J. Phys. Chem. B 2010, 114, 15773–15779
15773
Singlet Oxygen Imaging in Polymeric Nanofibers by Delayed Fluorescence Jirˇ´ı Mosinger,†,‡ Kamil Lang,‡ Jirˇ´ı Hostomsky´,‡ Jirˇ´ı Franc,§ Jan Sy´kora,§ Martin Hof,§ and Pavel Kuba´t*,§ Faculty of Science, Charles UniVersity in Prague, HlaVoVa 2030, 128 43 Praha 2, Czech Republic, Institute of Inorganic Chemistry, V.V.i., Academy of Sciences of the Czech Republic, 250 68 Rˇezˇ, Czech Republic, and J. HeyroVsky´ Institute of Physical Chemistry, V.V.i., Academy of Sciences of the Czech Republic, DolejsˇkoVa 3, 182 23 Praha 8, Czech Republic ReceiVed: June 23, 2010; ReVised Manuscript ReceiVed: October 18, 2010
Polymeric nanofiber materials loaded with photosensitizers exhibit significant antibacterial activity due to their generation of cytotoxic singlet oxygen O2(1∆g). A time-gated fluorescence imaging technique was used to monitor the photosensitized processes in polystyrene (PS) and gelatin (GE) nanofibers loaded with 0.1 wt % tetraphenylporphyrin (TPP) photosensitizer. The fluorescence decay of TPP at the periphery of the PS nanofibers was single exponential. Increased fluorescence quenching was observed in the domains with higher TPP loading, located in the center of the nanofibers, and added a shorter lifetime component to the kinetics. The domains exhibiting singlet oxygen activity within the nanofibers were visualized and analyzed by singlet oxygen-sensitized delayed fluorescence imaging (SODF). Whereas O2(1∆g) was produced in PS nanofibers, its production in GE nanofibers was limited. These results were confirmed by time-resolved phosphorescence measurements at 1270 nm. 1. Introduction The photosensitized production of singlet oxygen O2(1∆g) mediates the oxidative degradation of many molecules and is important in several areas of biology and medicine.1 As such, research efforts have been extended toward the design of novel hybrid materials in which photosensitizers are fixed to a solid support (i.e., polymers,2-5 silica,6,7 zeolites,7 alumina,8 layered double hydroxides,9 and clays10), which have been effective in killing bacteria and viruses.2 However, the ability to modulate the domains of activity in solid materials through spatial and temporal control of O2(1∆g) production11 requires more sensitive methods for imaging O2(1∆g). The standard method of imaging O2(1∆g) uses a spinforbidden transition between O2(1∆g) and O2(3Σg-).12 The phosphorescence spectrum of O2(1∆g) has a band maximum centered at ∼1270 nm with a small quantum efficiency range of 10-4 to 10-7, which is environment-dependent.13 In media containing O2(1∆g) quenchers, interfering luminescence from the photosensitizer can make detection of the singlet oxygen signal difficult. Indirect imaging strategies can be more sensitive, down to the single molecule level, than the direct monitoring of O2(1∆g) f O2(3Σg-) transitions. However, the addition of fluorescent probes is necessary for these methods.14,15 Recently, we prepared polyurethane nanofiber materials, loaded with porphyrin/phthalocyanine photosensitizers that produce O2(1∆g) and exhibit antibacterial properties.3-5 The electrospun nanofibers possess large surface to volume ratios, increased porosity, and good mechanical properties for photobiological applications. Herein, we present a polystyrene (PS) * Author to whom correspondence should be addressed. E-mail:
[email protected]. † Charles University in Prague. ‡ Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic. § J. Heyrovsky´ Institute of Physical Chemistry, Academy of Sciences of the Czech Republic.
Figure 1. Simplified energy level diagram for singlet oxygen-mediated delayed fluorescence produced by two TPP molecules and one oxygen molecule.5
nanofiber material that is loaded with 0.1 wt % tetraphenylporphyrin (TPP) photosensitizer. The PS nanofiber material was compared with a gelatin (GE) nanofiber material that, in contrast, produces significantly less O2(1∆g). The spatially dependent lifetimes and concentrations of O2(1∆g) reflect the unique chemical composition of given domains in the polymer nanofibers. Our approach of O2(1∆g) imaging is based on the singlet oxygen-sensitized delayed fluorescence (SODF) component of fluorescence lifetime imaging microscopy (FLIM). Strong SODF, described previously for our polyurethane nanofibers, occurs by the reaction of triplet photosensitizers with O2(1∆g) (Figure 1).5 Other mechanisms of delayed fluorescence, notably triplet-triplet annihilation and thermally activated reverse intersystem crossing, can be excluded because no signal of delayed fluorescence was measured in a vacuum or inert gas. Fluorescence lifetime imaging microscopy (FLIM)16-18 can be used for the quantification of photoinduced processes in nanofibers and spatially resolved imaging of porphyrin molecules and singlet oxygen O2(1∆g) in different environments. The time-gated approach19 allows for the effective separation of the individual processes. Both SODF and prompt fluorescence (PF) images at different time scales are reconstructed from the scanned data.
10.1021/jp105789p 2010 American Chemical Society Published on Web 11/15/2010
15774
J. Phys. Chem. B, Vol. 114, No. 48, 2010
Mosinger et al.
Figure 2. PS nanofiber material loaded with 0.1 wt % TPP: SEM micrograph (a), SEM micrograph of bundle of nanofibers (b), steady-state UV-vis (c), and fluorescence (excitation at 514 nm) spectra (d).
2. Experimental Section Chemicals. 5,10,15,20-Tetraphenylporphyrin, N,N-dimethylformamide (DMF), tetraethylammonium bromide (TEAB), gelatin (GE), polystyrene (PS, Mw ) 180 000), and formic and acetic acids were purchased from Sigma-Aldrich and used as received. Preparation of Samples. The nanofiber layers were produced using the Nanospider electrospinning industrial technology.5 The polymer solutions (15% in DMF) used for the PS nanofiber preparation contained 0.1 wt % TPP, 0.01 wt % TEAB, and 99.89 wt % PS. The gelatin solution (17% in formic acid:acetic acid, 1:3 v/v) used for the GE nanofiber preparation contained 0.1 wt % TPP and 99.9 wt % GE. The diameter statistics were obtained by the analysis of 80 independent fibers in scanning electron microscopy micrographs. A piece of the nanofiber material was peeled from the polypropylene-supporting textile and placed on a quartz plate in an evacuable cell. The cell was connected to a glass vacuum line and to oxygen and argon gas cylinders. The gases (oxygen, air, and argon) were delivered into the vacuum manifold through needle valves. Before each experiment, the cell was evacuated to as low as 1 Pa by a rotation pump for at least 5 min. The total pressure in the reaction cell was measured with capacitance manometers (MKS Baratron). Scanning Electron Microscopy (SEM). Micrographs were taken using field emission scanning electron microscopy (Hitachi S-4800) at a low acceleration voltage (1.5 kV) and a short working distance (1.5 mm). Steady-State and Time-Resolved Spectroscopy. The UV-vis absorption measurements were carried out with a PerkinElmer Lambda 35 spectrometer equipped with a Labsphere RSA-PE20 integration sphere. Time-correlated single photon counting (with laser diode NanoLED-405LH, λexc ) 405 nm, pulse width 750 ps, repetition rate 1 MHz) and steady-state fluorescence of TPP (450 W Xe lamp, λexc ) 418 nm) were performed on a Fluorolog 3 spectrometer (Horiba Jobin Yvon). Samples were excited by a FL 3002 dye laser (λexc ) 422 nm, pulse width ∼28 ns) for single-point time-resolved measurements of the
triplet states, O2(1∆g), and SODF. The time profiles of the TPP triplet states were probed using their absorption at 460 nm. The formation of singlet oxygen O2(1∆g) was monitored using its near-infrared phosphorescence at 1270 nm with a germanium diode (Judson J16-8SP-R05M-HS). Fluorescence (both PF and SODF) was measured at the maximum of the fluorescence emission band (650 nm). The details of the measurements have been previously described.5 Fluorescence Lifetime Imaging Microscopy (FLIM). FLIM was carried out on an inverted epifluorescence confocal microscope MicroTime 200 (PicoQuant, Germany). We used a configuration containing a pulsed diode laser (LDH-P-C-405, 405 nm, PicoQuant) that provides 120 ps pulses at a 250 kHz repetition rate. The average power of the laser was set to 100 nW. The fluorescence was separated from the excitation light by the appropriate filter set (clean up filter 405/10, dichroic mirror 405DRLP and long-pass filter LP580 (Omega Optical)). The excitation light was focused with a water immersion objective (1.2 NA, 60×) (Olympus) on a spot size of ∼200 nm. The fluorescence photons were detected with a singlephoton avalanche diode (PerkinElmer). The Picoharp 300 timecorrelated single-photon counting module (PicoQuant) recorded the photon events in a time-tagged time-resolved (TTTR) mode, enabling the reconstruction of the lifetime histogram and fluorescence kinetics for each pixel.20 3. Results and Discussion 3.1. Electrospun Nanofiber Materials and Their Characteristics. PS nanofiber materials (1.9-2.0 g m-2 area weight, 0.03 mm thickness, and 1.05-1.07 g cm-3 average density), loaded with 0.1 wt % TPP, consisted mostly of isolated nanofibers (200 ( 30 nm in diameter) and bundles with relatively flat surfaces and small inhomogeneities (Figure 2a,b). A high oxygen diffusion coefficient, ranging from (2.3-3.1) × 10-7 cm2 s-1, was measured for these materials.21 GE nanofiber materials (3.7 g m-2 area weight, 0.03 mm thickness, 1.3 g cm-3 approximate average density), loaded with 0.1 wt % TPP, consisted of nanofibers with a diameter of 340
Singlet Oxygen Imaging in Polymeric Nanofibers
J. Phys. Chem. B, Vol. 114, No. 48, 2010 15775
( 130 nm. Compared to the PS materials, the GE nanofibers typically exhibited a lower oxygen diffusion coefficient, on the order of 10-8 cm2 s-1.22 The UV/vis absorption (Figure 2c) and fluorescence (Figure 2d) spectra demonstrated that TPP in nanofibers exists predominantly in the monomeric state; no shifts, hypochromicity, or broadening of the bands was observed.23 3.2. Calculation of Photophysical Parameters from Measurements of TPP Triplet States, O2(1∆g), and SODF. The rate constants of individual photophysical processes were evaluated from the single-point kinetic measurements of the TPP triplet states, O2(1∆g), and SODF at 15 different oxygen pressures in the range 0-100 kPa (Supporting Information, Figures S1-S4). For kinetic analysis, we used a previously applied, simplified homogeneous model (eqs 1-7).5 Briefly, the excitation of TPP to the 1TPP* state is followed by fluorescence (quantum yield ΦF ∼ 0.1) and/or nonradiative processes to the ground state 1TPP (eq 1), or intersystem crossing to low-lying triplets 3TPP (ΦT g 0.5) (eq 2). The triplets, 3TPP, are deactivated spontaneously (eq 3) or by interaction with oxygen in the ground triplet state ( O2(3Σg-) (eqs 4a, 4b) to form predominantly O2(1∆g) (eq 4a). Singlet oxygen O2(1∆g) deactivates spontaneously (eq 7) and/or forms the collision complex (1TPP* · O2(3Σg-)) (eq 5) that either emits photons (SODF, eq 6a) or decays by other channels (eqs 6b and 6c). kF + knr
TPP* 98 1TPP + hνF
1
1
kisc
TPP* 98 3TPP
(1)
Figure 3. Stern-Volmer plot for the triplet state lifetimes τT (eq 9) (a) and the dependence of k1q [3TPP]0 on the oxygen pressure pO2 (b) for the PS nanofiber material loaded with 0.1 wt % TPP. 1 kO
O2(1∆g) 98 O2(3Σg)
(7)
Similarly as in the previous study,5 other mechanisms of delayed fluorescence in PS nanofibers, namely, triplet-triplet annihilation and thermally activated reverse intersystem crossing, were excluded because no signal of delayed fluorescence was observed in a vacuum or in inert gas. The reaction scheme (eqs 1-7) generates a set of differential equations that were solved, yielding the concentrations of 3TPP and O2(1∆g) and the intensity of SODF (ISODF) as a function of time (see Supporting Information):
[3TPP] ) [3TPP]0 exp(-t/τT)
(8)
1/τT ) 1/τT0 + (kq′ + kq′′)pO2
(9)
(2) where
kT0
TPP 98 1TPP
3
3
3
(3)
1 1 TPP + O2(3Σg ) 98 TPP + O2( ∆g)
TPP +
q
O2(3Σg)
k′′ q
{ ( ) (
( )}
t × τT t (n + 1)t exp - exp ∞ τ∆ τT 1 3 n (-kq[ TPP]0τT) n+1 1 n)0 n! τT τ∆
[O2(1∆g)] ) kq′ pO2[3TPP]0 exp kq1[3TPP]0τT exp -
k′
98 TPP + 1
O2(3Σg)
(4a)
∑
(
)
k1q
TPP + O2(1∆g) 98 (1TPP* · O2(3Σg ))
SO kDF
1 3 (1TPP* · O2(3Σg )) 98 TPP + O2( Σg ) + hνF
SO kisc
3 3 (1TPP* · O2(3Σg )) 98 TPP + O2( Σg )
SO kic
1 3 (1TPP* · O2(3Σg )) 98 TPP + O2( Σg )
(10)
(4b) ISODF ) ASODF[3TPP][O2(1∆g)]
3
)
(5)
(6a)
(6b)
(6c)
(11)
Here τT, τT0 ) 1/kT0 , and τ∆ ) 1/kO1 are the lifetimes of 3 TPP at given oxygen pressure, 3TPP in vacuum, and O2(1∆g), respectively, pO2 is the oxygen pressure, and ASODF is a scale constant. Time-resolved measurements of the absorbance changes at 460 nm (Supporting Information, Figure S1a) confirmed the monoexponential decay of 3TPP with the lifetime, τT (eq 8). The Stern-Volmer plot (eq 9) affords the rate constant of 3TPP quenching by oxygen, kq ) kq′ + kq′′ ) 2.0 Pa-1 s-1 (Figure 3a). The plot exhibits a slight downward curvature, indicating that the deactivation of 3TPP may be limited by oxygen transport. The individual triplet lifetimes, τT, were determined for each oxygen pressure by eq 8 (Supporting Information, Figures S1a and S2). Equations 10 and 11 were then applied for the simultaneous fitting of all O2(1∆g) and SODF traces (Supporting Information, Figures S1b, S1c, S3, and S4). The fitting procedure
15776
J. Phys. Chem. B, Vol. 114, No. 48, 2010
Figure 4. Time-resolved fluorescence of 0.1 wt % TPP in the PS nanofiber material (a) and in the GE nanofiber material (b) and the profile of the excitation pulse (c). The PF intensity image collected from 0 to 60 ns after excitation (d), PF intensity image 10-60 ns after excitation (e), and SODF intensity image 400-2000 ns after excitation (f). The corresponding fluorescence intensity profiles along the dotted line for 0.1 wt % TPP in PS nanofiber material are at the bottom.
using eq 10 produces τ∆ ) 17 ( 2 µs and the parameter kq1 [3TPP]0, which is a pseudomonomolecular apparent rate constant for the repopulation reaction (eq 5) with a constant concentration of triplet states equal to [3TPP]0. Additionally, the SODF traces were fitted in such a way that kq1 [3TPP]0 was determined individually for each value of the oxygen pressure, whereas τ∆ was fixed to 17 µs (Supporting Information, Figure S4). The individual kq1 [3TPP]0 values increased from 1.7 × 105 to 6.7 × 105 s-1 with increasing oxygen pressure (Figure 3b), indicating the effect of oxygen transport on the rate of the repopulation reaction (eq 5). 3.3. Fluorescence Lifetime Imaging Microscopy of Nanofibers. The imaging of porphyrin molecules in various materials and biological objects by confocal fluorescence microscopy is well-known.24 The time resolution of FLIM enables the observation of photosensitized processes in selected nanofibers. Taking into account the shape of the excitation pulse (Figure 4c), we were able to separate the prompt fluorescence (up to 60 ns after excitation) and the singlet-oxygen sensitized delayed fluorescence (400-2000 ns after excitation) components of TPP fluorescence in PS nanofiber materials (Figure 4a). The PF and SODF intensity images of PS nanofiber materials loaded with 0.1 wt % TPP are shown in Figure 4d-f. PF Images. The PF intensity images display the distribution of 1TPP* in the nanofibers (Figure 4d). Because the spatial resolution of FLIM in our arrangement is ∼200 nm, the PS nanofibers demonstrate a higher intrinsic diameter in PF images than in SEM micrographs (Figure 2a,b). We were unable to resolve all individual fibers, especially those in bundles. Assuming that the TPP molecules in the PS matrix (1.05 g.cm-3 density and 0.1 wt % TPP) are distributed regularly, a simple calculation generates an average intermolecular distance, linter, of ∼10 nm. With a laser spot diameter of ∼200 nm, hundreds to thousands of TPP molecules were excited by each laser pulse; therefore, the imaging of individual TPP molecules could not be achieved. The difference in the fluorescence kinetics within and at the periphery of the nanofibers can be seen in the PF images, where
Mosinger et al.
Figure 5. Fluorescence images (2.5 × 2.5 µm) of a polymeric node in PS nanofiber material loaded with 0.1 wt % TPP: PF intensity image for the photons that arrived during the time interval 0-50 ns after excitation and the intersection along the dotted line (a), corresponding fluorescence lifetime image calculated according to eq 12 for the interval 0-50 ns and the intersection of the lifetime values along the dotted line (b), fluorescence decay kinetics inside (c) and at the periphery of the node (d) and SEM micrograph of nodes and other irregularities in nanofiber material (e).
the fast kinetics (0-10 ns after excitation) was omitted (i.e., only the photons detected between 10-60 ns were imaged) (Figure 4e). The spots with lower fluorescence intensity (faster decay kinetics) within the nanofibers might correspond to higher concentrations of TPP. The shorter intermolecular distance between individual TPP molecules can be responsible for fluorescence quenching. Microenvironments with different fluorescence kinetics were clearly visualized in the larger polymeric objects. The PS nanofibers consisted of only sporadic irregularities, such as bundles of nanofibers (Figure 2) or nodes (Figure 5e). Parts a and b of Figure 5 show a polymeric node visualized by FLIM in a region of a PS nanofiber, in which irregularities are present (Figure 5e). To characterize the differences in the fluorescence decays recorded for each pixel, the average lifetime, τav, was determined as the difference between the barycenter of the fluorescence decay recorded at one position of the scanner and the offset. These values were extrapolated from the steepest rise of the decay and applied to the following:25
τav )
∑ Iiti - toffset ∑ Ii
(12)
where toffset is the offset time and Ii and ti are the intensity and time corresponding to the i-th channel in the time-correlated single photon counting histogram, respectively.
Singlet Oxygen Imaging in Polymeric Nanofibers The node possesses two subdomains with different τav. The fluorescence decays reconstructed from the pixels with the shortest τav within the PS matrix (violet spots, Figure 5b) exhibit faster kinetics (Figure 5c) than the 1TPP* decays extracted from the pixels at the periphery of the polymeric material (Figure 5d). The subdomain, where τav is as short as 7 ns (Figure 5b), correlated with a large number of TPP molecules (Figure 5a). The presence of different environments around the TPP molecules can be connected with the different polarity and accessibility of the quenchers, e.g., the close proximity of another TPP molecule critically influences the lifetime of the 1 TPP* by adding shorter components to the PF monoexponential decay (Figure 5c,d). Also, differences in the density of the PS matrix could contribute to the quenching of 1TPP*. Impurities, such as residual solvent (DMF) from the electrospinning process, should not change the PF lifetime of TPP molecules in the PS nanofiber material significantly (τF ) 10.4 ns in liquid DMF). SODF Image. Because SODF is based on the reaction of 3 TPP with O2(1∆g) (eqs 5 and 6a), the SODF intensity image shows only TPP molecules that are easily accessed by O2(1∆g). The amplitudes of SODF imaging signals (
