Beer-Lambert Law Calculator (Absorbance)
A more concentrated solution, a longer path through the cell: less light comes out the other side. That is the whole of the Beer-Lambert law. Absorbance on the spectrophotometer then tracks concentration as a straight line — which is why UV-Vis remains the workhorse assay in analytical chemistry, biochemistry and QC.
A = ε · l · c · T = I / I₀ = 10−A
The calculator rearranges the same relation four ways: absorbance, unknown concentration, molar absorptivity ε, or path length. Percent transmittance is reported every time and can be the starting reading.
Your data
An asterisk after a bench term opens its definition in the glossary at the bottom of the page.
Absorbance versus transmittance — why the log belongs in the assay
The detector reports a flux. Ten percent of the beam still reaching it means a transmittance* T of 0.10. That number is easy to picture. It is a poor quantity for an assay. Each successive slice of solution multiplies the remaining intensity by the same factor, so doubling the dye does not halve T: it squares it. Plot T against concentration and the curve sags. There is no useful slope to read an unknown from.
Absorbance* is the logarithm that turns that product into a sum. Each slice of path then adds the same increment, and A becomes linear in c.
T = I / I0 A = −log10 T = log10(I0 / I) T = 10−A
At A = 1, 10% of the incident light is still getting through. At A = 2, 1%. At A = 0, none of it was lost. While Beer–Lambert holds, twice the concentration means twice the absorbance. That is the whole reason analytical chemistry works in A. Older photometers still print percent transmittance; this page takes either and converts cleanly.
A wording trap is worth flagging once. Absorption is the physical process — photons gone. Absorbance is the dimensionless number on the display. Papers from before about 1980 still say optical density. Same A = −log10 T, not a third quantity. Using “absorption” for the reading is the usual slip in English, and it muddles the assay with the physics.
Molar absorptivity ε is a wavelength property
Molar absorptivity* ε states how strongly one mole of the species, in one centimetre of path, attenuates the beam — at one wavelength. The unit is L·mol⁻¹·cm⁻¹. The litre in the numerator is not decoration: l is in cm, c is in mol/L, and only then is ε·l·c dimensionless, as an absorbance must be. Mixing those units is the most common arithmetic accident on this law.
A = ε · l · c
ε is not a passport value for the molecule. It is a point on a spectrum. An azo dye can sit at 104–105 L·mol⁻¹·cm⁻¹ on its visible band; a simple ketone on its n→π* transition near 280 nm often stays below 100. Change the solvent, the pH, the temperature, and ε moves. NADH at 340 nm, 6,220 L·mol⁻¹·cm⁻¹, is one of the few figures stable enough to use as a bench check — and even then, only at 340 nm, not “somewhere in the near UV”.
The absorption spectrum is the plot of ε against λ. A round number such as 500 nm is not a method. The working wavelength is wherever the analyte absorbs and the rest of the flask does not.
Working wavelength: λmax, and when the peak is the wrong place
Scan A against λ and aim for the maximum. At λmax* a small change in c produces the largest change in A, so the calibration slope is steep. Neighbouring species seldom share the exact same peak, which trims interferences. On a plateau, a 1 nm drift of the monochromator barely moves A. On the shoulder of a narrow band, the same drift wrecks the assay.
That said, the peak is not always free. Acetate or nitrate buffers already eat 220 nm, so the assay has to move. A clean shoulder beats a crowded maximum. A wide slit averages the peak and lowers the apparent ε. Textbooks say “work at λmax”. The bench says “work where the assay holds”: the maximum when it is available, a documented alternative when it is not.
UV-Vis* splits naturally. The visible (400–800 nm) covers dyes, metal complexes, coloured extracts. The UV (200–400 nm) opens aromatics, carbonyls, nucleotides. Below about 200 nm, air oxygen and most solvents absorb on their own; that is no longer a routine assay.
What a blank actually subtracts
The blank* is not “water”. It is everything the sample contains except the analyte: solvent, assay reagents, buffer, the cuvette* itself, and whatever is smeared on the windows. Place it in the beam and zero. After that, A reports what was added, not the plastic of a disposable cell or the solvent’s own absorbance.
Skip the blank, or zero in distilled water while the samples sit in ethanol, and the whole series shifts. The calibration curve misses the origin, and the unknown is wrong from the first point. Repeating the blank after a change of cell, solvent or wavelength is not fussiness: it is what keeps A = 0 meaning “no analyte”.
A double-beam instrument leaves the reference in the second path for the whole run. A single-beam instrument needs the blank repeating, because the lamp drifts. In either design, the blank does not cancel a scratched window or a cloudy sample. It subtracts what is in the reference beam, not light scattered somewhere else.
A calibration curve that can be trusted
Rearranging A = ε·l·c with a handbook ε looks tidy and is often too optimistic. That ε was measured in another cell, another solvent, another year, sometimes at a nearby wavelength. Today’s calibration carries the instrument, the cell, the solvent and the room temperature in one go.
- Four or five standards that bracket the unknown — not a single point “because ε is in the CRC”.
- Same cell, same λ, same blank, same solvent.
- Plot A against c. Beer–Lambert promises a line through the origin. If the origin slips, the blank is wrong, the cell moved, or a residual absorbance was never subtracted.
- Read the unknown off that line, inside the span of the standards, not by stretching past the last point.
An r of 0.999 does not excuse a line that misses the origin by 0.04 absorbance units. The intercept is a diagnostic; it is not swept under the fit. And one “check standard” is not a calibration. It is a single measurement with no slope under control.
Why the plot bends once A is past about 1
At A = 1, 10% of the light still crosses the cell. At A = 2, 1%. The detector is then working near its floor, and a little light that never went through the sample — stray light from the monochromator, leaks, scatter off the walls — starts to matter. The instrument over-reports transmitted flux. Measured A is too low, so c comes out low. The plot droops.
Other effects pile on. Aggregates and ion pairing mean that twice the moles is no longer twice the independent chromophores. A wide slit averages the peak. Sample fluorescence adds photons at the detector, in the wrong direction. None of that sits inside A = ε·l·c, which assumes a dilute true solution, nearly monochromatic light, no scatter, no emission.
In practice the working window is 0.1–1. Below 0.05, baseline noise on most instruments eats the signal. Above 1, and certainly past 1.5–2, dilute rather than “read it anyway”. An A of 2.5 that the display is willing to print is not a reliable concentration. It is a hint to start again lower.
Dilute, read, scale back to the stock
A = 1.80 in a 1 cm cell? A tenfold dilution lands near 0.18, inside the linear range. Read cdiluted off the curve, then multiply by the factor. Dropping that factor is the classic transcription error of the assay: the number written down is the flask, not the original aliquot. Record Vtaken / Vflask in the notebook, not from memory.
Make up to the mark, the same way as any dilution of a stock solution. A splash of water in a beaker has no final volume. 1.00 mL of stock in a 10.00 mL volumetric flask is a factor of 10. 2.00 mL in 25.00 mL is 12.5 — and that 12.5 is written down before the run starts, not hunted afterwards.
A sample that is too pale (A < 0.05) is rarely concentrated on the bench. Lengthen the path (a 50 mm cell) or go back to the preparation. Forcing 0.02 into the calculator yields a number; that number is not an assay.
Mass absorptivity and A(1%, 1 cm)
Without a trustworthy molar mass — a polymer, a protein whose M is only an estimate, a plant extract — molar ε has nothing to hang on. Switch to mass absorptivity a, in L·g⁻¹·cm⁻¹, with c in g/L. The plot is still a straight line: A = a·l·c. Same law, different unit of stuff.
Biochemists often quote A(1%, 1 cm): the absorbance of 1 g/100 mL in a 1 cm cell. For bovine serum albumin at 280 nm, A1%280 is commonly about 6.6. One gram per litre in 1 cm then gives A ≈ 0.66. The conversions: if a is in L·g⁻¹·cm⁻¹, A(1%, 1 cm) = 10·a, because 1 g/100 mL = 10 g/L. And ε = a·M once M is known.
Mixing the three notations in one table, without saying so, throws concentrations out by a factor of 10 or by M. This calculator works in molar ε (L·mol⁻¹·cm⁻¹) and mol/L. An A(1%, 1 cm) value has to be converted first.
Scratches, bubbles, scatter and stray light
The equation assumes that every lost photon was absorbed by a molecule. Everything else — scatter, reflection, light that cheats the path — disguises itself as absorbance.
- Scratched or greasy cell. Fingerprints and micro-scratches scatter. A rises with no change in c. Hold optical windows by the top, wipe with optical paper, never a lab coat.
- Bubbles. A meniscus, a bubble stuck to the window, and the beam hits an air–liquid interface. Readings jump. Tap, wait, repeat.
- Turbidity. Particles, a forming precipitate, an emulsion: that is no longer a solution. Filter, centrifuge, or admit that Beer–Lambert does not apply. A blank will not subtract scatter that is only in the sample.
- Stray light. Instruments age, gratings get dirty. High A shows the bias first. A standard at A ≈ 2 that folds while the low points stay linear is often this.
- Temperature. Solution volume changes, so do equilibria, and ε with them. A pH indicator at 28 °C is not the standard calibrated at 20 °C. Enzyme assays are thermostatted; the room is not left to “do the job”.
A cell seated crooked, or rotated 90° with the frosted faces in the beam, gives an A that is both wild and repeatable — the worst kind, because it looks serious. Clear faces toward the beam, frosted faces to the sides. That is not a footnote in the manual.
Watching a reaction through the spectrophotometer
Park the wavelength where a reactant disappears, or a product appears, and the other species in the mix stay quiet. Convert A(t) to c(t) with the same day’s calibration, not with a supplier’s ε. The initial slope of c(t) is the initial rate v₀. A reaction order is read from a plot (ln c, 1/c…), not from a hunch.
If reactant and product both absorb at the chosen λ, total A is no longer one concentration. Move λ, work in difference, or treat the problem as a mixture. An isosbestic point — the wavelength where ε of the reactant equals ε of the product — is useful: it checks that only two spectra are in play, not a third intermediate walking in.
Mixing time, the delay before the first reading, baseline drift over a twenty-minute run: those weigh as much as the formula. A blank at the start and a blank at the end of a long kinetic keep a lamp drift from being mistaken for a slow reaction.
Two absorbers, two wavelengths
In a dilute medium, with no interaction, absorbances add. Two species, one λ: A = ε₁·l·c₁ + ε₂·l·c₂. One equation, two unknowns: not solvable. Two wavelengths, two equations, yes. That is the simultaneous assay, familiar for Fe(II)/Fe(III), a pair of food dyes, or two nucleotides. The ε values come from the pure compounds, at both wavelengths, in the same solvent.
A(λ1) = ε1,1 l c1 + ε2,1 l c2 A(λ2) = ε1,2 l c1 + ε2,2 l c2
This page does not solve the 2×2. It returns one A, one c, one species. As soon as the unknown’s spectrum no longer matches the standard, the single-species model has already failed. A third absorber, a charge-transfer complex, aggregation: additivity lets go. Beer–Lambert is not a mathematical identity. It is a law for dilute solutions.
A = ε·l·c assumes a true solution, nearly monochromatic light, no fluorescence, no scatter. An absorbance that “comes out right” outside those conditions is luck, not evidence that the law stretches.
Path length is not a habit
l sits in the formula. Double the path, double A, at the same c. A pale sample belongs in a 50 mm cell. A sample too dark to dilute belongs in 1 mm. The calculation has to be told. Typing l = 1 “because everyone does” while the cell is 0.50 cm makes c twice too large.
A “10 mm cuvette” is a 1.00 cm path. Entering 10 and leaving the unit on centimetres multiplies the computed A by ten, or divides c by ten, depending on the unknown. It is far and away the most frequent unit error on Beer–Lambert. The form converts mm, cm and m before it applies A = ε·l·c; a literature ε in L·mmol⁻¹·cm⁻¹ pasted next to a concentration in mol/L still throws the result by 1,000.
Quartz for the UV, polystyrene for the visible
The cell is part of the optical path. Fused silica (quartz) transmits down to about 190–200 nm: that is the cell for work below 300 nm. Ordinary glass already cuts a large part of the UV. Disposable polystyrene or PMMA is fine in the visible, useless for a 260 nm DNA reading. Measuring nucleic acids in a plastic cuvette “because it is simpler” is mostly measuring the plastic.
Dead volume, beam height, fill level: a 50 µL microcell only works if the beam actually crosses the window, not the meniscus. Still, the nominal path of a 1 mm cell — flow cell or insert — has to be the number typed in. Material and geometry are chosen before the run, not after a wild A.
NADH at 340 nm, and A280 for proteins
NADH absorbs at 340 nm; NAD⁺ barely does. That is why enzymology watches 340 nm for any reaction that makes or consumes NADH (dehydrogenases, glucose and lactate kits). The ε of 6,220 L·mol⁻¹·cm⁻¹ at 340 nm, 1 cm path, is the most copied coefficient in biochemistry manuals. A = 0.622 then corresponds to 100 µmol/L. A = 0.311, the default on this page, to 50 µmol/L.
Proteins are more often read at 280 nm, where tryptophan and tyrosine absorb. There the convenient figure is A(1%, 1 cm), not a molar ε, because M varies. Neither number is carved in stone. Solvent, pH and slit width nudge the value. For a careful assay, calibrate anyway, in the day’s buffer. The handbook figure is there to catch a factor-of-ten mistake, not to replace the curve.
The relation itself is short. The assay lives in the blank, the wavelength, the cell, the absorbance window, and the dilution factor written in the book.
Glossary
- Absorbance
- Dimensionless quantity A = −log10(I / I0) displayed by the spectrophotometer. Linear with concentration while the Beer–Lambert law holds. Not the same word as absorption, which names the physical process.
- Transmittance
- Ratio T = I / I0 of the flux leaving the cell to the incident flux. Often given as a percentage. T = 10−A.
- Molar absorptivity (ε)
- How strongly one mole of the species attenuates light, at a stated wavelength, per centimetre of path. Usual unit: L·mol⁻¹·cm⁻¹. Depends on solvent, pH and temperature. Also called the molar extinction coefficient.
- Cuvette
- A vessel with flat, parallel windows whose internal width is the path length l. Quartz for the UV, glass or plastic for the visible. “Cell” is the same object in many English-speaking labs.
- λmax
- Wavelength of the species’ absorption maximum, where ε is largest. Usually the working wavelength of the assay, unless an interference or the solvent absorbs there too.
- Blank
- Reference solution that contains the whole matrix except the analyte. Placed in the beam to zero the instrument, so the measured A no longer includes solvent, reagents or the cell.
- UV-Vis
- Spectrophotometry in the ultraviolet (about 200–400 nm) and the visible (400–800 nm). The usual working range of the Beer–Lambert law in analytical chemistry and biochemistry.