Photochemistry and Photophysics in Silica-Based Materials: Ultrafast

Photochemistry and Photophysics in Silica-Based Materials: Ultrafast...

0 downloads 92 Views 35MB Size

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Review Cite This: Chem. Rev. 2017, 117, 13639-13720

pubs.acs.org/CR

Photochemistry and Photophysics in Silica-Based Materials: Ultrafast and Single Molecule Spectroscopy Observation Noemí Alarcos,† Boiko Cohen,† Marcin Ziółek,‡ and Abderrazzak Douhal*,† †

Departamento de Química Física, Facultad de Ciencias Ambientales y Bioquímica, and INAMOL, Universidad de Castilla-La Mancha, Avenida Carlos III, S.N., 45071 Toledo, Spain ‡ Quantum Electronics Laboratory, Faculty of Physics, Adam Mickiewicz University, Umultowska 85, 61-614 Poznań, Poland ABSTRACT: Silica-based materials (SBMs) are widely used in catalysis, photonics, and drug delivery. Their pores and cavities act as hosts of diverse guests ranging from classical dyes to drugs and quantum dots, allowing changes in the photochemical behavior of the conﬁned guests. The heterogeneity of the guest populations as well as the conﬁnement provided by these hosts aﬀect the behavior of the formed hybrid materials. As a consequence, the observed reaction dynamics becomes signiﬁcantly diﬀerent and complex. Studying their photobehavior requires advanced laser-based spectroscopy and microscopy techniques as well as computational methods. Thanks to the development of ultrafast (spectroscopy and imaging) tools, we are witnessing an increasing interest of the scientiﬁc community to explore the intimate photobehavior of these composites. Here, we review the recent theoretical and ultrafast experimental studies of their photodynamics and discuss the results in comparison to those in homogeneous media. The discussion of the conﬁned dynamics includes solvation and intra- and intermolecular proton-, electron-, and energy transfer events of the guest within the SBMs. Several examples of applications in photocatalysis, (photo)sensors, photonics, photovoltaics, and drug delivery demonstrate the vast potential of the SBMs in modern science and technology.

CONTENTS 1. Introduction 2. Theoretical Studies 2.1. Solvation in Silica-Based Materials 2.1.1. Water and Silica Materials 2.1.2. Other Solvents and Silica Materials 2.1.3. Impact on Solvation Dynamics 2.2. Photoinduced Processes in Trapped Monomers and Aggregates 2.2.1. Impact of SBMs on Absorption and Emission Spectra of Chromophores 2.2.2. Impact on Excited-State Dynamics 2.2.3. Impact of Guest Aggregation 3. Ensemble Average Time-Resolved Studies of Photoinduced Processes in Silica-Based Materials 3.1. Proton-Transfer Reactions 3.1.1. Intermolecular Proton-Transfer Processes 3.1.2. Intramolecular Proton-Transfer Reactions 3.2. Electron-Transfer Reactions 3.2.1. Intramolecular Charge-Transfer Reactions 3.2.2. Intermolecular Electron-Transfer Reactions 3.3. Energy-Transfer Events 3.3.1. Energy Transfer in Dye-Doped Zeolites

© 2017 American Chemical Society

3.3.2. Energy Transfer in Dye-Doped Mesoporous Materials 3.3.3. Donor/Acceptor Concentration Eﬀects on Energy Transfer 3.3.4. Energy Transfer in Silica-Coated Quantum Dots and Metal Nanoparticles 3.4. Homo-Energy Transfer and Aggregates Formation 4. Single Molecule 4.1. Single Molecule Reactivity in Silica-Based Materials 4.2. Single Molecule Diﬀusion in Silica-Based Materials 4.3. Other Studies 4.4. Electronic Nanoconﬁnement 5. Recent Applications of Silica-Based Materials 5.1. Photocatalysis 5.2. Photonics 5.2.1. Sensors 5.3. Photovoltaics 5.4. Drug Delivery 6. General Conclusion and Outlook Author Information Corresponding Author ORCID

13640 13642 13643 13643 13644 13645 13647 13647 13648 13649 13650 13650 13651 13657 13659 13660 13665 13668 13669

13670 13673 13674 13676 13680 13681 13682 13682 13683 13684 13684 13689 13691 13695 13697 13700 13701 13701 13701

Received: July 13, 2017 Published: October 25, 2017 13639

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Scheme 1. Frameworks of the Silica-Based Materials (SBMs) Used in the Studies Commented in This Reviewa

a

(A) and (B) show the microporous and mesoporous hosts, respectively.

Notes Biographies Acknowledgments Dedication Abbreviations General Abbreviations Symbols and Rate Constants Material Structures Molecules References

guest dynamics and spectroscopy diﬀerent from those in solutions.2,7,44−49 In recent years, many studies of ensemble average ultrafast (femtosecond regime) processes within the conﬁned space provided by the nanochannels and nanocavities of the zeolite and mesoporous materials have been published. However, the most recent systematic review of photodynamical studies using SBM composites was published in 2003.4 Thus, there is a strong need for a detailed review of the recent contributions on the photoinduced fast and ultrafast dynamical properties of such SBMs interacting with chromophores and drugs. Here, by fast and ultrafast dynamics we mean events happening in picoand femtosecond (ps-fs) regimes. The present review mainly focuses on contributions since 2005. However, in a few places, earlier relevant contributions are also presented. We also review recent works using time-resolved ﬂuorescence microscopy applied to SBM composites. From the point of view of the framework materials, our interest is focused on microporous silica and aluminosilica sieves (e.g., zeolites), mesoporous silica sieves (e.g., MCM41, SBA15), silica nanoparticles, and silica shells. Scheme 1 shows the framework of these SBMs conﬁning media, which have been used in the published works commented in this review. As for the encapsulated chromophores, the main part is devoted to organic molecules, but quantum dots (QDs), metal nanoparticles, and perovskites are also presented. The chemical structure of these guests is shown in Schemes 2−8. Theoretical studies of the photophysical and photochemical processes in silica-based composites are, without doubt, much more diﬃcult in comparison to the calculations of isolated molecules or even those in a homogeneous environment. Caution must be taken when interpreting the results, due to the use of more approximated methods or base sets. Similarly, timeresolved experiments in such hybrid systems become a challenge against conventional studies in solutions because of their more complex and less repeatable samples preparation, diﬃculties in successful measurements (e.g., studies in solid state or in suspensions where the light is severely scattered),

13701 13701 13701 13701 13701 13701 13702 13702 13703 13704

1. INTRODUCTION Silica-based materials (SBMs) are one of the most studied framework systems in advanced chemical and biological applications, being used in catalysis, photonics, and drug delivery. They act as hosts for diverse organic molecules (dyes and drugs for example), allowing changes in the photochemical behavior of the conﬁned guests. The formed composites show a variety of photophysical and photochemical events that could be used to develop smart devices, such as drug nanocarriers, nanosensors, nanoOLEDs, nanolasers, energy storage nanospace, and nanophotocatalysts. Several review articles on the structural properties of SBMs and their composites with diﬀerent chromophores have been published.1−17 However, these works have not been oriented toward the photodynamics in the formed composites. They rather focus on the photochemistry at long time scale (nanosecond-millisecond regime). Therefore, a great knowledge on slow photochemical events and stability of the formed photoproducts has been reached and used for some application purposes.3,5,13−15,18−28 More than two decades ago, ultrafast laser-based techniques allowed one to witness the chemical bond and electronic events in real-time of reactions.29,30 These techniques are being used to characterize the molecular events in condensed phase, providing details on solvation, electron, proton, and energy transfer events. 16,27,31−43 Conﬁning molecules within chemical and biological cavities leads to 13640

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Scheme 2. Molecular Structures of the Compounds Interacting with SBMs and Discussed in the Theoretical Section 2

streak-cameras, and ﬂuorescence microscopy), the possibility of studying complex systems without having a detailed knowledge in ultrafast laser physics is growing. An analogous situation happens in the ﬁeld of theoretical calculations, where the rapid development of computational tools has empowered the ﬁeld in the two last decades. In addition to the Introduction and Conclusion, this review contains four sections. The review is organized as follows: the subsequent section (section 2) is devoted to the theoretical studies. Its ﬁrst section 2.1 examines the ultrafast dynamics of water and other solvents within the pores and cages provided by the SBMs, the solvation eﬀects on the solutes, and its impact on the observed dynamics and spectroscopy. The second section of the theoretical part (section 2.2) presents the studies of the electronic coupling between the guest and the framework of the host and its relevance to the photobehavior of the

and often a more complicated data analysis. We believe that the time of ultrafast studies of simple molecular systems dissolved in solutions is passing. Such systematic studies in the past 30 years resulted in profound knowledge about the photophysical processes that occur in solutions on the time scales of femtoseconds (fs), picoseconds (ps), and nanoseconds (ns). The rapid increase in novel technologies requires the scientists to focus more on systems that are application-oriented, including the composites formed by chromophores in SBMs. The basic studies of novel chromophores in solutions are still necessary (e.g., to get to know the eﬀects of surrounding polarity, viscosity, pH, and speciﬁc interactions) but rather as references to investigate more complex environments. With the recent development of turn-key ultrafast laser systems of high stability, and commercially available time-resolved spectrometers (e.g., transient absorption in UV, visible, NIR ranges, upconversion, time-correlated single-photon counting (TCSPC), 13641

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

in areas such as photocatalysis, photonics, sensors, photovoltaics, and drug delivery. A few results of ns−ms luminescence (that were not presented in parts 3−4) related to application-oriented studies are also commented in section 5. The review ends with the conclusions that can be drawn on the basis of the photophysical processes within SBMs, and we address future challenges and the possibility of how to face them. The goal that this review, among others, is to provide a guide of general changes in the photodynamics (with respect to homogeneous solvents) that one can expect when considering a given chromophore interacting with silica-based hosts and nanostructures. For example, excited guests, which have a signiﬁcant change in their dipole moments, should exhibit slower solvation (nanosolvation) with respect to that in solutions, and the extent of this slowing down depends on their position and adsoption on the silica pores and framework governing the speciﬁc and nonspeciﬁc interactions in the composites. When the guest nonradiative deactivation depends on its internal rotation or twisting, we anticipate that its ﬂuorescence lifetime in the composite will increase as a result of restriction in motion. On the other hand, other processes can make the photodynamics faster in SBMs than those in solutions. Excited state lifetime might become shorter due to self-quenching (homo energy transfer) upon high dye loading in the pores, and the size of host pores might favor or disfavor aggregates formation. On the other hand, metal doping of silica materials can trigger additional deactivation channels, due to charge transfer (e.g., electron injection). Small modiﬁcations in acidity or basicity inside silica pores might inﬂuence the protontransfer reaction rates and promote the deactivation route at the expense of another one. Often, it may even lead to the occurrence of a new molecular form (e.g., radical anions, cations, and CS state) in the ground state, which drastically changes the overall photophysics and photochemistry of the conﬁned system. These are the most common and simplest examples, with much more details, special cases, and competing factors outlined in the following parts. Our review of the fs−ps time-resolved studies of the photoinduced processes taking place in the hybrid SBM complexes should give a concise perspective on the clues and knowledge of the challenges found in the photocharacterization of these composites. One of its goals is to trigger further research to advance in the related subﬁelds and emerge with others in science and technology. We hope that the review will be of great interest to a wide scientiﬁc community (in chemistry, nanomaterials, spectroscopy, physics, nanomedicine, and pharmacology) that will beneﬁt from having at hand the state-of-the-art in the ultrafast photodynamics within silicabased composites, encouraging more and more scientiﬁc groups to join this fascinating and promising ﬁeld to boost the knowledge opening doors for new smart applications of SBM composites.

formed composites. The eﬀects of aggregation and energy transfer between closely laying guests are also considered. Section 3 (ensemble average time-resolved studies of photoinduced processes in SBMs) is the main and the longest section of the review, presenting and discussing the recent reports on the dynamics and spectroscopy of diﬀerent photoinduced processes taking place within silica-based hosts on the time scale from fs to ns. Section 3.1 deals with the excited-state proton-transfer reactions that can be either intra or intermolecular in nature. We review the eﬀects that the silica-based hosts have on these processes. The framework of the host can aﬀect directly or indirectly the excited-state proton-transfer reactions by modifying the electronic distribution within the guest and, thus, altering its dynamics and subsequent events. In speciﬁc cases, the framework can act as a proton acceptor/donor, inducing an intermolecular protontransfer reaction within the composites. The eﬀects of the pore size, metal doping, and functionalization of the host on the proton-transfer event of the guest are also discussed. Section 3.2 is devoted to electron-transfer reactions (one of the most important reactions in nature), and it includes both the passive and active role of the host. For the ﬁrst case, the host frameworks can aﬀect the charge separated (CS) states of the encapsulated guest, together with the rate constants of intraand intermolecular charge transfer reactions, and the nanosolvation dynamics. As for the active role, framework modiﬁcation using metal doping functionalization or pore size can further aﬀect the charge separation (within the guest) and can even facilitate electron injection to the doping metal. We also examine the eﬀects of speciﬁc and nonspeciﬁc interactions between the guest and the host on the electron-transfer events. SBMs are one of the most suitable hosts to build multichromophoric scaﬀolds for energy transfer, and the related reports dedicated to understanding these events and developing multichromophoric and hybrid systems by functionalization, doping, and pore size modiﬁcation are commented on in section 3.3. The eﬀects of energy donor and acceptor concentrations in the nanohosts on the processes in silicacoated QDs and metal nanoparticles are considered as well. Finally, section 3.4 analyzes the results from the point of view of molecular aggregates formation. The limited space of the nanochannels and nanocavities of the zeolites and mesoporous materials creates physical conditions for strong intermolecular interactions between the encapsulated guest molecules, thus giving rise to H- and J-aggregates of the caged monomers. We discuss how this type of guest−guest interactions shapes the photophysical behavior of the composites, and how doping and functionalization of the host can aﬀect the distribution of the aggregates and interacting monomers. Section 4 of the review shows the most recent studies oriented toward using the photophysical properties of the guests to characterize the hybrid complexes at single molecule/ particle level. Knowing the interaction of ﬂuorescent dyes with the silica-based host framework is paramount to deciphering the distribution of the guests and the type of speciﬁc and nonspeciﬁc interactions in the composites, by passing the ensemble average limitations. Special attention is paid to the electronic nanoconﬁnement, reﬂecting the eﬀect of the host framework on the photophysical properties of the guest from the point of view of orbitals coupling. Section 5 summarizes the most relevant applications of SBM composites that are related to the photophysical and photochemical processes presented in sections 2−4. Thus, we focused on the applications using light

2. THEORETICAL STUDIES There is a strong discrepancy between the number of theoretical and experimental works that examine ultrafast photoinduced processes in silica-based materials (SBMs). To predict the behavior of the system after light absorption, the interaction of the excited state of the organic molecules with silica-based structures must be optimized, which is diﬃcult not only because of the large systems that are involved but also due to the lack of proper methodology, which is still being 13642

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

shell model) divides solvent (water) molecules into two groups: (1) bound (surface) and (2) free (bulk) water. Surface, or shell, molecules are assumed to have signiﬁcantly modiﬁed properties due to their direct interaction with silica and their immobilization on its surface via H-bonding. In contrast, core molecules distant from the interface are considered to have bulklike properties. A given property of a conﬁned liquid should scale as 1/R and 1/R2 in cylindrical and spherical conﬁnements, respectively, due to the populations of the molecules in the bound and free states (R is the radius of the conﬁning space). A more complex model assumes that the inﬂuence of the interface extends further and more smoothly into the liquid than the step function used in a two-state model describes. A dynamic exchange model attributes the long-time dynamics to the exchange of bound and free water (a bound water molecule can only take part in solvation dynamics by exchanging with a free water molecule). This model leads to the prediction of a temperature dependence of the solvation dynamics that is due to an activation energy caused by a free energy barrier. This barrier is generated by the H-bonding character of water molecules with the silica surface. Finally, a model with chromophore diﬀusion assumes that changes occur in the position of the solute molecule when its dipole moment changes (excitation). We will discuss this model in the next section. More recently, the reorientation dynamics of water in nanoscale hydrophilic or hydrophobic silica pores were investigated using molecular dynamics simulations.51 The reorientation of water in hydrophilic conﬁnement occurs in more than one nanosecond (2 orders of magnitude longer than in bulk water) and exhibits a power-law decay. The hydrophobic pores (simulated as silica with no charges) aﬀect more modestly the dynamics, making water molecules reorientation only 2.5 times slower than that in the bulk. Moreover, the power-law decay of the correlation function in the hydrophilic pores was explained in terms of heterogeneity of the surface and demonstrates the incompleteness of the two-state model. The H-bond jump model was used to study the heterogeneity where H-bond acceptors on the pore surface (bridging oxygens or silanols) exchange H-bonding interactions with water molecules. This mechanism, when combined with the surface roughness, leads to a distribution of jump times. Obviously, water molecules cannot form H-bonds with the interface when they are in the hydrophobic pores, and the reorientation dynamics is governed only by the local excluded volume. Molecular dynamics simulations combined with analytic modeling were also used to explain why water reorientation dynamics does not vary monotonically with the surface hydrophilicity.52 The increasing hydrophilic character is accompanied by two competing eﬀects within the interfacial layer. The ﬁrst eﬀect is a rearrangement of water molecules. In this scenario, when the hydrophilicity of a surface increases, an increasing fraction of water molecules point their OH groups toward the surface. Further increases in hydrophilicity increases the interaction energy, slowing down the reorientation times. It is well-established that while fully hydroxylated surfaces (surface density of 4−5 Si−OH groups per nm2) have a hydrophilic character, highly dehydroxylated ones (approximately 1−2 Si−OH per nm2) manifest a hydrophobic character.11 The extended jump model in cases of hydrophilic conﬁnement was further studied by spatially resolving the Hbond jump dynamics among individual sites on the silica pore surface.53 The nonexponential dynamics was explained by a

developed (even in the case of small molecules). Another problem is the veriﬁcation of the calculated results: the calculations are usually performed on small molecules, while the experiments, involving dyes on SBMs, are usually performed with large molecules possessing the desired properties and generating strong enough signals to analyze (e.g., high extinction coeﬃcient in the visible range, strong dipole moment changes upon excitation, etc.). Compared with the case just described, there are relatively many theoretical reports that address ground-state interactions and adsorption of gases and small liquid molecules on SBMs. These studies, particularly relevant to catalysis, have been recently reviewed.15 Another area of theoretical studies of adsorption and diﬀusion of small molecules in mesoporous silica is devoted to sensing application (see section 5). For example, a physico-mathematical model for the related kinetics was recently proposed to predict the kinetic behavior of the explosive trinitrotoluene (TNT) in mesoporous ﬁlms.50 On the contrary, studying the adsorption of large biomolecules is important in, for example, drug delivery (see section 5). Such calculations have been recently commented on.11 The above reviews also outlined the main theoretical methods that are used to get the structures of SBMs and their interactions with the organic guest molecules.11,15 Therefore, these subjects will not be considered in this review. Below, we mainly focus on the theoretical works that are dedicated to ultrafast and fast dynamics in SBMs and the properties of the nanoconﬁned molecules in their excited states or, at least, those that predict their stationary absorption and emission spectra. The ﬁrst section is devoted to solvation dynamics that inﬂuences the relaxation of the excited states of the dyes. In the second section, the direct interaction of the dyes with SBMs will be considered with regards to photoinduced processes. 2.1. Solvation in Silica-Based Materials

Solvation dynamics is determined by the ability of solvent molecules to reorganize around a chromophore, thereby minimizing the energy of the resulting solvent−solute entity. The typical and most frequently used way to study solvation dynamics is the electronic excitation of a molecule and gating its ﬂuorescence signal at short and diﬀerent times after its excitation. A suitable chromophore should have diﬀerent dipole moments in the ground and excited states in a polar solvent. If the electronically ﬁrst excited state (S1) has signiﬁcant intramolecular charge transfer (ICT) character, solvent molecules around the solute move and rotate to assume the energetically optimum position. The stabilization of the excited solute in the new electronic conﬁguration is then accompanied by a redshift in its ﬂuorescence spectra (Stokes shift). This shift with the gating time is analyzed to get solvation time dynamics. 2.1.1. Water and Silica Materials. An important result from the studies of solvation dynamics is predicting the interactions of solvent molecules with SBMs, particularly in materials in which nanoconﬁnement occurs. Such studies will be brieﬂy discussed below, while the experimental ﬁndings are presented in section 3.2.1. Not surprisingly, water is the most frequently theoretically studied solvent because of its simple molecular structure, well-known strong interactions with SBMs, and relevance in biology and medicine, and to our life. Advances in the ﬁeld using theoretical tools were reviewed in 2011.33 Diﬀerent models describing the dynamical behavior of solvent molecules in conﬁned structures were summarized there. The simplest model (the so-called two-state or core− 13643

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 1. Picture of the molecular dynamics of water in MCM41, after (A) 3.5 ps (potential energy equilibration reached) and (B) 6.5 ps. The water shell (blue) is almost identical at both times. Reprinted with permission from ref 58. Copyright 2016 Royal Society of Chemistry.

correlation of the spectra at a distance of 2 Å, even until 1 ps. While at a distance of 5 Å, the OH groups lose the memory of their frequency much faster (although still slower than in the bulk liquid). Therefore, the theoretical report shows that it should be possible to probe the slower spectral diﬀusion of conﬁned water molecules (compared to the bulk liquid) by performing and analyzing 2D-IR spectra. Another recent work used density functional theory (DFT) based-molecular dynamics simulations to study water interacting with SBMs. Amorphous silica surfaces and quartz surfaces were compared when they were in contact with water.57 It was concluded that the water layer at the amorphous silica interface is less dipolar than the one at the crystalline silica interface. The layer of water molecules at the amorphous silica surface was found to be immobile, and an adsorbed water molecule would probably not be easily replaced by adsorbate organic molecules. Other composite models at the DFT level with explicit solvent water molecules were tested for the classical Mobil Catalytic Materials of Number 41 (MCM41).58 Here, ab initio molecular dynamics using an MCM41 channel having a silanol density of 5.8 Si−OH nm−2 showed that the ﬁrst water layer is strongly adsorbed on the silica wall (blue in Figure 1). The mobility of the ﬁrst shell of water is qualitatively very low, and water molecules stay close to the channel surface and do not diﬀuse during sampling times of up to 8 ps (Figure 1). 2.1.2. Other Solvents and Silica Materials. The interaction of nonaqueous solvents with SBMs was also theoretically studied. Molecular dynamics simulations of a silica/acetonitrile interface were performed.59 Acetonitrile molecules formed an arrangement at the silica surface that is reminiscent of a lipid bilayer. Such a conﬁguration is considerably diﬀerent from an antiparallel and dipole-paired structure in the bulk liquid. It was also found that the silica interface can aﬀect the bulk liquid structure over a distance of tens of angstroms. A similar bilayer structure was considered at the interface between silica and propionitrile.60 However, the CN vector of the cyanide group of propionitrile is more parallel to the surface than that of acetonitrile. As a result, the propionitrile bilayer is more compact and ordered than the corresponding acetonitrile one (Figure 2A).

broad distribution of rate constants for the exchange of Hbonding partners, and the distribution was due to a variety of local topographies at the silica surface. It was found that the key factors that determine the distribution of jump types were entropic: speciﬁcally, an excluded-volume eﬀect for the approach of a new partner and the elongation of the initial H-bond. However, enthalpy eﬀects arising from the chemical heterogeneity of the pore surface contributed only weakly to the distribution. The same research group investigated the hydrophobic conﬁnement in all-silica Linde type A (LTA) zeolites.54 Water molecules reorientation was found to be retarded by only a factor of 2−3 when compared with the bulk values. The slowdown of water dynamics was more pronounced at higher water loadings and was predominantly caused by an excluded-volume eﬀect (large fraction of water molecules lying at the interface within the zeolite matrix). The authors suggested that the presence of the interface and its chemical nature have a much greater impact on water dynamics than does the conﬁnement. Classical molecular dynamics simulations were used to study the interaction of water with the external surface of silicalite-1.55 Despite the hydrophobic behavior of the silicalite structure and the presence of hydrophilic OH groups on the surface, water enters the zeolite pore system and forms a layer. The adsorbed water molecules on this internal layer are H-bonded to each other and to a layer of molecules that are adsorbed on the surface and bound to the silanol groups. The surface layer has an ordered structure with a thickness of approximately 7 Å. The performed calculations were not only aimed at explaining experimental ﬁndings but also predicting new results and phenomena. Molecular dynamics calculations were recently used to simulate the results of a two-dimensional infrared spectroscopy (2D-IR) experiment, in which a novel ultrafast technique probed the correlation and coherence times of the molecular vibrations.56 The results predict that while the frequency correlation of OH vibrations is nearly complete in approximately 2 ps for bulk water, in a silica pore, this correlation is signiﬁcantly longer (6 ps). Moreover, the 2D description of the distance of OH vibrations from the pore surface drastically changed: there was little loss in the frequency 13644

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 2. (A) Comparison of the CN (vector of cyanine group) density (ρ) in acetonitrile and propionitrile (z is the center of the CN bond, ρB is the bulk density. Reprinted from ref 60. Copyright 2012 American Chemical Society. (B) Number of hydrogen bonds per methanol molecule as a function of distance from the silica surface and resultant polarity of the environment. Reprinted from ref 61. Copyright 2013 American Chemical Society.

Figure 3. (A) Variation of free energy proﬁles, ΔA(x), and probability distributions, P(x), of a solute of diﬀerent dipole moments: 5 D (black), 10 D (blue), and 15 D (green) in the hydrophilic pore. The ﬁgure shows that the solute of a larger dipole moment (e.g., in the excited state) is located further from the pore walls. Reprinted from ref 64. Copyright 2012 American Chemical Society. (B) A model system with a spherical solute in a Lennard-Jones ﬂuid conﬁned between a slit of parallel walls at the distance H, below the solvation time obtained for solvophilic and solvophobic walls (solute diameter is 3σ). Reprinted from ref 65. Copyright 2013 American Chemical Society.

methanol molecules that were dissolved in acetonitrile and conﬁned in nanoscale hydrophilic silica pores.63 Methanol molecules with speciﬁc interactions were found to exist in two distinct conformations that diﬀer in their H-bonding states, which resulted in two maxima in the position-dependent densities of the molecules with respect to the distance from the silica surface. Both entropy as well as H-bonding were found to play important roles in determining the location and orientation of the methanol molecules. 2.1.3. Impact on Solvation Dynamics. The conﬁgurations of solvent molecules and their reorientation times that are commented above have a direct eﬀect on the solvation dynamics of chromophores that interact with silica surfaces in the presence of solvent. Experimental observations from timedependent ﬂuorescence measurements (that were supported by theoretical calculations) were divided into three groups: (i) increase in solvation times observed in bulk solution, (ii) appearance of new, longer times in the solvation dynamics, and (iii) lack of any diﬀerences among the chromophores in bulk solutions and those within silica structures.33 The used chemical models (SBMs interacting with a solvent or a solute) before 2010 that were proposed to explain the experimental observations for nanoconﬁned silica structures have also been summarized.33 It was argued that the chemical models with

Another theoretical report showed that the Coulombic interaction of the π-electrons of benzene ring with the partial charges carried by the silica atoms is of great importance in the stability of the complex.62 The conﬁned benzene exhibits signiﬁcant layering, and the rings of molecules close to the interface tend to lie ﬂat on the silica surface. The rotational dynamics of this solvent in the center of the pore was found slightly slower when compared to that of the bulk, with more pronounced slowing in smaller pores. For example, the reorientation time at 293 K was predicted to be ∼6 and ∼7 ps in nanopores with diameters of 3.6 and 2.0 nm, respectively, while at the same temperature, this time was predicted ∼5 ps for bulk benzene. Molecular dynamics simulations were also used to characterize methanol.61 Strong H-bonds between the ﬁrst layer of methanol and the silanol groups of silica create a methylterminated surface that results in a second layer having signiﬁcantly less density and H-bonds than the bulk solution (Figure 2B). In agreement with the experimental ﬁndings, the interfacial environment in this layer resembles to a nonpolar solvent. Moreover, solvent reorientation times, in these ﬁrst two layers, are signiﬁcantly slower than those observed in the bulk solvent. Recently, replica exchange molecular dynamics simulations were used to investigate the location of acetone and 13645

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 4. (A) Coumarin 153 (C153) in methanol conﬁned in diﬀerent silica pores (from left to right: HOC, HIC, and RHOC). (B) Simulation of changes in solvation-response function, S(t), with the time. The inset shows the unnormalized energy gaps, ΔE(t), which represents the environment contribution (i.e., MeOH and pore walls), to the energy gap between the S1 and S0 Born−Oppenheimer potential energy surfaces of C153 at time t. Reprinted from ref 67. Copyright 2011 American Chemical Society.

restriction of solvent dynamics and faster in-plane solvent diﬀusion. More recently, studies of coumarin 153 (C153, Scheme 2) in ethanol and conﬁned in 2.4 nm hydrophilic amorphous silica pore have been reported.66 At the ground state, the most likely location of C153 when interacting with the SBMs is near the pore surface, but two possible H-bonding structures lead to diﬀerent orientations. Internal energy and entropy competed within the pore. The implication for solvation dynamics is that the two calculated orientations might lead to diﬀusions of diﬀerent rates upon excitation. These diﬀerences in rates are because the ﬂat orientation of C153 has a larger eﬀective surface area (slower diﬀusion) to move into the interior than the second, perpendicular orientation. In a subsequent work, the positions of C153 in a hydroxyl-terminated silica pore system was studied in both the ground and excited states.66 The predicted movement of C153 toward the pore interior upon excitation was attributed to more favorable nonspeciﬁc ethanol solvation of the large dipole moment in the excited C153, which occurred at the expense of H-bonding with the pore. Interestingly, the free energy variations were calculated to be comparatively small (1−3 kcal/mol), which suggests that relatively modest changes in the conﬁning framework or solute interactions may generally have a signiﬁcant impact on the distribution of solute positions. Solvation response functions of C153 were also simulated from averages of non-equilibrium independent trajectories obtained in molecular dynamic calculations.67 The investigated solvent was methanol, and three kinds of silica pores were examined (Figure 4A): hydrophobic cavities (HOC, in which wall-solvent interactions were exclusively of the dispersive, Lennard-Jones type), hydrophilic cavities (HIC, in which unsaturated oxygen sites at the wall were transformed into OH groups), and rugged pores (RHOC, in which 60% of the polar groups were transformed into trimethylsilyl moieties). The results of the simulation within the ﬁrst 30 ps are presented in Figure 4B. The overall responses were found to be between 2 and 4.5 times slower than the one observed in the bulk methanol, the one associated with RHOC being the fastest relaxation and the one corresponding to HIC the slowest. Finally, the kinetic processes occurring on the ps−ns time scale were also considered in the calculations of diﬀusion of diﬀerent molecules within SBMs. Molecular dynamics simulations were performed and visualized on pure silica zeolite

diﬀerent solvent layers were insuﬃcient to accurately reproduce the experimental results, and that the eﬀects of chromophore diﬀusion upon excitation must be considered. This diﬀusion occurs when the dipole moment of the chromophore increases as a result of the electronic distribution changes from the ground to the electronically excited state, and this diﬀusion should be responsible for the long-time solvation dynamics. In addition to that, the restriction of the solvent molecules to move at the interface decreases its polarity relative to the host interior. Therefore, the dependence of the location of a solute molecule (guest) on its electronic charge distribution during the time getting the photoresponse of the composite, and the gradient in eﬀective solvent polarity are likely important features to take into account in understanding the behavior of these systems. The solute (guest) diﬀusion model has been further used. A model molecule that was solvated by ethanol in a nanoscale silica pore was examined as a function of its position (Figure 3A).64 The guest distribution depends on both its dipole moment and its interaction with the surface of the silica pore, which can be hydrophilic or hydrophobic in nature. The electrostatic guest-pore interactions were responsible for the shifts in the ﬂuorescence spectrum of the guest molecules in some locations near the hydrophilic wall. However, this eﬀect was absent when the guest is in the hydrophobic pore. Therefore, while the guest-position distributions depend on its dipole moment in both hydrophilic and hydrophobic pores, the ﬂuorescence spectra are only sensitive to the dye position in the former case. As a result, time-dependent ﬂuorescence measurements of the hydrophobic pore system could be insuﬃcient to observe guest diﬀusion after excitation. However, the statedependent on its position might still have signiﬁcant importance in applications of porous materials. The diﬀerences between hydrophilic and hydrophobic conﬁnement were also studied using another approach.65 The authors generalized the case of water to any solvent, using the terms of solvophilic and solvophobic probe. A substantial slowing of solvation dynamics around a solute in strong solvophilic conﬁnement is due to the suppression of ﬂuid diﬀusion in the presence of the closely lying attractive walls, in addition to restricted solvent dynamics. The solvation in strong solvophobic conﬁnement is slower than in the bulk but is not slowed as signiﬁcantly as in the solvophilic case (Figure 3B). In this case, the eﬀect was explained by the competition between 13646

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Similar conclusions were drawn from calculating the preferred positions of FL in zeolite L.73 As commented upon above, the overall redshift of features in the absorption spectrum in hydrated zeolite L from their positions in cyclohexane samples was found to originate mainly from the interactions between the counterions of the zeolite and the carbonyl group of FL. It was found that the direct eﬀect of water molecules on the excitation energies of the dye is negligible, but they do aﬀect its orientation inside the zeolite channel. The role of water molecules in Ibu interactions with an amorphous silica surface was also studied.74 Simulations revealed that the silica surface exhibits a higher aﬃnity for water than for Ibu. Van der Waals forces are the leading contributor to its adsorption, independently of whether the drug is H-bonded directly to the surface or via water molecules. The eﬀect of immobilization of cationic Ir(III) complexes on sodium aluminosilicate mesoporous glass was experimentally and theoretically studied.75 The calculations considered the chemical environment of the Ir (III) complex by including both the counterion (representing the active site of the matrix) and the dielectric constant of the matrix. At low dielectric constants (ε < 20), the counterion has a strong inﬂuence on the energy and orbitals of the excited states. The inﬂuence of zeolite Y framework on a nickel(II) complex was also examined.76 The deviations that occur in the geometry of the complex upon encapsulation result in changes of the highest occupied molecular orbital and lowest unoccupied ones (HOMO and LUMO respectively) energy levels, and in the electronic properties of the system. The calculated chemical hardness conﬁrmed that the encapsulated nickel(II) complex is more stable than the free one, and it can undergo more easily electron transfer (ET) reactions, thus performing photocatalysis better than the free one. The calculations of absorption spectra of organic molecules interacting with SBMs were performed for possible applications in catalysis. For example, (poly)aromatic cationic compounds have characteristic absorption bands in the visible region, and the beneﬁts of using a theoretical approach to identify them as potential candidates was suggested.77 The origin of their absorption spectra has been attributed to structurally diﬀerent species on the basis of calculated excitation energies. Moreover, it was shown that the use of time-dependent density functional theory (TD-DFT) computations on gas-phase compounds was insuﬃciently accurate.78 It was suggested that molecular dynamic simulations are necessary to consider geometrical deformations of carbonaceous compounds as these lead to broad absorption bands: the results of such simulations were in excellent agreement with the experimental data. DFT/TD-DFT calculations were also used to explain the result of experiments on t-azobenzene in Linde type L zeolites (t-Ab and LTL in Scheme 2 and Scheme 1, respectively).79 In contrast to its planar structure in solution, t-Ab is twisted in LTL zeolites, as it is constrained by the framework. The photophysical properties in acidic (H-LTL) and potassium (KLTL) forms of the zeolites were explained on the basis of (n,π*) and (π,π*) transitions of the dye. In K-LTL zeolites, both transitions are present in the absorption spectra, with intensity maxima slightly red-shifted from their positions in solution samples. In contrast, in H-LTL zeolite, the (n,π*) transition was not observed as the lone electrons pair of the nitrogen atom forms a H-bond with a Brønsted acid site in the zeolite. These results conﬁrmed that t-Ab was protonated in H-

silicalite MFI framework. The microporous zeolite contains two channel systems that intersect tangentially, straight channels and zigzag or sinusoidal channels. The diﬀusion dynamics were simulated by modeling trapping in the subsequent cavities separated by thermally activated jumps through the channels.68,69 These simulations were performed maintaining both the framework ﬂexibility and the realistic account of electrostatic interactions with adsorbed water. Two types of adsorption, speciﬁc and weak unspeciﬁc, were predicted on the channel walls and at the channel intersection. The molecular diﬀusion of water in the silicalite found a barrier for crossing between the straight and the zigzag channels. Next, the analysis of the thermal motion predicted that at room temperature, framework oxygen atoms incurring into the zeolite channels signiﬁcantly inﬂuence the dynamics of adsorbed water.69 Further calculations were performed for ibuprofen (Ibu, Scheme 2) release from pH-gated silica nanochannels.70 The Ibu delivery process from cylindrical silica pores of 3 nm diameter, with polyamine chains anchored at the pore outlets, was investigated by means of massive molecular dynamics simulations. High, low, and intermediate pH environments were investigated. It was reported that the increment of the acidity of the environment leads to a signiﬁcant decrease of the pore aperture, yielding an eﬀective diameter, for the lowest pH case, that is 3.5 times smaller than the one associated with the highest pH one. The joint analysis of the corresponding Gibbs free energy proﬁles for the Ibu delivery process, the time evolution of its position within the channel, orientation of the molecule, and instantaneous eﬀective diameter of the gate suggest a three-step mechanism for Ibu delivery. A complementary analysis of its translational mobility along the axial direction of the channel revealed a subdiﬀusive dynamics in the low and intermediate pH cases.70 2.2. Photoinduced Processes in Trapped Monomers and Aggregates

In this section, results of theoretical predictions on the interactions of dyes with the SBMs will be commented on. First, we examine the eﬀects of the interaction with the silica environment on the absorption and emission spectra. Then, we address its impact on the excited state lifetimes and ﬁnish with the reports on interguests interaction that leads to energy transfer (EnT) between nearby trapped dyes within SBMs pore or cavity. 2.2.1. Impact of SBMs on Absorption and Emission Spectra of Chromophores. When trapped ﬂuorenone (FL, Scheme 2) in the channels of zeolite L was theoretically investigated, the interaction of its carbonyl group with the potassium cations of the zeolite framework was found to be responsible for its stabilization in the composites.71 The calculations predicted optical anisotropy in this composite, maintained upon contact with water, which can drastically alter the preferred position of the dye in the silica nanostructure. Such eﬀects were calculated too for oxonine and pyronine cationic dyes in zeolite L (Ox+ and Py+, respectively, Scheme 2).72 Diﬀerent preferred dye molecule orientations were obtained in the presence or absence of water, as a result of stabilization of either the host or the guest. In the hydrated composite, a perpendicular orientation of the Ox+ or Py+ molecules with respect to the zeolite channel axis was favored, while the order of preferred orientations was reversed under dry conditions. 13647

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 5. (A) Molecular structure of DEAC molecule showing the dihedral angle, θ. (B) Ground and excited state potential energy and oscillator strength as a function of θ. (C) Population density distributions as a function of time and θ. Reprinted with permission from ref 85. Copyright 2014 Elsevier B.V.

LTL, and that only this species can ﬂuoresce, as it does not connect with the dark (n,π*) state. In addition to organic chromophores as guests, absorption spectra of silica-encapsulated quantum dots (QDs) were also investigated.80 Strong coupling between Cd2Te2 QDs and silica shell was predicted, leading to the distortion of the silica nanocage. Therefore, a direct electronic transition from the occupied Cd2Te2 states to the outer silica nanocage excited states (core-shell electronic transitions) was predicted. As a result, the absorption peak of Cd2Te2 absorption spectrum is shifted from 584 nm (isolated) to 534 nm when the QDs is encapsulated by silica. In section 3.3, we comment on the experimental observations and the involved EnT processes. 2.2.2. Impact on Excited-State Dynamics. Tetramethylrhodamine isothiocyanate (TRITC, Scheme 2) in water and covalently encapsulated to silica nanoparticle was studied by a combination of TD-DFT and classical molecular mechanic simulations, in addition to ﬂuorescence spectroscopy.81 The authors studied the changes in the molecular structure upon electronic excitation and found that the variations were indeed smaller for the encapsulated dye than in water. A subsequent report explained the reason for the long excited state lifetime of TRITC in silica nanoparticles.82 It was found that the crossing from (π,π*) to (n,π*) state in water solutions is reduced by the rigid environment of the silica shell, decreasing the ﬂuorescence quenching of the bright (π,π*) state by the dark (n,π*) one. Two 7-aminocoumarin dyes that were incorporated into MCM41 were studied using DFT and TD-DFT calculations (coumarin 339 and 340, C339 and C340, respectively, Scheme 2).83 It was shown that the interaction of the carbonyl groups of these dyes with the silanol groups on the silica surface is responsible for stabilizing them in the composites and is strengthened upon photoexcitation. As a result, the computed

absorption and emission spectra of the incorporated dyes were red-shifted compared to the spectra in toluene. C339 and C340 molecules were further studied theoretically by comparing their photobehavior in ethanol and in Pluronic silica nanoparticles.84 Two molecules from that family were investigated: one of them with ﬂexible alkyl moieties and the other one has rigid alkyl groups. Electronic excitation results in a twisted-intramolecular charge-transfer (TICT) state in the former molecule, which leads to a short ﬂuorescence decay in solutions that is diﬀerent from that of the rigid molecule. However, both molecules showed similar photophysical behaviors when embedded into the nanoparticles, which was shown to be governed by the drastically hampered deactivation of the ﬂexible molecule in the silica matrix.84 Redshift in emission spectra (dye in MCM41 vs toluene solution) and their temperature eﬀects were calculated in a subsequent work by the same group.85 The authors used TD-DFT with a generalized Smoluchowski equation approach to compute time-resolved spectra of 7-diethylaminocoumarin3-carboxylic acid butylamine ester dye free in ethanol and encapsulated in silica nanoparticles (DEAC, Scheme 2 and Figure 5A). The predicted emission quenching occurred because of the conformational change from an ICT state to a TICT state (Figure 5). The rotation that leads to this change was signiﬁcantly hindered when the dye interacted with silica. The theoretically predicted restrictions in rotational motions of encapsulated molecules found conﬁrmation in experimental reports (see section 3). Possible ET reactions from the excited states of carotenoids to the mesoporous frameworks of MCM41 and its derivatives were studied, being of possible applications in dye-sensitized solar cells. Sections 3 and 5 comment on the involved events and their applications in photovoltaics. It was found that the formation of H-bonds stabilizes the canthaxanthin (CAN, 13648

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 6. (A) Kohn−Sham molecular orbitals of the dimer of 7-aminocoumarin. Reprinted with permission from ref 84. Copyright 2013 Royal Society of Chemistry. (B) Illustration of the distribution of guest (green bars) in a homogeneous (right) and inhomogeneous way (left) inside a zeolite L crystal for high loadings. Their impact on FRET dynamics is highlighted by the diﬀerent number of transfer steps (red circles) expected in a trajectory (indicated as black lines). The blue lines show the distribution of guests along the channels. Reprinted from ref 93. Copyright 2016 American Chemical Society.

a monomer to the binding LUMO of the dimer leads to a strong interaction between the two molecules. Interestingly, the results address two commonly observed dynamic properties of nanoconﬁned chromophores that have opposite eﬀects. On one hand, blocking the twisting of a molecule encapsulated in SBMs can lead to a longer excited state lifetime. On the other hand, the interaction of guests with the silica surface may induce the formation of their aggregates, which accelerates the deactivation of the excited state due to the EnT between the bright and the dark states in the aggregates. The orientation and photophysical properties of methylacridine dye in zeolite L were calculated (MeAcr+, Scheme 2).94 The most stable orientation of this guest results in both its long and short molecular axes being nearly perpendicular to the channel axis and was found mainly determined by dye-zeolite L electrostatic interactions in addition to that with the cosolvent, water. The optimized conformations were used to explain the observed EnT processes (self-absorption). FRET in dye-ﬁlled monolithic crystals of zeolite L was recently modeled using Monte Carlo and molecular dynamic simulations.93 Under ideal conditions (homogeneous dye distribution), very high exciton diﬀusion rates were expected in zeolite L. However, the inhomogeneity of the dye distribution along with the 1D channels led to a large change in the exciton dynamics that reduces the exciton lifetimes and modiﬁes the dimensionality of the energy transport (Figure 6B). FRET was also studied in LTL zeolites and interpreted with the help of theoretical calculations.95 Cyanine (1,1′-diethyl-2,2′cyanine iodide, PIC) and acriﬂavine hydrochloride (AF) dyes acted as energy donor and acceptor partners, respectively (Scheme 2). Energy was transferred from excited PIC to AF that were adsorbed into the internal surfaces of zeolite LTL. A redshift of AF emission band in the protonated LTL zeolite was explained in terms of a conversion from a cationic to a

Scheme 2) molecule in CuMCM41 and aﬀects its charge distribution and LUMO energy.86 As a result, CAN exhibits a lower photoinduced ET eﬃciency than β-carotene, for which no H-bond between the dye and the host is formed. This eﬀect was further studied in a subsequent work by the same group, predicting two types of H-bonds between the OH group of retinol and silanol ones on the surface of MCM41.87 One of these H-bonds decreases the LUMO energy of the interacting carotenoid and stabilizes its neutral species more than its radical cation, thus disfavoring photoinduced ET from the carotenoid to MCM41. The opposite behavior was found for the second type of H-bonds. 2.2.3. Impact of Guest Aggregation. Increasing the concentration of a guest in the silica framework usually leads to aggregates formation. Formation of these entities is closely related to the occurrence of EnT between nearby adsorbed guest molecules. EnT from photoexcited guest (donors) to acceptor molecules behaves according to the model of Förster resonance EnT (FRET).88 In particular, the ﬂuorescence kinetics for such a process should be described by a stretched-exponential function (sometimes also called Kohlrausch decay function).88,89 The equation has the following formula: f (t) = A exp [(−t/τs)α], where τs is a characteristic time of the decay and α is the heterogeneity parameter (α = 1 for monoexponential decay, and a smaller value indicates more dispersive kinetics and large sample heterogeneity).88,90−92 For the 7-aminocoumarin interacting with SBMs discussed previously, the eﬀect of the dye concentration in a matrix on the ﬂuorescence quenching was also theoretically explored.84 The calculations indicated that the strong quenching is attributable to the formation of excimers with a CT character upon excitation of the monomeric species. The LUMO of the calculated aggregate (dimer) has a strong bonding character (Figure 6A). The excitation of an electron from the HOMO of 13649

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Scheme 3. Molecular Structures of the Compounds Interacting with SBMs and Discussed in the Proton-Transfer Section 3.1

3. ENSEMBLE AVERAGE TIME-RESOLVED STUDIES OF PHOTOINDUCED PROCESSES IN SILICA-BASED MATERIALS

protonated AF form. The experimental reports on EnT in SBMs, and its application are highlighted, respectively, in section 3.3, and section 5. In this part, we have commented on how theoretical studies

3.1. Proton-Transfer Reactions

Proton-transfer (PT) reactions are intensively investigated since they play important roles in a wide variety of chemical and biological processes, such as acid−base reactions, biochemistry, and catalysis (see Scheme 3 for molecules involved in PT reactions discussed here).32,37−39,96−107 Moreover, the study of these reactions has resulted in the development of many applications, such as ﬂuorescent chemosensors,108,109 laser dyes,110−114 ultraviolet (UV) photostabilizers,115 photoswitches,116−119 and organic optoelectronic materials.120−122 These reactions occur between proton (or hydrogen atom) donor (typically −OH, −NH2, and −SH) and acceptor (CO, −N, base and some solvents) groups. The reaction can be inter- or intramolecular and can be activated thermally or by light; the process can occur through or without a

have become gradually important tools in explaining and predicting the short-time dynamics in SBMs. A large part of theoretical research is devoted to modeling solvation dynamics in silica-based hosts and the resulting eﬀects of diﬀerent local environment at diﬀerent positions in the silica pores. Other studies predicted how the interaction of the chromophore with the host changes its photophysical properties and how it aﬀects the stationary absorption/emission spectra and the photodynamics upon light excitation. 13650

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

barrier.32,97,101,105,107 Aromatic molecules that contain an acidic proton in their structures can undergo intermolecular PT reactions in both the ground (ground-state proton transfer, GSPT) and the electronically excited (excited-state proton transfer, ESPT) states.37,98,101,123,124 Upon electronic excitation, these molecules become strong acids (photoacids) that can undergo an ESPT reaction with a neighboring base molecule, such as water, alcohols, or amines.39,102,123−127 Since Weller’s and Eigen’s works,96,128−130 the mechanisms of PT reactions have been intensively investigated from experimental and theoretical points of view.100,101,107,124,127,131−133 The Smoluchowski-Collins-Kimball (SCK) model has been used to describe this type of reactions.134−137 This model shows how the diﬀusion of the acid and the base through the solvent plays an important role in the mechanism, as the reactants should approach each other at a speciﬁc distance for the PT reaction to occur. The surrounding environment may play a key factor in solute/solvent interactions and subsequent reorganization. Therefore, the global dynamics of these reactions are determined by diﬀusion, interactions with the solvent, and the adiabatic and nonadiabatic nature of the potential energy surfaces (PES), as well as by the type of the prototropic groups.123,124,138−141 In bifunctional organic molecules that contain both the proton donor and acceptor groups in close proximity, an intramolecular PT reaction in the ground (GSIPT) or electronically excited (ESIPT) state can occur.32,99,117,142 Because of the chemical structure and the positions of the partners in these bifunctional systems, an intramolecular Hbond (IHB) is formed in the ground state (S0), usually producing a species called an enol (E) form. Consequently, upon electronic excitation of E, an intramolecular electronic charge redistribution occurs in E* that leads to the formation of the keto (K*) or zwitterion (Z*) tautomer. When these species return to S0, they can be stabilized in this or other tautomeric forms (e.g., rotated K), or return to the E species.143−148 The mechanism and rate constant of the ESIPT reactions depend on the PESs of the E* and K* tautomers in the ﬁrst electronically ﬁrst excited single state (S1).101,149 The presence or absence of barriers for both processes in S0 and S1 opens the possibility of controlling the related spectroscopy and dynamics.44,150−155 Thus, for a PES without an energy barrier between the wells of E* and K* species, the ESIPT reaction is ultrafast (fs regime) and irreversible.101,156 When the PES has a barrier in the conversion of E* to K*, the reaction may be reversible or irreversible but will be slower (ps−ns regime) than in the barrierless situation.34,157−164 An accurate description of the reaction dynamics at the PESs of ESIPT processes in gas phase requires at least two reaction coordinates: (i) the proton motion within the corresponding IHB and (ii) the vibrational or torsional (angle) mode involving the distance between the partners; the angle between the involved moieties in the transfer could be important in the global reaction dynamics and subsequent processes.97,99,101,107 In the presence of an energy barrier in the PESs, the proton motion may occur via tunneling, which adds another (quantum) dimension to the reaction. In solution, the PESs might be complex, involving a solvent coordinate where its polarity and H-bonding could play an important role in shaping the spectroscopy and dynamics. The earliest studies of proton-transfer dyes mainly used ﬂuorescence spectroscopy because the photoproduced K* (or Z*) generally emits with a large Stokes shift (6000−10000 cm−1). Advances in laser spectroscopy gave details on PES’s reactions at the

f em t o s e c o n d t i m e s c a le i n g a s a n d c o n d e n se d phases.29,32,40,100,106,149,157,165−170 Coupled proton and electron motions in the molecular system, of great importance in many chemical and biological systems, have been experimentally and theoretically examined.34,171−173 The experimental and theoretical ﬁndings not only generate new knowledge but also open new research ﬁelds with possible applications in sensing, catalysis, drug delivery, photovoltaics, and photonics (section 5). Both inter- and intramolecular PT dynamics are sensitive to a variety of environmental properties, such as polarity, hydrophobicity, viscosity, acidity, basicity, and H-bonding ability.117,120 This sensitivity to the surrounding characteristics makes PT systems potential candidates as environmental sensors, an important aspect that has been explored by several groups.32,38,117,120 Researchers took advantage of this virtue by using molecules that transfer protons to characterize the nanoand microenvironments in cavities and pores of SBMs. In these systems, the speciﬁc and nonspeciﬁc molecular interactions between the trapped guest (the proton-transfer dye) and the host, as well as with the caged solvent molecules, can be modiﬁed to tune the spectroscopy and dynamics of the formed composites (dye/SBMs). 3.1.1. Intermolecular Proton-Transfer Processes. In the presence of bases (solvent or another solute), many molecules that have acidic protons undergo intermolecular PT processes in both S0 and S1 states (GSPT and ESPT, respectively).98,123,124,174−178 Central to this review, most of the SBMs have OH groups, and thus are interesting material models to study the PT reaction dynamics of guest molecules within their pores or channels. 3.1.1.1. Molecules Exhibiting Deprotonation with the Framework in S0 and S1 States. Upon encapsulation of chromophores in SBMs, they can establish an intermolecular H-bond with the surrounding framework; protonation or deprotonation in the ground and/or excited state are possible events. 2-Naphthol (2-NpOH, Scheme 3) is a classic example that demonstrates a PT reaction with a H-bonded acceptor group (water, alcohols, amines, etc.). 2-NpOH emission in water solutions is dominated by its anion’s (2-naphtholate) emission as a result of a large change in the acidity of 2-NpOH in S1 (pKa = 9.5 and pKa* = 2.8).124,179−182 Its spectroscopy and related dynamics are pH-dependent.183 In acidic and neutral media (pH = 1−7), which have relatively high [H+], the ESPT reaction is described by a reversible mechanism (Figure 7). In this accepted scheme, kPT and krec are the deprotonation and recombination rate constants, while kD and k−D are the forward and reverse diﬀusion-controlled rate constants that can be obtained from the Debye-Smoluchowski equation.124,183,184 In contrast, in alkaline conditions (pH > 7) and in the presence of sodium acetate, the ESPT reaction moves forward and is

Figure 7. General scheme of a reversible excited-state intermolecular proton transfer (ESPT) process in excited HPTS. Reprinted with permission from ref 186. Copyright 2016 Royal Society of Chemistry. 13651

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Figure 8. (A) Normalized (to the maximum intensity) ﬂuorescence spectra of HPTS (solid line), HPTS/MCM41 (dashed line), and HPTS/ MCM41 upon addition of 20 μL of water (dotted line) in dichloromethane suspensions. Schematic illustration (not in scale) of (B) HPTS/MCM41 in a dichloromethane suspension, and (C) HPTS trapped within a water drop adsorbed on the MCM41 surface. The insets of (B) and (C) show the wavelengths corresponding to the emission intensity maxima and the ﬂuorescence lifetimes (τ, τRO) of the formed species and of the values of proton transfer (kPT) and recombination (krec) rate constants. Adapted with permission from ref 186. Copyright 2016 Royal Society of Chemistry.

Figure 9. Emission decays of HPTS/MCM41 in dichloromethane suspension containing (A) water and (B) deuterium oxide. Reprinted with permission from ref 186. Copyright 2016 Royal Society of Chemistry.

irreversible.183 In neutral conditions, the proton motion occurs with a kPT = 7.6 × 107 s−1 and krec = 4.7 × 1010 M−1 s−1, generating an anion that has a ﬂuorescence lifetime of 9 ns.183,184 When 2-NpOH is caged in β-cyclodextrins (β-CDs), the ESPT reaction exhibits a reversible mechanism with kPT = 2.2 × 107 s−1 and krec = 2.3 × 1010 M−1 s−1, which are slower than those recorded in bulk water.182,185 The observed reduction of kPT and krec values is due to the encapsulating βCDs cage having a less polar and less protic character than the bulk water microenvironment does, reﬂecting the important role that the host plays in the PT dynamics.44,46,182 In contrast, the ﬂuorescence lifetime of the caged anion (τAN = 10.5 ns) is longer than that in aqueous solutions due to the protection provided by the CD cavity. 44,182 When 2-NpOH is encapsulated in solid NaX zeolites, a faster ESPT reaction (kPT = 3.3 × 109 s−1) and a slower recombination process (krec = 8.3 × 108 M−1 s−1) were reported.179 The larger value of kPT observed for these hybrid complexes, in comparison with that in water solutions (kPT = 7.6 × 107 s−1), is explained in terms of a stronger basic character of the NaX host. The formation of strong H-bonds between the silicate framework (Si−O−Si) and trapped 2-NpOH molecules leads to a fast deprotonation of the latter.179 When the trapped anion interacts with the NaX material, it lives 4.7 ns, which is signiﬁcantly shorter than in water solutions (9 ns), reﬂecting an enhanced intersystem crossing in the zeolite nanocavities.179 Another well-known photoacid molecule, 8-hydroxypyrene1,3,6-trisulfonate (pyranine, HPTS, Scheme 3), was investigated in the presence of mesoporous MCM41 materials.186

When HPTS in water solutions is excited, its acidity signiﬁcantly changes (pKa = 7.7 to pKa* = 0.6).187 In neutral water solutions, HPTS undergoes a reversible ESPT reaction that is characterized by kPT and krec of 1 × 1010 s−1 and 6 × 109 Å s−1, respectively.186,188 To understand the dynamics of the interaction of HPTS with SBMs, we brieﬂy discuss its behavior in other hosts such as CDs and the human serum albumin (HSA) protein. The ESPT reaction of HPTS within these hosts follows an irreversible mechanism with a rate constant (kPT) that is lower than that in bulk water.189−191 The change is attributed to the slow solvation dynamics of the encapsulated water inside or at the pore gates of these hosts (biological water).189−191 Within these hosts, HPTS exhibits two diﬀerent PT pathways that are characterized by times of 150 fs and 1.2 ns in HSA, and 140 ps and 1.4 ns in γ-CD, reﬂecting the heterogeneity of the formed complexes.190,191 By heterogeneity, we mean that the trapped dye molecules are found in locally diﬀerent environments and positions within the chemical (CD) and biological (HSA protein) pockets. For the HPTS:HSA complexes, the shortest time is due to a strong interaction of the dye with the host in which a direct reaction with the carboxylate groups of the protein’s amino acids is suggested. The slower process is assigned to a PT reaction with the biological encapsulated water molecules, which have slower dynamics.190,191 These results once again show the importance of the host’s nature on the photobehavior of trapped H-bonded aromatic molecules. However, HPTS does not undergo PT when it interacts with the surface of MCM41 materials in dichloromethane 13652

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

suspensions, and thus, the emissions of the protonated monomer and adsorbed aggregates were recorded (Figure 8).186 The emission spectrum is centered at 430 nm, and the ﬂuorescence lifetimes of these species are 350 ps and 2.5 ns, respectively. Adding water (1%) to these dichloromethane suspensions leads to a reversible ESPT reaction and a strong emission of the anion.186 In this case, the PT dynamics are slower than those observed in bulk water, with PT and recombination rate constants kPT = 5.4 × 109 s−1 and krec = 2.2 × 109 Å s−1, respectively.186 The decrease in the values of the rate constants is explained in terms of a decrease in the HPTS photoacidity due to a partial loss of the negative electronic charge of its SO3− groups that are H-bonded to the framework.186 Furthermore, the immobility of the HPTS anions on the silica frameworks leads to a slower recombination process.186 When normal water is replaced by deuterated one, the PT and recombination processes of these complexes are even slower, with kPT = 2.7 × 109 s−1 and krec = 1.7 × 109 Å s−1 (Figure 9). The study reported that the kinetic isotope eﬀects (KIEs) are 2 for (kHPT/kDPT) and 1.3 for (kHrec/kDrec), being larger for the PT process than the normal one in bulk water (1.4).186 Thus, the H-bonding interactions between HPTS and the MCM41 framework and the restricted diﬀusion space available to the adsorbed HPTS result in slower dynamics.186 However, a diﬀerent behavior was observed when HPTS interacts with other SBMs such as alumino-silica surfaces and hydrated porous silicon (10 nm pore size).188,192 For both complexes, the ESPT mechanism is similar to that observed for HPTS in bulk water. This similarity is because the HPTS molecules are covalently linked to the alumino-silica surfaces and retain their photoacid character, while in the hydrated porous silicon, the large pore size of the host does not aﬀect the ESPT dynamics.188,192 Clearly, the diﬀerent behaviors of HPTS when interacting with these materials reﬂect how the molecular environment and the restriction of the host aﬀect the ESPT reaction of the PT dye and its related dynamics. The hydroxyquinoline family, having amphoteric character, has been intensively studied due to the possibility of assisted ESPT reactions with protic media (water, alcohols, acetic acid, or ammonia, among others).138,139,193−198 As a result, hydroxyquinolines were selected as good candidates to study the catalytic properties of SBMs as single and double PT reactions can occur in the pores of these materials.143,144,199−201 In a neutral aqueous solution, 6-hydroxyquinoline (6-HQ, Scheme 3) exhibits an ESPT reaction in a stepwise fashion: an initial proton release from the OH group of 6-HQ that produces the anion (AN) in ∼20 ps (E*→ AN*) followed by a PT to the nitrogen atom of AN* that forms the ﬁnal tautomer (AN*→ T*) in ∼45 ps (Figure 10).202,203 When caged in β-CD, 6-HQ shows a similar mechanism.202 The forward reactions (k1 and k2), with time constants of ∼380 ps and ∼360 ps, are 19 and 8.4 times slower than those in water, respectively.202 However, the reverse PT events (k−1 and k−2) are slightly accelerated by CD’s encapsulation.202 When trapped in NaX and NaY zeolites, 6-HQ exhibits a similar reversible ESPT process.199 The UV absorption spectra show three bands centered at 325, 365, and 430 nm, indicating three possible prototropic species (E, AN, and Z, respectively) in the S0 state. In contrast, the emission spectra only exhibit two bands, which have intensity maxima at 400 and 510 nm, assigned to the emission of AN* and Z*, respectively. No emission signal was detected from E*, suggesting that almost all the molecules in this form readily deprotonate. The relative

Figure 10. Proposed diagram of proton transfer and relaxation of the prototropic species of excited-state 6-HQ in microporous catalytic faujasite zeolites. Adapted with permission from ref 199. Copyright 2010 Wiley-VCH.

intensities of AN* and Z* bands depend on the properties of the zeolite’s cage: AN* emission is predominant in NaX, while that of Z* in NaY.199 The authors explained this diﬀerence in terms of the zeolite basicity. Since NaX is more basic than NaY, the AN* population is larger in its cages. The ESPT times and the values of the ﬂuorescence lifetimes of the diﬀerent formed species of 6-HQ in these zeolites (kN−1, kAN−1, and kT−1 for the neutral, anion, and zwitterion forms, respectively) are shown in Table 1.199 Clearly, the zeolite conﬁnement decelerates the forward reactions (from 20 to 70 ps for k1−1 and 43 to 400 ps for k2−1) and accelerates the backward ones (from 380 to 140 ps for k−1−1 and 1 ns to 750 ps for k−2−1),199 when compared with the rate constants in water solutions.202,203 These photophysical changes reﬂect that the zeolite conﬁnes the contact ion pairs to be closer than they are in bulk water, thus enhancing the forward reactions.199 Note that within CDs, the forward reactions (380 and 360 ps) are even slower than in the zeolite complexes.202 Conﬁned water in the CD cavities leads to slower molecular dynamics.44 However, within zeolite cavities, where there is no conﬁned solvent (or its population is very weak), the dynamics are only aﬀected by the speciﬁc and nonspeciﬁc interactions with the acidic and basic sites of the framework. These examples nicely show the relevance of the medium (solution, suspension, or solid) in exploring the photodynamics of these conﬁned systems. On the other hand, the kinetic constants of the PT and recombination processes of 6-HQ within NaX and NaY zeolites are also aﬀected by the acid/base properties of these hosts (Table 1).199 The enol deprotonation process (k1−1) is slightly faster in NaX, a zeolite with a high basicity, than in NaY (67 vs 71 ps, respectively), while the reprotonation (k−1−1) is slower (180 vs 120 ps, respectively). Similarly, the anion protonation (k2−1) is slower in NaX than in NaY, and its redeprotonation (k−2−1) is faster.199 In NaX complexes, the predominant species is the anion as it is reﬂected by its large contribution (at ∼400 nm) to the timeresolved ﬂuorescence spectra (Figure 11).199 Thus, the host properties play an important role in determining the pathway and the rate constants of the ESPT reactions: the OH-group deprotonation processes are the most favored in a basic host (NaX), while the imine protonation processes are favored in more acidic ones (NaY).199 These results are in agreement with the fact that the protonation of the 6-HQ anion is largely favored in the presence of acetic acid.195 13653

DOI: 10.1021/acs.chemrev.7b00422 Chem. Rev. 2017, 117, 13639−13720

Chemical Reviews

Review

Table 1. Kinetic Constants of Excited-State Tautomerization and Relaxation of 6-HQ Encapsulated in Zeolitesa

a

zeolite

k1−1 (ps)

k−1−1 (ps)

k2−1 (ps)

k−2−1 (ps)

kN−1 (ps)

kAN−1 (ps)

kT−1 (ps)

NaX NaY

67 71

180 120

420 380

710 770

1100 1450

9600 12200

15600 18000

Reprinted with permission from ref 199. Copyright 2010 Wiley-VCH.

interact with the labile protons of the E form of 7-HQ, stabilizing the trapped AN.143 The emission spectra of these composites are rather complex, displaying signals from 325 to 650 nm. On the basis of the excitation-dependent emission spectra and the lifetime distributions, the authors assigned the emission to E*, AN*, and Z* that are all caged and H-bonded to the framework. Their ﬂuorescence lifetimes are 0.26, 1.5, and 5.5 ns, respectively.144 Femtosecond emission studies of caged E* revealed the deprotonation and protonation dynamics within the MCM41 and AlMCM41 composites (Figure 12).143,144 The obtained time constants are 0.3 and 3 ps,

Figure 11. Time-resolved ﬂuorescence spectra of 6-HQ trapped in (A) NaX and (B) NaY after a time delay of 0 (purple), 70 (blue), 200 (green), 500 (yellow), and 2000 ps (red). The inset in (A) is a cartoon (not to scale) of 6-HQ encapsulated in a faujasite zeolite nanocavity. Reprinted with permission from ref 199. Copyright 2010 Wiley-VCH.

7-Hydroxyquinoline (7-HQ, Scheme 3), a well-studied quinolone derivative, also exhibits single and stepwise double PT reactions.143,144,200 In water solutions, the OH-deprotonation and imine protonation processes occur in both S0 and S1 states and are reversible in the former.204 Upon electronic excitation of 7-HQ (to S1), the time constants of these acid and base processes are 40 and 160 ps, respectively.204 Several groups have reported that the photobehavior of 7-HQ prototropic forms is sensitive to the conﬁning environment.125,139,185,187,204,205 For example, when 7-HQ is caged in β-CD, its OH-deprotonation is slower (170 ps), while its imine protonation is faster (85 ps) than those observed in water solutions.39,204 Moreover, in these cavities, the enol population is the predominant form, while in pure water, the zwitterions dominate. This diﬀerence reﬂects the hydrophobic nature of the CD interior stabilizing the enol form and the slowing down of the double PT processes that occur at the two gates of CD cage.185 Similar behavior was observed in nonaqueous protic solvents, such as alcohols: an H-bond bridge is formed with the enol to function together as a proton relay.139 Solvent reorganization to reach an appropriate 7-HQ/solvent conﬁguration is an important factor to consider in PT reactions.139,184,206,207 When 7-HQ is within regular MCM41 materials in dichloromethane suspensions, its OH-deprotonation and imine protonation are irreversible.144,179 Contrary to the dynamics in bulk water, both the AN and Z/K forms are stabilized in S0 when the dye is caged in MCM41 pores, showing how the hydrophilicity, structure, and conﬁnement of this host aﬀect the formation and stability of these species.143,144 Moreover, the presence of Al atoms in MCM41 (AlMCM41) highly stabilizes the AN form in S0.143 The incorporation (doping) of Al atoms into the silica framework introduces negatively charged sites that are able to

Figure 12. Femtosecond (fs)-emission transients of 7-HQ interacting with MCM41 in a dichloromethane suspension. Reprinted from ref 144. Copyright 2010 American Chemical Society.

respectively, in MCM41, while in AlMCM41, both processes are slower (0.5 and 6 ps, respectively), showing the importance of speciﬁc and nonspeciﬁc interactions with the nanohosts in this guest’s dynamics and spectroscopy.143 Both processes are signiﬁcantly faster than those in water (40 and 160 ps) and CD (170 and 85 ps) solutions, which emphasizes direct PT events in robust composites, where conﬁnement, speciﬁc, and nonspeciﬁc interactions play a major role in its photobehavior. Because of these interactions with the host, even molecules that exhibit ultrafast ESIPT reactions in solutions can undergo ESPT processes when they are encapsulated in micro- or mesoporous materials, making very rich their spectroscopy. This possibility is illustrated by the cases of 2-(2′hydroxyphenyl)benzothiazole (HBT), 2-(2′-hydroxyphenyl)benzoxazole (HBO), and other derivatives when they interact with zeolites and mesoporous materials (Scheme 3).208−214 In solutions and upon electronic excitation, this kind of molecules exhibits ultrafast ESIPT reactions (

Photochemistry and Photophysics in Silica-Based Materials: Ultrafast

Recommend Documents