Chapter 2 IGZO-TFT Biosensors for Investigation of Biotin-Protein Interaction
2.1 Introduction
Y-type Microfluidic Channel
In many biomedical applications (e.g. biomedical analysis, drug delivery, chemical synthesis, and enzyme reactions), Y-type Microfluidic Channels, also known as micro-mixers [23, 24], are crucial and have many advantages [25-27]. Less consumption and contamination of bio-samples, as well as the time, are one of the merits.
The materials and fabrication of microfluidic channel have become more important because of the relevance to biology and chemistry during the development of microfluidic system. The characteristics of simple fabrication, optical transparency, and low cost make polymer materials such as PDMS and SU-8 negative photoresist widely used [28, 29].
In this part, PDMS layers is used as the Y-type microfluidic channels. The biological molecules diffuse and interact through the microfluidic channel and the mixture situations are determined by electrical signal and diffusion properties
Biotin-Avidin System
In the past decade, immunocytochemical techniques have a great advancement. For example, the Biotin-Avidin System (BAS), which has the following advantages, is one of them. First, biotin and avidin have an extraordinarily high affinity and specificity between each other. Second, many of the macromolecules can be conjugated to biotin and maintain the original biological activity. Third, the binding reaction is considered irreversible since the dissociation constant of the two is very small. Finally, four binding sites for biotin molecules of each avidin molecule makes it possible to be used to construct an effective
amplification system of biological reaction. BAS are widely used in biological and chemical applications such as tagging or the delivery of molecules, trace antigen, qualitative antibody and quantitative detection [30, 31]. Thus, understanding the mixture condition of the reaction is crucial.
Previous works have presented some optical methods such as fluorescence microscope and enzyme-linked immunosorbent assay (ELISA) [30],but these methods are not real-time measurements. Another study presented a real-time detection method by detecting optical shift of a resonant micro-cavity. In the study, biotinylated BSA was selected as the recognition element and immobilized on the sensing surface followed by the injection of streptavidin, which leads to a shift of the optical resonance wavelength.
The shift is due to binding interaction between introduced streptavidin and surface immobilized biotin [32]. Even though the study have preliminary determination ability of the occurrence of biotin and streptavidin binding reaction, a demand on further analyzing the internal mixture interaction in the BAS system still exists.
In this study, we demonstrated an IGZO-TFT biosensor for detecting the mixture situation of biotin and streptavidin. In addition, an analysis and experiment framework was discussed, which is useful on delving into further properties of the BAS reaction.
Moreover, bovine serum albumin (BSA) was applied as the control sample and several control experiments were conducted to verify the system specificity and the analysis theory.
2.2 Material and Methods
2.2.1 Fabrication of IGZO-TFT Biosensors Integrated with Y-type Microfluidic Channels
Fabrication of IGZO-TFT
The cross section of the double-gated IGZO-TFT is shown in Fig. 2.1. A Staggered bottom-gate TFT with the top gate was fabricated on the Corning Eagle 2000 glass substrate. The fabrication started with DC sputtering and reactive ion etching (RIE) molybdenum (Mo) metal as bottom gate. A silicon dioxide (SiO2) dielectric layer of 300 nm was deposited by plasma enhanced chemical vapor deposition (PECVD) at 300℃ as insulator. Then, 50-nm IGZO (In2O3:Ga2O3:ZnO =1:1:1) thin film was formed by RF sputtering at room temperature and etched by HCl wet etching to form the active layer with 100-μm channel width and 50-μm channel length . The 200-nm Mo source and drain contact layer was deposited by DC sputtering and etched by RIE. A SiO2 passivation layer was deposited by RF sputtering and followed by RIE etching to prevent the TFT from humidity and oxidation. The top gate with the thickness of 300-nm gold metal was evaporated by E-gun evaporation with lift-off technique. Finally, the fabrication process was finished with post-annealing under oxygen ambient at 270 ℃ for 30 minutes in the oven tube.
Fig. 2.1 The cross section of a staggered and double-gated IGZO-TFT.
Fabrication of Y-type Microfluidic Channels
The target biomaterials were applied on the microfluidic channels, which were fabricated on a separate glass substrate from the TFT chips. The Au sensing pad of 200 nm was first evaporated on the Corning Eagle 2000 glass substrate. In the next step, the PDMS prepolymer and curing agent (Sylgard 184 silicone elastomer kit, Dow Corning, Midland, MI) were mixed in the weight ratio of 10:1 and stirred well [33]. The PDMS viscous solution was then poured over the Y-shape acrylic patterned mold contained in a circular flat-bottomed polystyrene Petri dish followed by curing at 80 ℃ for 20 mins.
After curing, the PDMS replica of the patterned mold was peeled off the Petri dish and cut into a proper rectangular shape containing microfluidic channels. Fig. 2.2 shows the fabrication process of the PDMS replica.
The PDMS and glass substrate were brought together to form an irreversible seal after treated with oxygen plasma for 6 mins at 70 ℃ in the UV Ozone [34, 35]. The width and length of the microfluidic channel are 1000 and 10000 μm, respectively. Fig. 2.3 illustrates the PDMS bonding treatment flow. As illustrated in Fig. 2.4, the top gate contact of the TFT is wire bonded by Au wires to the Au metal pad of the microfluidic channel chip. The Y-type microfluidic channel contains two inlets and one outlet, the region of interest (ROI) was selected at the end of the microfluidic channel.
Fig. 2.2 Fabrication process of the PDMS replica.
Fig. 2.3 PDMS bonding treatment flow.
Fig. 2.4 Microscope up-view of TFT and the microfluidic channel and schematic of the reusable TFT biosensor. The TFT and a microfluidic channel chip were fabricated separately and then connected together by wire bonding.
2.2.2 Measurement and experiment flow
The experiment setup is shown in Fig. 2.5. The ROI was functionalized with 11-Mercaptoundecanoic acid (11-MUA) cross linker for 2 hours at room temperature and
washed by phosphate buffered saline (PBS). The analytes were introduced into the microfluidic channel by Chemyx Fusion 200 syringe pump.
Fig. 2.5 Experiment setup for detecting target analytes
The experiment aims at demonstrating the TFT sensor system as the platform for monitoring real-time biochemical reactions. The target analytes that diffuse through the microfluidic channel and reach the ROI will be captured by the cross linkers. Much proteins and biomaterials carry net charges. As illustrated in Fig. 2.6, additional charges will be induced in the TFT channel layer once the target proteins were sensed on the Au sensing electrode. The amount of drain current change is relevant to the concentration of the analytes.
Using the reaction of biotin and streptavidin as an example, the diffusion behaviors of biotin, streptavidin and biotin-streptavidin complex are characterized. The measurement consists of two steps. The diffusion times of biotin and streptavidin are first benchmarked by injecting the analyte at the inlet and meanwhile monitoring the real-time response of TFT drain current. In the second step, both biotin and streptavidin were applied into one single microfluidic channel chip. Signals of the individual analyte and the reaction complex were examined by the drain current response. Two delay
to understand the reaction dynamics of the mixture in the microfluidic channel. Finally, a control experiment is conducted. There is no nonspecific binding between bovine serum albumin (BSA) and streptavidin. It can be used to examine the reliability and the specificity of the previous experiments. In the control measurement, biotin is replaced with a 0.4mM BSA solution. The current response of the BSA and streptavidin mixture is compared with that of biotin and streptavidin.
Throughout the experiment, the electrical characteristics were characterized by Agilent 4155C semiconductor parameter analyzer. The TFT was biased at gate-source voltage, VGS, of 10 V, and drain-source voltage, VDS, of 5 V. The drain currents were sampled every 10 seconds.
Fig. 2.6 Illustration of the target analyte charge sensing of the TFT biosensor
2.3 Results and Discussion
2.3.1 Confirmation of diffusion dominant or flow dominant
In order to confirm the transmitting of biomolecules is diffusion dominant or flow dominant, the dye was injected into the microfluidic channel at the flow rate of 0.0003
ml/min, which is the same as that of the measurement. The cross-sectional area of the channel is 1 mm2 and the distance between the inlet and the electrode is 6 mm. If the transmitting mechanism is flow dominant, the flowing time will be 1200 s. In Table 1, the durations between the injection and the arrival of dye at electrode are much shorter than 1200 s. Therefore, the flow rate is low enough to make the transmitting mechanism being diffusion dominant.
Table 1. The Transmitting of the Dyes of Different Concentrations
Concentration Start of injection Arrival at electrode
1
t = 0 s t = 200 s
1/4
t = 0 s t = 340 s
2.3.2 Transient drain current responses by applying Biotin and Streptavidin separately
The streptavidin and biotin diffusion times are first benchmarked. The biotin concentration was selected to be 0.4mM and streptavidin concentration is 1.67μM. The PBS buffer was prefilled into microfluidic channel at t = 0 second. Biotin solution was then injected 60 second later. The drain currents of the TFT are sampled every 10 seconds
under VGS = 10 V, VDS = 5 V. The transient responses of TFT drain current for biotin and streptavidin are shown in Fig. 2.7 and Fig. 2.8. The measurements for each analyte were conducted twice and denoted as Measurement 1 and Measurement 2. The increases of the drain current are the results of the detection of biotin and streptavidin flowing through the ROI and bonded to the cross linker. Since biotin and streptavidin carry net charges, the increase of drain current is associated with the concentration of biomolecules. For biotin, the onset of the drain current was detected at t = 210 seconds for Measurement 1 and t = 240 for Measurement 2. The diffusion time is defined as the duration between the introduction of the target analyte and the onset of drain current increase. Therefore, the diffusion time of biotin in this measurement system is in the range between 150 and 180 seconds. As for streptavidin, the current increases at t = 460 and t = 480 seconds, so the diffusion time of streptavidin is in the range of 400 and 420 seconds.
Fig. 2.7 Drain current responses of 0.4mM biotin. PBS was applied at t=0 while biotin at t =60 seconds
Fig. 2.8 Drain current responses of 1.67μM streptavidin. PBS was applied at t=0 while streptavidin at t =60 seconds
2.3.3 Transient drain current responses by applying Biotin and Streptavidin mixture
The analysis of biotin-streptavidin biochemical reactions was next conducted.
Before looking into the responses of the mixture, mechanism of biotin and streptavidin interaction in the Y-shaped microfluidic channel is first discussed. The specific diffusion times t1, t2 and t3 are defined as the characteristics of biotin, streptavidin, and streptavidin-biotin complex, respectively. There are four possible scenarios as illustrated in Fig. 2.9 when the analytes are injected into inlet 1 and 2 simultaneously. Fig. 2.9(a) denotes the biotin, streptavidin and biotin-streptavidin complex biomolecules reach the ROI in sequence. Three species detected after the reaction is indicated by three obvious current incremental steps. In the second case, when the streptavidin is exhausted during the reaction process, only the biotin and complex signals are detected (see Fig. 2.9(b)). On
the other hand, Fig. 2.9(c) shows only streptavidin and complexes signals are observed if biotin molecules are all consumed. Finally, in the case of complete reaction, during which no biotin and streptavidin are left in the channel, only the complex biomolecules can be detected (see Fig. 2.9(d)).
In the experiment, PBS was introduced at 0 second and streptavidin (1.67μM) and biotin (0.4mM) were applied in inlet 1 and inlet 2 (see Fig. 2.4) 60 seconds later, respectively. They may be mixed and react in the microfluidic channel into complexes.
The experiment was conducted on two different devices, denoted as measurement 1 and measurement 2. There are two distinct current increment steps in Fig. 2.10. The first one at around t = 220 seconds (diffusion time of ~160 seconds) corresponds to the signal of biotin and the second response is at t = 740 seconds, which is longer than the diffusion time of streptavidin in Fig. 2.8. The increase at t = 740 seconds is attributed to the biotin-streptavidin complexes for the 1.67μM of streptavidin being all consumed in the reaction of biotin and streptavidin. The results in Fig. 2.10 correspond to the assumption in Fig.
2.9(b).
Fig. 2.9 Scenarios of drain current responses when the biotin and streptavidin are applied to the microfluidic channel
Fig. 2.10 Drain current responses of biotin and streptavidin interactions in the microfluidic channel
2.3.4 Delay experiment of Biotin and Streptavidin reaction in the microfluidic channel
In order to further understand the biochemical reaction of biotin and streptavidin, the experiments of either streptavidin or biotin being delayed to be injected into the microfluidic channel was designed. In the first case, defined as streptavidin delay experiment, after PBS prefilled at t=0, biotin and PBS were introduced into inlet 1 and 2 at t =60 seconds, respectively. At 660 seconds, the injection of PBS solution was replaced with streptavidin solution.
In Fig. 2.11, possible scenarios of streptavidin delay experiment is shown. For the first case in Fig. 2.11(a), the condition that streptavidin can't react with biotin sis assumed.
It means that even though they meet with each other in the channel, there is no reaction
occurs and no complex production. Thus, the drain current increases at t2 seconds after the injection of streptavidin and no third current increase is observed for the complexes.
For the second scenario, the biotin molecules are distributed in the microfluidic channel and thus all of streptavidin molecules are reacted with biotin. In such a case, the arrival signal of the complexes will be observed, while that of streptavidin will be missing (see Fig. 2.11(b)). For the last scenario as shown in Fig. 2.11(c), only part of the streptavidin molecules bind with biotin. There will be three increase steps in the transient response of drain current.
To verify the reaction when the streptavidin injection is delayed for 10 minutes, the experiment as described above is conducted. Biotin is introduced at t=60 seconds and streptavidin is at t =660 seconds. In Fig. 2.12, the results are shown. The drain current increase at 250 seconds indicates the arrival of biotin at ROI. Whereas, the diffusion signal of streptavidin is barely seen. A current increase at t =1320 seconds suggests the signal of the biotin-streptavidin complexes. The experiment coincides with the scenario described in Fig. 2.11(b), which indicates the reaction of biotin and streptavidin in the microfluidic channel.
Fig. 2.11 Illustration of the current response scenarios of the streptavidin delay experiment. Biotin solution was introduced in inlet 1 at t= 60 s and streptavidin solution was introduced in inlet 2 at t= 660 s
Fig. 2.12 Current response of the streptavidin delay experiment. PBS, biotin and streptavidin were introduced at t =0, 60 and 660 seconds
A similar experiment was conducted when the biotin solution was the one being delayed. As shown in Fig. 2.13, three possible scenarios that describe the interaction
the injection of the streptavidin, drain current increases in 400 seconds, which is owing to the arrival of streptavidin. The biotin was applied at t = 660 seconds. A small but observable current increase occurs 220 seconds after biotin injection. The increase is believed to be the signal of biotin. The biotin may be detected by binding either with the cross linkers or with the streptavidin. At t=1370 seconds, the biotin-streptavidin complexes are detected. The scenario corresponds to the case described in Fig. 2.13(c).
Since the concentration of the biotin solution (0.4mM) is much higher than that of the streptavidin solution (1.67μM), biotin molecules will not be consumed. There must be a third signal observed.
Fig. 2.13 Illustration of the current response scenarios for the biotin delay experiment
Fig. 2.14 Drain current response of the biotin delay experiment. PBS, streptavidin and biotin were introduced at t =0, 60 and 660 seconds
2.3.5 BSA control experiment
The specificity of the biotin-streptavidin biochemical reaction is next examined. For the biotin-streptavidin reaction, BSA is ideal as the reference group because of its non-specific property with streptavidin. In this study, in order to verify the non-specificity of the reaction in the microfluidic channel, biotin was replaced with BSA.
The diffusion behavior of BSA was first benchmarked. As shown in Fig. 2.15, an arrival time of 460 seconds is observed, which is regarded as the diffusion characteristic of BSA. Next, a delay experiment of BSA was conducted. After PBS prefilled at t=0, streptavidin and PBS were introduced into inlet 1 and inlet 2 at t =60 seconds, respectively.
The injection of PBS solution was replaced with BSA solution at 660 seconds.
Before the BSA delay experiment, several scenarios are considered. As streptavidin is introduced at the beginning, a diffusion time t2 of streptavidin molecules will be
observed. The injection of BSA will result in four possible conditions similar to the cases of biotin delay experiment as illustrated in Fig. 2.16. BSA molecules diffuse to ROI 460 seconds later. For case 1, BSA may be captured by the cross linkers. For case 2, BSA is captured by streptavidin at the ROI and form “non-specific” binding . In both case 1 and case 2, current increase will be observed when BSA reaches ROI, but case 2 is less likely.
BSA should not bind with streptavidin. If non-specific binding complexes of BSA-streptavidin will be produced in the microfluidic channel, the drain current increase at a specific time, which is the signature of the complex, will be obtained. There is also a scenario that no binding between BSA and cross linkers or streptavidin when streptavidin is immobilized on the cross linkers. In such a case, only current response of streptavidin is seen.
The BSA delay measurement is shown in Fig. 2.17. Only one obvious current increment step 460 seconds after the injection of streptavidin is observed. The increase is attributed to the arrival of streptavidin. The entire transient response shows a stable current level after the drain current increase. The result clearly shows no BSA and BSA-streptavidin complexes are detected by our TFT sensor system. High specificity of the streptavidin-biotin reaction is indicated in the control BSA delay experiment. Our TFT biosensor provides a platform for analyzing protein-ligand biochemical reactions.
Fig. 2.15 Transient response of drain current with PBS prefilled at t=0 and 0.4mM BSA injected at t =60 seconds
Fig. 2.16 Scenarios of possible BSA-streptavidin interactions once the analytes reach ROI
Fig. 2.17 Current response of BSA delay experiment. Streptavidin solution was introduced in inlet 1 at t= 60 s, and BSA solution was introduced in inlet 2 at t=660s
2.4 Summary
In this chapter, an IGZO TFT biosensor integrated with a Y-type external microfluidic channel chip is demonstrated. The device is designed for evaluating the mixing reaction situation of biomaterials. Biotin-streptavidin reaction is chosen to be detected and analyzed. First, the single biomaterial experiments are discussed. The diffusion time of the single biomaterials, i.e. biotin, streptavidin, BSA, are determined.
The diffusion times are defined as the duration between the injection time and arrival time of these biomaterials. Next, time difference assumptions and experiments of the mixing reaction conditions are investigated and discussed. To further understand the reaction property, the delay experiments are also demonstrated. Finally, BSA is chosen for a control experiment to verify the non-specific binding situation. With the analysis and experiment framework, biochemical reaction properties such as the diffusion times, the excess and exhaust conditions of the reactants, and the product formation are understood.
Chapter 3 IGZO-TFT Biosensors for Investigation of Protein-Ligand Kinetics
3.1 Introduction
Investigation of biochemical kinetics
Recently, several methods exist for protein-ligand kinetics assessment. Classified by working principles, label-based measurement and label-free measurement are two categories for biosensors. Label-based detection, such as fluorescent cAMP analog [36], requires labeling molecules with fluorescence. Though the label-based methods have lower detection limit, the labeling process is complex and difficult [37]. As for label-free detection, such as surface plasmon resonance (SPR) [5], mass spectrometric (MS) [38],
Recently, several methods exist for protein-ligand kinetics assessment. Classified by working principles, label-based measurement and label-free measurement are two categories for biosensors. Label-based detection, such as fluorescent cAMP analog [36], requires labeling molecules with fluorescence. Though the label-based methods have lower detection limit, the labeling process is complex and difficult [37]. As for label-free detection, such as surface plasmon resonance (SPR) [5], mass spectrometric (MS) [38],