Atomic spectroscopy and elemental analysis

An elemental signal answers an elemental question

Atomic spectroscopy uses responses associated with atoms or ions under the method’s conditions. Atomic absorption measures absorbed radiation, while atomic emission measurements observe emitted radiation from suitable excited species. Creating and observing those species changes the relation between the original sample forms and the measured signal.

For example, an inductively coupled plasma atomic-emission procedure, abbreviated ICP-AES, can support particular elemental measurements with appropriate sample preparation and calibration. The method name alone does not define whether a result is dissolved, total-recoverable, total elemental or species-specific. Those quantities depend on the defined preparation and measurand.

Spectral overlap, matrix effects, background, preparation recovery and contamination can affect a result. Measuring an element’s amount does not automatically identify its original oxidation state, molecular binding partner or biological role. The following values are invented quantity models, not a dissolution recipe or actual method validation.

Apply the relevant intercept before the dilution factor

Use a supplied linear response model A=b+mc for final elemental concentration c in mg L⁻¹. Assign intercept b=0.0150, slope m=0.0800 L mg⁻¹ in absorbance units, observed teaching input A=0.175 and dilution factor d=5.00. Then final concentration is (0.175−0.0150)/0.0800=2.00 mg L⁻¹ and original concentration is 10.0 mg L⁻¹.

Ignoring the relevant intercept gives 0.175/0.0800=2.1875 mg L⁻¹ in the final solution and 10.9375 mg L⁻¹ on the original basis, a +9.375% difference from the stipulated corrected value. A background already accounted for in a calibrated intercept should not be subtracted a second time.

This model assumes the applicable reference and matrix, fixed response and a valid linear range. An independently determined different intercept changes the calculation. Do not infer that an actual AAS or ICP response is linear merely because a linear equation has been supplied in a teaching example.

Solution concentration must be connected to sample inventory

Give a hypothetical dry sample mass 0.200 g, quantitative transfer into a supplied final solution volume 0.0500 L, further dilution factor 10.0 and diluted elemental concentration 0.300 mg L⁻¹. The undiluted final solution concentration is 3.00 mg L⁻¹, giving elemental mass 0.150 mg in that final volume.

Divide by dry sample mass 0.000200 kg to obtain 750 mg kg⁻¹ on the stipulated dry basis. This is the inferred elemental mass fraction in a unit convention; it is not the solution concentration 3.00 mg L⁻¹ or a fraction of a named compound. Moisture information would be needed to relate the dry basis to a received wet basis.

Quantitative transfer and valid response are given assumptions, not measured recovery evidence. If a separately established conditional recovery were 0.800 for this analyte and procedure, a justified pre-loss correction would give 750/0.800=937.5 mg kg⁻¹. It should not be applied blindly or multiplied again if the calibration already accounts for that recovery.

A standard-addition extrapolation needs common response and basis

Standard addition supplies known increments to the relevant sample response. In a stipulated normalized final-volume model, signal s=m(c_native,final+c_added,final), with constant matrix and equal native/spike response. The added concentrations must be on the same final-volume basis, rather than raw added amounts with changing volume left uncorrected.

Assign final added concentrations 0, 1.00 and 2.00 mg L⁻¹ with signals 3.00, 5.00 and 7.00 U. The line is s=3.00+2.00c_added in the stated units. With zero independent additive background, native final concentration is intercept/slope=1.50 mg L⁻¹; the x-intercept is −1.50 mg L⁻¹. If original sample makes half the final volume, original concentration is 3.00 mg L⁻¹.

The negative x-intercept is a formal extrapolation, not a prepared solution with negative concentration. The plot distinguishes selected positive additions from the extrapolated model interval. Its inference requires the stipulated linear response, constant matrix, volume normalization and shared native/spike behavior.

Standard addition does not automatically remove an additive signal

Add an unrecognized constant background 1.00 U to each of the three supplied signals. The slope stays 2.00 U per mg L⁻¹, while the intercept rises from 3.00 to 4.00 U. Naive extrapolation now gives final native concentration 2.00 mg L⁻¹ and original 4.00 mg L⁻¹ rather than the stipulated 1.50 and 3.00.

The apparent original concentration is a third higher. Independent knowledge of the additive contribution could allow an appropriate correction, but the standard-addition line alone does not establish it. An unchanged slope and a good straight-line fit cannot separate analyte response from an unknown constant interferent.

This illustrates why selectivity, blank evidence and a defined matrix are necessary alongside calibration. A spike can behave differently from an original bound or inaccessible analyte; supplied equality in this model is not evidence for every real preparation.

The same formula can belong to different molecular graphs

Ethanol and dimethyl ether both have formula C₂H₆O while differing in connectivity. In a nominal bookkeeping model assigning atom masses C=12, H=1 and O=16, each has assigned formula mass 46. Carbon, hydrogen and oxygen contributions are 24, 6 and 16, giving fractions 12/23, 3/23 and 8/23.

These values normalize to one and are identical for both candidate formulas. The nominal model is not an accurate isotope mass or an average molar mass calculation. Even exact knowledge of a formula does not make the two connectivities identical; elemental amounts alone cannot select between them.

Use evidence appropriate to the unresolved structural or chemical-form question, such as suitably interpreted spectra or separations. Keep elemental concentration, sample basis, formula and connectivity as separate claims, with complementary evidence connecting them.

Worked examples

Try it yourself

1. Use applicable intercept 0.0250 instead of 0.0150, keeping A=0.175, slope 0.0800 and dilution factor 5.00. Find original concentration.

   Show the answer and reasoning

   9.375 mg L⁻¹.

   Final concentration is 0.150/0.0800=1.875 mg L⁻¹; multiply by the stated dilution factor.

2. For the dry-basis example, suppose a separately established 0.800 recovery applies and was not already accounted for. Find the pre-loss dry-basis amount.

   Show the answer and reasoning

   937.5 mg kg⁻¹.

   Divide 750 by the stipulated recovery; multiplying back by 0.800 reconstructs the uncorrected inventory. This is not actual recovery-validation evidence.

3. An unrecognized constant 1.00 U is added to every standard-addition signal above. What original concentration would naive extrapolation give?

   Show the answer and reasoning

   4.00 mg L⁻¹.

   The intercept becomes 4.00 U while slope stays 2.00. Final apparent native concentration 2.00 mg L⁻¹ is divided by the sample volume fraction 0.500.

4. Within the same nominal atom-mass convention, do propanal and propanone with C₃H₆O have different elemental mass fractions? Give the fractions.

   Show the answer and reasoning

   No. C=36/58≈0.620690, H=6/58≈0.103448, O=16/58≈0.275862.

   Both have assigned formula mass 58 and the same elemental inventory; connectivity remains different.

What to carry forward

• Define which elemental quantity the preparation and response support.
• Apply the appropriate intercept and sample basis once.
• Separate solution concentration from sample mass fraction.
• Normalize standard additions and check additive interference.
• Shared elemental composition does not uniquely determine connectivity.