Flame Test

What actually happens in the flame

A flame test looks like magic — touch a salt to a flame and it blooms into a single vivid color — but the mechanism is pure quantum mechanics. Three things happen in quick succession: atomization, excitation, and emission.

First, the metal compound must break apart into free, gaseous atoms. A solid like sodium chloride cannot emit; its electrons are locked into a lattice. The flame supplies the thermal energy to vaporize the salt and dissociate it into individual atoms (and ions). This is why the wire loop is moistened with concentrated hydrochloric acid before sampling: HCl converts the salt into the corresponding metal chloride, which is more volatile and decomposes cleanly at flame temperature. The familiar reaction for an oxide or carbonate is simply conversion to MCl_n followed by thermal dissociation: MCl → M(g) + Cl(g) in the hot gas.

Second, collisions in the ~1500 °C flame (a roaring premixed Bunsen flame can reach ~1900 °C; a candle is far cooler) transfer enough energy to bump a valence electron from its ground-state orbital up to a higher, empty one. For sodium that means promoting the lone 3s electron into the 3p orbital. The atom is now electronically excited — an unstable, transient state that lasts only nanoseconds.

Third, the electron falls back down. Energy is conserved, so the exact energy difference between the two levels is carried away by a single photon. The photon's wavelength is fixed by the Planck relation:

E_photon = E_upper − E_lower = hc / λ

where h is Planck's constant (6.626 × 10⁻³⁴ J·s) and c is the speed of light. Rearranging, λ = hc/ΔE. For sodium's 3p→3s drop, ΔE ≈ 2.10 eV, which gives λ ≈ 589 nm — the intense yellow of a street sodium-vapor lamp, of salt thrown in a campfire, of a flame test gone slightly wrong. Because every element has a different ladder of energy levels, the set of gaps — and therefore the set of emitted wavelengths — is unique. That is the whole principle: the color is an atomic fingerprint.

The colors and the numbers behind them

The metals that give clean, bright flame colors are overwhelmingly the alkali metals (Group 1) and alkaline-earth metals (Group 2), plus copper and boron. These elements have low-lying excited states whose transitions land squarely inside the visible window (roughly 400-700 nm). The transition metals and heavier metals mostly disappoint: their lowest electronic transitions often sit in the ultraviolet, so the emitted photons are invisible to the eye and the flame stays colorless or merely orange-yellow from incandescent soot.

Metal ion	Flame color	Dominant wavelength(s)	Transition / note
Lithium (Li⁺)	Crimson red	671 nm	2p→2s; deep, slightly pink-red
Sodium (Na⁺)	Intense yellow	589.0 / 589.6 nm (D-doublet)	3p→3s; overwhelms everything
Potassium (K⁺)	Lilac / pale violet	766.5 / 769.9 nm + 404 nm	4p→4s; view through cobalt glass
Rubidium (Rb⁺)	Red-violet	780 / 794 nm	5p→5s; faint, near-IR component
Cesium (Cs⁺)	Blue-violet	455 / 459 nm	6p→6s; one of the few blue alkalis
Calcium (Ca²⁺)	Orange-red / brick	~622 nm + CaOH bands	molecular CaOH/CaO emission
Strontium (Sr²⁺)	Scarlet / crimson	~605-687 nm (SrCl, SrOH)	red fireworks; molecular bands
Barium (Ba²⁺)	Apple / yellow-green	~524 / 514 nm	green fireworks; weak, easily masked
Copper (Cu²⁺)	Blue-green / green	~510-530 nm (CuCl bands)	chloride sharpens the color
Boron (B)	Bright green	~518 / 546 nm	BO₂ molecular emission

Notice a subtlety: not every flame color comes from atomic line emission. Strontium's scarlet, calcium's brick-red, copper's blue-green, and boron's green are largely molecular band emission — short-lived species such as SrCl, SrOH, CaOH, CuCl, and BO₂ that form transiently in the flame and emit broad bands rather than razor-sharp lines. This is why these colors look like a smear of related hues instead of one pure tone, and why the choice of anion (chloride vs nitrate vs sulfate) shifts the exact shade. Sodium and lithium, by contrast, are nearly pure atomic-line emitters, which is why their colors are so clean and saturated.

The energetics are worth feeling. A 589 nm photon carries E = hc/λ = (1240 eV·nm)/589 nm ≈ 2.10 eV ≈ 3.4 × 10⁻¹⁹ J. A blue 455 nm cesium photon carries ≈ 2.73 eV; a red 671 nm lithium photon only ≈ 1.85 eV. Shorter wavelength means a bigger energy gap means more energetic photons — which is exactly why blue flame colors (cesium, copper) require larger electronic jumps than red ones (lithium, strontium).

The tyranny of sodium

If you run flame tests in a teaching lab you will quickly meet the single biggest practical headache: sodium contamination. The sodium D-line at 589 nm is one of the strongest emission lines in all of chemistry because the 3s→3p transition is fully allowed and sodium is everywhere — in glass, dust, fingerprints, sweat, and tap water. A few micrograms of Na are enough to flood the flame with yellow and bury the faint lilac of potassium or the apple-green of barium.

The classic countermeasure is optical, not chemical: view the flame through cobalt-blue glass or a didymium (neodymium/praseodymium) filter. Cobalt glass strongly absorbs around 589 nm, subtracting the sodium yellow while transmitting the blue and violet wavelengths, so potassium's hidden lilac suddenly becomes visible. This trick — selectively removing an interfering wavelength so a weaker one can be seen — is conceptually the same as the wavelength selection a monochromator performs in a real spectrometer.

Running the test cleanly

• Use the right wire. Platinum is ideal because it is inert and gives no color of its own; nichrome is a cheaper substitute that works for most ions.
• Clean between samples. Dip the loop in concentrated HCl and burn it off repeatedly until the flame shows no color — residual sodium especially clings.
• Moisten with concentrated HCl. This forms volatile chlorides, aids atomization, and produces the molecular chloride bands that sharpen copper and strontium colors.
• Use the hot, non-luminous (blue) flame. A yellow luminous flame is full of glowing soot that swamps faint colors; open the air hole to get the clean blue cone where the temperature peaks.
• Look quickly. The brightest emission appears the instant the sample volatilizes; a brief flash of crimson or green is the diagnostic event.

Why it still matters

The flame test is humble, but its physics underwrites a whole branch of analytical chemistry. Robert Bunsen and Gustav Kirchhoff turned exactly this observation into flame emission spectroscopy in the 1860s, discovering cesium (Latin caesius, sky-blue) and rubidium (rubidus, deep red) by their previously unseen flame lines. Replace the human eye with a monochromator and a photodetector and you have flame photometry, still used in clinical labs to measure sodium, potassium, and lithium in blood serum quickly and cheaply.

Invert the experiment — shine a lamp through the flame and measure the light the atoms absorb rather than emit — and you have flame atomic absorption spectroscopy (AAS), capable of parts-per-billion sensitivity for dozens of metals. The same energy-level diagram governs all three techniques; only the question (does the atom emit, or absorb?) changes.

Technique	Measures	Output	Sensitivity
Flame test (this page)	Visible color emitted	Qualitative — which metal?	~0.1-1% (eyeball)
Flame emission / photometry	Emission intensity at one λ	Quantitative — how much?	ppm (Na, K, Li, Ca)
Flame atomic absorption (AAS)	Light absorbed from a lamp	Quantitative, many metals	ppb
ICP-OES / ICP-MS	Plasma emission / ion mass	Multi-element, trace	ppt-ppb

And there is the spectacle. Every firework you have ever watched is a flame test scaled up to the sky: strontium and lithium salts for crimson reds, barium for green, copper for blue-green, sodium for gold-yellow, calcium for orange. The pyrotechnician's color palette is nothing more than a chart of metal-ion emission wavelengths, lit by a very large, very fast combustion flame.

Limits and pitfalls

• Mixtures confuse the eye. The flame test reports one dominant color; in a mix of ions the brightest emitter (usually sodium) wins, hiding the rest.
• Many metals give no color. Mg, Al, Zn, Fe, Pb and most transition metals have lowest transitions in the UV, outside the visible range — a colorless flame is not proof of absence.
• It is qualitative, not quantitative. Color tells you which ion, never how much; that requires instrumentation.
• Contamination is everywhere. Sodium especially produces false positives; rigorous cleaning and cobalt-glass viewing are essential.