SEARCH

SEARCH BY CITATION

Keywords:

  • chirality;
  • cysteine, gold;
  • second-harmonic generation;
  • surfaces

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

Chiral synthesis is an important route to biologically useful chemicals. Linear optical techniques like circular dichroism are too weak to detect chirality in monolayers of small molecules at surfaces, unless the surface area is very high and, in addition, the chiral response is annulled in reflection geometry. Chiral second-harmonic generation (SHG) from monolayer quantities of larger, more polarizable organic molecules has been detected at silica surfaces, where the presence of electronic resonances enhances the signal. Non-resonant chiral SHG has also been reported from multilayers of cysteine adsorbed on silica and silicon surfaces, where the strength of the response indicated potential monolayer sensitivity. Here, the chiral response of cysteine monolayers adsorbed on atomically clean, well-ordered Au(110) under ultra-high vacuum conditions is explored. Reproducible differences in the linear and nonlinear optical response from (R)-(+)-cysteine and (S)-(−)-cysteine monolayers are observed. Possible origins of the differences are discussed. Reflection anisotropy spectroscopy and SHG, in principle, provide complementary information on chiral structures at surfaces and interfaces and have potential technological application in the area of heterogeneous enantioselective catalysis.


1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

The molecules of life, such as amino acids, proteins and DNA are chiral. Chirality expresses the “handedness” of a molecule's three-dimensional structure and results in a rotation of the plane of polarized light either to the left or to the right. Circular dichroism (CD) detects the slight differences in optical absorption between left and right circularly polarized light interacting with chiral enantiomers in solution.

At surfaces, molecule–molecule and molecule–surface interactions can lead to chiral interfaces, which have no mirror or inversion symmetry elements. Two-dimensional (2D) chirality is currently being investigated using conventional surface science techniques, with much of the interest being driven by potential technological applications in the areas of heterogeneous enantioselective catalysis and biomolecular sensing and recognition [1]. However, probing catalysis and sensing at the molecular level at 2D solid–gas and solid–liquid interfaces, under realistic operating conditions, is generally not possible using conventional surface techniques.

Optical techniques can offer significant advantages in this area, due to the large penetration depth of the optical radiation [2]. The two challenges facing optical techniques in probing surface and interface structure are sensitivity and discrimination from the bulk. Two-dimensional chiral interfaces have weak optical activity that is difficult to detect, although progress has been made with very high surface area systems such as nanoparticles dispersed in solution. The optical activity of chiral ligand protected metal nanoparticles in solution have revealed novel phenomena at the nanoscale recently [3]. However, linear optical techniques cannot directly detect a chiral response in reflection geometry, as the chirality of the signal is reversed on reflection, annulling the chiral contribution to the optical response.

The nonlinear optical response, in contrast, offers the advantage of measurement in reflection from absorbing substrates, together with background-free discrimination, but sensitivity is always a concern with nonlinear optical probes. Optical second-harmonic generation (SHG) from large chiral molecules is readily detected and it has been applied over many years to problems of biophysical interest [4]. Macroscopic averaging does not remove the chiral dipolar SHG response, in contrast to solutions of achiral molecules, where the dipolar response averages to zero. Laser scanning microscopy, using fs pulsed laser systems, can produce live cell SH images when membrane-staining chiral dyes with large nonlinear optical coefficients are used to generate contrast [5].

In the 1990s, the Hicks group carried out pioneering work on ultrathin films and monolayers of easily polarizable molecules at fused silica and liquid interfaces, using a dye laser to tune into molecular resonances [6, 7]. More recently, this work has been extended to large molecules of biological interest, such as cytochrome c, and combined with surface-enhanced Raman spectroscopy [8].

In recent years, tuneable fs lasers have offered improved sensitivity and have allowed new systems to be probed. SHG has been used, for example, to probe the spontaneous formation of chiral Langmuir monolayers from the achiral molecule 5-octadecyloxy-2-(2-pyridylazo)phenol [9]. However, there appears to be no work reported on small chiral molecule adsorption on well-characterized single crystal surfaces, a necessary first step in probing chiral catalysis.

Chiral non-resonant SHG studies, in reflection geometry, of ultrathin films of the amino acid cysteine, HS[BOND]CH2[BOND]CH(NH2)[BOND]CO2H, deposited on native-oxide-covered Si(001) substrates have been reported, which indicated potential ML sensitivity [10]. Cysteine contains a thiol substituent that facilitates adsorption on gold surfaces and it has been shown recently, by STM, that a racemic mixture of cysteine, adsorbed on the single crystal Au(110) surface under ultra-high vacuum (UHV) conditions, dimerizes such that the molecular pairs that form on the surface are exclusively homochiral (Fig. 1) [11, 12]. This is compelling evidence for chiral discrimination in molecular interactions at crystalline surfaces.

image

Figure 1. STM images of cysteine adsorbed on Au(110)-1 × 2 (5 nm × 4 nm): (a) R-(+)-dimers and (b) S-(−)-dimers. (c) The R-(+)-dimer adsorption geometry (yellow, S; red O; blue, N; black, C; white, H) (after Ref. [12]).

Download figure to PowerPoint

Here, the linear and nonlinear optical response of cysteine monolayers adsorbed on atomically clean, well-ordered Au(110) under UHV conditions is explored. Reproducible differences in the response from the R-(+)- and S-(−)-monolayers are observed and possible origins of the differences are discussed.

2 Experiment

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

The samples were prepared under UHV conditions with a base pressure below 5 × 10−11 mbar. The Au(110) single crystal was cleaned by cycles of etching, using 500 eV Ar ions, and annealing to 700 K, the optimum temperature for large domain formation [13], until a sharp 1 × 2, “missing row” low energy electron diffraction (LEED) pattern appeared and no residual C and O contamination was observed with Auger electron spectroscopy (AES). After thorough degassing, R-(+)- or S-(−)-cysteine was sublimed onto the sample from a pyrex tube, at a pressure of 10−6 mbar in the UHV chamber. Quadrupole mass spectrometry confirmed that the cysteine sublimed without decomposition. Monolayer deposition was followed in situ and in real time by using reflection anisotropy spectroscopy (RAS) to monitor the attenuation of the 3.5 eV peak. RAS measures the difference in reflectance, at near normal incidence, of light linearly polarized in two orthogonal directions at the surface plane of a cubic material [14]:

  • display math(1)

where rx and ry are the complex Fresnel reflection coefficients for the surface for light polarized in the x and y directions. Spectra were compared with previous RAS studies of the R-(+)-enantiomer on Au(110) [15].

For the SHG measurements, a femtosecond laser system was used as the fundamental frequency light source, tuned to a wavelength of 800 nm. Both the fundamental and SH wavelengths are far from resonance, with the maximum CD response of cysteine occurring around 200 nm [16]. Unamplified 130 fs Ti:sapphire laser pulses of an average power of 200 mW were used at a repetition rate of 76 MHz. The beam size was ∼150 μm at the sample and the angle of incidence was 45°. A quarter-wave plate in the input beam was rotated and the s- and p-polarized SH response was measured as a function of the wave plate angle, using color filters and a monochromator to separate the SH photons from the fundamental.

3 Phenomenology

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

In addition to the usual electric-dipole moment, a magnetic dipole moment may also contribute to the SH response of chiral molecules, arising from their helicity. Chiral molecules also respond to the magnetic component of the driving field. The lowest order contribution of magnetic dipoles is where only one of these transitions is due to a magnetic dipole interaction. The magnetic dipole contribution at inline image to the second-order nonlinear surface polarization is

  • display math(2)

where e and m label the electric and magnetic contributions to the second order susceptibility tensor components inline image, and E and B are the fundamental electric field and magnetic fields, respectively. At inline image, the magnetic-dipole contribution to the second order nonlinear surface magnetization is

  • display math(3)

For molecules distributed isotropically in the surface plane, the usual electric dipole ijk terms, zzz, zxx = zyy, xxz = yyz are present, but chirality introduces additional tensor components of the form xyz and its permutations [17].

While the chiral response may be measured in a number of different ways using SHG, it has been shown that varying the input polarization state by rotating a quarter-wave plate simplifies the extraction of the tensor components [18]. For a rotating quarter-wave plate at an angle α,

  • display math(4)

where the real and imaginary parts of f, g, and h comprise combinations of Fresnel coefficients and tensor components [19]. For p-polarized output, f and g depend only on achiral components, while h involves only chiral components and thus will reverse sign between the (R)-(+)- and (S)-(−)-enantiomers; the converse holds for s-polarized output. The electric dipole terms are real and the magnetic dipole terms are imaginary for non-resonant measurements, resulting in a simplification of the formulae. This phase difference between chiral and achiral components enhances the sensitivity of the technique. Simultaneous nonlinear least-square fits of Eq. (4) to the enantiomers were used, with the same real and imaginary f, g, and h parameter values, but changing the sign of appropriate parameters between the enantiomers.

4 Results and discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

Figure 2 shows the RAS response from clean Au(110)-1 × 2, and with the cysteine MLs. The decrease in the RAS response at 3.5 eV was used to determine the completion of a ML of cysteine. The (R)-(+)-cysteine results agree well with previous work [15], including the effect of heating to ∼400 K, where phases of atomic S have been identified [12]. Figure 2 shows very good reproduciblity in the RAS response for the clean surface and the two enantiomer MLs when heated to 400 K. However, the room temperature spectra show a reproducible difference between the (R)-(+) and (S)-(−) MLs around 2.5 eV. RAS in electrolyte shows a similar decrease in the 2.5 eV signal as the interface structure changes from (1 × 2) to (1 × 1), as the electrochemical potential is increased [20]. This suggests that the enantiomers induce different achiral changes in the underlying Au structure that are detected in the RAS response.

image

Figure 2. RAS response of clean Au(110)-1 × 2 (blue line), with a ML of cysteine adsorbed (red line), and after heating to 400 K (black line): top, (R)-(+)-cysteine; bottom, (S)-(−)-cysteine.

Download figure to PowerPoint

In order to test this hypothesis, LEED images of clean Au(110)-1 × 2, and with cysteine MLs, were recorded and are shown in Fig. 3. Although no previous LEED studies appear to have been reported, reflection high-energy electron diffraction (RHEED) [21] and surface X-ray diffraction (SXRD) [13] studies of sub-MLs of the naturally occurring (R)-(+)-cysteine report a fourfold periodicity along [001], with the SXRD identifying the periodicity as arising from a reconstruction of the underlying Au atoms. The structure fades as the ML is completed [21]. Figure 3 shows that two different LEED patterns were obtained for a ML of the enantiomers: 1 × 2 for (R)-(+)-cysteine and c(2 × 4) for (S)-(−)-cysteine. These patterns were reproduced a number of times by depositing a new ML on the same substrate. The adsorption of the two enantiomers appears to affect the reconstruction of the Au substrate differently. The homochiral adsorption observed previously (Fig. 1) may be related to left- and right-handed kink sites in surface steps [22], and it may be that the relative density of these can affect the reconstruction observed from the different enantiomers. The LEED results are consistent with the hypothesis that adsorption of the two enantiomers produces different achiral reconstructions of the underlying Au surface, which are detected in the RAS response.

image

Figure 3. LEED images, taken at 129 eV, of Au(110)-1 × 2 (top), with an (R)-(+)-cysteine ML (middle), and with an (S)-(−)-cysteine ML (bottom).

Download figure to PowerPoint

The non-resonant SHG response was now measured and was found to be stable, with neither photo-bleaching nor photo-degradation being observed. The α-s response was more sensitive to the adsorption of cysteine than the α-p response. Figure 4 shows the s-polarized SHG data, together with simultaneous fits to Eq. (4), for MLs of the two enantiomers, together with the clean Au(110) response. The overall response was assumed to be the coherent sum of the substrate and cysteine ML contribution, allowing only the overall amplitude and phase of the measured Au(110) response to vary. These results were also found to be reproducible, with the (R)-(+)-enantiomer reducing the SHG response relative to the clean Au(110) surface, and the (S)-(−)-enantiomer increasing the SHG response.

image

Figure 4. α-s response of an (R)-(+)-cysteine ML (top), and an (S)-(−)-cysteine ML (bottom), adsorbed on Au(110) and measured in UHV (circles). The clean Au(110) response is shown for comparison (crosses). The solid lines are fits to Eq. (4).

Download figure to PowerPoint

Good fits are obtained, where the difference in the response of the enantiomers is modeled by simply changing the sign of the chiral components for s-polarized output, f and h, in Eq. (4). The difficulty, however, is that, due to the simple curve shape, good fits can also be obtained when the chiral components are excluded entirely and the overall intensities are simply scaled. Overall intensity variation is an unlikely explanation of the observed response, because the fs laser output is very stable and the intensities are normalized to the square of the input power. While the absolute intensities could vary by 20% depending on substrate preparation, the relative intensities varied by <10%. The changes in relative intensity between enantiomers in Fig. 4 are ∼40%.

While some contribution from the chiral tensor components cannot be excluded, the difference in this non-resonant SHG response is likely to be due mainly to the achiral tensor components, associated with the change in the underlying Au reconstruction, as observed in the LEED images and in the RAS response. Spectroscopic studies should be able to clarify the origin of the SHG signal, and such studies may be able to distinguish a direct chiral response when in the region of a cysteine resonance.

5 Conclusions

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

Reproducible differences in the linear and nonlinear response from (R)-(+)-cysteine and (S)-(−)-cysteine MLs adsorbed on Au(110) have been observed. While the origin of the differences remains to be resolved, RAS and SHG clearly have the potential to provide complementary information about chiral interactions at surfaces, the former probing achiral changes and the latter both achiral and chiral changes, with potential technological application in the area of heterogeneous enantioselective catalysis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References

The authors acknowledge the financial support of Science Foundation Ireland (SFI) provided under contract No. 08/RFP/MTR1043.

References

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experiment
  5. 3 Phenomenology
  6. 4 Results and discussion
  7. 5 Conclusions
  8. Acknowledgements
  9. References