Self-Terminated Electrodeposition of Ni, Co, and Fe Ultrathin Films

Subscriber access provided by GAZI UNIV

Article

Self-Terminated Electrodeposition of Ni, Co and Fe Ultrathin Films Rongyue Wang, Ugo Bertocci, Haiyan Tan, Leonid A. Bendersky, and Thomas P. Moffat J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.6b01901 • Publication Date (Web): 25 Mar 2016 Downloaded from http://pubs.acs.org on March 31, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Self-terminated Electrodeposition of Ni, Co and Fe Ultrathin Films Rongyue Wang, Ugo Bertocci, Haiyan Tan,‡ Leonid A. Bendersky, Thomas P. Moffat

*

Material Measurement Laboratory, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, Maryland 20899-1070, United States ‡

Theiss Research, La Jolla, California 92037, United States

Abstract Self-terminated Fe-group metal (Ni, Co, Fe) electrodeposition occurs at potentials negative of the onset of water reduction where OH- generation leads to the formation of a blocking hydroxide monolayer. Quenching of metal deposition is accompanied by an increase in dissipative energy loss in microbalance experiments attributed to increased hydrogen bonding to the adjacent double layer. Pulse deposition at -1.5 VSSCE, in 5 mmol/L (NiCl2, CoCl2, FeSO4) - 0.1 mol/L NaCl pH 3.0 electrolytes yields fully coalesced ultrathin films of Ni, Co, Fe, or alloys thereof, on Au. The film thickness is controlled by the nucleation, growth, and termination dynamics constrained by the electrochemical cell time constant. Precipitation of bulk Ni(OH)2 and related phases is minimized by using short deposition times and dilute metal cation concentrations to limit supersaturation. The rapid deposition of smooth, compact ultrathin Fe, Co, Ni films should facilitate mechanistic and durability studies of Fe-group metal catalysis and the fabrication of emerging microdevices. * E-mail: [email protected]

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 30

Introduction Thin films of iron group metals (Ni, Co, Fe), alloys and their oxide derivatives are of interest to a wide range of technical endeavors from energy conversion to microelectronics. The oxides and oxyhydroxides are among the best known oxygen evolution catalysts while the metal, alloys and hydroxides make effective cathodes for hydrogen production from alkaline water electrolysis.1-10 Likewise, the proton intercalation properties of the oxide and hydroxide phases are important in battery reactions and supercapacitors.11-13 In addition to energy conversion and storage technology, the magnetic properties of iron group alloys and oxides are central to information storage and sensing technologies and will be important building blocks in emerging spintronic applications.14-17 The metallurgical and thermal properties of iron group alloys are also important in microelectronics; from Co liners for Cu interconnects18,19

to

constituents

of

active

and

passive

components

in

microelectromechanical (MEMS) devices, such as low coefficient of thermal expansion alloys, inductors, etc.20 Thin films of iron group elements and alloys are routinely prepared on planar substrates using physical vapor deposition methods such as evaporation or sputtering or by less expensive electrochemical deposition. Electrodeposition has been used to grow high quality epitaxial iron group metal thin films on a range of metallic and semiconductor substrates including single crystal surfaces of Au, Pt, Ag, Si, GaAs, etc.14,15,21,22,23 Magnetic measurements indicate that the quality of the resulting electrodeposited films can be similar to those produced by vacuum deposition methods, including molecular beam epitaxy. A distinct advantage of electrodeposition, compared with physical vapor deposition methods, is its straightforward extension to growth on non-planar substrates that includes through-mask, template electroplating and additive-derived superconformal growth in recessed surface features. Morphological evolution during electrodeposition derives from the competition between nucleation and growth, where the ion charge transfer kinetics and surface 2


Page 3 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


diffusion dynamics are influenced by potential and the adsorption of anions, cations and solvent. For homo-epitaxial deposition, growth proceeds by a combination and competition between island nucleation and step flow with roughening dependent on substrate orientation, step-edge barriers and/or growth defects such as stacking faults. Heterodeposition can lead to a wider range of morphologies depending first and foremost on the wetting properties of the substrate and its impact on the nucleation process. For iron group metal deposition, film coalescence at small overpotentials is accomplished within the first few monolayers of deposition on Au.14,15,23,24 In contrast, deposition on non-wetting substrates usually requires operating at much larger overpotentials to favor higher nucleation densities although this often leads to undesirable porous deposits during subsequent growth. Added complications associated with electrodeposition of iron group metals from aqueous electrolytes are proton and water reduction reactions that result in alteration of the pH adjacent to the electrode surface.25,26,27 In particular, water reduction at higher overpotentials and extended times yields significant hydroxide generation and possible contamination of the metallic deposit by precipitation of the respective hydroxide phases. Buffered electrolytes are often used to counter such pH changes adjacent to the working electrode.28,29, 30 Strategies that favor heterogeneous nucleation include pulsed potential or pulsed current methods to transiently create conditions of high supersaturation of hydroxide species. A different approach involves using underpotential deposition to create a wetting monolayer film that also serves as the reducing agent in galvanic exchange reaction to produce a more noble metal overlayer.31 More recently, self-terminating electrodeposition reactions have been uncovered that enable effective deposition of ultrathin films of Pt and Ir.32,33 Quenching of the metal deposition arises from interactions between the metal reduction process and electrolyte breakdown; namely, a saturated Hupd coverage terminates the Pt or Ir deposition reaction. Pulsed potential can then be used to modulate the coverage of the chemically reversible Hupd passivation layer in order to grow multilayered Pt or Ir deposits. The process is 3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

analogous to atomic layer deposition with potential modulation being used in place of reactant flow to control the deposition process. Further generalization of self-terminated electrodeposition reactions is indicated by a recent study of the effect of H3BO3 on elemental Ni electrodeposition on TiN where growth termination occurred during galvanostatic deposition.34,35 Ni2+ cation and proton reduction proceed in parallel while Ni(OH)2 precipitation on the surface of Ni nuclei was proposed to inhibit further growth and stimulate the formation of new Ni clusters.34 However, once the potential reaches the onset of H2O reduction, further Ni deposition ceased (i.e., the process was quenched), an effect attributed to Ni(OH)2 precipitation within the diffusion boundary layer.34 For electroplating at 10 mA/cm2, from 5 mmol/L NiCl2 at pH 3.0, growth terminated before significant coalescence of Ni clusters occurs. However, increasing the concentration to 15 mmol/L NiCl2 enabled the TiN substrate to be covered by a Ni thin film with high continuity.35 In this contribution, the quenching of Ni electrodeposition process at Au electrodes is examined under potential controlled conditions, using electroanalytical and gravimetric measurements combined with surface analytical and structural studies of the deposited films. Electrochemical quartz microbalance (EQCM) studies indicate that Fe-group metal deposition is terminated by reaction between hydroxide generated at the onset of water reduction and available Ni2+ species at the interface. The resulting adsorbed hydroxide layer serves as a blocking layer that passivates the surface towards further metal deposition at potentials within the self-terminating region. For the electrolyte concentrations (5 to 50 mmol/L) examined, the blocking layer also prevents heterogeneous formation of bulk phase Ni(OH)2 although extended electrolysis eventually leads to its precipitation in the electrolyte adjacent to the electrode, in line with previous work that focused on Ni(OH)2 deposition.36,37 Nonetheless, the blocking behavior can be lifted by adjusting the potential to values congruent with conventional metal deposition. The generality of the self-terminated growth phenomenon and its application to other iron group metals and alloys is demonstrated and discussed. 4


Page 4 of 30

Page 5 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Experimental section Gold seeded wafers (Si-Au) were prepared by physical vapor deposition. Single side polished N/Phos Si wafers with resistivity of 1-5 ohm-cm and thickness of 475-575 µm were cleaned by soaking in acetone for 2 min and rinsing with ultrapure water (18.3 MΩ), followed by isopropanol, and water. After drying in flowing Ar, the Si wafers were loaded into e-beam evaporation chamber and pumped overnight. The metal ingots were cleaned and stabilized by e-beam evaporation prior to the deposition of 10 nm Ti adhesion layer followed by 150 nm thick Au film at 0.1 nm/s. The as-deposited Au films were strongly (111) textured as revealed by symmetric X-ray diffraction. The Au covered wafer fragments were sliced with a diamond scribe and cleaned with Caro’s solution (also known as piranha solution, 3:1 volume mixture of H2SO4 (conc):H2O2 (30 %)) before each experiment. (Warning: Piranha solution should be handled with caution both with regards to combustion with organic materials and to gas release that accompanies aging). For preliminary experiments the Au working electrode (WE) substrates were masked with 3M electroplater’s* tape to expose a circular electrode area that was 4 mm or 5 mm in diameter. All chemicals used in this study were analytical reagent grade without further purification. (*Product names are included for completeness and do not imply NIST endorsement). The initial electroanalytical experiments were performed in a beaker using a NaCl saturated calomel electrode (SSCE) as a reference electrode (RE) and a graphite rod as the counter electrode (CE). In order to examine the impact of the electrochemical time constant in the deposition process, an electrochemical cell with fixed dimension was

constructed

by

machining

a

rectangular

channel

in

a

block

of

polytetrafluoroethylene (PTFE). The working and counter electrode were fixed at opposite end of the channel while the reference electrode was positioned at the top surface of the electrolyte forming a slight meniscus at a fixed distance between the other two electrodes as shown in Figure S1. The working electrode was held firmly against the channel face such that the electroactive area corresponds to the 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 30

cross-sectional area of the electrolyte channel and a uniform primary current distribution is established across the specimen. Controlling the separation between the working and reference electrode, dref-wk, enabled the solution resistance to be adjusted and defined. For the smallest dref-wk examined, a Ag/AgCl/Cl- (0 V vs.

Ag/AgCl/0.1

mol/L Cl- = -0.05 VSSCE) electrode with external opening diameter of 2 mm was used in place of the SSCE in order to ensure that the tip of the reference electrode could be placed immediately adjacent to the surface of working electrode. Herein, the reference electrode position is defined by the center of the capillary opening, with dref-wk of 1 mm corresponding to the smallest spacing examined. Following electrodeposition, specimens were emersed and rinsed with H2-purged water, dried with flowing Ar, and transferred under an Ar atmosphere to the UHV chamber for surface analysis. X-ray photoelectron spectroscopy (XPS) measurements were performed under ultra high vacuum (UHV) condition (typically 10-9 torr) using a Kratos Ananlytical AXIS Ultra DLD spectrometer with a monochromatic Al Kα source and 20 eV pass energy. Further details of the XPS analysis are given in the supporting information. Electrochemical quartz crystal microbalance (EQCM) experiments were performed in a three compartment electrochemical cell. A Luggin-Haber capillary from the reference electrode was fixed approximately 2 to 3 mm in front of the working electrode. The 100 nm thick Au substrate and the underlying thin Ti adhesion layer were deposited by physical vapor deposition onto AT-cut quartz crystal blanks with 5 MHz resonance frequency. The counter electrode was a Pt foil while a KCl-saturated calomel reference electrode was used that differs by less than by 5 mV from the NaCl-saturated version, 0.2412 VNHE vs 0.2360 VNHE. High purity Ar was used to deaerate the electrolyte in the main compartment before the experiments, while a small overpressure was maintained inside the cell during the measurements. Prior studies of underpotential deposition reactions on Au-covered quartz crystal indicate a typical roughness factor between 1.3 and 1.6; however, in the present work, the projected geometric electrode area is used for normalization. 6


Page 7 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scanning electron microscopy (SEM) experiments were performed using a Hitachi S4700 field emission SEM with an accelerating voltage of 20 kV, a 20 µA beam current and a working distance of 12 mm. A scratch was made across the entire sample surface to assist focusing during imaging of the Au and Ni regions, respectively. Transmission electron microscopy (TEM) cross-sectional lamellae of the Ni/Au samples were prepared by focused ion beam milling (FEI Nova 600* NanoLabTM). The Ni film surface was coated with carbon and Pt to protect the Ni from ion damage during the milling operation. The lamellae were cut and cleaned by Ga ions to a typically thickness of 50 nm to ensure good electron transparency. The samples were protected by Ar gas during sample transfer to minimize its oxidation. High angle annular dark field - scanning transmission electron microscopy (HAADF-STEM) experiments were performed on the prepared lamellae using a probe-corrected FEI Titan 80-300 microscope operating at 300 kV with a spatial resolution of 0.1 nm. The probe convergence angle is 24 mrad and the HAADF inner and outer collection angles are 70 mrad and 400 mrad respectively. 2-D energy-dispersive X-ray spectroscopy (XEDS) mapping in STEM mode was acquired by an EDAX r-TEM Si(Li) detector with solid angle of 0.1 sr.

Results and Discussion EQCM with dissipation monitoring was used to examine the electrodeposition of iron group elements on (111) textured Au thin films during linear scan voltammetry. Experiments were performed in pH 3, 0.1 mol/L NaCl electrolyte containing 5 mmol/L of the respective divalent iron group metal (Ni2+, Co2+, Fe2+) salt. Voltammetry reveals at least three distinct reactions in the respective systems. For the NiCl2 solution (Figure 1A) proton reduction on Au is first evident near -0.55 VSSCE (E0H+/H2=-0.413 VSSCE). Ni deposition initiates near -0.6 VSSCE and then strongly accelerates below -0.9 VSSCE to reach its transport limited maximum near -1.1 VSSCE 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 30

followed by t-1/2 decay of the diffusion boundary layer. Below -1.3 VSSCE the onset of water reduction to hydrogen and hydroxide is evident with a monotonic increase in current with potential. Simultaneous gravimetric measurements reveal the mass gain associated with the Ni deposition wave (Figure 1A). Conversion of the gravimetric data to current density, based on 2 electron equivalents for Ni2+ reduction, yields favorable agreement with the voltammetric wave for Ni deposition.

However, as the

applied potential moves below -1.3 VSSCE, further metal deposition, evident as mass accumulation, is quenched coincident with the onset of water reduction. The same potential dependent sequence of reactions, namely, proton reduction, metal deposition, followed by growth-termination at the onset of water reduction, are also observed for the Co2+ (Figure 1B) and Fe2+ (Figure 1C) electrolytes. For Co deposition, an additional current peak at -0.83 VSSCE accompanied by a mass gain of 225 ng/cm2 is ascribed to underpotential deposition of Co prior to the onset of bulk Co deposition. For all iron group metal systems, complete quenching of metal deposition occurs at potentials below -1.4 VSSCE. Ni deposition from different NiCl2 electrolyte concentrations reveals the same general behavior as shown in Figure S2, although the larger currents associated with higher Ni2+ concentrations lead to an increased IR drop evident as the peak shift to a more negative potential. The peak voltammetric current and total mass of Ni deposited prior to the termination of the growth process scale linearly with NiCl2 concentration. The current efficiency for metal deposition at small overpotentials, i.e. near -1.0 VSSCE, increases with NiCl2 concentration consistent with the diffusional constraint on proton reduction in the pH 3 electrolyte. Shifts in the resonant frequency of the quartz crystal microbalance are most often attributed to mass changes as captured by the Sauerbrey equation while the width of the resonance reflects dissipative losses within the system. For EQCM, coupling between the electrode surface and adjacent electrolyte results in simultaneous alteration of the resonant frequency and dissipative losses. The effect of changes in solvation, hydrogen bonding, and double layer structure on slippage or viscous 8


Page 9 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


coupling have been examined for several systems.38-40 The motional frequency response of the QCM is usually modeled by a series combination of inductance-capacitance-resistance (L-C-R) elements, known as a Butterworth-von Dyke equivalent circuit, where damping losses are captured by the resistance term.41 During voltammetric cycling of Au in the base NaCl electrolyte, resistance changes are less than one ohm (Figure S3), with a slight drop near -0.6 Vssce coincident with the proton reduction process. In the presence of NiCl2, a small increase in resistance, on the order of an ohm, accompanies the onset of Ni deposition near -1.0 VSSCE, as evident in Figure S2 for the 10 mmol/L and 50 mmol/L Ni2+ electrolyte. The increase indicates a small change in the coupling between the electrolyte and Ni surface as compared to the original electrolyte Au interface. Minimal further changes in resistance accompany subsequent Ni deposition for these conditions. However, at more negative potentials where quenching of the metal deposition occurs, a distinct increase in resistance occurs as shown in Figure 1A and Figure S2. For the 5 mmol/L and 10 mmol/L NiCl2 electrolytes, termination of the growth process is accompanied with a 7 ohm increase in the damping resistance. In a subset of experiments, the increase in dissipation was accompanied by a small peak (arrow in Figure S2D) in the faradaic current response between -1.30 VSSCE and -1.36 VSSCE that corresponds to a sub-monolayer integrated charge of 89.1 µC/cm2 after subtraction (linear) of the background process. One explanation is that adsorbed hydroxide formed during incipient water reduction combines with Ni2+ to passivate the surface against further metal deposition while increased hydrogen bonding between the adsorbed hydroxide species and water in the adjacent double layer leads to increased damping of the EQCM. The increased coupling most likely also contributes, albeit slightly, to the downshift in the resonant frequency in addition to that associated with metal deposition and hydroxide adsorption. EQCM literature indicates that formation of nanoscale bubbles leads to the opposite trend, i.e. an upshift in peak frequency and narrowing of resonance due to replacement of electrolyte by lower density gas, thereby allowing this to be excluded as the source of increased dissipation.38-44 Furthermore, the increase in current at more negative potentials demonstrates that 9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

although the passivating layer effectively blocks metal deposition, it continues to support hydrogen production from water. This latter observation is consistent with recent work suggesting the importance of Ni(OH)x2-x and related hydroxide species in activating hydrogen production during water electrolysis in alkaline solutions.6,7,8,9,45 A similar increase in EQCM dissipative losses also accompanies self-termination of Co and Fe deposition (Figure 1B and 1C), demonstrating the generality of this quenching process for iron group metals. Gravimetric and chronoamperometric measurements were also used to examine Ni and Co deposition at fixed potentials corresponding to two potential regimes, namely, that of stable, steady-state metal deposition and that associated with higher overpotentials where growth termination occurs. Upon stepping from 0.385 VSSCE to -1.13 VSSCE, the current transient follows an initial t-1/2 decay for diffusion limited Ni deposition and concurrent proton reduction that after ≈ 50 s transitions to a steady-state value determined by natural convection as shown in Figure 2. The corresponding mass transient follows a t1/2 dependence before settling to a steady-state deposition rate of 60 ng/cm2/s. Close to 9000 ng/cm2, or ≈ 50 monolayers of Ni, accumulate within 100 s forming a 10 nm thick overlayer (based on 181.3 ng/cm2 for a monolayer on Ni (111) with a d111interplanar separation of 0.2035 nm). As with the voltammetric study only a slight change in the EQCM dissipation, less than an ohm, is observed under these conditions. Comparison between the steady state current and rate of mass gain (averaged between 80s and 100 s) indicates the current efficiency is 54% with the remainder going to hydrogen production via proton reduction. In contrast, stepping the potential to more negative values, from 0.23 VSSCE to -1.5 VSSCE, results in an increment of 850 ng/cm2corresponding to 4.7 monolayers of Ni deposition within 2 s, followed by complete cessation of growth as indicated in Figure 2C by the absence of further mass gain during the subsequent 100 s. Quenching of the deposition reaction is accompanied by a 7 ohm increase in the damping resistance consistent with that observed during growth termination in the voltammetric experiment. After termination the dissipative losses continue to increase 10


Page 10 of 30

Page 11 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


albeit at a much slower rate. Beyond 5 s, the net mass signal decreases slightly while the dissipative response saturates. After 100 s, potential control is released and the electrode relaxes to the open circuit potential. This accompanied by a slight mass gain and drop in resistance although the dissipative character associated with the self-terminated surface remains largely intact reflecting the persistence of the adsorbed hydroxide layer under open circuit potential. Performing the same potential step experiment in the absence of Ni2+ reveals negligible resistance change although a reversible ≈170 ng cm-2 mass change is evident as shown in Figure S4. Potential pulse experiments with Co2+ electrolytes reveal similar resistance changes accompanying the quenching of Co deposition. The subsequent open circuit stability of the dissipative losses associated with hydroxide layer is sensitive to the time held at negative potentials and the hydrodynamic conditions that define the extent of alkalinity at the interface, as shown in Figure S5, where the potential was poised for different times (1, 3, and 6 s) in both quiescent (stagnent) and stirred solutions.

Under forced convection (i.e., magnetic

stirring), the change in the dissipative, resistive losses are complete before the Co metal film is dissolved (Figure S5A) while under free convection conditions the two processes are convolved (Figure S5B). The difference is attributed to the acid-base chemistry being “jammed” at electrode/solution interface under the quiescent condition46 in contrast to the rapid perturbation of surface chemistry induced by fluid flow.47 In contrast to self-terminated deposition at high overpotentials, no resistance change is evident during potentiostatic polarization in the conventional Co deposition regime (Figure S6) consistent with voltammetric results (Figure 1B). XPS was used to examine the EQCM electrode immediately following self-terminated Ni deposition at -1.50 VSSCE. The Sauerbrey-derived deposition mass is equivalent to 4.7 monolayers of Ni, which corresponds to a 0.95 nm thick film. The film thickness was evaluated using a simple substrate/overlayer model based on the ratio of Au 4f and Ni 3p peaks, shown in Figure 2E, normalized by the attenuation length of the respective photoelectrons and elemental sensitivity factors as detailed 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

further in the supplemental information.48 Multiple sites on the emersed EQCM specimen were examined as indicated in Figure 2F. Averaging over thirteen locations gives an overlayer thickness of 0.95 ± 0.03 nm in good agreement with the gravimetric result. XPS was also used to examine the time dependence of self-terminating deposition of the iron group metals and alloys. An important aspect of this experiment was the use of an electrochemical cell with an adjustable distance, dref-wk, between the working and reference electrodes. For deposition at -1.50 VSSCE with dref-wk= 1 mm, the thickness of the Ni films saturate at 0.33 nm, 0.83 nm and 2.24 nm following 1 s of deposition from 5 mmol/L, 10 mmol/L and 20 mmol/L Ni2+ solutions, respectively, as shown in Figure 3A. No further growth occurs even for films held at -1.5VSSCE for 50 s. A monotonic increase in film thickness with Ni2+ concentration is evident that is slightly greater than expected for first order metal deposition reaction kinetics. For this cell configuration, termination of the deposition process only requires 1 s in contrast to 2 s observed in EQCM measurements that were performed in a different cell with dref-wk ≈ 3 mm. The difference in quenching time (1 s vs. 2 s for EQCM) and film thickness (0.33 nm vs. 0.95 nm for EQCM) arise from constraints on the rate of deposition and growth termination due to the finite RC time constant and ohmic electrolyte losses, associated with the respective electrochemical cells. The combined effect of the RC time constant and ohmic IR losses is demonstrated by varying dref-wk to adjust the value of the electrolyte resistance, R, in the rectangular parallel plate electrochemical cell. The results are summarized in Figure 3B where the saturated thickness for self-terminated deposition increases monotonically from 0.33 nm to 2.72 nm as dref-wk increases from 1 mm to 10 mm. The limiting film thickness is also reached at later times as dref-wk increases. The RC delay enables significant conventional overpotential deposition to occur as the potential requires a finite time to transit from 0 VSSCE to -1.5 VSSCE during potential step experiments. Likewise, transient ohmic losses in the electrolyte can hinder the electrode potential from reaching the value required for self-termination of the 12


Page 12 of 30

Page 13 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


deposition reaction. When dref-wk ≥ 20 mm the steady-state IR losses are such that the electrode potential never reaches the value required for quenching the metal deposition reaction. For configurations with lower R values, similar dynamics are operative when the transient currents are high enough to result in IR losses that delay the onset of the quenching reaction leading to the deposition of thicker films. For the smallest dref-wk configuration examined, self-termination of the metal deposition is independent of potential below -1.3 VSSCE as shown in Figure 3C. The -1.3 VSSCE threshold corresponds to the onset of water reduction and hydroxide generation consistent with self-termination by formation of a submonolayer hydroxide-derived blocking layer. This is distinct from homogeneous Ni(OH)2 precipitation, or variants thereof (e.g., Ni(OH-)x(Cl-)y2-x-y), that can occur with extended electrolysis. Closer examination of the XPS spectra of Ni, (Figure S7 and Figure S8) reveal the surface oxide/hydroxide overlayer is similar to literature reports for Ni films oxidized in the ambient.49,50 The similarity introduces some ambiguity as to the actual state of the self-terminated surface relative to the effects of oxidation that might occur during specimen transfer to the XPS chamber. No significant Cl- was observed on the emersed self-terminated growth surface. XPS analysis of Co and Fe films grown under the same processing conditions revealed self-terminated growth results in films of similar thickness as summarized in Figure 3D-F.

Furthermore, employing electrolytes of mixed divalent metal cations

enables self-terminated electrodeposition of iron group alloys such as NiCo and NiFe binary alloys. As shown in Figure S9 the alloy films are of similar thickness to the elemental forms with the alloy composition being a monotonic function of the ratio of the respective cations in the electrolyte. Ion scattering spectroscopy (ISS) was used to determine the surface coverage of self-terminated Ni deposits on Au. Films grown at -1.5 VSSCE for 10 s in 5 mmol/L NiCl2 with dref-wk = 10 mm yield a saturated Ni overlayer thickness of 2.73±0.05 nm 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

as determined by XPS measurements. The ISS spectra given in Figure 4A show that the as-deposited film is covered with hydrocarbon contamination (304 eV) and oxide (417 eV) species. After a brief 10 s etch by 500 eV Ar+ ion bombardment, the contribution of the O peak increases relative to C and the Ni peak appears without any indication of Au thereby demonstrating complete Ni coverage of the Au substrate. A second etching cycle leads to a further increase in the Ni to O ratio and exposure of the underlying Au substrate. Similar specimens were examined by field emission scanning electron microscopy (FE-SEM). Imaging conditions were first optimized on the perimeter of the Au substrate (Figure 4B) and then the specimen was translated to image the Ni overlayer (Figure 4C) with only minor adjustment of the focusing conditions. Crystallographic contrast between individual grains is readily evident for the clean Au substrate while for the oxidized Ni overlayer the grain contrast is significantly attenuated reflecting the uniformity of the hydroxide-terminated surface. Similar ISS and SEM results were observed for a film deposited for a 100 s under same experimental conditions. The structure and morphology of a Ni film deposited in 5 mmol/L NiCl2 at -1.5 VSSCE for 10 s with a dref-wk=10 mm was examined by transmission electron microscopy. A high-resolution TEM (HRTEM) image of the cross-section of such film oriented along the [011] zone axis of the Au substrate is shown in Figure 5A. Although the Au/Ni interface is not clear in this image, its approximate position can be recognized and is marked by the red dashed line, the (1-11) interplanar spacing is 0.204 nm above the interface and 0.235 nm beneath it. The inset is a Fast Fourier transform (FFT) of the imaged region and clearly shows two distinct diffraction patterns of the face centered cubic (FCC) [110] zone axis. The inner diffraction spots with red arrows correspond to the Au substrate with a lattice parameter of ao=0.408 nm while the Ni overlayer has ao = 0.352 nm (indicated by green arrows). The diffraction spots are elongated in the direction of the surface normal, an effect of the semicoherent Ni/Au interface. The conformal nature of the overlayer is clearly evident by the chemical contrast 14


Page 14 of 30

Page 15 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


between Ni and Au in the HAADF-STEM image shown in Figure 5B. The signal intensity of the HAADF-STEM images is proportional to a power law of the atomic number (Z1.6~1.9).51,52

The large difference in Z between Au and Ni enables the

Au/Ni interface to be clearly delineated as marked by the white dashed line (Figure 5B). The Ni film is 2.7±0.2 nm in thickness, agreeing well with the XPS data, and its (111) interplanar separation clearly differs from that of the Au substrate. The error bar represents the standard deviation between 17 measurements from different regions across the specimen. The effect of subsequent oxidation, due to aging, and specimen cleaning are shown Figure S10. X-ray energy dispersive spectroscopy mapping of the film confirms the compositional uniformity of the Ni film on the Au substrate (Figure 5B inset). The Ni signal was derived from its Kα1 peak at 7.48 kV(green) while the Au signal was derived from its Mα1 peak at 2.123 kV (red) The 2.5 nm measured Ni overlayer thickness is in agreement with the surface analytical results. Multipulse potential deposition was used to probe the nature of the adsorbed hydroxide layer on the self-terminated surface using the conventional Ni deposition, at -1.0 VSSCE, as a probe reaction. Following 10 s of self-terminated deposition at -1.50 VSSCE, the potential was stepped to -1.00 VSSCE for 600 s where steady Ni deposition is known to occur on a fresh surface. The pulse sequence was then repeated two more times and the film structure examined to see if the species that cause self-terminated deposition effect subsequent conventional Ni deposition. A cross-section TEM image in Figure S11 shows that a rough, 40-100 nm thick, film is formed with a significant density of internal voids. The defective regions were also evident from top view SEM images (Figure S12A). In contrast, no voids were present in control specimens grown at -1.00 VSSCE for 1800 s (Figure S12B). Despite the presence of the voids in the multipulse film, close inspection of the void–free areas at high magnification, Figure 5C, reveal numerous through-thickness columnar grains that are semi-coherent with the underlying Au. The interface between the Au substrate and Ni film is clearly evident by the high Z-contrast while the FFT (insert in Figure 5C) reveals the lattice spacing of the respective layers. In these regions, there is no 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

evidence of disruption of the crystal growth following the potential step from -1.5 VSSCE to -1.00 VSSCE (Figure 5C). This observation is evident for the first deposition cycle where, as demonstrated earlier, self-terminated growth results in a ≈ 2.7 nm Ni film that is subsequently deposited by conventional Ni deposition at -1.00 VSSCE. The continuity of the Ni lattice indicates that the quenching character of the adsorbed hydroxide layer can be lifted.

The strategy of using self-terminated deposition to

obtain full coalescence seed layer for subsequent film deposition should be a viable process provided the potential pulse sequence can be optimized to insure complete dissolution of the hydroxide layer to allow smooth void-free films to be formed. This potential controlled process is analogous to a recently detailed two-step galvanostatic process for Ni deposition.35 Self-terminated deposition of the iron group metals has been demonstrated in both voltammetric and potential step experiments whereby the freshly deposited metal serves as a catalyst for H2O reduction that generates the passivating OH- species. Quenching of metal deposition is accompanied by an increase in viscoelastic coupling between the electrode and electrolyte that can be ascribed to hydrogen bonding interactions between the OH- covered electrode and water in the adjacent electrolyte. This hydroxide layer blocks the metal deposition reaction while supporting on-going water reduction. The resulting pH shift in the adjacent boundary layer expands with electrolysis and a milky white precipitate, e.g. Ni(OH-)x(Cl-)y2-x-y, appears with prolonged time at the deposition potential. The propensity for metal hydroxide precipitation increases with metal cation concentration in line with thermodynamic expectations.37,53 The density and distribution of precipitates is sensitive to the electrodeposition conditions, and underestimation of the amount of precipitate due to rinsing for ex situ measurements is possible, perhaps even likely. Nevertheless, by using a low concentration of iron group cations and removing the specimens immediately after growth termination, significant hydroxide salt precipitation can be avoided. A previous study of heterogeneous Ni(OH)2 electrodeposition from a Ni(NO3)2 16


Page 16 of 30

Page 17 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


electrolyte dissolved in a 50 % (volume) ethanol and water mixture (1.5

Self-Terminated Electrodeposition of Ni, Co, and Fe Ultrathin Films

Recommend Documents