Broadband Two-Photon Absorption Characteristics of Highly Photostable Fluorenyl-Dicyanoethylenylated [60]Fullerene Dyads

We synthesized four C60-(light-harvesting antenna) dyads C60 (>CPAF-Cn) (n = 4, 9, 12, or 18) 1-Cn for the investigation of their broadband nonlinear absorption effect. Since we have previously demonstrated their high function as two-photon absorption (2PA) materials at 1000 nm, a different 2PA wavelength of 780 nm was applied in the study. The combined data taken at two different wavelength ranges substantiated the broadband characteristics of 1-Cn. We proposed that the observed broadband absorptions may be attributed by a partial π-conjugation between the C60 > cage and CPAF-Cn moieties, via endinitrile tautomeric resonance, giving a resonance state with enhanced molecular conjugation. This transient state could increase its 2PA and excited-state absorption at 800 nm. In addition, a trend of concentration-dependent 2PA cross-section (σ2 ) and excited-state absorption magnitude was detected showing a higher σ value at a lower concentration that was correlated to increasing molecular separation with less aggregation for dyads C60(>CPAF-C18) and C60(>CPAF-C9), as better 2PA and excited-state absorbers.


Materials Characterization
We have demonstrated the use of dialkyldiphenylaminofluorenyl-keto-[60]fullerene C60(>DPAF-Cn) dyads 2-Cn [16], branched triads C60(>DPAF-Cn)x (x = 2) [17], and the related starburst pentads (x = 4) [8] for the study of simultaneous 2PA phenomena under the photoexcitation of a 780-nm laser light in the fs region. In these compounds, DPAF-Cn was used as the light-harvesting antenna chromophore to compensate for the low optical absorption of the fullerene cage at wavelengths beyond 350 nm. It is also functioning as an electron donor to provide one electron capable of being transferred intramolecularly to the C60> cage moiety upon 2PA activation that increased largely the excited state absorptions leading to 2PA cross-section enhancement. As a general strategy to extend the active 2PA wavelength to the longer NIR region, we modified the bridging keto group in dyads 2-Cn by an electron-withdrawing 1,1-dicyanoethylenyl (DCE) unit, leading to the structure of C60(>CPAF-Cn) 1-Cn. The functional change resulted in the increase of molecular electronic polarization and bathochromic shift in the optical absorption from λmax 400 nm for DPAF to 500-550 nm for CPAF. One example was given by C60(>CPAF-C2M) dyad [18].
In the case of C60(>DPAF-C9)x (x = 1, 2, and 4), their 2PA σ2 values were found to be concentration-dependent [8] with a higher magnitude at a lower concentration than 10 −3 M. This revealed that a minimization of the molecular aggregation should be advantageous to prevent the loss of σ2 magnitude, especially at a high concentration for practical use. It is crucial since a

Materials Characterization
We have demonstrated the use of dialkyldiphenylaminofluorenyl-keto-[60]fullerene C 60 (>DPAF-C n ) dyads 2-C n [16], branched triads C 60 (>DPAF-C n ) x (x = 2) [17], and the related starburst pentads (x = 4) [8] for the study of simultaneous 2PA phenomena under the photoexcitation of a 780-nm laser light in the fs region. In these compounds, DPAF-C n was used as the light-harvesting antenna chromophore to compensate for the low optical absorption of the fullerene cage at wavelengths beyond 350 nm. It is also functioning as an electron donor to provide one electron capable of being transferred intramolecularly to the C 60 > cage moiety upon 2PA activation that increased largely the excited state absorptions leading to 2PA cross-section enhancement. As a general strategy to extend the active 2PA wavelength to the longer NIR region, we modified the bridging keto group in dyads 2-C n by an electron-withdrawing 1,1-dicyanoethylenyl (DCE) unit, leading to the structure of C 60 (>CPAF-C n ) 1-C n . The functional change resulted in the increase of molecular electronic polarization and bathochromic shift in the optical absorption from λ max 400 nm for DPAF to 500-550 nm for CPAF. One example was given by C 60 (>CPAF-C 2M ) dyad [18].
In the case of C 60 (>DPAF-C 9 ) x (x = 1, 2, and 4), their 2PA σ 2 values were found to be concentration-dependent [8] with a higher magnitude at a lower concentration than 10´3 M. This revealed that a minimization of the molecular aggregation should be advantageous to prevent the loss of σ 2 magnitude, especially at a high concentration for practical use. It is crucial since a sufficiently high 2PA material concentration may be required for the fabrication of NLO devices to bring in significant effects. Our logic approach is to modulate the compound's solubility by the variation of attached alkyl chain length and shape (linearly or sterically hindered branched structures) and to control the effective average separation distance among C 60 > cages when applied in highly concentrated solutions or solid films. Accordingly, we synthesized four samples, namely, C 60 (>CPAF-C n ) 1-C n (n = 4, 9, 12, and 18, Scheme 1) for the evaluation of their alkyl chain-dependent broadband 2PA characteristics. Since the 2PA activity of C 60 (>CPAF-C n ) analogous moiety by the 980-nm excitation in toluene was reported recently using examples of C 60 (>CPAF-C 9 ) and hybrid C 60 (>DPAF-C 18 ) (>CPAF-C 2M ) n (n = 1 or 2) [15], we investigated the σ 2 value and the corresponding nonlinear absorption efficiency of the compound 1-C n under the excitation wavelength of 780 nm to substantiate their broadband two-photon absorbing properties. sufficiently high 2PA material concentration may be required for the fabrication of NLO devices to bring in significant effects. Our logic approach is to modulate the compound's solubility by the variation of attached alkyl chain length and shape (linearly or sterically hindered branched structures) and to control the effective average separation distance among C60> cages when applied in highly concentrated solutions or solid films. Accordingly, we synthesized four samples, namely, C60(>CPAF-Cn) 1-Cn (n = 4, 9, 12, and 18, Scheme 1) for the evaluation of their alkyl chain-dependent broadband 2PA characteristics. Since the 2PA activity of C60(>CPAF-Cn) analogous moiety by the 980-nm excitation in toluene was reported recently using examples of C60(>CPAF-C9) and hybrid C60(>DPAF-C18) (>CPAF-C2M)n (n = 1 or 2) [15], we investigated the σ2 value and the corresponding nonlinear absorption efficiency of the compound 1-Cn under the excitation wavelength of 780 nm to substantiate their broadband two-photon absorbing properties. Synthesis of the compound C60(>CPAF-C9) 1-C9 followed the procedure described previously [18]. A similar synthetic sequence was applied for the preparation of C60(>CPAF-C4) 1-C4, C60(>CPAF-C12) 1-C12, and C60(>CPAF-C18) 1-C18. Formation of a fullerenyl monoadduct 1-Cn was evident by detection of its infrared spectrum displaying three typical fullerenyl signals at 753, 697, and 527 cm −1 corresponding to absorptions of an unfunctionalized half-cage sphere of C60>. The key chemical modification of a keto group of C60(>DPAF-C18) 2-C18 to a 1,1-dicyanoethylene (DCE) group of 1-C18 was made by using malononitrile as a reagent. Indication of the CPAF moiety attached on a C60> cage was seen clearly by a strong IR absorption band corresponding to cyano (-C≡N) stretching vibrations centered at 2222-2224 cm −1 with complete disappearance of the carbonyl stretching vibration of 2-C18 at 1680 cm −1 . It was also substantiated by its 13 C-NMR spectrum ( Figure 2 Synthesis of the compound C 60 (>CPAF-C 9 ) 1-C 9 followed the procedure described previously [18]. A similar synthetic sequence was applied for the preparation of C 60 (>CPAF-C 4 ) 1-C 4 , C 60 (>CPAF-C 12 ) 1-C 12 , and C 60 (>CPAF-C 18 ) 1-C 18 . Formation of a fullerenyl monoadduct 1-C n was evident by detection of its infrared spectrum displaying three typical fullerenyl signals at 753, 697, and 527 cm´1 corresponding to absorptions of an unfunctionalized half-cage sphere of C 60 >. The key chemical modification of a keto group of C 60 (>DPAF-C 18 ) 2-C 18 to a 1,1-dicyanoethylene (DCE) group of 1-C 18 was made by using malononitrile as a reagent. Indication of the CPAF moiety attached on a C 60 > cage was seen clearly by a strong IR absorption band corresponding to cyano (-C"N) stretching vibrations centered at 2222-2224 cm´1 with complete disappearance of the carbonyl stretching vibration of 2-C 18 at 1680 cm´1. It was also substantiated by its 13  152.18, and 149.82 along with all fullerenyl sp 2 carbon peaks located within δ 134-148, whereas two sp 3 C 60 > carbon (C F1 and C F2 ) peaks were assigned at δ 73.09. Direct confirmation of the molecular mass of 1-C 18 was made by a group of sharp molecular mass ions with the maximum mass intensity centered at m/z 1645 (M + ) and 1646 (MH + ) in its MALDI-TOF mass spectrum ( Figure 3). This was followed by several groups of ion peaks at m/z 1465-1550 with the group mass each separated by a -CH 2 -unit (m/z 14) indicating the consecutive loss of alkyl chain carbons from the M + peak. Full elimination of weaker aliphatic bonds of 1-C 18 led to a stable aromatic mass ion fragment at m/z 1079, matching with the structure assigned in the Figure. Further fragmentation gave stable C 60 + (m/z 720) and C 60 H 2 + (m/z 722) mass ion fragments.

Nonlinear Z-Scan Measurements
The open-aperture Z-scans of four C60(>CPAF-Cn) samples were carried out in THF with femtosecond laser pulses. Z-scan measurements carried out at an ultrafast time scale of 226 femtoseconds should be able to reduce potential accumulative thermal scattering effects, normally occurring at picosecond regions, at the wavelength of either 780 or 1000 nm. The transmittance of all compounds studied were collected, as shown in Figure 4B, indicating a consistent level of ~92.4% at 780 nm. The data reported in Figure 5a

Nonlinear Z-Scan Measurements
The open-aperture Z-scans of four C 60 (>CPAF-C n ) samples were carried out in THF with femtosecond laser pulses. Z-scan measurements carried out at an ultrafast time scale of 226 femtoseconds should be able to reduce potential accumulative thermal scattering effects, normally occurring at picosecond regions, at the wavelength of either 780 or 1000 nm. The transmittance of all compounds studied were collected, as shown in Figure 4B, indicating a consistent level of~92.4% at 780 nm. The data reported in Figure 5a,b were normalized to the linear transmittance for all Z-scans by the correction of the background transmittance, T(|Z| >> Z o ). The normalized transmittance ∆T(Z) was expressed as T(Z)/T(|Z| >> Z o ). Accordingly, the change in the normalized transmittance is indicative of the nonlinear (or light-dependent) part in the compound's absorption. Total absorption was described by the change in the absorption coefficient ∆α = βI, where β and I are the 2PA coefficient and the light intensity, respectively. The absorption coefficient can be extracted from the best fitting between the Z-scan theory [21] and the data. The 2PA cross-section value was then calculated from the coefficient by the formula σ 2 = βhω/N, wherehω is the photon energy and N is the number of the molecules. 2PA coefficient and the light intensity, respectively. The absorption coefficient can be extracted from the best fitting between the Z-scan theory [21] and the data. The 2PA cross-section value was then calculated from the coefficient by the formula σ2 = βħω/N, where ħω is the photon energy and N is the number of the molecules. Open-aperture Z-scans carried out under the irradiance of 220 GW/cm 2 at 780 nm were taken on the samples of 5, 1-C4, 1-C9, 1-C12, and 1-C18 in THF at the concentration of 5.0 × 10 −4 M with the profile plots shown in Figure 5a. These Z-scans displayed positive signs for absorptive nonlinearities with the decrease of light-transmittance in the order of C60(>CPAF-C18) < C60(>CPAF-C9) ≤ C60(>CPAF-C12) < C60(>CPAF-C4) in solution. As a result, the 2PA cross-section values of these compounds measured were summarized in Table 1. Table 1. Two-photon absorption cross sections (σ2) and excited state absorption (ESA) cross-sections (σESA) of CPAF-C12 and C60(>CPAF-Cn) measured using laser pulses working at 780 nm with a 226-fs duration and a repetition rate of 1.0 kHz. The light intensity I was 220 GW/cm 2 . Open-aperture Z-scans carried out under the irradiance of 220 GW/cm 2 at 780 nm were taken on the samples of 5, 1-C 4 , 1-C 9 , 1-C 12 , and 1-C 18 in THF at the concentration of 5.0ˆ10´4 M with the profile plots shown in Figure 5a. These Z-scans displayed positive signs for absorptive nonlinearities with the decrease of light-transmittance in the order of C 60 (>CPAF-C 18 ) < C 60 (>CPAF-C 9 ) ď C 60 (>CPAF-C 12 ) < C 60 (>CPAF-C 4 ) in solution. As a result, the 2PA cross-section values of these compounds measured were summarized in Table 1. Table 1. Two-photon absorption cross sections (σ 2 ) and excited state absorption (ESA) cross-sections (σ ESA ) of CPAF-C 12 and C 60 (>CPAF-C n ) measured using laser pulses working at 780 nm with a 226-fs duration and a repetition rate of 1.0 kHz. The light intensity I was 220 GW/cm 2 . It is interesting to observe a higher 2PA absorption cross-section value of 6.42ˆ10´4 8 cm 4¨sp hoton´1¨molecule´1 (or 642 GM) for C 60 (>CPAF-C 18 ) at a low concentration of 5.0ˆ10´4 M than that, 3.25ˆ10´4 8 cm 4¨s¨p hoton´1¨molecule´1 (or 325 GM), for C 60 (>CPAF-C 9 ) at the same concentration. A lower value of 1.28ˆ10´4 8 cm 4¨s¨p hoton´1¨molecule´1 (or 128 GM) for C 60 (>CPAF-C 4 ) than the un-fullerenized CPAF-C 12 (220 GM) was detected, perhaps owing to its higher particle aggregation tendency even at 10´4 M. Based on a 226-fs pulse duration is slightly longer than 130 fs required for the intramolecular energy-transfer from the photoexcited 1 (CPAF)*-C n antenna moiety to the C 60 > cage of C 60 (>CPAF-C n ). Completion of this energy-transfer event at the early time scale leads to the formation of excited 1 C 60 *(>CPAF-C n ) state. Therefore, the measured σ 2 values at 226-fs should cover partly two-photon absorptions of both CPAF-C n and C 60 > moieties in the fs region and the excited singlet state absorption (S 1 -S n ) of 1 (C 60 >)* cage moiety in subsequent subpicoseconds. The initial 2PA excitation process at 780 nm represents mainly the contribution of CPAF-C n moiety forming the transient C 60 (> 1 CPAF*-C n ) state. The argument is valid due to the fact of low linear and nonlinear C 60 > cage absorption at this wavelength as compared with the later of CPAF-C n moiety. In addition, the occurrence of transient conversion from 1 C 60 *(>DPAF-C 9 ) state to the corresponding 3 C 60 *(>DPAF-C 9 ) state via inter-system crossing was reported to be effective at a much longer time scale of~1.4 ns [17]. Therefore, the absorption contribution of 3 C 60 *(>CPAF-C 9 ) state can be excluded in this measurement. These nonlinear fs absorptions may be correlated to the following nonlinear absorption measurements.

Sample
We also investigated the intensity-dependent (70-420 GWcm´2) Z-scans using the compound 1-C 18 as an example in THF at a concentration of 2.0ˆ10´3 M by 780-nm excitation. The resulting data profiles were displayed in Figure 5b with the corresponding 2PA cross-sections plotted in Figure 6a (red triangle). At this concentration, the σ 2 values were higher, in general, and increased more rapidly than those taken at a higher concentration of 1.0ˆ10´2 M (Figure 6a, blue circle) at the same laser intensity. This trend of concentration-dependent σ 2 values having a higher quantity at a lower concentration consistent with that reported recently [8]. The intensity dependence on σ 2 values may also reveal higher order absorptions, such as excited state absorption (ESA) that can be effectively treated as three-photon absorption (3PA) for ESA, possibly taken place. In order to distinguish the contribution of 2PA from the higher order nonlinear absorption, the ln(1-T) vs. intensity (I) relationship was plotted, as shown in Figure 6b,c. These Z-scan curves were fitted with 2PA when the slope is~1.0 and fitted with ESA/3PA when the slope is~2.0 [22]. The fitting results confirmed that, at a low laser intensity of 74 GWcm´2 (Figure 6b), the event of 2PA process dominates, while at a high intensity of 400 GWcm´2 (Figure 6c), the photophysical processes of ESA/3PA became the major occurrence. Accordingly, the ESA cross-sections (σ ESA ) of 5, 1-C 4 , 1-C 9 , 1-C 12 , and 1-C 18 were determined and given in Table 1. They showed the similar trend of concentration dependence in magnitude to those of σ 2 (Figure 7a) having a higher value at a lower concentration. The trend was also coupled with their solubility where 1-C 18 with two linear octadecyl chains and 1-C 9 with two branched 3,5,5-trimethylhexyl chains exhibit better solubility in solvents and a higher magnitude of σ 2 . We have examined several solvents including CS 2 , THF, and toluene using C 60 (>CPAF-C 18 ) as the example under 780-nm excitation and found no significant difference in the value of σ 2 indicating no solvent effect for the case.
Nonlinear absorption properties of C 60 (>CPAF-C n ) were investigated by irradiance-dependent transmission measurements at the wavelength of 780 nm using the same setup as those applied in 2PA cross-section measurements conducted by fs-laser pulses. Nonlinear absorption of CPAF-C 12 , C 60 (>CPAF-C 4 ), C 60 (>CPAF-C 9 ), C 60 (>CPAF-C 12 ), and C 60 (>CPAF-C 18 ) in THF measured as a function of irradiance with 226-fs laser pulses operated at 780 nm were illustrated in Figure 7b. All the samples showed a linear transmission (T =~90%) with input intensity of up to 30 GW/cm 2 . When the incident intensity was increased above 70 GW/cm 2 , the transmittance (%) began to deviate from the linear transmission line and decrease indicating the initiation of nonlinear absorption. A systematic trend showing higher nonlinear absorption efficiency down to 50%, 57%, 60%, and 65% for the dyads 1-C 18 , 1-C 9 , 1-C 12 , and 1-C 4 , respectively, was observed with the increase of irradiance intensity up to

Spectroscopic Measurements
1 H-NMR and 13 C-NMR spectra were recorded on either a Bruker Avance  or Bruker AC-300 spectrometer (Bruker, Billerica, MA, USA). UV-vis spectra were recorded on a Hitachi U-3410 UV spectrometer (Hitachi, Chiyoda, Tokyo, Japan). Infrared spectra were recorded as KBr pellets on a Nicolet 750 series FT-IR spectrometer (Thermo Scientific Nicolet, Waltham, MA, USA). Mass spectroscopic measurements were performed by the use of positive ion matrix-assisted laser desorption ionization (MALDI-TOF) technique on a micromass M@LDI-LR mass spectrometer (Micromass, Cary, NC, USA). The matrix of 3,5-dimethoxy-4-hydroxycinnamic acid (sinapic acid) was used.