Hello everyone, I would like some collective brainstorming on my setup.
I have a laser providing a collimated, randomly polarized beam at 1064 nm and I run it at about 150 mW of power. I focus the beam with an aspheric lens onto the SMF-28-100 fiber. The fiber is fixed in an FC/PC holder mounted on an XYZ translational stage with 2 rotational knobs. I struggle to obtain a stable Gaussian output.
1. Are there any fibers with cut-off wavelength close to 1064 nm? It looks like at least Throlabs and Newport have cut-off wavelengths at 970-980 nm and then at 1300-1500 nm.
2. How much effect would alignment have? The stage is a little bit unstable and may move when I turn one of the knob, so I have to use the rotational knobs until the power is maximized. I am even thinking of ordering a custom metal holder to mount both lens and the FC/PC connector ring on it.
3. The fiber is about 1 m long and mostly covered with the acrylate coating so I don't think I should worry about the cladding modes.
4. How efficient would be folding the fiber into '8's to achieve mode mixing and more of a flat beam profile?
Thank you in advance!
Elena,
As Vincent Lecoeuche points out, at 1064 nm SMF 28 supports both the fundamental LP01 (HE11) mode and the higher order LP11 (HE21+TM01+TE01) modes. It will be difficult to achieve either a stable Gaussian mode, or a flat mode profile.
A stable and flat mode profile requires more than the six LP01 and LP11 modes. Regarding mode coupling, I suspect you will need rather severe bends to couple power between LP01 and LP11. On the other hand, small perturbations will be sufficient to change the relative optical phase delays and modify the near and far field intensity distributions at the fibre output.
For a stable Gaussian profile you must minimise the proportion of higher order LP11 modes present. If your laser source projects a circularly symmetric Gaussian spot onto the fibre core, fundamental LP01 mode power is maximised when the spot is centred on the core.
On the other hand, the LP11 mode is anti-symmetric, and an accurately centred excitation spot will not couple significant power. LP11 power will increase with an off-centre or tilted launch spot.
As the launch spot moves off-centre, fundamental LP01 power decreases, but higher order LP11 mode power increases. I have not attempted to quantify these opposing trends, but you may find that simply maximising the total coupled power is not the best way to achieve an intensity distribution which is close to Gaussian.
The specification of Thorlabs HI1060 and 1060XP fibres is almost identical to Corning HI1060, so that is probably where they procure it. It seems a good choice if you want a stable and more or less Gaussian mode profile. If you need less than 1 km lengths, Thorlabs may be a more convenient source.
Corning HI980 and Thorlabs 980HP fibre have a higher numerical aperture and smaller core size to maintain good bend performance at 1550 nm telecommunication wavelengths, while remaining single mode at 980 nm pump wavelengths used for erbium fibre amplifiers. It is probably easier to couple to HI1060 than to these, unless you have a particularly small launch spot. 780HP fibre fibres should work too, but will have higher bend loss at 1064 nm than the 980 or 1060 fibres.
Alan Robinson I expect my focused Gaussian to couple into the LP01 with the highest efficiency of all modes (I calculated 7 from V^2/2 and specified NA). But it looks like HI1060 may be a good choice. My fibers are only 1-1.5 m long. I also had to ask if the acrylate coating on SMF-28-100 is good at removing the cladding modes. This may not be the case as my wavelength is less than 1300-1500 nm where the fiber should behave as a single mode.
Elena Renzhiglova - I agree that LP01 will couple from a Gaussian beam with highest efficiency into the available modes. In fact, if the beam has a circularly symmetric Gaussian profile, there will be minimal excitation of the LP11 modes when the launch spot is accurately centred onto the fibre core.
My concern was that a small offset of the launch spot would decrease the LP01 power, but would increase the proportion of LP11 modes, so that simply looking at total launch power may not be the most accurate way to assess mode purity.
How did you calculate the V-number? Corning specify a nominal core diameter of 8.2 µm for SM28 Ultra, and a numerical aperture of 0.14 (measured as 1% far field intensity in a 1-D scan). V=3.39 at 1064 nm, so V²/2 = 5.7.
Strictly speaking, we should specify NA in terms of refractive index difference between core and cladding, rather than radiation angle (which varies with wavelength). For their earlier SMF 28e+ fibre, Corning specified nominal relative index difference, Δ=0.0036, so NA=0.124 and V=3 at 1064 nm.
In any case, the formula for the number of bound modes, V²/2, is an approximation which is most useful for larger V. At low V, it is simpler to count the number of modes directly.
- We have two orthogonally polarised fundamental modes.
- In addition there are four LP11 modes - two polarisations for each of the sine and cosine azimuthal variants. More accurately, we have TE01, TM01 and 2x HE21 hybrid modes, but in the weakly guided approximation, the LP11 modes are a linear combinations of the 4 vector modes. The LP modes are usually a more convenient description when the launch spot is linearly polarised, or has a uniform elliptical polarisation state.
If we ignore cladding modes, then the theoretical number of bound modes in a step index fibre is 6 for V numbers between 2.405 and 3.832. At shorter wavelengths (V>3.83), the four LP21 (HE31, EH11) modes are also supported.
https://www.thorlabs.com/thorproduct.cfm?partnumber=smf-28-100
https://www.corning.com/media/worldwide/coc/documents/Fiber/SMF-28%20Ultra.pdf
We can ignore most of this if you decide to use a fibre such as HI1060, which is single mode at 1064 nm. This would be my preferred solution.
I do not anticipate significant cladding mode propagation over 1 m of acrylate coated fibre. As Vincent Lecoeuche suggests, if cladding light is present, then a few loose bends in the fibre should couple to higher order cladding modes so they are more effectively absorbed by the cladding, without perturbing the bound modes (unless they are close to cut-off).
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials Chemistry
- Polymer Chemistry