QQ: 859818709

Thorlabs PF20-03-M01 Ø50.8 mm (2 in) Protected Gold Mirror, 800 nm – 20 µm

Thorlabs PF20-03-M01 Ø50.8 mm (2 in) Protected Gold Mirror, 800 nm – 20 µm

Price on request
Category:

Description

Product Overview

PF20-03-M01 is a 50.8 mm (2 in) protected gold mirror on a 12.0 mm fused silica substrate. Its specified band runs from 800 nm to 20 µm, its average reflectance across that entire span is 96 percent, and its surface figure is lambda/10 peak-to-valley at 633 nm. The number that matters commercially is not 96 – it is the fact that the same number applies at 1 µm, at 3 µm and at 10 µm, which means an infrared bench can be built once and re-used for a different experiment next month without touching the mounts.

One optic, four wavelength windows, no realignment

Most infrared laboratories do not run a single wavelength. They run a pump laser near 1 µm, they run an eye-safe or telecom beam somewhere between 1.3 and 1.6 µm, they generate mid-infrared light by difference frequency or in an optical parametric oscillator, and then they detect at 3 to 5 µm or illuminate a sample with a quantum cascade laser at 8 to 12 µm. A dielectric mirror optimised for any one of those windows is a poor reflector for the others, and every coating change is a realignment. A broadband metal coating removes that entire class of work from the schedule.

Window Typical source Why the same mirror works
0.8 – 1.1 µm Nd:YAG, fibre pump, Ti:sapphire Gold is already above 95 percent here; the 800 nm edge is the design boundary
1.3 – 1.6 µm Telecom diode, eye-safe source Flat region of the curve, no absorption feature to dodge
2 – 5 µm OPO idler, DFG output, QCL Metal reflectance is essentially wavelength-independent once past the interband edge
8 – 12 µm CO2 line, thermal source, FTIR 96 percent still applies at the far end of the band

What a flat curve is worth in a real build

An engineer choosing between coatings is often tempted to take the highest peak reflectance on the datasheet. That is the wrong optimisation for a multi-wavelength path. What you want is a reflectance you can predict without measuring, because every optic in the train is multiplied together: eight folds at 96 percent leave 72 percent of the light, eight folds at 99 percent leave 92 percent, but eight folds of a coating that is 99 percent at the design line and 85 percent two hundred nanometres away leave you guessing, and guessing is what costs a week of alignment time. A flat 96 percent is a budgeting tool. It lets you predict end-of-path power before you build, and it lets a second experimenter use the same hardware without re-deriving the loss budget.

Why the band starts at 800 nm and not lower

Gold reflects by free-electron response, and that response is screened by interband transitions at shorter wavelengths. Below roughly 800 nm the interband absorption rises steeply and the reflectance falls away, which is why this coating is not offered as a visible mirror and why nobody should expect it to fold a 633 nm alignment beam efficiently. The 800 nm boundary is a physical property of the metal, not a manufacturing choice, and no overcoat can move it.

The overcoat, and the price of being cleanable

Thorlabs deposits a silicon dioxide layer over the gold. It is what makes the surface robust enough to be handled and cleaned, and it is what allows a 96 percent figure to be specified as a repeatable product rather than a laboratory curiosity. The manufacturer is also candid about the cost, and the wording deserves to be repeated accurately: the protective overcoat introduces polarization changes that are difficult to quantify, and it can add absorption in the mid and far infrared. For work where the polarization state after reflection must be known to a fraction of a degree, or where every fraction of a percent of throughput in the thermal infrared counts, unprotected gold is the correct choice and this coating is not.

Key Features

  • 96 percent average reflectance from 800 nm to 20 µm – one flat specification covering pump, telecom, mid-infrared and thermal bands, so a single optic type can stock an entire infrared laboratory.
  • Silicon dioxide protective overcoat – a handling-tolerant surface, at the documented cost of some unquantified polarization change and possible mid-infrared absorption.
  • 50.8 mm diameter, 12.0 mm thick fused silica – the thick substrate resists mount-induced figure change and spreads thermal load laterally; CTE is 0.55 x 10-6 per degree C.
  • Lambda/10 peak-to-valley at 633 nm, 40-20 scratch-dig – a laser-grade figure and cosmetic grade, verified on the same fused silica blank used across the whole family.
  • Clear aperture 45.7 mm – 90 percent of the outside diameter, enough for a collimated 35 to 40 mm infrared beam with steering margin.
  • Pulsed damage threshold 2 J/cm² at 1.064 µm – measured at 10 ns, 10 Hz on a 1.000 mm spot, which corresponds to roughly 16 mJ on that spot.
  • CW thresholds 500 W/cm at 1.070 µm and 750 W/cm at 10.6 µm – stated per centimetre of spot, so they must be converted with your own beam diameter before use.
  • Angle of incidence 0 to 45 degrees, parallelism 3 arcmin or better – usable as a fold or a pickoff, with a wedge small enough that the rear-surface ghost does not return to the source.
  • Back-surface laser marking – the coating code is engraved on the rear face, which is how a gold mirror is distinguished from a silver one in a drawer.

Technical Specifications

Item Value Note
Coating Protected gold, SiO2 overcoat Cleanable, unlike unprotected gold
Band 800 nm – 20 µm Lower edge set by gold interband absorption
Average reflectance 96 percent Average across the band, not a peak value
Substrate Fused silica CTE 0.55 x 10-6 per degree C
Diameter 50.8 mm (2 in) Tolerance +0.0 / -0.1 mm
Thickness 12.0 mm (0.47 in) Tolerance plus or minus 0.2 mm
Clear aperture 45.7 mm 90 percent of diameter
Flatness lambda/10 P-V at 633 nm Over the clear aperture
Surface quality 40-20 scratch-dig Laser grade
Parallelism 3 arcmin or better Limits rear-surface ghost walk
Angle of incidence 0 to 45 degrees S and p reflectance diverge as angle opens
Marking Part number on rear face Standard from 1/2 in diameter upward

Damage thresholds as published, and what they convert to

Condition Published value Converted at that spot size
1.064 µm, 10 ns, 10 Hz, spot 1.000 mm 2 J/cm² about 15.7 mJ per pulse
1.070 µm, CW, spot 0.089 mm 500 W/cm about 4.5 W of total power
10.6 µm, CW, spot 0.339 mm 750 W/cm about 25 W of total power

The two CW figures differ by a factor of 1.5 on paper, but the test spots differ by a factor of 3.8 in diameter. Read as total power on the stated spot, 750 W/cm at 10.6 µm is about five and a half times the 1.070 µm figure, not 1.5 times. The threshold belongs to the coating, but the power you can safely put on a mirror depends on how large a spot you give it.

Applications

  1. Optical parametric oscillator and difference-frequency benches. The pump enters near 1 µm, the signal and idler leave between 1.5 and 5 µm, and the same mirror folds all three. No coating change between the pump arm and the idler arm.
  2. Thermal imaging and FTIR relay optics. Broadband blackbody or globar sources have no narrow line to optimise for, so a flat average reflectance is the only sensible specification.
  3. CO2 laser beam delivery in research settings. 10.6 µm sits comfortably inside the band, and the 12 mm substrate tolerates the modest thermal load of a laboratory-scale CO2 beam.
  4. Multi-source infrared microscopes. When a QCL, a thermal source and a 1.5 µm alignment laser share a single optical path, this is the only coating family that does not force a compromise.
  5. Laser rangefinder and eye-safe transceiver benches. 1.5 µm transmit and receive paths fold on the same mirror without a dielectric stack that would have to be angle-tuned.

Selection Guide

Work through these in order. The first question that gets a yes usually decides the coating.

If the requirement is Choose Because
Any wavelength below 450 nm PF20-03-F01, UV enhanced aluminum Gold and silver both absorb in the ultraviolet
450 nm to 2 µm, highest reflectance PF20-03-P01, protected silver 97 percent beats gold across the visible
Several bands from 800 nm to 20 µm PF20-03-M01, this optic One flat 96 percent curve, one alignment
Polarization state must be preserved PF20-03-M03, unprotected gold No overcoat, so no unquantified polarization change
Everything is above 2 µm PF20-03-M02, MIR enhanced gold 98 percent instead of 96 percent in the thermal infrared
One CO2 line at 10.6 µm only PF20-03-L01, line optimised gold 99 percent at that single line

Two secondary checks are worth making. First, does the beam ever run below 800 nm for alignment purposes? A helium-neon or diode pointer at 633 nm will be visibly dim on this mirror and should not be used to judge throughput. Second, will the optic be cleaned? If not, unprotected gold offers a percentage point more reflectance and no overcoat absorption; if it will be handled by students, the overcoat is worth more than the point.

Series Comparison

Same 50.8 mm diameter, same 12.0 mm fused silica substrate, same lambda/10 figure across all seven coatings. Only the coating differs.

Coating Band Ravg Trade-off you accept
-F01 UV enhanced aluminum 250 – 450 nm 90 percent Stops at 450 nm, no infrared use
-G01 protected aluminum 450 nm – 20 µm 90 / 95 percent Lower reflectance than gold above 2 µm
-P01 protected silver 450 nm – 20 µm 97 / 95 percent Silver tarnishes; overcoat is load bearing
-M01 protected gold 800 nm – 20 µm 96 percent Some polarization change, possible MIR absorption
-M02 MIR enhanced gold 2 – 20 µm 98 percent Near-infrared performance deliberately sacrificed
-M03 unprotected gold 800 nm – 20 µm 97 percent Surface is soft and must not be cleaned
-L01 unprotected gold, 10.6 µm 10.6 µm line 99 percent Narrow line only, soft surface

FAQ

Should a polarization-sensitive system use M01 or M03?

M03, unprotected gold, if the polarization state after reflection has to be known accurately. Thorlabs states that the protective overcoat on the protected gold coating causes polarization changes that are difficult to quantify, and that it may add absorption in the mid and far infrared. Difficult to quantify is the operative phrase: it does not mean the effect is large, it means you cannot compute it from a datasheet, which is unacceptable in an ellipsometry bench or a polarization-entangled source. M01 remains the right answer when the optic must survive cleaning, when the environment is dirty, or when the beam is unpolarised anyway.

Why not just use protected silver, which is one point higher in the visible?

Because 97 percent at 600 nm is irrelevant if your beam is at 4 µm, and silver’s advantage disappears in the infrared while its tarnish sensitivity does not. Gold is chemically inert; a protected gold mirror that sits in a drawer for three years is the same mirror. Silver depends on its overcoat for its entire life.

How much power can this mirror take?

The published CW values are 500 W/cm at 1.070 µm and 750 W/cm at 10.6 µm, both per centimetre of spot diameter, not per optic. Multiply by your own spot diameter in centimetres to get watts. A 3 mm spot at 10.6 µm therefore supports roughly 225 W by that linear rule, which is far beyond what a 2 inch optic can dissipate in most laboratory mounts. The practical limit is usually thermal distortion of the mount, not coating damage.

Does the 12 mm thickness cause any optical problem?

No, for the reflected beam. There is a weak reflection from the rear surface, and with parallelism held to 3 arcmin or better that ghost walks away from the incoming beam rather than back into the source. If the rear reflection is a problem in a sensitive interferometer, tilt the optic by a degree or two and it will leave the aperture.

Can it be used at 45 degrees?

Yes, the specified angle of incidence range is 0 to 45 degrees. Be aware that at 45 degrees the s and p reflectances separate, and for a metal the s component is the higher one. A beam that is folded several times at 45 degrees will accumulate a polarization-dependent loss that the average reflectance figure does not describe.

Is 800 nm a hard limit?

It is the edge of the specified band, and gold reflectance falls steeply below it because of interband absorption. Between 780 and 800 nm the coating still works, but it is outside the guarantee and the curve is moving fast. For a 780 nm diode, use protected aluminum or protected silver instead.

Usage Precautions

  • Clean only with reagent-grade solvents. The manufacturer’s handling guidance permits reagent-grade acetone, methanol or isopropanol applied with lens tissue; verify against the coating’s own handling notes before cleaning. Do not improvise with ethanol, household cleaners or abrasive wipes.
  • Drag, do not rub. One pass from centre to edge with a fresh tissue, then evaluate. Repeated rubbing on a soft metal film removes the overcoat before it removes the contamination.
  • Do not stack optics. Gold surfaces mark permanently at the slightest contact, and a mark inside the clear aperture is a permanent scatter site.
  • Mount on the edge, never on the face. A 2 inch, 12 mm optic weighs enough that a face-clamping mount will chip the edge; use a retaining ring with a nylon tip and finger torque.
  • Keep the alignment laser out of the specification. A 633 nm pointer reflected off this mirror looks dim and slightly coloured; that is the band edge, not a defective coating.
  • Record the orientation. The coated face is the front face; installing a mirror backwards in a fold produces a beam that appears to work but has a doubled ghost and a strange polarization.
  • Enclose invisible beams. At 1.5 µm and beyond the beam is not visible and the 4 percent that is not reflected still travels somewhere.

Disclaimer

All values on this page are taken from the manufacturer’s published specification for PF20-03-M01. Average reflectance is an average over the stated band and does not predict performance at a particular wavelength or angle. Damage thresholds are measured values obtained on clean optics under the conditions stated and do not guarantee survival in any specific system; beam profile, pulse duration, repetition rate, spot size, contamination and mounting all affect the real limit. The polarization and mid-infrared absorption behaviour of the overcoat is described qualitatively by the manufacturer and is not specified numerically. Gaofei Laser supplies this optic as a distributor; confirm current specifications with us before design freeze where a value is critical.

Contact Us

Tell us the wavelengths in your path, the pulse energy or CW power at the mirror, and the angle of incidence, and we will confirm the coating choice or recommend a different one from the family.

Hotline: +86 159-2722-0467
Email: 859818709@qq.com
Website: www.soar-laser.cn

Reviews

There are no reviews yet.

Be the first to review “Thorlabs PF20-03-M01 Ø50.8 mm (2 in) Protected Gold Mirror, 800 nm – 20 µm”

Your email address will not be published. Required fields are marked *

中文EN