AAS vs ICP Techniques

If you need to know how much lead is in drinking water, how much iron is in a supplement or how much cadmium is in rice, you’ll use an atomic spectroscopy technique. The three main options are atomic absorption spectroscopy (AAS), inductively coupled plasma optical emission spectroscopy (ICP-OES) and inductively coupled plasma mass spectrometry (ICP-MS). All three turn the sample into free atoms or ions and measure them, but they differ hugely in sensitivity, speed and cost.

Quick comparison

Feature	Flame AAS	Graphite furnace AAS	ICP-OES	ICP-MS
Atomisation / ionisation source	flame (about 2300–3000 °C)	electrically heated graphite tube	argon plasma (about 6000–10 000 K)	argon plasma
What is measured	light absorbed by ground-state atoms	light absorbed by ground-state atoms	light emitted by excited atoms and ions	ions sorted by mass-to-charge ratio
Elements per run	one (lamp per element)	one	many at once (dozens)	many at once (most of the periodic table)
Typical detection limits	mg/L to high μg/L	low μg/L	low μg/L	ng/L (parts per trillion) or lower
Sample throughput	fast for one element	slow	high	high
Isotope information	no	no	no	yes
Instrument cost	low	moderate	moderate to high	high
Running cost	low (acetylene, air)	moderate	higher (argon)	highest (argon, cones, expertise)

How each works

AAS

A lamp made of the element being measured (a hollow-cathode lamp) emits light at exactly the wavelengths that element absorbs. The sample is sprayed into a flame (or dried and heated in a graphite furnace), where it’s converted into free atoms. Ground-state atoms of the target element absorb some of the lamp’s light. The amount absorbed follows the Beer–Lambert law and is compared with standards on a calibration curve.

Because the lamp emits only that element’s lines, AAS is very selective. The trade-off is that you need a different lamp for each element, and usually measure one element at a time. See atomic absorption spectroscopy.

Graphite furnace AAS places a tiny drop of sample in a graphite tube and heats it in stages. All the atoms are released into the light path at once in a small volume, giving detection limits around a hundred times lower than a flame, at the cost of speed.

ICP-OES

The sample is sprayed into an argon plasma, an extremely hot, electrically conducting gas sustained by a radio-frequency coil. At these temperatures, atoms are not only freed but excited and often ionised, and they emit light at their characteristic wavelengths as electrons fall back to lower levels (the same principle as a flame test, but far hotter and more efficient). A spectrometer measures many emission lines simultaneously, so dozens of elements are measured in one run.

ICP-MS

The same argon plasma turns the sample into positive ions. Instead of measuring light, the ions are extracted through small cones into a mass spectrometer that separates them by mass-to-charge ratio. Because it counts ions directly, ICP-MS is extraordinarily sensitive, and it can measure individual isotopes. See ICP-MS and mass spectrometry.

Sensitivity: how low can each go?

Very roughly, for many metals:

• Flame AAS: tens to hundreds of μg/L
• ICP-OES: single-digit μg/L
• Graphite furnace AAS: below 1 μg/L for many elements
• ICP-MS: ng/L, sometimes lower

This matters for regulation. The WHO guideline for lead in drinking water is 10 μg/L, and many labs need to measure well below that to show compliance confidently. Flame AAS generally can’t do that for lead; graphite furnace AAS, ICP-OES (for some cases) and ICP-MS can. See how drinking water is tested.

Interferences

Every technique has problems to manage:

• AAS: chemical interferences (e.g. phosphate forming stable compounds with calcium in the flame, reducing free Ca atoms; fixed by adding lanthanum or strontium as a “releasing agent”); ionisation of easily ionised elements; background absorption from molecules and smoke.
• ICP-OES: spectral overlaps, where emission lines of different elements lie close together; solved by choosing alternative lines.
• ICP-MS: isobaric and polyatomic interferences, where different species have the same mass. For example, ⁴⁰Ar³⁵Cl⁺ from the argon plasma and chloride in the sample has mass 75, the same as the only stable isotope of arsenic, ⁷⁵As. Modern instruments use collision/reaction cells to remove such interferences.

Isotopes: ICP-MS’s unique advantage

Because ICP-MS measures mass, it can:

• measure isotope ratios, used in geology, nuclear forensics and food origin testing (for example, lead isotopes to trace pollution sources, or art authentication)
• use isotope dilution, adding a known amount of an enriched isotope as an internal standard, the most accurate quantitative method available
• detect elements whose lines are hard to measure optically

Choosing the right technique

Situation	Best choice	Why
A school or small lab measuring a few metals at moderate levels	flame AAS	cheap, simple, robust
Sodium, potassium, calcium in food or fertiliser at high levels	flame AAS or ICP-OES	concentrations are high; sensitivity not critical
Many elements in soil or wastewater	ICP-OES	multi-element, tolerates dirty samples well
Lead, arsenic, cadmium, mercury in drinking water at trace levels	ICP-MS (or graphite furnace AAS)	needs very low detection limits
Trace metals in blood or urine for clinical toxicology	ICP-MS	sensitivity and multi-element capability
Isotope ratios or isotope dilution	ICP-MS	only technique that measures isotopes
One element, very low level, limited budget	graphite furnace AAS	high sensitivity at lower cost

Worked example: choosing for a project

A local council wants to survey 300 garden soil samples for lead, cadmium, arsenic, copper and zinc, and report results against guideline values that are in the mg/kg range.

• Five elements per sample and 300 samples → multi-element, high throughput.
• Guideline values in mg/kg → after digestion and dilution, concentrations in solution are usually well within ICP-OES range.
• Choice: ICP-OES, with ICP-MS reserved for any samples where arsenic or cadmium is near the detection limit. Flame AAS would work but would mean changing lamps and running each sample five times.

Sample preparation (common to all)

Most samples must be converted into a clear acidic solution:

• Liquids such as water may just need acidifying and filtering.
• Solids such as food, soil and tissue are digested, usually with concentrated nitric acid (sometimes with hydrogen peroxide or other acids), often in sealed vessels heated by microwaves. This destroys organic matter and dissolves the metals.
• Blanks, certified reference materials and spiked samples are run to check for contamination and recovery.

See sampling for why the step before preparation matters most.

Key takeaways

• AAS measures light absorbed by atoms of one element at a time; it’s cheap, selective and robust.
• Graphite furnace AAS improves sensitivity greatly but is slower.
• ICP-OES measures light emitted from a hot argon plasma and handles many elements at once.
• ICP-MS measures ions by mass, giving the lowest detection limits and isotope information.
• Choose by required detection limit, number of elements, sample load and budget.