Chiral-discriminative amino acid films prepared by vacuum vaporization and/or plasma processing

Abstract

D-Phenylalanine (D-Phe) films capable of discriminating optical isomers were prepared by dry processing: radiofrequency sputtering or vacuum vaporization with and without the assistance of an inductively coupled plasma (ICP). Chemical and spectroscopic analyses revealed that the majority of the constituents of the vaporized film are D-Phe molecules. However, the film prepared by ICP-assisted vaporization contains scarcely any D-Phe. The chiral-sensing properties of these D-Phe films were confirmed by sorption measurements on cyclic monoterpenes. Quartz crystal resonators coated with these films respond to the (−)-forms of limonene and α-pinene preferentially to their (+)-forms at the ppm level. This chiral preference was not observed for carvone, which induced comparable frequency shifts to limonene and α-pinene even at sub-ppm levels. The strong electrostatic interactions of the carbonyl group in carvone probably overcome the weak interactions of the discriminative optical isomers.

---

Introduction

Chiral discrimination is one of the essential issues in chemical sensing. For biological and medical applications in particular, it can be regarded as an indispensable technology. Chromatographic studies on chiral separation can offer guiding principles for chiral sensing.¹ There have even been some reports of enantioselective chemical sensing (chiral sensing) by using representative chiral selectors developed for liquid chromatography.^2–4

α-Amino acids are the building blocks of biopolymers: as receptor and transport proteins, which interact with guest molecules, they are frequently regarded as a paradigm of the chemical-sensing approach. One of the most significant functions of these proteins is that they can discriminate between optical isomers. The interaction behaviors of these biopolymers are inherently governed by the electrostatic and steric environment generated by the combination of α-amino acids. We have taken the biomimetic approach in chiral sensing by using the α-amino acids as chiral selectors. The amino acid films can be considered inherently to contain the chiral-sensing capabilities originating in the raw materials (amino acids).

In the development of many chemical-sensing systems, the chemical-sensing films embedded in transducing devices play crucial roles in detecting target materials. The deposition procedure, therefore, is an essential issue in constructing chemical-sensing systems. We have already reported a method of preparing quartz crystal resonator (QCR) sensors that can detect fragrant odorants at the sub-ppm level based on the radiofrequency sputtering of amino acids.^5–7 This type of sensor is also called a quartz crystal microbalance (QCM), quartz microbalance (QMB), thickness share mode (TSM) acoustic wave or bulk acoustic wave (BAW) sensor. In this study, we developed a unique vacuum vaporization apparatus that is equipped with a plasma ion source to reduce the degradation of the raw amino acids during the preparation of chemical-sensing films. The resulting amino acid films are expected to be capable of chiral discrimination. We used an inductively coupled plasma (ICP) ion source,^8,9 which can generate a well-stabilized weak plasma even in high-vacuum conditions, where α-amino acids can vaporize at a relatively low temperature (<200 °C) needed to suppress thermal degradation and dehydration reactions.¹⁰

In the first part of this paper, we report the preparation and structures of amino acid films based on vacuum vaporization methods. Using these amino acid films as chemical-sensing films, we then demonstrate the chemical-sensing performance of the QCR sensors, focusing on the chiral discrimination of fragrant odorants.

Experimental

Film preparation

Fig. 1 shows the vacuum vaporization apparatus we have developed. It is equipped with a molecular beam generator (Knudsen cell) and an ICP generator. D-Phenylalanine (D-Phe) was placed in the graphite crucible placed in the Knudsen cell. After the deposition chamber had been evacuated to below 10⁻⁴ Pa by using a turbomolecular pump, the graphite crucible was heated to 200 °C. The vapor pressure was kept below 3 × 10⁻² Pa during the deposition by using a mechanical booster pump. In some cases, the vaporized flow of D-Phe was exposed to the ICP, generated by a 13.56 MHz rf power supply tuned by a π-impedance matching network.^11,12 The atmosphere in the deposition chamber was monitored with a quadrupole mass spectrometer (REGA-202R; ULVAC, Kanagawa, Japan) and was kept at 2 × 10⁻⁵ Pa by using a turbomolecular pump.

Fig. 1 Schematic diagram of deposition apparatus.

We deposited D-Phe films on both sides of the mass transducer of a 9 MHz AT-cut quartz crystal plate with Au electrodes, to measure their gas-sensing performance. For structural analysis, the films were also deposited on non-doped silicon chips and quartz blocks. All film substrates were placed on the temperature-controlled substrate holder kept at 20 °C. The D-Phe film produced by ICP-irradiated vaporization is referred to as Vap-ICP film and the D-Phe film prepared without ICP irradiation is referred to as Vap film. We also examined a sputtered film prepared by radiofrequency sputtering of D-Phe, which has been reported previously.⁵ The thickness of all of the D-Phe films on each side of the QCR substrate was controlled at about 0.5 μm. The film thickness was determined by the frequency shift of the bare QCR induced by the film coating. The thickness (t) is related to the frequency shift (Δf) as follows:¹³

t/Å = −205.44 Δf/kHz (1)

This equation is widely used for film thickness monitoring in vacuum deposition.

Gas sorption analysis

The gas-sensing properties were measured by the gas flow system in which organic vapors were supplied by a diffusion tube method.^14,15 Synthetic air was used as the carrier gas at a flow rate of 0.2 L min⁻¹ in most cases. The QCRs were set in a 30 mL acrylic resin cell, which was kept at 25 °C. The detailed structure of this measuring system has been described previously.¹⁶ We can estimate the amounts of sorbed gas by using Sauerbrey’s equation,¹⁷ which proportionally relates the frequency shifts (Δf) to the mass loading (Δm) on the QCRs as follows:^18,19

Δm = −[A (ρ_q μ_q)^1/2/2F₀ ²] Δf (2)

where F₀ is the fundamental resonant frequency of the unloaded QCR, A is the electrode area (0.13 cm²), ρ_q is the density of quartz (2.65 g cm⁻³) and μ_q is the shear modulus of quartz (2.95 × 10¹¹ dyn cm⁻²). With these constants, we obtain

Δm/ng = −1.05 Δf/Hz (3)

This Sauerbrey equation is theoretically valid if the acoustic impedance of the loaded material is identical with that of quartz. Even though the acoustic impedance of the loaded material differs from that of quartz, its validity is claimed when the loaded mass is extremely small.^20,21

For classification of the sensors’ responses, we performed principal components analysis (PCA). The PCA calculation was carried out using commercial software (STATISTICA™; StatSoft, Tulsa, OK, USA).

Film characterization

For analyzing the surface film morphology, we used an atomic force microscope (NV2000; Olympus, Tokyo, Japan). The film surface was traced by a silicon cantilever in the contact mode (tapping mode). The sharpened tetrahedral tip was 10 μm high and had a vertex angle of less than 35°.

Infrared spectra were recorded on a Fourier transform infrared spectrophotometer (FTIR-5M; JASCO, Tokyo, Japan), attached to a microscope apparatus (MICRO-10), which was equipped with an Hg–Cd–Te detector. The wavenumber resolution was 4 cm⁻¹ and the measurements were conducted by the reflection method, using the D-Phe films coated on the QCR.

X-ray photoelectron spectroscopy (XPS) was performed with an ESCA-3400 instrument (Shimadzu, Tokyo, Japan) using Mg Kα radiation as the excitation source. To prevent structural changes, the X-ray power was reduced to 100 W and surface cleaning by ion etching was not carried out. The measurement pressure during XPS analysis was kept at <1.5 × 10⁻⁶ Pa.

Results and discussion

We used an amino acid analyzer (L-8500; Hitachi, Tokyo, Japan) to evaluate the contents of the raw materials in the vaporized films. This analyzer was based on a high-performance liquid chromatograph in which a spectroscopic detector was tuned for the ninhydrin reaction. Table 1 summarizes the concentration of phenylalanine and the percentage of its D-form in both vaporized films. The Vap film mainly consisted of the intact D-Phe used as the raw material to be vaporized. However, very little D-Phe existed in the Vap-ICP film. The optical isomerization from the D-form to the L-form was negligible in both cases.

Table 1 Concentration and percentage of D-form of phenylalanine in the vaporized films

D-Phenylalanine film	[Phenylalanine]/ wt.%	D-form of phenylalanine (%)
Vap	71.6	98.8
Vap-ICP	1.7	97.3

---

Infrared spectroscopy also gave a clue to the structural changes in the bulk of the vaporized films (Fig. 2). The spectral changes were significantly larger for the Vap-ICP film than for the Vap film when compared with the standard spectrum of D-Phe. The spectrum of the Vap film contained an absorption band at 2150 cm⁻¹, a characteristic signal of amino acids. Compared with the standard spectrum, the disappearance of fingerprint patterns can be observed in the vaporized films. The broadening of characteristic bands is prominent for the Vap-ICP film, indicating that ICP irradiation was responsible for structural changes in the vaporized D-Phe. However, taking into account the maintenance of the main bands, we think that the degree of structural changes was not very large, restricted within the disproportionation and condensation reactions.

Fig. 2 Infrared spectra of the vaporized films.

The existence of D-Phe molecules during the vaporization without ICP irradiation was confirmed by using a quadrupole mass spectrometer. Fig. 3 shows the time-dependent ion currents of the fragment ions of D-Phe and the molecular ions of water and carbon dioxide. All of the signal intensities increased as the graphite crucible was heated; the vaporization started after about 30 min. After attaining maximum intensity, all signals gradually decreased to the equilibrium levels. These results suggest the vaporization of D-Phe molecules concomitant with water and carbon dioxide. Besides vaporization of adsorbed molecules from the inner wall of the Knudsen cell, water and carbon dioxide may have been thermally generated by the condensation and elimination reactions of D-Phe, inducing its polymerization.

Fig. 3 Mass chromatogram for depositing the Vap film.

The chemical structure of the film surface was analyzed by XPS. Fig. 4 shows the C1s XPS spectra and the C:O:N atomic ratios. The spectral patterns were roughly the same; the main- and sub-bands were at about 285.3 and 288.8 eV, respectively. However, the relative intensity of the latter with respect to the former increased in the order standard, followed by Vap film, followed by the two Vap-ICP films.

Fig. 4 C1s XPS spectra and atomic ratios of D-Phe films.

There were small differences in the atomic ratios; the carbon and oxygen ratios shifted by 2–3% in opposite directions between the standard and the three vaporized films. The slight decrease in the oxygen ratios in the vaporized films may have been caused by the dehydration and decarboxylation reactions induced by vaporization and plasma-irradiation effects. Compared with the Vap film, the two Vap-ICP films had slightly larger nitrogen ratios, which should strengthen the sub-band at 288.8 eV. This nitridation is probably derived from reactive nitrogen atoms produced by plasma excitation of residual nitrogen molecules in the deposition chamber.

Surface-morphological information about the vaporized films was obtained using scanning electron microscopy; the round sponge-like masses (2–6 μm diameter) were segregated in the Vap film as shown in Fig. 5. The mean roughness of the Vap film was measured as 82 nm by using an atomic force microscope (Fig. 6). Conversely, the Vap-ICP film was extremely smooth and had a roughness of 1 nm (Fig. 7). This remarkable planarization caused by the ICP may have been due to plasma-induced structural changes in D-Phe molecules, which are subject to elimination and polymerization reactions that form densified molecular networks, in combination with etching and surface migration effects by energetic plasma species.

Fig. 5 Surface scanning micrograph of the Vap film.

Fig. 6 Atomic force micrograph of the Vap film.

Fig. 7 Atomic force micrograph of the Vap-ICP film.

Table 2 summarizes the frequency shifts of the D-Phe film-coated QCR sensors for typical organic vapors at 20 ppm. Compared with the two vaporized films, the sputtered film had a high sorption capability for all types of organic vapors. Corresponding to the polar characteristics of the D-Phe films, all QCR sensors had higher affinities for polar organic gases characterized by oxygenated moieties: hydroxy and carbonyl groups. Among the vaporized films, the Vap-ICP film was more absorptive for polar small molecules, such as methanol and ethanol. This may be attributed to the more polar character of the Vap-ICP film, which has a flattened surface with a higher nitrogen concentration than the other films. These surface-structural characteristics stimulate the nitrogen-driven polar character, due to the reduction of the activated-edge effects at the rugged surface.

Table 2 Frequency shifts (Hz) of the D-Phe film-coated QCR sensors for 20 ppm organic vapors

Organic vapor	Vap	Vap-ICP	Sputtered
Methanol	3.8	7.5	214.2
Ethanol	3.1	6.3	183.0
Butan-1-ol	14.2	4.8	71.8
2-Methoxyethanol	27.0	16.6	414.4
Acetaldehyde	8.7	7.5	108.8
Butan-2-one	1.6	2.4	47.3
Ethyl acetate	2.0	1.4	33.1
Dipropyl ether	2.5	4.1	32.1
Toluene	1.9	0.4	17.1
Chloroform	1.6	1.6	22.7
Tetrachloromethane	0.7	0.6	9.6

---

We measured the sensing properties of our QCR sensors, focusing on discrimination of the optical isomers of fragrant odorants classified as cyclic monoterpenes: α-pinene, limonene ((+)-form, lemon; (−)-form, orange)²² and carvone ((+)-form, caraway; (−)-form, spearmint).²³

Fig. 8 shows the concentration-dependent frequency shifts of QCRs for limonene (the representative response curves are shown in Fig. 9). All of the QCRs had a higher sensitivity for the (−)-form than the (+)-form. This suggests that the D-Phe films preserved the optical activity originating in the raw amino acid; the D-form is inclined to have a higher affinity for the (−)-form. Compared with the vaporized films, the sputtered film had higher absorption capacities for the (−)-form over the range of our measurements at ppm levels. Among the vaporized films, the Vap film had a slightly higher sorption capacity. These tendencies in the chiral preference and sorption capacity of our QCR sensors were also observed for α-pinene (Fig. 10).

Fig. 8 Concentration dependence of frequency shifts of D-Phe film-coated QCR sensors for limonene.

Fig. 9 Sorption–desorption curves of the Vap film-coated QCR sensor for (+)- and (−)-forms of limonene.

Fig. 10 Concentration dependence of frequency shifts of D-Phe film-coated QCR sensors for α-pinene.

However, it is noteworthy that none of the D-Phe films exhibited significant differences between the (+)-form and (−)-form in their gas-sorption capacities for carvone (as shown in Fig. 11). The molecular structure of carvone is similar to that of limonene with respect to carbon frameworks, but carvone has a carbonyl group, which is a prominent difference causing strong dipolar interactions. The strong interactions derived from the carbonyl group may overcome the weak interactions inducing chiral preference. This expectation is based on the fact that the carvone-induced frequency shifts were comparable to those induced by limonene, even though the measured concentrations of carvone were lower than those of limonene.

Fig. 11 Concentration dependence of frequency shifts of D-Phe film-coated QCR sensors for carvone.

As shown in Fig. 12, we performed a PCA for all three of the fragrant odorants measured by our QCR sensors using three kinds of overlayers: sputtered film and two Vap films distinguished by the position of the substrate holder. We used the normalized maximum frequency shifts and time constants as the calculation parameters evaluated from the response curves. The percentage contributions of principal components 1 and 2 were 64.5 and 19.4%, respectively. The three kinds of odorants were roughly differentiated. Moreover, optical isomers of α-pinene and limonene were clearly distinguishable, although those of carvone were indistinguishable.

Fig. 12 Scores plot of principal component analysis. The numbers indicate the concentrations of odorants in ppm.

Based on the solubility approach,²⁴ the sorbed gas molecules may induce changes in the mechanical properties of the sorbing films. Hence mechanical damping^25–30 is essential for chemical sensing the overlayers used in mass-sensitive acoustic resonators. This mechanical damping is a critical problem in QCR using the liquid phase, and a comprehensive review has been published recently in this journal.³¹ Using a network analyzer (HP-4195A; Yokogawa-Hewlett-Packard, Tokyo, Japan), we performed impedance analysis of the D-Phe films to determine whether the sorption of fragrant odorants would induce mechanical damping of the films. The detailed procedure has been described previously.¹⁶Table 3 gives the changes in admittance of our QCR sensors due to sorption of each fragrant odorant over 3 h at the highest concentrations shown in Figs. 8, 10 and 11. All of the admittances of the D-Phe film-coated QCRs are smaller than that of the mean value of the bare QCR, indicating that mechanical damping occurs with the D-Phe film coating. Among the D-Phe films, the admittance of the Vap film decreases substantially. This may be attributable to its coarse structure derived from the sponge-like masses. The admittance in any of the D-Phe film-coated OCRs changes negligibly on sorbing the odorants. These results suggest that the D-Phe films are not likely to soften or swell during gas sorption of fragrant odorants at ppm levels.

Table 3 Admittances (mS) of the D-Phe film-coated QCR sensors before and after sorption of fragrant odorants

D-Phenylalanine film	(−)-Limonene^a before/after	(+)-Limonene^a before/after	(−)-α-Pinene^b before/after	(+)-α-Pinene^b before/after	(−)-Carvone^c before/after	(+)-Carvone^c before/after
a The average value for five bare QCRs is 115 mS (ranging from 78 to 145 mS). 16 ppm vapor.
b 1 ppm vapor.
c 10 ppm vapor.
Vap	12.5/12.7	12.6/12.7	12.3/12.4	12.3/12.6	13.0/13.0	12.9/13.1
Vap-ICP	89.5/90.4	89.9/90.3	89.3/89.9	90.3/90.3	89.0/90.0	89.0/89.9
Sputtered	75.5/76.3	74.8/75.7	75.1/75.6	74.7/75.3	74.6/75.4	74.9/75.7

---

Conclusions

We have developed quartz resonator sensors capable of discriminating optical isomers by using chemical-sensing films prepared by vacuum vaporization of D-Phe at 200 °C. The vaporized film consisted of round sponge-like masses (2–6 μm diameter) and the D-Phe raw material content was 72%. ICP irradiation during vaporization of D-Phe flattened the resultant film surface extremely well and reduced the content of D-Phe to 1.7%. However, the differences in atomic ratios between these vaporized films and the raw D-Phe were small, suggesting that the structural changes in the vaporized film molecules are not drastic. Quartz crystal resonators coated with these films responded preferentially to the (−)-forms of limonene and α-pinene over their (+)-forms at the ppm level. This chiral preference was not observed for carvone, however, which induced comparable frequency shifts to limonene and α-pinene even at the sub-ppm levels. The strong electrostatic interactions of the carbonyl group in carvone probably overcome the weak interactions of discriminative optical isomers.

References

1. M. C. Ringo and C. E. Evans, Anal. Chem., 1998, 70, 315A.
2. K. Bodenhöfer, A. Hierlemann, J. Seemann, G. Guaglitz, B. Christian, B. Koppenhoefer and W. Göpel, Anal. Chem., 1997, 69, 3058.
3. K. Bodenhöfer, A. Hierlemann, M. Juzza, V. Schuring and W. Göpel, Anal. Chem., 1997, 69, 4017.
4. K. Bodenhöfer, A. Hierlemann, J. Seemann, G. Guaglitz, B. Koppenhoefer and W. Göpel, Nature (London), 1997, 387, 577.
5. I. Sugimoto, M. Nakamura and H. Kuwano, Anal. Chem., 1994, 66, 4316.
6. I. Sugimoto, M. Nakamura and H. Kuwano, Sens. Actuators, B, 1996, 35–36, 342.
7. I. Sugimoto, M. Nakamura and M. Mizunuma, Sens. Mater., 1999, 11, 57.
8. J. Hopwood, Plasma Sources Sci. Technol., 1992, 1, 109.
9. M. M. Turner, Phys. Rev. Lett., 1993, 71, 1844.
10. S. W. Fox and K. Harada, J. Am. Chem. Soc., 1960, 82, 3745.
11. J. H. Keller, J. C. Forster and M. S. Barnes, J. Vac. Sci. Technol., A, 1993, 11, 2487.
12. A. J. Lamm, J. Vac. Sci. Technol., A, 1997, 15, 2615.
13. W. H. Lawson, J. Sci. Instrum., 1967, 44, 917.
14. J. M. H. Fortuin, Anal. Chim. Acta, 1956, 15, 521.
15. G. A. Lugg, Anal. Chem., 1968, 40, 1072.
16. I. Sugimoto, Analyst, 1998, 123, 1849.
17. G. Z. Sauerbrey, Z. Phys., 1959, 155, 206.
18. G. G. Guilbault and J. M. Jordan, CRC Crit. Rev. Anal. Chem., 1988, 19, 1.
19. J. J. McCallum, Analyst, 1989, 114, 1173.
20. W. Ward and D. Buttery, Science, 1990, 249, 1000.
21. R. Lucklum, C. Behling and P. Hauptmann, in Extended Abstracts of the 7th International Meeting on Chemical Sensors, Beijing, 1998, p. 121..
22. L. Friedman and J. G. Miller, Science, 1971, 172, 1044.
23. G. F. Russell and J. I. Hills, Science, 1971, 172, 1043.
24. J. W. Grate, S. J. Patrash, M. H. Abraham and C. M. Du, Anal. Chem., 1996, 68, 913.
25. S. J. Martin and G. C. Frye, in IEE Ultrasonics Symposium Proceedings, IEEE, New York, 1991, p. 393..
26. P. Hauptmann, R. Lucklum, J. Hartmann, J. Auge and B. Adler, Sens. Actuators, A, 1993, 37, 309.
27. T. Okajima, H. Sakurai, N. Oyama, K. Tokuda and T. Ohsaka, Electrochim. Acta, 1993, 37, 747.
28. D. James, D. V. Thiel, G. R. Bushell, K. Busfield and A. Mackay-Sim, Analyst, 1994, 119, 2005.
29. Y. Yin and R. E. Collins, Thin Solid Films, 1995, 257, 139.
30. F. A. M. Davide, C. D. Natale, A. D’Amico, A. Hieerlemeann, J. Mitrovics, M. Schweizer, U. Weimar, W. Göpel, S. Marco and A. Pardo, Sens. Actuators, B, 1995, 26, 275.
31. B. A. Cavic, G. L. Hayward and M. Thompson, Analyst, 1999, 124, 1405.

---