Open Access
Add to BibSonomy
Abstract
We introduce and experimentally demonstrate a miniaturized integrated spectrometer operating over a broad bandwidth in the short-wavelength infrared (SWIR) spectrum that combines an add-drop ring resonator narrow band filter with a distributed Bragg reflector (DBR) based broadband filter realized in a silicon photonic platform. The contra-directional coupling DBR filter in this design consists of a pair of waveguide sidewall gratings that act as a broadband filter (i.e., 3.9 nm). The re-directed beam is then fed into the ring resonator which functions as a narrowband filter (i.e., 0.121 nm). In this scheme the free spectral range (FSR) limitation of the ring resonator is overcome by using the DBR as a filter to isolate a single ring resonance line. The overall design of the spectrometer is further simplified by simultaneously tuning both components through the thermo-optic effect. Moreover, several ring-grating spectrometer cells with different central wavelengths can be stacked in cascade in order to cover a broader spectrum bandwidth. This can be done by centering each unit cell on a different center wavelength such that the maximum range of one-unit cell corresponds to the minimum range of the next unit cell. This configuration enables high spectral resolution over a large spectral bandwidth and high extinction ratio (ER), making it suitable for a wide variety of applications.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Spectroscopy is of fundamental importance to scientific disciplines ranging from physics [1] to chemistry [2] and biology [3]. The technological applications are just as vast, including medical research [4], telecommunications [5], and remote sensing [6]. Consequently, a commensurate effort has been made to refine and improve spectrometers [7].
The development of silicon photonics in the telecom NIR bands is a promising avenue for the development of miniaturized spectrometers, overlapping with many chemical absorption bands in the near- and mid-infrared spectral ranges [8,9]. There are many anticipated benefits of photonic integration, including reduced size and weight, reduced noise, increased sensitivity [7,10], simplified environmental stabilization, and reduced power consumption [7]. Finally, the availability of silicon photonics foundries provides a clear path for low cost commercialization and successful designs.
Recently a variety of integrated spectrometers have been demonstrated [7,10–12]. The most notable designs operate in ways that mimic conventional spectrometer designs employing dispersive optical elements [6,13], as well as several variations based on Fourier transform spectroscopy [10,14]. Nonetheless, the integrated platform introduces unique constraints that pose a challenge to a direct miniaturization of conventional designs. In particular, the presence of dispersion and thermo-optic nonlinearity in silicon complicates the analysis of Fourier transform based devices, whereas there is much more complexity addressing the high density integrated detector arrays required by grating based devices.
In this manuscript we introduce and demonstrate an integrated chip-scale spectrometer that has the advantage of being more resistant to thermal drifts. The ring-grating spectrometer presented here is made more resilient through the distributed waveguide Bragg reflector design (DBR) and its large bandwidth which will always match that of the ring resonator even in the event of some drift. The ring-grating spectrometer here is formed by combining a narrow-band filter realized by an add-drop ring resonator with a broadband filter that consists of a distributed waveguide Bragg reflector (DBR) as shown in Fig. 1 [15–17]. Compared to conventional spectrometers, this design realizes a chip-scale spectrometer with high spectral resolution, extinction ratio (ER) and spectral throughput. Moreover, the fabrication process for this spectrometer is compatible with processes available at global CMOS foundries, offering a direct route to cost-effective fabrication. The ER and spectral resolution characteristics of this device are determined by the ring transmission (at drop port) and the Q-factor, respectively. The device shares the advantages of other individually addresses pixel detector designs, in the sense that the gain of the individual pixels can be adjusted for optimal dynamic range. This compares favorably to conventional dispersive spectrometer designs that employ CCD detector arrays. This is due to the gain limit of the detector array is the same across all the pixels. In the common scenario that adjacent pixels rows are exposed to differences in power density (each row corresponding to a different wavelength), this manifests as a dynamic range limitation. This is because the maximum gain of the lower power pixel rows is limited by the highest power pixel row. Moreover, when comparing the work shown in this manuscript to a planar echelle grating spectrometer it is clear that the latter one consists of static arrays and the number of detectors required to realize high resolutions will not only result in a larger number of detectors but while also demand a large footprint [7]. In contrast, the ring-grating spectrometer shown here uses the same detector and is capable to tune over reasonable bandwidth (i.e., 30 nm) with the resolution of the ring resonator. Furthermore, the fundamental device unit cell (illustrated in Fig. 1) can be stacked in series to cover additional spectral bands in support of broad band applications. This can be done by centering each unit cell on a different center wavelength such that the maximum range of one-unit cell corresponds to the minimum range of the next unit cell. This is quite useful in the event that the maximum spectral range provided by the tuning mechanism is smaller than necessary [16]. The thermo-optic effect in both components is essentially identical therefore both elements can be thermally tuned simultaneously hence simplifying the overall design and enabling the capability to shift transfer functions of both filters thus apodization of the Bragg grating is unnecessary [16]. Finally, in comparison to FTS the ring-grating spectrometer shown here can offer sub-nm resolution with a small footprint design while FTS demands huge footprint design (i.e., mm’s of waveguide).
Fig. 1. Conceptual schematic diagram of the ring-grating integrated spectrometer: (a) Breakdown of a single unit cell (e.g. unit cell 1), where the distributed Bragg reflectors (DBRs) are designed such that the cross coupling band matches that of the ring resonator, thus a segment of the input signal is redirected towards the add-drop ring resonator while the remaining spectra passes through (DBR through port). The ring resonator then acts as narrowband filter hence permitting the resonance wavelength to couple into the ring drop port while the remaining wavelengths propagate towards the DBR drop port (red arrows indicate the path traveled by the optical signal). (b) Several unit cells can be cascaded in order to cover a wider spectral range, where each unit is designed to cover a portion of that spectral. (c) Lumerical MODE - 2.5D varFDTD simulation of two cascaded unit cells with approximately 300 period gratings at different center wavelengths (λc1 = 1542.1 nm and λc2 = 1563.9 nm). For completeness, the self-stop bands outside the operating bandwidth of unit cells one and two are 1588.8 nm and 1610.8 nm, respectively. These stop bands can be pushed further through dispersion engineering that maximizes the difference between neff1 and neff2 according to Eq. (2) and Eq. (3).
2. Operating principle of the ring-grating spectrometer
The miniaturized spectrometer shown here (see Fig. 1) employs a DBR in combination with and an add-drop ring resonator in order to perform a spectral interrogation over the short-wavelength infrared (SWIR) spectrum. The DBR functions as broadband filter and is designed by utilizing the cross-coupling Bragg condition [17]:
where ${\beta _1}$ is the propagation constant of waveguide one (WG1), ${\beta _2}$ is the propagation constant of waveguide two (WG2), $\mathrm{\Lambda }$ is the sidewall grating period for both waveguides, λc is the center wavelength of the drop band resulting from contra-directional coupling occurring between the two corrugated waveguides. In this design, the center wavelength of the contra-directional coupling in each waveguide is determined by the effective indices (n1 and n2) and can be engineered by modifying the width of WG1, the width of WG2, and the grating periods (Λ), such that the cross coupling band of the static Bragg grating centered at λc matches that of the ring resonator. The four port DBR configuration also has two backward coupling Bragg conditions (self-stop bands) [17]: where λ1 is the center wavelength of the back-ward coupling (stop band) in WG1, and λ2 is the center wavelength of the backward coupling (stop band) in WG2. When the left and right waveguides have different widths, the propagation constants ${\beta _1}$ and ${\beta _2}$ will be different, and these bands will be centered far above and below the cross coupling band. Consequently, they will not interfere with the operation of the device. Moreover, it should be noted that for operation with cascaded multiple unit cells all of them will need to be designed to have these self-stop bands outside the operating bandwidth of the spectrometer. The resonant wavelength (λres) and the free spectral range (FSR) of the add-drop ring resonator filter is determined by the effective indices and the ring size Eq. (4) and Eq. (5) [15]. where λres is the resonance wavelength of the add-drop ring resonator, neff is the effective index of the add-drop ring resonator, ng is the group index of the add-drop ring resonator, m is the mode order of the resonance wavelength (i.e., m=1, 2 …), and L is the length of single round trip in the ring resonator (defined as 2πr, where r is the radius of the ring).The overall performance of the DBRs, the add-drop ring resonator, and the DBRs combined with the ring resonator are designed using COMSOL to numerically determine the optical mode parameters of the structure. Through contra-directional coupling, the ideal two waveguide DBRs act as a broadband bandpass filter (λc = 1520 nm) according to Eq. (1)–(3), while the ideal add-drop ring resonator provides high resolution spectral lines that are spaced according to Eq. (5). When combining both elements the device behaves as a notch filter thereby isolating a single ring resonance, hence performing spectral interrogation. Figure 2 is an analytical model of how a device with the following parameters should ideally perform for λc = 1550 nm, Λ = 300 nm, nDBR = 2.583, λres = 1550 nm, Q-factor = 9.693 × 103, nRing = 2.5. The sidelobes in this case do not impact the operation of a cascaded device since they do not appear outside the maximum thermal tuning bandwidth. However, in a hypothetical design where they might be problematic, we note that the sidelobe suppression ratio (SLSR) can furtherly be optimized to become less than -25 dB by implementing an apodization design similar to that shown in Ref. [18].
Fig. 2. Analytical model of the spectral response of an ideal (a) Distributed Bragg reflector (DBR) in which the center wavelength of the contra-directional coupling is at 1550 nm and the effective index of the static Bragg grating is 2.583 (nDBR). (b) Add-drop ring resonator (radius = 5 µm) in which the resonance wavelength is at 1550 nm and the effective index of the ring is 2.5 (nRing). (c) Distributed Bragg reflector (DBR) and ring resonator (combined).
3. Device layout, fabrication and characterization
3.1 Silicon-on-Insulator photonic chip
The miniaturized ring-grating spectrometer investigated here was designed and fabricated in a similar manner to that described in Ref. [19], where a standard silicon on insulator (SOI) chip with a 250 nm device layer is employed. The designed structures are patterned into the device layer using a hydrogen silsesquioxane (HSQ) resist written by electron beam lithography (EBL) and etched using reactive ion etching (RIE). A 2-3 µm thick silicon dioxide (SiO2) top cladding layer is then deposited by plasma enhanced chemical vapor deposition (PECVD). As illustrated in Fig. 1, the miniaturized spectrometer consists of an add-drop ring resonator (radius = 5 µm) and static Bragg gratings. The width and the height of WG1 is chosen to be 500 and 250 nm, respectively. The width and the height of WG2 and the ring resonator-based add-drop filter waveguides are chosen to be 400 and 250 nm, respectively. By periodically modulating a portion of the WG1 sidewalls by amplitude of 50 nm and the WG2 sidewalls by amplitude of 40 nm, respectively the two DBR mirrors are created (see Fig. 3). The two DBR mirrors are separated by a 250 nm gap and are formed out of 914 periods with a grating period, Λ=314 nm. The effective indices (n1 and n2) for both DBR waveguides (WG1 and WG2) were numerically found to be 2.535352 and 2.311079, respectively. These values are in good agreement with the theory Eq. (1) in which the grating period was calculated and was found to be 313.6 nm. All input and output ports are tapered down to 170 nm in order to enhance the mode conversion during the edge coupling of the input signal. Finally, since the thermo-optic effect in both structures are essentially identical there is no need to deposit large heaters (e.g., titanium-tungsten alloy) in order to tune the wavelength instead the entire chip is heated.
Fig. 3. The uncladded scanning electron microscope (SEM) images of: (a) Distributed Bragg reflector filter (image scale 500 nm) (b) Add-drop ring resonator (image scale 10 µm).
3.2 Experimental setup
The characterization of the miniaturized spectrometer was performed using an upgraded version of the setup to that described in [19] (see Fig. 4). A multifunction I/O device (USB-6212 by National Instruments) was integrated into the setup thus enabling the capability towards faster automation of the measurements by means of sweeping the laser wavelengths through a MATLAB program, which controls the trigger of both the source and the photodetector simultaneously. In addition, a temperature control device (TED200C by Thorlabs) is introduced to monitor and control the overall temperature on the chip through the thermoelectric cooler (TEC) and a thermistor located in close proximity to the sample.
4. Measurements and discussions
Due to fabrication variations (e.g., sidewall roughness, geometric deviation, etc.) shifts in the center wavelengths (λc, λ1, and λ2) of the distributed Bragg reflectors and the resonance wavelength (λres) of the add-drop ring resonator are anticipated. Figure 5 shows a noticeable shift of the center wavelengths of the stop band of the contra-directional coupling (DBR drop port) centered at 1516.2 nm (λc) while the resonance wavelength (λres) of the add-drop ring (ring drop port) is found at 1515.1 nm. The overall experimental spectral resolution and ER were found to be 0.121 nm and 19 dB, respectively (illustrated in Fig. 6). Although the minimum insertion loss was found to be 27.7 dB in the current unpackaged device, it can be substantially decreased by implementing optical packaging with low loss couplers (i.e., to around 0.8-1.4 dB insertion loss [20]). Moreover, the time of acquisition of such a device can be made on the order of 1 ms as in Ref. [21]. The design has a rich optimization space. The overall resolution can be substantially improved by employing a higher Q-factor ring resonator into the design, although this must be balanced against a decrease in the SNR since the transmitted power will decrease. The ER aspect can be enhanced by lengthening the DBR elements or increasing the ring resonator Q-factor. Furthermore, depositing individual heaters on each element for future designs can rectify any local fabrication error for each of the individual elements hence enhancing the overall ER in the sense that the ring (i.e., λres) can be tuned towards the center of the DBR stop band thus enabling higher transmission (see Fig. 5).
Fig. 6. (a) Temperature dependence of the ring grating spectrometer response. (b) Spectrometer response at 54.98 Co showing a FWHM of 0.121 nm. Here the spectral response of the ring was fitted with a Lorentzian function hence xo and Γ were used as the ring resonance wavelength (i.e., 1515.1 nm) and its full width half maximum (i.e., 0.121 nm), respectively. (c) Spectrometer response around 54.98 Co showing an extinction ratio which is approximately 19 dB given optimal coupling.
The discrepancy between the analytical model and the fabricated DBR (e.g., drop port and through port intensity) arises due to one of the following reasons: (i) the realignment of the fiber during each measurement; (ii) the misalignment of the fiber due to the thermal expansion [10]; (iii) different propagation length for each port (e.g., Through port ≈ 0.525 mm and Drop port ≈ 0.671 mm) indicating different propagation losses in which the waveguide loss and DBR loss are known to be 4.5 dB/cm [22] and 6 dB/cm [23], respectively; (iv) possibility of small random fabrication defects on the waveguides in addition to rough facets from dicing the sample which cause losses.
We use the thermo-optic effect in silicon to thermally tune the central wavelength of the stop bands (λc, λ1, and λ2) and the resonance wavelength (λres) of the ring resonator. Since the mode is highly confined, the thermo-optic effect in both structures are essentially identical, where${\; }\frac{{d{\lambda _{DBR}}}}{{dT}} = 0.113{\; }\frac{{nm}}{K}{\; }and{\; }\frac{{d{\lambda _{ring}}}}{{dT}} = 0.115{\; }\frac{{nm}}{K}$, hence both elements can be simultaneously tuned. Moreover, CMOS compatible devices can tolerate high temperatures (i.e., 525 Co) [24] hence tuning the free spectral range (FSR) of ring grating spectrometer (i.e., 30 nm) is quite achievable.
The simulation results clarify that fabrication tolerance of ±20nm accounts for approximately -29.3 nm and 33.4 nm shifts in the stopband center, respectively. Similarly, the ring resonance shifts by -31.8 nm and +40.2 nm, respectively. To compensate for this an additional unit cell can be added above and below the nominal design to cover the desired spectrum, while the ring resonance could be compensated by a local heating element. Smaller errors could be compensated by local thermal tuning alone.
5. Conclusion
In this work we demonstrated a miniaturized integrated ring-grating spectrometer that operates in the short-wavelength infrared (SWIR) spectrum which enables high spectral resolution (i.e., 0.121 nm) and high extinction ratio (i.e., 19 dB). Although the minimum insertion loss was found to be 27.7 dB in the current unpackaged device, it can be substantially decreased by implementing optical packaging with low loss couplers (i.e., to around 0.8-1.4 dB insertion loss). The device layer consists of an ordinary add-drop ring resonator (radius = 5 µm) in addition to a distributed waveguide Bragg reflector. The Bragg reflector employs a pair of waveguide sidewall gratings that form a broadband filter (i.e., 3.9 nm) which redirects a specific spectrum into an add-drop ring resonator, which behaves as a narrowband filter (i.e., 0.121 nm). The thermo-optic coefficients in both elements are essentially close to each other therefore both elements can be simultaneously tuned to analyze the spectra. The device is far more robust to local temperature variations than CROW devices, which are the nearest alternative. Furthermore, several unit cells with different central wavelengths could be stacked in a cascade in order to cover a broader spectrum bandwidth. The spectral resolution and ER of the device can be further improved by employing a higher Q-factor ring resonator in addition to lengthening the DBR element.
Funding
Defense Advanced Research Projects Agency (DSO NLM and NAC Programs); Office of Naval Research; National Science Foundation (CBET-1704085, DMR-1707641, NSF ECCS-180789, NSF ECCS-190184, NSF ECCS-2023730); Army Research Office; San Diego Nanotechnology Infrastructure (SDNI) supported by the NSF National Nanotechnology Coordinated Infrastructure (ECCS-2025752); Quantum Materials for Energy Efficient Neuromorphic Computing-an Energy Frontier Research Center funded by the U.S. Department of Energy (DOE) Office of Science, Basic Energy Sciences (DE-SC0019273); LEED: A Lightwave Energy-Efficient Datacenter funded by the Advanced Research Projects Agency - Energy; Cymer Corporation; Keysight Technologies.
Acknowledgments
The authors would like to thank Greg Vanwiggeren and Ryan Scott at Keysight for their fruitful discussions, and the Nano3 Staff at UCSD for their assistance during sample fabrication. Naif Alshamrani would like to thank King Abdulaziz City for Science and Technology (KACST) for their support during his study.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Y. Masumoto and T. Takagahara, Semiconductor Quantum Dots, 1st ed (Springer, 2002). [CrossRef]
2. K. Kochan, E. Lai, Z. Richardson, C. Nethercott, A. Y. Peleg, P. Heraud, and B. R. Wood, “Vibrational Spectroscopy as a Sensitive Probe for the Chemistry of Intra-Phase Bacterial Growth,” Sensors 20(12), 3452 (2020). [CrossRef]
3. S. Guo, P. Rösch, J. Popp, and T. Bocklitz, “Modified PCA and PLS: Towards a better classification in Raman spectroscopy–based biological applications,” J. Chemom. 34(4), e3202 (2020). [CrossRef]
4. M. Ferrari and V. Quaresima, “A brief review on the history of human functional near-infrared spectroscopy (fNIRS) development and fields of application,” NeuroImage 63(2), 921–935 (2012). [CrossRef]
5. W. F. X. Frank and W. Goertz, “Measurement of transmission properties of optical fibers for telecommunications by FT-IR spectroscopy,” Microchim. Acta 94(1-6), 309–314 (1988). [CrossRef]
6. A. Z. Subramanian, E. Ryckeboer, A. Dhakal, F. Peyskens, A. Malik, B. Kuyken, H. Zhao, S. Pathak, A. Ruocco, A. De Groote, P. Wuytens, D. Martens, F. Leo, W. Xie, U. D. Dave, M. Muneeb, P. Van Dorpe, J. Van Campenhout, W. Bogaerts, P. Bienstman, N. Le Thomas, D. Van Thourhout, Z. Hens, G. Roelkens, and R. Baets, “Silicon and silicon nitride photonic circuits for spectroscopic sensing on-a-chip [Invited],” Photonics Res. 3(5), B47–B59 (2015). [CrossRef]
7. Z. Yang, T. Albrow-Owen, W. Cai, and T. Hasan, “Miniaturization of optical spectrometers,” Science 371(6528), eabe0722 (2021). [CrossRef]
8. K. B. Beć and C. W. Huck, “Breakthrough Potential in Near-Infrared Spectroscopy: Spectra Simulation. A Review of Recent Developments,” Front. Chem. 7, 48 (2019). [CrossRef]
9. M. Vainio and L. Halonen, “Mid-infrared optical parametric oscillators and frequency combs for molecular spectroscopy,” Phys. Chem. Chem. Phys. 18(6), 4266–4294 (2016). [CrossRef]
10. M. C. M. M. Souza, A. Grieco, N. C. Frateschi, and Y. Fainman, “Fourier transform spectrometer on silicon with thermo-optic non-linearity and dispersion correction,” Nat. Commun. 9(1), 665 (2018). [CrossRef]
11. X. Ma, M. Li, and J. He, “CMOS-Compatible Integrated Spectrometer Based on Echelle Diffraction Grating and MSM Photodetector Array,” IEEE Photonics J. 5(2), 6600807 (2013). [CrossRef]
12. M. Nedeljkovic, A. V. Velasco, A. Z. Khokhar, A. Delage, P. Cheben, and G. Z. Mashanovich, “Mid-Infrared Silicon-on-Insulator Fourier-Transform Spectrometer Chip,” IEEE Photonics Technol. Lett. 28(4), 528–531 (2016). [CrossRef]
13. M. Florjańczyk, P. Cheben, S. Janz, B. Lamontagne, J. Lapointe, A. Scott, B. Solheim, and D. Xu, “Development of a slab waveguide spatial heterodyne spectrometer for remote sensing,” Proc. SPIE 7594, 75940R (2010). [CrossRef]
14. M. Erfan, Y. M. Sabry, M. Sakr, B. Mortada, M. Medhat, and D. Khalil, “On-Chip Micro–Electro–Mechanical System Fourier Transform Infrared (MEMS FT-IR) Spectrometer-Based Gas Sensing,” Appl. Spectrosc. 70(5), 897–904 (2016). [CrossRef]
15. W. Bogaerts, P. De Heyn, T. Van Vaerenbergh, K. De Vos, S. K. Selvaraja, T. Claes, P. Dumon, P. Bienstman, D. Van Thourhout, and R. Baets, “Silicon microring resonators,” Laser Photonics Rev. 6(1), 47–73 (2012). [CrossRef]
16. B. Hong, Y. Fainman, and A. Grieco, 2020. Miniaturized And Integrated Photonic Spectrometers. US Patent App. 16/303,653.
17. K. Ikeda, M. Nezhad, and Y. Fainman, “Wavelength selective coupler with vertical gratings on silicon chip,” Appl. Phys. Lett. 92(20), 201111 (2008). [CrossRef]
18. J. A. Davis, A. Li, N. Alshamrani, and Y. Fainman, “Silicon photonic chip for 16-channel wavelength division (de-)multiplexing in the O-band,” Opt. Express 28(16), 23620–23627 (2020) [CrossRef]
19. A. Grieco, B. Slutsky, D. T. H. Tan, S. Zamek, M. P. Nezhad, and Y. Fainman, “Optical Bistability in a Silicon Waveguide Distributed Bragg Reflector Fabry–Pérot Resonator,” J. Lightwave Technol. 30(14), 2352–2355 (2012). [CrossRef]
20. K. Suzuki, R. Konoike, J. Hasegawa, S. Suda, H. Matsuura, K. Ikeda, S. Namiki, and H. Kawashima, “Low-Insertion-Loss and Power-Efficient 32 × 32 Silicon Photonics Switch With Extremely High-Δ Silica PLC Connector,” J. Lightwave Technol. 37(1), 116–122 (2019). [CrossRef]
21. P. Sun and R. M. Reano, “Low-power optical bistability in a free-standing silicon ring resonator,” Opt. Lett. 35(8), 1124–1126 (2010). [CrossRef]
22. A. Grieco, B. Slutsky, and Y. Fainman, “Characterization of waveguide loss using distributed Bragg reflectors,” Appl. Phys. B 114(4), 467–474 (2014). [CrossRef]
23. A. Grieco and Y. Fainman, “Characterization of Distributed Bragg Reflectors,” IEEE J. Quantum Electron. 50(6), 453–457 (2014). [CrossRef]
24. S. Sedky, A. Witvrouw, H. Bender, and K. Baert, “Experimental determination of the maximum post-process annealing temperature for standard CMOS wafers,” IEEE Trans. Electron Devices 48(2), 377–385 (2001). [CrossRef]
