Hydrogen Bond Formation between the Carotenoid Canthaxanthin

Article pubs.acs.org/JPCB

Hydrogen Bond Formation between the Carotenoid Canthaxanthin and the Silanol Group on MCM-41 Surface Yunlong Gao,*,† Dayong Xu,*,‡ and Lowell D. Kispert§ †

College of Science and ‡College of Agriculture, Nanjing Agricultural University, Nanjing, 210095, China Department of Chemistry, University of Alabama, Box 870336, Tuscaloosa, Alabama 35487, United States

§

ABSTRACT: The formation of one or two hydrogen bonds (H-bonds) between canthaxanthin (CAN), a dye, and the silanol group(s) on the MCM-41 surface has been studied by density functional theory (DFT) calculations and calorimetric experiments. It was found that the formation of the H-bond(s) stabilized the CAN molecule more than its radical cation (CAN•+). The charge distribution, bond lengths, and the HOMO and LUMO energies of CAN are also affected. The formation of the H-bond(s) explains the lower photoinduced electron transfer efficiency of CAN imbedded in Cu-MCM-41 versus that for β-carotene (CAR) imbedded in Cu-MCM-41 where complex formation with Cu2+ dominates. These calculations show that to achieve high electron transfer efficiency for a dye-sensitized solar cell, H-bonding between the dye and the host should be avoided.

■

INTRODUCTION A photovoltaic (PV) is a device that directly converts sunlight into electricity without pollution, sound, or moving parts, thereby making it long lasting, elegant, and dependable. The development of low-cost PV cells with high efficiency has been the topic of intensive research over the past 3 decades. Dyesensitized solar cell (DSSC) is one of these, and the performance of DSSC mainly relies on the photoinduced electron transfer (ET) efficiency of a dye as the sensitizer because photoionization is the first step in the light driven reactions related to the storage of light energy. Electron transfer (ET) of various dyes in different environments1−7 has been extensively studied to generate the basis for producing artificial photoredox systems for solar energy conversion and storage. The use of natural dyes has become of interest recently because of their ease of preparation, low cost, biodegradability, availability, purity, environmental friendliness, and most importantly, significant reduction of noble metal and chemical synthesis cost.8−10 Natural dyes are pigmentary molecules which are obtained with or without chemical treatments from plants and from animals or minerals. Numerous natural dyes use sensitizers in DSSC, such as chlorophyll,11,12 carotenoid,8,13,14 anthocyanin,15,16 flavonoid,9,17,18 cyanine,19−21 and tannin.22 Carotenoids are organic pigments that are found in plants and microorganisms. Carotenoids, flavonoids, andanthocyanin are often found in the same organs. Carotenoid pigments provide flowers and fruits with red, yellow, and orange color and are usually responsible for the yellow to orange petal colors.8,17,23 Carotenoids are involved in photosynthesis and contribute to light harvesting by absorbing light energy in a region of the visible spectrum where chlorophyll absorption is lower and by transferring the absorbed energy to chlorophyll. It © XXXX American Chemical Society

also provides protection from excess light via energy dissipation, free radical detoxification and limits damage to membranes.24,25 ET reactions of carotenoids imbedded in mesoporous molecular sieves MCM-41 and metal ion substituted MCM41 (Ni-MCM-41,26 Al-MCM-41,26 Fe-MCM-41,27 Ti-MCM41,28 and Cu-MCM-4129) have been studied. MCM-41 is a mesoporous silica containing a regular array of uniform cylindrical pores. The pore size ranges from 15 to 100 Å depending on the chain length of the template used in the synthesis.30 Previous studies26−29,31,32 have shown that such materials provide a microenvironment appropriate for retarding back ET and thus increase the lifetime of photoproduced radical ions. However, different carotenoids behave differently in metal ion substituted MCM-41. For example, β-carotene (CAR) interacts with Cu2+ to form a complex when it is imbedded in Cu-MCM-41, which was detected by EPR (electron paramagnetic resonance) measurements.29 The formation of the CAR-Cu2+ complex favors light-driven ET from CAR to Cu2+ and also permits thermal back ET from Cu+ to CAR•+.29 On the contrary, no CAN-Cu2+ complex was detected by EPR29 for canthaxanthin (CAN) imbedded in CuMCM-41. The photoinduced ET efficiency of CAN in CuMCM-41 was about 5 times lower than that of CAR in CuMCM-41.29 The much lower ET may be attributed to the absence of CAN-Cu2+ complex. Also, the photoinduced ET efficiency of CAN in MCM-41 is about 3 times lower than that of CAR in MCM-41,29 which suggests that the lack of the CAN-Cu2+ complex is not the only reason for the much lower Received: June 13, 2015 Revised: July 27, 2015

A

DOI: 10.1021/acs.jpcb.5b05645 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

effects result in the much lower photoinduced ET efficiency of CAN imbedded in MCM-41.

ET efficiency of CAN imbedded in Cu-MCM-41. The oxidation potential of CAN is about 0.1 V higher than that of CAR in CH2Cl23,33 (775 and 634 mV vs SCE, respectively), which may cause the ET efficiency of CAN to be lower than that of CAR in MCM-41. Whether there are other reasons for the lower ET efficiency of CAN is unknown. The structures of CAR and CAN (shown in Scheme 1) are very similar except

■

METHODS AND MATERIALS Synthesis of MCM-41. The procedure used for the preparation of the siliceous material (MCM-41) has been previously described.29 Chemicals. Canthaxanthin was purchased from Fluka (Buchs, Switzerland). β-Carotene was from Sigma, and anhydrous chloroform was from Aldrich. Calculations. All computations were performed with the Gaussian 0935 suite of programs. Calculations with basis set superposition error (BSSE) correction were carried out by the counterpoise method of Boys−Bernardi.36 In this study, the two-fragment BSSE correction is used. Details about the BSSE correction were described in a previous study.37 Microcalorimetry. Calorimetric experiments were performed using a LKB 2107 differential microcalorimetry system. The calorimeter consists of a rotating block placed inside an air bath maintained at a constant temperature. The block consists of a sample cell and a reference cell. Each cell has two compartments. Mixing of the reactants is achieved by rotating the block. The heat of reaction is measured in terms of area. Area−heat calibration is then carried out by passing a known current for a set time period through a calibration heater in either one of the vessels. Three separate measurements in a narrow range around the reaction area are done for each experiment to obtain a calibration curve. Before the experiment was carried out, two samples of MCM-41 of 1.00 g each were activated at 200 °C for 6 h, immediately transferred to a drybox, and allowed to cool to room temperature. The heat of reaction was measured at 25 °C by mixing 40 mL of CAN CH3Cl solution (5.0 mM) with one of the two MCM-41 samples in the sample vessel, and 40 mL of CAR CH3Cl solution (5.0 mM) with the other MCM-41 sample in the reference cell. After the measurement, the solid− liquid separation of the two mixtures was achieved using centrifugation at 1000 rpm for 10 min. A Shimadzu UV-1800 UV−vis spectrophotometer was used to measure the amount of CAN and CAR that remained in solution. Previous EPR Measurements and DFT Calculations. Previously, an EPR study29 showed that CAR interacts with Cu2+, and a CAR-Cu2+ complex was formed after CAR was imbedded in Cu-MCM-41; however, no CAN-Cu2+ complex was detected after CAN was imbedded in Cu-MCM-41. The CAR-Cu2+ complex was also recently examined by DFT calculations.37 The optimized structure was deduced using B3LYP with BSSE correction between two fragments (Cu2+ complex as one fragment and CAR as the other fragment). The calculated distances between Cu2+ and the two hydrogen atoms at the middle chain are almost the same, and the average Cu−H distance is about 3.16 Å, which fits very well with that obtained by pulse EPR experiment.38 It was also found that the interaction energy (IE) is the largest (−2.61 kcal/mol) when Cu2+ interacts at the middle double bond of CAR.37 To understand why CAN did not form a complex with Cu2+, the interaction between CAN and Cu2+ (Figure 1) was calculated using the Cu2+-complex model built in the previous study.37 The calculation method is the same as that in the DFT study,37 and the details are described in the caption of Figure 1. The calculated distances between Cu2+ and the two middle hydrogen atoms are almost the same, and the average Cu−H distance is about 3.18 Å, which is similar to that of the CAR-

Scheme 1. Structures of CAN and CAR

that CAN has one oxygen atom at each of the two cyclohexene rings. Therefore, CAN can form one or two hydrogen bonds with the silanol group(s) on the MCM-41 surface. To understand why no CAN-Cu2+ complex is formed when CAN is imbedded in Cu-MCM-41, density functional theory (DFT) calculations were carried out for CAN imbedded in CuMCM-41. The calculations show that the interaction energy (IE) between CAN and Cu2+ is much smaller than that between CAN and a silanol group (-SiOH) on the MCM-41 surface, suggesting that CAN prefers to form one or two hydrogen bonds with the silanol group(s) when CAN is imbedded in Cu-MCM-41. The hydrogen bond (H-bond) formed is due to the interaction between the oxygen atom in the cyclohexene ring of CAN and the proton of the -SiOH group. The calculated IE for H-bonding formation is about 11 kcal/ mol, which was further confirmed by our calorimetric experiments. This IE is higher than the IE (∼8 kcal/mol) between an acetone molecule and a -SiOH group, which was calculated here and measured by the NMR experiments in another study.34 To understand why the IE between CAN and -SiOH is higher than that between acetone and -SiOH, the charges on the oxygen atoms of CAN and acetone were compared. The high IE between CAN and -SiOH can be attributed to the large density of negative charge on the oxygen atom before and after the formation of the H-bond(s). The calculations also show that the negative charges on the conjugated chain of CAN migrate to the end because of the attraction of the proton of the -SiOH group; the lengths of the double bonds increase and the lengths of the single bonds decrease on the conjugated chain, suggesting that the formation of the hydrogen bond(s) increases the degree of conjugation of CAN. This in turn decreases significantly the HOMO and LUMO energies of CAN. The H-bonding properties between the radical cation (CAN•+) of CAN and -SiOH were also examined. The calculated H-bonding energy is ∼8 kcal/mol, indicating that the radical cation is less stabilized than the neutral species by the formation of the H-bond(s). All these B


Article


without counterpoise (CP) corrections (both single point and optimized on a CP-corrected surface) focusing on methods that might provide acceptable accuracy for large systems without excessive computational burden. By comparing with experimental value for the enthalpy of interaction for the water dimer and O···O distance obtained by XRD, the preferred choices became (in order of increasing size of the basis set) (1) D95(d,p) with B3LYP, B97D, M06, or MPWB1k; (2) 6311G(d,p) with B3LYP; (3) D95++(d,p) with B3LYP, B97D, or MPWB1K; (4) 6-311++G(d,p) with B3LYP or B97D; and (5) aug-cc-pvdz with M05-2X, M06-2X, or X3LYP.42 The accuracy of calculation can be further improved with basis set superposition error (BSSE) correction using the CP method. The optimization can be performed on the CP-corrected PES (CP-OPT) or on the “normal” PES methods followed by single point energy evaluation. This is a posteriori counterpoise correction (CP-SP).42 Although CP-OPT is slightly more accurate than CP-SP, the calculated interaction energies by B3LYP/6-311++G(d,p) are very similar (−5.17 and −5.20 kcal/mol, respectively) for the two methods.42 In this study, B3LYP with CP-SP is chosen because the systems are large and CP-OPT would be very expensive.

Figure 1. Optimized structures of Cu2+-CAN calculated by B3LYP with the BSSE correction using the mixed basis set (6-311+G(d) for Cu, 6-31G(d) for CAN and the three oxygen atoms coordinated to Cu2+, and 3-21G for the remaining atoms): H, light gray; O, red; Cu, orange; C, dark gray; Si, blue-gray.

Cu2+ complex. The calculated IE is −2.58 kcal/mol, which is also similar to that of the CAR-Cu2+ complex.37 The calculations suggest that CAN should also form a CAN-Cu2+ complex when CAN is imbedded in Cu-MCM-41. The only difference between the structure of CAR and that of CAN is that CAN has one oxygen atom at each of the two cyclohexene rings. The possible reason why CAN does not form a complex with Cu2+ is because CAN prefers to form one or two H-bonds with the silanol group(s) (-SiOH) on the surface of MCM-41 when it is imbedded in Cu-MCM-41. Definition of Hydrogen Bonding Using DFT Methods. Hydrogen bonding plays a fundamental role within chemistry and biochemistry. Many self-assembling systems such as nucleic acids, proteins, and nanomaterials owe much of their stabilities to H-bonds. Recent experimental and theoretical advances define the hydrogen bond as follows: the hydrogen bond is an attractive interaction between a hydrogen atom from a fragment or molecule D−H in which D is more electronegative than H, and an atom or a group of atoms A, in the same or a different molecule, where there is evidence of bond formation.39 The partial covalent nature of the hydrogen bond has been experimentally verified during the past decade by Compton scattering40 and NMR spin−spin coupling41 measurements. Today, it is well accepted that hydrogen bonding has contributions from electrostatic interactions between permanent multipoles, polarization or induced interactions between permanent and induced multipoles, dispersion arising from instantaneous multipoles-induced multipoles, charge transfer, and exchange correlation effects from short-range repulsion due to overlap of the electron distribution, in addition to partial covalent bonding.39 The strength of a H-bond can be measured by the bond length of H···A and ∠D-H···A. The shorter the bond length of H···A, the stronger the H-bond; the larger is the angle ∠D-H···A, the stronger is the H-bond. The IE between the two fragments or molecules or between H and A in the same molecule also indicates the strength of the H-bond: The lower the energy, the stronger is the H-bond. Quantum mechanical methods are usually applied to study these systems. Among those methods, accurate, yet economical molecular orbital (MO) methods are more frequently used to calculate hydrogen-bonding properties. Density functional theory (DFT) has become one of the best choices for studying large systems due to the balance of accuracy and efficiency.42 The B3LYP functional has generally been the method of choice for H-bonding interactions in self-assembling and biochemical materials such as studies42 of peptides. Plumley et al.42 calculated the H-bonding properties of the water dimer using a wide variety of functional/basis set combinations with and

■

RESULTS Model for H-Bond between SiOH and Acetone. One cannot assume that methods that work well for water dimers will work equally well elsewhere. To test whether the methods work for a H-bond between -SiOH on a siliceous surfaces, such as MCM-41 surface, and an organic compound, a model for the -SiOH group on a siliceous surface was built in this study (Figure 2). The formation of the H-bond between -SiOH and

Figure 2. Optimized structure of ACE-SiOH calculated by B3LYP using the mixed basis set (6-311++G(d,p) for the oxygen and hydrogen atoms of the silanol group and the oxygen atom of acetone, 6-31G(d) for the other atoms of acetone, the silica atom of the silanol group, and the three oxygen atoms connected to the silica atom, and 321G for the remaining atoms): H, light gray; O, red; C, dark gray; Si, blue-gray.

acetone was examined by the B3LYP method with a posteriori counterpoise correction (CP-SP). Acetone was chosen because the value of the interaction energy between -SiOH and acetone obtained by NMR experiments was known.34 The optimized structure of acetone-SiOH (ACE-SiOH) complex is shown in Figure 2. In order to perform efficient, yet sufficiently accurate calculations, a combination of basis sets was implemented. The mixed basis set includes 6-311++G(d,p) for the oxygen and hydrogen atoms of the silanol group and the oxygen atom of acetone, 6-31G(d) for the other atoms of acetone, the silica atom of the silanol group, and the three oxygen atoms C


Article

The Journal of Physical Chemistry B Table 1. Calculated H-Bonding Properties species

ACE-SiOH

CAN-SiOH

CAN•+-SiOH

SiOH-CAN-SiOH

SiOH-CAN•+-SiOH

O···H length (Å) ∠O···H−O (deg) IE (kcal/mol)

1.85292 165.52 −8.58

1.69461 177.24 −10.89

1.76412 172.21 −8.18

1.68653 176.12 −10.81

1.76422 172.56 −8.03

Figure 3. Optimized structures of CAN, CAN•+, CAN-SiOH, CAN•+-SiOH, SiOH-CAN-SiOH, and SiOH-CAN•+-SiOH calculated by B3LYP using the mixed basis set (6-311++G(d,p) for the oxygen and hydrogen atoms of the silanol group and the oxygen atom of CAN, 6-31G(d) for the other atoms of CAN, the silica atom of the silanol group, and the three oxygen atoms connected to the silica atom, and 3-21G for the remaining atoms): H, light gray; O, red; C, dark gray; Si, blue-gray.

CAN•+-SiOH are shown in Figure 3. B3LYP/6-31G(d) was used to calculate CAN and CAN•+. The method used to calculate the complexes with the formation of the H-bond(s) is the same as that for the calculation of ACE-SiOH. The mixed basis set used in the calculations is also the same as that for the calculation of ACE-SiOH (i.e., 6-311++G(d,p) for the oxygen and hydrogen atoms of the silanol group and the oxygen atom of CAN. 6-31G(d) is used for the other atoms of CAN, the silica atom of the silanol group, and the three oxygen atoms connected to the silica atom, and 3-21G is used for the remaining atoms). The single point energy calculations with two fragment BSSE correction after the optimization of CANSiOH and CAN•+-SiOH are similar to that for ACE-SiOH except that one fragment is CAN or CAN•+ instead of ACE. For the optimized structures of SiOH-CAN-SiOH and SiOHCAN•+-SiOH, the single point energy calculations with BSSE correction were also performed between two fragments (the -SiOH complex as one fragment and CAN-SiOH or CAN•+SiOH as the other fragment). The calculated H-bonding properties for the species are listed in Table 1. For SiOH-CANSiOH and SiOH-CAN•+-SiOH with the formation of two Hbonds, only the properties of one H-bond are listed because the structures are symmetrical, and the properties of the two Hbonds are the same. The H-bonding properties of CAN-SiOH and SiOH-CAN-SiOH are similar. The O···H length is about 1.69 Å and the ∠O···H−O is close to 180°, indicating a very strong H-bond. The calculated IE is about −11 kcal/mol (for the formation of two H-bonds, the IE is about −22 kcal/mol) between CAN and -SiOH which further proves the presence of strong H-bonding. This value is more than twice the IE (about −5.2 kcal/mol) for the water dimers.42 The calculated IE is also much lower than the calculated IE (−2.58 kcal/mol) between CAN and Cu2+, explaining why CAN prefers to form a H-

connected to the silica atom, and 3-21G for the remaining atoms. The large basis set 6-311++G(d,p) was used for the atoms of O···H−O to ensure that the calculated H-bonding properties are accurate because B3LYP/6-311++G(d,p) is a good method for the calculation of the H-bonding properties of a water dimer.42 After optimization of the structure, a single point energy calculation was carried out with the BSSE correction between two fragments (the -SiOH complex as one fragment and acetone as the other fragment) to calculate the interaction energy between the two fragments. The H-bonding properties of ACE-SiOH listed in Table 1 show that the calculated distance (O···H) between oxygen atom of acetone and hydrogen atom of -SiOH is 1.85292 Å and the ∠O···H−O is 165.52° indicating a strong hydrogen bond. The calculated IE value of −8.58 kcal/mol fits well with the experimental data of −37 kJ/mol34 (about −8.8 kcal/mol) for the formation of hydrogen bonded complexes of acetone with the surface silanol groups in the gas phase. H-Bonding between the Carotenoids and SiOH. The same model for the -SiOH group on the siliceous surface was used to calculate the H-bonding properties between CAN and -SiOH on the MCM-41 surface. Since CAN is located in the pores of MCM-41, it may form two H-bonds with the -SiOH groups because CAN contains two oxygen atoms at each of the two cyclohexene rings. The formation of the H-bond between CAN•+ and -SiOH after ET was also evaluated for comparison because it is important to know how the ET affects the Hbond(s). The complexes of CAN and -SiOH and that of CAN•+ and -SiOH are represented by CAN-SiOH and CAN•+SiOH, respectively. The complexes for the formation of two Hbonds are symbolized as SiOH-CAN-SiOH and SiOH-CAN•+SiOH, respectively. The optimized structures of CAN, CANSiOH, SiOH-CAN-SiOH, CAN•+, CAN•+-SiOH, and SiOHD


Article


Figure 4. (a) Calculated excess charge difference (DDexcess) between CAN-SiOH and CAN, (b) SiOH-CAN-SiOH and CAN, (c) CAN•+-SiOH and CAN•+, and (d) SiOH-CAN•+-SiOH and CAN•+. The calculated method is the same as that described in the caption of Figure 3.

bond(s) than to interact with Cu2+ when it is imbedded in CuMCM-41. The H-bonding properties of CAN•+-SiOH and SiOH-CAN•+-SiOH are also similar. The O···H length is about 1.76 Å, which is slightly longer than those for Can-SiOH and SiOH-CAN-SiOH. The ∠O···H−O is about 172°, which is slightly smaller than the angles for CAN-SiOH and SiOHCAN-SiOH, indicating a relatively weaker H-bond. The calculated IE between CAN•+ and -SiOH is about −8 kcal/ mol, which is less than those of the neutral species, suggesting that the radical cation is less stabilized than the neutral species by the formation of the H-bond(s). The H-bond between the radical cation and -SiOH is weaker than that between the neutral species and -SiOH because the negative charge on the oxygen atom of the radical cation is less than that for the neutral species due to the electron deficiency of the radical cation (the Mulliken charge on the oxygen atoms of CAN•+ is −0.454 and that on the oxygen atoms of CAN is −0.481). The H-bonding properties listed in Table 1 also show that the O···H length of ACE-SiOH is longer than that of CANSiOH, and the ∠O···H−O is smaller than that of CAN-SiOH, indicating that the H-bond strength of ACE-SiOH is smaller than that of CAN-SiOH. The calculated IE for CAN-SiOH is about 2.3 kcal/mol less than that for ACE-SiOH. To understand why the conjugation system of CAN interacts with -SiOH more strongly than that with acetone, the Mulliken charge distributions of CAN and acetone before and after the interaction with -SiOH were compared. Before the interaction,

the charges on the oxygen atoms of CAN and acetone are −0.481 and −0.426, respectively, showing that the oxygen atom of CAN has more negative charge. This is understandable because the electrons on the conjugated chain of CAN move freely and are attracted to the oxygen atoms because of the high electronegativity of the oxygen atom. The charges on C4, C5, and C6 of CAN are 0.413, 0.029, and 0.094, respectively. The charges on these carbons are positive because the electrons on these carbons are attracted to the oxygen atom. The more negative charge that the oxygen atom possesses, the stronger is the H-bond. After the interaction, the charges on the oxygen atoms of CAN and acetone are −0.497 and −0.471, respectively, and both charges increase slightly because of the attraction from the proton of the -SiOH group. The oxygen atom of CAN still possesses more negative charge than that of acetone, resulting in a stronger H-bond. Calorimetry Measurements. The large interaction energy between CAN and -SiOH was also examined by calorimetric measurements. Since the structure of CAR is very similar to that of CAN except that CAR has one less oxygen atom at each of the two cyclohexene rings, the difference between the adsorption heat of CAN on MCM-41 and that of CAR on MCM-41 can be approximately attributed to the interaction energy between CAN and the −SiOH group(s). The reaction heat of CAN CH3Cl solution mixing with MCM-41 was measured using the reaction heat of CAR CH3Cl solution mixing with MCM-41 as the reference in the experiments. The E


Article


Figure 5. Calculated HOMO and LUMO surfaces of CAN, CAN-SiOH, and SiOH-CAN-SiOH. The calculated method is the same as that described in the caption of Figure 3.

Since the charge distribution on the conjugated chain of CAN changes after the formation of the H-bond(s), the bond lengths should also change. The calculations show that the formation of the H-bond(s) increases the lengths of the double bonds and decreases the lengths of the single bonds, suggesting that the formation of the hydrogen bond increases the degree of conjugation of CAN. The degree of conjugation in a molecule can be evaluated by means of the bond length alternation (BLA) parameter, which is defined as the difference in total length between single bonds (C−C) and double bonds (CC). The BLA parameter increases as the degree of conjugation decreases. The BLA parameter was calculated from C5 to C6′. The calculated BLAs for CAN, CAN-SiOH, and SiOH-CAN-SiOH are 0.76563, 0.74246, and 0.72835, respectively, indicating that the degree of conjugation increases after the formation of the H-bond(s) and more for the formation of two H-bonds. HOMO and LUMO Energies. It is known that the HOMO (highest occupied molecular orbital) and LUMO (lowest unoccupied molecular orbital) of a conjugation system are related to the degree of conjugation. Thus, it is expected that the energies of the HOMO and LUMO of CAN will change after the formation of the H-bond(s) with the -SiOH group(s). The calculated HOMO and LUMO surfaces of CAN, CANSiOH, and SiOH-CAN-SiOH are shown in Figure 5. For the HOMO and LUMO of CAN-SiOH and SiOH-CAN-SiOH, the π orbitals are distributed over the whole conjugation system of CAN and there are no contributions from the atomic orbitals of the −SiOH group(s), suggesting that HOMO and LUMO are only affected by the degree of conjugation. The calculated HOMO and LUMO energy of CAN, CAN-SiOH, and SiOHCAN-SiOH are listed in Table 2. Both the energies of the

amount of solutions and MCM-41 samples and the concentrations of the carotenoids remained the same in the measurement. UV/vis measured the amounts of carotenoids remaining in solution after the experiments, at 450 nm. No carotenoids were detected, indicating that all carotenoids adsorbed on the MCM-41 surfaces. The adsorption heat of CAN is 15.1 kcal/mol more than that of CAR. This value is higher than the calculated interaction energy (∼11 kcal/mol) of CAN-SiOH and lower than twice (∼22 kcal/mol) the calculated interaction energy of SiOH-CAN-SiOH. It can be estimated that about one-half of CAN forms one H-bond and the other half forms two H-bonds when they adsorb on MCM41. Formation of H-Bonds: Change in Charge Distribution. It is useful to know how the formation of the H-bond(s) between CAN and −SiOH and that between CAN•+ and − SiOH affects the charge differences and bond lengths of CAN and CAN•+. The calculated excess charge differences (DDexcess) between CAN-SiOH and CAN and those between SiOH-CANSiOH and CAN are shown in Figure 4a and Figure 4b, respectively. “O” on the x-axis represents the oxygen atom, and the numbers represent the carbon atoms in the conjugated chain. It is shown in Figure 4a and Figure 4b that the negative charges migrate to the end of the chain because of the attraction of the SiOH proton of the −SiOH fragment after the H-bond is formed. The majority of the negative charges are located at C4, and only a small amount is located at the oxygen atom. The carbon C5 loses the most negative charge. The charge on individual carbons decreases the further it is from the end of the chain. The calculated excess charge differences (DDexcess) between CAN•+-SiOH and CAN•+ and those between SiOH-CAN•+−SiOH and CAN•+ are shown in Figure 4c and Figure 4d, respectively. The negative charges also migrate to the end of the chain similar to the neutral species. Nevertheless, the difference between the negative charges at the oxygen atom and those at the C4 is small compared with that in Figure 4a or Figure 4b. Like the neutral species, the carbon that loses negative charge the most is C5, and the charge on a carbon is less affected the further it is from the end of the chain.

Table 2. Calculated HOMO and LUMO Energy (eV) of CAN, CAN-SiOH, and SiOH-CAN-SiOH

HOMO LUMO F

CAN

CAN-SiOH

SiOH-CAN-SiOH

−4.754 −2.624

−4.892 −2.812

−5.033 −2.974



■

HOMO and LUMO of CAN decrease after the formation of the H-bond(s) and more for the formation of two H-bonds. A recent study43 showed that there is a very strong linear correlation of DFT-calculated HOMO/LUMO energies and the redox potentials from a series of similar compounds. The decrease of the HOMO energy of CAN after the formation of the H-bond(s) indicates that the oxidation potential of CAN Hbonding to −SiOH group is higher than that of the free CAN and would be more difficult to be oxidized than free CAN. The decrease of the LUMO energy of CAN after the formation of the H-bond(s) suggests that the reduction potential of CAN Hbonding to −SiOH group is lower than that of the free CAN and would be easier to be reduced than the free CAN. Photoinduced ET Efficiency. On the basis of the above calculations, it can be concluded that the much lower photoinduced ET efficiency of CAN imbedded in Cu-MCM41 compared with that of CAR is attributed to the following four facts: (1) the oxidation potential of CAN is higher than that of CAR. (2) CAR forms a complex with Cu2+, which favors ET from CAR to Cu2+ as a result of the short distance between CAR and Cu2+. However, a H-bond occurs between CAN and the −SiOH group on the surface of MCM-41 rather than forming a complex with Cu2+. (3) The formation of the Hbond(s) with the −SiOH group(s) stabilizes the CAN molecule more than the radical cation of CAN (more than −2.5 kcal/mol for the formation of one H-bond and more than −5 kcal/mol for the formation of two H-bonds), which causes the oxidation of CAN to be difficult. (4) the formation of the H-bond(s) with the −SiOH group(s) decreases the energy of the HOMO of CAN, which also causes CAN to be difficult to be oxidized. This study shows that to obtain high ET efficiency of a dye, the H-bonding between the dye and the host should be avoided. Such is the case in a previous study1 where CAN and CAR were deposited on indium tin oxide electrodes spincoated with a mixture of TiO2 nanoparticles, polyvinylcarbazole immersed in an aqueous KCl solution containing hydroquinone. In this system the photocurrent reached a steady state value of 90 and 40 nA/cm2 for CAN and CAR, respectively. The smaller photocurrent for CAR can now be understood in view of the above calculation since H-bonding was not present in this system.

■

Article

AUTHOR INFORMATION

Corresponding Authors

*Y.G.: e-mail, [email protected]. *D.X.: e-mail, [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by Faculty Research Grants Program at College of Sciences of Nanjing Agricultural University.

■

REFERENCES

(1) Gao, G.; Deng, Y.; Kispert, L. D. Photovoltaic Response of Carotenoid-Sensitized Electrode in Aqueous Solution: ITO Coated with a Mixture of TiO2 Nanoparticles, Carotenoid, and Polyvinylcarbazole. J. Chem. Soc., Perkin Trans. 2 1999, 1225−1230. (2) Gao, F. G.; Bard, A. J.; Kispert, L. D. Photocurrent Generated on a Carotenoid-Sensitized TiO2 Nanocrystalline Mesoporous Electrode. J. Photochem. Photobiol., A 2000, 130, 49−56. (3) Hapiot, P.; Kispert, L. D.; Konovalov, V. V.; Savéant, J.-M. Single Two-Electron Transfers vs Successive One-Electron Transfers in Polyconjugated Systems Illustrated by the Electrochemical Oxidation and Reduction of Carotenoids. J. Am. Chem. Soc. 2001, 123, 6669− 6677. (4) Kalyanasundaram, K. Photochemistry in Microheterogeneous Systems; Academic Press: New York, 1987. (5) Energy Resources through Photochemistry and Catalysis; Grätzel, M., Ed.; Academic Press: New York, 1983. (6) Grätzel, M. Heterogeneous Photochemical Electron Transfer; CRC Press: Boca Raton, FL, 1988. (7) Photochemical Conversion and Storage of Solar Energy; Connolly, J. S., Ed.; Academic Press: New York, 1981. (8) Kishiomoto, S.; Maoka, T.; Sumitomo, K.; Ohmya, A.; Ohmya, A. Analysis Of Carotenoid Composition in Petals of Calendula (Calendula Officinalis L). Biosci., Biotechnol., Biochem. 2005, 69, 2122−2128. (9) Sinha, K.; Saha, P. D.; Datta, S. Extraction of Natural Dye from Etals Flam of Forest (Butea Monosperma) Flower: Process Optimization Using Response Surface Methodology (RSM). Dyes Pigm. 2012, 94, 212−216. (10) Nishantha, M. R.; Yapa, Y. P.; Perera, P. S. Sensitization of Photoelectrochemical Solar Cells with a Natural Dye Extracted from Kopsia Flavida Fruit. Proc. Technol. Sess. 2012, 28, 54−58. (11) Hao, S.; Wu, J.; Huang, Y.; Lin, J. Natural Dyes as Photosinsetizers for Dye-Sensitized Solar Cell. Sol. Energy 2006, 80, 209−214. (12) Kumara, G. R. A; Kaneko, S.; Okuya, M.; Onwona-Agyeman, B.; Konno, A.; Tennakone, K. S. Shiso Leaf Pigments for Dye-Sensitized Solid State Solar Cell. Sol. Energy Mater. Sol. Cells 2006, 90, 1220− 1226. (13) Rühle, S.; Shalom, M.; Zaban, A. Quantum-Dot-Sensitized Solar Cells. ChemPhysChem 2010, 11, 2290−2304. (14) Gomez-Ortiz, N. M.; Vazquez-Maldonado, I. A.; Perez-Espadas, A. R.; Mena-Rejon, G. J.; Azamar-Barrios, J. A.; Oskam, G. DyeSensitized Solar Cells with Natural Dyes Extracted from Achiote Seeds. Sol. Energy Mater. Sol. Cells 2010, 94, 40−44. (15) Alhamed, M.; Issa, A. S.; Doubal, A. W. Studying of Natural Dyes Properties as Photo-Sensitizer for Dye Sensitized Solar Cells (DSSC). J. Electron Devices 2012, 16, 1370−1383. (16) Chang, H.; Lo, Y. J. Pomegranate Leaves and Mulberry Fruit as Natural Sensitizers for Dye-Sensitized Solar Cells. Sol. Energy 2010, 84, 1833−1837. (17) Davies, K. M. Plant Pigments and Their Manipulation, 14; Annual Plant Reviews;Blackwell Publishing Ltd.: Oxford, U.K., 2004; p 342. (18) Sutthanut, K.; Sripanidkulchai, B.; Yenjai, C.; Jay, M. Simultaneous Identification and Quantitation of 11 Flavonoid

CONCLUSIONS

The carotenoid CAN forms one or two H-bonds with the − SiOH groups on the MCM-41 surfaces when it is imbedded in MCM-41. The amount of the negative charges on the oxygen atoms on the cyclohexene rings increases because of the mobility of the electrons on the conjugated chain and the high electronegativity of the oxygen atoms. The formation of the Hbond(s) generates negative charge buildup toward the terminal end(s) of CAN, which increases the degree of conjugation. This in turn decreases the energies of the HOMO and LUMO of CAN. The formation of the H-bond(s) stabilizes the neutral species of CAN more than the radical cation. These effects plus the fact that CAN does not form a complex with Cu2+ in CuMCM-41 results in the much lower photoinduced ET efficiency of CAN than that of CAR. This study also shows that B3LYP/ 6-311++G(d,p) with a posteriori counterpoise correction (CPSP) is an accurate method for the calculation of the H-bonding properties. G


Article

The Journal of Physical Chemistry B Constituents in Kaempferia Parviflora by Gas Chromatography. J. Chromatogr A 2007, 1143, 227−233. (19) Fernando, J. M. R. C.; Senadeera, G. K. R. Natural Anthocyanins as Photosensitizers for Dye-Sensitized Solar Devices. Curr. Sci. 2008, 95, 663−666. (20) Hemandez-Martinez, A. R.; Vargas, S.; Estevez, M.; Rodriguez, R. Dye-Sensitized Solar Cells from Extracted Bracts Bougainvillea Betalain Pigments. First Int. Congr. Instrum. Appl. Sci. 2012, 1−15. (21) Ali, R.; Nayan, N. Fabrication and Analysis of Dye-Sensitized Solar Cell Using Natural Dye Extracted from Dragon Fruit. Int. J. Integr. Eng. 2010, 2, 55−62. (22) Kamel, M. M.; El-Shishtawy, R. M.; Yussef, B. M.; Mashaly, H. Ultrasonic Assisted Dyeing: III. Dyeing of Wool with Lac as a Natural Dye. Dyes Pigm. 2005, 65, 103−110. (23) Davies, K. Plant Pigments and Their Manipulation; Blackwell Publishing Ltd.: Oxford, U.K., 2004. (24) Yasushi, K. New Trends in Photobiology: Structures and Functions of Carotenoids in Photosynthetic Systems. J. Photochem. Photobiol., B 1991, 9, 265−280. (25) Koyama, Y.; Kuki, M.; Andersson, P. O.; Gillbro, T. Singlet Excited States and the Light Harvesting Function of Carotenoids in Bacterial Photosynthesis. Photochem. Photobiol. 1996, 63, 243−256. (26) Konovalova, T. K.; Gao, Y.; Schad, R.; Kispert, L. D.; Saylor, A.; Brunel, L.-C. Photooxidation of Carotenoids in Mesoporous MCM-41, Ni-MCM-41 and Al-MCM-41 Molecular Sieves. J. Phys. Chem. B 2001, 105, 7459−7464. (27) Konovalova, T. A.; Gao, Y.; Kispert, L. D.; van Tol, J.; Brunel, L.-C. Characterization of Fe-MCM-41 Molecular Sieves with Incorporated Carotenoids by Multifrequency Electron Paramagnetic Resonance. J. Phys. Chem. B 2003, 107, 1006−1011. (28) Gao, Y.; Konovalova, T. A.; Xu, T.; Kispert, L. D. Electron Transfer of Carotenoids Imbedded in MCM-41 and Ti MCM-41: EPR, ENDOR, and UV Vis Studies. J. Phys. Chem. B 2002, 106, 10808−10815. (29) Gao, Y.; Konovalova, T. A.; Lawrence, J. N.; Smitha, M. A.; Nunley, J.; Schad, R.; Kispert, L. D. Interaction of Carotenoids and Cu2 in Cu-MCM-41: Distance-Dependent Reversible Electron Transfer. J. Phys. Chem. B 2003, 107, 2459−2465. (30) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; et al. A New Family of Mesoporous Molecular Sieves Prepared with Liquid Crystal Templates. J. Am. Chem. Soc. 1992, 114, 10834−10843. (31) Krishna, R. M.; Prakash, A. M.; Kevan, L. Photoionization of NAlkylphenothiazines in Mesoporous SiMCM-41, AlMCM-41, and TiMCM-41 Molecular Sieves. J. Phys. Chem. B 2000, 104, 1796−1801. (32) Sung-Suh, H. M.; Luan, Z.; Kevan, L. Photoionization of Porphyrins in Mesoporous Siliceous MCM-41, AlMCM-41, and TiMCM-41 Molecular Sieves. J. Phys. Chem. B 1997, 101, 10455− 10463. (33) Focsan, A. L.; Pan, S.; Kispert, L. D. Electrochemical Study of Astaxanthin and Astaxanthin n-Octanoic Monoester and Diester: Tendency to Form Radicals. J. Phys. Chem. B 2014, 118, 2331−2339. (34) Pan, V. H.; Tao, T.; Zhou, J.-W.; Maciel, G. E. Hydrogen Bonding between Acetone and Silica Gel, as Studied by NMR. J. Phys. Chem. B 1999, 103, 6930−6943. (35) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009. (36) Boys, S. F.; Bernardi, F. Some Procedures with Reduced Errors. Mol. Phys. 1970, 19, 553−566. (37) Gao, Y.; Kispert, L. D. A DFT Study of the Interaction Between Olefins and Cu2+ on Silica and MCM-41 Model Surfaces. Dalton Trans. 2014, 43, 6221−6228. (38) Gao, Y.; Kispert, L. D.; van Tol, J.; Brunel, L.-C. Electron SpinEcho Envelope Modulation and Pulse Electron Nuclear Double Resonance Studies of Cu2+···Carotene Interactions in Cu-MCM-41 Molecular Sieves. J. Phys. Chem. B 2005, 109, 18289−18292.

(39) Arunan, E.; Desiraju, G. R.; Klein, R. A.; Sadlej, J.; Scheiner, S.; Alkorta, I.; Clary, D. C.; Crabtree, R. H.; Dannenberg, J. J.; Hobza, P.; et al. Definition of the Hydrogen Bond. Pure Appl. Chem. 2011, 83, 1637−1641. (40) Dingley, A. J.; Grzesiek, S. Direct Observation of Hydrogen Bonds In Nucleic Acid Base Pairs By Internucleotide 2JNN Couplings. J. Am. Chem. Soc. 1998, 120, 8293−8297. (41) Isaacs, E. D.; Shukla, A.; Platzman, P. M.; Hamann, D. R.; Barbiellini, B.; Tulk, C. A. Covalency of the Hydrogen Bond in Ice: a Direct X-Ray Measurement. Phys. Rev. Lett. 1999, 82, 600−603. (42) Plumley, J. A.; Dannenberg, J. J. A Comparison of the Behavior of Functional/Basis Set Combinations for Hydrogen-Bonding in the Water Dimer with Emphasis on Basis Set Superposition Error. J. Comput. Chem. 2011, 32, 1519−1527. (43) Méndez-Hernández, D. D.; Tarakeshwar, P.; Gust, D.; Moore, T. A.; Moore, A. L.; Mujica, V. Simple and Accurate Correlation of Experimental Redox Potentials and DFT-calculated HOMO/LUMO Energies of Polycyclic Aromatic Hydrocarbons. J. Mol. Model. 2013, 19, 2845−2848.

H


Hydrogen Bond Formation between the Carotenoid Canthaxanthin

Recommend Documents