The colour rendering index (CRI) is a measure that has been widely adopted and used by the lighting industry to characterise the quality of a light source. CRI is a specification to assist designers in making comparison between different lamp sources and is a relative comparison between a lamp source and a reference source. It is related to the idea that if a red object is illuminated with red light it appears red, but if it is illuminated with blue light it appears grey or black – the colour perceived will change with the colour of the source used. The CRI is used to measure these subtle shifts in colour of a source. In order to quantify this, the CRI uses 14 (or more) standard coloured tiles and measures the shift in colour when the illumination changes from a reference to the test source. The reference source used is the sun (or a black body), and by definition it has a CRI of 100. Everything else is measured from that point downwards. Sunlight is composed of a continuous spectrum of colours from the darkest reds through to the deepest violets with particular weighting at each wavelength, known as its spectral power distribution (SPD). If white light generated from an LED does not contain the same wavelength components then its SPD is different and it will have poor colour rendering properties.

Variation of CRI with different types of white LED

Deficiencies in the use of CRI for LEDs
The CRI was originally developed nearly 40 years ago, and even after a few revisions it still uses outdated formulae and methods, and it is known to have deficiencies:

• The problem is prominent with narrowband sources, and the use of CRI for RGB white LEDs or multichip white LEDs can mislead design directions.
• Studies carried out with white LED light sources have shown that CRI can have no correlation to people’s colour preferences.
CRI is just a figure of merit to show how close a source matches a standard.
• LED spectral power distribution is sensitive to temperature effects.

Effect of LED source characteristics on the CRI
Currently, white LED light can be generatedusing one of two basic methods:
• Phosphor conversion: where a phosphor is excited with light from a blue or near-UVlight LED.
• LED colour mixing: where light from three or more coloured LEDs is mixedtogether.
Each method has its merits and disadvantages, which are briefly discussed below.

Phosphor conversion
In the phosphor-conversion method, predominantly YAG:Ce phosphors are used to make white LEDs. The absorbed blue (or near- UV) light is re-emitted in the yellow to red region of the visible spectrum. If two phosphors are used, the CRI can be low and manufacturers can be tempted to use more phosphors to improve the CRI. However, this comes at a price; for example, the amount of absorbed power (measured in lumens) from the absorbed (blue) light would need to be increased. A high (90+) CRI necessitates the conversion of more blue lumens into red (phosphor) lumens. The result is a lower lumens output device.

Manufacturers are developing products along two paths for warm-white LEDs:
• High CRI (90+) at 20+ lumens (@350 mA) for colour-rendering critical applications, like halogen replacements.
• Average CRI (70+) at 30+ lumens (@350 mA) for luminance and less critical applications (eg landscape, cove, wall-wash).

Also, phosphor conversion suffers from a loss of energy (known as the Stokes shift) when the phosphors convert light from a shorter to a longer wavelength, which reduces the overall device efficiency.

One important source characteristic for blue LED phosphor conversion is that the CRI will depend on the colour tolerance of the blue ‘pump’ LED. Whereas, for near-UV phosphor conversion, the colour is predominantly determined by the stable phosphor colours, and a more consistent CRI will be obtained.

LED colour mixing
The LED colour-mixing method provides the greatest flexibility in terms of colour definition. An array of multicoloured LEDs can produce many shades of coloured light and white light by changing the output ratios of each coloured LED.

In principle this should be more efficient than the phosphor-conversion method since energy is not lost to the Stokes shift. In practice, however, additional optics are needed to mix the light, and these introduce extra losses. As a result, phosphor-converted white LEDs provide more lumens per watt than mixed colour systems.

Of course, for applications requiring a single source, phosphor-converted is the only choice too.

Source: DTI Global watch mission report
Author: Mike Bean, Optical Design Engineer at Carclo Technical Plastics Plc