Rectification Ratio and Tunneling Decay ... - ACS Publications

May 20, 2019 - where the electrical properties of the junctions are dominated by molecular effects .... the intensive quantities but not on the extens...
0 downloads 0 Views 11MB Size
Download PDF
Research Article www.acsami.org

Cite This: ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Rectification Ratio and Tunneling Decay Coefficient Depend on the Contact Geometry Revealed by in Situ Imaging of the Formation of EGaIn Junctions Xiaoping Chen,† Hongting Hu,† Jorge Trasobares,†,§ and Christian A. Nijhuis*,†,‡ †

Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543, Singapore Centre for Advanced 2D Materials and Graphene Research Centre, National University of Singapore, 6 Science Drive 2, Singapore 117546, Singapore

Downloaded via BUFFALO STATE on July 23, 2019 at 19:36:06 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: This paper describes how the intensive (tunneling decay coefficient β and rectification ratio R) and extensive (current density J) properties of Ag−S(CH2)n−1CH3//GaOx/EGaIn junctions (n = 10, 14, 18) and molecular diodes of the form of Ag−S(CH2)11Fc//GaOx/EGaIn depend on Ageo, the contact area between the self-assembled monolayer and the coneshaped EGaIn tip. Large junctions with Ageo ≥ 1000 μm2 are unreliable and defects, such as pinholes, dominate the charge transport characteristics. For S(CH2)11Fc SAMs, R decreases from 130 to unity with increasing Ageo due to an increase in the leakage current (the current flowing across the junction at reverse bias when the diodes block current flow). The value of β decreases from 1.00 ± 0.06 n−1 to 0.70 ± 0.03 n−1 with increasing Ageo which also indicates that large junctions suffer from defects. Small junctions with Ageo ≤ 300 μm2 are not stable due to the high surface tension of the bulk EGaIn resulting in unstable EGaIn tips. In addition, the contact area for such small junctions is dominated by the rough tip apex reducing the effective contact area and reproducibility significantly. The contact area of very large junctions is dominated by the relatively smooth side walls of the tips. Our findings show that there is an optimum range for the value of Ageo between 300−500 μm2 where the electrical properties of the junctions are dominated by molecular effects. In this range of Ageo, the value of J (defined by I/Ageo where I is the measured current) increases with Ageo until it plateaus for junctions with Ageo > 1000 μm2 in agreement with recently reported findings by the Whitesides group. In this regime reproducible measurements of J can be obtained provided Ageo is kept constant. KEYWORDS: molecular diodes, tunneling decay coefficient, EGaIn, molecular junctions, contact area

■

INTRODUCTION

the top-contacts without altering the (chemical) structure of the monolayer,1−5 but each technique has a certain set of ad- and disadvantages; most reports, however, emphasize on the advantages of a particular technique while the limitations and causes of junction failure are usually not documented. For example, the role of contact pressure has been investigated only for a few techniques (e.g., conductive probe based techniques,6 and, recently, for the EGaIn technique).7 Each type of large area

One of the major reasons that hampers molecular electronics to develop into a mature technology is that the reproducibility between laboratories and different junction platforms is notoriously poor: the measured tunneling rates across the same type of molecules (e.g., alkanthiols) differ by a factor of 108.1 The reasons behind this huge spread in currents are unclear, but how conformal the contact area between the top electrode and the self-assembled monolayer (SAM), or monolayers of covalently linked molecules, is, and the role of defects and how they scale with junction area, are usually unknown. A large variety of methods have been reported to form © 2019 American Chemical Society

Received: January 30, 2019 Accepted: May 20, 2019 Published: May 22, 2019 21018

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

the Aeff/Ageo ratio in our EGaIn junctions is close to 10−6.27 The value of Aeff for other types of junctions are unknown.

junction suffers from defects, but it is not always clear how the junction characteristics scale with junction area (see for examples refs 8−13). Junction failure is on its own an ill-defined term. Although catastrophic junction failure (e.g., a short) can be easily recognized and removed from a data set, junctions may be altered during the fabrication process in subtle ways resulting in leakage currents that are difficult to recognize.2,8,12,14−20 For these reasons, it is important to investigate how the fabrication process affects the junction properties and to elucidate for which fabrication parameters the junctions are dominated by molecular effects and not by leakage current across, for example, pinholes or other types of defects2,8,14,15,21 (which is a wellknown problem in other types of tunneling junctions).22−25 The “EGaIn technique” (EGaIn is a eutectic alloy of Ga and In) is one of the techniques that has produced consistent results between different research groups to measure currents across a large variety of systems which relies on molding a nonNewtonian liquid-metal alloy of Ga and In which spontaneously forms in air a surface layer of about 0.7 nm GaOx.26,27 This GaOx layer is inhomogeneous and amorphous in nature resulting in large uncertainties in the effective contact area (Aeff; see below).13,26−31 The GaOx layer reduces the surface tension of the bulk GaIn alloy and gives it non-Newtonian properties32,33 so that it can be shaped into cone-shape tips,13,32 or stabilized in through-holes in microfluidic networks,34,35 which then are used to form electrical contacts to SAMs. The GaOx layer itself has low resistivity and does not contribute significantly to the measured junction resistance.36,37 In a typical experiment, coneshaped EGaIn tips are made by simply pulling out a microneedle from a drop of EGaIn involving wrinkling and rupturing of the GaOx surface layer.13,32 Naturally, this process involves a certain level of “art and skill” causing variations in the measured junction properties for the same type of SAMs between different groups. Consequently, there is a general consensus between the various EGaIn techniques on the measured intensive quantities (e.g., the tunneling decay coefficient β or the rectification ratio R), but not on the extensive quantities (e.g., current density J, in A/cm2) that depend on the details of the contact area.13,28 For instance, within the various EGaIn techniques, the values of J across S(CH2)n−1CH3 SAMs reported by different groups vary by 2 orders of magnitude but the value of β is always close to 1.0 n−1.8,13,14,17,29,36,38 In a broader context, there is consensus on the intensive quantities but not on the extensive quantities across different types of molecular junctions: the measured values of J measured across alkanethiols vary by a factor of 108 yet β is around 1.0 n−1 across a large number of different types of junctions including single-molecule and large-area junctions (see ref 1 for a review). Therefore, it is important to identify the reasons for the observed discrepancies in the measured values of J. There are several reasons why the measured values of J vary across different platforms so greatly, but the major culprits seem to be differences in the nature of the molecule−electrode contacts and associated contact resistances, effective contact areas, and defects. For example, Frisbie et al. have shown that a chemisorbed contact (i.e., Au−S(CH2)nS−Au junctions) can be 10−102 times more conductive than physisorbed (i.e., Au− S(CH2)n−1CH3//Au junctions) contacts.39 Whitesides et al. reported that for EGaIn junctions the value of Aeff is 104 smaller than the geometrical contact area (Ageo) resulting in a gross underestimation of the value of J since J and the measured current I are interconverted using eq 1.29 Later, we showed that

J = I /Ageo

(1)

Each type of large area junction contains defects caused by, for instance, the roughness of the electrodes, defects in the SAMs, or contaminations that hamper the SAM formation process.8,14,15,17,21,40−42 These defects may decrease the separation between the electrodes (i.e., a pinhole), resulting in large local currents which dominate the device properties.2,8,15,16,41,43 Junctions with large values of Ageo are more likely to probe defects than junctions with small values of Ageo.2,5,8,16−18,41,43 For this reason it is desirable to minimize Ageo. On the other hand, increasingly small junctions are difficult to fabricate, and thus, the minimum value of Ageo is limited for practical reasons. In the case of EGaIn junctions, the minimum value of Ageo is limited by the high surface tension of the bulk EGaIn (624 mN/ m), resulting in stable tips only for tip curvatures of >10 μm.44,45 Whiteside and co-workers argued for this reason that coneshaped tip EGaIn junctions are not stable for junctions with Ageo < 1000 μm2, and therefore, they recommend using junctions with Ageo > 1000 μm2.13 They reached this conclusion based on the following three observations recorded from Ag−S(CH2)9CH3//GaOx/EGaIn junctions. (i) The value of J increases with increasing Ageo but then plateaus for Ageo > 1000 μm2. (ii) The contact pressure of the EGaIn tip with the SAM does not increase by “pushing” the EGaIn tip against the SAM surface to increase Ageo. (iii) The reproducibility (in terms of the Gaussian log-standard deviation σlog,G, see below) was higher (i.e., σlog,G was lower) in the plateau region for large junctions than for small junctions. Although these seem to be valid justifications to promote the use of large junctions, this recommendation goes against the general wisdom to use small junctions to minimize the probability to probe defects induced by grain boundaries, step edges, and other types of defects.2,5,8,15−18,22,41 The values of β (in n−1) and R are given by eqs 2 and 3 J = J0 e−βd = J0 10−βd /2.303 R=

|J(V )|for |J(V )|rev

(2)

(3)

where d is usually taken as the molecular length (or the number of CH2 units in case of aliphatic SAMs, n), J0 is a pre-exponential factor, and |J(V)|for and |J(V)|rev are the absolute current densities for a given voltage at opposite bias when the junction allows the current to flow (forward bias) or blocks the current (reverse bias). We elucidated how defects (mainly pinholes induced by exposed grain boundaries at which SAMs cannot pack well) in EGaIn junctions with S(CH2)n−1CH3 SAMs resulted in measured values of β ranging between 0.4 and 1.2 n−1 and a change of factor of 103 in the values of J0.8,25,40 For molecular diodes with S(CH2)11Fc SAMs, the value of R dropped from 100 to unity due to defects.15,17,42 These studies, in agreement with others,19,46 have shown that exposed grain boundaries (at which SAMs cannot pack well) are the major source of defects. Therefore, one would ideally aim for values of Aeff that are smaller than the average area of a grain to avoid sampling of exposed grain boundaries. We note that in the study by Whitesides and co-workers only Ag−S(CH2)n−1CH3//GaOx/EGaIn junctions with n = 10 were evaluated which are rather insensitive to defects.13 The 21019

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

J(V) Data Collection and Analysis Method. The EGaIn setup to fabricate junctions with cone-shaped tips has been previously described.48 We used two methods for recorded J(V) data as a function of Ageo. Method 1: using a single tip by gradually increasing Ageo by adjusting the manipulator. We first formed a small area junction and subsequently increased Ageo by adjusting the micromanipulator on which the syringe was mounted from which the EGaIn tip was suspended. Thus, a single junction was formed of which Ageo was increased from 1.2 × 102 to 4.0 × 103 μm2. For each Ageo, we measured 20 J(V) traces. This procedure was repeated three times, and the averaged values of log10 |J| and log10 R with σlog were plotted vs Ageo in the main text. In the Supporting Information we show one representative data set. Method 2: using 20 separate junctions (each junction 20 traces) with different values of Ageo. Here, we formed junctions with the desired Ageo directly, and each Ageo was obtained using a different tip. For each Ageo we recorded and analyzed the J(V) traces to determine ⟨log10 |J|⟩G, ⟨log10 R⟩G, and σlog,G following previously reported methods.48 Briefly, all values of log10 |J| for each measured bias were plotted in histograms to which a Gaussian (denoted by G) was fitted to obtain the Gaussian values of ⟨log10 |J|⟩G and σlog,G, which, in turn, were used to reconstruct the log-average J(V) curves. Optical Images and Videos. The optical images and videos were recorded using an inverted optical microscope (Nikon Eclipse Ti-E) with NIS-Elements BR 4.10.00 software (this setup has been described before30). The SAMs were formed on 20 nm thin, semitransparent template-stripped Ag substrates using an optically transparent adhesive (Norland 61). These thin Ag surfaces typically had a root-mean-square surface roughness of 1.2 ± 0.2 nm measured over an area of 1.0 × 1.0 μm2. The images and video (at 8× of their original speed) of the junction area and the junction formation process were recorded through the bottom electrode using a 60× oil immersed objective with a numerical aperture of 1.49 and a halogen lamp source to illuminate the junctions through the objective. On the microscope platform, we placed a small micromanipulator (Thorlab, 25 mm XYZ translation stage) equipped with the Hamilton syringe (10 μL, model 1701RN) loaded with EGaIn from which the EGaIn tips were suspended. A second micromanipulator (Crest innovation, S-725PLM) with a tungsten probe was used to ground the Ag surface. A USB camera (Edmund optics, EO-3112C color USB camera) equipped with a zoom ratio of 4:1 and 2.5−10× magnification zoom imaging lens (Edmund optics, VZM100DI) was mounted on a stand and used to image the junctions from the side and to monitor the movement of the tip. The junctions were illuminated with a LED light source (Shenzhen Selectech Electronics Co. Ltd.; SE-DM-500X). A Keithley (model 6340) was used to measure the J(V) and J(t) characteristics as described in the main text. Usually we prepare the EGaIn tips using a micromanipulator (Leica, Germany, cat. no. 11520137) as this manipulator allows for precise fine adjustment of tens of nanometers in the z direction needed to fabricate the tips. This micromanipulator is too heavy (∼2 kg) to mount directly on the microscope sample holder. For this reason, the tips were fabricated using the Leica micromanipulator and then transferred to the micromanipulator placed on the inverted microscope. A photo of the entire setup is provided in Figure S2. Scanning Electron Microscope Images of EGaIn Tips on Ag. A FEI Verios 460 Field Emission Scanning Electron Microscope (SEM) was used to record the images of an inverted EGaIn tip on a Ag substrate. We made a new EGaIn tip on a clean template-stripped Ag surface; the “inverted tip” standing on the Ag surface was used for imaging. The SEM images were recorded at different magnifications, and the SEM was operated at 2.00 kV and 0.10 nA in secondary electron mode. We note that it is important to handle the inverted tips carefully and avoid vibrations which could cause the tips to collapse and to make the tips small enough to ensure the whole structure is in the nonNewtonian regime and thus stable during handling and imaging. We note that these results are similar to those obtained by Barber et al.49

S(CH2)9CH3 SAM inside these junctions, however, is liquid like in character and therefore capable of self-repair,41 masking the presence of defects inside the tunneling junction as we have shown previously.14 Molecular dynamics simulations revealed that in liquid-like SAMs the molecules can readily rotate and compensate for defects that would have otherwise resulted in thin-area defects. In contrast, thick SAMs are more crystalline in nature with high packing energies, and therefore, the molecules in such SAMs have a lower degree of freedom and cannot rotate into defective regions which may result in thin area defects and associated leakage currents. By investigating the tunnelling behavior across Ag−S(CH2)n−1CH3//GaOx/EGaIn junctions with n = 2−18 (only even n) on rough and ultrasmooth Ag electrodes, we showed that thin SAMs (n < 10) yield better tunneling junctions than thick SAMs (n > 10) on rough electrodes.14 For these reasons we report here how Ageo of cone-shaped EGaIn junctions affects the measured values of J0 and β for junctions with SAMs of S(CH2)n−1CH3 with n = 10, 14, and 18 and R for junctions with SAMs of S(CH2)11Fc. By using semitransparent Ag electrodes and fabricating the junctions on an inverted microscope platform, we were able to measure the “optical Aeff”, Aeff,O, as a function of Ageo in situ with optical resolution. These measurements revealed that small coneshaped tip junctions with Ageo < 300 μm2 are not stable and that the values of R, J0, and β all vary with Ageo for large junctions with Ageo of >500 μm2. Therefore, the best trade-off between junction stability and quality is a foot print of 300 μm2 < Ageo < 500 μm2. Our experiments are important to understand the origins and justifications of the variations in the junction fabrication practices and the differences in the measured results between different laboratories that use EGaIn-based junctions. In extension, the major findings reported here also apply, at least to a certain degree, to other types of large area junctions since the effective contact areas or how the top-electrode interacts or forms contacts with the monolayer during the fabrication process are usually unknown.

■

EXPERIMENTAL SECTION

Materials. All regents were used directly as supplied unless specially stated. Deionized water was purified with an Elga Purelab option-Q system. Silica gel (high-purity grade, pore size 60 Å, 40−63 μm particle size) was obtained from Sigma-Aldrich. HS(CH2)11Fc was synthesized using the same method as reported elsewhere;47 the structure and purity were determined using 1H NMR and 13C NMR spectroscopy using a Bruker Avance 300 MHz spectrometer and gas chromatography-mass spectrometry (GC-MS). The n-alkanethiols were purchased from Sigma-Aldrich (with original purities of 96−98%) and purified using column chromatography followed by recrystallization from ethanol using previously reported methods.40 The SAM precursors were stored under an atmosphere of N2 at −50 °C. The purity of the precursor was monitored once every 2 weeks using TLC (thin layer chromatography). The Ag (purity 99.99%) was purchased from MOS Group Pte Ltd. (Singapore). Ethanol (assay 99.94% in V/V) was purchased from VWR Chemicals (France) and distilled before SAM incubation. Ag Substrates and SAM Formation. We used template-stripped Ag substrates that were obtained via a previously reported method.27 The Ag surfaces used in this study had a root-mean-square surface roughness of 0.6 ± 0.1 nm measured over an area of 5.0 × 5.0 μm2. The SAMs were formed by immersing a Ag surface immediately after template stripping in a freshly distilled ethanolic solution of 3 mM of the corresponding thiol for 3 h under N2. The Ag substrates were then taken out from the solution, rinsed with ethanol, and dried in a stream of N2.

■

RESULTS AND DISCUSSION In Situ Imaging of the EGaIn Junction Formation Process. Figure 1a shows the experimental platform that was 21020

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

Figure 1. (a) Schematic illustration of the experiment to image the junction formation process, and footprint of the EGaIn tip with the SAM in situ while measuring the electrical characteristics of the junctions. Camera 1 images the junction from the side; camera 2 images the junction via the objective through the bottom electrode. Contact area of the junction can be adjusted by lowering the tip using a micromanipulator as indicated by the red dashed arrow and light gray cone. Figure is not drawn to scale. Bottom electrode was grounded, and EGaIn electrode was biased. (b and c) Optical photographs of the junctions with two different diameters taken from the side of the junction (camera 1). Reflection of the tip in the Ag surface and an obvious wrinkle are indicated by black and white arrows. (d) Footprint of the EGaIn tip with a Ag−S(CH2)9CH3 SAM taken through the objective (camera 2). Dashed black ellipse indicates the tip apex. (e−g) SEM images of an inverted EGaIn tip with different magnifications; sample was 30° tilted around the x axis (see Supporting Information section 3). In panel e, an obvious wrinkle is indicated by the black arrow.

due to the rupture event corroborating the optical image of the SAM//EGaIn interface (Figure 1d), showing that this part of the tip does not form a good contact with the SAM. Figure 1d also shows that the EGaIn seems to form a more homogeneous contact in the area surrounding the dark feature; here the side walls of the cone-shaped tip are in contact with the SAM as the EGaIn tip is forced onto the SAM. The (broken) concentric rings in Figure 1d mirror the wrinkles on the sides of the EGaIn tip that can been seen in Figure 1e, resulting in a large reduction of the Aeff with the SAM. The magnified area of the tip (Figure 1f) shows that the roughness at the apex of the tip is indeed more than one order (∼33 times, see section S3 for more details) larger than the sides of the cones which explains why the tips form more efficient contacts with the SAM when the EGaIn tip is forced onto the SAM. In other words, this result explains why Whitesides and coworkers found that J increases with Ageo up to about Ageo = 1000 μm2 after which J scales with Ageo and why they recommend using large area junctions:13 by using large area junctions, the uncertainties in the contact area with the SAMs caused by the large roughness of the tip apex are compensated by the relatively smooth contacts of the side walls of the EGaIn tip. Although these observations explain the behavior of the junctions at the (sub)micrometer length scale, the SEM image shown in Figure 1g shows that the EGaIn tips also contains microscopic roughness that is present everywhere in the junction. This observation explains why we reported a lower value of Aeff of 10−6 (ref 27) than the value of 10−4 by Whitesides and coworkers.29 It is well known that the Aeff between two solid surfaces with microscopic roughness has a ratio of about 10−2−4;28,29,36,51,52small area cone-shaped EGaIn junctions also

used to image the formation of the contact between a SAM and a cone-shaped EGaIn top electrode. Normally, EGaIn junctions are only imaged from the side (camera 1 as indicated in Figure 1a) to record an image which is then used to determine the diameter of the junction from which then Ageo is calculated by assuming the footprint has a circular geometry. Figure 1b and 1c show two images with different values of Ageo of 1.2 × 102 and 4.0 × 103 μm2 by adjusting the diameter of the contact point by simply lowering the syringe from which the EGaIn tip was suspended (see Figure S1 and Videos S1 and S2). Considering the high roughness of the EGaIn surface and the large discrepancy between the values of Ageo and Aeff, it is important to directly measure the footprint of the EGaIn contact with the SAM. Therefore, we used an inverted microscope platform to image the junctions through the bottom electrode using an objective (camera 2) as indicated in Figure 1a (Video S3). We used 20 nm Ag and optically transparent adhesive for template stripping to ensure the bottom electrode was transparent enough (20 nm Ag has a 34.5% transparency50) to allow us to image the junction area. Figure 1d shows an example of an image of a cone-shaped EGaIn junction with a S(CH2)9CH3 SAM recorded through the bottom electrode with a Ageo of 4.65 × 103 μm2. As can be readily seen, the SAM//GaOx/EGaIn interface is inhomogeneous due to wrinkles formed during the tip fabrication process. The roughness is higher in the center part of the junction (dashed black circle) than in the outer parts as that is the location where the EGaIn ruptured resulting in a dark area in the image which indicates an area of noncontact. These noncontact areas are macroscopic, i.e., (sub)micrometer scale, as these images are diffraction limited. Figure 1e shows a wide field SEM image of the EGaIn tip which shows that the tip apex is indeed very rough 21021

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

diodes perform well for small junctions (Ageo < 1000 μm2) with values of R of 200 but not for large junctions where R dropped to unity. Hence, the leakage current is sensitive to Ageo, indicating that defects play a prominent role in large cone-shaped tip junctions. Figure 2c and 2d shows the same results but recorded with method 2 (all J(V) curves, histograms, and statistics are shown Figures S5 and S6 and Table S2). Table S2 shows that the fabrication yield and σlog,G are rather insensitive of Ageo. These observations reinforce our previous findings that the yield and σlog,G are not per se good indicators of the quality of junctions.8,14 On the basis of these results, we conclude that for junctions with 2.5 × 102 < Ageo < 1.0 × 103 μm2, the electrical characteristics of the molecular diodes are reproducible and stable and do not suffer from leakage currents, resulting in large rectification ratios. Performance of S(CH2)n−1CH3 SAMs Junctions vs Contact Area. To determine whether the values of J0 and β are sensitive to the value of Ageo, we recorded J(V) curves for junctions with S(CH2)n−1CH3, with n = 10, 14, and 18, as a function of Ageo using methods 1 and 2 described above. The J(V) curves and statistics are shown in Figures S7−S12 and Tables S3−S8. Figure 3 shows the results for the three junctions using both methods and that both methods yield indistinguishable results. For all junctions, the value of J increases with increasing Ageo, but J plateaus for Ageo > 1.0 × 103 μm2; this observation agrees with the results reported in ref 13. We notice that the data in Figure 3 shows a clear transition opposite to the data in Figure 2, which shows a more gradual change. The reason for this difference in behavior may be caused by a difference in wetting of the SAMs by the top electrode or the more liquid-like nature of the S(CH2)11Fc SAMs caused by the mismatch in diameter of the Fc head groups and the alkyl chains. Tables S6− S8 show that small junctions with Ageo ≈ 1.2 × 102 μm2 were not stable, resulting in large σlog,G. Thus, very small junctions with Ageo < 2.5 × 102 μm2 yield irreproducible junctions. As mentioned earlier, thin SAMs are less sensitive to defects than thick SAMs inside molecular junctions because thin SAMs are more dynamic and readily can self-repair defects than thick SAMs.14 This property of thin SAMs is also reflected in the data shown in Figure 3 since the values of J only change by 1.2 ± 0.1 order of magnitude for n = 10 but 1.6 ± 0.1 and 1.9 ± 0.1 orders of magnitude for n = 14 and 18, respectively. Therefore, it is important to examine how β and J0 depend on Ageo. Figure 4a shows how the values of β and ⟨log10 |J0|⟩G vary with Ageo. Here, the values of ⟨log10 |J|⟩G at +0.5 V along with 95% of confidence levels were obtained from method 2 for each value of Ageo (the results for −0.5 V are shown in Figure S13, and the values of β and J0 as a function of Ageo are summarized in Table S9). The solid lines are fits to eq 2. The value of β is constant around 1.0 n−1 for values of Ageo from 1.2 × 102 to 5.0 × 102 μm2, while log10 |J0| increases from 1.57 ± 0.05 to 2.23 ± 0.34 A/cm2. For junctions with Ageo > 5.0 × 102 μm2, β gradually decreases to 0.70 ± 0.03 n−1 and log10 |J0| to 1.48 ± 0.19 A/cm2 with Ageo. These small β values for large junctions indicate the presence of defects induced by exposed grain boundaries.8,40 On the basis of these observations we conclude that the safe region for charge transport measurements is 2.5 < Ageo < 5.0 × 102 μm2 to ensure optimal balance between junction stability while minimizing the role of defects and associated leakage currents. This range is very similar to that suggested above for the molecular diodes. The upper limit also agrees with ref 8 with EGaIn stabilized in through holes in PDMS; this upper limit value can be explained as follows. The template-stripped Ag electrodes have grains with

have macroscopic roughness causing an additional reduction of the correction factor of Aeff of 10−2 in our previous studies. Performance of Molecular Diodes vs Contact Area. As mentioned in the Introduction, the rectification behavior of S(CH2)11Fc SAMs inside EGaIn junctions is very sensitive to defects caused by, e.g., the roughness of the bottom electrode,17 purity of the SAM precursor,53 or disorder in the supramolecular structure of the SAM.15,48 As a group, these studies show that primarily the leakage currents (i.e., Jrev; eq 3) are affected by defects and the current at forward bias (Jfor; eq 3) is rather insensitive to defects.15 The advantage of exploring how R changes as a function of Ageo is that by examining R, Jrev, and Jfor as a function of Ageo we can determine whether leakage currents are important in large junctions. We used two different methods to increase Ageo by first making a small contact and subsequently increasing Ageo by simply lowering the cone-shaped EGaIn tip suspended from the syringe using a micromanipulator (method 1), i.e., here the effect of Ageo on a single junction is captured, or by making separate junctions with different Ageo (method 2). Figure 2a

Figure 2. Values of (a) log10 |J| at ±1.0 V (i.e., Jfor and Jrev) and (b) log10 R recorded from Ag−S(CH2)11Fc//GaOx/EGaIn junctions vs Ageo using method 1. Error bars represent σlog measured from 3 different junctions. Values of (c) ⟨log10|J|⟩G at ±1.0 V and (d) ⟨log10R⟩G recorded from Ag−S(CH2)11Fc//GaOx/EGaIn junctions using method 2. Error bars represent 95% confidence levels from 20 different junctions.

shows the values of Jrev and Jfor vs Ageo obtained by method 1 (Figure S4 shows the J(V) curves, and Table S1 summarizes the junction properties). Junctions with a small Ageo of 1.2 × 102 μm2 are not stable due to the high surface tension of the bulk EGaIn, resulting in large current fluctuations due to deformation of the tip; we were not able to make smaller junctions than Ageo < 1.2 × 102. For junctions with Ageo > 1.2 × 102 μm2, Jfor is constant. In sharp contrast, the value of Jrev is small for junctions Ageo < 1000 μm2 but rapidly increases for larger junctions by two orders of magnitude. Figure 2b reflects this behavior and shows that the 21022

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

Figure 3. Value of log10 |J| (determined with eq 1) of Ag−S(CH2)n−1CH3//GaOx/EGaIn junctions with n = 10, 14, and 18 as a function of Ageo using method 1 (a−c) and method 2 (d−f). n = 10 corresponds to panels a and d, n = 14 with b and e, and n = 18 with c and f. Error bar for panels a−c is the log standard deviation (σlog) from three separate data sets using method 1. For panels d−f the error bars represent the 95% confidence levels of log10 |J| from 20 different junctions using method 2.

an area Agr of 0.8−1.0 μm2.40 For Aeff/Ageo = 10−4, junctions with Ageo < 1000 μm2 have an Aeff value of 1000 μm2, the value of Aeff approaches Agr and Aeff ≈ Agr applies, resulting in a sharp decrease in the junction performance. Factoring in that small junctions can have Aeff/Ageo values as low as 10−6, Aeff ≪ Agr reduces the probability of probing defects even further. Footprint of the Junction. As mentioned above, the apex of the EGaIn tip is very rough while the sides of the tip are relatively flat (the apex is about 33 times rougher than those of the sides; see Figure 1 and section S3). Therefore, the contact geometry of small junctions will likely be dominated by the illdefined nature of the tip apex, while that of large junctions will be dominated by the relatively smooth sides of the tip. To investigate this hypothesis in more detail, we recorded optical images of the footprints of junctions of Ag−S(CH2)9CH3// GaOx/EGaIn with different Ageo using the same tip by gradually adjusting the micromanipulator. We used “ImageJ” software to enhance the contrast between the bright area of the image where the EGaIn is in physical contact with the SAM and the dark areas that indicate the locations the EGaIn is not in physical contact with the SAMs (the unprocessed images are shown in Figure S14). From these images the fraction of the contact area of EGaIn with the SAM can be directly determined to obtain the “optical Aeff”, Aeff,O, since this method only can capture

macroscopic details of the contact geometry due to the diffraction limit. Figure 5 shows the results for junctions with Aeff,O values ranging from 26.8 to 4.13 × 103 μm2. Figure 5i also shows a plot of Aeff,O/Ageo vs Ageo which indicates that Aeff,O/Ageo increases with Ageo when the junctions are small, but Aeff,O/Ageo seems to reach a plateau value for large junctions. The inset of Figure 5i shows the linear relation between the measured current, I (in A), with Aeff,O. In other words, the value of J determined with eq 4 (Jeff,O) is constant over all measured values of Ageo in this study (except for too small junctions which were not stable as mentioned earlier; see Figure S15 for the J(V) curves and a plot Jeff,O vs Ageo). These observations explain why the value of J determined by eq 1 varies with Ageo for small junctions as reported previously13 and, at least in part, the junction-to-junction fluctuations of the values of J as is common practice in the large-area junction community. These observations also explain why J is independent of Ageo for large junctions: the contact area of these junctions is dominated by SAM//EGaIn contact with EGaIn from the side walls of the tips, while the contact area of small junctions is dominated by the rough apex of the EGaIn tip. The tip apex area however does not increase with increasing Ageo, and therefore, its contribution in large junctions is insignificant. Jeff,O = I /Aeff,O 21023

(4) DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

previously13), and typically σlog,G values of 0.2−0.5 are obtained for both small and large junctions (see, for example, refs 14 and 54−59). Given the complex geometry of the junction where the area around the tip apex is essentially not in contact with the SAM but the side walls of the tip dominates the effective contact area of the SAM//EGaIn interface, why are reproducible experiments possible? Does the EGaIn tip “remember” its shape? To address these questions, we formed SAM//EGaIn contacts repeatedly while imaging the junction interface (camera 2 in Figure 1a) and recording the current−time, I(t), characteristic. Figure 6a shows the I(t) curve and that no current flowed when the EGaIn tip was retracted from the surface, but the current flow was restored once the EGaIn tip was lowered, and the SAM//EGaIn contact was formed again. Each time the junction was formed again; the current reached the same current plateau. These plateaus are labeled with corresponding panel labels; the Supporting Information includes the time lapse video (Video S3), and Figure 6b−g shows the corresponding images of the repeated footprint of the top electrode. The footprints of the EGaIn tip with the SAM are roughly the same: the positions of the tip apex, wrinkles, and contact and noncontact areas (two wrinkles are indicated by two black dashed circles) are similar in all images. These results indicate a “memory effect” of the GaOx/EGaIn tip which is caused by the non-Newtonian properties of the liquid GaOx/EGaIn alloy. These results also give new insights in the contact formation mechanism. The images show no evidence that the GaOx skin ruptures during contact formation and breaking (although the GaOx layer certainly ruptures in case a short formed as can be observed visually by the rapid alloying of the bulk EGaIn with the bottom electrode). Here we repeated the process 6 times which is typically used in an experimental setting. We have not studied for how long or how often this process can be repeated because here we only focused on experimental conditions typically used by the community. Although the foot print of the junctions did not change during repetitive formation and breaking of the junctions, the shape of the tips of large junctions did change. Figure 7a−d (snapshots taken from Video S1) shows the side view of small junctions and that the tip barely deformed after 3 cycles of junction formation and breaking. In contrast, for large Ageo Figure 7e−h (snapshots taken from Video S2), the fresh tip was significantly smaller than subsequent tips, resulting in an increase of the tip curvature. In other words, the tip deformed to increase the Ageo during contact formation. However, during the reverse process, when the tip was removed from the surface, the original shape of the fresh tip was not restored. This change in the tip curvature did not change the foot print of the junctions (Figure 6 and Video S3). This observation indicates that the tips deform without rupturing the GaOx layer during junction formation and breaking. The pressure on the SAM exerted by the EGaIn tip is small and remains constant as a function of Ageo.13 These findings can explain why the wrinkles and the original tip apex are not smoothened and remain present in the foot print of the junctions (as can be seen in Figures 5 and 6 and Video S3).

Figure 4. (a) ⟨log10|J|⟩G at +0.5 V vs n for Ag−S(CH2)n−1CH3//GaOx/ EGaIn junctions with different values of Ageo along with fits to eq 2 at 95% confidence levels. Error bars represent the 95% confidence levels. (b) β and log10|J0| vs Ageo of the corresponding junctions. Error bars represent the standard deviation of the fits shown in panel a.

To quantify the “noncontact area”, we performed the following analysis. Figure 5j shows a plot of Aeff,O/Ageo vs 1/ Ageo along with a fit to eq 5 with k = 0.91 and A0 = 3.4 × 102 μm2 which can be explained as follows. In the limit of very small junctions, the contact area is dominated by the rough tip apex and the junctions are not stable, therefore, the junction size needs to exceed a certain minimum value A0. The value of A0 extracted from the fit is close to the measured noncontact area around the tip apex of 3.2 × 102 μm2 (Figure 5h). We also determined A0 using 50 different EGaIn tips (Figure 5k; see Supporting Information section S6 for details) which gave a Gaussian average value of A0 of (3.0 ± 0.9) × 102 μm2. In the limit of very large junctions, the reduction of Ageo due to the tip apex becomes insignificant but the ratio k = Aeff,O/Ageo is smaller than 1 (for junctions with ideal contacts not suffering from roughness). In Figure 5j Aeff,O/Ageo reaches a maximum value of 0.91 (the intercept with the y axis), which indicates that even for very large junctions a significant portion of the junction is a noncontact area due to defects in the EGaIn tips (e.g., wrinkles). We also determined the value of k using 50 separate junctions (Figure 5l), resulting in a Gaussian average of k of 0.89 ± 0.06. Aeff,O /Ag eo = k(1 − A 0 /Ageo) with A 0 = Ageo − Aeff,O /k

■

(5)

CONCLUSIONS This works helps to reconcile the differences (of 2 orders of magnitude) in the observed values of J across the same SAMs measured with EGaIn tip-based junctions by different laboratories and establishes the range of junction areas, Ageo, to ensure that the junctions are dominated by molecular effects and not by leakage currents while having a good stability and

Memory Effect of the EGaIn Tip. In a typical J(V) experiment using GaOx/EGaIn tips, each tip is typically recycled to construct 5−7 junctions. Usually the junction-to-junction variation in J (determined with eq 1) of junctions with the same tip is very low provided Ageo is kept constant in the case where small area junctions are used (for large junctions J is independent of A geo as discussed earlier and reported 21024

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

Figure 5. (a−h) Images of the footprint of EGaIn on Ag−S(CH2)9CH3 by gradually increasing the Ageo. Values of Aeff,O are indicated in the upper left corner of the images. (i) Corresponding Aeff,O/Ageo vs Ageo curve. (Inset) Currents |I| vs Aeff,O. Error bars are the standard deviations σ from 20 traces. Dashed lines are visual guides. (j) Corresponding Aeff,O/Ageo vs 1/Ageo curve. Red solid line is a fit to eq 5. Vertical gray dashed line indicates 1/A0, and horizontal gray dashed line indicates k = 0.91. Histogram of A0 (k) and k (l) along with Gaussian fits (black solid line) to these histograms.

that J0, β, and R all decrease with Ageo when large junctions are used (Ageo > 500 μm2). Low values of β indicate the presence of defects in the junctions, and the value of R decreases because the currents at reverse bias (the leakage current) increases. On the basis of these observations, we conclude that the upper limit for Ageo is 500 μm2 to minimize the role of defects in junctions based on coneshaped EGaIn tips. (2) In agreement with Whitesides and co-workers, J depends on Ageo for small junctions. Therefore, it is important to ensure that from junction to junction and experiment to experiment, Ageo is kept constant and its value should always be reported. Provided the user keeps Ageo constant, small values of σlog,G of 0.2−0.5 can be readily obtained as has been demonstrated in various papers before;14,56,57,59 these small values of σlog are similar to those reported for large area junctions.13,60 Thus, both “small area” and “large area” junctions have good reproducibilities, albeit that small area junctions require careful attention to the junction area. (3) The origin of the variation of J with Ageo is caused by the rough tip apex which dominates the contact area in small junctions resulting in a correction factor for Ageo of 10−6

reproducibility. For reasons determined empirically, several groups (including us)14,56,57,59 use small area junctions of around 300−500 μm2 as smaller junctions ( 1000 μm2.8 The Whitesides group found that large EGaIn tip junctions are stable, and increasing the Ageo by lowering the EGaIn tip onto the SAM does not increase the pressure the EGaIn tip exerts on the SAM. They also observed that J is independent of Ageo for Ageo > 1000 μm2. For these reasons they promote using large areas as J is more reproducible than small area junctions and there is no risk in damaging the SAMs as no additional pressure is involved. The reasoning is perfectly valid; however, they did not report reasons to limit the upper boundary of Ageo contradicting our previous findings.8 The work reported here reaches the following 6 conclusions and recommendations. (1) By studying both intensive and extensive charge transport parameters (J0, β, and R) as a function of Ageo, we found 21025

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

Figure 6. (a) I(t) measurement of a Ag−S(CH2)9CH3//GaOx/EGaIn biased at +0.5 V with Ageo of ∼4.0× 103 μm2. After ∼30 s, the EGaIn tip was lifted from the SAM surface for ∼10 s after which the junction was formed again. This procedure was repeated six times (Video S3 shows the junction footprint played at 8× times the normal speed). Labels b−g correspond to the figure panels b−g showing the corresponding optical microscopy images of the footprints of the junction. Dashed black lines in panels b and g are guides to the eyes and highlight the area of the tip apex and one of the wrinkles that appear as concentric features.

due to macroscopic scale roughness, while the contact area in large junctions is dominated by the sides of the EGaIn tip resulting in a correction factor of 10−4. By imaging the SAM//EGaIn contact, the Aeff,O could be determined and J is independent of Aeff,O for both small and large area junctions. On the basis of this observation we conclude that differences in reported J0 values are caused by the macroscopic roughness of the junctions related to the rough tip apex. Both small and large area junctions, however, suffer from the same microscopic roughness of the GaOx/EGaIn surface which we quantified. The noncontact area caused by the tip apex (A0) is (3.0 ± 0.9) × 102 μm2, and the noncontact zone is about 10% (k = 0.89 ± 0.06) in large junctions. (4) The mechanism of contact formation involves deformation of the EGaIn tip. The EGaIn tip “remembers” its shape due to its non-Newtonian properties when contacts are made and subsequently broken. For this reason, EGaIn tips can be used to make 5−7 junctions with very similar SAM//EGaIn contact geometries (i.e., similar Aeff,O), resulting in an insignificantly small spread in J

values (provided Ageo is kept constant when small junctions are used). (5) Molecular effects dominate in small area junctions; leakage currents dominate in large area junctions. The Whitesides group claimed in a series of papers that a large number of functional groups in junctions with aliphatic SAMs did not result in a measurable effect in the electrical characteristics of the junctions.54,60−63 In sharp contrast, other groups, including Thuo, Chiechi, or Yoon, and us, reported clear molecular effects for a wide range of functional groups including (embedded) dipoles,56,64 polarizable groups,59,65 aromatic groups,18 donor− acceptor systems,66 and redox moieties.27,48 By investigating the rectifying properties of a well-established molecular diode as a function of Ageo, we found that mainly the currents at reverse bias increase diminishing the rectifying properties of the junctions. This experiment indicates that large area junctions risk being dominated by defects resulting in large leakage currents masking molecular effects. This finding could explain the lack of 21026

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

Figure 7. Optical photographs of two tips (tip 1 and tip 2) used to fabricate a small Ageo (∼2.5 × 102 μm2) and large Ageo (∼4.0 × 103 μm2) junction showing the shapes of the freshly made tips (panels a and e) and after 1, 2, or 3, cycles of junctions formation and breaking (panels b−d and f−h) as extracted from Videos S1 (for small Ageo) and S2 (for large Ageo).

■

molecular features in the large area junction experiments reported by the Whitesides group. (6) Too small junctions are unstable due to the high surface tension of EGaIn. The results reported here indicate that junctions with Ageo < 3.0 × 102 μm2 are not stable enough, resulting in a large junction to junction variation of the measured tunneling currents. The values of J0, β, and R all decrease substantially for junctions with values of Ageo of >5.0 × 102 μm2; therefore, we recommend using values of Ageo between 3.0 × 102 and 5.0 × 102 μm2 for cone-shaped EGaIn junctions to ensure an optimal balance between stability while keeping the role of defects to a minimum so that the mechanism of charge transport is dominated by molecular effects.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.9b02033.

In conclusion, this work reconciles the differences in reported tunneling rates across molecular junctions with EGaIn top electrodes and gives new insights in how and why EGaIn junctions work. In other types of large area junctions (e.g., with top contacts made of conductive polymer,1,11,20 printed metal,2 graphene/graphene oxides,3,67 or nanoparticles coated with stabilizing polymers or SAMs9,68) the value of Aeff, conformity of the top electrode with the SAM, role of macroscopic and microscopic roughness, role of wettability of the top electrode and the SAM, phase transitions18,69 and odd−even effects in SAMs,55,70 or “safe” Ageo values to minimize the role of defects are often unknown, complicating direct comparison of measured tunneling rates across different types of test beds. We hope that our efforts will encourage others to undertake similar investigations to elucidate the origins of the differences observed in both the intrinsic and the extrinsic variables between different experimental platforms.

■

Video recorded (8× faster than normal speed) from the junctions from the sides and through the bottom electrode for small Ageo (AVI) Video recorded (8× faster than normal speed) from the junctions from the sides and through the bottom electrode for large Ageo (AVI) Video recorded (8× faster than normal speed) from the junctions from the sides and through the bottom electrode (AVI) Optical images of EGaIn junctions with different contact diameters, EGaIn setup on inverted optical microscope, summary of the J(V) curves, and raw optical images of the EGaIn junctions (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Jorge Trasobares: 0000-0002-8711-1664 Christian A. Nijhuis: 0000-0003-3435-4600 Present Address

§ J.T.: IMDEA Nanociencia, Calle Faraday 9, Cantoblanco 28049, Madrid, Spain

Notes

The authors declare no competing financial interest. 21027

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces

■

(17) Yuan, L.; Jiang, L.; Thompson, D.; Nijhuis, C. A. On the Remarkable Role of Surface Topography of the Bottom Electrodes in Blocking Leakage Currents in Molecular Diodes. J. Am. Chem. Soc. 2014, 136, 6554−6557. (18) Kong, G. D.; Jin, J.; Thuo, M.; Song, H.; Joung, J. F.; Park, S.; Yoon, H. J. Elucidating the Role of Molecule−Electrode Interfacial Defects in Charge Tunneling Characteristics of Large-Area Junctions. J. Am. Chem. Soc. 2018, 140, 12303−12307. (19) Levine, I.; Weber, S. M.; Feldman, Y.; Bendikov, T.; Cohen, H.; Cahen, D.; Vilan, A. Molecular Length, Monolayer Density, and Charge Transport: Lessons from Al−AlOx/Alkyl−Phosphonate/Hg Junctions. Langmuir 2012, 28, 404−415. (20) Neuhausen, A. B.; Hosseini, A.; Sulpizio, J. A.; Chidsey, C. E. D.; Goldhaber-Gordon, D. Molecular Junctions of Self-Assembled Monolayers with Conducting Polymer Contacts. ACS Nano 2012, 6, 9920−9931. (21) Cannes, C.; Kanoufi, F.; Bard, A. J. Cyclic Voltammetry and Scanning Electrochemical Microscopy of Ferrocenemethanol at Monolayer and Bilayer-Modified Gold Electrodes. J. Electroanal. Chem. 2003, 547, 83−91. (22) Rabson, D. A.; Jönsson-Åkerman, B. J.; Romero, A. H.; Escudero, R.; Leighton, C.; Kim, S.; Schuller, I. K. Pinholes May Mimic Tunneling. J. Appl. Phys. 2001, 89, 2786−2790. (23) Miller, C. W.; Li, Z.-P.; Åkerman, J.; Schuller, I. K. Impact of Interfacial Roughness on Tunneling Conductance and Extracted Barrier Parameters. Appl. Phys. Lett. 2007, 90, 043513. (24) Chen, X.; Victora, R. H. Effect of Pinholes in Magnetic Tunnel Junctions. Appl. Phys. Lett. 2007, 91, 212104. (25) Zhang, Z.; Rabson, D. A. Diagnosis and Location of Pinhole Defects in Tunnel Junctions Using Only Electrical Measurements. J. Appl. Phys. 2004, 95, 199−203. (26) Regan, M. J.; Tostmann, H.; Pershan, P. S.; Magnussen, O. M.; DiMasi, E.; Ocko, B. M.; Deutsch, M. X-Ray Study of the Oxidation of Liquid-Gallium Surfaces. Phys. Rev. B: Condens. Matter Mater. Phys. 1997, 55, 10786−10790. (27) Chen, X.; Roemer, M.; Yuan, L.; Du, W.; Thompson, D.; del Barco, E.; Nijhuis, C. A. Molecular Diodes with Rectification Ratios Exceeding 105 Driven by Electrostatic Interactions. Nat. Nanotechnol. 2017, 12, 797−803. (28) Sangeeth, C. S. S.; Demissie, A. T.; Yuan, L.; Wang, T.; Frisbie, C. D.; Nijhuis, C. A. Comparison of DC and AC Transport in 1.5−7.5 nm Oligophenylene Imine Molecular Wires across Two Junction Platforms: Eutectic Ga−In versus Conducting Probe Atomic Force Microscope Junctions. J. Am. Chem. Soc. 2016, 138, 7305−7314. (29) Simeone, F. C.; Yoon, H. J.; Thuo, M. M.; Barber, J. R.; Smith, B.; Whitesides, G. M. Defining the Value of Injection Current and Effective Electrical Contact Area for EGaln-Based Molecular Tunneling Junctions. J. Am. Chem. Soc. 2013, 135, 18131−18144. (30) Du, W.; Wang, T.; Chu, H.-S.; Wu, L.; Liu, R.; Sun, S.; Phua, W. K.; Wang, L.; Tomczak, N.; Nijhuis, C. A. On-Chip Molecular Electronic Plasmon Sources Based on Self-Assembled Monolayer Tunnel Junctions. Nat. Photonics 2016, 10, 274−280. (31) Cademartiri, L.; Thuo, M. M.; Nijhuis, C. A.; Reus, W. F.; Tricard, S.; Barber, J. R.; Sodhi, R. N. S.; Brodersen, P.; Kim, C.; Chiechi, R. C.; Whitesides, G. M. Electrical Resistance of AgTS− S(CH2)n−1CH3//Ga2O3/EGaIn Tunneling Junctions. J. Phys. Chem. C 2012, 116, 10848−10860. (32) Chiechi, R. C.; Weiss, E. A.; Dickey, M. D.; Whitesides, G. M. Eutectic Gallium-Indium (Egain): A Moldable Liquid Metal for Electrical Characterization of Self-Assembled Monolayers. Angew. Chem., Int. Ed. 2008, 47, 142−144. (33) Xu, Q.; Oudalov, N.; Guo, Q.; Jaeger, H. M.; Brown, E. Effect of Oxidation on the Mechanical Properties of Liquid Gallium and Eutectic Gallium-Indium. Phys. Fluids 2012, 24, 063101. (34) Wan, A.; Jiang, L.; Sangeeth, C. S. S.; Nijhuis, C. A. Reversible Soft Top-Contacts to Yield Molecular Junctions with Precise and Reproducible Electrical Characteristics. Adv. Funct. Mater. 2014, 24, 4442−4456.

ACKNOWLEDGMENTS We acknowledge the Ministry of Education (MOE) for supporting this research under award no. MOE2015-T2-2134. The Prime Minister’s Office, Singapore under its Medium sized center program is also acknowledged for supporting this research. We thank Hong Zhang for extracting the effective contact area using “ImageJ”. We would like to acknowledge support by the National Research Foundation, Prime Minister’s Office, Singapore, under its Medium Sized Centre Programme.

■

REFERENCES

(1) Akkerman, H. B.; de Boer, B. Electrical Conduction Through Single Molecules and Self-Assembled Monolayers. J. Phys.: Condens. Matter 2008, 20, 013001. (2) Vilan, A.; Aswal, D.; Cahen, D. Large-Area, Ensemble Molecular Electronics: Motivation and Challenges. Chem. Rev. 2017, 117, 4248− 4286. (3) Wang, G.; Kim, Y.; Choe, M.; Kim, T.-W.; Lee, T. A New Approach for Molecular Electronic Junctions with a Multilayer Graphene Electrode. Adv. Mater. 2011, 23, 755−760. (4) Xiang, D.; Wang, X.; Jia, C.; Lee, T.; Guo, X. Molecular-Scale Electronics: From Concept to Function. Chem. Rev. 2016, 116, 4318− 4440. (5) Jeong, H.; Kim, D.; Xiang, D.; Lee, T. High-Yield Functional Molecular Electronic Devices. ACS Nano 2017, 11, 6511−6548. (6) Zhang, Y.; Qiu, X.; Gordiichuk, P.; Soni, S.; Krijger, T. L.; Herrmann, A.; Chiechi, R. C. Mechanically and Electrically Robust SelfAssembled Monolayers for Large-Area Tunneling Junctions. J. Phys. Chem. C 2017, 121, 14920−14928. (7) Dickey, M. D.; Chiechi, R. C.; Larsen, R. J.; Weiss, E. A.; Weitz, D. A.; Whitesides, G. M. Eutectic Gallium-Indium (EGaIn): A Liquid Metal Alloy for the Formation of Stable Structures in Microchannels at Room Temperature. Adv. Funct. Mater. 2008, 18, 1097−1104. (8) Jiang, L.; Sangeeth, C. S. S.; Wan, A.; Vilan, A.; Nijhuis, C. A. Defect Scaling with Contact Area in EGaIn-Based Junctions: Impact on Quality, Joule Heating, and Apparent Injection Current. J. Phys. Chem. C 2015, 119, 960−969. (9) Puebla-Hellmann, G.; Venkatesan, K.; Mayor, M.; Lörtscher, E. Metallic Nanoparticle Contacts for High-Yield, Ambient-Stable Molecular-Monolayer Devices. Nature 2018, 559, 232−235. (10) Ru, J.; Szeto, B.; Bonifas, A.; McCreery, R. L. Microfabrication and Integration of Diazonium-Based Aromatic Molecular Junctions. ACS Appl. Mater. Interfaces 2010, 2, 3693−3701. (11) Wang, G.; Yoo, H.; Na, S.-I.; Kim, T.-W.; Cho, B.; Kim, D.-Y.; Lee, T. Electrical Conduction Through Self-Assembled Monolayers in Molecular Junctions: Au/Molecules/Au Versus Au/Molecule/PEDOT:PSS/Au. Thin Solid Films 2009, 518, 824−828. (12) Van Hal, P. A.; Smits, E. C. P.; Geuns, T. C. T.; Akkerman, H. B.; De Brito, B. C.; Perissinotto, S.; Lanzani, G.; Kronemeijer, A. J.; Geskin, V.; Cornil, J.; Blom, P. W. M.; De Boer, B.; De Leeuw, D. M. Upscaling, Integration and Electrical Characterization of Molecular Junctions. Nat. Nanotechnol. 2008, 3, 749−754. (13) Rothemund, P.; Morris Bowers, C.; Suo, Z.; Whitesides, G. M. Influence of the Contact Area on the Current Density Across Molecular Tunneling Junctions Measured with Egain Top-Electrodes. Chem. Mater. 2018, 30, 129−137. (14) Jiang, L.; Sangeeth, C. S. S.; Yuan, L.; Thompson, D.; Nijhuis, C. A. One-Nanometer Thin Monolayers Remove the Deleterious Effect of Substrate Defects in Molecular Tunnel Junctions. Nano Lett. 2015, 15, 6643−6649. (15) Thompson, D.; Nijhuis, C. A. Even the Odd Numbers Help: Failure Modes of SAM-Based Tunnel Junctions Probed via Odd-Even Effects Revealed in Synchrotrons and Supercomputers. Acc. Chem. Res. 2016, 49, 2061−2069. (16) Wang, G.; Kim, T.-W.; Lee, T. Electrical Transport Characteristics Through Molecular Layers. J. Mater. Chem. 2011, 21, 18117− 18136. 21028

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029

Research Article

ACS Applied Materials & Interfaces (35) Wan, A.; Suchand Sangeeth, C. S.; Wang, L.; Yuan, L.; Jiang, L.; Nijhuis, C. A. Arrays of High Quality SAM-Based Junctions and Their Application in Molecular Diode Based Logic. Nanoscale 2015, 7, 19547−19556. (36) Sangeeth, C. S. S.; Wan, A.; Nijhuis, C. A. Equivalent Circuits of a Self-Assembled Monolayer-Based Tunnel Junction Determined by Impedance Spectroscopy. J. Am. Chem. Soc. 2014, 136, 11134−11144. (37) Suchand Sangeeth, C. S.; Wan, A.; Nijhuis, C. A. Probing the Nature and Resistance of the Molecule-Electrode Contact in SAMbased Junctions. Nanoscale 2015, 7, 12061−12067. (38) Sporrer, J.; Chen, J.; Wang, Z.; Thuo, M. M. Revealing the Nature of Molecule−Electrode Contact in Tunneling Junctions Using Raw Data Heat Maps. J. Phys. Chem. Lett. 2015, 6, 4952−4958. (39) Engelkes, V. B.; Beebe, J. M.; Frisbie, C. D. Length-Dependent Transport in Molecular Junctions Based on SAMs of Alkanethiols and Alkanedithiols: Effect of Metal Work Function and Applied Bias on Tunneling Efficiency and Contact Resistance. J. Am. Chem. Soc. 2004, 126, 14287−14296. (40) Yuan, L.; Jiang, L.; Zhang, B.; Nijhuis, C. A. Dependency of the Tunneling Decay Coefficient in Molecular Tunneling Junctions on the Topography of the Bottom Electrodes. Angew. Chem., Int. Ed. 2014, 53, 3377−3381. (41) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Self-Assembled Monolayers of Thiolates on Metals as a Form of Nanotechnology. Chem. Rev. 2005, 105, 1103−1170. (42) Jiang, L.; Yuan, L.; Cao, L.; Nijhuis, C. A. Controlling Leakage Currents: The Role of the Binding Group and Purity of the Precursors for Self-Assembled Monolayers in the Performance of Molecular Diodes. J. Am. Chem. Soc. 2014, 136, 1982−1991. (43) Weiss, E. A.; Chiechi, R. C.; Kaufman, G. K.; Kriebel, J. K.; Li, Z.; Duati, M.; Rampi, M. A.; Whitesides, G. M. Influence of Defects on the Electrical Characteristics of Mercury-Drop Junctions: Self-Assembled Monolayers of n-Alkanethiolates on Rough and Smooth Silver. J. Am. Chem. Soc. 2007, 129, 4336−4349. (44) Zrnic, D.; Swatik, D. S. On The Resistivity and Surface Tension of the Eutectic Alloy of Gallium and Indium. J. Less-Common Met. 1969, 18, 67−68. (45) Dickey, M. D. Emerging Applications of Liquid Metals Featuring Surface Oxides. ACS Appl. Mater. Interfaces 2014, 6, 18369−18379. (46) Haiss, W.; Martín, S.; Leary, E.; Zalinge, H. v.; Higgins, S. J.; Bouffier, L.; Nichols, R. J. Impact of Junction Formation Method and Surface Roughness on Single Molecule Conductance. J. Phys. Chem. C 2009, 113, 5823−5833. (47) Creager, S. E.; Rowe, G. K. Competitive Self-Assembly and Electrochemistry of Some Ferrocenyl-N-Alkanethiol Derivatives on Gold. J. Electroanal. Chem. 1994, 370, 203−211. (48) Nerngchamnong, N.; Yuan, L.; Qi, D.-C.; Li, J.; Thompson, D.; Nijhuis, C. A. The Role of van der Waals Forces in the Performance of Molecular Diodes. Nat. Nanotechnol. 2013, 8, 113−118. (49) Barber, J. R.; Yoon, H. J.; Bowers, C. M.; Thuo, M. M.; Breiten, B.; Gooding, D. M.; Whitesides, G. M. Influence of Environment on the Measurement of Rates of Charge Transport across AgTS/SAM// Ga2O3/EGaIn Junctions. Chem. Mater. 2014, 26, 3938−3947. (50) Lee, K.-S.; Kim, I.; Yeon, C. B.; Lim, J. W.; Yun, S. J.; Jabbour, G. E. Thin Metal Electrodes for Semitransparent Organic Photovoltaics. ETRI J. 2013, 35, 587−593. (51) Holm, R. Electrical Contacts, Theory and Applications, 4th ed.; Springer-Verlag: Berlin, Germany, 1967; pp 1−484. (52) Timsit, R. S. The True Area of Contact at A Liquid Metal-Solid Interface. Appl. Phys. Lett. 1982, 40, 379−381. (53) Jiang, L.; Yuan, L.; Cao, L.; Nijhuis, C. A. Controlling Leakage Currents: the Role of the Binding Group and Purity of the Precursors for Self-Assembled Monolayers in the Performance of Molecular Diodes. J. Am. Chem. Soc. 2014, 136, 1982−1991. (54) Liao, K.-C.; Yoon, H. J.; Bowers, C. M.; Simeone, F. C.; Whitesides, G. M. Replacing AgTSSCH2-R with AgTSO2C-R in EGaInBased Tunneling Junctions Does Not Significantly Change Rates of Charge Transport. Angew. Chem., Int. Ed. 2014, 53, 3889−3893.

(55) Chen, J.; Giroux, T. J.; Nguyen, Y.; Kadoma, A. A.; Chang, B. S.; VanVeller, B.; Thuo, M. M. Understanding Interface (Odd−Even) Effects in Charge Tunneling Using A Polished Egain Electrode. Phys. Chem. Chem. Phys. 2018, 20, 4864−4678. (56) Chen, J.; Gathiaka, S.; Wang, Z.; Thuo, M. Role of Molecular Dipoles in Charge Transport across Large Area Molecular Junctions Delineated Using Isomorphic Self-Assembled Monolayers. J. Phys. Chem. C 2017, 121, 23931−23938. (57) Zhang, Y.; Ye, G.; Soni, S.; Qiu, X.; Krijger, T. L.; Jonkman, H. T.; Carlotti, M.; Sauter, E.; Zharnikov, M.; Chiechi, R. C. Controlling Destructive Quantum Interference in Tunneling Junctions Comprising Self-Assembled Monolayers via Bond Topology and Functional Groups. Chem. Sci. 2018, 9, 4414−4423. (58) Kumar, S.; van Herpt, J. T.; Gengler, R. Y. N.; Feringa, B. L.; Rudolf, P.; Chiechi, R. C. Mixed Monolayers of Spiropyrans Maximize Tunneling Conductance Switching by Photoisomerization at the Molecule−Electrode Interface in EGaIn Junctions. J. Am. Chem. Soc. 2016, 138, 12519−12526. (59) Kong, G. D.; Kim, M.; Jang, H. J.; Liao, K. C.; Yoon, H. J. Influence of Halogen Substitutions on Rates of Charge Tunneling Across SAM-Based Large-Area Junctions. Phys. Chem. Chem. Phys. 2015, 17, 13804−13807. (60) Baghbanzadeh, M.; Pieters, P. F.; Yuan, L.; Collison, D.; Whitesides, G. M. The Rate of Charge Tunneling in EGaIn Junctions Is Not Sensitive to Halogen Substituents at the Self-Assembled Monolayer//Ga2O3 Interface. ACS Nano 2018, 12, 10221−10230. (61) Yoon, H. J.; Shapiro, N. D.; Park, K. M.; Thuo, M. M.; Soh, S.; Whitesides, G. M. The Rate of Charge Tunneling through SelfAssembled Monolayers Is Insensitive to Many Functional Group Substitutions. Angew. Chem., Int. Ed. 2012, 51, 4658−4661. (62) Yoon, H. J.; Bowers, C. M.; Baghbanzadeh, M.; Whitesides, G. M. The Rate of Charge Tunneling Is Insensitive to Polar Terminal Groups in Self-Assembled Monolayers in AgTSS(CH2)nM(CH2)mT//Ga2O3/ EGaIn Junctions. J. Am. Chem. Soc. 2014, 136, 16−19. (63) Thuo, M. M.; Reus, W. F.; Simeone, F. C.; Kim, C.; Schulz, M. D.; Yoon, H. J.; Whitesides, G. M. Replacing − CH2CH2− with − CONH− Does Not Significantly Change Rates of Charge Transport through AgTS-SAM//Ga2O3/EGaIn Junctions. J. Am. Chem. Soc. 2012, 134, 10876−10884. (64) Kovalchuk, A.; Abu-Husein, T.; Fracasso, D.; Egger, D. A.; Zojer, E.; Zharnikov, M.; Terfort, A.; Chiechi, R. C. Transition Voltages Respond to Synthetic Reorientation of Embedded Dipoles in SelfAssembled Monolayers. Chem. Sci. 2016, 7, 781−787. (65) Wang, D.; Fracasso, D.; Nurbawono, A.; Annadata, H. V.; Sangeeth, C. S. S.; Yuan, L.; Nijhuis, C. A. Tuning the Tunneling Rate and Dielectric Response of SAM-Based Junctions via a Single Polarizable Atom. Adv. Mater. 2015, 27, 6689−6695. (66) Souto, M.; Yuan, L.; Morales, D. C.; Jiang, L.; Ratera, I.; Nijhuis, C. A.; Veciana, J. Tuning the Rectification Ratio by Changing the Electronic Nature (Open-Shell and Closed-Shell) in Donor−Acceptor Self-Assembled Monolayers. J. Am. Chem. Soc. 2017, 139, 4262−4265. (67) Seo, S.; Min, M.; Lee, J.; Lee, T.; Choi, S.-Y.; Lee, H. SolutionProcessed Reduced Graphene Oxide Films as Electronic Contacts for Molecular Monolayer Junctions. Angew. Chem., Int. Ed. 2012, 51, 108− 112. (68) Chu, C.; Na, J.-S.; Parsons, G. N. Conductivity in Alkylamine/ Gold and Alkanethiol/Gold Molecular Junctions Measured in Molecule/Nanoparticle/Molecule Bridges and Conducting Probe Structures. J. Am. Chem. Soc. 2007, 129, 2287−2296. (69) Kong, G. D.; Kim, M.; Cho, S. J.; Yoon, H. J. Gradients of Rectification: Tuning Molecular Electronic Devices by the Controlled Use of Different-Sized Diluents in Heterogeneous Self-Assembled Monolayers. Angew. Chem., Int. Ed. 2016, 55, 10307−10311. (70) Wang, Z.; Chen, J.; Oyola-Reynoso, S.; Thuo, M. Empirical Evidence for Roughness-Dependent Limit in Observation of Odd− Even Effect in Wetting Properties of Polar Liquids on n-Alkanethiolate Self-Assembled Monolayers. Langmuir 2016, 32, 8230−8237.

21029

DOI: 10.1021/acsami.9b02033 ACS Appl. Mater. Interfaces 2019, 11, 21018−21029