2003_barlow and raval_complex organic molecules at metal surfaces bonding, organisation and...

8/12/2019 2003_Barlow and Raval_Complex Organic Molecules at Metal Surfaces Bonding, Organisation and Chirality

1/141

Complex organic molecules at metal surfaces:bonding, organisation and chirality

S.M. Barlow, R. Raval*

Department of Chemistry, Surface Science Research Centre and Leverhulme Centre for

Innovative Catalysis, University of Liverpool, Liverpool L69 7ZD, UKReceived in final form 15 January 2003

Abstract

Surface science techniques have now reached a stage of maturity that has enabled their successful deployment in

the study of complex adsorption systems. A particular example of this success has been the understanding that has

been gained regarding the behaviour of multi-functional organic molecules at metal surfaces. These organicmetal

systems show enormous diversity, starting from their local description which can vary in terms of chemical structure,

orientation and bonding. Additionally, in many cases, these complex organic molecules self-organise into beautiful,

ordered superstructures held together by networks of intermolecular bonds. Both these aspects enable a single

organic moleculemetal system to exhibit a wide-ranging and flexible approach to its environment, leading to avariety of adsorption phases, according to the prevailing temperature and coverage conditions. In this review we

have attempted to capture this complexity by constructing adsorption phase diagrams from the available literature

for complex carboxylic acids, amino acids, anhydrides and ring systems, all deposited under controlled conditions

onto defined metal surfaces. These provide an accessible, pictorial basis of the adsorption phases which are then

discussed further in the text of the review. Finally, interest has recently focused on the property of chirality that can

be bestowed at an achiral metal surface by the adsorption of these complex organic molecules. The creation of such

architectures offers the opportunity for ultimate stereocontrol of reactions and responses at surfaces. We have,

therefore, specifically examined the various ways in which chirality can be expressed at a surface and provide a

framework for classifying chiral hierarchies that are manifested at surfaces, with particular attention being paid to

the progression of chirality from a local to a global level.

# 2003 Elsevier B.V. All rights reserved.

Surface Science Reports 50 (2003) 201341

0167-5729/$ see front matter # 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0167-5729(03)00015-3

Abbreviations: AES, Auger electron spectroscopy; DFT, density functional theory; EELS, electron energy loss

spectroscopy; FTIR, Fourier transform infrared spectroscopy; HREELS, high resolution electron energy loss spectroscopy;

LEED, low energy electron diffraction; NEXAFS, near-edge extended absorption fine structure spectroscopy; PhD,

photoelectron diffraction; RAIRS, reflection absorption infrared spectroscopy; RAS, reflection anisotropy spectroscopy; STM,

scanning tunnelling microscopy; ToF-SIMS, time of flight secondary ion mass spectroscopy; TPD, thermal desorption

spectroscopy; UHV, ultra-high vacuum; UPS, ultraviolet photoelectron spectroscopy; XPD, X-ray photoelectron diffraction;

XPS, X-ray photoelectron spectroscopy; ML, monolayer; Sat. ML, saturated monolayer; y, coverage; R, rectus; S, sinister* Corresponding author. Tel.:44-151-794-6891; fax:44-151-794-3896.

E-mail address: [email protected] (R. Raval).


2/141

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2041.1. Complex adsorption phase diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

1.2. Manifestations of chirality at the organic/inorganic interface . . . . . . . . . . . . . . . . . . . . . . . . 206

1.2.1. Surface chirality from adsorption of non-chiral molecules . . . . . . . . . . . . . . . . . . . . 207

1.2.1.1. Point chirality: adsorption-induced chiral motifs . . . . . . . . . . . . . . . . . . . . 207

1.2.1.2. Organisational chirality: adsorption-induced chirally ordered domains . . . . . 207

1.2.2. Surface chirality from adsorption of chiral molecules . . . . . . . . . . . . . . . . . . . . . . . 209

1.2.2.1. Point chirality: molecule-induced chiral motifs . . . . . . . . . . . . . . . . . . . . . 209

1.2.2.2. Organisational chirality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

1.3. Systems of study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

2. Carboxylic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

2.1. Tartaric acid (L(+)-2,3-dihydroxysuccinic acid) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

2.1.1. R,R-tartaric acid on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2112.1.1.1. Low coverage (9 0, 1 2) phase: the bitartrate assembly . . . . . . . . . . . . . . . 212

2.1.1.2. Medium coverage (4 0, 2 3) phase: the monotartrate assembly . . . . . . . . . . 213

2.1.1.3. High coverage (4 1, 2 3) phase: the dimermonomer assembly . . . . . . . . . . 215

2.1.1.4. The emergence of global organisational chirality . . . . . . . . . . . . . . . . . . . . 217

2.1.1.5. Creation of chiral spaces within the chiral surfaces . . . . . . . . . . . . . . . . . . 218

2.1.1.6. Switching global organisational chirality. . . . . . . . . . . . . . . . . . . . . . . . . . 219

2.1.1.7. Sustaining a single rotational domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

2.1.2. R,R-tartaric acid on Ni(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

2.1.2.1. The room temperature phase: adsorption stresses and chiral reconstructions . 221

2.1.2.2. The nature of electronic chiral communication from molecule to metal . . . . 225

2.1.2.3. From local chiral reconstructions to global organisational chirality . . . . . . . 2252.2. Succinic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

2.2.1. Succinic acid on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

2.3. Benzoic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

2.3.1. Benzoic acid on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

2.3.1.1. Room temperature substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

2.3.1.2. Liquid nitrogen cooled substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

2.3.2. Benzoic acid on p(2 1)O/Cu(1 1 0) and Cu(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . 2352.3.3. Benzoic acid on Ni(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

2.4. 3-Thiophene carboxylic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

2.4.1. 3-Thiophene carboxylic acid on Cu(1 1 0) and p(2 1)O/Cu(1 1 0). . . . . . . . . . . . . 2392.5. 4-Trans-2-(pyrid-4-yl-vinyl) benzoic acid (PVBA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

2.5.1. PVBA on Pd(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

2.5.2. PVBA on Cu(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

2.5.3. PVBA on Ag(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

2.6. p-Aminobenzoic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

2.6.1. p-Aminobenzoic acid on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

3. Amino acids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

3.1. Glycine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

3.1.1. Glycine on Cu(1 1 0) and Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

3.1.2. Glycine on Pt(1 1 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

3.2. Alanine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

202 S.M. Barlow, R. Raval / Surface Science Reports 50 (2003) 201341


3/141

3.2.1. Alanine on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

3.2.1.1. Adsorption at room temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

3.2.1.2. High coverage at 420 K: the (2

2, 5 3) phase . . . . . . . . . . . . . . . . . . . . . 258

3.2.1.3. High coverage at 470 K: the (32) phase . . . . . . . . . . . . . . . . . . . . . . . . 2613.3. S-proline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

3.3.1. S-proline on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

3.4. S-(or L-)norvaline, methionine and serine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

3.4.1. S-norvaline on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

3.4.2. S-methionine on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

3.4.3. S-serine on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

3.5. R- and S-phenylglycine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

3.6. Tripeptides: tri-L-alanine and tri-L-leucine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268

3.6.1. Tri-L-alanine and tri-L-leucine on Cu(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

4. Aromatic anhydrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

4.1. Phthalic anhydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

4.1.1. Phthalic anhydride on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

4.1.2. Phthalic anhydride on p(2 1)O/Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2774.2. Pyromellitic dianhydride (PMDA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277

4.2.1. PMDA on Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

4.2.2. PMDA on p(21)O/Cu(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2794.2.3. PMDA on Cu(1 1 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

4.2.4. PMDA on Pt(1 1 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

4.3. Naphthalene-1,8-dicarboxylic anhydride (NDCA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

4.3.1. NDCA on Ni(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

4.3.2. NDCA on p(22)O/Ni(1 1 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2834.4. 1,4,5,8-Naphthalene-tetracarboxylic dianhydride (NTCDA) . . . . . . . . . . . . . . . . . . . . . . . . . 284

4.4.1. NTCDA on Ag(1 1 1), Ag(1 0 0) and Ag(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . 2854.4.2. NTCDA on Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

4.4.3. NTCDA on Ni(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

4.5. Perylene-3,4,9,10-tetracarboxylic acid-3,4,9,10-dianhydride (PTCDA). . . . . . . . . . . . . . . . . . 292

4.5.1. PTCDA on Ag(1 1 1) and Ag(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

4.5.2. PTCDA on Cu(1 0 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

4.5.3. PTCDA on Ni(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

4.6. Substituted derivatives of PTCDA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300

4.6.1. PTCDI on Ni(1 1 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

4.6.2. Me-PTCDI on Ag(1 1 1) and Ag(1 1 0). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

4.6.3. Me-PTCDI on Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303

4.6.4. DPP-PTCDI on Ag(1 1 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3035. Closed ring structures without additional functional groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

5.1. Planar molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

5.1.1. Perylene on Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

5.1.2. Coronene on Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

5.1.3. Pentacene, pentacenequinine or pentacenetetrone on Cu(1 0 0) . . . . . . . . . . . . . . . . . 308

5.2. Helical molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308

5.2.1. M- and P-heptahelicene (M-[7]-helicene and P-[7]-helicene) . . . . . . . . . . . . . . . . . . 310

5.2.1.1. M-[7]- and P-[7]-helicene on Cu(1 1 1) and Cu(3 3 2). . . . . . . . . . . . . . . . 310

5.2.1.2. M-[7]- and P-[7]-helicene on Ni(1 0 0) and Ni(1 1 1) . . . . . . . . . . . . . . . . 311

5.2.2. Thioheterohelicene on Au(1 1 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

S.M. Barlow, R. Raval / Surface Science Reports 50 (2003) 201341 203


4/141

6. Closed ring structures with additional functional groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

6.1. 1-Nitronaphthalene (1-NN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316

6.2. Benzotriazole and related molecules on Cu(1 0 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

6.3. 2,5-Dimethyl-dicyanoquinonediimine (DMe-DCNQI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3236.4. Hexabutyloxytriphenylene (HBT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

7. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332

Appendix A. Matrix notation for overlayer unit cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

1. Introduction

Organic and biological molecules provide an important means of introducing complex reactivefunctionalities and architectures at a metal surface. In most cases the organic layer serves to provide a

more sophisticated activity, passivation or selectivity function than would have been possible with a

bare metal surface. This is particularly true for the selectivity function, which requires thefinesse of the

organic system to be amalgamated to the robust and, generally, reactive inorganic substrate. Such

organic functionalisation of a metal surface has important applications, e.g. in catalysis, sensors,

adhesion, corrosion inhibition, molecular recognition, optoelectronics and lithography. Within such a

technological context, it is clear that the development of future organic/inorganic interfaces is critically

dependent on establishing a fundamental understanding of the various bonding and lateral interactions

that govern the ultimate orientation, conformation and two-dimensional (2D) organisation of these

molecules at a metal surface. Only then, will the generic factors emerge with which to design complex

interfaces with intelligent and sophisticated capabilities. It is only in recent years that this issue hasbeen tackled with a rigorous surface science approach, involving controlled dosing of known molecules

on defined surfaces under ultra-clean conditions and using an armoury of techniques to probe theinterface. This report will concentrate on such research only and the vast area of self-assembled

monolayers (SAMs) created under liquid environments will not be addressed here. We also only review

complex, multifunctional organic molecules and do not report work on the simpler organic molecules

for which excellent reviews already exist [14].An important point to note is that a number of developments have enabled the field of surface

science to progress from the study of simpler adsorbates to the interrogation of bigger and more

complicated molecules. An almost trivial advance was the demonstration that these compounds, which

are largely solid at room temperature, could be effectively sublimed intact into ultra-high vacuum(UHV) chambers using Knudsen cells without compromising subsequent vacuum integrity. In

addition, more significant advances in surface spectroscopies have accompanied and underpinned this

progress. Notable amongst these are improvements in sensitivity, spectral resolution and spectral span,

which have enabled the requisite depth of data collection and analysis to be attempted. Finally,

advances in data simulation and theoretical calculations have allowed these large adsorbed species to

be modelled more effectively. What stands out from this literature review is that the complexity of the

organic/metal interface calls upon a number of complementary surface spectroscopies to be combined

in order for detailed molecular level models to be constructed. For example, local structural details

such as the chemical nature of the adsorbed species, its bonding and orientation are best obtained by



5/141

the techniques of reflection absorption infrared spectroscopy (RAIRS), electron energy lossspectroscopy (EELS), X-ray photoelectron spectroscopy (XPS), photoelectron diffraction (PhD) and

near-edge extended absorptionfine structure spectroscopy (NEXAFS). A surprising discovery of the

research in this area is the revelation that, unlike smaller organic molecules, these complex moleculespossess an extraordinary capability for self-organisation and produce remarkably ordered, crystalline-

like structures at the metal surface. The details of these 2D assemblies have been best captured by low

energy electron diffraction (LEED) and scanning tunnelling microscopy (STM) experiments. We

attribute this behaviour directly to the multiple functionalities possessed by these molecules which

enable strong lateral interactions to be expressed.

Throughout the text and figures, coverage, y, at the surface is given either in terms of fractionalmonolayers (MLs), quoted with respect to the number density of surface metal atoms, or in terms of the

saturated monolayer (Sat. ML). The overlayer unit mesh is given in real space matrix notation as

follows and quoted in the text as (m11 m12, m21 m22)

aobo

m11m12

m21m22

as

bs

where ao, bo are the overlayer net vectors and as, bs the underlying metal surface mesh vectors. It

should be noted that there is widespread confusion within the surface science community (including

on occasion ourselves), on the use of crystallographic conventions to define overlayer and substrate

vectors and crystal directions. Thus although overlayer unit cells are generally correctly determined,

the overlayer matrices are often not consistent with the recommended guidelines in International

Tables for Crystallography [5]. This is discussed more fully in the chapter on SurfaceCrystallography by Unertl in [6] and in Appendix A here, where a comparison is made between

the overlayer matrix descriptions quoted in the text (from the original articles reviewed) and thosededuced using a more consistent approach. We would like to take this chance to urge authors to use

the criteria given in Appendix Ain future work to enable comparisons between different systems to

be more readily made.

A number of phenomena emerge from our review of these systems. We wish to highlight two here:

complex adsorption phase diagrams, manifestations of chirality at the interface.Both of these aspects are discussed briefly below.

1.1. Complex adsorption phase diagrams

The work reviewed here clearly demonstrates that the adsorbed molecules are dynamic and evolving

entities, displaying different characteristics in response to conditions such as adsorption temperature

and adlayer coverage. This range of response is directly attributable to the multifunctionality of these

molecules which can lead to a number of surfacemolecule and moleculemolecule interactions. Most

of these interactions are delicately balanced, leading to a rapid system response to varying conditions.

This complexity in behaviour is manifested at two levels: at the local level, variations of molecular

form, orientation and bonding are observed; while at the extended level, a variety of ordered assemblies

are observed. It can, therefore, be appreciated that even the combination of just one type of organic



6/141

molecule and one metal surface can unleash a cascade of phases, each possessing a different

combination of chemical, orientational and self-organisational behaviour (Fig. 1.1). This emphaticallyhighlights the sheer versatility of performance that can be generated. At present, only a few papers in

the literature attempt to summarise the multifaceted nature of these organic/inorganic interfaces. We

have, therefore, captured this information in terms of adsorption phase diagrams that provide anoverall and succinct perspective on the behaviour of the systems. These are not thermodynamic phase

diagrams but, rather, have been constructed to give chemical and structural information on the various

adsorption phases created as a function of coverage and temperature. The data for each system have

been collated by us from the reviewed references, so varies in the degree of detail presented, according

to the nature of the original work. In some cases, data is only available at one temperature and coverage

so the termphase diagramis not strictly appropriate. Nevertheless, we believe that this large pictorial

representation provides a valuable summary of the behaviour of each of the adsorption systemsreviewed. We have not attempted to give full accounts of the experiments reviewed but have selected

data which demonstrates particular aspects of the phase diagrams. We would like to point out that these

phase diagrams are, in fact, the essential backbone of this review, with the text providing supplementary

and supportive narrative.

1.2. Manifestations of chirality at the organic/inorganic interface

Of the various attributes that an organic molecule can bring to a metal surface, there is one that stands

out for special attention. This is the ultimate selectivity function of chirality. Chirality is simply a

Fig. 1.1. Schematic diagram to show how the adsorption of one simple molecule can unleash a cascade of phases on a

particular surface. Modifiers can take up many different combinations of molecular structure, bonding, orientation and 2D

order at a surface, yielding many different phases.



7/141

geometric property which dictates that the mirror transformation of an object is a non-identity

operation, i.e. the object and its mirror image are non-superimposable by any translation or rotation.

Clearly for this to hold, the object must not possess any inverse symmetry elements (i.e. centre of

inversion or reflection planes). As a result, a chiral object can exist in two distinguishable mirror, orenantiomeric, 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. In the organic system, chirality generally emerges at the tetrahedral carbon,

provided sufficient complexity is present, e.g. that all the four attached substituent groups are different.

The absolute configuration of such chiral centres can be labelled R (for rectus) or S (for sinister) asdetermined by the CahnIngoldPrelog rules[7,8].

Chiral expression at surfaces has only attracted increasing attention in recent years, despite the fact

that it is actually easier to create chirality in a 2D system since a surface cannot possess a centre of

inversion and can only maintain reflection mirror symmetry planes normal to the surface. Although

intrinsically chiral metal surfaces can be created by cutting to expose step and kink sites that are chiral,the interesting point for the organic/inorganic interface is how the adsorption of organic molecules

bestows chirality to a previously non-chiral surface. In fact, surface chirality can be manifested in a

number of ways and a hierarchy of surface chirality can be identified [9]. We suggest the following

classification of surface chiral systems that includes both the creation of local chiral motifs by singleadsorption events (i.e. point chirality) and the creation of chiral domains arising from the chiral

arrangements of the individual motifs (i.e. organisational chirality). We also differentiate between

molecule-induced chirality and adsorption-induced chirality and between expressions of local and

global chirality. A summary of the classification is shown inFig. 1.2and a description of how chirality

can be manifested at non-chiral surfaces is given below.

1.2.1. Surface chirality from adsorption of non-chiral moleculesThe adsorption of non-chiral molecules at non-chiral metal surfaces has been shown to lead

under certain conditions to expressions of chirality at a metal surface. The chirality is essentially

adsorption-induced and two major classes of chirality are expressed, described below. In both cases,

the chirality is strictly only expressed at a local level, and disappears at the global level.

1.2.1.1. Point chirality: adsorption-induced chiral motifs. This is the most basic form of chirality,

arising because the adsorption site symmetry of the molecule locally destroys all surface mirror planes.

For example, this can arise simply by adsorption of the molecule so that the molecular reflection planes

do not align with the surface mirror planes. Therefore, any system with adsorption site (or point group)

symmetry C1, C2, C3, C4or C6qualifies for this class of chirality, e.g. even a CO molecule tilted along anon-symmetry direction. What is very important to realise is that in such cases, energetically equivalent

reflectional configurations will always exist so that random adsorption will yield equal populations of

image and mirror image adsorption motifs. This means that the surface is a 50:50 racemic mixture and

possesses no overall chirality.

1.2.1.2. Organisational chirality: adsorption-induced chirally ordered domains. This type of chirality

arises when ordered adsorption structures are formed where the 2D organisationof molecules destroys

the reflection symmetry planes of the underlying surface. Such ordered domains belong to one of the five

possible chiral space groups (P1, P2, P3, P4or P6) that can exist at a surface. The organisational chirality



8/141

generally arises because local adsorption-induced chiral motifs of the type described above can lead to

asymmetries in lateral interactions, culminating in growth directions that lie along non-symmetry axes.

However, again, due to the inherent non-chirality of the initial molecule, there is equal probability of

nucleating reflectional chiral domains. As a result, these systems always consist of coexisting mirror

chiral domains, leading to an overall non-chiral, racemic surface.

Fig. 1.2. Classification of chirality at a surface.



9/141

1.2.2. Surface chirality from adsorption of chiral molecules

When a chiral molecule is adsorbed at a non-chiral surface, its very presence inevitably introduces

chirality at the surface. However, there are different levels of chiral expressions, ranging from point

chirality (molecule-induced) to highly organised, extended forms of chirality (molecule- andadsorption-induced). Crucially, the chirality of the adsorbed molecule enables chiral expression at

the surface to progress from a local to a global level.

1.2.2.1. Point chirality: molecule-induced chiral motifs. The adsorption of any chiral molecule at a

surface which leaves the molecular chiral centre intact will inevitably lead to a local chiral motif. Since

the inherent chirality of the molecule forbids the creation of its mirror image with all random adsorption

events, no mirror chiral motifs can be conceived. Therefore, an overall chiral system is always

produced.

1.2.2.2. Organisational chirality. The adsorption of chiral molecules on non-chiral surfaces can alsolead to a range of ordered structures. If one ignores the local chirality possessed by the molecule and,

instead, observes theorganisationof the adsorbates with respect to the surface, it is found that both non-

chiral and chiral arrangements can exist. Of the latter, two classes of chiral arrangements can exist, one in

which reflectional domains coexist, and the other in which they cannot.

Adsorption-induced chiral organisation(at the local level). In this class, the local chiral adsorption

motifs organise into a 2D chiral arrangement. However, we predict that in systems where lateral

interactions are mediated by groups that are non-chiral and sufficiently remote from the chiral centres,reflectional domain arrangements may also be nucleated and will coexist at the surface. Therefore,

organisationally both the image and mirror image chiral domains can exist. Overall, however, the

system is still chiral, because if the inherent chirality of the molecule is taken into account, then the two

domains are only pseudo-reflections of each other. At present, no published work on such systemsexists and this remains a hypothetical classification.

Adsorption-induced chiral organisation (at the global level). This is the highest expression of

chirality at a surface, involving both the creation of a molecule-induced chiral motif and an adsorption-

induced chiral organisation, with each present in only one of its two possible mirror arrangements. As a

result, only one unique chiral domain is nucleated and sustained over the entire interface so that a chiral

surface is created possessing both global point and global organisational chirality. The expression of

such global organisational chirality is difficult to attain and the first few examples have only beenrecently recorded. The constraints that govern the creation of a truly chiral organised array [10]can be

appreciated fromFig. 1.3which shows that the number of allowed space groups rapidly dwindles when

going from three-dimensional (3D) space (230 space groups) to 2D chiral space groups (only five spacegroups). More importantly, when a surface chiral space group is created, it can be expressed in either of

its mirror forms (Fig. 1.3). For a truly global organised chiral system, it is vitally important that the

manifestation of any of the five 2D chiral space groups is strictly restrained so that only one mirrorimage of the unit mesh is allowed to exist at the surface.

Clearly, chirality can be manifested at a surface in a number of ways. However, we note that only those

systems possessing global chirality can be considered to be truly chiral surfaces. To aid the reader,Table 1

shows the various combinations of chiral manifestations that can occur at surfaces and their various

outcomes in terms of exhibiting local or global chirality. We also note that there are some other aspects of

chirality that are particularly well demonstrated by the tartaric acid systems, discussed in Section 2.1.



10/141

1.3. Systems of study

In this review, we have concentrated solely on the following classes of complex organic molecules:

(i) multifunctional carboxylic acids,

(ii) amino acids,

(iii) aromatic anhydrides,(iv) closed and fused ring structures (without additional functional groups),

(v) closed and fused ring structures (with additional functional groups).

Table 1

Various combinations of chiral manifestations at surfaces along with the type of chiral surface created a

a Note that for a chiral molecule with a chiral adsorption motifwith no mirror image and a chiral organisation with no

mirror image this leads to the special case of global organisational chirality.

Fig. 1.3. Thefive chiral space groups allowed in 2D space. Each of the space groups can yield two mirror motifs at the

surface. A surface with global organisational chirality can only tolerate the existence of one of these two possible mirror

motifs.



11/141

Our choice represents systems for which a critical mass of literature has only recently been

accumulated, with sufficient multi-technique information available to enable a more holistic under-standing to be attempted.

2. Carboxylic acids

The simple carboxylic acids (functional group COOH) such as formic and acetic acid have been the

subject of much detailed earlier work and their properties at metal surfaces have been reviewed elsewhere

[1]. Here we concentrate on carboxylic acids that have more than one functionality that may be involved

in bonding to the metal surfaceeither additional rings or further COOH groups. As indicated in the text,

some of the carboxylic acids are truly chiral, others are achiral and some are prochiral, in the sense that

they can be adsorbed onto a surface with one of two possible faces uppermost, leading to a form of 2D

chirality. The varying interactions of these molecules are discussed for a range of metal surfaces.

2.1. Tartaric acid (L()-2,3-dihydroxysuccinic acid)

Techniques used: RAIRS, LEED, TPD, STM, periodic DFT calculations.

Preparation: Sublimation of pure enantiomers of tartaric acid (99%).

Tartaric acid occupies a special place in the scientific history of chirality with Pasteurs famousdiscovery that the molecule existed in two mirror crystalline forms[11]. More recently, it has been used in

chiral technology where there is a strong driver to establish enantioselective catalytic methods whereby

pure enantiomeric forms of pharmaceuticals, flavours, agrochemicals, etc. can be produced. One

successful way of creating heterogeneous chiral catalysts [1214]is to adsorb chiral organic modifier

molecules at metal surfaces in order to introduce asymmetry. The dicarboxylic acid, tartaric acid, which

possesses two chiral centres is one of the most successful chiral modifiers. Its presence on Ni, Cu andCo surfaces [13,14] endows significant discrimination to the hydrogenation of methylacetoacetate(MAA) to give methyl-3-hydroxybutrate (MHB) as shown in Fig. 2.1. For example, on a nickel surface

modified by R,R-tartaric acid, the reaction is stereodirected so that the R-product is produced in >90%

enantiomeric excess. Conversely, modification by S,S-tartaric acid favours the S-product. In suchmodifications, the need to create a globally chiral surface is self-evident. In order to gain a fundamentalunderstanding of the nature of the organic/metal interface created in such systems, the behaviour of the

modifierR,R-tartaric acid on defined Cu(1 1 0) and Ni(1 1 0) surfaces has been recently investigated. Aswe go to press, we also note that further adsorption studies have been carried out on the Ni(1 1 1) surface

[15,16].

2.1.1. R,R-tartaric acid on Cu(1 1 0)

The adsorption behaviour ofR,R-tartaric acid on Cu(1 1 0) under varying coverage and temperature

conditions reveals that the molecule/metal system occupies a complex and varied phase space (Fig. 2.2)



12/141

[10,1721]. Locally, the chemical, bonding and orientational structure adopted at the surface undergoesdynamic changes with conditions, causing the chiral organic molecule to adopt the monotartrate, the

bitartrate or the dimer forms (Fig. 2.2). Furthermore, these local units self-organise into a range ofordered, crystalline architectures at the surface, some of which possess the property of true global

organisational chirality. The adsorption of this molecule will be considered in some detail as this

system presents a particularly good example of the way one simple organic entity can lead to a

multitude of organic/metal interfaces, each possessing different characteristics. The phase diagram in

Fig. 2.2shows that, overall, at least six different types of monolayer phases are fashioned in the course

of adsorbing R,R-tartaric acid on Cu(1 1 0) [18] depending on adsorption temperature, coverage and

holding time. Of these, only three will be discussed in detail: the (9 0, 1 2) phase which is the

energetically preferred low coverage phase, the (4 0, 2 3) phase which dominates at intermediate

coverages and the (4 1, 2 3) phase that is created at the highest coverages.

2.1.1.1. Low coverage (9 0, 1 2) phase: the bitartrate assembly. The (9 0, 1 2) phase is the

thermodynamically preferred phase on Cu(1 1 0) at low coverages. However, a significant activation

barrier of greater than 70 kJ mol1 is associated with its creation[19]and it only forms spontaneously at

temperatures in excess of 400 K. Detailed RAIR spectroscopic data show that this adsorbed layer consists

entirely of the doubly deprotonated bitartrate species in which the two oxygen atoms in each COO unit

are held almost equidistant from the surface. Both carboxylate ends of the molecule are involved in

bonding to the surface, leaving the C2C3bond almost parallel to the surface and yielding a fairly rigidadsorption geometry. Recent periodic DFT calculations[22]have confirmed this general geometry and

have shown that the adsorption site of the bitartrate is one where the oxygens occupy on-top positions

Fig. 2.1. The stereodirected hydrogenation of MAA to give MHB on Ni catalysts chirally modified by tartaric acid (TA).



13/141

across the two short bridge sites with the CuO bonds possessing, on average, a length of 1.97 A

(Fig. 2.3c).

The 2D nature of this phase, shown inFig. 2.3d has been constructed from a comprehensive analysis

of the LEED and STM data. LEED data (Fig. 2.3a) reveal that the long-range ordering of tartaric acid at

the surface yields a (9 0, 1 2) structure which is consistent with a large unit cell possessing the

dimensions 23:047:68 A, a19:47. High resolution STM images, displayed in Fig. 2.3b, showthat there are three bitartrate molecules per unit cell, resulting in a fractional coverage of 1/6. The STMimages also reveal that rows of three bitartrate molecules assemble at the surface to form long chains,

which are aligned along theh1 1 4i surface direction. These growth directions are believed to bedictated by the presence of the a-hydroxy groups attached to both chiral centres of the molecule

[10,22].

2.1.1.2. Medium coverage (4 0, 2 3) phase: the monotartrate assembly. Although a medium coverage

phase, this structure only presents a low activation barrier for formation and so is created directly as

adsorption is carried out on a clean Cu(1 1 0) surface at 300 K. STM data show that the nucleation of this

Fig. 2.2. Phase diagram ofR,R-tartaric acid on Cu(1 1 0).



14/141

phase occurs in the early stages of adsorption and is preferentially located at step edges. Increasing

coverage leads to the steady growth of these islands until the entire surface is covered in this phase.

Again, detailed information on the nature of this phase has been constructed using a combination of STM,

LEED and RAIRS data. First, LEED photographs of this phase (Fig. 2.4a) show sharp diffraction spots

indicating that the islands possess very ordered arrangements of the modifier molecules in a (4 0, 2 3)

structure, on the Cu(1 1 0) surface. STM data (Fig. 2.4b) reveal that there are two other molecules within

this unit cell, giving a much more packed structure with a local coverage of 0.25 ML. The reason for theLEED inequivalence of these two extra molecules is not clear. However, there are a number of indications

that suggest that the details of the H-bonding network govern the overlayer symmetry and possibilities

are discussed in the original reference [18].

The chemical detail of the adsorbed entities is provided by RAIR spectra [18] obtained for this

phase (Fig. 2.4c). The presence of both the n(C=O) vibrations of the acid COOH functionality

at 1705 cm1 and the ns(COO) vibration of the carboxylate group at 1437 cm1, reveal that the

R,R-tartaric acid is adsorbed as a monotartrate species which is bound to the surface via the

deprotonated carboxylate group. In addition, the free and intact COOH acid group is held away from

the surface and the considerable downshift in frequency of then(C=O) vibration of this group suggests

Fig. 2.3. Details of the (9 0, 1 2) chiral phase created at low coverages and high temperatures showing: (a) the LEED pattern

obtained at 31 eV; (b) 150

200 ASTM image of the chiral surface showing trimers of bitartrate molecules aligned in

columns directed along the chiralh1 1 4i direction; (c) the local bonding description of the bitartrate unit; (d) a schematicmodel of the overlayer constructed from the STM, LEED and RAIRS.



15/141

that it is involved in intermolecular H-bonding interactions with the alcohol groups of neighbouring

monotartrate species, leading to a strong tendency for island growth in this phase. Combining all these

different pieces of information, a fairly complete description of this adsorbed phase is constructed

(Fig. 2.4d).

2.1.1.3. High coverage (4 1, 2 3) phase: the dimermonomer assembly. Further adsorption beyond a

coverage of 0.25 ML leads to the creation of a high coverage ordered phase at 300 K, which produces a

new (4 1, 2 3) LEED pattern. The RAIRS data of this (4 1, 2 3) phase[18]retain the overall fingerprint

of the monotartrate species, but reveal perturbation of the n(C=O) band which splits into two

contributions at 1759 and 1674 cm1, the latter frequency typical of H-bonded cyclic acid dimers

and the former indicative of a monomer or open-chain acid group which involves less H-bonding. In

contrast, the 1437 cm1 vibrational band due to the carboxylate functionality anchored to the Cu(1 1 0)

remains almost unchanged, indicating that this part of the adlayer is essentially similar to that of the

(4 0, 2 3) phases.

Fig. 2.4. The (4 0, 2 3) phase of R,R-tartaric acid on Cu(1 1 0): (a) LEED pattern obtained at 26 eV; (b) STM image

(8075 A; Vtip 1:52 V; It1:25 nA) showing the position of individual adsorbates; (c) RAIRS data monitoring thenature of the adsorbed species; (d) a structural model of the phase with the unit cell outlined.



16/141

STM images of this phase exhibit an ordered and dense adlayer, also with a (4 1, 2 3) repeat structure

(Fig. 2.5a). Using the chemical information and data from STM images taken at different tunnelling

conditions, a model of the (4 1, 2 3) adlayer has been constructed[18,21]. The STM image inFig. 2.5a

shows that along theh0 0 1i direction, repeat units consisting of a bright two-lobed structure and asmaller feature are seen. The RAIRS data for this structure show that cyclic carboxylic acid dimers andmonomer acid groups coexist in this structure, the former consistent with the two-lobed STM features,

and the latter associated with the single monotartrate unit (Fig. 2.5). Each cyclic dimer unit possesses

an OH group at each end and it has been proposed that lateral interactions (possibly H-bonding)

between the OH groups of adjacent dimer units naturally force the dimer chain alongh1 1 2idirection.The STM image inFig. 2.5b also shows strong lines alongh1 1 2i direction, which coincides with thealignments that need to be adopted by the dimer OH groups in order to facilitate strong H-bonding

intermolecular interactions. This STM image also shows that similar direct lines are also observed

along theh1 1 0idirection, suggesting that the OH groups on the single monotartrate molecules are alsoinvolved in H-bonding with the adjacent dimer units, tying the entire structure together by connecting

Fig. 2.5. STM images of the (4 1, 2 3) structure taken at different tunnelling conditions. (a) Molecular resolution showing

position of the adsorbedR,R-tartaric acid molecules (150150 A;Vtip 0:03 V;It0:51 nA); (b) lines along theh1 1 2iandh1 1 0idirections coinciding with the positions that would have to be adopted by OH groups in order to create a network ofH-bonds between molecules (150150 A; Vtip 1:11 V; It0:46 nA).



17/141

neighbouring dimer chains, as shown in Fig. 2.6c. In such a structure the acid group of the single

monotartrate molecules remains free, rationalising the emergence of the additional n(C=O) band at

1759 cm1. The overall pattern adopted by the dimer and monomer adsorbates on Cu(1 1 0) is shown in

Fig. 2.6a.

2.1.1.4. The emergence of global organisational chirality. For all the phases described above, the

inherent chirality ofR,R-tartaric acid always leads to the establishment of point chirality and a localchiral motif. However, in addition, a closer analysis reveals that for certain phases this local chirality

transfers to a truly global chiral organisation. Ignoring the local chirality possessed by the molecule and,

instead, observing the arrangement of the adsorbates with respect to the surface, it can be seen that the

(4 0, 2 3) pattern possesses two reflection symmetry elements but the (9 0, 1 2) and the (4 1, 2 3) templates

possess none (Fig. 2.7). Clearly, for the latter two cases, a global point and global organisational chiral

surface has been created, in which the arrangement of the molecules annihilates both reflection symmetry

planes of the underlying Cu(1 1 0). Importantly, for both phases, the same growth directions and

arrangements are maintained over the entire surface and as a result a perfect globally chiral surface is

created which is non-superimposable on its mirror image. A consideration of the (9 0, 1 2) and (4 1, 2 3)

Fig. 2.6. (a) Schematic diagram showing the unit cell and the position of theR,R-tartaric acid molecules in the (4 1, 2 3)

structure; (b) formation of cyclic dimers and hydrogen bonding interaction between dimers along theh1 1 2i direction; (c)depiction of single monotartrate molecules involved in further H-bonding interactions along the

h1 1 0

idirection, thus weaving

dimer chains together. Note that in all structures the carboxylate groups bonding with the surface are placed in short-bridged

sites which places each oxygen above a Cu atom.



18/141

structures at the molecular level enables the following three conditions to be identified as central to the

expression and sustainment of true chirality over an entire interface:

chirality of the modifier molecule, rigid and defined adsorption geometry, directional and anisotropic lateral interactions.

It can be seen that the (9 0, 1 2) bitartrate and the (4 1, 2 3) dimer phase fulfil all these requirements.The inherent chirality of these molecules and their two-point bonding at the surface uniquely de finesthe footprint the molecule casts at a surface and dictates the position of all its functional groups in

space. Once this is achieved, the intermolecular interactions between the modifiers control theplacement of neighbouring molecules. Here, the chirality of the adsorbates ensures that these lateral

interactions are anisotropic. For both cases, it is believed that the growth direction is dictated by the

spatial positioning of the a-hydroxy groups attached at the chiral centres of the adsorbed species

(Figs. 2.3 and 2.6), making the intermolecular interaction uniquely directional. Although, the detailed

nature of these interactions still remain a matter of debate[10,22]it is, nevertheless, clear that a better

description of these global organised chiral structures is that they are a supramolecular assembly [23]

of chiral modifiers. These lateral interactions essentially ensure that an energy difference existsbetween one chiral unit mesh and that of its mirror twin, so that only the energetically favoured chiral

domain is created. For the (9 0, 1 2) phase, DFT calculations show that an energy difference of

10 kJ mol1 exists betweenR,R-tartaric acid arranged in the (9 0, 1 2) mesh and its mirror (9 0,1 2)arrangement[22]. Therefore, at 300 K, over 95% of the adsorption would occur with one preferred

chiral arrangement.

2.1.1.5. Creation of chiral spaces within the chiral surfaces. The models presented of the surfaces

with global organisational chirality created byR,R-tartaric acid on Cu(1 1 0) lead to another notable fact,

namely, that the structure is open enough to reveal empty, nanosized chiral channels and spaces in which

Fig. 2.7. Overall adsorbate templates created by the (4 0, 2 3), the (9 0, 1 2) and the (4 1, 2 3) structures. Whereas the (4 0, 2 3)

pattern retains both reflection symmetry planes of the Cu(1 1 0) surface, the (9 0, 1 2) and the (4 1, 2 3) patterns do not, thus

creating surfaces with global organisational chirality.



19/141

the underlying metal is exposed. The genesis of these empty chiral spaces is still a matter of investigation,

with some indication that they are due to strain-breaks in the modifier overlayer arrangement and,

possibly, reconstruction of the underlying metal surface. These nanosized spaces combine the reactivity

of the underlying metal with the selectivity of the decorating chiral molecules and may be an importantroute to create molecular recognition or transformation sites, in which stereospecific docking and

reaction processes can be controlled.

2.1.1.6. Switching global organisational chirality. An important aspect of the asymmetric hydro-

genation of b-ketoesters on tartaric acid modified surfaces is that switching the chirality of the

tartaric acid switches enantioselectivity from the R- to the S-reaction product (Fig. 2.1). Therefore,

one would expect the mirror enantiomer ofR,R-tartaric acid to create a mirror surface structure. This

was first tested out by the adsorption of the opposite enantiomer, S,S-tartaric acid, on Cu(1 1 0) and

creating the parallel structures[10]. Turning first to the bitartrate structure, a sharp LEED pattern is

obtained from the equivalent S,S-tartaric acid phase, but the positions of the diffraction spots areswitched to the mirror (9 0, 1 2) structure. This chiral switching is better illustrated by the STM images(Fig. 2.8) where the (9 0,1 2) phase ofS,S-tartaric acid on Cu(1 1 0) is revealed to be a true mirror

Fig. 2.8. Switching global organisational chirality for the (9 0, 1 2) phase: R,R-tartaric acid versus S,S-tartaric acid. This

chiral switching is illustrated by the 108108 A STM images and the schematic models of mirror adlayers createdwhen R,R- andS,S-tartaric acid are adsorbed on Cu(1 1 0). Note the mirror positioning of the OH groups for the R,R- and the

S,S-tartaric acid adsorbates.



20/141

image of the (9 0, 1 2) phase obtained for R,R-tartaric acid, with every adsorbate position reflected inspace [10]. The preference for each enantiomer to describe a particular chiral domain is further

demonstrated when a mixture ofR,R- and S,S-tartaric acid is adsorbed at the Cu(1 1 0) surface and

shows the 2D version of the famous Pasteur experiment, i.e. distinct and separate R,R- andS,S-mirror

chiral domains are formed. This chiral switching is also observed for the (4 1, 2 3) phase where LEED

and STM images (Fig. 2.9) show a reversal of all adsorbate positions for the S,S-tartaric acid system to

create the chiral twin surface.

Surface chiral switching has been explained in terms of how opposite enantiomers guide the

supramolecular assembly directions. For both global organisationally chiral phases created by

tartaric acid, the growth direction of the adlayer is dictated by the conformation of the a-hydroxy

groups attached to the chiral centres of the adsorbed species. In both structures, the R,R- andS,S-adsorbed species possess defined and rigid adsorption geometries arising from their two-pointbonding with the metal surface. As a result they exhibit one major difference, namely the spatial

orientation of their OH groups,Figs. 2.8 and 2.9. Looking down at the surface, it can be seen that the

positions of the OH groups on R,R-tartaric acid in the (9 0, 1 2) structure are oriented in a mirror

configuration to those ofS,S-tartaric acid. Thus, the enantiomers adopt mirror growth directions, with

intermolecular interactions leading to a molecular chain alignment along theh1 1 4i direction forR,R-tartaric acid, but the mirrorh1 1 4i direction for S,S-tartaric acid. The same is true for the (4 1,2 3) phase, where the OH groups connecting the dimer and monomer units occupy mirror positions

for the R,R- and the S,S-tartaric acid systems, leading to the mirror chiral arrays being created.

Fig. 2.9. Switching global organisational chirality for the (4 1, 2 3) phase:R,R-tartaric acid versus S,S-tartaric acid. This

chiral switching is illustrated by the 110 100 ASTM images and the schematic models of mirror adlayers created whenR,R- and S,S-tartaric acid are adsorbed on Cu(1 1 0).



21/141

The central role played by the chiral OH groups in guiding a unique supramolecular assembly

direction is also nicely illustrated by comparing the (9 0, 1 2) and the (4 1, 2 3) structures. For the

(9 0, 1 2) structures the bitartrate species possesses OH groups along the bottom left to top right

diagonal, while for the cyclic dimer, these groups are now oriented along the bottom right to topleft diagonal. As a result, the assembly directions for the two structures also switch their alignment

sense accordingly.

2.1.1.7. Sustaining a single rotational domain. The importance of sustaining just one reflection, or

mirror, unit mesh in order to create a perfect global organised chiral array has been discussed. There is

another aspect of the (9 0, 1 2) and the (4 1, 2 3) structures that is also appealing, namely that a single

rotational domain is also sustained across the entire surface. Although rotational domains do not

threaten the chirality of a system and would not affect the heterogeneous enantioselective response,

some optical applications may require the stringent double-demand of a single reflection and rotational

domain to be maintained across an entire surface. However, adsorption is a random process and initialnucleation events cannot be controlled to occur in only one of many energetically equivalent rotational

positions. Generally, rotational domains occur when the overlayer surface mesh possesses a lower

symmetry than that of the clean substrate. Therefore, to ensure that all rotational domains that are

created are equivalent, rotational symmetry matching of the unit mesh with the surface is needed so that

all equally probable adsorption geometries result in an identical nucleation point. For example, it can be

seen that this is exactly the case for the (9 0, 1 2) and the (4 1, 2 3) structures ( Figs. 2.8 and 2.9). Both

possess unit meshes with C2point group symmetry, which rotationally matches the twofold axis of the

underlying Cu(1 1 0) surface.

2.1.2. R,R-tartaric acid on Ni(1 1 0)

R,R-tartaric acid modified nickel surfaces are perhaps the most successful heterogeneousenantioselective catalysts[13,14] and, therefore, the nature of chiral modification of Ni(1 1 0) by R,R-tartaric acid represents a particularly relevant system for study[2426]. As for Cu(1 1 0), the adsorbed

molecule shows significant chemical versatility with the bi-acid, the monotartrate and the bitartrate formsall being created at various points of the adsorption phase diagram (Fig. 2.10). However, there are two

main points of difference with respect to the Cu(1 1 0) surface. First, the barrier to the creation of the

bitartrate form is much lower so that it occurs spontaneously upon initial adsorption at room temperature.

Second, there is little evidence of the beautifully ordered supramolecular structures observed for

Cu(1 1 0). Here, the work on the room temperature phase only is reviewed in detail since it introduces new

aspects of chiral manifestation at the surface.

2.1.2.1. The room temperature phase: adsorption stresses and chiral reconstructions. Fig. 2.11shows

the STM data obtained with increasing coverage ofR,R-tartaric acid on Ni(1 1 0) at room temperature.

From these, it can be seen that no long-range ordered structures of the type observed on Cu(1 1 0) are

formed. However, despite the lack of 2D ordering, Fig. 2.11 shows that all the adsorbed molecules

exhibit a very preferred growth direction along the mainh1 1 0idirection, which also coincides with amirror symmetry direction of the surface. Again, this behaviour is very different from that observed

for Cu(1 1 0) where the molecules assemble along non-symmetry directions, thus annihilating all

the mirror planes. By the application of surface infrared spectroscopy and density functional

theory (DFT) it has been possible to obtain a fundamental understanding of the local adsorption



22/141

structures created on the Ni(1 1 0) surface and gain new and interesting insights on the nature of chiral

induction.

RAIRS data for low coverages ofR,R-tartaric acid adsorbed on Ni(1 1 0) at 300 K is consistent with

the creation of the bitartrate species, adsorbed so that both oxygen atoms in each carboxylate group are

approximately equidistant from the surface and the OCO planes are largely inclined towards the surface

plane[24,26]. However, the most valuable insight into the adsorption ofR,R-tartaric acid on Ni(1 1 0)

comes from consideration of its adsorption site using periodic DFT calculations [24]. For adsorption onthe bulk truncated Ni(1 1 0) surface, the most stable adsorption geometry is one in which two-point

bonding of the bitartrate occurs with the oxygen atoms of both carboxylate functionalities adsorbing on

top of adjacent Ni atoms, their positions in a plane parallel to the surface ( Fig. 2.12a). This adsorption

site is similar to that found for the bitartrate species on a bulk truncated Cu(1 1 0) surface [22].

Interestingly, this adsorption geometry is preferred largely due to the minimisation of repulsive

interactions between the molecular OH groups and the Ni surface atoms. However, such computed

configurations involving adsorption on the bulk truncated surface constrain the resulting groundstateseverely, as they do not allow for any change of the lateral distances between the Ni surface atoms.

Therefore, the adsorption energetics have been investigated [24] starting with a fully relaxed bitartrate-Ni4

Fig. 2.10. Phase diagram ofR,R-tartaric acid on Ni(1 1 0) from [26].



23/141

complex which is then free to optimise the positions of the Ni atoms comprising the adsorption site.

The most interesting outcomes of this are: (i) the interaction between the chiral OH groups of the

molecule and the Ni atoms leads to an acute distortion in the molecule whereby the C(2)C(3) bond isskewed by about 458 with respect to each of the other CC bonds, and (ii) the pairs of Ni atoms

constituting the adsorption site are now placed a significant distance apart and describe an oblique unitmesh at the surface for which all mirror planes are lost locally ( Fig. 2.12b). As a result, the adsorbed

bitartrate conveys its chirality not just simply by its presence at the surface, but more profoundly via

the chiral footprint it places onto the surface. Since the Ni footprint of the adsorbed complex requires

the long-bridge distance between the pairs of Ni atoms to be about 7.47 A, it can only be accommo-

dated on a bulk truncated Ni(1 1 0) surface via local paired-row and missing-row reconstructions

(Fig. 2.12c).

STM images (Fig. 2.11) support these general conclusions. The molecular structures in the STM

images occupy, on average, a space of 6:84:6 A, in good agreement with the calculated area of7:044:98 A encompassing the relaxed bitartrate-Ni4 complex (Fig. 2.12b). In addition, the smallscale structure of the molecular adsorption reveals severe distortions in the immediate vicinity(Fig. 2.11a and b), extending, on average, over a 15:512:5 A area for a single molecule. For thebulk truncated surface this represents an approximate 63:5 surface cell, consistent with either a localpaired-row or missing-row reconstructions, with the resulting perturbation being propagated a number

of atomic distances away from the adsorption centre. Schematic models depicting such local

reconstructions (Fig. 2.12c) show that the perturbation caused by the adsorbed bitartrate extends over at

least a 5 3 unit cell. For both geometries, the OCO planes of carboxylate units are signi ficantlyinclined towards the surface plane, consistent with the RAIRS data and the calculated value of 38.58for

the adsorbed bitartrate-Ni4complex. Finally, the driving force for this reconstruction is sufficient for itto be manifest even for the single molecule adsorption event. This is supported by the calculations

Fig. 2.11. STM images obtained with increasing coverage from (a) to (c) of R,R-tartaric acid on Ni(1 1 0) at room

temperature. (a) 300300 A image with Vtip 2:115 V and It1:16 nA; (b) 300300 A image withVtip 1:467 V and It1 nA; (c) 300300 A image with Vtip 1:76 V and It1 nA. In (a) areas offlat terracesare shown alongside step edges. In (b) and (c) the areas imaged as long dark stripes are due to co-adsorbed H atoms, which are

produced during the chemisorption process via a double deprotonation of the R,R-tartaric acid molecule. The H-atoms

segregate in separate islands and are thought to be adsorbed in a local missing-row structure.



24/141

Fig. 2.12. (a) Adsorption configurations for the bitartrate species adsorbed on Ni(1 1 0) aligned along the [0 0 1] axis;

(b) progressive distortions calculated for the C1C2C3C4 skeleton a bitartrate-Ni4 species adsorbed on a Ni(1 1 0) surface.

Note that the significant skewing of the molecular skeleton causes the four bonding nickel atoms to describe an oblique, chiral

footprint; (c) schematic models to demonstrate how the relaxed bitartrate-Ni4 footprint could be accommodated at the

Ni(1 1 0) surface by local reconstruction of the surface to (i) a paired-row structure and (ii) a missing-row structure.



25/141

which show that the adsorbed bitartrate-Ni4 species possesses a very large binding energy of

656 kJ mol1 at a Ni(1 1 0) surface. This value is consistent with DFT calculations of individual

carboxylate groups bonding to the surface, reported to be 357 kJ mol1 for formate on a Cu10(1 1 0)

cluster and 318 kJ mol1 for acetate on a Cu10(1 1 0) cluster[27].The strength of the bitartrate-surfacebond is also reflected in temperature programmed desorption (TPD) experiments [24] where anexplosive decomposition releasing the products, H2, H2O and CO2, is observed at 454 K, indicating that

the bitartratemetal interaction is so strong that intramolecular bonds break prior to metal moleculebonds.

2.1.2.2. The nature of electronic chiral communication from molecule to metal. An intriguing aspect

of theR,R-tartaric acid/Ni(1 1 0) system is that the chiral carbon centres of the molecule are two positions

removed from bonding oxygen groups which are attached to the non-chiral carboxylate carbons.

Nevertheless, a chiral restructuring of the metal occurs. Therefore, the question arises: how is the

chirality of the molecule communicated from the chiral centres to the bonding groups and thence to themetal surface itself? A detailed analysis of the charge distribution and electronic states of the bitartrate

species adsorbed on Ni(1 1 0) [25] reveals that adsorption involves a complex interplay of molecular

distortion, hybridisation of molecule and metal states, charge donation and back donation. First, the

charge distribution of all occupied states even in the gas phase bitartrate (Fig. 2.13a) shows that the chiral

properties of the molecule are actually manifest at the level of single electron states which show electron

distributions that connect the chiral OH groups directly to the carboxylate groups. Calculations of total

charge at each atomic position show that the carbons act as donors and oxygens as electron acceptors.

Crucially, the negative charge surplus at the four oxygens atoms turns out to be chiral in distribution,

ranging from0.27 for O(2) and O(3) and0.33 for O(1) and O(4). Upon attachment of the bitartratespecies to Ni atoms, this asymmetric charge distribution in the carboxylate oxygens vanishes and the

single electron states of the bitartarte-Ni4 entity are altered (Fig. 2.13b) due to hybridisation ofmolecular states with metal states. In fact, it is the donation and backdonation of charge via

hybridisation of molecular orbitals with nickel orbitals which removes the charge asymmetry in the

carboxylate groups and instead, propagates this chiral signature to the hybridised electron states of the

bitartarte-Ni4complex. This makes the four NiO bonds inequivalent, with the NiO(1) and NiO(4)bonds being 2.04 A long and NiO(2) and NiO(3) bonds being shorter at 1.94 A. The main

consequence of this is that the adsorption site, constituting the four Ni atoms now attains a chiral

electronic structure!

2.1.2.3. From local chiral reconstructions to global organisational chirality. At present, two separate

approaches have been utilised to create metal surfaces with global organisational chirality. One,demonstrated by the R,R-tartaric acid/Cu(1 1 0) system involves the assembly of chiral molecules

into organised chiral arrays at the surface. In such cases, the adlayer template possesses chiral growth

directions which serve to destroy the mirror planes of the surface and chirality is bestowed without

necessarily having to introduce chiral rearrangements of the metal surface atoms. The second approach

has been to utilise metal single crystals, cut to expose specific surface planes which are intrinsically chiral

exposing non-straight step edges with kink sites in which the metal atoms are arranged in a chiral

configuration [2830]. The research on R,R-tartaric acid on Ni(1 1 0) demonstrates that these twoattributes can be fused together, with the adsorption of the chiral molecule causing an attendant chiral

restructuring of the underlying metal atoms. We note that the underlying chiral reconstruction of metal



26/141

surfaces by chiral adsorbates may well be a more widely occurring phenomenon and there is now clear

need to re-examine other systems with appropriate structural techniques.

Again, the most important aspect to appreciate is that to progress from a local chiral structure to a

truly global organisationally chiral surface requires the creation of mirror chiral motifs to be severely

curtailed. This is a stringent requirement and a number of systems which exhibit local chirality, created

at surfaces by adsorption of molecules [31,32] or by restructuring of the metal [33], are unable to

advance to a truly global chiral surface because the mirror adsorption or the mirror reconstruction are

equally allowed, leading to an overall racemic system. In order to establish whether R,R-tartaric acidcan also create a mirror footprint at the surface, the adsorption energies for the relaxed bitartrate-Ni4complex in the twin chiral footprint geometries showed in Fig. 2.14 were computed. These spin-

polarised calculations showed that the adsorption ofR,R-tartaric acid in the mirror footprint geometry

(Fig. 2.14b) was less favoured energetically by about 6 kJ mol1. Although this is a small energy

difference, a simple consideration of the Boltzmann distribution law shows that, at 300 K, this difference

is sufficient to ensure that over 90% of the adsorbed bitartrate molecules adopt the preferred chiralfootprint (Fig. 2.14a) creating an overall highly chiral surface. As a comparison, it should be noted that in

homogeneous catalysis enantioselectivity is successfully achieved in systems operating with energy

differences of10 kJ mol1 between two reaction pathways leading to opposite enantiomers [34].

Fig. 2.13. Total charge distributions for single electron states of (a) gas phase bitartrate[25]and (b) bitartrate-Ni4complex[25].



27/141

2.2. Succinic acid

Techniques used: RAIRS, TPD, LEED.

Preparation: Thermal sublimation of solid powder under vacuum.

Succinic acid is a very similar molecule to tartaric acid, with the only difference being that the two

alcohol groups present in tartaric acid are replaced by hydrogen atoms leading to the consequent loss of

chirality. The adsorption behaviour of succinic acid on Cu(1 1 0) surface has been studied [35]in the

same manner as tartaric acid, allowing for a comparison of the behaviour of both to be made, especially

regarding the importance of the hydroxyl groups in tartaric acid which enable global organisational

chirality to be achieved.

2.2.1. Succinic acid on Cu(1 1 0)

As for tartaric acid, succinic acid is capable of existing in at least three different forms: the neutral bi-

acid form, the monosuccinate form, where one of the carboxylic groups has deprotonated, and the

bisuccinate form in which both acid groups have deprotonated. The adsorption behaviour is

summarised in the phase diagram (Fig. 2.15). The creation of the bi-acid form is confined to low

temperature adsorption, while at higher temperatures both the bisuccinate and the monosuccinate are

formed. Interestingly, the formation of bisuccinate is no longer kinetically hindered at room

temperature as seen for the bitartrate/Cu(1 1 0) system, and is formed spontaneously at low coverages.

It remains the preferred adsorption mode at low coverages, and only at high coverages is the

Fig. 2.14. Depiction of the relaxed bitartrate-Ni4 species adsorbed in twin mirror chiral footprints at the Ni(1 1 0) surface.



28/141

monosuccinate formed. Again, unlike tartaric acid, where the bitartrate form is converted to

the monotartrate form, here the bisuccinate is never transformed, but rather coexists with the newly

formed monosuccinate form. The monosuccinate form is largely unstable with respect to the bisuccinate

form and can be converted to the latter upon heating or with time, suggesting that its creation at high

coverages arises from kinetic factors. RAIRS data[35]show that the general adsorption geometry of the

bisuccinate is very similar to that of the bitartrate on Cu(1 1 0).

Information on the self-organisation of the monosuccinate and bisuccinate phases is provided by

LEED studies. The monosuccinate phase gives rise to a p(42) overlayer, whereas in thebisuccinate phase, coexisting (9 0,1 1) and (9 0, 1 1) chiral unit cells are observed ( Fig. 2.16).Whereas there is no chirality associated with the monosuccinate species, either at the local or theorganisational level, the bisuccinate does give rise to a local non-chiral motif organised in a chiral (9

0, 1 1) domain. However, the inherent symmetry of the molecule enables the reflection (9 0,1 1)domain to coexist leading to an overall racemic system. In conclusion, although the 2D organisation

of the bisuccinate adlayers can be chiral, two equivalent mirror image domains are always nucleated

and found at the surface (Fig. 2.16). This emphasises the importance of the alcohol groups of tartaric

acid in creating surfaces with global point and organisational chirality on the Cu(1 1 0). It is also

interesting to note from Fig. 2.17 that succinic acid molecules are adsorbed closer together than

tartaric acid molecules. This is again a direct consequence of the absence of the hydroxy groups in

succinic acid.

Fig. 2.15. Phase diagram of succinic acid on Cu(1 1 0).



29/141

TPD from the bisuccinate adlayer shows an explosive desorption at around 600 K arising from very

strong bonding interaction between the succinic acid molecules in the bisuccinate phase and the metal

surface. Similar explosive desorption has been observed in the bitartrate phase [19].

2.3. Benzoic acid

Fig. 2.16. LEED pattern observed when succinic acid is adsorbed on Cu(1 1 0) at 300 K, showing a mixture of p(42),(9 0,1 1) and (9 0, 1 1) structures.



30/141

Techniques used: RAIRS, TPD, LEED, STM, NEXAFS, XPS, HREELS.

Preparation: Solid powder sublimed under vacuum and exposed to crystal through leak valve.

Studies on benzoic acid at metal surfaces provided some of the first evidence that the adsorption ofmultifunctional molecules is not just dependent on the strength of interaction between the adsorbing

molecules and the metal atoms. The structureof the underlying metal surface and, thus, the mobility of

the metal atoms lead to different effects for benzoic acid adsorbed on different surfaces of the same

metal, e.g. Cu(1 1 0) and Cu(1 1 1). The strength of the metaladsorbate bond also leads to variations in

behaviour between different metals, e.g. Cu(1 1 0) and Ni(1 1 0). At all times there is an interplay

between preferred adsorption site, bonding functionality and mobility of both the underlying metal

atoms and the adsorbed species. In its acid state, benzoic acid is a prochiral molecule and thus capable

of adsorbing with a local chiral motif but incapable of sustaining global chirality. However, it is often

adsorbed in the deprotonated carboxylate form which has no inherent chirality. Benzoic acid is capable

of forming large ordered adsorbate structures and, for either chemical form, the possibility of creating

chirally organised domains exists in which both reflection domains are equally probable.

2.3.1. Benzoic acid on Cu(1 1 0)

Coverage and temperature dependent studies of this system have led to a very detailed understanding

of the way benzoic acid molecules interact with the Cu(1 1 0) surface [3639]. Most measurementshave been conducted with the copper held at room temperature but more recent studies have been

undertaken with the copper surface cooled by liquid nitrogen. The various forms of benzoic acid seen

on Cu(1 1 0) are collated in the phase diagram shown in Fig. 2.18, constructed from the results

presented in the above references. This is a modified version, to include molecular sketches, of the

phase diagram provided by the authors of[38].

Fig. 2.17. Schematic diagram showing the alignment of the succinic and tartaric acid molecules when they are adsorbed on a

Cu(1 1 0) surface with the bidentate orientation.



31/141

2.3.1.1. Room temperature substrate. At room temperature, in an analogous manner to the simplercarboxylic acids, both RAIRS and HREELS show no evidence for vibrations associated with the C=O or

OH groups but vibrations associated with the carboxylate COO ion are present, indicating that the

molecule adsorbs with the carboxylic acid group deprotonated. However, the presence of the aromatic

ring significantly modifies other aspects of the behaviour of the molecule compared to that of the simpler

carboxylic acid molecules such as formic acid and acetic acid. These latter molecules adopt an upright

stance on Cu(1 1 0) at all coverages with their carboxylate plane perpendicular to the surface. The

coverage dependent RAIR spectra (Fig. 2.19) of benzoic acid on this surface clearly show that at very

low coverage the only vibration observed is connected with the out-of-plane CH deformation, g(CH)

around 720740 cm1 which indicates that the aromatic ring is aligned parallel to the surface. With

increasing coverage new additional features appear, most notably the symmetric carboxylate stretchns(COO

) around 1440 cm1 which shows that at least some of the COO groups adopt a more upright

stance. This mode of vibration becomes increasingly strong and at saturation coverage all the molecules

appear to be aligned in an upright fashion. The Fourier transform infrared spectroscopy (FTIR) data

alone cannot determine whether or not the intermediate coverage spectra with vibrations due to both

g(CH) and ns(COO) arise as a result of the tilting of the original flat-lying molecules or the co-

adsorption of both flat-lying and perpendicular species. However, LEED and STM offer some clues

about the adsorption process.

At very low coverage no LEED pattern is seen and the system

2003_barlow and raval_complex organic molecules at metal surfaces bonding, organisation and...

Documents