Probing Molecular Recognition at the Solid–Gas Interface by Sum

Jul 20, 2016 - Identifying host–guest association in situ at the interface, between the substrate-bound receptors and the analyte-containing media, ...
0 downloads 4 Views 1MB Size
Letter pubs.acs.org/JPCL

Probing Molecular Recognition at the Solid−Gas Interface by SumFrequency Vibrational Spectroscopy Arianna Aprile,†,‡ Federica Ciuchi,‡ Roberta Pinalli,§ Enrico Dalcanale,§ and Pasquale Pagliusi*,†,‡ †

Department of Physics, University of Calabria, Ponte P. Bucci 31C, 87036 Rende, Cosenza, Italy CNR-Nanotec, LiCryL and Centre of Excellence CEMIF.CAL, Ponte P. Bucci 33B, 87036 Rende, Cosenza, Italy § Department of Chemistry, University of Parma, and INSTM, UdR Parma, Parco Area delle Scienze 17/A, 43124 Parma, Italy ‡

S Supporting Information *

ABSTRACT: Molecular recognition is among the most important chemical events in living systems and has been emulated in supramolecular chemistry, driven by chemical and biochemical sensing potential. Identifying host−guest association in situ at the interface, between the substrate-bound receptors and the analyte-containing media, is essential to predict complexation performances in term of the receptor conformation, orientation and organization. Herein, we report the first sum-frequency vibrational spectroscopy study of molecular recognition at the solid−gas interface. The binding capability of tetraquinoxaline cavitands toward volatile aromatic and aliphatic compounds, namely benzonitrile and acetonitrile, is investigated as test system. We prove the selective complexation of the receptors, organized in a solid-supported hybrid bilayer, toward aromatic compounds. Quantitative analysis allows to correlate the average orientations of the guest molecules and the host binding pockets, establishing “on-axis” complexation of benzonitrile within the cavitand cavity. The study is readily applicable to other receptors, molecular architectures, interfaces and analytes.

M

surface-specific nonlinear optical technique, which has the unique capability to study molecules at the interface with high sensitivity to polar order.7 In this context, SFVS can play a major role in providing a molecular-level comprehension of the host−guest complexation at the interfaces, resembling the realworld sensor operating conditions. Indeed, by deducing the average orientations of selected moieties through their vibrational spectra,8−12 SFVS could simultaneously provide conformational information on the surface-bound receptors and assess the orientation of the analytes upon inclusion at the interfacial receptor layer. Among the variety of organic macrocyclic receptors, cavitands,13 host-molecules with open-ended cavity of molecular dimensions, hold significant potential for chemical and biochemical sensing.14 The possibility to design them with proper cavity shape, depth, and chemical functionalities allows the selective complexation of a given class of analytes, through shape/size selection and multiple weak interactions (π−π, CH−π, H-bonding, dipole−dipole, etc.). In particular, tetraquinoxaline cavitands (QxCav) selectively bind airborne aromatic hydrocarbons, such as benzene, toluene, ethylbenzene, and xylenes (BTEX), whose detection at sub-part-per-billion level in the presence of aliphatic interferents represents a longstanding problem for air pollution monitoring in industrial

any natural events, such as DNA base pairing, enzymatic activity, antigen−antibody reaction and cellular signaling, are closely associated with specific binding interactions between biomolecules. To attain high specificity and efficiency, biological systems make use of recognition between molecules that complement each other in shape, size, charge and chemical functionalities, and self-organize in interfacial structures, like membranes. Inspired by nature, supramolecular chemists design and synthesize molecular receptors for chemical and biochemical sensing that mimic the exquisite selectivity of biomolecular receptors to the target analytes. Despite the vast body of scientific literature devoted to elucidate the host−guest complexes in the solid state,1,2 in solution,3 and in the gas phase,4,5 most promising sensor architectures require the receptors to be structured at the transducing interfaces, preserving the functional conformation and orienting the binding fragments toward the analyte-containing phase, in order to reduce nonspecific dispersion interactions.6 However, the mere presence of a preorganized cavity in the receptor does not guarantee selectivity in sensing because the analyte can also position itself outside the receptor cavity in the solid layer. Therefore, it is of paramount importance to complement host− guest association studies in solution, with analytical in situ investigations of the molecular recognition event directly at the sensor interfaces, by probing the molecular architecture of the host-decorated surfaces and correlating it with guest uptake efficiency and orientation upon complexation. Sum-frequency vibrational spectroscopy (SFVS) is a versatile analytical and © XXXX American Chemical Society

Received: June 13, 2016 Accepted: July 20, 2016

3022

DOI: 10.1021/acs.jpclett.6b01300 J. Phys. Chem. Lett. 2016, 7, 3022−3026

Letter

The Journal of Physical Chemistry Letters and urban areas.15−18 Inclusion of the aromatic guests in the deep, hydrophobic QxCav cavity is driven by CH-π and π−π interactions, as demonstrated by complexation studies in solid state,19 in solution,19 and in gas phase.20 In this paper, we report the first SFVS investigation of the complexation event at the solid−vapor interface, between the surface-bound receptor layer and aromatic or aliphatic guests in air. We prove SFVS detection of aromatic analytes, reversibly bound at the QxCavdecorated interface, and deduce their orientation within the cavity, by probing the molecular fragment involved in the complexation event. QxCav solid supported hybrid bilayers (SSHBs)21 have been arranged by transferring the C11H23-footed QxCav molecules by Langmuir−Schaefer (LS) method on alkyltrichlorosilanes self-assembled monolayer (SAM) on fused silica substrate (detailed sample preparation procedures in the Supporting Information, pages S1−S2 and Figures S1−S2). A QxCav transfer ∼30% of the Langmuir monolayer in the liquidcondensed phase (area per molecule 100 Å2) has been estimated from the drop in the trough area during the LS deposition carried out at the constant surface pressure 20 mN/ m. Both the alkylsiloxane SAM and the QxCav SSHB are studied in a gastight chamber where they are exposed to saturated vapors of benzonitrile (C6H5CN) or acetonitrile (CH3CN) in air (Figure 1). The nitrile (CN) group, shared by

basic SFVS theory7−12 and the procedures to analyze the SF spectra are given in the Supporting Information (pages S2−S4). Figure 2 shows the SF spectra of the QxCav SSHB at the solid−air interface in the spectral range 2800−3100 cm−1, where both the aliphatic and aromatic CH stretch modes are expected.

Figure 2. SF spectra of the QxCav-alkylsiloxane SSHB for the SSP, SPS and PPP polarization configurations in the CH stretching spectral range. The inset shows an enlargement of the SF spectra and their fitting curves in the aromatic CH stretching range. The vector diagrams for the ν2 and ν20 normal modes of the Qx wings, whose IR dipole moments are along ζ̂, are also reported.

The three SF−vis−IR polarization configurations SSP, SPS, and PPP (where S and P denote the polarizations perpendicular and parallel to the plane of incidence, respectively) are reported. The resonances observed in the aliphatic range (2800−3000 cm−1) arise from the alkyl chains, belonging to both the alkylsiloxanes and the QxCav molecules, and are assigned to the symmetric stretching, asymmetric stretching and Fermi resonance modes of the methylene (CH2) and methyl (CH3) groups.10,21 Three main resonances can be recognized in the aromatic range (3000−3100 cm−1) and are attributed to the asymmetric stretch ν7 (3038 cm−1) and the symmetric stretches ν2 and ν20 (3066 and 3080 cm−1) of the quinoxaline (Qx) wings.21,22 Quantitative information on the conformation of the QxCav cavity at the interface (i.e., aperture angle θm of the Qx wings)21,23,24 and on its average polar orientation θp (Figure S4) can be deduced from fitting of the SF spectra in the aromatic CH range (inset of Figure 2) and by analyzing the ν2 mode (detailed calculation in Supporting Information, pages S3−S4 and Figures S3−S4). Obtained values, θm = (13° ± 2°) and θp = (48° ± 11°), prove that the QxCav molecules preserve the vase conformation,24 which is the relevant one for guest complexation, and that their principal axis in the SSHB exhibits a significant deviation from the surface normal. Figure 3 shows the SF spectra of the QxCav SSHB and the alkylsiloxane SAM, exposed to vapors of C6H5CN in air. The spectral range 2200−2300 cm−1 has been chosen according to the CN stretching mode resonance of the analyte,25,26 in order to probe the QxCav-analyte complexation. A distinct peak occurs in the SSP configuration at ∼2240 cm−1, attributed to the CN stretch of the C6H5CN, which is significantly stronger for the QxCav SSHB (black circles, Figure 3a) than for the bare alkylsiloxane SAM (red circles). The SPS and PPP spectra are much weaker (Figure 3b,c). Without resorting to calculation,

Figure 1. Schematic view of the SFVS setup. The QxCav-alkylsiloxane SSHB is placed in a gastight chamber (triangle) and exposed to C6H5CN vapor in air. The chamber is equipped with two barium fluoride windows (in yellow), transparent in the spectral range of the input and output beams. SSP (S-polarized SF output, S-polarized visible, and P-polarized IR inputs) polarization combination is shown. The beams angles θi are reported versus the surface normal.

the two analytes, is an excellent vibrational probe for our purpose, because its IR and Raman active stretching resonance (∼2240 cm−1) does not overlap with the resonances of the QxCav SSHB. Thus, it allows to discriminate the analyte population adsorbed at the solid−vapor interface with a net polar orientation. The sum-frequency (SF) signal is generated by input beams at ωvis (532 nm) and ωIR (2800−3100 or 2200−2300 cm−1) that overlap at the solid−vapor interface and collected in reflection geometry (Figure 1). The SF intensity is proportional 2 (2) to |χ(2) eff (ωSF = ωvis + ωIR)| , where χeff is the effective surface nonlinear susceptibility and exhibits resonant enhancement as ωIR approaches a surface vibrational mode. Output/input polarization dependence of a resonant peak provides information on the average orientation of the moiety that contributes to the peak. A more detailed description of the 3023

DOI: 10.1021/acs.jpclett.6b01300 J. Phys. Chem. Lett. 2016, 7, 3022−3026

Letter

The Journal of Physical Chemistry Letters

molecular orientational distribution f(Ω). Assuming a Dirac’s delta function dependence f(Ω) = δ(θ − θCN)/4π2sin θCN in eq 1, we obtain θCN = (43° ± 15°). This value perfectly matches the average polar angle θp of the QxCav molecules in the SSHB, and is a strong indication of real complexation. It suggests “onaxis” inclusion of C6H5CN in the QxCav aromatic cavity, with a net polar orientation (i.e., ζ̂″ ≅ ζ̂′or ζ̂″ ≅ −ζ̂′) (inset in Figure 3b). In Figure 4, we compare the SF spectra of the QxCav SSHB exposed to air (solid circles) and to vapors of C6H5CN (open

Figure 3. SF spectra for the SSP (a), SPS (b), and PPP (c) configurations in the CN stretching range of the alkylsiloxane template (red circles) and QxCav SSHB (black circles) exposed to saturated vapors of C6H5CN in air. The black lines are the best curve fits of the QxCav SSHB spectra.

the spectra readily indicate that the QxCav SSHB promotes an analyte population at the solid−vapor interface with a net polar orientation. In order for the CN stretch to be excited, its IR dipole moment (parallel to ζ″ in the molecular frame, see inset in Figure 3a) must have a component along the polarization of the IR beam. Therefore, the prominence of the CN resonance in the SSP (but not in the SPS) configuration, which accesses modes with dipole moment components perpendicular (parallel) to the interface, indicates that the ζ̂″ axis of the C6H5CN has a sizable projection along the surface normal. Quantitative analysis from fitting of the SF spectra in Figure 3 allows to deduce the average orientation of the analytes adsorbed at the interface. The three independent nonlinear susceptibility amplitudes for the CN stretch mode are written in term of the two independent nonvanishing hyperpolarizability amplitudes aCN,ξ″ξ″ζ″ and aCN,ζ″ζ″ζ″, (eq S3 in Supporting Information) with ξ̂″ in the molecular plane and ζ̂″ along the CN bond (inset in Figure 3a)10 A CN, xxz =

Figure 4. SSP (a), SPS (b), and PPP (c) SF spectra of the QxCav SSHB exposed to air (solid circles) and to saturated vapors of C6H5CN (open circles), with their spectral fits (solid and dashed lines, respectively) in the aromatic CH stretching range. The red and blue lines, reported for the SSP spectra (a), are, respectively, the contribution of the ν2 and ν20 vibrational modes of the QxCav (solid lines) and of the QxCav exposed to C6H5CN (dashed lines), as determined from the fits. The negative solid blue curve (i.e., ν20 of the QxCav) denotes that it has a phase opposite to that of the positive peaks. The vector diagrams for the ν2 and ν20 normal modes of C6H5CN are reported in (a).

circles), in the aromatic CH stretching range 3000−3100 cm−1. The spectral analysis of the SSP polarization (Figure 4a) shows the two main resonant contributions, attributed to the ν2 (red) and ν20 (blue) modes, for both the QxCav SSHB exposed to air (solid lines) and to vapors of C6H5CN (dashed lines), as determined from fits (black lines). The ν2 mode (solid red line) dominates the SSP spectrum of the QxCav exposed to air, while ν20 (solid blue line) is much weaker and 180° out of phase with respect to ν2. When QxCav is exposed to C6H5CN, the ν2 peak (dashed red line) nearly vanishes and the ν20 mode (dashed blue line) becomes prominent. C6H5CN exhibits aromatic CH stretches, specifically the ν2 and ν20 modes (3066 and 3080 cm−1, inset in Figure 4a),22 which overlap in frequency with the corresponding ones of the Qx wings. Moreover, being SFVS a coherent second order nonlinear spectroscopy, each vibrational

Ns aCN, ζ ″ ζ ″ ζ ″⟨(1 + rCN)cos θ − (1 − rCN) 2 cos3 θ ⟩

A CN, xzx =

Ns aCN, ζ ″ ζ ″ ζ ″(1 − rCN)⟨cos θ − cos3 θ ⟩ 2

(1)

A CN, zzz = NSaCN, ζ ″ ζ ″ ζ ″⟨rCNcos θ + (1 − rCN)cos3 θ ⟩

where NS is the surface density of C6H5CN, θ is the polar angle between the molecular axis (ζ̂″) and the surface normal (ẑ), and the depolarization ratio rCN  aCN,ξ″ξ″ζ″/aCN,ζ″ζ″ζ″ ≅ 0.26.10,25,26 The angular brackets denote an average over the 3024

DOI: 10.1021/acs.jpclett.6b01300 J. Phys. Chem. Lett. 2016, 7, 3022−3026

Letter

The Journal of Physical Chemistry Letters

unique among surface analysis methods in determining the structure of the host−guest complexes. An alternative method to test molecular recognition at interfaces, namely X-ray photoelectron spectroscopy (XPS),29 is able only to assess the presence of the complexation event, but not to determine the orientation of the guest within the cavity. This additional information is pivotal for designing selective receptors for sensing and catalysis, where the correct orientation of the guest is necessary for recognition30,31 and reactivity.32,33

mode has a phase related to molecular orientation, which affects the SF signal through spectral interference. Therefore, depending on the actual polar orientation of C6H5CN with respect to the QxCav cavity, that is, the CN moiety pointing either outward or inward, one would expect destructive or constructive interference, respectively, in the aromatic CH stretching range. The suppression of the ν2 peak, as well as the phase change in the ν20 resonance, when QxCav SSHB is exposed to C6H5CN are experimental evidence of destructive interference between the modes of QxCav and C6H5CN, and suggest that QxCav-C6H5CN complexation at the solid−air interface occurs with the phenyl group pointing inward the QxCav cavity (i.e., ζ̂″ ≅ ζ̂′) (inset in Figure 3b). This is in agreement with complexation studies in solid state and in solution of QxCav toward monosubstituted benzene guests,19 according to which host−guest association is driven by aromatic CH−π interactions. In Figure 5, we compare the SSP spectra of the QxCav SSHB exposed to air and to saturated vapors of C6H5CN and CH3CN



EXPERIMENTAL METHODS The SFVS setup consists of a picosecond Nd:YAG laser (PL2251−20, Ekspla), that outputs 30 ps pulses (70 mJ @ 1064 nm) at 20 Hz repetition rate, an harmonics unit (SFGH1000-2H, Ekspla) and an OPG-OPA-DFG unit (PG501-DFG1P, Ekspla). The visible (532 nm) and infrared (1000−4300 cm−1) beams are overlapped temporally and spatially on the sample surface with angles θvis = 60° and θIR = 55° with respect to the sample surface normal (Figure 1), and the sum frequency beam is collected in reflection geometry at θSF ≅ 59°. Spectral resolution is