What to Check When Measuring an Infrared Filter’s Spectral Performance

An infrared filter is an optical component which selects or rejects wavelengths. It is only useful if its measured characteristic matches its intended use in an optical system. In addition to the peak values, a good test also measures the passband, the cutoff or edge, the blocking and any leakage that might reach the detector. The measurement is also recorded with the conditions under which the filter was measured.

At BoDian Optical, the practical starting point is to connect the spectral report to the source, detector, and unwanted radiation in the customer’s design. That makes the report useful during system integration, rather than a sheet of numbers filed after production.

Key takeaways

• Read the full transmission curve, not only the peak value.
• Check the transmission, reflectance, absorbance, blocking and the physical aperture of the filter together.
• Record wavelength range, resolution, angle, polarization, reference method, and sample orientation.
• Compare the measured result with the drawing or purchase specification before approving a batch.

What does spectral performance tell you about an infrared filter?

The spectral performance of a filter is expressed by its transmission curve, which indicates what is transmitted to the detector. The reflectance curve of a coated surface indicates what is reflected back from the surface of the filter. The absorbance of a filter indicates how much energy the filter is retaining. The various curves are related and have to be considered in a system with a strong source, a sensitive detector and a tight thermal budget.

This means that a detector which requires only a single narrow band can fail if it is sensitive to leakage through an adjacent window which is not monitored. The center wavelength defines the middle of the designed passband. The bandwidth is the width of the useful band. The cutoff or edge wavelength denotes the transition from transmission to rejection. The peak transmittance is the maximum value of the transmission curve, not to be confused with the average transmission in the band. The blocking region is the interval of wavelengths for which the unwanted radiation must be below a predetermined limit.

Which measurements belong on an infrared filter test report?

Transmittance across the working band

Scan the wavelength range of interest to a suitable length to see if there is any transmission outside this range that is unwanted. Measure the peak transmission, average transmission, passband width, slope and ripple. A narrow peak with steep shoulders may look good but transfer energy to a neighboring detector channel.

Reports for narrow-band products like the INBP10560 infrared narrow bandpass filter should clearly highlight the target band of interest and also show the adjacent regions of blocking. The product page describes INBP10560 with high peak transmission in the 10560nm region and sharp blocking on either side. The measurements must verify that the delivered item is equal to or better than the specifications.

Blocking, reflectance, and absorbance

The number that prevents background radiation from entering the measurement is usually the blocking. Verify the blocking band as stated for your filter, and verify the worst case leakage for that filter, as opposed to just reading the lowest number from a graph. Reflectance is important for returned light that can scatter to a source, to a cavity, or to nearby detectors. Absorbance is important for filters that can heat up under a strong source.

A report should clearly state the quantity that has been measured, the units in which it has been measured, the wavelength interval which was covered and the method of the instrument that was used to make the measurements. If a customer orders a transmission report, ask if he also needs reflectance or absorbance for high power or thermally sensitive applications.

Physical and aperture checks

Even the best spectral data cannot compensate for a damaged edge or a too small clear aperture. Check diameter or length and width, thickness, tolerance, surface quality, edge quality and orientation marks. Are there any scratches, cracks, chips or coating stains? Inspection has to be done according to agreement.

How should the measurement be set up?

Clearly define the test at the sample, not at the instrument entrance. List the wavelength range, spectral resolution, angle of incidence, polarization, reference baseline, etc. Even interference filters can shift if the beam is not incident normally. “Meas. at normal incidence” is not a meaningless phrase.

The entire range of work has to be covered by the instrument. A measurement which starts in the passband and immediately ends there does not reveal any hidden secondary window. Thin-film spectral measurements can be performed using spectrometers such as the Agilent Cary 5000, Cary 7000, and PerkinElmer Spectrum 3, depending on the measurement range and test requirement.

Pay attention to the fixture too. Keep the test sample clean. Center the useful part of the test aperture in the beam. Use the same face and orientation for subsequent test samples. A beam that only just clips the edge of the filter can look terrible in test even though the filter is perfectly acceptable. Note whether results come from one part, from several sampled parts, or from an average of the entire batch.

How do you read a transmission curve without misreading it?

Start with the application window

Add a plot to mark out the source wavelengths, the detector response, and the unwanted regions of sky. Then simply ask: Does the filter allow the system to get the light it needs whilst keeping all the other light below some minimum acceptable level? Don’t choose a filter because it sounds vaguely similar to what you are looking for from its catalog name.

Inspect the edges and blocking floor

A clean transition, some ripple in the passband and some leakage outside the edge of the passband are all things to look for when measuring a filter. The steepness of the edge of the passband can be more important than the difference in peak transmission between two channels near to each other. Measure both sides of a bandpass filter and measure the entire stopband for a longpass or shortpass filter.

Compare the curve with the specification

Use the same definitions as the purchase drawing. “Cutoff” may be defined at a percentage transmission, a half-power point, or another contractual value. If the test curve is close to a tolerance boundary, repeat the measurement and confirm the method before accepting or rejecting the part.

What commonly causes inconsistent infrared filter results?

Discrepancies in angle of incidence, spectral resolution, reference, scan range, polarization and surface contamination as well as incorrect sample orientation and beam size smaller than clear aperture aperture can all lead to misleading measurements and graphs.

Temperature and mounting stress need to be noted when you are running filters outside of a lab. A single room-temperature scan does not equal environmental test. State what was tested and if the application requires it, test further.

How can a buyer turn test data into a better filter specification?

Send the supplier a short, complete request:

• Target wavelength range and passband or cutoff
• Minimum transmission, blocking range, and test angle
• Substrate, clear aperture, dimensions, and report format target wavelength range; passband or cutoff; minimum transmission; blocking range and limit; incidence angle; polarization; substrate; dimensions and clear aperture; surface quality; operating temperature; quantity; and the format of the required spectral report.

The more precisely the system is described, the less likely the filter will be selected from one number. BoDian Optical supports filter work from customer materials, samples, or drawings. A technical review can then connect the requested curve to a suitable infrared narrow bandpass filter series or a custom design.

Frequently asked questions

What is the difference between transmittance and blocking?

The amount of light that is allowed to pass through at each wavelength is called transmittance. The amount of unwanted light of certain wavelengths that is blocked is called blocking. High peak transmission does not necessarily mean that a good filter, as the amount of leakage in the blocking region can be too high.

Why should a filter be measured beyond its passband?

The extra range shows leakage, secondary windows, and a less steep edge than expected. The details here can matter for how a detector responds outside of its nominal passband.

Does the test angle matter for an interference filter?

Yes, because the angle of incidence changes the optical path within the thin-film stack and thus the spectral response. Therefore, test angle should be the same as for the intended application or clearly stated to allow for correct interpretation of the measurement result.

Conclusion

A useful infrared filter test report answers three questions: what light passes, what light is rejected, and under what conditions the result was obtained. Read the curve with the source and detector in mind, then verify the physical part and the specification together. For a target wavelength, drawing, or sample, filter team can review the measurement requirements before production. For a basic reference on the infrared region, see NASA’s infrared spectrum overview.