Software … powerful tools for your research & development!

Ytterbium Fiber Lasers Emitting At 975 nm

Dr. Rüdiger Paschotta

It has been a longer while since I wrote the last software newsletter – mainly because we moved to Switzerland this year, which made me particularly busy. Anyway, I am now basically back to normal operation mode and present the next posting. Let us consider 975-nm Yb-doped fiber lasers because these exhibit some interesting peculiarities. Note that most Yb-doped fiber lasers operate at longer wavelengths between 1030 nm and 1100 nm, but lasing at 975 nm is possible.

Most fiber lasers are based on a quasi-three-level laser gain medium, and operated with quite high optical intensities – even in cases where the optical powers are quite moderate. Therefore, we do have particularly strong quasi-three-level effects – not only related to the pump and laser wave, but also with important implications on amplified spontaneous emission (ASE). This issue can even decide on whether or not a certain kind of laser can work or not.

Core-pumped 975-nm Laser

We begin with a core-pumped 975-nm ytterbium fiber laser. It is easy to explore the operation details of such devices using our RP Fiber Power software, simply starting with one of the nice demo files containing a custom form, i.e., a form which is defined in a script file and can thus be further tailored to the user's needs. (If you have the software but not yet that script, just tell me, and you will get it.) Here is a screen shot of that form:

RP Fiber Power fiber laser form

The input parameters are based on the selection of a commercially available standard single-mode gain fiber with a suitable length, end mirror reflectivities from narrow-band reflectors and the selection of the 975-nm laser wavelength. Basically, we assume that only the laser wave at 975 nm is reflected strongly at the ends (100% at the left one, 25% at the right end which is the output end), while the ASE gets only 4% reflectivity from a bare fiber end at one side and no reflection at the output end. (The output laser light may not go to free space, but in a fiber spliced to the laser, avoiding any Fresnel reflection.) Further operation parameters can be seen in the screen shot.

The first diagram obtained by executing this shows the optical powers vs. position in the fiber:

optical powers vs. position
Figure 1: Distribution of optical powers and Yb excitation in the core-pumped fiber.

We see that most pump light is absorbed (red curve), and the Yb excitation is quite uniform over the full fiber length.

The second diagram shows the ASE spectra at both fiber ends – with not much power, having no significant effect on the performance:

ASE spectrum
Figure 2: Output as functions of the pump power.

So far no problem at all – but we get that in the next case:

Cladding-pumped 975-nm Fiber Laser

We now try to similar things done at much higher power levels, using a double-clad fiber. (I selected CorActive DCF-YB-10-128.) We pump that with 30 W instead of only 300 mW. Due to the reduced overlap of pump light with the ytterbium-doped fiber core, we need a substantially longer length – let's try 5 m instead of 1 m, although the chosen double-clad fiber has an 8.3 times higher Yb concentration in the core! Even that length of fiber absorbs only roughly two thirds of the pump power, but more importantly we cannot even get it lasing. Instead, we produce a lot of ASE, mostly in forward direction:

variation of pump power
Figure 3: Outputs of the cladding-pumped laser as function of pump power: no lasing achieved, only generating ASE!

But how can you obviously get a lot of gain, generating ASE output on both sides, but not enough gain for lasing despite moderate resonator losses?

The answer is that the gain is wavelength-dependent, and is by far largest around 1030 nm where it produces ASE. To illustrate that, I have inserted a little script code for generating another diagram: gain at 975 nm (the laser wavelength) and at 1030 nm (near the ASE maximum) as functions of pump power. The result:

gain vs. pump power
Figure 4: Gain at two wavelengths versus pump power.

Here we see that the gain at 975 nm starts out strongly negative (−5000 dB!) due to the high absorption cross-section at that wavelength, while much less of such absorption occurs at 1030 nm. For increasing pump power and rising ytterbium excitation, both gains grow. The 975-nm gain grows substantially faster, but that growth essentially stops where ASE becomes so strong that it prevents a substantial further growth of Yb excitation.

Behind all this are the transition cross-sections (for absorption and stimulated emission) of Yb3+ in germanosilicate glass:

cross-sections of Yb-doped glass
Figure 5: Absorption and emission cross-sections of ytterbium-doped germanosilicate glass, as used in the cores of ytterbium-doped fibers (data from spectroscopic measurements: R. Paschotta et al., “Ytterbium-doped fiber amplifiers”, IEEE J. Quantum Electron. 33 (7), 1049 (1997);

You can see:

  • At 975 nm, we have a high emission cross-section, but also a similarly high absorption cross-section. Therefore, stimulated emission prevails only if we have at least ≈50% Yb excitation. But in our case we cannot reach that state because well before we get there, the gain at 1030 nm becomes large enough to produce strong ASE. Via stimulated emission, that saturates the gain and limits the average Yb excitation to well below 10%.
  • Around 1030 nm, reabsorption is much weaker. Therefore, the gain turns positive for far weaker excitation already.

It is also instructive to see the effective gain spectrum for different excitation levels between 0 (blue) and 100% (red):

Yb gain

You can see that the 1030-nm gain turns positive for moderate Yb excitation already, while for positive gain at 975 nm we need more than 50% excitation.

Some further questions wait to be addressed:

Why is that a problem specifically with double-clad fibers?

This is because there we need substantially more Yb in the system to absorb enough pump light. (Note that we had to increase the fiber length despite using a highly doped fiber.) We could partially solve the problem by using a much shorter fiber (e.g. 0.2 m only) and accepting the then quite poor pump absorption efficiency. That way, we can at least get it lasing at 975 nm without producing much ASE. But we get only 0.29 W for 30 W of pump power (without further parameter optimization).

There is actually another solution: using a ring-doped fiber (J. Nilsson et al., Opt. Lett. 22 (14), 1092 (1997)). That reduces the coupling of the dopant to the ASE and signal waves, but not to the pump light. In effect, one can have higher excitation densities, allowing for gain at shorter wavelengths.

Why is there so little backward ASE?

Let us go back to the long fiber and analyze that more closely. ASE has the same gain in forward and backward direction; how can it be that we see so much more ASE output in forward direction (15.9 W for 30 W pump power) than in backward direction (106 mW)? Well, this is because we have a 4% reflection at the left end but no reflection at the other end. So 4% of the ASE generated in backward direction is reflected and acts a seed for forward ASE, which thus becomes far stronger. Note that backward ASE has to start at the right end with zero initial power.

Without that 4% reflection, we even get some more backward ASE compared with forward ASE:

gain vs. pump power

This is basically due to the asymmetric distribution of Yb excitation. Note also that the single-pass ASE gain in the fiber gets substantially higher than with that reflection.

By the way, with ASE reflections of 4% at both sides, we would even get lasing in the 1030-nm region, preventing 975-nm lasing even more.

You may also want to know how the ASE spectra look (for 30 W pump power):

ASE spectra

They look very different for forward and backward ASE (on a logarithmic power scale):

  • Forward ASE is far stronger, as we have seen already, and this only at longer wavelengths. The 975-nm peak is not seen there because that radiation is reabsorbed on the way to the output end, where the Yb excitation is low enough for that.
  • Backward ASE does exhibit that 975-nm peak, although it is not as strong as the longer-wavelength ASE, where the gain is higher.

Note that the effect of the 4% reflection is not taken into account in the ASE spectra – it could be thought to occur only within a very small optical bandwidth and is thus ignored for simplicity.


First some technical/scientific conclusions:

  • Reabsorption by non-excited ytterbium can have highly important effects on such lasers – it can even lead to total failure of getting it lase at the wanted wavelength.
  • The problem for a 975-nm laser can get substantially worse with cladding pumping of double-clad fibers, essentially because there you need to put more Yb into the system for efficient pump absorption, with the consequence that gain at longer wavelengths (1030 nm) becomes strong well before you get any positive get at 975 nm. Then you can only generate ASE around 1030 nm.
  • You can in principle avoid that problem by using far less ytterbium, but then the pump absorption efficiency gets extremely poor. Ring doping of the fiber would be another solution, but such fibers are difficult to procure.

All sorts of similar problems can be encountered in other fiber lasers as well as in fiber amplifiers.

Some more thoughts:

  • The only effective way to learn such essential things in laser physics is to actively play with such models. By just passively digesting even the nicest pieces information (articles like this, textbooks etc.) you never get a similarly deep understanding. Also not by building such lasers, since you cannot look inside to see what really happens. Only a laser model is a kind of software modeling transparent laser, where you can inspect any internal things until you fully understand it.
  • Developing such lasers with trial & error is slow, frustrating and wasteful concerning time and money. You need a simulation model to understand it and design it properly. Even if you had a good understanding of such issues without having a model (not sure how to get that understanding then!), it would be very hard to anticipate all sorts effects with sufficient quantitative accuracy.
  • Understanding solid-state bulk lasers is often much less demanding because they exhibit lower gain (thus less ASE), often small excitation levels etc., resulting in simpler operation characteristics. Nevertheless, modeling can be quite important for doing effective work.

Finally, I want to remind you of my comprehensive tutorial on the modeling of fiber amplifiers and lasers.

This article is a posting of the RP Photonics Software News, authored by Dr. Rüdiger Paschotta. You may link to this page, because its location is permanent.

Note that you can also receive the articles in the form of a newsletter or with an RSS feed.

Questions and Comments from Users

Here you can submit questions and comments. As far as they get accepted by the author, they will appear above this paragraph together with the author’s answer. The author will decide on acceptance based on certain criteria. Essentially, the issue must be of sufficiently broad interest.

Please do not enter personal data here; we would otherwise delete it soon. (See also our privacy declaration.) If you wish to receive personal feedback or consultancy from the author, please contact him, e.g. via e-mail.

Spam check:

By submitting the information, you give your consent to the potential publication of your inputs on our website according to our rules. (If you later retract your consent, we will delete those inputs.) As your inputs are first reviewed by the author, they may be published with some delay.


Share this with your friends and colleagues, e.g. via social media:

These sharing buttons are implemented in a privacy-friendly way!