Photophysical properties and singlet oxygen generation efficiencies of water-soluble fullerene nanoparticles.

As various fullerene derivatives have been developed, it is necessary to explore their photophysical properties for potential use in photoelectronics and medicine. Here, we address the photophysical properties of newly synthesized water-soluble fullerene-based nanoparticles and polyhydroxylated fullerene as a representative water-soluble fullerene derivative. They show broad emission band arising from a wide-range of excitation energies. It is attributed to the optical transitions from disorder-induced states, which decay in the nanosecond time range. We determine the kinetic properties of the singlet oxygen ((1)O2) luminescence generated by the fullerene nanoparticles and polyhydroxylated fullerene to consider the potential as photodynamic agents. Triplet state decay of the nanoparticles was longer than (1)O2 lifetime in water. Singlet oxygen quantum yield of a series of the fullerene nanoparticles is comparably higher ranging from 0.15 to 0.2 than that of polyhydroxylated fullerene, which is about 0.06.

Introduction

Since its discovery in 1985, buckminsterfullerene (C60) has stimulated a large body of research due to its unique photophysical properties 1–4. Although the fluorescence of C60 is intrinsically week, chemical modification allows C60 to raise its fluorescence 3,4. Moreover, the color of n class="Chemical">fullerene nanoparticles may be tuned by varying the C60 concentration in the reaction solution 5. In general, certain carbon nanomaterials exhibit optical emission due to quantum confinement effects 6. In this way, nano-sized carbon materials have attracted much attention as they are expected to replace conventional cadmium-based quantum dots. Triplet state properties of modified C60 should also be promising, as pristine fullerene readily generates singlet oxygen (1O2) and other reactive oxygen species by illuminating ultraviolet or visible light 7,8. Energy transfer from the excited triplet state of fullerene to the ground state of oxygen gives rise to 1O2 as illustrated in the following scheme:

Singlet oxygen is able to irreversibly cause damage of various cellular organelles and biomolecules, including mitochondria, n class="Chemical">lipid, and nucleus, thus leading to damage of target diseased cells or tissue 9. Photosensitizer, a generator of singlet oxygen by irradiating light, is an important factor in photodynamic therapy. Although C60-fullerene is known to be a strong singlet oxygen generator (singlet oxygen generation quantum yield is 1), it has been limitedly used in photodynamic therapy due to its extremely low solubility in water. To overcome this drawback, many efforts have been explored to develop the water-soluble C60 by various approaches including reaction with hydrophilic moieties, grafting polymers, and applying surfactants 10. As a rule, such modification of C60 significantly affects its photophysical properties. For example, fullerol (C60(OH)24), a representative water-soluble fullerene, is known to have low efficiency of 1O2 generation compared to that of pristine fullerene 11. Therefore, it is a challenge to synthesize water-soluble fullerene derivatives with sufficient photosensitizing activity.

Recently, highly water-soluble fullerene nanoparticles (C60-TEGs) were prepared by conjugating C60 with tetraethylene glycol (TEG) using lithium hydroxide as a catalyst (Fig.1) 5.

Figure 1

Structure of tetraethyleneglycol-conjugated fullerene (C60-TEGn).

In current work, we investigated the photophysical properties of the C60-TEGs including time-resolved fluorescence, triplet state lifetime, and n class="Chemical">singlet oxygen generation quantum yield by combining various experimental approaches of conventional and time-resolved spectroscopy. In particular, we compared the photophysical properties of these new water-soluble fullerene nanoparticles with that of a representative water-soluble fullerene (polyhydroxylfullerene (C60(OH)n)) by measuring the 1O2 kinetic luminescence signals to demonstrate the potential of C60-TEGs as a photosensitizer.

Materials and Methods

Materials

Bukminsterfullerene (C60) and polyhydroxylfullerene (C60(OH)n) were obtained from MER Co. (AZ, USA). Tetraethylene glycol and lithium hydroxide were purchased from Sigma–Aldrich (St. Louis, MO, USA). 5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H,23H-porphine (TMPyP) tosylate salt was kindly provided by Dr. V.L. Malinovskii.

Synthesis of water-soluble fullerene nanoparticles

The water-soluble n class="Chemical">fullerene nanoparticles were prepared following previous procedure 5. Five mL of fullerene dissolved in toluene with various concentrations (0.12, 0.25, 0.5, 1.0, 2.0 mg mL−1) was mixed with 5 mL of tetraethylene glycol. Then, 20 mg of lithium hydroxide was added to the mixture solution, and the solution was stirred for 20 h at room temperature. After stirring, the mixture solutions were precipitated by adding excess of ethyl acetate and redissolved in ethanol. This step was repeated at least three times to remove the unreacted chemicals. The precipitates (C60-TEGs) were finally dried by evaporation and kept in storage until use. The nanoparticles were denoted as CT006, CT012, CT025, CT050, and CT100, according to initially used fullerene concentration in toluene/tetraethylene glycol mixture (0.06, 0.12, 0.25, 0.5, and 1.0 mg mL−1, respectively). The chemical characteristics of the fullerene nanoparticles were analyzed by FTIR, XPS, and NMR methods, and the results were similar to those previously reported. The analysis revealed that a large fraction of C60 is oxidized inside of the nanoparticles. During preparation, the concentration of C60 was the main factor in controlling the size and chemical composition of the C60-TEGs. Increasing initial concentration of C60 yielded the fullerene nanoparticles of greater size and correspondingly mass.

Steady-state absorption and fluorescence measurements

Absorption spectra were recorded on an MC 122 spectrophotometer (Proscan Special Instruments, Belarus) in quartz cuvettes. Fluorescence emission and excitation spectra were recorded on a CM 2203 spectrofluorimeter (SOLAR, Belarus) using the excitation wavelength at 350 nm and the emission wavelength at 550 nm, respectively. The fluorescence quantum yield, φ, was determined by a relative method using a solution of quinine sulfate in 1.0 n class="Chemical">N H2SO4 as a fluorescent standard, for which the quantum yield of 0.546 was obtained 12. The relative accuracy in determination of the φ values was 10%.

Time-resolved fluorescence measurements

The lifetimes of the excited singlet state, τ1 and τ2, of the fullerene derivatives were determined by measuring fluorescence decays using a modified PRA-3000 pulse fluorometer (Photochemical Research Associates, Canada) operating in the time-correlated single photon counting mode. The excitation source was a 378-nm pulsed light-emitting diode (PLS-380, PicoQuant GmbH, Germany), full width at half maximum (FWHM) of the diode pulses being ∽990 ps. The decays were analyzed by means of Edinburgh FLA900 software. The values of τi were determined by the nonlinear least-squares reconvolution analysis of fluorescence decay curves assuming a double-exponential fluorescence decay law F(t) = F1exp(−t/τ1) + F2exp(−t/τ2). The estimated relative errors in fluorescence lifetimes were about 5%.

Flash photolysis

Triplet state decay traces were recorded using a home-made laser transient absorption setup 13. The excitation light source was a LS-2134 Nd:YAG laser (JV LOTIS TII, Belarus) providing 355 nm pulses of 15 ns FWHM and 1 mJ energy. A 150 W xenon arc lamp (OSRAM AG, Germany) was used as a probe beam. Detection of the signal in the visible region was performed with a FEU-128 photomultiplier tube via an MS2004i monochromator (SOLAR TII Ltd., Belarus). The C60-n class="Chemical">TEG solutions were prepared with optical densities of the samples about 0.4 at the excitation wavelength. To evaluate triplet state lifetime, τT, signals were fitted by exponential decay function.

Singlet oxygen luminescence measurement

The luminescence of singlet oxygen was measured using a modified laser near-infrared lifetime spectrometer (detection range 950–1400 nm), which was created at the Institute of Physics of the n class="Chemical">National Academy of Sciences of Belarus 14,15. Samples were excited by the third harmonic (355 nm) of an Nd:YAG laser (LS-2134, JV LOTIS TII, Belarus). Typical parameters of the laser were as follows: the pulse width of 15 ns, the pulse energy of 20 μJ, and the repetition rate of 15 Hz. Luminescence radiation, collected with a high-throughput optical system, was spectrally isolated with an MS2004i monochromator (SOLAR TII Ltd., Belarus) and directed to a photomultiplier tube (model H10330-45, Hamamatsu Photonics K.K.), operated in the photon-counting mode. After amplification by C5594-44 unit (Hamamatsu Photonics K.K.), the output of the photomultiplier tube was sent to a multiscaler (P7888-2, FAST ComTec GmbH). The time channel width of the multiscaler was set to 32 ns. Singlet oxygen quantum yield, γΔ, is a ratio of number of produced singlet oxygen molecules to the number of absorbed quanta of excitation light by sensitizer molecules. The most common method of γΔ estimation is a comparative one. To evaluate γΔ from the C60-TEGs, we compared intensities of near-infrared (∽1270 nm) luminescence of 1O2 generated by the C60-TEGs and the TMPyP in water, the latter being a standard 16 with  = 0.77 ± 0.04.

The common expression for the time dependence of the number of 1O2 molecules may be derived under condition of pulsed laser excitation of the photosensitizer 17–20:

where [1O2] is the concentration of singlet oxygen molecules, γΔ is the singlet oxygen quantum yield, N is the number of quanta in the laser pulse, OD is the optical densities of the solute at the excitation wavelength, τΔ and τT are lifetimes of singlet oxygen and photosensitizer triplet state, respectively.

The intensity of 1O2 emission is proportional to the concentration of produced 1O2 molecules 19,20:

where α is a constant that contains geometrical and electronic factors of the detection system, kr is 1O2 radiative decay constant, n is the solvent refractive index,

Equation (6) may be used to determine singlet oxygen quantum yields. Under the identical experimental conditions the γΔ from the modified fullerenes was calculated according to:

In Eq. (7), the superscript “st” corresponds to the standard.

To observe singlet oxygen luminescence, optical densities of both standard and investigated samples were ∽0.2 at 355 nm excitation wavelengths. To minimize samples photodegradation, continuous magnetic stirring was used. Within the signal accumulation time, the optical density of all samples decreased no more than by 5%. Average values of optical densities were used to calculate n class="Chemical">singlet oxygen quantum yields.

Results and Discussion

Absorption and fluorescence characteristics

We have employed two different fullerene derivatives: (i) C60n class="Chemical">(OH)n, a covalent functionalized hydroxylated C60, and (ii) the C60-TEGs particles. Both of these fullerene derivatives are readily soluble in water. However, it has been reported that some functionalized fullerenes may form aggregates in polar media at high concentrations 11,21,22. To verify the formation of aggregates, UV-visible absorption spectra of different concentrations of the C60-TEGs and C60(OH)n in aqueous solution were recorded. As shown in Fig.2, the absorption of C60(OH)n in water showed a linear increase at 355 nm with increasing concentration up to 58 μm (Fig.2A, inset). The same linear tendency is observed for the CT006 (Fig.2B, inset) and for other C60-TEGs (not shown). For the separate sample, all normalized absorption spectra were coincided with each other. These results clearly showed that in aqueous solution both C60(OH)n molecules and the C60-TEGs particles were present in the monomeric state.

Figure 2

UV-vis absorption spectra of different concentrations of C60(OH)n (A) and CT006 (B) in aqueous solution. Spectra from bottom to top correspond to [C60(OH)n] 0.5, 0.8, 1.4, 2.4, 4.0, 7.0, 12, 20, 34, 58 μm and [CT006] 0.01, 0.02, 0.04, 0.07, 0.12, 0.21, 0.35, 0.59, 1.0, 1.7 mg mL−1. Insets: Plots of the absorbance at 355 nm against the concentration; path length, 1 cm.

UV-vis absorption spectra of different concentrations of C60(OH)n (A) and CT006 (B) in aqueous solution. Spectra from bottom to top correspond to [C60n class="Chemical">(OH)n] 0.5, 0.8, 1.4, 2.4, 4.0, 7.0, 12, 20, 34, 58 μm and [CT006] 0.01, 0.02, 0.04, 0.07, 0.12, 0.21, 0.35, 0.59, 1.0, 1.7 mg mL−1. Insets: Plots of the absorbance at 355 nm against the concentration; path length, 1 cm.

Previously, for the C60(OH)24 and C60(OH)36, deviation from the Beer-Lambert law was observed at micromolar concentrations and above 11,22, when the C60(OH)24 and C60(OH)36 became aggregated, possibly due to the formation of a hydrogen bond network. In present work, C60n class="Chemical">(OH)n was supplied by MER Co in the form of a salt with the following stochiometric formula: C60(OH)x(ONa)y, with x + y = 24 and y around 6–10. Thus, 24 carbon atoms are hydroxylated in our compound. The discrepancy between results of different scientific groups may be explained by the nature of hydroxylated fullerene. Synthetic routes in the works 11,22 may result in large number of OH-groups readily forming hydrogen bonds in water solutions, whereas our samples bear at least six repulsing O− group on their surface.

The broad absorption spectra extended up to 800 nm (Fig.3) indicate the absence of a well-defined bandedge in the UV-vis energy range both for the C60-TEGs and C60n class="Chemical">(OH)n.

Figure 3

Normalized fluorescent and absorption spectra of the C60TEGs and C60(OH)n water solutions. The fluorescent spectra were obtained for excitation at 350 nm (3.54 eV).

Normalized fluorescent and absorption spectra of the C60n class="Chemical">TEGs and C60(OH)n water solutions. The fluorescent spectra were obtained for excitation at 350 nm (3.54 eV).

In contrast to the broad absorption features, relatively narrow fluorescent peaks (FWHM ∽ 0.7 eV) centered around 580 nm are observed. Fluorescent spectra of both the C60-TEGs and C60n class="Chemical">(OH)n possess near-Gaussian shape, slight asymmetry being introduced by little shoulders. The fluorescent quantum yield of C60(OH)n is about 0.3%, while for the C60-TEGs the largest φ is about 1% as found previously 5.

The fluorescent excitation spectra obtained in the maximum of the emission spectrum agree well with the corresponding absorption spectra of the C60-TEGs and C60n class="Chemical">(OH)n. However, the fluorescent excitation spectra obtained for different wavelengths of the emission spectrum ranging from 430 to 660 nm slightly differ in shape between each other. The longer detection wavelength is used, the more excitation spectrum is shifted to “red” region. Also the fluorescence spectra undergo insignificant “red” shift, when excitation wavelength is increased from 290 to 450 nm. All these little differences in wavelength-dependent excitation and emission spectra are caused by small heterogeneity of particles’ size, confirming SEM measurements performed in 5. Therefore, each studied species actually possesses one dominant type of fluorophores.

Figure3 shows the absence of mirror symmetry between the absorption and fluorescent emission spectra. And this fact cannot be accounted by the small heterogeneity of nanoparticles’ size. Thorough analysis of experimental papers concerning carbon materials provides a number of luminescent nano-objects, which do not show mirror symmetry between excitation and emission luminescence spectra: n class="Chemical">graphene oxide 23, reduced graphene oxide 24, and graphene quantum dots 25. Although much progress has been achieved in graphene-related studies, the origin of the fluorescence of modified graphene is still a controversial issue with variety of models and theories 23. As discussed in Experimental section, the core of the C60-TEGs consists of conjugated fullerene molecules. So, the fullerene core of the studied particles brings together the C60-TEGs with fullerene films 26, carbon nanodots 27, and graphite nanoparticles, obtained by transformation of C60 molecules 28. At the same time, fullerene molecules are chemically modified by oxygen and TEG residuals on the surface of the C60-TEGs.

Chemical bonds of C60 with carbon, oxygen and hydrogen introduce sp3 hybridized sites into a sea of sp2 hybridized carbon of pristine fullerene. Therefore, the C60-TEGs may be similar to hydrogenated amorphous carbon 29 and to chemically derived graphene oxide 24, which contain both sp3 and sp2 sites. As was shown earlier 5, a large fraction (30–40%) of aromatic carbon in C60-TEGs is sp3 hybridized and covalently bonded with oxygen in form of epoxy and hydroxyl groups similar to graphene oxide 30. The remaining aromatic carbon is sp2 hybridized and bonded either with neighboring carbon atoms or with oxygen in the form of ketone group.

In carbon materials, containing a mixture of n class="Chemical">sp3 and sp2 bonding, the σ bonds of sp3 and sp2 sites give rise to σ valence and σ* conduction-band states 31. The π and π* states of the sp2 sites lie within the σ–σ* gap, and form the band edges and control the optical gap. Because of (i) the large energy difference between sp2 and sp3 sites and (ii) their disorder localization in the C60-TEGs, these particles may be treated just like conventional amorphous semiconductors, with disorder-induced states 32. As a result, in presented fullerene materials broad emission band arising from a wide range of excitation energies (Fig.3) may be mainly attributed to optical transitions from these disorder-induced states (Fig.4).

Figure 4

Schematic diagram showing the origin of emission in C60-TEGs from disorder-induced states.

As mentioned above, in amorphous semiconductors, sp2 sites control the optical gap. Robertson deduced empirical linear dependence between photoluminescence energy and n class="Chemical">sp2 carbon fraction for different samples of hydrogenated amorphous carbon 29. It is known that carbon can form double bonds with itself and different heteroatoms. This requires sp2 hybridization of its valence atomic orbitals. Based on the relative content of C=C and C=O bonding as measure of sp2 carbon inside the C60-TEGs 5, we obtain good linear relationship with maxima of fluorescence spectra of the particles, as shown in Fig.5.

Figure 5

Relative content of C=C and C=O bonding in the C60TEGs versus maxima of fluorescence spectrum of these particles.

To understand better the fluorescent behavior of the different C60-TEGs particles and the C60n class="Chemical">(OH)n, time-resolved photoluminescence decay were monitored at 550 nm, using 378 nm as excitation wavelength. Each decay curve was fitted to the double-exponential function (Table1), resulting in a fast component (∽1 ns) and a slow one (∽4 ns). Inherently, disorder-induced states are a set of separate states 32. They may be treated as a set of separate fluorophores, which both decay nonradiative and emit light with different rates. Therefore, double-exponential law to describe fluorescence of the C60-TEGs particles is a simplification and results listed in Table1 may be used only to compare different samples between each other.

Table 1

Summary of lifetimes and weighting factors*, WFi, from double-exponential fitting of fluorescence decay kinetics.

Sample	τ1, ns	WF1	τ2, ns	WF2
CT006	1.2	0.45	4.2	0.55
CT012	1.1	0.44	4.1	0.56
CT025	1.2	0.50	4.2	0.50
CT050	0.9	0.43	3.3	0.57
CT100	1.1	0.49	3.5	0.51
C60(OH)n	0.8	0.50	3.6	0.50

Weighting factors were calculated as WFi = Fiτi/(F1τ1 + F2τ2).

The samples with red shifted fluorescent spectra (Fig.3) emit light faster than the particles with blue shifted fluorescent spectra. However, for all samples contribution of fast decay component is almost equal to the slow one.

Triplet state properties

While the large body of calculations and measurements of optical properties of nanoparticles 23–25,27,29,31–33, the nature of the optically active states has remained controversial. Both fluorescence and phosphorescence of nanoparticles may be used in a variety of applications 34. Moreover, diatomic oxygen is able to quench nanoparticle's triplet states resulting in formation of n class="Chemical">singlet oxygen, which is the main cytotoxic agent during photodynamic therapy of malignant tumors 9. Therefore, nanoparticles with long-lived triplet state are promising for medical treatment and diagnostics. Flash photolysis experiments resulted in longer triplet state lifetime for the modified fullerenes than for the porphyrin (Table2). As mentioned in previous section, disorder-induced states may be treated as a set of separate fluorophores. Triplet state lifetime is a priori longer than singlet state lifetime and collisions with molecular oxygen define decay of triplet states in most cases. Therefore, conventionally different disorder-induced states inside separate particle possess the same triplet state lifetime resulting in mono exponential triplet decay (Table2).

Table 2

Lifetimes of triplet state of investigated sensitizers; lifetimes and quantum yields of singlet oxygen sensitized by the standard porphyrin and by the modified fullerenes in air equilibrated distilled water solutions.

Sample	τT*, μs	τT†, μs	TΔ, μs	γ Δ
TMPyP	1.9 ± 0.1	1.8 ± 0.1	3.7 ± 0.1	0.77
CT006	9.0 ± 1.2	13 ± 1	1.9 ± 0.4	0.16 ± 0.03
CT012	9.2 ± 1.1	11 ± 1	2.5 ± 0.4	0.15 ± 0.02
CT025	9.3 ± 1.7	11 ± 1	1.7 ± 0.4	0.18 ± 0.03
CT050	7.1 ± 0.8	10 ± 1	2.3 ± 0.3	0.20 ± 0.03
CT100	7.0 ± 1.4	8.9 ± 0.9	2.3 ± 0.4	0.19 ± 0.03
C60(OH)n	5.9 ± 1.3	6.2 ± 1.5	1.6 ± 0.8	0.06 ± 0.02

During flash photolysis, triplet state decay was monitored at 600 nm.

Triplet state lifetimes, obtained from measurements of singlet oxygen luminescence. Confidence intervals for all values are stated at the 80% confidence level.

It is known that oxygen quenching of sensitizer triplet states is mainly governed by diffusion-controlled bimolecular collisions and spin-statistical factor 35. The corresponding rate constant for quenching of the triplet state lies near 109 m−1s−1. Taking into account, the concentration of molecular n class="Chemical">oxygen in water (3·10−4 m), the typical triplet-state lifetime of water-soluble sensitizers is about 2–3 μs 14,18,20,36–38. As we can see from the Table2, in comparison with the porphyrin, the C60-TEGs and C60(OH)n are effectively shielded from the interaction with ground state oxygen. TEG residues at the C60-TEG particle's outer layer protect sufficiently excited triplet state from dissolved oxygen quenching. Also OH surface groups shield fullerene core of both the C60-TEGs and C60(OH)n. Similar shielding effect was found for the number of fulleropyrrolidine and bismethanofullerenedendrimers in aerated organic solvents 3. It was proofed that dendritic wedges are able to protect fullerene cores from the quenching of oxygen molecules.

Singlet oxygen generation efficiency

In most pure solvents under 1 atm. pressure due to high concentration and diffusion rate of dissolved oxygen, triplet state lifetime is generally much shorter than n class="Chemical">singlet oxygen lifetime. Therefore, the rising part of 1O2 signal from Eq. (5) is described by τT and the decay part by τΔ. But if the sensitizer is effectively shielded from interaction with ground state oxygen, τT may become larger than τΔ. Consequently, denominator in Eq. (5) will be negative leading to τrise = τΔ and τdecay = τT which is so called inversion of kinetic phases in detected signal 18,36. As was mentioned above, the modified fullerenes are effectively shielded from interaction with ground state oxygen. Therefore, when illuminated by UV light, the C60-TEGs and C60(OH)n produce clear kinetic signal of 1O2 luminescence at 1270 nm with inverted kinetic phases (Table2, Fig.6). Similar effect was observed for chlorin e6 binding with polyvinylpyrrolidone in aqueous solutions 37.

Figure 6

Kinetics of CT050-photosensitized luminescence of singlet oxygen in water at excitation wavelength 355 nm after 7.5 × 103 laser pulses. The time resolution is 32 ns channel−1. The average noise level equal to 80 counts was subtracted from the signal. A solid line is the two-exponential curve fitting by Eq. (4). 1O2 luminescence rise and decay times were found to be 2.3 ± 0.3 μs and 10 ± 1 μs, respectively. WR are the weighted residuals.

Kinetics of CT050-photosensitized luminescence of singlet oxygen in n class="Chemical">water at excitation wavelength 355 nm after 7.5 × 103 laser pulses. The time resolution is 32 ns channel−1. The average noise level equal to 80 counts was subtracted from the signal. A solid line is the two-exponential curve fitting by Eq. (4). 1O2 luminescence rise and decay times were found to be 2.3 ± 0.3 μs and 10 ± 1 μs, respectively. WR are the weighted residuals.

The biexponential fit of TMPyP signal at 1270 nm by Eq. (5) yields τT ∽ 1.8 μs, being the typical value of triplet state lifetime of this n class="Chemical">porphyrin 38, and τΔ ∽ 3.7 μs corresponding singlet oxygen lifetime in pure water 39–41. There is no literature evidence of time-resolved investigations of the C60-TEGs and C60(OH)n triplet state or their singlet oxygen generation in water. Data recorded at 1270 nm yielded time constants, τΔ (Table2), being smaller than the typical value 3.7 μs in pure water.

After 1O2 formation, it interacts both with solvent and with the C60-TEGs molecules. Within the accuracy of measurements, rate constants of 1O2 deactivation kΔ = 1/τΔ∽(0.5 ± 0.1) μs−1 were the same for all C60-TEGs.

Based on FTIR analysis, the C60-TEGs have high density of hydroxyl group (OH) which may act as an effective quencher of n class="Chemical">singlet oxygen luminescence 42. Moreover, 1O2 in C60(OH)n water solution has the same rate constants kΔ ∽ 0.6 μs−1similar in the solutions of the fullerene nanoparticles. Therefore, the presence of hydroxyl termini is the reason of faster 1O2 deactivation in solutions under investigation in comparison with water solution of porphyrinoid sensitizer, where singlet oxygen decays with a rate constant about 0.3 μs−1.

Time-resolved luminescence signals at 1270 nm were used to determine the efficiency of singlet oxygen generation by the modified n class="Chemical">fullerenes (Table2). Recent years many papers studying phototoxicity of different forms of water-soluble fullerene derivatives have been published 11,22,43,44. In particular, fullerol C60(OH)24, has been investigated and its photophysical properties including singlet oxygen generation are studied. Zhao et al. 11 reported the γΔ of singlet oxygen generation by fullerol as 0.08 in deuterated water at 366 nm excitation wavelength. Kong et al. 45 mentioned that the γΔ were ranging from 0.048 to 0.1 depending on pH. In this study, we also measured the γΔ of polyhydroxylfullerene (C60(OH)n) purchased from MER Corp. to compare its capability of singlet oxygen generation with the newly synthesized water-soluble fullerene derivatives. The γΔ value for C60(OH)n was equal to 0.06, which was lower than those obtained from the C60-TEGs. Therefore the C60-TEGs more efficiently generate singlet oxygen than polyhydroxylfullerene and fullerol.

Conclusions

We studied luminescent and photophysical properties of polyhydroxylfullerene and the new C60-n class="Chemical">TEGs nanoparticles. Broad emission band arising from a wide-range of excitation energies was attributed to optical transitions from disorder-induced states. For all the C60-TEGs particles, nanosecond photoluminescence kinetics does not decay monoexponentially. The fluorescent quantum yield of C60(OH)n is about 0.3%. For the first time, the 1O2 kinetic luminescence signals produced by polyhydroxylfullerene and the C60-TEGs nanoparticles were detected in water. The kinetic phases of singlet oxygen traces at 1270 nm from the modified fullerenes were inverted as their triplet state lifetimes were longer than singlet oxygen decay. Singlet oxygen quantum yield was obtained up to 0.2 for all the C60-TEGs particles.