Spectrophotometers most commonly report results in absorbance units (often abbreviated AU or simply A) and percent transmittance (%T), both of which describe how much light a sample absorbs or lets through at a given wavelength. But those two are just the starting point. Depending on the type of instrument and the application, you might also see optical density, relative fluorescence units, reflectance percentages, wavenumber shifts, or even direct concentration readouts in milligrams per milliliter. The diversity of units reflects the fact that “spectrophotometer” is an umbrella term covering instruments that do quite different things with light.
Absorbance and Transmittance
The two most fundamental units on any standard spectrophotometer are transmittance and absorbance, and they are mathematically linked. Transmittance is the simpler concept: shine a beam of light through a sample, and transmittance is the fraction of that light that makes it to the detector on the other side. It is usually expressed as a percentage. If your sample lets through 50% of the light, the transmittance is 50%T. A perfectly clear sample would be 100%T, and a completely opaque one would be 0%T.
Absorbance flips this relationship onto a logarithmic scale. An absorbance of 0 means no light was absorbed (100%T). An absorbance of 1 means 90% of the light was absorbed, leaving only 10%T. An absorbance of 2 means 99% was absorbed, leaving 1%T. The logarithmic scale is not just a mathematical preference. It exists because absorbance, unlike transmittance, has a direct linear relationship to the concentration of the absorbing substance. That linearity is what makes quantitative chemistry with a spectrophotometer practical.
Most modern instruments display both units and let you toggle between them. For qualitative work, where you just want to see whether a sample absorbs at a particular wavelength, either unit works. For quantitative work, where you need to figure out how much of something is in solution, absorbance is almost always what you want.
How Absorbance Connects to Concentration
The reason absorbance is so widely used is a relationship known as Beer’s Law (sometimes called the Beer-Lambert Law). It states that absorbance equals the product of a substance’s molar attenuation coefficient, the concentration of that substance, and the distance the light travels through the sample (the path length). In practical terms, if you double the concentration of a colored solution, the absorbance doubles. That proportionality makes it straightforward to build a calibration curve and read off unknown concentrations.
There is, however, a well-known limitation. Beer’s Law holds strictly only when the light hitting the sample is a single wavelength. Real instruments use a narrow band of wavelengths rather than a truly monochromatic beam, and at high concentrations the linear relationship can break down. A detailed analysis based on the Kramers-Kronig relations showed that using integrated absorbance across a spectral range rather than absorbance at a single peak frequency restores linear concentration dependence, with deviations smaller than 0.09% even for neat substances at very high concentrations.1PubMed Central. Beer’s Law‐Why Integrated Absorbance Depends Linearly on Concentration For everyday lab work, the practical message is simpler: stay in the absorbance range where your instrument behaves linearly (usually between about 0.1 and 1.0 AU), and Beer’s Law will serve you well.
Optical Density in Microbiology
Walk into a microbiology lab and you will hear people talking about “OD” constantly. Optical density is numerically identical to absorbance, but in microbiology it measures something quite different from the absorption of dissolved chemicals. When you put a bacterial culture into a spectrophotometer, most of the light that fails to reach the detector is not absorbed by the cells. It is scattered. The cloudier (more turbid) the culture, the more light is scattered away and the higher the OD reading. Microbiologists typically measure at 600 nm (OD600) because most culture media and cell components do not absorb strongly at that wavelength, so the reading reflects turbidity rather than pigment or media color.
The catch is that OD is only proportional to the actual number of cells in the culture up to roughly an OD of 0.1. Above that threshold, light scattered away from the detector by one cell can be scattered back toward it by another cell, so the OD reading climbs more slowly than the true cell count.2PubMed Central. Estimating microbial population data from optical density Researchers who need accurate growth curves typically build a calibration relating OD to viable cell counts (by plating), or they dilute samples back into the linear range before reading them. Optical density data can still yield useful growth-rate information when combined with known initial cell counts and appropriate mathematical models.3PubMed. Growth curve prediction from optical density data But treating a raw OD number as a direct cell count, especially above 0.1, will lead you astray.
The Wavelength Axis
Every spectrophotometer also reports the wavelength of light being used, and this axis carries its own set of units. Ultraviolet-visible (UV-Vis) spectrophotometers express wavelength in nanometers (nm), with the UV range spanning roughly 190 to 400 nm and the visible range extending from about 400 to 700 nm. Near-infrared instruments extend to about 2500 nm.
Infrared spectrophotometers, by contrast, traditionally use wavenumber rather than wavelength. Wavenumber is the reciprocal of wavelength, expressed in inverse centimeters (cm⁻¹). A wavenumber of 1000 cm⁻¹ corresponds to a wavelength of 10 micrometers. The reason infrared spectroscopists prefer wavenumber is that it is directly proportional to the energy of the light, which makes it easier to relate spectral features to molecular vibrations. A typical mid-infrared spectrum runs from about 4000 to 400 cm⁻¹.
The distinction matters when you are reading a spectrum or comparing results across instruments. If someone reports a peak at “1650,” you need to know whether they mean 1650 nm (near-infrared, close to the telecom window) or 1650 cm⁻¹ (mid-infrared, near the amide I band of proteins). Context usually makes it obvious, but the units themselves carry very different physical meaning.
Bandwidth and Spectral Resolution
Alongside the wavelength axis, spectrophotometers are characterized by their spectral bandwidth, which describes how narrow a slice of wavelengths the instrument isolates at any given setting. A spectrophotometer set to 500 nm with a bandwidth of 2 nm is actually passing light from about 499 to 501 nm. Narrower bandwidth gives finer spectral detail but reduces the amount of light reaching the detector, which can increase noise. Measurements on a clinical enzyme analyzer, for example, found a spectral bandwidth of about 3.5 to 4.0 nm depending on the method used to measure it.4Clinical Chemistry. Measurement of Spectral Bandwidth, as Exemplified with the Beckman “Enzyme Analyzer System TR Spectrophotometer”
Bandwidth is not a “measurement unit” the way absorbance or transmittance is, but it affects every unit the instrument reports. A spectrophotometer with a wide bandwidth will report a lower peak absorbance for a sharp spectral feature than one with a narrow bandwidth, simply because the wide-bandwidth instrument is averaging over a broader range. When comparing absorbance values between labs or between instruments, differences in bandwidth can explain small discrepancies that have nothing to do with the sample itself.
DNA and Protein Quantification
One of the most common uses of spectrophotometers in molecular biology is measuring the concentration of DNA, RNA, or protein in a sample. The instrument still fundamentally measures absorbance, but it reports a derived concentration, typically in nanograms per microliter (ng/µL) or milligrams per milliliter (mg/mL). It does this by applying Beer’s Law internally: DNA absorbs strongly at 260 nm with a known extinction coefficient, so a single absorbance reading can be converted directly to concentration.
Microvolume instruments like the NanoDrop changed how this is done in practice. Traditional spectrophotometers require placing the sample in a cuvette with a fixed path length, usually 10 mm. The NanoDrop holds a tiny droplet (as little as 0.5 to 2 microliters) between two measurement surfaces using surface tension, and the instrument adjusts the path length in real time.5PubMed Central. Microvolume protein concentration determination using the NanoDrop 2000c spectrophotometer This variable path length greatly expands the range of concentrations that can be measured without dilution.
Beyond the raw concentration number, these instruments report purity ratios that are just as important. The 260/280 ratio (absorbance at 260 nm divided by absorbance at 280 nm) indicates whether a DNA sample is contaminated with protein, while the 260/230 ratio flags contamination from salts or organic solvents. A validated method for NanoDrop DNA quantification showed linearity with correlation coefficients above 0.9950 and precision within 2% coefficient of variation, with samples stable for 60 days when stored at 2-4°C.6PubMed Central. Quantification of DNA through the NanoDrop Spectrophotometer: Methodological Validation Using Standard Reference Material and Sprague Dawley Rat and Human DNA The units here are dimensionless ratios, but they carry real diagnostic power about sample quality.
Fluorescence and Relative Fluorescence Units
A fluorescence spectrophotometer (or spectrofluorometer) works differently from a standard absorption instrument. Instead of measuring how much light passes through a sample, it excites the sample with one wavelength and measures the light emitted at a different, longer wavelength. The unit most commonly seen on these instruments is the relative fluorescence unit, or RFU. These values represent the amount of emitted light energy and are proportional to the quantity of fluorescent molecules in the sample.7ScienceDirect. Relative Fluorescence Units – an overview
The word “relative” is doing important work in that name. Unlike absorbance, which is defined by a clear physical ratio (light in versus light out), fluorescence intensity depends on the instrument’s detector sensitivity, optical geometry, lamp intensity, and gain settings. An RFU of 5000 on one instrument does not mean the same thing as an RFU of 5000 on a different model or even the same model with different settings. This is why fluorescence measurements almost always require calibration standards run under identical conditions. You compare your sample’s RFU to a standard curve, not to some universal reference value.
Fluorescence spectrophotometers also report excitation and emission wavelengths (in nm), and many generate three-dimensional spectra that map fluorescence intensity as a function of both excitation and emission wavelength simultaneously. The result is sometimes displayed as a contour plot with RFU on a color scale, which can fingerprint complex mixtures like dissolved organic matter in water or multicomponent drug formulations.
Reflectance and Color Measurement
Not all spectrophotometers work by shining light through a sample. Reflectance spectrophotometers bounce light off a surface and measure how much comes back at each wavelength. The primary unit is percent reflectance (%R), which functions analogously to percent transmittance but for opaque surfaces like paints, textiles, plastics, or tablets.
In industries where color matters, reflectance data is often converted into standardized color coordinates. The most widely used system is CIELab, which characterizes color using three components: L* for lightness (0 is black, 100 is white), a* for the red-green axis, and b* for the yellow-blue axis.8PubMed Central. Developments in image-based colorimetric analysis methods and applications of CIElab color space in pharmaceutical sciences: A narrative review These are not raw spectrophotometer outputs in the way absorbance is; they are computed from the full reflectance spectrum using mathematical transformations that model how the human eye perceives color. But they are the units that end up on pharmaceutical release specifications, paint formulation records, and food quality reports. A spectrophotometer measuring the color of a tablet coating is ultimately reporting in L*, a*, and b* values, even though the underlying measurement is reflectance at many wavelengths.
Raman Shift and Wavenumber Units
Raman spectrophotometers (more often called Raman spectrometers) measure light that has been scattered by a sample at slightly different wavelengths than the incoming laser beam. The difference between the laser wavelength and the scattered wavelength is called the Raman shift, and it is reported in wavenumber units (cm⁻¹). This shift is one of the most important parameters in Raman spectroscopy because it corresponds directly to the vibrational energy of molecular bonds.9Izmeritel`naya Tekhnika. A wide-range Raman shift wavenumber measure based on a polymer material
The intensity axis in Raman spectroscopy is typically reported in arbitrary units or counts, because (like fluorescence) the absolute signal strength depends heavily on laser power, collection optics, detector sensitivity, and integration time. Quantitative Raman work requires internal or external standards. The wavenumber shift axis, however, is highly reproducible and instrument-independent, which is why Raman spectra from different labs can be compared feature by feature. Calibration of the wavenumber axis uses reference materials with well-characterized Raman peaks, and the conversion from raw pixel positions on the detector to wavenumber shift involves the laser wavelength and the spectrometer’s optical geometry.10PubMed Central. Wavenumber Calibration Protocol for Raman Spectrometers Using Physical Modelling and a Fast Search Algorithm
What the Detector Actually Records
Behind every absorbance value or transmittance percentage, there is a more basic quantity: the raw signal from the detector. In older analog instruments, the electronics directly produced absorbance or transmittance values through circuit design. Modern instruments have shifted to digitizing the raw light intensity hitting the detector, then computing absorbance or transmittance in software. Over the last half century, “raw” data has moved from absorbance and transmittance values produced by analog electronics to digitized intensity values, though the latter are rarely made available to the user in favor of digitally transformed derived data.11ScienceDirect. Raw data and noise in spectrophotometry
For array-detector instruments (which capture a full spectrum at once rather than scanning wavelength by wavelength), the raw data is an array of intensity counts at each pixel, with each pixel corresponding to a narrow wavelength range. The counts themselves are in arbitrary digital units that depend on detector gain and integration time. Almost no one works with these raw counts directly. But understanding that they exist helps explain why spectrophotometer specifications list things like “dynamic range” and “signal-to-noise ratio.” These specs describe the quality of the raw signal that underlies every reported absorbance value. An instrument with a wider dynamic range can faithfully measure both very faint and very intense signals, which translates to accurate readings across a broader concentration range.
Why the Same Instrument Can Report Different Units
A common source of confusion is that a single spectrophotometer might display results in absorbance, transmittance, concentration, OD, or even custom units depending on the software mode selected and the application being run. The hardware is doing the same thing in every case: measuring how much light reaches the detector. Everything else is a calculation. Transmittance is the ratio of sample signal to reference signal. Absorbance is the negative logarithm of transmittance. Concentration is absorbance divided by the extinction coefficient and path length. OD in a microbiology context is absorbance measured at a specific wavelength under specific turbidity conditions.
This means that “what units does a spectrophotometer measure in?” depends partly on what you are doing with it. A chemist running a kinetic assay sees absorbance changing over time (AU vs. seconds). A microbiologist tracking bacterial growth sees OD600 values on a growth curve. A molecular biologist checking a DNA extraction sees ng/µL. A color scientist evaluating pill coatings sees L*, a*, b*. A Raman spectroscopist sees intensity versus wavenumber shift. All of these derive from the same fundamental act of measuring light, but the units that matter to the user vary enormously by field.
Stray Light and the Limits of Absorbance
One practical unit-related issue that trips up new users is the upper limit of reliable absorbance readings. Spectrophotometers are not equally accurate across the entire absorbance scale. At very high absorbance values (typically above 3 or 4 AU), so little light reaches the detector that stray light, meaning any light that reaches the detector without having passed properly through the sample, begins to dominate the signal. This puts a ceiling on meaningful absorbance readings.
Testing for stray light involves measuring the apparent absorbance of solutions that completely block light in a particular wavelength range. If the instrument reads an absorbance of 3.0 when it should read infinity, stray light is about 0.1% of the incident beam. Methods for characterizing stray light have used pure liquids and salt solutions that act as sharp cutoff filters in the ultraviolet region.12Optica Publishing Group. Test Methods in Spectrophotometry: Stray-Light Determination For routine work, the practical advice is straightforward: if your sample gives an absorbance reading above about 2.5 to 3.0, dilute it and re-measure. The number on the display may look precise, but at those extremes the underlying physics is working against you.
Stray light also explains why high-end research spectrophotometers cost so much more than basic teaching models. Much of the engineering goes into suppressing stray light through better optics, double monochromators, and detector shielding, all of which extend the useful absorbance range upward by perhaps one or two AU. That might sound like a small gain, but it corresponds to being able to measure samples ten to a hundred times more concentrated without dilution.