Turn-on delay time for Laser diode Do you know any simple explanation on the reason why the turn-on delay  time on a laser diode is reducing while we increase the bias current?
Turn on delay,is the time that the laser needs from the time that one applies the current until the time that the light goes out of the laser.This time is strongly depended to the input current density,the higher the bias current it is the less the turn on delay it is. That I don’t understand is the physics behind it,how that interaction occurs.
Is it something obvious, because I am trying to find a simple explanation and I can not.
Best Regards,
George
 A: There are a few factors that govern laser diode turn-on time.  
The first is the junction capacitance, which is the same thing that causes turn-on delay in an ordinary junction diode.  Under forward bias, the capacitance is proportional to the current and the diode transit time, just as with any pn-junction diode.
The second is unique to laser diodes.
Below a certain threshold at which lasing begins, the device acts like an ordinary LED.  When the current is increased to the point that the gain of the laser is equal to the loss of the cavity and mirrors, lasing starts and the light output increases suddenly.  It's at this point that the laser diode is considered to be turned on.  
What causes the delay is simply the time it takes to get to the threshold current $I_{th}$ For an ideal device with no capacitance, the turn on time is $$t_d = \tau_s \ln\frac{I}{I-I_{th}}$$ where $\tau_s$ is the recombination time.  (Remember that it's the recombination of electrons and holes that causes the emission of light in an LED or laser diode.)
A: Let's put a step function of voltage across a laser from a low impedance voltage source. The current rise is limited by the series inductance. Use multiple bond wires and good packaging to effectively eliminate this from consideration.  So we have a step function in current. [I have done this with multiple fast FET switches on the laser substrate.]
There are two processes that now limit how the laser turns on.  The first is how fast the carriers thermalize into the quantum well.  This time is the slowest semiconductor transport process, slower than diffusion of the current etc. (but still quite fast, ps).  The second, and more common effect particularly with standard laser diodes is the limited gain per round trip.
A 3 mm semiconductor laser of index of refraction of 3 has an optical length of 9 mm and a round trip time of 60 ps.  So, for example, a gain-switched laser has to wait until the current exceeds transparency, then wait for a lucky photon to head down the waveguide in the confined mode (primarily responsible for jitter in gain switched lasers) and then depending on the dynamics of the current (which sets the round trip gain per pass), how many passes are needed to get the laser output to be "on": something close to it's CW output power.  One can gain some feeling for what is needed by noting a 1 mW laser at 1.24 um (1 eV photon)  puts out 6.24 x 10^15 photons/sec. In a 60 ps round-trip time, that is 3.75 x 10^5 photons the gain media has to replace per round trip time. [CW lasers always operate at a gain of 1.]  If the output coupler is, say, 10% transmissive, the circulating light is just 10 times the output light.  So given the gain is not hugely greater than 1, how many passes does it take to build up enough photons in the cavity for you to consider it on.
From here you have to get directly into the carrier dynamics covered in any laser text. But to bound it, say you do get a round-trip gain of 2.  That's about 24 passes. (but the gain is not constant etc,] This makes the turn on time 24 x 60 ps or 1.44 ns.
So it really depends on how fast you can make extra carriers in the excited state resulting in gain > 1, and how many passes it takes to get the laser "on".  Typically the series inductance limits the rate of increase of carrier injection, so there is a complicated transient of rising current followed by delayed stimulated emission as the photon density in the cavity goes up.
In practice for telecom lasers, they are never turned off.  By keeping the space output at 5 - 10% of the mark output you only need a few round trip times, and smaller current change to get the laser on again.  The extra photons in the cavity also speed up the laser turning "off" as "off" is now 10% of "on" so spontaneous emission will dominate the carrier removal rather than the much longer spontaneous decay time (~ 1 ns).
Hope that helps.  I use Steigman's book "Lasers" but I am sure there are more recent treatments.
