Spectral characterization of passive optical devices
Optical spectrum analyzers (OSAs) use monochromators to transmit a specific range of light wavelengths. Conventional grating-based OSAs, however, have slow and moderate spectral resolution mechanisms that are incompatible with the requirements of modern sensing and bioengineering applications.1 In these applications, the amplitude, phase, and polarization responses of passive optical components in the systems must be measured with ultra-high spectral resolution and high speed. A novel technique is therefore needed to overcome these limitations.
Due to advances in electronic digital signal processing (DSP) technology, the receivers in optical fiber transmission systems can now be used to conduct optical channel estimation (OCE). These systems are used to measure the optical field response of cascaded or multiple components in an optical link. OCE algorithms are then used to compensate for distortion from the optical channel and to characterize the optical field response of a single component. The OCE methods must subsequently be adapted to meet the requirements of the specific instrumental application.
We have recently used DSP-based OCE and optical spectrum stitching (OSS) techniques to make an optical spectrum characterization measurement, with a resolution of ∼10MHz and a bandwidth of 250GHz.2 This type of measurement can potentially be made across the whole C-band (range of the electromagnetic spectrum used for long-distance radio telecommunications). Although OCE can be performed in either the time domain or the frequency domain, with the latter well-established and efficient algorithms (i.e., fast Fourier transform—FFT—and inverse FFT—IFFT) can be used. We therefore use DSP and digital-to-analog converters (DACs) in our system to produce a number of spectral lines with an IFFT algorithm (see Figure 1). After passing through the device under test (DUT), the transformed spectral lines are sampled by an analog-to-digital converter (ADC) and recovered via FFT analysis. We find the optical field response by collating the changes in characteristics of each probe before and after they pass through the DUT.
Our method has three notable features. First, by using IFFT analysis, we are able to efficiently generate a large number of narrowly spaced probes (1024 probes with sub-megahertz spacing) simultaneously.2, 3 Second, the calculation for each probe is very simple because they each occupy a very narrow frequency span, and we can assume that they have flat responses. Third, the measurement speed is close to the theoretical limit (bounded by the time-frequency uncertainty) because FFT is our main analysis tool and because we use real-time sampling. All these features can facilitate new sensor applications that require high resolution and fast speeds, such as biochemical binding kinetics and single-molecule reaction dynamics.
The spectral range of our technique is limited to about 10GHz (or 0.08nm) due to our use of ADCs and DACs. We have therefore replaced the single-wavelength laser with an optical frequency comb (OFC) that has a several gigahertz frequency spacing range. With this EFC-OFC combination we are able to significantly extend the spectral range of our measurements (see Figure 2) beyond the bandwidth of a single photodiode and ADC. We are also able to adjust our measurement bandwidth due to the tunability of the center wavelength and OFC line spacing. We use an array of photodiodes and ADCs to receive the extended optical spectrum and OSS to stitch all the separately measured segments together.
The measurement test bed in our system includes a fiber Bragg grating with fine structures. Measured amplitude responses obtained with a conventional OSA and with our technique are consistent to within 2nm (250GHz), as shown in Figure 2. However, with our method we are able to capture fine details that are not detected using the conventional OSA (see inset to Figure 2). Figures 3 and 4 also show high-resolution phase response, group delay, and polarization properties that were measured simultaneously (with the measurements shown in Figure 2) in our experiments. These parameters characterize the multiple dimensions of optical fields that contain independent and uncorrelated information. With these measurements we are therefore able to characterize the DUT in ways that are beyond the capability of the conventional OSAs. We have also been able to demonstrate that our method, using electro-absorption modulators and direct detection, is cost-effective.4
We have designed a new approach for making high-resolution and fast spectral measurements of passive optical components. The method incorporates DSP and OCE techniques, and will be particularly useful in future biochemical applications. We are currently building a new biochemical test bed so that we can adapt our method to meet practical requirements. We hope to then develop our technique at the instrumental level for the commercial market.
Zhaohui Li obtained his PhD from Nanyang Technological University in 2007. He joined the Institute of Photonics Technology as a professor in 2009.
Xingwen Yi obtained his PhD from the University of Melbourne in 2008 and has been in his current position since 2009.