A femtosecond laser amplifier puts 150-250 W of waste heat into a crystal the size of a fingernail, and every degree of that crystal's temperature shows up in the beam: focus shift, depolarisation, mode instability, lost output. The interface between crystal and heat sink is therefore not a packaging detail - it is part of the optical design. Here is how the choice actually resolves, including when a graphene pad is the wrong answer.
Why the crystal interface is brutal
- Flux, not total power, sets the problem. 250 W over a 10 × 10 mm face is 250 W/cm². At that flux the interface resistance is multiplied by the flux to give temperature drop - so 0.01 °C·cm²/W of interface is worth 2.5 °C.
- The crystal is the weakest mechanical part in the stack. It cannot be clamped hard, it must not be point-loaded, and it is usually the most expensive component in the system. Any mounting scheme has to be both thermally tight and mechanically gentle.
- Thermal lensing is a feedback loop. dn/dT turns a temperature gradient into a lens; the lens changes mode size, which changes the pumped volume, which changes the gradient. Small ΔT differences can flip a resonator from stable to marginal.
- Nothing may contaminate the optical path. Flux, solder, and silicone bleed are all unacceptable near coated surfaces.
The incumbent: indium solder or indium foil
Indium is used because it is genuinely excellent: soft enough to conform at modest pressure, high conductivity, and - when soldered - a metallic bond with essentially no interface resistance. Typical interface impedance lands around 0.006-0.012 °C·cm²/W. At 250 W/cm² that is roughly 1.5-3 °C of interface drop.
Its costs are real too:
- Process. Soldering a crystal means flux, a thermal cycle on an expensive optic, and a rework path that can destroy it.
- Creep and pump-out. Indium is soft in service as well as at assembly; under thermal cycling it can extrude and thin out at the edges.
- Replacement. A soldered crystal is not a field-serviceable part. Development iterations get expensive.
- Stress. The bond constrains the crystal; differential expansion is transferred straight into the optic.
What a graphene pad delivers - honestly
Our measured joint resistance for the 180 W/m·K pad, at real clamping pressures and 80 °C:
| Pressure (psi) | R 0.15 mm | R 0.20 mm | R 0.30 mm |
|---|---|---|---|
| 10 | 0.088 | 0.109 | 0.079 |
| 20 | 0.054 | 0.058 | 0.054 |
| 30 | 0.045 | 0.045 | 0.049 |
| 50 | 0.038 | 0.039 | 0.042 |
In °C·cm²/W. Compare that against the indium range of 0.006-0.012: the pad is roughly 3-8x the interface resistance. At 250 W/cm² that is 10-12 °C of interface drop instead of 1.5-3 °C. On a high-power amplifier with a tight wavefront budget, that difference is not acceptable - and no amount of salesmanship changes it.
So be precise about what the pad is good for:
| Scenario | Indium | Graphene pad |
|---|---|---|
| Highest-power oscillator, tight wavefront spec | Best choice | ΔT too large - not recommended |
| Prototype / iteration-heavy build | Slow, destructive rework | Re-mountable, replaceable, no flux |
| Crystal that must not see process heat | Requires solder cycle | No thermal process at all |
| Field-serviceable or modular head | Not serviceable | Pad replaced in minutes |
| Cycled hard, risk of indium creep | Extrudes over life | Solid, no pump-out (300-cycle data) |
| Medium power, relaxed optical spec | Overkill | Adequate and cheaper to assemble |
Mounting a pad under a crystal: the details that decide it
- Uniform pressure, low magnitude. The knee of our measured curve is at 20 psi; a crystal mount rarely wants more than 15-25 psi anyway, because stress birefringence grows with load. Design for the knee, not for maximum torque.
- Use the thin end of the range. 0.15 / 0.20 mm pads exist for exactly this: the interface should add as little bulk resistance as possible, and the gap under a lapped crystal on a lapped copper block is small.
- Flatness beats pressure. A crystal on a bowed heat sink has a wedge-shaped gap; no pad compensates for that without huge pressure. Lap the copper, check it, then choose the pad.
- Guard against edge loading. A hard clamp ring on the crystal edge creates a stress concentration. Distribute load through an elastomer pad or a compliant backing.
- Keep the optical path clean. Specify low-volatile material and avoid any silicone that can migrate onto coatings; our pad is a cured solid, which matters more here than in any other application.
When we tell customers not to use our product
If your interface budget is 1-2 °C at 250 W/cm², indium (or a soldered/AuSn joint) is the right engineering answer and we say so. Where the pad wins is when the real constraint is process, serviceability or risk: no flux near coatings, no thermal cycle on a €20k optic, a mount that must be re-opened during development, or a joint that must not creep over 10,000 cycles.
One more option worth considering: a hybrid - indium on the crystal side for the tight interface, a graphene pad between the copper block and the water-cooled baseplate where pressure is higher, flatness is easier and the resistance budget is looser. That is often the cheapest way to get most of the performance.