Description
Product Overview
The Thorlabs PF1011-F01 is a Ø25.4 mm (Ø1 in) plano mirror carrying a UV-enhanced aluminum film with an MgF₂ overcoat on a Zerodur® glass-ceramic substrate. Zerodur® is a registered trademark of Schott. Average reflectance is 90% across 250 nm to 450 nm. The disc is 6.0 mm thick with a tolerance of plus or minus 0.2 mm, the diameter tolerance is +0.0 / -0.1 mm, the coated face is figured to λ/10 peak to valley at 632.8 nm and polished to 40-20 scratch-dig, parallelism is held inside 3 arcmin, and the clear aperture is 90% of the diameter, which is Ø22.9 mm. The rear face is fine ground and carries a laser-engraved item number.
This part sits at the intersection of three constraints that pull in different directions, and understanding the compromise is most of understanding the product. The first is spectral: below 450 nm the metal films that dominate the visible and infrared all lose reflectance, aluminum least badly, and a coating optimised for the ultraviolet is what remains. The second is thermal: an ultraviolet photon carries more energy than a visible one, and the materials in a UV beam path absorb more of it, so a mirror in a UV system runs hotter per watt than the same mirror at 1064 nm. The third is mechanical: whatever the substrate does when it gets hot shows up as a change in the reflected wavefront, which in a UV lithography or inspection path is a change in the image. UV-enhanced aluminum answers the first, and Zerodur answers the third, which is why the two appear together on this part.
The reflectance figure needs to be read with the band in mind. 90% average across 250 to 450 nm is lower than the 97% a protected silver mirror delivers in the visible, and that is not a defect in the coating: it is what a metal film can do in the ultraviolet. Aluminum is the only common metal that holds up below 400 nm at all, because its plasma edge sits far enough into the ultraviolet that the free-electron behaviour survives, while silver falls away steeply below about 450 nm and gold is useless across the whole visible. The MgF₂ overcoat is what keeps the aluminum from growing a lossy oxide at the surface, and it is chosen over SiO₂ because fluoride films stay transparent further into the ultraviolet, where a silica overcoat would begin to absorb the very light the coating is there to reflect. The trade is that MgF₂ is the softer and more moisture-sensitive of the two, so a UV mirror is handled with more care than a visible one, not less.
The band edge deserves to be stated plainly because it decides whether this part can be used at all. The specification starts at 250 nm, so a 355 nm third harmonic of a Nd:YAG laser, a 351 nm xenon fluoride excimer, a 308 nm xenon chloride excimer and a 266 nm fourth harmonic all sit comfortably inside it. The two excimer lines that people ask about most often, 248 nm from krypton fluoride and 193 nm from argon fluoride, do not. A 248 nm line is at the edge of the specification at best and a 193 nm line is outside it by a wide margin, and neither should be planned onto this coating. If the process runs at 193 or 248 nm, the requirement is a different class of optic, usually a dielectric-coated or specially processed UV mirror, and asking this part to do the job will produce a short-lived mirror and an unexplained loss of throughput.
The thermal argument is where the Zerodur substrate earns its cost, and it is a shorter version of a longer story. Ultraviolet optics absorb more than infrared ones: the coatings are thinner, the materials are closer to their absorption edges, and contamination films that are transparent in the near infrared become absorbing in the deep ultraviolet. Absorbed power raises the temperature of the coated patch, the substrate beneath it expands, and a flat surface acquires a slight curvature. In a lithographic illuminator or an inspection objective that curvature is a focus error, and a focus error at 355 nm is a larger fraction of the wavelength than the same physical error at 1064 nm. Zerodur’s expansion coefficient of 0 ± 0.100 × 10⁻⁶/°C, against roughly 0.55 × 10⁻⁶/°C for fused silica, cuts that link in the chain. The coating still absorbs, the surface still warms, but the geometry barely moves.
One grade confusion is worth clearing up before it causes a rejected delivery. The coated part is specified at 40-20 scratch-dig, not 10-5. The uncoated PF1011 blank from the same family is quoted at 10-5 because it is inspected before anything is deposited on it; once the aluminum and the MgF₂ have been laid down and the part has been handled, mounted in a deposition fixture and inspected again, 40-20 is the grade that the finished mirror is held to. A purchase specification that asks for 10-5 on a coated UV mirror is asking for something the process cannot promise, and the mismatch usually surfaces as a dispute at goods-in rather than as a technical problem.
Key Features
- UV-enhanced aluminum, 90% average reflectance from 250 to 450 nm: the only common metal film that holds useful reflectance across the near and middle ultraviolet
- MgF₂ overcoat: chosen over SiO₂ because a fluoride layer stays transparent further into the ultraviolet, and it is what protects the aluminum from the lossy native oxide that would otherwise form
- Zerodur® substrate, 0 ± 0.100 × 10⁻⁶/°C: a coefficient of thermal expansion more than five times lower than fused silica, which is what keeps a UV-heated mirror from changing its figure
- λ/10 peak to valley at 632.8 nm: the figure is measured in the visible, but it is the number that governs how much wavefront error the mirror contributes at 355 nm, where every nanometre counts for more
- 40-20 scratch-dig on the coated face: the correct grade for a finished coated optic; the 10-5 figure belongs to the uncoated blank in the same family
- Ø25.4 mm body, Ø22.9 mm clear aperture: the standard one-inch mount format, with enough coated area for a beam of 10 to 15 mm without sampling the edge roll-off
- 6.0 mm thick, plus or minus 0.2 mm: substantial enough that a three-point seat does not bend it, thin enough for compact UV heads
- Parallelism inside 3 arcmin: keeps the rear-surface ghost close to the main beam, which matters in an imaging path where a stray reflection becomes a ghost image
- Fine-ground rear face with engraved item number: scatters the rear reflection instead of forming a clean ghost, and identifies the part after it leaves its bag
Technical Specifications
| Model | PF1011-F01 |
| Coating | UV-enhanced aluminum with MgF₂ protective overcoat |
| Wavelength range | 250 nm – 450 nm |
| Average reflectance | 90% across 250 nm – 450 nm |
| Substrate | Zerodur® glass-ceramic (registered trademark of Schott) |
| Coefficient of thermal expansion | 0 ± 0.100 × 10⁻⁶/°C |
| Diameter | Ø25.4 mm (Ø1 in) |
| 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 peak to valley at 632.8 nm over the clear aperture |
| Surface quality | 40-20 scratch-dig |
| Parallelism | 3 arcmin or better |
| Clear aperture | 90% of diameter (Ø22.9 mm) |
| Rear surface | Fine ground, laser engraved with the item number |
| Pulsed damage rating | 0.3 J/cm² at 355 nm in the family specification table; confirm test conditions against the current datasheet for your pulse length |
| Mounting | Standard Ø1 in optic mounts, kinematic seats and retaining rings |
| Source | Manufacturer published series specification table |
Applications
Everything below shares one property: the beam is ultraviolet and the layout has to stay aligned while the optics warm. Where either condition is absent, a cheaper part will do.
- Third-harmonic Nd:YAG at 355 nm: the most common use by a wide margin, because 355 nm sits in the middle of the band and the conversion stage that produces it is sensitive to any loss in the following fold
- Fourth-harmonic work at 266 nm: still inside the specification, with less margin at the short end and a stronger argument for keeping the surface clean, since a contamination film that is invisible at 532 nm absorbs at 266 nm
- Xenon chloride excimer at 308 nm and xenon fluoride at 351 nm: both inside the band; the 248 nm krypton fluoride line is not, and must not be planned onto this coating
- UV lithography and mask inspection illuminators: a fold inside an objective or an illuminator that must hold focus while the lamp or laser heats the housing, which is exactly the case the Zerodur substrate is there for
- Fluorescence excitation and UV curing delivery: long runs at moderate power, where the mirror’s own figure stability matters more than its peak reflectance
- Semiconductor and flat-panel inspection optics: UV illumination folded several times before the camera, where each fold adds wavefront error and a substrate that breathes with temperature adds it unevenly
- UV spectroscopy and beam conditioning: broadband ultraviolet sources where a film with a flat 250 to 450 nm response is worth more than a peak value at one line
Selection Guide
The first question is the wavelength, and it eliminates most of the catalogue immediately. The second is the power density on the surface, and the third is how much focus drift the process can tolerate. The table handles the first and the third.
| Situation | Part | Reason |
|---|---|---|
| 355 nm, 308 nm, 351 nm or 266 nm, focus must hold while warm | PF1011-F01 (this page) | UV band plus a near-zero expansion substrate |
| Same wavelengths, low power, bench work | PF10-03-F01 | Same coating on fused silica; cheaper, and thermal drift does not matter at milliwatt levels |
| 193 nm or 248 nm excimer | Dielectric UV mirror | This coating starts at 250 nm; the excimer lines below it are outside specification |
| 450 nm to 20 µm broadband | PF10-03-G01 or PF10-03-P01 | Protected aluminum or silver; both beat UV aluminum across the visible |
| Visible and near infrared, highest reflectance | PF10-03-P01 | 97% from 450 nm to 2 µm against 90% here |
| Uncoated substrate for a coating done elsewhere | PF1011 | Same Zerodur disc without a film, polished to 10-5 |
| UV plus a live power sample | Back-side-polished design, if available | Silver sampling mirrors do not cover 250 to 450 nm; a separate pickoff is the practical answer |
Sizing follows the beam as usual, with one ultraviolet-specific addition. A UV beam often has a small diameter and a high divergence, and the footprint on a 45 degree fold stretches by a factor of about 1.41, so a 6 mm beam covers roughly 8.5 mm of surface. That is comfortable inside 22.9 mm, but the coated zone is not the only limit: at these wavelengths a scratch, a pit or a dried droplet scatters far more efficiently than it does in the visible, and scatter that misses the target is lost throughput. Keeping the beam away from the edge, and keeping the surface clean, buys more performance here than moving up a reflectance class would.
Series Comparison
The PF1011 family puts several films on the same Zerodur disc, which makes the coating choice the only variable. The table is arranged from the shortest wavelength to the longest.
| Item | Coating | Band | Ravg | Typical role |
|---|---|---|---|---|
| PF1011-F01 | UV-enhanced aluminum, MgF₂ | 250 – 450 nm | 90% | Ultraviolet folding, thermal stability |
| PF1011-G01 | Protected aluminum, SiO₂ | 450 nm – 20 µm | 90% / 95% | General broadband folding |
| PF1011-P01 | Protected silver | 450 nm – 20 µm | 97% / 95% | Visible and near infrared, highest reflectance |
| PF1011-M01 | Protected gold | 800 nm – 20 µm | 96% | Infrared, chemically inert film |
| PF1011 | None (blank) | Not applicable | Not applicable | Substrate to be coated later |
Two comparisons matter in practice. Against PF1011-G01, the protected aluminum version, the difference is the band and the overcoat: the F01 reaches 100 nm further into the ultraviolet and uses a fluoride layer instead of silica, at the cost of a narrower total band and a softer surface. Against a dielectric UV mirror, the difference is angle and bandwidth: a metal film holds its reflectance over a wide range of incidence angles and a broad band, while a dielectric stack reaches higher reflectance inside a narrow design window and falls away outside it. A layout that folds a UV beam at 45 degrees with some wavelength spread is a metal-film job; a layout that folds one narrow line at normal incidence and needs the last few percent is a dielectric job.
FAQ
Q: Is the reflectance above 90% or exactly 90%?
A: The specification is 90% average across 250 to 450 nm, and it should be read as an average rather than as a floor at every wavelength. Individual points inside the band sit above and below it, and the curve typically dips near the short-wavelength end. For a link budget, use 90% and add margin if the working line sits near 250 nm.
Q: Will it work at 248 nm?
A: No, not within specification. The band begins at 250 nm, and the krypton fluoride line at 248 nm sits outside it, as does the argon fluoride line at 193 nm. Both need a different class of optic. This is the single most common mismatch in ultraviolet enquiries, and it is worth checking before anything else in the selection process.
Q: Why MgF₂ rather than the SiO₂ used on the visible coatings?
A: Because silica begins to absorb in the deep ultraviolet, and an overcoat that absorbs is worse than no overcoat at all. Magnesium fluoride stays transparent further into the ultraviolet, which is why it is the standard protective layer for UV aluminum. The price is mechanical: MgF₂ is softer and more sensitive to moisture than silica, so a UV mirror is cleaned less often and stored drier than a visible one.
Q: Why does the substrate matter so much in the ultraviolet?
A: Because two effects compound. A UV photon carries more energy, and the materials in the path absorb a larger fraction of the power, so the optic runs hotter per watt than it would at 1064 nm. At the same time, the wavelength is shorter, so the same physical surface distortion is a larger fraction of a wave. Heat produces more distortion and each unit of distortion costs more wavefront error, which is why a low-expansion substrate is worth its price here and often is not in the infrared.
Q: Can the coating be cleaned like a visible mirror?
A: More cautiously. The MgF₂ layer is softer than a silica overcoat, so the same drag-and-lift motion that is routine on a protected silver mirror is a risk here. Blow off dust first, use a solvent-moistened tissue with no pressure, work in one direction, and do not repeat the process as a habit. In a UV system, contamination control in the enclosure is worth more than any cleaning technique.
Q: Is the surface quality 10-5 or 40-20?
A: 40-20 on the coated mirror. The 10-5 figure that appears in the same family belongs to the uncoated PF1011 blank, which is inspected before deposition. A coated part is inspected after deposition and handling, and 40-20 is what the process delivers. Specifying 10-5 on a coated UV mirror is a common and avoidable purchasing error.
Q: How much power can it take at 355 nm?
A: The family table quotes 0.3 J/cm² at 355 nm for the pulsed rating, and the applicable number depends on pulse length, repetition rate and spot size, so confirm the test conditions against the current datasheet before designing to it. As with every metal film, the practical limit is often contamination rather than the coating: a UV-absorbing film of oil or dust on the surface will fail long before the coating reaches its rated fluence.
Usage Precautions
- Handle by the rim with gloves or finger cots. Skin oils are strongly absorbing in the ultraviolet, so a fingerprint that would be a cosmetic nuisance on a visible mirror is a local hot spot and a damage initiation site here.
- Keep the optic sealed until it is mounted, and keep the enclosure clean afterwards. Ultraviolet systems lose throughput to contamination faster than infrared ones, because films that are transparent at 1064 nm absorb at 355 nm.
- Clean sparingly and gently. The MgF₂ overcoat is softer than a silica layer, so blow off loose dust first and drag a solvent-moistened tissue in one direction without pressure.
- Store dry. Magnesium fluoride is more moisture-sensitive than silica, and a mirror stored in a humid room can show haze that no amount of cleaning will remove.
- Do not over-clamp. A 6 mm Zerodur disc in a metal ring bends, and the λ/10 figure is measured before the ring is tightened; a bent mirror in a UV imaging path shows up as astigmatism in the image.
- Block the rear scatter. The fine-ground rear face spreads the transmitted light rather than forming a clean ghost, but in a UV enclosure that stray light still reaches the walls and can fluoresce, which is a source of background in an imaging system.
- Ventilate or purge the beam path if the process generates ozone or organic vapour. A UV beam in air produces both, and both attack an aluminum film far faster than clean dry air does.
- Verify the wavelength before installation. A 248 nm excimer routed onto this mirror by mistake will produce a dim beam and a damaged coating, and the two symptoms are easy to confuse with a failing laser.
Disclaimer
Specifications on this page are compiled from the manufacturer published series tables and are offered for selection purposes only; the current official datasheet governs, including the test conditions behind the damage rating. Coatings, dimensions and markings may change without notice. Price and delivery are quoted separately by our sales desk.
Contact Us
Tell us the exact wavelength, the pulse length and repetition rate if the source is pulsed, the beam diameter at the mirror, the angle of incidence and how much focus drift the process tolerates, and we will confirm whether UV-enhanced aluminum is inside its band for your line or whether the wavelength has moved below 250 nm and needs a different coating. If the same layout also carries a visible alignment beam, say so, because that usually means two mirrors rather than one. Project and bulk quantities are quoted individually.









Reviews
There are no reviews yet.