Versatile Method for Chiral Recognition by the Quartz Crystal

Dec 23, 2008 - Quartz Crystal Microbalance Technique for Analysis of Cooling Crystallization. Li-Shang Liu , JongMin Kim , Sang-Mok Chang , Guang Jin ...
0 downloads 0 Views 930KB Size
Download PDF
648

Langmuir 2009, 25, 648-652

Versatile Method for Chiral Recognition by the Quartz Crystal Microbalance: Chiral Mandelic Acid as the Detection Model Hui-Shi Guo,†,⊥ Jong-Min Kim,‡ Seung-Jin Kim,‡ Sang-Mok Chang,‡ and Woo-Sik Kim*,§ Department of Chemistry, Shaoguan UniVersity, Shaoguan 512005, PR China, Department of Chemical Engineering, Dong A UniVersity, Busan 604-714, Korea, and Department of Chemical Engineering, Kyung Hee UniVersity, Kyungki-do 446-701, Korea ReceiVed October 10, 2008. ReVised Manuscript ReceiVed December 14, 2008 Chiral recognition is considered to be the most important, fundamental basis in the development of separation technology for chiral isomers in the pharmaceutical and biotechnology fields. However, the selective detection of individual enantiomers is still one of the most difficult analytical tasks because of the close similarity of the molecular configurations between chiral isomers. This study presents a versatile vapor-diffused molecular assembly (VDMA) reaction approach for chiral recognition by the quartz crystal microbalance (QCM). Chiral L/D-mandelic acid (MA) was used as the detection model, and L-phenylalanine (L-Phe) was used as the selector. The construction of the L-Phe-modified QCM sensor involved a four-step layer-by-layer assembly procedure. Each modification step was analyzed by cyclic voltammetry, the contact angle, and a resonance frequency measurement. The chiral recognizability of the L-Phe-modified QCM sensor to L-mandelic acid was then examined by resonance frequency measurement using the novel VDMA technique and also investigated by atomic force microscope (AFM) measurements. A chiral discrimination factor of up to ∼9 between L- and D-MA on the L-Phe-modified QCM sensor was obtained by using this gaseous-phase reaction technique. AFM results also showed obvious selective aggregation of L-MA on the L-Phe-modified surface but no noticeable aggregation of D-MA during the VDMA reaction. Both of the QCM and AFM results confirmed the usefulness of this proposed VDMA technique for the study of chiral recognition. The main advantage of the proposed method is that it offers a universal simple application scheme for the QCM detection of small resonance frequency changes due to chiral molecular recognition by a chiral selector immobilized on the QCM sensor surface.

Introduction Enantiomers can have very different physiological effects on human beings.1 Where one enantiomer is therapeutic, the other may be toxic. For example, (R)-thalidomide has an antinausea effect, yet the S enantiomer is teratogenic and was thought to be responsible for causing deformities in babies born after their mothers had taken the racemic drug.2 Therefore, chiral recognition plays a very important role in the fields of pharmaceuticals and biotechnology.3 During the past several decades, numerous studies have been focused on chiral analysis and the separation of chiral compounds resulting in various approaches to chiral recognition, including microcalorimetry,4 circular dichroism,5 UV-vis absorption spectrometry,6 infrared spectroscopy,7,8 various types of chromatography,9-11 capillary electrophoresis,12 nuclear magnetic resonance,13 and mass spectrometry.14 * Corresponding author. Phone: 0082-31-201-2970. Fax: 0082-31-2732971. E-mail: [email protected]. † Shaoguan University. ‡ Dong A University. § Kyung Hee University. ⊥ Current address: Kyung Hee University. E-mail: [email protected].

(1) Bentley, R. Chem. Soc. ReV. 2005, 34, 609–624. (2) Webster, W. S.; Brown-Woodman, P. D.; Ritchie, H. E. Int. J. DeV. Biol. 1997, 41, 329–335. (3) Izake, E. L. J. Pharm. Sci. 2007, 96, 1659–1676. (4) Liu, Y.; Yang, E. C.; Yang, Y. W.; Zhang, H. Y.; Fan, Z.; Ding, F.; Cao, R. J. Org. Chem. 2004, 69, 173–180. (5) Muranaka, A.; Yoshida, K.; Shoji, T.; Moriichi, N.; Masumoto, S.; Kanda, T.; Ohtake, Y.; Kobayashi, N. Org. Lett. 2006, 8, 2447–2450. (6) Gao, F.; Ruan, W. J.; Chen, J. M.; Zhang, Y. H.; Zhu, Z. A. Spectrochim. Acta, Part A 2005, 62, 886–895. (7) Bieri, M; Bu¨rgi, T. ChemPhysChem 2006, 7, 514–523. (8) Wirz, R.; Bu¨rgi, T.; Lindner, W.; Baiker, A. Anal. Chem. 2004, 76, 5319– 5330. (9) Roussel, C.; Del Rio, A.; Pierrot-Sanders, J.; Piras, P.; Vanthuyne, N. J. Chromatogr., A 2004, 1037, 311–328.

The quartz crystal microbalance (QCM) is well known for its remarkable mass sensitivity, low cost in analytical applications, and compactness for potential laboratory-on-a-chip use. As a material is adsorbed onto the surface of a quartz crystal, the oscillation frequency of the crystal is reduced because of an increase in the surface mass. This decrease in the frequency of an AT-cut quartz crystal is related via the Sauerbrey equation to the mass of the material adsorbed.15 QCMs have already been widely used in DNA analysis, pharmaceutical detection, microorganism assays, and nucleic acid and enzyme determination.16-20 Thus, considering its high sensitivity, the QCM is also a very powerful tool for the study of molecular recognition and chiral discrimination if chiral recognition of the sensitive layer can be realized.21,22 With respect to the facility and universal applicability, most QCM applications for biomacromolecules are fulfilled in the liquid phase.16-20 However, chiral recognition using the QCM (10) Xu, Y. L.; Liu, Z. S.; Wang, H. F.; Yan, C.; Gao, R. Y. Electrophoresis 2005, 26, 804–811. (11) Rizvi, S. A.; Shamsi, S. A. Anal. Chem. 2006, 78, 7061–7069. (12) Takatsy, A.; Hodrea, J.; Majdik, C.; Irimie, F. D.; Kilar, F. J. Mol. Recognit. 2006, 19, 270–274. (13) Wang, H.; Cao, R.; Ke, C. F.; Liu, Y.; Wada, T.; Inoue, Y. J. Org. Chem. 2005, 70, 8703–8711. (14) Di Tullio, A.; Reale, S.; De Angelis, F. J. Mass Spectrom. 2005, 40, 845–865. (15) Sauerbrey, G. Z. Z. Phys. 1959, 155, 206–222. (16) Takahashi, S.; Matsuno, H.; Furusawa, H.; Okahata, Y. Chem. Lett. 2007, 36, 230–231. (17) Dickert, F. L.; Lieberzeit, P.; Hayden, O. Anal. Bioanal. Chem. 2003, 377, 540–549. (18) Nie, L. B.; Yang, Y.; Li, S.; Wang, J. Q.; Hou, Q. L. J. Nanosci. Nanotechnol. 2007, 7, 2927–2929. (19) Yao, C. Y.; Chen, Q. H.; Chen, M.; Zhang, B.; Luo, Y.; Huang, Q.; Huang, J. F.; Fu, W. L. J. Nanosci. Nanotechnol. 2006, 6, 3828–3834. (20) Dickert, F. L.; Hayden, O.; Halikias, K. P. Analyst 2001, 126, 766–771. (21) Cao, L.; Zhou, X. C.; Li, S. F. Y. Analyst 2001, 126, 184–188. (22) Maier, N. M.; Lindner, W. Anal. Bioanal. Chem. 2007, 389, 377–397.

10.1021/la803364v CCC: $40.75  2009 American Chemical Society Published on Web 12/23/2008

Letters

Langmuir, Vol. 25, No. 2, 2009 649

in the liquid phase is inherently difficult because it is much more difficult to achieve a sufficiently stable resonance signal in the liquid phase than in the gas phase for two main reasons.23,24 First, solvents are denser and more viscous than air, leading to a greater transfer of acoustic energy from the crystal to the surrounding medium and resulting in higher-energy dissipation. Second, because the quartz crystal resonator is itself an electric device, some form of sealing is needed to prevent electrical short-circuiting between the electrodes on the opposite surfaces during measurement in the aqueous phase. However, any physical constraint imposed upon the free, uniform mechanical oscillation of the crystal will lead to further dissipation of energy. The higher the energy dissipation, the lower the signal-to-noise ratio (S/N) that can be achieved during QCM detection. Therefore, as far as the selectivity is concerned, chiral recognition using the QCM in the gas phase would seem to be a good choice. This study presents a vapor-diffused molecular assembly (VDMA) reaction approach for the highly selective recognition of chiral L/D-mandelic acid (MA) based on the QCM by using L-phenylalanine (L-Phe) as the selector. The proposed method offers a universal simple application scheme for the QCM detection of chiral molecular recognition by a chiral selector immobilized on the QCM sensor surface.

Figure 1. (A) Schematic diagram of QCM detection: (1) N2 gas supplier, (2) heat exchanger, (3) inlet valve, (4) detection chamber, (5) QCM sensor, (6) constant-temperature oven, (7) outlet valve, (8) oscillator, (9) QCM 922 frequency counter, and (10) computer. (B) Schematic diagram of the vapor-diffused molecular assembly reaction device: (1′) constant-temperature oven, (2′) reaction chamber, (3′) microsyringe, (4′) L-Phe-modified QCM sensor, and (5′) reaction solution.

Experimental Section Materials and Reagents. L-Mandelic acid (L-MA), D-mandelic acid (D-MA), 16-mercapto-hexadecanoic acid (MHA), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC), Nhydroxysuccinimide (NHS), tyramine, and p-amino-L-phenylalanine hydrochloride (4-NH2-L-Phe) were purchased from Sigma-Aldrich (ACS grade). All the other reagents were commercially available and of analytical reagent grade. Deionized water was used throughout all experiments. Apparatus. The QCM measurements were carried out using ATcut quartz crystals with gold electrodes (areas, 0.2 cm2) and a resonance frequency of 9 MHz (Seiko EG&G, Chiba, Japan). The QCM frequencies were measured using a quartz crystal microbalance QCM 922 (Princeton Applied Research). A conventional atomic force microscope (AFM) (Nanoscope 3a, Digital Instruments) was used to visualize the surface morphology. A normal tapping mode Si cantilever with an oscillation frequency of 127 kHz and a spring constant of 13 N/m (SI DF-20, SII Nanotechnology Inc.) was used in all AFM experiments, and the AFM measurements were performed using intermittent contact mode (tapping mode) with a scanning speed of 1 Hz (512 × 512 data format) in the regular atmosphere. Cyclic voltammetry (CV) was carried out using an E-Corder 401 interface and potentiostat (eDAQ Technology Corporation). A platinum wire was used as the counter electrode, and an Ag/AgCl electrode was used as the reference. The water contact angle was measured using a Phoenix 300 contact angle meter (Surface Electro Optics Co., Korea). Immobilization of Chiral Selectors on QCM Sensors. The key step for selectively sensing chiral compounds is to build a chiral selector surface with recognition sites for the enantiomers. In previous diastereomeric crystallization work, the current authors succeeded in separating L- and D-mandelic acid with L-Phe as the selector (unpublished results). Therefore, in this study, mandelic acid was selected as the chiral recognition model, and L-Phe was used as the chiral selector. The immobilization of L-Phe on the QCM sensors was achieved using a four-step layer-by-layer process, similar to that in a previous report.25 First, a monolayer of MHA was self(23) Sota, H.; Yoshimine, H.; Whittier, R. F.; Gotoh, M.; Shinohara, Y.; Hasegawa, Y.; Okahata, Y. Anal. Chem. 2002, 74, 3592–3598. (24) Kim, W. S.; Lee, H. Y.; Kawai, T.; Kang, H. W.; Muramatsu, H.; Kim, I. H.; Park, K. M.; Chang, S. M.; Kim, J. M. Sens. Actuators, B 2008, 129, 126–133. (25) Zhang, S.; Ding, J. J.; Liu, Y.; Kong, J. L.; Hofstetter, O. Anal. Chem. 2006, 78, 7592–7596.

Figure 2. Schematic process of vapor-diffused molecular assembly reaction of chiral mandelic acid on the L-Phe-modified QCM sensor surface. (A) L-Phe-modified QCM sensor surface, (B, C) solute molecular (L- or D-MA) evaporation with the solvent (H2O), diffusion to the L-Phemodified QCM sensor surface, and then selective chiral recognition reaction on the L-Phe-modified QCM sensor surface, and (D) final selectively assembled surface after the selective reaction of L-MA with immobilized L-Phe.

assembled on the QCM sensor surface through Au-S bonding. Then, the carboxyl groups of the immobilized MHAs were activated by using EDC/NHS. After that, tyramine was immobilized by forming an amide structure through its amino group. Finally, diazotizated 4-NH2-L-Phe was combined by using diazo coupling reaction. Refer to Supporting Information for details. QCM Measurements. The QCM detection was performed in a 100 mL detection chamber, which was placed in a constanttemperature oven. The oven temperature was controlled at 25 °C, and the detection was performed under N2 atmosphere conditions. The QCM sensor frequency was recorded until the baseline was stabilized with a fluctuation of (1 Hz. A schematic diagram of the QCM detection device is outlined in Figure 1A. The mass change on the QCM sensor was calculated according to the Sauerbrey equation.15 Experimental Setup for Vapor-Diffused Molecular Assembly Reaction. Because most chiral molecules exist in a solid or liquid state under normal conditions, a vapor-diffused molecular assembly (VDMA) technique is used to change the chiral molecules into a vapor to facilitate chiral recognition in the gas phase. As illustrated in Figure 2, the essential idea of the VDMA reaction technique is mass transfer through small solvent particles.24,26 Figure 2A shows the L-Phe-modified QCM sensor surface; Figure 2B,C shows the solute molecular (L- or D-MA) evaporation with the solvent (H2O) and diffusion to the L-Phe-modified QCM sensor surface and then selective chiral recognition reaction on the L-Phe-modified QCM sensor surface; and Figure 2D shows the final selectively assembled surface based on the selective reaction of L-MA with the QCMsensor-immobilized L-Phe. A schematic diagram of the VDMA reaction device is presented in Figure 1B. As shown, the modified QCM sensor was fixed in a (26) Kim, W. S.; Kim, S. J.; Park, J. J.; Chang, S. M.; Kim, J. M. J. Phys. Chem. Solids 2008, 69, 1422–1427.

650 Langmuir, Vol. 25, No. 2, 2009

Figure 3. Cyclic voltammograms initiated at -0.4 V for the MHAcoated gold sensor in 0.5 M KOH. (A) First scan and (B) second scan. The scan rate was 100 mV/s.

100 mL reaction chamber that was placed in an oven with a constant temperature of 25 °C. Next, 1 mL of the reaction solution was carefully injected into the reaction container using a microsyringe. The sample then evaporated during the VDMA reaction period and produced a chiral recognition reaction on the L-Phe-modified chiral surface. Investigation of Surface Morphology Using Atomic Force Microscopy. Mica is widely used as an AFM substrate because of its atomic-level flatness over areas of a few micrometers. Because there was no noticeable difference between the QCM gold surface and the mica surface with the self-assembled thiol-containing chemicals,27 a freshly cleaved mica surface was used as the sample substrate for AFM imaging instead of the QCM sensors. The immobilization of L-Phe on the mica surface and the VDMA reaction of L- or D-mandelic acid on the chiral-selector-modified mica surface were performed using the same procedure as described previously. After the modification and the VDMA reaction, the substrate was dried with N2 gas and used directly as the AFM sample.

Results and Discussion Characterization of 4-NH2-L-Phe-Modified QCM Sensors. To monitor the changes on the gold surface during the modification process, several techniques were used to characterize the surface after each modification step, as follows. Figure 3 shows the cyclic voltammetry curves for the MHAcoated gold sensor in the 0.5 M KOH supporting electrolyte. The first scan showed a large cathodic wave with a peak current at about -0.93 V resulting from the one-electron reduction desorption of MHA from the gold electrode surface.28 No corresponding anodic peak was found, indicating that the electron reaction was irreversible. Meanwhile, in the second scan, there was almost no cathodic peak, signifying that most of the immobilized MHA on the sensor had been desorbed from the electrode surface by the electrochemical reduction during the first scan. The charge associated with the electrode reaction in the first cathodic scan, Qrd, was found to be 6.2 × 10-5 C/cm2, which translated to a surface coverage,29 Γrd, of 6.4 × 10-10 mol/cm2. The fabrication of the 4-NH2-L-Phe-modified QCM sensor was also confirmed by cyclic voltammetry in a fairly reversible redox couple of [Fe(CN)6]3-/4- in a 5 mM K3Fe(CN)6/100 mM KCl solution. As seen in Figure 4, the cyclic voltammogram for (27) Kim, J.; Yamasaki, R.; Park, J.; Jung, H.; Lee, H.; Kawai, T. J. Biosci. Bioeng. 2004, 97, 138–140. (28) Imabayashi, S.-i.; Iida, M.; Hobara, D.; Feng, Z. Q.; Niki, K.; Kakiuchi, T. J. Electroanal. Chem. 1997, 428, 33–38. (29) Walczak, M. M.; Popenoe, D. D.; Deinhammer, R. S.; Lamp, B. D.; Chung, C.; Porter, M. D. Langmuir 1991, 7, 2687–2693.

Letters

Figure 4. Cyclic voltammograms recorded in solution of 5 mM [Fe(CN)6]3-/4- containing 0.1 M KCl with a bare QCM sensor (A), an MHA-coated sensor (B), and an MHA-coated sensor after electrochemical reduction in 0.5 M KOH (C). The scan rate was 100 mV/s, and the voltage ranged from 0 to -0.5 V (vs Ag/AgCl). Table 1. Frequency and Mass Changes Resulting from Stepwise Modification of the Sensor Surface, as Determined by QCM modified layera MHA tyramine 4-NH2-L-Phe

∆F (Hz)b

∆m (ng)c

RSD (%)

50.7 193.2 253.7

53.2 202.9 266.4

8.1 9.5 7.2

a The modification of the QCM sensor surface is shown in Figure S-1. Average results of six detections. c The mass change (∆m) was calculated using the Sauerbrey equation, as described in the Experimental Section.

b

the bare gold electrode showed clear redox peaks (curve A). In contrast, the self-assembled MHA monolayer on the gold electrode surface resulted in the disappearance of the redox peaks and lower current responses in both the anodic and cathodic processes (curve B), indicating that MHA had formed a covalently linked pinhole-free, densely packed monolayer on the gold electrode surface30 and that the electron transfer of [Fe(CN)6]3-/4- on the electrode was totally blocked. No redox peaks were found after the electrode was treated with EDC/NHS, tyramine, and 4-NH2L-Phe (figure not shown here). After the MHA-coated sensor had been treated by electrochemical reduction in 0.5 M KOH, the redox peaks of [Fe(CN)6]3-/4- appeared again (curve C), demonstrating the desorption of MHA and the renewal of the gold sensor surface. The contact angles obtained with deionized water after each modification step were also measured to show evidence of surface modification. The bare QCM sensor surface exhibited slight hydrophobicity with a contact angle (θ, mean ( S.D.) of 91.0 ( 0.9°. A contact angle decrease of about 38.7° was found after the gold surface was coated with MHA, indicating a much higher hydrophilic property due to the high hydrophilicity of the carboxyl end group of MHA. After tyramine modification, the contact angle increased again to 72.2 ( 1.6°, indicating an increase in the hydrophobicity of the surface due to the modification. Further modifications with 4-NH2-L-Phe produced a slight change in the hydrophilicity of the surface as the contact angle changed to 69.3 ( 1.3°. Table 1 shows the frequency changes detected by the QCM and the mass changes calculated by the Sauerbrey equation after the stepwise modification of the sensor surface. As seen, after MHA immobilization, the frequency of the QCM sensor decreased (30) Finklea, H. O.; Avery, S.; Lynch, M.; Furtsch, T. Langmuir 1987, 3, 409–413.

Letters

Figure 5. Response of the 4-NH2-L-Phe-modified QCM sensor exposure to (A) 1 mL of 0.05 mol/L L-mandelic acid, (B) 0.5 mL of 0.05 mol/L L-mandelic acid + 0.5 mL of 0.05 mol/L D-mandelic acid, (C) 1 mL of 0.05 mol/L D-mandelic acid, and (D) 1 mL of deionized H2O for different time periods at 25 °C. The volume of the reaction chamber was 100 mL, the detecting temperature was controlled at 25 °C, and the detection was performed under N2 atmosphere conditions.

50.7 Hz, corresponding to 53.2 ng of materials loaded onto the sensor surface and a surface coverage, ΓQCM, of 4.6 × 10-10 mol/cm2, which was close to the value detected by the electrochemical reduction method. As such, the frequency decrease and mass increase with every fabrication step, as determined by the QCM, confirmed the stepwise formation of the MHA, tyramine, and L-Phe layers on the sensor surface.25 All of these cyclic voltammetry, contact angle, and QCM detection results confirmed the successful modification of L-Phe on the sensor surface. Selective Response of L-Mandelic Acid on the Modified QCM Sensor. The selective response of L-MA on the L-Phemodified QCM sensor was tested using the VDMA method. The sensing and detection process was as follows. Before sensing, the frequency of the modified QCM sensor (F1) was detected in a detection chamber under N2 atmosphere conditions at 25 °C, as described in the Experimental Section. The modified QCM sensor was then transferred to the reaction chamber, as shown in Figure 1B, and 1 mL of the reaction solution was carefully injected into the 100 mL reaction container. Following a certain time period for the VDMA reaction, the sensor was moved back to the detection chamber to detect the frequency (F2), and the frequency change (∆Fsens) was calculated as ∆Fsens ) F2 - F1. Meanwhile, a bare gold sensor (unmodified QCM sensor) was also treated with the same sensing and detection procedure under the same conditions using 1 mL of deionized H2O as a substitute for the VDMA reaction solution, and the frequency change after the VDMA reaction in deionized H2O (∆F0) was used as the frequency change reference. The normalized frequency change (∆F) was defined as follows: ∆F ) ∆Fsens - ∆F0. Figure 5 shows a plot of the normalized frequency changes of the modified sensor (∆F) versus time with respect to the chiral mandelic acid (L-MA, D-MA, and a racemic mixture of MA). As shown in the Figure, the frequency response of the modified QCM sensor to L-MA decreased almost linearly for the first 5 days (curve A), indicating that before the VDMA equilibrium was reached a longer reaction time allowed more L-MA molecules to attach to the sensor surface, giving larger frequency changes. After a specific VDMA period (g5.5 days), the frequency changes tended to level off, indicating that reaction equilibrium had been achieved.

Langmuir, Vol. 25, No. 2, 2009 651

To confirm that the observed frequency changes were indeed based on the specific interaction between the sensor-immobilized L-Phe and the L-MA and not simply caused by nonspecific adsorption, experiments were performed with deionized H2O, D-MA, and the racemic mixture of MA. As shown by curve D in Figure 5, no frequency changes were detected with deionized H2O. For D-MA, the frequency response of the modified QCM sensor decreased only a little (about 9 Hz) during the first 2 days and then became almost stable after 2.5 days or more (Figure 5 curve C), indicating that only a very small amount of D-MA was attached to the modified sensor surface. For the racemic mixture of mandelic acid, the frequency response of the modified sensor also decreased continuously with VDMA reaction time (Figure 5 curve B), indicating that a longer reaction time was conducive to more MA molecules attaching to the sensor surface, resulting in larger frequency changes. However, the frequency change for the racemic mixture of MA was much slower than that for pure L-MA. The mechanism involved in the different frequency changes seems to be slightly complicated, and this should be carefully studied in the future. No obvious frequency changes were found when sensing L-MA, D-MA, and a racemic mixture of MA on the bare (unmodified) QCM sensor using the same sensing procedure (not shown here). Therefore, these results indicate that the frequency changes observed with L-MA were due to specific host-guest interactions rather than nonspecific adsorption and that the L-Phe-modified sensor surface had selective chiral recognition ability for chiral mandelic acid. The chiral discrimination factor between L- and D-MA, RL-MA/D-MA ) ∆FL-MA/∆FD-MA, was found to be about 9. These results are consistent with the results reported by Okamura et al.,31 where the hydrogen bonding force between L-MA and L-Phe was found to be much stronger than that between D-MA and L-Phe. The chiral recognizability of L- and D-MA on the L-Phemodified QCM sensor had also been examined in the aqueous phase by integrating QCM with a flow injection analysis (FIA) system as described previously.32 However, no differences in frequency shift were obtained after injecting L- and D-MA (not shown here). This was most probably due to the frequency shifts caused by the chiral recognition of L- or D-MA on the L-Phemodified QCM sensor in the solution phase that were too weak to be discriminated, which may be due to both the instability of the QCM resonance frequency caused by the high motional resistance in the liquid phase33 and the relative complexity of the binding mechanism between MA and L-Phe. For a better understanding of the selective binding of L-MA on the L-Phe-modified surface, a careful analysis of the relation between the VDMA reaction time and the surface topographic images was conducted. Figure 6 shows the topographic image changes in the L-Phe-modified mica surface after the VDMA reaction with L-MA for (A) 1 day, (B) 3 days, and (C) 6 days, and (D) shows the line profile changes of panel A (curve a), B (curve b), and C (curve c). As seen, the topographic images showed distinct changes for different VDMA reaction times. The root-mean-square surface roughness values (rms) were 1.03 nm after 1 day (A), 1.24 nm after 3 days (B), and 2.18 nm after 6 days (C), indicating a continuous progression of L-MA molecular aggregation on the chiral-selector-modified surface. In addition, the line profile analysis in Figure 6D also showed continuous L-MA aggregation on the chiral surface during the VDMA reaction (31) Okamura, K.; Aoe, K. I.; Hiramatsu, H.; Nishimura, N.; Sato, T.; Hashimoto, K. Anal. Sci. 1997, 13, 315–317. (32) Uttenthaler, E.; Ko¨sslinger, C.; Drost, S. Anal. Chim. Acta 1998, 362, 91–100. (33) Kim, J. M.; Chang, S. M.; Muramatsu, H. Polymer 1999, 40, 3291–3299.

652 Langmuir, Vol. 25, No. 2, 2009

Figure 6. Surface morphology obtained after VDMA reaction with L-MA for (A) 1 day, (B) 3 days, and (C) 6 days plus (D) line profiles for panels A (curve a), B (curve b), and C (curve c).

time. In this letter, the explanation of aggregation is still not clear. Several factors may be involved in the aggregation phenomena such as the role of volatile solvent, the physics of the isomers in the isolated solvent drops, and different molecular assembly structure in the 2-D surfaces.34 To confirm that the observed image changes were indeed based on the specific interaction between mica-immobilized L-Phe and L-MA, a control AFM imaging experiment was performed using D-MA as the VDMA reaction substance. As shown in Figure 7, no distinct changes in the surface morphology were observed after the VDMA reaction with D-MA for 1 day (A), 3 days (B), and 6 days (C). Plus, the line profile analysis (Figure 7D) and surface roughness values did not show any noticeable changes (rms ) 1.15 nm (A), rms ) 1.19 nm (B), and rms ) 1.02 nm (C)). Thus, the AFM results were consistent with those obtained using the QCM and confirmed the usability of VDMA reaction method in the study of chiral recognition. Reusability and Stability of Modified QCM Sensors. When washing the used sensors in deionized H2O after every sensing cycle to renew the modified sensor, after four sensing cycles, the sensors were found to retain 80% of their initial response to 0.05 M L-MA. One explanation is that L-Phe was immobilized on the sensor surface via chemical bonding whereas L-MA was attached to the surface by hydrogen bonding between L-MA and sensorimmobilized L-Phe. Thus, the chemical bonding force was strong enough to prevent the decomposition of L-Phe from the linker of the sensor surface. In contrast, the hydrogen bonding was very weak, so it was destroyed by washing, allowing the renewal of the modified sensor. The modified QCM sensor also remained effective for sensing chiral mandelic acid without a significant loss of its sensing (34) Zhang, J; Gesquie`re, A; Sieffert, M; Klapper, M; Mu¨llen, K; De Schryver, F. C.; De Feyter, S. Nano Lett. 2005, 5, 1395–1398.

Letters

Figure 7. Surface morphology obtained after VDMA reaction with D-MA for (A) 1 day, (B) 3 days, and (C) 6 days plus (D) line profiles for panels A (curve a), B (curve b), and C (curve c).

sensitivity after being stored dry at room temperature for at least 3 months. This good long-term stability may also be attributed to the strong chemical bonding force between gold-MHAtyramine and L-Phe. Furthermore, the complete renewal of the gold surface of the sensor was achieved by treatment with piranha solution or electrochemical reduction in 0.5 M KOH to remove the assembled molecular layers.

Conclusions A versatile vapor-diffused molecular assembly method was presented for chiral recognition using the QCM. A chiral sensor for mandelic acid was constructed by immobilizing L-Phe on a QCM sensor surface using a layer-by-layer assembly procedure. The chiral recognizability of the L-Phe-modified QCM sensor to L-mandelic acid was then examined using the proposed VDMA technique. The highly selective sensing of L-mandelic acid was also confirmed by AFM detection. As a result, the proposed method offers a universally simple solution for the QCM detection of small resonance frequency changes due to chiral molecular recognition by a chiral selector, which are relatively difficult to determine in the solution phase. Although this study examined only mandelic acid, this VDMA approach also has implications for more general use. In addition, because the sensing cycle in this study was time-consuming, further studies are now investigating the integration with a flow injection technique for the reaction process. Acknowledgment. This work was financially supported by funding from BK21, Korea. Supporting Information Available: Detailed process of the immobilization of L-Phe on the QCM sensor. This material is available free of charge via the Internet at http://pubs.acs.org.. LA803364V