Synthesis of Photoresponsive Dual NIR Two-Photon Absorptive [60]Fullerene Triads and Tetrads

Broadband nonlinear optical (NLO) organic nanostructures exhibiting both ultrafast photoresponse and a large cross-section of two-photon absorption throughout a wide NIR spectrum may make them suitable for use as nonlinear biophotonic materials. We report here the synthesis and characterization of two C60-(antenna)x analogous compounds as branched triad C60(>DPAF-C18)(>CPAF-C2M) and tetrad C60(>DPAF-C18)(>CPAF-C2M)2 nanostructures. These compounds showed approximately equal extinction coefficients of optical absorption over 400–550 nm that corresponds to near-IR two-photon based excitation wavelengths at 780–1,100 nm. Accordingly, they may be utilized as potential precursor candidates to the active-core structures of photosensitizing nanodrugs for 2γ-PDT in the biological optical window of 800–1,050 nm.


Introduction
Fullerenes are nanocarbon cages with all sp 2 carbons interlinked in a structure of hollow sphere. Highly strained curving regions of the cage surface consist of chemically reactive six fulvalenyl bridging olefins that can be utilized for making nucleophilic addition reactions. Chemical modification of C 60 on only a limit number of functionalization sites may not lead to much alternation of the cage's photophysical properties. Conversely, nucleophilic addition of one or two light-harvesting antenna chromophores will largely enhance the cage's ability to respond and perform various photoinduced electronic and energy-related events by acting as an electron-acceptor [1,2]. The development of broadband nonlinear optical (NLO) organic nanostructures exhibiting both ultrafast photo-response and high efficiency in two-photon absorption throughout a wide NIR spectrum to variable laser pulses with duration ranging from fs to ns remains as the focus of nonlinear biophotonic materials. The goal requires the design of sophisticated, hydrophilic and biocompatible multifunctional NLO materials for two-photon absorption (2PA) based photodynamic therapy (2γ-PDT) [3][4][5][6][7] against pathogens and cancer to minimize the damage to surrounding normal tissue. Photoresponsive complex fullerene derivatives [8][9][10][11][12][13][14][15] and a number of organic chromophores [16][17][18][19] have been found to exhibit enhanced nonlinear photonic behavior. The control of photodynamic effect is precise due to the fact that 2γ-PDT can only be practiced at the focal area of the laser beam that prevents side-effects arising from the undesirable photokilling of normal cells.
The most abundant [60]fullerene is more readily available commercially in up to kilogram quantities than a number of higher fullerenes. However, its visible absorption extinction coefficient is rather low. This limitation can be overcome by attaching highly fluorescent chromophores as light-harvesting antenna units, such as porphyrin [20,21] or dialkyldiphenylaminofluorene (DPAF-C n ), to enhance visible absorption of the resulting conjugates and, in the latter cases, 2PA cross-sections in the NIR wavelengths [10,13,14]. The absorbed photoenergy by the donor antenna was able to undergo efficient intramolecular transfer to the fullerene acceptor moiety, leading to the generation of excited triplet cage state 3 (C 60 >)* after the intersystem crossing from its excited singlet state 1 (C 60 >)*. Triplet energy transfer from 3 (C 60 >)* to molecular oxygen produces singlet oxygen ( 1 O 2 ) that gives the cytotoxic effect to the cells in the Type-II photochemistry [22,23]. In this paper, we report the synthesis and spectroscopic characterization of photoresponsive dual NIR two-photon absorptive [60]fullerene triads and tetrads using the extended synthetic method for the preparation of their corresponding monoadduct analogous C 60 (>DPAF-C 18 ) 1 and C 60 (>CPAF-C 2M ) 2, as shown in Scheme 1. These triads and tetrads are capable of undergoing 2PA-based photoexcitation process at either 780 or 980 nm making them potential precursor candidates to the active-core structures of nanodrugs for 2γ-PDT.

Results and Discussion
Structural design of hybrid [60]fullerene triads and tetrads was based on both linear and nonlinear optical characteristics of 9,9-dioctadecyl-2-diphenylaminofluorenyl-61-carbonylmethano[60]fullerene (1), C 60 (>DPAF-C 18 ) [24], and 9,9-di(2-methoxyethyl)-2-diphenylaminofluorenyl-61-(1,1-dicyanoethylenyl)methano[60]fullerene (2), C 60 (>CPAF-C 2M ) [25], to construct an unique nanostructure system with a shared C 60 cage. Specifically, covalent attachment of an antenna donor chromophore to a C 60 molecule (electron-acceptor) was accomplished via a periconjugation linkage with a physical separation distance of only <3.5 Ǻ between the donor and acceptor moieties. This led to the realization of ultrafast intramolecular energy-and/or electron-transfer from photoexcited antenna moiety to C 60 in <130-150 fs [14] that made this type of C 60 -antenna conjugates, C 60 (>DPAF-C n ) x , capable of exhibiting photoresponse in a nearly instantaneous time scale to protect against high-intensity radiation. By increasing the number of attached antennae to four per C 60 cage giving starburst pentad nanostructures, highly enhanced fs 2PA cross-section values were observed in a concentration-dependent manner [26]. Upon the chemical alteration of the keto group of C 60 (>DPAF-C n ) bridging between C 60 and the antenna moiety to a highly electron-withdrawing 1,1-dicyanoethylenyl (DCE) group, it was possible to extend the π-conjugation in the resulting C 60 (>CPAF-C n ) analogous chromophore molecules to a close contact with the cage current. This led to a large bathochromic shift of the linear optical absorption of C 60 (>CPAF-C 2 ) moving from 410 nm (λ max ) of the parent keto-compound to 503 nm with the shoulder band being extended beyond 550 nm in the UV-vis spectrum. The shift considerably increased its light-harvesting ability in visible wavelengths and caused a nearly 6-fold higher in the production quantum yield of singlet oxygen ( 1 O 2 ) from C 60 (>CPAF-C 2M ) as compared with that of C 60 (>DPAF-C 2M ). The mechanism of 1 O 2 production was originated from the intermolecular triplet-energy transfer from the 3 (C 60 >)* cage moiety to 3 O 2 . A large increase in the production of reactive oxygen species (ROS) by excited C 60 (>CPAF-C 2M ) explained its effective photokilling of HeLa cells in vitro, via 1γ-PDT [25]. The observation demonstrated the intramolecular/intramolecular interaction between the excited CPAF-C n donor antenna moiety and the acceptor C 60 cage that was also confirmed by transient absorption spectroscopic measurements using ns laser pulses at 480-500 nm [27]. The behavior resembles that of DPAF-C n antenna with transient photoexcitation at 380-410 nm reported previously [28]. By extending the same intramolecular photophysical properties to 2PA-based excitation applications, these C 60 -(antenna) x analogous nanostructures may be utilized as potential photosensitizers for 2γ-PDT at either 800 nm (with DPAF antenna) or 1,000 nm (with CPAF antenna) that is well-suited to the biological optical window of 800-1,100 nm.
Accordingly, selective attachment of these two antenna moiety types DPAF-C n and CPAF-C n in combination as hybrid chromophore addends to a single C 60 cage should result in the formation of new methano[60]fullerene triads, C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 3, and tetrads, C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 2 4, as shown in Scheme 1. The core chromophore moiety of 3 and 4 will then be capable of performing dual-band 2γ-PDT-based photoinduced biocidal effects with enhanced penetration depth at 800-1,100 nm. Synthetically, preparation of 3 and 4 was accomplished by the synthesis of a structurally well-defined monoadduct 1, followed by the attachment of one or two CPAF-C 2M antenna in sequence. A key intermediate precursor, 7-α-bromoacetyl-9,9-dioctadecyl-2-diphenylaminofluorene (BrDPAF-C 18 , 8) was prepared by a three-step reaction involving first palladium catalyzed diphenylamination of commercially available 2-bromofluorene at the C2 position of the fluorene ring to afford DPAF 5 (Scheme1). It was followed by dialkylation at the C9 carbon position of 5 using 1-bromooctadecane as the reagent in the presence of potassium t-butoxide, as a base, in THF at 0-25 °C to give the corresponding 9,9-dioctadecyl-2-diphenylaminofluorene (DPAF-C 18 ) in 97% yield. Friedel-Crafts acylation of DPAF-C 18 with α-bromoacetyl bromide and AlCl 3 in CH 2 Cl-CH 2 Cl at 0 °C for a period of 4.0 h afforded the compound 8 in a yield of 96%. Addition reaction of 8 to C 60 was carried out in the presence of 1,8-diazabicyclo [5.4.0]undec-7-ene (DBU, 1.0 eq.) at ambient temperature for 4.0 h to result in C 60 (>DPAF-C 18 ) 1 in 65% yield (based on recovered residual C 60 ) after column chromatographic purification. A similar reaction sequence was applied for the synthesis of the compound 2 by replacing two octadecyl groups with 2-methoxyethyl groups. Thus, 2-methoxyethyl methanesulfonate was used as a leaving group for dialkylation of DPAF 5 followed by Friedel-Crafts acylation with α-bromoacetyl bromide and AlCl 3 to yield 7-α-bromoacetyl-9,9-di(2-methoxy)ethyl-2-diphenylaminofluorene (BrDPAF-C 2M , 7), Subsequent conversion of the keto group of 7 to the corresponding 1,1-dicyanoethylenyl (DCE) group was carried out by the reaction using malononitrile as a reagent, pyridine as a base, and titanium tetrachloride as a deoxygenation agent in dry chloroform at ambient temperature for a short period of 5.0 min. The reaction resulted in the corresponding diphenylaminofluorene BrCPAF-C 2M 9 in a yield of 89% after chromatographic purification (PTLC, SiO 2 , CHCl 3 as the eluent). Attachment of a CPAF-C 2M antenna arm to a C 60 cage was carried out by identical reaction conditions as those for 1 with DBU (1.0 eq.) at room temperature for 4.0 h to afford 7-(1,2-dihydro-1,2-methano[60]fullerene-61-{1,1-dicyanoethylenyl})-9,9-di(2-methoxyethyl)-2-diphenylaminofluorene C 60 (>CPAF-C 2M ), 2) as orange red solids in 53% yield (based on recovered C 60 ). The bulkiness of DPAF-C 18 and CPAF-C 2M in size can prevent these two types of antenna moieties form locating in close vicinity to each other at the cage surface. By considering the regio-location of reactive bicyclopentadienyl olefin bonds on the fullerene surface, when the first antenna is bound at the northpole location, the second antenna arm is most likely to be pushed away to the equator area of the C 60 sphere. Therefore, only a very limited number of multiadduct regioisomers per C 60 are likely to form. Indeed, by controlling the reaction kinetic rate with two molar equivalents of CPAF-C 2M applied in the reaction with 1 in the presence of DBU (2.0 eq.), only two clear PTLC (SiO 2 , toluene-ethyl acetate/9:1 as the eluent) bands in the product mixtures were observed in addition to the starting 1 (~15%). The first less polar product band at R f = 0.5 was found to be the bisadduct C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 3 isolated as orange-brown solids in 28% yield. The second more polar product band at R f = 0.4 (toluene-ethyl acetate/4:1 as the eluent) was determined to be the trisadduct C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 2 4 isolated as red-brown solids in 40% yield.
Spectroscopic characterization of 1 and 2 was performed mainly by: (i) the clear detection of a group of molecular mass ion peaks with the maximum peak intensity centered at m/z 1,600 (MH + of 1) and 1,258 (MH + of 2) (Supporting Information) using positive ion matrix-assisted laser desorption ionization (MALDI-TOF) mass spectroscopy and (ii) analyses of 13 C-NMR spectra. The former spectra were also accompanied with two groups of fragmented mass ion peaks at m/z 720 and 734/735 corresponding to the mass units of C 60 and C 60 >, respectively, indicating high stability of the fullerene cage under MALDI-MS conditions. In addition to the IR spectral analysis (Figure 1 In the same spectra, the peaks at δ 40.14/41.22 and 72.48/72.30 were assigned to the cyclopropanyl or methanofullerene carbon C 61 (C 60 >) and fullerenyl sp 3 carbons of 1/2, respectively. The rest of aromatic carbon peaks were separated from each other into three different groups with assigned chemical shifts of (i) three aminoaryl carbons of 1/2 at δ (153.55, 151.20, 148.77)/(151.83, 150.31, 149.45) in close resemblance to those of 8 and 9, respectively, (ii) phenyl and fluorenyl carbons at δ 115-135, and (iii) fullerenyl sp 2 carbons located at δ 136-148, as shown in Figure 2. A total of 30 fullerenyl carbon (28 × 2C and 2 × 1C) signals, some with similar or slightly shifted δ, were accounted for 58 sp 2 fullerenyl carbons that fits well with a C 2 molecular symmetry of the compounds 1 and 2.   With well-characterized structures of 1 and 2, we were able to utilize their 1 H-NMR spectra for the correlation and identification of hybrid [60]fullerene triads 3 and tetrads 4. Upon the attachment of one CPAF-C 2M antenna arm to 1, a new cyano stretching band centered at 2,223 cm −1 in addition to the carbonyl stretching band at 1678 cm −1 were detected as expected. Intensity of characteristic half-fullerene cage absorption band at ~526 cm −1 was found to decrease significantly going from that of 1, 3, to 4 ( Figure 1) indicating the increasing percentage of regioisomers having at least one CPAF-C 2M addend located at more than 90° away the DPAF-C 18 arm (or the other side of the cage surface). Large difference of 1 H chemical shifts among alkyl groups of DPAF-C 18 (methyl and the most of methylene proton peaks at  0.69-1.29) and CPAF-C 2M (singlet terminal methoxy CH 3 -O-proton peak at  2.95 and triplet methylenoxy -CH 2 -O-proton peaks centered at  2.73) allowed us to measure a clear proton integration count to verify the structure of 3 and 4 as a bisadduct and trisadduct, respectively, as shown in Figure 3. A more branched structure of 4 was also evident by the detection of a higher aromatic proton integration ratio in the region of  7.5-7.8 and 8.10-8.15 [ Figure 3(b) and (e)] of CPAF moieties. The most distinguishable proton peaks at  5.5-5.7 in these spectra were assigned for α-protons each bound on the cyclopropanyl carbon located between either the keto (for DPAF) or DCE (for CPAF) group and the C 60 cage. Owing to the fullerenyl ring current, a large down-field shift of the   Therefore, detected α-H a peaks each in different intensities can be separately grouped into and accounted for two major regioisomer products and one minor regioisomer product. High similarity of molecular polarity among these regioisomers prohibited us to separate them chromatographically. However, we were able to confirm the identical composition mass of these regioisomers by detecting an group of sharp molecular mass ions with the maximum mass at m/z 2,136 (MH + ), as shown in Figure 5  Optical absorption of 1 and 2 [ Figure 6(d) and (c), respectively] was characterized by two distinguishable bands centered at 260 and 325-327 nm both arising from the C 60 > cage moiety that agrees with allowed 1 T 1u → 1 A g transition bands of pristine C 60 [29]. The third band with λ max at either 411 or 501 nm for 1 or 2, respectively, matches approximately with those of the corresponding precursor compound BrDPAF-C 18 [ Figure 6(a)] or BrCPAF-C 2M [ Figure 6(b)]. These bands are in the characteristic photoresponsive wavelength range of DPAF-C 18 or CPAF-C 2M antenna, respectively. When these two types of antenna were simultaneously attached to the same C 60 in 3, two absorption bands with λ max (ε) at 413 (3.9 × 10 4 ) and 494 nm (2.3 × 10 4 L/mol-cm) were observed in the spectrum showing extinction coefficient ε values matching roughly with those of 1 and 2. This clearly revealed a 1:1 ratio of DPAF-C 18 /CPAF-C 2M in 3 consistent with its composition. As the number of CPAF-C 2M antenna being increased to two in 4, the corresponding two bands remained in the same range with λ max (ε) at 417 (4.6 × 10 4 ) and 500 nm (4.6 × 10 4 L/mol-cm). The extinction coefficient ε value of the second band is nearly double to that of 3. The structural modification resulted in approximately equal visible absorption in intensity over the full wavelength range of 400-550 nm. Accordingly, these bands can be utilized for the corresponding near-IR two-photon absorption excitation at 800-1,100 nm, giving broadband characteristics of the materials while exhibiting good linear transparency beyond 800 nm [ Figure 6  It is noteworthy that excited state intramolecular energy-transfer resonance phenomena between the DPAF-C 18 and CPAF-C 2M antenna around the cage surface of 3 and 4 were observed. We first characterized the steady-state fluorescence (FL) emission of each antenna component using the model compound a-DPAF-C 2 10 and a-CPAF-C 2 11 (Scheme 1) in toluene as the spectroscopic reference. Upon photoexcitation of 10 at 410 nm to match with the optical absorption band of DPAF-C 18 , strong fluorescence emissions of 1 (a-DPAF-C 2 )* centered at 481 nm (λ max,em ) [ Figure 7A(a)] were detected. Likewise, strong FL emissions of 1 (a-CPAF-C 2 )* centered at 543 nm (λ max,em ) [ Figure 7B(a)] were observed when 11 was irradiated at 478 nm which matches with the optical absorption band of CPAF-C 2M . As expected, highly efficient intramolecular fluorescence quenching of these two bands by C 60 occurred when 1 and 2 were photoexcited at the same corresponding light wavelength, as shown in Figure 7A(b) and 7B(b), respectively. This photophysical event led to the subsequent emission from the 1 (C 60 >)* → 1 (C 60 >) o transition at 704 and 708 nm, respectively. The possible phosphorescence emission from 3 (C 60 >)* → 1 (C 60 >) o transition expected at ~800-850 nm was too weak to be detected. In the case of the bisadduct C 60 (>DPAF-C 18 ) 2 , two FL bands with λ max at 451 and 525 (shoulder) nm [ Figure 7A(c)] were shown, indicating incomplete quenching of C 60 [> 1 (DPAF)*-C 18 ] 2 by C 60 > when the number of antenna are more than one. Similarly, three fluorescence bands with λ max at 506, 531, and 615 (broad) nm [ Figure 7B(c)] were found for the bisadduct C 60 (>CPAF-C 2M ) 2 . Owing to high similarity on the structural moieties, these FL bands were used as the reference for the FL spectroscopic characterization of 3 and 4. Interestingly, upon photoexcitation of the triad 3 specifically on the DPAF-C 18 antenna moiety at 410 nm, the resulting FL spectrum [ Figure 7A(d)] displayed a weak broad FL band at 448 [from 1 (DPAF)*-C 18 ] and broad bands at 525-650 nm along with the 1 (C 60 >)* emission band centered at 708 nm. The latter broad bands fit in the similar range as those of C 60 [> 1 (CPAF)*-C 2M ] 2 . As the number of CPAF-C 2M antenna being increased by one to the structure of tetrad C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 2 4, the intensity of broad FL bands at 525-650 nm became more pronounced while retaining the same intensity of the 1 (C 60 >)* emission band at 709 nm [ Figure 7A(e)]. The data revealed intramolecular Förster energy-transfer resonance from the photoexcited C 60 [> 1 (DPAF)*-C 18 ](>CPAF-C 2M ) 2 state to both 1 C 60 *(>DPAF-C 18 )(>CPAF-C 2M ) 2 and C 60 (>DPAF-C 18 )[> 1 (CPAF)*-C 2M ] 2 states. The latter energy-transfer is possible since: (i) the energy level of 1 (CPAF)*-C 2M is lower than that of 1 (DPAF)*-C 18 , (ii) the energy of this FL band at 430-475 nm is slightly higher than that of the CPAF-C 2M absorption λ max at 500 nm, and (iii) there is a partial overlap of emission/absorption bands to enhance the energy-transfer efficiency. Conversely, photoexcitation of 3 specifically on the CPAF-C 2M antenna moiety at 478 nm, the resulting FL spectrum [ Figure 7B(d)] showed only a weak broad FL band at 540-660 nm along with the 1 (C 60 >)* emission band centered at 707 nm. Intensity of the former broad band was significantly increased using 4 [ Figure 7B(e)] with photoexcitation on both two CPAF-C 2M antenna moieties. This confirmed the band was contributed from the C 60 (>DPAF-C 18 )[> 1 (CPAF)*-C 2M ] 2 state, which was capable of inducing the 1 C 60 *(>DPAF-C 18 )(>CPAF-C 2M ) 2 state subsequently.
Data of femtosecond Z-scans and nonlinear light-intensity transmittance reduction measurements of C 60 (>CPAF-C 9 ), 3, and 4, performed as a function of irradiance intensity using 125-fs laser pulses at either 780 nm (corresponding to the two-photon absorption of DPAF moieties) or 980 nm (corresponding to the two-photon absorption of CPAF moieties) at the concentration of 5 × 10 −3 M in toluene, were provided in the supporting information. These data substantiated the nonlinear photonic characteristics of 3 and 4 showing dual NIR two-photon absorption capability that led to large nonlinear light-transmittance reduction in intensity in these two wavelength ranges up to the fs-laser fluence of 120 GW/cm 2 . Observed sufficiently large two-photon absorption cross-section values of 3 and 4 may allow their uses as the nanocarbon core of 2γ-PDT agents after the chemical modification with water-soluble side-chains and cationic targeting segments on the fluorene ring moiety.

Spectroscopic Measurements
Infrared spectra were recorded as KBr pellets on a Thermo Nicolet Avatar 370 FT-IR spectrometer. 1 H-NMR and 13 C-NMR spectra were recorded on a Bruker Avance Spectrospin-500 spectrometer.
UV-vis spectra were recorded on a Perkin Elmer Lambda 750 UV-vis-NIR Spectrometer. Photoluminescence (PL) spectra were measured using PTI Fluorescence Master Systems connected with a photomultiplier (914 Photomultiplier Detection System) with Xenon short arc lamp as the excitation source. Mass spectroscopic measurements were performed by the use of positive ion matrixassisted laser desorption ionization (MALDI-TOF) technique on a micromass M@LDI-LR mass spectrometer. The sample blended or dissolved in the matrix material was irradiated by nitrogen UV laser at 337 nm with 10 Hz pulses under high vacuum. Mass ion peaks were identified for the spectrum using the MassLynx v4.0 software. In a typical experiment, the samples of C 60 (>DPAF-C 18 ), C 60 (>CPAF-C 2M ), C 60 (>DPAF-C 18 )(>CPAF-C 2M ), or C 60 (>DPAF-C 18 )(>CPAF-C 2M ) 2 were dissolved in CHCl 3 in a concentration of 1.0 mg/mL. The matrix of 3,5-dimethoxy-4-hydroxycinnamic acid (sinapic acid) was dissolved in THF in a concentration of 10 mg/mL. The solution of matrix (1.0 mL) was taken and mixed with the sample solution (0.1 mL) prior to the deposition on a stainless-steel MALDI target probe. It was subsequently dried at ambient temperature.