Co-Adsorption of Formic Acid and Hexane Selenol on Cu

Abstract

Self-assembled monolayers of alkane thiolate and alkane selenolate have been proven to inhibit atmospheric corrosion, but upon prolonged exposure to the important constituents of indoor atmosphere, namely humidified air with formic acid, the protective layer eventually breaks, but the exact reason is not yet clear. In this paper, we report on an XPS study of co-adsorbed formic acid and hexane selenol on a Cu surface. Adsorption of hexane selenol at room temperature breaks the Se-C bond, leaving a monolayer of Se on the surface, whereas adsorption at 140 K leaves a layer of selenolate. Formic acid exposure to the selenolate-Cu surface leads to adsorbed formate on unprotected areas and absorption of formic acid within the alkane chain network. During heating, the formic acid desorbs and the Se-C bond breaks, but formic acid does not accelerate the Se-C scission, which occurs just below room temperature both with and without formic acid. Thus, formic acid alone does not affect the Se-C bond, but its presence may create disorder and open up the alkane carpet for other species. Selenol removes formate and oxide from the surface at room temperature. The Se-C bond breaks and the alkane chain reacts with surface oxygen to form carbon oxides and volatile hydrocarbons.

1. Introduction

Due to their nano-metric thickness, ease of application, and spontaneous formation, the protective self-assembled monolayers (SAMs) have attracted much attention [1,2,3]. Prof. Leygraf and his group [4,5,6,7,8], and other groups [9,10] have studied alkane thiols, alkane selenols, and other SAMs as protective layers against corrosion.

In [6] n-alkane, thiols (n = 4–18 is the number of C-atoms in the alkane chain) were used to protect Cu in humidified air with added formic acid, as one of the most important atmospheric corrosion stimulators on Cu [11,12] in indoor environments. Small concentrations of formic acid may pose corrosion problems on copper in various applications, such as in ant-nest corrosion [13,14,15]. It has been revealed that the selective diffusion of O₂, H₂O, and formic acid through the alkane chain alters the corrosion reaction path and delays the formation of certain corrosion products.

One key factor in the function of a protective layer is its robustness towards chemical interactions with the substrate and adsorbed molecules from the environment. For instance, in [7], hexane thiol and hexane selenol SAM were compared. Depending on their structure, chain length, and preparation, SAMs may form a well-ordered layer on the substrate with a dense carpet of carbon chains pointing out from the surface, which can hinder or at least slow down penetration of hostile corrosion promoters. It was found that hexane selenol was more easily removed from the surface, exhibiting a poorer corrosion protection capability compared to its hexane thiol counterpart, during exposure to a simulated corrosive environment.

Both thiol and selenol bind to metal surfaces as thiolate or selenolate by replacing the chalcogen–hydrogen bond with a metal–chalcogen bond [16,17]. The chalcogen–carbon bond is a weak point in the SAM formation that may break upon adsorption, leaving a chalcogen atom on the surface [18]. In a previous paper, we studied hexane selenol adsorption on Cu [19]. Room temperature adsorption resulted in Se-C cleavage and formation of a Se monolayer, on top of which a selenolate layer grew. When adsorbed on a layer of water and mildly heated to remove the water, an ordered layer of selenolate was formed without Se-C scission.

This paper reports on a study on the interaction between hexane selenol and formic acid on metallic Cu and oxidized Cu. Adsorption of hexane selenol at 140 K leaves a layer of selenolate on the Cu substrate. Exposure of this surface to formic acid leads to formate chemisorption on unprotected areas and absorption of formic acid within the alkane chain carpet. During heating, formic acid desorbs and the Se-C bond breaks. Nevertheless, formic acid does not accelerate the Se-C scission, which occurs just below room temperature, both with and without formic acid. Selenol removes formate and oxide from the surface at room temperature through carbon oxide and hydrocarbon desorption, as will be discussed here.

2. Materials and Methods

We use X-ray photoelectron spectroscopy (XPS), which is a method that distinguishes elements and also their chemical environment, such as chemical bonds and electron density. Our approach is to start from a clean Cu surface and adsorb controlled amounts of formic acid, hexane selenol, and oxygen in different orders. Thereafter, heat the sample. This allows detailed investigations of molecular interactions with the surface and also with each other without any unknown components.

The Cu sample was a polycrystalline disk (2 mm thick, 8 mm radius) polished by SiC paper and diamond paste, using ethanol as a lubricant. The sample was cleaned in a vacuum by Ar-ion bombardment (1 kV) until no traces of impurities (C, S, O) could be detected with XPS. The sample temperature was measured with a thermocouple in direct contact with the sample. Cooling was performed with liquid nitrogen, and heating was conducted by electron bombardment.

Hexane selenol (CH₃(CH₂)₅SeH) (AF ChemPharm) was deposited on the surface from the gas phase through a precision leak valve at 5 × 10⁻⁸ Torr. Doses are given in Langmuir (L), where 1 L = 10⁻⁶ Torr × s. Formic acid (HCOOH) (AF ChemPharm) and oxygen (O₂) were also introduced into the UHV chamber through separate leak valves. Both formic acid and hexane selenol were purified through repeated freeze-pump-thaw cycles until no impurities could be detected in mass spectra.

The XPS measurements were performed at beamline D1011 at MAXLAB in Lund, Sweden. D1011 was a bending magnet beamline in the energy range of 50–1500 eV. Energy selection was performed by a modified SX-700 plane grating monochromator (SX-700, Horiba Scientific, Kyoto, Japan). Electron spectra were recorded using a hemispherical SES200 analyzer (SES200, Surface Science Instruments (SSI), Auburn, CA, USA). The total energy resolution was in the range of 100–400 meV for the photon energies used in this study. The binding energy scale is referenced to the Fermi level recorded from the metal surface.

Generative artificial intelligence (GenAI) has not been used in this paper. Figures are made using the Origin Pro 2022, software.

3. Results and Discussion

Formic acid was dosed at 1 L, 5 L, and 10 L onto the clean Cu surface at 140 K and then heated in steps up to 300 K. C1s (left) and O1s (right) spectra from this series are presented in Figure 1. The top panels show spectra from adsorption, whereas the lower panels present spectra during heating. Several peaks are identified and marked; C1–C6 in the C1s spectra and O1–O4 in the O1s spectra. The origin of these peaks is largely based on the work by W. Osada et al. [20].

The first adsorption generates 4 C1s peaks. C1 at 283.2 eV represents atomic carbon and/or carbide [21], C2 at 285.0 eV represents hydrocarbon chains and graphitic carbon [21], and C3 is observed in the range of 288.0 to 288.5 eV. This binding energy range agrees well with the proposed species monodentate formate (288.4 eV) and bidentate formate (287.7 eV). These species may co-exist on the surface, and it was proposed that monodentate formate transforms to bidentate formate upon heating. C4 at 290.2 eV is assigned to formic acid within a complex network of formate and formic acid, as proposed by Osada et al. [20] and Yao [22]. At 5 and 10 L doses, the peak C5 appears at 292.0 eV, which is assigned to physisorbed formic acid.

During heating, the physisorbed layer desorbs first and gradually. In the temperature range of 230–250 K, C6 appears at 286.5 eV, which we identify as CO [23]. CO and water can be produced from the dissociation of formic acid. At room temperature, only the bidentate formate and carbon residues (C2) remain.

Formic acid has two inequivalent oxygen atoms, and depending on the adsorption geometry, the O1s spectrum may be very rich. Based on the temperature behavior and the identification by Osada [20], we associate O1 with bidentate formate, whereas O2 has its origin in monodentate formate and complex formation. O3 is from formic acid absorbed within the complex network, and finally, O4 originates from physisorbed formic acid. O4 disappears already at 160 K, whereas O3 and O2 intensities gradually decrease. At room temperature, only O1 remains. The O1s peak from water typically appears around 534 eV [24], but the exact value varies between different substrates.

At room temperature, only two species are left: bidentate formate and hydrocarbon chains/graphitic carbon. In the paper by Osada [20], the selected energy window did not include C1 and C2, and we cannot comment on those peaks, other than that Cu(977) and poly-crystalline Cu have similar surface chemistry towards formic acid.

To test the protective ability of hexane selenol, first 5 L of hexane selenol was deposited on Cu at 140 K, followed by 10 L of formic acid on top of it at 140 K. In Figure 2, we present (a) C1s, (b) O1s, and (c) Se3d core-level spectra recorded following adsorption at 140 K and heating in steps up to 380 K.

C1s from 5 L hexane selenol in Figure 2a has one strong peak, C2, just below 284.7 eV. The Se3d spectrum from the same preparation has a spin–orbit split peak at 54.8 eV. These values represent selenolate on Cu, according to our previous results [19].

After adsorption of 10 L of formic acid (red curves in all graphs) on top of the selenolate layer, two new peaks appear: C3 at 288.7 eV and C4 at 290.0 eV, representing formate and physisorbed formic acid, respectively. In addition, the C2 peak shifts by 0.35 eV to higher binding energy. O1s have three peaks: O1, O2, and O3, representing formate and formic acid, as described before.

The formic acid gradually desorbs up to 220–240 K, whereafter only selenolate and formate remain. As the formic acid-related peak (C4) decreases, C2 shifts back towards lower binding energy. We therefore suggest that the shift in C2 is perhaps not a chemical shift but rather reflects a perturbation of the alkane chains by absorbed formic acid within the chain network, which was previously observed by sum frequency generation [7].

At 260–280 K, the initial Se2 peaks shift to Se1, which previously was identified as Se-C scission [19]. Simultaneously, C2 is abruptly reduced, pointing to alkane desorption. Further, there are small but clearly visible C1 and C6 contributions from 240 K, which represent atomic C and CO. Thus, when the Se-C bond breaks, the alkane chain desorbs, but it also results in atomic C and CO on the surface. The O1 peak shifts gradually from the initial value 531.9 eV to 531.0 eV, indicating a smooth transition from a monodentate formate layer to bidentate formate. The CO contribution is located within the formate-related O1s peaks and cannot be resolved.

Thus, the initial selenolate/formate/formic acid layer desorbs and at the end of the heating series is transformed to bidentate formate and atomic selenium on Cu. Accordingly, selenol adsorption on Cu is found to leave a Se-layer on the surface at room temperature [19], and formic acid does not seem to stimulate Se-C bond scission at lower temperatures. However, the shift in C2 when adding formic acid to selenolate can be explained by formic acid penetrating the hexane chains. Interestingly, the presence of C3 already at low temperature suggests the formation of formate on selenolate-free areas.

In our previous paper, we demonstrated the ability of hexane selenol to remove a thin layer of copper oxide [19] through the formation of OH and H₂O with the aid of hydrogen released from selenol. Here, we prepare copper oxide by heating the Cu sample to 770 K during exposure to 1200 L O₂. XPS results are shown in Figure 3. A strong O1s peak at 529.6 eV is measured from the oxide, in agreement with previous results [19]. Exposure to 5 L of formic acid gives a second O1s peak at 531.0 eV, which may be assigned to bidentate formate. This identification is further supported by the C3 component in the C1s spectrum. However, when formate binds to the oxide, the binding energy may be different. In another study [25], formic acid adsorbed on Cu₂O(100) resulted in a broad O1s peak that was suggested to “hide” several components, such as formate in different geometries and hydroxyl formed by H adsorption to surface O. It is worth noting that the binding energy on Cu₂O is ~0.5 eV higher than what is measured here [25]. This is related to the fact that Cu₂O is a semiconductor with a surface-structure-dependent position of the Fermi level in the bandgap. Thus, we conclude that there is formate and hydroxyl on the surface, but we cannot determine the bond geometry. Adsorption of selenol on top of the oxide/formate structure reduces the intensities of both O1s components. After the 20 L selenol dose, only a fraction of the initial O1s intensity remains.

The C1s spectrum also changes; the C3 peak is weakened, and C2 appears strong. Further exposure to 20 L hexane selenole increases C2 even more and shifts the peak to higher binding energy.

The Se3d5/2 peak appears at 54.15 eV after the first dose of hexane selenol, which corresponds to atomic Se on the surface. At 20 L, a shoulder (Se2) appears on the higher binding energy side, representing selenolate.

4. Conclusions

We used XPS to study co-adsorbed layers of formic acid and hexane selenol on metallic Cu and oxidized Cu in UHV.

Adsorption of hexane selenol at room temperature breaks the Se-C bond, leaving a monolayer of Se on the surface. Conversely, adsorption at 140 K leaves a layer of selenolate. Upon exposure of Cu-selenole to formic acid, formic acid is absorbed within the alkane network, and formate is bound to unprotected areas. Following heating up to room temperature, formate remains untouched, formic acid desorbs, and the Se-C bond breaks.

Formic acid does not accelerate the Se-C scission, and we conclude that formic acid alone does not affect the Se-C bond, but its presence creates disorder and opens up the alkane carpet for other species. Selenol removes formate and oxide from the surface at room temperature. The Se-C bond breaks, and the alkane chain reacts with surface oxygen to form carbon oxides and volatile hydrocarbons.