Low Hydroxylated Fullerenes: Stability, Thermal Behavior, and

Subscriber access provided by University of Massachusetts Amherst Libraries

C: Physical Processes in Nanomaterials and Nanostructures

Low Hydroxylated Fullerenes: Stability, Thermal Behavior, and Vibrational Properties Marco Vinicio Velarde- Salcedo, Marco Gallo, and Ricardo A. Guirado-Lopez J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b01628 • Publication Date (Web): 30 May 2018 Downloaded from http://pubs.acs.org on May 30, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Low Hydroxylated Fullerenes: Stability, Thermal Behavior, and Vibrational Properties M. V. Velarde-Salcedo1, M. Gallo2, and R. A. Guirado-López1,*

1

Instituto de Física “Manuel Sandoval Vallarta”, Universidad Autónoma de San Luis Potosí, Álvaro Obregón 64, San Luis Potosí, S.L.P., 78000, México

2

Tecnológico Nacional de México/ITCJ, Av. Tecnológico 1340, Cd. Juárez, Chihuahua, 32500, México.

*Corresponding author: [email protected]

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 51

Abstract

Extensive density functional theory (DFT) calculations dedicated to analyze the stability, thermal behavior, as well as the infrared (IR) and Raman spectra of low hydroxylated C60(OH)12 fullerenols are presented. Adsorbed configurations in which OH groups form various types of molecular islands on the carbon surface are the preferred atomic arrays, while random distributions of hydroxyl species are the highest energy molecular structures. It is found that the formation of local networks of hydrogen-bonded OH groups plays a fundamental role in the stability of these complexes. The calculated dipole moments, polarizability values, optical gap, and Fukui functions of C60(OH)12 isomers strongly depend on the structure of the hydroxyl overlayer, being thus an important parameter to tune the material properties. DFT Born-Oppenheimer molecular dynamics calculations at T=300 K reveal that aggregated forms of OH groups on the fullerene surface show an interesting dynamical behavior, characterized by a continuous protonexchange process between neighboring hydroxyl molecules that modifies the structure and chemical nature of the molecular coating. From nudged elastic band studies analyzing OH diffusion on the C60 surface energy barriers opposing OH migration of ~1 eV are found. However, in the presence of surrounding H2O species, a water-assisted diffusion process is obtained which can reduce the energy barriers to values as low as 0.25 eV. The comparison between experimental and calculated IR spectra of various C60(OH)12 isomers shows welldefined spectral features which can be very helpful to identify the structure of these fullerene complexes. Finally, simulations of the wavelength dependent Raman spectra of hydroxylated fullerenes reveals that i) the intensities of the Raman active modes strongly depend on the excitation laser and ii) the importance of the wavelength dependent calculations to reveal precise features of different regions of the spectra. The combination of IR and Raman spectroscopies is


2

Page 3 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


an efficient approach to reveal the atomic structure of these sub nanometer-sized carbon nanostructures.

1. Introduction In the last years, it has been clearly established that the functionalization of both inner and outer surfaces of carbon nanostructures dramatically changes their chemical, electronic, and transport properties.1-3 In particular the attachment of hydrophilic substituents, such as OH groups, are of fundamental importance and provide an interesting route for the development of novel materials. For example, C60(OH)24 fullerenol structures can work as antioxidative agents,4 free radical scavengers5 and, recently, the potential application of fullerenols in cancer drug delivery has been reported.6 The functionalization of carbon nanotubes with polar groups can be used to improve dispersion as well as the chemical interaction with a polymer matrix.7 OH adsorption on single walled carbon nanotubes modifies their electronic properties which makes them good candidates for the design of novel electronic devices.8 Finally, hydroxylfunctionalized graphene has been recently synthesized.9 It shows good hydrophilicity and biocompatibility to human retinal pigment epithelium cells.10 It also provides and effective microenvironment for cell adhesion and proliferation.11 In addition, hydroxylated graphene sheets can self-assemble to form flexible films with high electrical conductivity which can be used for high power electronic applications.12 It is important to emphasize that, after hydroxyl adsorption, it is possible to link more complex reagents such as aminoacids, DNA, and various types of drugs, leading also to strong changes in the solubility of the different nano-carbons.


3


Page 4 of 51

In particular, hydroxylated fullerenes or fullerenols can be synthesized by various (and relatively simply) methods reported in the literature involving thermal treatments in the presence of HNO3/H2SO4 mixtures,13 different hydrolysis procedures,14 as well as synthetic approaches using alkaline media and halogens as intermediates.15-16 However, all these techniques generally lead to the formation of OH-covered C60 complexes with different structures and can be characterized by poor reproducibility. Actually, the atomic structure of fullerenols is still under investigation since their properties are expected to strongly depend on the number and disposition of OH groups on the carbon surface. In this respect, the use of mass spectrometry experiments allows to estimate the number of OH groups attached to C6017 while infrared spectrometry measurements plays an important role in the search for OH group location.18 Notice that the different experimental procedures employed for fullerene hydroxylation also produce carbon complexes with varied chemical composition, which introduces an additional source of complexity to understand the functionality of these materials. For example, C60(OH)nOx structures seems to be characterized by ring-opened carbon cages.19 Furthermore, the reaction of the C60 fullerene with tertrabutylammonium hydroxide (TBAH) in the presence of oxygen and water solution of NaOH leads to stable anion-radical Nan+[C60Ox(OH)y]n- (where n=2,3, x=7— 9, and y=12—15) complexes from which Na ions are very difficult to remove.20 Fullerenols having less than 12 OH groups show poor water solubility which reduces its range of applications. However, Kokubo et al.21 have demonstrated that these low-hydroxylated C60 cages are of fundamental importance to synthesize high-purity polyhydroxylated fullerenes. Using hydrogen peroxide (instead of sodium hydroxide) and a 12-hydroxylated fullerenol as a starting material, they synthesized highly water-soluble fullerenols with 36—40 OH groups. Infrared, UV-vis, thermogravimetric, and dynamic light scattering (DLS) experimental


4

Page 5 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


characterization reveal that these fullerene complexes contain well-defined C—O, C—O—H, C=C, and O—H vibrations, 8—9 secondary bound waters, a contrasting optical response when compared to bare C60, and a high dispersion property which gives a narrow particle size distribution within 0.7—2.0 nm. Even if it is possible to estimate the number of OH groups attached to C60 it has been always very difficult to know their surface location on these subnanometer sized carbon cages. The precise knowledge of the distribution of the 12 OH groups on the fullerene surface will be of fundamental importance to understand the observed macroscopic behavior, the initial stages of the polyhydroxylation process, and the role played by different local atomic environments in –OH addition. Clearly, without experimental guidance as to the isomer that is actually produced, it will be highly time consuming to theoretically explore the many configurations which can be adopted by the hydroxyl groups on the C60 surface. In this work we will take advantage of the measured infrared (IR) spectra reported in Ref. 21 for C60(OH)12 complexes (which reveals well-defined features) to present a systematic comparison between experimentally determined and theoretically calculated distribution of infrared active vibrational modes for various model C60(OH)12 fullerenes. It will be shown that a high-energy C60(OH)12 isomer, defined by a complete aggregated form of OH molecules on one side of the fullerene cage, is the one that has an IR spectra in good agreement with experiments. In addition, this structure has a large dipole moment, the highest number of hydrogen bonds between neighboring hydroxyl groups, four coordinated waters forming the first hydration shell, and a Raman spectrum with well-defined features, which are all facts that can be used to corroborate its abundance in real samples. The use of high-energy isomers to explain experimental data in closely related systems has been recently proposed by Marom et al.22 to understand the optical properties of small TixOy clusters,


5


Page 6 of 51

as well as by Harding and co-workers23 to explain gas-phase reactivity experiments on small rhodium particles. Here, the crucial role played by systematic theoretical studies in which the properties of both low and high-energy isomers are extensively analyzed will be underlined. The rest of the paper is organized as follows. In Sec. 2 we briefly describe the theoretical models used for the calculations. In Sec. 3 we present the discussion of our results and finally, in Sec. 4 the summary and conclusions are given.

2. Method of Calculation The structural, electronic, and vibrational properties of low hydroxylated C60(OH)12 fullerenes are obtained within the DFT approach with the help of the GAUSSIAN09 software.24 The KohnSham equations are solved employing the Perdew-Burke-Ernzerhof (PBE)25 expression for the exchange-correlation potential, together with the 6-31g(d,p) basis set (including double polarization functions) proposed by Hariharan and Pople,26 which is a good compromise between computational costs and accuracy. In all cases, a fully unconstrained structural optimization is performed, covering a large number of possible adsorbed configurations for the OH groups on the C60 surface. We expect the existence of a non-uniform electron density distribution within the structures that is obtained from a Bader charge analysis.27 The knowledge of the charge density distribution is of fundamental importance to understand the electronic properties and reactivity of fullerenol structures. The vibrational frequencies are determined by diagonalization of the full Hessian matrix within the harmonic approximation.28 This procedure generates all vibrational modes that can be directly compared with experimental infrared (IR) and Raman spectroscopy measurements. The resulting stick spectra are broadened with Gaussian line shape functions. The


6

Page 7 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Raman active modes are calculated for the two most commonly used excitation lasers, namely: 532 and 785 nm, in order to determine the one that more clearly reveals the fine details of the Raman spectra of the systems. The thermal behavior of our low-hydroxylated fullerenes as free-standing and hydrated species will be studied employing two theoretical approaches. For the first case, DFT Born-Oppenheimer molecular dynamics (BOMD) simulations29 for some representative fullerenol isomers are performed. For these more computationally tractable systems, we use an average time step of 0.33 fs and the simulations are run for 550 steps (or 180 fs) at a temperature of 300 K involving the PBE functional together with the 6-21 basis set.30-31 In the second case, the hydration properties of our fullerenols at ambient temperature are analyzed by employing classical NPT molecular dynamics simulations with the Maestro software.32-34 The simulations were carried at 300 K and a pressure of 1 bar using the Optimized Potentials for Liquid Simulations (OPLS) 2005 force field35 for the C60(OH)12 molecule placed in a solvated box containing simple point charge (SPC)36 water molecules. A time step of 2 fs was considered in the 2 ns molecular dynamics simulations using the Martyna-Tobias-Klein barostat37 with a 2.0 ps relaxation time and the Nose Hoover Chain Thermostat38 with a 1.0 ps relaxation time. For electrostatics we used the smooth particle mesh Ewald algorithm39 with a tolerance of 10-9 and a cutoff radius of 9 Å for short range interactions. During the simulation, the fullerenols as well as the surrounding waters are allowed to perform different rigid displacements and rotations. Finally, to analyze the diffusion properties of OH groups on the fullerene surface the nudged elastic band (NEB) methodology,40-42 as implemented in the Quantum-Espresso code,43 is used. The NEB calculation scheme is a chain-of-states method where a set of images between the initial and final states must be created to achieve a smooth curve. Here, the relatively small size


7


Page 8 of 51

of the model hydroxylated fullerenes ensures that the calculations for the diffusion paths remain computationally tractable. In the calculations, seven to thirteen images are employed to determine the energy profiles, which are enough to reveal the different stages of the possible structural changes that might occur. The direct (reverse) energy barriers between well-defined initial and final atomic configurations are obtained by calculating the energy difference between the initial (final) position and the saddle point of each one of the energy profiles.

3. Results and Discussion 3.1 Energetics, electronic properties, and thermal behavior In Figure 1 the optimized atomic configurations of the here-considered twelve C60(OH)12 isomers are presented. As is well known, determining the ground state configuration for the C60(OH)12 fullerenol is a very difficult problem due to the huge size of the configurational phase space. The problem is also very complex when increasing the OH-coverage on the fullerene surface. A common practice in the literature is to choose contrasting adsorbed configurations of chemisorbed OH groups and analyze the variations in the fullerenol properties for the different overlayers. Here, the model of fullerenol has been created by adding 12 OH groups to the C60 surface, guided by low energy atomic structures of various C60(OH)n fullerenes already reported in the literature.44-46 In agreement with these previous studies, aggregated forms of OH molecules on the fullerene surface define the lowest energy atomic arrays. The preferred isomer (isomer 1) is shown in Figure 1(a) where the twelve OH groups form two molecular islands in opposite sides of C60. The closest in energy structure (with an energy difference of 0.39 eV) is presented in Figure 1(b) (isomer 2) and is characterized by an adsorbed phase in which the


8

Page 9 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


hydroxyl molecules are arranged in a “linear-like” array stabilized, as in Figure 1(a), by strong hydrogen bonds between neighboring OH groups. Notice that these two types of molecular overlayers leave extended unprotected regions on the fullerene surface. Actually, the adsorbed configuration in which the OH’s are randomly distributed on the carbon cage [see Figure 1(l)] is the less stable C60(OH)12 fullerenol (isomer 12), with an energy difference of 5.9 eV when compared to Figure 1(a), where the large distance between hydroxyl species excludes the formation of H—bonding. As we will see in the following, the contrasting adsorbed configurations shown in Figure 1 will strongly affect the behavior of water molecules near these surfaces. The wetting properties, which can be determined by counting the number of first coordinated H2O species, will be also of fundamental importance to take into account in order to more precisely defined the C60(OH)12 fullerenol present in the experiments. It is important to emphasize that, in all cases, the OH species are adsorbed in an on-top configuration with C—O bond lengths that vary in the range of 1.41—1.46 Å and C—O—H angles of approximately 101—106 degrees. A notable C→OH charge transfer is found which is expected to strongly alter the electronic properties of the fullerenols. Our Bader analysis reveal that the total C60→(OH)12 charge transfer depends on the structure of the hydroxyl overlayer varying from 5.25—5.64e. Finally, the different adsorbed phases induce appreciable global deformations on the C60 cage yet maintaining the original closed-cage character [e.g.; isomers 1, 6, 8, and 10 shown in Figure 1 (a), (f), (h), and (j), respectively]. The previous deformations lead to a complex distribution of C—C bond lengths which, as we will see in the following, will affect the adsorption properties of the bare carbon regions. In Figure 2 calculated data for some relevant electronic and structural properties of the hereconsidered model C60(OH)12 systems are plotted. Notice that the values of the polarizability and


9


Page 10 of 51

dipole moments shown in Figure 2(a) and 2(b), respectively, depend on the fullerenol isomer. The behavior for the complex C60→(OH)12 charge transfer obtained as a function of the structure of the molecular over-layer is more strongly reflected in the dipole moment of the complexes [Figure 2(b)] changing from almost zero to 7.52 D. This range of variation can be compared with the ones reported by Rivelino et al. (0—5.815 D)45 as well as Dawid and co-workers (0.5—3.2 D)46 for C60 fullerenes with various degrees of OH-coverage, implying different degrees of solubility for the C60(OH)12 isomers considered in this work. In contrast, the polarizability [Figure 2(a)] shows less dramatic variations changing from 570—660 Bohr3 for the isomers presented in Figure 1. Interestingly, these changes are also similar to the ones calculated in Refs. 45 and 46, where variations of 469—698 Bohr3 and 654—714 Bohr3, respectively, are reported. We are confident in our estimated values since, for bare C60, we obtain a polarizability of 481.6 Bohr3, which differs by 7% when compared to the experimental value of 517 Bohr3 reported by Antoine et al.47 In Figure 2(c), the energy difference between the lowest energy fullerenol [Figure 1(a)] with each one of the considered isomers presented in Figure 1 is shown. We also include in Figure 2(d) the number of H-bonds characterizing the hydroxyl over-layer covering the C60 cage and, in Figure 2(e), we plot the number of isolated OH groups on the fullerenes surface. We must emphasize that there is no guarantee that the fullerenol isomer shown in Figure 1(a) defines the ground state atomic configuration. However, it is observed from Figure 2 that the stability of the fullerenols decreases as the number of H-bonds (isolated OH groups) decreases (increases), thus revealing a clear correlation between the energetics of the complexes and the existence of local networks of H-bonded OH groups on the C60 surface. However, it is important to emphasize that the formation of large domains of aggregated OH molecules define high-energy isomers. This is


10

Page 11 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


clearly observed from isomers 5 and 11 shown in Figure 1(e) and (k), respectively, where the 12 OH species are all adsorbed on one side of the C60 cage, and are separated by 2.6 and 3.7 eV when compared to the lowest energy structure presented in Figure 1(a). Several authors have considered the use of fullerenols as active species in solar cell devices48-49 and, consequently, the optical response of the different isomers shown in Figure 1 is important to analyze. In Figure 3(a) time dependent DFT PBE/6-31g(d,p) calculations of the first single dipole-allowed excitation for the here-considered C60(OH)12 fullerenols are presented. At this point we would like to emphasize that the best currently available methods to determine optical gaps take into account hybrids functionals. Consequently, and for the sake of comparison, we also calculate the first single dipole-allowed excitation using the hybrid HSE0650 functional. From Figure 3(a) we observe that, in both cases, this electronic excitation [which defines the optical gap (OG) of the structures] is located in the infrared region of the spectra, varying from 867—9794 nm for the pure PBE functional and, from 633—4319 nm, using the hybrid HSE06/631g(d,p) approach, which clearly defines our C60(OH)12 complexes as efficient carbon nanostructures for solar energy conversion applications. Furthermore, note that OG strongly depends on the disposition of the OH groups on the carbon surface. For both pure and hybrid functionals the smallest value is obtained for isomer 8 shown in Figure 1(h), made of a uniform distribution of neighboring pairs of OH molecules, while the largest one is found for the structure presented in Figure 1(l) (isomer 12) characterized by a random distribution of isolated OH species. It is interesting to remark how the optical response can be tuned by changing the structure of the molecular over-layer. From our calculations we obtain that the variations in the value of the electronic gap between the highest occupied and lowest unoccupied molecular orbitals (∆HL) for the different isomers shown in Figure 1 are very similar to the ones obtained


11


Page 12 of 51

for the optical gap. This is clearly seen from Figure 3(b) where, for PBE and HSE06 functionals, the optical gap versus the HOMO—LUMO gap (in eV) for each one of the C60(OH)12 considered isomers is plotted. Notice that redshifts (blueshifts) in the optical gap are strongly correlated (in an almost linear way) with a lowering (increase) of the HOMO—LUMO energy difference. The linear equations fitting this data specified as insets in Figure 3(b) yield values for R2=0.9499 (PBE) and R2=0.9746 (HSE06). We believe these linear relations can be used to predict, with high accuracy, the optical gap of any other C60(OH)12 fullerenol not considered in this work, avoiding thus the more demanding time dependent DFT calculations. The complex behavior of the dipole moment and polarizability values, together with the notable structural deformations and variations in the energy gap found for the different C60(OH)12 isomers anticipates also a complex behavior for the reactivity of OH-covered C60 fullerenes. Here, we analyze the reactivity of these complexes by calculating both and Fukui functions51 (obtained from single point calculations of the optimized structures shown in Figure 1) and then combining them to define the dual descriptor ∆ = − .52 The frozen orbital approximation is employed53 where = | | and = | | . In this cases, and define the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), respectively, of the molecules. This approximation is very convenient since allows the calculation of the Fukui function in just one-step, by only analyzing the properties of the neutral species.54-55 The sign of ∆ is an important criterion of reactivity. Molecular sites with ∆ > 0 are expected to be electrophilic, whereas molecular centers with ∆ < 0 are expected to be nucleophilic. We plot in turquoise green the zones with ∆ > 0 and in blue the areas with


12

Page 13 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


∆ < 0. In Figure 4(a), (b), and (c) the behavior of ∆ for the C60(OH)12 isomers presented in Figure 1(a), (e), and (l), respectively, is shown. It is observed that the sites with electrophilic and nucleophilic character change as function of the structure of the molecular coating, affecting the reactivity and aggregation behavior of the fullerene complexes. Notice from Figure 4(a) and (b) (which show a patchy behavior for the adsorption of the OH groups) that, while there is almost a complete absence of ∆ on the OH groups covering the C60 cage, the bare carbon regions present the most notable variations. These trends for the Fukui functions in conjunction with the molecular structure and the polarizability values presented in Figure 2(a) need to be taken into account in order to identify the fullerenol isomers that would facilitate the attachment of electrons and radicals to the surface of these carbon complexes. At this point, it is important to precise that our previous results have been obtained at zero temperature. However, it is also relevant to establish temperature-structure relationships that could provide additional physical insight to understand the observed macroscopic behavior. Consequently, we have decided to perform DFT BOMD for some representative fullerenol isomers. For the MD simulations, a time step of 0.33 fs is used and the simulations are run for 550 steps (or 180 fs) at a temperature of 300 K involving, as in previous cases, the PBE functional together with the 6-21 basis set.30-31 In Figure 5, selected snapshots of the MD simulations for isomers 1, 2, 5, and 9 shown in Figure 1(a), (b), (e), and (i), respectively, are presented. The snapshots shown in the figure reveal sizable contractions and expansions of the different C—C, C—O, and O—H bonds, together with notable variations (both opening and closings) in the C—O—H angle. At ambient temperature, no OH surface diffusion or desorption was found, a result that is consistent with the robustness of the C—OH bond. Interestingly, as time evolves, additional structural transformations consisting in proton exchange process


13


Page 14 of 51

between neighboring OH groups are obtained. This is clearly seen from the MD data reported for isomer 1 where, at t = 9.5 fs, isolated oxygen atoms (marked as a yellow spheres) appears on C60 together with a neighboring chemisorbed H2O species. Similar proton-exchange processes are observed for isomers 2, 5, and 9 where not only (isolated) oxygen atoms appear spontaneously on the carbon surface at t = 9.5, 25.8 and 39.4 fs, respectively, but also small domains of interconnected hydroxyl groups are formed at t = 78.8 and 52.6 fs for isomers 2 and 5, respectively. The previous adsorbed phases lead to a complex distribution of C—O bonds varying from 1.4—1.76 Å, the largest expansions obtained when the oxygen atoms belong to the interconnected regions on the surface. Clearly, the structure and chemical composition of the molecular over-layer can change as a function of time and these variations can alter the properties of the C60(OH)12 fullerenes. To test the validity of these predictions we have performed additional test calculations on isomer 5 (as a representative case) where the BOMD simulations at T=300 K are now run using the more accurate PBE/6-31g(d,p) calculation scheme. These results are included as supplementary material in Figure S1. From Figure S1 we notice that extending the basis set and including double polarization functions yields the same thermal behavior for the OH groups chemisorbed on the C60 surface. As in Fig. 5, proton exchange processes are not uncommon and this fact generates a varying chemical composition on the fullerene surface. In order to analyze the impact of the MD results shown in Figure 5 on the reactivity of the complexes we present in Figure 6, as a representative example, the behavior of the dual descriptor ∆ as a function of time for isomer 5 where the OH molecules are aggregated on one side of the fullerene cage. The MD atomic configurations presented for isomer 5 in Figure 5 for t = 0, 25.8, and 52.6 fs are selected to calculate ∆. In the left column, we show the top view of


14

Page 15 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the distribution of ∆ vs t while, in the right column, the side view of the same data is plotted (see the top of Figure 6). In both cases, selected zones on the fullerenol surface, labelled as z1, z2,…,z7, are specified where notable changes in ∆ are found. From the figure, we appreciate that the proton-exchange process reported in Figure 5 affects the reactivity of the molecular overlayer (see the regions in the left column marked as z1, z2, z3, and z4) since ∆ increases (decreases), mainly at the border (inner region) of the molecular island. Most interestingly, ∆ also changes its character in sites located on the bare carbon region (as specified in the right column by the zones z5, z6, and z7), a result that preclude us from deriving simple general rules. However, this behavior for ∆ could have an impact on the percentage concentration of reactive oxygen species around fullerenols, strongly influencing the scavenging activity. Even if the MD simulations shown in Figure 5 reveal that, at 300 K, no lateral displacement of the OH groups on the C60 surface are found, we believe that it is important to quantify the values of the energy barriers ∆ opposing OH diffusion in order to understand the stability of the hydroxyl over-layers at different experimental conditions. Our NEB calculations for a single OH molecule displacement on the C60 cage (not shown) yields a value for ∆ ~1.3 eV. It will be thus interesting to compare this value with the ones obtained when a single OH group moves in the presence of surrounding hydroxyl species to see if i) the lateral interactions with the coadsorbed molecules favors the here-predicted OH aggregation and ii) to quantify how the value of ∆ depends on the local atomic environment. In Figure 7 the diffusion paths (including seven images) followed by a selected test OH molecule (specified by a ball and stick model) adsorbed on isomers 2 [Figure 7(a)], 5 [Figure 7(b)], and 12 [Figure 7(c)] in different environments are plotted. Note that, when compared to isolated OH diffusion (∆ ~1.3 eV) the presence of co-adsorbed species (shown in a wireframe model) can lead to a reduction of 42% in


15


Page 16 of 51

the value of the direct energy barriers, ∆ . The largest reductions, as shown in Figures 7(a) (∆ = 0.75 eV) and 7(b) (∆ = 0.85 eV), are obtained when the OH group approaches to an OH-rich region of the C60 surface (images I→VII). Furthermore, from the reverse barriers ∆ we observe that, when the OH group belongs to a molecular island, higher energies need to be over-come to achieve a single lateral displacement. This is also clearly seen from Figures 7(a) (∆ = 1.76 eV) and 7(b) (∆ = 2.1 eV) when the test molecule moves out from a hydrogen bonded region of OH groups (images VII→I). This last result is of fundamental importance and shows the high stability of the hydroxyl domains once formed on the fullerene surface. Finally, from Figure 7(c) note that the direct (reverse) barriers for OH diffusion on a random distribution of hydroxyl groups are of the order of the ones found in bare C60.

3.2 Hydration behavior of C60(OH)12 fullerenols We must emphasize that, in biomedical applications, fullerenol structures will be in the presence of water and the formation of hydrogen bonds between surrounding H2O molecules and the OH sites of the C60 surface might alter the properties of the carbon complexes. In order to analyze solvent effects we will study the hydration properties of C60(OH)12 isomers to determine, on the one hand, the structure of the first hydration shell (that will be compared with experimental data)21 and, on the other hand, to evaluate how the stability of the molecular overlayer is perturbed by the presence of neighboring H2O species. To determine the structure of the first hydration shell we carried a Hydrogen Bond Analysis to our Molecular Dynamics Simulations trajectory of the fullerenol solvated in water with a 100 time equi-spaced frames during the simulation production period using the Maestro Software32-34. The geometric criteria


16

Page 17 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


used for defining a hydrogen bond in the Maestro Software is the following: the maximum distance from the hydrogen (H) atom to the acceptor atom (A) H—A for an H-bond to be identified is 2.8 Å. The minimum angle for a donor atom (D) bonded to hydrogen D–H—A is 120°, and the minimum H—A-B angle is 90° (B is another neighbor atom bonded to the acceptor atom A). In Figure 8(a) we show, as a representative example, an equilibrium configuration for the fullerenol isomer shown in Figure 1(e) when immersed in a water solvent of 868 H2O molecules, all of them confined in a simulation box having an edge length L=30 Å. As the starting array, the C60(OH)12 fullerenol was placed in the center of the box and the rest of the water molecules were spread evenly all over the simulation cell. No positional or orientational restrictions were applied to the fullerenol or the water molecules during the MD simulations, as clearly observed in the movie file included as supplementary material. In Figure 8(b) we present, for the previous equilibrium configuration, the corresponding number of H2O molecules defining the first hydration shell. The selected 4 water species form hydrogen bonds with the underlying OH groups located at the border of the hydroxyl island with H2O—OH intermolecular distances varying from 1.6-1.94 Å. In Figure 8(c) we show the same data as in Figure 8(b) but when the C60(OH)12 isomer characterized by a random distribution of OH groups [Figure 1(l)] is the one placed in the center of the simulation cell. From Figures 8(b) and (c) we note that the different structure of the molecular overlayer induces contrasting variations in the wettability of the fullerene compounds. While the C60(OH)12 isomer shown in Figure 8(b) is partially surrounded by an average of 4.246 ±2.12 H2O molecules, the fullerenol isomer presented in Figure 8(c) is more uniformly covered by an average of 12.078 ±5.627 water species.


17


Page 18 of 51

The adsorbed OH groups act as strong trapping sites for the neighboring water molecules and, consequently, it is reasonable to expect that water accommodation around fullerenols will be driven by the number and distribution of chemisorbed OH’s. However, we found that the existence of a hydrogen-bonded network of OH groups occurring on the C60 surface (shown in Figure 1) limits the ability of OH molecules to reorient and form hydrogen bonds with the surrounding waters, reducing thus the solubility of fullerenols. The present results can be useful to design new materials with specific wettabilities. Finally, it is important to remark that the elemental analysis performed on C60 covered with 12 OH groups in Ref. 21

reveals the

following composition: C60(OH)12·5H2O. The previous low number of H2O molecules forming the first hydration shell is very close to the one estimated in Figure 8(b), indirectly suggesting the existence of molecular domains on the C60 surface and defining the number of first coordinated waters as an important parameter to identify the presence of specific isomers in the samples. Finally, to end this section, we study the role played by the presence of water solvent on the diffusion barriers of chemisorbed OH molecules. Previous studies on metal oxide surfaces56 and graphene57 have revealed that the existence of water can alter the transport properties of adsorbates, being thus of fundamental importance to explore if the values for ∆ and ∆ obtained in Figure 7 can be reduced when simulating the properties of C60(OH)12 fullerenols in more realistic environments. A reduction in the diffusion barriers will be interesting to obtain since they would imply the possibility of having interconversion processes between different isomers shown in Figure 1 through a surface-reorganization mechanism of the adsorbates. To analyze the idea discussed in the previous paragraph we chose to work with the hydrated equilibrium configuration shown in Figure 8(b). As a second step, we re-optimize the C60(OH)12·4H2O complex at the DFT level of theory. Due to the absence of the rest of the water


18

Page 19 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


species included in Figure 8(a) and the use of DFT we obtain a reorganization of the four hydrogen-bonded waters forming the first hydration shell. Third, we select two test OH molecules belonging to the molecular hydroxyl island shown in Figure 8(b), one of them containing a physisorbed water and the second one adsorbed as a water-free species on the surface. Finally, using the NEB methodology, the energetics of single lateral displacements for the two selected test molecules is analyzed. At this point we would like to emphasize that, due to the previous assumptions, our calculated diffusion barriers might not correspond to those that would be found for a fullerenol in solution at room temperature. However, we believe that the contrasting (water assisted) diffusion behavior that has been obtained for hydrated OH groups chemisorbed on C60 could be present in real systems and might be at the origin of the observed macroscopic behavior. In the case of the water-free OH species we find (not shown) values for ∆ very similar to the ones already obtained in Figure 7. However, in the second case, we obtain a water-assisted diffusion process that lowers the energy barrier ∆ to values as low as 0.25 eV. This particular data is shown in Figure 9(a), together with some selected images of the diffusion path presented in Figure 9(b). This is a very time consuming calculation involving 13 images between the initial and final atomic configurations, which allow us to reveal a complex diffusion path having three maxima, defining three values for ∆ of 0.25, 0.35, and 0.11 eV. In the first transition state, specified by image III in Figure 9(a), an H2O molecule is found to be bonded to two neighboring OH groups of the C60 surface [see Figure 9(b)-III]. In the second transition state, specified by image X, two physisorbed waters appear on the surface together with an isolated oxygen atom (marked by a yellow sphere) bonded to the carbon network [see Figure 9(b)-X]. Here, the moving test OH group picks up an H atom from the chemisorbed water found on image III


19


Page 20 of 51

desorbing as an H2O molecule. Next, the as-formed physisorbed water slightly moves towards a neighboring carbon atom labelled as C2, which defines the end of the diffusion path. Then, this H2O species splits back into an OH group and a hydrogen atom where the former chemisorbs to C60 and the latter is transferred back to the neighboring H2O molecule. The as-formed H3O species donates one of his H atoms to the closest underlying OH group leading to a protonexchange process within the molecular island that transforms back the isolated O species into an hydroxyl molecule [see Figure 9(b)-XIII]. Clearly, our previous data demonstrates the possibility of having a (low barrier) cooperative migration of OH groups mediated by water that, depending on the experimental conditions, could lead to a molecular reorganization of the adsorbates on the C60 surface and to the coexistence of different fullerenol isomers in the samples.

3.3 Vibrational properties of C60(OH)12 fullerenols: IR vs Raman spectra As is well known, both infrared (IR) and Raman spectra of a molecule are intimately related to its atomic structure. Consequently, measurements of the vibrational properties of C60(OH)12 fullerenols can be useful to distinguish C60 cages with specific OH attachments. First, we would like to discuss infrared spectroscopy studies. In the last years, several IR spectra have been measured on C60(OH)x fullerenols with different degrees of surface coverage and composition.21, 58-59

From these studies it has been established that the IR spectra of fullerenols is characterized

by specific absorption bands located at 1080, 1370, 1620, and 3400 cm-1, defining C—O, C— O—H, C=C, and O—H vibrations which provide characteristic IR absorption peaks of C60(OH)x compounds. In addition, there are several reports where the IR spectra of model C60(OH)x complexes has been simulated using various theoretical approaches with different degrees of


20

Page 21 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


approximation. Here, we would like to take advantage of the measured IR spectra for C60(OH)12 systems reported by Kokubo et al.21 to perform a study of the vibrational properties of all the fullerenol isomers presented in Figure 1 and directly compare experimental and theoretical data. In Figure 10 we plot, as representative examples, the simulated IR spectra for isomers 2, 5, and 12 shown in Figures 1(b), (e), and (l), respectively. When comparing the three spectra we observe that the distribution of vibrational frequencies (see the continuous line) strongly depend on the disposition of the OH molecules on the carbon surface. In particular, it is observed that aggregated forms of OH groups [see Figures 10(a) and (b)] lead to the formation of a broad and asymmetric distribution of vibrational modes centered around ~3250 cm-1. In contrast, for a random distribution of OH molecules [Figure 10(c)] we note the existence of two prominent bands in the 3600—3400 cm-1, implying that we have almost independent vibrations of the chemisorbed hydroxyl species on the fullerene surface. In the 200—1700 cm-1 frequency range, we obtain also a broad distribution of IR active modes but, in all three cases, we appreciate the formation of two bands located at 1000 and 1350 cm-1, although their relative intensities differ. In order to compare with the experimental IR spectra on C60(OH)12 fullerenols reported in Ref. 21 we perform a Gaussian broadening of the distribution of vibrational frequencies shown in Figure 10 (see the dotted line). From the spectral broadening, we note that isomer 5 is the one that has an IR spectrum that closely resembles the measured IR data [see Figure 1(a) of Ref. 21] having well defined bands centered at 1054, 1381, 2689, and 3250 cm-1. In particular, the spectral feature located at 2689 cm-1, in the form of a small shoulder of the main absorption band placed at 3250 cm-1, is important to underline. Interestingly, we find that this frequency is defined by a stretching mode of an O—H bond corresponding to an hydroxyl molecule located at the center of the molecular over-layer [see the inset in Figure 10(b)], defining a unique local


21


Page 22 of 51

atomic environment. We must precise that the experimental band placed at 1620 cm-1 is apparently absent in the simulated spectra of Figure 10(b), however; by analyzing in detail the 1650—1500 cm-1 frequency domain (see the inset), we observe that a well-defined vibrational mode at 1612 cm-1 is indeed observed although with a reduced intensity. Based on the previous direct comparison between experiment and theory we conclude that i) broad and asymmetric bands around 3400 cm-1 are produced by the formation of hydroxyl molecular domains on the C60 surface, ii) the small shoulder calculated at 2689 cm-1 can be assigned to a localized vibration of an specific OH group belonging to an OH molecular island on isomer 5, and iii) that the experimentally determined three-peak structure in the 1000—1700 cm-1 range is also obtained in some of our model fullerenols but having different relative intensities. Finally, the distribution of the Raman active modes in our model C60(OH)12 fullerenes are also important to analyze. In contrast to IR spectroscopy, there are (at least to our knowledge) only a few studies in the literature reporting on the Raman spectra of fullerenols. In particular, from the theoretical side, only the wavelength independent (static limit) Raman spectra has been studied.18, 45-46 This is important to emphasize since, in general, the intensity of Raman active modes strongly depends on the wavelength of the excitation laser employed for measuring the Raman response of a sample. In order to analyze this further and also justify the use of our DFT PBE/6-31g(d,p) theoretical scheme we present in Figure 11 the simulated Raman spectra for the bare C60 fullerene. First, in Figure 11(a), we plot in the y-axis the calculated Raman intensities in the static limit. From this figure we obtain 10 Raman active modes located at 257, 420, 488, 719, 772, 1102, 1256, 1436, 1487, and 1570 cm-1 in line with previous theoretical studies.60 Interestingly, experimental measurements performed on C60 thin films,61, 62 single crystals,63 and dilute solutions64 also reveal the presence of 10 active modes placed at 273, 433, 498, 711, 775,


22

Page 23 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1101, 1251, 1427, 1470, and 1578 cm-1, being in good agreement with the simulated values. However, when comparing the different experimental reports, notable changes in the relative intensities of the Raman bands are found. We believe that these variations are mostly originated (among other factors) by the different excitation lasers used in the experiments. This has been clearly demonstrated by Feng et al.65 when analyzing the Raman spectra of nitrogen-doped graphene. In that case, the main Raman frequencies associated to graphene show prominent intensity reductions or enhancements when the samples are irradiated with different laser lines. Consequently, for an appropriate theoretical description of the Raman spectra of C60, the influence of the excitation laser on the simulated Raman intensities needs to be taken into account. In Figures. 11(b), (c), (d), and (e) we plot the wavelength dependent Raman spectra of bare C60, calculated as implemented in the GAUSSIAN09 software. Here the intensities of the active frequencies have been obtained including the effect of the most common 1064 [Figure 11(b)], 785 [Figure 11(c)], 532 [Figure 11(d)], and 488 nm [Figure 11(e)] excitation lasers. From these figures notice that, in general, sizable enhancements in the intensity of the Raman active modes are found, particularly for the 488 nm line which is the wavelength closer to the first excitation energy maxima of the simulated absorption spectra (not shown) of the C60 molecule (placed at 360 nm). In particular, the band located at 1487 cm-1 originated by symmetric stretching vibrations of C—C bonds forming the pentagonal rings is notably affected, presenting an enhancement factor of 12 when changing from 1064 to 488 nm excitation. We also observe notable variations in the relative intensities among the various Raman bands. This is particularly the case in Figures 11(d) and (e) where, when comparing with the measured data reported in Refs. 63 and 61 (for a 488 nm excitation), the correct relative intensity between the high


23


Page 24 of 51

frequency bands placed at 1436 and 1570 cm-1 is observed. However, in the simulated spectra of Figure 11(b), the use of a 1064 nm laser line changes this trend, the peak at 1570 cm-1 being now higher than the one placed at 1436 cm-1, a result that is consistent with the data reported in Ref. 63 obtained with the same laser. Clearly, in the case of the more complex C60(OH)12 fullerenol systems shown in Figure 1, we also expect a strong dependence of the Raman intensities with the excitation laser, a dependence that will be of crucial importance to perform an appropriate comparison between theoretical data and experimental measurements. In Figures 12(a), (b), and (c) we plot first, as representative examples, the wavelength independent Raman intensities for the fullerenol isomers shown in Figures 1(b), (e), and (l), respectively. Similar to infrared spectroscopy, it is observed that different arrays of OH groups on the fullerene surface leads to a contrasting distribution of Raman active modes, mainly at the high frequency region of the spectra. Notice from Figure 12 that the two well-defined radial breathing modes of bare C60 (see Figure 11), located at 257 and 488 cm-1 (marked as vertical doted lines), are found to be almost completely quenched upon functionalization of the fullerene surface. Furthermore, in the 1000—1700 cm-1 frequency range we obtain a complex distribution of Raman vibrational modes, due to the low symmetry of the here-considered fullerenols, whose intensities dominate the simulated spectra. The vibrations of the chemisorbed OH groups reveal interesting trends. As in the IR spectra, notice that aggregated forms of OH molecules [Figures 12(a) and (b)] are characterized by a broad distribution of Raman active frequencies covering the 2700—3700 cm-1 domain. In contrast, isolated hydroxyl groups lead to more localized bands where most of the OH molecules vibrate with a frequency of ~3650 cm-1. We believe that this data could be also used to corroborate the predicted formation of molecular islands on the C60 surface. Similar trends for the Raman activity in the static limit have been obtained by Rivelino


24

Page 25 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


et al.45 and Dawid and co-workers18,

46

when analyzing the vibrational properties of various

C60(OH)x isomers with different degrees of surface coverage. Following our data reported in Figure 11 for the bare C60 we plot in Figures 13 and 14 the wavelength dependent Raman spectra for the three isomers considered in Figure 12 but where the intensities are now calculated including the effect of 785 [Figure 13] and 532 nm [Figure 14] excitations. Notice that, when compared with the static spectra (Figure 12), the inclusion of both laser lines can increase the intensity of the Raman modes, but this enhancement in the signal not only depends on the laser wavelength but also on the structure of the molecular over-layer. For example for isomer 5, while the 785 nm excitation leads to a notable increase in the Raman intensities of the modes located in the 200—1500 cm-1 frequency domain [Figure 13(b)], the 532 nm line enhances instead the modes placed around 1500 cm-1 [Figure 14(b)]. In contrast, for isomer 12 with a 532 nm laser, we obtain a better resolution for the Raman active modes all along the frequency range [Figure 14(c)] while, for a 785 nm excitation [Figure 13(c)] the radial breathing mode region is notably reduced. Finally, for isomer 2, less dramatic variations are found the most important feature to remark is the formation, for a 532 nm excitation [Figure 14(a)], of a well defined band at 501 cm-1, almost completely absent in Figure 13(a). This band is originated by acomplex mixture of asymmetric radial breathing mode vibrations of the C60 cage and C—O—H angular displacements of some of the chemisorbed OH groups. This mode shows an enhancement factor of 9 when compared to the intensity of the corresponding vibration in the 785 nm spectra. We can thus conclude by saying that the use of different excitation lasers can allow us to reveal precise features of different regions of the Raman spectra, and that the calculation of the wavelength dependent Raman intensities plays a fundamental role when comparing theoretical data with experimental measurements.


25


Page 26 of 51

4. Conclusions In this work we have presented a comprehensive theoretical study dedicated to analyze the structure, thermal properties, reactivity, hydration behavior, as well as both the infrared and Raman spectra of low hydroxylated C60(OH)12 fullerenols. Various theoretical schemes with different levels of approximation have been employed and we believe interesting trends have been revealed. We predict a patchy behavior for the chemisorbed OH molecules on the fullerene surface, where the formation of local networks of hydrogen bonded neighboring hydroxyl groups plays a fundamental role on the stability of the molecular coating. Actually, at ambient temperature, we found a complex proton-exchange process within the molecular islands that can change the structure and chemical composition of the overlayer. Various C60(OH)12 isomers have been analyzed and we have found sizable variations on the electronic and optical properties, as well as on the hydration behavior, when varying the atomic structure of the complexes. In particular, we have demonstrated how the presence of neighboring water species can lower the diffusion energy barrier of the OH adsorbates on C60 through a water-assisted diffusion mechanism. The previous reduced barriers might allow for a possible reorganization mechanism of the molecules on the C60 surface that can lead to interconversion process between different isomers. The reactivity of our functionalized carbon cages reveal also interesting features. The here-reported variations of the Fukui functions should be taken into account to better understand the possible attachment of electrons and radicals to the surface of these carbon complexes. Finally, both simulated infrared and Raman spectra have been found to be efficient and complementary tools to analyze the structure of these subnanometer sized carbon materials. We have underlined the crucial role played by the wavelength dependent calculations of the Raman


26

Page 27 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


intensities to more precisely compare with experimental measurements, as well as the importance of the excitation laser to reveal precise features of different regions of the spectra.

ASSOCIATED CONTENT Supporting Information. The following files are available free of charge. Selected snapshots of the BOMD simulations performed at PBE/6-31g(d,p) level of theory. A movie of the Molecular Dynamics simulation (.mpeg)

ACKNOWLEDGMENT The authors would like to acknowledge the financial support from Consejo Nacional de Ciencia y Tecnología (CONACyT), México. Computer resources from the ACARUS supercomputer center from the Universidad de Sonora, México, as well as from the Laboratorio Nacional de Supercómputo del Sureste (LNS), México, are also acknowledged.

REFERENCES (1) Fürst, J. A.; Hashemi, J.; Markussen, T.; Brandbyge, M.; Jauho, A. P.; Nieminen, R. M. Electronic Transport Properties of Fullerene Functionalized Carbon Nanotubes: Ab Initio and Tight-Binding Calculations. Phys. Rev. B 2009, 80, 035427.


27


Page 28 of 51

(2) Peyghan, A. A.; Soleymanabadi, H.; Moradi, M. Structural and Electronic Properties of Pyrrolidine-Functionalized [60] Fullerenes. J. Phys. Chem. Solids 2013, 74, 1594–1598. (3) Shan, C.; Yena, H. J.; Wu, K.; Lin, Q.; Zhou, M.; Guo, X.; Wu, D.; Zhang, H.; Wu, G.; Wang, H. L. Functionalized Fullerenes for Highly Efficient Lithium Ion Storage: StructureProperty-Performance Correlation with Energy Implications. Nano Energy 2017, 40, 327–335. (4) Srdjenovic, B. U.; Slavić, M. N.; Stankov, K. M.; Kladar, N. V.; Jović, D. S.; Seke, M. N.; Bogdanović, V. V. Size Distribution of Fullerenol Nanoparticles in Cell Culture Medium and Their Influence on Antioxidative Enzymes in Chinese Hamster Ovary Cells. Hem. Ind. 2015, 69, 425–431. (5) Mirkov, S. M.; Djordjevic, A. N.; Andric, N. L.; Andric, S. A.; Kostic, T. S.; Bogdanovic, G. M.; Vojinovic-Miloradov, M. B.; Kovacevic, R. Z. Nitric Oxide-Scavenging Activity of Polyhydroxylated Fullerenol C60(OH)24. Nitric Oxide 2004, 11, 201–207. (6) Chaudhuri, P.; Paraskar, A.; Soni, S.; Mashelkar, R. A.; Sengupta, S. Fullerenol Cytotoxic Conjugates for Cancer Chemotherapy. ACS Nano 2009, 3, 2505–2514. (7) Maa, P. C.; Siddiqui, N. A.; Marom, G.; Kim, J. K. Dispersion and Functionalization of Carbon Nanotubes for Polymer-Based Nanocomposites: a Review. Composites: Part A 2010, 41, 1345–1367. (8) Lim, S. H.; Lin, J.; Liu, L.; Pan, H.; Pan, H. L.; Ji, W.; Feng, Y. P.; Shen, Z. Functionalization Effect on the Electronic Properties of Single Walled Carbon Nanotubes. Funct. Mater. Lett. 2008, 1, 1-6.


28

Page 29 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(9) Wu, L.; Zhang, B.; Lu, H.; Liu, C. Y. Nanoscale Ionic Materials Based on HydroxylFunctionalized Graphene. J. Mater. Chem. A 2014, 2, 1409–1417. (10) Lin, M.; Zou, R.; Shi, H.; Yu, S.; Li, X.; Guo, R.; Yan, L.; Li, G.; Liu, Y.; Dai, L. Ocular Biocompatibility Evaluation of Hydroxyl-Functionalized Graphene. Mater. Sci. Eng. C. 2015, 50, 300–308. (11) Kumar, S.; Chatterjee, K. Comprehensive Review on the Use of Graphene-Based Substrates for Regenerative Medicine and Biomedical Devices. ACS Appl. Mater. Interfaces 2016, 8, 26431−26457. (12) Ding, J.; Rahman, O.; Zhao, H.; Peng, W.; Dou, H.; Chen, H.; Yu, H. Hydroxylated Graphene-Based Flexible Carbon Film with Ultrahigh Electrical and Thermal Conductivity. Nanotechnology 2017, 28, 39LT01. (13) Chiang, L. Y.; Swirczewski, J. W.; Hsu, C. S.; Chowdhury, S. K.; Cameron, S.; Creegan, K. Multi-hydroxy Additions onto C60 Fullerene Molecules. J. Chem. Soc. Chem. Commun. 1992, 24, 1791-1793. (14) Chiang, L. Y.; Wang, L. Y.; Swirczewski, J. W.; Soled, S.; Cameron, S. Efficient Synthesis of Polyhydroxylated Fullerene Derivatives Via Hydrolysis of Polycyclosulfated Precursors. J. Org. Chem. 1994, 59, 3960–3968. (15) Troshin, P. A.; Astakhova, A. S.; Lyubovskaya, R. N. Synthesis of Fullerenols from Halofullerenes. Fuller. Nanotub. Car. N. 2005, 13, 331-343.


29


Page 30 of 51

(16) Djordjević, A.; Vojínović-Miloradov, M.; Petranović, N.; Devečerski, A.; Lazar, D.; Ribar, B. Catalytic Preparation and Characterization of C60Br24. Fuller. Sci. Technol. 1998, 6, 689-694. (17) Semenov, K. N.; Charykov, N. A.; Keskinov, V. N. Fullerenol Synthesis and Identification. Properties of the Fullerenol Water Solutions. J. Chem. Eng. Data 2011, 56, 230239. (18) Dawid, A.; Górny, K.; Gburski, Z. The Influence of Distribution of Hydroxyl Groups on Vibrational Spectra of Fullerenol C60(OH)24 Isomers: DFT study. Spectrochim. Acta Mol. Biomol. Spectrosc. 2015, 136, 1993-1997. (19) Xing, G.; Zhang, J.; Zhao, Y.; Tang, J.; Zhang, B.; Gao, X.; Yuan, H.; Qu, L.; Cao, W.; Chai, Z.; et al. Influences of Structural Properties on Stability of Fullerenols. J. Phys. Chem. B 2004, 108, 11473-11479. (20) Husebo, L. O.; Sitharaman, B.; Furukawa, K.; Kato T.; Wilson, L. J. Fullerenols Revisited as Stable Radical Anions. J. Am. Chem. Soc. 2004, 126, 12055-12064. (21) Kokubo, K.; Matsubayashi, K.; Tategaki, H.; Takada, H.; Oshima, T. Facile Synthesis of Highly Water Soluble Fullerenes More than Half Covered by Hydroxil Groups. ACS Nano 2008, 2, 327-333. (22) Marom, N.; Kim, M.; Chelikowsky, J. R. Structure Selection Based on High Vertical Electron Affinity for TiO2 Clusters. Phys. Rev. Lett. 2012, 108, 106801.


30

Page 31 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(23) Harding, D.; Ford, M. S.; Walsh, T. R.; Mackenzie, S. R.; Dramatic Size Effects and Evidence of Structural Isomers in the Reactions of Rhodium Clusters, Rhn±, with Nitrous Oxide. Phys. Chem. Chem. Phys. 2007, 9, 2130-2136. (24) 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. Gaussian09, revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (25) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865-3868. (26) Hariharan, P. C.; Pople, J. A. The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theoret. chim. Acta 1973, 28, 213–222. (27) Tang, W.; Sanville, E.; Henkelman, G. A grid-based Bader Analysis Algorithm Without Lattice Bias J. Phys.: Condens. Matter 2009, 21, 084204. (28) Lewars, E. G. The Concept of the Potential Energy Surface. Computational Chemistry, 2nd Ed.; Springer: Dordrecht, 2011. (29) Barnett, R. N.; Landman, U.; Nitzan, A.; Rajagopal, G. Born-Oppenheimer Dynamics Using Density-Functional Theory: Equilibrium and Fragmentation of Small Sodium Clusters. J. Chem. Phys. 1991, 94, 608-616. (30) Binkley, J. S.; Pople, J. A.; Hehre, W. J. Self-Consistent Molecular Orbital Methods. 21. Small Split-Valence Basis Sets for First-Row Elements. J. Am. Chem. Soc. 1980, 102, 939-947.


31


Page 32 of 51

(31) Gordon, M. S.; Binkley, J. S.; Pople, J. A.; Pietro, W. J.; Hehre, W. J. Self-Consistent Molecular-Orbital Methods. 22. Small Split-Valence Basis Sets for Second-Row Elements. J. Am. Chem. Soc. 1982, 104, 2797-2803. (32) Desmond Molecular Dynamics System, Version 3.9; D. E. Shaw Research: New York, 2014. (33) Maestro-Desmond Interoperability Tools, Version 3.9; Schrödinger: New York, 2014. (34) Bowers, K. J.; Chow, E.; Xu, H.; Dror, R. O.; Eastwood, M. P.; Gregersen, B. A.; Klepeis, J. L.; Kolossvary, I.; Moraes, M. A.; Sacerdoti, F. D.; et al. Scalable Algorithms for Molecular Dynamics Simulations on Commodity Clusters, Proceedings of the ACM/IEEE Conference on Supercomputing (SC06), Tampa, Florida, Nov 11-17, 2006. (35) Banks, J. L.; Beard, H. S.; Cao, Y.; Cho, A. E.; Damm, W.; Farid, R.; Felts, A. K.; Halgren, T. A.; Mainz, D. T.; Maple, J. R.; et al. Integrated Modeling Program, Applied Chemical Theory (IMPACT). J. Comp. Chem. 2005, 26, 1752- 1780. (36) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; Hermans, J. Interaction Models for Water in Relation to Protein Hydration, In Intermolecular Forces; Pullman, B. Ed.; D. Reidel Publishing Company: Dordrecht, 1981; 331–342. (37) Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant Pressure Molecular Dynamics Algorithms. J. Chem. Phys. 1994, 101, 4177-44189. (38) Martyna, G. J.; Klein, M. L.; Tuckerman, M. Nose-Hoover chains: The Canonical Ensemble Via Continuous Dynamics. J. Chem. Phys. 1992, 97, 2635-2643.


32

Page 33 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(39) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103, 8577-8593. (40) Sheppard, D.; Terrell, R.: Henkelman, G. Optimization Methods for Finding Minimum Energy Paths. J. Chem. Phys. 2008, 128, 134106. (41) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in the Nudged Elastic Band Method for Finding Minimum Energy Paths and Saddle Points. J. Chem. Phys. 2000, 113, 99789985. (42) Henkelman, G.; Uberuaga, B. P.; Jonsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 99019904. (43) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; et al. QUANTUM ESPRESSO: A Modular and OpenSource Software Project for Quantum Simulations of Materials. J. Phys.: Condens. Matter 2009, 21, 395502. (44) He, H.; Zheng, L.; Jin, P.; Yang, M. The Structural Stability of Polyhydroxylated C60(OH)24: Density Functional Theory Characterizations. Comp. Theor. Chem. 2011, 974, 16-20. (45) Rivelino, R.; Malaspina, T.; Fileti, E. E. Structure, Stability, Depolarized Light Scattering, and Vibrational Spectra of Fullerenols From All-Electron Density-Functional-Theory Calculations. Phys. Rev. A 2009, 79, 013201.


33


Page 34 of 51

(46) Dawid, A.; Górny, K.; Gburski, Z. The Water Solvent Effect on Infrared and Raman Spectra of C60(OH)24 Fullerenol Isomers: DFT Study. J. Phys. Chem. C 2017, 121, 2303−2315. (47) Antoine, R.; Dugourd, P.; Rayane, D.; Benichou, E.; Broyer, M.; Chandezon, F.; Guet, C. Direct Measurement of the Electric Polarizability of Isolated C60 Molecules. J. Chem. Phys. 1999, 110, 9771-9772. (48) Cao, T.; Wang, Z.; Xia, Y.; Song, B.; Zhou, Y.; Chen, N.; Li. Y. Facilitating Electron Transportation in Perovskite Solar Cells via Water-Soluble Fullerenol Interlayers. ACS Appl. Mater. Interfaces 2016, 8, 18284−18291. (49) Wang, N.; Sun, L.; Zhang, X.; Bao, X.; Zheng, W.; Yang, R. Easily-Accessible Fullerenol as a Cathode Buffer Layer for Inverted Organic Photovoltaic Devices. RSC Adv. 2014, 4, 2588625891. (50) Krukau, A. V.; Vydrov, O. A.; Izmaylov, A. F.; Scuseria, G. E. Influence of the Exchange Screening Parameter on the Performance of Screened Hybrid Functionals J. Chem. Phys. 2006, 125, 224106. (51) Parr, R. G.; Yang, W. Density Functional Approach to the Frontier-Electron Theory of Chemical Reactivity. J. Am. Chem. Soc. 1984, 106, 4049-4050. (52) Morell, C.; Grand, A.; Toro-Labbe, A. New Dual Descriptor for Chemical Reactivity. J. Phys. Chem. A 2005, 109, 205-212. (53) Koopman, T. Über die Zuordnung von Wellenfunktionen und Eigenwerten zu den Einzelnen Elektronen Eines Atoms. Physica 1934, 1,104-113.


34

Page 35 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(54) Fuentealba, P.; Florez, E.; Tiznado, W. Topological Analysis of the Fukui Function. J. Chem. Theory Comput. 2010, 6, 1470–1478. (55) Martínez-Araya, J. I. Why is the Dual Descriptor a More Accurate Local Reactivity Descriptor than Fukui Functions? J. Math Chem. 2015, 53, 451-465. (56) Merte, L. R.; Peng, G.; Bechstein, R.; Rieboldt, F.; Farberow, C. A.; Grabow, L. C.; Kudernatsch, W.; Wendt, S.; Lægsgaard, E.; Mavrikakis, M.; et al. Water-Mediated Proton Hopping on an Iron Oxide Surface. Science 2012, 336, 889-893. (57) Zhao, Y.; Gennett, T. Water-Mediated Cooperative Migration of Chemisorbed Hydrogen on Graphene. Phys. Rev. Lett. 2014, 112, 076101. (58) Indeglia, P. A.; Georgieva, A.; Krishna, V. B.; Bonzongo, J. J. Physicochemical Characterization of Fullerenol and Fullerenol Synthesis By-Products Prepared in Alkaline Media. J. Nanopart. Res. 2014, 16, 2599. (59) Kokubo, K.; Shirakawa, S.; Kobayashi, N.; Aoshima, H.; Oshima, T. Facile and Scalable Synthesis of a Highly Hydroxylated Water-Soluble Fullerenol as a Single Nanoparticle. Nano Res. 2011, 4, 204-215. (60) Choi, C. H.; Kertesz, M.; Mihaly, L. Vibrational Assignment of All 46 Fundamentals of C60 and C606-: Scaled Quantum Mechanical Results Performed in Redundant Internal Coordinates and Compared to Experiments. J. Phys. Chem. A 2000, 104, 102-112. (61) Akers, K. L.; Douketis, C.; Haslett, T. L.; Moskovits, M. Raman Spectroscopy of C60 Solid Films: A Tale of Two Spectra. J. Phys. Chem. 1994, 98, 10824-10831.


35


Page 36 of 51

(62) Bethune, D. S.; Meijer, G.; Tang, W. C.; Rosen, H. J. The Vibrational Raman Spectra of Purified Solid Films of C60 and C70. Chem. Phys. Lett. 1990, 174, 219-222. (63) Kuzmany, H.; Pfeiffer, R.; Hulman, M.; Kramberger, C. Raman Spectroscopy of Fullerenes and Fullerene–Nanotube Composites. Phil. Trans. R. Soc. Lond. A 2004, 362, 23752406. (64) Scharff, P.; Risch, K.; Carta-Abelmann, L.; Dmytruk, I. M.; Bilyi, M. M.; Golub, O. A.; Khavryuchenko, A. V.; Buzaneva, E. V.; Aksenov, V. L.; Avdeev, M. V.; et al. Structure of C60 Fullerene in Water: Spectroscopic Data. Carbon 2004, 42, 1203–1206. (65) Feng, S.; dos Santos, M. C.; Carvalho, B. R.; Lv, R.; Li, Q.; Fujisawa, K.; Elías, A. L.; Lei, Y.; Perea-López, N.; Endo, M.; et al. Ultrasensitive Molecular Sensor Using N-Doped Graphene Through Enhanced Raman Scattering. Sci. Adv. 2016, 2, 1-12.


36

Page 37 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Optimized atomic configurations for the here-considered model C60(OH)12 fullerenols.


37


Page 38 of 51

Figure 2. Calculated electronic and structural properties for the here-considered model C60(OH)12 fullerenols.


38

Page 39 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. (a) Calculated first single dipole-allowed excitation (optical gap) of model C60(OH)12 fullerenols. (b) Optical gap vs. electronic gap for the here-used C60(OH)12 fullerenols.


39


Page 40 of 51

Figure 4. Distribution of the dual descriptor = − for different isomers. In turquoise green (blue) we show the zones with > 0 ( < 0).


40

Page 41 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Selected snapshots of the BOMD simulations performed at 300 K for some representative C60(OH)12 fullerenols. Isolated oxygens atoms on the C60 surface are specified as yellow spheres.


41


Page 42 of 51

Figure 6. Distribution of the dual descriptor = − as a function of time for isomer 5 [Figure 1(e)]. In the left (right) column we present the top (side) view of the considered fullerenol isomers. Selected zones on the surface are specified as z1, z2,…, z7. In turquoise green (blue) we show the zones with > 0 ( < 0).


42

Page 43 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7. Calculated diffusion paths followed by a test OH molecule (specified by a ball and stick model) absorbed on isomer (a) 2, (b) 5, and (c) 12. The initial (final) atomic configuration are shown as insets in the figures.


43


Page 44 of 51

Figure 8. (a) Selected snapshots of the classical NPT MD simulations for the C60(OH)12 isomer specified in Figure 1(e) immersed in 868 H2O molecules. In (b) and (c) we show the location of the H2O species forming the first hydration shell on isomer 5 and 12, respectively.


44

Page 45 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 9. (a) Calculated diffusion path followed by a test OH molecule containing a physisorbed H2O molecule. The initial and final atomic configurations are specified as insets in the figures. (b) Selected images of the diffusion path shown in (a). The carbon atom defining the end of the diffusion path is defined as C2.


45


Page 46 of 51

Figure 10. Simulated infrared (IR) spectra for selected C60(OH)12 isomers. The dotted line corresponds to a Gaussian broadening of the discrete distribution of vibrational frequencies. As insets we include the atomic configuration of isomer 5 as well as its corresponding distribution of frequencies on the 1500—1650 cm-1 range.


46

Page 47 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 11. (a) Simulated Raman spectra (static limit) of the C60 molecule. In (b), (c), (d), and (e) we plot the wavelength dependent Raman spectra of C60 including the effect of 1064, 785, 532, and 488 nm laser lines, respectively.


47


Page 48 of 51

Figure 12. Simulated Raman spectra (static limit) of three representative C60(OH)12 isomers, shown as insets in the figures. The radial breathing modes of bare C60 located at 257 and 488 cm1

as well as the C=C stretching at 1487 cm-1 (see Figure 11) are specified by vertical dotted lines.


48

Page 49 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 13. Simulated wavelength dependent Raman spectra (785 nm) of three representative C60(OH)12 isomers, shown as insets in the figures. The radial breathing modes of bare C60 located at 257 and 488 cm-1 as well as the C=C stretching at 1487 cm-1 (see Figure 11) are specified by vertical dotted lines.


49


Page 50 of 51

Figure 14. Simulated wavelength dependent Raman spectra (532 nm) of three representative C60(OH)12 isomers, shown as insets in the figures. The radial breathing modes of bare C60 located at 257 and 488 cm-1 as well as the C=C stretching at 1487 cm-1 (see Figure 11) are specified by vertical dotted lines.


50

Page 51 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TABLE OF CONTENTS IMAGE


51

Low Hydroxylated Fullerenes: Stability, Thermal Behavior, and

Recommend Documents