Nonlinear Optical and Structural Properties of Langmuir−Blodgett

1940

J. Phys. Chem. B 2008, 112, 1940-1945

Nonlinear Optical and Structural Properties of Langmuir-Blodgett Films of Thiohelicenebisquinones Mikael Siltanen,*,† Elina Vuorimaa,‡ Helge Lemmetyinen,‡ Petri Ihalainen,§ Jouko Peltonen,§,⊥ and Martti Kauranen† Institute of Physics, Tampere UniVersity of Technology, Tampere, Finland, Institute of Materials Chemistry, Tampere UniVersity of Technology, Tampere, Finland, and Department of Physical Chemistry, Åbo Akademi UniVersity, Turku, Finland ReceiVed: October 30, 2007; In Final Form: NoVember 22, 2007

We provide a detailed investigation of the second-order nonlinear optical and structural properties of LangmuirBlodgett (LB) films of nonracemic thiohelicenebisquinone (THBQ). We prepare both X- and Y-type films of different thicknesses and characterize them using optical second-harmonic generation and atomic-force microscopy (AFM). We find that the overall nonlinear properties of the samples are essentially independent of the film thickness and the deposition type and arise from susceptibility tensor components associated with chirality. Both X- and Y-type films can be described by D2 symmetry, which is a higher symmetry than the previously assumed C2 of LB films of THBQ and a similar helicenebisquinone (HBQ). However, the two types of films are shown to differ significantly with respect to the orientation of the in-plane axis. For Y type, the axis follows the direction of vertical sample deposition, but for X type, the direction of the axis varies randomly and significantly between different samples. The Y-type samples are therefore more ordered than the X-type samples. This was confirmed by AFM measurements in which the Y type exhibits uniform ordering into columnar structures. Similar structures in X type, on the other hand, are shorter and more randomly oriented, like those earlier observed for racemic samples of HBQ [Verbiest, T., et al. Science 1998, 282, 913]. The common nonlinear properties and different high-level ordering observed here for two different types of nonracemic samples reinforces that the nonlinearity of THBQ (and probably HBQ, as well) originates from the low-level columnar aggregation of the molecules with the higher-level structures playing a lesser role. In addition, within the columns, the molecules likely assume fairly random azimuthal orientations so that the columns themselves exhibit approximate D∞ symmetry.

1. Introduction The nonlinear optical response of materials is closely related to their structural symmetry. This is particularly the case for second-order processes, which are forbidden in centrosymmetric materials within the electric-dipole approximation of the lightmatter interaction.1 This fact justifies the use of second-order processes as a tool to study surfaces, interfaces, and thin films, where the symmetry is usually broken.2 In more general terms, the symmetry of the material determines the structure of its nonlinear susceptibility tensor. The values of the components of the susceptibility tensor can provide valuable information about the structural order of materials.3 Of particular interest has been the determination of average molecular orientations in organic thin films,4-6 recently also in microscopic spatial scale.7 Finding the correct orientation angle is also a good example of how important it is to obtain reliable experimental data and to properly interpret it.8,9 Organic molecules have received significant attention as nonlinear optical materials.10 Such materials are convenient to investigate as thin-film samples. Of particular interest is film preparation using the Langmuir-Blodgett (LB) technique,11 †

Institute of Physics, Tampere University of Technology. Institute of Materials Chemistry, Tampere University of Technology. § Åbo Akademi University. ⊥ Current address: Laboratory of Paper Coating and Converting, Åbo Akademi University, Turku, Finland. ‡

which is applicable to amphiphilic molecules and works well with a large number of organic compounds. The thickness and composition of LB films and whether the molecules point in or out of the substrate can be precisely controlled, which is a significant advantage in many situations over spin-coated, cast, and evaporated films. Nonlinear properties of chiral molecules, which occur in two different forms that are mirror images of each other, have received considerable attention from two complementary points of view. First, nonlinear techniques can provide new probes of chirality,12-21 and second, chiral molecules provide a new class of nonlinear materials with unique properties. Such molecules themselves and even their isotropic, randomly oriented collections lack a center of symmetry and therefore possess a secondorder response, which greatly facilitates the construction of more ordered nonlinear structures. In addition, magnetic nonlinearities may play a role in chiral materials,22,23 and such materials provide a new way for quasi-phase-matching.24 Chiral molecules can also be ordered into anisotropic structures,25,26 which results in samples with very low symmetry (lack of both reflection/inversion and rotational symmetry) and a large number of independent nonvanishing components of the nonlinear susceptibility tensor.27 The nonlinear characterization of such samples is particularly challenging, because the separation of anisotropy and chirality requires great care.28 Nevertheless, the basic principles for this have been demonstrated. 25,26,29

10.1021/jp710476k CCC: $40.75 © 2008 American Chemical Society Published on Web 01/30/2008

LB Films of Thiohelicenebisquinones In addition, we have recently developed new measurement techniques that allow very accurate and reliable characterization of thin films of very low symmetry.30,31 Chiral helicenebisquinone (HBQ) and thiohelicenebisquinone (THBQ) have been shown to possess very interesting secondorder nonlinear properties.29,32,33 The nonracemic bulk samples from those materials, such as cooled isotropic melt, can assemble into macroscopic fiberlike higher-level structures consisting of hexagonally packed columnar aggregates.34,35 In LB films, fibers are not formed, but the molecules are still organized into columns36 that are similar to the ones observed in the bulk. The nonlinear response of the helicene LB films is dominated by tensor components associated with chirality.25,29 In addition, molecular aggregation has been shown to play an important role in the high nonlinear response. The supramolecular chiral aggregates are therefore crucial for obtaining a large nonlinear response. However, indirect evidence from comparing racemic and nonracemic samples suggests that ordering into higher-level superstructures plays a lesser role in the nonlinear properties,29 but this issue deserves further attention. We also note that other types of supramolecular structures may also have beneficial nonlinear properties.37,38 In this paper, we provide a detailed investigation of the nonlinear and structural properties of LB films of nonracemic THBQ prepared in different ways. In particular, we prepare both X- and Y-type films of different thicknesses and characterize them using optical second-harmonic generation (SHG) and atomic-force microscopy (AFM). Our results show that the overall nonlinear properties are essentially independent of the film thickness and the preparation technique. More specifically, both X- and Y-type films can be well-described by D2 symmetry, which is a higher symmetry than the C2 symmetry previously assumed for similar LB films of HBQ and THBQ molecules25,26,29,30,39 and includes an in-plane rotation axis as an additional symmetry element. The relative values of the tensor components are found to be the same for all films, and all of the significant components are associated with chirality. In addition, their absolute magnitudes for a given thickness are equal for X and Y types. However, the X- and Y-type films do differ significantly with respect to the orientation of the inplane symmetry axis. For Y type, the direction of the axis follows, within experimental uncertainty, the direction of vertical sample deposition. For X type, on the other hand, the direction of the axis varies significantly between the films of different thicknesses. This suggests that the Y-type samples are more ordered than the X-type samples. This was confirmed by AFM measurements in which the Y type exhibits uniform ordering into columnar structures, whereas similar structures in the X type are shorter and more randomly oriented, like those earlier observed for racemic samples of similar materials.29 The similar nonlinear properties and different high-level ordering, observed here for nonracemic samples of two types, reinforces the interpretation that the nonlinearity of the helicene originates from the low-level columnar aggregation of the molecules, with the higher-level structures playing a lesser role. In addition, within the columns, the molecules likely assume fairly random azimuthal orientations so that the columns themselves exhibit approximate D∞ symmetry. 2. Theoretical Background Our nonlinear optical technique is based on SHG, in which the polarization properties of both the incoming fundamental and second-harmonic (SH) signal light are used to extract information from the LB film samples. Within the electric-dipole

J. Phys. Chem. B, Vol. 112, No. 7, 2008 1941

Figure 1. The geometry of the thin film second-harmonic generation. The z-axis is normal to the sample surface, and the in-plane symmetry axis defines the angle of orientation φ of the x and y axes. The p and s components of the fundamental beam at frequency ω and secondharmonic beams at frequency 2ω are defined by the plane of incidence.

approximation, the macroscopic SH response of a material is 40 described by the susceptibility tensor, χ(2) ijk , defined as

P(2) i (2ω) )

χ(2) ∑ ijk Ej(ω) Ek(ω) jk

(1)

where P(2)(2ω) is the second-order polarization of the medium that acts as the source for the SH radiation, E(ω) is the electric field of the fundamental light, and subscripts i, j, and k denote components of the fields corresponding to the Cartesian coordinate axes x, y, and z. The second-harmonic susceptibility tensor has 27 components, but several of them often vanish or have equal or opposite values, depending on the symmetry and other properties of the sample. For example, the symmetry groups C2 and D2, which are of particular interest in this work, have eight and three independent nonzero components, respectively.41 Nevertheless, full understanding of the macroscopic SH response of a sample requires that the tensor be known. Unfortunately, the components of the tensor cannot in general be measured directly, but must be determined from a number of measurements for which the polarization properties of the fundamental and SH light are varied.42 In surface and thin-film geometries and for the case of collimated or weakly focused beam, the fields are most naturally expanded in terms of their p and s polarized components, which are parallel and perpendicular to the plane of incidence, respectively. Figure 1 shows the plane of incidence, fundamental and second-harmonic generated beams and their p and s components, and the xyz coordinate system of the sample with the azimuthal orientation characterized by the angle φ. The intensity of any measured SH signal must therefore be of the general form43

I(2ω) ) |fEp2(ω) + gEs2(ω) + hEp(ω) Es(ω)|2

(2)

where f, g, and h are the quantities that can be directly measured in an experiment. These quantities are complex-valued linear combinations of the tensor components and their expressions. In particular, these quantities depend on the polarization of the detected SH signal (e.g., s or p); on the linear optical properties of the sample; and on the experimental geometry, including the angles of incidence, azimuthal orientation of a sample with inplane anisotropy, and whether a SH signal is detected in the transmitted or reflected direction. We have earlier demonstrated a technique31 that is able not only to yield the relative values of the susceptibility components but also to estimate their accuracies and to check the validity

1942 J. Phys. Chem. B, Vol. 112, No. 7, 2008

Siltanen et al.

of the theoretical model. The main idea behind the technique is as follows: All of the measured f, g, and h in eq 2 for a sample are equated with their theoretical expressions that are linear combinations of the unknown tensor components. Because the overall magnitude and phase of the SH light can vary between individual measurements, a complex-valued scaling factor is introduced to bring all measurements to the same scale. This results in a linear group of equations of the form t (2) t (2) cF m ) f t1 χ (2) 1 + f 2 χ 2 + ... + f q χ q

(3)

where c is a scaling factor; Fm is the measured value of f, g, or h; and fit are the coefficients of the unknown susceptibility tensor components, χ(2) i . The coefficients are obtained from the physical model of the sample and the experimental geometry corresponding to the measurement. By performing more than the required minimum number of independent measurements, the group of equations becomes over-determined and can be solved using standard matrix manipulation and regression methods. For best results, we have chosen to use the total leastsquares (TLS) method44 to obtain the solution. The solution to the group of equations yields not only the relative values of the tensor components but also statistical information and error estimates on the basis of the additional information available due to the over-determined group of equations. Additionally, the solved scaling factors can be compared to the power levels of the incident laser beam used for each measurement to see if they match. All this provides important additional information, not traditionally available, when comparing different physical models to see which one is appropriate for a specific sample. An additional detail deserves attention when the sample has in-plane anisotropy. The above procedure yields values for the susceptibility tensor components in the coordinate system chosen before the experiment. However, this choice may not be the proper one to display all the symmetry properties of the sample. Therefore, a high-symmetry sample having only few nonzero components in a properly oriented coordinate system may appear to have very low symmetry in another, misaligned coordinate system in which many more tensor components have nonzero values. The possibility of such wrong assignment of symmetry can be removed by transforming the tensor into a new, rotated coordinate system by (2) ) χi′j′k′

Ri′iRj′jRk′k χ(2) ∑ ijk ijk

(4)

where the new coordinate system is denoted by primed subscripts and the matrix for rotation by angle ∆φ about the surface normal is

[

cos ∆φ sin ∆φ 0 R ) -sin ∆φ cos ∆φ 0 0 0 1

]

Figure 2. The schematic structure of the thiohelicenebisquinone molecule used in the Langmuir-Blodgett samples.

(5)

A higher symmetry where some of the tensor components must vanish can then be recognized, provided that the rotation angle ∆φ can be adjusted in such a way that the values of those tensor components become vanishingly small as compared to the dominant components. 3. Experimental Methods The Langmuir-Blodgett films used as samples contained several layers of the chiral THBQ molecules deposited on glass (for optical measurements) or silicon (for AFM measurements)

Figure 3. The optical setup used for the nonlinear measurements. GP is a Glan polarizer, QWP is a quarter-wave plate, and PMT is a photomultiplier tube. VIS and IR are filters blocking visible and infrared light, respectively. IF is a narrowband interference filter centered around 532 nm.

substrates. The structure of the molecule is shown in Figure 2. Only the (+) enantiomer of the THBQ was used. The plates were precoated with an optically inactive monolayer of octadecylamine (ODA) to improve the deposition of THBQ molecules on the substrates. Our series of samples consists of both X- and Y-type LB films, with the number of THBQ layers ranging from 4 to 16. The LB films were prepared with a KSV minialternate system (KSV Instruments ltd.). Water purified by a Milli-Q system (Millipore) was used as a subphase. The temperature of the subphase was 18.0 ( 0.5 °C. THBQ and ODA were spread on the subphase from chloroform solutions. The monolayers were compressed at a rate of 550 mm/min to the deposition pressure of 25 mN/m. All films were prepared with the vertical deposition technique. For Y-type films, substrate was repeatedly dipped first down and then up through the surface, and the deposition took place in both directions. For X-type film, the deposition took place during the downstroke, and the substrate was taken out through the clean subphase surface. In both cases, the transfer rate was 10 mm/ min and the film was allowed to dry for at least 20 min before the next deposition. The transfer ratio of the monolayers was close to unity at all times. The transfer of the films was also monitored by absorption spectroscopy (Shimadzu UV 2501PC spectrophotometer). For both film types, the absorbance was linearly proportional to the number of layers and matched closely the absorption in solution measured by Phillips et al. (see ref 35, Figure 1b). A weak absorption band peaking at about 465 nm and other bands growing in strength toward the shorter wavelengths were detected. Optical microscopy did not reveal any features in addition to random impurities and imperfections, which together with the steady transfer ratio and increase in absorption during the deposition process indicates good sample quality. After deposition, the molecular film from one side of the substrate was removed, resulting in samples with film on one side of the substrate only. SHG from the samples was measured using a setup similar to what has been described earlier31 and schematically shown in Figure 3, except that the laser pulse length and energy were now about 70 ps and 100 µJ, respectively. The fundamental laser beam at the wavelength of 1064 nm was weakly focused using a lens (focal length 30 cm) and applied on the film side of the samples at a 45° angle of incidence. The fundamental

LB Films of Thiohelicenebisquinones

Figure 4. A typical dependence of the p-polarized second-harmonic signal as a function of the quarter-wave plate rotation angle. In this measurement, the sample was an X-type LB film with 16 layers of THBQ molecules and rotated -15° with respect to the orientation when the film was deposited.

beam was initially p-polarized, but its state of polarization was varied by a continuously rotating quarter-wave plate. The detection of p- or s-polarized SH signals at 532 nm in the transmitted direction behind the sample was selected by placing a Glan polarizer in front of a photomultiplier tube detector. Appropriate spectral filters were used before and after the sample to discard SH light generated in optical components before the sample and the fundamental beam after the sample. In addition, a 532 nm interference filter was used before the photomultiplier. The measurements were repeated for several azimuthal orientations of the sample. The optical setup was carefully aligned and tested beforehand to ensure that the measurement results could be reliably combined. For tests and alignment, we used a high-quality LB film of terthiophene-vinylbenzoate and a spin-coated and poled polymer film of octadecylamino cyanoazostilbene in PMMA. Both of these films are known to be achiral and isotropic in the plane of the substrate.45 We performed a series of four measurements, which yielded the values of f, g, and h in eq 2 and allowed one of them to be calculated in two different ways for this type of sample.46 Any differences must be due to measurement errors and reveal even small misalignments of the polarizers. In addition, since the samples are achiral, any handedness detected via different SH signal levels for left- and right-hand circularly polarized incident beams reveals misalignments that would make the optical setup to have handedness as a whole. Naturally, other basic tests, such as the quadratic dependence of the detected signal on the power of the fundamental beam, were performed to confirm that SHG was being detected. The polarization-dependent SHG from each sample was measured for eight different azimuthal orientations. Both s- and p-polarized detections were used, giving a total of 16 measurement curves per sample. A typical measurement curve is shown in Figure 4. Each curve was fitted to yield the measured f, g, and h values in eq 2. The data corresponding to each sample was combined with a physical model of the sample, including details such as film thickness and indices of refraction of both the glass substrate and the LB film. The thickness was estimated to be ∼40 nm for 16 molecular layers as measured by an optical profilometer, resulting in about 2.5 nm per molecular layer. For this particular molecule, the index of refraction of the LB film is not known precisely and was therefore assumed to be ∼1.55 at both wavelengths on the basis of previous measurements24 on LB films of slightly different HBQ molecules (see ref 35, structure 1). We verified through extensive simulations that our final results are not sensitive to small variations of the exact value of the index of refraction.

J. Phys. Chem. B, Vol. 112, No. 7, 2008 1943

Figure 5. AFM phase images of X-type (a) and Y-type (b) LB films with four layers of chiral THBQ molecules on silicon. The imaged area is 0.2 µm × 0.2 µm in both images. The phase scale is 2°s and 1° for images a and b, respectively. The X-type sample (a) clearly has more small-scale variation in the film and the orientation of the columnar aggregated structures.

The AFM measurements for imaging the sample surfaces were performed with a Nanoscope IIIa scanning probe microscope equipped with a JVT scanner in MultiMode (Digital Instruments, Inc, USA). The microscope was placed on an active vibration isolation table, which was further placed on a massive stone table to eliminate external vibrational noise. Silicon cantilevers were used for imaging. All images (512 × 512 pixels) were captured using the intermittent contact mode under ambient conditions (RH 35 ( 5%, T ) 26.3 ( 0.7 °C) without filtering. The free amplitude of the oscillating cantilever (off contact) was 30 ( 5 nm. The measurement procedure caused a shift in the resonance frequency, which was taken into account. The new resonance frequency for the cantilever in contact was determined and used as the operating frequency. A damping ratio (contact amplitude/free amplitude) of ∼0.6-0.8 was used for imaging. The scanning probe image processor software was used for the analysis of images (Image Metrology, Denmark). 4. Results and Discussion Initially, we assumed C2 symmetry with the y-axis along the dipping direction in our physical models. This symmetry is a reasonable assumption, at least for the X-type films, where deposition of all molecular layers in the same orientation could lead to top-bottom asymmetry (polar ordering). The C2 symmetry is also the one that has previously been commonly assumed when determining the nonlinear response of LB films of HBQ and THBQ molecules.25,26,29,30 The other possible symmetry is D2, which would be expected for Y-type films24,25 and includes 180° rotation about an in-plane axis as an additional symmetry operation. The differences in the susceptibility tensors between the two types are that only three independent tensor components can exist for D2, all associated with chirality, whereas C2 allows five additional components. In addition, the presence of the in-plane rotation axis fixes the directions of the in-plane coordinate axes for D2, whereas no symmetry operation exists to do this for C2. However, our results show that for samples of both X and Y types, the tensor components specific to C2 can be set simultaneously close to zero when the coordinate system of the determined tensor is rotated by angle ∆φ to optimal orientation using eqs 4 and 5. If D2 symmetry is assumed and the in-plane coordinates are properly oriented, the results obtained are almost as accurate as for C2. The relative values of the susceptibility components in the optimized coordinates, normalized to χ(2) xyz ) 1, are shown in Table 1. It is seen that only the components specific to the D2 symmetry have significant nonzero values.

1944 J. Phys. Chem. B, Vol. 112, No. 7, 2008

Siltanen et al.

TABLE 1: Relative SH Signal Amplitudes Erel(2ω),a the Azimuthal Differences ∆φ of the Dipping Direction and the y-Axis in Optimally Oriented Coordinates, and the Relative Values of the Nonvanishing Susceptibility Componentsb for X- and Y-Type Samples with Varying Numbers of Molecular Layersc thickness and type 16 layers X Erel(2ω) ∆φ χ(2) xyz χ(2) yxz χ(2) zxy χ(2) xxz χ(2) yyz χ(2) zxx χ(2) zyy χ(2) zzz

1.00 9° 1 -0.16 + 0.01i ((0.01) -0.84 - 0.23i ((0.01) 0.00 + 0.00i ((0.01) 0.00 + 0.02i ((0.01) 0.03 + 0.04i ((0.02) 0.05 + 0.00i ((0.04) -0.12 - 0.12i ((0.08)

8 layers Y

0.98 -1.5° 1 -0.16 + 0.01i ((0.01) -0.83 - 0.22i ((0.02) 0.01 + 0.01i ((0.01) 0.01 + 0.02i ((0.01) 0.01 + 0.01i ((0.02) 0.01 - 0.01i ((0.03) -0.10 - 0.09i ((0.06)

X

4 layers Y

0.51 17° 1

0.50 -3° 1

-0.14 + 0.01i ((0.01) -0.87 - 0.15i ((0.01) 0.02 + 0.01i ((0.01) 0.01 + 0.03i ((0.02) - 0.05 + 0.01i ((0.02) -0.01 - 0.02i ((0.04) 0.04 + 0.01i ((0.08)

-0.26 + 0.01i ((0.01) -0.78 - 0.14i ((0.03) 0.00 + 0.01i ((0.02) 0.03 + 0.03i ((0.03) 0.00 + 0.08i ((0.08) 0.05 + 0.03i ((0.14) -0.05 - 0.36i ((0.35)

X 0.29 2° 1 -0.33 + 0.02i ((0.01) -0.72 - 0.08i ((0.06) 0.00 - 0.02i ((0.06) 0.06 + 0.01i ((0.06) -0.01 - 0.02i ((0.06) 0.01 - 0.09i ((0.06) N/A

Y 0.14 5° 1 -0.55 + 0.00i ((0.17) -0.48 - 0.04i ((0.09) 0.17 + 0.09i ((0.02) 0.18 + 0.11i ((0.03) 0.14 + 0.02i ((0.10) 0.13 + 0.06i ((0.11) N/A

c The strongest signal is normalized to unity. b χ(2) xyz is normalized to unity, and C2 symmetry is assumed in analysis. For samples with only (2) four molecular layers, the most inaccurate component, χzzz, could not be reliably determined and was forced to vanish in the analysis (N/A). a

The numbers in the parentheses represent the estimated errors associated with the regression-based data analysis only. The actual errors including possible systematic imperfections in the setup are expected to be somewhat larger. The relative amplitudes of the SHG and the orientation of the in-plane symmetry axis with respect to the direction of vertical deposition for different samples is also given in the table. The first result to note is that the SH signal amplitudes have the expected linear dependence on the film thickness. Only the four-layer samples deviate from this, X type giving somewhat more, and the Y type, less signal than expected. Second, our results show that, independent of whether the LB film is of X or Y type, the overall symmetry of the film is not influenced and always appears to be D2. The results of the tensor analysis for all samples are remarkably similar; in particular, the tensors of the 8- and 16-layer samples with strong SHG signal levels are virtually identical. Only the 4-layer Y-type sample deviates from the other samples in quantitative terms. We believe that this is due to its low signal level and possible interference with a background signal from the substrate, which also leads to increased uncertainty in the values of the tensor components for this sample. Nevertheless, even this sample exhibits the same qualitative trend as the others. Note also the facts that the (2) component χ(2) xyz is not equal to -χyxz and that the component (2) χzxy is nonvanishing, which are associated with the strong anisotropy of the samples. In addition, all these components arise from the chirality of the sample, with any resonance effects favoring such chiral tensor components being unlikely because there is no significant absorption at the fundamental or secondharmonic wavelengths. The results of Table 1 also reveal an important difference in the structure of the X- and Y-type films. For Y-type films, the direction of the in-plane axis follows the direction of deposition within experimental uncertainty. However, for X-type films the orientation of the in-plane axis can deviate significantly from the dipping direction and is generally more random. Further information regarding the structural differences between the X- and Y-type films was obtained from AFM

measurements, whose results are shown in Figure 5. The columnar structure is visible in both images, but there is a striking difference in the macrostructure: in X-type samples, the orientations of the columns and precise thickness of the film varies in small domains, whereas in Y-type samples, the columns are all oriented in the same direction, and the surface appears to be very smooth. In this respect, the X-type films resemble those of horizontally deposited racemic films of HBQ,29 whereas Y-type films form even more-ordered structures than the nonracemic HBQ. Note also that the irregular domains of the X-type films need not induce the lower C2 symmetry, because the area illuminated with the fundamental beam is much larger than the domains. Therefore, if the high-level macrostructure and its domains have no effect on the nonlinear response, the nonuniformity of the higher-level ordering of the columns is averaged and does not induce lower symmetry. The width of a single column for both X- and Y-type films is ∼5 nm, slightly more than the spacing of the columns in LB films of HBQ molecules.32,34,36 This together with the thickness data from the profilometer measurements supports the conclusion that the molecules are formed in columnar aggregates that can be seen in the images, and that the columns must be stacked in a flattened hexagonal structure somewhat similar to the one earlier depicted for HBQ.36 The alignment of the columns is basically known to be sensitive to the environmental factors. Thus, the differences between the X- and Y-type samples can be explained by the details of the sample manufacturing process. Both types of samples are allowed to dry for at least 20 min after translating them up from the subphase, but the dipping procedure for the X-type samples accumulates only one new molecular layer on the substrate whereas for the Y-type samples, there will be two new layers. Apparently the Y-type deposition allows the THBQ columns in the two layers to mutually align themselves during the deposition, leading to a well-organized overall structure of the film. For X-type deposition, each layer is first allowed to dry for 20 min before the second layer is deposited on it. Thus, the alignment of the previous layer is already fixed, and no

LB Films of Thiohelicenebisquinones mutual orientation can take place during deposition, leading to a less organized overall structure of the film. The main result regarding the nonlinear response of THBQ LB films is that it is similar in magnitude and tensorial properties for both X- and Y-type films and independent of the differences of the two types in terms of their high-level ordering into aligned columns. The origin of the nonlinearity must therefore be the low-level stacked aggregation of the molecules into chiral columns whose macroscopic packing plays a lesser role. Note that the nonlinear properties of certain chiral polymers, such as polyisocyanides and polypeptides, can also arise from the helical stacking of the nonlinearly active side chains.12,18 5. Conclusions We have performed a detailed study of the second-order nonlinear optical and structural properties of X- and Y-type LB films of nonracemic thiohelicenebisquinone. The overall properties were found to be essentially independent of the film thickness and the deposition type. Both X- and Y-type films can be well-described by D2 symmetry, higher than the previously assumed C2 symmetry of similar samples. The relative values of the susceptibility tensor components were found to be nearly the same for all films, and all of them are associated with chirality. In addition, their absolute magnitudes for a given thickness are equal for X and Y types. This suggests that the nonlinear response of both types of films originates from similar molecular superstructures. However, the X- and Y-type films were found to differ significantly with respect to the orientation of the in-plane axis. For Y type, the direction of the axis follows the direction of vertical sample deposition. For the X type, on the other hand, the direction of the axis varies randomly and significantly between the films of different thicknesses, suggesting a higher level of structural ordering in Y-type samples as compared to X-type samples. This was confirmed by AFM measurements, which showed that the Y type exhibits uniform ordering into columnar structures, whereas similar structures in the X type are shorter and more randomly oriented. The similar nonlinear properties but different high-level ordering reinforces that the nonlinearity of the THBQ films originates from the low-level columnar aggregation of the molecules, with the higher-level structures playing a lesser role. In addition, within the columns, the molecules likely assume fairly random azimuthal orientations so that the columns themselves exhibit approximate D∞ symmetry. Acknowledgment. We greatly appreciate T. J. Katz for providing us with the THBQ material. The work was financially supported by the Academy of Finland (107009), by the Graduate School of Modern Optics and Photonics in Finland, and by the Nanophotonics Research and Development Program of the Ministry of Education in Finland. References and Notes (1) Armstrong, J. A.; Bloembergen, N.; Ducuing, J.; Pershan, P. S. Phys. ReV. 1962, 127, 1918. (2) Bloembergen, N.; Chang, R. K.; Jha, S. S.; Lee, C. H. Phys. ReV. 1968, 174, 813. Errata: Bloembergen, N.; Chang, R. K.; Jha, S. S.; Lee, C. H. Phys. ReV. 1968, 178, 1528. (3) Shen, Y. R. IEEE J. Select. Top. Quantum Electron. 2000, 6, 1375. (4) Heinz, T. F.; Tom, H. W. K.; Shen, Y. R. Phys. ReV. A: At., Mol., Opt. Phys. 1983, 28, 1883. (5) Plocinik, R. M.; Everly, R. M.; Moad, A. J.; Simpson, G. J. Phys. ReV. B: Condens. Matter Mater. Phys. 2005, 72, 125409. (6) Mitchell, S. A. J. Chem. Phys. 2006, 125, 044716. (7) Anceau, C.; Brasselet, S.; Zyss, J. Chem. Phys. Lett. 2005, 411, 98.

J. Phys. Chem. B, Vol. 112, No. 7, 2008 1945 (8) Simpson, G. J.; Rowlen, K. L. J. Am. Chem. Soc. 1999, 121, 2635. (9) Zhang, W.-K.; Wang, H.-F.; Zheng, D.-S. Phys. Chem. Chem. Phys. 2006, 8, 4041. (10) Bosshard, C.; Sutter, K.; Preˆtre, P.; Hullinger, J.; Flo¨rsheimer, M.; Kaatz, P.; Gu¨nter, P. In AdVances in Nonlinear Optics; Kajzar, F.; Garito, A. F., Eds.; Gordon and Breach: Basel, Switzerland, 1995; Vol. 1. (11) Langmuir-Blodgett Films; Roberts, G., Ed.; Plenum: New York, 1990. (12) Kauranen, M.; Verbiest, T.; Maki, J. J.; Persoons, A. J. Chem. Phys. 1994, 101, 8193. (13) Verbiest, T.; Kauranen, M.; Maki, J. J.; Teerenstra, M. N.; Schouten, A. J.; Nolte, R. J. M.; Persoons, A. J. Chem. Phys. 1995, 103, 8296. Erratum: Verbiest, T.; Kauranen, M.; Maki, J. J.; Teerestra, M. N.; Schouten, A. J.; Nolte, R. J. M.; Persoons, A. J. Chem. Phys. 1996, 104, 5713. (14) Kauranen, M.; Verbiest, T.; Persoons, A. J. Mod. Opt. 1998, 45, 403. (15) Hicks, J. M.; Petralli-Mallow, T. Appl. Phys. B: Lasers Opt. 1999, 68, 589. (16) Sioncke, S.; Verbiest, T.; Persoons, A. Mater. Sci. Eng. R 2003, 42, 115. (17) Fischer, P.; Hache, F. Chirality 2005, 17, 421. (18) Mitchell, S. A.; McAloney, R. A.; Moffatt, D.; Mora-Diez, N.; Zgierski, M. Z. J. Chem. Phys. 2005, 122, 114707. (19) Belkin, M. A.; Shen, Y. R. Int. ReV. Phys. Chem. 2005, 24, 257. (20) Kriech, M. A.; Conboy, J. C. Appl. Spectrosc. 2005, 59, 746. (21) Ji, N.; Ostroverkhov, V.; Belkin, M.; Shiu, Y. J.; Shen, Y. R. J. Am. Chem. Soc. 2006, 128, 8845. (22) Kauranen, M.; Maki, J. J.; Verbiest, T.; Van Elshocht, S.; Persoons, A. Phys. ReV. B: Condens. Matter Mater. Phys. 1997, 55, R1985. (23) Schanne-Klein, M. C.; Hache, F.; Brotin, T.; Andraud, C.; Collet, A. Chem. Phys. Lett. 2001, 338, 159. (24) Busson, B.; Kauranen, M.; Nuckolls, C.; Katz, T. J.; Persoons, A. Phys. ReV. Lett. 2000, 84, 79. (25) Kauranen, M.; Van Elshocht, S.; Verbiest, T.; Persoons, A. J. Chem. Phys. 2000, 112, 1497. (26) Sioncke, S.; Van Elshocht, S.; Verbiest, T.; Persoons, A.; Kauranen, M.; Phillips, K. E. S.; Katz, T. J. J. Chem. Phys. 2000, 113, 7578. (27) Guyot-Sionnest, P.; Chen, W.; Shen, Y. R. Phys. ReV. B: Condens. Matter Mater. Phys. 1986, 33, 8254. (28) Verbiest, T.; Kauranen, M.; Van Rompaey, Y.; Persoons, A. Phys. ReV. Lett. 1996, 77, 1456. (29) Verbiest, T.; Van Elshocht, S.; Kauranen, M.; Hellemans, L.; Snauwaert, J.; Nuckolls, C.; Katz, T. J.; Persoons, A. Science 1998, 282, 913. (30) Siltanen, M.; Cattaneo, S.; Vuorimaa, E.; Lemmetyinen, H.; Katz, T. J.; Phillips, K. E. S.; Kauranen, M. J. Chem. Phys. 2004, 121, 1. (31) Siltanen, M.; Kauranen, M. Opt. Comm. 2006, 261, 359. (32) Van Elshocht, S.; Verbiest, T.; de Schaetzen, G.; Hellemans, L.; Phillips, K. E. S.; Nuckolls, C.; Katz, T. J.; Persoons, A. Chem. Phys. Lett. 2000, 323, 340. (33) Verbiest, T.; Van Elshocht, S.; Persoons, A.; Nuckolls, C.; Phillips, K. E.; Katz, T. J. Langmuir 2001, 17, 4685. (34) Lovinger, A. J.; Nuckolls, C.; Katz, T. J. J. Am. Chem. Soc. 1998, 120, 264. (35) Phillips, K. E. S.; Katz, T. J.; Jockush, S.; Lovinger, A. J.; Turro, N. J. J. Am. Chem. Soc. 2001, 123, 11899. (36) Nuckolls, C.; Katz, T. J.; Verbiest, T.; Van Elshocht, S.; Kuball, H.-G.; Kiesewalter, S.; Lovinger, A. J.; Persoons, A. J. Am. Chem. Soc. 1998, 120, 8656. (37) Kauranen, M.; Verbiest, T.; Boutton, C.; Teerenstra, M. N.; Clays, K.; Schouten, A. J.; Nolte, R. J. M.; Persoons, A. Science 1995, 270, 966. (38) Verbiest, T.; Samyn, C.; Boutton, C.; Houbrechts, S.; Kauranen, M.; Persoons, A. AdV. Mater. 1996, 8, 156. (39) Van Elshocht, S.; Verbiest, T.; Busson, B.; Kauranen, M.; Snauwaert, J.; Hellemans, L.; Persoons, A.; Nuckolls, C.; Phillips, K. E.; Katz, T. J. Synth. Met. 2000, 115, 201. (40) Boyd, R. W. Nonlinear Optics; Academic: San Diego, 1992. (41) Popov, S. V.; Svirko, Y. P.; Zheludev, N. I. Susceptibility Tensors for Nonlinear Optics; IOP Publishing: Bristol, UK, 1995. (42) Maki, J. J.; Kauranen, M.; Verbiest, T.; Persoons, A. Phys. ReV. B: Condens. Matter Mater. Phys. 1997, 55, 5021. (43) Maki, J. J.; Kauranen, M.; Persoons, A. Phys. ReV. B: Condens. Matter Mater. Phys. 1995, 51, 1425. (44) Golub, G. H.; Van Loan, C. F. SIAM J. Numer. Anal. 1980, 17, 883. (45) Cattaneo, S.; Rouhento, C.; Vuorimaa, E.; Efimov, A.; Lemmetyinen, H.; Kauranen, M. Chem. Phys. Lett. 2003, 377, 306. (46) Wang, F. X.; Siltanen, M.; Kauranen, M. Phys. ReV. B: Condens. Matter Mater. Phys. 2007, 76, 085428.