Reshaping and induction of optical activity in gold@silver nanocuboids by chiral glutathione molecules

The interaction of chiral molecules with achiral plasmonic metal nanostructures has been attracting the attention of many scientists for over a decade, in the context of basic molecule–metal surface interaction as well as practical biomolecular sensing applications.¹ The model for induction of circular dichroism (CD) in localized plasmonic resonances through dipolar interaction of molecular excitation with plasmon resonance was originally described by Fan and Govorov.² Since then, different experimental and theoretical works described similar effects,^3–6 including studies with various types of metals,⁷ and creation of plasmonic hot-spots at inter-particle gaps in order to further enhance the chiroptical effects.^6,8,9 Several system parameters are important for optimizing the strength of the induced plasmonic CD effects: 1. It is essential to have a well-defined orientation of the molecules with respect to the particle surface; otherwise, the induced CD might average out to zero. 2. The effect increases with particle size, within the size range where particle size is still < light wavelength. 3. The induced CD effect should be inversely proportional to the spectral separation between the molecular and plasmon resonances. According to the last point, it is beneficial to use plasmonic resonances in the UV range, which would be close to biomolecular electronic transitions typically occurring at wavelengths <300 nm.

While the induction of CD at plasmon resonances of spherical Au or Ag nanoparticles is a weak effect,² typically undetectable with regular CD spectrometers, it has been shown that the effect becomes stronger in non-spherical particles,⁴ and particularly strong in certain high-order multipolar plasmon resonances in cuboidal Ag nanoparticles.^10–12 However, the small number of experimental results and lack of a model to describe the chiroptical effects involving non-spherical particles and high-order plasmon modes limit the possibility to further optimize such effects as a chiroptical sensing platform. One of the more interesting plasmonic systems for obtaining significant plasmon resonances in the near UV is using nanometric silver slab resonances, typically observed at different geometries of core–shell Au@Ag nanoparticles.¹³ Such systems were shown to have fairly significant transverse silver slab plasmon resonances at ∼350 nm, which is probably the blue-most localized resonance that one could obtain with silver (or gold) nanostructures. An additional advantage of such resonances is the relatively high photon energy absorbed by them, which could be useful for inducing hot-electron/hole photochemical effects at the surface of the particles,¹⁴ which is beyond the topic of this work.

The effect of biomolecules on plasmon resonances^15,16 and, in particular, on induced CD at such resonances has been a topic of increasing interest recently, with potential applications in chiral bio-sensing.^17–19 One of the model chiral molecules used by us and others for induction of plasmonic CD is l-glutathione (GSH).^11,12,20,21 This natural tripeptide has important biological functions, including as an anti-oxidant (reducing agent) in cellular processes, through transformation into its oxidized dimeric form (GSH + GSH = GSSG).

In this work, we explore near UV induced plasmonic CD effects in Au@Ag cuboidal nanoparticles interacting with GSH. In contrast to the simple cubic silver nanoparticles previously used, we chose to use shape-tunable elongated gold nanocuboids, which can be fine-tuned to become concave with sharp edges, and coat such nanoparticles by a silver shell of variable thickness. We used the chiral biomolecule GSH and found a surprising role that this molecule has not only in inducing CD activity at plasmon resonances associated with the silver shell but also as a shape modifier of the silver shell. We compared the GSH to its closely related chiral biomolecule l-cysteine and the thiolated chiral drug molecules, l-penicillamine and Captopril.

Gold nanorods were prepared following a previously published protocol.²² The preparation of a colloidal Au seed solution for nanorod growth is as follows: 10 ml of 0.25 mM HAuCl₄ and 0.1M cetyltrimethylammonium bromide (CTAB) solutions was prepared in a 20 ml glass vial. Then, 1 ml of ice-cold, freshly prepared 6 mM NaBH₄ was quickly injected under vigorous stirring, which resulted in the formation of brownish color. The solution was stirred for 2 min and kept undisturbed at room temperature for 30 min before use. To prepare an Au nanorod growth solution, 77 mM CTAB and 16.2 mM sodium oleate were prepared in 250 ml water at 60 °C. After the solution was cooled to 30 °C, 18 ml of a 4.0 mM AgNO₃ solution was added, and the mixed solution was kept undisturbed at 30 °C for 15 min. Then, 250 ml of a 1 mM HAuCl₄ solution was added, and after 90 min of stirring at 700 rpm, the solution became colorless. Then, 1.78 ml of HCl (32 wt. %, 10.2M) was added while stirring at 700 rpm for 15 min, followed by addition of 1.25 ml of 0.064M ascorbic acid. Finally, 0.8 ml of an Au seed solution was injected into the growth solution and was vigorously stirred for another 30 s and then left undisturbed at 30 °C for 12 h. The resulting Au nanorods were collected by centrifugation at 7000 rpm for 20 min followed by removal of the supernatant and redispersion of the precipitate in 30 ml of 0.02M CTAB. Formation of CTAB foam during the synthesis needs to be avoided to obtain a high degree of size uniformity.

Au concave-shaped nanocuboids (CCNCBs) were prepared via controllable overgrowth of Au nanorods following a previously published protocol, with minor modifications.²³ First, 200 μl of the nanorod solution was precipitated via centrifugation at 7000 rpm for 20 min and redispersed in 100 μl of 0.1M CTAB. 8.9 ml of a growth solution was prepared in a 20 ml glass vial with the following concentrations: 56 mM CTAC (cetyltrimethylammonium chloride), 2.2 mM CTAB, 0.23 mM HAuCl₄, and 5.6 μM of Cu(NO₃)₂. 2 ml of 0.1M ascorbic acid was then added to the vial. After gently mixing for 15 s, the growth of Au CCNCBs was initiated by adding 100 μl of the nanorod solution (in 0.1M CTAB). The reaction solution was gently mixed for 15 s immediately after the addition of nanorods and then left undisturbed at 25 °C for 25 min. The obtained Au CCNCBs were washed with water twice through centrifugation/redispersion cycles and finally redispersed in 50 μl of water.

Ag coating of Au CCNCBs was performed in the presence of polyvinylpyrrolidone (PVP), CTAC, and ascorbic acid following a previously published procedure,²⁴ with modifications: 50 μl of Au CCNCBs was added to a solution containing 48.4 mg PVP (MW 40,000) and 0.018M CTAC. 25–200 μl of a 4 mM AgNO₃ solution was added to this solution followed by 100 μl of 0.1M ascorbic acid. The reaction proceeded at 30 °C for 12 h. A color change could be observed after the reaction was completed and depended on the amount of added AgNO₃. The obtained Au–Ag core–shell CCNCBs were washed via centrifugation at 7000 rpm for 5 min and redispersed in 50 μl of water.

800 μl of the Au nanorod solution was centrifuged; then, the supernatant was decanted, and the precipitate was dispersed in an 80 mM CTAC solution and then centrifuged again and redispersed in 800 µl of a 80 mM CTAC solution. The nanorod/CTAC solution was then added to an Erlenmeyer flask with 100 ml of a 80 mM CTAC solution and heated to 70 °C with mild stirring. After 30 s, 5 ml of a 0.1M ascorbic acid solution was added. 5 min later, a 0.5M NaOH solution was added under vigorous stirring to adjust the pH to ∼3, and then, 2.0 ml of a 5 mM AgNO₃ solution was added. The solution was kept in the dark at 70 °C under vigorous stirring for 180 min. The resultant Au@Ag NCBs were centrifuged and redispersed with H₂O twice and kept at 4 °C for future use.

Solutions of Au@Ag CCNCBs or NCBs and thiolated bio-molecules (or drug molecules) were prepared in 20 ml glass vials: 50 μl of the Au@Ag nanoparticle solution was added into 2.8 ml of a 1.0 mM CTAC solution and mixed gently to obtain a homogenous solution. Then, a small volume of the chiral biomolecules stock solution was added into the vial to obtain a final biomolecule concentration of 30 μM and mixed gently for 10 s. The solution was kept undisturbed at 25 °C for 14–48 h. HNO₃ (0.001M) and NaOH (0.01M) solutions were used in order to adjust pH values of the solutions.

SEM imaging was done using a ZEISS Gemini 300 microscope. TEM imaging was done using a Technai F200 microscope, and scanning transmission electron microscopy - energy dispersive spectroscopy (STEM-EDS) chemical mapping of the nanoparticles was performed using an aberration corrected STEM Themis Z.

CD measurements were performed in a Chirascan (Applied Photophysics, UK) CD spectrometer. Absorption measurements were done in a fiber-optic based StellarNet Black Comet spectrometer.

The basic concept of this work was to find new nanoparticle platforms that will be used to extend the preliminary results on chiral molecule induced CD at the near UV previously obtained on silver nanocubes to other silver-based nanoparticle morphologies. Since a larger nanoparticle shape variety exists for gold nanostructures, and in particular, for gold nanorod-based shapes, combined with the need to obtain significant Ag slab resonances in silver shells, we decided to work with elongated colloidal Au nanostructures and then coat them with a silver shell, which would have a larger propensity to produce slab/multipolar plasmon resonances at near-UV wavelengths. Hence, we explored recently developed syntheses to obtain gold CCNCBs, as shown in Fig. 1, and coated them with silver shells of varying thickness. As can also be seen in Fig. 1(d), the transformation of the gold nanorods into CCNCBs done by over-growth of gold on the nanorods in the presence of CTAC, CTAB, HAuCl₄, Cu(NO₃)₂, and ascorbic acid resulted in the red-shift of both the transverse (from 520 to ∼600 nm) and longitudinal (from 870 to ∼1000 nm) surface plasmon resonance peaks and in a relative increase of the intensity of the transverse mode. The formation of a concave shape is driven by stabilization of high-index facets of the gold surface during the over-growth, which is contributed by the CTAC/CTAB surfactants together with the presence of Cu²⁺ (also Cu⁺ found on the surface), as extensively discussed in Ref. 23. A thin silver coating (∼5 nm) caused an opposite, blue shift in the two resonances, to 560 and 890 nm.

In addition, a small spectral feature appeared at ∼350 nm for the silver coated CCNCBs, indicating the appearance of a weak plasmon resonance. Resonance modes at this wavelength would typically correspond to a (dipolar) silver slab mode¹³ and, depending on the sharpness of the edges, could also exhibit a multipolar resonance nature with their associated electric field mainly concentrated at edges and apexes of the nanocuboids.²⁵ 

Theoretical modeling of the electromagnetic responses has been performed in order to allow for a correct interpretation of the optical data obtained experimentally. To do so, we have solved the classical electrodynamic response of the plasmonic particles using a solver that uses finite element methods (FEMs), COMSOL Multiphysics, and from these results, we have derived the relevant extensions for the different shapes and coatings (Fig. 2). The local dielectric constants used by us are those of Johnson and Christy.²⁶ 

Figure 2 displays a simulated extinction spectrum of a model CCNCB coated with a uniform thin silver film. The two different resonances can be clearly observed in the simulation results: 1. a strong transverse plasmonic extinction peak (T-plasmon) at ∼550 nm associated with the transverse dipole oscillation of the whole CCNCB and 2. a weak near UV peak at ∼350 nm corresponding to the silver slab dipole resonance. Hence, the simulation results are similar to the experimental (green) curve in Fig. 1(d). The simulation also reveals a very weak feature at ∼400 nm corresponding to the gold’s s–d inter-band transition, which is sometimes observable in the experimental spectra as a very weak, broad feature.

The addition of a small amount of GSH molecules caused slow changes (over many hours; see the supplementary material, Fig. S1) to the extinction spectra of the Ag-coated CCNCBs: The two main dipolar plasmon resonances have slightly red-shifted [see Fig. 1(d)], possibly indicating a slight morphological change, including some increase in the effective aspect ratio, and a near UV plasmonic feature has increased into a significant peak at 350 nm. Figure 3 displays extinction and CD spectra for a series of the same CCNBCs coated with an increasing Ag coating thickness, after incubation with a fixed amount of GSH molecules. Several interesting trends can be observed in the spectra: 1. There is a relatively strong negative CD peak associated with the Ag slab plasmon resonance at 350 nm and the magnitude of the CD peak is roughly proportional to the magnitude of the extinction peak. 2. The longitudinal plasmon peak gradually and monotonically shifts to the blue with increasing coating thickness, as expected from the decreasing aspect ratio of the particles. 3. There is an optimal thickness for which the slab plasmon resonance extinction peak and its associated CD peak are at maximum. 4. Additional positive and negative CD lines that scale similar to the peak at ∼350 nm occur at ∼290 and 220–230 nm, respectively. Those CD signals cannot be of plasmonic origin, as they are positioned below ∼320 nm, which is the bulk surface plasmon resonance value for silver. Hence, they should be related to molecular CD, which could be enhanced by the surface plasmon resonances of the metal nanostructure and, in particular enhanced by the resonances supported by the silver shell that may have a dominant contribution to local field enhancement.¹⁴ Pure GSH has CD lines only below ∼230 nm; hence, it can be excluded as the source of the 290 nm peak (see the supplementary material, Fig. S2). 5. A very small, noisy CD signal appears at the position of the transverse plasmon resonance around ∼600 nm for the Ag coating thickness, which produces the maximal slab CD signal at 350 nm. This is similar to the previously observed induced CD effects in silver nanocubes.^10,11

Additional useful information was obtained by changing the pH of the solution where the Ag-coated CCNCBs were incubated with GSH (see Fig. 4). The natural pH value of this solution is about 5. Changing the pH to ∼9 resulted in basically zero CD spectrum, also for the 350 nm resonance. However, on changing to more acidic pH values, it is possible to observe a significant change in the line shape of the CD peak at ∼350 nm. At pH ∼ 2, the peak becomes roughly antisymmetric, bisignate in shape. A similar effect occurs also for the weak induced CD at the transverse plasmonic peak at ∼600 nm. The relative intensity changes of the CD feature at ∼350 nm in the pH = 3–5 range were sample dependent, as different particle batches had slight morphology changes, which can yield different CD intensities after the GSH treatment.

This type of line shape change is known to occur when the conformation/orientation of the biomolecule changes with respect to the metal surface,² which would be expected to occur, as the pH value changes from ∼5 to 2. At this range, the two carboxylate groups of the GSH peptide become fully protonated, and thus, their interaction with the silver (oxide) surface might change. In addition, the silver surface itself might change the composition and charging state with this pH change.

Surprisingly, the CD peak at 290 nm, which must be associated with some molecular species, perhaps the GSH-Ag complex or dimer of oxidized GSH (GSSG-Ag),¹¹ does not change with the changes in the pH value. On the other hand, the CD line at ∼220 nm, associated with carbonyl groups, does change significantly with the pH value in the range of 2–5 either.

A comparison of GSH to other thiolated biomolecules and drug molecules: l-cysteine, l-penicillamine, and Captopril (thiolated chiral molecule based on l-proline), shown in Fig. 5, demonstrates the uniqueness of GSH. Incubation with all other thiolated chiral biomolecules at the same concentration as glutathione did not show any effect on the plasmon resonances nor appearance of the near UV CD line. It should be noted that all the molecules caused induction of a very small induced CD signal at the transverse dipolar plasmon resonance at ∼600 nm [see the inset of Fig. 5(b)], which serves as a proof of their adsorption to the silver surface.

Hence, it seems that the changes in the extinction and CD spectra on incubation with GSH must be related to unique physical changes in the silver coating due to the presence of the GSH. This assumption led us to perform high-resolution STEM-EDS chemical mapping of the Ag-coated CCNCBs before and after incubation with GSH, shown in Fig. 6 (for additional images, see the supplementary material, Fig. S3). It can be seen that before incubation with GSH, the CCNCBs are coated with a fairly uniform thin layer of silver of several nm thickness. After interaction with GSH, the coating seems less uniform, where the silver seems to be removed from the apexes of the nanoparticles and redeposited along facets, both at the long facets and at the ends of the nanorods. These thicker Ag film parts apparently give rise to the increased near-UV silver slab plasmon resonances and associated strong induced CD peaks. The weak CD signals appearing in the transverse resonance peak are not sensitive to the morphology of the silver coating, as expected from the regular dipole resonances involving the nanoparticle’s core. The sulfur distribution was also imaged using the STEM-EDS technique and revealed roughly uniform, low sulfur content across the surface of CCNCBs (supplementary material, Fig. S4). Since the GSH forms a monolayer on the surface of CCNCBs, such a low signal would be expected and would not permit for drawing further conclusions about its precise distribution between various positions on the surface.

It is now left to ask why is GSH so unique in its ability to rearrange the silver shell and obtain the increased slab plasmon resonances? It could be possible that binding of the GSH’s thiol group to Ag⁺ is energetically favorable (see Fig. 7), possibly with additional stabilization from the carboxylate groups, thus implying that glutathione molecules can pull out silver ions from the slightly oxidized silver surface. This may be even further facilitated at acidic pH. At basic pH, oxidation of the GSH’s thiol group to form the disulfide form might inhibit the possibility to remove silver ions from the surface. The GS(H)–Ag complex may then adsorb at different positions at the nanoparticle surfaces, possibly at more stable adsorption sites on the abovementioned facets, and re-deposit the silver ion and perhaps also reduce it while oxidizing to the disulfide state, possibly aided by the catalytic activity of the silver surface for this process.¹¹ It should be noted that GSH is considered a relatively potent reducing agent, compared to the other molecules that we tried, with the general reducing potential trend: GSH > cysteine > penicillamine^27,28 and also GSH > captopril, penicillamine.²⁹ 

Another open question is regarding the assignment of the pH independent CD peak at 290 nm. It might correspond to either GS-Ag complex or the oxidized GSSG form, which might be pH insensitive at the range of pH = 2–5. The CD spectrum of the GSH-Ag⁺ complex (see the supplementary material, Fig. S2, pH = 5) extends up to ∼380 nm, but its peaks are much broader than the observed 290 nm line. However, it is possible that the surface-adsorbed form of this complex, enhanced by the metal surface, could be the source of this peak, as also suggested for a similar signal observed in Ref. 11 for pH < 5. The concentration of the dissolved GSH-Ag⁺ complex in the CCNCB samples is probably too low (≪30 µM) to be observed by CD spectroscopy.

Consequently, we believe that while most of the discussed thiolated biomolecules could form stable complexes with silver ions,³⁰ only GSH has strong enough reduction activity to reduce silver ions to the metallic state (possibly even enhanced catalytically by the metal surface), leading to the unique function of GSH to restructure the silver shell around the CCNCBs, which leads to the appearance of a substantial Ag slab resonance peak. This peak, in turn, would be active in enhancing the molecular-Ag⁺ complex CD at ∼290 nm, which appears only for the GSH.

Another question refers to the comparison of a previous work on silver-glutathione chiroptical response,¹¹ where the induced plasmonic CD spectra differ from the current work, including in the pH dependence. The low pH results (pH = 4.5; see Fig. 3 of Ref. 11) had opposite CD polarity relative to the current work at the same pH value. In the previous work, the metal nanostructures were pure silver nanocubes, where the near UV plasmon resonances were thoroughly analyzed and found to be of highly multipolar origin.³¹ In the present work, the silver forms a very thin layer (slab) on top of the cuboidal gold nanoparticles and the associated slab resonances seem to be of highly dipolar origin. This should probably make a difference in the resulting induced plasmonic CD response.

Electromagnetic simulations of model Ag-coated CCNCBs reveal some interesting information about our system. Several schematic shapes were modeled, as shown in Figs. 8(a) and 8(b) where initially concave gold cuboids with the silver shell of varying uniform thickness were simulated. Then, the shape was modified to have non-uniform shell, thicker at the centers of the facets, to qualitatively match the observation in Fig. 6. Figures 8(c) and 8(d) show the simulated extinction spectra. Panel (d) of Fig. 8 has the spectra shown in panel (c) of Fig. 8 normalized to the height of the transverse plasmon peak appearing at ∼550–650 nm. It can be seen that as in the experimental case [panel (e)], increasing the thickness of the silver shells causes a blue shift in the main plasmonic peak. A small spectral feature associated with silver slab resonance appears at ∼350 nm with increasing thickness (Δ) of the silver shell. Transforming the silver shell to non-uniform, similar to the STEM observation in Fig. 6(b), further enhances the silver peak at ∼350 nm, as seen for b-NCB (NCB = nanocuboid with straight faces), i.e. thick enough shell, in accordance with the observation of the effect of silver shell redistribution after GSH treatment. In this case, a shoulder also develops to the blue side of the main plasmon peak, below 500 nm, bearing some similarity to the experimental case, where a shoulder appears around 520 nm.

We compared the effect of GSH on the silver coated CCNCBs to its effect on Au@Ag NCBs with straight facets. Figure 9 shows the CD and absorption spectra of a series of such nanocuboids with a thicker silver shell (about 8 nm thick) grown around gold nanorods to form elongated cuboidal shape. The absorbance spectrum before interaction with GSH already shows a series of slab plasmon resonances in the wavelength range of 350–390 nm. After GSH treatment, also here, induced plasmonic CD appears around 350 nm, including varying line shape with the pH value [Fig. 9(a)]. The line shape at pH = 3 is particularly complex and hints at CD contributions of multiple resonances. In the present case, also the molecular CD peak at ∼300 nm changes with pH. In addition, as in the CCNCBs, a very weak, broad CD signal appears for the main dipolar resonances around 650 nm.

Figure 10 compares the CD dissymmetry factor spectra for both Ag-coated CCNCBs and straight NCBs after interaction with 30 µM GSH for about a day with optimal pH conditions. It is evident that the CCNCBs produce stronger plasmonic CD signals at 350 nm, with much thinner silver shells. The dissymmetry in the case of the Ag-coated CCNCBs is approaching a value of 10⁻³, which is considered fairly large for such induced CD effects. Note that with increased silver shell thickness also in the CCNCBs, the 350 nm peak diminishes (Fig. 3, the case of 200 µl AgNO₃ added). The diminishing 350 nm peak for thicker Ag coating also broadened and slightly blue-shifted, relative to the peak observed for the thinner silver coating. It seems that the same effect occurs in the NCBs (which have much thicker silver coating) and in the simulated NCBs with thicker silver coating (Fig. 8).

It would be interesting to use the electromagnetic simulation results to better understand the evolution of the magnitude of the 350 nm extinction peak and the corresponding CD peak with silver shell thickness. Figures S5–S8 of the supplementary material analyze the local field distributions around the various simulated geometries described in Fig. 8. It can be generally seen that with straight facets of the Ag coating (b-NCB), the field intensity across the facets becomes significantly higher, giving rise to a stronger slab resonance at 350 nm, as observed in Fig. 8. However, this does not produce a stronger CD signal at 350 nm, compared to the CCNCB shapes. In addition, the thinner Ag coating of the CCNCBs (2.5 nm thickness) produces stronger fields along the edges compared to the thicker coating (5 nm). The situation after the GSH treatment (Fig. 6) was such that the coating became thicker at part of the CCNCBs surface area, giving rise to the appearance of the 350 nm extinction peak. This conforms to the simulation results, where thicker Ag coating produces stronger slab resonance at 350 nm. The magnitude of the 350 nm CD peak does grow in this case with silver thickness, up to a certain thickness, which is somewhere between 2.5 and 5 nm; however, the comparison to the lower CD intensity in the NCBs indicates that the UV CD signal not only depends on the slab thickness but also on the particular geometry. Unfortunately, a model for the induced CD signal in such a complex geometry is not available yet.

This work sheds light on the special behavior of chiral GSH molecules with silver surfaces and, in parallel, also on the topic of optimizing gold@silver nanostructure geometry for obtaining relatively strong near UV induced plasmonic CD lines. Geometric parameters, such as silver film’s thickness, spatial distribution, and surface curvature, seem to play an important role in determining the magnitude of the induced CD effect. Further work on the chemical analysis of the chemical species prevalent in solution during the GSH treatment is required in order to formulate a conclusive mechanism for the redistribution of silver by GSH. In addition, modification of the silver coating procedure would be required in order to imitate the effect of GSH treatment and produce Au@Ag particles that would produce induced plasmonic CD of any chiral molecules that adsorbs to the surface of the nanoparticles.

The supplementary material contains additional spectra for the molecular GSH and GSH-Ag+ species and evolution of CCNCB extinction spectra with time on reaction with GSH. It also contains additional STEM-EDS images and electric field distribution maps obtained from simulation for the 350 nm plasmon resonance for different Au@Ag CCNCB geometric models.

A.O.G. and G.M. acknowledge the generous support from the United States–Israel Binational Science Foundation (BSF) under Grant No. 2018050.

The authors have no conflicts to disclose.

Yael Levitan Engel: Data curation (lead); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – original draft (equal). Daniel Feferman: Data curation (supporting); Formal analysis (supporting); Investigation (supporting). Monika Ghalawat: Data curation (supporting); Formal analysis (supporting); Investigation (supporting). Eva Yazmin Santiago: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Software (equal). Oscar Avalos-Ovando: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Software (equal). Alexander O. Govorov: Conceptualization (supporting); Formal analysis (equal); Funding acquisition (equal); Investigation (supporting); Methodology (supporting); Project administration (equal); Software (equal); Supervision (equal); Writing – original draft (supporting). Gil Markovich: Conceptualization (lead); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (lead); Project administration (lead); Supervision (lead); Writing – original draft (equal).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.