Failures in quartz and crystal glass used in quantum photonics tend to follow a small number of patterns, and each pattern has a signature. A part can pass every dimensional check and still destroy a single-photon measurement, because the quantities that matter here are measured in counts per second rather than in millimetres. This article walks through the common failure modes, the measurements that reveal them, and the practical fixes that keep a quantum optical bench reproducible.
Optical Quartz and Crystal Glass in Quantum Photonics
In quantum photonics, that translates into a specific set of requirements rather than a general preference for quality. A squeezed-light experiment, a spontaneous parametric down-conversion source or a trapped-ion readout chain all work at the level of individual photons, so any stray light emitted by the optic itself competes directly with the signal. Bulk fluorescence from fused silica, scatter at a subsurface damage layer, stress-induced birefringence that rotates a polarisation state, and absorption that heats the part are all first-order problems rather than second-order ones. The specification is therefore written around loss and background: transmission at the working wavelength, fluorescence background under a defined pump, polarisation extinction through the part, and the wavefront error that the beam actually sees.
Inspection and measurement
Standard dimensional inspection does not catch any of these failures. Fluorescence background is measured by pumping the optic at the intended wavelength and counting emitted photons through a band-pass filter that rejects the pump; UV-grade fused silica typically sits one to two orders of magnitude below standard grades in this measurement, which is why the material grade matters more than the polish. Scatter is quantified with an integrating sphere or a scanning scatterometer, and the result is dominated by subsurface damage rather than by the final surface finish, so a part that looks clean under a microscope can still scatter. Stress birefringence is measured between crossed polarisers or with a polarimeter, and it is reported as a retardance in nanometres per centimetre rather than as a pass or fail. Wavefront error is measured interferometrically at the working wavelength, because a part that meets a visible-wavelength specification frequently does not meet the same specification in the near infrared.
Coating and deposition considerations
Every coating in a single-photon path has to be specified in terms of loss and background, not only reflectance. Absorption in the coating adds a thermal load and a noise floor, and ion-assisted deposition with dense oxide layers typically produces lower total loss than electron-beam evaporation, at the cost of a longer cycle. Where the stack sits at normal incidence, the residual reflection can be brought below a tenth of a percent, but that number is quoted per surface, so a window with two coated faces still returns about two tenths of a percent of the beam. Fluorescence from the coating itself is a separate specification and should be requested explicitly, because a stack that is optically correct can still glow under the pump. Where the optic is a crystal rather than fused silica, the deposition temperature has to stay below the point at which the material begins to lose its optical quality.
Design rules that reduce cost
Keeping the design simple reduces both cost and background. Use a single flat window wherever the optical function allows, because every additional surface adds scatter, reflection and a cleaning step. Choose the material grade from the fluorescence requirement rather than from habit: standard fused silica is cheaper, but if the pump wavelength sits in the ultraviolet the grade is not negotiable. Keep coatings to one face when the second face can be left uncoated without breaking the budget of reflected light, and avoid specifying both a hard anti-reflective coating and a separate hydrophobic layer unless the environment genuinely requires the hydrophobic finish. Finally, define the wavefront requirement at the working wavelength and to the beam diameter actually used, because specifying a full-aperture figure that the beam never touches adds polishing time without adding performance.
Common failure modes and how they show up
The first pattern is a rising dark-count floor that tracks the pump power: the optic is fluorescing, and the fix is a higher material grade or a different pump geometry. The second is a loss of polarisation extinction after the part is mounted: the optic is being stressed by its own holder, and the fix is a compliant mount or a thicker part rather than a better polish. The third is scatter that appears after cleaning: aggressive solvents and wiping pull material out of the subsurface layer, and the fix is a validated cleaning procedure with a defined solvent and no dry wiping. The fourth is drift in transmission over hours: the coating or the bulk is absorbing, and the fix is a lower-loss stack or improved thermal contact to the mount. Each of these leaves a different signature in the data, which is why the measurement plan matters as much as the specification.
Requirements specific to Quantum Photonics
Quantum experiments add constraints that ordinary optical specifications do not cover. The bench is usually pumped continuously for days, so long-term drift and thermal stability matter as much as initial performance. Polarisation is frequently the information carrier, which makes stress birefringence a specification rather than a footnote. Because the signal is counted rather than observed, background from the optic competes directly with the measurement, and a part that is acceptable in an imaging system can be unusable here. Documenting the material grade, the fluorescence measurement and the coating loss on the first article is therefore normal practice rather than an extra request.
- Fluorescence background measured under the actual pump wavelength and power
- Stress birefringence quoted as retardance, not as a visual pass or fail
- Coating loss and coating fluorescence specified separately from reflectance
- Wavefront error defined at the working wavelength and beam diameter
ISO 10110 drawings, laser-induced damage threshold testing per ISO 21254 where relevant, and material certificates stating the fused silica grade and hydroxyl content.
Selection data at a glance
| Parameter | Typical value | Notes |
|---|---|---|
| Transmission | over 93 % at 1064 nm uncoated, over 99 % coated | Grade dependent |
| Fluorescence background | UV-grade 10-100x below standard grade | Under the working pump |
| Stress birefringence | ≤ 2 nm/cm standard, ≤ 0.5 nm/cm on request | Measured by polarimeter |
| Surface quality | 20-10 scratch-dig achievable | Per ISO 10110-7 |
| Wavefront error | λ/4 standard, λ/10 on request | At the working wavelength |
| Subsurface damage | ≤ 5 µm achievable with controlled polishing | Drives scatter |
| Coating loss | ≤ 0.1 % per surface, ion-assisted deposition | Absorption and scatter included |
| Flatness | λ/4 per 25 mm over the clear aperture | Beam-limited |
| Material | UV-grade fused silica, quartz crystal, calcium fluoride | Certificate with the batch |
Frequently asked questions
Why does the same optic behave differently in a quantum setup?
Because the measurement is counting photons rather than looking at an image. A part with a perfectly acceptable transmission can still emit enough fluorescence to raise the dark-count floor, and that background scales with pump power. The material grade and the coating chemistry therefore have to be chosen against the fluorescence specification, not against transmission alone.
Can cleaning damage the optic?
Yes, and it is one of the most common causes of rising scatter in service. Aggressive solvents and dry wiping pull material out of the subsurface damage layer and leave the surface rougher than the polish produced. A validated procedure with a defined solvent, controlled wiping and a documented drying step prevents most of it.
How do I stop polarisation from drifting after mounting?
Treat the mount as part of the optical design. Stress birefringence introduced by a rigid holder can exceed the intrinsic birefringence of the material several times over. A compliant mount, a thicker window, or a deliberate strain-relief geometry usually solves it more cheaply than specifying a lower-birefringence material.
Need a quotation for optical quartz and crystal glass?
Send your drawing, target specification or coating requirement and an optical engineer will reply within 24 hours with a price and a lead time.
Request a Quote Email an engineer