Connecting Solution-Phase to Single-Molecule Properties of Ni

Letter pubs.acs.org/JPCL

Cite This: J. Phys. Chem. Lett. 2019, 10, 3525−3530

Connecting Solution-Phase to Single-Molecule Properties of Ni(Salophen) Yi C. Zhang,† Bhaskar Chilukuri,‡ Tanner B. Hanson,‡ Zachariah M. Heiden,*,‡ and David Y. Lee*,† †

Downloaded via BUFFALO STATE on July 31, 2019 at 08:15:45 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Department of Chemistry and Materials Science & Engineering Program, Washington State University, Pullman, Washington 99164, United States ‡ Department of Chemistry, Washington State University, Pullman, Washington 99164, United States S Supporting Information *

ABSTRACT: We present a strong correlation of the Ni(salophen) structure and properties measured in single-molecule vs bulk quantities and in ultra high vacuum vs solution phase. Under a scanning tunneling microscope (STM), Ni(salophen) forms a self-assembled monolayer (SAM) on Au(111) at 23 °C with molecular structure identical to that of the X-ray crystallographic measurement. The HOMO and LUMO levels are determined using elastic tunneling spectroscopy at the single-molecule level with confirmation by monolayer-quantity ultraviolet photoelectron spectroscopy (UPS) and by cyclic voltammetry (CV) measurements. The STM-determined HOMO− LUMO gap of 3.28 eV and (HOMO−1)−HOMO gap of 0.36 eV form a new foundation for the selection of hybrid functionals with a simple basis set to be effective in accurately calculating single-molecule Ni(salophen) frontier MO levels. Our results suggest that microscopy-based experiments on a surface, along with free-molecule gas-phase calculations, can provide useful insights into the physical properties of metal(salen) complexes, especially when such direct measurements are not available in solution.

T

particular, for metal−salen, intermediate species (such as metal hydrides, metal oxides, and metal nitrides) have been proposed to exist in solution-phase reaction mechanisms,23−25 but their short lifetimes have made it experimentally unfeasible to probe these intermediate species in solution. The ability to place metal complexes on surfaces followed by localized microscopy/spectroscopy measurements of stepwise gas−surface or solution−surface interactions at the singlemolecule level26−32 holds great promise to provide experimental insights into reaction mechanisms not obtainable in solution-phase measurements. However, the validity of reaction steps probed on a surface, as compared to the mechanisms proposed in a solution environment, depends heavily on the initial states and structures of the molecules before reaction. In other words, one cannot simply assume that solution-phase properties can be directly transposed onto a surface environment for the same molecule.33 In this work, using a combination of experimental and computational techniques, we demonstrate that the structures and physical properties of Ni(salophen) can be comprehensively correlated in the solution phase vs in a self-assembled monolayer (SAM) on top of a Au(111) surface in ultrahigh-vacuum (UHV). Detailed experimental and computational procedures are provided in the Supporting Information (SI), and all of the experiments were performed at 23 °C for the purpose of

he diverse functionality of metal−salen complexes (where salen is the N,N′-bis(salicylidene) ethylenediamine motif; Figure 1) have inspired the use of these molecules in numerous

Figure 1. Comparison of the general metal−salen motif to the metal− salophen motif investigated in this study.

disciplines including inorganic,1−4 organic,5,6 biological,7−10 and materials11,12 chemistry. For example, metal−salen complexes have been employed as conducting agents in lithium-ion batteries,13 as molecular sensors in bioimaging applications,4 in medicinal applications,4,14 in optical applications,15 as catalysts in the epoxidation of olefins,5 or in the production of polycarbonates, 1 to name a few. The attractiveness of implementing salen ligands in molecular reactions, besides their ease of synthesis,16 is their ability to provide a relatively rigid tetradentate environment (similar to porphyrins17,18) that can be utilized to influence the steric and electronic properties of their respective metal complexes in a predictable manner. In general, for molecules with unique properties, their chemical reactivities have been widely studied using solution-based or computational methods.19−22 In © 2019 American Chemical Society

Received: May 14, 2019 Accepted: June 5, 2019 Published: June 5, 2019 3525

DOI: 10.1021/acs.jpclett.9b01381 J. Phys. Chem. Lett. 2019, 10, 3525−3530

Letter

The Journal of Physical Chemistry Letters

geometry34 is illustrated in Figure 2d for Ni(salophen) and revealed little distortion upon forming a stable SAM on Au(111) at 23 °C. Although STM is powerful to resolve submolecular structures, computational analysis is needed to provide insights into the Ni(salophen)−Au interaction. Because calculations using non-van der Waals (vdW) DFT functionals severely underestimate molecular adsorption energies on metal surfaces,36,37 periodic calculations employing vdW DFT functionals are used to determine the Ni(salophen) adsorption energy on Au(111). Computational analysis reveals a −2.74 eV adsorption energy, which is about 2/3 of the adsorption energy of a metal porphyrin molecule on the same surfaces.38,39 The DFT-optimized Ni(salophen)−Au geometries are shown in Figure 3. From the perspective that either the adsorption

effective comparison between the different reaction environments. Structures. Although it has been reported that a halogenated Co(salophen) complex can be isolated and imaged in UHV at a temperature of a few Kelvin using a scanning tunneling microscope (STM),35 we demonstrate within the white dashed boundary in Figure 2a that a SAM of Ni(salophen) (molecular

Figure 3. (a) Top and (b) side views of the DFT-optimized physisorption configuration of a single Ni(salophen) molecule on Au(111).

energy of a metal(salen) or a metal(porphine) complex is an accumulation of vdW interactions of the constituent atoms within each molecule with the gold surface (i.e., physisorption without bond formation to the gold), the computed adsorption energy agrees with the fact that a Ni(salophen) complex can be approximated as 3/4 of a Ni(porphyrin) in chemical structure. Because the computed adsorption energy ratio of 2/3 is less than this structural ratio of 3/4, the adsorption for Ni(salophen) on Au(111) is even weaker than this simple cumulative vdW picture. The DFT-optimized physisorption distance between Ni(salophen) and Au(111) is 2.6 ± 0.1 Å, which is comparable to the previously reported ∼2.9 Å for Cooctaethyl porphyrin (CoOEP) physisorption on the same substrate.36 Physical Properties. On the basis of reports describing the physisorption of metal(porphyrins) and metal(phthalocyanines) on a Au(111) surface,40−44 molecular orbital (MO) levels of the adsorbed species in the ±2 eV range relative to the Au(111) Fermi energy (Ef) can be directly measured at the single-molecule level using STM-based orbital-mediated tunneling spectroscopy (STM-OMTS) with a tip−sample separation of >500 MΩ. Furthermore, the STMOMTS-measured occupied MO levels can be confirmed using ultraviolet photoemission spectroscopy (UPS) measured for an ensemble quantity of ∼1 surface monolayer (∼1012 molecules). The unoccupied MO levels can also be confirmed with electrochemical measurements in a solution environment that exhibit an offset value of 4.7 eV relative to the saturated calomel electrode (SCE), i.e., E1/2(vacuum) = E1/2(SCE) + 4.7 eV.45−47 Due to the similarity between the STM-probed molecular structure on Au(111) and the XRC geometry, in addition to the computed adsorption energy being comparable to that of a metal(porphyrin), the Ni(salophen) frontier MO

Figure 2. (a) A 1200 Å × 1200 Å STM image of a partial Ni(salophen) SAM on Au(111) with its chemical structure drawn in the lower left corner. STM parameters: −300 mV and 350 pA. (b) A 150 Å × 150 Å closer view of the SAM. Colored features are discussed in the text. STM parameters: −30 mV and 350 pA. (c) A 80 Å × 80 Å image showing the SAM unit cell. STM parameters: −10 mV 350 pA. (d) Comparison of molecular geometries measured using STM vs XRC.34 All STM images were collected in UHV at 23 °C.

structures illustrated in the lower left corner) can be stably formed on top of Au(111) at 23 °C. The Ni(salophen) SAM is shown in Figure 2b to be a highly ordered striped pattern with individual phenyl rings and the central Ni atom resolved for each molecule. A significant feature of this SAM is that there are two “rows” of Ni(salophen) molecules within each stripe, with the row directions alternating between adjacent stripes. A semitransparent red arrow is superimposed on the molecular structure in Figure 2a, with similar arrows drawn in panel (b) to illustrate the alternating molecular surface pattern. In addition, one of the gaps between two adjacent stripes is highlighted with a bold red line in both panels (a) and (b) of Figure 2 as visual aids. The angle between the SAM stripes and the Au(111) herringbone shown in Figure 2a is 44.8°; however, other angles (between 37.0 and 43.1°) have been observed on various samples, which imply that there is no strict relationship between the SAM and Au surface structures. Due to the above-mentioned alternating pattern of the Ni(salophen) SAM, the unit cell spans across two stripes, with its dimensions drawn and indicated in blue in Figure 2c. Each unit cell contains 12 phenyls and 4 Ni atoms, corresponding to a total of 4 Ni(salophen) molecules. The direct comparison of the STM-measured single-molecule structure on Au(111) vs the X-ray crystallography (XRC)-determined molecular 3526

DOI: 10.1021/acs.jpclett.9b01381 J. Phys. Chem. Lett. 2019, 10, 3525−3530

Letter

The Journal of Physical Chemistry Letters

energy level is readily observed as a shoulder in the STMOMTS spectrum at ∼6.66 eV. The solution-phase CV result in Figure 4b shows that the first reduction level is at −1.38 eV and the first oxidation level is at +1.23 eV (2nd oxidation at +1.38 eV) vs SCE. When 4.7 eV is added, the reduction level (−1.38 + 4.7 = 3.32 eV) agrees well with the STM-OMTS LUMO peak observed at 3.37 eV, as indicated in Figure 4a. However, the CV oxidation level of 1.23 + 4.7 = 5.93 eV lies shallower than the STM-OMTS/ UPS-measured HOMO level of ∼6.66 eV. Armstrong and coworkers52 have determined a correction factor of 1.71 ± 0.15 that works well to correct this discrepancy for metal(phthalocyanine) and metal(porphyrin) complexes.40,42,43 In our case for Ni(salophen), we can also find the exact match between the solution-phase CV measurement and the UHV surface STM-OMTS/UPS data by simply multiplying the CVmeasured oxidation level by this correction factor (1.23 × 1.594 + 4.7 = 6.66 eV). We can further assign the STM-OMTS peak in Figure 4a at 7.01 eV to be HOMO−1. This peak, together with the previously discussed HOMO and LUMO levels, indicates an experimentally determined HOMO−LUMO gap of 3.28 eV and a (HOMO−1)−HOMO gap of 0.36 eV. These UHV single-molecule measurements, in comparison to the recent solution-phase absorption spectroscopy and computation study,53 provide a more direct and solvent-free determination of frontier MO energy gaps. Because each Ni(salophen) molecule experiences little distortion or interaction with Au(111), we further expect that these STM-determined MO gaps to be readily accounted for using gas-phase free-molecule calculations. In fact, metal(salen) complexes have been intensely investigated computationally, with several reports describing that one computational method is better than the others.20,53−57 Our STM-OMTS measurement thus provides a new foundation to re-examine these computational methods. The agreement between gas-phase single-molecule computations using hybrid functionals and the STM-based singlemolecule measurement is presented in Table 1. Calculations

levels are expected to be unambiguously determined using the same experimental techniques as those of metal(porphyrin) complexes. The combined STM-OMTS, UPS, and cyclic voltammetry (CV) results for Ni(salophen) are shown in Figure 4, with the origin of the absolute energy scale set at the vacuum level.

Table 1. Comparison of Single-Molecule Gas-Phase Calculations vs Single-Molecule Surface Measurements for Ni(salophen)a

Figure 4. (a) Single-molecule STM-OMTS (shown in black) and ensemble UPS measurements of Ni(salophen) on Au(111) shown in red ○ (vs bare Au(111) shown in red ■) with the origin of the absolute energy scale set at the vacuum level. (b) Cyclic voltammogram of a 1 mM solution of Ni(salophen) in dichloromethane. Scan rate = 100 mV/s; electrolyte = nBu4NPF6 (0.1 M).

The black spectrum in Figure 4a shows the UHV STMOMTS measurement for Ni(salophen) on Au(111) at 23 °C scanned between −2 and +2 eV relative to the Au(111) Ef (the bottom abscissa). Overall, the “U”-shaped spectrum clearly indicates a substantial band gap, and the flatness in the middle of the gap further indicates that Ni charge transfer (especially for the Ni(II) → Ni(III) + e− oxidation) does not occur in this ±2 eV range.48 For UPS, although various Fermi energies have been reported for Au(111) between 5.20 and 5.43 eV,40,49−51 we add the value of 5.20 eV (which was previously determined using the same UPS/XPS instrument here at WSU40) to the UPS binding energy (the top abscissa) for direct comparison to the STM-OMTS result. The difference in UPS data for Ni(salophen) (red ○) and for a clean Au(111) substrate (red ■) clearly shows the HOMO level at ∼6.65 eV for the Ni(salophen) monolayer on Au(111) (red arrow), and this

a

functional

basis set

HOMO−1 to HOMO

HOMO to LUMO

B3LYP B3P86 B97-1 PBE1PBE (PBE0) experiment

6-31G(d,p) 6-31G(d,p) 6-31G(d,p) 6-31G(d,p)

0.3901 0.4332 0.3717 0.3562

3.0542 3.0384 3.1244 3.4292

STM-OMTS

0.36

3.28

Energy values are shown in eV.

using pure functionals are included in the SI. In general, we find that hybrid functionals reproduce experimental values better than the pure ones, and the simple 6-31G(d,p) level sufficient with no further improvements using the more timeconsuming coupled-cluster basis sets (also see the SI). Because the chemical reactivity of a molecular complex is largely dependent on the location of its frontier MO energy levels, our results suggest that surface STM measurements and gas-phase calculations can be used to provide insights at the single3527

DOI: 10.1021/acs.jpclett.9b01381 J. Phys. Chem. Lett. 2019, 10, 3525−3530

The Journal of Physical Chemistry Letters

■

molecule level into future homogeneous and heterogeneous reactions for metal(salen) complexes.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected].

EXPERIMENTAL METHODS A more detailed description of experimental methods is provided in the SI. Essentially, Au(111) films were deposited on mica in a cryo-pumped chamber with a base pressure of 2 × 10−9 Torr. Ni(salophen) was prepared by the procedure reported previously by Teixeira and co-workers,58 which matched the characterization data reported by Kleij and coworkers.59 A 25 μ L of 5 × 10−4 M Ni(salophen) in chloroform droplet was dropcast onto a freshly annealed Au(111) film before introduction into the UHV STM chamber. The sample was further annealed in UHV at 92 °C for 1 h, with subsequent annealing at 130 °C for 1 additional hour. The UHV STM (model PanScan) and controller (model R9) were purchased from RHK technologies. All images were acquired in constant current mode at room temperature using chemically etched Pt0.8/Ir0.2 tips. STM-OMTS measurements were performed by placing the tip on top of a Ni(salophen) molecule in the middle of the monolayer followed by measuring dI/dV as a function of bias voltage at a fixed sample−tip separation of 2 V and 800 pA. UPS measurements were conducted by following the previously reported procedure40 using a home-built machine with a He lamp producing two resonance lines via cold cathode capillary discharge at a working pressure of