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

Buy Site Supporter Role (remove some ads) | LPF Donations

Links below open in new window

FrozenGate by Avery

single-mode and multi-mode laser diodes

ixfd64

0
Joined
Sep 12, 2007
Messages
1,176
Points
48
As we know, multi-mode laser diodes generally have worse specs than their single-mode counterparts. For example, the specs of most multi-mode 650 nm diodes are something like 5 mrad at 5 mm.

However, the recent 445 nm diodes seem to have much better specs, even though they are also multi-mode. Does anyone know what's so special about those diodes? I know that lasers with shorter wavelengths generally have better specs, but I don't think a difference 200 nm would have that much of an impact. I've heard that laser diodes can be multi-mode in two ways (lateral and transverse), but I have no idea what the difference is.

Also, DPSS lasers generally have very good specs, even though they are pumped by multi-mode diodes. What exactly do DPSS lasers have that diode lasers don't, aside from crystals?
 





Phew! what a question...

As we know, multi-mode laser diodes generally have worse specs than their single-mode counterparts. For example, the specs of most multi-mode 650 nm diodes are something like 5 mrad at 5 mm.

However, the recent 445 nm diodes seem to have much better specs, even though they are also multi-mode. Does anyone know what's so special about those diodes? I know that lasers with shorter wavelengths generally have better specs, but I don't think a difference 200 nm would have that much of an impact. I've heard that laser diodes can be multi-mode in two ways (lateral and transverse), but I have no idea what the difference is.

Also, DPSS lasers generally have very good specs, even though they are pumped by multi-mode diodes. What exactly do DPSS lasers have that diode lasers don't, aside from crystals?

So modes....

First, there are lateral, transverse, and longitudinal modes. It's all about orientation relative to the diode. Longitudinal is in the direction of the propagation of the light, the "length" of the diode. Transverse is perpendicular to the diode/substrate, the "height" of the diode. Lateral is parallel to the diode/substrate, the "width" of the diode. The mode is the volume/shape of light within the cavity. The cavity is defined by: the facets of the cavity in the longitudinal direction, and waveguides in the lateral and transverse directions.

For edge-emitting laser diodes, which is pretty much all the laser diodes we see on this forum, we don't worry too much about longitudinal modes. A lot of people do, but not as much here. So optical storage laser diodes (DVD and Blu-Ray drive diodes) are always nice single lateral and single transverse modes, giving you one nice Gaussian ellipse. That is because you need the one, nice round dot that gives consistent performance for reading and writing data.

That single "moded-ness" come from the waveguide, defined by the structure of the diode, which is built small enough in both the lateral and transverse directions that only one fundamental mode "fits" in the waveguide. You start to run into multimode when you expand the size of the waveguide, and it becomes big enough that multiple modes "fit". The transverse waveguide isn't really changed as much, and pretty much stays single mode. However, to get more power out of a diode, you expand the size of the diode in the lateral direction, getting more current in and more light out, but allowing multiple mode operation in the lateral direction. And that's exactly what we see in these new 445nm diodes: a single transverse mode, and multiple lateral modes.

So multimode is relative, it's all about the size of the lateral waveguide, the wider it is, the more multimode it is likely to be. But suffice to say, in general, multimode is relative and one multimode diode may be much better or much worse than another multimode diode. There's something else going on with these 445nm diodes that makes them better than some others as well, but I'm not ready to go into what that might be. Then you get into multi-emitter diodes, which have multiple waveguides in addition to possible multiple modes in each waveguide, and it goes all to h*** quickly with such lasers.

DPSS is a whole other animal: the diode is only pumping the first crystal, so the first crystal lasing all on its own independent of the diode's optical modes. Since it is its own laser, its own cavity defines the output, and you can get better specs, with things like circular output instead of elliptical, and the different cavity dimensions change things like divergence as well.
 
Last edited:
This is an excellent description. Very technical and you have to think about it quite a bit. Might wanna see if you can add some simple illustrations to help us understand. Like a simple x,y,z representation of the lateral, transverse and longitudinal orientations.
 
No problem. Here is a decent-enough illustration, courtesy of wikipedia.

Here is the most basic design. This view is looking straight at the facet of the diode, the light would be coming straight out at you, and the red is a representation of the shape of the optical mode. In this diode, the longitudinal direction is into and out of the screen, lateral is horizontal, and transverse is vertical.

Notice I mentioned waveguides above. This laser would be referred to as "broad area", "gain-guided", or "loss-guided" because it has NO lateral waveguide. There's no physical construct there to act as a lateral waveguide, so this diode would definitely be multiple lateral modes. Also, there is no transverse waveguide, the mode goes all the way to the metal. This would never be done because it would be hugely inefficient and likely wouldn't work at all, but such a theoretical diode may or may not be multimode in such a configuration, but would likely still be single mode because the dimensions are still small.

531px-Simple_qw_laser_diode.svg.png


Next, this image shows a more typical design. Notice it is still a broad-area laser, with no waveguide in the lateral direction, but that it does have a transverse waveguide, which "holds" the light into a more-confined space vertically. So this would be a single transverse mode, multiple lateral mode, broad-area laser diode, likely similar to the 445nm diodes in many ways. This is still a very simple case, but it does illustrate a good representation of a transverse waveguide.

The light is held into a smaller dimension vertically than it is horizontally, which is why the vertical/transverse direction diverges faster when you're looking at the output of a bare diode with no lens.

563px-Simple_sch_laser_diode.svg.png


Nice single mode diodes would, in addition to the transverse waveguide, add a lateral waveguide. The most common method would probably be etching material off of the diode to create a ridge, with the light confined to an area somewhat within the ridge. For single mode operation, the ridge has to be quite thin, like single-digit microns wide.
 
Last edited:
I too love the descriptions and information. Thanks Kenom for suggesting the pictures and thanks pullbangdead for putting them up and further explaining - it makes much more sense now.
 
"You must spread some Reputation around before giving it to pullbangdead again." Bah! Great post. I have a question, though I am having trouble finding the right words.

So, in your post, you explained that the cavity is generally formed by etching a very very thin ridge out of the waveguide material at that point in construction. This ridge is obscured from view on the finished diode by other layers, but it's in there.

From what I understand, solid state and dye lasers sort of amplify the energy of the pump by absorbing the energy and releasing the energy as a photon when another photon hits it. This occurs in the cavity, which is usually filled with, or made of a gain medium. It is this special material that amplifies the light.

So I was kinda curious if the ridge in a diode, the laser cavity, is filled with a gain medium too, like a classical laser? Would it be pure GaN deposited in that ridge?

Also, I remember reading about the new process of growing GaN that can be cut along a different plane to provide better light extracting properties. Would that just be the big chunk of material on the bottom of the diode in the Wiki diagrams? Or is the entire diode made of GaN with various dopants? I don't understand how this new GaN would help unless it would be used in the active region, but that looks like it is grown on the substrate.

I tried to give this question some structure, but it all went to crap at the end. It also managed to multiply into 4 or 5 more questions.... Sorry 'bout that.
 
Haha, it's all good.

"You must spread some Reputation around before giving it to pullbangdead again." Bah! Great post. I have a question, though I am having trouble finding the right words.

So, in your post, you explained that the cavity is generally formed by etching a very very thin ridge out of the waveguide material at that point in construction. This ridge is obscured from view on the finished diode by other layers, but it's in there.

Yep, it's in there for most diodes. Not for a broad area diode, but it's generally there for the optical storage diodes we see most on here. And it may very well be visible in a microscope, it all just "depends".

From what I understand, solid state and dye lasers sort of amplify the energy of the pump by absorbing the energy and releasing the energy as a photon when another photon hits it. This occurs in the cavity, which is usually filled with, or made of a gain medium. It is this special material that amplifies the light.

So I was kinda curious if the ridge in a diode, the laser cavity, is filled with a gain medium too, like a classical laser? Would it be pure GaN deposited in that ridge?

It is a gain medium. The semiconductor itself IS the gain medium: the active material, in the active region. Touching on more below....

Also, I remember reading about the new process of growing GaN that can be cut along a different plane to provide better light extracting properties. Would that just be the big chunk of material on the bottom of the diode in the Wiki diagrams? Or is the entire diode made of GaN with various dopants? I don't understand how this new GaN would help unless it would be used in the active region, but that looks like it is grown on the substrate.

"Most" of the diode is made up of "alloys" of GaN in the cases of violet, blue, and green laser diodes: GaN, InGaN, AlGaN, in some cases even InAlGaN. All of these are then doped as well, typically Si-doped for n-type and Mg-doped for p-type. With GaN especially, there is also what we call UID, "unintentionally doped", because it's pretty difficult or impossible to grow intrinsic/undoped GaN, it'll always come out a little bit n-type. Or you can just call it "undoped" because it's as close to undoped as possible.

So for laser diodes, you start with a piece of GaN. With LEDs that are a lot easier to make, they usually start with sapphire or silicon carbide. But yeah, that's the substrate. On top of that you grow layers of semiconductor to "build" the diode: GaN, AlGaN, InGaN, etc. With a laser diode, it could easily be dozens of layers, or even hundreds of layers with certain designs. The active region is a layer of lower bandgap material. With GaN laser diodes, the active region, where gain happens, is InGaN. For red and Ir laser diode,s the materials are different, but the ideas are the same. You trap the carriers there, allowing for more gain. The current you push in pumps the semiconductor, allowing for more gain and more light out.

In processing, you add additional materials like facets coatings and metal for contacts, but the semiconductors materials are what I'm talking about here. Also in processing, after you have grown all these semiconductor layers, is where you can, for instance, remove some material to make the waveguide, trapping/confining the laser light into a smaller area inside the semiconductor.

So yeah, umm...the "new GaN". That's all about crystal orientation. The subtrate matters because the orientation of the crystal substrate determines the orientation of the rest of the semiconductor layers you grow. GaN is a hexagonal prism, and think of the diode as always growing on a plane sliced out of that hexagonal prism. Growing on a slice that is differently oriented with respect to the hexagonal prism gives you different properties. So starting on a different crystal plane changes everything throughout the laser diode.

I tried to give this question some structure, but it all went to crap at the end. It also managed to multiply into 4 or 5 more questions.... Sorry 'bout that.

No problem at all.

Did all of that make sense? Any more explanations needed? This one was hard for me to turn my visualizations into words.
 
Yep, it's in there for most diodes. Not for a broad area diode, but it's generally there for the optical storage diodes we see most on here. And it may very well be visible in a microscope, it all just "depends".
I noticed your thread with the amazing pictures of the abused 445nm diode. Looks like the top few layers were blown off, leaving a visible representation of what I'm assuming is this gain region with waveguides on either side of it and on top of it (before it was zapped off).



"Most" of the diode is made up of "alloys" of GaN in the cases of violet, blue, and green laser diodes: GaN, InGaN, AlGaN, in some cases even InAlGaN. All of these are then doped as well, typically Si-doped for n-type and Mg-doped for p-type.

...

So for laser diodes, you start with a piece of GaN. On top of that you grow layers of semiconductor to "build" the diode: GaN, AlGaN, InGaN, etc. With a laser diode, it could easily be dozens of layers, or even hundreds of layers with certain designs.

I see, that clears up how they have so many different "materials" in the wiki diagrams further up this thread, and how the same material can have P and N types. So the diode is more or less entirely made of different mixtures of GaN and other materials, then doped to be either P or N type, and these layers are grown on the beginning substrate of usually "pure" GaN.


The active region is a layer of lower bandgap material. With GaN laser diodes, the active region, where gain happens, is InGaN. For red and Ir laser diode,s the materials are different, but the ideas are the same. You trap the carriers there, allowing for more gain. The current you push in pumps the semiconductor, allowing for more gain and more light out.
Materials in the active region kind of match up with LED materials. Is it that in that layer, light is produced, but the active region bounded by waveguides is what actually allows gain to occur, or is this something else entirely? Semiconductors always confused me with their charge carriers, electron holes, terminology that I can't quite wrap my brain around yet.

In processing, you add additional materials like facets coatings and metal for contacts, but the semiconductors materials are what I'm talking about here. Also in processing, after you have grown all these semiconductor layers, is where you can, for instance, remove some material to make the waveguide, trapping/confining the laser light into a smaller area inside the semiconductor.
Ok, up until now I've been thinking the waveguides are added material to keep the light where it needs to be and going in the direction that it should be. Now I see it is formed by removing material, is this only so that the next layer will take the place and form the final waveguide? I've been kind of thinking of the waveguides as walls that hold the light in and keep crooked light from gaining or something like that. Is there a different, better way to think of them?

So yeah, umm...the "new GaN". That's all about crystal orientation. The subtrate matters because the orientation of the crystal substrate determines the orientation of the rest of the semiconductor layers you grow.
Eureka! That clicked in my brain! Sometimes it just needs a little prod.

Well, let me ask you this: what are your favorite reference materials for diode lasers, besides Mr. Sam's Laser FAQ? It seems like every time I read the LaserFAQ, I end up getting linked to a whole new section that I haven't seen before and I get distracted and read that part too.
 
I noticed your thread with the amazing pictures of the abused 445nm diode. Looks like the top few layers were blown off, leaving a visible representation of what I'm assuming is this gain region with waveguides on either side of it and on top of it (before it was zapped off).

Most of what is gone there is the metal, it looks like most of the semiconductor is still there. Especially looking at the end, you can still see the ridge just fine.


I see, that clears up how they have so many different "materials" in the wiki diagrams further up this thread, and how the same material can have P and N types. So the diode is more or less entirely made of different mixtures of GaN and other materials, then doped to be either P or N type, and these layers are grown on the beginning substrate of usually "pure" GaN.

Yep, sounds right.

Materials in the active region kind of match up with LED materials. Is it that in that layer, light is produced, but the active region bounded by waveguides is what actually allows gain to occur, or is this something else entirely? Semiconductors always confused me with their charge carriers, electron holes, terminology that I can't quite wrap my brain around yet.

Yep. A laser diode is basically an LED + a cavity, and the cavity is the waveguide + the mirrors/facets. You need the waveguide/cavity for feedback, to produce gain.

Ok, up until now I've been thinking the waveguides are added material to keep the light where it needs to be and going in the direction that it should be. Now I see it is formed by removing material, is this only so that the next layer will take the place and form the final waveguide? I've been kind of thinking of the waveguides as walls that hold the light in and keep crooked light from gaining or something like that. Is there a different, better way to think of them?

It's more like both: the transverse waveguide is generally made by adding layers above and below. The lateral waveguide, if there is one, is often formed by removing material, to the sides. You have to separate the lateral and transverse to see how each are made.

And the waveguide is basically as you describe: partial mirrors above/below/to the sides that keep light going in the direction you want it to go in.

Eureka! That clicked in my brain! Sometimes it just needs a little prod.

Good, glad it's working for you.

Well, let me ask you this: what are your favorite reference materials for diode lasers, besides Mr. Sam's Laser FAQ? It seems like every time I read the LaserFAQ, I end up getting linked to a whole new section that I haven't seen before and I get distracted and read that part too.

Yikes, well..."Diode Lasers and Photonic Integrated Circuits" by Coldren is a big one. Pretty technical and 80%+ of it is way beyond anything you need on here, but I keep it around for the ~20% of it that I use. There are others, but that's "the laser diode book". The first few chapters are great reference.

And then just papers, lots and lots of papers. But those are expensive to come by unless you're at a university or company that subscribes to the journals.
 
Bah... PBD.. What the hell does he know?





















Only volumes about laser diodes!!!!!! Really.

Nice, man. You've enlightened us all greatly.. amazing posts. Can't give you any more rep though..
 
So does this mean if I buy a laser diode labeled Single Mode, it will produce a truly circular beam instead of eliptical? It also would be a simpler design and thus be cheaper right?
 
Not of the normal breeds of laser diodes that we use unfortunately. All typical laser diodes produce astigmatic output. There are special types of laser diodes that produce a circular beam, but I've never seen one used or discussed in detail here. Generally speaking though, the oval produced by a singlemode diode is "close enough" to TEM00 that it can be collimated extremely well, it has Gaussian power distribution, and it can be focused down to just as small of a spot. The only thing it lacks is the perfect roundness.

If it's pure TEM00 with good beam quality you're after you'll need to look at solid-state (DPSS, flash pumped YAG etc..), gas, metal halide and ion lasers. The vast majority of laser diodes aren't capable of it.
 
Last edited:





Back
Top