Design and Demonstration of Tunable Amplified Sensitivity of AlGaN

Subscriber access provided by IDAHO STATE UNIV

Article

Design and demonstration of tunable amplified sensitivity of AlGaN/GaN high electron mobility transistor (HEMT)-based biosensors in human serum Tse-Yu Tai, ANIRBAN SINHA, Indu Sarangadharan, Anil Kumar Pulikkathodi, ShinLi Wang, Geng-Yen Lee, Jen-Inn Chyi, Shu-Chu Shiesh, Gwo-Bin Lee, and Yu-Lin Wang Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b00353 • Publication Date (Web): 17 Apr 2019 Downloaded from http://pubs.acs.org on April 17, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1

Design and demonstration of tunable amplified sensitivity of AlGaN/GaN high

2

electron mobility transistor (HEMT)-based biosensors in human serum

3

Tse-Yu Taia, Anirban Sinhaa, Indu Sarangadharana, Anil Kumar Pulikkathodia, Shin-Li Wanga ,

4

Geng-Yen Leed, Jen-Inn Chyid, Shu-Chu Shieshe, Gwo-Bin Leea,b,c**, Yu-Lin Wanga,b*

5

aInstitute

of Nanoengineering and Microsystems, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C.

6 7

bDepartment

of Power Mechanical Engineering, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C.

8 9

cInstitute

of Biomedical Engineering, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C.

10 11

dDepartment

of Electrical engineering, National Central University, Zhongli District, Taoyuan City 320, Taiwan, R.O.C.

12 13 14

eDepartment

of Medical Laboratory Science and Biotechnology, National Cheng Kung University, Tainan City 701, Taiwan, R.O.C

15

*Correspondence to [email protected]

16

**Co-correspondence to [email protected]

17 18

Abstract

19

We have developed a swift and simplistic protein immunoassay using aptamer functionalized

20

AlGaN/GaN high electron mobility transistors (HEMTs). The unique design of the sensor

21

facilitates protein detection in physiological salt environment overcoming charge screening

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 23

22

effects, without requiring sample pre-processing. This study reports a tunable and amplified

23

sensitivity of solution-gated electric double layer (EDL) HEMT-based biosensors, which

24

demonstrates significantly enhanced sensitivity by designing a smaller gap between the gate

25

electrode and the detection, and by operating at higher gate voltage. Sensitivity is calculated by

26

quantifying NT-proBNP, a clinical biomarker of heart failure, in buffer and untreated human serum

27

samples. The biosensor depicts elevated sensitivity and high selectivity. Furthermore, detailed

28

investigation of the amplified sensitivity in increased ionic strength environment is conducted and

29

it is revealed that a high sensitivity of 80.54 mV/decade protein concentration can be achieved,

30

which is much higher than previously reported FET biosensors. This sensor technology

31

demonstrates immense potential in developing surface affinity sensors for clinical diagnostics.

32 33

1. Introduction:

34

Molecular diagnostics aim to analyze biomarkers such as proteins for early diagnosis of diseases

35

and risk assessment1, 2. Protein based biomarkers have been routinely used for the diagnosis of

36

various cardiovascular diseases (CVDs) which contribute to the most number of deaths in the

37

world today3, 4, 5. Among CVDs, heart failure (HF) progresses slowly and results in excessive cost

38

burden in terms of treatment and disease management. HF can be better managed or even

39

prevented if an individual’s blood-based protein biomarkers are analyzed, which provides

40

additional risk stratification beyond the conventional disease markers5. Brain natriuretic peptides

41

(BNP) have been used in clinical diagnostics for the diagnosis and prognosis of heart failure6, 7, 8.

42

The amine terminated fragment of the proBNP molecule called NT-proBNP is relatively more

43

stable natriuretic peptide with longer half-life in circulation9,

44

emerged as an important clinical marker of HF in the recent years.


10.

Therefore, NT-proBNP has

Page 3 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


45

Current clinical diagnostics for NT-proBNP are mainly optical based detection methodologies

46

such as electrochemiluminiscence (Roche cobas e 411) and enzyme immunoassay (Biomedica

47

Gruppe). They require the use of sophisticated bench-top instruments, trained laboratory personnel

48

to carry out sample pre-treatments and long turnaround times, due to which these assays can be

49

performed only in clinical settings. Therefore, the present clinical diagnostics are not accessible to

50

people widely and are highly inconvenient which pose problems such as delayed diagnosis and

51

poor prognosis. The drawbacks of current techniques can be overcome by developing point of care

52

and home care tests that assay protein biomarkers in automated fashion while retaining or

53

enhancing the sensitivity of protein detection. Field effect transistor (FET) for biomolecule sensing

54

have shown high sensitivity, low cost implementation and very short response times and are thus

55

highly desired for implementing rapid immunoassays using portable systems11, 12, 13. However,

56

complex automation is required to carry out sample pre-treatment methods such as purification,

57

dilution and desalting while testing clinical samples, due to the Debye screening or charge

58

screening effect in physiological salt environment14,

59

negatively impact the cost and convenience of FET based protein assays and sample processing

60

such as dilution affect the immunogenic reactions resulting in large variations16. More recently

61

organic transistors including electrochemical transistors and electrolyte gated organic FETs have

62

demonstrated very sensitive protein detection17, 18. However, use of extra reagents like labels to

63

effect redox reactions, add to the complexity of the system and stable, mass production of organic

64

FETs is not as mature as the semiconductor based counterparts.

65

In this work, we have developed an aptamer functionalized III-V semiconductor-based FET

66

biosensor for the rapid and highly sensitive detection of NT-proBNP in clinical serum, without the

67

requirement for additional sample pre-processing techniques. AlGaN/GaN high electron mobility

15.

The pressing demands of automation



68

transistor (HEMT) is used for transduction as it is highly sensitive, biocompatible (unlike Si) and

69

thermally stable19, 20. The unique sensing structure implemented in our HEMT biosensor facilitates

70

protein detection in high salt environment, by overcoming the limitations of charge screening

71

effects. Aptamer specific to NT-proBNP is used as the receptor which provides advantages over

72

traditional protein-based receptors such as stability, longer shelf-life and ease of surface

73

functionalization. The sensing characteristics in physiological buffer and untreated human serum

74

samples are demonstrated. The sensor demonstrates high sensitivity and selectivity, over the

75

clinical range of detection of NT-proBNP, which indicates the potential applications in clinical

76

diagnostics. Furthermore, we demonstrate a method to amplify and modulate the sensitivity of

77

protein detection by means of varying sensor design. Higher sensitivity is obtained through the

78

sensor design adopted in this work compared to traditional FET biosensors previously reported.

79

This technology has promising future in the in-vitro diagnostics industry particularly home-care

80

diagnostics owing to the short response time (5 minutes), low sample requirement (2.5-5 µL) and

81

no extensive sample pre-treatments.

82 83 84

2. Experimental: 1. AlGaN/GaN HEMT fabrication

85

The HEMT fabrication process begins with creating mesa structures on the epi-wafer

86

which is composed of 3 µm thick undoped GaN layer and 150 Å thick Al0.25Ga0.75N layer on Si.

87

The AlGaN/GaN is deposited by Metal Organic Chemical Vapor Deposition (MOCVD). The

88

device fabrication processes are described in supplementary Figure S1. Inductively coupled

89

plasma (ICP) etching is performed to create device active area in the presence of Cl2/BCl3 gases

90

at an ICP power of 300 W and RF bias of 120 W at 2 MHz. Then ohmic metal contacts are


Page 4 of 23

Page 5 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


91

deposited on the active area with the composition of Ti/Al/Ni/Au, with a thickness of

92

200/400/800/1000 Å, respectively, using electron beam evaporator (E-beam). Following this, two-

93

step rapid thermal annealing (RTA) process is performed: first step of annealing is carried out at

94

200 °C for 25 s and second step at 850 °C for 40 s in an inert environment of N2. The metal

95

interconnections and the gate electrode (which is spaced apart from the channel) consisting of 200

96

Å of Ti and 1000 Å of Au, are deposited using e-beam evaporator. The whole device is then

97

passivated using photoresist and using photolithography, the sensing region, i.e., the transistor

98

channel and gate electrode are selectively opened. The schematic illustration and real-view of the

99

device are shown in Figure 1. The miniaturized HEMT device can be embedded in a thermo-

100

curable polymer substrate and using photolithography and e-beam evaporator, metal interconnects

101

can be laid, which can then be used for measurement and sensor read-out with a portable biosensor

102

system21.

103 104 105 106 107 108

Figure 1 (a) The schematic diagram of aptamer functionalized AlGaN/GaN HEMT biosensor (b) Real view image of the HEMT chip showing the gate electrode and channel openings. 2. Aptamer selection



109

The aptamer specific to NT-proBNP was screened and selected using Systematic Evolution

110

of Ligands by Exponential Enrichment (SELEX) process, which was performed on an integrated

111

microfluidic system22. Briefly, magnetic beads surface-coated with NT-proBNP were mixed with

112

a single-stranded DNA (ssDNA) library (concentration= 1 M, Medclub Scientific Co., Ltd.,

113

Taiwan) such that the ssDNA with high affinity towards NT-proBNP could bind with the magnetic

114

beads. A magnetic field was then applied to collect the ssDNA-magnetic bead complexes,

115

following which unbound ssDNA was washed away. The high-affinity, captured ssDNA was

116

heated and released for subsequent nucleic acid amplification step using polymerase chain reaction

117

(PCR). After several repeated cycles, an aptamer sequence with high affinity towards NT-proBNP

118

was screened and selected22. The dissociation constant (Kd) of the aptamer specific to NT-proBNP

119

was measured to be 2.89 nM.

120

3. Surface functionalization

121

The surface functionalization process is schematically described in supplementary Figure

122

S2. The NT-proBNP specific aptamer is thiolated at the 5-prime side. The thiolated aptamer and

123

thiol reducing agent, Tris(2-carboxyethyl) phosphine (TCEP) are mixed in the molar ration 1:1000

124

in tris-ethylenediaminetetraacetic acid (EDTA) buffer (TE buffer). After incubation in room

125

temperature for 15 minutes, the aptamer containing solution is heated to 95 ºC and flash cooled

126

and dropped on the sensor for incubation at room temperature for 24 hours.

127

4. Protein samples

128

In the present work, two protein sample solutions, in 1X PBS and untreated clinical serum

129

have been used. Purified NT-proBNP protein obtained from Abcam (catalog #ab51403, United

130

Kingdom) is diluted to different concentrations in 1X PBS that contains 4% Bovine Serum albumin


Page 6 of 23

Page 7 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


131

(BSA) to simulate human blood serum. Clinical samples of human serum with NT-proBNP

132

received from patients are collected as per National Cheng-Kung University Hospital institutional

133

review board (IRB. No. B-ER-104–116) and under the guidance of National Tsing Hua University

134

(IRB. No. 10405HE014). The clinical samples are used for electrical measurements without

135

performing any further processing.

136

5. Measurement of sensor

137

The electrical characterization of the HEMT sensor are performed using Agilent

138

B1530/B1500A semiconductor parameter analyzer. The I-V characteristics of the sensor are

139

provided in the supplementary section in Figure S3. A DC bias of 2 V is applied as the drain

140

voltage and a short duration (100 µs on-time) pulse of 1.5 V is applied to the separated gate

141

electrode as the gate voltage. When test solution is placed on the sensor and as the gate pulse is

142

turned on, the potential drops in the solution thereby changing the channel conductivity.

143 144

6. Protein elution

145

After testing every sample containing NT-proBNP, the sensor is repeatedly washed in DI

146

water and then incubated at 40 ºC in DI water to disrupt the aptamer-protein binding and thereby

147

elute the bound and unbound proteins away from the sensor surface. The sensor is later thoroughly

148

rinsed in 1X PBS and sensor baseline is verified using electrical measurements to confirm that the

149

sensor has been regenerated.

150 151

3. Results and Discussion:

152

The device structure is depicted in Figure 1. The gold electrode is spaced apart from the

153

transistor active area and functionalized with aptamer specific to NT-proBNP (Figure 1 (a)). The



Page 8 of 23

154

top view image (Figure 1 (b)) depicts the sensing area which is composed of the openings on the

155

gate electrode (100 x 120 µm2) and the FET active area or the channel (10 x 60 µm2), which are

156

spaced apart at a distance of 65 µm. The sample solution placed on the sensor covers the two open

157

areas, simulating a liquid capacitor, with two conductive plates (the FET channel and gate

158

electrode) sandwiching a dielectric medium (test solution). Our previous works have demonstrated

159

the sensing mechanism in the separated gate FET structure23. Briefly, the gate electrode is

160

connected to a voltage source that applies a short duration gate pulse Vg (100 µs; 1.5 V) which

161

drops across the solution placed in the gap between the gate electrode and the transistor dielectric

162

(AlGaN). By using pulsed gate bias operation, heat induced FET baseline shifting commonly

163

observed in DC bias operation of FET sensors, can be avoided. Moreover, the pulsed operation

164

ensures that there are no redox currents generated and hence displays only capacitive effects. This

165

is confirmed by monitoring the gate electrode leakage current which quickly relaxes to zero after

166

pulse application, indicating the lack of resistive response due to redox currents. Therefore, the Vg

167

applied can be represented mathematically as:

168

V g  Vs  Vd

169

Where VS, Vd refer to the voltages that drop in the test liquid and the transistor dielectric,

170

respectively.

171

The pulse application re-distributes the electrical double layer (EDL) occurring at the solid/liquid

172

interfaces, i.e, the interface between the sample solution and the gate electrode and transistor

173

dielectric. The changes in the double layer at the gate electrode interface will be mirrored in the

174

double layer at the transistor dielectric interface. This generates a capacitance across the solution,

175

Cs which controls the voltage drop in the transistor dielectric. This potential that drops in the

(1)


Page 9 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


176

dielectric (Vd) eventually varies the channel drain current as the 2-dimensional electron gas (2-

177

DEG) in the interface between AlGaN and GaN which forms the transistor channel is highly

178

responsive to the changes in the surface potential. Therefore, the impedances in the sensing system,

179

in the solution and the transistor dielectric modulate the sensor output response. By assuming these

180

impedances are in series combination, we can denote the potential that drops across the dielectric

181

as: 1 j C d

Cs  Vg Cd  Cs

(2)

182

Vd 

183

where Cd, Cs and ω are capacitances across the dielectric, solution and angular frequency,

184

respectively.

185

This shows that changes to the solution capacitance will result in changes in the potential drop

186

across the dielectric, leading to changes in drain current response. The solution capacitance can

187

change under different scenarios: when the ionic strength of the test medium is varied, the change

188

in surface property due to functionalization of the gate electrode, and the modification of the

189

electrostatic interaction at the gate electrode EDL via receptor-ligand binding.

190

3.1 Protein detection in buffer: NT-proBNP tested in 1X phosphate buffered saline

191

The liquid capacitor like operation of the GaN HEMT biosensor facilitates direct protein detection

192

in high ionic strength media. This is demonstrated by testing for NT-proBNP prepared in 1X PBS

193

that contains 4% BSA. In the human plasma, albumin occurs as the most abundant protein and 4%

194

BSA is added to the buffer to simulate the physiological plasma proteome conditions. The purified

195

NT-proBNP is diluted to desired test concentrations in albumin containing buffer (0.25, 0.5, 1, 2,

1 1  j C d j C s

 Vg 



196

5, 10 ng/mL). Prior to testing the target proteins, the sensor baseline is measured with zero target

197

concentration in the albumin containing buffer solution. The sample was placed on the sensor

198

region with 5 min incubation at room temperature, after which electrical measurements were

199

conducted. The drain current response for different NT-proBNP test concentrations in albumin

200

containing buffer is shown in Figure 2 (a). When the concentration of NT-proBNP increases, the

201

drain current reduces. This is due to the changes in the electrostatic interaction in the vicinity of

202

the double layer at the electrode/solution interface, where the immobilized aptamer captures the

203

target protein from the test solution. The intense electrostatic interaction of the aptamer and protein

204

induces charge re-distributions in the EDL leading to changes in the capacitance across the solution

205

and hence the HEMT current response. In the present work, we utilize differential drain current as

206

the sensor signal rather than the absolute drain current. The differential drain current hereon

207

mentioned as the current gain or gain is defined as the difference in drain current response before

208

and after Vg application (illustrated in Figure 2 (a)). The current gain offers better stability as the

209

absolute drain current is prone to variations resulting from external sources and/or thermal effects.

210

The sensor response curve can be obtained by plotting current gain versus NT-proBNP

211

concentrations, as seen in Figure 2 (b). The sensor shows elevated sensitivity across a vast dynamic

212

range of detection from 0-10 ng/mL. Figure 2 (c) shows the gate electrode leakage current (IG)

213

measured for our sensor. When the pulse is turned on, IG peaks and then quickly relaxes to zero

214

which indicates that there are no charge transfer reactions. This eliminates the resistive component

215

and we can assume a purely capacitive model for our sensor response. In fact, the gate electrode

216

leakage current is evidence that our sensor offers non-faradaic response.


Page 10 of 23

Page 11 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


217 218

Figure 2 NT-proBNP detection in albumin containing buffer (a) Drain current vs. time graph (b)

219

Sensor calibration curve depicting current gain vs. NT-proBNP concentration (n=3) (c) Gate

220

electrode leakage current depicting the absence of charge transfer reactions.

221

3.2 Tunable sensitivity in increased ionic strength

222

The main feature of our aptamer functionalized GaN HEMT biosensor is the ability to directly

223

sense the target proteins in increased ionic strength environments without dilution or filtering. FET

224

sensor being a surface affinity type of sensor exhibits very high sensitivity as well. However,

225

traditional FET biosensor design suffers from the drawbacks of charge screening effects, to

226

overcome which deliberate test medium dilution has to be conducted. Using our unique sensor

227

design, we have been able to detect proteins in 1X PBS or in other words, have been able to detect



228

proteins beyond the Debye length. In this section, we will explore the phenomenon that facilitates

229

the detection circumventing the charge screening effects and enhanced sensitivity in high salt

230

concentration containing liquids. To illustrate the differences in transduction methodology of our

231

GaN HEMT biosensor and traditional FET biosensors, we used different gap spacings between the

232

gate electrode and transistor channel region. Different gaps in the sensing region were formed

233

using photolithography that allowed the gate electrode opening to be positioned at 65, 500, 1000,

234

3000, 5000 and 10,000 µm from the transistor channel opening (Figure 3 (a)). The sample solution

235

is placed on the sensor, covering the gate electrode and channel openings. After 5 mins of sample

236

incubation, drain current response is recorded. The test results are shown in Figure 3, which depicts

237

sensor response curves obtained for NT-proBNP detection in albumin containing buffer, for

238

different gaps (from 65 to 10,000 µm). From the results in Figure 3 (b) through (g), it can be seen

239

that larger change in current gain corresponding to NT-proBNP test concentrations are observed

240

for smaller gaps. This means that sensitivity is larger when the spacing between the gate electrode

241

and the channel is smaller.


Page 12 of 23

Page 13 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


242 243 244 245

Figure 3 (a) Sensor with different spacings between channel and gate electrode (b) through (g) NT-proBNP detection in albumin containing buffer in different gaps of 65, 500, 1000, 3000, 5000 and 10,000 µm, respectively (n=3).

246

A quantitative comparison of sensitivity can be obtained by expressing the sensor signal in terms

247

of the change in effective Vg due to protein binding with the aptamer. This is depicted in Figure 4

248

as Vg versus protein concentration plots. The method of extrapolation of effective Vg versus

249

concentration is shown in detail in supplementary Figure S4. Briefly, the current gain is expressed

250

in terms of applied Vg and when current gain versus concentration curves are obtained for different

251

gap configurations (as in Figure 3), the y axis can be converted to represent effective Vg that



252

corresponds to different NT-proBNP concentrations. The plots in Figure 4 are expressed in

253

logarithmic scale to elucidate the linearity over a wide range of protein concentrations. Figures 4

254

(a) through (f) correspond to Vg versus log concentration curves for sensing structures with gaps

255

of 65, 500, 1000, 3000, 5000 and 10,000 µm, respectively. Linear fitting of the response curve is

256

carried out to obtain the sensitivity, which is consolidated and depicted in Figure 4 (g). The

257

sensitivities obtained from response curves are 80.54, 75.54, 56.26, 33.53 and 25.38 mV/decade

258

concentration at gaps of 65, 500, 1000, 3000 and 5000 µm, respectively. At 10,000 µm gap, there

259

is no protein detection as indicated by the minimal change in current gain (as in Figure 4 (f)). Table

260

1 summarizes the sensitivity values obtained at different gaps and shows the quality of fitting used

261

to obtain the sensitivity values. The results in Figure 4 (g) demonstrate quantitatively that enhanced

262

sensitivity can be obtained with smaller gaps between the gate electrode and the channel. The

263

smaller gap in the sensing region leads to less potential drop in the solution, which effectively

264

leads to a larger drain current gain. In the design of traditional FET biosensors, the reference

265

electrode is either unbiased or given a very low bias and arbitrarily positioned at extremely far

266

away from the transistor active region24. This sensing structure is similar to our experimental

267

design of 10,000 µm gap, which generates large potential drop in the sensing region, leading to a

268

low drain current gain, which is difficult to distinguish different protein concentrations. In that

269

scenario, the potential changes induced by receptor-ligand binding on the active area is largely

270

reduced in high salt concentrations, due to the strong charge screening, which limits the detection

271

of the surface potential changes.

272

P. Bergveld had developed a model for MOSFET-based pH sensor based on Boltzmann

273

distribution and electric double layer theory to predict the sensitivity, which is not larger than 59.2

274

mV/decade25, 26. This predicted ideal sensitivity is exactly the same as the ideal sensitivity in the


Page 14 of 23

Page 15 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


275

Nernst Equation, which is frequently used in potentiometry25. This is very interesting and a

276

question arises. Is it possible to generate the sensitivity higher than the ideal sensitivity for ion-

277

selective FET? Previously our group has published lead and mercury ion-selective FET sensors,

278

which exhibit sensitivity higher than the ideal Nernst sensitivity, based on our new methodology27,

279

28.

280

based protein sensors. Here we have shown the enhancement in sensitivity, which can be as high

281

as 80 mV/decade for protein detection and it is significantly higher than the reported sensitivity in

282

the range of 10-30 mV/decade29, 30, 31, 32, 33. Our sensor design thus offers a technique of tuning and

283

amplifying the sensitivity for FET biosensors. The enhancement in sensitivity reported in this work

284

has been quantitatively correlated with the sensor design, i.e., the gap between the gate electrode

285

and FET active area (Figure 4). In the future more study in the theoretical model will be needed to

286

explain the enhancement in sensitivity with strong basis of physics. However, through this work,

287

we are able to depict that by modulating the FET sensor design, we can indeed elevate the

288

sensitivity of protein detection. Furthermore, we can obtain high sensitivity in physiological salt

289

environment without performing any pre-treatments of the test sample. By tuning the gate voltage

290

bias and the gap between the gate electrode and channel, we can enhance the sensitivity as desired

291

by the application of the sensor.

In this study, we would like to prove the enhanced sensitivity can also be achieved with FET-



Page 16 of 23

292 293 294 295

Figure 4 (a) through (f) Vg versus NT-proBNP concentration curves for different gaps of 65, 500, 1000, 3000, 5000 and 10,000 µm, respectively. (g) Consolidated Vg versus NT-proBNP concentration curves.

296 297 298

Gap

65 µm

500 µm

1000 µm


3000 µm

5000 µm

Page 17 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Sensitivity

R-square

80.54

75.54

56.26

33.53

25.38

mV/decade

mV/decade

mV/decade

mV/decade

mV/decade

0.994

0.974

0.920

0.950

0.940

299

Table 1 Sensitivity values corresponding to different gaps between channel and electrode.

300

3.3 Detection of NT-proBNP in human serum

301

The direct protein detection capabilities demonstrated in high salt containing solutions like 1X

302

PBS can be further extended to real physiological fluids such as human serum. The human serum

303

samples obtained from patients contain NT-proBNP. The concentration of NT-proBNP in serum

304

samples is determined using standard clinical diagnostics. The NT-proBNP concentrations so

305

determined used in our study are 0.2239, 0.5076, 1.1050, 2.7470, 4.6210 and 11.4970 ng/mL. The

306

assay protocols for testing serum samples remain essentially the same. The serum sample was

307

allowed to incubate on the sensor for 5 mins before the measurement of electrical response. The

308

test results are shown in Figures 5 (a) and (b). Similar to the sensor response in 1X PBS with 4%

309

BSA, the drain current decreases as the protein concentration increases. The sensor response curve

310

(Figure 5 (b)) reveals good NT-proBNP concentration dependency. This highlights an important

311

feature of GaN HEMT biosensor: selectivity towards the target protein. The aptamer is specific to

312

NT-proBNP and captures the protein from the test solution during the sample incubation period.

313

However, in human serum there are various matrix proteins that occur in concentrations several

314

orders higher in concentration than the target protein. During sample incubation, the background

315

proteins may also exert weak forces on the aptamer and/or sensor surface, but they are non-specific

316

interactions and hence relatively much weaker than specific interactions. The strong electrostatic

317

interaction between the aptamer and target protein will dominate during the short sample solution

318

incubation duration of 5 minutes, leading to charge re-distribution within the EDL and thus



319

modulating solution capacitance and eventually the sensor current signal. The electrical test results

320

in Figure 5 demonstrate that even in the existence of very high matrix protein concentration, the

321

sensor is able to deliver good target protein concentration dependent drain current response. A

322

comparison of sensitivity in buffer and serum is carried out in figures 5 (c) and (d). The gain versus

323

concentration curves obtained by testing in albumin containing buffer and serum are converted to

324

effective Vg versus log concentration curves. In this way, we can normalize the sensitivity

325

obtained from different tests. In buffer, the sensitivity is 78.31 mV/decade and in serum, it is 72.71

326

mV/decade NT-proBNP concentration. The sensitivity values are very similar, and this

327

experimentally proves that in our sensing methodology, there is minimal interference of non-

328

specific binding in sensor signal which contributes to high selectivity even in untreated clinical

329

serum samples. As seen from Figure 4, shortening the gap between the electrode and channel,

330

thereby creating less potential drop across the sample solution under test, we can amplify the

331

sensitivity and detect very low concentrations of target protein. The elevated sensitivity obtained

332

under high electric field generation can facilitate the detection of proteins in increased ionic

333

strength environments such as clinical serum, without any additional pre-processing of the test

334

sample. This is indeed an important feature of our technology which will greatly scale down the

335

cost and complexity to obtain superior sensing characteristics required for point of care/home care

336

applications. The features of high sensitivity and selectivity exhibited by the aptamer

337

functionalized HEMT biosensor are crucial in developing clinical applications for this technology.


Page 18 of 23

Page 19 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


338 339 340 341

Figure 5 NT-proBNP detection in untreated human sera (a) Drain current versus time graph (b) Sensor calibration curve depicting current gain versus NT-proBNP concentration (c) and (d) Vg versus concentration graphs for buffer and serum, respectively (n=3).

342 343

To control the test background and reduce accumulation of non-specific binding, the sensor chip

344

is washed in DI water and 1X PBS, after testing each sample. The detailed washing procedure is

345

described in the methods section. The very low salt concentration in DI water is used to disrupt

346

the receptor-ligand binding as well as other non-specific interactions on the sensor surface. After

347

washing with elution buffer, 1X PBS is used to and restore the test background to native conditions.

348

In this way, the sensor baseline characteristics are restored to initial conditions prior to each protein



349

sample test. The baseline regeneration results are depicted in supplementary Figure S5. We can

350

see that the sensor baseline can be restored between each test in albumin containing buffer and

351

serum environments. This demonstrates that the decrease in current observed in the response

352

curves in Figures 2 and 5 are indeed resultant from the specific aptamer-protein interactions and

353

not from accumulation of non-specific binding. The protein elution is an effective way to tightly

354

control and maintain similar test background for assaying each protein sample.

355

4. Conclusion:

356

In this study, we realized an aptamer functionalized AlGaN/GaN HEMT biosensor for

357

detecting NT-proBNP from human serum, with amplified and tunable sensitivity. The sensor is

358

operated under high electric field such that a liquid capacitor is generated at the sensing region

359

which modulates the FET channel conductivity with high sensitivity. Using the high field

360

modulation technique, we can eliminate the charge screening effect in high salt environment, thus

361

enabling direct detection of protein biomarkers in physiological fluids such as serum (untreated

362

clinical samples). Since highly selective aptamer is employed as opposed to conventional protein-

363

based receptor, we can obtain better stability and longer shelf-life for the biosensor, making our

364

sensor methodology highly desirable for in-vitro diagnostics. We demonstrate the sensing and

365

selectivity characteristics of GaN HEMT biosensor using purified protein samples in albumin

366

containing buffer and clinical serum samples. A detailed and systematic investigation of the

367

enhanced sensitivity of our FET biosensor is carried out. It is revealed that by optimizing the sensor

368

design (gap and applied potential), we can modulate the sensitivity. This technology can also be

369

implemented for rapid, multiplexed cardiac marker detection in clinical samples, for CVD risk

370

assessment and prevention.


Page 20 of 23

Page 21 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


371 372

Acknowledgement

373

This work was partially supported by research grants from Ministry of Science &

374

Technology (MOST 107-2218-E-007-021), (MOST 106-2218-E-007-015-MY2) and National

375

Tsing Hua University (107Q2526E1, 107Q2713E1). We thank the technical support from National

376

Nano Device Laboratories (NDL) in Hsinchu and the Center for Nanotechnology, Materials

377

science, and Microsystems (CNMM) at National Tsing Hua University.

378

Supporting Information. Additional schematics and experimental results.

379

References

380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408

1. Song, Y.; Huang, Y.-Y.; Liu, X.; Zhang, X.; Ferrari, M.; Qin, L., Point-of-care technologies for molecular diagnostics using a drop of blood. Trends in biotechnology 2014, 32 (3), 132-139. 2. Johnson, C. J.; Zhukovsky, N.; Cass, A. E. G.; Nagy, J. M., Proteomics, nanotechnology and molecular diagnostics. PROTEOMICS 2008, 8 (4), 715-730. 3. Gilstrap, L. G.; Wang, T. J., Biomarkers and Cardiovascular Risk Assessment for Primary Prevention: An Update. Clinical Chemistry 2012, 58 (1), 72-82. 4. van Holten, T. C.; Waanders, L. F.; de Groot, P. G.; Vissers, J.; Hoefer, I. E.; Pasterkamp, G.; Prins, M. W. J.; Roest, M., Circulating Biomarkers for Predicting Cardiovascular Disease Risk; a Systematic Review and Comprehensive Overview of Meta-Analyses. PLOS ONE 2013, 8 (4), e62080. 5. Dhingra, R.; Vasan, R. S., Biomarkers in cardiovascular disease: Statistical assessment and section on key novel heart failure biomarkers. Trends in Cardiovascular Medicine 2017, 27 (2), 123-133. 6. Lin, D. C. C.; Diamandis, E. P.; Januzzi, J. L.; Maisel, A.; Jaffe, A. S.; Clerico, A., Natriuretic Peptides in Heart Failure. Clinical Chemistry 2014, 60 (8), 1040-1046. 7. Sharma, A.; Felker, G. M., Brain Natriuretic Peptide Treatment and Heart Failure Prevention. Reliving the Mistakes of the Past or Charting a New Course for the Future? 2016, 4 (7), 548-550. 8. Savarese, G.; Musella, F.; D’Amore, C.; Vassallo, E.; Losco, T.; Gambardella, F.; Cecere, M.; Petraglia, L.; Pagano, G.; Fimiani, L.; Rengo, G.; Leosco, D.; Trimarco, B.; Perrone-Filardi, P., Changes of Natriuretic Peptides Predict Hospital Admissions in Patients With Chronic Heart Failure. A Meta-Analysis 2014, 2 (2), 148-158. 9. Taylor, C. J.; Roalfe, A. K.; Iles, R.; Hobbs, F. D. R., The potential role of NT-proBNP in screening for and predicting prognosis in heart failure: a survival analysis. BMJ Open 2014, 4 (4), e004675. 10. Kemperman, H.; van den Berg, M.; Kirkels, H.; de Jonge, N., B-Type Natriuretic Peptide (BNP) and N-Terminal proBNP in Patients with End-Stage Heart Failure Supported by a Left Ventricular Assist Device. Clinical Chemistry 2004, 50 (9), 1670-1672. 11. Wang, C.; Cui, X.; Li, Y.; Li, H.; Huang, L.; Bi, J.; Luo, J.; Ma, L. Q.; Zhou, W.; Cao, Y.; Wang, B.; Miao, F., A labelfree and portable graphene FET aptasensor for children blood lead detection. Scientific Reports 2016, 6, 21711. 12. Chen, J.-C.; Chou, J.-C.; Sun, T.-P.; Hsiung, S.-K., Portable urea biosensor based on the extended-gate field effect transistor. Sensors and Actuators B: Chemical 2003, 91 (1), 180-186. 13. Fernández, D. G.; Blanco, A.; Durán, A.; Jiménez-Jorquera, C.; Fuentes, O. A.-d., Portable measurement system for FET type microsensors based on PSoC microcontroller. Journal of Physics: Conference Series 2013, 421 (1), 012015.



409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461

14. Elnathan, R.; Kwiat, M.; Pevzner, A.; Engel, Y.; Burstein, L.; Khatchtourints, A.; Lichtenstein, A.; Kantaev, R.; Patolsky, F., Biorecognition Layer Engineering: Overcoming Screening Limitations of Nanowire-Based FET Devices. Nano Letters 2012, 12 (10), 5245-5254. 15. Stern, E.; Vacic, A.; Rajan, N. K.; Criscione, J. M.; Park, J.; Ilic, B. R.; Mooney, D. J.; Reed, M. A.; Fahmy, T. M., Label-free biomarker detection from whole blood. Nat Nano 2010, 5 (2), 138-142. 16. Reverberi, R.; Reverberi, L., Factors affecting the antigen-antibody reaction. Blood Transfusion 2007, 5 (4), 227-240. 17. Fu, Y.; Wang, N.; Yang, A.; Law, H. K.-w.; Li, L.; Yan, F., Highly Sensitive Detection of Protein Biomarkers with Organic Electrochemical Transistors. Advanced Materials 2017, 29 (41), 1703787. 18. Wang, N.; Yang, A.; Fu, Y.; Li, Y.; Yan, F., Functionalized Organic Thin Film Transistors for Biosensing. Accounts of Chemical Research 2019, 52 (2), 277-287. 19. Lee, H. H.; Bae, M.; Jo, S.-H.; Shin, J.-K.; Son, D. H.; Won, C.-H.; Jeong, H.-M.; Lee, J.-H.; Kang, S.-W., AlGaN/GaN High Electron Mobility Transistor-Based Biosensor for the Detection of C-Reactive Protein. Sensors (Basel, Switzerland) 2015, 15 (8), 18416-18426. 20. Huang, C.-C.; Lee, G.-Y.; Chyi, J.-I.; Cheng, H.-T.; Hsu, C.-P.; Hsu, Y.-R.; Hsu, C.-H.; Huang, Y.-F.; Sun, Y.-C.; Chen, C.-C.; Li, S.-S.; Andrew Yeh, J.; Yao, D.-J.; Ren, F.; Wang, Y.-L., AlGaN/GaN high electron mobility transistors for protein– peptide binding affinity study. Biosensors and Bioelectronics 2013, 41, 717-722. 21. Hsu, C.-P.; Chen, P.-C.; Pulikkathodi, A. K.; Hsiao, Y.-H.; Chen, C.-C.; Wang, Y.-L., A Package Technology for Miniaturized Field-Effect Transistor-Based Biosensors and the Sensor Array. ECS Journal of Solid State Science and Technology 2017, 6 (5), Q63-Q67. 22. Sinha, A.; Gopinathan, P.; Chung, Y.-D.; Lin, H.-Y.; Li, K.-H.; Ma, H.-P.; Huang, P.-C.; Shiesh, S.-C.; Lee, G.-B., An integrated microfluidic platform to perform uninterrupted SELEX cycles to screen affinity reagents specific to cardiovascular biomarkers. Biosensors and Bioelectronics 2018, 122, 104-112. 23. Chu, C.-H.; Sarangadharan, I.; Regmi, A.; Chen, Y.-W.; Hsu, C.-P.; Chang, W.-H.; Lee, G.-Y.; Chyi, J.-I.; Chen, C.C.; Shiesh, S.-C.; Lee, G.-B.; Wang, Y.-L., Beyond the Debye length in high ionic strength solution: direct protein detection with field-effect transistors (FETs) in human serum. Scientific Reports 2017, 7 (1), 5256. 24. Kaisti, M., Detection principles of biological and chemical FET sensors. Biosensors and Bioelectronics 2017, 98, 437-448. 25. Bergveld, P., ISFET, Theory and Practice. IEEE Sensor Conference Toronto 2003, 1-26. 26. Parizi, K. B.; Xu, X.; Pal, A.; Hu, X.; Wong, H. S. P., ISFET pH Sensitivity: Counter-Ions Play a Key Role. Scientific reports 2017, 7, 41305-41305. 27. Chen, Y.-T.; Sarangadharan, I.; Sukesan, R.; Hseih, C.-Y.; Lee, G.-Y.; Chyi, J.-I.; Wang, Y.-L., High-field modulated ion-selective field-effect-transistor (FET) sensors with sensitivity higher than the ideal Nernst sensitivity. Scientific Reports 2018, 8 (1), 8300. 28. Chen, Y.-T.; Hseih, C.-Y.; Sarangadharan, I.; Sukesan, R.; Lee, G.-Y.; Chyi, J.-I.; Wang, Y.-L., Beyond the Limit of Ideal Nernst Sensitivity: Ultra-High Sensitivity of Heavy Metal Ion Detection with Ion-Selective High Electron Mobility Transistors. ECS Journal of Solid State Science and Technology 2018, 7 (9), Q176-Q183. 29. Cheng, S.; Hotani, K.; Hideshima, S.; Kuroiwa, S.; Nakanishi, T.; Hashimoto, M.; Mori, Y.; Osaka, T., Field Effect Transistor Biosensor Using Antigen Binding Fragment for Detecting Tumor Marker in Human Serum. Materials 2014, 7 (4), 2490-2500. 30. Lee, D.-W.; Lee, J.; Sohn, I. Y.; Kim, B.-Y.; Son, Y. M.; Bark, H.; Jung, J.; Choi, M.; Kim, T. H.; Lee, C.; Lee, N.-E., Field-effect transistor with a chemically synthesized MoS2 sensing channel for label-free and highly sensitive electrical detection of DNA hybridization. Nano Research 2015, 8 (7), 2340-2350. 31. Hideshima, S.; Fujita, K.; Harada, Y.; Tsuna, M.; Seto, Y.; Sekiguchi, S.; Kuroiwa, S.; Nakanishi, T.; Osaka, T., Signal amplification in electrochemical detection of buckwheat allergenic protein using field effect transistor biosensor by introduction of anionic surfactant. Sensing and Bio-Sensing Research 2016, 7, 90-94. 32. Münzer, A. M.; Seo, W.; Morgan, G. J.; Michael, Z. P.; Zhao, Y.; Melzer, K.; Scarpa, G.; Star, A., Sensing Reversible Protein–Ligand Interactions with Single-Walled Carbon Nanotube Field-Effect Transistors. The Journal of Physical Chemistry C 2014, 118 (31), 17193-17199. 33. Gao, N.; Gao, T.; Yang, X.; Dai, X.; Zhou, W.; Zhang, A.; Lieber, C. M., Specific detection of biomolecules in physiological solutions using graphene transistor biosensors. Proceedings of the National Academy of Sciences 2016, 113 (51), 14633-14638.


Page 22 of 23

Page 23 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

462


For Table of Contents Only

463


Design and Demonstration of Tunable Amplified Sensitivity of AlGaN

Recommend Documents