The nitrile stretching frequency of pure MBN SAMs on Ag and Au electrodes, both roughened on the 100-nm scale (see experimental section) was measured by SER and SEIRA spectroscopy, respectively. In the absence of any solvent, i.e., for the SAM/air interfaces, the C≡N stretching frequency was determined to be 2230.0 ± 1.0 and 2225.8 ± 0.2 cm⁻¹ for Ag/MBN and Au/MBN, respectively (Figure 2). In the case of Ag/SAM, the frequency increases upon addition of DMSO/buffer solutions with increasing amounts of aqueous buffer (from 0 to 100%) up to a value of 2235.1 ± 1.0 cm⁻¹ at 100% aqueous buffer. For Au/SAM in contact with DMSO/buffer solutions, the C≡N stretching frequency first drops considerably (from 0% to 10% buffer) and then steadily increase up to a value 2228.6 ± 1.0 cm⁻¹ in 100% buffer solution. This value is essentially the same as that determined also for MBN SAM on Au in 100% water (2228.4 ± 1 cm⁻¹).

Formation of a MBN SAM on Au and Ag electrode surfaces within the electrochemical cell was monitored by CV (Supplementary Information). The nitrile stretching mode of MBN was probed as a function of the electrode potential, which was varied from −0.4 to +0.6 V and from −0.5 to +0.1 V for Au/SAM and Ag/SAM, respectively. The potential limits were defined by the onset of metal oxidation and reduction of the metal-sulfur bond. A selection of SEIRA and SER spectra is shown in Figure 3. The frequency linearly increases with increasing potential, both for the Au and the Ag electrode. The slope and intercept of the linear fits to the experimental data are similar for the MBN SAMs on Au and Ag (Figure 4). There is no indication of potential-dependent variations in the band profiles. However, in both cases, peak intensity decreases significantly as the potential becomes more positive (Figure 5). This potential-dependent intensity variation has been found to be reversible.

To probe the behavior of the MBN nitrile stretching mode within an environment of charged residues, SEIRA-active Au electrodes were coated with SAMs of mercaptohexanoic acid (MHA). Subsequently, MBN was added to the solution under controlled potential. To integrate MBN into a MHA monolayer, a potential of −0.4 V was applied to remove a fraction of MHA molecules from the surface such that binding of MBN from the bulk solution to the vacant adsorption sites became possible. The process was monitored by SEIRA spectroscopy on the basis of the decreasing band intensities for the carboxylate stretching modes of MHA and increasing intensity of the nitrile stretching of MBN (Figure 6). After ca. 12 hours, the SEIRA intensity of the nitrile stretching has reached a value of ca. 4% of that for the pure MBN monolayer at the same potential, reflecting the relative contribution of MBN to the composition of the mixed monolayer. Subsequently, potential-dependent SEIRA measurements were carried out to probe the nitrile stretching frequency of MBN within an environment dominated by MHA with protonated and deprotonated head groups. Similar to the pure MBN, the frequency follows a linear relationship with the electrode potential, albeit with slightly different slope and intercept (Figure 4). The larger slope, compared to Au/MBN, may imply that the nitrile groups are less exposed to the aqueous phase [16] and are instead partly buried within the SAM.

Incorporation of MBN into a thiophenol SAM on an Au electrode was achieved in a similar way, however, at an electrode potential of 0.0 V. The nitrile stretching was detected by SEIRA spectroscopy. This band remains after removal of excess MBN from the cell and careful rinsing the electrode (Supplementary Information). Detachment of MBN from the electrode was only possible by reductive desorption, i.e., upon applying an electrode potential of −1.5 V that led to an irreversible loss of the band. Potential-dependent SEIRA measurements demonstrated again a behavior similar to that of a pure MBN SAM (Figure 4). In this case, however, the slope is lower than for the pure SAM, possibly indicating that in this system the C≡N groups are more exposed to the solvent [16].

For the analysis of the VSE of MBN we use the linear dependence of the C≡N stretching frequency ν on the electric field E F → according to Equation (1):

where Δμ⃗ is the difference dipole moment (between the ground and excited vibrational states) and ν₀ is the frequency in the absence of an external electric field. Neglecting the higher terms in the electric field dependence is justified a posteriori (vide infra). For MBN and other nitrile-containing VSE labels, Δμ⃗ has been determined both in frozen 2-methyl-tetrahydrofurane glass and after attachment to a protein [8,16], indicating that this quantity is lowered by about 20% in aqueous solutions compared to a hydrophobic environment [16]. Accordingly, we have scaled the value of Δμ⃗ = 7.4 × 10⁻⁹ cm⁻¹·V⁻¹·m, determined for MBN in the hydrophobic solvent 2-methyl-tetrahydrofurane [8], by a factor of 0.8 to obtain an approximate value for MBN SAMs with the nitrile function exposed to the aqueous solution, i.e., Δμ⃗ = 6.0 × 10⁻⁹ cm⁻¹·V⁻¹·m. The value for the zero-field frequency ν₀ determined for a protein-attached MBN in an aqueous environment, as well as that determined for MBN in a frozen 2-methyl-tetrahydrofurane glass cannot be adopted for MBN SAMs, as electric field and hydrogen bonding interactions, as well as binding to the different metals and the surface potential of the SAM may affect the C≡N stretching frequency.

The Au/MBN/air and Ag/MBN/air systems present C≡N stretching frequencies at 2225.8 and 2230.0 cm⁻¹, respectively. For comparison, the C≡N stretching frequency of benzonitrile in the gas phase is observed at 2238 cm⁻¹ [17], indicating a decrease in the C≡N stretching frequency upon metal binding and SAM formation.

Upon inserting the MBN-coated metal electrodes into a DMSO solution, the polar solvent molecules tend to orient in the SAM/solvent interface. The resulting surface potential causes a downshift of the C≡N stretching frequency by 2.6 cm⁻¹ in the case of Au. Taking into account the potential-dependence of the C≡N stretching frequency (Figure 4), a frequency downshift is equivalent to a negative shift of the electrode potential. Accordingly, we conclude that the DMSO molecules are preferentially aligned with the partially negatively charged oxygen atoms facing the SAM surface. Conversely, a 2-cm⁻¹ upshift is observed in the case of Ag/MBN in contact with 100% DMSO, pointing to the opposite orientation of the DMSO molecules with respect to the SAM surface.

In addition to the above observations, the response of the C≡N stretching frequency to the admixture of aqueous buffer is different for Au/MBN and Ag/MBN. The further decrease of the frequency at small amounts of buffer for Au/MBN may be attributed to the parallel orientation of DMSO and the even more polar water molecules, both with the oxygen atoms pointing to the SAM, such that surface potential becomes more negative. However, with increasing buffer content, hydrogen bonding interactions of water molecules with the nitrile group gain importance. Such interactions have been shown to exert the opposite effect on the C≡N stretching frequency [9,12,13] which eventually dominates and thus leads to an increase of the frequency with the buffer content. In contrast, in the Ag-bound MBN SAM the nitrile stretching frequency increases monotonically, since in this case the solvent molecule (DMSO, H₂O) orientation and hydrogen bonding affect the nitrile stretching in the same direction. It is interesting to note that the C≡N stretching frequency of MBN in DMSO solution was recorded at 2227.7 cm⁻¹ and the corresponding value for an aqueous solution was estimated to be 2233.6 cm⁻¹ (Supplementary Information). These values lie between the respective frequencies for the Au- and Ag-bound MBN SAM, consistent with the conclusion that the SAM-coated metal surfaces are oppositely charged at open circuit.

On the basis of these results we may now estimate the zero-field frequency ν₀ of the Ag/SAM/buffer and Au/SAM/buffer systems, i.e., the frequency that includes the effect of metal-sulfur bond formation and the hydrogen bonding interaction in the aqueous solution while excluding the effect of the (open circuit) metal electrode potential and thus that of the surface potential at the SAM/solution interface. This value is experimentally not accessible. However, we may assume that the effect of hydrogen bonding on the C≡N stretching frequency in an aqueous environment is independent of the existence of the metal-sulfur bond. Correspondingly, the frequency upshift of 7.6 cm⁻¹ (Δν_HB) from solid MBN (2226 cm⁻¹ [11]) to MBN in aqueous solution (2233.6 cm⁻¹) (Supplementary Information) should be approximately the same as the frequency shift arising from the transfer of the Ag/MBN/air (2230.0 cm⁻¹) and Au/MBN/air (2225.8 cm⁻¹) systems into an aqueous solution in the absence of any effects due to the electrode charge. Correspondingly, the addition of the Δν_HB increment of 7.6 cm⁻¹ to the frequencies of the metal/MBN/air interfaces is considered to be a reasonable approximation for the zero-field frequencies and we thus obtain ν₀ = 2237.6 cm⁻¹ and ν₀ = 2233.4 cm⁻¹ for the MBN SAMs on Ag and Au, respectively.

On the basis of an electrostatic description of the interfacial potential distribution for SAM-coated devices by Smith and White [4] and its further developments [14,15], the electric field at the SAM/solution interface is given by Equation (2):

where ɛ_c and ɛ_s are the dielectric constants of the SAM and aqueous solution, respectively; d_c is the thickness of the MBN monolayer, ɛ₀ denotes the permittivity, and κ is the reciprocal Debye length. The term σ_c refers to the charge density on the SAM surface which for carboxyl-terminated SAMs has been related to apparent pK_a of the head groups. In a more general sense, σ_c may be regarded as the charge density in the inner Helmholtz layer reflecting the contributions of charged SAM head groups as well as those of the ions and water dipoles aligned in the immediate vicinity of the SAM surface. This surface charge density, originally assumed to be constant, varies with the electrode potential E, or more precisely, with the difference between the electrode potential E and the potential of zero charge of the metal E_pzc. Here, we assume in a first approximation a linear relationship according to Equation (3):

where k is a constant and σ₀ is the charge density at E_pzc. Equation (3) may be combined with Equation (2) to afford Equation (4)

and after rearrangement:

Note that Equation (5) indicates a linear relationship between the electrode potential and the electric field which, taking into account the observed linear correlation between the nitrile stretching frequency and E (Figure 4), justifies neglecting higher order terms in the analysis of the VSE (cf. Equation (1)). For the electric-field dependence of the C≡N stretching frequency, we have to take into account the definition of the nitrile dipole axis pointing from the carbon to the nitrogen such that Δμ⃗ has a negative sign [18]. Thus combining Equation (5) and Equation (1) affords:

where θ refers to the angle formed by the vector of the nitrile dipole moment and that of the electric field, arising from the scalar multiplication (see Equation (1)). Rearranging Equation (6) leads to:

where the first two terms on the right hand side of Equation (7) account for the intercept (b) of the linear fits in Figure 4, and the last term corresponds to the slope (m). From the slope term one may determine the quantity k for the various metal/SAM interfaces (Table 1).

For these calculations, estimates for the tilt angle of the nitrile group are required. For MBN on Au θ was determined to be ~40° [24], which is similar to that for a thiophenol monolayer on Au (49°–54°) [25,26]. Reported angles for MBN and thiophenol on Ag, however, differ significantly, i.e., 24°–28° for thiophenol on Ag [27,28], and, most surprisingly, ca. 0° for MBN [29]. In view of this striking discrepancy between thiophenol and MBN on Ag, which is in sharp contrast to the findings for Au, one has to consider the reported value for Ag/MBN with caution. Nevertheless, for the present calculations we first adopt 49° and 0° for Au/MBN and Ag/MBN, respectively. These values are also included in determining the thickness of the SAM, d_c, using an Ag-S bond length of 2.4 Å and Au-S bond length of 1.9 Å [30,31], a sulfur-nitrogen distance of 6.5 Å as well as the respective values of θ discussed above. As a result, d_c is evaluated to be 6.8 × 10⁻¹⁰ m and 8.9 × 10⁻¹⁰ m for Au/MBN and Ag/MBN, respectively.

Now we consider the potential at which the electric field at the position of the nitrile group is zero (Equation (5)). This potential may be considered as the effective potential of zero-charge E₀ for the entire metal/SAM/solution system. Expressing Equation (7) for E = E₀, in terms of the slope m and intercept b of the linear fits to the experimental data in Figure 4, we then obtain

Note that E₀ may, therefore, differ substantially from E_pzc, which refers to the potential of zero-charge of the bare metal. These experimentally determined E₀ values are listed in Table 1. We further use these E₀ values obtained in this way to calculate the charge densities in the inner Helmholtz layer at zero-field (σ₀) according to Equation (9):

which is derived from Equation (7) when E = E₀ and ν = ν₀.

Equation (5) may now be used to calculate the variation of the electric field strength at the SAM surface, i.e., at the position of the nitrile group, as a function of the electrode potential (Figure 7). For all metal/SAM devices, the modulus of the field strength increases with decreasing potential, covering a range from 0.9 × 10⁸ V/m to −2.5 × 10⁹ V/m. Whereas similar values are obtained for Au/MBN and Au/thiophenol/MBN, the modulus of the field strength is larger in the mixed MHA/MBN SAM on Au.

As judged from the SEIRA intensity, this mixed SAM includes only 4% MBN, however, it is not known if the incorporation of MBN into a preformed MHA SAM leads to a largely homogeneous MBN distribution or to MBN islands within the MHA SAM. In the latter case, one would expect an inhomogeneity of the SEIRA bands, given that there are different electric field strengths at the surface of the dominant MHA part of the SAM and the much smaller MBN islands, leading to different nitrile stretching frequencies in the interior of the MBN island compared to the MBN/MHA boundary. Such inhomogeneities are not observed in the SEIRA spectra of Au/MHA/MBN, implying that the electric field at the SAM surface does not display significant gradients parallel to the surface or that the distribution of MBN within the MHA SAM is largely homogeneous. In any case, we may conclude that (i) MBN probes the electric field on the MHA SAM surface, and (ii) the magnitude of the electric field on a MHA SAM on Au at pH 7.0 is only larger by ca. 15% compared to Au/MBN, despite the partial dissociation of the carboxylate head groups at pH 7.0 [5,14,32,33]. This conclusion is consistent with the finding that the magnitude of the electric field on Ag/MBN, which is ca. twice as large as that found for Au/MBN, is quite similar to that estimated for Ag/MHA monolayers [14].

The quantity E₀ derived from the experiments reflects the change in the potential of zero charge of the pure metal (E_pzc) due to binding of the dipolar SAM to the metal surface. The effective potential of zero charge E₀ can also be estimated by calculating the change in the work function of the metal (ΔΦ) due to the coating by the SAM. In one approach, suggested by Wang et al. [22], the only parameters affecting ΔΦ are assumed to be the direction of the dipole moment of the SAM and its angle with respect to the surface normal (β), which can be evaluated from the tilt angle θ for the nitrile function, taken to be 49° and 0° for Au/MBN and Ag/MBN, respectively (vide supra). On the basis of the ΔΦ vs. cos(β) relationship determined by Wang et al. [22], one then obtains ΔΦ = 0.15 and 2.6 eV for Au/MBN and Ag/MBN respectively.

For bare metals, the work function Φ_M is related to the potential of zero charge according to:

where here E_pzc is expressed in V vs. NHE and K is a constant that was determined to be 5.01 and 4.61 eV for Ag and Au, respectively [34,20]. Recommended values for the work function of bare polycrystalline Ag and Au are 4.3 and 4.88 eV [21,35] such that one obtains −0.71 V and +0.27 V for E_pzc of polycrystalline Ag and Au, respectively. Correspondingly, Equation (10) allows evaluating the effective potential of zero charge from the work function of the SAM-coated metals Φ_M/SAM according to Equation (11):

such that one obtains, after translating the reference potential scale from NHE to Ag/AgCl, E₀ = 0.215 and 1.658 V for Au/MBN and Ag/MBN, respectively. The values are in reasonable agreement with the data derived from the VSE analysis (Table 1), specifically in view of the underlying approximations of both the electrostatic model and the E₀ calculations.

An alternative for calculating ΔΦ, suggested by Heimel et al. [23,36], includes the metal-sulfur bond dipole, BD, and the potential difference across the dipolar SAM, ΔV_vac, according to Equation (12):

here ΔV_vac is given by Equation (13):

where μ_⊥ is the dipole moment component of the SAM in the direction of the surface normal, A is the area of the unit cell, and ε_eff is a correction coefficient suggested to be 2.3 for a SAM of 4-mercaptobiphenylnitrile [23]. Estimating μ_⊥, even if β is known, is not straightforward as it differs from the mere projection of the single MBN molecular dipole onto the surface normal, due to the depolarization of the MBN dipole by neighboring molecules in the SAM [37]. According to Natan et al. [37], we adopt a value of 3.6 Debye for the dipole moment of the MBN SAM, which is then multiplied by cos(β) to afford μ_⊥. On the basis of calculated values for a series of related systems [23], BD is taken to be −1.05 eV and −0.27 eV for Au/MBN and Ag/MBN, respectively. The difference between these two values is in good agreement with experimental data reported by Alloway et al. [38]. Thus, one obtains 0.3 eV (Au/MBN) and 2.9 eV (Ag/MBN) for ΔΦ and eventually E₀ = 0.402 and 1.936 V vs. Ag/AgCl for Au/MBN and Ag/MBN, respectively. Here, E₀ for Au/MBN is much closer to the experimentally derived value (Table 1) than in the previous approach, while there is a stronger deviation from E₀ for Ag/MBN. The failure to reproduce the experimentally derived value for Ag/MBN in a satisfactory manner does not necessarily argue against the approach proposed by Heimel et al. [23]. In view of the very good agreement for Au/MBN, one may also question the most uncertain parameter in the calculation for Ag/MBN, i.e., the tilt angle of the nitrile function (θ) and thus the angle of the dipole moment of the entire molecule with respect to the surface normal (β). One may, therefore, take the approach suggested by Ballav et al. [16] to estimate the angles β and θ starting with E₀ = 1.277 V, i.e., the experimentally derived value. Under these conditions, a β angle of 38° is predicted, corresponding to a tilt angle (θ) of 26°, which is in good agreement with those reported for thiophenol on Ag [27,28].

Either way, it should be noted that previous experimental and theoretical studies on SAM structures and specifically on the tilt angle refer to the SAMs on smooth metals in air or in vacuo. Possible structural differences of the SAMs due to the nanostructured metal surfaces or the contact with an aqueous solution or in an electrochemical environment are not considered. Moreover, the present analysis tacitly assumes that the tilt angle of the nitrile function is potential-independent. The variation in the SEIRA peak intensities in Figure 5 allows one to estimate the underlying error of this assumption. In the ATR SEIRA set-up, the IR signals are enhanced via the electric field component of the electromagnetic radiation perpendicular to the surface |E⃗_rad|_⊥, such that the maximum enhancement is achieved for molecular oscillators that are oriented in that direction. Since the SEIRA intensity I_SEIRA scales with the square of |E⃗_rad|_⊥, it follows that I_SEIRA ∝ (cosθ)². Assuming that the SEIRA intensity for a pure MBN SAM on an Au electrode, measured at an electrode potential of +0.1 V, would refer to a tilt angle of the nitrile function with respect to the surface normal of 49° (vide supra) the increased and decreased intensities at −0.4 V and +0.6 V then corresponded to a tilt angle of 45° and 57°, respectively.

6-Mercaptohexanoic acid (MHA), 4-mercaptobenzonitrile (MBN), and thiophenol were purchased from Dojindo, Apin Chemicals, and Sigma-Aldrich, respectively. All other chemicals and solvents were of the highest purity grade available. Water was purified by a Millipore system and had a resistivity > 18 MΩ·cm.

Prior to SAM deposition, Au films and Ag ring electrodes were prepared and electrochemically roughened as described previously [39,40]. MBN SAMs on Ag were obtained by immersing the electrode in a solution of 1 mM MBN in dimethyl sulfoxide (DMSO):H₂O (3:1 v/v) at 4 °C for 18 h; for SAM formation on the Au surface, a 1 mM MBN solution in anhydrous DMSO was employed. Mixed thiophenol/MBN SAMs on Au were prepared in two steps. First, a pure thiophenol SAM was formed from a 2 mM solution of thiophenol in DMSO:H₂O 3:1 (v/v), using the same procedure as for MBN SAM formation in the spectro-electrochemical cell. Subsequently, the cell was washed successively with DMSO, ethanol, and water. Afterwards a 10 mM phosphate buffer solution (pH = 7) was added to the cell. Finally, the addition of MBN dissolved in DMSO:H₂O 3:1 (v/v) to the cell under potential control (0 V vs. Ag/AgCl) resulted in a final concentration of 100 nM. In a similar way, a mixed MHA/MBN SAM was formed, starting with the pure MHA SAM [11,39] followed by addition of MBN (10 μM) to the solution at −0.4 V. The formation of the SAMs was monitored by CV and SEIRA spectroscopy (Supplementary Information).

SEIRA measurements were carried out in the Kretschmann-ATR configuration which was integrated in an electrochemical cell for potential-controlled studies. Details of the set-up are given elsewhere [39]. SEIRA spectra were recorded with a spectral resolution of 4 cm⁻¹ on a Bruker IFS66v/s spectrometer, using a liquid nitrogen cooled photoconductive MCT detector. 400 scans were co-added for each spectrum; three to five spectra were recorded for each experiment and averaged. Each SEIRA experiment was repeated at least twice.

SER spectroelectrochemical experiments were carried out with the set-up described previously [41]. The spectra were obtained with the 413-nm excitation line of a Krypton ion laser using a confocal Raman spectrometer equipped with a liquid nitrogen cooled CCD detector. The spectral resolution was better than 2 cm⁻¹. The power of the focused laser beam on the sample was 1 mW. The accumulation time was 60 s. All experiments were repeated four times.

CV measurements were carried out in the same cells used for SEIRA and SER spectroelectrochemistry [39,41]. In all electrochemical and spectroelectrochemical measurements potassium phosphate buffer (10 mM, pH = 7.0) was used. Unless otherwise indicated, all potentials cited in this work refer to the Ag/AgCl electrode (3 M KCl).