PAPER

www.rsc.org/materials | Journal of Materials Chemistry

Efficiency of singlet oxygen production from self-assembled nanospheres of molecular micelle-like photosensitizers FC4S Chi Yu,a Taizoon Canteenwala,b Mohamed E. El-Khouly,c Yasuyuki Araki,c Kenneth Pritzker,a Osamu Ito,c Brian C. Wilsond and Long Y. Chiang*b Received 10th January 2005, Accepted 17th March 2005 First published as an Advance Article on the web 29th March 2005 DOI: 10.1039/b500369e Direct detection of singlet oxygen (1O2) production under irradiation of molecular micelle-like hexa(sulfo-n-butyl)[60]fullerene (FC4S) self-assembled nanospheres at 500–600 nm was obtained by the measurement of its near-infrared luminescence at 1270 nm. This photocatalytic effect makes FC4S a potential alternative sensitizer to TiO2 and feasible for use in the visible region in addition to its intrinsic high UV efficiency. Despite having a relatively low optical absorption of FC4S at 600 nm, appreciable 1O2 signal was detected comparable to that of hematoporphyrin derivatives Photofrin at the same molar concentration, but less than sulfonated aluminium phthalocyanine, AlS4Pc. The quantum yield of FC4S for the generation of 1O2 in H2O was roughly estimated to be 0.36 using the relative correlation to that of C60/c-CD. The absolute value is not available. These results demonstrated efficient triplet energy transfer from 3FC4S* to molecular oxygen in the nanosphere structure. We also confirmed certain retention between photocatalytic characteristics of underivatized C60 with that of FC4S, a C60 hexaadduct containing a single covalent bond between each addend and the fullerene cage, in contrast to other Bingel-type hexamalonate adducts of C60 in the literature with a low 1O2 yield.

Introduction High photostability and oxidative power makes titanium dioxide (TiO2) a good photocatalytic semiconductor for use as a photodynamic agent in destroying cancer cells and viruses.1 However, its relatively large band gap of 3.2 eV for the anatase form makes TiO2 photoactive only at light wavelengths shorter than 380 nm which utilizes less than 3% of the solar irradiance.2 Photoexcitation of C60 and fullerene derivatives induces a singlet fullerenyl excited state that is transformed to the corresponding triplet excited state, via intersystem energy crossing, with nearly quantitative efficiency.3,4 Subsequent energy transfer from the triplet fullerene derivatives to molecular oxygen produces singlet molecular oxygen in aerobic media. This photocatalytic effect becomes one of the mechanisms in photodynamic treatments using fullerene derivatives as photosensitizers,5–8 complementary to TiO2. However, a high degree of functionalizaton on C60 for enhancement of solubility and compatibility results in a progressive decrease of the singlet oxygen production quantum yield [W(1O2)]. Examples were given by Bingel-type malonic acid, C60[C(COOH)2]n, and malonic ester, C60[C(COOEt)2]n, [60]fullerene adducts,9 showing a decreasing trend of W(1O2) as the number of addends (n) increases. When the number n reaches six for a hexaadduct, its W(1O2) value declines to only 13% or less of that for C60.10 Our unexpected observation of relatively high singlet oxygen production quantum yield of hexa(sulfo-n-butyl)[60]fullerenes (FC4S), as a fullerene hexaadduct, revealed its unique electronic features in contrast *[email protected]

This journal is ß The Royal Society of Chemistry 2005

to other malonic hexaadducts of C60 synthesized by Bingel reaction. Fullerene molecules are highly hydrophobic. It can be dispersed into an aqueous medium in a micelle form with the application of surfactants. For increasing the structural stability in H2O, we made direct chemical conversion of C60 into its water-soluble derivatives FC4S. The synthesis involves recently developed hexaanionic C60 (C6026) chemistry for attaching six surfactant-like sulfobutyl arms on C60 forming a molecular micelle-like structure.11,12 This facilitates the utility of fullerenes in vitro showing a free radical scavenging effect in an aqueous medium. Accordingly, FC4S self-assembled nanospheres exhibit biological activities for several disease-related treatments.13 The use of FC4S nanospheres as photodynamic therapeutic (PDT) photosensitizers was reported by in vitro and in vivo antitumoral activities.8 The observed PDT efficacy was proposed in correlation with the presence of several reactive oxygen species (ROS) generated photochemically. One of these, superoxide radical (O22?), was detected in media containing FC4S and ferricytochrome (Fe+3) under photoexcitation and confirmed by trapping of it by superoxide dismutase (SOD) enzyme. The production of O22? in media increases the optical absorption of ferrocytochrome (Fe+2) at 550 nm. The method can be used for indication of O22? generation. An electron-transfer mechanism involving the excited state of FC4S, molecular oxygen, and donor residues of the cytochrome protein was proposed for the presence of O22?.8 Here, we report direct evidence of efficient singlet oxygen (1O2) production under irradiation of the FC4S selfassembled nanospheres at 500–600 nm by the detection of 1 O2 emission at 1270 nm. The observation allows utilization of FC4S to harvest light energy extending from UV to J. Mater. Chem., 2005, 15, 1857–1864 | 1857

saline (PBS), distilled water, methanol, dimethyl sulfoxide, and dimethylformamide. Detection of singlet oxygen Fig. 1 Chemical composition of hexa(sulfo-n-butyl)[60]fullerene 1 and protonated FC4S 2.

the visible wavelength region for performing photocatalytic effects.

Materials and methods General Photofrin was obtained from QLT Phototherapeutics, Vancouver, Canada. Tetrasulfonated aluminium phthalocyanine (AlS4Pc) was obtained from Porphyrin Products, Utah, USA. Phosphate buffered saline (PBS) was purchased from Sigma, St. Louis. All solvents including methanol, dimethyl sulfoxide (DMSO), and N,N-dimethylformamide (DMF) were purchased from Merck, Darmstadt, Germany. Synthesis of hexa(sulfo-n-butyl)[60]fullerene (FC4S, 1) A direct one-pot method was used for the preparation of micelle-like fullerene derivatives. Thus, water-soluble hexa(sulfo-n-butyl)[60]fullerene, C60(CH2CH2CH2CH2SO3Na)6 as shown in Fig. 1, was synthesized at 80–85% yield by treating C60 in dimethoxyethane (DME) with sodium naphthalenide (10.0 equiv.) at 25 uC, followed by reacting the resulting hexaanionic fullerene intermediates, C6026, with an excess of 1,4-butane sultone (15.0 equiv.).11 After purification by filtration and repeated reprecipitation in MeOH from an aqueous solution, the FC4S product gave a single HPLC peak using a reverse-phase C-18 column eluted with H2O. Acidification of FC4S with HCl (4 N) gave the corresponding hexasulfonic acid (FC4SH, 2). The solution of sodium naphthalenide in DME was titrated by succinic acid prior to use. The hexaanionic C60 generation reaction was carried out inside a glove-box conditioned to ,50 ppm O2 to minimize partial oxidation or oligomerization of fullerenic anions by dissolved O2. A slight excess of sodium naphthalide was found to be necessary to achieve the generation of hexaanionic fullerene intermediates. Similar reaction conditions were also carried out using a hindered halide instead of 1,4-butane sultone in reaction with C6026 to yield a highly symmetrical hexaadduct of C60 with each addend linked on the fullerene cage by a single covalent bond.{12 Optical absorption measurements Optical absorption spectra (250–850 nm) were determined in a quartz cuvette with a UV–vis spectrophotometer (Model UV160U, Shimadzu, Kyoto, Japan), with the photosensitizers dissolved in different solvents, including phosphate buffered { The structure of one hindered product, denoted emerald green fullerene EF-6MC4, was identified by the X-ray single crystal structural analysis and 13C NMR spectroscopic methods to reveal a highly symmetrical structure and a single covalent bond between each addend arm and the connecting carbon of the C60 cage.

1858 | J. Mater. Chem., 2005, 15, 1857–1864

Direct measurement of singlet oxygen was carried out by the detection of its luminescence emission at 1270 nm, 14 corresponding to 1Dg A 3S2 The measurement g transition. was accomplished using a photon multiplier tube (PMT) detector (Model R5509-42, Hamamatsu Corp., Bridgewater, NJ) with high near-IR sensitivity. The laser instrumentation and technique were described in detail elsewhere.15 Concisely, four bandpass filters (1200, 1270, 1300, and 1330 nm) placed sequentially in front of the detector were used for measurements of the luminescence spectrum. The PMT output was amplified and converted to a voltage pulse using a high-speed current preamplifier (Model SR445, Stanford Research Systems, Sunnyvale, CA). A multichannel scaler (Model SR430, Stanford) was applied for time-resolved single photon counting, giving a typical temporal resolution of 80 ns. Samples of photosensitizer (10 mM) were irradiated by a tunable pulsed laser system (OPO Rainbow 355, OPOTEK Inc., Carlsbad, CA), with an irradiation spot size of 3.0 mm in diameter. The pulse energy applied at the sample was approximately 1.0 mJ. Emission from the sample at each detection wavelength was recorded as total photon counts over 2400 laser pulses. Time-resolved measurements were made using a laser with 20 ns pulse length and 100 ms pulse interval. Transient absorption measurements Collection and analysis of nanosecond transient absorption spectra in the near-IR region (600–1600 nm) were made using third-harmonic generation (HG, 355 nm) of a Nd–YAG laser (Spectra-Physics, Quanta-Ray GCR-130, fwhm 6 ns) as an excitation light source. Light intensity from a pulsed Xe lamp was monitored and detected with a Ge-APD device (Hamamatsu Photonics, B2834). For measurement of the spectrum in the visible region of 400–1000 nm, a Si-PIN photodiode (Hamamatsu Photonics, S1722-02) was used as the detector. Triplet lifetime was estimated by using a photomultiplier (Hamamatsu Photonics). All the samples in a quartz cell (1.0 6 1.0 cm) were deaerated by bubbling Ar through the solution for 15 min.

Results and discussion Chemical composition of FC4S was determined by various spectroscopic methods. It was confirmed by elemental analyses giving a hydrated hexasulfobutylated fullerene sodium salt product with the weight ratio of each element summarized as follows. Calculated for C84H88O38S6Na6 as C60(CH2CH2CH2CH2SO3Na)6?20H2O: C, 50.45; H, 4.20; O, 28.82; S, 9.60; Na, 6.91, Found: C, 49.56; H, 4.33; S, 9.44; Na, 6.78, where analysis of sodium was carried out on a Varian Spectro AA-100 plus analyzer. Direct mass spectroscopic (MS) measurement of watersoluble fullerene derivatives was difficult in general with an extremely low intensity of the molecular mass ion peak obtainable in most experiments. The use of non-volatile polar This journal is ß The Royal Society of Chemistry 2005

Molecular self-assembly of FC4S for nanosphere formation in water

Fig. 2 Negative ion desorption chemical ionization (DCI2) mass spectrum of FC4S showing a series of fragmentation groups with up to six consecutive weight increases of m/z 136, the mass of a sulfo-n-butyl group, confirming the molecular composition of 1 and 2.

solvent was also problematic due to slow and inefficient solvent removal after deposition on the probe prior to the MS measurement. After evaluating various MS detection methods, we were able to obtain adequate results by desorption chemical ionization (DCI) and matrix-assisted laser desorption ionization (MALDI). The negative ion DCI mass spectrum (DCI2MS) of FC4S, as shown in Fig. 2, displayed a maximum relative intensity at a peak of the C60 ion fragmentation (m/z 720). This was followed by groups of peaks at m/z 776, 832, 887 and 944, having a consecutive weight increase of 56 mass units that matches the gain of a butyl group to the preceding ion fragment, corresponding to C60Bu1, C60Bu2, C60Bu3, and C60Bu4, respectively. In the higher mass region, a series of fragmentation groups was observed, with a consecutive weight increase of 136 mass units, corresponding to a sulfobutyl group (CH2CH2CH2CH2–SO3). The ion fragment masses agree with those of di-, tri-, tetra-, penta- and hexasulfobutylated fullerenes. In the case of the negative-ion MALDI mass spectrum (MALDI2-MS) of FC4SH, a molecular ion at m/z 1542 was detected that is conceivably C60(CH2CH2CH2CH2–SO3H)6. It was followed by two lower mass groups at m/z 1405 and 1269, corresponding to the ion mass of C60(CH2CH2CH2CH2–SO3H)5 and C60(CH2CH2CH2CH2–SO3H)4, respectively. These data provide evidence of the chemical compositions of hexa(sulfo-n-butylated) C60 compounds 1 and 2. In addition, conversion of sulfonic moieties (–SO2–ONa) of 1 to the corresponding hydrophobic diethylsulfamide (–SO2–NEt2) derivatives allowed the detection of a peak group with the maximum intensity located at m/z 1873 (negative ion DCI-MS) corresponding to the molecular ion mass of hexa(diethylaminosulfo-n-butyl)[60]fullerene. That gave further indirect confirmation of the FC4S composition. Recent success in obtaining the single-crystal molecular structure of fully symmetrical [6,6] hexasubstituted C60 derivative, synthesized from the same C6026 chemistry under similar reaction conditions, indirectly suggested a relatively even distribution of sulfo-n-butyl arms on the C60 cage.{12 This journal is ß The Royal Society of Chemistry 2005

Aggregation behaviours of highly water-soluble micelle-like hexa(sulfo-n-butyl)fullerenes were investigated using small angle neutron scattering (SANS) in D2O and small angle X-ray scattering (SAXS) in H2O.16 Molecular self-assemblies of FC4S resulted in formation of nearly monodisperse spheroidal nanospheres with the sphere radius of gyration Rg ˚ , where the major axis y29 A ˚ and the minor axis y 19 A ˚ y21 A for the ellipsoid-like aggregates, or an estimated long ˚ [the radius 5 (5/3)1/2Rg] for the sphere diameter of 60 A aggregates. Interestingly, this radius of gyration was found to remain relatively constant over a wide concentration range from 0.35 mM to 26 mM in aqueous solutions. The mean number of FC4S molecules for the nanosphere was determined to be 6.5 ¡ 0.7.16 From these data we propose that a strong hydrophobic interaction between core fullerene cages overcomes loose charge repulsion at the molecular surface of the micelle-like FC4S structure. It allows the nanosphere formation at a low concentration in spite of steric hindrance and high hydrophilicity arising from six sulfo-n-butyl arms surrounding C60. Based on the mean FC4S molecule number of slightly more than six for each small uniform coalescent nanosphere and the ˚, fully stretched molecular length of FC4S being roughly 15 A the dimension of the nanosphere fits well with six cage molecules packed in close vicinity, perhaps each at the vertex of an octahedron shape, forming a hydrophobic core region, as shown in Fig. 3. High hydrophobicity at the central core results in the bending of butylsulfonate arms outward into the aqueous phase. The hydrodynamic volume revealed a large number of H2O molecules trapped inside the nanosphere. This suggests possible formation of small hydrophilic domains surrounded by a main hydrophobic spherical structure, with the C60 cages composing the latter. Accordingly, a densely-packed nanosphere may still allow high singlet oxygen generation by the fullerene cage. A higher percentage of fullerene cage moieties located at the interfacial region between the sphere and the aqueous medium at both internal and external regions are susceptible to photoexcitation and the photodynamic effect. Solvation effects on FC4S may play a role in its aggregation

Fig. 3 Schematic representation of a FC4S-derived nanosphere in aqueous solution based on the aggregation size determined by SANS and SAXS.

J. Mater. Chem., 2005, 15, 1857–1864 | 1859

behaviour and make the concentration.

1

O2 generation non-linear with

Singlet oxygen generation by FC4S self-assembled nanospheres in solution under laser irradiation Hexafunctionalization of the fullerene cage changes the local conjugation structure from that of C60 and thus modifies the corresponding efficacy of singlet oxygen generation, which is a function of several main factors, including the excited tripletstate FC4S (3FC4S*) quantum yield, the intermolecular energy transfer efficiency between 3FC4S* and triplet molecular oxygen, the stability and lifetime of 3FC4S* in various solvents, and the competitive thermal decay. Direct measurement of singlet oxygen produced in situ in the system is the most reliable method for characterizing this energy transfer process. Apart from chemical considerations, physical aggregation of FC4S molecules and the 1O2 lifetime in different solvents are dominant factors in the overall luminescence signal intensity measured. The fullerene cage of FC4S tends to aggregate into solubilized nanospheres in a solvent-dependent manner as a function of solubility. Respectively, UV–vis absorption spectrum of FC4S in dilute solutions (10 mM) displayed an overall higher extinction coefficient of FC4S in H2O than in MeOH (Fig. 4) at the same concentration. The relative peak intensities correlated well with the solubility of 1 in H2O (high) and MeOH (low), with those in DMSO and DMF being intermediate, possibly due to the solvent dependence of the molecular nanosphere size. The main optical absorption peak is located below 300 nm with a shoulder around 340 nm. The spectrum profile decreases monotonically up to 650 nm. Better molecular dispersion of 1 in H2O increased its 350–550 nm optical absorption. In all solvents, the extended absorption at all wavelengths of 250– 650 nm reveals the existence of fullerenyl olefin conjugation moieties of various lengths on the C60 cage. Variation of the olefinic conjugation length, each with a slightly different energy environment and a different band gap in a number of resonance forms, leads to effective broadening of the

Fig. 4 Steady state UV–visible absorption spectra of FC4S (10 mM) in (a) H2O, (b) MeOH, (c) DMSO and (d) DMF.

1860 | J. Mater. Chem., 2005, 15, 1857–1864

Table 1 The molar extinction coefficient of FC4S (10 mM) in different solvents at 430, 512 and 600 nm Solvent

e430/M21cm21

e512/M21cm21

e600/M21cm21

PBS H2O MeOH DMSO DMF

13 160 14 470 7930 8870 7910

5720 5990 3460 3300 3060

2000 1950 1230 920 1060

corresponding energy absorption band. A systematic decrease of the signal intensity at longer wavelengths revealed fewer domains with a long conjugation length. Table 1 summarizes the relative absorption extinction coefficients (e) of FC4S in various solvents at three wavelengths 430, 512, and 600 nm, applied in photoexcitation studies below. These coefficients were also used in the calculation of relative 1O2 production efficiency. To correlate the singlet oxygen production effectiveness to the photodynamic tumour treatment,8 a similar concentration of 10 mM was applied in the study. The time-resolved luminescence emission from the solution of 1 at 1270 nm upon the laser excitation at either 512 or 600 nm is shown in Fig. 5. Even though the optical absorption of FC4S is not high at these wavelengths, the 1O2 luminescence signal was easily measurable, indicating an overall high quantum yield and triplet energy-transfer efficiency of excited FC4S (3FC4S*) to molecular oxygen. Marked differences were seen in these timedependent 1O2 luminescence data between the solvents, showing decrease in the integrated signal strength in the order of DMF . DMSO & MeOH . H2O y PBS. In any given solvent, the peak intensity was roughly 3.6 times higher with 512 nm than with 600 nm excitation, after normalization for the laser output energy at these two wavelengths. This correlated well with the corresponding higher (estimated 3.5 times in average) extinction coefficient of FC4S at 512 nm (Table 1). For example, the relative time-integrated luminescence signals generated by FC4S solution (10 mM) in DMSO under 512, 600, and 430 nm excitation were found to be 10.2, 2.5, and 1.0, respectively. The origin of the much lower normalized luminescence intensity at 430 nm excitation is not clear. The measured 1O2 signal intensity is a function of the quantum yield of FC4S in a nanosphere structure in solution, the lifetime of singlet oxygen, and the lifetime of excited FC4S transient triplet state. The method of curve fitting and the calculation of 3FC4S* and singlet oxygen lifetimes were described in detail elsewhere.15 Based on curve-fitting of the singlet oxygen peak profiles in Fig. 5 and the reported O2 concentration in air-saturated solution, the 1O2 lifetime was calculated to be 18.9, 5.7, and 9.7 ms in DMF, DMSO, and MeOH, respectively (Table 2). These are comparable to published values of 18–25 ms in DMF17 and 9.8 ms in MeOH,18 but lower than 30 ms reported for DMSO.19 The low signal in H2O (Figs. 5Ad and 5Bd) and PBS (Figs. 5Ae and 5Be) limited accurate lifetime calculation. However, the value was roughly 2–3 ms in water, which is similar to the published value of 3.2 ms.18 Lifetime of the triplet-state of FC4S was calculated as 0.7–0.9, 1.2–1.4 and 0.5–0.7 ms in DMF, DMSO and MeOH, respectively. These This journal is ß The Royal Society of Chemistry 2005

600 nm excitation, respectively, in DMSO. As seen in Fig. 4, the optical absorption of FC4S decreases monotonically from 400 to .700 nm (relative extinction coefficient ratio 5 8.0 and 9.6 for e430/e600 at 430 nm and 3.0 and 3.5 for e512/e600 at 512 nm in DMF and DMSO, respectively). In principle, the higher absorption and excitation energy at 430 nm should lead to a higher production of 1O2 after normalization to the laser output power. However, this is not the case, unlike the results at 512 and 600 nm excitation, which showed good correlation with the relative absorption extinction coefficients. Regardless of this difference, the observed overall data revealed good quantum yields of FC4S under excitation over a range of wavelengths. Characterization of energy-transfer from triplet 3FC4S* state to molecular oxygen for 1O2 generation by nanosecond transient absorption measurements

Fig. 5 Time-resolved 1O2 luminescence emission at 1270 nm produced upon laser irradiation of FC4S solution (10 mM) at (A) 512 nm and (B) 600 nm in different solvents: (a) DMF, (b) DMSO, (c) MeOH, (d) H2O, and (e) PBS in the air-saturated solution. Table 2 The lifetime of 3FC4S* and singlet oxygen generated from FC4S solution (10 mM) upon photoexcitation at (a) 600 nm and (b) 512 nm in different solvents Solvent

Lifetime of 3FC4S*/ms

(a), DMF (a), DMSO (a), MeOH (b), DMF (b), DMSO (b), MeOH

0.7 1.2 0.7 0.9 1.4 0.5

¡ ¡ ¡ ¡ ¡ ¡

0.1a 0.1 0.1 0.1 0.1 0.0

Lifetime of 1O2/ms 17.9 5.2 9.2 18.9 5.7 9.7

¡ ¡ ¡ ¡ ¡ ¡

0.2a 0.1 0.1 0.2 0.3 0.0

The transient absorption spectrum observed under nanosecond laser (355 nm, 6.0 ns fwhm) excitation of FC4S in deaerated DMSO is shown in Fig. 6. The experiment was carried out at a concentration (0.1 mM) close to that used in the SAXS and SANS measurements for the nanosphere characterization. A clear new broad transient absorption band with a peak maximum centered at roughly 700 nm was detected. The absorption intensity of this transient band was quenched by the well-known triplet quenchers, such as O2, via triplet energy transfer in O2-saturated DMSO (inserted time profile (c) of Fig. 6). That allowed us to assign it to the triplettriplet (T–T) transition of FC4S. Although the peak position of this T–T absorption band is almost the same as those (700–740 nm) of pristine C60 and other fullerene derivatives containing weakly influencing functional groups, the band width is quite broad, probably due to the presence of a number of regio-isomers. The quantum yield W(1O2) of FC4S in the generation of 1O2 was estimated by the direct comparison of its luminescence

a

Data 5 mean ¡ SD (n 5 3), calculated based on the reported method.18

short lifetimes imply a high energy transfer rate from 3FC4S* to 3O2. It is noted that the maximum 1O2 luminescence emission decreased in the order DMF (11) . DMSO (7.5) & MeOH (2.5) . H2O y PBS (1.0), as shown in Fig. 5. Hence, the lowest emission was in water, despite its having the highest FC4S solubility. Although aggregation characteristics in DMF, DMSO and MeOH were not determined here, the solventdependent luminescence signal seems to depend more on the lifetime of 1O2. Taking into account the 1O2 lifetime calculated in Table 2, and the reported value18 in H2O, the relative 1O2 concentration corresponds roughly to the relative lifetime ratios, except for DMSO. In any given solvent, the 1O2 luminescence signal can be approximately proportional to the product of laser pulse energy and optical absorption extinction coefficient. We observed the 1O2 luminescence signal of y10 and 3.6 times higher at 512 nm excitation than at 430 and This journal is ß The Royal Society of Chemistry 2005

Fig. 6 Transient absorption spectra observed by 355 nm laser excitation of FC4S solution (1.0 6 1024 M) in Ar-saturated DMSO. Insert: decay time profiles of triplet 3FC4S* at 700 nm in (a) Ar, (b) air, and (c) O2-saturated DMSO.

J. Mater. Chem., 2005, 15, 1857–1864 | 1861

emission characteristics at 1270 nm with that produced from c-cyclodextrin encapsulated C60 (C60/c-CD) in water.20 As a result, the emission intensity produced from 3FC4S* was found to be roughly a half in value indicating an approximately 50% efficacy of 3FC4S* formation from the corresponding singlet excited state of FC4S. This conclusion is based on the arguments of (a) the intersystem crossing process taking place in a nearly quantitative yield for C60/c-CD and (b) the triplet energy-transfer rates from 3C60/c-CD* and 3FC4S* to the molecular oxygen being the same. Accordingly, on the basis of W(1O2) being 0.98 for C60 in organic solvent,21 the quantum yield of FC4S in the generation of 1O2 in H2O was roughly estimated to be 0.36 using this relative correlation method to that of C60/c-CD.20 The absolute value is not available in the present study. Comparison of decay-time profiles (the inset in Fig. 6) of transient absorption band at 700 nm in Ar-, air- and O2saturated DMSO solutions indicated effective quenching of 3 FC4S* by O2. Based on the pseudo-first-order plots, the triplet-quenching rate constant by O2 [kq(O2)] was calculated to be 1.68 6 109 M21 s21 in DMSO. The value is in a similar range to the reported kq(O2) value of the pristine C60 in benzonitrile [kq(O2) 5 1.4 6 109 M21 s21].22 These kq(O2) values are slightly smaller than the diffusion controlled limit of 5–6 6 109 M21 s21 in DMSO and benzonitrile. Decay time profiles of the excited triplet 3FC4S* intermediate in O2saturated solutions are shown in Fig. 7. The decay rates increase in the order of H2O (pH 5 7.0 and pH 5 7.8) , DMSO , DMF , MeOH, dependent on the quenching rate constants and O2 concentrations. By employing saturated O2 concentrations reported in the literature, the quenching rate constants kq(O2) were calculated as 2.19, 1.68, 1.05, 1.28, and 1.04 6 109 M21 s21 for DMF, DMSO, MeOH, H2O, and PBS, respectively. The value of these quenching rate constants may inversely correspond to the lifetime of 3FC4S* in Table 2, in the same trend as the solvent dependence. In principle, the energy-transfer rate constant [ken(O2)] can be evaluated by the equation, ken(O2) 5 kq(O2) 6 quantum yield of energy transfer. However, the absolute values of the quantum yields were not obtained in the present study, the rate constants of the energy transfer can not be determined.

Comparison of singlet oxygen production efficiency of FC4S nanospheres with Photofrin and AlS4Pc

Fig. 7 Decay profiles of triplet state of 1 (3FC4S*) at 700 nm observed by 355 nm laser excitation in O2-saturated (a) DMSO, (b) DMF, and (c) MeOH solutions.

Fig. 8 Time-resolved 1O2 luminescence of (A) Photofrin (10 mM), (B) AlS4Pc (2.0 mM), and (C) AlS4Pc (10 mM) at 600 nm excitation in different solvents: (a) H2O, (b) MeOH, (c) DMSO, and (d) DMF.

1862 | J. Mater. Chem., 2005, 15, 1857–1864

Photofrin, as hematoporphyrin derivatives,23 and sulfonated aluminium phthalocyanine (AlS2Pc)24 are effective photodynamic therapeutic agents that exhibit photoinduced cytotoxicity associated with the formation of 1O2 and O22?.25,26 Regardless of the low absorption of FC4S at 600 nm, 1O2 was generated at levels comparable with that of Photofrin, as shown in Fig. 8A. At the same concentration (2.0 or 10 mM), the luminescence signal from FC4S solution is higher in DMF than in DMSO, MeOH and H2O, in agreement with the short lifetime of singlet oxygen in H2O. At a concentration of 10 mM

This journal is ß The Royal Society of Chemistry 2005

Table 3 Relative 1270 nm luminescence intensity (I) of different photosensitizers at a concentration of 10 mM in DMF Dye

e600

1

AlS4Pc Photofrin FC4S FC4S

10 680 1780 1060 3060a

1138 490 389 1150

a

O2 luminescence I (600 (600 (600 (512

nm nm nm nm

ex.) ex.) ex.) ex.)

I/e

I/e (rel.)

0.107 0.275 0.367 0.376

0.29 0.75 1.00 1.02

e512 instead.

for all samples, the luminescence level from FC4S in DMF (Fig. 5Ba) reached more than 79 and 34% of that generated by Photofrin (Fig. 8Ad) and AlS4Pc (Fig. 8Cd), respectively, under photoexcitation at 600 nm. This signal intensity was y80% of that of 2.0 mM AlS4Pc (Fig. 8Bd). In the case of DMSO (Fig. 5Bb), the luminescence level produced from 10 mM FC4S was nearly equal to that of 2.0 mM AlS4Pc (Fig. 8Bc). Furthermore, a lower AlS4Pc concentration (2.0 mM) in different solvents gave reduced luminescence (Fig. 8B), in general non-proportionally to the concentration decrease. Higher luminescence per molecule in dilute solution may be due to the drug being less aggregated. The luminescence intensity at 600 nm of all three photosensitizers in DMF, relative to their optical absorption, is summarized in Table 3. This is then a measure of the singlet oxygen quantum yield relative to that in DMF (assuming that the longer lifetime of 1O2 in DMF should give a more accurate measurement of the 1O2 level). High luminescence, corresponding to the optical absorption, was obtained for FC4S (I/e 5 0.367) and Photofrin (I/e 5 0.275), with a lower correlated value (I/e 5 0.107) for AlS4Pc aggregates in DMF due mainly to high optical absorbance of AlS4Pc in this region. A similar trend was observed in DMSO, showing higher correlated luminescent efficiency for FC4S. A high I/e value obtained for FC4S is apparently due to its low optical absorption intensity (e) in the measured wavelength range. As the low absorption of FC4S at longer wavelengths is still able to produce a comparable signal of 1O2, the photon energy conversion process of triplet AlS4Pc aggregates in the generation of 1O2 may be interpreted as less effective than the triplet fullerene nanospheres in both solvents, on the basis of total light energy involved in the molecular process. In the case of the aqueous solution, the short lifetime of 1O2 and the low normalized luminescence intensity made it difficult to derive a meaningful and accurate peak profile of both FC4S and Photofrin for comparison.

Conclusion Singlet oxygen is widely believed to be the major cytotoxic agent during the PDT processes with many photosensitizers.27 Direct detection of the 1O2 luminescence during irradiation of FC4S self-assembled nanospheres at 500–600 nm has substantiated its photocatalytic property in the visible region. The results also provide evidence of the efficient energy transfer from the excited 3FC4S* to triplet molecular oxygen at a low concentration of 0.1 mM which is close to that used in the SAXS and SANS measurements for the nanosphere characterization. We concluded that functionalization of C60 to its This journal is ß The Royal Society of Chemistry 2005

molecular micelle-like derivative FC4S with six hydrophilic addends, each being bound on the C60 cage by a single covalent bond, does not appear to markedly alter the photoexcitation process, triplet state formation, and energy transfer efficiency of the fullerene cage in the form of a nanosphere. That agrees with a quantum yield of FC4S for the generation of 1O2 in H2O estimated to be roughly 0.36, using the relative correlation to that of C60/c-CD. Despite having relatively low optical absorptions at 600 nm, FC4S nanospheres produce appreciable 1O2 emission signal in solution comparable to that of the hematoporphyrin derivative Photofrin at the same molar concentration. Unlike Photofrin, FC4S is photostable. This photocatalytic effect makes FC4S a potential alternative sensitizer to TiO2 and feasible for use in the visible region in addition to its intrinsic high UV efficiency.

Acknowledgements The authors thank M. Niedre and R. Weersink for their assistance on singlet oxygen measurements. Development of the singlet oxygen technique was supported by the Canadian Cancer Society under a grant from the National Cancer Institute of Canada. Transient absorption measurements were supported by the Ministry of Education, Culture, Sports, Science and Technology, a Grant-in-Aid for the Center of Excellence (COE) Project, Giant Molecules and Complex Systems, 2004. Chi Yu,a Taizoon Canteenwala,b Mohamed E. El-Khouly,c Yasuyuki Araki,c Kenneth Pritzker,a Osamu Ito,c Brian C. Wilsond and Long Y. Chiang*b a Pathology and Laboratory Medicine, Mount Sinai Hospital and University of Toronto, Canada. E-mail: [email protected]; Fax: (416) 586 8589; Tel: (416) 586 4453 b Department of Chemistry, Institute of Nanoscience and Engineering Technology, University of Massachusetts, Lowell, MA, USA 01854. E-mail: [email protected]; Fax: (978) 934 3013; Tel: (978) 934 3663 c Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Katahira, Aoba-ku, Sendai, 980-8577, Japan. E-mail: [email protected]; Fax: (81) 22 617 5608; Tel: (81) 22 617 5608 d Department of Medical Biophysics, Ontario Cancer Institute and University of Toronto, Ontario, Canada. E-mail: [email protected]; Fax: (416) 946 6529; Tel: (416) 946 2952

References 1 D. M. Blake, P. Maness, Z. Huang, E. J. Wolfrum, J. Huang and W. A. Jacoby, Sep. Purif. Methods, 1999, 28, 1. 2 A. Shiga, A. Tsujiko, S. Yae and Y. Nakato, Bull. Chem. Soc. Jpn., 1998, 71, 2119. 3 R. C. Dunn, J. L. Anderson, C. S. Foote and J. D. Simon, J. Am. Chem. Soc., 1993, 115, 5307. For a review: D. M. Guldi and M. Prato, Acc. Chem. Res., 2000, 33, 695. 4 J. L. Anderson, Y. Z. An, Y. Rubin and C. S. Foote, J. Am. Chem. Soc., 1994, 116, 9763; C. Taliani, G. Ruani, R. Zamboni, R. Danieli, S. Rossini, V. N. Denisov, V. M. Buriakov, F. Negri, G. Oriandi and F. Zerbetto, J. Chem. Soc., Chem. Commun., 1993, 220. 5 D. M. Guldi and P. V. Kamat, in Fullerenes, Chemistry, Physics and Technology, ed. K. M. Kadish and R. S. Ruoff, WielyInterscience, New York, 2000, pp. 225–281; M. Maggini and D. M. Guldi, in Molecular and Supramolecular Photochemistry, ed. V. Ramamurthy and K. S. Schanze, Marcel Dekker, New York, 2000, vol. 4, pp. 149–196.

J. Mater. Chem., 2005, 15, 1857–1864 | 1863

6 E. Nakamura and H. Isobe, Acc. Chem. Res., 2004, 36, 807; I. Nakanishi, S. Fukuzuami, T. Konishi, K. Ohkubo, M. Fujitsuka, O. Ito and N. Miyata, J. Phys. Chem. B, 2002, 106, 2372. 7 J. Hirayama, H. Abe, N. Kamo, T. Shinbo, Y. Ohnishi-Yamada, S. Kurosawa, K. Ikebuchi and S. Sekiguchi, Biol. Pharm. Bull., 1999, 22, 1106; J. P. Kamat, T. P. A. Devasagayam, H. Mohan, L. Y. Chiang and J. P. Mittal, Fullerene Sci. Technol., 1998, 6, 663. 8 C. Yu, T. Canteenwala, H. H. C. Chen, B. J. Chen, M. Canteenwala and L. Y. Chiang, Proc.–Electrochem. Soc., 1999, 12, 234. 9 D. M. Guldi and K. D. Asmus, J. Phys. Chem., 1997, 101, 1472. 10 R. V. Bensasson, M. N. Berberan-Santos, M. Bretreich, J. Frederiksen, H. Go¨ttinger, A. Hirsch, E. J. Land, S. Leach, D. J. McGarvey, H. Scho¨nberger and C. Schro¨der, Phys. Chem. Chem. Phys., 2001, 3, 4679. 11 J. B. Bhonsle, C. Yu, J. P. Huang, J. Shiea, B. J. Chen and L. Y. Chiang, Chem. Lett., 1998, 465. 12 T. Canteenwala, P. A. Padmawar and L. Y. Chiang, J. Am. Chem. Soc., 2005, 127, 26. 13 Y. T. Lee, L. Y. Chiang, W. J. Chen and H. C. Hsu, Proc. Soc. Exp. Biol. Med., 2000, 224, 69; H. C. Hsu, L. Y. Chiang, W. J. Chen and Y. T. Lee, J. Cardiovasc. Pharmacol., 2000, 36, 423; S. S. Huang, S. K. Tsai, C. L. Chih, Y. L. Chiang, H. M. Hsieh, C. M. Teng and M. C. Tsai, Free Radical Biol. Med., 2001, 30, 643; S. S. Huang, L. H. Chih, C. H. Lin, L. Y. Chiang, T. Mashino, M. Mochizuki, K. Okuda, T. Hirota and M. C. Tsai, Fullerene Sci. Technol., 2001, 9, 375; Y. L. Lai, W. Y. Chiou, F. J. Lu and L. Y. Chiang, Br. J. Pharmacol., 1999, 126, 778; S. C. Chueh, M. K. Lai, M. S. Lee, L. Y. Chiang, T. I. Ho and S. C. Chen, Transplant Proc., 1999, 31, 1976; M. J. Chen, L. Y. Chiang and Y. L. Lai, Toxicol. Appl. Pharmacol., 2001, 171, 165. 14 A. U. Khan, Chem. Phys. Lett., 1980, 72, 112. 15 M. Niedre, M. S. Patterson and B. C. Wilson, Photochem. Photobiol., 2002, 75, 382.

1864 | J. Mater. Chem., 2005, 15, 1857–1864

16 U. S. Jeng, T. L. Lin, C. S. Tsao, C. H. Lee, T. Canteenwala, L. Y. Wang, L. Y. Chiang and C. C. Han, J. Phys. Chem. B, 1999, 103, 1059; U. Jeng, T.-L. Lin, W.-J. Liu, C.-S. Tsao, T. Canteenwala, L. Y. Chiang, L. P. Sung and C. C. Han, Physica A, 2002, 304, 191. 17 P. R. Ogilby and J. Sanetra, J. Phys. Chem., 1993, 97, 4689; B. Aveline, O. Delgado and D. Brault, J. Chem. Soc., Faraday Trans., 1992, 88, 1971. 18 M. S. Patterson, S. J. Madsen and B. Wilson, J. Photochem. Photobiol. B, 1990, 15, 69. 19 R. Nilsson and D. R. Kearns, J. Phys. Chem., 1974, 78, 1681. 20 T. Konishi, M. Fujitsuka, H. Luo, Y. Araki, O. Ito and L. Y. Chiang, Fullerenes, Nanotubes, Carbon Nanostruct., 2003, 11, 237. 21 J. W. Arbogast, A. P. Darmanyan, C. S. Foote, Y. Rubin, F. N. Diederich, M. M. Alvarez, S. J. Anz and R. L. Whetten, J. Phys. Chem., 1991, 95, 11; M. Terazima, H. Noburo, H. Shinohara and Y. Saito, J. Phys. Chem., 1991, 95, 9080; R. Stackow, G. Schick, T. Jarrosson, Y. Rubin and C. S. Foote, J. Phys. Chem. B, 2000, 104, 7914. 22 S. D.-M. Islam, M. Fujitsuka, O. Ito, A. Ikeda, T. Hatano and S. Shinkai, Chem. Lett., 2000, 78. 23 T. J. Dougherty, G. B. Grindey, R. Fiel, K. R. Weishaupt and D. G. Boyle, J. Natl. Cancer Inst., 1975, 55, 115. 24 S. G. Bown, C. J. Tralau, P. D. Coleridge Smith, D. Akdemis and T. J. Wieman, Br. J. Cancer, 1986, 54, 43; E. Ben-Hur and I. Rosenthal, Int. J. Radiat. Biol., 1985, 47, 145; W. S. Chan, R. Svensen, D. Phillips and I. R. Hart, Br. J. Cancer, 1986, 53, 255; N. Brasseur, H. Ali, D. Autenrieth, R. Langlois and J. E. van Lier, Photochem. Photobiol., 1985, 42, 515. 25 J. Moan, E. O. Pettersen and T. Christensen, Br. J. Cancer, 1979, 39, 398. 26 B. W. Henderson and T. J. Dougherty, Photochem. Photobiol., 1992, 55, 145. 27 K. R. Weishaupt, C. J. Gomer and T. J. Dougherty, Cancer Res., 1976, 36, 2326.

This journal is ß The Royal Society of Chemistry 2005