Selecting the Appropriate Mirror for your System

High-performance optical mirrors are a key component to most system designs.  From bench-top setups to high-end lab instrumentation, mirrors can enable a reduced system footprint, combine multiple sources, help to form images, or other purposes.  This article provides a high-level, basic overview of some of the key specification parameters to consider when selecting the appropriate mirror for your optical system.

Light Source

Most mirrors used in high-performance instrumentation or laboratory setups will typically involve a laser (or multiple lasers) bring present.  Focusing on applications where the laser wavelength falls in the near-UV, visible, and near-IR spectral region, there are many common mirror requirements, such as tight cosmetic, polarization, laser induced damage threshold (LIDT), and flatness performance levels.

In addition to lasers, some applications require the mirror to direct or focus broadband light.  These applications can see an overlap in some of the laser mirror requirements, such as cosmetic performance, but additional requirements, such as a wide bandwidth, become important as laser mirrors typically have a narrow bandwidth requirement due to the laser’s single wavelength of operation.

Angle of Incidence (AOI) & Polarization

Sometimes overlooked, the angle of incidence is an important specification consideration.  Mirrors designed for 0 degrees AOI will typically hold specification for shallow angles (typically up to ~10 degrees) but as one moves to a high off-angles, such as up to 45 degrees (needed for making 90 degree turns), performance can degrade significantly.  The reflection region of the mirror will shift to a shorter wavelength range, the bandwidth will narrow, and the polarization performance will suffer.

When going to higher angles, p-polarized light exhibits poorer reflection performance and has an inherent narrower bandwidth than s-polarized light.  Usually for laser mirrors this narrower reflection region is acceptable as only a small region is required to reflect the laser, but if the mirror is used under non-optimized conditions, or if a tunable laser is present, having a mirror with a wide reflection region is ideal to simplify the system setup or require fewer component swaps when switching wavelengths.

Wavelength & Reflectivity

When designing the bandwidth the mirror is going to reflect over, typically the wider the better as this will require fewer mirrors to accomplish the job.  Increasing the bandwidth is non-trivial, as it requires many additional layers of coating material, leading to a much more complicated design to optimize & manufacture.  In addition to bandwidth, the reflection level should be “high”, where > 95% and even upwards of > 98% is acceptable in most applications.  But, in certain high-end applications, like creating a laser cavity or performing cavity ring-down spectroscopy (CRDS) measurements, mirrors with reflectivity of > 99% aren’t good enough.  Reflection levels of > 99.95% and even > 99.999% can be required to provide the necessary system performance.  When calculating reflection (R) at this level, all losses need to be considered, including transmission (T), absorption (A), and scatter (S), such that:

As one can see, to get 99.999% reflectivity, this requires each component to be controlled down to the parts-per-million level.  Performance to this level is accomplished by proper mirror design to minimize transmission (typically at the expense of bandwidth), careful control of the deposition of the coating to minimize absorption, and a super-polished substrate to minimize scatter sites.

Flatness & Curvature

Most mirrors in the industry are manufactured on a flat substrate, typically 6mm and up to ¼” thick.  This thickness allows for the fabrication of the glass blanks to meet pre-coating flatness levels of λ/10 to λ/20 per inch at 633nm.  For most mirrors, this flatness value is typically held through the coating process.  Some applications relating to laser cavities or CRDS may require there to be a radius of curvature, say from 50cm to 100 cm, and if imaging is being done, a customized curvature would need to be specified to provide the proper focus.

Cosmetics & Laser Induced Damage Threshold (LIDT)

Most laser systems require 10-5 Scratch-Dig cosmetic specifications.  This cosmetic performance is required as there can be a reduction in LIDT performance, additional system scatter can occur, and wavefront distortion could be introduced.  But tight cosmetic performance down to this level (or beyond) can result in increased cost of the mirror, due to additional inspection and associated yield fall-out during manufacture of the mirrors.

Laser induced damage threshold is strongly dependent on many factors, including the power, beam size, pulse duration & repetition rate of the laser.  For more information on how these parameters relate to LIDT and how to convert between pulsed, CW, and quasi-CW lasers, refer to the Semrock tech note on laser damage.

Dispersion

If you have an ultrafast laser (those with laser pulses > 1 picosecond), pulse distortion can be introduced by the mirror’s coating, causing undesired pulse broadening, which reduces performance.  As pulse widths become shorter and shorter, additional constraints on the mirror design must be considered, including the Group Delay (GD), Group Delay Dispersion (GDD) and Third-order Dispersion (TOD) performance.  As additional coating layers and coating thickness can result in poorer dispersion performance, optimizing for dispersion can limit the bandwidth and reflectivity levels of the mirror.  Coating technology can also drive performance of dispersion vs. these other parameters, as explained in IBS Coatings for Ultrafast Lasers and Applications.  For most applications, ultrafast lasers are not present, so dispersion performance of the mirror does not impact one’s choice in selecting the appropriate mirror.

Cost

With all the parameters to consider when making a mirror, a designer can look to focus on a subset of parameters to simplify the design of the mirror, and thus minimizes the cost to produce the mirror.  Common examples of this include limiting parameters to only consider a single wavelength (narrow bandwidth), or a moderately high reflectivity level, a moderate LIDT, or even specifying flatness “pre-coating” to avoid needing to compensate for any power introduced by the coating.  While this can impact the mirror’s performance outside the specified ranges, this allows for a mirror to be produced at a much lower cost, and in some instances the mirror can be considered “

As coating technology has improved and understanding of how highly-durable thin-film coatings can be used to address the cost-performance trade-off, mirror designs can be modified to include many more parameters while maintaining a reasonable price point.  Even with these advances in design & coating technology, some parameters such as ultra-high reflectivity levels or tight dispersion performance can still drive costs higher.  But, with this higher cost comes a higher level of performance, allowing applications which previously were not achievable.

Conclusion

A seemingly simple component most of us use at some point early every day, a mirror used in a high-performance system requires a significant number of considerations to ensure proper selection.  While no single mirror can cover every single parameter noted above, continued advances in coating design & processing can allow for mirrors to work over a larger number of parameters, allowing the researcher or system designer additional flexibility for their task at hand.