The Diode Laser

danger.gif (10071 bytes)

My final lab project involved observing the effects changing the temperature and the driving current of a diode laser had on its output.  To begin, I will describe how a diode laser operates, starting with... what is a semi-conductor?

Semiconductors

When atoms combine to form a solid, the previously distinct energy levels of free atoms overlap and thus broaden due to neighboring interactions.  These broadened energy levels are called energy bands.  The outermost electron's energy states broaden and combine to form the valence band.  Excited energy bands are the result of excited states combining and broadening.  At absolute zero, all these states are empty.  The collection of the lowest, unoccupied states forms the conduction band.

semi-cond.gif (1752 bytes)

The energy spacing, Es, between the top of the highest filled energy band (the valence band) and the bottom of the lowest non-full band is referred to as the band gap.  Characteristic of semi-conductors, Es ~ kT (actually a few times larger).

When an electric field is applied across a semi-conductor it conducts.   Electrons in the conduction band are accelerated in the direction opposing the direction of the applied field.  When these electrons fall from the conduction to the valence band, a photon of wavelength is emitted.  When the electron jumps, it leaves behind a 'hole' which acts like a positive charge in the presence of an electric field.

Except at absolute zero, there are holes in the valence band and electrons in the conduction band.  The numbers in these bands is determined by Fermi-Dirac statistics because no two electrons can occupy the same energy state.  The probability of finding and electron at the energy En is given by: p(En) = (e (En-u) / kT + 1)-1.

If a semi-conducting material is 'doped,' it has been coated with some impurity.  An n-type semi-conductor has been doped with an impurity which has an excess of electrons and wants to release them.  A p-type is doped with atoms that lack, and want to gain, an electron - thus has an excess of holes.

As was previously stated, it is the recombination of electrons and holes that results in the emission of radiation.  In diode lasers, the junction is a p-n junction.  If we attain a density of electrons in the conduction band and holes in the valence band, by forward biasing such a junction, the electrons will want to recombine with the holes.  Thus, stimulated and spontaneous emission can occur.  This leads to laser action if the gain in the system is greater than the loss.

diode-dia.gif (2640 bytes)

The laser's output is controlled/determined by the band gap (emitted wavelength = ch / Es).  No mirrors are needed for gain increase.  Two sides of the diode are polished.  Because the index of refraction within the diode is so much greater than that of air, there is enough reflection at the interface to allow resonance.

What I did with the Diode Laser

The following is a picture of the setup for my experimenting:

di-setup'.jpg (44097 bytes)

I experimented to determine whether or not the diode laser would single-mode by increasing the current powering the diode and the temperature of the diode.  In order to learn how the spectrometer, its motor, and the photo-multiplier tube work as a system, I used a sodium light source.  The slits which allow light into the spectrometer had to be adjusted (as with the laser when I used it) so that the PMT would not experience a current greater than 0.1 mA.  The guts of the spectrometer are simplified in the diagram below. 

spect-dia.gif (5187 bytes) Light incident on M1 through S1 reflects off the adjustable diffraction grating.  After reflecting off M2, this light is output through S2.  The diffraction grating pivots about an axis at a rate and direction determined by the motor.   If the grating is rotated sufficiently, the output light will show the spectral lines/modes of the light source.  In my case, I used a photo-multiplier tube to analyze the output light.  For the Sodium light, the following spectrum was the output:

wpe69.jpg (11671 bytes) The characteristic peaks (5890 and 5896 ) in the spectrum are apparent and at the correct wavelengths.  Notice that the peaks are negative.  This is due a negative potential difference applied across the PMT.

So, now I was ready to use this device with a light source that cannot be detected by the human eye - the diode laser.  It's output is in the infrared.  This in conjunction with the fact that I was not sure of the laser's emission spectrum encouraged me to test the spectrometer for accuracy.

The procedure for my data collection involved selecting a temperature on the Thermoelectric Temperature Controller to remain constant for each of the data collection series.  I would then select a particular current to drive the laser and, while holding this constant, allow the grating to sweep across an estimated spectral range of the laser.  I would then vary the current (usually by increments of 20 mA), take another run, etc., until the maximum operating current was reached.

I followed this procedure for three different temperatures in hopes of finding modal characteristic being dependent on temperature.   The following were the three temperatures I used: 20.3 Co (a resistance setting of 12.0k Ohms on the temperature controller), 25 Co (9.777k Ohms) and 35 Co (6.392k Ohms).  With each run, the voltage was read from the PMT through a Pasco interface and displayed for manipulation on Mathematica.

For all the following, the current was at maximum operating level: 120 mA:

When the temperature was 20.3 Co, the following was seen - the first peak occurs at a wavelength of 7780.31 ; the second occurs at 7825.47 and with a slightly greater voltage magnitude:

 wpe85.jpg (24931 bytes)

When looked at more closely, the peaks look like:

wpe8D.jpg (18211 bytes)

When the temperature was 25 Co, the following was seen - the peak occurs at a wavelength of 7824.46 :

wpe87.jpg (25285 bytes)

with a close-up of the peak looking like:

wpe88.jpg (18753 bytes)

When the temperature was 35 Co, the following was seen - the peak of the greatest magnitude occurs at a wavelength of 7837.26 .

wpe89.jpg (24662 bytes)

It can be seen in this close-up of the peak 'region' that the maximum peak is hardly distinguishable from the surrounding peaks.  The other peaks occur at 7831.28 , 7834.27  and 7840.42 , respectively

wpe8C.jpg (11058 bytes)

In Conclusion...

I could find no apparent pattern between runs with the laser at constant temperatures.  The data yielded while the laser was driven at less than the maximum operating current was not interesting and showed no dominant modes as does the previously shown data.  Perhaps this is a fault of the laser (it's been used by underclassman!) as it was predicted that at high driving currents, the laser should approach single-mode operation.  Furthermore, there was no pattern in the shift of the dominant mode between different temperatures.  Again, perhaps this is due to a less-than-perfect laser.  However, this experiment was handy in that it may help the physics department here.  My data show that the laser approaches a single-mode operation at a wavelength between ~7824 and 7840.42 .  This allows me to assume the laser that I was using was a Thor Labs LT027MD Laser Diode.  In their specifications sheet, they claim an approximate radiation emission of about 780 nm with a maximum of about 790 nm.


Back to Table of Contents
Back to my Lab Page


References:

Davis, Christopher C.  Lasers and Electro-Optics.   New York: Cambridge University Press, 1996.

Laurence, Clifford L.  The Laser Book.  New York: Prentice Hall Press, 1986.

Milonni, Peter W. and Joseph H. Eberly. Lasers.  New York:  John Wiley & Sons, 1988.