Procedures/ Methods
Three categories of testing were performed to measure the resistance of each type of Moxtek X-ray window to harsh environments: 1. Exposure to high temperatures and then to 1% hydrochloric acid 2. Exposure to a variety of acids at room temperature 3. Exposure to a variety of petroleum based fuels at room temperature Bare uncoated beryllium, DuraBeryllium, and DuraBeryllium Plus windows were used in each test. Each window consisted of a 100 μm thick beryllium foil. The DuraBeryllium had a chemically resistive coating (DuraCoat) applied to the beryllium foil. The DuraBeryllium Plus had two different coatings applied to the beryllium foil (DuraCoat and Plus).High Temperature Bake Followed by 1% HCl Exposure Tests
Most applications require the X-ray windows to be baked at high temperatures in assembling X-ray detectors. The first test was designed to see how chemically resistive the windows were after being baked. Windows were baked at 400°C, 450°C, and 500°C for five hours. One group was baked in air and another in vacuum. After being baked, a solution of 1% hydrochloric acid was dripped on top of the window for a period of time and then checked for leaks on a helium leak detector. A window was considered to have failed if a leak rate greater than 1 x 10-10 mbar*L/s was observed. The windows were repeatedly exposed to acid and checked for leaks until they failed.
5% Acid Exposure Tests
5% concentrations of hydrochloric acid, sulfuric acid, and hydrofluoric acid were dripped on X-Ray window for a time period followed by a leak check then repeated until found leaky (>1 x 10-10 mbar*L/s).
Petroleum Based Fuels Test
The final test consisted of exposing the X-ray windows to a variety of common petroleum based fuels. Windows were dipped into regular unleaded gasoline, diesel fuel, and kerosene for month long intervals and then checked for leaks on a helium leak detector.Experimental Results
High Temperature Bake Followed by 1% HCl Exposure Results
The following two graphs show the results of the X-ray windows exposed to high temperatures and to 1% hydrochloric acid.
5% Acid Exposure Results
The graphs below show the results of the windows exposed to 5% concentrations of hydrochloric acid, sulfuric acid, and hydrofluoric acid.

Petroleum Based Fuels Results
X-ray windows have been exposed to petroleum based fuels for a total of eight months thus far in which no measurable change in leak rate has been observed in any of the X-ray windows tested.Conclusions
DuraBeryllium Plus windows demonstrated the greatest resistance to a variety of corrosive elements. This resistance was seen even after the parts have been baked at temperatures up to 400°C. In detector manufacturing processing under 400°C will fully preserve the chemical resistivity of the DuraBeryllium Plus windows. DuraBeryllium windows, while not as resistive as DuraBeryllium Plus windows, do demonstrate significant resistance to acid corrosion than do bare beryllium windows in non-baking applications and in those baked up to 400°C in vacuum. Both coated and non-coated X-ray windows have relatively high resistance to corrosion from petroleum based fuels. Applications that require windows to be exposed to high temperatures or to be exposed harsh environments will find the best performance from Moxtek’s DuraBeryllium Plus windows due to its greater chemical and temperature resistance.Performance Improvement
Figure 1, depicts the light path for a simplified spectrophotometer with polarizing analyzer. This type of arrangement is useful when examining dichroic and birefringent samples, diffraction gratings, or when characterizing samples in reflectance. Typical polarizers used in this analyzer application are based on the Glan-Taylor (GT) or Glan-Thompson (GTh) designs, which generally utilize Calcite prisms. Unfortunately, inherent absorption in calcite and scattering from inclusions and impurities can severely limit UV transmittance and achievable signal to noise ratio in GT and GTh designs. The ProFlux UVD Series polarizers by Moxtek offer an alternative with dramatically improved deep UV transparency for better signal to noise ratio performance. Figure 2a compares the UV passing state transmittance for a standard calcite GTh analyzer and the Moxtek UVD260, while figure 2b compares their performance when measuring the blocking state transmittance of a reference polarizer. The GTh analyzer shows a dramatic reduction in UV light throughput resulting in a poor signal to noise ratio, while the UVD260 is fabricated on fused silica and maintains excellent performance well into the deep UV.
Figure 2a: Performance comparison of ProFlux UVD260 and Glan-Thompson polarizers in an analyzer application. Analyzer passing state UV
transmittance comparison.

Figure 2b: Performance comparison of ProFlux UVD260 and Glan-Thompson polarizers in an analyzer application. Reference part measurement.
Infrared Performance
The Glan-Taylor (GT) and Glan-Thompson (GTh) polarizer designs consist of two birefringent prisms mounted with their diagonal faces either separated with a small air gap (GT), or filled with an optical cement (GTh). The separation of an incoming beam into orthogonal polarization states relies on total internal reflection and thus imparts strict requirements on beam collimation and entrance angle. Since refractive index generally decreases in the infrared, this results in a decrease in the allowed deviation from normal incidence for the GT and GTh polarizer designs. For spectroscopic applications, where beam collimation is usually far from ideal, this results in significant leakage of the unwanted polarization state. While the GT and GTh polarizer designs can typically only tolerate a few degrees of entrance angle misalignment or a narrow field (cone) angle before performance deteriorates, wire-grid polarizer performance is relatively angle and wavelength insensitive and shows practically no IR leakage of the unwanted polarization state. The ProFlux UVD designs can easily accommodate ±20 degree variations from normal incidence with minimal performance variation. This corresponds to a field angle of 40 degrees, which allows for dramatically improved light utilization when using poorly collimated sources and eases any alignment concerns. Figure 3a depicts the dramatic difference in infrared performance between a GTh polarizer and the ProFlux UVD260 when analyzing the same part (another UVD260). In addition to the increased leakage at longer wavelengths due to the limited field of view of the GTh design (see inset), there are strong peaks, likely due to IR absorption resonances from calcite and impurities (e.g. moisture). The Kramer-Kronig relationship dictates that any absorption resonance is also accompanied by an incongruity in the index of refraction, which should manifest itself in the field acceptance angle. By comparison, the UVD260 analyzer shows no such infrared leakage or resonance peaks due to its sub-wavelength grating design and fused silica substrate. The transmittance of the GTh and UVD260 polarizers in the passing state are depicted in Figure 3b. The same absorption features responsible for the IR leakage peaks in the GTh blocking state measurement are also apparent here in the passing state. The moisture absorption line at ~2725 nm is much stronger in the hygroscopic calcite material of the GTh than in the UVD260, which is composed of aluminum and infrared grade fused silica. For enhanced short-wavelength infrared throughput and signal to noise, the UVD260 is clearly superior.
Figure 3a: Infrared performance comparison of UVD260 and calcite Glan-Thompson analyzers. Blocking state transmittance measurements of the same reference polarizer using GTh (—-) and UVD260 (—-) analyzers. Inset shows magnified scale away from absorption resonances.

Figure 3b: Infrared performance comparison of UVD260 and calcite Glan-Thompson analyzers. Blocking state transmittance measurements of the same reference polarizer using GTh (—-) and UVD260 (—-) analyzers. Inset shows magnified scale away from absorption resonances.

Figure 4: Contrast ratio between passing and blocking state transmittances for ProFlux UVD260 (—) and Glan-Thompson (—) polarizers.
Environmental and Form Factor Considerations
The UVD260 utilizes similar materials to Moxtek’s standard visible spectrum wire-grid polarizer products, which are recognized for their excellent sustained performance in high temperature and high humidity projection display applications. Furthermore, the buried nanowire design of the UVD series helps protect against handling damage and environmental contamination. Whereas the infrared performance of calcite GT and GTh polarizers can noticeably degrade with time due to moisture uptake, the UVD Series wire-grid polarizers are fabricated on non-hygroscopic fused silica and do not show significant performance degradation in humid environments. Due to the total internal reflection operating principal and large critical angle, GT and GTh polarizers have large aspect ratios between their length and clear aperture dimension. As aperture size increases, this aspect ratio requirement ensures the prism-based designs take up much more space in an optical system than the planar wire-grid polarizer configuration, which has a fixed thickness (usually 2.1 mm) set by the substrate choice and spacer thickness. For the UVD series wire-grid polarizers, the physical space required along the beam propagation direction remains fixed, regardless of the choice of aperture size. As a side effect, for larger aperture size GT and GTh designs, the length of the prisms also increases, which degrades performance in the UV and short-wave IR regions due to absorption and scattering. ProFlux UVD Series polarizers are capable of covering the entire spectral range of most spectrometers from deep UV through shortwave IR using the same part and without having to discard light from the larger field angles required when utilizing finite-sized broadband light sources. For GT and GTh polarizer designs, in order to dump the internally reflected beam, the prism sides have either an absorptive coating or are highly polished and housed in an absorptive case. However Fresnel reflections can still occur, leading to leakage of the unwanted polarization state through the exit face of the polarizer. ProFlux UVD Series polarizers separate the beams at the wire-grid surface by an anisotropic absorption and reflection mechanism, and do not rely on birefringence and total internal reflection. This eliminates the long optical path length and absorptive surface/casing requirements inherent in GT and GTh products as well as the unwanted IR performance variation with wavelength. Table 1 summarizes the design, form factor and environmental differences between wire-grid and Glan-prism based designs.| Feature | ProFlux UVD260 | ProFlux UVD240 | Glan-Taylor / Glan-Thompson |
|---|---|---|---|
| Angle of Incidence, AOI | ±20° | ±20° | ±4 / ±6° |
| Length | 2.1mm (aperture independent) | 2.1mm (aperture independent) | Scales with aperture size |
| IR Disturbance | none | none | Dispersion & absorption induced |
| Fresnel Disturbance | none | none | Need absorptive coating / case |
| Spectral Range | 260-3300nm | 240-3300nm | Broadband performance often requires two sets of polarizers |
Conclusion
When compared to Glan-Taylor and Glan-Thompson designs, the ProFlux UVD Series polarizers provide superior broadband performance for polarization sensitive spectroscopic applications. The aluminum Nanowire grid design on fused silica substrate is deeply sub-wavelength, providing excellent contrast and outstanding passing state transmission with minimal performance variation from the UV to the short-wave IR. The wide acceptance angle and space-saving form factor improve performance and efficiency while easing system design. See Figure 5 below for typical broadband performance plots.Introduction
This technical brief compares imaging needs and how Moxtek is improving its products. It explains how competing technologies in LCoS projectors compare in terms of brightness, performance, durability and reliability.PBS Efficiency
Efficiency (Tp*Rs) for a polarizing beamsplitter is a measure of how perfectly a polarizer converts randomly polarized light into (reflected) s and (transmitted) p polarized light. If all of the light is converted, then the beamsplitter would be 100% efficient. In reality, some of the light is absorbed, some ‘s’ is transmitted and some ‘p’ is reflected, reducing the efficiency. The Moxtek standard PBS is typically 81% efficient at 550nm wavelength. Recently, with improved manufacturing techniques and new wire-grid technology, this efficiency has been dramatically improved. Figure 1 shows the improvement in efficiency for the Moxtek High Efficiency PBS (HEPBS) versus standard PBS. This represents a 10% improvement.| Element | ProFlux Wire Grid | Dichroic Polarizer | MacNeille Polarizer |
|---|---|---|---|
| Polarizing Mechanism | Polarizer using a nano- structured thin film | Resonant absorption in thick stretched film | Reflects s-polarization and transmits p-polarization |
| Significance | A perfect polarization match to the LC imager, even skew rays can align to the LC. PBS: f/1.5 and below. PPL: f/1.0 and below. | Thick polarizing film results in beam depolarization | Polarization direction depends on incoming ray: skew rays are not aligned to LC |
| Contrast | Excellent brightness, contrast, and on-screen uniformity | Reduced contrast and uniformity | Reduced contrast and uniformity |
| Durability | Made from all inorganic materials for high durability | Polymer film degrades over time and in high flux conditions | Requires polymer waveplate for correct polarization |
Comparison of Other Technologies
Table 1, “Comparison of Technologies”, shows a comparison of wire-grid polarizer, stretched polymer absorbing films (dichroic polarizers), and Brewster’s angle polarizers (MacNeille cubes). Flatter response of ProFlux polarizer across both wavelength and angle creates a more uniform picture from the projector. The wide acceptance angle allows for small f/# optical designs while maintaining color and contrast uniformity. Figure 2 shows an angular map comparison of the Proflux Beam Splitter and the MacNielle Cube.
Figure 2: Angular Map Comparison between a Moxtek Wire-Grid Beamsplitter measured in reflection at 550nm (left), and a MacNeille Cube measured in reflection at 550nm (right).





