Smallest Spot Diameter: Focusing High Wattage IR Laser Diodes

[metalbot]

Hello Everyone!

I would like to ask some specific questions regarding the optics required to focus a high wattage laser diode.

To give some background to the question: I started an open source initiative to design and prototype what is known as a 3D Metal Printer.

The machine we are designing will use a powerful laser beam to melt very thin layers of metal powder* untill the object is materialized layer by layer. Simple.

*For reference, the powder layers are 50-30um in thickness

Reasons for a Very Tight Focus: To melt the metal powder the laser beam has to reach a critical energy density. This depends on two things:

* (1) The wattage of the laser
* (2) The area of the laser spot on the surface of the powder.

To give an idea of the energy levels required to melt metal powder, current 3D Printers use 200W Fiber Lasers with a focus of 500 microns down to 100 microns.

The single biggest challenge for us is the sheer cost of a high wattage laser like that, so our solution is to go smaller.

Here is the logic: Instead of using a 200W Fiber laser with a focus of 500 microns, why don't we use a 20W laser with a focus of 50 microns? We will have the same critical energy density needed to melt the powder, but use a laser a fraction of the cost.

Questions:

* What is the best focus you can expect to achieve from a high wattage (20-50W) 'laser diode'. for example this fiber coupled one: http://bwt-bj.en.alibaba.com/product...time&edm_ver=e
From what I have read, high power laser diodes have some pretty bad beam quality.

* What lenses will we need in order to get the best focus possible out of the diode.

Keep in mind that the laser beam will travel through a laser scanner.

From my limited knowledge, the raw laser beam goes into a beam expander then into the scanner where it is reflected off the two galvo mirrors then down through the F-Theta lens which focuses the expanded beam down to the work surface.

This part of the project is very similar to a laser engraver.

We have a budget estimate of $3000.

Many thanks in advance

Jethro.

[planters]

I am not an optical engenier, but I have been working for some time with free space diodes (as opposed to fiber pigtailed diodes) to produce small, optimally collimnated 808nm beams with up to 20W for pumping DPSS lasers.

One issue you may have to contend with is that due to the heat affected zone and there may be a non-linear relationship between power density based on spot size and local heating.

You will probably need a NA for your focusing lens in the range of .10 to .20 to reach a corrected spot size of a few hundred um.

You will have to correct the beam for astigmatism as well as for focus and so more than radially symmetrical lenses will be needed.

Got to go. I will think about this some more.

[danielbriggs]

Don't forget the beam profile...
Tophat? Gaussian?

[andythemechanic]

Hi,

a focused fiber coupled diode laser hs a more or less tophat profile. The focal spot size is the fiber diameter times the ratio of collimating lens focal length to focusing lens focal length. So you have to choose a laser which is coupled to the smallest possible fiber diameter (current standard in the 20W range is 100µm e.g. http://www.oclaro.com/product_pages/...9xx-01-R03.php). One thing is that is good is, that energy density goes with the sqare of the focus diameter so for a thenth of the power you only need a ~3.2 times smaller spot.
In principle by playing with the focal length ratio you can achieve a pretty small spot but: your depth of focus is dependent on the 1/NA^2 where NA is the numerical aperture of your focused beam which is beam diameter/(2*focal length). There is a practical limit somewhere at a NA of 0.1 to 0.15 over which it gets very hard to practically realize the small spot because you have to place your part inside the focus to some µm or sub-µm precision. With a laser module NA of 0.12 to 0.15 this limits your practically achievable spot size to roughly the fiber diameter.

Andreas

[planters]

OK, I have thought about this a little more. I agree with the issues presented by Dan and the rough calculations by Andreas. I have actually seen one of these machines in operation and it is neat to watch, but it is pretty slow and that is another consideration here. As you try to minimize the focused spot size the performance of the telecentric lens needs to be quite good with such a limited depth of focus, but even if it preforms well enough your scan speed may need to be reduced in proportion to the average power of the laser. The heat affected zone area might be difficult to predict and not just power/surface area because as the process proceeds the melt zone will be conducting heat out in three dimensions. Before proceeding, you might see if an existing machine (with more typical laser melting) has a PO vs performance table and a spec. sheet on the beam parameters. Also, you need to clarify if the 200 W lasers they are using are actually fiber based vs only fiber pigtailed diode lasers. There is a world of difference in the beam quality of these two laser sources.

[VDX]

Hi Jethro,

I'll hope to get free time at weekend to do some measuring - with the fibercoupled (0.1mm-fiber) 9W-IR-diodes I've measured roughly 0.15mm wide cutting traces with my single-lens head, the 50W-fiberlaser and single focussing lens too was around 0.04mm, but could be 0.02mm too, depends on exposure time and heat dissipation.

Will try to measure the spot diameter of the fiberlaser through the scanning head and the F-theta lens ... my hope is, that this values can be used for the laserdiodes too when having a good collimating setup ...

Viktor

[metalbot]

Thanks for your informed replies! It' taken a few reads to digest .

I want to get specific so I will take the laser diode linked to below as the example.

http://www.alibaba.com/product-gs/34...er_module.html

The reason is that it can lase 50W for a fiber core of 105um, this is the best power density for a laser diode that I have seen so far.

Doing some rough calculations. Taking the lowest laser power density specifications of the commercial laser sintering machine: 200W for a laser spot of 500um, a ratio of 200/500 = 0.4...

...And comparing that with the laser diode: 50W for a laser spot of 105um, a ratio of 50/105 = 0.476.


We can say, at least theoretically, that the laser diode will deliver the critical energy density required for sintering.

Tophat? Gaussian?

Please understand that these terms are very foreign to me at the moment. I enquired about the laser diode linked to above, specifically concerning the best possible focus achievable, and received the following response.

You mean you will use it directly in machining? The beam is of about 10 degree divergence angle. And it is Gaussian distribution. So in our opinion, the beam diameter depends on the distance from fiber tip to the aiming target. Besides, some of my other customers always shape the beam by lens system. It is for your reference. But I suggest you purchase one piece to do a test. After all, different customer has different application.

----

The focal spot size is the fiber diameter times the ratio of collimating lens focal length to focusing lens focal length

I have drawn a simple diagram (attached) that will show my level of understanding far better than any words...

I don't understand what the collimating lens focal length would be.

There is a practical limit somewhere at a NA of 0.1 to 0.15 over which it gets very hard to practically realize the small spot because you have to place your part inside the focus to some µm or sub-µm precision

I digress, but...would that be to do with the accuracy of the galvo scanner or practically focusing the F-theta lens?

The heat affected zone area might be difficult to predict and not just power/surface area because as the process proceeds the melt zone will be conducting heat out in three dimensions.

That is why we need a high energy density, if we draw the laser spot over the surface slowly, the heat will be drawn away from the area.

Unfortunately the commercial systems are very closed when it comes to data about their machines, hence we are starting from the ground up, but I can confirm that they use Yb-fiber lasers. At $680k a machine they sure can afford it.

Which is why it would be ideal if we could get a spot of 100um from a 50W laser diode.

---

Viktor, you have been a help from day one, thank you!

As a side question, how well would a F-Theta lens - designed for focusing a 1064nm Nd:YAG laser - cope with focusing a 970nm beam from a laser diode? In other words, can we bang the laser diode into one of these http://www.alibaba.com/product-gs/30..._marking.html?

Best regards,

Jethro.

[macona]

There is a very good chance that the fiber lasers they are using are q-switched with a 200w average power. This will give a significantly higher power pulse. And really freaking expensive.

You would be better off trying one of the cheap chinese CO2 laser tubes, you can get those up to at least 150watts and will have a much better beam quality. Though your material choice may effect the wavelength needed. If you are using the typical bronze/stainless SLS process CO2 might not work so well.

[planters]

There is a very good chance that the fiber lasers they are using are q-switched with a 200w average power. This will give a significantly higher power pulse. And really freaking expensive.

You may be right, but for a melting operation with a continuous feed of powder I'm not so sure. Nevertheless, even if it is a CW, true fiber laser @ 200W this is still very expensive.

[VDX]

... I'll soon assemble some laserdiode-toolheads for engraving+sintering, so can make some tests with different powders with the IR-diodes.

Attached the diode types I've actually here - the 975nm-types are the best candidates for metal powder for their 0.1mm-fiber ... the 808nm@25W is a diodearray with 19 single emitters bundled to a 0.8mm output, so around 60 times lower energy density than the 975nm@25W!

The 808@1W-diode has a fiber with 0.05mm core diameter, so in energy density comparable with the 975nm@5W-diode for sintering plastic powder -- but won't be usable with metal powder, as here the much higher heat dissipation in the metal needs more overall energy ...

Viktor