The Roles of Precursor Conformation and Adatoms in Ullmann

21 hours ago - Surface diffusion, molecular conformation, and on-surface coupling reactions are key processes for building tailored molecular nanostru...
1 downloads 0 Views 8MB Size
Subscriber access provided by IDAHO STATE UNIV

Article

The Roles of Precursor Conformation and Adatoms in Ullmann Coupling: An Inverted Porphyrin on Cu(111) Juan Carlos Moreno-López, Duncan John Mowbray, Alejandro Perez Paz, Rodrigo Cezar de Campos Ferreira, Alisson Ceccatto dos Santos, Paola Ayala, and Abner de Siervo Chem. Mater., Just Accepted Manuscript • Publication Date (Web): 26 Mar 2019 Downloaded from http://pubs.acs.org on March 27, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

The Roles of Precursor Conformation and Adatoms in Ullmann Coupling: An Inverted Porphyrin on Cu(111) Juan Carlos Moreno-López,∗,†,‡ Duncan John Mowbray,†,¶ Alejandro Pérez Paz,†,¶ Rodrigo Cezar de Campos Ferreira,§ Alisson Ceccatto dos Santos,§ Paola Ayala,†,‡ and Abner de Siervo∗,§ † School

of Physical Sciences and Nanotechnology, Yachay Tech University, 100119 Urcuquí, Ecuador, ‡ University of Vienna, Faculty of Physics, 1090 Vienna, Austria, ¶ Nano-Bio Spectroscopy Group and ETSF Scientific Development Center, Departamento de Física de Materiales, Universidad del País Vasco UPV/EHU, E-20018 San Sebastián, Spain, § Instituto de Física “Gleb Wataghin”-Universidade Estadual de Campinas-UNICAMP, 13083-859 Campinas, Brazil ABSTRACT: Surface diffusion, molecular conformation, and on-surface coupling reactions are key processes for building tailored molecular nanostructures such as graphene nanoribbons, polycyclic aromatic hydrocarbons and 1D/2D polymers. Here, we study the surface diffusion and coupling in situ of a chlorinated porphyrin, namely 5,10,15,20-Tetrakis(4-chlorophenyl)porphyrin (Cl4 TPP), using a combined scanning tunneling microscopy (STM), density functional theory (DFT), and x-ray photoelectron spectroscopy (XPS) approach. Using STM, we obtain surface migration and rotation barriers ∆E of 0.77 ± 0.09 and 0.93 ± 0.28 eV, respectively, indicative of covalent binding to the surface. In fact, we find that the precursors as well as all the reaction species exclusively (≈100%) adopt a peculiar “inverted” conformation covalently bonded to the Cu(111). Using DFT, we have mapped two coupling reaction pathways: direct dechlorination and Cu adatom–mediated Ullmann coupling. We find the latter is essentially barrierless, whereas the former faces a barrier of about 0.9 eV for inverted Cl4 TPP on Cu(111). Our STM measurements show that C–Cu–C organometallic species are the main final products in the presence of Cu adatoms, which is explained by our DFT reaction profile when heat dissipation to the substrate is taken into account. This work not only highlights the relevance of surface adatoms in selecting the reaction pathway, but also opens the possibility of precisely tailoring 2D molecular assemblies by controlling the supply of Cu adatoms.

n-surface synthesis has become one of the most interesting approaches to building complex molecular nanostructures, 1,2 generally under ultra-high vacuum conditions. 3–5 Typically, the molecular precursors are sublimated onto a well-defined substrate and subsequently, via the stabilization of transition states by the catalyst, the chemical reaction takes place. 6–13 Even though multiple on-surface reactions are achievable, Ullmann-coupling reactions 14–17 have been shown to be among the most effective ways to build promising covalent-bonded nanostructures. For example, this approach has recently been used to synthesize 1D polymers, 11,18 polycyclic aromatic hydrocarbons, 12 2D conjugated aromatic polymers which can be exfoliated into micrometer-sized 2D sheets, 13 and to precisely synthesize graphene nanoribbons (GNRs), 19–21 among others. In general, an Ullmann-coupling reaction begins with the cleavage of the halogens on the molecular precursors, proceeds via organometallic intermediates, and terminates with the formation of C–C covalent bonds. 17 Depending on the molecular precursor, substrate, coverage, and annealing temperature and time, it may be possible to observe the precursors, organometallic intermediate species, C–C bonded species, radicals, etc. 10 In order to obtain a deeper insight into the Ullmann-coupling on-surface reaction it is necessary to characterize the precursor’s conformation and the chemical intermediates formed, as well as the most stable final product of the reaction. Herein, we have studied the molecular self-assembly of a chlorinated porphyrin, 5,10,15,20-Tetrakis(4-chlorophenyl)porphyrin

O

Figure 1. (a) Top and (b) side views of optimized gas-phase structure for 5,10,15,20-Tetrakis(4-chlorophenyl)porphyrin (Cl4 TPP). C, H, N, and Cl atoms are marked in grey, white, blue, and green, respectively. Partial outof-plane rotations of the chlorophenyl and pyrrole rings are denoted by arrows. The separations between Cl atoms are 1.34 and 1.36 nm, parallel and perpendicular to the N–H bonds, respectively.

(Cl4 TPP), depicted in Figure 1, on Cu(111). By means of a combined experimental and computational study using scanning tunneling microscopy (STM), x-ray photoelectron spectroscopy (XPS)

ACS Paragon Plus Environment 1

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and density functional theory (DFT), we are able to image and unequivocally characterize the initial, intermediate, and final species after deposition of Cl4 TPP on Cu(111). Moreover, by comparing the calculated energy barriers for a Cu adatom assisted Ullmanncoupling reaction and a direct dechlorination of the molecule, we are able to determine the preferred pathway for this reaction. These calculations, in agreement with our experimental data, show that the organometallic C–Cu–C bond is the expected final product of the reaction, when heat dissipation to the Cu substrate is taken into account. 1. RESULTS AND DISCUSSION

Our STM measurements of Cl4 TPP deposited on Cu(111) show individual molecules with a “rectangular” symmetry characterized by four bright protrusions at the corners of the molecule and two bright central ovals (Figure 3a). In order to precisely measure individual molecules, we have performed a statistical analysis of several STM images, including atomically resolved images where porphyrin molecules are imaged simultaneously with Cu surface atoms (Figure S2 in SI). From this analysis, the dimensions of a single molecule are determined to be a = 1.47 ± 0.05 nm and b = 1.11 ± 0.05 nm with a separation between the two bright central ovals of d = 0.46 ± 0.05 nm (Figure 3a). This indicates surface adsorption on Cu(111) elongates the porphyrin, reducing the symmetry from square to rectangular.

In gas phase, the Cl4 TPP precursor adopts a square non-planar shape, with the pyrrole and chlorophenyl rings rotated around the C–C bonds, as indicated in Figure 1. After Cl4 TPP deposition on Cu(111), the adsorbed isolated species are mobile on the surface, while linear chains are less mobile, at room temperature (Figure S1 in Supporting Information (SI)). By cooling the samples down to 233 K, we drastically reduce the mobility and are able to precisely resolve individual molecules. This already indicates there is a high activation barrier for surface migration.

Figure 3. STM images of an individual species of Cl4 TPP as deposited on Cu(111) a) high resolution measurement (T = 233 K, A = 2.0 × 1.5 nm2 , U = −1.5 V, I = 130 pA), and simulations in b) “inverted” and c) “saddle” configurations. d) STM simulation and schematic e) top and f) side views of Cl4 TPP on Cu(111) with Cu adatoms below diagonally opposite Cl atoms. C, H, N, Cl, and Cu atoms marked in grey, white, blue, green, and brown, respectively. Arrows in e) and f) indicate the h110i, h111i and h112i crystallographic directions of Cu(111).

Figure 2. Arrhenius plot for the surface migration (filled squares) and rotation (open circles) of Cl4 TPP on Cu(111) with activation barriers ∆E of 0.77 ± 0.09 and 0.93 ± 0.28 eV and attempt frequencies ν of 1013±2 and 1015±5 s−1 , respectively.

In Figure 2 we analyze the surface diffusion behavior of Cl4 TPP on Cu(111) via an Arrhenius plot of the temperature-dependent rates. Similarly to other tetraphenylporphyrins, 22 Cl4 TPP moves only along the three main h110i directions and exhibits discrete rotations of ±60◦ . The data exhibits a clear linear dependence, consistent with the Arrhenius equation for surface diffusion,  ν exp − k∆E , 22–25 where ν is the attempt frequency and ∆E is the BT activation barrier. From the slope, we obtain the migration barrier for unidirectional diffusion of ∆E = 0.77 ± 0.09 eV, and from the intercept an attempt frequency of ν = 1013±2 s−1 . These values are comparable to ∆E = 0.71 ± 0.08 eV and ν = 1010.9±1.4 s−1 for 2H-tetraphenylporphyrin (2HTPP). 22 The slightly higher migration barrier suggests a stronger molecule–substrate interaction for Cl4 TPP compared to 2HTPP. Similarly, from the linear fit to the surface rotation data we extract the rotation barrier ∆E = 0.93 ± 0.28 eV and the corresponding attempt frequency ν = 1015±5 s−1 . The higher barrier for rotation compared to migration is again suggestive of a covalent interaction between Cl4 TPP and Cu(111).

Since STM measurements only provide geometrical information indirectly via the local electronic density of states, often their interpretation is not straightforward. In order to obtain deeper insight into the molecular conformation, we have performed structural relaxations of an individual Cl4 TPP molecule on Cu(111) for the two main conformations of surface adsorbed porphyrins 26–28 (Figures 3b and c). Figure 3b shows our relaxed DFT structure (wire-frame model) and the resulting STM simulation. The overall structure of the molecule is almost parallel to the surface, except for the two opposing pyrrole rings (NC4 ). These pyrrole rings undergo a drastic rotation of ∼100◦ with respect to the surface plane, with their iminic (=N–) atoms pointing towards the Cu(111) surface. The simulated STM image (Figure 3b) shows four bright protrusions at the corners and two bright ovals in the middle of the molecule, clearly reproducing the main features observed by STM. The calculated (measured) separations between the external Cl atoms are a = 1.56 (1.47 ± 0.05) nm and b = 1.07 (1.11 ± 0.05) nm and between the rotated pyrrole groups are d = 0.52 (0.46 ± 0.05) nm, showing a very good agreement between theory and experiment. In this distorted configuration, the distance between the two rotated pyrrole rings is too large to have significant face-to-face π–π interactions. 29–31 Therefore, this drastic rotation of the pyrrole rings must be stabilized mainly by the interaction between the two iminic (=N–) atoms and the Cu surface. Indeed, our DFT relaxed structures show that covalent N–Cu bonds are formed in a bridge position with a length

ACS Paragon Plus Environment 2

Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials of about 2.3 Å. This is consistent with our high measured activation barrier for surface migration and rotation arising from chemisorption rather than physisorption (Figure 2). Moreover, the separation between the rotated pyrrole rings (d = 0.52 nm) is commensurate with Cu(111), i.e., the next nearest neighbour separation of 0.51 nm along the h110i directions. This differs from the other coinage metals, where the next nearest neighbour separation of 0.58 nm would require further rotation leading to steric hindrance between the rotated pyrrole rings. Lepper et al. recently reported this extremely distorted configuration for CN and H terminated porphyrins, 26–28 as seen previously by Albrecht et al. 32 This configuration was named the “inverted” structure by Lepper et al., and from hereon we will use their nomenclature. In previous works 26–28 the most commonly reported conformation of surface adsorbed porphyrins was the so-called “saddle” structure. 33–35 In order to check this possibility, in Figure 3c we show our relaxed DFT structure (wire-frame model) and the resulting STM simulation for Cl4 TPP adsorbed on Cu(111) in the saddle conformation. Comparing Figures 3a, b, and c, we find the STM simulation for inverted is in better agreement with the measurement than saddle. In particular, Figure 3a does not feature the central gap, two opposing “banana” structures, and more squarish shape of Figure 3c, which are, however, in excellent agreement with the saddle configuration rarely observed here (Figure S3 in SI) but frequently reported for similar molecules. 33,34 Moreover, our DFT calculations show that Cl4 TPP is physisorbed on Cu(111) when in the saddle conformation, with a N· · · Cu separation of 3.0 Å. This physisorption is inconsistent with the high activation barriers we observe for surface migration and rotation (Figure 2). Moreover, according to the Sabatier principle of heterogeneous catalysis, the binding of the reagents must be neither too strong nor too weak, but “just right”. 36 On the one hand, this implies that when Cl4 TPP is physisorbed in the rarely observed saddle conformation it is probably too weakly bound to have significant reactivity at low coverage. On the other hand, it also implies that when Cl4 TPP is covalently bound in the inverted conformation it is both mobile enough for species to meet, and sufficiently bound for species to react, as seen in our STM measurements. This underscores the important role played by the adsorption conformation in determining a precursor’s reactivity. Finally, a careful observation of Figure 3a finds a brighter STM contrast in the diagonally opposite Cl atoms. To explain this observation, we have performed DFT calculations for Cl4 TPP on Cu(111) with Cu adatoms underneath the Cl atoms which exhibit brighter contrast. The resulting STM simulation is shown in Figure 3d, with top and side views of the structure shown in Figures 3e and f, respectively. Moreover, the presence of the two anchoring Cu adatoms explains why this particular molecule was stationary during 20 consecutive STM measurements. Note that a variety of contrasts were observed, e.g., one, two, three, or four bright spots, consistent with a corresponding number of Cu adatoms underneath the Cl terminations. However, our STM measurements of Cl4 TPP deposited on Cu(111) most commonly exhibited a uniform contrast between the Cl terminating groups, resembling Figure 3b. Overviews of the samples show that Cl4 TPP molecules are oriented in three equivalent crystallographic directions with their long axis oriented along the h110i main directions of the substrate (Figure 4a). To obtain a deeper insight into the molecule–molecule interactions, we have analyzed the main bonding motifs observed between neighboring molecules. On one hand, some molecules are bonded via single “side-on” interactions (Figure 4b). In this configuration the Cl atoms of neighboring molecules are placed beside each other. This configuration is consistent with a non-covalent type I van der Waals (vdW)

Figure 4. a) STM measurement after deposition of Cl4 TPP on Cu(111) (T = 300 K, A = 7.0 × 7.0 nm2 , U = −1.5 V, I = 130 pA). Van der Waals (dotted oval) and covalent (dashed oval) single “head-on” interactions and a single “side-on” (dashed rectangle) interaction are marked. b) STM simulation of the latter (dashed rectangle in a)), a type I van der Waals interaction (θ1 ≈ θ2 ), as shown in c). The white arrows shown in a) indicate the h110i crystallographic directions of Cu(111).

halogen· · · halogen interaction, where the C-Cl· · · Cl angles are approximately equal (θ1 ≈ θ2 ), as depicted in Figure 4c. 37,38 For our relaxed DFT structure we obtain θ1 = 97◦ ≈ θ2 = 107◦ and an intermolecular Cl· · · Cl distance of 3.54 Å, consistent with the sum of their vdW radii of 3.50 Å. 39 Moreover, our STM simulation 40 (Figure 4b) is in excellent agreement with our STM measurement (dashed rectangle of Figure 4a), both in the dimensions and in the overall appearance. On the other hand, some molecules are bonded via single “headon” interactions (dotted and dashed ovals). In these configurations the “arms” of neighboring molecules are collinear, that is, directly facing each other. From the STM measurements, two different kinds of single “head-on” interactions can be clearly distinguished, with one (dotted oval) 0.54 nm longer than the other (dashed-oval). The longer single “head-on” interaction (dotted oval) is consistent with a non-covalent type I vdW Cl· · · Cl interaction (θ1 ≈ θ2 ≈ 180◦ ), whereas the shorter single “head-on” interaction is covalent in nature. This result suggests that after deposition at room temperature some molecules undergo cleavage of the C–Cl bonds as part of the on-surface Ullmann-coupling reaction. However, the relative ratio between “head-on” and “side-on” interactions is time and temperature dependent. As mentioned, after deposition of Cl4 TPP on Cu(111), the molecules are quite mobile at room temperature and some changes occur in the samples as a function of time. A “freshly” 41 deposited sample typically has many isolated molecules, “side-on” interactions, and relatively few “head-on” interactions (∼ 25%). However, after 24 hours at room temperature, isolated molecules are extremely difficult to observe and almost all (∼ 99%) of the molecules are involved in “head-on” interactions, with a majority being covalent in nature. These statistics are taken from observations of approximately 300 molecules in representative images of the entire surface (see Figure S4 in the SI). Taking into account that the growth of molecular structures on surfaces is intrinsically a non-equilibrium process governed by a competition between kinetics and thermodynamics, 42 our results clearly suggest that some molecules do not react after deposition

ACS Paragon Plus Environment 3

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

due to the kinetics, i.e., the activation barrier that they need to overcome or a limiting reagent for the process (e.g. Cu adatoms).

Page 4 of 9

a)

b)

Figure 5. a) STM measurement after annealing of Cl4 TPP deposited on Cu(111) (T = 350 K for t = 30 min, A = 30.0 × 30.0 nm2 , U = −2.0 V, I = 200 pA). A double “head-on” configuration is marked (dashed rectangle). Direct comparison of STM b) measurement (A = 3.0 × 3.5 nm2 ) and c) simulation of the double “head-on” configuration (dashed rectangle in a)), with a distance between pyrrole rings of L = 1.51 and 1.52 nm, respectively.

Figure 5 shows a STM measurement of Cl4 TPP molecules after annealing at 350 K for 30 min. At this stage, the molecules are mainly involved in “head-on” interactions forming various 1D chains. A double “head-on” configuration is showed in Figure 5b, where a small gap in the charge density between the molecules is clearly observed. For this arrangement, the distance between the closest pyrrole rings of neighboring molecules is L = 1.51 ± 0.05 nm, following the procedure described in Figure S5 of the SI. Figure 5c shows our STM simulation 40,43 where the distance between the neighboring pyrrole rings (L = 1.52 nm), the small gap in the charge density, and the overall structure of the molecules is consistent with our experimental results. From our relaxed DFT structure (Figure 5c) we obtain intermolecular Cl· · · Cl separations of ≈ 4.6 Å, which is much larger than the sum of their vdW radii of 3.50 Å 39 and a negligible binding energy between species. These results are in agreement with the observed instability of this structure in consecutive STM images acquired at 274 K (Video 1 in SI). For these reasons, we attribute the double “head-on” structure to confinement effects. At this point, it is worth highlighting the different behavior we have observed in the cleavage of the Cl atoms on the molecules. As we mentioned before, remarkably, even at room temperature, some molecules indicate cleavage of Cl atoms, whereas after annealing at 400 K for 30 min other molecules are still intact. This result might suggest that the rate-limiting step for cleavage of the Cl atoms is a limiting reactant (e.g. Cu adatoms) rather than a kinetic activation barrier. To gain additional insight into the chemical environment of the Cl atoms as a function of the annealing temperature, we have performed XPS measurements. Figure 6 shows a summary of the dependence on annealing temperature of the XPS data for Cl atoms. XPS measurements of a “freshly” 41 deposited sample of Cl4 TPP on Cu(111) shows a main set of two peaks in the Cl region, which represent ∼61% of the total integrated area (blue fitting, Cl(1)). These peaks have a binding energy of 200.2(5) eV and 201.9(5) eV for the Cl 2p3/2 and 2p1/2 peaks, in excellent agreement with the reference values for Cl atoms in chlorobenzene (C6 H5 Cl). 47 This allows us to assign the Cl(1) peaks to Cl atoms bonded to the molecules. As we increase the annealing temperature of the sam-

c)

Figure 6. Cl 2p XPS spectra (open symbols) after deposition of Cl4 TPP on Cu(111) a) as grown at T = 300 K (circles), and after performing postdeposition annealing at b) T = 400 K (squares) and c) T = 450 K (triangles) for t = 30 min. Experimental fits (red lines) and individual components from Cl in porphyrin (blue lines) and as adatoms (orange lines) on Cu(111), with the fitted Cl 2p1/2 binding energies (dashed lines) at 200.2(5) eV and 198.2(5) eV, respectively, are shown. The area ratios of Cl in porphyrin to Cl adatoms obtained from the deconvolutions are a) 61:39, b) 17:83, and c) 11:89.

ples, the Cl(1) peaks start to decrease while the Cl(2) peaks (orange peaks in Figure 6) increase in magnitude. The binding energies of the Cl(2) peaks are 198.2(5) eV and 199.7(5) eV for the 2p3/2 and 2p1/2 components, in quantitative agreement with the reference values for CuCl2 . 47 Based on this, we can attribute the Cl(2) peaks to Cl atoms that are cleaved from the molecule and are strongly interacting with the Cu substrate. XPS shows that, as the samples are annealed, Cl atoms are gradually cleaved from the molecules. The presence of Cl(2) peaks at room temperature confirms the cleavage of Cl atoms observed by STM (Figure 4a). Moreover, the nearly complete dechlorination observed after annealing at 450 K is also confirmed by our STM measurements, which show molecules almost completely dechlorinated but preserving their organometallic (C-Cu-C) nature (Figure S6 in SI). In order to determine if after the Cl cleavage the molecules are stabilized by C–C covalent bonds or by C–Cu–C organometallic bonds, 48 we performed DFT simulations of these structures. Figures 7b and c show the DFT relaxed structure (wire-frame model) for C–Cu–C organometallic bonds and C–C covalent bonds, respectively. While the STM simulation of C–Cu–C organometallic bonds shows a bright protrusion in the bond between the molecules and a distance between the two pyrrole rings of neighbor molecules of L = 1.54 nm (Figure 7b), the STM simulation of C–C co-

ACS Paragon Plus Environment 4

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Figure 7. STM images of a covalent single “head-on” interaction for Cl4 TPP deposited on Cu(111) a) measured (T = 400 K for t = 30 min) and b,c) simulated for b) a C–Cu–C organometallic interaction and c) a C–C covalent bond between two Cl3 TPP molecules.

valent bonds is characterized by a uniform STM contrast in the bond with a distance between the two neighboring pyrrole rings of L = 1.33 nm (Figure 7c). The better agreement between Figures 7a and b allows us to assign the covalent single “head-on” interaction to a C–Cu–C organometallic bond. For a comparison between the C–Cu–C organometallic bond and the C–Cl· · · C configuration see Figure S7 in SI. For a 3D high-resolution image of a structure containing a C–Cu–C organometallic bond see Figure S8. In a final effort to shed light on the kinetics and thermodynamics of the processes, we performed DFT energy calculations for the relevant chemical species following two different reaction pathways: direct dechlorination (red lines) and Cu adatom-mediated Ullmann coupling (blue lines), as shown in Figure 8. Here, we reference all energies to that of Cl4 TPP adsorbed on Cu(111) in the inverted conformation. The relative stability of saddle versus inverted conformations of Cl4 TPP (Figure 8(i,ii)) is highly sensitive to the choice of exchange and correlation (xc) functional. When vdW interactions are neglected (PBE 44 solid lines in Figure 8), inverted and saddle conformations are isoenergetic, whereas, if vdW interactions are included at the Grimme’s D3 level (PBE-D3 45 dashed lines in Figure 8), the inverted conformation is 0.6 eV more stable than saddle, as expected. If instead vdW interactions are included at the selfconsistent level (vdW-DF2 46 dotted lines of Figure 8), we find the opposite behavior. 49 This suggests PBE-D3 is the more robust xc functional for describing porphyrins on metal surfaces. 26,27 Nonetheless, the remainder of the pathway, referenced to the experimentally observed inverted conformation of Cl4 TPP, is in semiquantitative agreement amongst all three xc functionals. This suggests our vdW-DF2 calculations benefit from an error cancellation between the various inverted structures in the pathway, and we may have confidence in all these results. From hereon, we shall refer

specifically to PBE-D3 energy values. In order to form a stable C–C covalent bond between molecules via the direct dechlorination path (Figure 8(a)), two isolated molecules (2Cl4 TPP) first find a slightly more stable configuration interacting via a single “side-on” Cl· · · Cl interaction (Cl4 TPP)2 (Figure 8(a)(iii)). Then, after overcoming an energy barrier of 0.9 eV, a Cl atom is cleaved from one molecule to form the complex Cl4 TPP· · · Cl3 TPP (Figure 8(a)(iv)). 50 Finally, after overcoming a similar barrier, the second Cl atom is also cleaved from the other molecule and the final C–C covalent bond is formed between molecules (Cl3 TPP)2 with a binding energy of 2.8 eV (Figure 8(a)(v)). This C–C covalent bond is the energetically most favorable structure in a direct dechlorination process. In contrast, the reaction pathway to form a stable C–C covalent bond via a Cu adatom (Cuad ) mediated Ullmann-coupling reaction is facile with no overall barrier (Figure 8(b)). In this case, one of the two isolated molecules first interacts with a Cu adatom (Figure 8(b)(iii)), spontaneously forming a dimer complex with a partially dissociated C–Cl bond (CuCl4 TPP+Cl4 TPP) and a binding energy of 0.2 eV. Both molecules are then dechlorinated with no overall activation barrier due to the stabilization brought about by the Cuad underneath the dissociating C–Cl bond. This results in the formation of a C–Cu–C organometallic bond with a binding energy of 2.9 eV (Cl3 TPP-Cu–Cl3 TPP) (Figure 8(b)(iv)). These theoretical reaction pathways allow us to explain why some Cl4 TPP molecules on Cu(111) at room temperature show cleavage of Cl atoms while after annealing at 400 K other molecules are still intact. In fact, even though the C–Cu–C organometallic bond is thermodynamically more stable than the adsorbed precursor molecules (either isolated or interacting via “side-on” Cl· · · Cl interactions) and the Ullmann-coupling reaction is barrierless, not all molecules reach this equilibrium state due to the limited availability of Cuad on the surface, i.e., Cuad acts as a limiting reagent. In a direct dechlorination process, the most stable species is by far the C–C covalent bonded (Cl3 TPP)2 , while in a Cuad mediated Ullmann-coupling reaction the C–Cu–C organometallic bond is isoenergetic with the C–C covalent bond between two molecules. To understand why the C–Cu–C organometallic species (Figure 8(b)(iv)) is strongly preferred over the final C–C product (Figure 8(b)(v)) under our experimental conditions, we must consider the heat dissipation to the Cu(111) surface, which is an excellent thermal conductor. The organometallic C–Cu–C intermediate features strong coupling to the Cu surface via its C–Cuad and N–Cu bonds. In this way, before the C–Cu–C organometallic intermediates can be demetallated, a heat transfer to the substrate occurs that is sufficient to kinetically trap the organometallic species. 51 Without invoking heat dissipation, one would expect a nearly 1:1 ratio of C–Cu–C to C–C species, something which is not observed in our measurements. Finally, it is important to highlight that the behavior of Cl4 TPP on Cu(111) is a quite rich process that cannot be fully understood without considering the two reaction pathways described herein. Since the Cuad mediated Ullmann-coupling reaction is the facile and thermodynamically most stable process, the direct dechlorination process will only take place when no Cu adatoms are available. However, heating the Cu(111) surface naturally produces more Cu adatoms, so that the Cuad mediated Ullmann process remains facile even under conditions where the barrier for direct dechlorination could be overcome. For this reason we do not expect the direct dechlorination pathway to play a major role under ordinary conditions.

ACS Paragon Plus Environment 5

Chemistry of Materials

Energy [kcal/mol]

40

(ii)

saddle

inverted

(iii)

(v)

(iv)

side-on

head-on

Clad 2Cl4TPP

2Cl4TPP

2Clad

Cl4TPP Cl3TPP + Clad (Cl3TPP)2 + 2Clad

(Cl4TPP)2

20

1

0

0

-20 -60 -80

(b)

-1

PBE PBE-D3 vdW-D2

-40

Energy [eV]

Direct Dechlorination

(a)

(i)

-2 -3

2Cl4TPP + Cuad saddle

2Cl4TPP + Cuad

inverted

CuCl4TPP + Cl4TPP

Cuad

Cl3TPP−Cu−Cl3TPP + 2Clad head-on

Ullmann Coupling

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 9

Cu(Cl3TPP)2 + 2Clad

head-on

Cu 2Clad

Cu

Figure 8. STM simulations and DFT reaction profile for the (a) direct dechlorination reaction (red lines) and (b) Cu adatom-mediated Ullmann coupling reaction (blue lines). For each process we plot the energy of the (i,ii) precursors, (iii, iv) intermediates, (v) and final species (gray circles) and two transition states (white circles) using PBE 44 (solid lines), PBE-D3 45 (dashed lines) and vdW-DF2 46 (dotted lines) xc functionals. Structural schematics are provided in Figure S9.

2. CONCLUSIONS

We have studied the surface migration, conformation, and Ullmann-coupling reaction profile for 5,10,15,20-Tetrakis(4chlorophenyl)porphyrin deposited on Cu(111). We obtain significant surface migration and rotation barriers of 0.77 ± 0.09 and 0.93 ± 0.28 eV, respectively, indicative of covalent binding to the surface. Surprisingly, we find all adsorbates exclusively (≈100%) adopt a highly distorted molecular configuration where the pyrrole rings of the molecules are rotated ∼100◦ , the so-called “inverted” structure. Moreover, based on this conformation, we are able to unequivocally identify the bonding motifs: vdW single “head-on” and “side-on” Cl· · · Cl interactions, and Cu-Cl species; and the product: the organometallic C–Cu–C species, by our combined STMDFT study. Finally, by simulating two different chemical reaction pathways, we have shown that a Cu adatom-mediated Ullmanncoupling reaction is essentially barrierless, and the organometallic C–Cu–C species is the final product of the reaction under mild conditions, when heat dissipation to the Cu(111) surface is taken into account. To summarize, we report (1) the first system to have chlorinated TPP molecules as building blocks for on-surface synthesis; (2) the first system to adsorb almost 100% in the inverted structure conformation; (3) the first system to remain exclusively in the inverted conformation throughout the Ullmann reaction pathway, (4) the first elucidation of the Ullmann-coupling reaction profile

for inverted porphyrins, and (5) the first measurements of surface mobility of an inverted porphyrin. Altogether, these results blaze the trail towards the production of precisely tailored 2D molecular assemblies 48 via the controlled application of Cu adatoms to the molecular precursors. 3. METHODS Experimental Setup. The experiments were performed in two

connected chambers. One chamber was equipped with a STM and the other one with an XPS, low energy electron diffraction (LEED), home-made Knudsen cell for molecule sublimation, and standard cleaning facilities. The base pressure in the XPS chamber was in the low 10−10 mbar range and in the STM in the middle 10−11 mbar range. The STM microscope used was a SPECS Aarhus 150 equipped with a SPECS SPC 260 Controller. The STM measurements were performed in constant current mode with a tungsten tip cleaned in situ by Ar+ sputtering. XPS was performed with a SPECS Phoibos 150 hemispherical analyzer with multi channeltron detection. The photons employed in XPS were obtained from an Al-Kα anode. The Cu(111) single crystal was prepared by several cycles of Ar+ sputtering (1000 V @ ∼5 µAcm−2 ) for 30 minutes with subsequent annealing at 850 K for 20 minutes following a slow cooling ramp to ensure large and well-ordered terraces. Cl4 TPP molecules (5,10,15,20-Tetrakis(4-chlorophenyl)porphyrin) were sublimed in situ using a home-made Knudsen cell from a

ACS Paragon Plus Environment 6

Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials quartz crucible at a temperature of about 570 K while the sample was kept at room temperature. Coverage was determined using STM images. STM bias voltages are with respect to the sample. The images were analyzed using the WSXM software. 52 Cl4 TPP molecules were supply by Porphyrin Chemical & Engineering. 53 We obtained the molecular displacements from sequences of consecutive STM images for a given temperature. This yielded the mean square displacement h∆x2 i and in turn the migration rate of h∆x2 i/(λ2 τ), where λ is the jump length, i.e., the lattice parameter of 2.55 Å along the h110i direction, and τ is the corresponding time interval. Using the temperature-dependent rates we determined the   surface mobility using the Arrhenius formula Rate = ν exp − k∆E , T B where ν is the attempt frequency and ∆E is the activation barrier. The measurements and analysis were performed as described previously. 22,23,25 We adhered to the following experimental protocol to ensure proper measurements: 23 (1) STM measurements were performed in the tunneling resistance range of 1 to 10 GΩ or above (U ≈ 1.5 V, I ≈ 150 pA) to avoid tip-induced effects; 22 (2) a low coverage was maintained on the surface to avoid molecule– molecule interactions; (3) a large number of events were accumulated to validate our statistical analysis (∼ 2000); and (4) measurements were excluded when molecules were already trapped, e.g., at step-edges. Our measurements followed ∼ 20 molecules in 900 nm2 images in intervals of τ ≈ 400 s per sequence. Several sequences of STM images in different and representative areas were used for each temperature, where each sequence contains around 30 frames and in total, more than 2000 independent events for each temperature. Computational Details. All DFT calculations were performed using linear combinations of atomic orbitals (LCAOs) 54 within the projector-augmented wave method (PAW) 55 code gpaw. 43 For the adsorbed species we employed a double-zeta-polarized (DZP) basis set and a single-zeta-polarized (SZP) basis set for the Cu(111) surface. We employed three different types of xc functionals: the generalized gradient approximation (GGA) for the exchange and correlation (xc) functional as implemented by Perdew, Burke, and Ernzerhof (PBE), 44 including vdW interactions at the Grimme’s D3 level (PBE-D3), 45 and including vdW interactions self-consistently (vdW-DF2). 46 We used a grid spacing of h ≈ 0.2 Å, and performed structural relaxation of the relevant species until a maximum force . 0.03 eV/Å was obtained. To model the Cu(111) surface we have employed a 15 × 14 × 3 supercell of 3.823 × 3.095 × 2.417 nm3 , frozen to the experimental coordinates (a = 0.361 nm), with more than 1.5 nm of vacuum between repeated images. STM simulations have employed the Tersoff-Hamann approximation 56 in constantcurrent mode with a bias of −1.5 V relative to the Fermi level as implemented in the code ase. 57 To estimate the C–Cl and C–C bond distance in the transition states during direct dechlorination and a Cu adatom mediated Ullmann-coupling reaction, we have performed nudged elastic band calculations for a chlorinated benzene with the anchoring C atom constrained to the same height and/or position as the relevant atom in Cl4 TPP above the Cu(111) surface. The transition state energy for Cl4 TPP was then estimated by fixing either the C–Cl or C–C bond distance and performing an otherwise unconstrained surface relaxation on the frozen Cu(111) surface slab. ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.xxxxxxxx.

Additional STM simulations, measurements and videos; atomic coordinates and side views of all structures.

AUTHOR INFORMATION Corresponding Author

*E-mail (J. C. Moreno-López): [email protected]. *E-mail (A. de Siervo): [email protected]. Author Contributions J.C.M.-L., R.C.C.F., A.C.S and A. S. performed the STM and XPS measurements at Instituto de Fisica Gleb Wataghin in Campinas, Brazil. D.J.M and A.P.P. performed the DFT calculations. J.C.M.-L, A.S., R.C.C.F, D.J.M. and A.P.P. analyzed the data. J.C.M.-L., A.S., D.J.M., P.A. and A.P.P co-wrote the paper. A.S. and J.C.M.-L. conceived the study. All authors have read, and given approval to the final version of the manuscript. Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS

Funding by Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP) projects No. 2017/08846-7 and 2007/54829-5, CAPES and CNP from Brazil are gratefully acknowledged. This work used the Quinde I supercomputer of Public Company Yachay E. P., which was implemented under contract No. 0051-2015, corresponding to Component No. 7 of Group No. 2, Re-YACHAY018-2015, and the Imbabura cluster of Yachay Tech University, which was purchased under contract No. 2017-024 (SIE-UITEY007-2017). J.C.M.-L. gratefully acknowledges Yachay Tech University and its authorities for their support in the performance of the experiments at Instituto de Fisica Gleb Wataghin in Campinas, Brazil. REFERENCES (1) Liang, B.; Wang, H.; Shi, X.; Shen, B.; He, X.; Ghazi, Z. A.; Khan, N. A.; Sin, H.; Khattak, A. M.; Li, L.; Tang, Z. Microporous membranes comprising conjugated polymers with rigid backbones enable ultrafast organicsolvent nanofiltration. Nature Chemistry 2018, 10, 961–967. (2) Treier, M.; Pignedoli, C. A.; Laino, T.; Rieger, R.; Müllen, K.; Passerone, D.; Fasel, R. Surface-assisted cyclodehydrogenation provides a synthetic route towards easily processable and chemically tailored nanographenes. Nature Chemistry 2010, 3, 61. (3) Zhang, Y.-Q.; Paszkiewicz, M.; Du, P.; Zhang, L.; Lin, T.; Chen, Z.; Klyatskaya, S.; Ruben, M.; Seitsonen, A. P.; Barth, J. V.; Klappenberger, F. Complex supramolecular interfacial tessellation through convergent multistep reaction of a dissymmetric simple organic precursor. Nature Chemistry 2018, 10, 296. (4) Bartels, L. Tailoring molecular layers at metal surfaces. Nature Chemistry 2010, 2, 87. (5) Wang, X.-Y.; Urgel, J. I.; Barin, G. B.; Eimre, K.; Di Giovannantonio, M.; Milani, A.; Tommasini, M.; Pignedoli, C. A.; Ruffieux, P.; Feng, X.; Fasel, R.; Müllen, K.; Narita, A. Bottom-Up Synthesis of HeteroatomDoped Chiral Graphene Nanoribbons. Journal of the American Chemical Society 2018, 140, 9104–9107. (6) Xiong, Z.; Chenguang, W.; Yajie, Z.; Fang, C.; Yang, H.; Qian, S.; Jian, S.; Xiang, S.; Wei, J.; Wei, C.; Guoqin, X.; Kai, W. Steering Surface Reaction Dynamics with a Self-Assembly Strategy: Ullmann Coupling on Metal Surfaces. Angewandte Chemie International Edition 56, 12852–12856. (7) Shi, K. J.; Zhang, X.; Shu, C. H.; Li, D. Y.; Wu, X. Y.; Liu, P. N. Ullmann coupling reaction of aryl chlorides on Au(111) using dosed Cu as a catalyst and the programmed growth of 2D covalent organic frameworks. Chem. Commun. 2016, 52, 8726–8729. (8) Lafferentz, L.; Eberhardt, V.; Dri, C.; Africh, C.; Comelli, G.; Esch, F.; Hecht, S.; Grill, L. Controlling on-surface polymerization by hierarchical and substrate-directed growth. Nature Chemistry 2012, 4, 215–250. (9) Weigelt, S.; Busse, C.; Bombis, C.; Knudsen, M.; Gothelf, K.; Lægsgaard, E.; Besenbacher, F.; Linderoth, T. Surface Synthesis of 2D Branched Polymer Nanostructures. Angewandte Chemie International Edition 2007, 47, 4406–4410. (10) Zint, S.; Ebeling, D.; Schlöder, T.; Ahles, S.; Mollenhauer, D.; Wegner, H. A.; Schirmeisen, A. Imaging Successive Intermediate States of the On-Surface Ullmann Reaction on Cu(111): Role of the Metal Coordination. ACS Nano 2017, 11, 4183–4190. (11) Pham, T. A.; Tran, B. V.; Nguyen, M.-T.; Stöhr, M. Chiral-Selective Formation of 1D Polymers Based on Ullmann-Type Coupling: The Role of the Metallic Substrate. Small 2016, 13, 1603675. (12) de Oteyza, D. G.; Gorman, P.; Chen, Y.-C.; Wickenburg, S.; Riss, A.; Mowbray, D. J.; Etkin, G.; Pedramrazi, Z.; Tsai, H.-Z.; Rubio, A.; Crommie, M. F.; Fischer, F. R. Direct Imaging of Covalent Bond Structure in Single-Molecule Chemical Reactions. Science 2013, 340, 1434–1437.

ACS Paragon Plus Environment 7

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(13) Liu, W.; Luo, X.; Bao, Y.; Liu, Y. P.; Ning, G.-H.; Abdelwahab, I.; Li, L.; Nai, C. T.; Hu, Z. G.; Zhao, D.; Liu, B.; Quek, S. Y.; Loh, K. P. A twodimensional conjugated aromatic polymer via C–C coupling reaction. Nature Chemistry 2017, 9, 563–570. (14) F., U.; Jean, B. Ueber Synthesen in der Biphenylreihe. Berichte der deutschen chemischen Gesellschaft 1901, 34, 2174–2185. (15) Xi, M.; Bent, B. E. Mechanisms of the Ullmann coupling reaction in adsorbed monolayers. Journal of the American Chemical Society 1993, 115, 7426–7433. (16) Björk, J.; Hanke, F.; Stafström, S. Mechanisms of Halogen-Based Covalent Self-Assembly on Metal Surfaces. Journal of the American Chemical Society 2013, 135, 5768–5775. (17) Sambiagio, C.; Marsden, S. P.; Blacker, A. J.; McGowan, P. C. Copper catalysed Ullmann type chemistry: from mechanistic aspects to modern development. Chem. Soc. Rev. 2014, 43, 3525–3550. (18) Di Giovannantonio, M.; Tomellini, M.; Lipton-Duffin, J.; Galeotti, G.; Ebrahimi, M.; Cossaro, A.; Verdini, A.; Kharche, N.; Meunier, V.; Vasseur, G.; Fagot-Revurat, Y.; Perepichka, D. F.; Rosei, F.; Contini, G. Mechanistic Picture and Kinetic Analysis of Surface-Confined Ullmann Polymerization. Journal of the American Chemical Society 2016, 138, 16696–16702. (19) Ruffieux, P.; Wang, S.; Yang, B.; Sánchez-Sánchez, C.; Liu, J.; Dienel, T.; Talirz, L.; Shinde, P.; Pignedoli, C. A.; Passerone, D.; Dumslaff, T.; Feng, X.; Müllen, K.; Fasel, R. On-surface synthesis of graphene nanoribbons with zigzag edge topology. Nature 2016, 531, 485–492. (20) van der Lit, J.; Boneschanscher, M. P.; Vanmaekelbergh, D.; Ijäs, M.; Uppstu, A.; Ervasti, M.; Harju, A.; Liljeroth, P.; Swart, I. Suppression of electron–vibron coupling in graphene nanoribbons contacted via a single atom. Nature Communications 2013, 4, 2023. (21) Jacobse, P. H.; Kimouche, A.; Gebraad, T.; Ervasti, M. M.; Thijssen, J. M.; Liljeroth, P.; Swart, I. Electronic components embedded in a single graphene nanoribbon. Nature Communications 2017, 8, 119. (22) Buchner, F.; Xiao, J.; Zillner, E.; Chen, M.; Röckert, M.; Ditze, S.; Stark, M.; Steinrück, H.-P.; Gottfried, J. M.; Marbach, H. Diffusion, Rotation, and Surface Chemical Bond of Individual 2H-Tetraphenylporphyrin Molecules on Cu(111). The Journal of Physical Chemistry C 2011, 115, 24172–24177. (23) Eichberger, M.; Marschall, M.; Reichert, J.; Weber-Bargioni, A.; Auwärter, W.; Wang, R. L. C.; Kreuzer, H. J.; Pennec, Y.; Schiffrin, A.; Barth, J. V. Dimerization Boosts One-Dimensional Mobility of Conformationally Adapted Porphyrins on a Hexagonal Surface Atomic Lattice. Nano Letters 2008, 8, 4608–4613. (24) Barth, J. V. Transport of adsorbates at metal surfaces: from thermal migration to hot precursors. Surf. Sci. Rep. 2000, 40, 75 – 149. (25) Marbach, H.; Steinrück, H.-P. Studying the dynamic behaviour of porphyrins as prototype functional molecules by scanning tunnelling microscopy close to room temperature. Chem. Commun. 2014, 50, 9034– 9048. (26) Lepper, M.; Köbl, J.; Schmitt, T.; Gurrath, M.; de Siervo, A.; Schneider, M. A.; Steinrück, H.-P.; Meyer, B.; Marbach, H.; Hieringer, W. “Inverted” porphyrins: a distorted adsorption geometry of free-base porphyrins on Cu(111). Chem. Commun. 2017, 53, 8207–8210. (27) Lepper, M.; Schmitt, T.; Gurrath, M.; Raschmann, M.; Zhang, L.; Stark, M.; Hölzel, H.; Jux, N.; Meyer, B.; Schneider, M. A.; Steinrück, H.P.; Marbach, H. Adsorption Behavior of a Cyano-Functionalized Porphyrin on Cu(111) and Ag(111): From Molecular Wires to Ordered Supramolecular Two-Dimensional Aggregates. The Journal of Physical Chemistry C 2017, 121, 26361–26371. (28) Lepper, M.; Köbl, J.; Zhang, L.; Meusel, M.; Hölzel, H.; Lungerich, D.; Jux, N.; de Siervo, A.; Meyer, B.; Steinrück, H.-P.; Marbach, H. Controlling the Self-Metalation Rate of Tetraphenylporphyrins on Cu(111) via Cyano Functionalization. Angewandte Chemie International Edition 2018, 57, 10074–10079, doi:10.1002/anie.201803601. (29) Janiak, C. A critical account on π–π stacking in metal complexes with aromatic nitrogen-containing ligands. J. Chem. Soc., Dalton Trans. 2000, 3885–3896. (30) Hunter, C. A.; Lawson, K. R.; Perkins, J.; Urch, C. J. Aromatic interactions. J. Chem. Soc., Perkin Trans. 2 2001, 651–669. (31) Xu, D.; Ma, Y.; Jing, Z.; Han, L.; Singh, B.; Feng, J.; Shen, X.; Cao, F.; Oleynikov, P.; Sun, H.; Terasaki, O.; Che, S. π–π interaction of aromatic groups in amphiphilic molecules directing for single-crystalline mesostructured zeolite nanosheets. Nature Communications 2014, 5, 4262. (32) Albrecht, F.; Bischoff, F.; Auwärter, W.; Barth, J. V.; Repp, J. Direct Identification and Determination of Conformational Response in Adsorbed Individual Nonplanar Molecular Species Using Noncontact Atomic Force Microscopy. Nano Letters 2016, 16, 7703–7709. (33) Jarvis, S. P.; Taylor, S.; Baran, J. D.; Thompson, D.; Saywell, A.; Mangham, B.; Champness, N. R.; Larsson, J. A.; Moriarty, P. Physisorption Controls the Conformation and Density of States of an Adsorbed Porphyrin. Journal of Physical Chemistry C 2015, 119, 27982–27994. (34) Mielke, J.; Hanke, F.; Peters, M. V.; Hecht, S.; Persson, M.; Grill, L. Adatoms underneath Single Porphyrin Molecules on Au(111). Journal of the American Chemical Society 2015, 137, 1844–1849. (35) Auwärter, W.; Écija, D.; Klappenberger, F.; Barth, J. V. Porphyrins at interfaces. Nature Chemistry 2015, 7, 105. (36) Rothenberg, G. Catalysis: Concepts and Green Applications; Wiley-VCH: Weinheim, 2008; p 65. (37) The Nature of Halogen· · · Halogen Interactions: A Model Derived from Experimental Charge-Density Analysis. Angew. Chem. Int. Ed. 2009, 48, 3838–3841. (38) Bosch, E.; Barnes, C. L. Triangular Halogen–Halogen–Halogen Interactions as a Cohesive Force in the Structures of Trihalomesitylenes. Crystal Growth & Design 2002, 2, 299–302.

(39) Bondi, A. van der Waals Volumes and Radii. J. Phys. Chem. 1964, 68, 441–451. (40) In this DFT analysis, two Cl4 TPP molecules were placed on a Cu(111) surface where they were brought close together to form the desired bond and then allowed to relax. (41) We defined a freshly sample as a sample in which the deposition was done a maximum of 6 hours before the measurement. (42) Zhang, Z.; Lagally, M. G. Atomistic Processes in the Early Stages of ThinFilm Growth. Science 1997, 276, 377–383. (43) Enkovaara, J.; Rostgaard, C.; Mortensen, J. J.; Chen, J.; Dułak, M.; Ferrighi, L.; Gavnholt, J.; Glinsvad, C.; Haikola, V.; Hansen, H. A.; Kristoffersen, H. H.; Kuisma, M.; Larsen, A. H.; Lehtovaara, L.; Ljungberg, M.; Lopez-Acevedo, O.; Moses, P. G.; Ojanen, J.; Olsen, T.; Petzold, V.; Romero, N. A.; Stausholm-Møller, J.; Strange, M.; Tritsaris, G. A.; Vanin, M.; Walter, M.; Hammer, B.; Häkkinen, H.; Madsen, G. K. H.; Nieminen, R. M.; Nørskov, J. K.; Puska, M.; Rantala, T. T.; Schiøtz, J.; Thygesen, K. S.; Jacobsen, K. W. Electronic Structure Calculations with GPAW: A Real-Space Implementation of the Projector Augmented-Wave Method. J. Phys.: Condens. Matter 2010, 22, 253202. (44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865. (45) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFTD) for the 94 elements H-Pu. The Journal of Chemical Physics 2010, 132, 154104. (46) Klimeš, J.; Bowler, D. R.; Michaelides, A. Chemical accuracy for the van der Waals density functional. Journal of Physics: Condensed Matter 2010, 22, 022201. (47) Wagner, C. D. Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Data for Use in X-ray Photoelectron Spectroscopy; Perkin-Elmer: Eden Prairie, 1979. (48) Fan, Q.; Zhu, J.; Gottfried, J. In Encyclopedia of Interfacial Chemistry; Wandelt, K., Ed.; Elsevier: Oxford, 2018; pp 343 – 353. (49) We attribute this opposite relative stability to our use of PAW pseudopotentials that are not norm-conserving, something which is assumed by the vdW-DF2 implementation, but are currently unavailable for the gpaw code. This leads to an overestimation of metallic bond lengths with vdW-DF2 and an underestimation of the N–Cu bond energy in the inverted configuration (see pyrrole tests in Table S1 of Supporting Information). Although correcting this issue is beyond the scope of this article, we believe that it is important to show this discrepancy. (50) We use Cl3 TPP to denote a Cl4 TPP molecule which has lost one Cl atom. (51) Riss, A.; Pérez Paz, A.; Wickenburg, S.; Tsai, H.-Z.; De Oteyza, D. G.; Bradley, A. J.; Ugeda, M. M.; Gorman, P.; Jung, H. S.; Crommie, M. F.; Rubio, A.; Fischer, F. R. Imaging single-molecule reaction intermediates stabilized by surface dissipation and entropy. Nature Chemistry 2016, 8, 678–683. (52) Horcas, I.; Fernández, R.; Gómez-Rodríguez, J. M.; Colchero, J.; GómezHerrero, J.; Baro, A. M. WSXM: A software for scanning probe microscopy and a tool for nanotechnology. Review of Scientific Instruments 2007, 78, 013705. (53) https://shop.porphychem.com/products/5-10-15-20-tetra-4-chlorophenylporphyrin.html (accessed April 19th, 2019). (54) Larsen, A. H.; Vanin, M.; Mortensen, J. J.; Thygesen, K. S.; Jacobsen, K. W. Localized Atomic Basis Set in the Projector Augmented Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 195112. (55) Mortensen, J. J.; Hansen, L. B.; Jacobsen, K. W. Real-Space Grid Implementation of the Projector Augmented Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 71, 035109. (56) Tersoff, J.; Hamann, D. R. Theory of the scanning tunneling microscope. Phys. Rev. B 1985, 31, 805–813. (57) Larsen, A. H.; Mortensen, J. J.; Blomqvist, J.; Castelli, I. E.; Christensen, R.; Dułak, M.; Friis, J.; Groves, M. N.; Hammer, B.; Hargus, C.; Hermes, E. D.; Jennings, P. C.; Jensen, P. B.; Kermode, J.; Kitchin, J. R.; Kolsbjerg, E. L.; Kubal, J.; Kaasbjerg, K.; Lysgaard, S.; Maronsson, J. B.; Maxson, T.; Olsen, T.; Pastewka, L.; Peterson, A.; Rostgaard, C.; Schiøtz, J.; Schütt, O.; Strange, M.; Thygesen, K. S.; Vegge, T.; Vilhelmsen, L.; Walter, M.; Zeng, Z.; Jacobsen, K. W. The atomic simulation environment—a Python library for working with atoms. Journal of Physics: Condensed Matter 2017, 29, 273002.

ACS Paragon Plus Environment 8

Page 8 of 9

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Graphical TOC Entry

ACS Paragon Plus Environment 9