1 Introduction

One of the effective methods for altering the properties of self-assembled monolayers (SAMs) of (poly)aromatic molecules is the exposure to electron irradiation. This method can have practical applications in SAM-based nanofabrication and nanolithography [1], fabrication of mechanically stable, carbon-based nanomembranes [2, 3], development of complex chemical gradients for biomedical applications [4,5,6] and functionalization of metallic surfaces [7]. Electron irradiation is also known as an effective tool to tune the properties of other type of materials [8] together with electromagnetic irradiation [9]. However, the interaction of electron radiation with molecular SAMs is a complex process which depends not only on the parameters of the incident electron beam, but also on the properties of the molecules (e.g., molecular backbone (aromatic or aliphatic), terminal groups and molecular proximity) and environment (e.g., properties of the substrate). For example, enhanced stability can be obtained for aromatic molecules due to electron-induced intermolecular cross-linking [10], whereas significant structural damages can follow electron bombardment in the case of aliphatic molecules [11, 12].

For certain type of applications, such as fabrication of carbon nanomembranes from SAMs of organic molecules by electron irradiation [2, 13, 14], it is desired to remove the docking groups of the molecules in order to achieve high quality of the resulting products. Recently, Neumann et al. [15] demonstrated successful creation of electron irradiation-induced nanomembranes from molecular SAMs with complete removal of anchoring groups. These chemically inert nanomembranes were created using aromatic SAMs with carboxylic acid terminal groups. Following experiments showed that structural parameters and purity of the resulting nanomembranes can be controlled by introducing aliphatic (CH2) linker between the aromatic molecular backbone and the docking group [16]. However, the situation becomes more complicated in the case of thiolate SAMs, as the remaining thiol groups may still decorate the resulting nanomembranes and increase their chemical reactivity, thus reducing their inertness. On the other hand, it is a common practice to use dithiol molecules to have well-ordered and closely packed SAMs on metallic surfaces due to the strong metal-sulfur bonds [17,18,19,20,21,22,23].

Here, we use X-ray photoelectron spectroscopy (XPS) measurements to study the effect of methylene spacer on the properties of aromatic molecular SAMs subjected to electron irradiation. We consider biphenyl-4,4′-dithiol (BPN) (Fig. 1a) and 5,5′-bis(mercaptomethyl)-2,2′-bipyridine (BPD) (Fig. 1b) molecules self-assembled on gold (111) surface. We show that BPN molecules, where the thiol group is directly bonded to the phenyl core of the molecule, are more resistant to the electron beam than BPD molecules for which the thiol end-group is separated from the phenyl rings by electron saturated methylene group. The experimental results are supplemented by density functional theory (DFT) calculations to determine the role of edge groups in the stability of the considered molecules under electron irradiation. It is known that DFT can be effectively used to describe complex structures and usually provide a good agreement with the experiment [24,25,26,27]. The simulation results show that the presence of the methylene unit weakens the bonding of the -SH unit to the molecule, which suggests that significantly larger doses of electron radiation are required to break the thiol bonds when the thiol unit is in direct conjugation with the phenyl group.

Fig. 1
figure 1

Optimized structures of biphenyl-4,4′-dithiol (BPN) (a) and 5,5′-bis(mercaptomethyl)-2,2′-bipyridine (BPD) (b) molecules adsorbed on Au (111) surface

Full size image

2 Methods

2.1 Sample Preparation

We have purchased all the solvents and chemicals from Sigma-Aldrich and used them without further purification. Gold (111) substrate on mica was purchased from PHASIS Switzerland. Pure hexane was degassed for 30 minutes under nitrogen gas flow for sample preparation. The gold substrate was rinsed with ethanol and dried under nitrogen gas flow. Both BPN and BPD molecules were prepared using 1 mM solution in hexane at 60 °C. The solution was degassed for 30 minutes at 60 °C before gold substrate was immersed. The systems remained in this environmental condition for 15 min under reduced light condition. The resulting gold substrate was washed first with hexane and 3 times with absolute ethanol. After that the samples were dried with nitrogen gas. The samples were immediately transferred to XPS chamber (under ultrahigh vacuum of 10−10 mbar) for photoemission measurements to avoid oxidation.

2.2 Photoemission

Photoemission measurements are conducted using the standard Thermo Fisher ESCALAB 250XI type XPS platform. Energy resolution of a monochromatic Al Kα anode X-ray beam (of energy 1486.6 eV) was 0.5 eV. X-ray beam incident angle was 45° to the surface normal, and the XPS spectra were measured with a normal emission. Energy values in the spectra were calibrated with respect to the Au4f7/2 located at 84.0 eV. The electron irradiation was performed using reflective electron energy loss spectroscopy (REELS) mode with energy of 1000 eV. The irradiation dose is evaluated by a faraday cup. The dose of 70×10−7 C/(cm2s) and exposure time of 30 mins are used during the experiments.

2.3 DFT Calculations

All simulations (geometry optimizations and electronic structure calculations) are conducted using DFT within the generalized gradient approximation of Perdew–Burke–Ernzerhof (PBE) to describe the exchange-correlation energy [28]. Non-bound van der Waals interactions are taken into account using Grimme’s empirical correction to PBE [29]. The Brillouin zone was integrated using 5×5×1 Monkhorst-Pack k-points [30]. All atoms are described using norm-conserving PseudoDojo pseudopotential with medium basis set [31]. For the geometry optimization, the convergence criterion for Hellmann–Feynman forces was 0.01 eV/˚A. All simulations are conducted using the computational package Atomistix toolkit [32, 33].

3 Results and Discussions

In situ XPS measurements are conducted in the same environmental conditions (e.g., vacuum level) for both systems in order to investigate the effect of electron irradiation. As a main result, we present in Fig. 2 C 1s (a,b) and S 2p (c,d) core-level spectra of BPN (a,c) and BPD (b,d) SAMs on Au (111) substrate before (solid-black curves) and after (dashed-red curves) the electron beam irradiation. For the BPN SAM system, C 1s spectrum has the main peak at 284.34 eV originating from the biphenyl moiety of the molecules (solid-black curve in Fig. 2a). BPN SAMs system is found to be resistive to the considered dose of the electron irradiation: the position of the maxima in C1s spectrum remains the same and the intensity of the XPS signal reduces only slightly (∼ 2%). The effect of electron irradiation becomes more pronounced in the case of BPD SAM system which includes methyl sulfide anchor group (see Fig. 2b). First, the main peak in the C1s signal (located at 284.86 eV) shifts to higher binding energy (285.14 eV) after electron bombardment, indicating possible structural changes in the backbone of the molecules (possible changes from sp2 bonding to sp3-like bond formation). Second, the electron irradiation results in reduction of the carbon content (∼ 5%) in the C1s spectrum, which is related either to partial desorption of the molecules or reduction in the effective SAM thickness [11]. However, as the drop of carbon intensity signal is not significant, (only∼ 5%), the probability of partial SAM desorption might be excluded [16]. Our estimate for the reduction in the effective monolayer thickness from the intensity ratios of carbon and gold signal [34] is about 5%, which can be achieved by removing methyl sulfide terminal group.

Fig. 2
figure 2

C 1s (a,b) and S 2p (c,d) core-level spectra of biphenyl-4,4′-dithiol (BPN) (a,c) and 5,5′-bis(mercaptomethyl)-2,2′-bipyridine (BPD) (b,d) SAMs on Au (111) substrate before (solid-black curves) and after (dashed-red curves) the electron beam irradiation. The intensities are normalized to the corresponding Au4f signal. The XPS spectra are fitted with a Voigt profile (GL(30)) after a Shirley background subtraction. Insets in (a) and (b) show optimized structures of BPN (a) and BPD (b) molecules on gold surface

Full size image

Figure 2c, d shows a direct comparison of the S2p core level spectra of BPN (c) and BPD (d) SAM systems before (solid-black curves) and after (dashed-red curves) electron irradiation. In both cases, XPS measurements give well-defined splitting of the S2p spectra with the main maximum located at 163.32 eV and 163.42 eV, respectively, for BPN and BPD systems. This signal corresponds to the S-H bonding [12]. We also observed additional peak in the S2p spectrum around 167.9 eV, which related to sulfur oxidation compound [35]. The latter becomes more pronounced in the case of BPN SAM. Effect of electron irradiation on the S2p signal of the BPN system is negligible (see dashed-red curves in Fig. 2c). On the contrary, significant reduction of sulfur content corresponding to the main peak in the spectrum is obtained in the case of BPD SAM (dashed-red curve in Fig. 2d). This indicates that only the thiol group on top of the BPD SAM is affected by the electron bombardment for the considered dose of the beam. Reduced signal is also found for sulfur oxide content. Thus, introducing saturated CH2 unit between the thiol group and the phenyl ring reduces the resistance of aromatic dithiol molecules to the electron irradiation.

To give qualitative description to the experimental results, we have conducted bond length-dependent total energy calculations for the considered molecular systems. This explicit energy-distance approach is useful in practice, especially for complex metal-molecule systems. Figure 1 shows the optimized structures of our model systems consisting of BPN (a) and BPD (b) molecules adsorbed on Au (111) surface. Note that we assume the release of the hydrogen atom from the -SH unit [36] and the adsorption of the molecule on the gold surface through Au-sulfur bond. The gold substrate contains six layers of gold atoms, and bottom two layers were kept fixed during the geometry optimization. The simulation cell has vacuum spacing of 50 Å along the molecules. Figure 3 shows the total energies of the considered systems as a function of C-S (a) and C-C (b) bond distances. Calculations are conducted taking into account the dispersive interactions. For C-S bonds, the minimum energy is obtained at 1.77 Å and 1.84 Å, respectively, for BPN and BPD molecules. At smaller distances, the system energy increases rapidly with decreasing distance due to the repulsive interaction between the overlapping molecular orbitals. The system energy increases with further increase in the bond length starting from the minimum in the energy curve. In the case of BPN molecule, when the sulfur atom makes a covalent bond directly with the phenyl ring (see inset 2 in Fig. 3a), 4.87 eV energy is required to break the C-S bond. However, when introducing methylene unit between the phenyl ring and the -SH group (see inset 1 in Fig. 3a), the energy required to desorb the -SH unit reduces by 0.8 eV. This can be used to explain the reduced sulfur content in the XPS spectra of the BPD SAM (Fig. 2d). Figure 3b shows the variation of the total energy of the BPD system by changing C-C distance between the phenyl ring and the CH2 unit as highlighted in the inset of Fig. 3b. In this case, 5.63 eV is needed to beak the C-C bond.

Fig. 3
figure 3

Total energies of BPD (solid-black curves) and BPN (dashed-red curve) as a function of C-S (a) and C-C (b) bonds. The energy for the bond distance of 12 Å is taken as reference point in each case. Insets show the optimized structures of the molecules with highlighted bonds for total energy calculations

Full size image

Bond distance-dependent total energy calculations show strong covalent bonds (see insets of Fig. 3) indicating enhanced stability of the considered molecules on gold surface. To model how electron irradiation affects the stability of the molecules, we have conducted the same bond distance-dependent total energy calculations but with injecting extra electrons to the system. For the sake of simplicity and to avoid time-consuming calculations, we conducted simulations for isolated molecules without the presence of the gold substrate. Figure 4a shows the variation of the total energy of BPD molecule as a function of C-S bond distance (see the inset) for neutral molecule (solid-black curve), for 1 (dashed-red curve) and 2 (dotted-blue curve) injected electrons. In the case of the neutral BPD molecule, the energy of 4.02 eV is required to break the C-S bond, which is 1.7% smaller than the energy required for the same reaction in the presence of the gold substrate (solid-black curve in Fig. 3a). The bond dissociation energy decreases dramatically when a single electron is injected to the molecule (1.91 eV, see dashed-red curve in Fig. 4a). The position of the minimal energy shifts to higher distances. With injecting the second electron (dotted-blue curve in Fig. 4a), the minimal energy value increases to − 0.25 eV, which indicates the dissociation of the molecules at room temperature. The injection of the external electrons increases the energy of the molecules and bring the molecule to the antibonding state. The molecule relaxes to the state with dissociated C-S bond. This effect can be used to describe qualitatively the reduced sulfur content in the XPS spectra (dashed-red curve in Fig. 2d). The situation changes dramatically in the case of BPN molecule (Fig. 4b). First, due to strong interaction of the thiol unit with the phenyl ring, a finite energy (1.83 eV) is required to break C-S bond even for two injected electrons (dotted-blue curve in Fig. 4b). This means that larger dose of the electron irradiation is required in the experiment to remove the terminal thiol group. Second, the position of the minimum total energy value shifts to smaller distances with injecting the electrons to the system. To see if these changes in the response of the molecule to the external electrons are due to the nature of the molecular backbone (i.e., BPD and BPN molecules are structurally different) or due to the difference in the terminal group, we have conducted simulations for 4,4′-Bis(mercaptomethyl)biphenyl (BMP), which differs from the BPN molecule only by the presence of the methylene group (see the inset of Fig. 4c). Figure 4c shows the variations of the total energy of the BMP molecule as a function of C-S bond distance for different number of the injected electrons. These findings are very similar to the results obtained for the BPD molecule (Fig. 4a), which also contains methyl sulfide anchor group (see the inset in Fig. 4a). Namely, the total energy increases rapidly with injecting external electrons and the position of the energy minimum shifts to larger bond distances. These results indicate the importance of the edge groups on the response of the molecular system to electron irradiation.

Fig. 4
figure 4

Total energies of BPD (a), BPN (b) and BMP (c) as a function of C-S bonds length. Insets show the optimized structures of the molecules with highlighted bonds for total energy calculations

Full size image

To find the origin for such different behaviors of the considered molecules during the electron injection, we have conducted additional structural and electronic structure analysis. Figure 5 shows the optimized structures of the considered molecules without (a, c, e) and with one external electron (b, d, f). As a general trend, the dihedral angle between the phenyl rings of the molecules increases (i.e., molecules become more planar) with injecting external electrons. This effect is more pronounced in the case of the BPN molecule (see Fig. 5a, b). However, the C-S bond distance (dC-S) in the molecule remains unchanged after electron adsorption (dC-S=1.77 Å). On the contrary, dC-S increases considerably due to the electron injection for the other two molecule (see Fig. 5c–f). To see how these structural changes affect the electron distribution in the system, we present in Fig. 6 isosurface plots of the density of the electrons in the considered molecules without (a, c, e) and in the presence of a single external electron (b, d, f). In the case of the BPN molecule (Fig. 6a, b), the finite electron density is obtained between the carbon and sulfur atoms without and with the external electron. However, for the other two molecules (Fig. 6c–f), no reduced electron density is obtained after electron injection (Fig. 6d, f). This indicates weakening of the C-S bond after electron injection and explains the bond distance dependence of the total energy of the molecules as shown in Fig. 4. In particular, the shift of the minimum in the total energy curve to larger distances can be explained by the increase of the C-S bond distance after electron injection.

Fig. 5
figure 5

Optimized structures of BPN (a, b), BPD (c, d), and BMP (e, f) molecules without (a, c, e) and with one injected external electron (b, d, f). The numbers show the C-S bond distance (numbers at the top of the molecules) and the dihedral angle between the phenyl rings

Full size image
Fig. 6
figure 6

Isosurface plots of the electron density (isovalue 1.21 Å3) in BPN (a, b), BPD (c, d), and BMP (e, f) molecules without (a, c, e) and with one injected external electron (b, d, f). Finite values of the electron density between carbon and sulfur atoms are highlighted

Full size image

To explain the strong bonding of the thiol group to the backbone of the molecule in the absence of the CH2 unit, we have also conducted electron localization calculations, which defines the probability of locating the electron in the vicinity of another electron with the same spin-state. Figure 7 shows the isosurface plots of the electron localization function (isovalue 0.95) in BPN (a), BPD (b), and BMP (c) molecules with one injected external electron. We obtained finite probability for the electrons to be located between the carbon atom of the phenyl ring and the thiol group (indicated by an arrow in Fig. 7a) indicating stronger interaction between the edge group and the molecular backbone. However, in the other two cases (see Fig. 7b, c), the electrons are localized more in the CH2 and SH units. This indeed indicates the weaker interaction of the thiol unit with the molecule.

Fig. 7
figure 7

Isosurface plots of the electron localization function (isovalue 0.95) in BPN (a), BPD (b), and BMP (c) molecules with one injected external electron

Full size image

4 Conclusions

We performed XPS measurements to study the effect of electron irradiation on the structural properties of BPN and BPD molecules self-assembled on Au (111) surfaces. We found that the presence of insulating CH2 unit in the case of BPD molecule has an important effect on the structural properties of such thiol-terminated molecules during electron beam irradiation. Namely, the methylene unit weakens the thiol group attachment to the molecule as evidenced from XPS spectra of the BPD SAMs. For example, the main peak in the carbon signal shifts to higher binding energies, and the carbon content in the XPS spectra reduces after electron bombardment. These indicate the structural changes in BPD SAM during the electron irradiation which are not very pronounced in case of the BPN SAM. The experimental results are supplemented with first principles DFT calculations for structural and electronic properties of the considered molecular systems to have a qualitative description for the experimental findings. We found that the binding energy of the thiol unit becomes significantly weaker due to the presence of the CH2 unit. This suggests that larger doses of electron irradiation are required to remove the thiol unit when it is in direct conjugation with the phenyl core of the molecule. These finding show the importance of the presence of saturated CH2 units for structural properties and stability of dithiol molecules.