- Email: [email protected]
Higher-order mode photonic crystal based nanofluidic sensor
Optics Communications 382 (2017) 105–112
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Higher-order mode photonic crystal based nanofluidic sensor Wang Peng a,b,c, Youping Chen a,n, Wu Ai a a b c
School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA Micro and Nanotechnology Laboratory, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 9 June 2016 Received in revised form 7 July 2016 Accepted 8 July 2016 Available online 6 August 2016
A higher-order photonic crystal (PC) based nanofluidic sensor, which worked at 532 nm, was designed and demonstrated. A systematical and detailed method for sculpturing a PC sensor for a given peak wavelength value (PWV) and specified materials was illuminated. It was the first time that the higher order mode was used to design PC based nanofluidic sensor, and the refractive index (RI) sensitivity of this sensor had been verified with FDTD simulation software from Lumerical. The enhanced electrical field of higher order mode structure was mostly confined in the channel area, where the enhance field is wholly interacting with the analytes in the channels. The comparison of RI sensitivity between fundamental mode and higher order mode shows the RI variation of higher order mode is 124.5 nm/RIU which is much larger than the fundamental mode. The proposed PC based nanofluidic structure pioneering a novel style for future optofluidic design. & 2016 Elsevier B.V. All rights reserved.
Keywords: Photonic crystals Optofluidic biosensors Label-free detection Sensitivity
1. Introduction PC (Photonic crystal) is a periodic grating structure, with high refractive index guided layer surrounded by lower refractive index layers. Since the initial work of Yablonovitch [1], PCs have attracted increasing interest due to their intrinsic advantages, such as low cost [2], large electrical field enhancement [3], and high sensitivity [4–7]. PCs are usually designed to couple strongly with special optical wavelength according to the materials and the modulated grating period. When matched with the wavelength of resonance, light will be confined strongly to the guided layer and results in an electric field related to the PC many times larger than the field of incident light [8,9]. Since the changing of the ambient medium parameters can induce resonance wavelength shift due to the effective RI variation of the PC, it can vary the distribution of the related evanescent filed. As a result, PC has been used in many different fields including label-free biosensing [10,11], lasing [12], PCEF (photonic crystal enhanced fluorescence) [13], PCEM (photonic crystal enhanced microscopy) [14], and PC micro-fluidics [15,16]. Recently, PC have been utilized to sense ambient refractive index [17,18], protein detection [19], and antibody in the micro fluidic [15] and cell imaging [20]. However, nearly all the reported sensing schemes adopt the detecting area on the surface of guided layer, and the enhanced electromagnetic field is located inside or on the corner of the guided layer [8–20]. In the past, researchers n
Corresponding author. E-mail address: [email protected] (Y. Chen).
http://dx.doi.org/10.1016/j.optcom.2016.07.019 0030-4018/& 2016 Elsevier B.V. All rights reserved.
were emphasized on confining the guided layer thin enough to suppress the arising of higher-order mode [21–23]. Even high value of enhanced electrical field value can be acquired, the distribution of the enhanced field is not good enough. Also, PCs have certain up limits in the field of sensing since a large fraction of the optical power is coupled into the guided layer instead of propagates in the evanescent field outside the physical boundary of the guided layer. By this way, the enhanced electrical field is separated from the analytes on the surface and cannot make a wholly contribution to the sensor. As the enhancement factor of evanescent field is exponentially decay on the above of PC surface [20,24], the sensing area of traditional PCs is very limited. Meanwhile, microfluidic PC structures have been used for Reflective index sensing related area, like the detection of dissolved avidin concentrations on slotted photonic crystal cavities [25], measurement of the kinetic binding interaction of protein A with IgG molecules on polymer microfluidic channels [26], and microfluidic refractive index sensors [27] in the past few years. As far as we know, all these PC based microfluidics in the above references [24–27] are worked on the fundamental mode of resonance, in which the enhanced evanescent field will be mostly distributed on the edge of the grating angle or inside the guided layer, and cannot interact significantly with the detection sample. Nowadays, nanofluidic sensors, which can be used for biochemical, medical and life science detection, have been investigated worldwide [28]. Most of these nanofluidic sensors are fabricated as a single fluidic channel, and the analytes inside the channel need to be detected with aided tags, such as fluorescence [29], current [30]. To design a nanofluidic sensor that can test analytes with its own characteristics is of key importance.
106
W. Peng et al. / Optics Communications 382 (2017) 105–112
In this paper, a novel PC based nanofluidic sensor has been presented, which can work on the higher-order mode and has a detection area set inside the PC. To the best of our knowledge, there has been no publish to date of PC based nanofluidic sensor which is working on higher-order mode with liquid channel inside the PC structure. A well designed structure has been proposed, which is working on the 532 nm wavelength on resonance. The resonance peak induced by the higher-order mode is used to sense ambient RI (refractive index), instead of the fundamental mode, and the electric field distributes on the center of detection area while working in higher-order mode. The proposed PC based nanofluidic sensor possesses high sensitivity, compact size, and multi channels which can be used to detect different samples at the same time. This paper gives detailed method about how to design a PC sensor and optimize the related parameters to obtain the ideal structure. Based on FDTD simulation results, the superior performance of higher-order TE01 mode had been analyzed and demonstrated by simulation results. A brief way of how to fabricate this sensor was given. Also, the upcoming issues when passing from the theoretical part to the experiments were discussed. In the future, we will focus on fabricating and testing its performance with experiments.
2. Physical principle and device scheme The physical phenomenon behind the narrowband reflection from PC has been described previously [31,32] as the periodic index differences enabling the exciting of guided mode resonance. The resonance wavelength occurs when the incident light satisfies the phase matching condition [33]:
⎛ 2π ⎞ ⎟ = β k 0nc sin θ ± m⎜ ⎝ Λ⎠
(1)
where k0 = 2π /λ is the wave vector in free space, θ is the incident angle, nc is the refractive index of the cladding area (in this structure, nc is decided by the refractive index of Pyrex and liquid in nanofluidic channel), m = 0, ± 1... is the diffraction order, and 2π β = λ neff is the propagation wave vector in the PC structure, and neff is the effective index. The PC based nanofluidic sensor in this research has a Si3N4 based periodic structure built on the Pyrex substrate and covered by glass, while the nanofluidic channel is made by the gap between gratings. The grating period (Λ) and grating depth (d) of the sensor structure are depicted in Fig. 1. Since the gap between the Si3N4 based periodic structures will be used as nanofluidic channels, we plan to set it in a reasonable way, which means the size of the channel should be possible to fabricate with normal cleanroom equipment. For this PC based nanofluidic sensor, the detection of light is under normal incidence (θ ¼0) on the PC plane, and the targeted resonance wavelength will be 532 nm. Thus, θ ¼0, and we can rewrite the formula (1) as follows:
mλ = Λneff
(2)
For a certain period, when there is a PWV according to the effective index changes resulting from ambient RI perturbation, we can get:
mΔλ = ΛΔneff
(3)
From (2) and (3), we can deduce that the sensitivity of PC based nanofluidic sensor is decided by the perturbation of effective index, which is mainly due to the interaction between the electrical field and the nanofluidic channels, as we can calculate from Eq. (4). The resonance wave couples into the guided layer due to the perturbation of the grating structure. The ambient refractive index
Fig. 1. PC based nanofluidic design comprised on a Pyrex substrate, a Si3N4 grating layer, and a SiO2 superstrate. (a) PC based nanofluidic sensor schematic diagram; (b) Parameters and refractive indexes (RI) of PC based nanofluidic sensor; d: grating depth, ng: RI of Si3N4, ntest: RI of test analyte (e.g. water), ns: RI of substrate Pyrex, nc: RI of cladding layer, SiO2.
vary leads to redistribution of electric field and concomitant spectral changes of the reflected far fields which can be detected by a spectrometer. The effective index can be regard as the average of the refractive indices of the materials at resonance mode, and can be written as [34]: ∞
2 neff
=
Λ
∫−∞ ∫0 ε( x, y) E(x, y) 2 dxdy ∞
Λ
∫−∞ ∫0 E(x, y) 2 dxdy
(4)
where ε(x, y) is the dielectric permittivity, and E(x, y) is the electrical field intensity. In order to improve the sensitivity of the sensor, the enhanced electrical field need to be confined mostly in the channel, where the interaction is exist.
3. Design method and results analysis In this section, an analysis method for optimizing the electrical field enhancement within the channels in the sensor and predicting the absolute resonance wavelength at 532 nm have been proposed and discussed. 3.1. Design method For a certain period, there is only fundamental mode when the grating depth is very thin. Whereas the higher-order mode occurs as the grating layer is thick enough. For any specific PC, there are two types of input parameters: structure-determined parameters and user-intent parameters [35]. The structure-determined parameter include grating depth d, refractive index of the cladding layer nc, grating layer ng, substrate ns, and test sample area ntest, grating period Λ and fillfactor f which are determined by the structure of PC. The user-intent parameters include resonance wavelength value λ, incident angle θ, and polarization (TE, TM), which can be tuned by designer after the PC structure have been fabricated. In this paper, the structure-determined parameters were used to design the PC for PWV at 532 nm, and the user-intent
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 2. Design flow of the PC based nanofluidic structure.
parameters were used to optimize the electromagnetic field and sensitivity of this sensor. The PC designs demonstrated in this paper are based on a 500 um thick Pyrex substrate and a guiding layer about 200– 300 nm. The two-dimensional finite-difference time-domain (FDTD) method was used to simulate the resonance wavelength and its related electrical field distribution. FDTD is the finite-difference time-domain algorithm, which divides space and time into a regular grid and simulates the time evolution of Maxwell equations [36]. In this section, different structures resonance at a certain wavelength value, 532 nm, had been simulated, since 532 nm is a general used light source which can be achieved easily when set biodetection experiments in the future. The peak wavelength value, peak reflectance, and bandwidth of the PC spectrum were all decided by the grating depth d, period Λ, fillfactor f and polarization direction. Although tuning the fillfactor of the grating can modulate the peak wavelength value [37], grating depth is the most commonly adapt parameter, which can be finely tuned using deposition, lithography and etching process [38]. In order to simplify the analysis process, the fillfactor was set as 0.5 which can be obtain easily in experiments [39] and incident angle as 0° due to most of the PC based biosensor structure. Instead of continuously changing grating period, several discrete values based on the equality for resonance regime of GMR filter had been chosen [38]:
max {nc , ns , ntest } ≤
mλ ≤ ng Λ
(5)
where m is the order of diffracted wave, nc, ng, ns and ntest are the refractive indices of the cladding layer, grating layer, substrate, and test sample area respectively. The refractive indices were set as nc ¼1.5, ng ¼2.02, ns ¼1.5 and ntest ¼ 1.33 at 532 nm [40]. Thus, from Eq. (5), the range of the grating period will be 264 nm o Λ o 355 nm, we choose discrete periods Λ ¼310, 320, 330, 340, 350, 360 nm. The design flow is given in Fig. 2. The design flow is as follows: firstly the PC structure was proposed as in Fig. 1, with Si3N4 as the high refractive index guiding layer material, SiO2 as the superstrate, and Pyrex as substrate. Then the resonance wavelength for this sensor was defined
107
as 532 nm, with normal incidence, and fillfactor as 0.5 which is easy to fabricate. Also, a set of discrete grating periods, Λ1, Λ2… based on the Eq. (5) had been set. Then, simulation with FDTD was used to get the expected result. In the FDTD solutions, when a plane wave is interacting with the PC, the simulation area is confined in a grating lattice. As with many periodic grating structures, a mesh override region is required at the grating. The boundary condition of top and bottom side is PML (perfect matched layer), which will absorb light waves with minimum reflection. While the boundary condition of left and right side is Bloch boundary conditions. The Bloch boundary condition replicate the fields at one edge of the simulation region and re-inject at the other edge with a phase correction of the fields. Before the simulation start, the polarization of the light source was chosen, TE or TM. The TE mode is defined as the electrical field which is perpendicular to the incident plane and parallel to the direction of gratings. The TM mode is defined as the magnetic field which is perpendicular to the incident plane and parallel to the direction of gratings. Then, a specific period was used, and gradually tuned the depth of the grating, to build contours of PWV, and reflection of PWV. Finally, the 532 nm resonance with a specific grating was there for a certain period and polarized direction. When increase the grating depth step by step, there were 3 different which resonant on 532 nm, which are single mode resonance, fundamental mode resonance in multi-mode and higher-order mode resonance in multi-mode. The single mode resonance of PC is indicated as there is only one peak exist among the whole reflection spectrum. Fundamental mode resonance in multi-mode is the zero order diffraction among the phase matching condition in Eq. (1), which is usually the rightmost peak in the reflection spectrum. Higherorder mode resonance in multi-mode is the first order diffraction which is the peak wavelength next to the fundamental mode peak. All these definitions are effective when the light is under normal incidence. Also, the electrical field for each result was plotted. Finally, the optimized design was selected according to three factors: (1) electrical field enhancement. (2) The ratio of enhancement in the detection area. (3) The sensitivity of the selected mode. The way of how to using PWV contour to determine the grating depth for 532 nm resonance had been explained clearly in a former paper from our group [18], so it will not show it any more. Instead, the main research was focused on how to confine the enhanced electrical field inside the detection area and get a reasonable channel depth, about 150–300 nm. The TE and TM mode resonance had been analyzed independently. 3.2. Results analysis Based on the design method of PC based nanofluidic structure, a certain grating depth and related parameters for PWV (peak wavelength value) on 532 nm were obtained with FDTD Lumerical. As shown in supplementary materials, channel area was represented by dash line and red arrow showed the location of PWV at 532 nm. In each TE or TM mode polarized direction, three different ways of PWV centered at 532 nm had been analyzed, which is single mode 532 nm resonance, fundamental mode 532 nm resonance in multimode, and higher-order mode 532 nm resonance in multimode. Basically, these three modes occurred sequentially as the grating depth increasing for a certain grating period. However, the modes will be cut off for a specified grating, since the PWV will be away from 532 nm when the grating period is too small, according to Eq. (2). All the possibilities of PWV at 532 nm were shown in Fig. 3 and its related electromagnetic filed in TM mode were shown in supplementary material. TM0 curve in Fig. 3 was used to analysis the potential single mode PC structure which will resonance on 532 nm, which indicated that the single peak resonance was not
108
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 3. Grating depth vs grating period for TM mode PWV at 532 nm.TM0: single TM mode PC resonance at 532 nm; TM00: fundamental TM mode PC resonance at 532 nm in multi-mode structure; TM01: higher-order mode PC resonance at 532 nm in multi-mode structure.
Fig. 4. Grating depth vs grating period for TE mode PWV at 532 nm. TE0: single TE mode PC resonance at 532 nm; TE00: fundamental TE mode PC resonance at 532 nm in multi-mode structure; TE01: higher-order mode PC resonance at 532 nm in multi-mode structure.
arising at 532 nm and only exist on Λ ¼340 nm for the given discrete periods from 310 nm to 360 nm. The 532 single peak resonance was cutoff when the period reaches to 350 nm or 330 nm. In supplementary material Fig. 1(a) shows the reflection spectrum and electromagnetic field while PWV at 532 nm. Although the electrical field was enhanced largely, the normalized reflection
intensity was only 0.3 at 532 nm. The fundamental mode (TM00) of multi-mode PC structures resonance on 532 nm indicate that the fundamental mode resonance on 532 nm was cutoff when the period reaches 340 nm, and there were resonances while the period was from 310 nm to 330 nm. In supplementary material Fig. 1(b), the electrical field mostly been largely enhanced. However, the enhanced area is in the corner of gratings, and mostly are inside the Si3N4, with little enhanced E field can interact with the detection area. According to TM01 curve, when the period is around 330 – 350 nm, the TM01 mode in multimode was generated. The related normalized reflection spectrums and electromagnetic fields were shown on supplementary material Fig. 1 (c) when the period was located on 330 nm, 340 nm and 350 nm respectively. The E field enhancement area was located around the channel area, yet its E field enhancement was not high, about 5 times of the normalized incident E field averagely (Emax/E0). In conclusion, for all the simulated TM mode, the enhanced E fields are usually gathered around the corners of grating structure, and cannot interact well with the detection area. So, we may speculate that the TM mode will not be the best style for PC based nanofluidic sensor in this structure. Fig. 4 gave all the possibilities of PWV at 532 nm in TE mode. TE0 curve was used for the analysis of single mode PC structure. For the given discrete periods, the single mode structures had peak at 532 nm resonance when the period was tuning from 330 nm to 350 nm, as shown in Supplementary material Fig. 2(a). From the comparison of normalized reflection spectrums and E-field plots, it indicated that the band width of resonance peak became much narrow and normalized reflection intensity decreased gradually as the grating period increased from 330 nm to 350 nm, but its E field value enhanced gradually as the period increased. For the E-field enhancement of the dashed line represented nanochannel area, half of the enhanced area was distributed in the detection area. The second set of pictures, in supplementary material Fig. 2(b) was focused on the study of fundamental mode (TE00) resonance on 532 nm in multimode sensor structures. The peak occurred when the period is between 320 nm and 340 nm, as shown from TE00 curve in Fig. 4. The bandwidth of the resonance peak at 532 nm was gradually increased, as shown in supplementary material Fig. 2(b). The E field enhancement value was around 10, and the enhanced E field area mostly upon the detection area. The last set of figures, as shown in f in supplementary material Fig. 2(c), was the higher-order mode resonance on 532 nm which is TE01 mode sensor structures. As the grating depth increased to a certain extent (less than 1 um), the TE01 mode with periods vary from 320 to 360 nm is always exist, as shown in Fig. 4, TE01 curve. The TE01 mode will be cutoff at
Fig. 5. Sensor structure: Λ 350 nm, d 297 nm, TE01 mode. Left: spectrum; right: electrical field intensity. (a) TE higher-order mode structure, Λ ¼350 nm, d ¼297 nm. (b) TE fundamental mode, Λ ¼ 320 nm, d ¼323 nm. (c) TE single mode, Λ ¼ 330 nm, d ¼80 nm.
W. Peng et al. / Optics Communications 382 (2017) 105–112
109
Table 1 Summary of Designed PC nanofluidic sensor with at PWV 532 nm for various modes.
Fig. 7. Comparison the reflection wavelengths shift on the ambient RI between higher-order mode, fundamental mode and single mode resonance PWV at 532 nm. The higher order mode at TE1 is much sensitive, which means using higher order mode can improve the performance of the PC based nanofluidic sensor.
Fig. 6. Normalized reflection spectrum shift due to the change of RI in detection area. Black arrow indicates the red shift of PWV initially from 532 nm, when tuning the ambient RI.
310 nm, where the grating depth is too large in compare with grating. When the grating depth is much larger than the period, the wavelength shift is slow which decrease sensitivity of the sensor. However, in a certain range of grating depth, the sensitivity of this higher-order mode is significant. From the E field plots in supplementary material Fig. 2(c), it can be indicated that the TE01 mode was confined in the detection area and its E field enhancement was about 15, especially in the 350 nm period with grating depth 297 nm, as shown in Fig. 5. In conclusion, for all the simulated TE mode, the enhanced E fields are usually gathered around the center of gratings or detection area, which can interact well with the detection area. For the TE01 mode, the enhanced E filed is
110
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 8. Schematic of fabrication process.
4. Optimized structure for refractive index (RI) sensing and fabrication process design 4.1. Optimized structure for refractive index (RI) sensing
Fig. 9. Schematic of PC based nanofluidic sensor.
centered in the detection area, and the E field is also significant, which may achieve a better sensitivity for the sensor. We tried to target the channel size in a reasonable range, which means the grating depth should not be too low in order to let the sample flow smoothly in the nanofluidic channel. With the related spectrum and E field information from Fig. 3, Fig. 4 and Supplementary material, the parameters were reorganized, and the whole parameters satisfy the requirements are shown in Table 1. In Table 1, all the modes and parameters related to this sensor which can arise 532 nm resonance have been given and sort by category. In order to decide which method is the best for this sensor, the last three parameters: peak reflection value, E field enhancement, and E field distribution in the detection area had been analyzed. With this three factors, the optimized structure is the TE higher-order mode, with period 350 nm and grating depth 297 nm. Thus, we can deduce that the TE mode, higher-order resonance at 532 nm with period 350 nm and grating depth 297 nm has the best performance.
After the optimized process, the optimized PC based nanofluidic sensor have been designed, which is the higher-order mode in multimode, TE polarized, with a grating period Λ ¼350 nm, and grating depth d¼ 297 nm. Then, its reflection spectrum with different refractive index among the detection area from 1.33 to 1.40 was analyzed. As shown in Fig. 6(a), it can be easily discovered that two main reflected peaks occur in the reflection spectrum. The PWV at long wavelength part is due to the coupling between the propagating fundamental TE0 mode and grating structure, and the other PWV which resonance at 532 nm is due to the coupling between higherorder TE01 mode and grating structure. When the refractive index in the detection area change from 1.33 to 1.40 gradually, the red shift of TE00 and TE01 mode was simultaneously. When noticing the red shift distance, it can be deduced that the higher-order mode's red shift is much larger than the fundamental mode, since the enhanced electrical field of the TE01 mode is mainly concentrated on the channels area as can be seen from Fig. 6(a) and Fig. 5. Then, the other two different styles has a PWV at 532 nm, which is fundamental mode resonance in multimode as Fig. 6(b) and single mode resonance as Fig. 6(c), were also analyzed. As it can be seen from these two figures, the shift at 532 nm is slow when the RI in the detection area varies. From the related E field plot, the conclusion can be given that RI sensitivity is related to the E field enhancement, especially when the enhance area is centered in the detection area. The RI sensing performance is demonstrated with numerical simulations by compare three different structures with PWV at 532 nm, as shown in Fig. 5. From Fig. 7, it shows that the higherorder mode has a significant sensitivity than the other two modes. Also, it indicates that the maximum RIU of TE1 mode is about 124.5 nm/RIU (RI Unit), while the RIU of TE0 mode and single peak mode are 60 nm/RIU and 45.9 nm/RIU, respectively. 4.2. Fabrication process design In this section, a planned fabrication process was given. In order to fabricate such a PC based nanofluidic sensor with a grating period Λ ¼ 350 nm, and grating depth d ¼297 nm, a schematic fabrication process was shown in Fig. 8. The whole process includes six parts: (1) SiO2 deposition, (2) Photo resist (PR) coating, (3) E-beam lithography, (4) SiO2 etching, (5) Si3N4 etching,
W. Peng et al. / Optics Communications 382 (2017) 105–112
(6) Anodic bonding between Pyrex glass and PC. A common polished Si or SiO2 glass can be selected as substrate (Si wafer was selected as substrate in this design). Then, 2 um can be deposited on the polished side of Si wafer with PECVD (Plasma Enhanced Chemical Vapor Deposition) [41]. Since the feature size of the grating is only 350 nm, e-beam lithography was selected as the feature sculpturing method [42]. For the e-beam lithography process, a thin layer of related PR will be coating on the SiO2 surface initially, and then the e-beam will write the required 350 nm feature on the PR. After e-beam lithography, the PR feature will be used as mask to etch 297 nm SiO2 with RIE (reactive ion etching) tool [43]. Then, the PR will be cleaned by acetone, IPA, water and IPA in sequence. A thickness of 297 nm Si3N4 will be deposited on the etched SiO2 grating surface with PECVD tool [44]. With this fabricated PC structure, anodic bonding process can be used to seal the PC gratings with Pyrex glass [45]. In this way, the PC based nanofluidic channel can be realized. During the fabrication process, the parameters may be not as accurate as designed. The reason can be classified as follows: (1) the performance of related fabrication equipment, (2) the metrology device used to measure the related size, (3) the ambient influence, such as particle level and temperature. Therefore, there may be limitations when passing from theoretical approach to the experiments. The schematic of the PC based nanofluidic sensor was shown in Fig. 9. The PC based nanofluidic sensor is located in the center of the chip and connected with inlets and outlets. A white light source is initially collimated by a convex lens and then TE polarized by a polarizer. After this, the TE polarized collimation light to reach the sensor surface at normal incident angle. A resonance light will be reflected back and collected by a spectrometer. The peak wavelength of the reflection spectrum will red shift as the concentration of analytes in the nanochannels gradually increase.
5. Conclusion In conclusion, a novel PC based nanofluidic sensor has been designed, and proposed a systematic and detail way of how to design a PC for a given PWV and related materials. Also, a method to optimize structure has been created, and the way of how to select the high performance. Furthermore, the higher-order mode in PC as sensor is the first time been proposed, and its high performance was verified by simulation results, and a RIU of 124.5 nm/RIU has been acquired. Higher-order mode as a new method for PC sensing, can be largely used in biosensing detection and nanofluidic system in the future.
Acknowledgments This work was supported by the China Scholarship Council (No. 201306160029). WP also acknowledges support from Professor Brian T. Cunningham and his Nano Sensors Group.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.optcom.2016.07.019.
References [1] Eli Yablonovitch, Inhibited spontaneous emission in solid-state physics and electronics, Phys. Rev. Lett. 58 (20) (1987) 2059. [2] Nikhil Ganesh, Brian T. Cunningham, Photonic-crystal near-ultraviolet reflectance filters fabricated by nanoreplica molding, Appl. Phys. Lett. 88 (7) (2006) 071110.
111
[3] Leo L. Chan, et al., A label-free photonic crystal biosensor imaging method for detection of cancer cell cytotoxicity and proliferation, Apoptosis 12 (6) (2007) 1061–1068. [4] Wei Zhang, et al., High sensitivity photonic crystal biosensor incorporating nanorod structures for enhanced surface area, Sens. Actuators B: Chem. 131 (1) (2008) 279–284. [5] Lijun Huang, Huiping Tian, L. Huang, H. Tian, J. Zhou, et al., Label-free optical sensor by designing a high-Q photonic crystal ring–slot structure, Opt. Commun. 335 (2015) 73–77. [6] D. Yang, H. Tian, Y. Ji, The study of electro-optical sensor based on slotted photonic crystal waveguide, Opt. Commun. 284 (20) (2011) 4986–4990. [7] H. Liu, L. Leng, H. Ma, et al., Temperature-independent strain sensing characteristics of coupled photonic crystal waveguides, Opt. Commun. 367 (2016) 17–22. [8] Brian T. Cunningham, Photonic crystal surfaces as a general purpose platform for label-free and fluorescent assays, J. Assoc. Lab. Autom. 15 (2) (2010) 120–135. [9] X.C. Yang, Y. Lu, M.T. Wang, et al., A photonic crystal fiber glucose sensor filled with silver nanowires, Opt. Commun. 359 (2016) 279–284. [10] Dustin Gallegos, et al., Label-free biodetection using a smartphone, Lab. Chip 13 (11) (2013) 2124–2132. [11] A.S. Kuchyanov, P.A. Chubakov, A.I. Plekhanov, Highly sensitive and fast response gas sensor based on a light reflection at the glass-photonic crystal interface, Opt. Commun. 351 (2015) 109–114. [12] Meng Zhang, et al., Plasmonic external cavity laser refractometric sensor, Opt. Express 22 (17) (2014) 20347–20357. [13] Hsin-Yu Wu, et al., Magnification of photonic crystal fluorescence enhancement via TM resonance excitation and TE resonance extraction on a dielectric nanorod surface, Nanotechnology 21 (12) (2010) 125203. [14] Weili Chen, et al., Photonic crystal enhanced microscopy for imaging of live cell adhesion, Analyst 138 (20) (2013) 5886–5894. [15] Tan Yafang, et al., Photonic crystal enhancement of a homogeneous fluorescent assay using submicron fluid channels fabricated by E-jet patterning, J. Biophoton. 7 (3–4) (2014) 266–275. [16] E.P. Furlani, R. Biswas, A.N. Cartwright, et al., Antiresonant guiding optofluidic biosensor, Opt. Commun. 284 (16) (2011) 4094–4098. [17] D.F. Dorfner, et al., Silicon photonic crystal nanostructures for refractive index sensing, Appl. Phys. Lett. 93 (18) (2008), 181103–181103. [18] J. Zhang, X. Zhang, Biomolecular binding dynamics in sensors based on metallic photonic crystals, Opt. Commun. 320 (2014) 56–59. [19] Sherine George, et al., Sensitive detection of protein and miRNA cancer biomarkers using silicon-based photonic crystals and a resonance coupling laser scanning platform, Lab. Chip 13 (20) (2013) 4053–4064. [20] Weili Chen, et al., Enhanced live cell imaging via photonic crystal enhanced fluorescence microscopy, Analyst 139 (22) (2014) 5955–5964. [21] Anil K. Kodali, et al., Narowband mid infrared reflectance filters using guided mode resonance, Anal. Chem. 82 (13) (2010) 5697–5706. [22] Nikhil Ganesh, et al., Leaky-mode assisted fluorescence extraction: application to fluorescence enhancement biosensors, Opt. Express 16 (26) (2008) 21626–21640. [23] D. Dobbs, B.T. Cunningham, Optically tunable photonic crystal reflectance filters, Appl. Opt. 45 (28) (2006) 7286–7293. [24] X. Lu, F. Chi, T. Zhou, et al., Optical properties of high-order band gaps in onedimensional photonic crystal, Opt. Commun. 285 (7) (2012) 1885–1890. [25] M.G. Scullion, A. Di Falco, T.F. Krauss, Slotted photonic crystal cavities with integrated microfluidics for biosensing applications, Biosens. Bioelectron. 27 (1) (2011) 101–105. [26] Charles J. Choi, T. Brian Cunningham, A 96-well microplate incorporating a replica molded microfluidic network integrated with photonic crystal biosensors for high throughput kinetic biomolecular interaction analysis, Lab. Chip 7 (5) (2007) 550–556. [27] Jing Wu, Day Daniel, Min Gu, A microfluidic refractive index sensor based on an integrated three-dimensional photonic crystal, Appl. Phys. Lett. 92 (7) (2008) 071108. [28] M. Napoli, J.C.T. Eijkel, S. Pennathur, Nanofluidic technology for biomolecule applications: a critical review, Lab. Chip 10 (8) (2010) 957–985. [29] M.L. Kovarik, S.C. Jacobson, Nanofluidics in lab-on-a-chip devices, Anal. Chem. 81 (17) (2009) 7133–7140. [30] S.J. Kim, Y.C. Wang, J.H. Lee, et al., Concentration polarization and nonlinear electrokinetic flow near a nanofluidic channel, Phys. Rev. Lett. 99 (4) (2007) 044501. [31] S.S. Wang, R. Magnusson, Theory and applications of guided-mode resonance filters, Appl. Opt. 32 (14) (1993) 2606–2613. [32] R. Shiri, A. Bananej, E. Safari, Compression of ultra-short light pulses using the graded refractive index one-dimensional photonic crystals, Opt. Commun. 375 (2016) 23–28. [33] Jui-Nung Liu, et al., Optimally designed narrowband guided-mode resonance reflectance filters for mid-infrared spectroscopy, Opt. express 19 (24) (2011) 24182–24197. [34] Ian D. Block, et al., A sensitivity model for predicting photonic crystal biosensor performance, Sens. J. IEEE 8 (3) (2008) 274–280. [35] Wang, Yun, et al., Universal grating coupler design, Photonics North 2013, International Society for Optics and Photonics, 2013. [36] A.F. Oskooi, D. Roundy, M. Ibanescu, et al., MEEP: A flexible free-software package for electromagnetic simulations by the FDTD method[J], Comp. Phys. Commun. 181 (3) (2010) 687–702. [37] Wenxing Liu, et al., Guided-mode resonance filters with shallow grating, “Opt. Lett. 35 (6) (2010) 865–867. [38] Jui-Nung Liu, et al., Sculpting narrowband Fano resonances inherent in the large-area mid-infrared photonic crystal microresonators for spectroscopic imaging, Opt. Express 22 (15) (2014) 18142–18158.
112
W. Peng et al. / Optics Communications 382 (2017) 105–112
[39] Justin E. Hujdic, et al., Lead selenide nanowires prepared by lithographically patterned nanowire electrodeposition, J. Phys. Chem. Lett. 1 (7) (2010) 1055–1059. [40] 〈http://refractiveindex.info〉. [41] S.C. Deshmukh, E.S. Aydil, Investigation of SiO2 plasma enhanced chemical vapor deposition through tetraethoxysilane using attenuated total reflection Fourier transform infrared spectroscopy, J. Vac. Sci. Technol. A 13 (5) (1995) 2355–2367. [42] C. Vieu, F. Carcenac, A. Pepin, et al., Electron beam lithography: resolution limits and applications, Appl. Surf. Sci. 164 (1) (2000) 111–117. [43] J. Ding, N. Hershkowitz, Symmetric rate model for fluorocarbon plasma etching of SiO2, Appl. Phys. Lett. 68 (12) (1996) 1619–1621.
[44] C. Iliescu, F.E.H. Tay, J. Wei, Low stress PECVD—SiNx layers at high deposition rates using high power and high frequency for MEMS applications, J. Micromech. Microeng. 16 (4) (2006) 869. [45] G.W. Hsieh, C.H. Tsai, W.C. Lin, Anodic bonding of glass and silicon wafers with an intermediate silicon nitride film and its application to batch fabrication of SPM tip arrays, Microelectron. J. 36 (7) (2005) 678–682.
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Higher-order mode photonic crystal based nanofluidic sensor Wang Peng a,b,c, Youping Chen a,n, Wu Ai a a b c
School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA Micro and Nanotechnology Laboratory, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 9 June 2016 Received in revised form 7 July 2016 Accepted 8 July 2016 Available online 6 August 2016
A higher-order photonic crystal (PC) based nanofluidic sensor, which worked at 532 nm, was designed and demonstrated. A systematical and detailed method for sculpturing a PC sensor for a given peak wavelength value (PWV) and specified materials was illuminated. It was the first time that the higher order mode was used to design PC based nanofluidic sensor, and the refractive index (RI) sensitivity of this sensor had been verified with FDTD simulation software from Lumerical. The enhanced electrical field of higher order mode structure was mostly confined in the channel area, where the enhance field is wholly interacting with the analytes in the channels. The comparison of RI sensitivity between fundamental mode and higher order mode shows the RI variation of higher order mode is 124.5 nm/RIU which is much larger than the fundamental mode. The proposed PC based nanofluidic structure pioneering a novel style for future optofluidic design. & 2016 Elsevier B.V. All rights reserved.
Keywords: Photonic crystals Optofluidic biosensors Label-free detection Sensitivity
1. Introduction PC (Photonic crystal) is a periodic grating structure, with high refractive index guided layer surrounded by lower refractive index layers. Since the initial work of Yablonovitch [1], PCs have attracted increasing interest due to their intrinsic advantages, such as low cost [2], large electrical field enhancement [3], and high sensitivity [4–7]. PCs are usually designed to couple strongly with special optical wavelength according to the materials and the modulated grating period. When matched with the wavelength of resonance, light will be confined strongly to the guided layer and results in an electric field related to the PC many times larger than the field of incident light [8,9]. Since the changing of the ambient medium parameters can induce resonance wavelength shift due to the effective RI variation of the PC, it can vary the distribution of the related evanescent filed. As a result, PC has been used in many different fields including label-free biosensing [10,11], lasing [12], PCEF (photonic crystal enhanced fluorescence) [13], PCEM (photonic crystal enhanced microscopy) [14], and PC micro-fluidics [15,16]. Recently, PC have been utilized to sense ambient refractive index [17,18], protein detection [19], and antibody in the micro fluidic [15] and cell imaging [20]. However, nearly all the reported sensing schemes adopt the detecting area on the surface of guided layer, and the enhanced electromagnetic field is located inside or on the corner of the guided layer [8–20]. In the past, researchers n
Corresponding author. E-mail address: [email protected] (Y. Chen).
http://dx.doi.org/10.1016/j.optcom.2016.07.019 0030-4018/& 2016 Elsevier B.V. All rights reserved.
were emphasized on confining the guided layer thin enough to suppress the arising of higher-order mode [21–23]. Even high value of enhanced electrical field value can be acquired, the distribution of the enhanced field is not good enough. Also, PCs have certain up limits in the field of sensing since a large fraction of the optical power is coupled into the guided layer instead of propagates in the evanescent field outside the physical boundary of the guided layer. By this way, the enhanced electrical field is separated from the analytes on the surface and cannot make a wholly contribution to the sensor. As the enhancement factor of evanescent field is exponentially decay on the above of PC surface [20,24], the sensing area of traditional PCs is very limited. Meanwhile, microfluidic PC structures have been used for Reflective index sensing related area, like the detection of dissolved avidin concentrations on slotted photonic crystal cavities [25], measurement of the kinetic binding interaction of protein A with IgG molecules on polymer microfluidic channels [26], and microfluidic refractive index sensors [27] in the past few years. As far as we know, all these PC based microfluidics in the above references [24–27] are worked on the fundamental mode of resonance, in which the enhanced evanescent field will be mostly distributed on the edge of the grating angle or inside the guided layer, and cannot interact significantly with the detection sample. Nowadays, nanofluidic sensors, which can be used for biochemical, medical and life science detection, have been investigated worldwide [28]. Most of these nanofluidic sensors are fabricated as a single fluidic channel, and the analytes inside the channel need to be detected with aided tags, such as fluorescence [29], current [30]. To design a nanofluidic sensor that can test analytes with its own characteristics is of key importance.
106
W. Peng et al. / Optics Communications 382 (2017) 105–112
In this paper, a novel PC based nanofluidic sensor has been presented, which can work on the higher-order mode and has a detection area set inside the PC. To the best of our knowledge, there has been no publish to date of PC based nanofluidic sensor which is working on higher-order mode with liquid channel inside the PC structure. A well designed structure has been proposed, which is working on the 532 nm wavelength on resonance. The resonance peak induced by the higher-order mode is used to sense ambient RI (refractive index), instead of the fundamental mode, and the electric field distributes on the center of detection area while working in higher-order mode. The proposed PC based nanofluidic sensor possesses high sensitivity, compact size, and multi channels which can be used to detect different samples at the same time. This paper gives detailed method about how to design a PC sensor and optimize the related parameters to obtain the ideal structure. Based on FDTD simulation results, the superior performance of higher-order TE01 mode had been analyzed and demonstrated by simulation results. A brief way of how to fabricate this sensor was given. Also, the upcoming issues when passing from the theoretical part to the experiments were discussed. In the future, we will focus on fabricating and testing its performance with experiments.
2. Physical principle and device scheme The physical phenomenon behind the narrowband reflection from PC has been described previously [31,32] as the periodic index differences enabling the exciting of guided mode resonance. The resonance wavelength occurs when the incident light satisfies the phase matching condition [33]:
⎛ 2π ⎞ ⎟ = β k 0nc sin θ ± m⎜ ⎝ Λ⎠
(1)
where k0 = 2π /λ is the wave vector in free space, θ is the incident angle, nc is the refractive index of the cladding area (in this structure, nc is decided by the refractive index of Pyrex and liquid in nanofluidic channel), m = 0, ± 1... is the diffraction order, and 2π β = λ neff is the propagation wave vector in the PC structure, and neff is the effective index. The PC based nanofluidic sensor in this research has a Si3N4 based periodic structure built on the Pyrex substrate and covered by glass, while the nanofluidic channel is made by the gap between gratings. The grating period (Λ) and grating depth (d) of the sensor structure are depicted in Fig. 1. Since the gap between the Si3N4 based periodic structures will be used as nanofluidic channels, we plan to set it in a reasonable way, which means the size of the channel should be possible to fabricate with normal cleanroom equipment. For this PC based nanofluidic sensor, the detection of light is under normal incidence (θ ¼0) on the PC plane, and the targeted resonance wavelength will be 532 nm. Thus, θ ¼0, and we can rewrite the formula (1) as follows:
mλ = Λneff
(2)
For a certain period, when there is a PWV according to the effective index changes resulting from ambient RI perturbation, we can get:
mΔλ = ΛΔneff
(3)
From (2) and (3), we can deduce that the sensitivity of PC based nanofluidic sensor is decided by the perturbation of effective index, which is mainly due to the interaction between the electrical field and the nanofluidic channels, as we can calculate from Eq. (4). The resonance wave couples into the guided layer due to the perturbation of the grating structure. The ambient refractive index
Fig. 1. PC based nanofluidic design comprised on a Pyrex substrate, a Si3N4 grating layer, and a SiO2 superstrate. (a) PC based nanofluidic sensor schematic diagram; (b) Parameters and refractive indexes (RI) of PC based nanofluidic sensor; d: grating depth, ng: RI of Si3N4, ntest: RI of test analyte (e.g. water), ns: RI of substrate Pyrex, nc: RI of cladding layer, SiO2.
vary leads to redistribution of electric field and concomitant spectral changes of the reflected far fields which can be detected by a spectrometer. The effective index can be regard as the average of the refractive indices of the materials at resonance mode, and can be written as [34]: ∞
2 neff
=
Λ
∫−∞ ∫0 ε( x, y) E(x, y) 2 dxdy ∞
Λ
∫−∞ ∫0 E(x, y) 2 dxdy
(4)
where ε(x, y) is the dielectric permittivity, and E(x, y) is the electrical field intensity. In order to improve the sensitivity of the sensor, the enhanced electrical field need to be confined mostly in the channel, where the interaction is exist.
3. Design method and results analysis In this section, an analysis method for optimizing the electrical field enhancement within the channels in the sensor and predicting the absolute resonance wavelength at 532 nm have been proposed and discussed. 3.1. Design method For a certain period, there is only fundamental mode when the grating depth is very thin. Whereas the higher-order mode occurs as the grating layer is thick enough. For any specific PC, there are two types of input parameters: structure-determined parameters and user-intent parameters [35]. The structure-determined parameter include grating depth d, refractive index of the cladding layer nc, grating layer ng, substrate ns, and test sample area ntest, grating period Λ and fillfactor f which are determined by the structure of PC. The user-intent parameters include resonance wavelength value λ, incident angle θ, and polarization (TE, TM), which can be tuned by designer after the PC structure have been fabricated. In this paper, the structure-determined parameters were used to design the PC for PWV at 532 nm, and the user-intent
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 2. Design flow of the PC based nanofluidic structure.
parameters were used to optimize the electromagnetic field and sensitivity of this sensor. The PC designs demonstrated in this paper are based on a 500 um thick Pyrex substrate and a guiding layer about 200– 300 nm. The two-dimensional finite-difference time-domain (FDTD) method was used to simulate the resonance wavelength and its related electrical field distribution. FDTD is the finite-difference time-domain algorithm, which divides space and time into a regular grid and simulates the time evolution of Maxwell equations [36]. In this section, different structures resonance at a certain wavelength value, 532 nm, had been simulated, since 532 nm is a general used light source which can be achieved easily when set biodetection experiments in the future. The peak wavelength value, peak reflectance, and bandwidth of the PC spectrum were all decided by the grating depth d, period Λ, fillfactor f and polarization direction. Although tuning the fillfactor of the grating can modulate the peak wavelength value [37], grating depth is the most commonly adapt parameter, which can be finely tuned using deposition, lithography and etching process [38]. In order to simplify the analysis process, the fillfactor was set as 0.5 which can be obtain easily in experiments [39] and incident angle as 0° due to most of the PC based biosensor structure. Instead of continuously changing grating period, several discrete values based on the equality for resonance regime of GMR filter had been chosen [38]:
max {nc , ns , ntest } ≤
mλ ≤ ng Λ
(5)
where m is the order of diffracted wave, nc, ng, ns and ntest are the refractive indices of the cladding layer, grating layer, substrate, and test sample area respectively. The refractive indices were set as nc ¼1.5, ng ¼2.02, ns ¼1.5 and ntest ¼ 1.33 at 532 nm [40]. Thus, from Eq. (5), the range of the grating period will be 264 nm o Λ o 355 nm, we choose discrete periods Λ ¼310, 320, 330, 340, 350, 360 nm. The design flow is given in Fig. 2. The design flow is as follows: firstly the PC structure was proposed as in Fig. 1, with Si3N4 as the high refractive index guiding layer material, SiO2 as the superstrate, and Pyrex as substrate. Then the resonance wavelength for this sensor was defined
107
as 532 nm, with normal incidence, and fillfactor as 0.5 which is easy to fabricate. Also, a set of discrete grating periods, Λ1, Λ2… based on the Eq. (5) had been set. Then, simulation with FDTD was used to get the expected result. In the FDTD solutions, when a plane wave is interacting with the PC, the simulation area is confined in a grating lattice. As with many periodic grating structures, a mesh override region is required at the grating. The boundary condition of top and bottom side is PML (perfect matched layer), which will absorb light waves with minimum reflection. While the boundary condition of left and right side is Bloch boundary conditions. The Bloch boundary condition replicate the fields at one edge of the simulation region and re-inject at the other edge with a phase correction of the fields. Before the simulation start, the polarization of the light source was chosen, TE or TM. The TE mode is defined as the electrical field which is perpendicular to the incident plane and parallel to the direction of gratings. The TM mode is defined as the magnetic field which is perpendicular to the incident plane and parallel to the direction of gratings. Then, a specific period was used, and gradually tuned the depth of the grating, to build contours of PWV, and reflection of PWV. Finally, the 532 nm resonance with a specific grating was there for a certain period and polarized direction. When increase the grating depth step by step, there were 3 different which resonant on 532 nm, which are single mode resonance, fundamental mode resonance in multi-mode and higher-order mode resonance in multi-mode. The single mode resonance of PC is indicated as there is only one peak exist among the whole reflection spectrum. Fundamental mode resonance in multi-mode is the zero order diffraction among the phase matching condition in Eq. (1), which is usually the rightmost peak in the reflection spectrum. Higherorder mode resonance in multi-mode is the first order diffraction which is the peak wavelength next to the fundamental mode peak. All these definitions are effective when the light is under normal incidence. Also, the electrical field for each result was plotted. Finally, the optimized design was selected according to three factors: (1) electrical field enhancement. (2) The ratio of enhancement in the detection area. (3) The sensitivity of the selected mode. The way of how to using PWV contour to determine the grating depth for 532 nm resonance had been explained clearly in a former paper from our group [18], so it will not show it any more. Instead, the main research was focused on how to confine the enhanced electrical field inside the detection area and get a reasonable channel depth, about 150–300 nm. The TE and TM mode resonance had been analyzed independently. 3.2. Results analysis Based on the design method of PC based nanofluidic structure, a certain grating depth and related parameters for PWV (peak wavelength value) on 532 nm were obtained with FDTD Lumerical. As shown in supplementary materials, channel area was represented by dash line and red arrow showed the location of PWV at 532 nm. In each TE or TM mode polarized direction, three different ways of PWV centered at 532 nm had been analyzed, which is single mode 532 nm resonance, fundamental mode 532 nm resonance in multimode, and higher-order mode 532 nm resonance in multimode. Basically, these three modes occurred sequentially as the grating depth increasing for a certain grating period. However, the modes will be cut off for a specified grating, since the PWV will be away from 532 nm when the grating period is too small, according to Eq. (2). All the possibilities of PWV at 532 nm were shown in Fig. 3 and its related electromagnetic filed in TM mode were shown in supplementary material. TM0 curve in Fig. 3 was used to analysis the potential single mode PC structure which will resonance on 532 nm, which indicated that the single peak resonance was not
108
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 3. Grating depth vs grating period for TM mode PWV at 532 nm.TM0: single TM mode PC resonance at 532 nm; TM00: fundamental TM mode PC resonance at 532 nm in multi-mode structure; TM01: higher-order mode PC resonance at 532 nm in multi-mode structure.
Fig. 4. Grating depth vs grating period for TE mode PWV at 532 nm. TE0: single TE mode PC resonance at 532 nm; TE00: fundamental TE mode PC resonance at 532 nm in multi-mode structure; TE01: higher-order mode PC resonance at 532 nm in multi-mode structure.
arising at 532 nm and only exist on Λ ¼340 nm for the given discrete periods from 310 nm to 360 nm. The 532 single peak resonance was cutoff when the period reaches to 350 nm or 330 nm. In supplementary material Fig. 1(a) shows the reflection spectrum and electromagnetic field while PWV at 532 nm. Although the electrical field was enhanced largely, the normalized reflection
intensity was only 0.3 at 532 nm. The fundamental mode (TM00) of multi-mode PC structures resonance on 532 nm indicate that the fundamental mode resonance on 532 nm was cutoff when the period reaches 340 nm, and there were resonances while the period was from 310 nm to 330 nm. In supplementary material Fig. 1(b), the electrical field mostly been largely enhanced. However, the enhanced area is in the corner of gratings, and mostly are inside the Si3N4, with little enhanced E field can interact with the detection area. According to TM01 curve, when the period is around 330 – 350 nm, the TM01 mode in multimode was generated. The related normalized reflection spectrums and electromagnetic fields were shown on supplementary material Fig. 1 (c) when the period was located on 330 nm, 340 nm and 350 nm respectively. The E field enhancement area was located around the channel area, yet its E field enhancement was not high, about 5 times of the normalized incident E field averagely (Emax/E0). In conclusion, for all the simulated TM mode, the enhanced E fields are usually gathered around the corners of grating structure, and cannot interact well with the detection area. So, we may speculate that the TM mode will not be the best style for PC based nanofluidic sensor in this structure. Fig. 4 gave all the possibilities of PWV at 532 nm in TE mode. TE0 curve was used for the analysis of single mode PC structure. For the given discrete periods, the single mode structures had peak at 532 nm resonance when the period was tuning from 330 nm to 350 nm, as shown in Supplementary material Fig. 2(a). From the comparison of normalized reflection spectrums and E-field plots, it indicated that the band width of resonance peak became much narrow and normalized reflection intensity decreased gradually as the grating period increased from 330 nm to 350 nm, but its E field value enhanced gradually as the period increased. For the E-field enhancement of the dashed line represented nanochannel area, half of the enhanced area was distributed in the detection area. The second set of pictures, in supplementary material Fig. 2(b) was focused on the study of fundamental mode (TE00) resonance on 532 nm in multimode sensor structures. The peak occurred when the period is between 320 nm and 340 nm, as shown from TE00 curve in Fig. 4. The bandwidth of the resonance peak at 532 nm was gradually increased, as shown in supplementary material Fig. 2(b). The E field enhancement value was around 10, and the enhanced E field area mostly upon the detection area. The last set of figures, as shown in f in supplementary material Fig. 2(c), was the higher-order mode resonance on 532 nm which is TE01 mode sensor structures. As the grating depth increased to a certain extent (less than 1 um), the TE01 mode with periods vary from 320 to 360 nm is always exist, as shown in Fig. 4, TE01 curve. The TE01 mode will be cutoff at
Fig. 5. Sensor structure: Λ 350 nm, d 297 nm, TE01 mode. Left: spectrum; right: electrical field intensity. (a) TE higher-order mode structure, Λ ¼350 nm, d ¼297 nm. (b) TE fundamental mode, Λ ¼ 320 nm, d ¼323 nm. (c) TE single mode, Λ ¼ 330 nm, d ¼80 nm.
W. Peng et al. / Optics Communications 382 (2017) 105–112
109
Table 1 Summary of Designed PC nanofluidic sensor with at PWV 532 nm for various modes.
Fig. 7. Comparison the reflection wavelengths shift on the ambient RI between higher-order mode, fundamental mode and single mode resonance PWV at 532 nm. The higher order mode at TE1 is much sensitive, which means using higher order mode can improve the performance of the PC based nanofluidic sensor.
Fig. 6. Normalized reflection spectrum shift due to the change of RI in detection area. Black arrow indicates the red shift of PWV initially from 532 nm, when tuning the ambient RI.
310 nm, where the grating depth is too large in compare with grating. When the grating depth is much larger than the period, the wavelength shift is slow which decrease sensitivity of the sensor. However, in a certain range of grating depth, the sensitivity of this higher-order mode is significant. From the E field plots in supplementary material Fig. 2(c), it can be indicated that the TE01 mode was confined in the detection area and its E field enhancement was about 15, especially in the 350 nm period with grating depth 297 nm, as shown in Fig. 5. In conclusion, for all the simulated TE mode, the enhanced E fields are usually gathered around the center of gratings or detection area, which can interact well with the detection area. For the TE01 mode, the enhanced E filed is
110
W. Peng et al. / Optics Communications 382 (2017) 105–112
Fig. 8. Schematic of fabrication process.
4. Optimized structure for refractive index (RI) sensing and fabrication process design 4.1. Optimized structure for refractive index (RI) sensing
Fig. 9. Schematic of PC based nanofluidic sensor.
centered in the detection area, and the E field is also significant, which may achieve a better sensitivity for the sensor. We tried to target the channel size in a reasonable range, which means the grating depth should not be too low in order to let the sample flow smoothly in the nanofluidic channel. With the related spectrum and E field information from Fig. 3, Fig. 4 and Supplementary material, the parameters were reorganized, and the whole parameters satisfy the requirements are shown in Table 1. In Table 1, all the modes and parameters related to this sensor which can arise 532 nm resonance have been given and sort by category. In order to decide which method is the best for this sensor, the last three parameters: peak reflection value, E field enhancement, and E field distribution in the detection area had been analyzed. With this three factors, the optimized structure is the TE higher-order mode, with period 350 nm and grating depth 297 nm. Thus, we can deduce that the TE mode, higher-order resonance at 532 nm with period 350 nm and grating depth 297 nm has the best performance.
After the optimized process, the optimized PC based nanofluidic sensor have been designed, which is the higher-order mode in multimode, TE polarized, with a grating period Λ ¼350 nm, and grating depth d¼ 297 nm. Then, its reflection spectrum with different refractive index among the detection area from 1.33 to 1.40 was analyzed. As shown in Fig. 6(a), it can be easily discovered that two main reflected peaks occur in the reflection spectrum. The PWV at long wavelength part is due to the coupling between the propagating fundamental TE0 mode and grating structure, and the other PWV which resonance at 532 nm is due to the coupling between higherorder TE01 mode and grating structure. When the refractive index in the detection area change from 1.33 to 1.40 gradually, the red shift of TE00 and TE01 mode was simultaneously. When noticing the red shift distance, it can be deduced that the higher-order mode's red shift is much larger than the fundamental mode, since the enhanced electrical field of the TE01 mode is mainly concentrated on the channels area as can be seen from Fig. 6(a) and Fig. 5. Then, the other two different styles has a PWV at 532 nm, which is fundamental mode resonance in multimode as Fig. 6(b) and single mode resonance as Fig. 6(c), were also analyzed. As it can be seen from these two figures, the shift at 532 nm is slow when the RI in the detection area varies. From the related E field plot, the conclusion can be given that RI sensitivity is related to the E field enhancement, especially when the enhance area is centered in the detection area. The RI sensing performance is demonstrated with numerical simulations by compare three different structures with PWV at 532 nm, as shown in Fig. 5. From Fig. 7, it shows that the higherorder mode has a significant sensitivity than the other two modes. Also, it indicates that the maximum RIU of TE1 mode is about 124.5 nm/RIU (RI Unit), while the RIU of TE0 mode and single peak mode are 60 nm/RIU and 45.9 nm/RIU, respectively. 4.2. Fabrication process design In this section, a planned fabrication process was given. In order to fabricate such a PC based nanofluidic sensor with a grating period Λ ¼ 350 nm, and grating depth d ¼297 nm, a schematic fabrication process was shown in Fig. 8. The whole process includes six parts: (1) SiO2 deposition, (2) Photo resist (PR) coating, (3) E-beam lithography, (4) SiO2 etching, (5) Si3N4 etching,
W. Peng et al. / Optics Communications 382 (2017) 105–112
(6) Anodic bonding between Pyrex glass and PC. A common polished Si or SiO2 glass can be selected as substrate (Si wafer was selected as substrate in this design). Then, 2 um can be deposited on the polished side of Si wafer with PECVD (Plasma Enhanced Chemical Vapor Deposition) [41]. Since the feature size of the grating is only 350 nm, e-beam lithography was selected as the feature sculpturing method [42]. For the e-beam lithography process, a thin layer of related PR will be coating on the SiO2 surface initially, and then the e-beam will write the required 350 nm feature on the PR. After e-beam lithography, the PR feature will be used as mask to etch 297 nm SiO2 with RIE (reactive ion etching) tool [43]. Then, the PR will be cleaned by acetone, IPA, water and IPA in sequence. A thickness of 297 nm Si3N4 will be deposited on the etched SiO2 grating surface with PECVD tool [44]. With this fabricated PC structure, anodic bonding process can be used to seal the PC gratings with Pyrex glass [45]. In this way, the PC based nanofluidic channel can be realized. During the fabrication process, the parameters may be not as accurate as designed. The reason can be classified as follows: (1) the performance of related fabrication equipment, (2) the metrology device used to measure the related size, (3) the ambient influence, such as particle level and temperature. Therefore, there may be limitations when passing from theoretical approach to the experiments. The schematic of the PC based nanofluidic sensor was shown in Fig. 9. The PC based nanofluidic sensor is located in the center of the chip and connected with inlets and outlets. A white light source is initially collimated by a convex lens and then TE polarized by a polarizer. After this, the TE polarized collimation light to reach the sensor surface at normal incident angle. A resonance light will be reflected back and collected by a spectrometer. The peak wavelength of the reflection spectrum will red shift as the concentration of analytes in the nanochannels gradually increase.
5. Conclusion In conclusion, a novel PC based nanofluidic sensor has been designed, and proposed a systematic and detail way of how to design a PC for a given PWV and related materials. Also, a method to optimize structure has been created, and the way of how to select the high performance. Furthermore, the higher-order mode in PC as sensor is the first time been proposed, and its high performance was verified by simulation results, and a RIU of 124.5 nm/RIU has been acquired. Higher-order mode as a new method for PC sensing, can be largely used in biosensing detection and nanofluidic system in the future.
Acknowledgments This work was supported by the China Scholarship Council (No. 201306160029). WP also acknowledges support from Professor Brian T. Cunningham and his Nano Sensors Group.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.optcom.2016.07.019.
References [1] Eli Yablonovitch, Inhibited spontaneous emission in solid-state physics and electronics, Phys. Rev. Lett. 58 (20) (1987) 2059. [2] Nikhil Ganesh, Brian T. Cunningham, Photonic-crystal near-ultraviolet reflectance filters fabricated by nanoreplica molding, Appl. Phys. Lett. 88 (7) (2006) 071110.
111
[3] Leo L. Chan, et al., A label-free photonic crystal biosensor imaging method for detection of cancer cell cytotoxicity and proliferation, Apoptosis 12 (6) (2007) 1061–1068. [4] Wei Zhang, et al., High sensitivity photonic crystal biosensor incorporating nanorod structures for enhanced surface area, Sens. Actuators B: Chem. 131 (1) (2008) 279–284. [5] Lijun Huang, Huiping Tian, L. Huang, H. Tian, J. Zhou, et al., Label-free optical sensor by designing a high-Q photonic crystal ring–slot structure, Opt. Commun. 335 (2015) 73–77. [6] D. Yang, H. Tian, Y. Ji, The study of electro-optical sensor based on slotted photonic crystal waveguide, Opt. Commun. 284 (20) (2011) 4986–4990. [7] H. Liu, L. Leng, H. Ma, et al., Temperature-independent strain sensing characteristics of coupled photonic crystal waveguides, Opt. Commun. 367 (2016) 17–22. [8] Brian T. Cunningham, Photonic crystal surfaces as a general purpose platform for label-free and fluorescent assays, J. Assoc. Lab. Autom. 15 (2) (2010) 120–135. [9] X.C. Yang, Y. Lu, M.T. Wang, et al., A photonic crystal fiber glucose sensor filled with silver nanowires, Opt. Commun. 359 (2016) 279–284. [10] Dustin Gallegos, et al., Label-free biodetection using a smartphone, Lab. Chip 13 (11) (2013) 2124–2132. [11] A.S. Kuchyanov, P.A. Chubakov, A.I. Plekhanov, Highly sensitive and fast response gas sensor based on a light reflection at the glass-photonic crystal interface, Opt. Commun. 351 (2015) 109–114. [12] Meng Zhang, et al., Plasmonic external cavity laser refractometric sensor, Opt. Express 22 (17) (2014) 20347–20357. [13] Hsin-Yu Wu, et al., Magnification of photonic crystal fluorescence enhancement via TM resonance excitation and TE resonance extraction on a dielectric nanorod surface, Nanotechnology 21 (12) (2010) 125203. [14] Weili Chen, et al., Photonic crystal enhanced microscopy for imaging of live cell adhesion, Analyst 138 (20) (2013) 5886–5894. [15] Tan Yafang, et al., Photonic crystal enhancement of a homogeneous fluorescent assay using submicron fluid channels fabricated by E-jet patterning, J. Biophoton. 7 (3–4) (2014) 266–275. [16] E.P. Furlani, R. Biswas, A.N. Cartwright, et al., Antiresonant guiding optofluidic biosensor, Opt. Commun. 284 (16) (2011) 4094–4098. [17] D.F. Dorfner, et al., Silicon photonic crystal nanostructures for refractive index sensing, Appl. Phys. Lett. 93 (18) (2008), 181103–181103. [18] J. Zhang, X. Zhang, Biomolecular binding dynamics in sensors based on metallic photonic crystals, Opt. Commun. 320 (2014) 56–59. [19] Sherine George, et al., Sensitive detection of protein and miRNA cancer biomarkers using silicon-based photonic crystals and a resonance coupling laser scanning platform, Lab. Chip 13 (20) (2013) 4053–4064. [20] Weili Chen, et al., Enhanced live cell imaging via photonic crystal enhanced fluorescence microscopy, Analyst 139 (22) (2014) 5955–5964. [21] Anil K. Kodali, et al., Narowband mid infrared reflectance filters using guided mode resonance, Anal. Chem. 82 (13) (2010) 5697–5706. [22] Nikhil Ganesh, et al., Leaky-mode assisted fluorescence extraction: application to fluorescence enhancement biosensors, Opt. Express 16 (26) (2008) 21626–21640. [23] D. Dobbs, B.T. Cunningham, Optically tunable photonic crystal reflectance filters, Appl. Opt. 45 (28) (2006) 7286–7293. [24] X. Lu, F. Chi, T. Zhou, et al., Optical properties of high-order band gaps in onedimensional photonic crystal, Opt. Commun. 285 (7) (2012) 1885–1890. [25] M.G. Scullion, A. Di Falco, T.F. Krauss, Slotted photonic crystal cavities with integrated microfluidics for biosensing applications, Biosens. Bioelectron. 27 (1) (2011) 101–105. [26] Charles J. Choi, T. Brian Cunningham, A 96-well microplate incorporating a replica molded microfluidic network integrated with photonic crystal biosensors for high throughput kinetic biomolecular interaction analysis, Lab. Chip 7 (5) (2007) 550–556. [27] Jing Wu, Day Daniel, Min Gu, A microfluidic refractive index sensor based on an integrated three-dimensional photonic crystal, Appl. Phys. Lett. 92 (7) (2008) 071108. [28] M. Napoli, J.C.T. Eijkel, S. Pennathur, Nanofluidic technology for biomolecule applications: a critical review, Lab. Chip 10 (8) (2010) 957–985. [29] M.L. Kovarik, S.C. Jacobson, Nanofluidics in lab-on-a-chip devices, Anal. Chem. 81 (17) (2009) 7133–7140. [30] S.J. Kim, Y.C. Wang, J.H. Lee, et al., Concentration polarization and nonlinear electrokinetic flow near a nanofluidic channel, Phys. Rev. Lett. 99 (4) (2007) 044501. [31] S.S. Wang, R. Magnusson, Theory and applications of guided-mode resonance filters, Appl. Opt. 32 (14) (1993) 2606–2613. [32] R. Shiri, A. Bananej, E. Safari, Compression of ultra-short light pulses using the graded refractive index one-dimensional photonic crystals, Opt. Commun. 375 (2016) 23–28. [33] Jui-Nung Liu, et al., Optimally designed narrowband guided-mode resonance reflectance filters for mid-infrared spectroscopy, Opt. express 19 (24) (2011) 24182–24197. [34] Ian D. Block, et al., A sensitivity model for predicting photonic crystal biosensor performance, Sens. J. IEEE 8 (3) (2008) 274–280. [35] Wang, Yun, et al., Universal grating coupler design, Photonics North 2013, International Society for Optics and Photonics, 2013. [36] A.F. Oskooi, D. Roundy, M. Ibanescu, et al., MEEP: A flexible free-software package for electromagnetic simulations by the FDTD method[J], Comp. Phys. Commun. 181 (3) (2010) 687–702. [37] Wenxing Liu, et al., Guided-mode resonance filters with shallow grating, “Opt. Lett. 35 (6) (2010) 865–867. [38] Jui-Nung Liu, et al., Sculpting narrowband Fano resonances inherent in the large-area mid-infrared photonic crystal microresonators for spectroscopic imaging, Opt. Express 22 (15) (2014) 18142–18158.
112
W. Peng et al. / Optics Communications 382 (2017) 105–112
[39] Justin E. Hujdic, et al., Lead selenide nanowires prepared by lithographically patterned nanowire electrodeposition, J. Phys. Chem. Lett. 1 (7) (2010) 1055–1059. [40] 〈http://refractiveindex.info〉. [41] S.C. Deshmukh, E.S. Aydil, Investigation of SiO2 plasma enhanced chemical vapor deposition through tetraethoxysilane using attenuated total reflection Fourier transform infrared spectroscopy, J. Vac. Sci. Technol. A 13 (5) (1995) 2355–2367. [42] C. Vieu, F. Carcenac, A. Pepin, et al., Electron beam lithography: resolution limits and applications, Appl. Surf. Sci. 164 (1) (2000) 111–117. [43] J. Ding, N. Hershkowitz, Symmetric rate model for fluorocarbon plasma etching of SiO2, Appl. Phys. Lett. 68 (12) (1996) 1619–1621.
[44] C. Iliescu, F.E.H. Tay, J. Wei, Low stress PECVD—SiNx layers at high deposition rates using high power and high frequency for MEMS applications, J. Micromech. Microeng. 16 (4) (2006) 869. [45] G.W. Hsieh, C.H. Tsai, W.C. Lin, Anodic bonding of glass and silicon wafers with an intermediate silicon nitride film and its application to batch fabrication of SPM tip arrays, Microelectron. J. 36 (7) (2005) 678–682.