This paper reports a digital micro-mirror device (DMD)-enabled real-time multi-channel biosensing system based on angular interrogation surface plasmon resonance (SPR). In the experiments, angular scanning is achieved by a DMD that facilitates SPR measurements using a single-point photodetector. In the four-channel measurement setup, real-time monitoring of bovine serum albumin (BSA) and anti-BSA binding interactions is performed at various concentration levels. The experimental results have verified that the system has a resolution of 3.54 × 10−6 RIU (refractive index unit); and a detection limit of 9 ng/mL. The new DMD-based SPR interrogation system presents a new design route for practical solid-state SPR biosensing with a user-selectable range of interrogation, enhanced signal-to-noise ratio, and fast data throughput.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Surface plasmon resonance (SPR) sensor has been widely demonstrated as a practical label-free sensing technology for analyte detection and characterization of biomolecular interactions in the past two decades . It utilizes the evanescent field of a special mode of electromagnetic field propagating at a metal/dielectric interface, i.e., surface plasmon, to measure variations in the refractive index of the dielectric in the proximity of the interface. Most biological species have higher refractive index (RI) comparing to that of water. An increase of RI is typically associated with biomolecular binding as it results in the formation of an ultra-thin organic layer on the metal film . This phenomenon has led to wide applications of SPR in various fields of health and biological sciences, e.g., environmental monitoring , clinical diagnostics , food safety screening , drug discovery , and measurement of optical constants .
The quest for better detection performance has resulted in the development of numerous SPR sensor platforms. The schemes of optical excitation of surface plasmons include attenuated total reflection (prism coupling) , diffraction on periodic metallic gratings (grating coupling)  and nanoparticle-enhanced localized SPR . The modulation approaches used in SPR sensors are based on wavelength , angle  and phase [13–15] interrogation. However, no matter which SPR configuration is used, measurements are basically performed by measuring intensities of light; therefore the performance of SPR sensors are predominantly limited by photon statistics .
Existing SPR solutions utilize an array detector such as CCD to record the intensity data . This imaging array is similar to that used in a scientific camera. However, the use of array detectors presents a significant challenge. They are either quite expensive or typically of less than optimal performance in terms of signal-to-noise (S/N) ratio, rate of response and issues related to inter-pixel cross-talks. In some cases, dead pixels and pixel-to-pixel non-uniformity will further prohibit the possibility of achieving ultimate sensing performance close to the theoretical limit.
It has been predicted that future improvement of SPR sensors will depend mostly on the development of optical detectors with a higher S/N ratio and faster data throughput. The ultimate detection noise is limited by the statistical fluctuations corresponding to the number of detected photons, which suggests that higher photon flux levels are always desired. This objective can only be achieved by using a detector that has large size, wide dynamic range and the capacity of taking temporal average of intensity measurements . Such attributes suggest that single-point photodetectors may be a better light detecting device in future SPR sensors. It has been shown that with a photodiode array detector, the maximum S/N ratio is about 14 times higher than that of the CCD . Single-point photodetector has been used in angular interrogation SPR system to measure SPR response in different angles [19,20]. However, the angular scanning was achieved by mechanical rotation of a laser source or prism coupling system, which is slow, mechanically unstable, and prone to hysteresis.
To realize a high performance SPR system, we have developed a digital micro-mirror device (DMD)-enabled angular interrogation-based SPR sensor configuration using a single-point photodetector. A DMD is a high-speed 2-D light modulation device (max. pattern rate: 32.5 kHz) containing several millions of micromirrors. The unique capability of DMD enables a high-speed, scanless, and programmable SPR sensor design that uses a single-point detector with high S/N ratio, wide dynamic range, and fast response, thereby achieving high-resolution SPR measurements with substantially reduced cost.
All chemicals used in SPR biosensing are purchased from Sigma Aldrich. To generate a monolayer with carboxyl termination, 1 mM 11-mercaptoundecanoic acid (MUA) solution is kept at room temperature to settle overnight. Excess MUA is rinsed with ethanol and followed by deionized water. Freshly prepared 2-(N- Morpholino) ethane sulfonic acid (MES) from mixing equal volume of 400 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and 150 mM N-hydroxysuccinimide (NHS) is then added and incubated for 10 min at room temperature in order to activate the carboxyl terminal on the chip and form covalent bonding with the amine group of the target molecules. For the biosensing experiments that monitor antigen-antibody interaction, 1 mg/mL bovine serum albumin (BSA) is incubated on the chip for 30 min at room temperature. The BSA immobilized on the chip surface is considered as an antigen, while the anti-BSA antibody is added to bind specifically on the BSA. Various concentrations (0.2, 2, 20 µg/mL) of the antibody are injected and incubated for the indicated time at room temperature. Phosphate-buffered saline (PBS) solution is then flown into the reaction chamber to remove the excess antibody. 50 mM NaOH, an alkaline that denatures antibody to remove its binding affinity on antigen, is injected and then PBS is added again. SPR signal is recorded at 20-sec intervals.
2.2 Preparation of the SPR sensor chip
The SPR cell is composed of a right-angle prism made of BK7 glass, a SPR gold chip, and a flow chamber device. The gold chip, fabricated by sputtering, has a gold film with a thickness of approximately 50 nm (refractive index matching oil is used to connect the prism and the chip). The flow chamber device is mounted on the chip surface allowing different sample solutions to make contact with the gold film. For the simultaneous multiple sensing experiments, the flow chamber device contains four channels; each channel acts as an independent cell. The width and length of each flow channel is 0.9 mm and 10.0 mm respectively, i.e., the contact area between the fluid and the sensing surface is 9 mm2 in the calculation of mass sensitivity. The distance between adjacent channels is 0.8 mm.
2.3 Optical configuration
Figure 1 presents the optical configuration of the SPR system, where a He-Ne laser is used as the light source. The laser beam is expanded by a telescope system and directed to the DMD (1140 × 912 pixels, DLP4500VIS, Texas Instrument). The scanning speed of the DMD device is 4.2 kHz. The selection of individual incident angle for SPR excitement is accomplished by selectively switching the columns of micromirrors on the DMD to the “on” or “off” positions. Accordingly, only light rays with the desired angle are reflected to the SPR cell and further to the detector. The single-point photodetector in the system is OPT101 (Monolithic Photodiode and Single-supply Transimpedance Amplifier, Texas Instruments). The size of the detection area is 2.29 × 2.29 mm2 and the measurement bandwidth is 14 kHz. Adjacent columns of micromirrors are turned on successively to conduct angular scanning; and angular interrogation SPR responses are measured successively by the single-point detector. In this respect, angular interrogation of SPR is realized in the time domain. This architecture, with its programmable operation range enabled by the DMD, offers the advantages of angle selection agility, repeatability, and mechanical stability. Light rays selectively reflected by the DMD are subsequently divided by a polarizing beam-splitter. The s-polarized light is focused onto the point detector D1 through a spherical lens (fL3 = 100 mm) to serve as the reference signal. The p-polarized light passes through a cylindrical lens (fCL1 = 50 mm) to form a convergent beam with an incident angle ranging from 70.2° to 76.2° with respect to the sensor surface. After passing through the SPR cell, the reflected light is collimated by another cylindrical lens (fCL2 = 50 mm) to form a collimated light beam, which is subsequently collected by a spherical lens (fL4 = 100 mm) before finally entering the point detector D2.
The proposed SPR sensor architecture offers the capability to perform simultaneous multi-channel biosensing. To realize this, each column of micromirrors on the DMD may be further divided into multiple regions in the horizontal direction. The SPR excitation in different biosensor cells (arranged in the horizontal direction) within the linear array is achieved by selectively activating different micromirror clusters, i.e., turning rows of micromirrors on the DMD to the “on” or “off” positions to excite only the desired SPR cell. The reflected light rays from different SPR channels are focused onto the same point and measured by the single-point detector. This ensures that all sensing elements are addressed sequentially at high speed.
2.4 System control and data analysis
An Arduino board is used to control the SPR system. The output trigger signal of the DMD is utilized for synchronizing the two detectors (D1 and D2) with the pattern sequence. During the period of a single DMD frame, the Arduino board reads ten data points (in voltage) from the detector D1 and D2 in the reference arm and the probe arm respectively. We average the 10 data points from each arm as the final output. Then, the voltage outputs are converted into the radiant powers. The ratio of the power in the probe arm to that in the reference arm is calculated as the final response signal of the SPR system, i.e., power ratio. The collected data are analyzed using a custom-developed software in MATLAB. This process is repeated until all the patterns are displayed successively. Depending on the need, a pattern defined by a rectangle of arbitrary size can be uploaded to the DMD. For real-time multi-channel sensing experiments, the pattern exposure time of the DMD is set to 10 ms, during which the Arduino board reads 10 voltage data points from each detector and calculates the power ratio for the current pattern, or equivalently, the incident angle that it represents. There are 45 angular points in total. Therefore, each complete angular scan for one channel can be completed in 450 ms, which means that for 4 channels the total measurement time is 1.8 s. The total required time for performing the measurement is approximately 20 seconds. This time resolution is currently limited by the data acquisition and transmission speed of the Arduino board, which can be further improved by using a more powerful microcontroller board.
3. Results and discussion
3.1 Characterization of the SPR system
For an angular interrogation SPR system, an SPR response curve indicates the reflected intensity of light versus the incident angle, which can be derived from the three-layer Fresnel equation . The angle yielding the minimum light intensity on the SPR curve is denoted as the resonance angle (RA). Accordingly, we deﬁne that the sensor sensitivity (S) as the RA shift () caused by a small refractive index change (), which can be mathematically expressed in Eq. (1):
Sensor resolution (R) is defined as the lowest detection limit that the SPR sensor can resolve. This parameter is affected by the measurement uncertainty standard deviation (S.D.), which is described in Eq. (2):
To estimate the performance of the SPR system, experiments are performed to measure the RA shift with refractive index variation caused by dissolving glycerol in water. Glycerol-water mixtures with various solute concentrations, i.e., 0% (water), 0.0625%, 0.125%, 0.25%, 0.5%, 1%, 2%, 4% and 8% in weight ratio are injected into the prism coupling system as samples. The corresponding refractive indexes range from 1.3330 to 1.3424 RIU. As shown in Fig. 2(a), the SPR curves are fitted by 3rd-order polynomials [22,23] to the theoretical Fresnel equation , which provides a close fit with an optimal sensitivity . The minimum values of the fitting curves are then determined as RAs. Within the most sensitive region of the SPR response plot, we find a RA shift of 0.0735° for the refractive index change from 1.3330 to 1.3331 RIU (from water to 0.0625% solution), as shown in Fig. 2(b). The fluctuation of the measurement is estimated by monitoring the S.D. of 10 data points collected from a sample containing 8% of glycerol. The monitoring period is 10 min with each data point collected in every minute. The best experimental S.D. is found to be 0.0026°. Taking this as the minimum angular measurement that the system can resolve, the estimated sensor resolution is 3.54 × 10−6 RIU.
For the multi-channel sensing experiments, water is injected to all channels and RAs are recorded as the baseline. Experimental data, i.e., the 15 lowest data points, are fitted by 3rd-order polynomials. Then glycerol-water mixtures with various solute concentrations, i.e., 1%, 2%, 4% and 8% in weight ratio are injected into channel 1, 2, 3, 4 as samples. RAs are recorded and presented in Table 1. The slightly different responses among different channels for the same solution are attributed to the thickness variation of the gold film across the different sensing surfaces. Relative shifts of RAs are plotted in Fig. 3; the results confirm the sensor array’s capability for simultaneously detecting multiple analytes.
To demonstrate multi-channel scanning capability of the DMD-enabled SPR system, a sensor head containing four channels is used to independently monitor BSA antigen-antibody binding at different concentration levels. The experimental results are presented in Fig. 4. Channel 1 is set as the negative control with distilled water flowing inside the chamber during the entire immobilization of BSA. In step 1, surface activation with MES is performed in Channel 2 - 4, where a sharp increase in SPR signal is observed (Fig. 4). The increase in SPR signal is due to a large refractive index difference between the MES and distilled water. Afterwards, the BSA is injected into Channel 4, where in Fig. 4 a corresponding gradual increase of SPR signal can be observed. In step 5, Channel 2 is not blocked with excess MES-activated group, thereby leaving the possibility of downstream covalent protein conjugation. In step 8, the SPR signal of Channel 4 is increased quickly within the first few minutes when the anti-BSA is added, indicating rapid binding of the anti-BSA antibody on the coated BSA surface by molecular binding affinity. The response curves saturate in around 15 min, suggesting that the binding has reached equilibrium. In step 9, the injection of the PBS solution, removing all unbound anti-BSA antibody, reveals that the true SPR signal change is solely due to the antigen-antibody interaction, rather than the sum of antigen-antibody interaction plus reflective index difference in anti-BSA antibody solution. The small change in SPR signal in Channel 1 - 3 suggests that there is non-specific adsorption of anti-BSA antibody on the chip surface. To prepare the SPR sensing surface for reuse, SPR signal is recorded during and after NaOH circulation. In step 9, the signal returns to the original value in Channel 1, 3 and 4, suggesting the adsorbed and bound anti-BSA antibody are dissociated from the BSA-coated surface (The final signal in Channel 4 is slightly higher than the original value; this is due to the fact that the adsorbed and bound anti-BSA antibody are not fully dissociated from the BSA-coated surface during the NaOH circulation), while the signal remains unchanged in Channel 2, suggesting the anti-BSA antibody is covalently bound on the unblocked MES-activated SPR surface and the anti-BSA antibody in Channel 2 cannot be dissociated simply by the reduction of binding affinity with a pH change.
As a demonstration, we apply our SPR system to measure quantitative real-time antigen-antibody interaction in four channels simultaneously. Concentration-dependent SPR responses obtained from the four channels are presented in Fig. 5. After surface functionalization by the BSA antigen, samples of anti-BSA antibody at various concentration levels (0, 0.2, 2, 20 µg/mL) are injected into the different channels, as shown in Fig. 5(a); and Fig. 5(b) is the summary of the net change in SPR signal of anti-BSA antibody addition in Fig. 5(a).
The measured SPR angular shift of 0.2 μg/mL (or 200 ng/mL) is 0.0456 as shown in Fig. 5(b). On the other hand, the steady baseline of PBS injection shows an S.D. of 0.002. The detection limit of the anti-BSA antibody detection is calculated according to the established equation: Detection Limit = Concentration of Biomolecule / Sensor Response Measurement Stability. Therefore, it is 200 ng/mL / 0.0456 × 0.002 = 9 ng/mL, which is similar to other SPR systems for bio-sensing on BSA antigen-antibody binding [24,25]. Since each flow channel in our device has a volume of 0.0108 mL, the corresponding system mass sensitivity is calculated to be (9 ng/mL 0.0108 mL) / 9 mm2 = 10.8 pg/mm2. The mass sensitivity ranges from 0.5 to 10 pg/mm2 in other reported SPR systems with a similar RIU of 10−6, where lower mass sensitivity (0.5 pg/mm2) is indeed achieved by pre-coating biomolecules in a 3-D dextran matrix (CM5 chip) instead of monolayer format reported here [26,27]. To illustrate the repeatability of the sensing system, the anti-BSA antibody and NaOH are injected into the system for measurement; and the experiments are repeated three times. The results are presented in Fig. 6. From Fig. 6, one may observe that the system is highly repeatable, where measured curves in all three experiments have similar rising and decreasing traits. For the analysis of reproducibility, coefficient of variation (CV) is introduced. CV is the analysis of the intermediate precision of a single system over multiple detections, and measure variability of CV percentage (CV%) < 5% is considered as an accurate detection . CV% is calculated as (Standard Deviation / Mean) 100 = (0.008/0.473) 100 = 1.7%, which supports high reproducibility in our biosensing measurement.
In summary, we have developed a DMD-enabled real-time multi-channel biosensor system based on angular interrogation of the Kretchmann configuration. Notably, the system employs a single-point detector that simultaneously optimizes the system S/N ratio, provides a wide dynamic range and fast response. The experimental results have confirmed that the system has a resolution of 3.54 × 10−6 RIU; and a detection limit of 9 ng/mL, calculated based on real-time monitoring of BSA/anti-BSA binding interactions at various concentration levels using a four channel biosensor array. Further enhancement of measurement resolution can be achieved by increasing the laser excitation power or optimizing the shot-noise limited performance of the single-point detector.
HKSAR Research Grants Council (RGC) General Research Fund (GRF) (412613 and 14206517); CUHK Direct Grant (CUHK 3132798 & 4053100); AoE Scheme Funding (AoE/P-0/12); HKSAR Innovation and Technology Commission (ITS/428/17FP).
1. J. Homola, “Electromagnetic theory of surface plasmons,” in Surface plasmon resonance based sensors (Springer, 2006), pp. 3–44.
2. Y. Huang, H. P. Ho, S. K. Kong, and A. V. Kabashin, “Phase‐sensitive surface plasmon resonance biosensors: methodology, instrumentation and applications,” Ann. Phys. 524(11), 637–662 (2012). [CrossRef]
5. L. D. Mello and L. T. Kubota, “Review of the use of biosensors as analytical tools in the food and drink industries,” Food Chem. 77(2), 237–256 (2002). [CrossRef]
6. J.-P. Renaud, C. W. Chung, U. H. Danielson, U. Egner, M. Hennig, R. E. Hubbard, and H. Nar, “Biophysics in drug discovery: impact, challenges and opportunities,” Nat. Rev. Drug Discov. 15(10), 679–698 (2016). [CrossRef] [PubMed]
7. G. Hu, H. He, A. Sytchkova, J. Zhao, J. Shao, M. Grilli, and A. Piegari, “High-precision measurement of optical constants of ultra-thin coating using surface plasmon resonance spectroscopic ellipsometry in Otto-Bliokh configuration,” Opt. Express 25(12), 13425–13434 (2017). [CrossRef] [PubMed]
8. E. Kretschmann and H. Raether, “Radiative decay of non radiative surface plasmons excited by light,” Z. Naturforsch. A 23(12), 2135–2136 (1968). [CrossRef]
9. R. W. Wood, “XLII. On a remarkable case of uneven distribution of light in a diffraction grating spectrum,” Lond. Edinb. Dublin Philos. Mag. J. Sci. 4(21), 396–402 (1902). [CrossRef]
10. P. Singh, “SPR biosensors: historical perspectives and current challenges,” Sens. Actuators B Chem. 229, 110–130 (2016). [CrossRef]
11. K. Johansen, H. Arwin, I. Lundström, and B. Liedberg, “Imaging surface plasmon resonance sensor based on multiple wavelengths: Sensitivity considerations,” Rev. Sci. Instrum. 71(9), 3530–3538 (2000). [CrossRef]
12. R. Karlsson and A. Fält, “Experimental design for kinetic analysis of protein-protein interactions with surface plasmon resonance biosensors,” J. Immunol. Methods 200(1-2), 121–133 (1997). [CrossRef] [PubMed]
13. H. Ho and W. Lam, “Application of differential phase measurement technique to surface plasmon resonance sensors,” Sens. Actuators B Chem. 96(3), 554–559 (2003). [CrossRef]
15. F.-C. Loo, S.-P. Ng, C.-M. L. Wu, and S. K. Kong, “An aptasensor using DNA aptamer and white light common-path SPR spectral interferometry to detect cytochrome-c for anti-cancer drug screening,” Sens. Actuators B Chem. 198, 416–423 (2014). [CrossRef]
18. G. G. Nenninger, M. Piliarik, and J. Homola, “Data analysis for optical sensors based on spectroscopy of surface plasmons,” Meas. Sci. Technol. 13(12), 2038–2046 (2002). [CrossRef]
19. X. Guo, “Fe3O4@ Au nanoparticles enhanced surface plasmon resonance for ultrasensitive immunoassay,” Sens. Actuators B Chem. 205, 276–280 (2014). [CrossRef]
20. A. R. Sadrolhosseini, A. S. Noor, A. Bahrami, H. N. Lim, Z. A. Talib, and M. A. Mahdi, “Application of polypyrrole multi-walled carbon nanotube composite layer for detection of mercury, lead and iron ions using surface plasmon resonance technique,” PLoS One 9(4), e93962 (2014). [CrossRef] [PubMed]
21. E. Kretschmann, “Die bestimmung optischer konstanten von metallen durch anregung von oberflächenplasmaschwingungen,” Zeitschrift für Physik A Hadrons and nuclei 241(4), 313–324 (1971). [CrossRef]
22. K. Johansen, R. Stålberg, I. Lundström, and B. Liedberg, “Surface plasmon resonance: instrumental resolution using photo diode arrays,” Meas. Sci. Technol. 11(11), 1630–1638 (2000). [CrossRef]
23. R. Khrystosenko, “Optimization of the surface plasmon resonance minimum detection algorithm for improvement of method sensitivity,” Semiconductor Physics Quantum Electronics & Optoelectronics (2015).
24. C. L. Wong, G. C. K. Chen, B. K. Ng, S. Agarwal, Z. Lin, P. Chen, and H. P. Ho, “Multiplex spectral surface plasmon resonance imaging (SPRI) sensor based on the polarization control scheme,” Opt. Express 19(20), 18965–18978 (2011). [CrossRef] [PubMed]
25. H. Ho, W. Law, S. Wu, X. Liu, S. Wong, C. Lin, and S. Kong, “Phase-sensitive surface plasmon resonance biosensor using the photoelastic modulation technique,” Sens. Actuators B Chem. 114(1), 80–84 (2006). [CrossRef]
26. U. Jönsson, L. Fägerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, S. Löfås, B. Persson, H. Roos, and I. Rönnberg et al.., “Real-time biospecific interaction analysis using surface plasmon resonance and a sensor chip technology,” Biotechniques 11(5), 620–627 (1991). [PubMed]