% for flame
Accuracy: Largely determined by calibration with standards
Applicability: Limited for certain elements for which flame or
furnace is not hot enough (e.g. W, Ta, Nb).
Flow rates of flame are compatible with HPLC flow rates.
Speed: Multielement analysis with multiple HCLs may require lamp exchanges to select desired elements. This is tedious and also costs light because of beam splitters.

14. Method of Standard Additions
In order to quantitate the element of interest in a sample, it is necessary to calibrate with the method of standard additions.
The analytical signal for the sample, Sx, is obtained (after measuring the blank signal).
A small volume, Vs, of a concentrated standard solution of known concentration, cs, is added to a relatively large volume, Vx of the analytical sample.
The analytical signal for the standard addition solution, Sx+s, is obtained.
cx = (SxVscs)/[Sx+s(Vx+Vs) – SxVx]
if Vs << Vx
cx = (SxVscs)/[(Sx+s – Sx)Vx]

15. Are you getting the concept?
The determination of Pb in a brass sample is done with AAS. The 50 mL original sample was introduced into the instrument and an absorbance of was obtained. To the original solution, 20 mL of a 10.0 mg/mL Pb standard was then added. The absorbance of this solution was Find the concentration of Pb in the original sample. What assumption has been made in order to use a single standard addition?

16. AAS: Figures of Merit
Detection limits: * Generally lower LOD for very volatile elements
* Higher LOD for carbide-forming elements (e.g. Ba, B, Ca, Mo, W, V, Zr)
* Concentration in GF up to 1000 times higher than in flame; much lower LOD for GF.
* Lower LOD for GF-AAS than ICP-AES unless
atomization requires high temperature
* Generally similar LOD for flame-AAS and ICP-AES
* Generally higher LOD for very volatile elements
Chemical * HCl often avoided as acid in GF-AAS because
interferences: metal chlorides are more volatile than sulfates or
phosphates.
* Addition of Cs salt to sample suppresses ionization.
* La precipitates phosphate, facilitating Ca analysis.
* Proteins may clog burners and are precipitated with
trichloroacetic acid.

17. mA: Concentration giving rise to 1% absorption.