- Email: [email protected]
Chiral metal surfaces from the adsorption of chiral and achiral molecules
Applied Surface Science 241 (2005) 150–156 www.elsevier.com/locate/apsusc
Chiral metal surfaces from the adsorption of chiral and achiral molecules V. Humblot, R. Raval* Department of Chemistry, Surface Science Research Centre, University of Liverpool, Liverpool L69 3BX, UK Available online 11 November 2004
Abstract Chiral surfaces, capable of existing in two distinguishable mirror forms that cannot be superimposed, are attracting worldwide attention. The adsorption of complex organic molecules provides a means of introducing the ultimate discrimination function of chirality to a metal surface. Here, a comparison of the chiral tartaric acid (HOOC–CHOH–CHOH–COOH) molecule and the achiral succinic acid (HOOC–CH2–CH2–COOH) molecule on a Cu(110) surface is presented. For both molecules, twodimensional assembly is found to depend strongly on molecule–metal bonding interactions, whereas the presence/absence of the OH groups causes subtler, second-order effects on the self-assembled structure. The driving force for creating chiral organisations is shown to arise from adsorption-induced asymmetrisation, via molecular distortion and/or metal reconstruction of the local adsorption unit. The macroscopic chirality of the surface is then determined by whether nucleation points of both chirality can be equally created, or whether non-degeneracy can be introduced to favour one chirality. # 2004 Elsevier B.V. All rights reserved. PACS: 68.43.h; 87.15.By Keywords: Chirality; Surfaces; Chemisorption; Carboxylates
1. Introduction Chirality is a simple geometric property that dictates that the mirror transformation of an object that does not possess any inverse symmetry elements, is a non-identity operation, i.e. the object and its mirror image are non-superimposable. As a result, a chiral object can exist in two distinguishable mirror, or * Corresponding author. Tel.: +44 151 794 6891; fax.: +44 151 794 3896. E-mail address: [email protected] (R. Raval).
enantiomeric, forms. The property of chirality has profound effects in physics, chemistry and biology, ranging from parity violations for weak forces, to the exclusive use of one mirror-form of amino acids by all life-forms on earth. Turning specifically to phenomena in two dimensions, chiral surfaces offer fascinating possibilities in the technological fields of non-linear optical materials, heterogeneous enantioselective catalysis and sensor devices. Yet, the fundamental aspects of constructing two-dimensional chirality is only just beginning to be understood with the aid of sophisticated surface science techniques [1–15].
0169-4332/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2004.09.058
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
A particularly successful way of endowing nonchiral metal surfaces with chirality is via the adsorption of complex organic molecules, which provide a more sophisticated selectivity and asymmetrisation function to be introduced. This paper discusses the creation of surface chirality via adsorption of organic molecules in two systems: one, utilising (R,R)-tartaric acid, in which the chirality function is inherently present in the adsorbing organic molecule; and, the other, utilising succinic acid, in which chirality is spontaneously created as a direct result of adsorption. This allows a neat comparison and contrast of the underlying forces and mechanisms that generate chirality at surfaces. Succinic acid is a very similar molecule to tartaric acid, with the only difference being that the two hydroxyl groups present in tartaric acid are replaced by hydrogen atoms leading to the loss of both chiral centres (Figs. 1 and 2). To compare and contrast these two systems, the surfaces were probed using three main techniques: reflection absorption infrared spectroscopy (RAIRS), low energy electron diffraction (LEED) and scanning tunnelling microscopy (STM). Each technique provides complementary information on the interface. The high sensitivity, high resolution vibrational data obtained by Fourier-transform RAIRS provides direct information on the chemical nature of the adsorbed molecules, their perturbation by the surface, and their orientation at the surface. LEED monitors the longrange two-dimensional periodicity of the adlayer while STM provides information on the local arrangements of the molecules within the domains formed at the surface.
2. Experimental Experiments were carried out in an Omicron Vakuumphysik variable temperature-STM chamber with facilities for STM, LEED, AES and sample cleaning. All STM experiments were carried out with the sample at room temperature. The images were acquired in constant current mode. Complementary RAIRS experiments (not presented here) identified the nature of the adsorbed species imaged by STM. The Cu(110) crystal was cleaned by cycles of Ar+ ion sputtering, flashing and annealing to 700 K. The surface ordering and cleanliness were monitored by
151
LEED and AES. Succinic acid (99%)and (R,R)tartaric acid (99%) were obtained from Sigma-Aldrich and used without further purification. The adsorbate was introduced into the chamber via sublimation from a Knudsen cell. During sublimation the main chamber pressure was typically 2 109 mbar. Adsorbate coverage at the surface is given in terms of fractional monolayers (ML), quoted with respect to the number density of surface metal atoms. The adlayer unit mesh is given in standard matrix notation and quoted in the text as (G11 G12, G21 G22): ao G11 G12 as ¼ bo G21 G22 bs where ao, bo are the overlayer unit vectors, as, bs the substrate unit vectors and jasj < jbsj with the vectors all defined by a right-handed axis system.
3. Results and discussion There are two major manifestations of chirality at surfaces. The first, leading to the creation of chiral motifs at surfaces, arises because adsorption of the molecule essentially creates a molecule–metal complex which locally destroys all mirror planes. Such point chirality is the most basic manifestation of chirality and leads to a local chiral motif that is restricted to possessing the point group symmetries C1, C2, C3, C4 or C6. Alternatively, chirality arises when adsorbates assemble into clusters, 1D chains or ordered 2D structures where the self-assembly of individual motifs destroys the reflection symmetry planes of the underlying surface, leading to the creation of organisational chirality. Such chirally ordered domains belong to one of the five possible chiral space groups, C1, C2, C3, C4, C6, that can exist at a surface. Here, we discuss these two aspects for the bicarboxylate phases of (R,R)-tartaric acid (the R,Rbitartrate) and succinic acid (the bisuccinate) on Cu(110). Detailed RAIR spectroscopic data [2,4] for these phases show that the doubly deprotonated form of the molecules is thermodynamically preferred on Cu(110) at low coverages. Application of the RAIRS selection rule also allows one to determine that both carboxylate ends of the molecule are involved in bonding to the surface and that the two oxygen atoms
152
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
Fig. 1. Depiction of the (1 2, 9 0) and (9 0, 1 2) bitartrate phases formed for the adsorption of tartaric acid on Cu(110): (a and b) STM images ˚ 108 A ˚ ) [V = 1.7 V; I = 1.18 nA] showing the (R,R)-bitartrate (1 2, 9 0) phase and (108 A ˚ 108 A ˚ ) [V = 2.73 V; I = 1.02 nA] (108 A showing the (S,S)-bitartrate (9 0, 1 2) phase; (c and d) Structural models for each of the tartrate phases: (R,R)-bitartrate (9 0, 1 2) and (S,S)-bitartrate (1 2, 9 0).
in each COO unit are held almost equidistant from the surface, leaving the C2–C3 bond almost parallel to the surface and yielding a fairly rigid adsorption geometry. This general geometry has recently been confirmed by periodic density functional theory calculations for both bitartrate and bisuccinate on
Cu(110) [14]. These calculations also show that, in both cases, the molecule is adsorbed across the longbridge site, and that the carboxylate oxygens of the bitartrate and the bisuccinate occupy the on-top Cu positions across the two-fold short bridge site (Figs. 1 and 2).
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
153
˚ 170 A ˚ ) [V = 0.21 V; I = 0.15 nA] showing the coexistence of the Fig. 2. Depiction of the bisuccinate phases: (a) STM images (260 A (2 2, 9 0) and (9 0, 2 2) phases; (b) structural models are presented for each of the phases. The (2 2, 9 0) and (9 0, 2 2) unit cells of the overall structure are shown on the top part, while the lower part presents the model that could lead to the (1 1, 9 0) and (9 0, 1 1) periodicities by molecular distortions of the bisuccinate adsorbate.
The two-dimensional nature of these two bicarboxylate structures can be constructed from a comprehensive analysis of the LEED and STM data. The (R,R)-bitartrate phase [1] possesses a (1 2, 9 0) structure in which the very large unit cell has the ˚ 7.68 A ˚ , a = 19.478. High resodimensions 23.04 A
lution STM images, displayed in Fig. 1(a), show that there are three bitartrate molecules per unit cell, resulting in a fractional coverage of 1/6 ML. The STM images also reveal that rows of three bitartrate molecules assemble at the surface to form long chains, which are aligned along the [1 1 4] surface
154
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
direction. Significantly, this structure annihilates both reflection symmetry planes of the underlying Cu(110) and, as a result, creates a system where chirality is present at both the molecular adsorption motif (point group) level and at the organisational (space group) level. Perhaps the most important aspect of this phase is that the same growth directions and arrangements are maintained over the entire surface and, as a result, truly global organisational chirality is expressed where no reflection domains of the opposite chirality are tolerated. In fact, the chirality of the system can only be switched by creating the same phase but using the twin enantiomer, (S,S)-tartaric acid, which leads to a perfect image chiral surface in which all molecular positions and growth directions are reflected in space [1], Fig. 1(b). STM images from the bisuccinate phase, Fig. 2(a), clearly show that it produces two domains at the surface. Detailed structural models on the organisation of adsorbates show that each domain consists of rows of three bisuccinate molecules that assemble at the surface to form long chains. This structure, with unit ˚ 8.86 A ˚ with a = 35.38, cell dimensions of 23.04 A also possesses a chiral organisation which destroys the mirror symmetry elements of the Cu(110) surface. Importantly, the overall similarity in the trimer chain structures adopted by both the bitartrate and the bisuccinate phases, suggests that it is the molecule– metal interactions that determine the general nature of the self-organisation and that the presence/absence of the chiral OH groups does not affect the overall type of supramolecular assembly that is adopted. However, a closer inspection reveals the two systems show a divergence of behaviour. First, there is a slight difference in the growth directions and the molecular packing adopted by the bitartrate and bisuccinate systems. This is essentially a second-order difference and arises from subtle changes in molecule–metal and molecule–molecule interactions in the absence/presence of the OH groups. A more detailed understanding requires development of current density functional theory (DFT) codes to allow these large adsorption structures on extended surfaces to be represented accurately. The second difference that arises is critical from the viewpoint of chirality at surfaces. From the adsorption models presented in Figs. 1 and 2, it is clear that the manifestation of organisational surface chirality is not just restricted to
the chiral bitartrate molecule, but also extends to the achiral bisuccinate. However, here the pertinent difference between the two systems emerges: whereas for tartaric acid only one chiral handedness exists which is sustained over the entire surface, for succinic acid both the chiral (9 0, 2 2) phase and its mirror image (2 2, 9 0) phase coexist at the surface as shown by STM data, Fig. 2(a), so that integrating over the entire surface one obtains an overall racemic conglomerate. This is a crucial difference in the expression of chirality between the two systems in that the bitartrate possesses global chirality while the bisuccinate is locally chiral but globally non-chiral. The reasons underlying this difference in behaviour can be surmised from considering the bisuccinate system. The succinic acid molecule possesses no inherent chirality and, therefore, any chiral behaviour it displays is adsorption induced. Important clues on this spontaneous creation of chirality in the bisuccinate phase comes from the LEED data which show diffraction patterns with a slightly different repeat structure to that picked up by the STM experiment, namely a (1 1, 9 0) and (9 0, 1 1) structure with a halving of the repeat distance in the [1 1 2] and [1– 12] directions, respectively. Such a result could be attributed to one of two possibilities. First, the LEED scattering could be dominated by the bonding carboxylate groups, with each molecule contributing two scattering centers. In fact, DFT calculations on this system [14] show that the molecule straddles straight across the close-packed rows with an attendant deformation of the molecular skeleton, as shown in Fig. 3(d). Such a deformation would naturally lead to inequivalence within each COO bonding unit, halving the LEED scattering repeat. Another possibility is that the LEED scattering is dominated, not by the bonding carboxylate units, but by reconstructions of the underlying metal. It is known that the succinate and tartrate structures possess significant surface interaction energies, calculated at 166 and 164 kJ mol1, respectively [14], and are stable beyond 480 K at the surface, suggesting strong enough metal–molecule interactions that could trigger reconstructions. In addition, DFT calculations [14] show that adsorption of both molecules lead to a compressive strain along the [1–10] direction, with ˚ which is both preferring a Cu–Cu distance of 2.63 A greater than the bulk truncation close-packed Cu–Cu
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
155
Fig. 3. (a) Distorted skeleton of the (R,R)-bitartrate molecule and the local reconstruction of the Ni atoms to describe an oblique chiral footprint, for the (R,R)-bitartrate/Ni(110) system; (b and c) depiction of the (R,R)-bitartrate species adsorbed with mirror chiral distortions and reconstructions at the Ni(110) surface; (d and e) depiction of the bicarboxylate species with a skeletal chiral distortion adsorbed on Cu(110). Note that the energy degeneracy of the two chiral pairs is preserved for the bisuccinate but broken for the bitartrate; (f and g) depiction of the distorted bicarboxylate species adsorbed on a chirally reconstructed Cu(110) surface. Note that the energy degeneracy of the two chiral pairs is preserved for the bisuccinate but broken for the bitartrate.
distance. This strain could be relieved by lateral relaxations of the underlying Cu surface as, for example, depicted in Fig. 3(f and g), which would naturally lead to halving of periodicity in the required direction. For both scenarios, we note that the halving of the periodicity along the [1 1 2] and [1–12] directions requires the deformation and/or reconstruction to be essentially chiral so that the two diagonal Cu–O1 and Cu–O4 units are equivalent, and differ from the other two Cu–O2 and Cu–O3 units comprising the opposite diagonal. Chiral reconstructions and deformations of this type which force the adsorbate to describe a chiral footprint at the surface have been reported for the bitartrate/Ni(110) system [3,15] where the adsorbed bitartrate molecule forces the bonding metal atoms to twist out of their rectangular unit mesh into a chiral oblique arrangement, Fig. 3(a). At present, the two types of structural models shown in Fig. 3(d–g) cannot be differentiated on the basis of our data. However, the LEED and STM data clearly indicate that the halving of the molecular repeat distance can only arise if asymmetry is
introduced for the local adsorption unit, i.e. the achiral bisuccinate adopts a local chiral adsorption motif. Hence, the spontaneous creation of chirality is intimately related to the nature of the metal–molecule bonding and, it follows from the general similarity of the trimer chain structure adopted by both molecules, that a similar influence must also be present when the bitartrate is adsorbed. The above discussion also explains why the bitartrate system exhibits global chirality, while the bisuccinate can only maintain chirality at the local domain level. For the bisuccinate system, the achirality of the adsorbing molecule makes the nucleation of point chirality of one-handedness (via a chiral molecular distortion/reconstruction) energetically degenerate to creating point chirality of the opposite handedness (via a mirror chiral molecular distortion/reconstruction), so random adsorption processes will produce equal numbers of each from which local chiral domains will be propagated. Thus, the succinate system can only maintain chirality at the local level, but is globally a racemic conglomerate.
156
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
For the bitartrate adsorption, the same type of chiral molecular distortion/reconstruction is expected to arise from essentially the same type of molecule– metal interaction. However, here, the rigid adsorption of the bitartrate unit in which both carboxylate functionalities are bonded to the surface forces the chiral OH groups to adopt a uniquely defined direction with respect to the surface. This breaks the degeneracy of one chiral molecular distortion/reconstruction visa-vis its mirror chiral molecular distortion/reconstruction. This is clearly illustrated in DFT calculations of the R,R-bitartrate/Ni(110) system [3,15] where an energy difference of 7 kJmol1 was found to exist between one chiral adsorption distortion and its mirror distortion, Fig. 3(b and c), which would ensure that at room temperature over 90% of nucleation points would adopt the favoured chirality. On Cu(110), the role of the OH groups in breaking this degeneracy is also clearly illustrated when adsorption of the (S,S)bitartrate unit is examined, Fig. 1(b). Here, the rigid adsorption structure of the (S,S)-bitartrate unit again forces the OH groups to lie in a uniquely defined direction, but one which is reflected in space compared to the (R,R)-bitartrate unit. As a result, the mirror local adsorption motif is preferred, leading to the propagation of the mirror domain across the entire surface.
4. Conclusions A detailed comparison of the adsorption of the chiral bitartrate and the achiral bisuccinate on Cu(110) shows a similar trimer chain structure which adopts a non-symmetry growth direction that destroys the mirror planes associated with the Cu(110) surface, i.e. both systems give rise to two-dimensional organizational chirality. Overall, this general similarity points to the dominance of molecule–metal bonding in dictating the overall type of superstructures adopted. STM and LEED data show that the creation of chirality from the achiral succinate arises from adsorption-induced asymmetrisation via molecular distortion and/or local reconstruction that creates point chirality. This chiral adsorption unit then acts as a nucleation point for generating the chiral organization. Since the succinate is inherently achiral,
nucleation points of either chirality are generated with equal probability, so overall a racemic conglomerate is created. For the (R,R)-bitartrate system, the presence of the OH groups at the molecular chiral centers breaks this degeneracy, and forces the molecular distortion/reconstruction to adopt a favored chirality, which then generates one favored chiral domain only, bestowing the system with global chirality at the macroscopic scale. For the (S,S)enantiomer, the OH group alignment is reflectionally flipped and the mirror distortion/reconstruction is created, generating the mirror chiral organization.
Acknowledgements It is a great pleasure to acknowledge the substantial contributions of our co-researchers, Maria OrtegaLorenzo, Sam Haq, Chris Baddeley, Chris Muryn and Paul Murray, have made to this research effort.
References [1] M. Ortega Lorenzo, C.J. Baddeley, C. Muryn, R. Raval, Nature 404 (2000) 376. [2] M. Ortega Lorenzo, S. Haq, T. Bertrams, P. Murray, R. Raval, C.J. Baddeley, J. Phys. Chem. B 103 (1999) 10661. [3] V. Humblot, S. Haq, C. Muryn, W. Hofer, R. Raval, J. Am. Chem. Soc. 124 (2002) 503. [4] S.M. Barlow, R. Raval, Surf. Sci. Rep. 50 (2003) 201. [5] J. Weckesser, J.V. Barth, C. Cai, B. Muller, K. Kern, Surf. Sci. 431 (1999) 168. [6] M. Schunack, L. Petersen, A. Ku¨ hlne, E. Laegsgaard, I. Stensgaard, I. Johannsen, F. Besenbacher, Phys. Rev. Lett. 86 (2001) 456. [7] X. Zhao, J. Am. Chem. Soc. 122 (2000) 12584. [8] T.E. Jones, C.J. Baddeley, Surf. Sci. 513 (2002) 453. [9] A. Kuhnle, T.R. Linderoth, B. Hammer, F. Besenbacher, Nature 415 (2002) 891. [10] R. Fasel, M. Parschau, K.-H. Ernst, Angew. Chem. Int. Ed. 42 (2003) 5177. [11] R. Raval, Curr. Opin. Solid State Mater. Sci. 7 (2003) 67. [12] M. Ortega Lorenzo, V. Humblot, P. Murray, C.J. Baddeley, S. Haq, R. Raval, J. Catal. 205 (2002) 123. [13] R. Raval, CATTECH 5 (2001) 12. [14] L.A.M.M. Barbosa, P. Sautet, J. Am. Chem. Soc. 123 (2001) 6639. [15] W. Hofer, V. Humblot, R. Raval, Surf. Sci. 554 (2004) 141.
Chiral metal surfaces from the adsorption of chiral and achiral molecules V. Humblot, R. Raval* Department of Chemistry, Surface Science Research Centre, University of Liverpool, Liverpool L69 3BX, UK Available online 11 November 2004
Abstract Chiral surfaces, capable of existing in two distinguishable mirror forms that cannot be superimposed, are attracting worldwide attention. The adsorption of complex organic molecules provides a means of introducing the ultimate discrimination function of chirality to a metal surface. Here, a comparison of the chiral tartaric acid (HOOC–CHOH–CHOH–COOH) molecule and the achiral succinic acid (HOOC–CH2–CH2–COOH) molecule on a Cu(110) surface is presented. For both molecules, twodimensional assembly is found to depend strongly on molecule–metal bonding interactions, whereas the presence/absence of the OH groups causes subtler, second-order effects on the self-assembled structure. The driving force for creating chiral organisations is shown to arise from adsorption-induced asymmetrisation, via molecular distortion and/or metal reconstruction of the local adsorption unit. The macroscopic chirality of the surface is then determined by whether nucleation points of both chirality can be equally created, or whether non-degeneracy can be introduced to favour one chirality. # 2004 Elsevier B.V. All rights reserved. PACS: 68.43.h; 87.15.By Keywords: Chirality; Surfaces; Chemisorption; Carboxylates
1. Introduction Chirality is a simple geometric property that dictates that the mirror transformation of an object that does not possess any inverse symmetry elements, is a non-identity operation, i.e. the object and its mirror image are non-superimposable. As a result, a chiral object can exist in two distinguishable mirror, or * Corresponding author. Tel.: +44 151 794 6891; fax.: +44 151 794 3896. E-mail address: [email protected] (R. Raval).
enantiomeric, forms. The property of chirality has profound effects in physics, chemistry and biology, ranging from parity violations for weak forces, to the exclusive use of one mirror-form of amino acids by all life-forms on earth. Turning specifically to phenomena in two dimensions, chiral surfaces offer fascinating possibilities in the technological fields of non-linear optical materials, heterogeneous enantioselective catalysis and sensor devices. Yet, the fundamental aspects of constructing two-dimensional chirality is only just beginning to be understood with the aid of sophisticated surface science techniques [1–15].
0169-4332/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2004.09.058
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
A particularly successful way of endowing nonchiral metal surfaces with chirality is via the adsorption of complex organic molecules, which provide a more sophisticated selectivity and asymmetrisation function to be introduced. This paper discusses the creation of surface chirality via adsorption of organic molecules in two systems: one, utilising (R,R)-tartaric acid, in which the chirality function is inherently present in the adsorbing organic molecule; and, the other, utilising succinic acid, in which chirality is spontaneously created as a direct result of adsorption. This allows a neat comparison and contrast of the underlying forces and mechanisms that generate chirality at surfaces. Succinic acid is a very similar molecule to tartaric acid, with the only difference being that the two hydroxyl groups present in tartaric acid are replaced by hydrogen atoms leading to the loss of both chiral centres (Figs. 1 and 2). To compare and contrast these two systems, the surfaces were probed using three main techniques: reflection absorption infrared spectroscopy (RAIRS), low energy electron diffraction (LEED) and scanning tunnelling microscopy (STM). Each technique provides complementary information on the interface. The high sensitivity, high resolution vibrational data obtained by Fourier-transform RAIRS provides direct information on the chemical nature of the adsorbed molecules, their perturbation by the surface, and their orientation at the surface. LEED monitors the longrange two-dimensional periodicity of the adlayer while STM provides information on the local arrangements of the molecules within the domains formed at the surface.
2. Experimental Experiments were carried out in an Omicron Vakuumphysik variable temperature-STM chamber with facilities for STM, LEED, AES and sample cleaning. All STM experiments were carried out with the sample at room temperature. The images were acquired in constant current mode. Complementary RAIRS experiments (not presented here) identified the nature of the adsorbed species imaged by STM. The Cu(110) crystal was cleaned by cycles of Ar+ ion sputtering, flashing and annealing to 700 K. The surface ordering and cleanliness were monitored by
151
LEED and AES. Succinic acid (99%)and (R,R)tartaric acid (99%) were obtained from Sigma-Aldrich and used without further purification. The adsorbate was introduced into the chamber via sublimation from a Knudsen cell. During sublimation the main chamber pressure was typically 2 109 mbar. Adsorbate coverage at the surface is given in terms of fractional monolayers (ML), quoted with respect to the number density of surface metal atoms. The adlayer unit mesh is given in standard matrix notation and quoted in the text as (G11 G12, G21 G22): ao G11 G12 as ¼ bo G21 G22 bs where ao, bo are the overlayer unit vectors, as, bs the substrate unit vectors and jasj < jbsj with the vectors all defined by a right-handed axis system.
3. Results and discussion There are two major manifestations of chirality at surfaces. The first, leading to the creation of chiral motifs at surfaces, arises because adsorption of the molecule essentially creates a molecule–metal complex which locally destroys all mirror planes. Such point chirality is the most basic manifestation of chirality and leads to a local chiral motif that is restricted to possessing the point group symmetries C1, C2, C3, C4 or C6. Alternatively, chirality arises when adsorbates assemble into clusters, 1D chains or ordered 2D structures where the self-assembly of individual motifs destroys the reflection symmetry planes of the underlying surface, leading to the creation of organisational chirality. Such chirally ordered domains belong to one of the five possible chiral space groups, C1, C2, C3, C4, C6, that can exist at a surface. Here, we discuss these two aspects for the bicarboxylate phases of (R,R)-tartaric acid (the R,Rbitartrate) and succinic acid (the bisuccinate) on Cu(110). Detailed RAIR spectroscopic data [2,4] for these phases show that the doubly deprotonated form of the molecules is thermodynamically preferred on Cu(110) at low coverages. Application of the RAIRS selection rule also allows one to determine that both carboxylate ends of the molecule are involved in bonding to the surface and that the two oxygen atoms
152
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
Fig. 1. Depiction of the (1 2, 9 0) and (9 0, 1 2) bitartrate phases formed for the adsorption of tartaric acid on Cu(110): (a and b) STM images ˚ 108 A ˚ ) [V = 1.7 V; I = 1.18 nA] showing the (R,R)-bitartrate (1 2, 9 0) phase and (108 A ˚ 108 A ˚ ) [V = 2.73 V; I = 1.02 nA] (108 A showing the (S,S)-bitartrate (9 0, 1 2) phase; (c and d) Structural models for each of the tartrate phases: (R,R)-bitartrate (9 0, 1 2) and (S,S)-bitartrate (1 2, 9 0).
in each COO unit are held almost equidistant from the surface, leaving the C2–C3 bond almost parallel to the surface and yielding a fairly rigid adsorption geometry. This general geometry has recently been confirmed by periodic density functional theory calculations for both bitartrate and bisuccinate on
Cu(110) [14]. These calculations also show that, in both cases, the molecule is adsorbed across the longbridge site, and that the carboxylate oxygens of the bitartrate and the bisuccinate occupy the on-top Cu positions across the two-fold short bridge site (Figs. 1 and 2).
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
153
˚ 170 A ˚ ) [V = 0.21 V; I = 0.15 nA] showing the coexistence of the Fig. 2. Depiction of the bisuccinate phases: (a) STM images (260 A (2 2, 9 0) and (9 0, 2 2) phases; (b) structural models are presented for each of the phases. The (2 2, 9 0) and (9 0, 2 2) unit cells of the overall structure are shown on the top part, while the lower part presents the model that could lead to the (1 1, 9 0) and (9 0, 1 1) periodicities by molecular distortions of the bisuccinate adsorbate.
The two-dimensional nature of these two bicarboxylate structures can be constructed from a comprehensive analysis of the LEED and STM data. The (R,R)-bitartrate phase [1] possesses a (1 2, 9 0) structure in which the very large unit cell has the ˚ 7.68 A ˚ , a = 19.478. High resodimensions 23.04 A
lution STM images, displayed in Fig. 1(a), show that there are three bitartrate molecules per unit cell, resulting in a fractional coverage of 1/6 ML. The STM images also reveal that rows of three bitartrate molecules assemble at the surface to form long chains, which are aligned along the [1 1 4] surface
154
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
direction. Significantly, this structure annihilates both reflection symmetry planes of the underlying Cu(110) and, as a result, creates a system where chirality is present at both the molecular adsorption motif (point group) level and at the organisational (space group) level. Perhaps the most important aspect of this phase is that the same growth directions and arrangements are maintained over the entire surface and, as a result, truly global organisational chirality is expressed where no reflection domains of the opposite chirality are tolerated. In fact, the chirality of the system can only be switched by creating the same phase but using the twin enantiomer, (S,S)-tartaric acid, which leads to a perfect image chiral surface in which all molecular positions and growth directions are reflected in space [1], Fig. 1(b). STM images from the bisuccinate phase, Fig. 2(a), clearly show that it produces two domains at the surface. Detailed structural models on the organisation of adsorbates show that each domain consists of rows of three bisuccinate molecules that assemble at the surface to form long chains. This structure, with unit ˚ 8.86 A ˚ with a = 35.38, cell dimensions of 23.04 A also possesses a chiral organisation which destroys the mirror symmetry elements of the Cu(110) surface. Importantly, the overall similarity in the trimer chain structures adopted by both the bitartrate and the bisuccinate phases, suggests that it is the molecule– metal interactions that determine the general nature of the self-organisation and that the presence/absence of the chiral OH groups does not affect the overall type of supramolecular assembly that is adopted. However, a closer inspection reveals the two systems show a divergence of behaviour. First, there is a slight difference in the growth directions and the molecular packing adopted by the bitartrate and bisuccinate systems. This is essentially a second-order difference and arises from subtle changes in molecule–metal and molecule–molecule interactions in the absence/presence of the OH groups. A more detailed understanding requires development of current density functional theory (DFT) codes to allow these large adsorption structures on extended surfaces to be represented accurately. The second difference that arises is critical from the viewpoint of chirality at surfaces. From the adsorption models presented in Figs. 1 and 2, it is clear that the manifestation of organisational surface chirality is not just restricted to
the chiral bitartrate molecule, but also extends to the achiral bisuccinate. However, here the pertinent difference between the two systems emerges: whereas for tartaric acid only one chiral handedness exists which is sustained over the entire surface, for succinic acid both the chiral (9 0, 2 2) phase and its mirror image (2 2, 9 0) phase coexist at the surface as shown by STM data, Fig. 2(a), so that integrating over the entire surface one obtains an overall racemic conglomerate. This is a crucial difference in the expression of chirality between the two systems in that the bitartrate possesses global chirality while the bisuccinate is locally chiral but globally non-chiral. The reasons underlying this difference in behaviour can be surmised from considering the bisuccinate system. The succinic acid molecule possesses no inherent chirality and, therefore, any chiral behaviour it displays is adsorption induced. Important clues on this spontaneous creation of chirality in the bisuccinate phase comes from the LEED data which show diffraction patterns with a slightly different repeat structure to that picked up by the STM experiment, namely a (1 1, 9 0) and (9 0, 1 1) structure with a halving of the repeat distance in the [1 1 2] and [1– 12] directions, respectively. Such a result could be attributed to one of two possibilities. First, the LEED scattering could be dominated by the bonding carboxylate groups, with each molecule contributing two scattering centers. In fact, DFT calculations on this system [14] show that the molecule straddles straight across the close-packed rows with an attendant deformation of the molecular skeleton, as shown in Fig. 3(d). Such a deformation would naturally lead to inequivalence within each COO bonding unit, halving the LEED scattering repeat. Another possibility is that the LEED scattering is dominated, not by the bonding carboxylate units, but by reconstructions of the underlying metal. It is known that the succinate and tartrate structures possess significant surface interaction energies, calculated at 166 and 164 kJ mol1, respectively [14], and are stable beyond 480 K at the surface, suggesting strong enough metal–molecule interactions that could trigger reconstructions. In addition, DFT calculations [14] show that adsorption of both molecules lead to a compressive strain along the [1–10] direction, with ˚ which is both preferring a Cu–Cu distance of 2.63 A greater than the bulk truncation close-packed Cu–Cu
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
155
Fig. 3. (a) Distorted skeleton of the (R,R)-bitartrate molecule and the local reconstruction of the Ni atoms to describe an oblique chiral footprint, for the (R,R)-bitartrate/Ni(110) system; (b and c) depiction of the (R,R)-bitartrate species adsorbed with mirror chiral distortions and reconstructions at the Ni(110) surface; (d and e) depiction of the bicarboxylate species with a skeletal chiral distortion adsorbed on Cu(110). Note that the energy degeneracy of the two chiral pairs is preserved for the bisuccinate but broken for the bitartrate; (f and g) depiction of the distorted bicarboxylate species adsorbed on a chirally reconstructed Cu(110) surface. Note that the energy degeneracy of the two chiral pairs is preserved for the bisuccinate but broken for the bitartrate.
distance. This strain could be relieved by lateral relaxations of the underlying Cu surface as, for example, depicted in Fig. 3(f and g), which would naturally lead to halving of periodicity in the required direction. For both scenarios, we note that the halving of the periodicity along the [1 1 2] and [1–12] directions requires the deformation and/or reconstruction to be essentially chiral so that the two diagonal Cu–O1 and Cu–O4 units are equivalent, and differ from the other two Cu–O2 and Cu–O3 units comprising the opposite diagonal. Chiral reconstructions and deformations of this type which force the adsorbate to describe a chiral footprint at the surface have been reported for the bitartrate/Ni(110) system [3,15] where the adsorbed bitartrate molecule forces the bonding metal atoms to twist out of their rectangular unit mesh into a chiral oblique arrangement, Fig. 3(a). At present, the two types of structural models shown in Fig. 3(d–g) cannot be differentiated on the basis of our data. However, the LEED and STM data clearly indicate that the halving of the molecular repeat distance can only arise if asymmetry is
introduced for the local adsorption unit, i.e. the achiral bisuccinate adopts a local chiral adsorption motif. Hence, the spontaneous creation of chirality is intimately related to the nature of the metal–molecule bonding and, it follows from the general similarity of the trimer chain structure adopted by both molecules, that a similar influence must also be present when the bitartrate is adsorbed. The above discussion also explains why the bitartrate system exhibits global chirality, while the bisuccinate can only maintain chirality at the local domain level. For the bisuccinate system, the achirality of the adsorbing molecule makes the nucleation of point chirality of one-handedness (via a chiral molecular distortion/reconstruction) energetically degenerate to creating point chirality of the opposite handedness (via a mirror chiral molecular distortion/reconstruction), so random adsorption processes will produce equal numbers of each from which local chiral domains will be propagated. Thus, the succinate system can only maintain chirality at the local level, but is globally a racemic conglomerate.
156
V. Humblot, R. Raval / Applied Surface Science 241 (2005) 150–156
For the bitartrate adsorption, the same type of chiral molecular distortion/reconstruction is expected to arise from essentially the same type of molecule– metal interaction. However, here, the rigid adsorption of the bitartrate unit in which both carboxylate functionalities are bonded to the surface forces the chiral OH groups to adopt a uniquely defined direction with respect to the surface. This breaks the degeneracy of one chiral molecular distortion/reconstruction visa-vis its mirror chiral molecular distortion/reconstruction. This is clearly illustrated in DFT calculations of the R,R-bitartrate/Ni(110) system [3,15] where an energy difference of 7 kJmol1 was found to exist between one chiral adsorption distortion and its mirror distortion, Fig. 3(b and c), which would ensure that at room temperature over 90% of nucleation points would adopt the favoured chirality. On Cu(110), the role of the OH groups in breaking this degeneracy is also clearly illustrated when adsorption of the (S,S)bitartrate unit is examined, Fig. 1(b). Here, the rigid adsorption structure of the (S,S)-bitartrate unit again forces the OH groups to lie in a uniquely defined direction, but one which is reflected in space compared to the (R,R)-bitartrate unit. As a result, the mirror local adsorption motif is preferred, leading to the propagation of the mirror domain across the entire surface.
4. Conclusions A detailed comparison of the adsorption of the chiral bitartrate and the achiral bisuccinate on Cu(110) shows a similar trimer chain structure which adopts a non-symmetry growth direction that destroys the mirror planes associated with the Cu(110) surface, i.e. both systems give rise to two-dimensional organizational chirality. Overall, this general similarity points to the dominance of molecule–metal bonding in dictating the overall type of superstructures adopted. STM and LEED data show that the creation of chirality from the achiral succinate arises from adsorption-induced asymmetrisation via molecular distortion and/or local reconstruction that creates point chirality. This chiral adsorption unit then acts as a nucleation point for generating the chiral organization. Since the succinate is inherently achiral,
nucleation points of either chirality are generated with equal probability, so overall a racemic conglomerate is created. For the (R,R)-bitartrate system, the presence of the OH groups at the molecular chiral centers breaks this degeneracy, and forces the molecular distortion/reconstruction to adopt a favored chirality, which then generates one favored chiral domain only, bestowing the system with global chirality at the macroscopic scale. For the (S,S)enantiomer, the OH group alignment is reflectionally flipped and the mirror distortion/reconstruction is created, generating the mirror chiral organization.
Acknowledgements It is a great pleasure to acknowledge the substantial contributions of our co-researchers, Maria OrtegaLorenzo, Sam Haq, Chris Baddeley, Chris Muryn and Paul Murray, have made to this research effort.
References [1] M. Ortega Lorenzo, C.J. Baddeley, C. Muryn, R. Raval, Nature 404 (2000) 376. [2] M. Ortega Lorenzo, S. Haq, T. Bertrams, P. Murray, R. Raval, C.J. Baddeley, J. Phys. Chem. B 103 (1999) 10661. [3] V. Humblot, S. Haq, C. Muryn, W. Hofer, R. Raval, J. Am. Chem. Soc. 124 (2002) 503. [4] S.M. Barlow, R. Raval, Surf. Sci. Rep. 50 (2003) 201. [5] J. Weckesser, J.V. Barth, C. Cai, B. Muller, K. Kern, Surf. Sci. 431 (1999) 168. [6] M. Schunack, L. Petersen, A. Ku¨ hlne, E. Laegsgaard, I. Stensgaard, I. Johannsen, F. Besenbacher, Phys. Rev. Lett. 86 (2001) 456. [7] X. Zhao, J. Am. Chem. Soc. 122 (2000) 12584. [8] T.E. Jones, C.J. Baddeley, Surf. Sci. 513 (2002) 453. [9] A. Kuhnle, T.R. Linderoth, B. Hammer, F. Besenbacher, Nature 415 (2002) 891. [10] R. Fasel, M. Parschau, K.-H. Ernst, Angew. Chem. Int. Ed. 42 (2003) 5177. [11] R. Raval, Curr. Opin. Solid State Mater. Sci. 7 (2003) 67. [12] M. Ortega Lorenzo, V. Humblot, P. Murray, C.J. Baddeley, S. Haq, R. Raval, J. Catal. 205 (2002) 123. [13] R. Raval, CATTECH 5 (2001) 12. [14] L.A.M.M. Barbosa, P. Sautet, J. Am. Chem. Soc. 123 (2001) 6639. [15] W. Hofer, V. Humblot, R. Raval, Surf. Sci. 554 (2004) 141.