Interface Electrostatics dictates the Electron Transport via Bio

Subscriber access provided by University of Sunderland

Organic Electronic Devices

Interface Electrostatics dictates the Electron Transport via Bio-electronic Junctions Kavita Garg, Sara Raichlin, Tatyana Bendikov, Israel Pecht, Mordechai Sheves, and David Cahen ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b16312 • Publication Date (Web): 30 Oct 2018 Downloaded from http://pubs.acs.org on November 4, 2018

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 28 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

ACS Applied Materials & Interfaces

Interface Electrostatics dictates the Electron Transport via Bio-electronic Junctions

Kavita Garg,a* Sara Raichlin,b Tatyana Bendikov,c Israel Pecht,d* Mordechai Sheves,a * David Cahenb* a Department

of Organic Chemistry; b Department of materials and interfaces; c Department of Chemical Research Support; d Department of Immunology Weizmann Institute of Science, Rehovot- 7610001, Israel

E-mail: [email protected]; [email protected]; [email protected]; [email protected] ABSTRACT Different batches of Si wafers with nominally the same specifications were found to respond differently to identical chemical surface treatments aimed at re-growing Si oxide on them. We now found that the oxides produced on different batches of wafer, do differ electrically, thereby affecting solid-state electron transport (ETp) via protein films, assembled on them. These results led to the another set of experiments, where we studied this phenomenon using two distinct chemical methods to regrow oxides on the same batch of Si wafers. We have characterized the surfaces of the regrown oxides and of monolayers of linker molecules that connect proteins with the oxides and examined ETp via ultra-thin layers of the protein bacteriorhodopsin, assembled on them. Our results illustrate the crucial role of (near) surface charges on the substrate in defining the ETp characteristics across the proteins. This is expressed most strikingly in the observed current’s temperature dependences, and we propose that these are governed by the electrostatic landscape at the electrode-protein interface rather than by intrinsic protein properties. This study’s major finding, relevant to protein bioelectronics, is that protein-electrode coupling in junctions is a decisive factor in ETp across them. Hence, coupling can create a barrier that dominate charge transport and control the transport mode across the junction. Our findings’ wider importance lies in their relevance to hybrid junctions of Si with (polyelectrolyte) biomolecules, a likely direction for future bioelectronics. A remarkable corollary of presented results is that once an electron is

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 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 28

injected into the protein, transport within the proteins is so efficient that it does not encounter a measurable barrier down to 160K. KEYWORDS: electron transport, temperature dependence, Bacteriorhodopsin, coupling, electrode-protein interface, bio-electronics.

Introduction Measurements of electron transport (ETp), i.e. electronic conductance in the solid-state, allow assessing the efficiency of electronic charge flow within, across and out of proteins in a way that differs markedly from the more common electrochemical or optical spectroscopic measurements of electron transfer (ET).1–3 Over the past decade, we have studied ETp across proteins that are held in a solid-state junction between two electrodes. One of the electrodes is a (conducting) substrate onto which the protein is assembled in a monolayer, often by an ionic4 or covalent bond.5 The other electrode is deposited in a non-destructive way onto the protein, to preserve their native functional conformation.5,6 For the latter, we primarily employ Au pads or rods, or Hg, while others also employed eutectic In-Ga (covered with some Ga oxide),7 to measure reproducibly I-V curves. We use mostly low (< 0.5 V) bias to stay close to equilibrium and avoid changes in protein conformation or even proteins damage. Highly doped single crystal Si wafers are a convenient and commonly used substrate. Such wafers are among the most reproducible commercially available materials in terms of spatial dimensions, mechanical, optical and electronic specifications (over narrow ranges) and controlled smoothness (down to 0.2 nm over several μm2 areas). Due to their electronic properties they can also be suitable, if highly doped (~ ≥ 1018 cm-3), as conductive substrates, and can then function as contact electrode.8–11 As Si readily oxidizes in air, tight control over the surface chemistry is required for its use in molecular electronics,12–15 where current passes perpendicularly through the wafer, from top to bottom surface, as is the case in our work.16 We note that such direction of current flow is not the usual way in which Si crystals are used in electronics, because mostly the oxide serves as an insulator between gate and channel in the much more common transistor configuration.17–20 The proteins that have been studied include globular ones, such as azurin (Az), Cytochrome C, human and bovine serum albumin (HSA and BSA, resp.), and membranal ones such as bacteriorhodopsin (bR) and halorhodopsin (phR). Some proteins like bR,4,5,21 BSA5 and HSA,22


2

Page 3 of 28 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


showed temperature-dependent ETp with an activation energy (Ea) >200 meV, whereas others exhibited temperature-independent currents (Ea < 90 meV), all at small (50-100 mV) bias voltages, where we take Ea > 3kT (~75 meV) as minimum to indicate (significant) temperature dependence. In the present work we use Bacteriorhodopsin. Junctions based on this protein have shown very significant temperature dependence of the ETp. Bacteriorhodopsin produced consistently highquality layers once octylthioglucoside (OTG) vesicles (bR-OTG) formed by treatment of native purple membranes this detergent were used.21 We report here that temperature-dependence of the currents measured via Bacteriorhodopsin junctions originates in the electrostatic landscape of the electrode-protein interfaces, and is unlikely to reflect intrinsic protein properties, at least at low bias (< |1| V). The central finding which has implications beyond protein electronics, as it can be relevant for junctions with any polyelectrolyte, is that this re-grown silicon oxide can induce changes, depending on oxide quality and thickness,23,24 in the I-V characteristics of the junction, formed by applying a second contact to the exposed side of the protein layer. Results and discussions: The results of two sets of experiments are presented in this report: In the first set we used different batches of Si wafers with nominally the same specifications, that behave differently to the same chemical treatment, leading to different transport mechanisms through biomolecular devices, made with these wafers. The experimental results of this part are presented in detail in the Supporting Information. These results led to the second set of experiments. There we used identical wafers, to which we applied different chemical surface treatments to grow the ultra-thin Si-oxide film onto the Si wafer, viz., piranha and nitric acid etches. The results of that second set are reported below. Different batches of wafer with oxide grown by same chemical method We examined ETp via p-Si (100) wafers with the same specifications in terms of dopant, resistivity range, polish and thickness, from the same supplier (and foundry). To be able to use highly doped Silicon as substrate and as one of the electrodes, uncontrolled oxide formation during protein deposition (which has to be from aqueous solution) has to be minimized. This is achieved by first removing the native Si oxide using diluted HF solution and then carefully regrow ultrathin with


3


Page 4 of 28

piranha solution (1 nm) Si oxide.25 We noticed that for one of the Si batches (I) growing 1 nm oxide required 5 seconds, whereas for another batch, II, 180 seconds were required to grow the same thickness. The work function measured by UPS of Batch (I) with regrown silicon oxide was 140 meV higher than the analogous sample with Batch (II), which indicates that the batch (I) oxide surfaces are less positive than batch (II) ones. The junctions were made by depositing OTG-bR monolayers electrostatically using APTMS (3AminoPropyl)TriMethoxysilane) linker on SiO2 and top electrode as Au-LOFO (lift-off-float-on)). We found that the differences between the batches paralleled different behavior of currents as function of temperature in OTG-bR junctions: with batch I, currents across the junction made with these wafers were temperature-dependent, whereas for batch II they were temperature-independent (Fig. S3). The detailed description of the differences between the batches of Si is given in the supporting Information (Section 1). We found this behavior for various proteins like bR, Myglobin, HSA, PhR, cryptochrome and Cu-depleted Azurin. This result became the seed of the present study, as the wafers, which the whole scientific community uses as standard ones, may not be standard for molecular electronic measurements. Slight differences between the wafers, can cause significant differences after chemical treatments, leading to pronounced changes in electrical behavior. Different oxide growth procedures on same batch wafers Controlled experiments with different batches of wafers were impractical (more batches were needed to get statistically significant differences than we could afford). Therefore, we used two different procedures to grow ultra-thin (~ 1 nm) silicon oxide on 4 types of Si wafers (each type from same batch, viz. p- and n-Si (100) and p- and n-Si (111), using the piranha (H2SO4+H2O2 (3:1)) or the NAOS method (Nitric Acid Oxidation of Silicon; 61% HNO3). The native oxide was removed prior to controlled growth of oxide. In case of Si (100) we use 2% HF and for Si (111) we use NH4F to remove native oxide, to get atomically smooth interface. There are no literature reports on growing oxides on Si (111) using piranha. We found from our results indicating that piranha treatment forms a non-uniform, unstable and irreproducible oxide on both n- and p-Si (111). Hence, we used only NAOS for oxide re-growth on Si (111), but both methods for doing so on Si (100). NAOS-grown oxides on Si (100) are slightly thicker (1.30.1 nm) than piranha-grown ones (1.10.1 nm), as deduced from ellipsometry measurements.


4

Page 5 of 28 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


APTMS is used as a linker for electrostatic attachment of the negatively charged OTG-bR to the Si oxide. OTG-bR is a vesicular bR suspension, prepared, as mentioned above, by a mild treatment of native bR with OTG detergent, yielding vesicles, made up of the protein and a mixture of the original lipids and OTG. After oxide formation an APTMS monolayer was deposited by 20 min sonication of the wafer in the APTMS solution,26 followed by 40 min incubation with OTG-bR solution and then 3h in DI water.27 All the samples showed similar morphology of OTG-bR monolayers (Fig. 1.) by AFM. The thickness of the N2-dried OTG-bR layer was 7-7.5 nm as deduced from ellipsometry and 9-9.5 nm from AFM (by scratching Fig. 2.). This thickness exceeds the largest dimension of bR, as derived from its known structure (~5 nm), and indicates that the vesicles are transformed into protein bi-layers upon drying (in earlier reports we did not enter into this issue of mono- vs. bilayer; in the SI, Section 2, we discuss bilayer formation; its implications for ETp across OTG-bR films will be reported and discussed in a forthcoming publication). Junction formation and current measurements Junctions for I-V measurements were made by depositing 6-8 Au pads by LOFO on OTG-bR bilayers, as top contacts and by contacting backside the Si wafer with Ag paste. All junctions, made with Si (100) (n & p type; piranha- & NAOS-treated) showed similar I-V curves. All junctions with Si (111) wafers (n & p type, NAOS treated) showed ~3 orders of magnitude lower currents than those with Si (100) ones. These results indicate that the regrown oxide film on Si (111) is more insulating and/or the protein coupling to it is poorer (more resistive) than with Si (100) (Fig. 3). Temperature-dependent I-V measurements between 160-340K showed that junctions with NAOSformed oxides on p-Si (100) and on p- and n-Si (111) all show temperature-dependent currents with 240-500 meV activation energies (Ea) (Fig. 4), in rough agreement with our earlier results.4,5 However, piranha-treated p-Si (100) wafers and, NAOS- and piranha-treated n-Si (100) wafers show temperature-independent I-V. All these results indicate that the temperature dependence of currents across these junctions depends on the oxide, grown on the Si substrate. Analysis of differences between NAOS and piranha-grown oxides.


5


As suggested by Lucovsky et al.

28,29

Page 6 of 28

and others30,31 both sample preparation and orientation are

critical for defining the nature of the SiO2/Si interface, which in turn dictates the electrical properties of the interface+oxide.28,29 Because earlier work on ultra-thin Si oxides on Si focused on as-grown and annealed oxides, prepared thermally or by plasma treatment,28,29,32 we first analyze the SiO2/Si interface, formed by Piranha and NAOS treatments, for the highly doped (100) and (111) Si wafers that we use here. AFM

AFM analysis showed that the oxide surfaces formed by the piranha method are

rougher than those grown by the NAOS treatment (Fig. 5). This is consistent with earlier reports that piranha-formed oxide layers are rough, due to decomposition of H2O2 on the surface.33 Differences in roughness between the oxides grown by both the treatments on the different wafers are given in Table 1. All NAOS grown oxides are found to be smoother than piranha ones and nSi (100) were rougher than those on p-Si (100) (fig. 5 and Table 1). Correlating the data in Table 1 with temperature dependence data implies that the Si oxide surfaces of junctions that show temperature-dependent ETp are smoother that those that show temperature-independent ETp.31 Different surface roughness can lead to different surface coverage and tilt of the APTMS linker molecules, which can induce work function, WF, changes.34 Work Function

The WF of the surfaces of the wafers onto which oxides were regrown by

different treatments were obtained from the contact potential difference (CPD), measured by Kelvin probe, (in air) and by UPS (in UHV) (Table 2). We used for these measurements oxide surfaces which were stabilized with the APTMS monolayer because it stabilizes the Si oxide surfaces in ambient (vide infra), and allows more reproducible and reliable CPD measurements and comparison with UPS. The CPD measurements indicate that the NAOS-treated Si (100) wafers have a higher (deeper, i.e., farther from the vacuum level) WF than the piranha-treated ones (by ~0.06 eV) (fig. 6 and Table 2), n-Si (100) wafers have also lower WF (by ~ 0.015 eV) than that of p-Si (100) and of p- or n-Si (111) (~ 4.1 eV). The WF derived from the CPD is higher by 0.3-0.5 eV than that from UPS measurements (Figure 6), but trends with these two methods are similar. The differences in UPS and CPD values can be due to measurement conditions, i.e., UPS measurements are done in (ultra-high vacuum) UHV, while our CPD measurements were done in ambient, where polar oxidizing species can cause a higher WF.† All the CPD- and UPS-derived


6

Page 7 of 28 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


WF values are given in Table 2. If we correlate the roughness (Table 1) and WF (Table 2) values measured by CPD, we see that smooth surfaces (roughness < 0.3 nm) have higher WF (> 4.1 eV), and rough surfaces (roughness > 0.3 nm) lower WF (< 4.0 eV). Because a lower work function implies more positive (or less negative) surface charges, we used XPS to look for a possible correlation between charges on the APTMS-stabilized Si oxide surfaces and the WF data. XPS

XPS of N 1s signal (from the terminal, exposed amine group on the APTMS) shows

two types of N, a charged and a neutral one (fig. 7), which fit NH3+ and NH2 (r = NH3+/NH2 ratio).  For n-Si (100) the r is >1.7 after both the treatments, indicating that these (near)surfaces are highly positively charged.  For p-Si (100) after Piranha treatment r = 1, while for NAOS-treated wafers NH3+/NH2 = 0.5, i.e., the surface is 2× more positively charged after piranha treated ones.  NAOS-treated Si (111) wafers all show r ~ 0.8, consistent with that these are weakly charged. Hence, the smaller WF of piranha-treated surfaces than of NAOS ones can be explained by the piranha surfaces bearing more positive charges than NAOS ones, i.e., the NAOS treatment results in smoother SiO2 surfaces with less positive surface charge than piranha-treated ones (for Si (100)). Thus, we find a correlation between surface morphology and surface charge, which, we argue, in turn influences ETp. Previous studies suggest that the higher the SiO2 roughness at an interface, the smaller the tunneling barrier.31 Origin of charges on/in the surface APTMS layer The positive charges at Si/Si oxide interface or/and charges within the oxide itself are often created during the fabrication process. These are localized charges at the interface, created due to trapping to electrons or holes in interface states. These interface states are located at or very close to the Si/SiO2 interface with energies in the bandgap of the semiconductor.35,36 APTMS is a neutral polar molecule that can act as a bipolar one upon binding with a charged substrate. In our case, the substrate is positively charged because SiOx has occluded positive charges. One of the methoxy groups of the trimethoxy Silane (of the APTMS) reacts chemically with -OH groups on the SiOx surface, to form Si-O-Si links, leaving dimethoxy silane at that end of (of what was originally APTMS) and NH2 at the other end. Under ambient conditions, some of the methoxy groups convert to hydroxyl ones (-OCH3 + H2O -OH + CH3OH).37 To neutralize


7


Page 8 of 28

the occluded charge in SiOx, these hydroxy groups will transfer their protons to NH2 groups to form NH3+.37,38 So, the higher the density of positive charges on/in the substrate, the higher the density of NH3+ species. Based on Si 2p XPS data, as reported by Kobayashi et al., these positive charges are located at/near the SiO2 interface.39 For our APTMS-stabilized silicon oxide substrates the Si and SiO2 peak components were well-resolved (see SI, Section 3.).40 Percentage of Si+, Si2+ and Si3+ with respect to Si0 were calculated from the XPS data. The relative concentration of each of the Si+ and Si 3+ species on NAOS-treated Si wafers is less than on Piranha-treated ones (figure S6 and Table 3) except for n-Si (100). Moreover, the oxides on Si (111) (p- & n-) show less Si+ and Si 3+ states than oxides on p-Si (100), this can be due to the fact that SiO2 /Si(111) interface is atomically sharp and in fact comprises a single monoatomic layer, which contains Si atoms in the 4 possible positive oxidation states.32 This justifies the lesser charges on APTMS for Si (111) (r~ 0.8) than Si (100) (r~1-1.7). Kobayashi et al. suggested that interface state density strongly depends on the density of the Si+ and Si3+ species, as these species can become interface states when the Si—O bond of these species is ruptured (at the interface).39,41 It was further suggested that oxides formed by HNO3 (NAOS) have low leakage currents, and have lower interface density than oxides from piranha treatment and are, thus, better dielectrics.42–44 All these results are consistent with our proposal that the smoother oxide surfaces possess lower positive charge, and show larger temperature dependence. Let’s understand how these charges affect electron transport. Since, proteins are polyelectrolytes with many charges, and that for membrane proteins, like bR, the two accessible surfaces have significant negative charge (the ones that are exposed to the cytoplasm and extracellular environment, which are also the exposed protein surfaces in the OTGbR vesicles), they will adsorb so as to neutralize positive charges on the surface, to minimize the system’s energy. Let’s consider individual cases of temperature dependence and independence: In case of a more positively charged surface, the negative charge of the protein can be completely neutralized, and the junction (Si/SiO2/APTMS/ OTG-bR/Au) shows barrier-less transport i.e., temperature independent currents. If the surface is less positively charged, the negative charge on the OTG-bR can be neutralized only partially. In addition to an electrostatic barrier between the Si/Si oxide and APTMS/protein interfaces, excess charges can induce a space charge layer, which


8

Page 9 of 28 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


even in our highly doped Si can easily be some 10 nm,45 The latter scenario may explain part or all of what we measure with the rather smooth NAOS-treated Si (111), but the situation with the rougher Si (100) surfaces is likely more complex. In any case, each of these factors can create a barrier that will lead to temperature-dependent I-V curves. This hypothesis of charge distribution (neglecting the space charge in the Si) is illustrated pictorially in figure 8. Experimental: Oxide formation: The original batches of Si wafers that led to this study were all from Virginia wafers; these were p-Si(100) with specified resistivities of (batch (1)

Interface Electrostatics dictates the Electron Transport via Bio

Recommend Documents