Microphotonic sensors have been actively studied with increasing demands for label-free biosensing in medical diagnoses and life sciences. For high-throughput and low-cost sensing, a high sensitivity is crucial for eliminating the pre-concentration process, while a simple setup of sensors is also desirable. This paper demonstrates a super-sensitivity for protein, which satisfies these requirements. The key device is a photonic crystal nanolaser, in particular with a nanoslot. Even using a simple setup, the nanolaser achieves an extraordinary-low detection limit for BSA protein, i.e. 255 fM on an average, which cannot be explained by its bulk index sensitivity. The specific adsorption of the protein is observed only around the nanoslot with strong laser intensity. This suggests that the super-sensitivity arises from the effective trapping of protein in the nanoslot.
© 2011 OSA
In the field of biosensing, the detection of protein is becoming important for tumor markers, allergy testing, cytoscopy and proteomics [1,2]. Current methods, however, require complicated screening and concentration processes prior to sensing. Direct detection is more desirable for high-throughput analyses, but high selectivity and high sensitivity giving a low detection limit (DL) (pM order for tumor markers ) are required. The selectivity is mainly attributed to antigen-antibody reactions and surface chemistry, while the sensitivity is attributed to the performance of sensors as transducers. The sensor chip and the overall system must be simple for cost reduction and disposable use, and the chip should be small enough to integrate with advanced inspection kits such as μ-TAS.
Now, we focus on the sensitivity of sensors as transducers. Fluorescent labels, which achieve a low DL of pM order [4,5], are widely used as transducers in biosensing. However, they are suspected to denature the targets and their functionalization complicates the overall process. Therefore, label-free methods using surface plasmon resonance (SPR)  and passive optical cavities [7–9] have been studied as alternatives. They detect the adsorption of bio-molecules through a resonant wavelength shift Δλs (or resonant angle). In general, however, they have issues such as large size, low sensitivity, wide resonant spectrum (linewidth Δλw), and/or complicated setup. Being proportional to Δλw/Δλs, DL has been limited to nM regime in most cases; a lower regime is achieved only by using an ultrahigh Q cavity with complicated equipment , which may not be low-cost or disposable.
Recently, nanolasers have also been studied for sensing [10–16]. They are photopumped and the emission is detected through the same free space optics, hence they operate remotely without the need for any electrical contacts. Cost reduction, disposable use and integration are possible when the nanolaser as a sensor chip is isolated from other equipment. A small size and low DL can be achieved simultaneously, because Δλw under lasing is reduced independently of the Q factor when the spectral broadening due to thermal chirping  is suppressed. As a device specifically suitable for this purpose, we have reported the nanoslot (NS) photonic crystal nanolaser , as shown in Fig. 1 . In high-index-contrast waveguides and cavities, the modal electric field is confined strongly in the NS , resulting in enhanced light-matter interaction and high sensitivity in liquids [12,18]. Furthermore in water, our NS nanolaser has low thermal chirping and is particularly effective for a narrow Δλw of 18 pm with a high sensitivity of 410 nm/RIU, giving a liquid index resolution of 4.5 × 10−5 .
In this paper, we discuss protein sensing using this NS nanolaser and show that the advantage of the NS is not limited to the narrow spectrum. In this experiment, it exhibits a super-sensitivity giving a remarkably low DL. It is confirmed to originate from the accelerated adsorption of the protein at the NS, which may be occurring due to optical trapping. In the following, we briefly describe the device structure, fabrication, and the procedure of sensing a standard protein—bovine serum albumin (BSA). We then show the details of the sensing characteristics, particularly focusing on the unique behaviors in the regime of ultra-low BSA concentration. We present simple theoretical analyses and discuss the advantage of this device compared to others.
2. Device structure and protein adsorption process
As illustrated in Fig. 1, two airholes in a triangular-lattice air-bridge photonic crystal slab are shifted outwards (H0 structure [11,13,19]), and a NS with width w NS is perforated. The laser mode is localized inside and around the NS. The laser wavelength shifts with the surrounding index n. The detailed device structure and the sensitivity against liquids, Δλs/Δn, have been reported in . Devices with w NS = 30−70 nm are fabricated into commercial GaInAsP/InP quantum-well wafer by using e-beam lithography, HI inductively-coupled plasma etching and HCl wet etching. Since the footprint of a single device is only ~10 μm square, high-throughput fabrication is possible even using e-beam lithography. The device is operated by pulsed photopumping at 980 nm (duty-ratio of 200, spot diameter of 20 μm) to avoid the excess heating and the subsequent laser emission is detected using a fluorescence microscope setup. The device does not show any degradation even after operating for 8 hours in water. Since the single sensing was completed within 1 minute in this experiment, the device does not deteriorate at all after hundreds of measurements.
Figure 2(a) shows the procedure of BSA adsorption. First, we functionalized the device with a hydroxyl group with 9% HCl / deionized water at 3°C, then with an amino group with 10% N-2-(aminoethyl)-3-aminopropyltrimethoxysilane (APTES) and an aldehyde group with 2.5% glutaraldehyde (GA) at 20°C. We used BSA (molecular weight = 68 kDa)  as a target protein which nonspecifically cross-links with GA. Figure 2(b) shows scanning electron micrograph (SEM) and atomic force micrograph (AFM) around the NS after soaking in high-concentration BSA aqueous solution of 10 μM for more than 30 minutes. More than 10-nm roughness is observed on the surface and in the NS, indicating the BSA adsorption, which was not observed before the process. Since BSA is known to have an elliptical shape and a molecular size larger than ~3 nm × 8 nm , the larger roughness observed suggests that the high-concentration BSA molecules agglutinate due to hydrophobic interaction and/or cross-linking through GA.
3. Observation of super-sensitivity
Figure 3(a) shows laser spectra for devices in water before and after the BSA adsorption. Here, the BSA concentration was 10 μM so that the device surface is covered with BSA almost completely. The clear redshift of the spectrum appears for both devices with and without the NS. Without the NS, the spectra are broadened by thermal chirping (Δλw > 600 pm). With the NS, the spectra are markedly narrowed due to athermalization  (Δλw < 30 pm), and the resolution is drastically improved. Figure 3(b) summarizes the dependence of Δλs on the NS width. The data scattering does not depend on the NS width. Therefore, it is thought to be due to the nonuniform adsorption of BSA between devices. The average Δλs with and without the NS are 4.4 and 2.2 nm, respectively. We theoretically estimated the overlap of the laser mode with the BSA film of a constant thickness using finite-difference time-domain (FDTD) simulation (see Supplements). It predicted a 1.2-fold enhancement of Δλs by the localized field in the NS and could not explain the 2-fold enhancement in Fig. 3(b). These results suggest that the adsorption is accelerated by the NS and the excess adsorption occurs nonuniformly around the NS.
The accelerated adsorption is also suggested in the wide range of BSA concentrations C. Figure 4(a) shows spectral shifts for different C. The large shift in the high-concentration regime of ~10 nM, ShiftH, is observed for both with and without the NS. ShiftH is not saturated even in the high concentration regime of over 100 μM. This might be due to the physisorption, aggregation and phase transition  at particularly high concentrations (this could be a reason of the data scattering in Fig. 3(b)). On the other hand, the shift in the low concentration regime, ShiftL, is observed clearly only with the NS. This was observed repeatedly for all devices in different fabrication lots, but the concentration, at which ShiftL started, was distributed widely from <100 fM to >100 pM for different devices. We can consider that ShiftH arises from the adsorption over all other surfaces, while ShiftL from the adsorption inside the NS (see Supplements), and that the different concentration for ShiftL is due to the nonuniform adsorption as well as different NS width. Provided that each adsorption process occurs independently, Δλs is expressed as the following linear summation of three different Langmuir adsorption isotherms :
where K A1 and K A2 are the affinity constants of the chemisorption inside and outside of the NS, respectively, and K A3 is that of the physisorption. The latter two do not change in the presence of the NS. Δλmaxi (i = 1, 2, 3) is the maximum shift when each adsorption is saturated. The circular plots in Fig. 4(b) shows Δλs averaged for many devices at different C. Error bars for the devices without the NS denote the temporal fluctuation in the spectral shift due to the broadened spectrum. For the devices with the NS, the temporal fluctuation is negligible. The solid line of Eq. (1) fit well with the experimental plots, when parameters in Table 1 are assumed. Here, K A2 = 1.5 × 107 M−1 is a standard value for BSA . On the other hand, K A1 is more than four orders of magnitude larger than K A2. This means that huge enhancement in BSA adsorption, which enhances the effective sensitivity, occurs at the NS. The DL is equivalent to C at Δλs = Δλw, and is given by
DLs obtained by substituting the parameters into Eq. (2) are also presented in Table 1. The DL for the averaged plots without the NS is calculated from K A2 and Δλmax2 to be 33 nM ( = 2.6 μg/ml, considering the molecular weight of BSA). The DL for those with the NS, calculated from K A1 and Δλmax1, is drastically improved to be 255 fM ( = 17 pg/ml). For the best NS device, the saturation in Δλs occurred at less than 1 pM, indicating that the lowest DL is less than 10 fM.
4. Possibility of trapping effect
Figure 5(a) shows the SEM picture at the NS after the measurement in the low-concentration regime (C < 1 pM). The specific adsorption is observed clearly under strong pumping (irradiation power at pulse peak P irr = 15 mW), and not under weak pumping (P irr = 6 mW). Figure 5(b) compares spectral shifts between three different conditions for the same C. Without the NS and under the strong pumping, a significant shift is not observed. With the NS, the shift increased from 0–0.1 to 0.5–0.8 nm when the pump power is switched from weak to strong. These results show clearly that the adsorption of BSA is accelerated by the specific trapping at the NS , and it consistently explain the large enhancement of Δλs in Fig. 3(b) and the ultralow DL in Fig. 4(b).
where k B T is the thermal energy. This equation and results in Table 1 gives U trap = 9.7k B T. Possible mechanisms of the trapping include the optical gradient force from the localized laser mode and/or from pump light, and also from the convection of water and/or thermal gradient force due to heating . For example, let us apply the expression for the optical gradient force  to the laser mode. Then,
where D is the diameter of the protein when it is approximated as a sphere, n p and n w are the indices of the protein and water, respectively, ξ is the enhancement ratio of the modal energy in the NS to the average value, P is the radiated laser mode power, and V m is the modal volume in water. The highest Q in the ideal cavity is calculated to be 15,000–7,000 for w NS = 20–40 nm, respectively, also leading to V m = 0.011–0.018 μm3 and ξ = 74–46. Assuming n p = 1.5, n w = 1.321, and D = 10 nm, the trapping threshold P satisfying U trap > k B T at T = 300 K becomes 10–59 μW. On the other hand, P in the experiment is roughly estimated to be 0.01P irr = 150 μW under the strong pumping, which ensures a reasonable U trap = 15–2.5k B T in comparison with that from Eq. (3). Thus this force is a possible trapping mechanism. Other mechanisms are also worth considering, because they could explain the continuous supply of dispersed BSA molecules to the NS.
5. Comparison with other label-free biosensors
Table 2 compares the performance of label-free biosensors including this work. Since targets are different between reports, standard values of K A in literatures change from 105–1011 M−1. In general, the detection becomes easier and DL is improved for larger K A. Therefore, the performance should be evaluated from the product of DL and K A. Let us define the figure-of-merit (FOM) factor as (DL × K A)−1 normalized by that for the SPR sensor . The nanolaser achieves a FOM of 230. To date, FOMs higher than this value have been reported only for Au nanoparticles  and ultrahigh-Q micro-toroidal cavity  although they have drawbacks such as unstable DL, large sensing area, and complicated I/O. Other passive cavities and interferometers do not meet the requirement for the low DL, and their optical I/O can be a large bottleneck in cost reduction. In comparison, the nanolaser satisfies the requirements of simple fabrication, easy sensing and excellent DL. By using nanolasers as sensor chips isolated from other equipment, the cost reduction and disposable use will be possible. In addition, further investigations and controlled trapping mechanisms will lead to a unique tool for the efficient manipulation and detection of targets in biosensing.
We performed two FDTD calculations that help the understanding of the experimental results. The first one concerns the spectral shift with and without the NS, approximating the adsorbed BSA as a 20-nm thick film. Since the biosensing only targets the film, its index sensitivity Δλs/Δn f decreases from the bulk index sensitivity Δλs/Δn b. We calculated Δλs/Δn f to be 96 and 80 nm/RIU with and without the NS, respectively, which are 29 and 23% of Δλs/Δn b in . The enhancement due to the NS is only 1.2 times. Even though the mode appears to be localized inside the NS in Fig. 1(a), its small intensity is also distributed widely outside, and this reflects the small enhancement. In any case, the super-sensitivity in the experiment cannot be explained by this enhancement.
The second one is the spectral shift for the local adsorption inside the NS. It was calculated to be 26–6% of the adsorption over the whole device surface when w NS = 10–50 nm, respectively. This roughly corresponds to the experimental ratio Δλmax1/ΣiΔλmaxi = 11%, and supports the discussion that ShiftL is due to the selective adsorption inside the NS.
References and links
2. X. D. Fan, I. M. White, S. I. Shopova, H. Y. Zhu, J. D. Suter, and Y. Z. Sun, “Sensitive optical biosensors for unlabeled targets: a review,” Anal. Chim. Acta 620(1-2), 8–26 (2008). [CrossRef] [PubMed]
3. P. R. Srinivas, M. Verma, Y. M. Zhao, and S. Srivastava, “Proteomics for cancer biomarker discovery,” Clin. Chem. 48(8), 1160–1169 (2002). [PubMed]
4. D. A. Lashkari, J. L. DeRisi, J. H. McCusker, A. F. Namath, C. Gentile, S. Y. Hwang, P. O. Brown, and R. W. Davis, “Yeast microarrays for genome wide parallel genetic and gene expression analysis,” Proc. Natl. Acad. Sci. U.S.A. 94(24), 13057–13062 (1997). [CrossRef] [PubMed]
6. J. Homola, S. S. Yee, and G. Gauglitz, “Surface plasmon resonance sensors: review,” Sens. Actuators B Chem. 54(1-2), 3–15 (1999). [CrossRef]
7. K. De Vos, I. Bartolozzi, E. Schacht, P. Bienstman, and R. Baets, “Silicon-on-Insulator microring resonator for sensitive and label-free biosensing,” Opt. Express 15(12), 7610–7615 (2007). [CrossRef] [PubMed]
8. A. M. Armani, R. P. Kulkarni, S. E. Fraser, R. C. Flagan, and K. J. Vahala, “Label-free, single-molecule detection with optical microcavities,” Science 317(5839), 783–787 (2007). [CrossRef] [PubMed]
10. M. Lončar, A. Scherer, and Y. M. Qiu, “Photonic crystal laser sources for chemical detection,” Appl. Phys. Lett. 82(26), 4648–4650 (2003). [CrossRef]
11. S. Kita, K. Nozaki, S. Hachuda, H. Watanabe, Y. Saito, S. Otsuka, T. Nakada, Y. Arita, and T. Baba, “Photonic crystal point-shift nanolaser with and without nanoslots—design, fabrication, lasing and sensing characteristics,” IEEE J. Sel. Top. Quantum Electron. (to be published).
12. S. Kita, S. Hachuda, K. Nozaki, and T. Baba, “Nanoslot laser,” Appl. Phys. Lett. 97(16), 161108 (2010). [CrossRef]
14. M. A. Dündar, E. C. I. Ryckebosch, R. Nötzel, F. Karouta, L. J. van Ijzendoorn, and R. W. van der Heijden, “Sensitivities of InGaAsP photonic crystal membrane nanocavities to hole refractive index,” Opt. Express 18(5), 4049–4056 (2010). [CrossRef] [PubMed]
15. S. Kita, Y. Nishijima, H. Misawa and T. Baba, "Label-free biosensing utilizing ultrasmall photonic crystal nanolaser," in Integrated Photonics and Nanophotonics Research and Applications, OSA Technical Digest (CD) (Optical Society of America, 2009), paper IMB3.
16. T. W. Lu, P. T. Lin, K.-U. Sio, and P.-T. Lee, “Optical sensing of square lattice photonic crystal point-shifted nanocavity for protein adsorption detection,” Appl. Phys. Lett. 96(21), 213702 (2010). [CrossRef]
18. A. Di Falco, L. O'Faolain, and T. F. Krauss, “Chemical sensing in slotted photonic crystal heterostructure cavities,” Appl. Phys. Lett. 94(6), 063503 (2009). [CrossRef]
19. K. Nozaki, S. Kita, and T. Baba, “Room temperature continuous wave operation and controlled spontaneous emission in ultrasmall photonic crystal nanolaser,” Opt. Express 15(12), 7506–7514 (2007). [CrossRef] [PubMed]
20. C. Silva, F. Sousa, G. G. Bitz, and A. Cavaco-Paulo, “Chemical modifications on proteins using glutaraldehyde,” Food Technol. Biotechnol. 42, 51–56 (2004).
21. C. Pacholski, M. Sartor, M. J. Sailor, F. Cunin, and G. M. Miskelly, “Biosensing using porous silicon double-layer interferometers: reflective interferometric Fourier transform spectroscopy,” J. Am. Chem. Soc. 127(33), 11636–11645 (2005). [CrossRef] [PubMed]
22. M. Noto, D. Keng, I. Teraoka, and S. Arnold, “Detection of protein orientation on the silica microsphere surface using transverse electric/transverse magnetic whispering gallery modes,” Biophys. J. 92(12), 4466–4472 (2007). [CrossRef] [PubMed]
23. H. J. Butt, K. Graf, and M. Kappl, Physics and Chemistry of Interfaces (Wiley-VCH, 2003). p. 195.
24. A. H. J. Yang, S. D. Moore, B. S. Schmidt, M. Klug, M. Lipson, and D. Erickson, “Optical manipulation of nanoparticles and biomolecules in sub-wavelength slot waveguides,” Nature 457(7225), 71–75 (2009). [CrossRef] [PubMed]
25. K. Shigemori, S. Nishizawa, T. Yokobori, T. Shioya, and N. Teramae, “Selective binding of very hydrophilic H2PO4- anion by a hydrogen-bonding receptor adsorbed at the 1,2-dichloroethane-water interface,” N. J. Chem. 26(9), 1102–1104 (2002). [CrossRef]
26. R. Piazza, “‘Thermal forces’: colloids in temperature gradients,” J. Phys. Condens. Matter 16(38), S4195–S4211 (2004). [CrossRef]
28. W. C. Law, K. T. Yong, A. Baev, R. Hu, and P. N. Prasad, “Nanoparticle enhanced surface plasmon resonance biosensing: application of gold nanorods,” Opt. Express 17(21), 19041–19046 (2009). [CrossRef] [PubMed]
29. D. S. Grubisha, R. J. Lipert, H. Y. Park, J. Driskell, and M. D. Porter, “Femtomolar detection of prostate-specific antigen: an immunoassay based on surface-enhanced Raman scattering and immunogold labels,” Anal. Chem. 75(21), 5936–5943 (2003). [CrossRef] [PubMed]
31. A. Densmore, M. Vachon, D. X. Xu, S. Janz, R. Ma, Y. H. Li, G. Lopinski, A. Delâge, J. Lapointe, C. C. Luebbert, Q. Y. Liu, P. Cheben, and J. H. Schmid, “Silicon photonic wire biosensor array for multiplexed real-time and label-free molecular detection,” Opt. Lett. 34(23), 3598–3600 (2009). [CrossRef] [PubMed]
33. S. Zlatanovic, L. W. Mirkarimi, M. M. Sigalas, M. A. Bynum, E. Chow, K. M. Robotti, G. W. Burr, S. Esener, and A. Grot, “Photonic crystal microcavity sensor for ultracompact monitoring of reaction kinetics and protein concentration,” Sens. Actuators B Chem. 141(1), 13–19 (2009). [CrossRef]