For LEDs in the visible region, the total radiant flux (watt) can be converted from the luminous flux value and LED’s relative spectral distribution. However, the uncertainty increases, especially at near the wings of the V(l) function. This technique cannot be used for UV and IR LEDs.
Figure 1. High intensity LEDs of wide range of colors.
Various new types of Light Emitting Diodes (LEDs) are being developed in the industry for solid state lighting and other applications, and there are increasing needs for measuring various optical parameters of LEDs with high accuracy. Traditional standard lamps do not suffice the specific needs for LED measurement. NIST has recently developed capabilities for calibrating LEDs for photometric, radiometric, and colorimetric quantities and established NIST calibration services for LEDs. This page outlines NIST measurement facilities for Averaged LED Intensity, total luminous flux, total radiant flux, and color quantities of LEDs.
Application of LEDs is expanding rapidly since high intensity LEDs of wide range of colors including white LEDs have been made available, which enabled application of LEDs in lighting applications as well as signaling applications. As the LED application develops, accurate specifications of LED characteristics are increasingly important. However, large discrepancies of measurements are reported among LED manufacturers and users. LEDs are very different from traditional lamps in terms of physical size, flux level, spectrum, and spatial intensity distribution. Thus, a transfer of photometric scales from traditional luminous intensity standard lamps to LEDs is not a trivial task, and large uncertainties are involved. The temperature-dependent characteristics and a large variety of optical designs of LEDs make it even more difficult to reproduce measurements. To assure high accuracy LED measurements, standard LEDs and calibration services have been in high demand . To address such urgent needs in industry, NIST has recently established calibration services for photometric, colorimetric, and radiometric quantities of LEDs so that the industry can establish standard LEDs in their laboratory to calibrate their instruments or verify the uncertainty of their measurements.
The luminous intensity (unit: candela) of LEDs can measured with the conventional photometric bench and the standard photometers  under a far field condition, at a distance far enough so that the test LED can be regarded as a point source (typically 2 m or longer). However, a common practice in the LED industry has been to measure LEDs at much shorter distances, such as 10 cm to 50 cm. The tradition presumably came from the time when LEDs were very dim and also photometers were not very sensitive. The same practice is still prevailing even when LEDs are much brighter and photometers are sensitive enough. Measuring luminous intensity of LEDs at short distances is a problem because many of LEDs have an epoxy lens, and it does not work as a point source. The effective center of LED emission canshift from the physical center of the LED. This causes variations in measured luminous intensity when measured at different distances, especially when the distance is short. This was determined to be one of major causes of measurement variation for luminous intensity.
To solve this problem, Commission Internationale de l'Éclairage (CIE) standardized the measurement distances (100 mm and 316 mm) for LED intensity measurements as published in CIE 127 (1997) . This publication also standardized that the area of the photometer aperture is to be 1 cm2 circular, that the distance is measured from the tip of the LED encapsulation, and that the direction of measurement is the mechanical axis of the LED. This CIE geometry is shown in the figure below. The luminous intensity measured under these standardized conditions is called Averaged LED Intensity, since the value can be slightly different from the real (far-field) luminous intensity of the LED. This CIE recommendation should be used for intensity specification of individual LEDs. This recommendation does not apply for LED clusters, arrays, and fixtures made with LEDs.
Figure 2. Geometry for CIE averaged LED intensity.
NIST has developed standard photometers in compliance with this CIE recommendation, and has established a calibration service for Averaged LED Luminous Intensity in either Condition A or Condition B. Typical NIST calibration uncertainty for these calibrations is 1 % to 3 % (k=2) depending on test LEDs. See references for the details [4-7].
Total Luminous Flux
The total luminous flux (unit: lumen) of LEDs is probably the most important quantity for LEDs for lighting applications. Compared with the measurement of traditional incandescent lamps, the uncertainty of LED measurements tend to be much larger primarily due to narrowband spectral distribution and narrow beam pattern of LEDs with lenses. NIST has studied these uncertainty components for LED measurements using the NIST 2.5 m integrating sphere (which is used for lamp calibrations), and established calibration procedures for LED total luminous flux measurement. Even with a very large size of the sphere, the NIST sphere has sufficient sensitivity for LED luminous flux measurement. The spectral throughput of the NIST sphere has been precisely determined and spectral mismatch correction is applied with reduced uncertainties. The error due to spatial nonuniformity of the sphere responsivity, associated with difference in LED angular intensity distributions, has also been analyzed and uncertainty component determined. The 2.5 m sphere system uses a detector-based calibration method, where the total flux of a test LED is measured against known amount of luminous flux introduced from an external source. The details of the NIST 2.5 m sphere  and the LED calibration procedures for luminous flux are found in references [4,5,6,9].
Figures 3 and 4. NIST 2.5 m integrating sphere configured for LED measurements.
The luminous flux part of CIE 127 (1997) is inappropriate and is being revised in the CIE Technical Committee 2-45. A particular issue is the treatment of backward emission of LEDs. The sphere geometries shown in CIE 127 (1997) ignore backward and sideward emission. When total luminous flux is measured, radiation emitted in all directions must be measured. NIST recommends the following sphere geometry for total luminous flux measurement, which are considered for inclusion in the revision of CIE 127. For those applications where only forward emission is of interest, a new quantity, partial LED flux, is considered in the draft.
Figure 5. Sphere geometries for total flux measurement recommended by NIST.
Total Radiant Flux
Total radiant flux (unit: watt) is spectrally and spatially integrated total radiant flux of a source. This quantity is used to specify LEDs in the UV or IR region. Particularly, LEDs in the near UV region is being actively developed in the industry to produce phosphor-type white LEDs, and accurate measurement of total radiant flux is increasingly important. Even for visible region, total luminous flux (lumen) is not very useful for deep blue and deep red LEDs, and total radiant flux is often used.
Figure 6. Configuration of the NIST 2.5 m sphere for UV LED radiant flux calibration.
Total Spectral Radiant Flux
Integrating spheres equipped with a spectroradiometer is increasingly used in the industry to measure not only color quantities of LEDs but also for total luminous flux and total radiant flux of LEDs. Such a spectroradiometer/sphere system has an advantage that photometric quantities can be measured theoretically with no spectral mismatch error, thus is a convenient way of accurate measurement of LEDs in the industry Such a sphere system needs to be calibrated against a total spectral radiant flux standard. NIST has recently established the total spectral radiant flux scale for the 360 nm to 830 nm region, and started a calibration service for several different types of incandescent standard lamps including small halogen lamps suitable for smaller integrating spheres used for LED measurement. See the Link for this service.
Figure 7. Arrangement of a sphere for total spectral radiant flux measurement
Color quantities such as chromaticity coordinates, dominant wavelength, correlated color temperature, are used to specify LEDs used for many applications. The uncertainty in color measurement of LEDs is often unexpectedly large or unknown even when using a spectroradiometer, and calibrated reference LEDs are needed in the industry to verify the accuracy of measurements. NIST has established a reference spectroradiometer, tailored for LED color measurement, using a double-grating monochromator with irradiance input optics. The spectroradiometer is tuned to have a constant triangular bandpass of 5 nm width (FWHM) throughout the visible region and scanned typically at 2.5 nm interval (optionally 1 nm interval), and has uncertainty of LEDs of any color to be within 0.001 in (u’', v'’). The measured spectra of LEDs are corrected for bandpass. Further details on the reference spectroradiometer is found in ref  and also in the section of Spectral Color Measurement.
Figures 8 and 9. NIST reference spectroradiometer for LED color measurement.
Figure 10. Comparison of a red LED measurement with a diode-array instrument (red line) and the NIST reference spectroradiometer (blue line), showing significant stray light of the diode-array instrument.
Strategy on standard LEDs in NIST calibration services
Some NIST calibration services issue calibrated artifacts, and others calibrate artifacts submitted by customers. We decided not to prepare and issue "standard LEDs" because there are so many types of LEDs and new types of LEDs are coming out constantly, and thus, any standard LED we may develop would not satisfy many customers and also these will be obsolete quickly. We are rather committed to provide calibrations for any type of LEDs submitted by the customers, which can be used as standard LEDs of the type needed in the customer's lab. Customers are responsible to ensure the quality of the LEDs to submit to NIST.
Figure 11. Bandpass function of the NIST spectroradiometer (5 nm and 1 nm).
NIST Calibration Services on LEDs:
For LEDs in the visible region, the total radiant flux (watt) can be converted from the luminous flux value and LED’s relative spectral distribution. However, the uncertainty increases, especially at near the wings of the V(l) function. This technique cannot be used for UV and IR LEDs.NIST recently established a new instrumentation for LED total radiant flux measurement for the 360 nm to 500 nm region using the 2.5 m integrating sphere. The figure below shows the configuration of the NIST 2.5 m sphere, in which a spectroradiometer (corrected for stray light) is used as the detector and the total spectral radiant flux of the test LED is measured against the flux introduced from the external source. For the details see reference .