UV Radiometer

What is a UV Radiometer

A UV radiometer measures ultraviolet optical power arriving at a surface. At its core sits a photodiode, a semiconductor element that converts incident photons directly into a small electrical current. That current is amplified and digitized by an analog-to-digital converter, then scaled by a calibration factor into engineering units of irradiance, typically milliwatts per square centimeter (mW/cm2), microwatts per square centimeter (uW/cm2), or watts per square meter (W/m2). When the instrument integrates that irradiance over time, it reports a dose, also called energy density or radiant exposure, in millijoules or joules per square centimeter (mJ/cm2 or J/cm2).

The distinction between irradiance and dose is the single most important concept in UV measurement. Irradiance is an instantaneous quantity: how brightly the source illuminates the sensor right now. Dose is the time integral: how much total energy has accumulated. A curing process cares about both, because it needs enough peak irradiance to initiate the photochemistry and enough total dose to drive it to completion. A disinfection process cares mostly about dose, because microbial inactivation depends on accumulated germicidal energy. An occupational safety assessment also cares about dose, because exposure limits are stated as energy per workday.

A complete UV radiometer has three functional layers. First, an input optic: a diffuser or aperture that defines the collection geometry and, ideally, gives the sensor a cosine angular response so it reads true irradiance on a flat plane. Second, a spectral selection layer: optical bandpass and blocking filters, or for spectroradiometers a dispersing grating, that determine which wavelengths the instrument actually weights. Third, the detector and electronics that convert, amplify, digitize, and display. The quality of any UV reading is set by how faithfully these three layers reproduce a defined spectral and angular response, not by the resolution of the display.

Ultraviolet metrology is harder than visible photometry for physical reasons. UV photons carry more energy and degrade the very materials used to measure them: filters solarize, diffusers yellow, and adhesives darken, all faster at shorter wavelengths. Air, quartz, and detector windows absorb UV differently across the band. Many UV sources, such as medium-pressure mercury arcs and pulsed xenon lamps, emit complex line-plus-continuum spectra that shift as lamps age. As a result, a UV radiometer reading is meaningful only when the instrument calibration spectrum matches the source being measured, a constraint that does not exist in the same way for broadband white-light photometry.

The application scale spans many orders of magnitude. Solar erythemal UV at ground level produces an irradiance of roughly a few milliwatts per square centimeter at peak; UV curing lamps at the focal line deliver several watts per square centimeter, a thousand times higher; while UVC germicidal validation looks at doses from a few mJ/cm2 to hundreds of mJ/cm2. Human-exposure safety limits, by contrast, sit near 3 mJ/cm2 per eight-hour workday for broadband 254 nm UV. No single sensor head spans this full range, which is why UV radiometer systems are built around interchangeable, application-matched detectors.

UV Bands and Instrument Types

Ultraviolet radiation is divided into bands by wavelength, each with distinct physical effects and measurement needs. The CIE and common engineering convention splits the actinic UV range into UVA, UVB, and UVC, with a longer UVV region defined by curing-equipment makers and a far-UVC sub-band that has become important for occupied-space disinfection. The table below summarizes the bands a radiometer is built to target.

Note that band edges vary slightly between standards bodies, and curing-equipment makers define their own band windows. For example, EIT 2.0 curing radiometers report UVA as roughly 320 to 390 nm, UVB as 280 to 320 nm, UVC as 250 to 260 nm, and UVV as 395 to 445 nm, which differs from the CIE 315 nm UVA/UVB boundary. When comparing readings across instruments, always confirm the exact band window each meter integrates, because a number labeled simply UVA can correspond to different wavelength limits.

Instruments split into three architectures by how they select wavelength. Broadband radiometers, spectroradiometers, and integrating dose meters each suit different work. The table below compares them on the dimensions that drive selection.

Broadband radiometers place an optical filter in front of a photodiode so the combined response approximates a target band. They report one summed irradiance value and are fast, rugged, and inexpensive, which makes them the workhorse for production checks and field measurements. Their weakness is spectral mismatch: a broadband head is calibrated against one source spectrum, and reading a different spectrum (for example, a 365 nm LED with a mercury-calibrated head) introduces error because the filter weights wavelengths differently from how the source emits them.

Profiling dose meters, such as puck-style curing radiometers, are broadband instruments that travel through the process on the conveyor or web, sampling irradiance at a high rate and integrating it into a dose while also capturing the peak irradiance. They are optimized for repeatability rather than absolute accuracy: the value of a curing radiometer is that today reads the same as last week if the lamp is healthy, so a drop in reading flags a degrading lamp or reflector.

Spectroradiometers disperse the incoming light with a grating or prism across a CCD or CMOS detector array, reporting irradiance at each wavelength, often at 1 nm resolution. Because they capture the full spectrum, they can compute any weighted quantity (erythemal, germicidal, actinic hazard) for any source, and they are the only correct instrument when the source spectrum is unknown or shifting. They cost several times more than broadband meters and are slower, so they are used for source characterization, calibration transfer, and research rather than fast line QA.

Detector Technologies

The detector material sets the instrument's intrinsic spectral window, out-of-band rejection, and long-term stability. Four detector approaches dominate UV radiometry: filtered silicon, silicon carbide, gallium nitride and AlGaN, and thermal detectors. The choice matters most in the UVC band, where stray visible and near-IR light can swamp a weak UV signal unless the detector is intrinsically blind to it.

Filtered silicon photodiodes respond across roughly 200 to 1100 nm, well into the visible and near-infrared. Silicon cannot be made intrinsically visible-blind, so a silicon UVC head depends entirely on bandpass and blocking filters to reject out-of-band light. This is workable and economical, especially for UVA and broadband curing where the target band overlaps silicon's strong response, but the filter stack solarizes and degrades over time, shifting calibration. It is the most common and lowest-cost detector in handheld curing and UVA meters.

Silicon carbide (SiC) is a wide-bandgap semiconductor that responds only from about 210 to 380 nm and is naturally insensitive to visible and infrared light. This intrinsic visible blindness gives SiC heads a high solar-blind ratio and good long-term stability, which is why they are favored for UVB and UVC measurement where rejecting daylight and lamp continuum is essential. SiC detectors cost more than silicon but reduce reliance on fragile filters.

GaN and AlGaN detectors offer a tunable cutoff: by adjusting the aluminum fraction in AlxGa1-xN alloy from 0 to 1, the bandgap moves from 3.4 eV to 6.2 eV, shifting the long-wavelength cutoff from about 365 nm down to about 200 nm. As a direct-bandgap material, AlGaN has a high absorption coefficient and a sharper spectral cutoff than SiC, and modern doping control allows engineered heterojunction devices. This makes AlGaN attractive for demanding UVC and far-UVC (222 nm) instruments that need a clean, solar-blind response, at higher cost.

Thermal detectors such as thermopiles absorb optical power and measure the resulting temperature rise, giving a flat, wavelength-independent response across a very wide band. Because they respond to all wavelengths equally, they need a defining filter or aperture to isolate UV, but their flat response makes them valuable as absolute power references and for measuring total radiant power where spectral weighting is not required. They are slower and less sensitive than photodiodes, so they appear in laboratory reference setups rather than handheld field meters.

One detail that separates a credible instrument from a toy is the diffuser and input optic. To read true irradiance on a plane, the sensor must collect off-axis light in proportion to the cosine of the incidence angle, per Lambert cosine law. Real diffusers deviate from the ideal, and that deviation is quantified as a cosine error (often the f2 metric in percent). A meter intended for diffuse fields, solar UV, or multi-lamp chambers must specify a cosine-corrected diffuser with low f2; a meter for a single collimated lamp held normal to the sensor is far less sensitive to this error.

Calibration, Traceability and Standards

A UV radiometer reading is only as trustworthy as its calibration chain. Absolute UV calibration is traceable to a national metrology institute such as NIST (United States), PTB (Germany), or an equivalent body, usually through a standard 1000 W FEL quartz-halogen lamp that transfers a spectral irradiance scale, with documentation under ISO/IEC 17025 accreditation. At the top of the chain, NIST maintains primary scales such as the Primary Optical Watt Radiometer and spectral irradiance facilities; commercial calibration laboratories transfer that scale down to working standards and finally to your instrument.

Two corrections sit inside any honest calibration. The first is the absolute responsivity factor that converts detector current into irradiance. The second is the spectral correction: because a broadband detector weights wavelengths according to its own filter response rather than the target action spectrum, a function converts the detector-weighted value into the desired weighted quantity, for example the CIE erythemal value. A cosine-correction function handles angular response. A reading that omits any of these is not metrologically complete.

UV detectors drift, and they drift faster at shorter wavelengths because filter solarization and diffuser degradation are worse in the UVC. For this reason, manufacturers commonly recommend annual recalibration of working instruments, while NIST-traceable reference instruments are often recalibrated near six-month intervals. Production radiometers exposed to intense curing lamps should be verified more often, ideally against a reference unit kept off the line. The table below maps the principal standards and references a buyer encounters when specifying a UV radiometer.

For solar and environmental work, the UV index deserves a note because it is a weighted derived quantity, not a raw reading. The measured UV spectrum is multiplied by the CIE McKinlay-Diffey erythemal action spectrum, integrated, and scaled: the daily UVI is the maximum 30-minute moving average of the erythemal irradiance multiplied by 40 m2/W and rounded to the nearest integer. A radiometer that reports a UV index must therefore embed the erythemal weighting in its detector response or compute it from a measured spectrum, which is why broadband UVI meters are calibrated specifically against the erythemal curve rather than a flat UVA or UVB window.

Key Specification Parameters

Reading a UV radiometer spec sheet is a core skill for buyers, because two meters with the same headline accuracy can behave very differently in the field. Eight parameters drive selection: target band and spectral response, measurement range, irradiance versus dose capability, accuracy and source matching, cosine response, dynamic range and sample rate, calibration traceability, and operating environment. Each is explained below.

Target band and spectral response is the first filter on any candidate. Confirm the exact wavelength window the meter integrates (for example UVA 320 to 390 nm versus 315 to 400 nm) and, critically, the source spectrum it is calibrated against. A head calibrated to a medium-pressure mercury lamp is not interchangeable with one calibrated to a 365 nm or 395 nm LED. For 222 nm far-UVC work, the head must be specifically designed and calibrated for that excimer line.

Measurement range must bracket your real values with margin. Curing radiometers are commonly offered in full-scale ranges such as 100 mW/cm2, 1 W/cm2, or 10 W/cm2, chosen to match lamp intensity. Germicidal validation instruments must resolve from single mJ/cm2 doses up to hundreds of mJ/cm2, while a meter that also checks worker exposure must reach down near the few-mJ/cm2 daily limits. A range that is too coarse loses resolution; one that is too narrow saturates and reads a false ceiling.

Irradiance versus dose capability decides whether the instrument simply reads instantaneous intensity or integrates over time. Profiling curing radiometers report both peak irradiance (W/cm2) and total energy density (J/cm2) as they pass under a lamp. Germicidal meters emphasize dose (mJ/cm2). Always read the spec carefully to confirm which quantity a number represents, because peak irradiance and integrated dose are physically different and are not convertible without the exposure time.

Accuracy and source matching are coupled in UV. A broadband meter may carry a 5 percent calibration uncertainty against its reference spectrum, but real-world error is dominated by spectral mismatch when the measured source differs from the calibration source. Spectroradiometers state absolute accuracy (often 5 to 10 percent) plus wavelength accuracy (often better than 1 nm). For curing QA, repeatability is frequently more valuable than absolute accuracy, because the goal is to detect lamp degradation by comparing today against a baseline.

Cosine response matters whenever light arrives off-axis. A cosine-corrected diffuser with low f2 error is required for diffuse fields, solar UV, scattered germicidal light, and multi-lamp chambers; for a single collimated lamp held normal to the sensor it is less critical. Dynamic range and sample rate define whether the instrument can follow fast conveyors and pulsed sources: research radiometers such as the ILT5000 specify a wide current range (on the order of 100 fA to 1 mA) and sampling up to about 100 Hz, while puck-style profilers offer user-selectable sample rates for static or fast-moving lines.

The remaining parameters are practical but decisive over a service life. Calibration traceability should be NIST or equivalent under ISO/IEC 17025, with a stated recalibration interval. Operating environment covers temperature rating (curing zones run hot, so a puck must survive repeated thermal cycling), ingress protection for washdown or outdoor use, battery life and data logging, and connectivity to a PC for profile analysis. A meter that cannot survive the heat near a curing lamp, or cannot log a conveyor pass, is the wrong tool regardless of its accuracy spec.

Selection Decision Factors

To turn the preceding chapters into a specific instrument, follow the decision sequence below. Most selection mistakes come not from one wrong step but from deciding the model before defining the spectrum and quantity it must measure. These eight steps form a reusable RFQ template.

1. Define band and source spectrum: Identify whether you measure UVA, UVB, UVC, far-UVC, or UVV, and the exact source (medium-pressure mercury, 365/395 nm LED, 254 nm low-pressure, 222 nm excimer). The instrument calibration spectrum must match this source, or readings will be biased by spectral mismatch.
2. Choose irradiance, dose, or both: Curing typically needs peak irradiance plus integrated dose; disinfection validation needs dose; safety assessment needs dose against an exposure limit. This decides between an instantaneous radiometer and an integrating profiler.
3. Set the measurement range: Pick a full-scale range that places normal readings comfortably below saturation with resolution to spare. Curing ranges (100 mW/cm2 to 10 W/cm2) differ greatly from germicidal dose ranges (single to hundreds of mJ/cm2).
4. Pick broadband versus spectroradiometer: Use a broadband radiometer for a known, stable spectrum and fast QA; use a spectroradiometer when the spectrum is unknown, shifting, or must yield multiple weighted quantities (erythemal, germicidal, actinic).
5. Specify detector and cosine response: For UVC and far-UVC, prefer intrinsically solar-blind SiC or AlGaN over filtered silicon. For diffuse or multi-lamp fields, require a cosine-corrected diffuser with low f2; for a single collimated lamp this is less critical.
6. Confirm calibration and standards: Require NIST-traceable, ISO/IEC 17025-accredited calibration with a stated interval, and check applicable references (IEC 62471 and 62471-6 for safety, IES guidance for germicidal dosing, CIE erythemal weighting for UVI).
7. Match the operating environment: Verify temperature rating for hot curing zones, ingress protection for washdown or outdoor use, sample rate for conveyor speed, and data logging or PC connectivity for profile analysis.
8. Total cost of ownership: Add the recurring annual recalibration cost, spare detector heads, and the cost of process risk from drift. A meter that drifts undetected can pass bad cures or under-dosed disinfection, a cost that dwarfs the instrument price.

One dimension buyers routinely overlook is serviceability and ecosystem: interchangeable application-matched heads, in-region calibration turnaround, availability of lamp-specific and LED-specific calibration, and software for profile analysis. For UV curing of coatings, inks, and adhesives, EIT 2.0 (UV Power Puck II and UviCure Plus II) profiles up to four bands on a conveyor, and Henkel LOCTITE PM20 serves adhesive-cure checks. For laboratory and field metrology, Gigahertz-Optik (X1 optometer with interchangeable UV heads and the BTS256-UV spectroradiometer) and International Light Technologies (ILT2400 handheld and ILT5000 research radiometer) provide NIST-traceable calibration and dedicated 254 nm and 222 nm heads. Opsytec Dr. Groebel and Miltec UV supply process radiometers, while Linshang and similar vendors cover lower-cost handheld UVA and UVC checks. Whichever brand you choose, match the calibration spectrum to your lamp before comparing readings across instruments.