Haitz's law is an observation and forecast about the steady improvement, over many years, of light-emitting diodes (LED).
It states that every decade, the cost per lumen (unit of useful light emitted) falls by a factor of 10, and the amount of light generated per LED package increases by a factor of 20, for a given wavelength (color) of light. It is considered the LED counterpart to Moore's law, which states that the number of transistors in a given integrated circuit doubles every 18 to 24 months. Both laws rely on the process optimization of the production of semiconductor devices. There is an important difference between the two laws: There is practically unlimited need for computing power so that the transistor count per integrated circuit will continue to rise until the physical limits are reached. In contrast, certain lumen-packages are well-suited for certain applications (~1 klm for residential lighting, ~10 klm for commercial lighting) and these lumen packages can be addressed already with today's technology. In principle it will be possible to go beyond these lumen packages in the years ahead, but this will be constrained to special applications and will have only limited commercial relevance in contrast to integrated circuit technology where there is always market demand for higher computing power.
Haitz's law is named after Dr. Roland Haitz, a now-retired scientist at Agilent Technologies. It was first presented to the larger public at Strategies in Light 2000, the first of a series of annual conferences organized by Strategies Unlimited. Besides the forecast of exponential development of cost per lumen and amount of light per package the publication also forecast that the efficiency of LED-based lighting could reach 200 lm/W (lumen per Watt) in 2020 crossing 100 lm/W in 2010. This would be the case if enough industrial and government resources were spent for research on LED-lighting. More than 50% of the electricity consumption for lighting (20% of the totally consumed electrical energy) would be saved reaching 200 lm/W. This prospect and other stepping-stone applications of LEDs (e.g. mobile phone flash and LCD-backlighting) led to a massive investment in LED-research so that the LED efficiency did indeed cross 100 lm/W in 2010. If this trend continues, LEDs will become the most efficient light source by 2020.
In 2010, Cree Inc., developed and marketed the XM-L LED that claimed 1000 lumens at 100/W efficiency and 160 lm/W @ 350mA and 150 lm/W @ 700mA. They also claimed to have broken the 200 lm/W barrier in R&D with a prototype producing 208 lm @ 350mA. In May 2011, Cree announced another prototype with 231 lm/W efficiency @ 350mA.
The theoretical max for white light is 251 lm/W. The high efficiency LEDs that come near this number (or beat it) are not for white light, and are not commercially useful.
- Haitz's law. doi:10.1038/nphoton.2006.78. Retrieved 2009-03-04.
- Solid-state lighting: ‘The case’ 10 years after and future prospects. doi:10.1002/pssa.201026349.
- "Cree’s New Lighting-Class LEDs Shatter Industry Performance Standards". 10 November 2010. Retrieved 27 July 2011.
- "Cree Breaks 200 Lumen Per Watt Efficacy Barrier". 3 February 2010. Retrieved 27 July 2011.
- "Cree 231 Lumen Per Watt LED Shatters LED Efficacy Records". 9 May 2011. Retrieved 27 July 2011.
- "Maximum Efficiency of White Light". Retrieved 2011-07-31.
- Roland Haitz, Fred Kish, Jeff Tsao, Jeff Nelson. The Case for a National Research Program on Semiconductor Lighting. Page 5. Date of Access: 4 Feb. 2008.