It is a challenge to choose the correct tunable filter for a specific experiment, as there are a wide range of companies offering similar products, and a large number of advanced technologies available. The challenge is tougher when manufacturers define requirements differently.
Figure 1. Illustration of the out-of-band rejection of a volume holographic grating at λc = 632nm. Bands of ±45nm are presented and an out-of-band rejection of -60dB is obtained
This article focuses on defining the specifications of widely tunable filters and informing customers how these critical specifications need to be measured.
The ability of optical filter’s to isolate narrow light bands is easier to appreciate once the out-of-band rejection and optical density are understood clearly. However, consistent and accurate measurement of these parameters requires highly sensitive and ultra-low noise setups. Hence, before actually measuring specifications the terminology standards need to be defined. This will help in an optimal comparison of a range of technologies.
The term "band rejection", as used in signal processing, can be defined as a filter’s ability to differentiate the desired and the undesired signal on a particular frequency band.
A bandpass filter enables the transmission of a particular frequency range and does not allow the rest to pass through. Stopband filters perform the opposite function, preventing just a specific band of frequencies from passing without affecting the rest of the signal.
The out-of-band rejection of a bandpass filter is the attenuation applied to any of the signal outside of the chosen frequency band. This is sometimes incorrectly referred to as optical density, or isolation. Out-of-band rejection specifies a rejection value centered around λc, which is normally in dB. It is defined in relation to a filter’s peak efficiency.
This is shown in Figure 1. For example, a band of ±45 nm is shown. Outside this band, a rejection level (for unwanted wavelengths) of -60 dB is obtained. On a specification sheet, the corresponding property appears as: out-of-band rejection <-60 dB @ λc± 45 nm.
Two sets of bands are defined in the case of the Laser Line Tunable Filter (LLTF) Contrast™ from Photon etc. (see Fig. 2) The first is λc ± 40 nm for the visible (VIS) version of the instrument (400-1000 nm), and the second λc ± 80 nm for the short wavelength infrared (SWIR) version (1000-2300 nm).
The greatest challenge with respect to this specification is its measurement. These limitations normally arise from the dynamic range and the measurement setup’s noise limit.
Highly sensitive instruments with a high dynamic range, combined with a comparatively high power broadband source or tunable laser, are used to overcome this.
Figure 2. Photon etc’s Laser Line Tunable Filter (LLTF) Contrast
In spectroscopy, the optical density is the logarithmic ratio between the radiated power of the incident beam and the radiated power transmitted through the instrument. In terms of this transmittance definition, OD is given by:
Where T is the transmittance (T=It/Io), It is the transmitted light intensity and Io is the incident light intensity.
The optical density is defined relative to the unfiltered light source, this is in contrast to the out-of-band rejection which is defined relative to the peak efficiency.
Optical density describes an optical filter’s ability to obstruct unwanted wavelengths.
Measurement of Out-of-Band Rejection and OD
For determining OD and out-of-band rejection, the ideal instrument must have a high sensitivity and a high dynamic range. There is also a need for a sufficient spectral resolution to resolve the tunable filter's diffraction efficiency, and to prevent convolution and sampling. This is normally less than 25% of the FWHM (full width at half maximum).
This ideal setup can be assembled using commercially available instrumentation; however, a few adjustments need to be made to ensure measurement accuracy.
The specifications with regards to the measuring instruments are:
- Dynamic Range > 70 dB
- Sensitivity < -80 dBm
- Wavelength span > out-of-band limits
It is important the dynamic range is higher than the defined out-of-band rejection (< -60 dB) and optical density (> OD 6.5). In instances, a low dynamic range will result in either saturated measurement if the detector is too sensitive or in noise limited measurement if the filter and a high-power source are connected. Also, it is important that the sensitivity is as high as possible, to enable measurements of power densities as low as a few mW/nm with good signal-to-noise performance.
Lastly, the spectral range must at least encompass a window with the out-of-band wavelength limits (λc ± X nm) for viewing the complete spectral range of the grating. The rejection span is validated by performing measurements exterior to these out-of-band limits.
Based on these requirements, complementary metal oxide semiconductor cameras (CMOS) or charge coupled devices (CCD) combined with spectrometers are good prospects, however they would be pushed to their limits.
Another approach is to combine a monochromator and a low noise amplification chain photodiode. This setup enables power measurements in a 20 dBm to -85 dBm range with off-the-shelf commercial devices. As a result an optical spectrum analyzer is the best choice as it has the optimum sensitivity and wavelength span.
Figure 3 shows a practical representation of the measurement setup deployed by Photon for determining out-of-band rejection and the OD, where a supercontinuum (SC) laser is the external source. Within the LLTF, the equivalent of two light paths are created. One light path passes through the Volume Bragg Gatring (VBG) for the collection of filtered light and the other is transmitted via the mirrors shown by the M1-4 (dotted line), where the collection of all power from the SC takes place.
Figure 3. Realistic representation of the measurement setup used to measure the out-of-band rejection and the optical density. SC : Fianium’s supercontinuum source SC-400-4, FOC : Fiberoptic Output Coupler (using achromatic 10mm focal lens, multimode 50/62.5µm), OSA: Optical Spectrum Analyser ANDO AQ6315E, M1-4: mirrors, VBG: Volume Bragg Grating. Two light paths are created within the LLTF: one goes through the VBG to collect the filtered light, while another through mirrors, M1-4 (dotted line), and where all the power from the SC is collected.
In order to obtain the out-of-band rejection (Figure 3), first a supercontinuum source is injected through the filter (LLTF™ Contrast) and it is then recombined in a fiber (achromatic 10 mm focal lens, multimode 50/62.5 µm). Then the signal is sent to an OSA (ANDO AQ6315E) for detection. This setup is used to obtain the out-of- band rejection for various supercontinuum sources, using the same filter, as shown in Figure 4.
Figure 4. LLTF Contrast™ output power spectrum for different SC laser. The noise floor of the measurement is around -80dBm. Green: Fianium’s WL SC480-10 (10 W total output power, ˜ 8 mW/nm @700 nm), red: Fianium’s WL SC400-4 (4 W total output power, ˜2.1 mW/nm @700 nm), blue: Leukos SM30 (100 mW total output power, ˜ 0.03 mW/nm @700 nm).
As previously mentioned, the higher the power the source provides, it is less likely that the measurements will be restricted by the noise floor of the detector (around -80 dBm in this case).
Figure 4 explains the impact of this drawback. As seen in the blue curve in Figure 4, the determined out-of-band rejection using a low power SC attains -40 dB. This measurement does not impact the filter’s real performance but only shows the measuring instrument’s limitations. So it is very important to use a significantly powerful source.
When a medium power SC laser is used as seen in the red curve in Figure 4, it is possible to determine the specified -60 dB out-of-band rejection at ± 25 nm from the central wavelength.
Finally, when a high-power SC laser is used as depicted by the green curve in Figure 4, the specified -60 dB out-of-band rejection can be measured, but at very close proximity from the central wavelength (±21 nm).
Optical Density Measurement
Determining the OD (log10(I0/It)) is a greater challenge when compared to out-of-band rejection as for any specified wavelength, it is important to determine the supercontinuum source intensities (I0) and the filtered beam (It) with precisely the same conditions. As shown in Figure 5, the 1064 nm pump residual of the SC laser is an ideal reference mark to help in determining the real OD. It is the only spectral feature where there is a signal external to the out-of-band rejection span and above the measurement noise limit.
Figure 5. Measurements of the optical density (OD = log10 (I0/It)) at the pump energy (1064nm). Blue: LLTF Contrast™ output (It), Red: unfiltered supercontinuum output (I0). Those measurements were realised with the setup shown in Figure 3.
In Figure 5 the noise floor has a value around -76 dBm with an emerging pump residual at 1,064 nm. In order to obtain the OD, the experimental configuration explained in Figure 3 is used. A sequential repetition of the measurements is performed, once with the light path passing through the VBG and once using the mirrors’ path. An OD of 6.8 is obtained by comparing the two intensity peaks at 1064 nm (Figure 5).
This information has been sourced, reviewed and adapted from materials provided by Photon Etc.
For more information on this source, please visit Photon Etc.