1. Performance Specification

The performance specification of a filter defines its capabilities in a way that can be easily understood by system designers, customers, and manufacturers. Sometimes, the manufacturer creates the specification based on known achievable performance, either for a specific customer or as part of a standard product catalog. More often, the specification is drafted by the system designer to detail the required filter performance to achieve the desired system performance.

To develop a specification, the question of the filter’s purpose must first be clearly and concisely addressed. This forms the foundation of the performance specification. While some systems require a specific performance level to function effectively, others may simply aim for the best possible performance within cost or complexity constraints. In such cases, filter performance must be balanced against cost and system complexity to determine a reasonable specification. This process often involves design input, manufacturing knowledge, and collaboration between customer and manufacturer.

Example: Spectral Line Filtering

Consider the example of isolating a spectral line against a continuum. A narrowband filter is necessary, but the required bandwidth and filter type depend on various factors:

1. Energy Transmission:
   • The transmitted energy from the spectral line depends on the peak transmittance, assuming the filter can always be tuned to the line.
   • Energy from the continuum depends on the total area under the transmission curve, including the rejection regions far from the peak.
2. Bandwidth and Contrast:
   • Narrower passbands increase contrast between the line and continuum while typically improving rejection. However, narrower bandwidths:
     • Increase manufacturing difficulty and cost.
     • Heighten sensitivity to collimation errors, requiring larger tolerable focal ratios.
     • Necessitate larger filters for the same field of view, further complicating system design and manufacturing.
3. Passband Shape:
   • Steeper passband edges or rectangular shapes offer higher contrast and better rejection in regions far from the peak compared to simple Fabry–Perot filters.
   • Edge steepness may be quantified using metrics like 10% or 1% bandwidth, but achieving steeper edges increases cost and manufacturing complexity.

Key Parameters for Narrowband Filters

Filters cannot be manufactured to exact specifications, so tolerances must be stated for critical parameters:

1. Peak Wavelength:
   • Uniformity Across the Filter Surface: Grading across the surface should be limited to one-third of the half-width to minimize uniformity errors.
   • Mean Peak Wavelength: Measured over the entire filter area, the tolerance is typically positive to allow tuning via tilt. The tolerable tilt depends on system aperture and field requirements.
2. Bandwidth:
   • Specify the bandwidth with a tolerance, but keep it wide enough to accommodate manufacturing challenges (e.g., no less than 20% of the nominal bandwidth).
3. Rejection in Stopping Zones:
   • May be defined as either the average transmittance over a range or the absolute transmittance at specific wavelengths. For line sources, specify relevant wavelengths.

Alternative Specification Methods

Filters can also be specified using transmittance envelopes that define maximum and minimum permissible values across the wavelength range. While more precise, this approach can be overly stringent, as it requires absolute compliance, which is difficult to test given the finite bandwidth of measurement apparatus. Including a note that performance is defined as an average over a specified interval is advisable.

Final Considerations

The relative importance of these factors varies depending on the application, so each case must be evaluated individually. Collaboration between the system designer and the filter manufacturer is crucial to creating a practical and effective performance specification.

2. Manufacturing Specification

The manufacturing specification includes the filter design and detailed manufacturing instructions, primarily for the use of machine operators. It serves as a practical guide to ensure consistency and quality during the production process.

Filter Design

The specification begins with the filter design, which outlines the materials used. Most filters utilize up to three thin-film materials with low, medium, and high refractive indices. Designs are typically expressed in terms of quarter-wave optical thicknesses at a reference wavelength \( \lambda_0 \), using the symbols \( L \), \( M \), and \( H \). For example, common designs may be written as:

\[
L \; LHLHHLHLH \quad \text{or} \quad M \; MHLHHLHL
\]

The materials might correspond to:

– \( L = \text{Si} \)
– \( M = \text{CaF}_2 \)
– \( H = \text{ZnS} \)

The substrates are indicated by symbols, such as \( \vert \text{Ge} \vert \) or \( \vert \text{Si} \vert \).

Constructional Details

The manufacturing specification also includes constructional details, such as:

– Monitoring Methods:

• The wavelengths used.
• The form of monitoring signals.

– Other Details:

• Substrate temperature.
• Special types of evaporation sources or processes.

Specification as a Checklist

To ensure clarity and ease of use, it is helpful to format the manufacturing specification as a table or checklist for the operators. The checklist should allow operators to:

– Follow the required steps.
– Record observations or anomalies during the process.

Operators should be encouraged to critically observe the machine operation to identify faults early. Comments and observations can be recorded directly on the form.

Batch Management

Each filter production batch should be assigned a unique reference number. After production, the completed specification forms should be filed as part of the machine logbook. Additional details, such as pumping performance, may also be recorded for maintenance purposes.

Layer Ordering

For calculation purposes, there is no standard practice regarding whether the incident medium appears at the top or bottom of a design table. However, for manufacturing instructions, layers are typically listed from the innermost (next to the substrate) to the outermost layer, as this sequence aligns with the deposition process.

Software Integration

Software tools can streamline the creation of manufacturing specifications, particularly for generating the sequence of monitoring signals. In some cases, these specifications can be directly integrated with the deposition controller, ensuring automated and precise execution. A printed copy of the specification can serve as a reference and record for operators.

3. Test Specification

The test specification is arguably the most important document in the filter manufacturing process. It defines the complete set of tests to be conducted on the filters to assess their performance and serves as the ultimate guarantee of a filter’s quality. Even though filters are designed to meet specific performance specifications, only the performance defined in the test specification is guaranteed.

Key Components of the Test Specification

1. Purpose and Scope:
– The test specification should align closely with the performance requirements outlined in the original performance specification.
– In some cases, a test specification may simply state that the filter will pass a predefined set of tests.

2. Detailed Test Definition:
– It is crucial to define the test equipment and its make or type to ensure reproducibility, even at remote testing locations.
– The specification should outline the required tests and their acceptance levels.

Uniformity Testing

Absolute uniformity is impractical to measure due to the need for infinitesimally small beams across the filter’s entire surface. A practical approach is to measure performance at specific points:
– Example: Measure the peak wavelength at the center and four equally spaced points around the filter circumference.
– The spread in performance values across these measurements indicates uniformity.

Measurement Challenges

1. Rejection Testing:
– Rejection must often be measured over wide spectral regions, requiring fast scanning speeds and broad bandwidths.
– This approach may not accurately capture features of narrow-line rejection. For line sources, careful testing at specific wavelengths is necessary.

2. Fourier Transform Spectrometer:
– Suggested by Bousquet and Richier for measuring rejection.
– While difficult to apply in the visible spectrum, it is feasible for infrared filters using commercial Fourier transform spectrometers.

Testing Levels and Costs

– Batch Testing:
– For low-cost filters, batch testing may involve basic checks on only a few samples.
– Buyers should recognize that low-cost catalog filters may not undergo extensive testing and may not guarantee superlative performance.

– Extensive Testing:
– Individual filter testing, particularly for high-performance components, increases costs significantly.

Subjective Quality Testing

While optical performance can be measured directly, subjective properties related to the quality and finish of films and substrates are harder to quantify.

1. Substrates:
– Specifications for flatness, curvature, polish, and allowable blemishes are usually provided.
– Standards like MIL-E-13830A are often used for optical components.

2. Coating Defects:
– Pinholes:
– Small uncoated or partially coated areas reduce rejection in stop bands and can be unsightly.
– Acceptance criteria often specify a maximum number and size of pinholes per unit area.
– Spatter Marks:
– Caused by ejected material from sources; these are typically aesthetic issues unless excessive, causing nodular growth or scattering losses.
– Stains:
– Can result from substrate defects, poor cleaning, or drying residues. They are judged on appearance and may not significantly impact optical performance unless severe.

Uncoated Areas

Filters must be held in jigs during coating, leading to uncoated areas, usually forming a ring around the edge:
– These areas, often about 1.0 mm wide, include both a taper and an uncoated strip.
– Uncoated edges serve to protect the filter during mechanical mounting and help prevent delamination.

Rejection Criteria

Certain defects demand immediate rejection, such as:
– Blisters: Indicators of adhesion failures.
– Uncoated Areas within Filter Boundaries: Typically due to contamination or moisture penetration.
– Damage Near Edges: Initiates delamination and can compromise the filter’s integrity.