Programmable ion-sensitive transistor interfaces. II. Biomolecular sensing and manipulation

Krishna Jayant, Kshitij Auluck, Mary Funke, Sharlin Anwar, Joshua B. Phelps, Philip H. Gordon, Shantanu R. Rajwade, and Edwin C. Kan

Phys. Rev. E 88, 012802 – Published 1 July 2013

Abstract

The chemoreceptive neuron metal-oxide-semiconductor transistor described in the preceding paper is further used to monitor the adsorption and interaction of DNA molecules and subsequently manipulate the adsorbed biomolecules with injected static charge. Adsorption of DNA molecules onto poly-L-lysine–coated sensing gates (SGs) modulates the floating gate (FG) potential $ψ_{O}$ , which is reflected as a threshold voltage shift measured from the control gate (CG) $V_{th_CG}$ . The asymmetric capacitive coupling between the CG and SG to the FG results in $V_{th_CG}$ amplification. The electric field in the SG oxide $E_{SG_ox}$ is fundamentally different when we drive the current readout with $V_{CG}$ and $V_{ref}$ (i.e., the potential applied to the CG and reference electrode, respectively). The $V_{CG}$ -driven readout induces a larger $E_{SG_ox}$ , leading to a larger $V_{th_CG}$ shift when DNA is present. Simulation studies indicate that the counterion screening within the DNA membrane is responsible for this effect. The DNA manipulation mechanism is enabled by tunneling electrons (program) or holes (erase) onto FGs to produce repulsive or attractive forces. Programming leads to repulsion and eventual desorption of DNA, while erasing reestablishes adsorption. We further show that injected holes or electrons prior to DNA addition either aids or disrupts the immobilization process, which can be used for addressable sensor interfaces. To further substantiate DNA manipulation, we used impedance spectroscopy with a split ac-dc technique to reveal the net interface impedance before and after charge injection.

Received 26 February 2013

DOI:https://doi.org/10.1103/PhysRevE.88.012802

Authors & Affiliations

Krishna Jayant^1,*, Kshitij Auluck¹, Mary Funke^2,†, Sharlin Anwar^3,†, Joshua B. Phelps¹, Philip H. Gordon¹, Shantanu R. Rajwade¹, and Edwin C. Kan¹

¹School of Electrical and Computer Engineering, Cornell University, Ithaca, New York 14853, USA
²Department of Chemistry, High Point University, North Carolina 27262, USA
³Department of Biomedical Engineering, City College of New York, New York 10031, USA

^*Corresponding author: kj75@cornell.edu
^†Contributed equally to this work.

Programmable ion-sensitive transistor interfaces. I. Electrochemical gating

Krishna Jayant, Kshitij Auluck, Mary Funke, Sharlin Anwar, Joshua B. Phelps, Philip H. Gordon, Shantanu R. Rajwade, and Edwin C. Kan

Phys. Rev. E 88, 012801 (2013)

Programmable ion-sensitive transistor interfaces. III. Design considerations, signal generation, and sensitivity enhancement

Krishna Jayant, Kshitij Auluck, Sergio Rodriguez, Yingqiu Cao, and Edwin C. Kan

Phys. Rev. E 89, 052817 (2014)

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 88, Iss. 1 — July 2013

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
Schematic of the CνMOS transistor with independent control and sense gates: (a) DNA immobilization on the SG with FG charge erased (hole injection), resulting in DNA diffusion towards the poly-L-lysine–coated surface and (b) DNA manipulation upon programming (electron injection).Reuse & Permissions
Figure 2
(a) Comparison between reference electrode and CG readouts during DNA immobilization and hybridization. A 10-V CG bias during $V_{ref}$ readout renders $E_{SG_ox}$ in the SG oxide, while a −10-V CG bias renders $- E_{SG_ox}$ . The CG-driven readout with $V_{ref}$ at 0.2 V shows a larger $ψ_{O}$ shift prior to hybridization mainly due to different $E_{SG_ox}$ conditions. During hybridization, however, $Δ ψ_{O}$ (∼60 mV) is only marginally different between CG and $V_{ref}$ readout. At $- E_{SG_ox}$ conditions a slight reversal and diminished $ψ_{O}$ is observed, suggesting that the underlying field affects the net charge at the interface. At such field magnitude (0.05 V/nm) DNA desorption does not occur, but the ionic screening can be perturbed. (b) Effect of electron and hole injection into the FG prior to dsDNA (24mer) addition. With electrons injected, a very small shift in $V_{th_CG}$ is observed, which for the given capacitive ratio of ∼15 implies a $ψ_{O}$ shift of approximately 10–15 mV. With hole storage the shift in $ψ_{O}$ is ∼150 mV.Reuse & Permissions
Figure 3
(a) The CνMOS with a poly-L-lsyine–coated sensing gate is exposed to buffer and subsequent tunneling operations are performed. The red (dashed) arrow indicates programming, while the black (solid) arrow indicates erase. (b) DNA strands C1 and C2 are added to the chip under the erased conditions (electron tunneled out), which results in marked $V_{th_CG}$ shifts. (c) Buffer exchange after step (b) indicates an unchanged surface state. (d) Programming (electron tunneled in) the device after step (c) indicates the SG surface state is now similar to when pure buffer was present. Subsequent buffer exchange and erasing creates a refreshed interface.Reuse & Permissions
Figure 4
(a) Simplified capacitive model representing the FG-DNA interface. (b) DNA-SG model representing the various interfaces, potentials, and fields. Here $ψ_{O}$ and $ψ_{β}$ represent the potentials at the SG interface and DNA-electrolyte interface, respectively; $E_{O}$ and $E_{β}$ are the respective fields across the SG interface and DNA electrolyte interface, respectively; and the numbers 1 and 2 represent the discontinuity in the electric field across the DNA-electrolyte interface due to permittivity differences. (c) Potential profile across the capacitive network shown in (a) for various ionic screening models within the DNA membrane. Debye Hückel (DH) screening represents the linearized Poisson-Boltzmann (PB) approximation. Notice that when ionic screening both within and outside the membrane is low, the $ψ_{O}$ shift is maximum. The nonlinear PB approximation results in a much lower shift in $ψ_{O}$ . (d) Potential profile including the partition energy barrier to account for the ion charge density within the DNA membrane. The self-energy of ions $Δ G_{m}$ is lowered in the DNA membrane represented by varying $ɛ_{eff}$ . This leads to a lower charge density within the DNA membrane and larger change in $ψ_{O}$ . The inset depicts the orientation of DNA considered in the simulation. (e) Comparison of $Δ ψ_{O}$ hybridization signals between a PB approximation with $ɛ_{eff} = 80$ and an approximate $ɛ_{eff}$ extracted for CG-driven experimental data. Experimental evidence indicates tight packing of DNA at the surface, resulting in ion exclusion and a more pronounced $ψ_{O}$ shift.Reuse & Permissions
Figure 5
(a) DNA strands D1 and D2 are added to the chip in sequence and the frequency response before and after hybridization is monitored. A clear relaxation $Z_{1}$ is observed after hybridization, indicating that dispersion mechanisms are possibly tied to the structure and stiffness of the DNA strand. Charge injection is shown to refresh the surface with a recovery of impedance. (b) Step (a) repeated for strands C1 and C2, showing the molecular weight dependence on the formation of $Z_{1}$ , which is very weak. The initial shift in $P_{1}$ is attributed to shift in resistance due to an inefficient relaxation at extremely small molecular length scales.Reuse & Permissions

Physical Review E

covering statistical, nonlinear, biological, and soft matter physics