Atomic Absorption Spectrometry: Flame, Graphite Furnace, Cold Vapour

Strip AAS down to the physics and there is one idea. A free, ground-state atom absorbs a photon whose energy exactly matches an electronic transition out of that ground state. The element is the fingerprint, the wavelength is the lock. Lead absorbs at 283.3 nm. Copper at 324.7. Zinc at 213.9. Cadmium at 228.8. Mercury at 253.7. These resonance lines are sharp, narrow and unique enough that a single-wavelength readout is workable.

Beer-Lambert turns absorbance into concentration. A equals log of I-zero over I, and over the linear range A is proportional to the number of absorbing atoms in the light path. The instrument's job is to put free atoms in that path, hold them long enough to read, and isolate the resonance line.

The classical optical chain is short. A hollow-cathode lamp emits the element's resonance lines. The beam passes through the atom cell (flame, graphite tube, or quartz vapour cell). A monochromator (usually a Czerny-Turner with a holographic grating) picks out the resonance line. A photomultiplier reads the intensity, and the absorbance follows.

The hollow-cathode lamp is worth examining in detail. The cathode is made of the analyte element, filled with argon or neon at a few torr. Strike a few hundred volts across it, the fill gas ionises, ions sputter the cathode, and the sputtered atoms emit characteristic lines. One lamp per element is conventional, though multi-element lamps exist for combinations like Ca-Mg or Cu-Fe-Mn. Modern continuum-source instruments like the Analytik Jena ContrAA replace the HCL with a xenon arc, but most working benches in India still run the classical HCL design.

The atomiser is what changes between AAS modes. Same lamp, same monochromator, same detector. Different way of turning a liquid sample into free gas-phase atoms.

Flame AAS uses a slot burner fed with a fuel and an oxidant. Air-acetylene at about 2300 degrees C handles the easy elements: Pb, Cu, Zn, Cd, Ca, Mg, Fe, Mn, Ni. Refractory elements that hold their oxide bonds at that temperature need nitrous oxide-acetylene at about 2800 degrees C: Al, Cr, V, Ti, Si, B, Mo. The sample is pulled up by a pneumatic nebuliser, mixed with the fuel and oxidant in a spray chamber, and a few percent of the mist actually reaches the flame. Detection sits in the µg/mL (ppm) range, which is fine for acute high-dose forensic cases.

Graphite-furnace AAS, also called electrothermal AAS, swaps the flame for a hollow graphite tube about 28 mm long. A 5 to 50 µL aliquot is pipetted into the tube. The tube is heated electrically through a programmed temperature ramp: dry at about 110 degrees C to evaporate the solvent, ash at 400 to 1200 degrees C to char away matrix organics, atomise at 2000 to 2700 degrees C to release the analyte as free atoms, then a clean step above the atomisation temperature to burn off residue. The atomic vapour sits in the light path for one to two seconds, long enough for a sharp absorbance peak. Sensitivity is roughly 100 times better than flame, with detection in the ng/mL (ppb) range, which is the regime needed for paediatric blood lead, chronic arsenic and routine cadmium and selenium work.

Cold-vapour AAS exploits an oddity of mercury. Mercury is the only metal with a measurable vapour pressure at room temperature, which means you do not need a flame or a furnace to make Hg vapour. The digest is mixed in a reaction vessel with stannous chloride (or with sodium borohydride for greater sensitivity), Hg2+ is reduced to Hg-zero, and an argon stream sweeps the elemental mercury vapour through a long quartz absorption cell. Absorbance is read at 253.7 nm. Detection at the µg/L range without any heating, and the method is fast, cheap and the global standard for mercury in blood, urine and fish.

Hydride generation AAS extends the same trick to elements that form volatile hydrides. Sodium borohydride in dilute hydrochloric acid reduces As(III), Sb(III), Se(IV), Bi, Te, Sn and Ge to their gaseous hydrides (arsine AsH3, stibine SbH3, selenium hydride H2Se, etc). An argon stream carries the hydride into a heated quartz cell at about 900 degrees C, where the hydride dissociates back to free atoms. Detection in the ng/mL range, with arsenic at 193.7 nm being the workhorse application for the Bengal-Bihar arsenicosis caseload.

Reference resonance wavelengths: Pb 283.3 (or 217.0), Cu 324.7, Zn 213.9, Cd 228.8, Hg 253.7, As 193.7, Mn 279.5, Fe 248.3, Ca 422.7, Mg 285.2 nm. The shorter UV lines (As 193.7, Zn 213.9) need careful background correction because matrix scatter rises sharply below 220 nm.

AAS looks like a clean technique on paper, but a real digest carries a lot more than the analyte. Four interference families show up routinely.

Spectral interference is the rarest in AAS because the resonance lines are so narrow and the HCL emits only the element's lines. The classical example is the iron 271.903 nm line sitting close to platinum 271.904 nm, which only matters when both are present at high concentration. The much more common form is broadband molecular absorption from the matrix: a sodium chloride matrix in seawater or in saline-soaked tissue absorbs continuously across the UV at high concentration. The fix is background correction, covered in the next section.

Chemical interference is the most common nuisance. The hot atomiser does not always break every bond. Calcium with phosphate forms refractory calcium phosphate, Ca3(PO4)2, which does not atomise efficiently in an air-acetylene flame and depresses the calcium reading. Aluminium does the same to many alkaline-earth elements. The classical fix is a releasing agent, usually a large excess of lanthanum or strontium chloride, which preferentially binds the interferent (lanthanum phosphate forms in place of calcium phosphate) and frees the analyte. For nitrous oxide-acetylene work the higher temperature solves most chemical interferences directly.

Ionisation interference shows up with the easily ionised elements, the alkali and alkaline-earth metals. In a 2800 degree N2O-acetylene flame, sodium and potassium ionise rather than atomise, the absorption signal at the resonance wavelength drops, and the calibration goes nonlinear. The fix is an ionisation suppressor, typically 1000 mg/L of caesium chloride or potassium chloride added to every standard and sample. The high concentration of an even more easily ionised cation floods the flame with electrons and pushes the analyte's ionisation equilibrium back toward the neutral atom.

Physical interference is the most operationally important category. Two solutions of the same lead concentration but different viscosity (a clean aqueous standard versus a heavily acidic urine digest) nebulise at different rates, deliver different amounts of lead per second to the flame, and read different absorbances. The fix is matrix matching: standards should match the sample in acid concentration, total dissolved solids and density. Where matrix matching is impractical, standard addition does the job by building the calibration line within the sample matrix itself.

Modern AAS instruments operate with background correction as standard practice, particularly in the UV below 250 nm, where matrix scatter is most significant. Four correction methods are in routine use.

The deuterium-lamp continuum-source method is the simplest. A deuterium arc lamp emitting a continuum across the UV is alternately pulsed through the same atom cell. The HCL signal carries atomic absorption plus background, the deuterium signal carries background only, and subtraction leaves the atomic signal. It works only in the UV (about 190 to 350 nm) and is the default on entry-level flame instruments.

The Zeeman effect method is the high-end correction for graphite furnace work. A magnetic field around the atom cell splits the absorption line into a central pi component and two flanking sigma components. With a polariser in the beam, switching the magnet on and off lets the analyst alternately read total absorbance and background-only absorbance. The PerkinElmer PinAAcle 900Z and the Analytik Jena ZEEnit are the two Zeeman GFAAS instruments most often seen at NABL-accredited Indian FSLs. Zeeman handles structured background that deuterium misses and works across the UV-visible range.

The Smith-Hieftje method pulses the HCL at very high current so the emission line self-reverses, hollowing out the line centre where the analyte absorbs. The high-current pulse reads background only. Rare on new instruments but still seen on older Buck Scientific units in some state SFSLs.

High-resolution continuum-source AAS, in the Analytik Jena ContrAA 800, replaces the HCL with a high-intensity xenon arc and a high-resolution echelle spectrometer with a CCD detector. CCD pixels around the resonance line read background continuously, the analyte absorbs only at line centre, and correction is built into every measurement. In India, continuum-source instruments are in use at NFSU Gandhinagar and a number of newer NABL-accredited laboratories.

The instrument is the easy part. Sample preparation decides whether the result is real.

Biological matrices need digestion. Closed-vessel microwave digestion in concentrated nitric acid plus 30 percent hydrogen peroxide at 180 to 200 degrees C for 15 to 30 minutes is the modern standard at any NABL-accredited Indian FSL. About 0.5 to 1 g of liver, kidney or whole blood gives a clear aqueous digest in 25 to 50 mL final volume. Older labs still run open-vessel hot-plate wet digestion in nitric-perchloric mixtures, which loses volatile mercury and arsenic and carries an explosion risk microwave digestion does not.

Calibration follows the matrix. A clean digest takes external calibration, four to six standards bracketing the expected concentration plus a reagent blank, with linearity at R-squared above 0.999. A complex digest (urine, incompletely digested viscera) takes standard addition: spike at three or four concentrations and read from the negative x-intercept.

Method validation under ISO/IEC 17025 sets the floor. Linearity R-squared above 0.999. LOD by the 3-sigma method on seven blank replicates, LOQ by 10-sigma. Recovery 90 to 110 percent on a spiked matrix-matched sample. RSD below 5 percent on duplicates. Each batch then runs with a method blank, reagent blank, CRM (SRM 955c blood lead, SRM 1577c bovine liver) and duplicates on at least 10 percent of samples.

1. 1. Acid digestion of the matrix
   0.5 to 1 g of biological matrix into a closed PTFE microwave vessel with 5 to 8 mL concentrated nitric acid and 1 to 2 mL 30 percent hydrogen peroxide. Microwave-assisted digestion at 180 to 200 deg C for 15 to 30 minutes. Cool, dilute to 25 or 50 mL with deionised water, filter if needed.
2. 2. HCL warm-up and instrument setup
   Mount the element-specific HCL, set the lamp current to manufacturer recommendation, allow 15 to 30 minutes warm-up. Set the resonance wavelength, slit width, burner height for flame work or the temperature programme for GFAAS. Light the flame on air-acetylene or N2O-acetylene.
3. 3. Calibration
   External calibration with four to six standards in matrix-matched acid for clean digests, plus a reagent blank. Standard addition for urine or incompletely digested matrices: spike at three to four concentrations and read from the negative x-intercept. Require R-squared above 0.999.
4. 4. Sample analysis with QC
   Run method blank, reagent blank, CRM and duplicates with every batch. Recovery 90 to 110 percent on the CRM. Spike recovery on at least 10 percent of samples. RSD below 5 percent on duplicates. Background correction (deuterium for flame UV, Zeeman for GFAAS) on throughout.
5. 5. Reporting
   Result with matrix, method (e.g., GFAAS at 283.3 nm with Zeeman correction), LOD, LOQ, measurement uncertainty, CRM result against certified value, and the relevant medico-legal cut-off (paediatric blood lead 5 µg/dL screen, occupational 40 µg/dL, etc).

The case mix at any working Indian SFSL keeps AAS busy. Five recurring scenarios:

Blood lead is the highest-volume AAS workload at AIIMS Forensic Toxicology, PGI Chandigarh and most state SFSLs. Paediatric screen at 5 µg/dL by GFAAS or LeadCare-II venous draw, intervention at 10 µg/dL, adult occupational action at 30 to 40 µg/dL under the Factories Act for battery, smelter and lead-paint workers, acute encephalopathy above 80 µg/dL. The reference method is NIOSH 8003 and the Patancheru and Howrah battery-recycling clusters drive continuing paediatric caseload.

Arsenic in hair and nails is the chronic-exposure assay on HG-AAS or GFAAS at 193.7 nm. Hair arsenic above 1 ppm flags chronic exposure, and segmental hair analysis at 1 cm sections gives a rough timeline. The West Bengal arsenic groundwater monitoring programme, led by WBPCB and the Central Groundwater Board, feeds AAS work at Kolkata FSL Belgachia and at Jadavpur University SOES.

Mercury work is almost entirely CVAAS at 253.7 nm. Urine mercury above 50 µg/L flags organic mercury exposure. The WHO drinking-water guideline of 1 µg/L sets the regulatory frame, with dental amalgam, fish (Kerala coastal cohorts) and small-scale gold mining as recurring exposure routes.

Cadmium and copper round out the routine panel. Blood cadmium above 5 µg/L in non-smokers is occupational exposure. Copper is the Wilson versus copper sulphate workup at PGI Chandigarh: serum ceruloplasmin (low in Wilson, normal in poisoning) and 24-hour urinary copper after a penicillamine challenge settle it. Thallium has no physiological level, any detection is presumptively poisoning, GFAAS at 276.8 nm with Zeeman is the standard method.

The instrument map at major Indian labs: