Open Access
Add to BibSonomy
Abstract
The terahertz (THz) metamaterial biosensor has great potential for label-free and rapid specificity testing. Here, we designed two highly sensitive structures to detect the carcinoembryonic antigen (CEA) of the cancer biomarker in early stages. There was about 29 GHz (500 ng/ml) resonance shift for CEA with an insert grate metamaterial, which was consistent with simulation results. Moreover, the concentration of CEA was gained through the relationship between the cancer marker concentration and frequency shift (Δƒ). Our design and detection methods may provide a potential route for the early warning stages of cancer.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
There is no obvious disease signal in the early stage of cancer in most case, which is often ignored by patients; consequently, the fatality rate of the cancer is high in humans. Therefore, the early warning of cancer has an important role in the diagnosis. Lots of researches show that cancer cells will produce different cancer marker proteins when they are formed, which can help people distinguish whether there are corresponding cancer cells in the body [1]. Among them, CEA is a broad-spectrum cancer marker, which rarely exists in the serum of healthy individuals. But in the serum of cancer patients (colorectal cancer, breast cancer, etc.), the concentration of CEA increases significantly. Thus, it’s a clinically essential marker for cancer detection.
Generally, methods for detecting protein concentration such as enzyme-linked immunosorbent assay (ELISA) [2] and fluorescence-based method [3] in serum samples are mainly used in biochemistry and molecular biology. However, those methods are time-consuming and may cause reaction between relevant molecules [4], which are unable to meet the demand of early warning of cancer. There are also some optic methods for biosensor detection, such as surface enhanced Raman scattering (SERS) [5], fluorescence-based microbial detection and surface plasmon resonance (SPR) [6]. It should be noticed that those detection methods maybe have drawbacks in sensitivity, volume of sample and detection limit. Therefore, it is highly desired to develop new biosensors that are more sensitive and easily operated.
Recently, terahertz sensors based on metamaterials (MMs) have received great attention due to the immediacy, label-less and non-destructive testing properties [7]. By changing the structure and size of the MMs resonance unit, it is possible to have a high response rate to the change of micro-environmental media. In the past decades, researchers have designed different MMs structures to improve their sensing performance, including SRRs type [8–10], cross type [11–13] and coupling type [14–16]. Among various structures, kinds of asymmetric MMs which stimulate the Fano resonance have raised the public concern due to their sharp spectra shape and higher Q resonance [17]. However, most structures being reported only used a single structure repeating MMs [8–10], which limited the tuning performance of MMs. There are some reports of asymmetrical MMs in the structure of composition unit such as a set of rotating cross [13] and adding grate in structures [15,18] that can create Fano resonances and have high THz wave tuning potential. In addition, reports on the detection of THz MMs biosensors mainly focused on nucleic acids [19], cellular [20–22], bacteria [23], small molecule solutions [24] and skin [25]. However, those researches mainly focused on the non-specific type of detecting which was not suitable for specific biomarker analysis. Furthermore, there were also some reports on the detection of specific proteins [26–28], but few works have been devoted to the quantitative detection and serum test, which plays a vital role in the detection of cancer marker.
In this paper, we propose the asymmetry MMs group based THz biosensors and demonstrate their use for quantitative THz time-domain spectroscopy (THz-TDS) detection of CEA. The optimal designs were obtained through comparing the transmission characteristics of different structures and substrates. We also compare the sensitivity of unimodal and multi-peak resonance mode in CEA protein, where the asymmetry MMs structures with multi-peak resonance exhibit a strong Fano resonance and show high sensitivity and low detection error for minute detection of CEA. Moreover, the relationship between CEA concentration and the frequency shift of transmission spectrum was demonstrated as a function by detailed measurement and simulation. According to the results of serums detection, we demonstrated the accuracy of quantitative relationship between concentration and the shifts of spectrum. This method might be very helpful for quantitative detection, cancer warning, medicine and biological industry.
2. Experiment
2.1 Metamaterial fabrication
To research the effect of asymmetry MMs group in resonance, we designed three asymmetry structures based on the classic double spilt-ring resonators (DSRRs) structures [10]. Those series of DSRRs structures (DSRRs, rotating DSRRs, grating DSRRs and inner grating DSRRs) were fabricated on silicon substrate or quartz substrate using a conventional photo-lithography. First, 50 nm Ti was deposited on the substrates, then 200 nm Au was deposited on the Ti through the electron beam evaporation. Finally, the lift-off technology was implemented to remove the metallic layer on the photoresist.
The structural diagrams and microscopic images of MMs are shown in Fig. 1. The basic structure consists of two concentric square SRRs, with two square gaps situating oppositely. Besides, to study the effect of rotating symmetric structure on THz transmission spectrum, we rotated the two adjacent DSRRs structures clockwise and countered clockwise by 22.5° respectively, and then obtained the rotating DSRRs structure. We also added the long transverse grating and short lengthways grating respectively, which were placed outside and inside the structure, to explore the effect of different grating on the THz transmission spectrum.
Fig. 1. Schematic diagram and microscopic image of we designed metamaterial structures. (a) DSRRs; (b) rotating DSRRs; (c) grating DSRRs; (d) inner grating DSRRs. (T = 50 µm, L = 45 µm, d1 = 7.5 µm, d2 = 8 µm, w = 6 µm, g = 2 µm, θ = 22.5°, H1 = 900 µm, H2 = 250 µm.)
2.2 Samples and measurement
Mercapto acetic acid in aqueous solution (MPA, 70 µL), 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide-HCI (EDC, 35 µL), N-hydroxy succinimide (NHS, 35 µL) and phosphate buffer solution (PBS, PH 7.4) were purchased from Sigma. Deionized water, Carcinoembryonic antigen (CEA) and bovine serum albumin (BSA) were purchased from Abcam (Britain). The serum of patients, with the concentration of CEA 392 ng/mL, was offered by Shandong University Qilu Hospital. Part of the serum is diluted ten times as comparison test.
The THz biosensor was put on the MPA (70 µL) solution in room temperature (25 °C) for about 24 h. During that period of time, the native sulfydryl groups in the MPA molecule were used to covalently bind to the Au surface of the DSRRs, which made it possible for CEA antibody modified on the Au layer.
The mixture solution of NHS (35 µL) and EDC (35 µL) was dropped on the surface of DSRRs at room temperature (25 °C) maintaining 15 min. They were functioned as activator to accelerate the rate of antibody modification. Afterwards, the CEA antibody was coupled to the surface of the DSRRs through Au-S bond by 12 h incubation period in refrigerator at 4 °C.
To prevent the non-specific binding, redundant active carboxyl groups were enclosed by 3% BSA at 4 °C maintaining 40 min. The BSA protein is commonly used for blocking in biomedical experiment [29] and make it more precise in detection. Figure 2(a) depicts the schematic diagram of THz MMs biosensor chip. The CEA protein sample is incubated on the DSRRs at 25 °C for one hour prior to spectrum testing. Additionally, to eliminate the effect of water absorption in THz beam, before spectral measurement, nitrogen gas was used to purge the remaining protein solution on the chip and to remove water vapor from the THz beam path. PBS was used to flush the test area 3 times between the two concentration measurements. We tested the transmission curve and sensing performances of the group sensors in a high precision THz time-domain spectrometer (0-3 THz spectrum range, 5 GHz resolution). The experiment measurement was under the temperature of 23 ± 0.5 °C, and the relative humidity was less than 2.0%. The devices were fixed in the testing region of the THz spectrometer. The transmitter transmission THz wave perpendicularly launched into the fabricated biosensors. The THz field incidentally paralleled to the DSRRs, forming a THz wave resonance in the MMs. Finally, the THz spectrum was obtained and displayed on the computer. Then we could get the time-domain THz pulse wave.
Fig. 2. (a) Schematic diagram of THz MMs biosensor chip; (b) Detection principle and the equivalent circuit for the DSRRs. (The capacitance of substrates and sample are expressed as $C\textrm{sub}$ and $C\textrm{sam}$, respectively. $Ls$ is expressed as inductance of DSRRs.)
We defined the THz time-domain spectrum measured in the air under the same environmental conditions as reference signal. After getting the reference signal in the air, the samples were measured subsequently. The sample and reference amplitude spectra were obtained from the time-domain data by Fourier transform and denoted as ${\tilde{E}_S}({\omega })$ and ${\tilde{E}_R}({\omega })$. The transmission spectra $\tilde{t}({\omega })$ can be obtained by:
Figure 2(b) shows the equivalent circuit diagram of the biosensor during protein detection. The DSRRs equivalent circuit is regarded as the LC circuit based on fundamental circuit theory [8,9]. Before detecting protein, the capacitance of MMs is regarded as $Csub$. After the protein binding to the surface antibody, the total capacitance has increased $Csam$, which changed the surface wave of MMs [30], causing the frequency shift in the transmission spectrum.
3. Results and discussion
3.1 Time-domain waveforms and transmission spectra
We first describe the THz time-domain and transmission spectra experiment for four kinds of DSRRs. As shown in Figs. 3(a) and 3(b), the detection spectrum of grating DSRRs is significantly different from the other three structures. After the spectral results of the frequency domain are processed by Fourier transform, the transmission spectra resonance characteristic mode of those DSRRs structures are close to 1.55 THz. As for rotating DSRRs, there is 0.05 THz red-shift and the narrowest half-peak width, which decreased by 0.1 THz when compared with original DSRRs. This is mainly because rotating the single DSRRs structure can increase the asymmetry of the overall structure, which results in a stronger Fano resonance and high Q value. There is no obvious change in frequency resonant mode between DSRRs and inner grating DSRRs, showing the adding of short grating in the inner ring does not affect the resonance characteristics of the structure. Because there are few effects in the capacitive characteristics of the structure. However, for grating DSRRs structure, except for generating the resonant mode of DSRRs at 1.55 THz, adding long grating outside the DSRRs structure can generate a new resonant peak at 0.93 THz. The grating array can be excited by electric field, acting as a new resonance mode.
Fig. 3. THz test results of different structures and substrates. (a)Time-domain waveforms; (b) Transmission spectra of four kinds DSRRs on quartz substrate.
In addition, in the process of fabrication, inner grating DSRRs is hard to be stripped. Therefore, we choose rotating DSRRs and grating DSRRs as the main research structure.
3.2 Substrate influence
To evaluate the impact of the substrate on the sensitivity of those chips, we selected rotating DSRRs and conducted the following two sets of experiments. Si substrate or quartz substrate were used in fabrication. Then, the THz transmission spectra of MMs, with and without CEA antibody (10 µg/mL), were measured for comparison purpose. The results are shown in Figs. 4(a) and 4(b), respectively. It can be seen that after modified CEA antibody on the surface, the silicon substrate THz biosensor has a displacement of 11 GHz, and the quartz substrate THz biosensor has a displacement of 25 GHz compared with the unmodified CEA antibody. It should be noticed that the frequency shift becomes more obvious when using quartz substrate THz biosensor. It suggests that simply replacing the silicon substrate (${\varepsilon } = 11.56$) with quartz (${\varepsilon } = 4.0$) [5] can increase the frequency shift (Δƒ), which shows that the lower dielectric constant of substrate can improve the response sensitivity of THz biosensor. Then, we must take the substrate into account to obtain a higher frequency shift.
Fig. 4. (a) Silicon substrate rotating DSRRs resonance frequency shift after modified CEA antibody. The inset shows the 11 GHz shift; (b) Quartz substrate rotating DSRRs resonance frequency shift after modified CEA antibody. The inset shows the 25 GHz shift.
3.3 Sensing performance simulation
To give an insight into the characteristics of the THz MMs biosensor, the CST simulation software is applied to explore the THz biosensing sensitivity. In this simulation, the surfaces of the quartz based biosensors are covered with a 5 µm-thick analyte which refractive index (n) from $n = 1.0$ to $n = 1.4$. The simulated transmission frequency of two kinds of DSRRs resonators under different ambient refractive index are shown in Figs. 5(a) and 5(c), and their resonance frequency shift curves are compared in Figs. 5(b) and 5(d), respectively. The results show that when the refractive index of the analyte increasing from 1.1 to 1.4, the resonance frequency decreases by about 82, 104 and 42 GHz in rotating DSRRs Dip1, grating DSRRs Dip1 and Dip2, respectively. The sensitive (S) of MMs biosensor is defined as $\Delta f/\Delta n$, where Δƒ is the resonance frequency shift, and Δn is the change of the analyte refractive index. In Figs. 5(b) and 5(d), the frequency shift shows good linearity with the refractive index (n), and the sensitivity of the rotating DSRRs Dip1 reaches 312 GHz/refractive index unit (RUI). As to grating DSRRs the sensitivity are 387 GHz/RIU and 150 GHz/RIU at Dip1 and Dip2, respectively. Compared with rotating DSRRs Dip1, grating DSRRs Dip1 shows a higher sensitive in the same refractive index change, confirming that the resonators with multi-peak resonance have a higher sensing capability.
Fig. 5. The simulated transmission spectra and sensitivity of THz MMs biosensor chip. (a) Transmission spectra of rotating DSRRs under different refractive index; (b) Frequency shift changes with refractive index of rotating DSRRs; (c) Transmission spectra of grating DSRRs under different refractive index, with the magnified red-shift shown in inset; (d) Frequency shift changes with the refractive index of grating DSRRs Dip1 and Dip2.
As to grating DSRRs, the sensitivity of Dip1 is much higher than Dip2 in grating DSRRs. By considering the oscillation frequency of the LC oscillation and the plasmonic oscillation, the difference in their sensitivity is analysed. By using the LC circuit model, the resonance frequency of the LC oscillation mode (Dip2) at low frequency resonance modes (LFRM) can be defined as:
where C and L are the capacitance and inductance of the DSRRs resonator, respectively [7]. According to Hara’s suggestion [31], the total capacitance C can be divided into five parts. Besides the capacitance properties affected by geometric factors, the change in $Csam$ only accounts for a small part of all the influencing factors, so the LFRM is usually reflected a small red-shift.As for high frequency resonance modes (HFRM), which is suggested as a plasmonic oscillation mode (Dip1), the resonance frequency (ω) is described as [7]:
3.4 Cancer marker protein and serum test
To further test the effectiveness of THz biosensor, two kinds of DSRRs (rotating DSRRs and grating DSRRs) were used to detect cancer biomarker CEA. In order to obtain the relationship between CEA concentration and frequency shift, we diluted CEA with PBS to a known concentration for measurement (50, 100, 250 and 500 ng/mL). The CEA protein samples were measured in the order of concentration from low to high, and reacted one hour at room temperature (25°C) to bind the antibodies sufficiently. Different concentrations of CEA proteins binded to antibodies, which have already modified on the surface of the MMs. The increase in CEA’s concentration causes the change of refractive index and the surface dielectric constant, which presents frequency red-shift in THz transmission spectra. The experimental transmission spectrum results of using rotating DSRRs structures are shown in Fig. 6(a). After adding the CEA antigen with different concentrations (50, 100, 250 and 500 ng/mL), the frequency red-shifts are 6.2, 13.6, 20.0 and 28.8 GHz, respectively (compared with the transmission mode after adding the antibody). The red-shift and concentration changes show a similar nonlinear relationship, and the rate of frequency displacement slows down. The fitting function is described by:
where the x represents the concentration of CEA. Figure 6(b) is the detection results of grating DSRRs structure. After adding different concentrations of CEA antigen, the frequency red-shift are 8.6, 12.0, 20.2 and 29.0 GHz in Dip1. As for Dip2, the shift, 4.3, 7.5, 15.0 and 22.0 GHz, is slightly less than the values observed from Dip1, respectively. The fitting function of grating DSRRs Dip1 and Dip2 are described as follows:As mentioned previously, the resonators with multi-peak resonance have a higher sensing capability then the single-peaks resonators. By comparison, it can be found that the displacement of the grating DSRRs Dip1 is larger than that of the rotating DSRRs Dip1 at the same concentration. The results of those CEA proteins detection are consistent with the simulation in part 3.3. Due to the combination of the double resonance effects, the device can achieve higher detection sensitivity among the current comparision tests.
Fig. 6. The typical THz detection results for different CEA protein concentrations and serum samples. (a) Frequency shift with different concentration of CEA protein based on rotating DSRRs; (b) Frequency shift with different concentration of CEA protein based on grating DSRRs; (c) The results of serum detect based on rotating DSRRs; (d) The results of serum detect based on grating DSRRs. The fitting function of Δƒ and x, i.e., the CEA concentration, is shown in (a) and (b), respectively.
To further evaluate it’s sensitive as a biosensor, the error of serum detection was investigated through the THz biosensor. Based on previous results, two kinds of the DSRRs were used to detect the content of CEA in two serum samples. The measurement results are shown in Figs. 6(c) and 6(d), respectively. It is known that the concentration of CEA in serum sample 2 is 10 times larger than serum sample 1. Substituting the Δƒ into Eqs. (5)–(7), the CEA concentration of different dip could be calculated.
The measurement results and calculated concentration are shown in Table 1. As shown in Table 1, Δƒ1 and Δƒ2 are the terahertz transmission frequency shift of serum sample 1 and 2, while C1 and C2 are the calculated CEA concentration of serum samples 1 and 2, respectively, according to the previously fitted formula [as shown in Figs. 6(a) and 6(b)]. As for HFRM (Dip1), the multiple relationships between C2 and C1 in rotating DSRRs and grating DSRRs are 9.39 and 9.48 times, respectively. According to the relationship between the concentration of serum 1 and 2 (10 times), the relative errors of CEA concentration calculated by this method in the two structures are 6.54% and 5.46%, respectively. Compared with rotating DSRRs, the CEA concentration identified by HFRM (Dip1) in grating DSRRs structure is more accurate. In addition to HFRM, there is also an LFRM (Dip2) in grating DSRRs, and the relative error of calculation based on LFRM migration is 12.9%. Moreover, the concentration of CEA in serum 2 (original serum) is 392 ng/mL. The error result of serum 2 measured by rotating DSRRs Dip1, grating DSRRs Dip1 and Dip2 are 14.3%, 10.7% and 49.6%, respectively. Obviously, when using grating DSRRs Dip 1 for analyzing, it has the minimum error of CEA detection. As mentioned earlier, the HFRM is more sensitive than LFRM. However, due to the emergence of LFRM, the quantification of HFRM will be more accurate. Therefore, for the design of MMs, multi-peak resonance should be considered. Among them, the concentration of tumour markers in serum is extracted by using HFRM as the standard of quantitative detection.
Table 1. The test results of CEA content in serum.
4. Conclusion
In conclusion, we designed four kinds of DSRRs MMs THz biosensors, which have been fabricated on silicon or quartz substrates. The structures of different MMs were characterized by THz-TDS system, and the detection of CEA concentration by MMs biosensors was analyzed. From the CEA protein measuring results, we obtained the relationship between Δƒ and CEA concentration was compliance with the logarithm function formula. The CEA concentration in the serum could be calculated from the function. When MMs structure is designed, the multi-peak resonance patterns should be considered, which can quantify serum more accurately. Moreover, the structure will be further optimized by MMs and combined with microfluidic technology in more detecting field, such as distinguishing chiral proteins, high accuracy detection of serum, liquid sensing and so on. The THz biosensor has promising applications in bio-molecules, disease diagnosis, and medicine.
Funding
Capital Health Research and Development of Special (2020-2-4084); Youth Innovation Promotion Association of the Chinese Academy of Sciences (Y201925); National Natural Science Foundation of China (61874105); National Key Research and Development Program of China (2017YFB0405400).
Acknowledgments
The authors would like to acknowledge Dr. Weihao Fang for the simulation support, and Mr. Jin Li in Beijing Daheng photoelectric technology co. LTD for the assistance in THz-TDS test.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. J. A. Ho, Y. C. Lin, L. S. Wang, K. C. Hwang, and P. T. Chou, “Carbon nanoparticle-enhanced immunoelectrochemical detection for protein tumor marker with cadmium sulfide biotracers,” Anal. Chem. 81(4), 1340–1346 (2009). [CrossRef]
2. L. Ma, J. Zhang, H. Chen, J. Zhou, Y. Ding, and Y. Liu, “An overview on ELISA techniques for FMD,” Virol. J. 8(1), 419 (2011). [CrossRef]
3. D. M. Alloway, M. Hofmann, D. L. Smith, N. E. Gruhn, and A. L. Graham, “Interface Dipoles Arising from Self-Assembled Monolayers on Gold: UV− Photoemission Studies of Alkanethiols and Partially Fluorinated Alkanethiols,” J. Phys. Chem. B 107(42), 11690–11699 (2003). [CrossRef]
4. R. Chen, R. H. Guo, H. X. Zhu, and Z. D. Shi, “Development of a sandwich ELISA for determining plasma prolactin concentration in domestic birds,” Domest. Anim. Endocrinol. 67(1), 21–27 (2019). [CrossRef]
5. A. Zybin, V. Shpacovitch, J. Skolnik, and R. Hergenroder, “Optimal conditions for SPR-imaging of nano-objects,” Sens. Actuators, B 239(1), 338–342 (2017). [CrossRef]
6. H. Etayash, K. S. Jiang, S. Azmi, T. Thundat, and K. Kaur, “Real-time detection of breast cancer cells using peptide-functionalized microcantilever arrays,” Sci. Rep. 5(1), 13967 (2015). [CrossRef]
7. X. Wu, B. Quan, X. Pan, X. Xu, X. Lu, C. Gu, and L. Wang, “Alkanethiol-functionalized terahertz metamaterial as label-free, highly-sensitive and specific biosensor,” Biosens. Bioelectron. 42(1), 626–631 (2013). [CrossRef]
8. O. Sydoruk, E. Tatartschuk, E. Shamonina, and L. Solymar, “Analytical formulation for the resonant frequency of split rings,” J. Appl. Phys. 105(1), 014903 (2009). [CrossRef]
9. R. Singh, I. A. Al-Naib, M. Koch, and W. Zhang, “Asymmetric planar terahertz metamaterials,” Opt. Express 18(12), 13044–13050 (2010). [CrossRef]
10. T. J. Yen, W. Padilla, N. Fang, D. Vier, D. Smith, J. Pendry, D. Basov, and X. Zhang, “Terahertz magnetic response from artificial materials,” Science 303(5663), 1494–1496 (2004). [CrossRef]
11. F. B. Zarrabi, M. N. Moghadasi, S. Heydari, M. Maleki, and A. S. Arezomand, “Cross-slot nano-antenna with graphene coat for bio-sensing application,” Opt. Commun. 371(1), 34–39 (2016). [CrossRef]
12. Z. Vafapour and H. Ghahraloud, “Semiconductor-based far-infrared biosensor by optical control of light propagation using THz metamaterial,” J. Opt. Soc. Am. B 35(5), 1192–1199 (2018). [CrossRef]
13. B. Reinhard, K. M. Schmitt, and V. Wollrab, “Metamaterial near-field sensor for deep-subwavelength thickness measurements and sensitive refractometry in the terahertz frequency range,” Appl. Phys. Lett. 100(22), 221101 (2012). [CrossRef]
14. J. Yuan, H. Z. Cao, Y. Y. Xie, Z. X. Geng, Q. Kan, X. M. Duan, and H. D. Chen, “Gold Elliptic Nanocavity Array Biosensor With High Refractive Index Sensitivity Based on Two-Photon Nanolithography,” IEEE Photonics J. 7(1), 1–8 (2015). [CrossRef]
15. X. Liu, J. Gu, R. Singh, and Y. Ma, “Electromagnetically induced transparency in terahertz plasmonic metamaterials via dual excitation pathways of the dark mode,” Appl. Phys. Lett. 100(13), 131101 (2012). [CrossRef]
16. T. T. Yeh, S. Genovesi, A. Monorchio, E. Prati, F. Costa, T. Y. Huang, and T. J. Yen, “Ultra-broad and sharp-transition bandpass terahertz filters by hybridizing multiple resonances mode in monolithic metamaterials,” Opt. Express 20(7), 7580–7589 (2012). [CrossRef]
17. R. Singh, I. A. Al-Naib, M. Koch, and W. Zhang, “Sharp Fano resonances in THz metamaterials,” Opt. Express 19(7), 6312–6319 (2011). [CrossRef]
18. F. Hu, Y. Qian, Z. Li, J. Niu, K. Nie, X. Xiong, W. Zhang, and Z. Peng, “Design of a tunable terahertz narrowband metamaterial absorber based on an electrostatically actuated MEMS cantilever and split ring resonator array,” J. Opt. 15(5), 055101 (2013). [CrossRef]
19. Y. Yang, D. Xu, and W. Zhang, “High-sensitivity and label-free identification of a transgenic genome using a terahertz meta-biosensor,” Opt. Express 26(24), 31589–31598 (2018). [CrossRef]
20. C. Zhang, L. Liang, L. Ding, B. Jin, Y. Hou, C. Li, L. Jiang, W. Liu, W. Hu, and Y. Lu, “Label-free measurements on cell apoptosis using a terahertz metamaterial-based biosensor,” Appl. Phys. Lett. 108(24), 241105 (2016). [CrossRef]
21. X. Yan, M. Yang, Z. Zhang, L. Liang, D. Wei, M. Wang, M. Zhang, T. Wang, L. Liu, J. Xie, and J. Yao, “The terahertz electromagnetically induced transparency-like metamaterials for sensitive biosensors in the detection of cancer cells,” Biosens. Bioelectron. 126(1), 485–492 (2019). [CrossRef]
22. S. J. Park, J. T. Hong, S. J. Choi, H. S. Kim, W. K. Park, S. T. Han, J. Y. Park, S. Lee, D. S. Kim, and Y. H. Ahn, “Detection of microorganisms using terahertz metamaterials,” Sci. Rep. 4(1), 4988 (2015). [CrossRef]
23. W. Zhang, E. R. Brown, L. Viveros, K. P. Burris, and C. N. Stewart, “Narrow terahertz attenuation signatures in Bacillus thuringiensis,” J. Biophotonics 7(10), 818–824 (2014). [CrossRef]
24. W. Xu, L. Xie, J. Zhu, W. Wang, Z. Ye, Y. Ma, C. Y. Tsai, S. Chen, and Y. Ying, “Terahertz sensing of chlorpyrifos-methyl using metamaterials,” Food Chem. 218(1), 330–334 (2017). [CrossRef]
25. Q. Sun, E. P. J. Parrott, Y. He, and E. P. MacPherson, “In vivo THz imaging of human skin: Accounting for occlusion effect,” J. Biophotonics 11(2), e201700111 (2018). [CrossRef]
26. R. Zhang, Q. Chen, K. Liu, Z. Chen, K. Li, X. Zhang, J. Xu, and E. P. MacPherson, “Terahertz Microfluidic Metamaterial Biosensor for Sensitive Detection of Small-Volume Liquid Samples,” IEEE Trans. Terahertz Sci. Technol. 9(2), 209–214 (2019). [CrossRef]
27. Y. Li, X. Chen, F. Hu, D. Li, H. Teng, Q. Rong, W. Zhang, J. Han, and H. Liang, “Four resonators based high sensitive terahertz metamaterial biosensor used for measuring concentration of protein,” J. Phys. D: Appl. Phys. 52(9), 095105 (2019). [CrossRef]
28. Z. Geng, X. Zhang, Z. Fan, X. Lv, and H. Chen, “A Route to Terahertz Metamaterial Biosensor Integrated with Microfluidics for Liver Cancer Biomarker Testing in Early Stage,” Sci. Rep. 7(1), 16378 (2017). [CrossRef]
29. I. Sarangadharan, A. Regmi, Y. Chen, C. Hsu, P. Chen, W. Chang, G. Lee, J. Chyi, S. Shiesh, G. Lee, and Y. Wang, “High sensitivity cardiac troponin I detection in physiological environment using AlGaN/GaN High Electron Mobility Transistor (HEMT) Biosensors,” Biosens. Bioelectron. 100(1), 282–289 (2018). [CrossRef]
30. P. Mehrotra, B. Chatterjee, and S. Sen, “EM-Wave Biosensors: A Review of RF, Microwave, mm-Wave and Optical Sensing,” Sensors 19(5), 1013 (2019). [CrossRef]
31. J. F. O’Hara, R. Singh, I. Brener, E. Smirnova, J. Han, A. J. Taylor, and W. Zhang, “Thin-film sensing with planar terahertz metamaterials: sensitivity and limitations,” Opt. Express 16(3), 1786–1795 (2008). [CrossRef]
32. E. Cubukcu, S. Zhang, Y. S. Park, G. Bartal, and X. Zhang, “Split ring resonator sensors for infrared detection of single molecular monolayers,” Appl. Phys. Lett. 95(4), 043113 (2009). [CrossRef]
