WP's Foto Gallery


Light Heating

Heat with light bulbs

Measuring device, instrumental set-up

Entirely touch-free temperature measurement with an ellipsometer. Reference dimension is the change of polarization of a laser beam caused by change in temperature. Twice it runs through a 1 mm thin quartz pane, which can be heated by halogen lamps. Its refraction angle of the inner total reflection changes with temperature. The 180° turn of the beam after the first run through the glass pane and the entry to the detector after the second run on the way back are done by prisms.
A calibration curve with polarization change and temperature over time is the goal.



Measuring temperature with an ellipsometer

The power of light - almost like in a tanning salon


Measurement set-up in working condition

Extraordinary effort! - Why? Thermocouples need a good contact to the measuring object. Bad then, if the surface pressure is limited. Together with a low heat conduction capability of the measuring object or its surface (e.g. due to coating) the thermocouple moreover acts like a cooling fin or under fast temperature changes like a thermal load. The deviation can easily reach values beyond ±10 °C (± 18 °F) or in the two-digit per cent range.
And pyrometers? They measure the heat radiation of an object and calculate the temperature out of the Stefan-Boltzmann Law, refer below, and the area of the measurement spot. This radiation power, however, strongly depends of the object and is incorporated by the correcting factor emissivity 0 < ε ≤ 1. Shiny metal surfaces show an ε much lower than 0.1, the black body (approximately e.g. carbon black) a value of 1. Usually more than 0.7 is set. Besides one should not underestimate the distortion caused by external irradiation (e.g. reflections and other sources in the field of view analog to a video camera) and standing waves in transparent layers.
Therefore the results of hectically fidgeting with pyrometers in those countless pseudo-scientific »science magazines« are to treat with caution. (Among other »info« provided there. ;-)


The light bulb as heating element

Existing light bulb - the better stove: Seven 500 W halogen bulbs easily achieve 1 200 °C (2 192 °F) within seconds. In industry and semiconductor processing this setup is often used for RTP (rapid thermal processing).

The main connection between irradiation and heat is given by Planck's law. In wave-length form (wave-length λ in m) it provides for the temperature depending spectral irradiation energy density of a black body (the object with the highest possible emissivity)

r(λ, T) dλ = 2 π h c25 (exp (h c / (k T) λ-1) -1)-1

Planck constant h = 6.6262 10-34 Js
speed of light c = 2.99792 108 m
Boltzmann constant k = 1.3807 10-23 J/K
absolute temperature T in Kelvin

Planck's Law of Radiation
Efficiency of sun and light bulbs in comparison

Displayed in the figure is the irradiation of a black body after Planck with the temperature of the sun surface (about 6 100 °C, 11 012 °F), that of a filament of a standard light bulb (about 2 800±200 °C, 1 472±392 °F) and that of a halogen bulb (about 3 500±300 °C, 6 332±572 °F).
The visible light region is highlighted in yellow.


Integrated over all wave-lengths one gets with the Stefan-Boltzmann law the temperature depending irradiance

R(T) = ∫ r(λ, T) dλ = σ T4

Stefan-Boltzmann constant σ = 5.670 10-8 W/(m2 K4)
absolute temperature T in Kelvin
(Putting the area A [m2] of the black body as multiplier into the formula above yields the radiation power.)

When (numerically) integrating over the visible light wave-length region 380 - 780 nm of the light bulb (temperature of the filament 2 500 K) one calculates a efficiency of about 6%. With the halogen bulb (temperature of the filament 3 200 K) one receives an efficiency of about 16%. Taking in account the low sensitivity of the human eye at the etches of the visible region, here especially the infrared, the »literature values« of about 3% for a 100 Watt bulb and maximal 8 to 10% for halogen bulbs, resp., seem to be comprehensive.


Fluorescent tubes, energy saving tubes and LEDs

Emitting of these light media is not done by a heated source, but by electron pumping, refer to Franck-Hertz experiment. In the in principle comparable fluorescent tubes and energy saving tubes free electrons accelerated by the mains voltage hit electrons circling around their nucleus and with sufficient energy kicking them on a higher orbital. After the relaxation time of 10-8 seconds these electrons return to their former orbital by emitting an ultraviolet photon with energy hf. The opaque special coating of the glass tube converts these UV irradiation to visible white light. These lamps show a more or less continuous spectrum, thus covering all wave-lengths.

The semiconductor material of LEDs (light emitting diodes, luminescence diodes) is directly electrically pumped. The emitted photons with the energy hf are closely arranged around a frequency or a wave-length, resp., about some tenth nanometers, thus they are monochromatic.

Spectrum of a red LED
Measured spectrum of a red super-bright LED

The spectrum of the 640 nm (sic) GaAlAs super-bright Kingbright L-53SRC-E LED measured with a (grating) monochromator in arbitrary units shows good accordance with the data sheet values - i.a. the maximum is found at about 660 nm. (Maybe due to the not calibrated measurement setup ;-)
The completely depicted section lies in the red. The often circulated »tail« into the UV section were neither noticed here nor with other monochrome LEDs. That is the efficincy improvement compared to bulbs (incl. halogen) with their wide spread spectrum shown above.
By the way: In this plot the spectrum of a laser LED would degenerate to a single small (but very high ;-) impulse line instead of the bell-shaped curve.

For white light one has to bundle a red, a green and a blue LED, either in separate housings or integrated in one and the same. (Often there are two blue LEDs due to their low efficiency.) A discontinuous spectrum is generated by this.
There is a similar possibility compared to fluorescent tubes. Often a efficient blue LED illuminates a special yellowish coating, which then together emit white light with a halfway continuous spectrum.

For conspiracy addicts - yes, industry and state earn much more with energy saving tubes and LED lamps. And the political pressure considering that only less than 2% of the energy consumption of a private household is caused by illumination is rather high. But there is no doubt that these show a significant higher efficiency of more than 20% compared to standard light bulbs, even if the advantages are beautified sometimes.(And in opposite to some energy saving tubes light bulbs make no noise. ;-)

So one just compares the occasionally narrow light beam of a LED (e.g. with just ±6° field angle only), outside it is dark, with the almost complete 360° illumination of a light bulb. Thus one calculates as if the LED would illuminate 360° with their full intensity, too.
Additionally one conceals that LEDs as substitute for household bulbs need a transformer with rectifier, because they do not work with mains alternating voltage. (LEDs work in flow direction only at about 2 to 5 V depending on type.) It almost completely get lost that the installation of high energy LEDs can turn problematic. While a light bulb gets rid of its power dissipation by irradiation in infrared in large extend, LEDs just cool down by convection and heat conduction. A heat sink or an fan might get necessary.
Concerning energy saving tubes one often disregards the filling gas (among others mercury), the plastic parts and the hard to rot printed circuit board with electronics. Fluorescent tubes and energy saving lamps and LED lamps as well are hazardous waste, light bulbs can be dropped into the household waste.

Bottom line: To declare LEDs as general 1:1 substitute for all light bulbs is as if one carries on stacking bridges with steel ingots instead of stones after steel construction was discovered. Energy saving tubes are more suitable here, or just halogen bulbs then. As spot or as large-scale (programmable?) lights LEDs just make more sense. And they will go their way. Not to mention OLEDs.