Extraction Of Contact Resistance Research Articles

A Specific Contact Resistivity Extraction Scheme With Strong Variation Immunity Customized for Thin-Film Semiconductors: Bridge Transmission Line Method

A Specific Contact Resistivity Extraction Scheme With Strong Variation Immunity Customized for Thin-Film Semiconductors: Bridge Transmission Line Method

Extraction of contact resistivity for a partial contact solar cell

Partial contacts are used in solar cells and other electronic devices. The contact resistivity and contact resistance of these partial contacts play a significant role in the performance optimization of these devices. In this work, we propose a strategy for accurate extraction of the partial contact resistivity (ρPC) of a solar cell with Ohmic (or linear) and Schottky (or non-linear) contacts. We demonstrate how the Cox–Strack method can accurately estimate the ρPC of a solar cell using Sentaurus TCAD-based simulations. Our simulation predicts that for linear contact solar cells, ρPC is approximately constant when the contact fraction is varied, whereas for non-linear contact, ρPC decreases with contact fraction as opposed to the previously reported works. The experimentally reported data is also shown to support the claim for the linear contacts further. Despite the larger full-contact resistivity of non-linear contacts, the contact resistivity and fill factor (FF) of both linear and non-linear partial contact solar cell is almost similar to ∼3.5−5mΩcm−2 and ∼ 75% respectively at a 0.3% contact fraction. This is attributed to the reduction in contact resistivity for non-linear contacts with contact fraction. We finally show the generality of this study for any contact material system. This work can be applied to any device with partial contacts.

A Universal Method for Extracting and Quantitatively Analyzing Bias‐Dependent Contact Resistance in Carbon‐Nanotube Thin‐Film Transistors

AbstractA single‐device method is reported for extracting gate‐ and/or drain‐voltage‐dependent contact resistance of thin‐film transistors (TFTs). An extended transition‐voltage method is proposed and verified by experiments of all‐carbon‐nanotube thin‐film transistors (ACNT‐TFTs), which can extract gate‐ and/or drain‐voltage‐dependent contact resistance at source and drain independently. By measuring the output and transfer characteristics of a single‐device and extracting the basic parameters with the aid of mature Y‐function method, the contact resistance can be calculated directly. The results show that although a slight Schottky contact behavior is exhibited at very small drain voltages, good electrical contact characteristics can still be obtained in ACNT‐TFTs, exhibiting quasi‐Ohmic contacts. Compared with the existing single‐device methods, this method is suitable for both Ohmic and Schottky contact scenarios without requiring a complex iteration process, which greatly improves the universality and efficiency of the contact resistance extraction. Besides, this method reveals the physical essence of the complex interface contacts and enables researchers to quantitatively analyze the contact performance, not only for network carbon nanotube TFTs but also for the other emerging transistors.

Accurate parametrization revealing an extremely low disorder in polymer field-effect transistors

In this paper, a robust self-consistent parameter extraction method is applied to high-performance p-type printed polymer field-effect transistors. Simultaneous extraction of contact resistance and intrinsic channel mobility with their full gate-voltage dependence is achieved through an analytically reinforced transmission-line method. The proposed method yields a minimum width-normalized contact resistance of 12 kΩ cm and a maximum hole mobility of 1.9 cm2 V−1 s−1. Moreover, the gate-voltage-dependent mobility is interpreted in the framework of trap-and-release transport through double-exponential density of states, unveiling a disorder energy near the transport orbital as low as 29 meV.

Extraction of intrinsic contact resistance in organic thin-film transistors with single channel length and high capacitance density

Accurate characterization of the intrinsic contact resistance is important for the optimization of organic thin-film transistors (OTFTs), especially in devices with high gate capacitance density. This study reports on a method for the extraction of the intrinsic, or gate-voltage independent, contact resistance from OTFTs. In contrast to the transfer length method (TLM), this method can be applied to devices with single geometry. The method is an extension of the previously reported Y-function method and combines the model proposed by Luan and Neudeck to analyze the effect of the gate voltage on the contact resistance. This method is first applied to OTFTs with a bottom-gate top-contact geometry having varying gate capacitance density values. The intrinsic contact resistance is found to decrease with increased gate capacitance density with values ranging between 4 kΩ cm and under 1 kΩ cm when the gate capacitance is changed from 36.6 to 231.7 nF/cm2, respectively. These values are in good agreement with those obtained using TLM experiments. This method is also shown to be applicable to OTFTs with top-gate bottom-contact geometry.

Rapid MOSFET Contact Resistance Extraction From Circuit Using SPICE-Augmented Machine Learning Without Feature Extraction

It is desirable to monitor the degradation of integrated circuits (ICs) or perform their failure analysis through their electrical characteristics [such as the voltage-transfer characteristic (VTC), of an inverter]. Such a method is nondestructive, low-cost, and can be applied to a large number of samples. Machine learning is naturally an excellent tool to perform this task. However, it is very expensive, in terms of time and cost, to generate enough experimental data with well-controlled defects to train a reliable machine. Moreover, IC defect signatures and features are usually embedded in the hyperspace of their electrical characteristics and are difficult to extract. In this article, we propose to use dimensionality reduction to extract the defect signature from the IC electrical characteristics using data generated through simulations. A CMOS inverter is used for demonstration. The drain contact resistances, which can increase due to defect or degradation, of the nMOSFET and pMOSFET in an inverter are extracted using a machine based on autoencoder (AE). The machine is trained using data generated from SPICE simulation. The machine is then tested using experimental data and high accuracy is obtained ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${R}^{\,2} &gt;0.9$ </tex-math></inline-formula> ). In particular, for the first time, through the analysis of the hidden variables, we demonstrate that the machine has effectively extracted the features automatically which obviates the cumbersome feature extraction process.

Spreading Resistance Modeling for Contact Resistivity Extraction in Ohmic Multilayer Structures With Circular Electrodes

Contacts featuring multiple layers of materials in combination with metal electrodes enabled record-high power conversion efficiencies for solar cells in recent years. Minimizing the contact resistivity in these multilayer structures is critical to achieve a high device performance and rapid characterization methods are crucial for the experimental analysis of this aspect. However, commonly used test methods and the associated mathematical formulations to extract contact resistivity are not designed for samples with multiple layers. In this work, we analyze formulations describing the resistance measured in ohmic multilayer test structures, which are used to fit and extract an unknown contact resistivity. We focus on a contacting scheme with circular electrodes, commonly named after Cox and Strack (CS), and analyze a material configuration consisting of a thick Si substrate fully coated with a thin conductive layer (CL). We demonstrate the applicability and accuracy of various spreading resistance models (SRMs) from the literature to calculate the measured resistance and compare the results with those obtained by the finite element method (FEM). Moreover, we present a new SRM with a closed-form expression for the measured resistance, which is computationally more efficient than the existing SRMs. Last, we demonstrate the applicability of this model in experiments with ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${n}$ </tex-math></inline-formula> )c-Si/( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${i}$ </tex-math></inline-formula> )a-Si:H/( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${n}$ </tex-math></inline-formula> )a-Si:H/indium tin oxide (ITO) test structures.

Spreading resistance modeling for rapid extraction of contact resistivity with a four-point probe

Extraction of contact resistivity (ρc) with the four-point probe (4PP) method requires considerably fewer fabrication and measurement steps compared to conventional alternatives (e.g. the transfer length method, TLM). Yet, the method is rarely used, mainly due to the requirement of 3D simulations used in the data fitting procedure. Moreover, a work analyzing the details of the method, which can provide a quantitative guideline to the users, is lacking in the literature. In this work, we show that the 3D simulations can be replaced with a spreading resistance model (SRM) implemented in a numerical calculation platform (e.g. MATLAB). Using the SRM, we achieve drastically short calculation durations (<1 s) compared to durations of 3D simulations based on the finite element method. We show that the resolution and reliability of the extracted ρc depends critically on geometrical and electrical parameters of the system, such as the probe tip spacing and sheet resistances of the layers. We also present an alternative measurement procedure that is fully based on 4PP measurements of only one substrate, which eliminates the need for an eddy-current measurement or an additional set of dedicated samples. We demonstrate the viability of the 4PP method in experiments by comparing the results with those obtained by the TLM, for ρc between 5 × 10−3 to 1 × 100 Ω cm2. Lastly, we provide a free software that can execute the calculations discussed herein.

Accuracy of Contact Resistivity Measurements for Electron-Selective Titanium Oxide Contacts in n-Type c-Si Solar Cell

Contact resistivity quantifies the charge transport barrier, which is one of the key parameters for choosing the carrier selective contact for silicon solar cells. Optically transparent and electrically selective contacts such as transition metal oxide (TMO)-based contacts are employed in solar cell applications. Therefore, the extraction of contact resistivity for such Schottky contacts requires apposite validation. In this article, we have extracted the contact resistivity of TiOx/LiFx/Al stack over n-type c-Si wafer using two conventional techniques: 1) Shockley's transfer length method (TLM) and 2) Cox and Strack method (CSM). The extracted contact resistivity is validated by comparing with the solar cell's contact resistivity ( true contact resistivity) and fill factor (FF). We found that TLM overestimates the contact resistivity for highly resistive TiOx films, due to the asymmetrical nature of the TMO barrier and correlated with the reported experimental data. In contrast, CSM extracts more accurate contact resistivity for TiOx contacts when total resistance is extracted closer to the J mpp (38 mA cm−2). Besides, we recommend including the passivating layer along with the contacts (in the case of passivating contacts) to extract the actual contact resistivity felt by the solar cell.

Contact resistance assessment and high-frequency performance projection of black phosphorus field-effect transistor technologies

In this work, an evaluation of the contact quality of black phosphorus field-effect transistors from different technologies previously reported is performed by means of an efficient and reliable contact resistance extraction methodology based on individual device practical characteristics. A good agreement is achieved between the extracted values with the Y-function method used here and reference values obtained with other methods considering internal values as well as with more expensive methods involving fabricated test structures. The method enables a direct evaluation of different steps in the same technology and it embraces the temperature dependence of the contact characteristics. Channel phenomena have no impact on the extracted contact resistance values. High-frequency performance projections are obtained for fabricated devices based on the extracted contact resistance.

Contact resistance extraction of graphene FET technologies based on individual device characterization

Straightforward contact resistance extraction methods based on electrical device characteristics are described and applied here to graphene field-effect transistors from different technologies. The methods are an educated adaptation of extraction procedures originally developed for conventional transistors by exploiting the drift–diffusion-like transport in graphene devices under certain bias conditions. In contrast to other available approaches for contact resistance extraction of graphene transistors, the practical methods used here do not require either the fabrication of dedicated test structures or internal device phenomena characterization. The methodologies are evaluated with simulation-based data and applied to fabricated devices. The extracted values are close to the ones obtained with other more intricate methodologies. Bias-dependent contact and channel resistances studies, bias-dependent high-frequency performance studies and contact engineering studies are enhanced and evaluated by the extracted contact resistance values.

A Ladder Transmission Line Model for the Extraction of Ultralow Specific Contact Resistivity—Part II: Experimental Verification

In the first part of this two-part article, a ladder transmission line model (LTLM), featuring elimination of parasitic metal resistances and high accuracy for extracting specific contact resistivity $\rho _{\text {c}}$ , is developed and verified by simulation. In Part II, the LTLM is applied experimentally to Ni contacts formed on the heavily Ga-doped Ge0.95Sn0.05 film. Two types of Ni/p+-Ge0.95Sn0.05 LTLM structures, single-cell and multicell LTLMs, are fabricated. Ultralow $\rho _{\text {c}}$ values of ${8.6} \pm {2} \times {10}^{-{10}}\,\,\Omega {-}\text {cm}^{{2}}$ and ${1.1} \pm {0.3} \times {10}^{-{9}}\,\,\Omega {-}\text {cm}^{{2}}$ are extracted from the single-cell and multicell LTLMs, respectively. The good agreement of the extracted $\rho _{\text {c}}$ from the single-cell and multicell structures validates the $\rho _{\text {c}}$ extraction by the LTLM. $\rho _{\text {c}}$ is also extracted by using refined transmission line model (RTLM). In contrast, the extracted $\rho _{\text {c}}$ by the RTLM is more than two times larger due to the ignored parasitic metal resistance. The proposed LTLM is a useful tool for an accurate assessment of $\rho _{\text {c}}$ in advanced contact technology.

A Ladder Transmission Line Model for the Extraction of Ultralow Specific Contact Resistivity—Part I: Theoretical Design and Simulation Study

A ladder transmission line model (LTLM) that features capability to eliminate parasitic metal resistance from contact metal and access electrodes, a simple fabrication process, and a sub- $10^{-{10}}\,\, \Omega \cdot $ cm2 resolution is demonstrated for accurate extraction of the specific contact resistivity $\rho _{c}$ . The proposed LTLM is verified by technology computer-aided design (TCAD) simulation and a two-layer distributed-resistor network model. After eliminating the parasitic metal resistance, the current spreading and crowding effects are the main factors that contribute to the extraction error in LTLM. The effects of current spreading and crowding on the extraction of $\rho _{c}$ in LTLM are further quantitatively analyzed and design guidelines to achieve a sub- $10^{-{10}}\,\, \Omega \cdot $ cm2 resolution of $\rho _{c}$ extraction are provided. The details of theoretical modeling and simulation of LTLM are documented in Part I. The experimental demonstration of extracting sub- $10^{-{9}}\,\, \Omega \cdot $ cm $^{{2}} \rho _{c}$ using the proposed LTLM is provided in Part II.

Elimination of the Parasitic Metal Resistance in Transmission Line Model for Extraction of Ultralow Specific Contact Resistivity

A metal-resistance-aware transmission line model (MRA-TLM) is developed to eliminate the parasitic metal resistance from the extraction of specific contact resistivity $\rho _{c}$ in the metal/semiconductor contact. The proposed MRA-TLM is verified by Synopsys technology computer-aided design (TCAD) simulation, and a superior extraction accuracy as compared to the conventional TLM-based method is achieved. The proposed MRA-TLM is experimentally demonstrated in the metal/p+-Ge0.95Sn0.05 (p+-GeSn) contacts. Ultralow $\rho _{c}$ values of ${7.5} \times {10}^{-{10}}$ and ${2.1} \times {10}^{-{9}} \,\, \Omega \cdot \text {cm}^{{2}}$ are extracted by MRA-TLM for Ni/p+-GeSn and Ti/p+-GeSn contacts, respectively, which are in good agreement with ${9.5} \times {10}^{-{10}}$ and ${2.6} \times {10}^{-{9}} \,\,\Omega \cdot \text {cm}^{{2}}$ extracted numerically by the two-layer distributed-resistor-network model (TDM), qualifying the proposed MRA-TLM for the extraction of ultralow $\rho _{c}$ down to sub- $10^{-{9}}\,\,\Omega \cdot \text {cm}^{{2}}$ regime. In contrast, the $\rho _{c}$ extracted by the conventional TLM-based method is more than three times higher due to the parasitic metal resistance.

An Expanded Cox and Strack Method for Precise Extraction of Specific Contact Resistance of Transition Metal Oxide/n-Silicon Heterojunction

Specific contact resistance $(\rho _{c})$ , commonly extracted by Cox and Strack method (CSM) and transfer length method (TLM), is one of the most important properties of carrier-selective contacts. However, in most cases, the hole-selective contacts (HSCs) deposited on n-type silicon (n-Si) substrate are Schottky heterojunction other than Ohmic contact, impeding the accurate extraction of $\rho _{c}$ by CSM and TLM. In this paper, an expanded CSM is proposed to precisely extract the $\rho _{c}$ of MoOx/n-Si heterojunction, achieving a generally lower coefficient of variation. The current transport characteristic and the $\rho _{c}$ value of MoOx/n-Si heterocontact are further verified by technology computer aided design (TCAD) simulation. The results demonstrate that the expanded CSM enables a more precise $\rho _{c}$ extraction, a better preparation technology compatibility, and a wider range of application, compared to TLM.

Effect of interfacial Ni between graphene and Pt/Au on reducing specific contact resistivity

Low contact resistance between graphene and metals is of critical importance to fabricate high-performance graphene-based devices. In this work, the effect of interfacial Ni of different thicknesses, i.e. 0 nm, 1 nm, 2 nm, 4 nm and 5 nm, on the contact resistance of graphene and Pt/Au was investigated systematically. As-transferred graphene grown on Cu foil using a PMMA-assisted process was characterized by optical, Raman and atomic force microscopy. As to the extraction of contact resistance, conventional transmission line method structures were fabricated and measured sophisticatedly. Rapid thermal annealing at different temperatures, i.e. 400 °C, 450 °C, 500 °C, 550 °C and 600 °C, was also performed to determine the optimum temperature and thickness of Ni as an interfacial layer. It was found that an interfacial Ni with proper thickness and thermal treatment is beneficial in reducing contact resistance between graphene and Pt/Au. The role of interfacial Ni is also discussed.

Extraction of Contact Resistance and DC Modeling in Metal Oxide TFTs

Based on the device physics, a simple and fast method of extracting contact resistance in metal oxide thin-film transistors (MOTFTs) is proposed through the I–V characteristics. This method divides the channel into two parts: the contact channel and the intrinsic channel, and assumes the electrons injected into the active layer at the source side are completed in the line injection. Using the I–V characteristics, the contact voltage can be obtained, and then the contact resistance can be extracted. The results indicate that contact resistance in MOTFTs depends on Vg , Vd , and L . Applying the extraction results, a dc drain current model considering contact resistance is proposed. Using this model, we can accurately describe the measurements of MOTFTs with channel lengths scaling from 50 $\mu \text{m}$ to 10 $\mu \text{m}$ . Through the extensive comparisons between the model results and the numerical iteration or experimental data, the validity of the method is strongly supported. This extraction method uses non-numerical iteration, which is simple, fast, and accurate.

An Improved Methodology for Accurate Extraction of Ultra-Low Specific Contact Resistivity of Alloyed Contacts Using Nanoscale Transmission Line Method

An effective transfer length (ETL) model is developed to accurately extract the specific contact resistivity $\rho _{c}$ for alloyed or recessed contacts by accounting for the modification of sheet resistance of semiconductor under the contact ( ${R}_{{\textit {shc}}}$ ) and uneven contact interface due to the reaction of the contact metal and the semiconductor. This method is applied to an example where the formation of NiGeSn extends into p+-Ge0.95Sn0.05 region to form a recessed contact structure. An ultra-low $\rho _{c}$ of $\textsf {2.6}\pm \textsf {0.2}\times \textsf {10}^{-9} \Omega $ -cm $^{\textsf {2}}$ is extracted by the ETL model, which is consistent with $\rho _{c}$ of $\textsf {2.9}\times \textsf {10}^{\textsf {-9}} \Omega $ -cm $^{\textsf {2}}$ extracted by the two-layer resistive-network model. The conventional nano-scale transmission line method extracted a $\rho _{c}$ , that is, ~86% higher, which is due to an underestimated ${R}_{{\textit {shc}}}$ . The proposed ETL model can be generally used for $\rho _{c}$ extraction in any alloyed or recessed contacts.

Improved Charge Injection of Metal Oxide Thin‐Film Transistors by Stacked Electrodes of Indium Tin Oxide Nanoparticles and Silver Nanowires

AbstractSolution processed transparent electrodes are developed combining indium tin oxide nanoparticles (ITO‐NPs) and silver nanowires (Ag NWs) at low process temperature of 100 °C, compatible with solution processed indium gallium zinc oxide (IGZO) thin‐film transistors as source and drain electrodes. Four structures with various stacking of ITO‐NPs and Ag NWs are investigated. Although sole ITO‐NPs cannot provide the conductivity electrical properties of devices, including the mobility, threshold voltage, and subthreshold swing, are improved using ITO‐NPs over/under the Ag NWs. The mobility is enhanced by employing a hybrid electrode structure from 0.004 cm2 V−1 s−1 (single layer) to 3.42 cm2 V−1 s−1 (triple layer). The ultraviolet photoelectron spectroscopy and UV–vis measurements reveal the electronic structure to confirm the increase of charge carrier concentration at the interface, when ITO‐NP/IGZO heterojunction is formed. The ITO‐NP layer acts as a buffer layer between the IGZO and Ag NWs layer to facilitate charge injection to reduce the contact resistance, while the Ag NWs supply conductivity. These are also confirmed by Hall measurement and contact resistance extraction. Hybrid electrodes developed here are promising approach for solution processed transparent electrodes toward transparent and flexible electronics with advantages of low‐cost process excluding vacuum system.

Analysis of charge injection and contact resistance as a function of electrode surface treatment in ambipolar polymer transistors

Ambipolar organic field-effect transistors (OFETs) have both of hole and electron enhancements in charge transport. The characteristics of conjugated diketopyrrolopyrrole ambipolar OFETs depend on the metal-contact surface treatment for charge injection. To investigate the charge-injection characteristics of ambipolar transistors, these devices are processed via various types of self-assembled monolayer treatments and annealing. We conclude that treatment by the self-assembled monolayer 1-decanethiol gives the best enhancement of electron charge injection at both 100 and 300 °C annealing temperature. In addition, the contact resistance is calculated by using two methods: One is the gated four-point probe (gFPP) method that gives the voltage drop between channels, and the other is the simultaneous contact resistance extraction method, which extracts the contact resistance from the general transfer curve. We confirm that the gFPP method and the simultaneous extraction method give similar contact resistance, which means that we can extract contact resistance from the general transfer curve without any special contact pattern. Based on these characteristics of ambipolar p- and n-type transistors, we fabricate inverter devices with only one active layer.