S-shaped Negative Differential Resistance Research Articles

Cu TSV Filling with Cetyltrimethyl Ammonium Cations (CTA+) and Br-

Through silicon via (TSV) technology is essential for the three-dimensional packaging of electronic devices. After creating high-aspect-ratio vias, Cu electrodeposition is performed to achieve defect-free Cu interconnection. The most important purpose of electrodeposition process is to fill the features without any defects, and this process is called superfilling, superconformal deposition, or bottom-up filling. Since TSV is a key component for the packaging of semiconductor devices, various research on the electrodeposition process for TSV filling have been studied. Defect-free TSV filling can be achieved by introducing appropriate organic/inorganic additives into the electrolyte and the filling performance can be promoted by pulse reverse electrodeposition.1-3 Generally, it has been reported that Cu bottom-up filling at TSVs can be obtained by three-additive combinations (suppressor, accelerator, and leveler) or by using a single suppressor. Tetronic 701 is the most representative suppressor in a single-additive filling method, which was based on S-shaped negative differential resistance (S-NDR) behavior.4-6 In the case of three-additive systems, various levelers that contained ammonium functional groups, a polymeric suppressor, and an accelerator have been researched.7,8 Furthermore, it has been reported that halide ions (Br- and I-) can act as levelers for inducing Cu bottom-up filling at TSVs.9,10 This study investigated the effects of cetyltrimethyl ammonium cation (CTA+) and halide ion on TSV filling in the presence of an accelerator and polymeric suppressor. The influences of two halide ions, Cl- and Br-, on the adsorption of CTA+ and their convection dependency were studied. This presentation will introduce the electrochemical behaviors of halide ions and CTA+ and defect-free filling using the optimum combination of these additives. References M. Hayase, and K. Otsubo. "Copper deep via filling with selective accelerator deactivation by a reverse pulse" J. Electrochem. Soc.,157, D628 (2010).Q.S. Zhu, X. Zhang, C.Z. Liu, and H.Y. Liu. "Effect of reverse pulse on additives adsorption and copper filling for through silicon via" J. Electrochem. Soc.,166, D3006 (2019)Q.S. Zhu, A. Toda, Y. Zhang, T. Itoh, and R. Maeda. "Void-free copper filling of through silicon via by periodic pulse reverse electrodeposition" J. Electrochem. Soc.,161, D263 (2014)T. P. Moffat, and D. Josell. "Extreme bottom-up superfilling of through-silicon-vias by damascene processing: suppressor disruption, positive feedback and turing patterns" J. Electrochem. Soc.,159, D208 (2012)D. Joell, and T. P. Moffat. "Superconformal copper deposition in through silicon vias by suppression-breakdown." J. Electrochem. Soc.,165, D23 (2018)M. Dong, Y. Zhang, T. Hang, and M. Li. "Structural effect of inhibitors on adsorption and desorption behaviors during copper electroplating for through-silicon vias." Electrochim Acta372 (2021)M. J. Kim, Y. Seo, J. H. Oh, Y. Lee, H. C. Kim, Y. G. Kim, and J. J. Kim. "Communication—halide ions in teg-based levelers affecting TSV filling performance." J. Electrochem. Soc.,163, D185 (2016)V. Q. Dinh, K. Kondo, V. H. Hoang, and T. Hirato. "Communication—Bottom-up TSV filling using sulfonated diallyl dimethyl ammonium bromide copolymer as a leveler." J. Electrochem. Soc.,166, D505 (2019)M. J. Kim, H. C. Kim, and J. J. Kim. "The influences of iodide ion on Cu electrodeposition and TSV filling." J. Electrochem. Soc.,163, D434 (2016)M. Sung, Y. Yoon, J. Hong, M. J. Kim, and J. J. Kim. "Bromide ion as a leveler for high-speed TSV filling." J. Electrochem. Soc., 166, D546 (2019) Figure 1

Lifting Polyether Suppression during Cu Electrodeposition

The manufacturing of semiconductor devices involves electrodeposition of copper interconnects from additive containing electrolytes. The process depends on additives that affect the local deposition rate to yield void-free superconformal or bottom–up filling of trenches and vias. Useful quantitative surfactant conservation models are available for describing the filling of high aspect ratio features ranging from nanometers (on-chip wiring) to micrometers (through-silicon-vias) to millimeters (printed circuit boards). This includes the curvature enhanced adsorbate coverage (CEAC) model as well as the additive derived S-shaped negative differential resistance (S-NDR) description of film growth and feature filling. Nonetheless, much remains to be known about the operational limits of these constructs as well as the molecular nature of the competitive, co-adsorption dynamics of the additives and their impact on metal deposition. This lecture will detail recent experiments that reveal the important role of competitive, co-adsorption of polyether, Cl-, SPS and LEV additives in controlling the hydrophilicity of the interface and thereby the rate of metal deposition and its morphological and microstructural evolution.

(Invited) Through-Silicon Via Copper Plating and Endpoint Detection

Heterogenous, 2.5D, and 3D integration provide solutions for making microelectronics systems more compact and improving their performance. Cu-filled through-silicon vias (TSVs) are an important interconnect technology for these advanced integration techniques. TSV design and geometry (overall size, aspect ratio, and pitch) depend on the intended microelectronics application. Standard integrated circuit (IC) approaches often utilize high density vias that are achieved through substrate thinning; however, some applications need large-scale vias (≥ 600 µm deep). For example, some microelectromechanical systems (MEMS) applications need full wafer thickness vias due to die thickness specifications, and high-power devices can either employ large-scale vias or via arrays for current requirements and thermal management constraints. Cu deposition fill profile is sensitive to applied voltage or current, electrolyte composition and concentrations, and feature geometry; therefore, deposition processes must be tuned specifically for each via geometry. This work outlines Cu deposition process development and filling results for full wafer thickness TSV geometries in a suppressor-based electrolyte consisting of copper sulfate, sulfuric acid, potassium chloride, and Tetronic 701 suppressor, [i] according to the S-shaped Negative Differential Resistance (S-NDR) mechanism. [ii],[iii] During these TSV filling experiments, unique features in the electrical data were observed that signify completion of Cu plating in the vias. A novel endpoint detection method was derived from these data to determine when vias are filled. [i] Analogous endpoint detection through electrical signals has been previously observed for bottom-up Au filling of high aspect ratio trenches. [iv] During voltage-controlled via filling, as the Cu electrodeposit approaches the top of the TSVs, current can be monitored for a slight rise (due to increased metal ion transport and plating rate high up in the features), followed by a current local minimum (caused by adsorbed suppressor molecules that impede Cu deposition at the top of the vias). These characteristic features in the electrical data can be used to determine via filling completion. A similar technique can be used during current-controlled filling, where voltage data are monitored for a voltage maximum, which indicates that the Cu electrodeposit has reached the top of the features. Current work involves development of an automated method to recognize these electrical endpoint signals and terminate deposition when the vias are filled. This endpoint detection method has potential to greatly enhance microelectronics manufacturing by providing geometry-independent process control, improving manufacturing repeatability, decreasing sensitivity to changes in electrolyte over time, and reducing the time involved in subsequent processing steps through minimizing Cu overburden formation.Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. SAND2021-5013 A [i] R. P. Schmitt et al 2020 J. Electrochem. Soc. 167 162517. [ii] D. Josell and T. P. Moffat 2016 ECS Trans. 75 15. [iii] D. Josell and T. P. Moffat 2018 J. Electrochem. Soc. 165 D23. [iv] S. Ambrozik et al 2019 J. Electrochem. Soc. 166 D443.

Equilibrium analysis of Mott memristor reveals criterion for negative differential resistance

Two-terminal electronic devices that exhibit voltage-controlled threshold switching (TS) via negative differential resistance (NDR) are important for many emerging applications. Pickett and Williams developed what has become a well-known physics-based model for nanoscale devices exhibiting NDR due to a reversible insulator-metal phase (Mott) transition. The Mott memristor model couples changes in electrical resistance and Joule heating to the phase of the material using one dynamic state variable, u, that describes the volume fraction of metal in the cross section of the device. The model formulation involves one nonlinear first-order ordinary differential equation and eight physical parameters. New equilibrium analysis reveals a simple condition that determines whether the model predicts NDR required for current–voltage (i–v) hysteresis in a voltage-controlled operation. We show that S-shaped NDR (also called current-controlled NDR) arises only above a critical ratio, Mc, of insulator to metal resistivity. Specifically, hysteresis in the i–v plane cannot occur below Mc=e2+1≈8.39 (i.e., e≈2.718…; Euler's number), but above this value hysteresis appears. This understanding enables tuning of hysteretic features, including threshold voltages for resistive switching, which benefit the use of TS memristors as memory storage elements, as well as excitable devices mimicking neural action potentials.

Mathematical Modeling of Cyclic Voltammogram Curves of Copper Deposition Involving Multiple Additives

The manufacturing of interconnects and the packaging of integrated circuits are achieved with electrodeposition of copper or other metals. In order to increase the rate of deposition, especially for the large features in packaging, forced convection is provided with certain agitation mechanisms. Although this reduces deposition time, it leads to non-uniform mass transport within each feature and between different features. Special organic additives are used in the solution during the process in order to tune the nucleation and growth of metal, as well as to modify the deposition rate and improve the uniformity. A mathematical model to describe the behavior of organic additives in conjugation with fluid flow and features of various geometry and dimensions is very much desired to facilitate chemistry and process development. In order to achieve this, the physiochemical kinetics of additive and their influence on the Cu deposition rate need to be described precisely.This presentation focuses on a method to extract the kinetic parameters describing the combined effect of multiple additives during copper deposition using rotating disk electrode (RDE). The one-dimensional steady state convection-diffusion equation for each of the chemical species including copper is solved by a semi-analytical method for a range of potentials. The boundary conditions of these differential equations are coupled on the surface of the RDE through the surface coverage of the absorbed species. The steady state of surface coverage of the species represents a dynamic equilibrium of three key processes i.e., adsorption, desorption, and consumption (incorporation). When equilibrium is achieved, the net rate of adsorption and desorption becomes equal to the rate of consumption. At each value of potential, the surface coverage of the additives is solved. At first, the solution is obtained with only one species known as suppressor and it was found that in a specific range of voltage and kinetic parameter multiple solutions of the surface coverage exist at same applied potential. This mathematically explains the S-shaped negative differential resistance (NDR) feature in experimental Cyclic Voltammogram (CV) curves. Figure 1 shows three such experimental S-shaped curves for different concentration of suppressors. The NDR region obtained in the theoretical CV curve is sensitive to the kinetic parameters of the additives. It is possible to match the theoretical and the experimental CV curves by optimizing the kinetic parameters. Determination of the kinetic parameters by particle swarm optimization using experimental data for multiple additive concentration will be discussed in detail in this talk. Figure 1

Simulating Cu electrodeposition in high aspect ratio features: Effect of control mode and uncompensated resistance in S-NDR systems

Void-free Cu electrodeposition in high aspect ratio features requires, at a minimum, an additive package containing micromolar halide and polyether that combine to form a co-adsorbed adlayer that inhibits metal deposition on the electrode interface. Successful feature filling relies on preferential growth proceeding from the most recessed surfaces where sustained breakdown of the polyether-halide suppressor layer occurs. Localization is the result of positive feedback between inhibitor breakdown and metal deposition subject to transport limitations of the suppressor precursor(s). This gives rise to a S-shaped negative differential resistance (S-NDR) that, convolved with uncompensated ohmic resistance, results in electrode bifurcation into active and passive zones. The interplay between the additive derived S-NDR behavior, uncompensated cell resistance, and potentiostatic regulation is explored in comparison to galvanostatic feature filling. Uncompensated resistance arises from the working electrode contact and electrolyte between the working and reference electrodes. For a CuSO4 – H2SO4 electrolyte containing 80 µmol/L Cl− and 40 µmol/L polyether, simulations of potentiostatic deposition with minimal uncompensated resistance reveal a narrow window between fully passive and voided feature filling; bottom-up filling terminates prematurely even under the most favorable conditions. In contrast, optimized galvanostatic operation enables void-free feature filling with termination dictated by coulometry. Increasing the uncompensated resistance along with application of accordingly more negative applied potentials produces filling dynamics that blend the positive attributes of galvanostatic and potentiostatic deposition to enable complete, void-free feature filling with spontaneous passivation near the feature opening. Importantly, these beneficial filling effects are also evident for trench arrays having variable widths or heights.

(Electrodeposition Division Research Award Address) Superconformal Deposition: Where We Are and Where We Might Go

Superconformal deposition processes for void-free filling of features have developed remarkably over the first two decades of this century. Superconformal copper deposition, aka “superfill”, enabled void-free copper filling of features with conductive surfaces having aspect ratios unimaginable when implemented in on-chip interconnects in 1997. The only known mechanism, already more than 40 years old, provided only more modest superconformal feature filling through a gradient of deposition rate that scaled with a gradient of suppressor concentration arising from incorporation in the depositing metal. A new mechanism, for the copper “superfill”, was determined in 2001: local coverage of a strongly bound but surfactant accelerator arose through area change that accompanied deposition on nonplanar surfaces. Copper “superfill” through this Curvature Enhanced Accelerator Coverage, or CEAC, mechanism accompanied and accounted for the decrease of feature size in on-chip interconnects over the next ten years. Along the way the CEAC mechanism was shown to be generic to the coinage metals of silver and gold. In all cases, feature filling was shown to be predicted a priori by CEAC-based models using only kinetics derived from studies on planar substrates.In the meantime, the microelectronics industry began to pursue wafer stacking that hinged upon void-free copper filling of much larger Through Silicon Vias, TSV, for which the CEAC mechanism was poorly suited. Demonstrations of bottom-up copper filling incompatible in both geometry and additive-chemistry with the CEAC mechanism required a new mechanistic understanding. In 2012 the old, linear model for suppressor containing electrolytes was replaced by a model reflecting suppressors whose deactivation was so non-linear in potential and concentration that suppression breakdown yielded increasing negative current with increasing positive potential during cyclic voltammetry. Models based on this S-shaped negative differential resistance (S-NDR) predicted the extreme bottom-up nature of the copper filling. Subsequent years brought demonstrations of generality with gold, nickel, cobalt, zinc and copper, albeit manifesting in two very different filling geometries: activation of only the bottom-up surface and activation of the bottom surface with a portion of the adjacent sidewalls. Filling of features having much larger dimensions as well as geometries such as through holes was also demonstrated. In all cases, the filling evolution was shown to be a-priori predictable by S-NDR based models using only kinetics derived from studies on planar substrates.Today we find ourselves with the next superconformal filling process requiring new mechanistic understanding. Gold filling exhibits the bottom-up geometry of suppressor-derived S-NDR filling but with a bismuth additive that accelerates as in CEAC filling. It is thereby explained by neither. Void-free, bottom-up filling has been achieved in trenches whose depths range from a few micrometers to a quarter of a millimeter. Trenches with aspect ratios as low as unity and as high as fifty-six (!) have been successfully bottom-up filled, and surface profiles indicate the limit has not been reached. Adsorption and deposition exhibit a unique nonlinear dependence on transport, and bottom-up filling halts automatically at a depth within patterned features defined by potential that, equivalently, yields a tell-tale potential signature upon filling under galvanodynamic control. Uniformity of filling is excellent across patterned substrates. The new process satisfies a need outside the area of microelectronics: gold filling of trenches in x-ray diffraction gratings. As in the past, understanding will enable superconformal electrodeposition to meet the evolving needs of a new technology. Figure 1

Copper Electrodeposition in High Aspect Ratio Mesoscale Through-Silicon Vias: Scaling from Die Level to Wafer Level Plating

Three-dimensional (3D) microsystem integration offers solutions for improved device functionality, increased bandwidth per volume, simplified design and assembly, and system miniaturization. Through-silicon vias (TSVs) are a key technology for incorporating devices into 3D integration schemes. Microelectromechanical systems (MEMS) applications often utilize full thickness silicon substrates, creating a need for mesoscale TSVs (> 0.5 mm thick) at high aspect ratios (> 5:1) to facilitate demands of increased interconnect density. Copper (Cu) is a preferred material for TSV filling because of its high electrical and thermal conductivity and can be deposited through electrochemical deposition (ECD) to fill high aspect ratio features. Although Cu ECD has been historically performed using three additive systems, recent work has demonstrated that Cu deposition can be achieved in a simplified CuSO4-H2SO4 electrolyte system with only a single suppressor additive and halide salt.[i],[ii] The cyclic voltammetry (CV) scan for this electrolyte exhibits s-shaped negative differential resistance (S-NDR), observed as hysteresis in the CV voltage-current relationship between forward and reverse scans. Hysteretic voltammetry indicates bifurcating behavior in the electrolyte system, where bottom-up filling is possible in the vias, while the field surface remains passivated. In systems with limited suppression, performing ECD with a sustained potential leads to localized Cu deposition at specific depths within the vias. An increasingly negative potential waveform can be implemented to progressively move deposition vertically through the vias. Achieving a bottom-up, void-free fill with these additive-based electrolytes is typically performed through potentiostatic plating with a reference electrode. Scaling these processes from die level plating experiments, in a small electrochemical cell, to a production-scale full-wafer plating tool requires development of a current-controlled plating regime, because production tools are not equipped with reference electrodes. Furthermore, the developed plating process cannot be highly sensitive to sample rotation rate, because vias at various radial positions across a wafer move at different speeds. In this work, we describe the development of a current-controlled deposition process for bottom-up, void-free filling and further scale this plating process from 1 cm2 sample sizes to a full 150 mm diameter wafer.Voltage-controlled ECD parameters were first established with die level plating experiments. This voltage-controlled process was used to derive a current-controlled ECD procedure, which was followed by sample size scaling experiments and rotation rate studies. The TSV geometries for these experiments had a 10:1 aspect ratio, with via diameters of 62.5 µm, etched into a 625 µm thick silicon substrate. These vias were successfully filled with a potential stepping regime in which the voltage started at -500 mV (MSE) and was gradually incremented to -560 mV (MSE) using -10 mV steps. Time duration at each potential ranged from 2 to 5 hours to achieve a void-free fill. During these voltage-controlled Cu deposition experiments, the resulting current was measured. Subsequent plating experiments were executed by applying current, as opposed to voltage, to mimic the measured current values, and this current-controlled ECD process also achieved void-free filling. Figure 1 shows the development process from voltage-controlled to current-controlled ECD and corresponding mechanical cross sections. As shown in the sample cross sections, these electroplating processes result in fully suppressed deposition on the field area, even though the field is part of the total conducting surface. It was hypothesized that current density scales with via area, as opposed to total conductive surface area; therefore, scaling experiments were performed to understand the relationship between applied current density, total conductive surface area, and active via area. Additionally, with a wafer rotating about its center, solution replenishment within each via is dependent on its radial position. To understand the dependence on solution replenishment and applied bias in a given electrolyte, a series of investigations were performed on samples rotated at different rates. This work presents the results of these studies for establishing a wafer level plating process for 10:1 aspect ratio TSVs. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. SAND2020-5515 C [i] D. Josell and T. P. Moffat, Journal of The Electrochemical Society, 165 (2) D23-D30 (2018). [ii] L. A. Menk, D. Josell, T. P. Moffat, E. Baca, M. G. Blain, A. Smith, J. Dominguez, J. McClain, P. D. Yeh, and A. E. Hollowell, Journal of The Electrochemical Society, 166 (1) D3066-D3071 (2019). Figure 1

Estimation of Kinetic Parameters of Additives By Semi-Analytical Solution

The electrodeposition of Cu and other metals is an essential step in the interconnect fabrication, 3-D integration, and various packaging technologies. Such electrodeposition processes are enabled by a group of organic additives. Furthermore, strong agitation is often used to expedite the electrodeposition rate when the processing time becomes critical for large features, which however introduces non-uniform mass transport within a single feature and between different features. A numerical description of how organic additives behave in conjunction with external flow at various feature geometry and dimensions is a challenging task, but much needed for chemistry and process development. Among the various requirements, the additive behavior or their electrochemical impacts on Cu deposition needs to be precisely described.This talk reports an effort to obtain the numerical parameters that describe the additive behaviors during Cu deposition on rotating disc electrodes (RDE). A one-dimensional semi-analytical model is developed for Cu electrodeposition from a two-component additive system. The steady-state mass transport differential equations of each chemical species, including diffusion and convection, are solved analytically at different potentials, while the boundary conditions at the RDE surface are coupled with each other through the surface coverages of adsorbed species. The steady state of such surface coverage results from a dynamic equilibrium between three co-existing processes, adsorption, desorption and incorporation (consumption), namely, Adsorption Rate – Desorption Rate – Consumption Rate = 0 A set of solutions, including the concentration profile of each species in electrolyte as well as the surface coverage of each adsorbate, can be obtained at a fixed potential. The equations are first solved considering only the suppressor for a range of over-potential. It was found that, for a specific range of voltages and kinetic parameters, multiple non-trivial solutions of surface coverage exist. In other words, multiple steady states exist at a same applied potential. As shown in Figure 1, this translates to a S-shaped negative differential resistance (NDR) feature in cyclic voltammogram (CV), consistent with experimental observation reported in literature. The numerically calculated NDR region was found sensitive to the additive concentration and therefore extremely useful in determining the electrochemical parameters using multiple experimental datasets with different concentrations. The swarm optimization algorithm used for parameter fitting and the correlation between multiple numerical solutions with the NDR region will be discussed in details in the talk. Figure 1

Bottom-up Copper Electrodeposition in 20:1 Aspect Ratio, 20 Μm Pitch through Silicon Vias

Heterogeneous integration (HI) offers a path to increase bandwidth density and reducing power consumption per area for advanced microelectronic systems by minimizing fan-in/out structures and reducing parasitics. To achieve state-of-the-art areal bandwidth density, 3D HI is required with through silicon vias (TSVs) with extremely tight interconnect pitches (< 50 µm) and high aspect ratios. Fabrication and fill of these proposed TSVs pose significant challenges including substrate thinning, high aspect ratio deep reactive ion etching (DRIE), conformal thin film deposition to deposit insulating layers as well as seed metals for electrodeposition, and void-free fill of the TSVs. Here, we target an interconnect pitch of 20 µm and TSV diameter of 5 µm. At these dimensions, DRIE lag limits the TSV depth to 100 µm. The TSVs are integrated using a via-last scheme with atomic layer deposition (ALD) Al2O3 as the insulating liner and ALD Pt serving as both the barrier and electrochemical deposition (ECD) seed liner. In this work, we describe the integration scheme for these high density TSVs and focus on the ECD process for achieving a void-free fill with Cu as well as characterization approaches using focus ion beam (FIB) cross sectioning.Recent work on Cu ECD has demonstrated the efficacy of a two-additive acid sulfate electrolyte system based on a chemistry exhibiting a S-shaped negative differential resistance voltammogram1. The electrodeposition system presented is a 1.0 mol/L CuSO4 – 0.5 mol/L H2SO4 acid sulfate electrolyte containing a polyether surfactant and halide additive which promotes a bottom-up Cu growth profile via positive feedback between localized disruption of adsorbed polyether-halide layer and metal deposition. This electrolyte has been applied to potentiostatic filling of a variety of geometries with both micro-scale1 and mm-scale2 TSVs as well as current controlled bottom-up filling from a conformally conductive seed metal3. This system relies on fine tuning of applied bias, adsorbate concentrations, convection, and deposition time to ensure a bottom-up deposition regime is maintained in a particular via geometry.Previous work developed parameters for filling 600 µm deep 5:1 aspect ratio TSVs and the same parameters were applied to 600 µm deep 10:1 aspect ratio TSVs with drastically different filling results.2 In this work, we further adjust the relevant ECD parameters to fill 20:1 aspect ratio TSVs that are 100 µm deep.In fine pitch, high aspect ratio ECD, characterization of the deposited fill profile is important and is difficult to obtain using mechanical cross-sectioning. Here, we use a characterization approach which relies on focused ion beam (FIB) cross-sectioning rather physical grinding and polishing. The FIB mills into the surface of the sample at a 38º angle with respect to the surface of the sample while the electron beam collects a series of images looking down at the sample, as shown in Figure 1.The series of images collected throughout the FIB milling process are then reconstructed to generate a cross-sectional image. FIB cross-sectioning is also used to provide granular detail about deposition morphology and grain size of the ECD Cu that can be lost during mechanical cross-sectioning. Figure 2, shows a top down scanning electron microscopy image of the TSV sample after milling at 38º and the resulting reconstructed cross-sectional image. T. P. Moffat and D. Josell, J. Electrochem. Soc., 159, D208 (2012).L. A. Menk, D. Josell, T. P. Moffat, E. Baca, M. G. Blain, A. Smith, J. Dominguez, J. McClain, P. D. Yeh, and A. E. Hollowell, Journal of the Electrochemical Society, 166 (1) D3066-D3071 (2019).L. A. Menk, E. Baca, M. G. Blain, J. McClain, J. Dominguez, A. Smith, and A. E. Hollowell, Journal of the Electrochemical Society, 166 (1) D3226-D3231 (2019) Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government. This work was funded by the DARPA Photonics in the Package for Extreme Scalability (PIPES) program under award number HR001119S0004. The views, opinions, and/or findings expressed are those of the author(s) and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. Figure 1

An interface-controlled Mott memristor in α -RuCl3

Memristor devices have history-dependent charge transport properties that are ideal for neuromorphic computing applications. We reveal a memristor material and mechanism in the layered Mott insulator α-RuCl3. The pinched hysteresis loops and S-shaped negative differential resistance in bulk crystals verify memristor behavior and are attributed to a nonlinear coupling between charge injection over a Schottky barrier at the electrical contacts and concurrent Joule heating. Direct simulations of this coupling can reproduce the device characteristics.

Anisotropic Copper Electrodeposition on Microelectrodes

Copper electrodeposition is critical to electronics manufacturing, ranging across lengthscales from nanometer damascene plating to 100s of micrometers for through-silicon-vias in 3D interconnect packaging and even millimeter-scale in printed circuit board manufacturing. Essential to this process is formation of void-free copper structures in high aspect ratio cavities, which is achieved with various additives that enable superconformal or bottom-up electrodeposition. When the additive package consists of only a suppressor (polyether) and chloride, the superconformal phenomenon is driven by S-shaped negative-differential resistance (S-NDR), which promotes spatial bifurcation of the electrode surface into active and passive depositing regions. All S-NDR systems involve kinetic suppression, polyether-chloride adsorption in this case, and kinetic activation, potential-driven disruption of the polyether-chloride adlayer. The term S-NDR is derived from the s-shape the i-V curve expresses when cyclic voltammetry (CV) is post-experimentally corrected for ohmic resistance, resulting in a system that expresses a multiplicity of reaction rates (current) for a given driving force (potential).Electroanalytical measurements using microelectrodes (UMEs) offer an avenue for exploring S-NDR systems without significant iR-drop in solution. In other reaction systems, the reduced electrode dimension of UMEs has shown the ability to frustrate active-passive bifurcation, resulting in a homogeneously reacting electrode surface. Nonetheless, potential-controlled measurements such as cyclic voltammetry restrict S-NDR behavior from fully expressing itself, as the potential is the control parameter throughout the experiment. Current-controlled techniques, however, allow the thermodynamic state (i.e., potential) to dynamically shift as the polyether-chloride suppressing layer is formed and disrupted. In this work, current-controlled measurements in the form of linear galvanodynamic sweeps (GDS) are utilized to explore S-NDR behavior and electrode bifurcation during copper electrodeposition. GDS and CV measurements on UMEs are compared across a range of additive concentrations and polyether types. Influence of the operating mode on pattern formation is evaluated by in-situ optical microscopy on UMEs ranging from 10 μm to 25 μm in diameter. Spatial bifurcation during GDS measurements is further exploited to grow anisotropic copper deposits.

Development of a Scalable Two-Additive Electrodeposition System for Bottom-up Copper Filling of Sub-Millimeter through Silicon Vias

A two-additive electrodeposition system for Cu feature filling of sub-millimeter scale through silicon vias (TSVs) is discussed. This 1.0 mol/L CuSO4 – 0.5 mol/L H2SO4 acid sulfate electrolyte contains a polyether surfactant and halide additive which together promote a bottom-up Cu growth profile through positive feedback between localized disruption of the adsorbed polyether-halide layer and metal deposition.1 The functionality of this system relies on a tightly controlled additive concentration gradient down the length of the feature at the onset of metal deposition to allow bottom-up Cu growth to propagate. Careful consideration of convective contributions is also made to ensure an appropriate boundary layer thickness is maintained for non-linear bottom-up filling, where localized active and passive zone bifurcation of rapid electrodeposition is present. Initial experiments demonstrate the efficacy of potentiostatic control in determining the applied current window by which galvanostatic control can ultimately be achieved for a sub-millimeter feature geometry. Nominally void-free bottom-up deposition of multiple recessed feature geometries is presented.One advantage of the two-additive bottom-up electrodeposition system is that it has been demonstrated across multiple length scales and with multiple materials,2-9 making it suitable for heterogeneous integration (HI) of a wide array of devices which can require unique feature geometries and interconnect materials. Depending on the application, the two-additive bottom-up electrodeposition system can replace the more traditional three or four-additive system in damascene processing10 which can yield superconformal deposition profiles in TSVs as explained by the curvature enhanced adsorbate coverage (CEAC) mechanism.11 The CEAC mechanism relies upon high curvature areas and is thus most appropriate for micro and nanoscale features which exhibit proximity of locally high curvature areas as deposition proceeds due to area change affects. Void-free superconformal deposition is more difficult to repeatably achieve on larger, mesoscale features. Thus, a bottom-up deposition technique that relies upon the non-curvature dependent S-shaped negative differential resistance (S-NDR) between just two constituent additives to fill conductively seeded features is desirable.The issue of scalability is of significant interest after developing an electrodeposition process at the die level. While this two-additive system has been successfully implemented using potentiostatic control, industrial full wafer plating tools rely exclusively on galvanostatic control as they are not equipped with reference electrodes. In addition, they combine wafer rotation and forced fluid flow to maintain a controlled convective contribution across the wafer. So, to demonstrate scalability of this two-additive chemistry, the effect of the field surface area, feature geometry, and sample rotation rate on the deposition profile at a given applied potential or current is investigated. Altering the feature quantity and field area for a given deposition experiment can provide a means to estimate the effective current density required to fill features at the same rate across different sample sizes. By altering the sample rotation rate in the otherwise quiescent electrolyte, the impact of the transport phenomena of the constituent surfactant and halide additives are assessed. Cross-sectional images along with computed tomography (CT) produced images are produced to obtain Cu fill profiles across features arrays to demonstrate TSV uniformity. 1. P. Moffat and D. Josell, J. Electrochem. Soc., 159, D208 (2012).2. A. Menk et al., J. Electrochem. Soc. 166, D3066-D3071 (2018).3. Josell, M. E. Williams, S. Ambrozik, C. Zhang, T. P. Moffat, J. Electrochem. Soc. 166, D487-D495 (2019).4. Josell, M. Silva, T. P. Moffat, ECS Trans. 75, 25-30 (2016).5. Josell, T. P. Moffat, J. Electrochem. Soc. 166, D3022-D3034 (2018).6. Josell, T. P. Moffat, ECS Trans. 75, 19-24 (2016).7. Josell, T. P. Moffat, J. Electrochem. Soc. 162, D129-D135 (2015).8. Josell, T. P. Moffat, J. Electrochem. Soc. 160, D3035-D3039 (2013).9. Josell, L. A. Menk, A. E. Hollowell, M. Blain, T. P. Moffat, J. Electrochem. Soc. 166, D3254-D3258 (2019).10. C. Andricacos, C. Uzoh, J. O. Dukovic, J. Horkans, and H. Deligianni, IBM J. Res. Dev., 42, 567 (1998).11. P. Moffat, D. Wheeler, M. Edelstein, and D. Josell, IBM J. Res. Dev., 49, 19 (2005). Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525.

(Invited) Bottom-up Electrodeposition

Superconformal electrodeposition that enables the fabrication of high aspect ratio Cu interconnects is not unique to copper (or for that matter even to electrodeposition). I will detail two known mechanisms underlying such processes: Curvature Enhanced Accelerator Coverage (CEAC) and S-shaped Negative Differential Resistance (S-NDR). I will describe how voltammetric measurements on planar substrates can quantify kinetics for models based on both mechanisms, enabling accurate, a-priori prediction of filling from a given electrolyte in features of various size and aspect ratio. I will summarize the chemistries and associated mechanisms for superconformal Ag, Au, Co, and Ni deposition for fabrication of defect-free nanometer scale damascene to millimeter scale through silicon via interconnects, including model predictions. I will finish by detailing truly bottom-up Au filling of high aspect trenches.

Bottom-up Gold Filling of High Aspect Ratio Trenches :Toward a Mechanistic Understanding

Cu is the standard choice for the formation of multi-level 3-D interconnects allowing for microelectronic chip-stacking (1). This is predicated on the well-defined additive chemistries that drive the superconformal Cu filling of through silicon vias (TSVs) and trenches (2). To transition to Au filling of complex 3-D Si architectures for optoelectronic applications, analogous chemistries that enable preferential filling within recessed silicon features must be developed. To date the state of the art for void free Au filling has come from the use of a Pb additive in both KAu(CN)2 (3, 4) and Na3Au(SO3)2 (5) electrolytes. Both of these systems yield superconformal filling via a curvature enhanced accelerator coverage (CEAC) mechanism. However, due to the toxicity of CN- and Pb2+ containing electrolytes, alternatives must be explored. Recently, it has been demonstrated that the addition of Bi to alkaline Na3Au(SO3)2 electrolytes yields preferential bottom-up Au electrodeposition in Si trenches, with limited deposition on the free surface (6). The addition of Bi to the intrinsically suppressed Na3Au(SO3)2 electrolyte clearly accelerates deposition(6, 7). However, growth in this system does not propagate via the typical CEAC mechanism demonstrated in other accelerated systems (3). Instead the growth propagates via a mechanism more closely resembling that of S-shaped negative differential resistance (S-NDR) based on suppression breakdown, typical for suppressor based additive systems(8). This talk will explore the utilization of this Bi, Na3Au(SO3)2 electrolyte system for the bottom-up filling of high aspect ratio silicon trenches. The effect of; potential, additive concentration and electrolyte transport will be explored on the filling of features with aspect ratios ranging from 1.5 to 17. Additionally, the chemical and morphological nature of the accelerated bottoms of features, and of the passive free surface, will be compared via x-ray photoelectron spectroscopy, scanning electron microscopy and anodic stripping voltammetry to gain mechanistic insight into the role of Bi in deposition acceleration and feature localization. P. C. Andricacos, C. Uzoh, J. O. Dukovic, J. Horkans and H. Deligianni, IBM J. Res. Dev., 42, 567 (1998).T. P. Moffat, D. Wheeler, M. D. Edelstein and D. Josell, IBM J. Res. Dev., 49, 19 (2005).D. Josell, C. R. Beauchamp, D. R. Kelley, C. A. Witt and T. P. Moffat, Electrochem. Solid-State Lett., 8, C54 (2005).D. Josell, D. Wheeler and T. P. Moffat, J. Electrochem. Soc., 153, C11 (2006).D. Josell and T. P. Moffat, J. Electrochem. Soc., 160, D3009 (2013).D. Josell and T. P. Moffat, J. Electrochem. Soc., 166, D3022 (2019).I. D. E. McIntyre and W. F. Peck, J. Electrochem. Soc., 123, 1800 (1976).S.-K. Kim, J. E. Bonevich, D. Josell and T. P. Moffat, J. Electrochem. Soc., 154, D443 (2007).

Non-Linear Self-Heating in Organic Transistors Reaching High Power Densities

The improvement of the performance of organic thin-film transistors is driven by novel materials and improved device engineering. Key developments are a continuous increase of the charge carrier mobility, a scale-down of transistor dimensions, and the reduction of contact resistance. Furthermore, new transistor designs such as vertical devices are introduced to benefit from drastically reduced channel length while keeping the effort for structuring moderate. Here, we show that a strong electrothermal feedback occurs in organic transistors, ultimately leading to output characteristics with regions of S-shaped negative differential resistance. For that purpose, we use an organic permeable-base transistor (OPBT) with outstanding current densities, where a strong and reproducible, non-linear electrothermal feedback is revealed. We derive an analytical description of the temperature dependent current-voltage behavior and offer a rapid investigation method for material systems, where a temperature-activated conductivity can be observed.

A Deeper Look at Bottom-up Copper Deposition in High Aspect Ratio through Silicon Vias

Superconformal copper electrodeposition in larger features, as in smaller features, is influenced by both suppressor and chloride in the copper electrolyte. Different behaviors are observed in electroanalytical measurements of low and high chloride electrolytes. One significant difference is the dependence of suppression breakdown and associated S-shaped negative differential resistance (S-NDR) on transport assessed using a rotating disk electrode. Superconformal filling of 56 μm tall, annular Through Silicon Vias (TSVs) from the electrolytes can exhibit the bottom-up evolution previously noted for the Cu system as well as transient or steady-state passivation breakdown along the TSV sidewall, with uniform deposition below, previously demonstrated with Au, Ni and Co S-NDR systems. As both filling geometries are predicted by S-NDR based models, I will discuss the factors that we believe determine the experimental filling geometry.

Exploiting the Spatial Homogeneity of Adlayer Breakdown on Microelectrodes to Develop a Kinetic Model for S-NDR Copper Electrodeposition

Copper electrodeposition is a major component in integrated circuit manufacturing, ranging across lengthscales from nanoscale damascene plating to 100s of microns for thorugh-silicon-vias used in 3D interconnect packaging.1 Critical to this process is electrodeposition of void-free copper structures in high aspect ratio cavities, which is achieved with various additives that enable superconformal bottom-up filling. This physicochemical phenomenon is driven by an S-shaped negative-differential resistance (S-NDR), which promotes spatial bifurcation of the electrode surface into active and passive deposition regions. 2 Electrode bifurcation is dependent on additive adsorption/desorption kinetics, chemical species transport, and electrode geometry. Prior studies have demonstrated that the active/passive bifurcation is dependent on electrode size and that reduced electrode dimensions will permit a homogeneous reaction across the electrode.3 In this work, microelectrodes are used to experimentally probe the critical breakdown behavior in acid copper-sulfate electrolytes in the presence of micromolar concentrations of chloride and a single polymeric additive. In-situ optical microscopy indicates that the entirety of the microelectrode surface (either 12.5 or 25 μm in diameter) is either active or passive during cyclic voltammetry measurements, permitting current to be normalized to the geometric electrode area. Experimental electrolyte parameter sweeps cover chloride concentration, polymer concentration, copper concentration, and polymer adsorption kinetics (i.e. polymer type). These measurements are compared to static macroscopic electrodes and rotating disk electrodes at dimensions known to exhibit active/passive bifurcation. Experimental measurements are fit to finite element method computations using a two-additive model tracking the adsorption and rate-dependent desorption for chloride and polymer species. This work is an extension of previous computational models for describing S-NDR behavior in single additive systems.4-5 Kinetic constants for copper electrodeposition as well as adsorption and desorption dependent rate constants for chloride and polymer are fit to experimental cyclic voltammograms. Electrochemical kinetic constructs using microelectrodes and a 2-additive adsorption/consumption model provide a more accurate representation of critical breakdown phenomenon in S-NDR electrochemical systems. References T.P. Moffat, J.E. Bonevich, W.H. Huber, A. Stanishevsky, D.R. Kelly, G.R. Stafford, and D. Josell, “Superconformal Electrodeposition of Copper in 500-90 nm Features”, J. Electrochem. Soc., 147 (2000) 4524.T.P. Moffat and D. Josell, “Extreme Bottom-Up Superfilling of Through-Silicon-Vias by Damascene Processing: Suppressor Disruption, Positive Feedback and Turing Patterns”, J. Electrochem. Soc., 159 (2012) D208.S. Bozdech, K. Krischer, DA. Crespo-Yapur, E. Savinova, and A. Bonnefont, “1/f2 noise in bistable electrocatalytic reactions on mesoscale electrodes”, Faraday Discuss., 193 (2016) 187.D. Wheeler, T.P. Moffat, and D. Josell, “Spatial-temporal modeling of extreme bottom-up filling of through-silicon vias”, J. Electrochem. Soc., 160 (2013) D3260.D. Josell and T.P. Moffat, “Superconformal bottom-up nickel electrodeposition in high aspect ratio through silicon vias”, J. Electrochem. Soc., 163 (2016) D322.

Drop-in Electrodeposition Processes for Void-Free, High Aspect Ratio Structures of Five Metals

Superconformal electrodeposition processes enable void- and seam-free metal filling of high aspect ratio features from tens of nanometers to hundreds of micrometers in size. There are two known mechanisms underlying such processes: Curvature Enhanced Accelerator Coverage (CEAC) and S-shaped Negative Differential Resistance (S-NDR). Voltammetric measurements on planar substrates can quantify kinetics in models based on both mechanisms, enabling accurate, a-priori prediction of filling from a particular electrolyte in a particular size and aspect ratio feature. I will summarize the chemistries and associated mechanisms for superconformal Cu, Ag, Au, Co, and Ni feature filling, including experiment and model predictions, for processes that are drop-in compatible with those presently used for fabrication of Cu damascene and through silicon via interconnects.

Oscillations in an array of bistable microelectrodes coupled through a globally conserved quantity.

The dynamical behavior of an array of microelectrodes is investigated under controlled current conditions during CO electrooxidation, a bistable electrochemical reaction with an S-shaped negative differential resistance (S-NDR) current-potential curve. Under these conditions, the total current constitutes a globally conserved quantity, thus coupling all microelectrodes globally. Upon increasing the total current, the microelectrodes activate one by one, with a single microelectrode being on its intermediate S-NDR current branch and the other ones being either on their passive or their active branches. When a few coupled microelectrodes are activated, the electrochemical system exhibits spontaneous potential oscillations. Mathematical analysis shows that oscillations arise already in a two group approximation of the dynamics, the two groups consisting of 1 electrode and n - 1 electrodes with n ≥ 3, respectively, with each group being described by a single evolution equation. In this minimal representation, oscillations occur when the single electrode is on the intermediate branch and the larger group is on the active branch.