Software … powerful tools for your research & development!

RP Fiber Power – Simulation and Design Software for Fiber Optics, Amplifiers and Fiber Lasers

Case Study: 975-nm Fiber Lasers

Definition of Task

We want to explore how one can realize ytterbium-doped fiber lasers emitting at 975 nm. This laser wavelength is quite special: while we have the maximum emission cross-section of ytterbium there, we also have strong reabsorption:

cross-sections of Yb-doped glass
Figure 1: Absorption and emission cross-sections of ytterbium-doped germanosilicate glass, as typically used in the cores of ytterbium-doped fibers (data from spectroscopic measurements by R. Paschotta, Ref. [1]).

Therefore, we get positive gain only if the Yb excitation density is over 50% (strong three-level characteristics). That has important implications for the laser design, which turn out to be substantially more problematic for designs based on double-clad fibers.

We conveniently use the Power Form for continuous-wave laser and amplifiers. We just use only the laser part and leave the amplifier stages deactivated.

Core-pumped Laser

The easy part is that for a normal core-pumped single-mode fiber. We select the commercial fiber Yb 103 from CorActive, enter some end reflectance values (10% and 99%) to define the laser resonator and configure a pump input with 250 mW at 940 nm. Note that the pump wavelength has to be shorter than the laser wavelength, of course; only that way, we can get sufficiently high Yb excitation to get positive gain at the laser wavelength.

We find that it works well e.g. with 3 m fiber length, e.g. inspecting how the optical powers and excitation density evolve in the fiber:

Powers vs. position in the Yb-doped fiber
Figure 2: Optical powers and excitation density vs. position in the fiber for a simple fiber laser setup, pumped at 940 nm.

We can use a variation diagram to see the influence of the fiber length:

Laser output power vs. fiber length
Figure 3: The laser output power as a function of fiber length for pumping at 940 nm.

We see that 3 m length is only slightly less than ideal, and that the achieved power efficiency is reasonable, although not perfect as even for the optimum length pump absorption is somewhat incomplete. Essentially, the problem is that although you could get more pump absorption by increasing the fiber length, that would also introduce more reabsorption at the laser wavelength. In effect, more ytterbium needs to be kept in the excited state, and that leads to increasing loss through spontaneous emission, which also implies a higher laser threshold pump power.

It works somewhat better for pumping at 920 nm, where the absorption is stronger (see Figure 1) – despite the higher quantum defect:

Laser output power vs. fiber length
Figure 4: The laser output power as a function of fiber length for pumping at 920 nm.

In any case, amplified spontaneous emission remains very weak, despite the low output coupler reflectance. By the way, the model does not have any reflectance for ASE at the output end, only at the back mirror. (That is realistic, for example, if output coupling is done with a fiber Bragg grating.) We can also inspect the ASE output spectra and the gain spectrum:

ASE output spectra and gain spectrum
Figure 5: ASE output spectra and gain spectrum of the 975-nm fiber laser.

We can see that the gain around 1030 nm is a bit higher than the laser gain, but not enough to cause excessive ASE output.

It would be easy to optimize the output coupler reflectivity as well, but that would not have a strong impact in this case.

In conclusion, we see that it is not difficult to get such a fiber laser designed with reasonable power efficiency. Only, the efficiency is slightly lower than in cases with longer laser wavelength, strongly reducing the reabsorption.

Cladding-pumped Laser

Now we select a double-clad fiber – the DCF-YB-6/128 from CorActive –, trying to realize a 975-nm fiber laser with substantially higher pump power. Let us assume that we now pump with 50 W at 920 nm.

We then quickly find complete failure of our attempt. When trying with a relatively short fiber length of 3 m, we find quite incomplete pump absorption, and only get strong forward ASE but no laser output (being below the laser threshold):

double-clad fiber laser with 3 m length
Figure 6: Double-clad fiber laser with 3 m length.

So we see that ASE is now a big challenge. So far, we still assumed the full 99% reflectance of the back mirror for ASE, but we can change this, limiting that reflectance to a region with 0.1 nm bandwidth around 975 nm. That indeed changes a lot, but without solving the problem:

double-clad fiber laser with 3 m length
Figure 7: Double-clad fiber laser with 3 m length.

Now, we get strong forward and backward ASE, and can again not reach the laser threshold.

If we increase the fiber length to 10 m, we get better pump absorption, but only more efficiently generate ASE:

double-clad fiber laser with 10 m length
Figure 8: Double-clad fiber laser with 10 m length, now with narrow-band reflection at the pump input end.

In this case, we have ASE basically only at long wavelengths beginning at 1030 nm:

ASE output spectra of 10 m long laser
Figure 9: ASE output spectra of the 10 m long laser. ASE is now basically only at long wavelengths.

The problem can be explained as follows:

  • In order to reach sufficiently strong pump absorption (despite the small spatial overlap of pump light and the doped core in the double-clad fiber), we need to have quite a lot of ytterbium in our fiber (using a much longer length than required for core pumping).
  • As a consequence, when ramping up the ytterbium excitation, we soon get rather strong gain in the 1030-nm region, and consequently strong ASE – well before we get positive gain at 975 nm as required for lasing.

As a result, we really cannot realize such a laser with that type of double-clad fiber.

Using a Ring-doped Fiber

However, there is a solution based on a modified fiber design with ring doping [2]:

  • We do not place the Yb doping in the fiber core as usual, but in a ring just around the fiber core. The dopant is thus placed in the outer wing of the laser intensity profile.
  • While the overlap of the pump light with the Yb doping remains small, we now also get a reduced overlap of the laser light with the core.
  • As a result, we can have a higher degree of Yb excitation without getting excessive gain beyond 1030 nm wavelength.

The challenge is just to procure such a fiber; there is probably no commercial supplier offering it, as there may not be a substantial demand. If we wanted it, we could at least work out a fiber design with RP Fiber Power and hopefully find someone producing this as a custom fiber. Or you make it yourself, if you happen to have a fiber drawing tower and everything else needed for fiber fabrication.

In our demonstration, we start with parameters of the previously tried ordinary double-clad fibers and make a data file for a modified ring-doped variant. Here, we assume constant Yb doping density within a ring with radial coordinates from 4.5 μm to 5 μm. That turns out to work reasonably well with 20 m fiber length:

powers vs. position for ring-doped fiber
Figure 10: Powers vs. position in the ring-doped fiber.

Here, ASE is now largely suppressed. The power efficiency is still not perfect; we have substantial power losses due to two factors:

  • Pump absorption is reasonable, but not perfectly efficient. That aspect could be optimized somewhat further.
  • We have a relatively high laser threshold power since we have a lot of ytterbium in the system, which we need to keep at an excitation above 50%. That causes a lot of spontaneous emission, i.e., a high threshold pump power. That problem is harder to solve; essentially, it results from the weak overlap between pump light and dopant – in the end, from the large diameter of the pump cladding (here: 128 μm). A smaller pump cladding could be realized either by just making the glass fiber thinner (which may not be practical) or by introducing a depressed index ring around the pump cladding. In any case, you then also need a pump source with higher beam quality.

Conclusions

You can learn various things from this demonstration:

  • The modeling of fiber lasers is quite straightforward with the RP Fiber Power software: simply enter its parameters and inspect the resulting numerical and graphical outputs. You can rapidly explore parameters and configurations. The Power Form gives you plenty of freedom, e.g. for using multiple stages, multiple signals and pump sources, filters between the stages, etc., besides offering many useful diagrams.
  • Even in seemingly simple cases, fiber amplifiers can exhibit quite unexpected behavior, which can have drastic effects on the amplifier performance. With a simulation model, you identify such effects quickly and have the best chances to find out how things really work. (If you have difficulties, we will help you within our diligent support!)
  • With such a model, you can quickly solve the problem by optimizing your design. If you would instead order the parts, build the amplifier, test it and try to analyze and solve the issues there, this would be far more tedious, costly and time-consuming!

Bibliography

[1]R. Paschotta et al., “Ytterbium-doped fiber amplifiers”, IEEE J. Quantum Electron. 33 (7), 1049 (1997); https://doi.org/10.1109/3.594865
[2]J. Nilsson, J. D. Minelly, R. Paschotta, A. C. Tropper, and D. C. Hanna, “Ring-doped cladding-pumped single-mode three-level fiber laser”, Opt. Lett. 23 (5), 355 (1998); https://doi.org/10.1364/OL.23.000355