Stereochemical Recognition of Helicenes on Metal Surfaces

Jun 2, 2016 - In 1995, he founded the Molecular Surface Science Group, specializing in chirality of two-dimensional molecular crystals, functional sur...
3 downloads 11 Views 3MB Size
Article pubs.acs.org/accounts

Stereochemical Recognition of Helicenes on Metal Surfaces Karl-Heinz Ernst*,§,† Nanoscale Materials Science, Empa, Swiss Federal Laboratories for Materials Science and Technology, Ü berlandstrasse 129, CH-8600 Dübendorf, Switzerland † Department of Chemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland §

S Supporting Information *

CONSPECTUS: The chiral recognition among biomolecules is fundamentally important for many processes of life, including the stereochemistry of evolution. Of special interest is chiral recognition during crystallization of racemates, when either homochiral recognition leads to a conglomerate of homochiral crystals or heterochiral recognition dominates resulting in a racemic compound. The complex nature of molecular recognition at the level of nucleation and crystal growth renders it difficult to understand and calls for manageable model systems. Notably, the approach of studying aggregation of molecules at surfaces under well-defined conditions includes the benefit of the availability of a multitude of highly sensitive investigation methods, of which scanning tunneling microscopy (STM) with its submolecular resolution is tremendously valuable. Heterogeneous nucleation at surfaces is strongly favored over homogeneous nucleation in solution; hence, surfaces are significantly involved in stereochemical recognition during crystallization. Helicenes are a fascinating class of chiral compounds with outstanding optical activity. These π-conjugated, ortho-fused, aromatic hydrocarbons are promising candidates for organic electronic devices such as sensors, circular dichroic photonics, liquid crystal displays or spin filters. But in particular the defined footprint of their terminal benzo rings on a surface makes them interesting for studying stereochemical recognition with different single crystalline surfaces and the impact this has, in turn, on intermolecular recognition. In this Account, we describe the self-assembly of helicenes on metal surfaces with the focus on stereochemical recognition in twodimensional structures. Using the isomeric all-carbon helicenes, heptahelicene and dibenzohelicene as examples, different aggregation phenomena on different surfaces of single crystalline copper, silver, and gold are investigated. By means of STM different modes of transmission of molecular handedness from single molecules into extended two-dimensional supramolecular structures are identified. For the problem of racemate versus conglomerate crystallization, the impact of surface and molecular structure and their interplay are analyzed. This leads to detailed conclusions about the importance of the match of molecular and surface binding sites for long-range self-assembly. The absence of polar groups puts emphasis on van der Waals interaction and their maximization by steric overlap of molecular parts in enantiomeric and diastereomeric interactions. With STM as a manipulation tool, dimers are manually separated in order to analyze their chiral composition. And finally, new nonlinear cooperative effects induced by small enantiospecific bias are discovered that lead to single enantiomorphism in two-dimensional racemate crystals as well as in racemic multilayered films. By means of these model studies many details that govern chiral recognition at surfaces are rationalized.

1. INTRODUCTION In 1848, Louis Pasteur separated left- and right-handed ammonium sodium tartrate crystals manually and noted opposite rotation angles of plane polarized light for their aqueous solutions.1 His insight that the origin of chirality is based on molecular structure sparked the development of stereochemistry in the 19th century. Pasteur’s experiment was successful due to two aspects that allowed the manual separation: (1) the molecules crystallized from racemic solution as a conglomerate to yield homochiral crystals, and (2) molecular handedness was transmitted into the macroscopic shape of the crystal (hemihedral crystallization). Although more than 160 years have passed since Pasteur’s observation and resolution of enantiomers by crystallization is the most important means,2 it is still not understood why heterochiral recognition is strongly preferred in crystallization of racemates.3 An early explanation was provided by Liebisch in 1894 by showing © 2016 American Chemical Society

that racemate crystals tend to have a higher density than their pure enantiomer analogues.4,5 The tendency toward denser racemate crystals has been confirmed later for many substances,6 but amino acids, for example, do not follow this rule.7

2. A MODEL FOR CHIRAL CRYSTALLIZATION: HELICENES ON SURFACES A modern way for better understanding crystallization in general is based on surface science, which successfully unraveled fundamental steps in heterogeneous catalysis.8 The complexity of the problem requires well-defined model conditions, that is, ultrahigh vacuum and use of single crystal substrates. Another advantage Received: February 29, 2016 Published: June 2, 2016 1182

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

pure enantiomers, racemates (Figure 1e), or enantioenriched samples (ee ≠ 0) were studied.24,25

of this approach lies in the availability of analytical tools providing structural information with atomic resolution. Notably scanning tunneling microscopy (STM) has contributed inestimable insight.9 Because of the potential to serve in organic electronic devices, like light sensors or spin filters,10,11 surface self-assembly of helicene films has attracted much interest recently.12,13 In terms of chiral two-dimensional (2D) crystallization, they provide an interesting alternative to the commonly studied polar compounds like amino acids or long aliphatic chain compounds at the solid−liquid interface.14−16 Figure 1 presents several aspects

3. TRANSMISSION OF CHIRALITY IN TWO DIMENSIONS: FROM SINGLE MOLECULE TO EXTENDED ADSORBATE LATTICES The 2D self-assembly of enantiopure compounds on surfaces often leads to enantiomorphism of the adsorbate lattice.14,15 That is, the lattice vectors of the molecular 2D crystals have an oblique tilt angle with respect to the underlying substrate crystal lattice. Molecular handedness is thereby transmitted by stereochemical recognition among the molecules as well as between molecules and specific surface sites. Consequently, the other enantiomer shows a tilt in the opposite direction with respect to a high-symmetry direction of the metal surface. Self-assembly of enantiopure (M)-[7]H on single crystalline surfaces of gold, silver, and copper leads to long-range order only at coverages close to the saturated monolayer. All observed adsorbate lattices show the mentioned oblique tilt relative to the substrate lattice (Figure 2). Interestingly, similar chiral motifs are observed on all surfaces. For the (111) surfaces of Cu, Ag, and Au, a typical three-dot appearance is observed in STM at monolayer saturation coverage (Figure 2a−c).26−28 Each bright dot represents a single molecule (indicated by a circle in Figure 2b), so the three-dot appearance represents a three-molecule cluster (triplet). Such STM appearance reflects essentially different azimuthal alignments of the molecules in a quasi-hexagonal packing. That is, the three molecules in the unit cell are rotated by 120° with respect to each other, causing the distal parts to be closer to each other than their helical axes. The dark voids between the bright dots in the STM images are therefore not empty but represent lower parts of the molecules. Surprisingly, 5-amino[6]helicene shows the same structure at the Au(111)−trichlorobenzene interface.29 Models for the relative alignment of the (M)-[7]H triplets on Cu(111) and Ag(111), as obtained from molecular modeling calculation, are presented in Figure 2e. Due to identical interatomic distances for silver and gold (d = 2.88 Å), the three-dot structures of Au(111) (not shown) and Ag(111) are assumed to be identical. They have lower areal density than the one observed for Cu(111) (d = 2.55 Å). The relative arrangement of (M)-[7]H molecules in the unit cell on Cu(111) (Figure 2f) is similar to that observed for 3D crystals of isotactic polypropylene helices,30 which demonstrates the validity of 2D model systems for studying crystallization phenomena. Identical motifs on different surfaces suggest that the aggregation is dominated by lateral intermolecular interaction. However, the difference of intermolecular distances in otherwise identical motifs highlights the importance of favored adsites provided by the substrate. At about 90% of the monolayer saturation coverage (θ = 0.90), two other chiral cluster motifs are observed on Cu(111) and Au(111) (Figure 2c,d). A “6 + 3-structure” is formed on Cu(111).26,27 Here the unit cell includes a six-molecule pinwheel appearance that actually reflects the handedness of the single molecule. The distal parts of the six molecules appear either in a clockwise or counterclockwise sense (Figure 2c), which exemplifies nicely the transmission of single-molecule handedness into self-assembled molecular layers by distinct azimuthal orientations. On Au(111), a long-range ordered phase with a quadruplet motif is formed Figure 2d).28 Remarkably, the same motif has been observed for (M)-[7]H on the C4v-symmetric (100) surfaces of Cu and Ag (Figure 2g).31,32 Submolecular resolution STM images suggest that within the quadruplet, the molecules are rotated by 90°; corresponding models for

Figure 1. All-carbon helicenes and their stereochemical recognition with metal surfaces. (a) Ball-and-stick-models for the two [7]H enantiomers and (P)-db[5]H. (b) Model of adsorbate complex of (M)-[7]H on Cu(111). (c) Comparison of the stereochemical recognition of phenanthrene on different surfaces. (d) Principle setup of STM. A piezo scanner controlled by tunneling current moves a sharp metal tip in defined distance across a surface. (e) Sketch for rac-[7]H on a surface.

discussed in this Account. We focus here on heptahelicene ([7]H, C30H18, Figure 1a), a compound that crystallizes (like [6]H, [8]H, and [9]H) as a conglomerate,6,17 and its isomer dibenzopentahelicene (db[5]H; CAS, dibenzo[f,j]picene). Due to the lack of polar functional groups, the helicenes are expected to interact via van der Waals (vdW) forces. On most metal surfaces, [7]H adsorbs such that as many C6 rings as possible become aligned parallel to the surface (Figure 1b).18 Depending on the lattice constant of the metal surface or surface symmetry, the proximal part of the helicene is then responsible for the stereochemical molecule−surface recognition (Figure 1c).19 The distal part, on the other end, is responsible for the contrast in STM, because in constant current mode, the tip is there retracted most and becomes coded as brightest part in gray scale images. As early studies showed, the affinity of [7]H to metal surfaces can be substantial, suppressing the mobility as required for good long-range order.20−22 A higher mobility is achieved for [7]H on Cu, Ag, and Au surfaces, but this requires cryogenic temperatures for 2D aggregation. If not noted otherwise, all shown STM images were obtained after sublimation of the compound in vacuo onto the substrate at room temperature and cooling to 60 K. After successful enantioseparation of racemic [7]H (rac-[7]H) and determination of absolute configuration,23 either 1183

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

Figure 2. Triplets, quadruplets, and “pinwheels” formed by (M)-[7]H on the (111) and (100) surfaces of Au, Ag, and Cu. (a) STM image (20 × 20 nm2) of (M)-[7]H on Au(111) showing three-molecule (triplet) clusters as packing motif in the saturated monolayer (θ = 1.0). (b) The STM image (20 × 20 nm2) of the saturated monolayer of (M)-[7]H on Ag(111) shows the same triplet motif. A single molecule in a triplet is marked with a white circle. Unit cells are indicated as red diamonds. Green lines highlight the tilt of the adlattice with respect to the [110̅ ] direction (arrow) of the substrate. (c) STM image (50 × 50 nm2) at θ = 0.96 of (M)-[7]H on Cu(111), showing the coexistence of the “triplet” phase and a “pinwheel” phase. The unit cell (inset) of the “pinwheel” phase appears as a six-molecule pinwheel plus a triplet. The pinwheel handedness is opposite for (M)- and (P)-[7]H. (d) STM image (20 × 20 nm2) of (M)-[7]H on Au(111) at a coverage of θ = 0.95, showing a four-molecule (quadruplet) motif as building block. (e) Models of triplets on Cu(111) (top) and Ag(111) (bottom). White filled circles mark the distal parts of the molecules that dominate the STM contrast. (f) Unit cell of the (M)-[7]H triplet phase on Cu(111). (g) Quadruplets are building blocks in the (M)-[7]H monolayer on Ag(100) (50 × 50 nm2; inset, 4.1 × 4.1 nm2). (h,i) Models for the quadruplet motif of (M)-[7]H on Ag(100) and Au(111). (For STM parameters, see Supporting Information Table S1.) Reproduced with permission from refs 27 and 28. Copyright 2008 and 2014 American Chemical Society.

extended Hückel simulation and molecular modeling results based on Amber force field calculations, the zigzag rows were clearly identified as alternation of M/P enantiomers.33 Interestingly, cyano-functionalized rac-[7]H undergoes spontaneous resolution on Cu(111).34 Formation of racemic mirror domains is based on the fact that there are two equal but enantiomorphous alignments of the enantiomers in an M−P pair (Figure 3c,d). On Ag(111) and Au(111) zigzag rows of rac-[7]H were observed as well,28 but with two pronounced differences: On both surfaces, the zigzag rows were aligned parallel to the high-symmetry [11̅0] direction, so no enantiomorphism due to oblique adsorbate lattice vector tilt was present, and a relatively low degree of long-range order was achieved. While [7]H on the (111) surfaces of the group 11 elements shows almost identical adsorbate structures, the (100) surfaces are quite different in case of racemate aggregation (Figure 4). Quadruplet structures are observed at very low coverages on Cu(100) and on Ag(100). On Cu(100), these quadruplets prevail at all coverages. Submolecular resolution STM (Figure 4b,c) and STM-appearance modeling confirm the homochiral composition of the quadruplets.31 Consequently, a 2D conglomerate forms at high coverage (Figure 4a). On Ag(100), on the other hand, the homochiral quadruplets coexist with zigzag rows up to high coverage, but the quadruplets disappear completely at monolayer saturation coverage. Whereas zigzag rows are allowed to grow longer and longer, no additional attachment of molecules to quadruplets is observed. As force field simulations show, at the stage of nucleation, homochiral quadruplets are more stable than heterochiral 2M−2P structures. But already for five-molecule

quadruplets on Ag(100) and Au(111) are shown in Figure 2h,i. Such quadruplets are formed already at low coverages and stay intact until monolayer saturation. All observed quadruplet phases show basically the same molecular density per unit area (Supporting Information Table S2). It is speculated that the combination of favored adsorption sites and the 4-fold lateral interaction causes this stable 90° arrangement.

4. TWO-DIMENSIONAL CRYSTALLIZATION OF HELICENE RACEMATES 4.1. Homochiral versus Heterochiral Recognition

Equipped with the knowledge of the enantiopure structures, it is relatively easy to judge whether racemic crystals or conglomerates form upon deposition of a racemic mixture. That is, coexistence of mirror domains with the same local crystal structure as observed for the pure enantiomers strongly suggests 2D conglomerate crystallization. However, it is not unlikely that racemic 2D crystals show enantiomorphism in the plane. A representative example is rac-[7]H/Cu(111). At monolayer saturation coverage rotational domains (due to the 3-fold symmetric substrate) as well as mirror domains are observed.27 Only one type of mirror domain is observed on the same terrace (Figure 3a). Hence, mirror domain boundaries (MDBs) coincide with step edges of the metal surface. Higher resolution STM images reveal zigzag rows with an oblique tilt with respect to the [110̅ ] direction of the Cu(111) surface (Figure 3, Table S2). The STM contrast is again dominated by the distal part of the molecular helix. Considering that such structure was not observed for pure enantiomers, through modeling of STM appearance by 1184

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

clusters, the heterochiral combinations are energetically favored.32 Once the stable homochiral quadruplet is formed, attaching more molecules to this structure is energetically not favorable. This observation is again explained by a better match of preferred molecular and surface binding sites on Cu(100), but not on Ag(100) (Figure 4h). This stabilizes the quadruplet on Cu(100), but a homochiral extended lattice on Ag(100) would induce substantial stress in the molecular layer. Hence, although lower in density, only the zigzag row motif supports further growth into extended structures. 4.2. Pasteur at the Nanoscale: [7]H/Cu(111) Revisited

The transition from homochiral nuclei to racemic 2D domains on Ag(100) brings us back to the (111) surfaces, where only higher coverage results were available. Could it be that the initial aggregates of [7]H on Cu(111), for example, are homochiral, but these grow into heterochiral structures with larger cluster size? Earlier force field calculations favored heterochiral dimers and heterochiral small clusters.27 Due to high mobility at the level of dimers or trimers, lower temperatures are required to study their chiral recognition. Figure 5 shows the situation on an atomically flat terrace for very low coverage at 6 K.35 Predominantly, dimers are detected (Figure 5a). They reveal broken mirror symmetry in STM (Figure 5b). From such submolecular contrast, however, the chiral composition can be hardly judged directly. On the other hand, single molecules show either a clockwise or counterclockwise decrease of brightness in the helix (Figure 5c), therefore revealing their absolute configuration. Consequently, after manual separation of the dimers, their composition should become definable. There are numerous examples for single molecule manipulations with the STM;36 some of them are concerned with chirality.37,38

Figure 3. Two-dimensional self-assembly of rac-[7]H on Cu(111). (a) The STM image (150 × 150 nm2) of the monolayer saturation phase shows mirror domains that have zigzag rows as structural motif (inset, 5 × 5 nm2). (b) Higher-magnification STM image (10 × 10 nm2) that suggests alternation of both enantiomers in a single zigzag row. Double row distances (ρ and ρ′) vary at intermediate coverages. (c,d) Structure models for both mirror domains. White filled circles mark the distal parts of the molecules that dominate the STM contrast. Circular arrows mimic the increasing brightness in the STM image (b). The [110̅ ] direction of the substrate surface is indicated as arrow. Reproduced with permission from refs 27 and 28. Copyright 2008 and 2014 American Chemical Society.

Figure 4. Two-dimensional aggregation of rac-[7]H on Cu(100) and Ag(100). (a) STM image (35 × 35 nm2) of a close-packed layer of rac-[7]H on Cu(100) near a mirror-domain boundary. The enantiomorphous domains (framed in blue and green) are homochiral within each domain and consist of quadruplets of molecules. The angle between both domains is 29°. A unit cell is indicated as white square. (b) STM image (20 × 20 nm2) at low coverage of the racemate. Quadruplets appear as four-bladed propeller. High-resolution images (6 × 6 nm2) of the two mirror domains at high coverage show the same propeller structure of the quadruplets. In the model, the distal rings are colored yellow and orange, which fits the appearance in the images. (c) High resolution STM image (6 × 6 nm2) of one mirror domain on Cu(100) confirming homochirality by the identical sequence of submolecular brightness. The sense of molecular helicity is indicated by a green arrow pointing to the proximal ends (M-helicity). (d) Homochiral quadruplets are also identified at low coverage of rac-[7]H on Ag(100) (50 × 50 nm2; insets, 5 × 5 nm2). (e) At slightly higher coverage, zigzag motifs appear (45 × 45 nm2). (f) With further increasing coverage longer zigzag rows form, still coexisting with quadruplets (50 × 50 nm2; inset, 8 × 4.5 nm2). (g) At full monolayer coverage only the zigzag motif prevails (50 × 50 nm2). (h) Sketch of the difference in molecular footprints of [7]H on Cu(100) (top) and Ag(100) (bottom). Reproduced with permission from refs 31 and 32. Copyright 2014 Royal Society of Chemistry (a−c) and 2015 American Chemical Society (d−h). 1185

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

Figure 6. Two-dimensional self-assembly of rac-db[5]H on Au(111). (a) Ordered mirror domains are surrounded by nonordered areas (150 × 150 nm2). (b) STM image of a region near a mirror domain boundary (blue transparent line) emphasizing the opposite tilt of the domains (30 × 30 nm2). (c) STM image of a single domain (10 × 10 nm2). Submolecular-resolution allows direct determination of absolute handedness of molecules: The clockwise sequence of brightness, indicated by redblue-green-yellow ellipses or 1−2−3−4 in the inset (2 × 2 nm2, averaged 30×), identifies (P)-enantiomers. (d) STM image (7.1 × 6.3 nm2, averaged 30×) and molecular model of a homochiral dimer and the four-molecule arrangement in the unit cell. Reproduced with permission from ref 39. Copyright 2013 American Chemical Society.

Figure 5. Separation of vdW dimers via STM manipulation. (a) At 6 K and very low coverage, predominantly dimers are observed on a flat Cu(111) terrace (50 × 50 nm2). (b) STM image (5 × 5 nm2) of two enantiomorphous dimers. (c) STM image (5 × 5 nm2) of [7]H enantiomers. The decrease in brightness (1 → 2 → 3) runs clockwise for a (P)-enantiomer and counterclockwise for the (M)-enantiomer. (d) Space-filling model of the dimer configuration using the vdW radii of the atoms. (e) Sketch of the dimer separation and chiral evaluation, including: target dimer location (1), modification of the STM tip with a single [7]H molecule (2,3), target dimer retrieving (4), dimer separation (5,6), reconditioning of STM tip, and chiral analysis of monomers (8). Reproduced with permission from ref 35. Copyright 2015 American Chemical Society.

A thorough analysis of STM contrast revealed that the system crystallizes as a 2D conglomerate. Interestingly, two molecules form a homochiral dimer and the unit cell contains two of such dimers, one rotated by 180° to the other. There is substantial overlap of benzo rings within the homochiral pair and a smaller overlap between the pairs Figure 6d).39

Figure 5e shows a sketch of the STM manipulation procedure that finally revealed the chiral nature of the dimers. After location of a target dimer the STM tip has to be modified with a single [7]H molecule. This step is necessary, because a bare metal tip has a too strong attraction, rather moving complete dimers instead of their separation. The [7]H-modified tip, on the other hand, allows π−π repulsion at short distances, which leads with a high probability to the separation of the dimer. Identification of handedness, however, requires again a bare metal tip and therefore a reconditioning of it by poking into the metal surface somewhere away from the target. Returning to the separated dimer allows the analysis of both monomers. Of 52 analyzed dimers, 51 were identified to be heterochiral. Further manipulations, including creating and switching M−P pairs finally allowed reasonable conclusions on the relative alignment of the enantiomers in a dimer (Figure 5d).35

5. COOPERATIVELY INDUCED HOMOCHIRALITY IN UNBALANCED LAYERS 5.1. Sergeant-and-Soldiers and Majority Rule in 2D

There are many examples of cooperative response to small chiral bias, leading to large magnification effects in supramolecular systems.40,41 The so-called “sergeant-and-soldiers” and “majority-rule” effects, leading to single-sense of helicity in polymers due to small homochiral bias or enantiomeric excess in polymer side chains, respectively, are prominent examples.42 Similar effects have been described for 2D systems.43 The first 2D “sergeant-and-soldiers” examples were reported for polar organic acids on Cu(110),44,45 followed shortly by [7]H.33 Although still close to racemic composition, single enantiomorphism of [7]H on Cu(111) is observed either at ee larger than +8% (M/P = 54:46) or lower than −8% (M/P = 46:54).27 As molecular modeling calculations confirmed, the composition within the domains is strictly racemic.33 The chiral bias allows only one enantiomorphous arrangement of the M−P pair. The excess of the majority enantiomer is located outside the domains, preferably at step edges of the substrate. At sufficient positive ee (excess of M),24 only λ-domains are observed. Modeling the situation at the edge of a λ-domain under close-packing conditions shows that excess of (P)-[7]H quickly becomes unstable, whereas

4.3. Footprint or Overlap: Dibenzopentahelicene

Different helicene species, in particular if modified with polar groups, can show different modes of chiral recognition. Already slight modification of the helicene backbone may have consequences, as the following example of the isomer of [7]H, racdb[5]H shows. The main difference compared with [7]H lies in the larger helix opening angle, allowing substantial more pronounced intermolecular overlap and a potentially more intense footprint due to the additional benzo groups. On Au(111), only 50% of the surface showed ordered areas at monolayer saturation coverage, clearly displaying mirror domains (Figure 6).39 1186

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

Figure 7. Two-dimensional single mirror-domain-enantiomorphism for [7]H on Cu(111). (a,b) Exclusively ρ- or λ-domain formation is observed at ee of −8% or +8%, respectively. (c) Molecular modeling of chiral bias at a λ-domain edge. With increasing close packing, i.e., decreasing space or number of available surface sites, excess of P becomes quickly unstable. (d) Sketch of the interplay of energies at domain edges (black arrows) and MDBs (white arrows). Reproduced with permission from refs 27 and 33. Copyright 2008 American Chemical Society (a,b) and 2006 Nature Publishing Group (c,d).

Figure 8. Chiral conflict between (M)-[7]H and rac-db[5]H on Au(111). (a) STM image (100 × 100 nm2) showing only one large domain of (M)-db[5]H. (b) Models of (M)-[7]H/P-db[5]H diastereomers. (c) Plot of fraction of (M)-db[5]H domains and overall ordered domains versus (M)-[7]H content in area % and mol %. Reproduced with permission from ref 39. Copyright 2013 American Chemical Society.

(M)-[7]H prevails at close-packing (Figure 7c). This result shows actually which enantiomorph will form under ee. However, nucleation of both enantiomorphs at different places of a terrace may occur. MDBs are identified to be higher in energy, because they do not allow dense packing (Figure 7d). But rearrangement of M−P pairs near MDBs requires only small site changes of the enantiomers, so the MDB will be “expelled” through one domain. This leads to spontaneous local symmetry breaking on a single terrace with an equal chance of forming either enantiomorph. Due to chiral bias at the domain edges under ee, one domain becomes more stable, favoring the majority domain during restructuring. One can look at the chiral bias at the domain edge as a sergeant, who forces the M−P pairs at the edge (soldiers) into the favored arrangement. In order to avoid MDB formation, these pairs steer other pairs into the same arrangement. Because both enantiomers are present in the excess area, the initial arrangement at the domain edge falls under the “majority rule”.

been reported for tartaric acid/malic acid mixtures on Cu(110), where instead of vdW forces hydrogen bonds are at work.46 5.3. Single Enantiomorphs and Enantiospecific Dewetting

Single enantiomorphism has also been identified in multilayered samples of [7]H on Cu(111).47 Upon exceeding the closepacked monolayer coverage (θ = 1) by only a few percent, a new situation is encountered. Double-layer nucleation is favored over the first nucleation such that double layer islands are observed, embedded in low-ordered first layer areas (Figure 9a). Again mirror domains are observed. At complete second layer coverage, and similar to the first layer, single mirror domains completely cover single terraces (Figure 2b). Submolecular STM images reveal homochirality of the second layer for a given domain (Figure 9c,d). At positive ee, only one enantiomorph is observed with the second layer containing exclusively the (P)-enantiomer (Figure 9e−g). With the second layer completely filled and small ee, that is, close to racemic composition (Figure 9f,g), this means that the first layer must contain the other enantiomer. Under enantioenriched conditions, the minority enantiomer is always expelled from the first layer, and the entire top layer is homochiral. This effect has been observed for up to four layers. The transition from monolayer to multilayer is characterized by turning a 2D racemate into a 3D racemate with homochiral layers (Figure 9h).

5.2. Chiral Conflict and Diastereomers

We have seen that an enantiomorphous racemic lattice can be switched or suppressed by chiral bias. For a conglomerate, any excess is expected to linearly shift the area toward the majority. But what happens when the chiral imbalance stems from a different species, that is, diastereomeric interactions are at work? The example for the db[5]H racemate mixed with increasing amount of (M)-[7]H is presented in Figure 8.25 At 26% molar content of (M)-[7]H, only the (M)-db[5]H enantiomorph is observed, whereas all other constituents are located in a large disordered area. One of the enantiomers undergoes preferentially diastereomer formation, leaving the other still able to form homochiral domains. In order to maximize vdW forces, spatial overlap between both species should be most. The strongest overlap for all pair combinations is found by molecular modeling for (M)-[7]H/(P)-db[5]H (Figure 8b). The same effect has

6. SUMMARY AND OUTLOOK Heterogeneous nucleation is strongly favored over homogeneous nucleation, so surfaces play an important role for initial steps of crystallization. Studying crystallization on well-defined surface 1187

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research

Figure 9. Single enantiomorphism in double-layered rac-[7]H on Cu(111). (a) STM image (260 × 260 nm2, θ = 1.03) of a 2nd layer island (red) surrounded by low-ordered 1st layer (gray). (b) STM image of the completed 2nd layer. Single mirror domains completely cover terraces (150 × 150 nm2, θ = 1.60). (c,d). Submolecular-resolution STM images (10 × 10 nm2, θ = 1.70) of both mirror domains in the second layer. All molecules of a single domain exhibit the same mirror-breaking fore-lobe contrast. (e) STM image (400 × 400 nm2, θ = 1.05) at 7% ee. (f) STM image (200 × 200 nm2, θ = 1.60) at 8% ee with exclusively (P)-[7]H in the second layer. (g) STM image at −8% ee, the second layer consisting of only (M)-[7]H molecules (200 × 200 nm2, θ = 1.61). (h) Structure model of a multilayered racemic crystal including homochiral layers. Reproduced with permission from ref 47. Copyright 2015 Wiley and Sons.



model systems bodes well for new insight and leads to discovery of interesting new effects. Unfunctionalized helicenes interacting via vdW binding provide here a wealth of interesting phenomena, including cooperative amplification into single enantiomorphism. Essentially, two processes govern the outcome of molecular 2D crystallization: the stereochemical recognition between substrate and molecule and the lateral interaction between the molecules. The latter occurs under the limitation that the molecules prefer certain surface binding sites. A good match of a preferred molecular lattice with the substrate lattice results then in a higher stability of the lattice, but a mismatch will introduce stress into the extended structure and lowers the stability. In general, such “binding site grid” puts a limit to the available intermolecular distances and strongly influences whether homochiral or heterochiral recognition is favored. There have been first surface studies of functionalized helicenes or heterohelicenes with interesting results.29,34,48−52 A systematic study, including a tuning of the polarity of functional groups, for example, will complete our picture of the factors that govern stereochemical recognition of helicenes on surfaces. The next step, however, must be to use these 2D systems as templates to study the stereochemistry during further growth into three-dimensional structures and construction of useful devices based on this fascinating class of molecules.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Funding

This work was funded by the Swiss National Science Foundation (Grant 200020_163296). Notes

The author declares no competing financial interest. Biography Karl-Heinz Ernst, originally trained as Chemical Technical Assistant (CTA), studied Chemical Engineering and Chemistry at the TFH Berlin and the Freie Universität Berlin, respectively. He conducted his Ph.D. studies at the Berlin Electron Storage Ring for Synchrotron Radiation (BESSY I) and the Freie Universität Berlin and joined (after postdoctoral research at the University of Washington in Seattle) Empa, the Swiss Federal Laboratories for Materials Science and Technology. In 1995, he founded the Molecular Surface Science Group, specializing in chirality of two-dimensional molecular crystals, functional surfaces, and single molecule surface dynamics. He was visiting researcher at the Physics Department of UC Berkeley, at the Department of Bioengineering at the University of Washington in Seattle and at IBM Almaden Research Center, San José, CA. Currently he holds an adjunct faculty position at the Department of Chemistry at the University of Zurich and was recently promoted at Empa to Distinguished Senior Scientist.

ASSOCIATED CONTENT



S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.accounts.6b00110.

REFERENCES

(1) Pasteur, L. Recherches Sur Les Relations Qui Peuvent Exister Entre La Forme Cristalline: La Composition Chimique Et Les Sens De La Polarisation Rotatoire. Ann. Chim. Phys. 1848, 24, 442−459. (2) Coquerel, G.; Tamura, R.; Yoshioka, R.; Faigl, F.; Kellogg, R. M.; Sakai, K.; Sakurai, R.; Murakami, H. Novel Optical Resolution Technologies. Top. Curr. Chem. 2007, 269, 1−313.

STM electrical acquisition parameters of images and structural parameters of [7]H adsorbate lattices and their matrix notations53 (PDF) 1188

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research (3) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates, and Resolutions; Krieger Pub. Co., Malabar, FL, 1994. (4) Ernst, K.-H. On the validity to call Wallach’s rule Wallach’s rule. Isr. J. Chem. 2016, (in press) DOI: 10.1002/ijch.201600029. (5) Wallach, O. Zur Kenntniss Der Terpene Und Der Ä therischen Oele 34. Liebigs Ann. Chem. 1895, 286, 119−143. (6) Brock, C. P.; Schweizer, W. B.; Dunitz, J. D. On the Validity of Wallach’s Rule: on the Density and Stability of Racemic Crystals Compared with Their Chiral Counterparts. J. Am. Chem. Soc. 1991, 113, 9811−9820. (7) Dunitz, J. D.; Gavezzotti, A. Proteogenic Amino Acids: Chiral and Racemic Crystal Packings and Stabilities. J. Phys. Chem. B 2012, 116, 6740−6750. (8) Ertl, G. Reactions at Surfaces: From Atoms to Complexity (Nobel Lecture). Angew. Chem., Int. Ed. 2008, 47, 3524−3535. (9) De Feyter, S.; De Schryver, F. C. Two-Dimensional Supramolecular Self-Assembly Probed by Scanning Tunneling Microscopy. Chem. Soc. Rev. 2003, 32, 139−150. (10) Yang, Y.; da Costa, R. C.; Fuchter, M. J.; Campbell, A. J. Circularly Polarized Light Detection by a Chiral organic Semiconductor Transistor. Nat. Photonics 2013, 7, 634−638. (11) Kiran, V.; Mathew, S. P.; Cohen, S. R.; Hernández Delgado, I.; Lacour, J.; Naaman, R. Helicenes-a New Class of Organic Spin Filter. Adv. Mater. 2016, 28, 1957−1962. (12) Shen, Y.; Chen, C.-F. Helicenes: Synthesis and Applications. Chem. Rev. 2012, 112, 1463−1535. (13) Gingras, M. One Hundred Years of Helicene Chemistry. Part 3: Applications and Properties of Carbohelicenes. Chem. Soc. Rev. 2013, 42, 1051−1095. (14) Raval, R. Chiral Expression From Molecular Assemblies at Metal Surfaces: Insights From Surface Science Techniques. Chem. Soc. Rev. 2009, 38, 707−721. (15) De Feyter, S.; Iavicoli, P.; Xu, H. Expression of Chirality in Physisorbed Monolayers Observed by Scanning Tunneling Microscopy. In Chirality at the Nanoscale: Nanoparticles, Surfaces, Materials and more. Amabilino, D. B., Ed.; Wiley-VCH, Weinheim, Germany, 2009; pp 215− 245. (16) Ernst, K.-H. Molecular Chirality at Surfaces. Phys. Status Solidi B 2012, 249, 2057−2088. (17) Martin, R. H.; Marchant, M. J. Tetrahedron 1974, 30, 343−345. (18) Fasel, R.; Cossy, A.; Ernst, K.-H.; Baumberger, F.; Greber, T.; Osterwalder, J. Orientation of Chiral Heptahelicene C30H18 On Copper Surfaces: an X-Ray Photoelectron Diffraction Study. J. Chem. Phys. 2001, 115, 1020−1027. (19) We use distal and proximal here for the parts of a molecule farthest away from and closest to the surface, respectively. (20) Ernst, K.-H.; Böhringer, M.; McFadden, C. F.; Hug, P.; Müller, U.; Ellerbeck, U. Nanostructured Chiral Surfaces. Nanotechnology 1999, 10, 355−361. (21) Ernst, K.-H.; Kuster, Y.; Fasel, R.; McFadden, C. F.; Ellerbeck, U. Adsorption of Helical Aromatic Molecules: Heptahelicene on Ni(111). Surf. Sci. 2003, 530, 195−202. (22) Ernst, K.-H.; Neuber, M.; Grunze, M.; Ellerbeck, U. NEXAFS Study on the Orientation of Chiral P-Heptahelicene on Ni(100). J. Am. Chem. Soc. 2001, 123, 493−495. (23) Bürgi, T.; Urakawa, A.; Behzadi, B.; Ernst, K.-H.; Baiker, A. The Absolute Configuration of Heptahelicene: a VCD Spectroscopy Study. New J. Chem. 2004, 28, 332−334. |M | − |P| (24) For helicenes ee is defined as ee[%] = |M| + |P| × 100. (25) Because of the low inversion barrier and the relatively high sublimation temperature, separation of enantiomers was not pursued for the dibenzopentahelicene derivative. (26) Fasel, R.; Parschau, M.; Ernst, K.-H. Chirality Transfer From Single Molecules Into Self-Assembled Monolayers. Angew. Chem., Int. Ed. 2003, 42, 5178−5181. (27) Parschau, M.; Fasel, R.; Ernst, K.-H. Coverage and Enantiomeric Excess Dependent Enantiomorphism in Two-Dimensional Molecular Crystals. Cryst. Growth Des. 2008, 8, 1890−1896.

(28) Seibel, J.; Parschau, M.; Ernst, K.-H. Two-Dimensional Crystallization of Enantiopure and Racemic Heptahelicene on Ag(111) and Au(111). J. Phys. Chem. C 2014, 118, 29135−29141. (29) Balandina, T.; W. van der Meijden, M.; Ivasenko, O.; Cornil, D.; Cornil, J.; Lazzaroni, R.; Kellogg, R. M.; De Feyter, S. Self-Assembly of an Asymmetrically Functionalized [6]Helicene at Liquid/Solid Interfaces. Chem. Commun. (Cambridge, U. K.) 2013, 49, 2207−2209. (30) Stocker, W.; Schumacher, M.; Graff, S.; Thierry, A.; Wittmann, J.C.; Lotz, B. Epitaxial Crystallization and AFM Investigation of a Frustrated Polymer Structure: Isotactic Poly(Propylene), B Phase. Macromolecules 1998, 31, 807−814. (31) Seibel, J.; Zoppi, L.; Ernst, K.-H. 2D Conglomerate Crystallization of Heptahelicene. Chem. Commun. (Cambridge, U. K.) 2014, 50, 8751− 8753. (32) Seibel, J.; Parschau, M.; Ernst, K.-H. From Homochiral Clusters to Racemate Crystals: Viable Nuclei in 2D Chiral Crystallization. J. Am. Chem. Soc. 2015, 137, 7970−7973. (33) Fasel, R.; Parschau, M.; Ernst, K.-H. Amplification of Chirality in Two-Dimensional Enantiomorphous Lattices. Nature 2006, 439, 449− 452. (34) Stöhr, M.; Boz, S.; Schär, M.; Nguyen, M.-T.; Pignedoli, C. A.; Passerone, D.; Schweizer, W. B.; Thilgen, C.; Jung, T. A.; Diederich, F. Self-Assembly and Two-Dimensional Spontaneous Resolution of Cyano-Functionalized [7]Helicenes on Cu(111). Angew. Chem., Int. Ed. 2011, 50, 9982−9986. (35) Ernst, K.-H.; Baumann, S.; Lutz, C. P.; Seibel, J.; Zoppi, L.; Heinrich, A. J. Pasteur’s Experiment Performed at the Nanoscale: Manual Separation of Chiral Molecules, One by One. Nano Lett. 2015, 15, 5388−5392. (36) Morgenstern, K.; Lorente, N.; Rieder, K.-H. Controlled Manipulation of Single Atoms and Small Molecules Using the Scanning Tunnelling Microscope. Phys. Status Solidi B 2013, 250, 1671−1751. (37) Böhringer, M.; Morgenstern, K.; Schneider, W. D.; Berndt, R. Separation of a Racemic Mixture of Two-Dimensional Molecular Clusters by Scanning Tunneling Microscopy. Angew. Chem., Int. Ed. 1999, 38, 821−823. (38) Parschau, M.; Passerone, D.; Rieder, K.-H.; Hug, H. J.; Ernst, K.H. Switching the Chirality of Single Adsorbate Complexes. Angew. Chem., Int. Ed. 2009, 48, 4065−4068. (39) Seibel, J.; Allemann, O.; Siegel, J. S.; Ernst, K.-H. Chiral Conflict Among Different Helicenes Suppresses Formation of One Enantiomorph in 2D Crystallization. J. Am. Chem. Soc. 2013, 135, 7434−7437. (40) Palmans, A. R. A.; Meijer, E. W. Amplification of Chirality in Dynamic Supramolecular Aggregates. Angew. Chem., Int. Ed. 2007, 46, 8948−8968. (41) Zepik, H.; Shavit, E.; Tang, M.; Jensen, T. R.; Kjaer, K.; Bolbach, G.; Leiserowitz, L.; Weissbuch, I.; Lahav, M. Chiral Amplification of Oligopeptides in Two-Dimensional Crystalline Self-Assemblies on Water. Science 2002, 295, 1266−1269. (42) Green, M. M.; Park, J. W.; Sato, T.; Teramoto, A.; Lifson, S. S.; Selinger, R.; Selinger, J. V. The Macromolecular Route to Chiral Amplification. Angew. Chem., Int. Ed. 1999, 38, 3138−3154. (43) Ernst, K.-H. Amplification of Chirality at Solid Surfaces. Origins Life Evol. Biospheres 2010, 40, 41−50. (44) Parschau, M.; Romer, S.; Ernst, K.-H. Induction of Homochirality in Achiral Enantiomorphous Monolayers. J. Am. Chem. Soc. 2004, 126, 15398−15399. (45) Parschau, M.; Kampen, T.; Ernst, K.-H. Homochirality in Monolayers of Achiral Meso Tartaric Acid. Chem. Phys. Lett. 2005, 407, 433−437. (46) Roth, C.; Passerone, D.; Ernst, K.-H. Pasteur’s Quasiracemates in 2D: Chiral Conflict Between Structurally Different Enantiomers Induces Single-Handed Enantiomorphism. Chem. Commun. (Cambridge, U. K.) 2010, 46, 8645−8647. (47) Parschau, M.; Ernst, K.-H. Disappearing Enantiomorphs: Single Handedness in Racemate Crystals. Angew. Chem., Int. Ed. 2015, 54, 14422−14426. (48) Ascolani, H.; van der Meijden, M. W.; Cristina, L. J.; Gayone, J. E.; Kellogg, R. M.; Fuhr, J. D.; Lingenfelder, M. X. Van Der Waals 1189

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190

Article

Accounts of Chemical Research Interactions in the Self-Assembly of 5-Amino[6]Helicene on Cu(100) and Au(111). Chem. Commun. (Cambridge, U. K.) 2014, 50, 13907− 13909. (49) Shchyrba, A.; Nguyen, M.-T.; Wäckerlin, C.; Martens, S.; Nowakowska, S.; Ivas, T.; Roose, J.; Nijs, T.; Boz, S.; Schär, M.; Stöhr, M.; Pignedoli, C. A.; Thilgen, C.; Diederich, F.; Passerone, D.; Jung, T. A. Chirality Transfer in 1D Self-Assemblies: Influence of H-Bonding vs Metal Coordination Between Dicyano[7]Helicene Enantiomers. J. Am. Chem. Soc. 2013, 135, 15270−15273. (50) Rybácě k, J.; Huerta-Angeles, G.; Kollárovič, A.; Stará, I. G.; Starý, I.; Rahe, P.; Nimmrich, M.; Kühnle, A. Racemic and Optically Pure Heptahelicene-2-Carboxylic Acid: Its Synthesis and Self-Assembly Into Nanowire-Like Aggregates. Eur. J. Org. Chem. 2011, 2011, 853−860. (51) Taniguchi, M.; Nakagawa, H.; Yamagishi, A.; Yamada, K. STM Observation of Thia[11]heterohelicene on Gold(111) and Gold(110) Surface. Surf. Sci. 2002, 507-510, 458−462. (52) Chaunchaiyakul, S.; Krukowski, P.; Tsuzuki, T.; Minagawa, Y.; Akai-Kasaya, M.; Saito, A.; Osuga, H.; Kuwahara, Y. Self-Assembly Formation of M-Type Enantiomer of 2,13-Bis(Hydroxymethyl)[7]Thiaheterohelicene Molecules on Au(111) Surface Investigated by STM/CITS. J. Phys. Chem. C 2015, 119, 21434−21442. (53) Merz, L.; Ernst, K.-H. Unification of the Matrix Notation in Molecular Surface Science. Surf. Sci. 2010, 604, 1049−1054.

1190

DOI: 10.1021/acs.accounts.6b00110 Acc. Chem. Res. 2016, 49, 1182−1190