Redox-activity and self-organization of iron-porphyrin monolayers at a copper/electrolyte interface Thanh Hai Phan and Klaus Wandelt Citation: The Journal of Chemical Physics 142, 101917 (2015); doi: 10.1063/1.4906892 View online: http://dx.doi.org/10.1063/1.4906892 View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/142/10?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Redox-active on-surface polymerization of single-site divalent cations from pure metals by a ketonefunctionalized phenanthroline J. Chem. Phys. 142, 101913 (2015); 10.1063/1.4906894 In operando observation system for electrochemical reaction by soft X-ray absorption spectroscopy with potential modulation method Rev. Sci. Instrum. 85, 104105 (2014); 10.1063/1.4898054 Diffusional motion of redox centers in carbonate electrolytes J. Chem. Phys. 141, 104509 (2014); 10.1063/1.4894481 Selective detection of Cr(VI) using a microcantilever electrode coated with a self-assembled monolayer J. Vac. Sci. Technol. A 23, 1022 (2005); 10.1116/1.1943456 Electrical characterization of redox-active molecular monolayers on SiO 2 for memory applications Appl. Phys. Lett. 83, 198 (2003); 10.1063/1.1584088
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
THE JOURNAL OF CHEMICAL PHYSICS 142, 101917 (2015)
Redox-activity and self-organization of iron-porphyrin monolayers at a copper/electrolyte interface Thanh Hai Phan1,2,3,a) and Klaus Wandelt1,4 1
Institute of Physical and Theoretical Chemistry, University of Bonn, Wegelerstr. 12, D-53115 Bonn, Germany Division of Molecular Imaging and Photonics, Department of Chemistry, KU Leuven—University of Leuven, Celestijnenlaan 200F, B 3001 Leuven, Belgium 3 Physics Department, Quynhon University, 170 An Duong Vuong, Quynhon, Vietnam 4 Institute of Experimental Physics, University of Wroclaw, MaxaBorna 9, 50-204 Wroclaw, Poland 2
(Received 5 November 2014; accepted 14 January 2015; published online 12 February 2015) The electrochemical behaviour and molecular structure of a layer of water-soluble 5,10,15,20Tetrakis-(N-methyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate, abbreviated as FeTMPyP, on a chloride modified Cu(100) electrode surface were investigated by means of cyclic voltammetry (CV) and in-situ electrochemical scanning tunneling microscopy. Voltammetric results of HOPG in an electrolyte containing FeTMPyP molecules indicate three distinguishable redox steps involving both the central iron metal and the π-conjugated ring system. However, only the first two reduction steps are observable within the narrow potential window of CVs of Cu(100) measured in the same electrolyte. In the potential range below the first reduction peak, at which the [FeIIITMPyP]5+ molecules are reduced to the corresponding [FeIITMPyP]4+ species, in-situ scanning tunneling microscopy (STM) images revealed, for the first time, a highly ordered adlayer of this reduced porphyrin species on the chloride terminated Cu(100) surface. The ordered adlayer exhibits a (quasi)square unit cell with the lattice vectors |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm and an angle of 93◦ ± 2◦ between them. A model is proposed based on the STM observation illustrating the arrangement of the [FeIITMPyP]4+ molecules at the electrolyte/copper interface. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4906892]
I. INTRODUCTION
Porphyrins and related metal porphyrins are of great interest because of their crucial roles in many biological processes, like photosynthesis,1 electrocatalysis,2,3 as well as in molecular devices.4 Due to their importance in both fundamental research and applied areas, the formation and characterization of ordered layers of porphyrin molecules have, besides in ultrahigh vacuum (UHV), also been extensively studied in the recent decades by electrochemical and spectroscopic5–8 methods as well as in-situ scanning tunneling microscopy (STM).9–24 Itaya’s group reported on highly ordered porphyrin arrays formed either on bare gold surfaces13–15 or on iodide and/or sulfur modified Au surfaces.9–12 Iron- and zinc-protoporphyrin adlayers formed on graphite in 0.1 M Na2B4O7 solution were also reported by Tao et al.16,17 Recently, potential controlled manipulation of surface mobility and redox activity of metal free watersoluble porphyrin derivatives such as TPyP on Au(111), and H2TMPyP and H2TTMAPP on copper electrodes was performed by Borguet et al.,18–20 Hai et al.,21 and in our previous works.22–24 Among these studies, only very few papers addressed the electrochemistry and self-organization of adlayers of iron porphyrin molecules, which are, for instance, applied as catalysts for the reduction of oxygen in developing efficient fuel a)Author to whom correspondence should be addressed. Electronic mail:
[email protected] 0021-9606/2015/142(10)/101917/8/$30.00
cells.2,3 In particular, Itaya et al. successfully demonstrated an influence of an iron octaethylporphyrin adlayer on the electrocatalytic reduction of O2 on an Au(111) electrode.13 However, investigations on the electrochemical properties and the formation of ordered adlayers of the iron porphyrin on other surfaces carried out under electrochemical environment are still missed. Therefore in the present study, we choose to study the adsorption and ordering of a redox active watersoluble iron porphyrin, namely, 5,10,15,20-Tetrakis-(Nmethyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate (abbreviated as FeTMPyP), on a Cu(100) surface in hydrochloric acid (10 mM HCl) as the supporting electrolyte. Under these conditions, the chloride modified Cu(100) electrode has proven to be excellent templates for the ordered deposition of organic cations from solution.21–24 The chemical structure of the FeTMPyP molecule is shown in Fig. 1. A copper substrate was employed as the working electrode with a more negative potential window compared to the noble electrodes as Au and Pt in order to be able to study further reduction processes. II. EXPERIMENTAL
All cyclic voltammetry (CV) and scanning tunneling microscopy experiments were carried out in-situ with the use of an electrochemical scanning tunneling microscopy (ECSTM) apparatus designed at the University of Bonn which has been described in detail elsewhere.25,26 In short, this instrument combines a STM with an electrochemical cell
142, 101917-1
© 2015 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-2
T. H. Phan and K. Wandelt
FIG. 1. Chemical structure of iron porphyrin (FeTMPyP) molecule.
of a volume of 2.5 ml. This combination enables a direct comparison of in-situ STM derived structure data with electrochemical measurements, specifically cyclic voltammograms, in the same cell. Furthermore, by means of a specifically coupled design of the bipotentiostate with the electronics of the STM, it is possible to register tunneling currents in three different detection modes, namely, (i) at constant electrode potential and constant bias voltage (potentiostatic imaging mode), (ii) during changing electrode potential, i.e., during the scan of a cyclic voltammogram, and constant bias (potentiodynamic imaging mode), and (iii) at a fixed surface position with constant electrode potential but varying bias voltage (local spectroscopic mode). All functions of the STM are fully software controlled including safety measures against tip-sample collisions. All STM images shown in the following were recorded in the constant current mode while the sample potential was kept constant with respect to the reference electrode. The tunneling tips used in our experiments were electrochemically etched from a 0.25 mm tungsten wire in 2M KOH solution, rinsed with water, dried, and subsequently coated by piercing the tip through a lamella of hot-melt glue. For all solutions, high purity water (Milli-Q purification system, conductivity >18 MΩ.cm, TOC < 4ppb), halidefree redox-active water-soluble iron porphyrin (5,10,15,20Tetrakis-(N-methyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate [FeTMPyP]) purchased from Frontier Scientific Company, and other reagent grade chemicals were used. Before use, all electrolyte solutions were purged with oxygen free argon gas for several hours. The potential of the copper electrode is quoted with respect to a reversible hydrogen electrode (RHE), and a Pt wire is employed as counter-electrode. In order to guarantee a reproducibly smooth surface even after several electro-polishing cycles, a surface orientation of less than 0.5◦ off the Cu(100) plane was required (MaTeck Company, Julich, Germany). Before each STM experiment, the Cu(100) sample was electropolished in 50% orthophosphoric acid in order to remove the native oxide film formed in
J. Chem. Phys. 142, 101917 (2015)
air as well as other contaminations on the surface. After this etching procedure of the sample at an anodic potential of 2 V applied between the copper electrode and a platinum foil for about 20 to 40 s, the copper surface was rinsed with degassed 10 mM hydrochloric acid solution and then mounted into the electrochemical cell of the home-built EC-STM. In order to eliminate the influence of oxygen as well as acoustic and electromagnetic interferences from all cyclic voltammogram and EC-STM measurements, the whole ECSTM system is housed within a sealed aluminum chamber with electrical and liquid feedthroughs and filled with suprapure argon gas.26 Initial CV measurements were first performed in the pure supporting electrolyte (10 mM HCl) because this significantly reduces the high density of surface defects deriving from the electropolishing process. This so-called “electrochemical annealing” mechanism is verified by the subsequent insitu STM measurements. For the adsorption of the redoxactive porphyrin molecules on the chloride precovered copper electrode, the pure supporting electrolyte was substituted by a 10 mM HCl solution containing 0.1 mM FeTMPyP. This electrolyte exchange procedure was performed under potential control within the range of the double layer regime of the Cu(100) in hydrochloric acid, e.g., between E = +100 mV ÷ −100 mV and RHE.
III. RESULTS AND DISCUSSION A. Chloride modified Cu(100) surface
In this work, a chloride modified Cu(100) electrode serves as the template for the adsorption of the FeTMPyP molecules. Chloride anions are well known to adsorb strongly on the bare Cu(100) surface forming a well ordered c(2 × 2) superstructure.27,28 This structure is stable in the potential regime between the anodic copper dissolution reaction (CDR) and close to the onset of the cathodic hydrogen evolution reaction (HER) (see black-dotted CV in Fig. 4). Typical medium scale and high resolution STM images of the fourfold symmetric chloride adlayer covering the positively polarized Cu(100) electrode surface are presented in Fig. 2. First of all, the STM results show the significant change in the surface morphology of the Cu(100) electrode due to the “electrochemical annealing effect,” namely, there are large defect-free (100) terraces covered by the c(2 × 2)-Cl structure, and very long chloride anion stabilized copper steps along the [001] and the [010] directions of the copper substrate as depicted in Fig. 2(a).29,30 This new topographic equilibrium of the surface is a consequence of the strong electrostatic copperchloride interaction. The step-edges are aligned parallel to the close packed chloride rows whose direction is rotated by 45◦ with respect to the main symmetry axes of the copper substrate underneath. This results in the perfectly straight and atomically smooth appearance of the step edges, i.e., defects such as kink sites are drastically disfavored. In contrast to larger halides such as bromide and iodide which are known to be either less charged (Br) or uncharged (I), respectively, the chloride anions still remain to a large extent negatively charged upon adsorption.31 A chloride modified
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-3
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 3. CVs of HOPG in pure supporting 10 mM HCl electrolyte (dotted black curve) and in solution containing [FeIIITMPyP]5+ cations (solid red curve). Note the set of redox peak pairs P1/P′1, P2/P′2, and P3/P′3 due to the presence of the porphyrin cations. Inset: the dashed blue curve is a CV of Cu(100) in the same solution containing [FeIIITMPyP]5+ cations.
FIG. 2. Medium scale and high resolution STM images of a chloride covered Cu(100) electrode surface: (a) copper steps are preferentially aligned parallel to the close packed chloride rows: 9.92 nm × 9.92 nm, Ub = +40 mV, It = 5 nA, E = +20 mV; (b) c (2 × 2)-Cl structure on Cu(100): 4.37 nm × 4.37 nm, Ub = +20 mV, It = 2.5 nA, E = −10 mV; (c) structural model of the c (2 × 2)-Cl adlayer forming on a Cu(100) surface.
Cu(100) electrode is, therefore, from the electrochemistry point of view, a suitable template for the adsorption of positively charged organic cations due to the electrostatic interaction enhancement between the organic adlayer and the chloride lattice underneath.21,22 A high-resolution STM image of the chloride terminated Cu(100) electrode at (an √ atomic √ ) level with the typical quadratic features of the 2× 2 R45◦ (or c(2 × 2)) superstructure as well as the corresponding hard sphere model illustrating two successive layers and the unit cell are presented in Figs. 2(b) and 2(c). These √ STM measurements yield lattice constants which are 2 times longer than the nearest neighbor distance of copper atoms on the Cu(100) surface, namely, 3.6 ± 0.1 Å.29 B. Electrochemical characterization
Fig. 3 shows representative steady state CVs of HOPG in the absence (pure electrolyte—dotted black curve) and in the presence of iron porphyrin species in solution (working electrolyte—solid red curve). The CV of the HOPG in pure electrolyte is very broad, featureless, and limited by two characteristic reactions, namely, the reductive HER at the negative limit and the oxidative oxygen evolution reaction (OER) determining the positive limit. This CV is an ideal background to study the redox behaviour of organic molecules over a wide potential range. Indeed, in the FeTMPyP containing electrolyte, the CV of the HOPG exhibits pronounced changes manifested by
additional cathodic and anodic pairs of peaks. Based on the observed CVs and previous reports,32–35 the reversible pair of peaks with P1 = +240 mV and P1′ = +320 mV vs RHE can be assigned to the first one-electron transfer step resulting in a reduction/oxidation of the central iron atom from [FeIIITMPyP]5+ to [FeIITMPyP]4+ and vice versa as described in Eq. (1), [FeIIITMPyP]5+ + e− [FeIITMPyP]4+.
(1)
These results are in good agreement with former reports.32–35 Forshey and Kuwana32 found that in acidic solution, the [FeIIITMPyP]5+ species undergoes a one-electron reduction in order to form the corresponding [FeIITMPyP]4+ species at E = +0.18 V vs NHE (Normal Hydrogen Electrode) at a highly polished glassy carbon electrode. Both the oxidized and the reduced form of iron porphyrin remained stable as evidenced by the reproducibility of the respective spectral features when the redox states were cycled and characterized spectroelectrochemically by using an Au minigrid electrode.33 This was also found by Chen et al.,34 who reported on the electrochemistry of the same molecule in phosphate buffer solution (PBS) on a glassy carbon working electrode. The first reduction observed at E = −0.18 V vs Ag/AgCl was consistent with what was reported by Forshey and Kuwana.32 At more negative potentials closer to the HER, further reduction peaks P2 and P3 are detected at P2 = −455 mV and P3 = −680 mV, respectively (Fig. 3). It seems likely that these waves are further reduction steps which relate to the πconjugated ring system of the porphyrin ligand. This hypothesis is supported by Van Caemelbecke et al.,36 who reported that there were three additional reduction steps including six electrons in total after the initial Fe(III)/Fe(II) reduction process as indicated by corresponding peaks in the CV of [(TMPyP)FeCl]4+(Cl−)4 in non-aqueous dimethylformamide (DMF) medium. The first three reduction peaks mentioned in this work are almost congruent with P1, P2, and P3 observed
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-4
T. H. Phan and K. Wandelt
FIG. 4. CVs of Cu(100) in the pure supporting 10 mM HCl electrolyte (dotted black curve) and in electrolyte containing [FeIIITMPyP]5+ cations (10 mM HCl + 0.1 mM FeTMPyP) (solid blue curve).
in our red CV of HOPG (see Fig. 3), whereas a fourth reduction peak with P4 = −1.01 V vs SCE (Saturated Calomel Electrode) is probably hidden in our CV by the hydrogen evolution reaction. Here we present, for the first time, the CV of a Cu(100) surface recorded in an aqueous chloride and FeTMPyP molecules containing electrolyte. Fig. 4 compares typical CVs of a Cu(100) electrode in the pure supporting electrolyte (10 mM HCl) given by the dotted black curve and in [FeIITMPyP]4+ molecules containing electrolyte (10 mM HCl + 0.1 mM FeTMPyP) represented by the blue curve. The CV of the Cu(100) electrode in the pure supporting electrolyte is characterized by only two typical reactions, namely, the reductive hydrogen evolution reaction at the cathodic limit and the oxidative copper dissolution reaction at the anodic limit. The negative peak in the black-dotted curve around +200 mV corresponds to the reductive redeposition of copper during the scan towards negative potentials. After replacing the supporting electrolyte by the [FeIIITMPyP]5+ containing electrolyte, some characteristic changes within the potential window of the Cu(100) are observed. The assignment of P1 and P2 to reduction peaks results from a comparison to the CV of HOPG in the same electrolyte as shown in Fig. 3. Therefrom, it is deduced that the [FeIIITMPyP]5+ molecules react immediately at the interface after being introduced into solution at a potential between +100 mV and −100 mV (see Sec. II), i.e., below a value of the first reduction peak P1 (Fig. 3), leading to the reduction of the iron centre and formation of the product [FeIITMPyP]4+ according to Eq. (1). Consequently, the highly ordered adlayer on the chloride terminated Cu(100) surface, as described in Sec. III C, is due to the adsorbed cationic [FeIITMPyP]4+ species. C. Structural characterization of the iron porphyrin adlayer
For the controlled iron porphyrin adsorption, the pure supporting electrolyte (10 mM HCl) is replaced at a potential value of E = −50 mV vs RHE by an electrolyte containing
J. Chem. Phys. 142, 101917 (2015)
[FeIIITMPyP]5+ cationic species. As a result, the whole electrode surface is covered by a highly ordered monolayer of the reduced product, i.e., [FeIITMPyP]4+ molecules, as discussed in Sec. III B. A typical large-scale STM image showing steps and rotational domains marked by I and II and I′ and II′, respectively, is presented in Fig. 5(a). The molecules are arranged in rows whose direction is rotated by an angle of 17◦ with respect to the step directions. Interestingly, the angle of 34◦ ± 1◦ between the two rotational domains on the same terrace is twice as big as the angle between a step direction and the adjacent molecular row. This indicates that these observed domains are mirror domains. It is well documented that the step directions of chloride pre-covered Cu(100) electrode surfaces are aligned along the close-packed chloride rows which run parallel to the 001 directions of the Cu(100) surface underneath.28,37,38 Consequently, the direction of molecular rows in mirrored domains are rotated by ±17◦ with respect to a [100] direction of the surface. Thus, for reasons of symmetry, there are four mirror domains of the organic molecules in total co-existing on the chloride terminated Cu(100) substrate. The step angle of 90◦ (Fig. 5(a)) characteristic for the chloride precovered Cu(100) surface remains stable in the presence of the adsorbed [FeIITMPyP]4+ species, indicating that the adsorption of the [FeIITMPyP]4+ molecules has no significant impact on the underlying chloride lattice. A close-up of the ordered [FeIITMPyP]4+ structure in domain II is presented in Fig. 5(b). Obviously, the molecules are organized in rows forming a quasi-quadratic lattice on the chlorinated template. Individual molecules are recognized as propellers and arranged in the same orientation. The symmetry axis of the [FeIITMPyP]4+ molecules (dotted yellow line in Fig. 5(c)) is rotated by an angle of 22◦ with respect to the molecular row direction (dashed-dotted white line in Fig. 5(c)). In contrast to the case of metal-free porphyrins,22,36 the molecular centres of the [FeIITMPyP]4+ species appear as protrusions instead of hollows. This is illustrated by the line profile in Fig. 5(d) recorded along the white solid line in Fig. 5(c). The bright central dots are, thus, assigned to the iron atoms in the porphyrin molecules (see the molecular model in Fig. 5(c)). This observation is in agreement with previous reports which investigated the adsorption of iron porphyrin (FeTPP) and phthalocyanine molecules (FePcs) on metal39–42 and on HOPG43 under UHV environment. Buchner et al. reported on the incorporation of iron atoms into the molecules of a self-assembled metal-free porphyrin (2HTPP) layer on a Ag(111) surface;39 the Fe atoms coordinated into the 2HTPP molecules via a replacement of the central hydrogen atoms made the porphyrin centres appear much brighter like in our case of direct adsorption of iron porphyrin. Bai et al.40 and Ahlund et al.,43 found that FePcs adsorbed on either a bare Ag(111) or a bare HOPG substrate form a quasiquadratic lattice in which the individual molecule appears like a propeller enclosing a pronounced protrusion in the centre. Recently, Manandhar et al. also succeeded in studying FePcs based thin films on a bare Ag(111) substrate and made the same observation.41 A reasonable interpretation for the observed protrusions is based on the electron configuration of the central iron atom. For 3d transition metal compounds, the highest
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-5
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 5. (a) Typical large scale STM image of an [FeIITMPyP]4+ adlayer on a chloride modified Cu(100) surface with two mirror domains enclosing an angle of 34◦ between each other: 36.1 nm × 36.1 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (b) medium scale STM image showing the molecules forming a square lattice: 14.4 nm × 14.4 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (c) high resolution STM image in which each molecule can be recognized as a propeller shape including a central protrusion and being rotated by an angle of 22◦ with respect to the direction of the molecular rows (dotted yellow and white dashed-dotted lines): 3.7 nm × 3.7 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (d) the line profile measured along the solid white line in Fig. 5(c) indicates the Fe atom at the molecular center.
occupied (HOMO) and the lowest unoccupied (LUMO) electronic levels are dominated by the 3d orbitals.44,45 By changing the metal in a metal-porphyrin/phthalocyanine system leads to a profound change in the STM images. In particular, for iron atoms, the Fe 3d states are partially occupied and only 0.4 eV apart from the energy level of the porphyrin HOMO46 resulting in a significant contribution to the ligand HOMO of iron porphyrin. This also becomes evident in the case of cobalt containing porphyrin and/or phthalocyanine molecules.44,47 In contrast, a central depression is imaged with Ni, Cu, and Zn atoms within these molecules44,48,49 due to the fact that the d orbitals of these metal atoms are further separated from the ligand HOMO.50 In conclusion, the bright protrusions observed at the center of the [FeIITMPyP]4+ molecules in our STM images can be assigned to a tunneling effect from the partially filled d orbital of Fe. By varying of the tunneling conditions both the iron porphyrin overlayer and the underlying c(2 × 2)-Cl lattice can be observed separately. Fig. 6(a) shows one STM image, scanned from the upper left to the lower right corner, covering three different surface regions 1-3. At low tunneling current and high bias voltage, the [FeIITMPyP]4+ adlayer, actually consisting of two mirror domains (I and II), is imaged as shown in the upper section 1 of Fig. 6(a). After increasing the tunneling current and/or lowering the bias voltage, the image changes as demonstrated by section 2; the molecular structure is gone and the c(2 × 2)Cl underlayer is visible. Re-increase of the bias voltage and/or re-lowering of the tunneling current lead to the reappearance of the molecular structure in section 3, now in the form of one homogeneous domain. In principle, there are three possible explanations for the structure changes
between the three sections. (i) The three structurally different regions exist on the surface prior to imaging, and it is pure accident that the changes in tunneling conditions coincide with the boundaries 1-2 and 2-3, respectively. (ii) The surface area corresponding to the total image in Fig. 6(a) is fully covered with the [FeIITMPyP]4+ adlayer (also in section 2), but under the different tunneling conditions in section 2 (compared to 1 and 3), the organic adlayer is “transparent,” i.e., the electrons tunnel through the molecules which remain invisible. (iii) Under the “moderate tunneling conditions” in sections 1 and 3, the tunneling tip is further away from the measured surface (Fig. 6(b)) and yields an image of the undisturbed [FeIITMPyP]4+ adlayer. Conversely, under the more “drastic tunneling conditions” in section 2, the tunneling tip is much closer to the surface (Fig. 6(c)) and acts as a brush sweeping the [FeIITMPyP]4+ molecules away but leaving the c(2×2)Cl− lattice behind. While the first explanation is highly unlikely, a distinction between the two other explanations would require further investigations, either theoretical calculations on the electronic structure of the adlayer and simulations of the resulting STM images or experiments in which the tipinduced removal of the molecules under “drastic tunneling conditions” followed by the restoration of the adlayer after return to “moderate tunneling conditions” is clearly shown by scanning the same surface region three times under the conditions of the sections 1, 2, and 3 in Fig. 6(a). Even though we favor the third explanation, a commitment to one of the two explanations (ii) or (iii) is actually not necessary for the purpose Fig. 6(a) is shown for. Figures like Fig. 6(a) serve to correlate the structures of organic overlayer and Cl− underlayer, no matter what the ultimate
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-6
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 6. (a) Imaged surface structure at the different tunneling conditions given next to panels (b) and (c): 14.4 nm × 14.4 nm, It = 0.2-3.6-0.1 nA, Ub = 200-29-175 mV, E = −50 mV; (b)-(c) illustration of the different tunneling conditions applied in the individual sections 1-3 in Fig. 6(a).
reason is for the co-appearance of both structures in the same STM image. Fig. 6(a) enables a direct correlation between the two lattices of the [FeIITMPyP]4+ adlayer and the c(2 × 2)Cl underlayer. This correlation is presented in Fig. 7, both panels of which are recorded successively at the same surface area but under different, i.e., moderate (a) and drastic (b) tunneling conditions, respectively. In panel (a), individual molecules can be recognized by their propeller like shape enclosing the metal induced protrusion in their center. Conversely, in panel (b) under the drastic tunneling conditions, the [FeIITMPyP]4+ molecules are no longer visible, only the c(2 × 2) chloride lattice is left behind. From a superposition ( ) of the lattices in panels (a) and (b), the unit cell ⃗a2, ⃗b2 of the molecular overlayer lattice containing one [FeIITMPyP]4+ molecule can be described by a 41 −1 4 transformation matrix with respect to the underlying chloride lattice, with |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm enclosing an angle of 93◦ ± 2◦. Based on results like those shown in Figs. 6(a) and 7 suggesting commensuracy between the lattices of the organic overlayer and the Cl− - underlayer, a model is proposed in Fig. 8 in which two mirror domains are rotated with respect to each other by an angle of 32◦, i.e., slightly smaller than that (34◦) measured in the STM image. On the one hand, this small deviation of 2◦ might arise from slight drift during measurement and small metric errors in the image analysis. On the other hand, the condition of commensuracy is not unreasonable in view of a strong interaction between the iron centre and a chloride anion underneath. Based on this model,
the surface coverage per molecular domain is calculated to be 0.059 Ml with respect to the c(2 × 2)-Cl underlying lattice or 4.27 × 1013 molecules/cm2. It is interesting to compare the structure obtained here with [FeIITMPyP]4+ to that found for the iron-free [TMPyP]4+ porphyrin on the same substrate, i.e., a chloride precovered Cu(100) surface. Table I summarizes the values for the orientation of the molecular rows with respect to high symmetry substrate directions, for the rotational orientation of individual molecules with respect to the row directions as well as for the lattice parameters of the porphyrin layer. We note that there are only very small differences, if at all. A plausible explanation for this similar behaviour is the same charge state of both species, namely, 4+. As a result, two important driving forces in the self-organization process, namely, the electrostatic interactions between the adsorbed molecular cations, on the one hand, and between these cations and the negative anion layer underneath, on the other hand, are very similar. Finally, a word needs to be said about a possible influence of the tosylate counter-anions of the [FeIIITMPyP]5+ cations in solution. In fact, in an earlier publication,52 we could show ex situ with S(2p) XPS spectra that sulfate anions (SO42−) do coadsorb with 3-pyridyloxy-appended phthalocyanine on
FIG. 7. Structural relation between an [FeIITMPyP]4+ adlayer and the chloride lattice underneath, (a) 7.74 nm × 7.74 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (b) 7.74 nm × 7.74 nm, It = 5 nA, Ub = 40 mV, E = −30 mV.
FIG. 8. Possible structure model consisting of two mirror domains of an [FeIITMPyP]4+ adlayer enclosing an angle of 32◦ on a chloride terminated Cu(100).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-7
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
TABLE I. Adsorbed structure comparison between [FeIITTMAPP]4+ and [TMPyP]4+ on Cl/Cu(100). [FeIITMPyP]4+ Rotation of molecular rows with respect to substrate [100] directions Angle between mirror domains Rotation of individual molecules with respect to molecular row directions Lattice parameters Molecular density
[TMPyP]4+ (Ref. 51)
17◦
14◦
34◦ 22◦
28◦ 17◦
⃗ 2 = 1.53 ± 0.1 nm |a ⃗ 2| = b
⃗ 2 = 1.49 ± 0.1 nm |a ⃗ 2| = b
α = 93◦ ± 2◦ 4.27 × 1013 molecules/cm2
an iodide-modified Cu(100) electrode surface. Likewise, it was shown with S(2p) XPS spectra that tosylate anions coadsorb with tetra-methylpyridyl-porphyrin on a chloride precovered Cu(100) surface.53 This indicates at least that the tosylate anions are transferred into UHV and show up in the XPS measurements. It is therefore more than likely that also in the present case of the adsorbed [FeIITMPyP]4+ layer, tosylate anions are coadsorbed with the Fe-porphyrin cations. Regrettably in the latter case, we do not have XPS data. Moreover, in none of the cases we have explicit information about an influence of the coadsorbed tosylate anions on the structure of the organic layer. Available XPS data suggest their presence, but in situ STM images do not show them explicitly. It remains therefore reserved to future measurements (of other groups) to study this possible influence by varying systematically the counter-ion of the respective porphyrins/phthalocyanines.
IV. CONCLUSION
Self-organization and electrochemical behaviour of iron porphyrin molecules on a chloride modified Cu(100) electrode surface were studied by cyclic voltammetry and scanning tunneling microscopy in acidic solution. The [FeIIITMPyP]5+ molecules undergo two reduction processes within the narrow potential window of the Cu(100) electrode in 10 mM HCl solution (from −450 mV to +300 mV vs RHE). The first reduction step related to the transfer of one electron is assigned to the reduction of the [FeIIITMPyP]5+ to the corresponding [FeIITMPyP]4+ species. The ligand related reduction due to further electron transfer steps, as verified by CV measurements with a HOPG electrode, occurs at potentials within or even below the cathodic limit of the CV with Cu(100). Thus, investigations of the structural arrangement of the organic cations on the chloride precovered Cu(100) surface were only possible with the first reduction product, [FeIITMPyP]4+. A highly ordered layer of the [FeIITMPyP]4+ species could be observed by in-situ STM on the Cl/Cu(100) substrate with sub-molecular resolution. Individual molecules are lying flat on the electrode surface and are recognized by their propeller shape with a bright central spot due to the iron atom contribution. The molecules form rows in directions ±17◦ off the high symmetry [001] directions of the copper substrate, with all molecules being uniformly rotated by 27◦ with respect to the row direction. As a consequence, by reasons of symmetry, the overlayer consists of domains of four equivalent
α = 90◦ ± 2◦ 4.29 × 1013 molecules/cm2
rotational orientations. Based on a direct correlation between the [FeIITMPyP]4+ adlayer and the underlying chloride lattice, the unit cell containing one [FeIITMPyP]4+ molecule can be described by the 14 −1 4 transformation matrix with respect to the underlying chloride lattice resulting in the lattice constants |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm and an angle of 93◦ ± 2◦ between them. The surface coverage per domain is 4.27 × 1013 molecules/cm2. All in all the structure of the adsorbed [FeIITMPyP]4+ layer turns out to be very similar to the one found for the same iron-free porphyrin molecules on the same substrate, most likely due to the same charge state of both species. A possible influence of the tosylate counterions on the structure of both the Fe-containing and the metal-free tetra-methylpyridyl-porphyrin layers remains to be studied. ACKNOWLEDGMENTS
The authors gratefully acknowledge the Deutsche Forschungsgemeinschaft (DFG) for financial support via the SFB 624 project. 1Electron
Transfer in Chemistry, edited by V. Balzani (Wiley–VCH, Weinheim, 2001), Vol. 3. 2E. Yeager, Electrochim. Acta 29, 1527 (1984). 3J. P. Collman, P. S. Wagenknecht, and J. E. Hutchison, Angew. Chem., Int. Ed. 33, 1537 (1994). 4Molecular Electronics, edited by J. Jortner and M. Ratner (IUPAC, Oxford, 1997). 5S. Ye, A. Kathiravan, H. Hayashi, Y. Tong, Y. Infahsaeng, P. Chabera, T. Pascher, A. P. Yartsev, S. Isoda, H. Imahori, and V. Sundström, J. Phys. Chem. C 117, 6066 (2013). 6S. Cherian and C. C. Wamser, J. Phys. Chem. B 104, 3624 (2000). 7N. C. Berner, Y. N. Sergeeva, N. N. Sergeeva, M. O. Senge, A. A. Cafolla, and I. T. McGovern, Phys. Status Solidi C 9, 1404 (2012). 8V. Arima, R. I. R. Blyth, F. Della Sala, R. Del Sole, F. Matino, G. Mele, G. Vasapollo, R. Cingolani, and R. Rinaldi, Mater. Sci. Eng., C 24, 569 (2004). 9K. Itaya, Prog. Surf. Sci. 58, 121 (1998). 10M. Kunitake, N. Batina, and K. Itaya, Langmuir 11, 2337 (1995). 11N. Batina, M. Kunitake, and K. Itaya, J. Electroanal. Chem. 405, 245 (1996). 12M. Kunitake, U. Akiba, N. Batina, and K. Itaya, Langmuir 13, 1607 (1997). 13S. Yoshimoto, J. Inukai, A. Tada, T. Abe, T. Morimoto, A. Osuka, H. Furuta, and K. Itaya, J. Phys. Chem. B 108, 1948 (2004). 14S. Yoshimoto, A. Tada, K. Suto, R. Narita, and K. Itaya, Langmuir 19, 672 (2003). 15S. Yoshimoto, K. Sato, S. Sugawara, Y. Chen, O. Ito, T. Sawaguchi, O. Niwa, and K. Itaya, Langmuir 23, 809 (2007). 16N. J. Tao, G. Cardenas, F. Cunha, and Z. Shi, Langmuir 11, 4445 (1995). 17N. J. Tao, Phys. Rev. Lett. 76, 4066 (1996). 18Y. He, T. Ye, and E. Borguet, J. Am. Chem. Soc. 124, 11964 (2002). 19T. Ye, Y. F. He, and E. Borguet, J. Phys. Chem. B 110, 6141-6147 (2006). 20Y. He and E. Borguet, Angew. Chem., Int. Ed. 46, 6098 (2007).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-8 21N.
T. H. Phan and K. Wandelt
T. M. Hai, K. Wandelt, and P. Broekmann, J. Phys. Chem. C 112, 10176 (2008). 22T. H. Phan and K. Wandelt, Surf. Sci. 607, 82 (2013). 23T. H. Phan and K. Wandelt, Int. J. Mol. Sci. 14, 4498 (2014). 24T. H. Phan and K. Wandelt, Surf. Sci., 631, 207-212 (2015). 25M. Wilms, Ph.D. thesis, University of Bonn, Germany, 1999. 26M. Wilms, M. Kruft, G. Bermes, and K. Wandelt, Rev. Sci. Inst. 70, 3641 (1999). 27D. W. Suggs and A. J. Bard, J. Phys. Chem. B 99, 8349 (1995). 28M. R. Vogt et al., J. Surf. Sci. 399, 49 (1998). 29P. Broekmann, N. T. M. Hai, and K. Wandelt, Surf. Sci. 600, 3971 (2006). 30J. Inukai, Y. Osawa, and K. Itaya, J. Phys. Chem. B 102, 10034 (1998). 31M. T. M. Koper, J. Electroanal. Chem. 450, 189 (1998). 32P. A. Forshey and T. Kuwana, Inorg. Chem. 20, 3 (1981). 33R. F. Pasternack and E. G. Spiro, J. Am. Chem. Soc. 100, 968 (1978). 34J. Chen, O. Ikeda, T. Hatasa, A. Kitajima, M. Miyake, and A. Yamatodani, Electrochem. Commun. 1, 274 (1999). 35M. S. Rana and K. Tamagake, J. Electroanal. Chem. 581, 145 (2005). 36E. Van Caemelbecke, A. Derbin, P. Hambright, R. Garcia, A. Doukkali, A. Saoiabi, K. Ohkubo, S. Fukuzumi, and K. M. Kadish, Inorg. Chem. 44, 3789 (2005). 37O. M. Magnussen, Chem. Rev. 102, 679 (2002). 38D.-T. Pham, K. Gentz, C. Zorlein, N. T. M. Hai, S. L. Tsay, B. Kirchner, S. Kossmann, K. Wandelt, and P. Broekmann, New J. Chem. 30, 1439 (2006).
J. Chem. Phys. 142, 101917 (2015) 39F.
Buchner, K. Flechtner, Y. Bai, E. Zillner, I. Kellner, H. P. Steinrück, H. Marbach, and J. M. Gottfried, J. Phys. Chem. C 112, 15485 (2008). 40Y. Bai, F. Buchner, M. T. Wendahl, I. Kellner, A. Bayer, H. P. Steinrück, H. Marbach, and J. M. Gottfried, J. Phys. Chem. C 112, 6087 (2008). 41K. Manandhar, K. T. Park, S. Ma, and J. Hrbek, Surf. Sci. 603, 636 (2009). 42D. Écija, M. Trelka, C. Urban, P. de Mendoza, E. Mateo-Martí, C. Rogero, J. A. Martín-Gago, A. M. Echavarren, R. Otero, J. M. Gallego, and R. Miranda, J. Phys. Chem. C 112, 8988 (2008). 43J. Ahlund, J. Schnadt, K. Nilson, E. Gothelid, J. Schiessling, F. Besenbacher, N. Martensson, and C. Puglia, Surf. Sci. 601, 3661 (2007). 44X. Lu and K. W. Hipps, J. Phys. Chem. B 101, 5391 (1997). 45M. S. Liao and S. Scheiner, J. Comput. Chem. 23, 1391 (2002). 46M. S. Liao and S. Scheiner, J. Chem. Phys. 114, 9780 (2001). 47S. Yoshimoto, A. Tada, K. Suto, and K. Itaya, J. Phys. Chem. B 107, 5836 (2003). 48T. N. Krauss, E. Barrena, H. Dosch, and Y. Wakayama, ChemPhysChem 10, 2445 (2009). 49S. Yoshimoto, E. Tsutsumi, K. Suto, Y. Honda, and K. Itaya, Chem. Phys. 319, 147 (2005). 50A. Rosa and E. Baerends, J. Inorg. Chem. 33, 584 (1994). 51N. T. M. Hai, Ph.D. thesis, University of Bonn, Germany, 2007. 52T. H. Phan, S. Breuer, D. T. Pham, U. Hahn, T. Torres, and K. Wandelt, J. Phys. Chem. C 118, 457 (2014). 53S. Breuer, Ph.D. thesis, University of Bonn, Germany, 2014.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
THE JOURNAL OF CHEMICAL PHYSICS 142, 101917 (2015)
Redox-activity and self-organization of iron-porphyrin monolayers at a copper/electrolyte interface Thanh Hai Phan1,2,3,a) and Klaus Wandelt1,4 1
Institute of Physical and Theoretical Chemistry, University of Bonn, Wegelerstr. 12, D-53115 Bonn, Germany Division of Molecular Imaging and Photonics, Department of Chemistry, KU Leuven—University of Leuven, Celestijnenlaan 200F, B 3001 Leuven, Belgium 3 Physics Department, Quynhon University, 170 An Duong Vuong, Quynhon, Vietnam 4 Institute of Experimental Physics, University of Wroclaw, MaxaBorna 9, 50-204 Wroclaw, Poland 2
(Received 5 November 2014; accepted 14 January 2015; published online 12 February 2015) The electrochemical behaviour and molecular structure of a layer of water-soluble 5,10,15,20Tetrakis-(N-methyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate, abbreviated as FeTMPyP, on a chloride modified Cu(100) electrode surface were investigated by means of cyclic voltammetry (CV) and in-situ electrochemical scanning tunneling microscopy. Voltammetric results of HOPG in an electrolyte containing FeTMPyP molecules indicate three distinguishable redox steps involving both the central iron metal and the π-conjugated ring system. However, only the first two reduction steps are observable within the narrow potential window of CVs of Cu(100) measured in the same electrolyte. In the potential range below the first reduction peak, at which the [FeIIITMPyP]5+ molecules are reduced to the corresponding [FeIITMPyP]4+ species, in-situ scanning tunneling microscopy (STM) images revealed, for the first time, a highly ordered adlayer of this reduced porphyrin species on the chloride terminated Cu(100) surface. The ordered adlayer exhibits a (quasi)square unit cell with the lattice vectors |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm and an angle of 93◦ ± 2◦ between them. A model is proposed based on the STM observation illustrating the arrangement of the [FeIITMPyP]4+ molecules at the electrolyte/copper interface. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4906892]
I. INTRODUCTION
Porphyrins and related metal porphyrins are of great interest because of their crucial roles in many biological processes, like photosynthesis,1 electrocatalysis,2,3 as well as in molecular devices.4 Due to their importance in both fundamental research and applied areas, the formation and characterization of ordered layers of porphyrin molecules have, besides in ultrahigh vacuum (UHV), also been extensively studied in the recent decades by electrochemical and spectroscopic5–8 methods as well as in-situ scanning tunneling microscopy (STM).9–24 Itaya’s group reported on highly ordered porphyrin arrays formed either on bare gold surfaces13–15 or on iodide and/or sulfur modified Au surfaces.9–12 Iron- and zinc-protoporphyrin adlayers formed on graphite in 0.1 M Na2B4O7 solution were also reported by Tao et al.16,17 Recently, potential controlled manipulation of surface mobility and redox activity of metal free watersoluble porphyrin derivatives such as TPyP on Au(111), and H2TMPyP and H2TTMAPP on copper electrodes was performed by Borguet et al.,18–20 Hai et al.,21 and in our previous works.22–24 Among these studies, only very few papers addressed the electrochemistry and self-organization of adlayers of iron porphyrin molecules, which are, for instance, applied as catalysts for the reduction of oxygen in developing efficient fuel a)Author to whom correspondence should be addressed. Electronic mail:
[email protected] 0021-9606/2015/142(10)/101917/8/$30.00
cells.2,3 In particular, Itaya et al. successfully demonstrated an influence of an iron octaethylporphyrin adlayer on the electrocatalytic reduction of O2 on an Au(111) electrode.13 However, investigations on the electrochemical properties and the formation of ordered adlayers of the iron porphyrin on other surfaces carried out under electrochemical environment are still missed. Therefore in the present study, we choose to study the adsorption and ordering of a redox active watersoluble iron porphyrin, namely, 5,10,15,20-Tetrakis-(Nmethyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate (abbreviated as FeTMPyP), on a Cu(100) surface in hydrochloric acid (10 mM HCl) as the supporting electrolyte. Under these conditions, the chloride modified Cu(100) electrode has proven to be excellent templates for the ordered deposition of organic cations from solution.21–24 The chemical structure of the FeTMPyP molecule is shown in Fig. 1. A copper substrate was employed as the working electrode with a more negative potential window compared to the noble electrodes as Au and Pt in order to be able to study further reduction processes. II. EXPERIMENTAL
All cyclic voltammetry (CV) and scanning tunneling microscopy experiments were carried out in-situ with the use of an electrochemical scanning tunneling microscopy (ECSTM) apparatus designed at the University of Bonn which has been described in detail elsewhere.25,26 In short, this instrument combines a STM with an electrochemical cell
142, 101917-1
© 2015 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-2
T. H. Phan and K. Wandelt
FIG. 1. Chemical structure of iron porphyrin (FeTMPyP) molecule.
of a volume of 2.5 ml. This combination enables a direct comparison of in-situ STM derived structure data with electrochemical measurements, specifically cyclic voltammograms, in the same cell. Furthermore, by means of a specifically coupled design of the bipotentiostate with the electronics of the STM, it is possible to register tunneling currents in three different detection modes, namely, (i) at constant electrode potential and constant bias voltage (potentiostatic imaging mode), (ii) during changing electrode potential, i.e., during the scan of a cyclic voltammogram, and constant bias (potentiodynamic imaging mode), and (iii) at a fixed surface position with constant electrode potential but varying bias voltage (local spectroscopic mode). All functions of the STM are fully software controlled including safety measures against tip-sample collisions. All STM images shown in the following were recorded in the constant current mode while the sample potential was kept constant with respect to the reference electrode. The tunneling tips used in our experiments were electrochemically etched from a 0.25 mm tungsten wire in 2M KOH solution, rinsed with water, dried, and subsequently coated by piercing the tip through a lamella of hot-melt glue. For all solutions, high purity water (Milli-Q purification system, conductivity >18 MΩ.cm, TOC < 4ppb), halidefree redox-active water-soluble iron porphyrin (5,10,15,20Tetrakis-(N-methyl-4-pyridyl)-porphyrin-Fe(III) pentatosylate [FeTMPyP]) purchased from Frontier Scientific Company, and other reagent grade chemicals were used. Before use, all electrolyte solutions were purged with oxygen free argon gas for several hours. The potential of the copper electrode is quoted with respect to a reversible hydrogen electrode (RHE), and a Pt wire is employed as counter-electrode. In order to guarantee a reproducibly smooth surface even after several electro-polishing cycles, a surface orientation of less than 0.5◦ off the Cu(100) plane was required (MaTeck Company, Julich, Germany). Before each STM experiment, the Cu(100) sample was electropolished in 50% orthophosphoric acid in order to remove the native oxide film formed in
J. Chem. Phys. 142, 101917 (2015)
air as well as other contaminations on the surface. After this etching procedure of the sample at an anodic potential of 2 V applied between the copper electrode and a platinum foil for about 20 to 40 s, the copper surface was rinsed with degassed 10 mM hydrochloric acid solution and then mounted into the electrochemical cell of the home-built EC-STM. In order to eliminate the influence of oxygen as well as acoustic and electromagnetic interferences from all cyclic voltammogram and EC-STM measurements, the whole ECSTM system is housed within a sealed aluminum chamber with electrical and liquid feedthroughs and filled with suprapure argon gas.26 Initial CV measurements were first performed in the pure supporting electrolyte (10 mM HCl) because this significantly reduces the high density of surface defects deriving from the electropolishing process. This so-called “electrochemical annealing” mechanism is verified by the subsequent insitu STM measurements. For the adsorption of the redoxactive porphyrin molecules on the chloride precovered copper electrode, the pure supporting electrolyte was substituted by a 10 mM HCl solution containing 0.1 mM FeTMPyP. This electrolyte exchange procedure was performed under potential control within the range of the double layer regime of the Cu(100) in hydrochloric acid, e.g., between E = +100 mV ÷ −100 mV and RHE.
III. RESULTS AND DISCUSSION A. Chloride modified Cu(100) surface
In this work, a chloride modified Cu(100) electrode serves as the template for the adsorption of the FeTMPyP molecules. Chloride anions are well known to adsorb strongly on the bare Cu(100) surface forming a well ordered c(2 × 2) superstructure.27,28 This structure is stable in the potential regime between the anodic copper dissolution reaction (CDR) and close to the onset of the cathodic hydrogen evolution reaction (HER) (see black-dotted CV in Fig. 4). Typical medium scale and high resolution STM images of the fourfold symmetric chloride adlayer covering the positively polarized Cu(100) electrode surface are presented in Fig. 2. First of all, the STM results show the significant change in the surface morphology of the Cu(100) electrode due to the “electrochemical annealing effect,” namely, there are large defect-free (100) terraces covered by the c(2 × 2)-Cl structure, and very long chloride anion stabilized copper steps along the [001] and the [010] directions of the copper substrate as depicted in Fig. 2(a).29,30 This new topographic equilibrium of the surface is a consequence of the strong electrostatic copperchloride interaction. The step-edges are aligned parallel to the close packed chloride rows whose direction is rotated by 45◦ with respect to the main symmetry axes of the copper substrate underneath. This results in the perfectly straight and atomically smooth appearance of the step edges, i.e., defects such as kink sites are drastically disfavored. In contrast to larger halides such as bromide and iodide which are known to be either less charged (Br) or uncharged (I), respectively, the chloride anions still remain to a large extent negatively charged upon adsorption.31 A chloride modified
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-3
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 3. CVs of HOPG in pure supporting 10 mM HCl electrolyte (dotted black curve) and in solution containing [FeIIITMPyP]5+ cations (solid red curve). Note the set of redox peak pairs P1/P′1, P2/P′2, and P3/P′3 due to the presence of the porphyrin cations. Inset: the dashed blue curve is a CV of Cu(100) in the same solution containing [FeIIITMPyP]5+ cations.
FIG. 2. Medium scale and high resolution STM images of a chloride covered Cu(100) electrode surface: (a) copper steps are preferentially aligned parallel to the close packed chloride rows: 9.92 nm × 9.92 nm, Ub = +40 mV, It = 5 nA, E = +20 mV; (b) c (2 × 2)-Cl structure on Cu(100): 4.37 nm × 4.37 nm, Ub = +20 mV, It = 2.5 nA, E = −10 mV; (c) structural model of the c (2 × 2)-Cl adlayer forming on a Cu(100) surface.
Cu(100) electrode is, therefore, from the electrochemistry point of view, a suitable template for the adsorption of positively charged organic cations due to the electrostatic interaction enhancement between the organic adlayer and the chloride lattice underneath.21,22 A high-resolution STM image of the chloride terminated Cu(100) electrode at (an √ atomic √ ) level with the typical quadratic features of the 2× 2 R45◦ (or c(2 × 2)) superstructure as well as the corresponding hard sphere model illustrating two successive layers and the unit cell are presented in Figs. 2(b) and 2(c). These √ STM measurements yield lattice constants which are 2 times longer than the nearest neighbor distance of copper atoms on the Cu(100) surface, namely, 3.6 ± 0.1 Å.29 B. Electrochemical characterization
Fig. 3 shows representative steady state CVs of HOPG in the absence (pure electrolyte—dotted black curve) and in the presence of iron porphyrin species in solution (working electrolyte—solid red curve). The CV of the HOPG in pure electrolyte is very broad, featureless, and limited by two characteristic reactions, namely, the reductive HER at the negative limit and the oxidative oxygen evolution reaction (OER) determining the positive limit. This CV is an ideal background to study the redox behaviour of organic molecules over a wide potential range. Indeed, in the FeTMPyP containing electrolyte, the CV of the HOPG exhibits pronounced changes manifested by
additional cathodic and anodic pairs of peaks. Based on the observed CVs and previous reports,32–35 the reversible pair of peaks with P1 = +240 mV and P1′ = +320 mV vs RHE can be assigned to the first one-electron transfer step resulting in a reduction/oxidation of the central iron atom from [FeIIITMPyP]5+ to [FeIITMPyP]4+ and vice versa as described in Eq. (1), [FeIIITMPyP]5+ + e− [FeIITMPyP]4+.
(1)
These results are in good agreement with former reports.32–35 Forshey and Kuwana32 found that in acidic solution, the [FeIIITMPyP]5+ species undergoes a one-electron reduction in order to form the corresponding [FeIITMPyP]4+ species at E = +0.18 V vs NHE (Normal Hydrogen Electrode) at a highly polished glassy carbon electrode. Both the oxidized and the reduced form of iron porphyrin remained stable as evidenced by the reproducibility of the respective spectral features when the redox states were cycled and characterized spectroelectrochemically by using an Au minigrid electrode.33 This was also found by Chen et al.,34 who reported on the electrochemistry of the same molecule in phosphate buffer solution (PBS) on a glassy carbon working electrode. The first reduction observed at E = −0.18 V vs Ag/AgCl was consistent with what was reported by Forshey and Kuwana.32 At more negative potentials closer to the HER, further reduction peaks P2 and P3 are detected at P2 = −455 mV and P3 = −680 mV, respectively (Fig. 3). It seems likely that these waves are further reduction steps which relate to the πconjugated ring system of the porphyrin ligand. This hypothesis is supported by Van Caemelbecke et al.,36 who reported that there were three additional reduction steps including six electrons in total after the initial Fe(III)/Fe(II) reduction process as indicated by corresponding peaks in the CV of [(TMPyP)FeCl]4+(Cl−)4 in non-aqueous dimethylformamide (DMF) medium. The first three reduction peaks mentioned in this work are almost congruent with P1, P2, and P3 observed
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-4
T. H. Phan and K. Wandelt
FIG. 4. CVs of Cu(100) in the pure supporting 10 mM HCl electrolyte (dotted black curve) and in electrolyte containing [FeIIITMPyP]5+ cations (10 mM HCl + 0.1 mM FeTMPyP) (solid blue curve).
in our red CV of HOPG (see Fig. 3), whereas a fourth reduction peak with P4 = −1.01 V vs SCE (Saturated Calomel Electrode) is probably hidden in our CV by the hydrogen evolution reaction. Here we present, for the first time, the CV of a Cu(100) surface recorded in an aqueous chloride and FeTMPyP molecules containing electrolyte. Fig. 4 compares typical CVs of a Cu(100) electrode in the pure supporting electrolyte (10 mM HCl) given by the dotted black curve and in [FeIITMPyP]4+ molecules containing electrolyte (10 mM HCl + 0.1 mM FeTMPyP) represented by the blue curve. The CV of the Cu(100) electrode in the pure supporting electrolyte is characterized by only two typical reactions, namely, the reductive hydrogen evolution reaction at the cathodic limit and the oxidative copper dissolution reaction at the anodic limit. The negative peak in the black-dotted curve around +200 mV corresponds to the reductive redeposition of copper during the scan towards negative potentials. After replacing the supporting electrolyte by the [FeIIITMPyP]5+ containing electrolyte, some characteristic changes within the potential window of the Cu(100) are observed. The assignment of P1 and P2 to reduction peaks results from a comparison to the CV of HOPG in the same electrolyte as shown in Fig. 3. Therefrom, it is deduced that the [FeIIITMPyP]5+ molecules react immediately at the interface after being introduced into solution at a potential between +100 mV and −100 mV (see Sec. II), i.e., below a value of the first reduction peak P1 (Fig. 3), leading to the reduction of the iron centre and formation of the product [FeIITMPyP]4+ according to Eq. (1). Consequently, the highly ordered adlayer on the chloride terminated Cu(100) surface, as described in Sec. III C, is due to the adsorbed cationic [FeIITMPyP]4+ species. C. Structural characterization of the iron porphyrin adlayer
For the controlled iron porphyrin adsorption, the pure supporting electrolyte (10 mM HCl) is replaced at a potential value of E = −50 mV vs RHE by an electrolyte containing
J. Chem. Phys. 142, 101917 (2015)
[FeIIITMPyP]5+ cationic species. As a result, the whole electrode surface is covered by a highly ordered monolayer of the reduced product, i.e., [FeIITMPyP]4+ molecules, as discussed in Sec. III B. A typical large-scale STM image showing steps and rotational domains marked by I and II and I′ and II′, respectively, is presented in Fig. 5(a). The molecules are arranged in rows whose direction is rotated by an angle of 17◦ with respect to the step directions. Interestingly, the angle of 34◦ ± 1◦ between the two rotational domains on the same terrace is twice as big as the angle between a step direction and the adjacent molecular row. This indicates that these observed domains are mirror domains. It is well documented that the step directions of chloride pre-covered Cu(100) electrode surfaces are aligned along the close-packed chloride rows which run parallel to the 001 directions of the Cu(100) surface underneath.28,37,38 Consequently, the direction of molecular rows in mirrored domains are rotated by ±17◦ with respect to a [100] direction of the surface. Thus, for reasons of symmetry, there are four mirror domains of the organic molecules in total co-existing on the chloride terminated Cu(100) substrate. The step angle of 90◦ (Fig. 5(a)) characteristic for the chloride precovered Cu(100) surface remains stable in the presence of the adsorbed [FeIITMPyP]4+ species, indicating that the adsorption of the [FeIITMPyP]4+ molecules has no significant impact on the underlying chloride lattice. A close-up of the ordered [FeIITMPyP]4+ structure in domain II is presented in Fig. 5(b). Obviously, the molecules are organized in rows forming a quasi-quadratic lattice on the chlorinated template. Individual molecules are recognized as propellers and arranged in the same orientation. The symmetry axis of the [FeIITMPyP]4+ molecules (dotted yellow line in Fig. 5(c)) is rotated by an angle of 22◦ with respect to the molecular row direction (dashed-dotted white line in Fig. 5(c)). In contrast to the case of metal-free porphyrins,22,36 the molecular centres of the [FeIITMPyP]4+ species appear as protrusions instead of hollows. This is illustrated by the line profile in Fig. 5(d) recorded along the white solid line in Fig. 5(c). The bright central dots are, thus, assigned to the iron atoms in the porphyrin molecules (see the molecular model in Fig. 5(c)). This observation is in agreement with previous reports which investigated the adsorption of iron porphyrin (FeTPP) and phthalocyanine molecules (FePcs) on metal39–42 and on HOPG43 under UHV environment. Buchner et al. reported on the incorporation of iron atoms into the molecules of a self-assembled metal-free porphyrin (2HTPP) layer on a Ag(111) surface;39 the Fe atoms coordinated into the 2HTPP molecules via a replacement of the central hydrogen atoms made the porphyrin centres appear much brighter like in our case of direct adsorption of iron porphyrin. Bai et al.40 and Ahlund et al.,43 found that FePcs adsorbed on either a bare Ag(111) or a bare HOPG substrate form a quasiquadratic lattice in which the individual molecule appears like a propeller enclosing a pronounced protrusion in the centre. Recently, Manandhar et al. also succeeded in studying FePcs based thin films on a bare Ag(111) substrate and made the same observation.41 A reasonable interpretation for the observed protrusions is based on the electron configuration of the central iron atom. For 3d transition metal compounds, the highest
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-5
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 5. (a) Typical large scale STM image of an [FeIITMPyP]4+ adlayer on a chloride modified Cu(100) surface with two mirror domains enclosing an angle of 34◦ between each other: 36.1 nm × 36.1 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (b) medium scale STM image showing the molecules forming a square lattice: 14.4 nm × 14.4 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (c) high resolution STM image in which each molecule can be recognized as a propeller shape including a central protrusion and being rotated by an angle of 22◦ with respect to the direction of the molecular rows (dotted yellow and white dashed-dotted lines): 3.7 nm × 3.7 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (d) the line profile measured along the solid white line in Fig. 5(c) indicates the Fe atom at the molecular center.
occupied (HOMO) and the lowest unoccupied (LUMO) electronic levels are dominated by the 3d orbitals.44,45 By changing the metal in a metal-porphyrin/phthalocyanine system leads to a profound change in the STM images. In particular, for iron atoms, the Fe 3d states are partially occupied and only 0.4 eV apart from the energy level of the porphyrin HOMO46 resulting in a significant contribution to the ligand HOMO of iron porphyrin. This also becomes evident in the case of cobalt containing porphyrin and/or phthalocyanine molecules.44,47 In contrast, a central depression is imaged with Ni, Cu, and Zn atoms within these molecules44,48,49 due to the fact that the d orbitals of these metal atoms are further separated from the ligand HOMO.50 In conclusion, the bright protrusions observed at the center of the [FeIITMPyP]4+ molecules in our STM images can be assigned to a tunneling effect from the partially filled d orbital of Fe. By varying of the tunneling conditions both the iron porphyrin overlayer and the underlying c(2 × 2)-Cl lattice can be observed separately. Fig. 6(a) shows one STM image, scanned from the upper left to the lower right corner, covering three different surface regions 1-3. At low tunneling current and high bias voltage, the [FeIITMPyP]4+ adlayer, actually consisting of two mirror domains (I and II), is imaged as shown in the upper section 1 of Fig. 6(a). After increasing the tunneling current and/or lowering the bias voltage, the image changes as demonstrated by section 2; the molecular structure is gone and the c(2 × 2)Cl underlayer is visible. Re-increase of the bias voltage and/or re-lowering of the tunneling current lead to the reappearance of the molecular structure in section 3, now in the form of one homogeneous domain. In principle, there are three possible explanations for the structure changes
between the three sections. (i) The three structurally different regions exist on the surface prior to imaging, and it is pure accident that the changes in tunneling conditions coincide with the boundaries 1-2 and 2-3, respectively. (ii) The surface area corresponding to the total image in Fig. 6(a) is fully covered with the [FeIITMPyP]4+ adlayer (also in section 2), but under the different tunneling conditions in section 2 (compared to 1 and 3), the organic adlayer is “transparent,” i.e., the electrons tunnel through the molecules which remain invisible. (iii) Under the “moderate tunneling conditions” in sections 1 and 3, the tunneling tip is further away from the measured surface (Fig. 6(b)) and yields an image of the undisturbed [FeIITMPyP]4+ adlayer. Conversely, under the more “drastic tunneling conditions” in section 2, the tunneling tip is much closer to the surface (Fig. 6(c)) and acts as a brush sweeping the [FeIITMPyP]4+ molecules away but leaving the c(2×2)Cl− lattice behind. While the first explanation is highly unlikely, a distinction between the two other explanations would require further investigations, either theoretical calculations on the electronic structure of the adlayer and simulations of the resulting STM images or experiments in which the tipinduced removal of the molecules under “drastic tunneling conditions” followed by the restoration of the adlayer after return to “moderate tunneling conditions” is clearly shown by scanning the same surface region three times under the conditions of the sections 1, 2, and 3 in Fig. 6(a). Even though we favor the third explanation, a commitment to one of the two explanations (ii) or (iii) is actually not necessary for the purpose Fig. 6(a) is shown for. Figures like Fig. 6(a) serve to correlate the structures of organic overlayer and Cl− underlayer, no matter what the ultimate
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-6
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
FIG. 6. (a) Imaged surface structure at the different tunneling conditions given next to panels (b) and (c): 14.4 nm × 14.4 nm, It = 0.2-3.6-0.1 nA, Ub = 200-29-175 mV, E = −50 mV; (b)-(c) illustration of the different tunneling conditions applied in the individual sections 1-3 in Fig. 6(a).
reason is for the co-appearance of both structures in the same STM image. Fig. 6(a) enables a direct correlation between the two lattices of the [FeIITMPyP]4+ adlayer and the c(2 × 2)Cl underlayer. This correlation is presented in Fig. 7, both panels of which are recorded successively at the same surface area but under different, i.e., moderate (a) and drastic (b) tunneling conditions, respectively. In panel (a), individual molecules can be recognized by their propeller like shape enclosing the metal induced protrusion in their center. Conversely, in panel (b) under the drastic tunneling conditions, the [FeIITMPyP]4+ molecules are no longer visible, only the c(2 × 2) chloride lattice is left behind. From a superposition ( ) of the lattices in panels (a) and (b), the unit cell ⃗a2, ⃗b2 of the molecular overlayer lattice containing one [FeIITMPyP]4+ molecule can be described by a 41 −1 4 transformation matrix with respect to the underlying chloride lattice, with |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm enclosing an angle of 93◦ ± 2◦. Based on results like those shown in Figs. 6(a) and 7 suggesting commensuracy between the lattices of the organic overlayer and the Cl− - underlayer, a model is proposed in Fig. 8 in which two mirror domains are rotated with respect to each other by an angle of 32◦, i.e., slightly smaller than that (34◦) measured in the STM image. On the one hand, this small deviation of 2◦ might arise from slight drift during measurement and small metric errors in the image analysis. On the other hand, the condition of commensuracy is not unreasonable in view of a strong interaction between the iron centre and a chloride anion underneath. Based on this model,
the surface coverage per molecular domain is calculated to be 0.059 Ml with respect to the c(2 × 2)-Cl underlying lattice or 4.27 × 1013 molecules/cm2. It is interesting to compare the structure obtained here with [FeIITMPyP]4+ to that found for the iron-free [TMPyP]4+ porphyrin on the same substrate, i.e., a chloride precovered Cu(100) surface. Table I summarizes the values for the orientation of the molecular rows with respect to high symmetry substrate directions, for the rotational orientation of individual molecules with respect to the row directions as well as for the lattice parameters of the porphyrin layer. We note that there are only very small differences, if at all. A plausible explanation for this similar behaviour is the same charge state of both species, namely, 4+. As a result, two important driving forces in the self-organization process, namely, the electrostatic interactions between the adsorbed molecular cations, on the one hand, and between these cations and the negative anion layer underneath, on the other hand, are very similar. Finally, a word needs to be said about a possible influence of the tosylate counter-anions of the [FeIIITMPyP]5+ cations in solution. In fact, in an earlier publication,52 we could show ex situ with S(2p) XPS spectra that sulfate anions (SO42−) do coadsorb with 3-pyridyloxy-appended phthalocyanine on
FIG. 7. Structural relation between an [FeIITMPyP]4+ adlayer and the chloride lattice underneath, (a) 7.74 nm × 7.74 nm, It = 0.2 nA, Ub = 240 mV, E = −30 mV; (b) 7.74 nm × 7.74 nm, It = 5 nA, Ub = 40 mV, E = −30 mV.
FIG. 8. Possible structure model consisting of two mirror domains of an [FeIITMPyP]4+ adlayer enclosing an angle of 32◦ on a chloride terminated Cu(100).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-7
T. H. Phan and K. Wandelt
J. Chem. Phys. 142, 101917 (2015)
TABLE I. Adsorbed structure comparison between [FeIITTMAPP]4+ and [TMPyP]4+ on Cl/Cu(100). [FeIITMPyP]4+ Rotation of molecular rows with respect to substrate [100] directions Angle between mirror domains Rotation of individual molecules with respect to molecular row directions Lattice parameters Molecular density
[TMPyP]4+ (Ref. 51)
17◦
14◦
34◦ 22◦
28◦ 17◦
⃗ 2 = 1.53 ± 0.1 nm |a ⃗ 2| = b
⃗ 2 = 1.49 ± 0.1 nm |a ⃗ 2| = b
α = 93◦ ± 2◦ 4.27 × 1013 molecules/cm2
an iodide-modified Cu(100) electrode surface. Likewise, it was shown with S(2p) XPS spectra that tosylate anions coadsorb with tetra-methylpyridyl-porphyrin on a chloride precovered Cu(100) surface.53 This indicates at least that the tosylate anions are transferred into UHV and show up in the XPS measurements. It is therefore more than likely that also in the present case of the adsorbed [FeIITMPyP]4+ layer, tosylate anions are coadsorbed with the Fe-porphyrin cations. Regrettably in the latter case, we do not have XPS data. Moreover, in none of the cases we have explicit information about an influence of the coadsorbed tosylate anions on the structure of the organic layer. Available XPS data suggest their presence, but in situ STM images do not show them explicitly. It remains therefore reserved to future measurements (of other groups) to study this possible influence by varying systematically the counter-ion of the respective porphyrins/phthalocyanines.
IV. CONCLUSION
Self-organization and electrochemical behaviour of iron porphyrin molecules on a chloride modified Cu(100) electrode surface were studied by cyclic voltammetry and scanning tunneling microscopy in acidic solution. The [FeIIITMPyP]5+ molecules undergo two reduction processes within the narrow potential window of the Cu(100) electrode in 10 mM HCl solution (from −450 mV to +300 mV vs RHE). The first reduction step related to the transfer of one electron is assigned to the reduction of the [FeIIITMPyP]5+ to the corresponding [FeIITMPyP]4+ species. The ligand related reduction due to further electron transfer steps, as verified by CV measurements with a HOPG electrode, occurs at potentials within or even below the cathodic limit of the CV with Cu(100). Thus, investigations of the structural arrangement of the organic cations on the chloride precovered Cu(100) surface were only possible with the first reduction product, [FeIITMPyP]4+. A highly ordered layer of the [FeIITMPyP]4+ species could be observed by in-situ STM on the Cl/Cu(100) substrate with sub-molecular resolution. Individual molecules are lying flat on the electrode surface and are recognized by their propeller shape with a bright central spot due to the iron atom contribution. The molecules form rows in directions ±17◦ off the high symmetry [001] directions of the copper substrate, with all molecules being uniformly rotated by 27◦ with respect to the row direction. As a consequence, by reasons of symmetry, the overlayer consists of domains of four equivalent
α = 90◦ ± 2◦ 4.29 × 1013 molecules/cm2
rotational orientations. Based on a direct correlation between the [FeIITMPyP]4+ adlayer and the underlying chloride lattice, the unit cell containing one [FeIITMPyP]4+ molecule can be described by the 14 −1 4 transformation matrix with respect to the underlying chloride lattice resulting in the lattice constants |⃗a2| = ⃗b2 = 1.53 ± 0.1 nm and an angle of 93◦ ± 2◦ between them. The surface coverage per domain is 4.27 × 1013 molecules/cm2. All in all the structure of the adsorbed [FeIITMPyP]4+ layer turns out to be very similar to the one found for the same iron-free porphyrin molecules on the same substrate, most likely due to the same charge state of both species. A possible influence of the tosylate counterions on the structure of both the Fe-containing and the metal-free tetra-methylpyridyl-porphyrin layers remains to be studied. ACKNOWLEDGMENTS
The authors gratefully acknowledge the Deutsche Forschungsgemeinschaft (DFG) for financial support via the SFB 624 project. 1Electron
Transfer in Chemistry, edited by V. Balzani (Wiley–VCH, Weinheim, 2001), Vol. 3. 2E. Yeager, Electrochim. Acta 29, 1527 (1984). 3J. P. Collman, P. S. Wagenknecht, and J. E. Hutchison, Angew. Chem., Int. Ed. 33, 1537 (1994). 4Molecular Electronics, edited by J. Jortner and M. Ratner (IUPAC, Oxford, 1997). 5S. Ye, A. Kathiravan, H. Hayashi, Y. Tong, Y. Infahsaeng, P. Chabera, T. Pascher, A. P. Yartsev, S. Isoda, H. Imahori, and V. Sundström, J. Phys. Chem. C 117, 6066 (2013). 6S. Cherian and C. C. Wamser, J. Phys. Chem. B 104, 3624 (2000). 7N. C. Berner, Y. N. Sergeeva, N. N. Sergeeva, M. O. Senge, A. A. Cafolla, and I. T. McGovern, Phys. Status Solidi C 9, 1404 (2012). 8V. Arima, R. I. R. Blyth, F. Della Sala, R. Del Sole, F. Matino, G. Mele, G. Vasapollo, R. Cingolani, and R. Rinaldi, Mater. Sci. Eng., C 24, 569 (2004). 9K. Itaya, Prog. Surf. Sci. 58, 121 (1998). 10M. Kunitake, N. Batina, and K. Itaya, Langmuir 11, 2337 (1995). 11N. Batina, M. Kunitake, and K. Itaya, J. Electroanal. Chem. 405, 245 (1996). 12M. Kunitake, U. Akiba, N. Batina, and K. Itaya, Langmuir 13, 1607 (1997). 13S. Yoshimoto, J. Inukai, A. Tada, T. Abe, T. Morimoto, A. Osuka, H. Furuta, and K. Itaya, J. Phys. Chem. B 108, 1948 (2004). 14S. Yoshimoto, A. Tada, K. Suto, R. Narita, and K. Itaya, Langmuir 19, 672 (2003). 15S. Yoshimoto, K. Sato, S. Sugawara, Y. Chen, O. Ito, T. Sawaguchi, O. Niwa, and K. Itaya, Langmuir 23, 809 (2007). 16N. J. Tao, G. Cardenas, F. Cunha, and Z. Shi, Langmuir 11, 4445 (1995). 17N. J. Tao, Phys. Rev. Lett. 76, 4066 (1996). 18Y. He, T. Ye, and E. Borguet, J. Am. Chem. Soc. 124, 11964 (2002). 19T. Ye, Y. F. He, and E. Borguet, J. Phys. Chem. B 110, 6141-6147 (2006). 20Y. He and E. Borguet, Angew. Chem., Int. Ed. 46, 6098 (2007).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
101917-8 21N.
T. H. Phan and K. Wandelt
T. M. Hai, K. Wandelt, and P. Broekmann, J. Phys. Chem. C 112, 10176 (2008). 22T. H. Phan and K. Wandelt, Surf. Sci. 607, 82 (2013). 23T. H. Phan and K. Wandelt, Int. J. Mol. Sci. 14, 4498 (2014). 24T. H. Phan and K. Wandelt, Surf. Sci., 631, 207-212 (2015). 25M. Wilms, Ph.D. thesis, University of Bonn, Germany, 1999. 26M. Wilms, M. Kruft, G. Bermes, and K. Wandelt, Rev. Sci. Inst. 70, 3641 (1999). 27D. W. Suggs and A. J. Bard, J. Phys. Chem. B 99, 8349 (1995). 28M. R. Vogt et al., J. Surf. Sci. 399, 49 (1998). 29P. Broekmann, N. T. M. Hai, and K. Wandelt, Surf. Sci. 600, 3971 (2006). 30J. Inukai, Y. Osawa, and K. Itaya, J. Phys. Chem. B 102, 10034 (1998). 31M. T. M. Koper, J. Electroanal. Chem. 450, 189 (1998). 32P. A. Forshey and T. Kuwana, Inorg. Chem. 20, 3 (1981). 33R. F. Pasternack and E. G. Spiro, J. Am. Chem. Soc. 100, 968 (1978). 34J. Chen, O. Ikeda, T. Hatasa, A. Kitajima, M. Miyake, and A. Yamatodani, Electrochem. Commun. 1, 274 (1999). 35M. S. Rana and K. Tamagake, J. Electroanal. Chem. 581, 145 (2005). 36E. Van Caemelbecke, A. Derbin, P. Hambright, R. Garcia, A. Doukkali, A. Saoiabi, K. Ohkubo, S. Fukuzumi, and K. M. Kadish, Inorg. Chem. 44, 3789 (2005). 37O. M. Magnussen, Chem. Rev. 102, 679 (2002). 38D.-T. Pham, K. Gentz, C. Zorlein, N. T. M. Hai, S. L. Tsay, B. Kirchner, S. Kossmann, K. Wandelt, and P. Broekmann, New J. Chem. 30, 1439 (2006).
J. Chem. Phys. 142, 101917 (2015) 39F.
Buchner, K. Flechtner, Y. Bai, E. Zillner, I. Kellner, H. P. Steinrück, H. Marbach, and J. M. Gottfried, J. Phys. Chem. C 112, 15485 (2008). 40Y. Bai, F. Buchner, M. T. Wendahl, I. Kellner, A. Bayer, H. P. Steinrück, H. Marbach, and J. M. Gottfried, J. Phys. Chem. C 112, 6087 (2008). 41K. Manandhar, K. T. Park, S. Ma, and J. Hrbek, Surf. Sci. 603, 636 (2009). 42D. Écija, M. Trelka, C. Urban, P. de Mendoza, E. Mateo-Martí, C. Rogero, J. A. Martín-Gago, A. M. Echavarren, R. Otero, J. M. Gallego, and R. Miranda, J. Phys. Chem. C 112, 8988 (2008). 43J. Ahlund, J. Schnadt, K. Nilson, E. Gothelid, J. Schiessling, F. Besenbacher, N. Martensson, and C. Puglia, Surf. Sci. 601, 3661 (2007). 44X. Lu and K. W. Hipps, J. Phys. Chem. B 101, 5391 (1997). 45M. S. Liao and S. Scheiner, J. Comput. Chem. 23, 1391 (2002). 46M. S. Liao and S. Scheiner, J. Chem. Phys. 114, 9780 (2001). 47S. Yoshimoto, A. Tada, K. Suto, and K. Itaya, J. Phys. Chem. B 107, 5836 (2003). 48T. N. Krauss, E. Barrena, H. Dosch, and Y. Wakayama, ChemPhysChem 10, 2445 (2009). 49S. Yoshimoto, E. Tsutsumi, K. Suto, Y. Honda, and K. Itaya, Chem. Phys. 319, 147 (2005). 50A. Rosa and E. Baerends, J. Inorg. Chem. 33, 584 (1994). 51N. T. M. Hai, Ph.D. thesis, University of Bonn, Germany, 2007. 52T. H. Phan, S. Breuer, D. T. Pham, U. Hahn, T. Torres, and K. Wandelt, J. Phys. Chem. C 118, 457 (2014). 53S. Breuer, Ph.D. thesis, University of Bonn, Germany, 2014.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Mon, 23 Feb 2015 04:40:04
