Extreme atomic-scale surface roughening: Amino acids on Ag on Au(111)

Surface modification has become a progressively important field of study as the demand for y specialized technology at the nanometer scale has increased.^1,2 Studying materials at this scale requires specialized techniques for surface imaging and analysis. Scanning tunneling microscopy (STM) has been extremely useful in the acquisition of surface data due to its enhanced sensitivity and ability to obtain topographical data at the atomic scale.^3–5 However, this technique is most commonly performed at low temperatures (LT) and under ultrahigh vacuum (UHV)^6–8 in order to achieve the ideal parameters for atomic resolution. While these conditions provide a pristine environment, they can lack an applicability to real-world conditions. To remediate this, electrochemical scanning tunneling microscopy (EC-STM) can be used to image surfaces under ambient conditions to obtain a greater understanding of surface chemistry and reactivity.^9–12 The novel EC-STM technique used not only allows for the electrochemical control of the system but it also permits the simultaneous observation of the structural changes occuring on the surface.¹³ EC-STM is especially applicable for understanding the interaction between metals and molecules, and recently, specifically biological molecules.^14–19 

Thin films, a form of surface modification, have been extensively studied due to their ability to alter the physical and chemical properties of a surface by applying an extremely thin layer of material onto a substrate.^20–22 The ability of thin films to affect surface properties was previously demonstrated on Au(111) through the under potential deposition (UPD) of a Ag (1 × 1) monolayer to the surface, which resulted in a thermally stable film that could withstand temperatures as high as 1000 K.²³ Although the thermal stability of this monolayer was heavily studied, the chemical stability was not directly investigated. As detailed below, amino acids can be used to probe the chemical stability of the Ag layer thanks to their ability to form bonds with the surface, which can change based on the exact termination of said surface.

Understanding the interactions occurring at the metal–molecule interface are particularly important in evaluating the biocompatibility and feasibility of thin films in biosensors, medical implants, and other molecular nanoelectronics.^24–27 These interactions can completely transform a substrate and form unique topographical features. Creating new methods to etch or roughen surfaces, especially with a specific degree of isotropy, is being investigated for use in solar cells^28–30 and microelectronics.^31–34 However, this type of etching often uses harsh acids or toxic chemicals like cyanide, thiourea, and concentrated HCl.^35,36 Throughout the paper, the term “harsh” will be used to indicate strong acids/bases or chemicals that pose a significant health hazard. Thus, it is advantageous to investigate ways in which surfaces can be safely and effectively roughened, using mild chemicals while still maintaining selectivity.

Previous studies have demonstrated the ability of amino acids (AAs) and nucleic acids to etch metallic nanoparticles (NPs) without harsh and toxic chemicals.^37–39 It is suggested that this etching occurs through the nitrogen atoms in the molecules chelating with the atoms of a noble metal to remove them from the surface.^37–39 The literature also suggests that the bonding between AAs and Ag may be weaker than that of Au due to decreased polarizability and shortened 4d orbitals.⁴⁰ This difference in bonding can be investigated through a comparison of AA bonding on bulk Au versus bonding on an Ag monolayer. Bulk Ag is not used for ambient STM experiments because it cannot be cleaned under ambient conditions.

Surface investigations involving the etching of Au(111) frequently include the use of thiols and other molecules with sulfur containing groups, which are known to remove Au atoms from the top surface layer.^41–45 In contrast, this current study investigates the interaction between a modified metallic substrate and amino acids barren of sulfur. Furthermore, similar past studies investigating amino acids on bare metallic substrates found that the interaction resulted in the growth of metallic islands above the surface rather than the formation of etch pits.^14–16,46 Instead, the system described herein composed of a single layer of Ag adsorbed on Au(111) displayed a significantly different surface morphology after the adsorption of AAs when compared to those same molecules on a bare Au surface. The presence of the atomically thin Ag layer dramatically changed how those molecules interacted with the surface, and multilayer, atomic, surface roughening was observed. This work demonstrates how modifying a metal surface with even one monolayer of a different metal can dramatically alter the expected outcomes.

The data were collected using a Keysight/Agilent PicoScan 5500 EC-STM at ambient conditions. The STM was equipped with an internal bipotentiostat to apply a potential to the surface and to allow for cyclic voltammograms (CVs) to be measured. Apiezon wax coated, Pt_0.8Ir_0.2 tips (Keysight Technologies) with less than 70 pA of leakage current were used. An Au(111) sample with a surface roughness <0.01 μm, <0.1° orientation accuracy, and 99.999% purity (Princeton Scientific Corp.) was used as the working electrode and was housed in an electrochemical fluid cell. Nanoscience Instruments, Pt_0.8Ir_0.2 wire acted as the counter electrode. Due to size constraints of the fluid cell, a Pt wire (Alfa Aesar, ≥99.997%) was used as the pseudoreference electrode. All potentials listed were ultimately referenced to Ag/AgCl using saturated KCl. Platinum was experimentally determined to have an open circuit potential of +0.63 V vs Ag/AgCl. Because of its well documented interaction with Au(111), 0.1M HClO₄ (Fisher Scientific, ACS Optima Grade) was used as the electrolyte.^47–50 Images and CVs were collected with Keysight PicoView software and image processing was performed using Gwyddion version 2.49 (Czech Metrology Institute, Department of Nanometrology).

Prior to assembling the electrochemical cell, the Au(111) crystal was placed on a quartz disk situated on a stainless-steel plate. The crystal was then flame annealed at ∼1000 K in a H₂(g) flame for 15 min. Afterward, the electrochemical cell was assembled and inserted into the EC-STM. In order to prevent possible templating effects upon molecular adsorption, a +0.73 V potential was then applied to the surface for 2 h to remove the $22×3$ herringbone reconstruction.^47,49

The Ag thin film was applied to the Au(111) surface from a saturated AgCl solution in 0.1M HClO₄ that was sonicated with a Sonics VibraCell probe sonicator prior to use to ensure maximum solubility.²³ UPD was used to apply the thin Ag film due to its ability to electrochemically deposit a single layer of Ag atoms through a linear potential sweep, as observed in Fig. S1 in the supplementary material.⁷⁰ A CV was taken after Ag deposition to verify a change in surface chemistry consistent with the presence of a Ag monolayer [Fig. 1(b)]. The AgCl solution was then removed from the fluid cell and replaced with a 0.55 mM solution of L-isoleucine (Sigma Aldrich, ≥98%) in 0.1M HClO₄ which was deaerated with ultrahigh purity N₂(g) for 10 min prior to use. To determine the potential at which the amino acid interacted with the surface, a CV was taken prior to scanning. The potential indicated by the reduction peak, Fig. 1(c), was then applied to the surface while STM imaging occurred.¹⁴ 

This report examines the disturbance of the flat Ag(1 × 1)–Au(111) surface morphology under mild electrochemical conditions and in the presence of biological molecules. The change in surface morphology was observed after the UPD of a Ag monolayer followed by the adsorption of a simple amino acid to the surface through a constant applied external potential. After the initial flattening process, the bare Au(111) surface was imaged in order to confirm that the surface was devoid of surface structures like islands or etch pits, which would be indicative of contamination within the system [Fig. 1(a)]. A CV of the surface was also taken as seen in Fig. 1(a), showing no faradaic peaks. The absence of peaks indicated that no redox chemistries occurred within the electrochemical window under investigation.⁵¹ For reference, a full scale CV of bare Au(111) is also included in Fig. S2 in the supplementary material.⁷⁰ 

After the surface was deemed clean for investigation, the Ag monlayer was applied via UPD from a Ag source of saturated AgCl in 0.1M HClO₄. Although two distinct UPD potentials are possible for Ag on Au(111), the more negative potential (+0.3 V), referred to in Ref. 23 as region 2, was used. Region 2 was found to result in a close packed, (1 × 1) Ag layer as well as increased thermal stability of the system.²³ In contrast, the region 1 deposition at +0.87 V did not result in a thermally stable layer. To apply the Ag layer, the potential was swept from +0.63 to +0.38 V at a scan rate of 0.5 mV/s. The external potential was held at the final voltage for 20 min, ensuring maximun Ag coverage and resulting in a (1 × 1) layer.^23,52 The presence of the Ag monolayer was confirmed via CV, showing a reduction peak at +0.34 V [Fig. 1(b)]. Upon imaging the surface with STM, it was confirmed that there was no significant change in the surface morphology, as seen in Fig. 1(b). Because the UPD of Ag on Au(111) has been extensively studied in the past, a flat surface was expected.^23,52 Due to the enhanced thermal stability imparted by the Ag monolayer to the Au surface, it was concluded that the saturated Ag solution could be removed from the fluid cell without altering the monolayer deposited on the surface and further experimentation could follow.^23,52

Prior to the introduction of the AAs into the fluid cell, the solution was deaerated with ultrahigh purity N₂(g) for 10 min. This practice is a common electrochemical procedure and allows for O₂(g) and other contaminants to be purged from the solution.^53–55 L-isoleucine, a nonpolar, essential amino acid with a branched hydrocarbon sidechain, was chosen due to overlap with past/control studies.^14–16 After the addition of L-isoleucine into the fluid cell, a CV was taken in order to determine the potential at which the molecules interacted with the surface. An external, applied potential can be used to force an interaction of molecules with a surface that would otherwise be difficult at room temperature and in a liquid.^{14–16,18,56} In Fig. 1(c), a faradaic peak can be seen at ∼0.45 V, indicating a change in surface energetics from that of the Ag monlayer alone. Interestingly, past EC-STM studies have shown this potential to indicate a chemical mechansim involving the deprotonation of the molecule to form a bond with the surface.⁵⁷ Once this potential was applied, the surface was imaged in order to examine how the existing Ag monolayer paired with the adsoption of the AAs affected the system [Fig. 1(c)]. The adsorption of L-isoleucine proved to drastically perturb the existing, flat Ag monolayer. This perturbation was significant enough to cause an extreme layered modification to the surface, as observed in Figs. 2(c) and 2(d), suggesting a lack of chemical stability of the Ag film. Although modifications of crystalline metallic substrates by biomolecules have been observed in Refs. ,58–60, the atomic alterations observed in this study have yet to be seen under such mild solvent and electrochemical conditions.^61,62 While many investigations have examined the interactions between biomolecules and metallic substrates, most resulted in growth on top of the (111) substrate.^14–16,46 In the case of simple amino acids, they are known to immobilize and trap diffusing adatoms atop metallic surfaces, coalescing them into small clusters/islands [Figs. 2(a) and 2(b)].^14,46 The formation of metal islands on the surface has been the only notable surface deformation observed for the AA/metal systems for (111) morphologies. Surprisingly, the interaction of the molecules with the surface is extremely different in the presence of a single layer of Ag. The molecular interaction in this system resulted in a surface morphology most similar to surface roughening.

Previous studies involving the roughening/etching of metallic surfaces, specifically NPs, under different, often harsh conditions, with amino acids and other biomolecules suggest that the etching process involves the chelation of atoms removed from the surface by the nitrogen atoms in the AAs or nucleic acids.^37,38 One study hypothesized that the amines of the molecules coordinated with the Au atoms in a manner that allowed them to be removed from certain surfaces including Au nanoparticles.³⁸ Additionally, the etching of Au surfaces by thiols is a well-known and understood process, which results in small pits across the surface, caused by the removal of atoms in the top surface layer.^41–43 Importantly, in these previous studies, the observed etch pits were the depth of a single step and did not penetrate past the initial Au layer.^41,63–66

The roughening observed in the AA on Ag on the Au(111) system aggresively disturbed the Ag thin film and had the strength to penetrate past the surface and into the sublayers of the Au(111) crystal. The unique environment of the AA system not only allowed for possible penetration into the sublayer but also for the layered three-dimensional growth above the surface. Notably, the adsorption of AAs on bare Au(111), which served as control experiments, revealed that multilayer roughening did not occur in the absence of Ag.^14–16 Because this phenomenon was not seen with AAs on bare Au, Figs. 2(a) and 2(b), it can be deduced that the existing Ag monolayer played some role in destabilizing the metallic bonding of the Au(111) surface in addition to interacting with the AA molecules.

Figure 3 shows height profiles taken over various areas of the modified surface. Of experimental note, the reported height measurements have been calibrated using the known step heights of bare Au(111) and HOPG. The dashed line represents the height of a single Au step while the dotted-line represents the average depth of the layers. The small peak on one side of the step edges observed in the line profiles in Fig. 3(a) likely represent the small height contribution of molecules adsorbed at the step edge. The average depth of the steps was 0.25 ± .08 nm, well within the acceptable range of a gold and/or silver step, suggesting that this form of roughening attacks the surface layer-by-layer on an atomic scale. Importantly, the height difference of a Au step versus that of a Ag step is within ±0.004 nm, making it impossible to say whether a step is composed entirely of Au or Ag atoms, rather it is most likely a combination of both.^67,68 While the Ag remains segregated at the surface during the UPD process,⁵² the extreme amount of surface roughening observed presents the possibility that Au and Ag do not remain segregated upon the addition of the amino acid. It is important to note that the continuous application of the external potential did not result in the interminable roughening of the surface. This may indicate that the mechanism is dependent upon the availability of both Ag and AAs. The roughening possibly ceased due to the minuscule amount of Ag present in the system, acting as the limiting reactant. The explicit role Ag plays is unclear and further investigations into the exact mechanism will be addressed in future studies. Furthermore, the detailed chemical reactions describing the AA adsorption are also still under investigation but may involve a deprotonation reaction.⁵⁷ The growth mechanism of the islands/layers appears to not fall within Ostwald or Smoluchowski ripening as the islands do not appear to coalesce and larger islands do not grow at the expense of the smaller islands.⁶⁹ The area of the islands/layers remains steady over many scans indicating that they reach a stable size and do not change significantly over time once formed. While an exact time value is not known, the timescale for the restructuring appears to be quite fast as all of the restructuring takes place and is complete as soon as the STM tip approaches the surface after AA deposition.

The potential to transform a substrate and tailor its functionality or morphology for specific needs in optical electronics, molecular recognition, solar cells, biocatalysis, and biomedical advancements puts surface modification at the forefront of technological advancements. Exploring the interactions between a modified metallic substrate and various biomolecules can provide a better understanding of the fundamental interactions that occur at the complex molecule–substrate interface. To better understand the properties of electrochemically deposited Ag thin films, L-isoleucine was used to investigate the chemical stability of Ag monolayers on Au(111). Generally, thin films follow one of three known growth modes; however, isoleucine adsorbed on a Ag monolayer induces an alternative growth mode which is otherwise absent when Ag is not present. The resulting surface morphology was markedly different than that of a bare Ag thin film or AAs on bare Au(111), which suggests that the mere presence of an atomically thin layer of Ag alters surface reactivity, inducing a divergent interaction between L-isoleucine and the surface. This interaction resulted in the atomic-scale roughening of the surface beyond that of the first surface layer. The average depth/height of each layer was 0.25 ± 0.08 nm which is within the acceptable range of both an atomic step of Ag and Au. In the realm of surface modification, roughening has proven to be an effective technique to distinctly alter the surface morphology while increasing surface area and complexity. Importantly, this modification traditionally requires the use of harsh conditions, whereas this study demonstrates that it is possible for this process to occur at much milder conditions using biological molecules and at extremely low potentials. It is also a clear demonstration of how atomically thin layers can drastically alter observed bulk properties.

The authors thankfully acknowledge the financial support from the University of Tulsa. Specifically, this work was provided additional support by the National Science Foundation under Award No. OIA-1833019. K.P.S.B. was supported in part through the Graduate Research Grant Program through the Office of Research and Sponsored Programs at The University of Tulsa. E.A.C. and M.A.P. were supported through both the Chemistry Summer Undergraduate Research Program and the Tulsa Undergraduate Research Challenge offered through The University of Tulsa.

The data that support the findings of this study are available within the article and its supplementary material.