Description
Thorlabs PF2011-M01 Ø50.8 mm (2 in) Zerodur® Protected Gold Mirror, 800 nm – 20 µm
Product Overview
Ask an engineer what limits an infrared fold mirror and the answer is usually “the coating”. In a high-power infrared system that answer is often wrong, and getting it wrong sends money in the wrong direction. The useful exercise is to account for the heat instead of guessing at it: list every place in the optic where a watt can be absorbed, and then ask which of those places the coating choice actually controls.
Take the PF2011-M01 as the subject. It is a round mirror 50.8 mm (2 in) in diameter and 12.0 mm (0.47 in) thick, Zerodur® substrate, protected gold coating with a SiO₂ overcoat, average reflectance 96 percent from 800 nm to 20 µm, front surface flatness λ/10 at 632.8 nm, surface quality 40-20 scratch-dig, clear aperture 45.7 mm, parallelism 3 arcmin or better, angle of incidence 0° to 45°, model number laser-engraved on the fine-ground back face. The thermal expansion coefficient of the substrate is 0 ± 0.100 x 10⁻⁶/°C, against 0.55 x 10⁻⁶/°C for fused silica.
Now the heat budget. The four percent that this coating does not reflect is not all absorbed in the gold: some of it is absorbed in the SiO₂ overcoat, some in the metal, some at the interface between them, and a further fraction of the transmitted power passes into the substrate and is absorbed in the 12.0 mm of glass-ceramic behind the film. In the near infrared, where the substrate is still reasonably transparent, that last term is small. In the far infrared it is not negligible, because the glass-ceramic materials used for mirror substrates absorb strongly beyond a few micrometres and a 12 mm path is a long path. The practical conclusion is uncomfortable for anyone who assumed the coating is the whole story: at 10.6 µm a meaningful share of the absorbed power sits in the substrate, tens of micrometres to millimetres behind the reflecting surface, where no coating improvement can reach it.
That is why the substrate specification on this part is not a footnote. Whatever heat ends up in the body has to be conducted to the mount, and while it travels it sets up a temperature distribution. A distribution that is uniform scales the part and changes nothing optically. A distribution with a gradient bends it, and the bending is proportional to the expansion coefficient. At 0 ± 0.100 x 10⁻⁶/°C the Zerodur body converts a given gradient into about one fifth of the figure change that a fused silica body of the same geometry would show, and that ratio is the whole reason a protected gold mirror is offered on this substrate rather than only on the standard one.
The coating itself is a well-behaved broadband reflector. Gold’s reflectance rises steeply through the visible and flattens beyond roughly 800 nm, which is why the specification starts there and not lower: below that edge the metal absorbs through interband transitions and the reflectance falls away quickly, so there is no useful band to specify. From 800 nm to 20 µm the average is 96 percent, and the word that matters is average. A flat 96 percent across nearly one and a half decades of wavelength is what makes a broadband infrared system predictable: the same optic behaves the same way when the source is changed, so the link budget does not have to be rebuilt for every new line.
The SiO₂ overcoat is the difference between the M01 and the unprotected M03, and Thorlabs describes that trade-off precisely. The protected gold, silver and aluminum mirrors have a durable SiO₂ overcoat, and each protective overcoat helps to protect the metallic coating from tarnishing and prevents damage to the fragile metallic surface. By contrast, unprotected gold offers higher reflectance than protected gold but is more delicate. The same source also states the cost of the protection: the overcoat of the protected mirrors results in polarization changes that can be difficult to quantify and sometimes induces absorption in the mid- to far-IR. Both halves of that sentence matter here. The M01 is the right part when the mirror has to be handled, cleaned and shared; it is the wrong part when the polarization state has to be preserved to a known tolerance, or when the last octave of the infrared has to be free of overcoat absorption.
Zerodur is a registered trademark of Schott.
Key Features
- Ravg of 96 percent from 800 nm to 20 µm. A single average covering more than four octaves, which is the property that makes one optic usable across a multi-source infrared bench.
- SiO₂ overcoat for handling durability. The overcoat protects the metal from tarnish and from contact damage, and in exchange it introduces the polarization and mid-infrared absorption effects described above. It is a deliberate trade, not an upgrade.
- Zerodur® substrate, 0 ± 0.100 x 10⁻⁶/°C. The residual heat that the coating cannot prevent still has to cross the substrate, and the substrate coefficient decides how much of it becomes curvature.
- Pulsed threshold of 2 J/cm² at 1.064 µm. Measured with 10 ns pulses at 10 Hz over a Ø1.000 mm spot, which corresponds to about 15.7 mJ in a single pulse at that fluence and that spot size. A large test spot makes this a conservative number rather than an optimistic one.
- CW thresholds of 500 W/cm at 1.070 µm and 750 W/cm at 10.6 µm. Two different spot sizes and two different wavelengths, so the pair cannot be read as a ranking; the arithmetic in the specifications section shows why.
- λ/10 at 632.8 nm over a 45.7 mm clear aperture, 40-20 scratch-dig, 3 arcmin parallelism. The parallelism specification matters more than usual on a thick infrared optic, because a wedge between the faces sends a back reflection off in a different direction than the front reflection.
- Twelve millimetres of body. A thickness-to-diameter ratio of about 0.24 makes the part stiff enough that a hand-tightened ring holds the figure, and thick enough that a short thermal transient is spread through the body before it reaches the mount.
Technical Specifications
| Parameter | Specification |
|---|---|
| Part Number | PF2011-M01 |
| Coating | Protected gold |
| Overcoat | SiO₂ |
| Spectral Range | 800 nm to 20 µm |
| Average Reflectance | Ravg of 96 percent from 800 nm to 20 µm |
| Substrate | Zerodur® (substrate grade) |
| Coefficient of Thermal Expansion | 0 ± 0.100 x 10⁻⁶/°C (fused silica reference: 0.55 x 10⁻⁶/°C) |
| Diameter | 50.8 mm (2 in), tolerance +0.0 / -0.1 mm |
| Thickness | 12.0 mm (0.47 in), tolerance ±0.2 mm |
| Parallelism | 3 arcmin or better |
| Front Surface Flatness | λ/10 at 632.8 nm |
| Surface Quality | 40-20 scratch-dig |
| Clear Aperture | 90 percent of diameter (45.7 mm) |
| Angle of Incidence | 0° to 45° |
| Back Surface | Fine ground, model number laser-engraved |
| Pulsed Damage Threshold | 2 J/cm² at 1.064 µm, 10 ns, 10 Hz, Ø1.000 mm |
| CW Damage Threshold (1.07 µm) | 500 W/cm at 1.070 µm, Ø0.089 mm |
| CW Damage Threshold (10.6 µm) | 750 W/cm at 10.6 µm, Ø0.339 mm |
| Polarization Behaviour | Overcoat introduces polarization changes that are difficult to quantify; choose unprotected gold where the state must be preserved |
| Typical Mounting | Two-inch optic mount or kinematic mount with retaining ring |
The damage thresholds belong to the coating, and Thorlabs notes that the measured value does not change with the diameter of the optic or with the substrate, because the coating itself does not change. The measurements are made on clean optics, and contamination lowers the threshold substantially, which matters more in the infrared than the numbers alone suggest because a contaminated surface also absorbs more of the beam that is supposed to pass through the coating.
The two CW figures need the same treatment as the pulsed one. At 500 W/cm over a Ø0.089 mm beam, the power is about 4.5 W into 6.2 x 10⁻⁵ cm², an intensity near 72,000 W/cm². At 750 W/cm over a Ø0.339 mm beam, the power is about 25.4 W into 9.0 x 10⁻⁴ cm², an intensity near 28,000 W/cm². Unlike some other coatings in this family, these two do not collapse onto a single intensity: the near-infrared test is 2.5 times harsher in W/cm² than the far-infrared test. The reason is that the two tests are not testing the same failure mode. At 1.07 µm the dominant risk is heating of the gold film and the overcoat, and the small spot concentrates it; at 10.6 µm the substrate absorbs a much larger share of the transmitted power and the failure is a mixed film-and-substrate event that the larger spot distributes differently. Treat the two numbers as separate ceilings for separate regimes and do not interpolate between them.
| Thermal Regime | Typical Source | Where the Heat Lands | What the Substrate Buys |
|---|---|---|---|
| Near infrared, high average power | Yb fibre, 1.03 to 1.08 µm | Mostly the film and the overcoat, small share in the substrate | Keeps the residual substrate absorption from turning into curvature |
| Far infrared, continuous | CO₂, 10.6 µm | A significant share in the 12.0 mm of glass-ceramic behind the film | The only lever that acts on heat the coating cannot reflect |
| Pulsed, nanosecond | Q-switched 1 µm | A thin layer at the film interface, in a time too short for the body to respond | Little during the pulse; a great deal in the milliseconds after it |
Applications
- CO₂ laser processing heads. At 10.6 µm the mirror is exposed to continuous power and to back-reflected light from the workpiece, and the fold that sends the beam into a focusing lens is usually the hottest optic in the head. The 750 W/cm figure applies at that line, and the substrate keeps the heat that does get through from walking the focus.
- Fibre laser beam delivery with infrared monitoring. A 1 µm process beam plus a longer-wavelength pyrometer or camera channel can share one coating, and the flat 96 percent average means the monitoring channel does not need its own optic.
- Mid-infrared spectroscopy and gas sensing. Quantum cascade and interband cascade sources between 3 and 12 µm are low power but the paths are long and multipass cells fold the beam dozens of times; a predictable average reflectance is what makes a multipass throughput calculation meaningful.
- Thermal imaging optics for furnace and process monitoring. A fold mirror inside a camera housing sits in a warm enclosure and sees the enclosure’s thermal radiation; a low-expansion substrate keeps the image from creeping as the housing heats.
- Free-space infrared communication and ranging benches. Where a 1.5 µm source is combined with a visible boresight, the gold coating handles the infrared and a separate dielectric handles the visible, and the gold fold is chosen for stability rather than for peak reflectance.
- High-power infrared attenuation and beam dumping. A gold fold used to redirect a beam into a dump sees the full power at grazing angles; the 0° to 45° specified range and the metal’s smooth angular behaviour make the geometry forgiving.
Selection Guide
The decision between the three gold codes on this body is not about reflectance, because all three are within a few points of each other. It is about what the coating must not do. The table below reduces the choice to the requirement that is actually driving it.
| If the requirement is | Then the answer is |
|---|---|
| A mirror that can be handled, cleaned and shared between users | M01, this part: 96 percent average with a SiO₂ overcoat |
| Maximum reflectance in the 2 to 20 µm window | M02, mid-infrared enhanced gold at 98 percent average, with the near infrared given up |
| A polarization state that must survive reflection, or no overcoat dispersion | M03, unprotected gold at 97 percent average, with strict handling rules |
| Maximum reflectance at the 10.6 µm line and nothing else | L01, line-optimised bare gold at 99 percent average |
| Coverage from the visible through the infrared on one optic | G01 protected aluminum, at the cost of a lower infrared average |
Two further points belong to selection rather than to the coating table. First, aperture: the clear aperture is 45.7 mm, and at 45° incidence the footprint stretches by about 1.41 along one axis, so a beam wider than roughly 32 mm begins to see the edge. Second, mounting: with a substrate this dimensionally quiet, the mount becomes the dominant thermal term, and a rigid aluminium clamp will impose its own expansion on a Zerodur optic. Use a compliant retainer and let the part slide.
Series Comparison
The PF2011 body carries seven metallic coatings. The columns below are chosen for the infrared buyer: band, average, overcoat and the polarization consequence of that overcoat.
| Code | Coating | Spectral Range | Average Reflectance | Overcoat | Polarization |
|---|---|---|---|---|---|
| F01 | UV-enhanced aluminum | 250-450 nm | Ravg 90% | MgF₂ | Overcoat present |
| G01 | Protected aluminum | 450 nm-20 µm | Ravg 90% (450 nm-2 µm), 95% (2-20 µm) | SiO₂ | Overcoat present |
| P01 | Protected silver | 450 nm-20 µm | Ravg 97% (450 nm-2 µm), 95% (2-20 µm) | SiO₂ | Overcoat present |
| M01 | Protected gold | 800 nm-20 µm | Ravg 96% | SiO₂ | Overcoat present, change not quantifiable |
| M02 | MIR enhanced gold | 2-20 µm | Ravg 98% | SiO₂ | Overcoat present |
| M03 | Unprotected gold | 800 nm-20 µm | Ravg 97% | None | No overcoat, state preserved |
| L01 | Unprotected gold, line optimised | 10.6 µm line | Ravg 99% | None | No overcoat, state preserved |
The pattern in the last column is the pattern to remember: every code with a dielectric overcoat carries the polarization caveat, and every code without one carries a handling penalty instead. There is no member of the family that offers both protection and an exactly preserved polarization state, because the overcoat that provides the first is the thing that causes the second. The M01 sits in the protected group with the highest broadband average in that group, which is the precise definition of a general-purpose infrared fold.
FAQ
Q1. At high power, should I worry about the coating or about the substrate?
Both, but in sequence, and the sequence decides where the money goes. Start with the coating, because it is what the beam hits first and because the failure at the coating is catastrophic and fast: a coating damaged at fluence fails in a single pulse and the optic is scrap. Once the coating is comfortably inside its threshold, the remaining problem is slow and reversible: the substrate warms, the figure changes, and the beam walks. That second problem is where the substrate specification acts, and it is the one that a low-expansion body solves. In a well-designed system the coating is never near its threshold, so the substrate is the term that governs long-term stability; in a badly designed system the coating is near its threshold and no substrate will save it.
Q2. Does the SiO₂ overcoat change the polarization of my beam?
Yes, and Thorlabs says so explicitly: the overcoat of the protected mirrors results in polarization changes that can be difficult to quantify. The operative phrase is “difficult to quantify”, not “small”. A dielectric layer of finite thickness has a slight birefringence under oblique incidence and a phase shift that depends on angle and wavelength, and because the exact thickness is a process parameter rather than a published number, you cannot compute the effect from the datasheet. If your experiment depends on a known polarization state at the output, either characterise the actual part you were shipped or move to the unprotected gold, which has no overcoat to introduce the effect.
Q3. The 1.070 µm CW value is 500 W/cm and the 10.6 µm value is 750 W/cm. Is the mirror better at 10.6 µm?
Not in the way that comparison implies. The two figures come from different spot sizes, 0.089 mm and 0.339 mm, and converting both to intensity gives about 72,000 W/cm² and about 28,000 W/cm², so the near-infrared test was actually the harsher of the two in intensity terms. The two also probe different failure mechanisms: at 1.07 µm the heat is mostly in the film, at 10.6 µm a substantial share of the transmitted power is absorbed in the substrate. Use each number inside its own regime, and if your application sits between the two, measure rather than interpolate.
Q4. Can this mirror be used below 800 nm at all?
It will reflect, but the specification does not cover it and the reflectance falls quickly as the wavelength shortens. Gold’s useful band begins where interband absorption stops dominating, which is the reason the specification starts at 800 nm. If a visible alignment beam has to share the path with a 10.6 µm beam, the protected aluminum G01 is the correct choice because it is specified from 450 nm, and if the visible beam is only for set-up, a separate visible fold on a flip mount is cheaper than compromising the infrared optic.
Q5. Why is the flatness quoted at 632.8 nm on an optic that is used at 10.6 µm?
Because flatness is a physical height, and the interferometer that measures it works at a visible helium-neon wavelength. A λ/10 deviation at 632.8 nm is a physical deviation of about 0.063 µm. At 10.6 µm that same deviation is a tiny fraction of a wave, roughly λ/168, so the surface is optically far smoother at the working wavelength than the specification suggests. The figure specification is therefore conservative for infrared use, and the number that actually limits an infrared system is usually the thermal figure change, not the polished figure.
Q6. What does the overcoat do to the far-infrared end of the band?
It can add absorption. Thorlabs states that the overcoat of the protected mirrors sometimes induces absorption in the mid- to far-IR, which is one of the reasons the unprotected gold is offered at all. In practice the effect is a slow decline toward the long-wavelength end rather than a sharp cutoff, and it is already inside the 96 percent average over 800 nm to 20 µm. If your work sits at the extreme long-wavelength end and every percent matters, the M03 or the line-optimised L01 is the better starting point, at the cost of handling discipline.
Usage Precautions
- Handle by the edge with gloves, and keep the coated face out of contact with anything. The overcoat is durable, not indestructible; a fingerprint left on the surface will be cleaned off eventually and will leave a cosmetic trace that never fully disappears.
- Clean by flooding and a single tissue drag, never by rubbing a spot. Use clean solvent, drag once in one direction under the weight of the tissue, and stop. Repeated wiping on the same area is what damages overcoats.
- Do not assume the overcoat seals the metal. It protects against tarnish and contact damage; it is not a hermetic barrier. Store the optic in its case in a dry place, and keep a purged enclosure around any optic that runs hot.
- Watch the back reflection. A 3 arcmin parallelism specification still leaves a back-surface reflection that can travel a few milliradians away from the front reflection; in a system with a sensitive detector, either angle the optic slightly or block the ghost.
- Respect the difference between a large-spot and a small-spot threshold. The published 2 J/cm² was measured over a 1 mm spot on a clean optic. A dirty optic in a 100 µm spot has a lower and unknown threshold, and alignment at full power into a small spot is the most common way infrared folds are destroyed.
- Let the mount move with the optic. A Zerodur body held in a rigid aluminium ring will be bent by the ring, not by its own expansion. Use a compliant retainer, tighten by hand, and re-check the figure after the first thermal cycle rather than before it.
Disclaimer
The specifications on this page are compiled from Thorlabs’ publicly published product documentation and are provided for selection reference only. Where this page differs from the manufacturer’s current official datasheet, the datasheet prevails. Specifications, coatings and appearance may change without notice. Damage thresholds are measured values under the stated test conditions and are not a guarantee of performance in any particular setup; users must verify suitability at their own wavelength, power, spot size and pulse format. The intensity conversions and thermal reasoning on this page are engineering arithmetic from published figures, not manufacturer specifications. Prices and lead times are confirmed at quotation.
Contact Us
Send us the wavelength, the power, the spot size on the optic and the number of folds, and we will tell you whether the coating threshold or the substrate stability is your binding constraint, and whether the protected or unprotected gold is the better fit for how the bench is actually used.
Hotline: +86 159-2722-0467
E-mail: 859818709@qq.com | Website: www.soar-laser.cn
We supply evaluation units as well as volume orders, and the same coating is available on Ø1/2 in and Ø1 in bodies for systems that need a matched set.









Reviews
There are no reviews yet.