Dropping a DPSS is generally very bad. If the any of the optics, or pump sources get misaligned by even a fraction of a mm, you can lose phasing, power density, mode control, and proper collimation.
A YAG/KTP based 532nm DPSS system will typically land between 20-25% efficient via light power out over light power in. So a 1W pump diode can typically be the source of 200-250mws of green light. On a high end system, 300mws is not unheard of.
I will directly quote an explanation I once read, that best describes why YAG is still used, and why 405nm laser diodes are still a better option than directly doubling an 808nm laser diode. You can buy a 1064nm laser diode, but it can't pump a KTP because of its relatively wide bandwidth, non gaussian beam profile, and the lack of the ability to achieve a 50um wide beam with geometric optics.
The explanation excludes the fact that laser diodes are not perfectly monochromatic, whereas to get doubling, two photons with the exact same wavelength must be phased within the KTP before you convince any green photons to come out.
"Unfortunately, with KTP, it isn't simply dumping power into one end and watching it poop[word modified] out a stream of photons through the other.
Several issues come into play here; power density is one of the most obvious. You might think that 2.5W seems like a lot of power, but frankly, it isn't. Second harmonic generation (frequency doubling) has an extremely non-linear efficiency curve- the higher the power, the higher the conversion efficiency becomes. If the power density is too low, you won't get anything out at all.
To put it into perspective, your dinky 5mW green pen laser has WATTS of intracavity power at any given point. Said power is concentrated as a spot that's micrometers wide.
Shooting a diode through a KTP doubler won't get the power density you need, and not only because it makes one pass through the cavity.
Remember that power density is a combination of area and power- in the pen laser, the 1064 beam is only micrometers wide. With a diode, the spot will be significantly bigger.
Now, assuming you got the diode down to a small enough point, you'll also have issues with phase-matching the diode to the KTP. Considering that most diodes at that power are horrible multi-mode, multi-emitter devices, it certainly won't be easy. And assuming you could phase-match it perfectly (which is damn near impossible), it'd be just as hard to get the power density needed.
Anyone remember the directly-doubled Novalux Proteras? They used a special VESCEL diode (which produced a clean, Gaussian beam with very little astigmatism), which also had an external cavity. The cavity didn't just cover the laser diode itself; the doubling optic, the LBO crystal, was also part of the cavity. Consequently, the power within the cavity was higher by magnitudes than not only what was being pumped in, but also what was being emitted. Those diodes are extremely sensitive, but it was the only way to achieve direct doubling.
So, assuming you had an external-cavity, single-emitter 2.5W 808nm diode, you have a slightly bigger problem.
KTP is a glitch. By some quirk, it's brutally efficient at doubling 1064nm. By the same quirk, it can't be phase-matched to double below approximately 500nm, and it won't double at all below 490nm (if I remember correctly).
Consequently, you're left with other options including LBO (which is annoying to work with for more reasons than one) and BBO. Neither of them have the efficiency of KTP."
-Goninanblood from Laserpointerforums from thread "DPSS violet?" Post #3 on 4-26-11
I hope this covers it reasonably well..
mp3car said:
this reminds me of the other physorg article i read a few days ago about a recent advancement in green LED manufacturing... maybe it will help improve the green laser diode... I suspect that would increase the battery life significantly of green pointers.
Green diodes are being developed in the 515-529nm range.
The only info I really have so far, shows that the 515nm diodes are single mode, maxing out at 50mws at 400mA...
horrible efficiency still, but it will get better.