Welcome to Laser Pointer Forums - discuss green laser pointers, blue laser pointers, and all types of lasers



The DPSS Primer...(and how lasers work)

Status
Not open for further replies.

goninanbl00d

New member
Joined
Jun 12, 2010
Messages
892
Points
0
I've compiled a short (if you could call it short) primer on the workings of DPSS lasers and DPSS laser systems here. This guide aims to answer the common questions and misconceptions about DPSS laser systems. A basic knowledge of lasers is essential, and the Wikipedia article on lasers (linked here) is a good primer for this article.

'DPSS' stands for Diode Pumped Solid State. In this case, 'solid state' refers to the gain medium, not to the electronics involved. Many simply associate DPSS with a 808nm pump diode, and a crystal set. However, there is much more to it.

Firstly, DPSS lasers are not 'filters' or 'optics' that change the colour, shape or any other attributes of a beam, nor are they fluorescent materials. Instead, they are a separate laser system, independent on the pump laser diode, hence the name diode-pumped solid state. This is the reason why a DPSS laser will have completely different beam specifications from it's pump diode.

To understand how the DPSS process works, you will need to understand how a basic laser works. In this case, I will be using the flashlamp-pumped Ruby laser to explain, as it is one of the simplest laser systems.


As you can see, the laser resonator is composed of 4 basic elements- the high reflector, output coupler, gain medium, and pump.

In this laser, the flashlamp is used to excite atoms in the ruby into a higher energy level. Once in this higher energy state, they will return to a lower energy level by emitting a photon. This photon travels along the cavity, between the two mirrors. When this photon is travelling, it may collide with another atom, causing it to release a photon, and return to a lower energy level. This process is the 'stimulated emission' in the acronym LASER. These photons resonate in the cavity, triggering more emissions, with some photons escaping the cavity every pass.

In a laser, a pump can be anything, from a flashlamp, a chemical reaction, or even another laser. In a DPSS laser, the pump source is replaced with a high-powered laser diode, which has it's wavelength matched to one of the major absorption bands of the gain medium.


In this image of a DPSS system, you can see the elements of the laser resonator, as described above. You will also see that the flashlamp has been replaced with three diode bars, and that there is a KTP and a Pockels Cell. These intracavity optics will be discussed later.

The principal of operation is the same as the ruby laser described in the section above. The diode bars are used to pump the neodymium atoms in the Nd:YAG into an excited state. The 1064nm photons resonate between the cavity mirrors, and are doubled by the KTP crystal into 532nm. These 532nm photons can proceed to leave the cavity through the cavity mirrors.

The main difference here, apart from the pump source, is that this laser is capable of running in Continuous Wave mode, as opposed to the flashlamp-pumped Ruby laser, which is only capable of emitting laser radiation in pulses. This is partly due to the change in pump sources, as well as limitations of ruby which prevent it operating in CW mode.

Although not a typical representation of your 'average' DPSS system (this is a multi-watt Q-Switched system), this cavity layout makes it much easier to see the similarities between it, and the ruby laser above. With this layout, the Nd:YVO4 gain medium and KTP no longer appear to be 'filters', instead, it is now apparent that they are a separate laser system.


The above diagram shows a DPSS laser that is much more commonly encountered (as opposed to the multi-watt Q-Switched system shown in the previous diagram). As you can see, it has been simplified, but the cavity components are mostly the same.

In this case, however, the High Reflector is not independent, but is a coating on the Nd:YVO4 crystal. The laser is also end-pumped, with the die of the pump diode in direct contact with the face of the Nd:YVO4.

Although the pump diode is a multimode, multiemitter 808nm diode, no FAC or corrective optics are generally used in cheap modules. In many cases, the diode is of an open can type, and the diode's die is pressed against the surface of the Nd:YVO4 (which, along with the KTP, is coated to form a cavity). When closed can diodes and/or lower pump powers are used, an ultra-short focal length lens is placed between the diode and crystal set. This lens serves to focus the pump light to a pinpoint just below the surface of the Nd:YVO4, and achieve the power densities necessary for lasing (as opposed to just fluorescence)

It may be interesting to note here that the infrared leakage is mainly 808nm, not 1064nm. The 1064nm is reflected back into the Nd:YVO4 by the output coupler, however, the output coupler is not coated for 808nm, and this radiation passes through unhindered. As a result, the majority of IR laser radiation leaking from cheap laser pointers and modules is 808nm, not 1064nm. The only time 1064nm will leak out is if the coatings on the output coupler are damaged.


As you can see, there are many small components that need to be aligned before the laser will work. It is not unheard of that modules and laser pointers arrive emitting IR only, with the crystals misaligned after damage incurred during shipping or manufacturing.

With these lasers, realignment is possible, but can be difficult. With a bonded crystal set, alignment is simplified as all cavity optics are already self aligned, and it is only a matter of phase-matching the pump laser diode to the Nd:YVO4 for optimal power output.


Appendix 1: Creating different wavelengths

Although I have covered the basic 1064nm/532nm Nd:YAG/YVO4 laser, there are many other wavelengths available through the DPSS process.

Many of these other wavelengths use the different lasing lines available from Nd:YAG and Nd:YVO4. Although the 1064nm lasing line is dominant in these two crystals, other, weaker lasing lines do exist, and with the correct optics (such as the corresponding output couplers and high reflectors), these other lines can be made to lase.

The main difficulty, however, is that the dominant line (or lines) will try to compete with the selected lines for energy. This is most commonly seen in Argon-Ion lasers, where the majority of the lines will appear as tube current is turned up.


In a 473nm laser, for example, the 946nm line of Nd:YAG is used. In this case, both the 1064nm and 1319nm compete with the 946 line for energy, and overall efficiency is very low as a result.

In a gain medium where one line is significantly more 'prominent' than other lines (such as in Nd:YAG), only one line will be actively lasing given a moderate pump power. There will be side-emissions caused by fluorescence, however, these emissions are not coherent (as there is no resonance at this wavelength), nor do they have the energy necessary to undergo SHG or SFG.

You can see that the coatings have been changed to promote resonance at 946nm, while suppressing other frequencies such as 1064nm and 1319nm.

In other cases, different gain media with different lasing wavelengths used. In 515nm lasers, for example, a Yb:YAG or Yb:KGW gain medium is used. This lases at 1030nm, which is then doubled to 515nm.

A change of gain medium will often mean a change of pump wavelength. In the case with Yb:YAG, it has an absorption spectrum at 940nm, and an appropriate pump diode is used.

This is the only influence the pump wavelength has on the final wavelength, the pump wavelength must match an absorption spectra of the gain medium. Often, a gain medium will have more than one absorption spectra (Nd:YVO4, for exmaple, has one at 808nm, and another at 780nm). If this is the case, the more efficient one is used.

Below is a list of common gain media, their absorption bands, and their output wavelengths (listed in order of dominance). This is by no means a comprehensive list, but I am aiming to cover all DPSS wavelengths CNI offer, at the very least.

Nd:YAG (Neodymium-doped Yttrium Aluminium Garnet)

Absorption bands: 808nm (+/- <1nm)

1064nm (SHG to 532nm via KTP)
1319nm (SHG to 659.5nm via KTP)
1342nm (SHG to 671nm via KTP)
946nm (SHG to 473nm via LBO/BiBO)
1123nm (SHG to 561nm via KTP)

Nd:YVO4 (Neodymium-doped Yttrium Orthovanadate)

Absorption bands: 808nm, 780nm (808 is used due to higher efficiency)

1064nm (SHG to 532nm via KTP)
1319nm (SHG to 659.5nm via KTP)
1342nm (SHG to 671nm via KTP)
914nm (SHG to 457nm via LBO/BiBO)

Nd:YLF (Neodymium-doped Yttrium Lithium Fluoride)

Absorption bands: 792nm

1047nm (SHG to 523nm via KTP)
1053nm (SHG to 526nm via KTP)

Yb:YAG (Ytterbium-doped Yttrium Aluminium Garnet)

Absorption bands: 940nm, 970nm (940nm is used due to higher efficiency)

1030nm (SHG to 515nm)
1050nm (SHG to 525nm)

Yb:KGW (Ytterbium-doped Potassium Gadolinium Tungstate)

Absorption bands: 981nm (+/- <1nm)

1030nm (SHG to 515nm via KDP)
1050nm (SHG to 525nm via KTP or KDP)


Appendix 2: Non-linear Optical processes

In all of the lasers above, there has been mentions of frequency doubling. Frequency doubling, otherwise known as Second Harmonic Generation (SHG), falls into a category of processes known as Non-Linear optical processes.

These processes also include Third Harmonic Generation (frequency tripling), and Sum Frequency Generation (where two photons are combined to produce a single photon with the energy of both photons). However, the most common process is still frequency doubling.

In commonly-available DPSS systems, there are two different SHG crystals used. The first of these is KTP, while the second is LBO. Both of these crystals have respective scenarios where they are used, and others where their use is contraindicated.

KTP requires no critical phase matching to the gain medium, and at low-to-medium power, requires little to no temperature stabilisation. KTP also has a high degree of non-linearity, meaning that the threshold required for doubling to occur is lower, and a smaller crystal can be used as a result.

Due to a 'quirk' in the physical properties of KTP, it cannot be effectively phase-matched below 500nm, and as a result, cannot be used to double wavelengths below 500nm. However, it is this very quirk that gives KTP it's abnormally high efficiency (compared to other SHG media) when doubling 1064nm Nd:YAG and Nd:YVO4 lasers.

LBO requires critical phase matching in order to achieve maximum efficiency (or any doubling effect at all), as well as requiring temperature stabilisation to 0.1 of a degree. A larger crystal must be used, and temperature stabilisation must be used in order to facilitate SHG. However, LBO is one of the few SHG media that can double <500nm light.

There have been developments concerning other SHG media for use, mainly BBO and BiBO. These crystals are used for doubling into the deep-UV range, and will not be covered here.

Sum Frequency Generation is an entirely different process altogether, where two photons are summed, and the average taken. This process is most commonly seen in 593.5nm lasers, where Nd:YVO4 is made to simultaneously lase at 1064nm and 1342nm. These are then combined in a KTP to form 593.5nm


Appendix 3: Oops, what went wrong?

It's not unheard of around here that someone will buy a laser, and it will emit nothing but IR light. This is usually the result of a dying pump diode (unable to output above threshold), or due to misaligned/loose crystal sets.


NEVER look into the laser at the dim red light- you run the risk of inflicting permanent retinal burns


To achieve even a low-power output requires large amounts of 808nm IR pump radiation. For example, the pump diodes found in a 50mW green laser pointer will often average around 300mW IR output.

The human eye isn't very sensitive to 808nm IR- and it's barely visible. It takes several watts of 808nm to achieve visual parity with 1mW of 650nm. As a result, what is mistaken as a dim red light is easily capable of causing instantaneous eye damage.

Although the diagram used earlier shows an IR filter (which would cut out all of the infrared from the green beam), most cheap pointers and modules lack this IR filter.

The second most commonly experienced anamoly with DPSS laser pointers and modules is transverse mode-hopping or mode-shifting. When this happens, the laser no longer operates in a single transverse mode (TEM00), and instead operate in a higher-order transverse mode.



The above table shows some of the most common modes, there may be more higher-order modes or modes superimposed onto other modes.

The most common cause of mode-hopping is a weak pump diode (caused by a dying battery), or temperature instability. The properties of both the gain medium and SHG crystal change with temperature, and a stabilisation period is required just after startup.


Appendix 4: Q-Switching


Although Q-Switching is applicable to much more than just DPSS laser systems, there have been many questions in the past about what it is, what it does, and how it is used.

Q-Switching is the use of a modulator (either passive or active) to modulate the Q-factor of a cavity. The Q-factor defines the losses in the optical resonator; a high Q-factor means that there are extremely low losses per round trip, and vice versa. By modulating the Q-factor, lasing can be controlled and extremely short pulses with extremely high peak intensities can be created.


In a Q-switched laser, the Q-switch is closed to begin with. The pump source will continue to raise the energy states of the atoms within the gain medium, but as the Q-switch is closed, lasing is unable to occur. In this diagram, a Q-Switched Nd:YAG laser is show. The beam path is shown in yellow, neodymium atoms in the ground state are shown in blue, while neodymium atoms in the excited state are shown in yellow.


Eventually, a population inversion will be achieved, but lasing action cannot commence. Due to losses caused by spontaneous emission (shown here as the yellow beam stopping at the Q-Switch and output coupler), the gain medium will reach a maximum energy state, after which pump input will have no further effect on the energy level inside the gain medium.


At this point, the Q-switch opens, and the Q-factor of the cavity will rise. Now unobstructed, lasing can commence. As a population inversion has already been achieved in the gain medium, this energy is suddenly released in one short, high-intensity burst.



Q-switching is trading pulse length for peak power; Q-switching does not imply a higher average output power, or a higher average power for the duration of the pulse. In fact, many Q-switched laser systems have a lower average output power than their CW or pulsed non-Q-switched equivalents, however, their peak power during the pulse is much higher, often by a factor of 10 to 100.

Q-switching is not a method of modulating the laser's output in the same sense as TTL or analog modulation (the equivalent of this would be an extracavity AOM).

Passive and active Q-switching referrs to the operational method of the Q-Switch.

A passive Q-Switch is often a saturable absorber such as a dye, or a crystal such as Cr:YAG. These substances bleach once a certain power density has been achieved, allowing the cavity to lase. Pulse repetition rates in a passive system are dependant on the Q-switching medium used, and on the pump power.

An active Q-switch is usually an AOM or EOM (such as the Pockels Cell mentioned earlier). These are fully controllable, and pulse rates can be fully controlled.
 
Last edited by a moderator:



csshih

Active member
Joined
Sep 27, 2010
Messages
431
Points
43
I vote for a sticky.

Excellent read for the budding laserist!
 

Eudaimonium

Well-known member
Joined
Feb 5, 2008
Messages
6,311
Points
83
I vote for a sticky too! Great work!

Gotta bookmark this to show it to people who come asking me about green lasers.

Thanks, man!
 

Leodahsan

New member
Joined
Aug 7, 2010
Messages
2,088
Points
0
just a thought: the output infrared light on the inexpensive dpss method - isn't it 808nm instead of 1064nm? or both?

+1 to sticky, great work.
 

Meatball

New member
Joined
Feb 1, 2008
Messages
2,987
Points
0
just a thought: the output infrared light on the inexpensive dpss method - isn't it 808nm instead of 1064nm? or both?

+1 to sticky, great work.
The Red should be 1064nm, the yellow 808nm.
 

Eudaimonium

Well-known member
Joined
Feb 5, 2008
Messages
6,311
Points
83
The Red should be 1064nm, the yellow 808nm.
Wait what?

I don't understand his question nor your answer.

But in case ya'all mean waveleght of pump diode, it's always 808nm I think, for both green and 473nm blue. Not sure about yellow and 671nm red.
 

goninanbl00d

New member
Joined
Jun 12, 2010
Messages
892
Points
0
I've added some revisions, clarifications, and two new appendices. Hope this helps.
 

Trevor

Well-known member
Joined
Jul 17, 2009
Messages
4,512
Points
113
Appendix 2: Non-linear Optical processes

In all of the lasers above, there has been mentions of frequency doubling. Frequency doubling, otherwise known as Second Harmonic Generation (SHG), falls into a category of processes known as Non-Linear optical processes.

These processes also include Third Harmonic Generation (frequency tripling), and Sum Frequency Generation (where the average of two wavelengths taken). However, the most common process is still frequency doubling.
It's not averaging of the wavelengths, per se.

Second harmonic generation and the other frequency-multiplication processes are all special cases of sum frequency generation. In all cases of sum frequency generation, two photons collide and they combine into a single photon with the energy of both photons. The new wavelength reflects this different energy, but it is not an average of the wavelengths.

The most common example is a 1064nm photon and 1342nm photon being rammed together to yield a photon of the sum of the energies - at a wavelength of 593.5nm. :)

-Trevor
 
Last edited:

goninanbl00d

New member
Joined
Jun 12, 2010
Messages
892
Points
0
It's not averaging of the wavelengths, per se.

Second harmonic generation and the other frequency-multiplication processes are all special cases of sum frequency generation. In all cases of sum frequency generation, two photons collide and they combine into a single photon with the energy of both photons. The new wavelength reflects this different energy, but it is not an average of the wavelengths.

The most common example is a 1064nm photon and 1342nm photon being rammed together to yield a photon of the sum of the energies - at a wavelength of 593.5nm. :)

-Trevor
I was tired, and couldn't think of a better way to describe it. Thanks for the heads-up, it's since been fixed.
 

goninanbl00d

New member
Joined
Jun 12, 2010
Messages
892
Points
0
It's not averaging of the wavelengths, per se.

Second harmonic generation and the other frequency-multiplication processes are all special cases of sum frequency generation. In all cases of sum frequency generation, two photons collide and they combine into a single photon with the energy of both photons. The new wavelength reflects this different energy, but it is not an average of the wavelengths.

The most common example is a 1064nm photon and 1342nm photon being rammed together to yield a photon of the sum of the energies - at a wavelength of 593.5nm. :)

-Trevor
I was tired, and I could have expressed it better. I've since changed it.

Also, seeing as ElektroFreak liked the graphics so much, I decided to add them in for the section about Q-Switching as well. That should really help clear things up, it can be hard explaining without diagrams. :)
 

LaserBen

New member
Joined
Jan 1, 2011
Messages
1
Points
0
Very well written post goninanbl00d, very informative and factual.

A couple things that you may consider adding to it.

Focusing the pump light from a diode down into a point serves two purposes, one, as you mentioned, is to get he power density to a high enough level to induce lasing, another important reason is to control the mode of the final output, this is called mode-matching. By pumping a small, conical portion of the lasing medium, you force the laser to lase only through that space, this kills off any random modes that may occur, generally it is done to keep the desired mode, sometimes it is just to prevent the mode from shifting.

Many of the common SHG and THG frequencies can be achieved with more than one type of doubling medium. Probably the most relevant wavelength to use as an example is 1064nm to 532nm. Several different materials can be used, the most common are KTP, LBO, and KDP. KTP is used for general applications where extreme stability, conversion efficiency, and high energy density are not critical. Many CW scientific lasers use either critically phase matched or non critically phase matched LBO as it provides a very efficient conversion and high stability. I won't get into the difference between critical and non-critical phase matching, there is a good article about it somewhere. Non Critically phase matched LBO is very temperature sensitive, the conversion efficiency follows a sinusoidal type plot with the temperature. On high energy pulsed lasers, KDP or KD*P are the desired doubling mediums as they provide very high efficiency and extreme damage thresholds, necessary for the intra-cavity power densities.

I'll add some more stuff later.

Oh, and Dave, I know you banned me because you disagree with my business, I know I am violating the rules of the forum by making another account, just ignore it, nothing is meant to cause harm here...
 

goninanbl00d

New member
Joined
Jun 12, 2010
Messages
892
Points
0
Very well written post goninanbl00d, very informative and factual.

A couple things that you may consider adding to it.

Focusing the pump light from a diode down into a point serves two purposes, one, as you mentioned, is to get he power density to a high enough level to induce lasing, another important reason is to control the mode of the final output, this is called mode-matching. By pumping a small, conical portion of the lasing medium, you force the laser to lase only through that space, this kills off any random modes that may occur, generally it is done to keep the desired mode, sometimes it is just to prevent the mode from shifting.

Many of the common SHG and THG frequencies can be achieved with more than one type of doubling medium. Probably the most relevant wavelength to use as an example is 1064nm to 532nm. Several different materials can be used, the most common are KTP, LBO, and KDP. KTP is used for general applications where extreme stability, conversion efficiency, and high energy density are not critical. Many CW scientific lasers use either critically phase matched or non critically phase matched LBO as it provides a very efficient conversion and high stability. I won't get into the difference between critical and non-critical phase matching, there is a good article about it somewhere. Non Critically phase matched LBO is very temperature sensitive, the conversion efficiency follows a sinusoidal type plot with the temperature. On high energy pulsed lasers, KDP or KD*P are the desired doubling mediums as they provide very high efficiency and extreme damage thresholds, necessary for the intra-cavity power densities.

I'll add some more stuff later.

Oh, and Dave, I know you banned me because you disagree with my business, I know I am violating the rules of the forum by making another account, just ignore it, nothing is meant to cause harm here...
Hey Ben,

Welcome back.

Firstly, it's a very dumbed-down guide, as it's to help newbies understand. Hence, it's missing stuff and some other stuff has been over-simplified. But it's a start, and if anyone wants to learn more, there's plenty of resources that pick up where this left off.

If I ever get around to doing it (maybe after April SPL 2011, maybe :banghead:), I'll do a PDF version with much more info as a download.

The reason why I didn't mention mode-matching is because with cheap DPSS modules, it's not used in most cases, the focal point isn't small enough. With some larger modules, it is used effectively, but mostly, pump focusing is just to achieve a higher power density.

About the NLO media, once again, didn't want to confuse anyone just yet. Will do when I get what's in front of me sorted.


Cheers,

Andrew
 

gklecker

Banned
Joined
Feb 15, 2011
Messages
4
Points
0
Well. I am new to this forum and this was extremely helpful to me. I got a generic answer about Frequency doubling at work and no one could explain it for me. It's sad, now I'm the expert here after reading this. :crackup:

So where do I find a link that explains what types of setups are used to create the IR lasers?

Glenn
 

MrB

New member
Joined
Mar 22, 2011
Messages
17
Points
0
Hi Andrew,
Is it possible to do this without the Q-Switch for CW operation?
If so, what sort of efficiency could one expect?
Say you use 3x 20W side pump diodes, would it be reasonable to expect say 15W CW @ 1064nm (using a 25% efficiency guesstimate)
I would like to do something like this for a CNC engraver/router, but starting to think a cheap chinese CO2 tube, even with it's lifetime issues.. may be easier.

Edit: Oh and what is going on in your avatar? Too small to see...
 
Last edited:
Status
Not open for further replies.




Top