Sponsored by AlluxaDec 20 2016
Thin film optical filters have a complex nature and diverse end-use applications and therefore, they have always been a challenge to specify. This article discusses the major types of filters and provides guidelines on how to specify key features, while also optimizing the cost.
The types of filters discussed in this article include
- Dichroic and polychroic tilted beamsplitter filters
- Band pass types of ultra-square high cavity count, ultra-narrow and soft coating replacements
Substrates
NBK7, Borofloat 33, and fused silica (or equivalents) are the most commonly used substrates for thin film filters. Alluxa strongly discourages using “green” float glass and color glass except for extremely low cost or specialty applications. For most applications, the thin film equivalent of a color glass is lower cost and higher performance.
Color glass is available in limited sizes and shapes, which limits the effective load sizes of each coating lot. Green glass is affected by scatter, absorption and auto-fluorescence and other “low quality” flaws. It is recommended to use laminated filters, which sandwich multiple green glass substrates between fragile and environmentally sensitive coatings, only for laboratory scale demonstrations.
Precision filters have to be deposited on flat substrates. Although curved substrates can be used, the introduction of thickness variations by them degrade performance.
Background: Light Sources, Detectors
It is obvious that a filter need not be specified in wavelength for blocking range, reflection range, or transmission range beyond either the detector range or the light source range. For instance, photomultipliers and silicon detectors have very defined response curves, and nothing is gained other than added cost by specifying past its response curve.
Likewise, there is typically no advantage to specifying the blocking range past the relatively narrow emission spectrum for LED and laser diode “clean up” filter applications.
Blocking and transmission must be specified as average levels unless a discrete light source such as a laser is required to be blocked or transmitted. Thin films filters often exhibit blocking levels and very narrow spikes can easily rise 1 or even 2 OD above the average blocking levels.
However, spikes have a very narrow spectral bandwidth and therefore, they do not have a meaningful effect on blocking level performance or signal to noise ratio.
Bandpass Filters
Bandpass filters allow a spectral band to transmit while rejecting light outside of that band. Traditionally, these filters were specified with half-power bandwidth (HPBW), center wavelength, and peak transmission. Providing a range of wavelengths to transmit, the level of transmission, and a range of wavelengths to reject is a more precise and modern way to specify these filters.
The gap between these ranges is used as a single specification to capture the tolerance of the filter for centering, uniformity, filter slope, and manufacturing margin.
Bandpass filters are typically used at near zero angle of incidence (AOI). Angles greater than zero deteriorate performance due to the angle sensitivity of the filter. Filters shift with angle approximately proportional to the square of the angle and the inverse of the square of the effective index.
The larger the angle of incidence, the greater the issue the angle creates. They also create polarization splitting where the S & P polarization states exhibit different performances as a function of angle within the filter design. As each polarization state creates a different filter function with increasing angle, polarization splitting generally degrades filter performance.
Bandpass filters are available in two types, a long pass/short pass structure or a resonant cavity Fabry-Perot structure. ‘Squareness’ of a bandpass filter is a function of number of cavities for resonant cavity type filters and the number of layers for a LWP and SWP type filter.
The impact of increasing the number of cavities is shown in Figure 1 and the typical performance for a wide band high cavity count filter is shown in Figure 2. A flat top narrow band cavity filter is depicted in Figure 3.
Figure 1. Squareness of narrow band pass filter is a function of number of cavities.
Figure 2. Typical high cavity count thin film cavity band pass filter.
Table 1. Bandpass filter specification guidelines
Term/ Parameter |
Description |
High Performance |
Standard |
Lowest Cost |
Center wavelength |
Center of passband. This should be used only as a nominal value |
+/- 0.25% of cwl |
+/- 0.5% of cwl |
+/- 1% of cwl |
Transmission |
Average transmission over desired band |
>95% |
>90% |
>85% |
Blocking ranges |
Range of wavelengths required to suppress |
200 nm - 1200 nm |
300 nm - 1100 nm |
Optimized for detector and light source |
Passband to blocking band delta |
Spectral gap between the passband and blocking band giving tolerance for slope, centering, etc. |
<1% of wavelength |
1% to 3% of wavelength |
>3% of wavelength |
Blocking levels |
Blocking suppression levels in log units average over the band |
6 OD average, with 10 OD in specified bands |
5 OD average |
4 OD average integrated over light and detector |
AOI and cone angle |
Range of angles about the primary AOI |
Normal AOI with <10 degrees cone angle |
Normal AOI with <10 degrees cone angle |
Normal AOI |
Ultra-Narrow Bandpass Filters
Ultra-narrow bandpass filters are typically the type of bandpass filters with bandwidths less than 1% of wavelength. Key attributes are blocking levels, bandwidth, and transmission. For narrow bandpass filters, most applications are designed to transmit a single wavelength such as an emission line or a laser.
Figure 3. Ultra-narrow and flat top band pass filter with bandwidth <1 nm at 532 nm.
Table 2. Ultra-narrow band pass filter specification guidelines
Term/ Parameter |
Description |
High Performance |
Standard |
Lowest Cost |
Center wavelength |
Center of passband. This should be used only as a nominal value |
+/- 0.1% of cwl (nominal) |
+/- 0.25% of cwl |
+/- 0.5% of cwl |
Transmission |
Transmission at the desired line |
>95% |
>90% |
>80% |
Bandwidth |
FWHM |
<0.2% of CWL |
<0.5% of CWL |
<1% |
Blocking ranges |
Range of wavelengths required to suppress |
200 nm - 1200 nm |
300 nm - 1100 nm |
Optimized for detector and light source |
Passband to blocking band delta |
Spectral gap between the passband and blocking band giving tolerance for slope, centering, etc. |
<0.5% of wavelength |
<1% of wavelength |
>2% of wavelength |
Blocking levels |
Blocking suppression levels in log units average over the band |
6 OD average |
5 OD average |
4 OD average integrated over light and detector |
AOI and cone angle |
Range of angles about the primary AOI |
Normal AOI with <10 degrees cone angle |
Normal AOI with <10 degrees cone angle |
Normal AOI |
Multi-Bandpass Filter
A multi-bandpass filter is a band pass filter capable of transmitting multiple spectral regions or multiple pass bands, with blocking regions in between the transmitted regions as illustrated in Figure 2. The guidelines outlined for single bandpass filters can be followed to specify multi-bandpass filters. However, it is necessary to carefully take into consideration the tolerances in order to reduce cost.
In general, depositing a multi-bandpass filter is a more challenging task than creating two single bandpass filters separately.
Figure 4. Multi-bandpass filter theory and measured.
Table 3. Multi-band pass filter specification guidelines
Term/ Parameter |
Description |
High Performance |
Standard |
Lowest Cost |
Center wavelength |
Center of passband. This should be used only as a nominal value |
+/- 0.4% of cwl |
+/- 0.75% of cwl |
+/- 1% of cwl |
Transmission |
Average transmission over desired band |
>95% |
>90% |
>85% |
Blocking ranges |
Range of wavelengths required to suppress |
200 nm - 1200 nm |
300 nm - 1100 nm |
Optimized for detector and light source |
Passband to blocking band delta |
Spectral gap between the passband and blocking band giving tolerance for slope, centering, etc. |
<1% of wavelength |
1% to 3% of wavelength |
>3% of wavelength |
Blocking levels |
Blocking suppression levels in log units average over the band |
6 OD average, with 10 OD in specified bands |
5 OD average |
4 OD average integrated over light and detector |
AOI and cone angle |
Range of angles about the primary AOI |
Normal AOI with <10 degrees cone angle |
Normal AOI with <10 degrees cone angle |
Normal AOI |
Soft Coating Replacement Filters
“Soft coatings”, also called “laminated coatings,” have high optical scatter, low transmission, poor temperature stability, poor environmental durability, and low blocking levels, but they are still widely used due to their ready availability, low cost, and legacy status.
With proprietary high speed plasma deposition technology, Alluxa is able to deliver the durability and optical performance of hard coated thin film optical filters at laminated soft coating pricing.
Figure 5 shows the typical performance of soft coating replacement filters for various wavelengths from the blue to the NIR. These filters are normally deposited on borofloat for cost purposes. For these filters, price is generally the key parameter and Table 1 price sensitive column is most appropriate.
Figure 5. A variety of soft coating replacement filters. The bandwidth is typically between 10 and 50 nm with blocking > OD4. Substrates are borofloat to reduce cost.
Tilted Filters – Dichroic and Trichroic Filter
Dichroic filters are the class of filters that are tilted, and transmit and reflect. Specifying dichroics and polychroics filters is more like multi-bandpass filters, and primarily involves choosing the desired blocking bands and pass bands, and substrate type and thickness.
The spectral gaps between the bands are the key driver of coating complexity and cost. Tilted filters have inferior performance compared to the filters used at normal incidence because of the complexity of suppressing polarization splitting as well as other effects.
Figure 6. Multiband dichroic beamsplitter used at 45 degrees vs. theory
The standard angle for mounting is 45 degrees, but it is easier to achieve performance of slope, transmission, and blocking when the angle is lower. Similarly, the lower the angular range of the beam, and more collimated, the better the performance in general. A summary of recommended specifications for polychroic and dichroic filters is presented in Table 4.
Table 4. Summary of specifications for dichroics and polychroics
Term/ Parameter |
Description |
High Performance |
Standard |
Lowest Cost |
Center wavelength |
Center of pass band. This should be used only as a nominal value |
+/- 1% (nominal) |
+/- 2% (nominal) |
+/- 3% (nominal) |
Pass band Transmission |
Transmission average across pass band |
>95% |
>90% |
>85% |
Pass band width |
Range of wavelengths required to transmit |
>1% wide 310 – 1100 nm |
>2% wide 310 nm – 1100 nm |
>4% wide 310 – 1100 nm |
Blocking band |
Range of wavelengths required to suppress |
>1% wide 310 – 1100 nm |
300 nm - 1100 nm |
Optimized for detector and light source |
Passband to blocking band delta |
Spectral gap between the passband and blocking band giving tolerance for slope, centering, etc. |
<1% of wavelength |
>2% wide 310 nm – 1100 nm |
Optimized for detector and light source |
Blocking levels |
Blocking suppression levels in log units average over the band |
1% average across blocking band |
5% average across blocking band |
10% average across blocking band |
Pass band to blocking band gaps |
Spectral gap between the pass band and blocking band giving tolerance for slope, centering, etc. |
<1% of wavelength |
1% to 3% of wavelength |
>3% of wavelength |
Substrate |
Thickness and type |
1 mm or less on fused silica |
2 mm on fused silica or polished borofloat |
Borofloat |
Flatness |
RMS flatness measured at 632nm using interferometer |
<0.1wave RMS/inch |
0.5 wave RMS/inch |
1 wave RMS/inch |
Conclusion
Modern hard coated optical thin film filters are nearly infinitely flexible in capability. Single band, multiband, and dichroic tilted filters all come with high transmission and outstanding blocking. However, thin film optical filters have always been a challenge to specify due to their complex nature and diverse end-use applications. Guidelines discussed in this article on how to specify key attributes and optimize the cost are helpful to designers in selecting specifications.
This information has been sourced, reviewed and adapted from materials provided by Alluxa.
For more information on this source, please visit Alluxa.