Open Access
Add to BibSonomy
Abstract
We demonstrate a novel, to the best of our knowledge, temperature-insensitive optical sensor based on two cascaded identical microring resonators (CIMRR) in this Letter. The structural parameters of the reference ring and sensing ring are designed to be identical. The upper cladding in the sensing windows of the two rings is removed. With different microfluidic channels, the reference ring and sensing ring are exposed to the reference solution and reagent sample, respectively. For wavelength interrogation experiments in the transmission spectrum contrast ratio and low-cost intensity interrogation experiments, the sensitivities of refractive index (RI) sensing are 3402.4 dB/RIU and 1087.3 dB/RIU, respectively, while the temperature sensitivities are as low as 0.023 dB/K and 0.0124 dB/K, respectively.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
Microring resonator (MRR) structures have been widely used in optical filters, microring lasers, optical switches, and microring sensors [1–5] due to their simple structure, diversified functional design, and high integration [6], and an optical waveguide sensor based on an MRR can be combined with microfluidic technology [7]. This has broad application prospects in environmental monitoring, food processing, medical diagnosis, and other fields [8–11].
Sensors based on an MRR often employ the wavelength interrogation method for sensing. To further improve the sensitivity of sensors based on a single MRR sensor, a reference ring is introduced to form a cascaded microring resonators (CMRR) structure [12–14]. The free spectral ranges (FSRs) of the two rings are designed to be different to produce the Vernier effect, which greatly improves the sensing sensitivity. Compared with wavelength interrogation, intensity interrogation uses a low-cost broadband light source (BLS) combined with an optical power meter, and does not require wavelength information for the transmission spectrum [14–16]. A refractive index (RI) change of the sample is measured by detecting the output optical power change. When the FSRs of the two rings are identical, the transmission spectrum envelope of the CMRR sensor disappears for wavelength interrogation. However, the sensitivity is highest for intensity interrogation [14]. Recently, a highly sensitive optical biosensor based on a CMRR with equal FSRs was reported [16]. The radii of the two rings were designed to be 128 µm and 125.5 µm to achieve similar FSRs ranging from 1550 to 1560 nm. The sensitivity of this sensor reached 1579 dB/RIU in intensity interrogation, and there was an RI-dependent detection limit (DL) of 9.7 × 10−6 RIU. Because of the large thermo-optical coefficient of silicon (1.86 × 10−4 K−1) [17], the effective RI (neff) of the microring waveguide changes with environmental temperature fluctuations, which reduces the accuracy of the detection results. Also, the influence of the temperature on the CMRR sensor is magnified by the Vernier effect. To overcome this problem, one method is to utilize a thermoelectric cooler (TEC) for the CMRR sensing system, but the use of a TEC will increase the cost and device size [16]. Another method is to use the reference ring to monitor the temperature change, which needs extra data processing [18]. Therefore, Zhang et al. demonstrated an optical waveguide sensor based on a ring cascaded with a Mach–Zehnder interferometer (MZI) [19]. A temperature sensitivity of less than 4 pm/K was achieved by appropriately designing the structural parameters of the MZI, but the design and the fabrication process were complex because of the small tolerance. In order to achieve a temperature-insensitive and highly sensitive sensor, a cascaded identical microring resonators (CIMRR) sensor is proposed in this Letter.
As shown in Fig. 1, the reference ring and sensing ring are cascaded by a tapered bus waveguide to form a temperature-insensitive CMRR sensor. The upper cladding layers in the sensing window areas of the two rings are removed. By using two microfluidic channels, the reference ring and sensing ring are exposed to the reference solution and reagent sample, respectively.
The waveguide widths of the two rings are W1 and W2, respectively, corresponding to different waveguide sensitivities. Due to the effect of the thermo-optic coefficient of silicon, the temperature sensitivity of the MRR is
Here, Sw is the temperature-dependent waveguide sensitivity. ng is the group RI of the fundamental mode. The FSR of the MRR is expressed as
Here, L is the perimeter of the microring. In the general case, the two rings have slightly different FSRs to produce the Vernier effect. The RI sensitivity of the CMRR sensor is F = FSRr/(FSRr − FSRs) times that of the MRR sensor. The subscripts r and s represent the reference ring and sensing ring, respectively. However, due to the Vernier effect, the temperature sensitivity of the CMRR sensor is also amplified accordingly. The envelope peak position of the total transmission spectrum shifts with the temperature change. The temperature sensitivity of a conventional CMRR sensor is [19]
According to Eq. (1) and Eq. (2), Eq. (3) can be further simplified as
If ∂λenv/∂T in Eq. (4) is equal to zero, we get the temperature-insensitive condition of the CMRR sensor as Swr·Lr = Sws·Ls. When Lr·ngr = Ls·ngs, the FSRs of the two rings are identical and the transmission spectrum envelope of the CMRR sensor doesn’t exist. With the RI change of the reagent sample in the sensing ring and the RI invariance of the reference solution in the reference ring, the spectral change is extracted from the contrast ratio (η) of the spectrum curve. The CIMRR sensor satisfies the temperature-insensitive condition and its spectral envelope doesn’t exist. Here, the waveguide widths W1 and W2 of the two rings are equal, and the structural parameters of the two rings are identical. Figure 2(a) shows the transmission spectrum change of the CIMRR sensor during RI sensing. The two rings have the same FSR of 0.83 nm in the simulation. When the neff of the sensing ring changes from 1.80505 to 1.80605 and the neff of the reference ring is set to 1.80505, the peak power P0 decreases and the valley power P1 increases. η = (P0 − P1)/(P0 + P1) can be derived by using the relative intensities of the two adjacent extrema P0 and P1 in the transmission spectrum. Then the linear part of the contrast ratio curve is extracted to calculate the RI sensitivity for the wavelength interrogation of the CIMRR sensor [see Fig. 2(b)]. The derived sensitivity (∂η/∂neff) reaches 6739 dB/RIU. Because the RI-dependent waveguide sensitivity of the waveguide structure used in this Letter is 0.44 according to Lumerical simulation, the RI sensitivity is 0.44 × 6739 dB/RIU = 2965.16 dB/RIU.
Fig. 2. (a) Transmission spectra of the CIMRR sensor with different neff values during RI sensing. (b) Simulated contrast ratio versus the neff change during RI sensing. (c) Transmission spectra of two identical MRR sensors and their cascaded CIMRR sensor at different temperatures during temperature sensing.
According to Eq. (2), Eq. (1) can also be expressed as
Two identical rings have equal Sw·L/λ values. When the resonant wavelength shifts one FSR with temperature, the temperature-dependent detection ranges (DR) of two single rings are equal to DR=λ/(Sw·L).
Figure 2(c) shows the simulation results for the transmission spectra (Tdrop) of the CIMRR sensor at different temperatures. When the neff of the sensing ring is 1.80505, η is exactly at the maximum position, and the transmission spectra (Tr, Ts) of the two single rings are completely coincident [see Fig. 2(c)]. The wavelength shifts of two single rings are identical. Here, Tdrop = Tr × Ts. Because of the equal temperature sensitivities of two single rings, the temperature change does not influence the contrast ratio of the transmission curve. The wavelength difference between the adjacent component peaks of the transmission curve is Δλpeak = (FSRr + FSRs)/2 [13]. With a temperature change of one DR, the transmission curve of the CIMRR sensor shifts by Δλpeak. Therefore, the change in the transmission spectrum of the CIMRR sensor with different temperatures is reflected in the shift of the component peak. Here, the temperature sensitivity of the component peak is
To measure the transmission spectrum of the CIMRR sensor, the light source uses a tunable laser source (Agilent 81642B) (see Fig. 3). To achieve the maximum coupling of the TM mode, a polarization controller is used to adjust the polarization of the input light. The light is coupled into and out of the waveguides by input/output grating couplers. Finally, the output signal is connected to a photoelectric detector (Agilent 81635A). The CIMRR sensing system is used by packaging the chip with the fiber array. An optical stripe waveguide with a width of 550 nm and a depth of 220 nm is fabricated on a silicon-on-insulator (SOI) platform with 2 µm of buried oxide and a 220 nm top silicon layer. Compared to the TE mode, the TM mode has a higher RI sensitivity due to the larger mode-field overlap with the reagent solution of the upper cladding [16]. In the CIMRR sensor, the TM mode is used for sensing.
The whole fabrication process is compatible with the complementary metal-oxide semiconductor (CMOS) process in the Integrated Circuit Advanced Process Center of the Institute of Microelectronics of the Chinese Academy of Science (IMECAS) in Beijing. The patterns of the optical waveguides, the grating couplers, and the microrings are formed by a stepper and inductively coupled plasma (ICP) etching. Two sensing windows are opened by dry and wet etching, and 2 µm of SiO2 upper cladding are deposited by plasma-enhanced chemical vapor deposition (PECVD). The radii of the two rings are 123 µm. In Fig. 4, the measurement setup is seen to consist of three parts: the CIMRR chip with the fiber array, the microfluidic system, and the temperature-controlled system. A thermistor in the platform is used to monitor the temperature. A TEC is embedded between the chip and a copper sheet to control the temperature. The measured reagent and the reference solution are injected into two microfluidic channels by the syringe pump, respectively.
Fig. 4. (a) Measurement setup. (b) Microscope image of the sensing ring. (c) Microscope image of the CIMRR sensor.
In all of the following experiments, the reference ring is exposed to the de-ionized water while the sensing ring is exposed to NaCl solutions with different concentrations, corresponding to an RI change of about 1.8 × 10−3 RIU/% [12]. Figure 5(a) shows measured transmission spectra of the CIMRR sensor, where the FSRs of the reference ring with a Q factor of 6736 and the sensing ring with a Q factor of 6692 are 1015.66 pm and 1017.39 pm, respectively. The contrast ratio of the transmission spectrum changes dramatically with the concentration of the NaCl solution. For wavelength interrogation, the RI sensitivity of the CIMRR sensor is found to be S = 3402.4 dB/RIU by extracting the linear part of the curve in Fig. 5(b). This sensitivity is much higher than the sensitivity of 2500 dB/RIU found by using the ratio between the intensities of the two central peaks for the conventional CMRR sensor reported in Ref. [14]. The power measurement accuracy of the RI sensing of the CIMRR sensor is σ ≈ 0.01 dB. Further, we get DL = σ/S = 2.9 × 10−6 RIU.
Fig. 5. (a) Measured transmission spectra when the concentration of the NaCl solution is changed from 1.2% to 2.3%. (b) The measured contrast ratio η versus the RI change of the NaCl solution.
Then, the sensing ring is exposed to an NaCl solution with a concentration of 1.6%. The temperature of the platform is tuned to 26°C, 30°C, and 34°C, respectively. Figure 6 shows measured transmission spectra from the through and drop ports of the CIMRR sensor. The resonant wavelength shifts upon shifting between the three different temperatures, while the transmission curve contrast ratio of the drop port is almost unchanged, varying by only 0.0234 dB/K. For the wavelength shift caused by the temperature change, the sensitivities measured by the through and drop ports are ∂λr/∂T = 0.0288 nm/K and ∂λpeak/∂T = 0.0267 nm/K. For the CIMRR sensor, ∂λpeak/∂T ≈ ∂λr/∂T is consistent with Eq. (6), but the contrast ratio used as the RI sensing signal is temperature insensitive.
Fig. 6. Measured transmission spectra when the sensing ring is exposed to NaCl(aq) with a concentration of 1.6% and the temperature of the platform is set to 26°C, 30°C, and 34°C.
To further verify the temperature-insensitive condition of the CIMRR sensor for wavelength interrogation, the sensing ring is exposed to NaCl solutions with concentrations of 0% and 0.8%, respectively. Figure 7 shows the transmission curve shifts at the drop port of the CIMRR sensor corresponding to three different temperatures. Due to the fabrication tolerance, two fabricated rings with an identical design will be slightly different. Thus, the contrast ratio will not be largest for both the reference ring and the sensing ring with de-ionized water used as the upper cladding layer. In the transmission curve contrast ratio experiments, the temperature sensitivities in Fig. 7 are 0.079 dB/K and 0.062 dB/K, respectively.
Fig. 7. The measured transmission spectra when the sensing ring is exposed to NaCl solution with concentrations of (a) 0% and (b) 0.8%, respectively, at different temperatures.
For intensity interrogation in the RI sensing process, the incident BLS is filtered by an optical tunable filter (Santec, OTF-930) with a central wavelength of 1562 nm and a 3 dB bandwidth of 3 nm. The output power of the drop port is received by a power meter (Keysight N7745A). In order to eliminate the influence of the input light source fluctuation, the output power is normalized by the output power of the through port. In Fig. 8(a), the RI sensitivity of the CIMRR sensor in the linear part reaches 1087.3 dB/RIU, which is twice the value of 450 dB/RIU reported for the TE mode of the CMRR sensor in Ref. [12]. The Q factors of the sensing ring in our Letter and Ref. [12] are around 6700. For intensity interrogation, the RI-dependent DL of the CIMRR sensor is around 9.19 × 10−6 RIU.
Fig. 8. (a) Measured power ratio between the drop port and the through port versus the RI change of the aqueous solution of NaCl. (b) Measured power ratio when the temperature of the platform is set at values ranging from 22°C to 44°C.
For intensity interrogation in the temperature sensing process, the sensing ring of the CIMRR sensor is exposed to de-ionized water. The temperature of the platform is set at values ranging from 22°C to 44°C. The power ratio changes with the temperature [see Fig. 8(b)]. It can be seen that there is a small change in the power ratio with an environmental temperature change because two exactly identical rings cannot be fabricated using the CMOS process. For intensity interrogation, the experimental results show that the temperature sensitivity is 0.0124 dB/K.
In summary, a temperature-insensitive optical sensor based on CIMRR has been demonstrated theoretically and experimentally in this Letter. Two identical rings were exposed to the reference solution and the reagent sample via sensing windows. For wavelength interrogation, a high RI sensitivity of 3402.4 dB/RIU with a DL of 2.9 × 10−6 RIU is achieved from the contrast ratio of the transmission spectrum, while its temperature sensitivity is only 0.023–0.079 dB/K. For low-cost intensity interrogation, the RI sensitivity is 1087.3 dB/RIU and the temperature sensitivity is 0.0124 dB/K. After surface functionalization of the waveguide in the sensing ring, the CIMRR sensor holds great potential for application in biological research, environment monitoring, healthcare, food quality control, and so on.
Funding
Natural Science Foundation of Jilin Province (20220101130JC); National Key Research and Development Program of China (2020YFC2004600); National Natural Science Foundation of China (62027825); 111 Project of China (D21009).
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. T. Makino, T. Gotoh, R. Hasegawa, T. Arakawa, and Y. Kokubun, J. Lightwave Technol. 29, 2387 (2011). [CrossRef]
2. W. Liu, M. Li, R. S. Guzzon, E. J. Norberg, and J. Yao, Nat. Commun. 8, 15389 (2017). [CrossRef]
3. J. K. Rakshit and J. N. Roy, Opt. Appl. 46, 517 (2016). [CrossRef]
4. X. Tu, S.-L. Chen, C. L. Song, T. Y. Huang, and L. J. Guo, IEEE Photonics J. 11, 4200110 (2019). [CrossRef]
5. V. Zamora, P. Lützow, M. Weiland, and D. Pergande, Sensors 13, 14601 (2013). [CrossRef]
6. W. Bogaerts, P. De Heyn, T. Van Vaerenbergh, K. De Vos, S. Kumar Selvaraja, T. Claes, P. Dumon, P. Bienstman, D. Van Thourhout, and R. Baets, Laser Photonics Rev. 6, 47 (2012). [CrossRef]
7. Y. Chen, F. Yu, C. Yang, J. Song, L. Tang, M. Li, and J. J. He, Opt. Commun. 344, 129 (2015). [CrossRef]
8. C.-Y. Chao and L. J. Guo, J. Lightwave Technol. 24, 1395 (2006). [CrossRef]
9. J. H. Wade, A. T. Alsop, N. R. Vertin, H. Yang, M. D. Johnson, and R. C. Bailey, ACS Cent. Sci. 1, 374 (2015). [CrossRef]
10. H.-T. Kim and M. Yu, Opt. Express 24, 9501 (2016). [CrossRef]
11. N. Jokerst, M. Royal, S. Palit, L. Luan, S. Dhar, and T. Tyler, J. Biophoton. 2, 212 (2009). [CrossRef]
12. L. Jin, M. Li, and J.-J. He, Opt. Lett. 36, 1128 (2011). [CrossRef]
13. T. Claes, W. Bogaerts, and P. Bienstam, Opt. Express 18, 22747 (2010). [CrossRef]
14. L. Jin, M. Li, and J.-J. He, J. Lightwave Technol. 30, 1994 (2012). [CrossRef]
15. C.-Y. Chao, W. Fung, and L. J. Guo, IEEE J. Sel. Top. Quantum Electron. 12, 134 (2006). [CrossRef]
16. Z. Xie, Z. Cao, Y. Liu, Q. Zhang, J. Zou, L. Shao, Y. Wang, J.-J. He, and M. Li, Opt. Express 25, 33193 (2017). [CrossRef]
17. S. Dwivedi, H. D’heer, and W. Bogaerts, IEEE Photonics Technol. Lett. 25, 2167 (2013). [CrossRef]
18. D.-X. Xu, M. Vachon, A. Densmore, R. Ma, S. Janz, A. Delâge, J. Lapointe, P. Cheben, J. H. Schmid, E. Post, S. Messaoudène, and J.-M. Fédéli, Opt. Express 18, 22867 (2010). [CrossRef]
19. Y. Zhang, J. Zou, Z. Cao, and J.-J. He, Opt. Lett. 44, 299 (2019). [CrossRef]
