Description
Product Overview
The PF10-03-M02 is the mid-infrared specialist of the Thorlabs 1 inch metal mirror family. It is a Ø25.4 mm fused silica disc, 6.0 mm thick to within plus or minus 0.2 mm, carrying a dielectric-enhanced gold coating specified at 98 percent average reflectance and 95 percent minimum absolute reflectance across 2 µm to 20 µm, for incidence angles from 0 to 45 degrees. The front face is figured to λ/10 at 633 nm with 40-20 scratch-dig, parallelism is 3 arcmin or better, the clear aperture is 90 percent of the diameter at Ø22.9 mm, and the rear face is fine ground and laser engraved with the item number.
The 98 percent figure is the highest average reflectance anywhere in this family. Bare gold on its own is already a good mid-infrared reflector, but its curve is not flat: it dips and rolls across the 2 to 20 µm window depending on where you look. The added dielectric layers are what lift the floor to 98 percent average and hold a 95 percent absolute minimum, which is the number that actually protects a power budget. This is why a 10.6 µm carbon dioxide beam path or a 3 to 5 µm optical parametric oscillator belongs on this coating rather than on plain gold.
The price of that performance is paid entirely at short wavelengths, and the payment is absolute. Below 2 µm this coating carries no specification of any kind: not a reduced reflectance, not a derated figure, nothing at all. It is a mid-infrared design, and the same stack that lifts the curve at 10 µm collapses it in the near infrared. A beam path that carries a 633 nm alignment laser and a 10.6 µm working beam through the same fold cannot use this optic. The alignment beam will come back dim, and worse, its brightness will change with angle in a way that no datasheet describes, which turns alignment into guesswork.
The second price is the damage threshold, and it is the point most often misread. At 1.064 µm the pulsed rating is 0.1 J/cm² and the continuous rating is 25 W/cm, roughly an order of magnitude below protected gold. Those numbers are alarming only if the wavelength is ignored: they are measured where the coating is not designed to work. At 10.6 µm, where it is designed to work, the ratings are 3 J/cm² for 100 ns pulses at 1 Hz and 450 W/cm continuous, which sits comfortably above what a carbon dioxide cutting or marking head delivers at the fold mirror. Read the rating that belongs to your wavelength, and read nothing else.
Key Features
- 98 percent average reflectance from 2 to 20 µm: the highest of any metal coating in this format, and the reason to accept its restrictions
- 95 percent absolute minimum, not just an average: a floor rather than a mean, so a wavelength that happens to land in a dip still returns 95 percent
- Specified from 0 to 45 degrees: the reflectance figure holds across the usual 45 degree fold and the steering offsets around it, rather than only at normal incidence
- Dielectric-enhanced gold: a gold film with a designed stack over it, giving both the high reflectance and a surface that resists physical damage better than bare gold
- λ/10 figure at 633 nm: a reflected wavefront clean enough to focus or to interfere downstream
- 40-20 scratch-dig: the standard grade for a coated optic, adequate for imaging and for beam delivery
- 6.0 mm fused silica: thermally stable and stiff enough that a retaining ring does not bend it
- 3 J/cm² and 450 W/cm at 10.6 µm: ratings that match the wavelength the coating was built for
Technical Specifications
| Model | PF10-03-M02 |
| Coating | MIR-enhanced gold (dielectric-enhanced gold film) |
| Wavelength range | 2 µm – 20 µm |
| Average reflectance | Ravg 98 percent from 2 to 20 µm |
| Absolute reflectance | Rabs 95 percent minimum from 2 to 20 µm |
| Angle of incidence | 0 to 45 degrees |
| Specification note | No reflectance specification exists below 2 µm |
| Substrate | Fused silica |
| Diameter | Ø25.4 mm (1 inch) |
| Diameter tolerance | +0.0 mm / -0.1 mm |
| Thickness | 6.0 mm (0.24 in) |
| Thickness tolerance | plus or minus 0.2 mm |
| Surface flatness | λ/10 at 633 nm, peak to valley |
| Surface quality | 40-20 scratch-dig |
| Parallelism | 3 arcmin or better |
| Clear aperture | 90 percent of diameter (Ø22.9 mm) |
| Rear surface | Fine ground, laser engraved with the item number |
| Damage threshold, pulsed at 1.064 µm | 0.1 J/cm² (10 ns, 10 Hz, Ø1.06 mm) |
| Damage threshold, pulsed at 10.6 µm | 3 J/cm² (100 ns, 1 Hz, Ø1.29 mm) |
| Damage threshold, CW at 1.07 µm | 25 W/cm (Ø1.04 mm) |
| Damage threshold, CW at 10.6 µm | 450 W/cm (Ø1.18 mm) |
| Source | Manufacturer published series specifications |
Applications
- Carbon dioxide laser beam delivery: every fold between the source and the cutting head, at 10.6 µm, where the coating returns 98 percent and tolerates 450 W/cm
- Mid-infrared optical parametric oscillators: signal and idler routing in the 2 to 5 µm region, where the absolute 95 percent floor protects the conversion efficiency
- Quantum cascade and interband cascade laser benches: 4 to 12 µm spectroscopy setups where several steering mirrors in series would otherwise eat the signal
- Fourier transform infrared spectrometers: external beam paths and accessory modules, where a flat response across the whole instrument band matters more than peak reflectance at one wavelength
- Thermal imaging and radiometry: 8 to 12 µm collection optics where a metal coating avoids the chromatic behaviour of a refractive element
- Gas sensing and process analysers: absorption cells at 3 to 5 µm and 8 to 12 µm with multiple passes and therefore multiple reflections
Selection Guide
Four coatings cover almost every request in this format. The table is the honest comparison, including the columns where the M02 loses.
| Coating / item | Band | Ravg | AOI | Pulsed threshold | Cleanable | Polarization |
|---|---|---|---|---|---|---|
| Protected gold M01 | 800 nm – 20 µm | 96% | 0 – 45° | 2 J/cm² at 1.064 µm | Yes | Overcoat shifts it, hard to quantify |
| MIR-enhanced gold M02 | 2 – 20 µm | 98% | 0 – 45° | 0.1 J/cm² at 1.064 µm | Yes | Overcoat shifts it, hard to quantify |
| Unprotected gold M03 | 800 nm – 20 µm | 97% | 0 – 45° | 4 J/cm² at 10.6 µm | No | Preserved, no overcoat |
| Protected silver P01 | 450 nm – 20 µm | 97% / 95% | 0 – 45° | 3 J/cm² at 1.064 µm | Yes | Overcoat shifts it, hard to quantify |
Read down the band column first. If the answer is “2 µm and above only”, the M02 wins on reflectance and nothing else comes close. If the answer includes any visible or near-infrared light, the M02 is eliminated before reflectance is discussed, and the choice falls to M01, M03 or P01 depending on power, polarization and whether the surface will be touched.
Series Comparison
The -M02 coating is available across the full range of round and square sizes, so a mid-infrared beam path can be built from one coating family rather than from whatever happened to be in stock. Damage thresholds are constant for a given coating and do not scale with size.
| Item | Format | Diameter or face | Thickness | Clear aperture |
|---|---|---|---|---|
| PF03-03-M02 | Round | Ø7.0 mm | 2.0 mm | Ø6.3 mm |
| PF05-03-M02 | Round | Ø12.7 mm | 6.0 mm | Ø11.4 mm |
| PF07-03-M02 | Round | Ø19.0 mm | 6.0 mm | Ø17.1 mm |
| PF10-03-M02 | Round | Ø25.4 mm | 6.0 mm | Ø22.9 mm |
| PF20-03-M02 | Round | Ø50.8 mm | 12.0 mm | Ø45.7 mm |
| PFSQ10-03-M02 | Square | 25.4 x 25.4 mm | 6.0 mm | 90 percent of face |
The same optic is also sold as the PF10-03-M02-10, a pack of ten, which is the sensible purchase for a machine with several fold points rather than for a single bench.
FAQ
Q: Why does a 98 percent mirror have a lower damage threshold than a 96 percent one?
A: Because the two numbers describe different wavelengths. The 98 percent is achieved in the mid infrared by a dielectric stack that is tuned for that band. At 1.064 µm that stack is far from its design point, absorbs strongly and fails early, which is what the 0.1 J/cm² figure records. At 10.6 µm the same coating is on its design point and the rating rises to 3 J/cm² and 450 W/cm.
Q: Can I use it at 1064 nm if my power is low?
A: Technically the optic will reflect something there, but you would be operating a part with no specified reflectance and a low damage threshold, and you would have no way to predict the result. Use the protected gold M01 for 1064 nm work. The M02 exists for the band it was designed for.
Q: What is the difference between Ravg 98 percent and Rabs 95 percent?
A: The average is the mean across the band; the absolute figure is the worst point you are guaranteed to meet anywhere in the band. A coating can average 98 percent while dipping to 90 percent at one wavelength. The 95 percent floor is the number to use when a single laser line sits inside the band and you do not know exactly where the dips are.
Q: Can I align the beam path with a helium-neon laser through this mirror?
A: Not through this mirror. A 633 nm beam is below the 2 µm start of the specified band and will not return a usable reflection. If the system needs a visible alignment beam and a mid-infrared working beam on the same fold, the fold has to be a different coating: M01 or P01, both of which cover 633 nm and 10.6 µm in one film.
Q: Is the coating cleanable?
A: Yes, more so than bare gold. The dielectric stack is a hard layer and will tolerate a gentle solvent drag wipe after loose dust has been blown off. It is not armour plate: no scrubbing, no repeated wiping, no abrasive tissue, and no solvent that leaves a residue.
Q: What happens at 45 degrees?
A: The 98 percent average and 95 percent absolute figures are specified over 0 to 45 degrees, so a 45 degree fold is inside the specification rather than outside it. Beyond 45 degrees the reflectance is not specified, and the polarization split between s and p becomes large enough to matter in a power budget.
Usage Precautions
- Confirm the working wavelength before the optic goes anywhere near a mount. The failure mode of this part is being fitted to a 1 µm beam path because it looked like the best mirror in the catalogue.
- Keep every visible and near-infrared beam off this surface. If a red pilot laser shares the enclosure, give it its own fold mirror rather than a dichroic trick on this one.
- Do not use the 1.064 µm rating as a safety margin calculation. It is a measurement at a wavelength the coating was not built for, and it says nothing useful about your 10.6 µm beam.
- Respect the 450 W/cm continuous rating at 10.6 µm with the beam diameter in the calculation, not the total power. A 100 W beam focused to 0.3 mm is a very different number from the same beam at 5 mm.
- Clean rarely and gently. Blow first, wipe second, one direction, no pressure. A dielectric stack that is wiped weekly will be a dielectric stack with a shortened life.
- Keep the part dry and let a cold optic reach room temperature sealed before opening it. Condensation on a mid-infrared coating is the usual route to a hazy surface and a lost percent.
- Dump the rear scatter. The rear face is fine ground and scatters rather than forming a clean ghost, but at 10.6 µm that scatter is still heat landing on something.
Disclaimer
The figures on this page are taken from the manufacturer published series tables and are supplied for selection purposes only. The current official datasheet governs. Coatings, dimensions and markings may change without notice. Price and delivery are quoted separately by our sales desk.
Contact Us
Send us the working wavelength, the beam diameter at the optic, the incidence angle and the total power, and we will tell you whether the M02 is the right film or whether the presence of an alignment beam rules it out. Where a path mixes bands, we will quote the combination rather than a single part number.









Reviews
There are no reviews yet.