With the variety of choices available, it can be a daunting task to specify and choose the right optic for a given application. Manufacturers offer most optical components in an array of shapes and sizes and even off-the-shelf catalogue components have many features which should be taken into consideration before an order is placed. Specifications of significant concern include outer diameter, damage threshold, bandwidth, angle of incidence, polarization and reflectivity.
Figure 1. Round optics are easier and more economical to manufacture.
What may seem to be an easy decision can end up being the wrong decision if factors such as clear aperture, angle of incidence and access to suitable mounts are not taken into consideration. A typical 1-in.-diameter optic, for example, will have a clear aperture of about 22-mm at normal incidence, accommodating any beam smaller than that. The same optic used at an incident angle of 45° will only accommodate a round beam up to 15.5-mm in diameter, however. The obvious choice might be to use elliptical or rectangular substrates for all applications at 45°, but for substrates in shapes other than round, polishing costs are higher and the selection of mounts is limited, making them a poor choice for many applications (see figure 1). Unless there is a spatial constraint to the optical set-up, a suitably large round substrate can generally meet the requirements with lower cost and faster lead time.
Lenses and other curved substrates add another factor to the equation because radius of curvature also has to be considered. Using the simple lens equation for a plano-convex singlet lens where focal length f = R/(n-1) it is clear that for any material with an index near 1.5, like crown glass or fused silica, the focal length will be approximately twice the radius of curvature. Also, because the physical diameter of a positive singlet lens cannot be more than twice the radius of curvature, it follows that the diameter of a lens cannot be more than the approximate focal length of the lens, unless you switch to a higher index glass material like SF11. And again, fabrication costs (and therefore selling price) will vary depending on the size and shape of the lens. A short focal length lens which has to be polished one-at-a time will cost more than a lens with a longer radius of curvature that can be manufactured in lot sizes of 19 or more. Similarly it may be more economical to polish a batch of bi-convex lenses where R1 and R2 approximately equal the focal length of the lens than a batch of plano-convex lenses with a shorter radius that limits the lot size to only three or seven parts per polishing block. These are important concepts to remember when considering the price, quality and performance of a lens.
For imaging and laser focusing applications, surface quality and surface figure need to be clearly defined to meet damage threshold, wavefront, and scatter loss requirements. Most high energy laser systems require substrates with a surface quality of 10-5 or better, meaning that the polished surfaces will have no scratch greater than 10 μm in width and no dig greater than 5 μm in diameter. An uncoated fused-silica lens polished to a 10-5 surface can withstand energies of more than 40 J/cm2 at 1064 nm. Commercial grade optics with scratch/dig specs of 60-40 or 40-20 cost less and may be perfectly suitable for research and low-energy applications. Materials and coatings
The performance of an optic is also affected by reflected and/or transmitted wavefront distortion (TWD), which are primarily determined by the surface figure. For lenses, a true calculation cannot be done without taking into consideration diffraction limits and spherical aberration theory, but in general, a long focal length singlet lens will have less spherical aberration and therefore a better TWD than one with a short radius of curvature on one or both sides. Most laser-quality optics are polished to a surface figure of λ/10 waves p-v at 632.8 nm, or tenth-wave flatness. To achieve this on a window or other flat substrate, it is important to allow for enough physical thickness to maintain a suitably low diameter-to-thickness aspect ratio; for BK7 (borosilicate crown glass) the preferred maximum value is 6:1 and for fused silica it is 10:1. TWD is dependent on both material homogeneity and surface flatness so thinner substrates are better; one inch diameter substrates made from grade A BK7 glass are typically available with λ/10 TWD in thicknesses ranging from 1.5 to 10 mm.
Because of availability and optical quality, BK7 and fused silica have emerged as the most common substrate materials used in the ultra-violet (UV) to near-infrared (NIR) wavelength regions. For uses in the visible and NIR, BK7 is a low-cost glass with good homogeneity and good manufacturability, suitable for many applications as either a transmissive or reflective component. When an application requires increased thermal stability, very high damage threshold or exposure to UV light, fused silica provides the better option. Fused silica is an amorphous form of silicon dioxide glass and is readily available in a variety of grades and material classes with high transmittance as low at 193 nm. VUV grade calcium fluoride (CaF2) transmits from 157 nm to about 8 μm in the IR spectral region so is a good choice for either IR or deep UV applications; care must be taken when specifying this material however since issues such as birefringence, cost and surface quality limitations must all be taken into consideration.
Coatings can be used to enhance the performance and lifetime of an optical component. For transmissive optics such as lenses, windows, waveplates, and prisms, a thin-film dielectric anti-reflection (AR) coating can increase the overall transmission through the optic by as much as 4% per surface, while minimizing stray light and back reflections throughout the system.
The Fresnel equation gives the reflectance R in air (n = 1.0) off an uncoated glass surface at normal incidence as:
R = ((n - 1)/(n + 1))2
where n is the refractive index of the substrate material. For BK7 glass at 633 nm, n = 1.5151 so R = 4.2%; this results in a total loss of more than 8% of the initial power. An AR coating can reduce the reflectance per surface to 0.1 to 2.0%, depending on the design chosen.
Figure 2. Plot of reflection versus wavelength shows the performance of the three most common AR coating types: V-type, single layer magnesium fluoride, and broadband.
Three types of AR coating designs are most often used. Multilayer V-coatings, called V-coatings due to the V-shape of the reflectance are narrowband, durable and damage resistant (see figure 2). They can be designed for use at almost any UV/near-IR wavelength to achieve reflectance below 0.25% per surface at normal incidence. Broadband multilayer AR coatings are used to optimize transmission over several wavelengths or wavelength regions within a limited range. These coatings provide an average reflectance of less than 0.5% per surface over 100 to 200 nm, depending on the substrate material and wavelength region being covered, which improves overall transmission by more than 7% through two surfaces. The most broadband of the three readily available options is a single layer magnesium fluoride (SLMF) AR coating. With n = 1.38, the SLMF coating is an excellent and inexpensive coating choice for use on high-index materials such as sapphire and SF10 glass. Normal incidence reflectance of less than 0.25% can easily be achieved on these materials without sacrificing either durability or bandwidth.
Dielectric and metal coatings can be applied to flat and curved substrates in order to reflect, redirect or focus light. Metal coatings such as gold, silver and aluminum are primarily useful for reflecting very-broadband light sources with low energies (less than 100 mJ/cm2) over incident angles ranging from 0 to 60º. These coatings often have a protective dielectric layer to increase the durability and lifetime of the coating, and can be enhanced with a multilayer dielectric stack to increase reflectivity in the UV or visible regions. Metal coatings are still sensitive to moisture and/or scratches so are generally not used in abrasive environments or systems in which the optics require frequent cleaning or handling.
Figure 3. Gold, silver, and aluminum coatings provide broadband performance while dielectric coatings achieve very high reflectivity at a single wavelength or over a narrow bandwidth and incident angle.
High energy laser mirrors use all-dielectric coating stacks to achieve very high reflectivity at a single wavelength or over a narrow bandwidth and incident angle. Depending on the substrate material and wavelength of operation, damage thresholds of 10 to 20 J/cm2 and reflectivity of greater than 99.5% are standard for normal-incidence mirrors (see figure 3). At 45°, S-polarized light reflects more and is more broadband than P-polarized light, so it is important to identify the correct component and polarization state when ordering or determining optic specifications.
When bandwidth is a higher priority than damage threshold, alternate coating materials can be used in the visible and NIR to create highly reflective mirrors with low dispersion and minimal pulse distortion. Coating designs can also be modified to create mirrors with high reflectivity at two distinct wavelengths or over much broader bandwidths or angles of incidence than can be achieved with "single-stack" designs. Polarization Optics
Figure 4. A birefringent polarizer such as this beamsplitting cube divides an incident beam into two different polarizations.
When polarization is a concern, linear and birefringent polarizers can be used to remove, reflect, or refract one of the polarization components (see figure 4). The choice between a calcite polarizer, a thin film plate polarizer, or a polarizing beamsplitter cube depends on the bandwidth and energy levels of the input light. A cemented beamsplitter cube works well for low energy applications in which both components of light are to be utilized. These are available at many wavelengths from the UV to the NIR, in both narrow and broadband varieties with damage thresholds up to 100 mJ/cm2 and transmitted extinction ratios of 1000:1. Higher extinction ratios on the order of 105:1 over very broad bandwidths can be achieved using birefringent calcite or magnesium fluoride (MgF2) Rochon polarizers. Damage thresholds on these vary from 10 mJ to 1 J/cm2 depending on whether the rejected beam is absorbed or removed through an escape window, and whether the assembly design is cemented or air-spaced.
For fluences of a joule or more, a plate polarizer or an optically contacted polarizing beamsplitter cube can be used to avoid damage due to cement or tracking within calcite material; both types of polarizers reflect the S-polarized light and can be used for high energy laser applications. Plate polarizers feature exhibit group-velocity dispersion so are often used in ultrashort femtosecond systems, but usually need to be angle tuned to maximize transmission and extinction ratio. Polarizing beamsplitter cubes require no angle tuning and the transmitted and reflected beams are separated by 90° but are often more expensive and limited to dimensions of an inch or less.
To get the most from your optical component budget, give consideration to functionality and manufacturability as well as the mechanical and optical specifications of each component. oe
Emily Kubacki is a senior engineer in the sales and marketing group at CVI Laser LLC Albuquerque, NM.