The purpose of this presentation is to describe an optical set up developed to measure axial stress in optical fibers and all fiber devices. It is believed that knowledge and control of residual stresses will impact the future development of all fiber components. Single mode fibers used in optical devices usually have two vectorial propagations modes. A stressed induced non-uniform index distribution will affect the following parameters: polarization mode dispersion, polarization dependant loss and mode coupling ratio in optical devices. Those effects have to be well characterised due to recent increases in bit rate of optical transmission systems requiring better component performances. In this paper, we will explain how our measurement system works and show some preliminary results.
A CW kilowatt fiber laser numerical model has been developed taking into account intracavity stimulated Raman scattering (SRS). It uses the split-step Fourier method which is applied iteratively over several cavity round trips. The gain distribution is re-evaluated after each iteration with a standard CW model using an effective FBG reflectivity that quantifies the non-linear spectral leakage. This model explains why spectrally narrow output couplers produce more SRS than wider FBGs, as recently reported by other authors, and constitute a powerful tool to design optimized and innovative fiber components to push back the onset of SRS for a given fiber core diameter.
A new method is presented for the analysis of the modal content of a beam travelling in a waveguide. This method uses a
simple optical set up to record beam images. Depending on the application, the source can be broad band (BBS) or a
tunable laser. The method uses the eigenmode profiles of the waveguide under test, either theoretical or experimental
ones. In this case, the technique is applied to characterize the modal content of few moded large mode area (LMA)
fibers. Such LMA fibers are typically used in high power fiber lasers and amplifiers to reduce sensitivity to non-linear
effects. By calculating the scalar products of the unfolded experimental and theoretical 2D profiles, the modal content is
obtained. Access to such cost effective and easy to implement diagnosis tool will greatly help improving modal quality
preservation in components and systems based on the fundamental mode operation of few moded LMA fibers. The high
precision and performance of the method is evaluated using both computer generated and experimental data sets.
As overall power increases in fiber lasers and amplifiers, the amount of optical power which must be dealt
with in order to obtain high core to core and core to cladding isolation also increases. This unwanted light
can represent hundreds of watts and must be managed adequately. By combining a proper termination (end
cap) design and cladding stripping techniques it is possible to obtain a robust output beam delivery
component. The cladding stripping techniques are inspired by previous work done on high power cladding
strippers. All measurement presented here are done with a flat end cap. Both core to core and core to
cladding isolation will be better with an angled end cap. A core-to-core isolation of over 25dB was
measured, while core to cladding was over 30dB. Power handling was characterized by the capability of
the device to handle optical power loss, rather than transmitted power. The component dissipated over 50
watts of optical power due to isolation. The above results show that understanding the mechanisms of
optical loss for forward and backward propagating light in a end cap and the heat load that these losses
generate is the key to deliver kilowatts of optical power and protect the integrity of the system.
High-power combiner designs (such as kilowatt-class combiners and beyond) are increasingly aggressive on brightness
conservation in order to reduce the brightness loss of the pumps as much as possible in both direct diode combining
and pump and signal coupling, especially with the advent of next-generation high-power pumps. Since most of the pump
loss is due to brightness loss across the combiner, tighter designs (close to the brightness limit) are considerably more
sensitive to variations in the input power distribution as a function of numerical aperture; for instance, next-generation,
high-power multi-emitter pumps are likely to have larger numerical apertures than conventional single-emitter diodes. As
a consequence, pump insertion loss for a given combiner design sitting close to the brightness limit should be dependant
on the input power distribution. Aside from presenting a manufacturing challenge, high brightness combiners also imply
more sophisticated testing to allow a deeper understanding of the loss with respect to the far-field distribution of the pump
inputs and thus enable the extrapolation of loss for an arbitrary, cylindrically symmetric radiant intensity distribution. In
this paper, we present a novel test method to measure loss as a function of numerical aperture (NA) fill factor using a
variable NA source with square-shaped far field distributions. Results are presented for a range of combiners, such as 7x1
and 19x1 pump combiners, with different brightness ratio and fiber inputs. Combiners violating the brightness conservation
equation are also characterized in order to estimate the loss as a function of input power vs. NA distribution and fill factor.
Most of the current large mode area (LMA) fibers are few-moded designs using a large, low numerical aperture (N.A.) core, which promotes mode coupling between core modes and increases bending losses (coupling with claddingmodes), which is undesirable both in terms ofmode area and beamquality. Furthermore, short LMA fiber lengths and small cladding diameters are needed to minimize nonlinear effects and maximize pump absorption respectively in high-power pulsed laser systems. Although gain fiber coiling is a widely used technique to filter-out unwanted modes in LMA fibers, coupling between modes can still occur in component leads and relay fibers. In relay fiber, light coupled into higher-order modes can subsequently be lost in the coiling or continue as higher-order modes, which has the overall effect of reducing the effective transmission of the LP01 mode and degrading the beam quality. However, maximum transmission of the LP01 mode is often required in order to have the best possible beam quality (minimal M2). Launching in an LMA fiber with a mode field adapter (MFA)1 provides an excellent way of ensuring maximum LP01 coupling, but preservation of this mode requires highmodal stability in the output fiber. Small cladding, low N.A. LMA fibers have the disadvantage of being extremely sensitive to external forces in real-life applications, which is unwanted for systems where highly sensitive mode coupling can occur. In this paper, we present a detailed experimental and theoretical analysis of mode coupling sensitivity in LMA fibers as a function of fiber parameters such as N.A., core diameter and cladding diameter. Furthermore, we present the impact of higher N.A. as a solution to increase mode stability in terms of its effect on peak power, effective mode area and coupling efficiency.
The ability to strip cladding light from double clad fiber (DCF) fibers is required for many different reasons, one example is to strip unwanted cladding light in fiber lasers and amplifiers. When removing residual pump light for example, this light is characterized by a large numerical aperture distribution and can reach power levels into the hundreds of watts. By locally changing the numerical aperture (N.A.) of the light to be stripped, it is possible to achieve
significant attenuation even for the low N.A. rays such as escaped core modes in the same device. In order to test the power-handling capability of this device, one hundred watts of pump and signal light is launched from a tapered fusedbundle (TFB) 6+1x1 combiner into a high power-cladding stripper. In this case, the fiber used in the cladding stripper and the output fiber of the TFB was a 20/400 0.06/0.46 N.A. double clad fiber. Attenuation of over 20dB in the cladding was measured without signal loss. By spreading out the heat load generated by the unwanted light that is stripped, the
package remained safely below the maximum operating temperature internally and externally. This is achieved by
uniformly stripping the energy along the length of the fiber within the stripper. Different adhesive and heat sinking
techniques are used to achieve this uniform removal of the light. This suggests that these cladding strippers can be used
to strip hundreds of watts of light in high power fiber lasers and amplifiers.
We present an all-fiber monolithically integrated fiber laser based on a custom tapered fused bundle pump combiner
with 32 inputs ports connected to a double clad gain fiber. The pump combiner is designed to provide high isolation
between signal and pumps fibers providing intrinsic pump protection. This configuration can generate more than 100W
of continuous wave (CW) laser light using single-chip multimode pumps enabling long term reliability.
In order to test power-handling at 1kW, a special splitter component had to be developed to make use of available
sources. A tapered fused-bundle (TFB) 1X7 splitter using a 1.00mm core diameter 0.22NA input fiber coupled to seven
400 micron core 0.22 NA output fibers was tested up to 860W at 976nm. Surface temperature rise was measured to be
less than 15°C with active heat sinking. The above results suggest that understanding the mechanisms of optical loss for
forward and backward propagating light in a TFB and the heat load that these losses generate is the key to producing
multi kW components, and demonstrates that reliable kW-level all fiber devices are possible.
Light absorption in structural adhesives constitutes the main source of heat in tapered fused bundle (TFB) devices.
Efficient heat dissipation solutions were developed by studying these thermal loads. The relative merits of transparent
vs. opaque package designs were established experimentally. In the former, light escapes without being absorbed by the
package walls, whereas in the latter, the spurious optical signal is directly absorbed and dissipated. The fact that heat is
generated directly in the adhesive largely favors the opaque package, which offers more efficient heat extraction. By
using a thermally conductive package, a temperature rise of 1.1°C per Watt of dissipated power was measured. These
numbers demonstrate that passive heat sinking at 20°C is sufficient to allow reliable operation up to 45Watts of
dissipated power (1kW with 0.2dB optical loss) without compromising long-term reliability.
Fiber lasers have shown extraordinary progress in power level, reaching the kilowatt range. These results were achieved with large mode area fibers pumped with high power laser diodes coupled with bulk-optics. To enable the commercial development of these high power fiber lasers, we have demonstrated several All-Fiber components, which replace the bulk-optic interface in the present laser configurations. These components include multimode fused fiber bundle combiners with or without signal fiber feed-through, Bragg gratings and mode field adaptors. The multimode fibers are used to couple several fiber pigtailed pump diodes to a double-clad fiber. Such combiners may contain a signal fiber to provide an input or output for the core modes of the double-clad fiber. Mode field adaptors perform fundamental mode matching between different core fibers. Bragg gratings are used as reflectors for the laser cavity. These components exhibit low-loss and high power handling of 200 Watts has been demonstrated. They enable the design of true high power single-mode All-Fiber lasers that will be small, rugged and reliable.