Article pubs.acs.org/JPCA
Singlet Oxygen Phosphorescence Lifetime Imaging Based on a Fluorescence Lifetime Imaging Microscope Wenming Tian,† Liezheng Deng,*,† Shengye Jin,† Heping Yang,† Rongrong Cui,† Qing Zhang,‡ Wenbo Shi,† Chunlei Zhang,‡ Xiaolin Yuan,‡ and Guohe Sha† †
State Key Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, Liaoning 116023, China ‡ Research Center, Affiliated Zhongshan Hospital of Dalian University, Dalian, Liaoning 116001, China ABSTRACT: The feasibility of singlet oxygen phosphorescence (SOP) lifetime imaging microscope was studied on a modified fluorescence lifetime imaging microscope (FLIM). SOP results from the infrared radiative transition of O2(a1Δg → X3Σg−) and O2(a1Δg) was produced in a C60 powder sample via photosensitization process. To capture the very weak SOP signal, a dichroic mirror was placed between the objective and tube lens of the FLIM and used to divide the luminescence returning from the sample into two beams: the reflected SOP beam and the transmitted photoluminescence of C60 (C60-PL) beam. The C60-PL beam entered the scanner of the FLIM and followed the normal optical path of the FLIM, while the SOP steered clear of the scanner and directly entered a finely designed SOP detection channel. Confocal C60-PL images and nonconfocal SOP images were then simultaneously obtained by using laser-scanning mode. Experimental results show that (1) under laser-scanning mode, the obstacle to confocal SOP imaging is the infrared-incompatible scanner, which can be solved by using an infrared-compatible scanner. Confocal SOP imaging is also expected to be realized under stage-scanning mode when the laser beam is parked and meanwhile a pinhole is added into the SOP detection channel. (2) A great challenge to SOP imaging is its extraordinarily long imaging time, and selecting only a few interesting points from fluorescence images to measure their SOP time-dependent traces may be a correct compromise. oxygen absorption microscope (SOAM)2,5 technique which is based on the absorption transition of O2(a1Δg → b1Σg+); and the other is the singlet oxygen phosphorescence microscope (SOPM) technique which is based on SOP. As it is not in the category of SOP, SOAM is not detailed here. The distinct feature of their SOPM is that the image of a sample is acquired by recording SOP signal. To produce SOP, a sensitizer, Sens, is added beforehand into the sample. Under light irradiation, Sens is excited to the singlet excited state of 1Sens*; 1Sens* then becomes the triplet excited state of 3Sens* thought intersystem crossing; finally, 3Sens* transfers its energy to O2(X3Σg−) to form O2(a1Δg). Virtually, their SOPM is similar to a fluorescence microscope in which the imaging is fulfilled by recording fluorescence, the difference only lies in the luminophor. In a fluorescence microscope, the luminophor is 1 Sens*, while in their SOPM it is O2(a1Δg). However, for the two following obvious reasons, SOMP is faced with a much greater challenge than a fluorescence microscope. (1) Because the production of O2(a1Δg) from 1Sens* needs to undergo two
I. INTRODUCTION Singlet oxygen, O2(a1Δg), is the molecular oxygen lying at the lowest electronically excited state (the ground state of O2 is X3Σg−). Since it was discovered in the 1930s, O2(a1Δg) has attracted sustained attention from scientists in the fields of physics, chemistry, photochemistry, and biology, etc., and is still a hot topic.1 Out of various and numerous O2(a1Δg) related researches, one important branch is the detection technique of singlet oxygen phosphorescence (SOP), which results from the infrared radiative transition of O2(a1Δg → X3Σg−) and has the central wavelength of 1268 nm in gaseous phase. Because it contains a lot of useful information about the studied object, directly detecting SOP has become the most commonly used method in the research of O2(a1Δg). Therefore, it is very significant to develop novel SOP detection techniques to further the research of O2(a1Δg). In the past ten years, a most compelling achievement in the field of the detection technique of O2(a1Δg) is the singlet oxygen microscope (SOM) technique proposed and developed by Ogilby et al.2−4 By combining the optics for the detection of O2(a1Δg) with microscopy, micron-order spatial resolution for the detection of O2(a1Δg) was achieved in SOM. Two types of SOM technique were studied by Ogilby et al.: one is the singlet © XXXX American Chemical Society
Received: February 13, 2015 Revised: March 13, 2015
A
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A additional steps, namely, intersystem crossing and energy transfer, and SOP has an extraordinarily long spontaneous radiative lifetime (about 9 s in water6 and 72 min or so in gas phase7) compared to the fluorescence of 1Sens* whose spontaneous radiative lifetime is generally 10−9sec order, SOP is far weaker than that of the fluorescence of 1Sens*, with the former being only about one 109th of the latter or even less. (2) SOP is infrared and the fluorescence of 1Sens* is often visible; for a photoelectric detector, the response in infrared region is generally two orders lower than that in visible region. In 2002, the first SOPM was successfully demonstrated on a modified inverted microscope with Xe-lamp as excitation light source by Ogilby et al. By using this SOPM, they obtained a SOP intensity image with 2.5-μm resolution for the sample of phase-separated toluene/water mixture.8 This SOPM is the same as a general fluorescence microscope in the aspect of structure except that the dichroic mirror and filter used in it are customized for SOP and its imaging sensor at image plane is an InGaAs linear array detector. The SOP intensity image is obtained by piecing together a series of one-dimensional “slices” recorded by the linear array detector when the sample on the microscope stage is translated. Soon after, they used this SOPM again to study the effect of O2(a1Δg) diffusion across the interfacial boundary in phase-separated liquids and polymers,9 and before long they realized SOP intensity imaging for single nerve cells.10 They have also developed a dithered sampling technique to increase the spatial resolution of SOP images.11 In addition, based on the two-photon excitation of a sensitizer, they have also studied the feasibility to construct a two-photon SOPM.2,12 However, SOP intensity image obtained by a SOPM can only represent the distribution of SOP intensity, or O2(a1Δg) concentration in a sample. In a study that concerns the interaction of O2(a1Δg) with the sample, the distribution of SOP lifetime, or a SOP lifetime image is in fact more interesting, for people can extract more important dynamics information about O2(a1Δg) from the SOP lifetime image. The significance of SOP lifetime imaging is to SOP intensity imaging what the significance of fluorescence lifetime imaging is to fluorescence intensity imaging. In fact, to study the O2(a1Δg) dynamics in cell, Ogilby et al. have already developed a subcellular time-resolved SOP detection technique by combining SOPM with single photon counting technique,13−16 but they only performed the measurements of the SOP traces and lifetimes at a few points or some parts (e.g., nucleus and cytoplasm) in a cell, not the SOP lifetime imaging. So far, SOP lifetime imaging has not been reported. Just like a fluorescence lifetime microscope (FLIM), which is the upgrade of a fluorescence microscope, theoretically, as long as it has the ability of point-scanning and the ability to record the time-dependent SOP signal trace for every point, a SOPM will be allowed to be upgraded to a singlet oxygen phosphorescence lifetime imaging microscope (SOPLIM) and SOP lifetime imaging will be realized. These two abilities are very readily found in a commercially mature FLIM. Thus, in this work we used a modified FLIM to study the feasibility of SOP lifetime imaging.
Figure 1. Schematic of experimental SOPLIM setup.
picosecond pulsed laser, the detectors D1, D2, and D3, the ultrafast time-correlated single photon counting (TCSPC) system (B&H Simple-Tau-153-DX-7, Germany), and the PC and software (B&H SPC-Image, Germany). The inverted microscope provides the stage and the objective. The objective focuses the excitation laser (bold solid line) coming out of the scanner onto the sample lying on the stage, and the resulted fluorescence (bold and thin dash line) and SOP (thin solid line) are then collected by the same objective and sent into the scanner along the light path of the excitation laser. Controlled by the scan controller card that is installed in the PC, the scanner guides the excitation laser emitted from the 405-nm picosecond pulsed laser to the objective, and the resulted fluorescence (bold and thin dash line) and SOP (thin solid line) to the detectors D1, D2, and D3 respectively. Detector D1 is a liquid-nitrogen-cooled infrared photomultiplier tube (IRPMT R5509-43, Hamamatsu, Japan) and is used together with an 800-nm cutoff filter glass to specifically detect the SOP in the wavelength range of 1240−1340 nm. Detectors D3 (HPM100-50, B&H, Germany) and D2 (HPM-100-40, B&H, Germany), with their response wavelength ranges being 380− 890 nm and 300−720 nm, respectively, are high-speed hybrid PMT detectors and used to detect the UV−visible fluorescence from the sample. Point-scanning function is realized when the galvanometer mirror of the scanner moves the focus of laser on the focal plane of the objective by altering the direction of the laser beam. After entering into the scanner, the fluorescence (bold and thin dash line) and the SOP (thin solid line) pass through the pinholes and arrive at the detectors. The resulted electric signals of the detectors are then acquired by the TCSPC system. The scan controller card and the TCSPC system are driven by the PC and software, and the data
II. EXPERIMENTAL SETUP As shown in Figure 1, the experimental SOPLIM setup is in fact a modified laser-scanning confocal FLIM. It consists of six subsystems: the inverted microscope (Olympus IX81, Japan), the scanner (B&H DCS120 system, Germany), the 405-nm B
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A acquired by the TCSPC system are handled by the software SPC-Image to simultaneously generate the intensity and lifetime images of the sample, as well as the time-dependent luminescence trace for every point of these images. Two schemes, nonconfocal SOPLIM and confocal SOPLIM, were designed to perform SOP imaging. In nonconfocal SOPLIM scheme, detector D1 is placed at location 1, and meanwhile, a dichroic mirror DM1 that reflects infrared and transmits visible light is placed between the objective and the tube lens of the scanner to separate fluorescence and SOP; then, the SOP is focused by lens 1 onto the detector D1, and the fluorescence is conducted to detectors D2 and D3 along the optical path of the scanner. The function of lens 1 is to ensure that all the SOP reflected by the dichroic mirror DM1 can be collected by detector D1. In confocal SOPLIM scheme, the dichroic mirror DM1 and the reflecting mirror RM1 that appear in nonconfocal SOPLIM scheme are removed, and detector D1 is placed at location 2 and behind the pinhole PH1 of the scanner, and SOP needs to be transmitted in the scanner before it is detected by the detector D1. This is the simplest scheme for upgrading a FLIM to a SOPLIM, for the confocal SOPLIM is no more than a general FLIM in which one of its detection channels that should have been used to detect fluorescence is changed to detect SOP. Compared to the confocal SOPLIM, the advantage of the nonconfocal SOPLIM lies in that the SOP can be detected as much as possible because the SOP loss in the scanner has been avoided. The sample on the stage is a layer of loose C60 powder (purity 99.9%, Sigma-Aldrich), formed by randomly splashing C60 powder over a slide (see Figure 2). The reason C60 is selected lies in that (1) solid C60 itself is a high efficient sensitizer17 and very easy to produce O2(a1Δg) under ambient atmosphere. (2) According to ref 18, the O2 molecules in C60 solid are trapped in the octahedral interstices of the facecentered-cubic lattice, and the time needed for O2 molecules diffusing out of a C60 film with 200 Å thickness is as long as about 40 s (half-life) at room temperature. Thus, it can be considered that the number of O2(a1Δg) molecules produced in a point is “fixed” during the period (28.8 ms, see below) when the SOP lifetime is measured for this point. Therefore, the spatial heterogeneity of C60 sample can be labeled with the distribution of O2(a1Δg), or in other words, the singlet oxygen imaging for C60 sample becomes possible. The photoluminescence of C60 (C60-PL) is a kind of fluorescence. The C60-PL in 600−900 nm range is detected by the detector D3, and other luminescence shorter than 600 nm is detected by the D2. The numerical aperture and magnification of the objective (UPLSAPO, Olympus, Japan) are 0.16 and 4×, respectively. The pulse width and repetition rate of the picosecond pulsed laser are about 100 ps and 80 MHz, respectively. Measured with a power meter, the laser power at the focus of the objective is 0.187 mW, corresponding to 2.34 pJ/pulse. The spectra of SOP and C60-PL were acquired in nonconfocal SOPLIM scheme by using an IR-CCD spectrograph that substituted for the detector D1 at location 1. The IRCCD spectrograph consists of a monochromator (SpectraPro2500i, Acton Research Co., USA) and a liquid-nitrogen-cooled infrared charge coupled device (IR-CCD) camera (7498-0001, Acton Research Co., USA). When the spectrum of SOP was acquired, an 800-nm cutoff filter glass was placed in front of the IR-CCD spectrograph to eliminate the interference of the second-order diffraction of visible light on grating; and when the spectrum of C60-PL was acquired, the filter was removed
Figure 2. Lifetime and intensity images of SOP and C60-PL. Images A and C are the intensity image and lifetime image of SOP, respectively; histogram E is the lifetime distribution of image C. Images B and D are the intensity image and lifetime image of C60-PL, respectively; histogram F is the lifetime distribution of image D.
and meanwhile the dichroic mirror DM1 was replaced by a 50/ 50 beam splitter. The C60-PL lifetime was measured by TCSPC technique. Note that the laser pulse period is 12.5 ns. It is out of question to measure the C60-PL lifetime (several hundred picoseconds, see below) and a general fluorescence lifetime (several nanoseconds order) under the repetition rate of 80 MHz. However, TCSPC technique is not suitable for the SOP lifetime measurement because the SOP lifetime in C60 powder is as long as several milliseconds (see below), much longer than the pulse period. Fortunately, the technique that simultaneously records the lifetime images of short-lifetime fluorescence and long-lifetime phosphorescence in a laser scanning system has been maturely commercialized at present.19 Here it is briefly described as follows. Suppose that the resident time of laser beam at a point is Tpxl. Tpxl can be divided into two parts, Ton and Toff, that is Tpxl = Ton + Toff, and Toff is much greater than Ton. Within the time of Ton, the laser is turned on and the short fluorescence lifetime is measured with TCSPC technique, and meanwhile the luminophor of the phosphorescence is accumulated on the irradiation of the laser; then, within the time of Toff, the laser is turned off and the luminophor of the phosphorescence accumulated within the time of Ton begins to decay, and the corresponding phosphorescence decay trace is C
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
complementary relationship means that coupling may have taken place between the first and second pathways, and the factor that leads to the coupling is likely that the energy transfer from 3C60* to O2 can affect the ISC of 1C60* and thus indirectly affect the fluorescence emission of 1C60*. As mentioned above, the O2 molecules in C60 solid are trapped in the octahedral interstices of the face-centered-cubic lattice. Very close contact should exist between the O2 and C60 molecules, and it can not be hard to imagine that their interaction would certainly have effect on the ISC of 1C60*. Of course, this supposition can also be further confirmed from C60-PL lifetime (see below). Similar complementary relationship also exists between the SOP intensity distribution (Figure 2A) and the SOP lifetime distribution (Figure 2C), as well as the C60-PL intensity distribution (Figure 2B) and the C60-PL lifetime distribution (Figure 2D). For the SOP images, the high intensity regions are basically the short lifetime regions and vice versa; for example, the bright white (high intensity) and the dark (low intensity) regions in Figure 2A correspond, respectively, to the orange (short lifetime) and blue (long lifetime) regions in Figure 2C. The reason probably lies in that the high intensity regions contain more O2 so that O2(a1Δg) is easier to be quenched through the process O2(a1Δg) + O2(X3Σg−) → 2 O2(X3Σg−), and in addition, the produced higher O2(a1Δg) concentration also quickens the quenching process 2O2(a1Δg) →O2(b1Σg+) + 2O2(X3Σg−), and therefore the SOP lifetime is shortened. For the C60-PL images, their complementary relationship is right converse to that of the SOP images, that is, the high intensity regions are roughly the long lifetime regions and vice versa; for example, the brightest (high intensity) and the dark (low intensity) regions in Figure 2B correspond respectively to the blue (long lifetime) and yellow−brown (short lifetime) regions in Figure 2D. This is because the bright regions mean less O2 and lower possibility of the energy transfer from 3C60* to O2, therefore, the loss of 1C60* through the second pathway will be mitigated and the C60-PL lifetime be prolonged. Further, the fact that the C60-PL lifetime has relationship with the concentration of O2 also indicates that O2 may affect the ISC of 1C60*. To further confirm that Figure 2A and 2C indeed result from SOP imaging, the SOP lifetime and spectrum obtained in this work are compared with those reported in literature. Figure 2E shows that the SOP lifetime in this work distributes over the range of 3.00−4.60 ms with its mean value being 3.70 ms. These values agree with the SOP lifetime of about 4 ms in crystalline C60 at the temperature of 300 K20 and the SOP lifetime of 2−3 ms in aqueous C60 suspensions.21 Figure 3 is the SOP spectrum aquired by placing the IR-CCD spectrograph at location 1. The shape and peak position of this SOP spectrum is basically the same as that of the SOP spectrum reported in refs 18 and 21. These comparisons clearly indicate that Figure 2A and 2C are indeed SOP images. Similar comparisons are also applied in C60-PL. Figure 2F shows that the C60-PL lifetime distributes over the range of 310−400 ps with its mean value being 347 ps. Reference 22 has shown that (1) the fluorescence lifetime of 1C 60* is predominantly determined by ISC, and the ISC of 1C60* is independent of temperature. (2) The fluorescence lifetime of 1 C60* is 1.5 ns in Ar matrices at 4 K, 1.45 ns in room temperature toluene, hexane, and benzene solutions, 1.2 ns in solid C60 both at room and at low temperature, and dramatically drops to 0.9 ns in Ne matrices at 4 K for the reason that the ISC is affected by Ne. Note that the above
recorded by the multichannel scaler integrated in the TCSPC system.
III. RESULTS AND DISCUSSION The confocal SOPLIM scheme was first tested. But none of SOP signal was detected, even the pinhole PH1 had been fully opened. The reason should be that the SOP has been completely decayed by the inner optical elements of the scanner. Therefore, the nonconfocal SOPLIM scheme had to be used to avoid the loss of SOP in the scanner. In the nonconfocal SOPLIM scheme, the SOP images were acquired by detector D1, and the C60-PL images longer and shorter than the wavelength of 600 nm were aquired by detectors D3 and D2, respectively. The experimental results are shown in Figure 2. No C60-PL images shorter than the wavelength of 600 nm were detected. For convenience of comparison, the C60-PL images and the SOP images are put together. The pixel number of these images is 64 × 64 and corresponds to a 1150 μm × 1150 μm region of the sample. The resident time of laser beam on a pixel is Tpxl = 28.8000 ms, within which the laser-on time is Ton = 5.0625 ms and the laser-off time is Toff = 23.7375 ms, and 117.965 s is needed to complete a single-frame image. To improve the signal-to-noise ratio, the images in Figure 2 were all generated by the accumulation of five frames, which leads to the acquisition time for an image is 589.824 s or 9.8304 min. From Figure 2, one can see that SOP intensity image (Figure 2A) and SOP lifetime image (Figure 2C) have been successfully obtained by the nonconfocal SOPLIM. Compared with the corresponding C60-PL images (Figure 2B and 2D), the definition of SOP image is obviously lower than that of C60PL image, but the shape of C60 granules still can be recognized in the SOP images. It should be emphasized that the low definition of the SOP image is resulted from the nonconfocal imaging and should be able to be improved by using the following confocal imaging method. That is the laser beam is kept stationary and meanwhile a pinhole is added between lens 1 and detector D1 to block the SOP that does not come from the focus of the objective, and then confocal SOP imaging is performed by using stage-scanning mode instead of laserscanning mode. The essence of this method is that the definition of SOP image is compensated by sacrificing its imaging speed, for stage-scanning is much slower than laserscanning. The SOP intensity distribution in Figure 2A and the C60-PL intensity distribution in Figure 2B approximately complement each other, that is, the relatively bright regions in Figure 2A correspond to the relatively dark regions in Figure 2B and vice versa. This phenomenon can be interpreted by the conservation of energy. According to ref 17, after the singlet excited state 1 C60* is produced by the excitation of C60, the formed 1C60* will release its energy through the following two pathways. The first pathway is to emit fluorescence (C60-PL), and the second is that the 1C60* first forms the triplet excited state 3C60* through intersystem crossing (ISC) and then the 3C60* transfer its energy to O2 to produce O2(a1Δg). These two pathways compete between each other and lead to the complementary relationship between the intensity distributions of SOP and C60-PL. The relatively bright regions in Figure 2A (or the relatively dark region in Figure 2B) are just the regions that have high O2(a1Δg) concentrations. These regions may contain more O2 so that the energy of 3C60* is more conducive to accepted by O2, and meanwhile the fluorescence (C60-PL) emissions are more reduced. The phenomenon of the D
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
infrared-incompatible scanner that is suitable only for UV− visible light. Therefore, it can be reasonable to anticipate that if the optical elements in the scanner are all infrared-compatible, the SOP should be detectable at location 2 and a confocal SOPLIM should be able to be established. Given that the lenses, reflecting mirrors, and galvanometer mirror in the scanner are all more or less infrared-compatible, the key to an infrared-compatible scanner is likely the dichroic mirror DM2. The requirement for this mirror is that it can highly reflect the UV−visible excitation light and meanwhile highly transmit SOP, just like the dichroic mirror that is used by Ogilby et al. in their SOM.2 It needs to be pointed out that even if the confocal SOPLIM scheme can be realized, SOP lifetime imaging is still faced with a great challenge, which is imaging time. As mentioned in the Introduction, SOP is very weak, and ultralong acquisition time has to be consumed in order to get a clear SOP image. For example, the imaging time for the SOP images in Figure 2 is 9.8304 min; for hydrous samples such as cells or other organisms, it is easy to predict that the imaging time will be unendurably long and reach 9830 min based on the fact that the SOP lifetime in H2O is only 3.5 μs,24 over 1000 times shorter than the lifetime of 3.7 ms in C60 solid and resulting in an about 1000 times decrease in the quantum efficiency of SOP. Moreover, given that the concentration and sensitization efficiency of a hydrous-environment-used sensitizer are generally lower than those of C60 solid, the needed imaging time will become even longer. In addition to extraordinarily long imaging time, another disadvantage for SOP imaging is its relatively low image resolution that results from the relatively long wavelength of SOP (about 2 times longer than that of visible light). To avoid long SOP imaging time, a useful tactic is that fluorescence images are first acquired, and then SOP lifetime images are acquired for only a few interesting small domains selected from the fluorescence images, or SOP trace and lifetime are measured for only a few interesting points. For example, obvious difference of SOP lifetime may exist between different areas of a cell such as nucleus and cytoplasm, but this difference may be very slight at a roughly uniform area, so much so that measurements of SOP trace and lifetime for only a few points at different areas may be enough for a specific study. Figure 5 is just a case in point. Viewed from the C60-PL images, the small areas around the points a, b, c, and d look much
Figure 3. SOP spectrum from C60 powder.
lifetime values in ref 22 were all measured in oxygen-free environments, and in any case, it can be sure that our value of 347 ps is much smaller than these values. One explanation for such large lifetime difference, we think, is that the ISC is affected by O2 and thus shortens the fluorescence lifetime of 1 C60*. After all, according to ref 18, the concentration of O2 in C60 solid is as high as 0.233M; O2 should have the ability to affect the ISC just like Ne matrix can affect the ISC. In a word, by affecting the ISC, O2 is able to indirectly affect the fluorescence lifetime of 1C60*. Figure 4 is the C60-PL spectrum acquired by placing the IRCCD spectrograph at location 1. This C60-PL spectrum agrees
Figure 4. C60-PL spectrum from C60 powder.
well with the luminescence spectrum of the C60 powder at room temperature reported in ref 23, corresponding to the broad peak of the luminescence spectrum of C60 in the wavelength range of 750−820 nm (1.65−1.51 eV), with their maximums being both at the position of about 780 nm or 1.59 eV. In addition, ref 23 has shown that at short-wavelength direction the luminescence spectrum of C60 has decayed to zero at the position of 653 nm (1.9 eV), which also explains why no light signal was detected by the detector D2. As shown in Figure 1, when detecting SOP at location 2 in the confocal SOPLIM scheme, SOP needs to propagate in the scanner, and when at location 1 in the nonconfocal SOPLIM scheme, it does not. The fact that SOP can be detected at location 1 and can not be detected at location 2 shows that the scanner has greatly attenuated the SOP and is obviously an
Figure 5. SOP trace and decay lifetime for a single point. Traces a, b, c, and d correspond respectively to the points a, b, c, and d in Figure 2. E
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
■
different from each other but yet roughly uniform within every individual area, and their differences can indeed be seen from the corresponding SOP images; so the SOP dynamics in these areas can be represented by the SOP traces and lifetimes at points a, b, c, and d in Figure 2. These SOP traces, in fact, can also be acquired by using the single-point measurement function in a FLIM; this function allows one to park the laser beam at a specific point to acquire the fluorescence or phosphorescence trace for this point. This tactic in fact has been first used by Ogilby et al. when they measured the SOP trace and lifetime for small subcellular domains (nucleus and cytoplasm) selected from the fluorescence image of a cell.3,13,14 Figure 5 also clearly shows that with the excitation laser being turned on, O2(a1Δg) is accumulated during the time of Ton, which leads to the SOP intensity going up; then with the excitation laser being turned off, the SOP intensity begins to decay during the time of Toff. The lifetime image of SOP in Figure 2 is just built up by using the decay lifetime for every single point. Note that the ISC of 1C60* and the energy transfer from 3C60* to O2 (with the rate constant of 2 × 109 M−1 s−1)17are both fast processes, and O2(a1Δg) can be quickly produced upon the laser excitation of C60. During the laser-on period, O2(a1Δg) is produced at the arrival of every laser pulse and then decays with the lifetime of about 3.70 ms until the next arrival of laser pulse. The pulse interval is 12.5 ns, obviously far shorter than the decay lifetime time of 3.70 ms, so the concentration of O2(a1Δg) can be continually accumulated and the SOP intensity continually rises during the laser-on period.
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: +86-411-84379010; e-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by a grant from the National Natural Science Foundation of China (grant 21173217). REFERENCES
(1) Ogilby, P. R. Singlet oxygen: there is indeed something new under the sun. Chem. Soc. Rev. 2010, 39, 3181−3209. (2) Snyder, J. W.; Zebger, I.; Gao, Z.; Poulsen, L.; Frederiksen, P. K.; Skovsen, E.; Mcilroy, S. P.; Klinger, M.; Andersen, L. K.; Ogilby, P. R. Singlet oxygen microscope: from phase-separated polymers to single biological cells. Acc. Chem. Res. 2004, 37, 894−901. (3) Snyder, J. W.; Skovsen, E.; Lambert, J. D. C.; Poulsen, L.; Ogilby, P. R. Optical detection of singlet oxygen from single cells. Phys. Chem. Chem. Phys. 2006, 8, 4280−4293. (4) Breitenbach, T.; Kuimova, M. K.; Gbur, P.; Hatz, S.; Schack, N. B.; Pedersen, B. W.; Lambert, J. D. C.; Poulsen, L.; Ogilby, P. R. Photosensitized production of singlet oxygen: spatially-resolved optical studies in single cells. Photochem. Photobiol. Sci. 2009, 8, 442−452. (5) Andersen, L. K.; Ogilby, P. R. Time-resolved detection of singlet oxygen in a transmission microscope. Photochem. Photobiol. 2001, 73 (5), 489−492. (6) Poulsen, T. D.; Ogilby, P. R. Solvent effects on the O2(a1Δg)− O2(X3Σg−) radiative transition: Comments regarding charge-transfer interactions. J. Phys. Chem. A 1998, 102, 9829−9832. (7) Miller, H. C.; McCord, J. E.; Choy, J.; Hager, G. D. Measurement of the radiative lifetime of O2(a1Δg) using cavity ring down spectroscopy. J. Quant. Spectrosc. Radiat. Transfer 2001, 69, 305−325. (8) Andersen, L. K.; Gao, Z.; Ogilby, P. R.; Poulsen, L.; Zebger, I. A Singlet oxygen image with 2.5 μm resolution. J. Phys. Chem. A 2002, 106, 8488−8490. (9) Zebger, I.; Poulsen, L.; Gao, Z.; Andersen, L. K.; Ogilby, P. R. Singlet oxygen images of heterogeneous samples: Examining the effect of singlet oxygen diffusion across the interfacial boundary in phaseseparated liquids and polymers. Langmuir 2003, 19, 8927−8933. (10) Zebger, I.; Snyder, J. W.; Andersen, L. K.; Poulsen, L.; Gao, Z.; Lambert, J. D. C.; Kristiansen, U.; Ogilby, P. R. Direct optical detection of singlet oxygen from a single cell. Photochem. Photobiol. 2004, 79, 319−322. (11) Snyder, J. W.; Gao, Z.; Ogilby, P. R. Application of a dithered sampling technique to increase the spatial resolution of singlet oxygen images. Rev. Sci. Instrum. 2005, 76, 013701. (12) Skovsen, E.; Snyder, J. W.; Ogilby, P. R. Two-photon singlet oxygen microscopy: The challenges of working with single cells. Photochem. Photobiol. 2006, 82, 1187−1197. (13) Skovsen, E.; Snyder, J. W.; Lambert, J. D. C.; Ogilby, P. R. Lifetime and diffusion of singlet oxygen in a cell. J. Phys. Chem. B 2005, 109, 8570−8573. (14) Snyder, J. W.; Skovsen, E.; Lambert, J. D. C.; Ogilby, P. R. Subcellular, time-resolved studies of singlet oxygen in single cells. J. Am. Chem. Soc. 2005, 127, 14558−14559. (15) Hatz, S.; Lambertb, J. D. C.; Ogilby, P. R. Measuring the lifetime of singlet oxygen in a single cell: addressing the issue of cell viability. Photochem. Photobiol. Sci. 2007, 6, 1106−1116. (16) Kuimova, M. K.; Yahioglu, G.; Ogilby, P. R. Singlet oxygen in a cell: Spatially dependent lifetimes and quenching rate constants. J. Am. Chem. Soc. 2009, 131, 332−340. (17) Arbogast, J. W.; Darmanyan, A. P.; Foote, C. S.; Rubin, Y.; Diederich, F. N.; Alvarez, M. M.; Anz, S. J.; Whetten, R. L. Photophysical properties of C60. J. Phys. Chem. 1991, 95, 11−12. (18) Howells, S. C.; Black, G.; Schlie, L. A. O2(a1Δg) production and oxygen diffusion in C60 films. Synth. Met. 1994, 62, 1−7.
IV. CONCLUSION Nonconfocal SOP imaging, including intensity and lifetime imaging, is very easy to be realized on a general FLIM by using a sample of C60 powder. Analyses show that the obstacle to confocal SOP imaging is the great attenuation of SOP resulted from the scanner. Therefore, it is reasonable to anticipate that the two following solutions can be adopted to fulfill confocal SOP imaging. The first solution is that the general UV−visible scanner of a FLIM is replaced by an infrared-compatible scanner, and then the SOP detector is placed behind one of the pinholes of the infrared-compatible scanner. The SOPLIM in this solution in fact is no more than an infrared-compatible FLIM, in which both laser-scanning mode and stage-scanning mode can be used to perform confocal SOP imaging. The second solution is that as shown in Figure 1, the SOP returning from the objective is directly introduced into an optical path that is specifically designed for SOP detection by placing a dichroic mirror that reflects SOP and transmits visible light between the objective and the tube lens of a FLIM; meanwhile, a pinhole is added into the optical path and placed behind lens 1 to block the SOP emitted from off-focus positions. Finally, in the condition that the laser beam is kept stationary, confocal SOP imaging is performed by using stage-scanning mode. For the reason that the conjugate relationship between the two focal points of the objective and lens 1 needs to be maintained, only stage-scanning mode can be used for the SOPLIM in the second solution. For hydrous samples, another obstacle to confocal SOP imaging is its ultralong imaging time; and selecting only a few interesting points from fluorescence images to measure their SOP time-dependent traces may be a correct compromise. F
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A (19) Becker & Hickl GmbH. Microsecond decay FLIM: Combined fluorescence and phosphorescence lifetime imaging, Application Note; 2010. www.becker-hickl.com. (20) Nissen, M. K.; Wilson, S. M.; Thewalt, M. L. W. Highly structured singlet oxygen photoluminescence from crystalline C60. Phys. Rev. Lett. 1992, 69, 2423−2426. (21) Wang, J.; Leng, J.; Yang, H.; Sha, G.; Zhang, C. Long-lifetime and asymmetric singlet oxygen photoluminescence from aqueous fullerene suspensions. Langmuir 2013, 29, 9051−9056. (22) Sassara, A.; Zerza, G.; Chergui, M.; Ciulin, V.; Ganière, J. D.; Deveaud, B. Picosecond studies of the intramolecular relaxation processes in isolated C60 and C70 molecules. J. Chem. Phys. 1999, 111, 689−697. (23) Byrne, H. J.; Maser, W.; Rühle, W. W.; Mittelbach, A.; Hönle, W.; Schnering, H. G. V.; Movaghar, B.; Roth, S. Time-resolved photoluminescence of solid state fullerenes. Chem. Phys. Lett. 1993, 204, 461−466. (24) Egorov, S. Y.; Kamalov, V. F.; Koroteev, N. I.; Krasnovsky, A. A., Jr.; Toleutaev, B. N.; Zinukov, S. V. Rise and decay kinetics of photosensitized singlet oxygen luminescence in water. Measurements with nanosecond time-correlated single photon counting technique. Chem. Phys. Lett. 1989, 163, 421−424.
G
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Singlet Oxygen Phosphorescence Lifetime Imaging Based on a Fluorescence Lifetime Imaging Microscope Wenming Tian,† Liezheng Deng,*,† Shengye Jin,† Heping Yang,† Rongrong Cui,† Qing Zhang,‡ Wenbo Shi,† Chunlei Zhang,‡ Xiaolin Yuan,‡ and Guohe Sha† †
State Key Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, Liaoning 116023, China ‡ Research Center, Affiliated Zhongshan Hospital of Dalian University, Dalian, Liaoning 116001, China ABSTRACT: The feasibility of singlet oxygen phosphorescence (SOP) lifetime imaging microscope was studied on a modified fluorescence lifetime imaging microscope (FLIM). SOP results from the infrared radiative transition of O2(a1Δg → X3Σg−) and O2(a1Δg) was produced in a C60 powder sample via photosensitization process. To capture the very weak SOP signal, a dichroic mirror was placed between the objective and tube lens of the FLIM and used to divide the luminescence returning from the sample into two beams: the reflected SOP beam and the transmitted photoluminescence of C60 (C60-PL) beam. The C60-PL beam entered the scanner of the FLIM and followed the normal optical path of the FLIM, while the SOP steered clear of the scanner and directly entered a finely designed SOP detection channel. Confocal C60-PL images and nonconfocal SOP images were then simultaneously obtained by using laser-scanning mode. Experimental results show that (1) under laser-scanning mode, the obstacle to confocal SOP imaging is the infrared-incompatible scanner, which can be solved by using an infrared-compatible scanner. Confocal SOP imaging is also expected to be realized under stage-scanning mode when the laser beam is parked and meanwhile a pinhole is added into the SOP detection channel. (2) A great challenge to SOP imaging is its extraordinarily long imaging time, and selecting only a few interesting points from fluorescence images to measure their SOP time-dependent traces may be a correct compromise. oxygen absorption microscope (SOAM)2,5 technique which is based on the absorption transition of O2(a1Δg → b1Σg+); and the other is the singlet oxygen phosphorescence microscope (SOPM) technique which is based on SOP. As it is not in the category of SOP, SOAM is not detailed here. The distinct feature of their SOPM is that the image of a sample is acquired by recording SOP signal. To produce SOP, a sensitizer, Sens, is added beforehand into the sample. Under light irradiation, Sens is excited to the singlet excited state of 1Sens*; 1Sens* then becomes the triplet excited state of 3Sens* thought intersystem crossing; finally, 3Sens* transfers its energy to O2(X3Σg−) to form O2(a1Δg). Virtually, their SOPM is similar to a fluorescence microscope in which the imaging is fulfilled by recording fluorescence, the difference only lies in the luminophor. In a fluorescence microscope, the luminophor is 1 Sens*, while in their SOPM it is O2(a1Δg). However, for the two following obvious reasons, SOMP is faced with a much greater challenge than a fluorescence microscope. (1) Because the production of O2(a1Δg) from 1Sens* needs to undergo two
I. INTRODUCTION Singlet oxygen, O2(a1Δg), is the molecular oxygen lying at the lowest electronically excited state (the ground state of O2 is X3Σg−). Since it was discovered in the 1930s, O2(a1Δg) has attracted sustained attention from scientists in the fields of physics, chemistry, photochemistry, and biology, etc., and is still a hot topic.1 Out of various and numerous O2(a1Δg) related researches, one important branch is the detection technique of singlet oxygen phosphorescence (SOP), which results from the infrared radiative transition of O2(a1Δg → X3Σg−) and has the central wavelength of 1268 nm in gaseous phase. Because it contains a lot of useful information about the studied object, directly detecting SOP has become the most commonly used method in the research of O2(a1Δg). Therefore, it is very significant to develop novel SOP detection techniques to further the research of O2(a1Δg). In the past ten years, a most compelling achievement in the field of the detection technique of O2(a1Δg) is the singlet oxygen microscope (SOM) technique proposed and developed by Ogilby et al.2−4 By combining the optics for the detection of O2(a1Δg) with microscopy, micron-order spatial resolution for the detection of O2(a1Δg) was achieved in SOM. Two types of SOM technique were studied by Ogilby et al.: one is the singlet © XXXX American Chemical Society
Received: February 13, 2015 Revised: March 13, 2015
A
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A additional steps, namely, intersystem crossing and energy transfer, and SOP has an extraordinarily long spontaneous radiative lifetime (about 9 s in water6 and 72 min or so in gas phase7) compared to the fluorescence of 1Sens* whose spontaneous radiative lifetime is generally 10−9sec order, SOP is far weaker than that of the fluorescence of 1Sens*, with the former being only about one 109th of the latter or even less. (2) SOP is infrared and the fluorescence of 1Sens* is often visible; for a photoelectric detector, the response in infrared region is generally two orders lower than that in visible region. In 2002, the first SOPM was successfully demonstrated on a modified inverted microscope with Xe-lamp as excitation light source by Ogilby et al. By using this SOPM, they obtained a SOP intensity image with 2.5-μm resolution for the sample of phase-separated toluene/water mixture.8 This SOPM is the same as a general fluorescence microscope in the aspect of structure except that the dichroic mirror and filter used in it are customized for SOP and its imaging sensor at image plane is an InGaAs linear array detector. The SOP intensity image is obtained by piecing together a series of one-dimensional “slices” recorded by the linear array detector when the sample on the microscope stage is translated. Soon after, they used this SOPM again to study the effect of O2(a1Δg) diffusion across the interfacial boundary in phase-separated liquids and polymers,9 and before long they realized SOP intensity imaging for single nerve cells.10 They have also developed a dithered sampling technique to increase the spatial resolution of SOP images.11 In addition, based on the two-photon excitation of a sensitizer, they have also studied the feasibility to construct a two-photon SOPM.2,12 However, SOP intensity image obtained by a SOPM can only represent the distribution of SOP intensity, or O2(a1Δg) concentration in a sample. In a study that concerns the interaction of O2(a1Δg) with the sample, the distribution of SOP lifetime, or a SOP lifetime image is in fact more interesting, for people can extract more important dynamics information about O2(a1Δg) from the SOP lifetime image. The significance of SOP lifetime imaging is to SOP intensity imaging what the significance of fluorescence lifetime imaging is to fluorescence intensity imaging. In fact, to study the O2(a1Δg) dynamics in cell, Ogilby et al. have already developed a subcellular time-resolved SOP detection technique by combining SOPM with single photon counting technique,13−16 but they only performed the measurements of the SOP traces and lifetimes at a few points or some parts (e.g., nucleus and cytoplasm) in a cell, not the SOP lifetime imaging. So far, SOP lifetime imaging has not been reported. Just like a fluorescence lifetime microscope (FLIM), which is the upgrade of a fluorescence microscope, theoretically, as long as it has the ability of point-scanning and the ability to record the time-dependent SOP signal trace for every point, a SOPM will be allowed to be upgraded to a singlet oxygen phosphorescence lifetime imaging microscope (SOPLIM) and SOP lifetime imaging will be realized. These two abilities are very readily found in a commercially mature FLIM. Thus, in this work we used a modified FLIM to study the feasibility of SOP lifetime imaging.
Figure 1. Schematic of experimental SOPLIM setup.
picosecond pulsed laser, the detectors D1, D2, and D3, the ultrafast time-correlated single photon counting (TCSPC) system (B&H Simple-Tau-153-DX-7, Germany), and the PC and software (B&H SPC-Image, Germany). The inverted microscope provides the stage and the objective. The objective focuses the excitation laser (bold solid line) coming out of the scanner onto the sample lying on the stage, and the resulted fluorescence (bold and thin dash line) and SOP (thin solid line) are then collected by the same objective and sent into the scanner along the light path of the excitation laser. Controlled by the scan controller card that is installed in the PC, the scanner guides the excitation laser emitted from the 405-nm picosecond pulsed laser to the objective, and the resulted fluorescence (bold and thin dash line) and SOP (thin solid line) to the detectors D1, D2, and D3 respectively. Detector D1 is a liquid-nitrogen-cooled infrared photomultiplier tube (IRPMT R5509-43, Hamamatsu, Japan) and is used together with an 800-nm cutoff filter glass to specifically detect the SOP in the wavelength range of 1240−1340 nm. Detectors D3 (HPM100-50, B&H, Germany) and D2 (HPM-100-40, B&H, Germany), with their response wavelength ranges being 380− 890 nm and 300−720 nm, respectively, are high-speed hybrid PMT detectors and used to detect the UV−visible fluorescence from the sample. Point-scanning function is realized when the galvanometer mirror of the scanner moves the focus of laser on the focal plane of the objective by altering the direction of the laser beam. After entering into the scanner, the fluorescence (bold and thin dash line) and the SOP (thin solid line) pass through the pinholes and arrive at the detectors. The resulted electric signals of the detectors are then acquired by the TCSPC system. The scan controller card and the TCSPC system are driven by the PC and software, and the data
II. EXPERIMENTAL SETUP As shown in Figure 1, the experimental SOPLIM setup is in fact a modified laser-scanning confocal FLIM. It consists of six subsystems: the inverted microscope (Olympus IX81, Japan), the scanner (B&H DCS120 system, Germany), the 405-nm B
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A acquired by the TCSPC system are handled by the software SPC-Image to simultaneously generate the intensity and lifetime images of the sample, as well as the time-dependent luminescence trace for every point of these images. Two schemes, nonconfocal SOPLIM and confocal SOPLIM, were designed to perform SOP imaging. In nonconfocal SOPLIM scheme, detector D1 is placed at location 1, and meanwhile, a dichroic mirror DM1 that reflects infrared and transmits visible light is placed between the objective and the tube lens of the scanner to separate fluorescence and SOP; then, the SOP is focused by lens 1 onto the detector D1, and the fluorescence is conducted to detectors D2 and D3 along the optical path of the scanner. The function of lens 1 is to ensure that all the SOP reflected by the dichroic mirror DM1 can be collected by detector D1. In confocal SOPLIM scheme, the dichroic mirror DM1 and the reflecting mirror RM1 that appear in nonconfocal SOPLIM scheme are removed, and detector D1 is placed at location 2 and behind the pinhole PH1 of the scanner, and SOP needs to be transmitted in the scanner before it is detected by the detector D1. This is the simplest scheme for upgrading a FLIM to a SOPLIM, for the confocal SOPLIM is no more than a general FLIM in which one of its detection channels that should have been used to detect fluorescence is changed to detect SOP. Compared to the confocal SOPLIM, the advantage of the nonconfocal SOPLIM lies in that the SOP can be detected as much as possible because the SOP loss in the scanner has been avoided. The sample on the stage is a layer of loose C60 powder (purity 99.9%, Sigma-Aldrich), formed by randomly splashing C60 powder over a slide (see Figure 2). The reason C60 is selected lies in that (1) solid C60 itself is a high efficient sensitizer17 and very easy to produce O2(a1Δg) under ambient atmosphere. (2) According to ref 18, the O2 molecules in C60 solid are trapped in the octahedral interstices of the facecentered-cubic lattice, and the time needed for O2 molecules diffusing out of a C60 film with 200 Å thickness is as long as about 40 s (half-life) at room temperature. Thus, it can be considered that the number of O2(a1Δg) molecules produced in a point is “fixed” during the period (28.8 ms, see below) when the SOP lifetime is measured for this point. Therefore, the spatial heterogeneity of C60 sample can be labeled with the distribution of O2(a1Δg), or in other words, the singlet oxygen imaging for C60 sample becomes possible. The photoluminescence of C60 (C60-PL) is a kind of fluorescence. The C60-PL in 600−900 nm range is detected by the detector D3, and other luminescence shorter than 600 nm is detected by the D2. The numerical aperture and magnification of the objective (UPLSAPO, Olympus, Japan) are 0.16 and 4×, respectively. The pulse width and repetition rate of the picosecond pulsed laser are about 100 ps and 80 MHz, respectively. Measured with a power meter, the laser power at the focus of the objective is 0.187 mW, corresponding to 2.34 pJ/pulse. The spectra of SOP and C60-PL were acquired in nonconfocal SOPLIM scheme by using an IR-CCD spectrograph that substituted for the detector D1 at location 1. The IRCCD spectrograph consists of a monochromator (SpectraPro2500i, Acton Research Co., USA) and a liquid-nitrogen-cooled infrared charge coupled device (IR-CCD) camera (7498-0001, Acton Research Co., USA). When the spectrum of SOP was acquired, an 800-nm cutoff filter glass was placed in front of the IR-CCD spectrograph to eliminate the interference of the second-order diffraction of visible light on grating; and when the spectrum of C60-PL was acquired, the filter was removed
Figure 2. Lifetime and intensity images of SOP and C60-PL. Images A and C are the intensity image and lifetime image of SOP, respectively; histogram E is the lifetime distribution of image C. Images B and D are the intensity image and lifetime image of C60-PL, respectively; histogram F is the lifetime distribution of image D.
and meanwhile the dichroic mirror DM1 was replaced by a 50/ 50 beam splitter. The C60-PL lifetime was measured by TCSPC technique. Note that the laser pulse period is 12.5 ns. It is out of question to measure the C60-PL lifetime (several hundred picoseconds, see below) and a general fluorescence lifetime (several nanoseconds order) under the repetition rate of 80 MHz. However, TCSPC technique is not suitable for the SOP lifetime measurement because the SOP lifetime in C60 powder is as long as several milliseconds (see below), much longer than the pulse period. Fortunately, the technique that simultaneously records the lifetime images of short-lifetime fluorescence and long-lifetime phosphorescence in a laser scanning system has been maturely commercialized at present.19 Here it is briefly described as follows. Suppose that the resident time of laser beam at a point is Tpxl. Tpxl can be divided into two parts, Ton and Toff, that is Tpxl = Ton + Toff, and Toff is much greater than Ton. Within the time of Ton, the laser is turned on and the short fluorescence lifetime is measured with TCSPC technique, and meanwhile the luminophor of the phosphorescence is accumulated on the irradiation of the laser; then, within the time of Toff, the laser is turned off and the luminophor of the phosphorescence accumulated within the time of Ton begins to decay, and the corresponding phosphorescence decay trace is C
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
complementary relationship means that coupling may have taken place between the first and second pathways, and the factor that leads to the coupling is likely that the energy transfer from 3C60* to O2 can affect the ISC of 1C60* and thus indirectly affect the fluorescence emission of 1C60*. As mentioned above, the O2 molecules in C60 solid are trapped in the octahedral interstices of the face-centered-cubic lattice. Very close contact should exist between the O2 and C60 molecules, and it can not be hard to imagine that their interaction would certainly have effect on the ISC of 1C60*. Of course, this supposition can also be further confirmed from C60-PL lifetime (see below). Similar complementary relationship also exists between the SOP intensity distribution (Figure 2A) and the SOP lifetime distribution (Figure 2C), as well as the C60-PL intensity distribution (Figure 2B) and the C60-PL lifetime distribution (Figure 2D). For the SOP images, the high intensity regions are basically the short lifetime regions and vice versa; for example, the bright white (high intensity) and the dark (low intensity) regions in Figure 2A correspond, respectively, to the orange (short lifetime) and blue (long lifetime) regions in Figure 2C. The reason probably lies in that the high intensity regions contain more O2 so that O2(a1Δg) is easier to be quenched through the process O2(a1Δg) + O2(X3Σg−) → 2 O2(X3Σg−), and in addition, the produced higher O2(a1Δg) concentration also quickens the quenching process 2O2(a1Δg) →O2(b1Σg+) + 2O2(X3Σg−), and therefore the SOP lifetime is shortened. For the C60-PL images, their complementary relationship is right converse to that of the SOP images, that is, the high intensity regions are roughly the long lifetime regions and vice versa; for example, the brightest (high intensity) and the dark (low intensity) regions in Figure 2B correspond respectively to the blue (long lifetime) and yellow−brown (short lifetime) regions in Figure 2D. This is because the bright regions mean less O2 and lower possibility of the energy transfer from 3C60* to O2, therefore, the loss of 1C60* through the second pathway will be mitigated and the C60-PL lifetime be prolonged. Further, the fact that the C60-PL lifetime has relationship with the concentration of O2 also indicates that O2 may affect the ISC of 1C60*. To further confirm that Figure 2A and 2C indeed result from SOP imaging, the SOP lifetime and spectrum obtained in this work are compared with those reported in literature. Figure 2E shows that the SOP lifetime in this work distributes over the range of 3.00−4.60 ms with its mean value being 3.70 ms. These values agree with the SOP lifetime of about 4 ms in crystalline C60 at the temperature of 300 K20 and the SOP lifetime of 2−3 ms in aqueous C60 suspensions.21 Figure 3 is the SOP spectrum aquired by placing the IR-CCD spectrograph at location 1. The shape and peak position of this SOP spectrum is basically the same as that of the SOP spectrum reported in refs 18 and 21. These comparisons clearly indicate that Figure 2A and 2C are indeed SOP images. Similar comparisons are also applied in C60-PL. Figure 2F shows that the C60-PL lifetime distributes over the range of 310−400 ps with its mean value being 347 ps. Reference 22 has shown that (1) the fluorescence lifetime of 1C 60* is predominantly determined by ISC, and the ISC of 1C60* is independent of temperature. (2) The fluorescence lifetime of 1 C60* is 1.5 ns in Ar matrices at 4 K, 1.45 ns in room temperature toluene, hexane, and benzene solutions, 1.2 ns in solid C60 both at room and at low temperature, and dramatically drops to 0.9 ns in Ne matrices at 4 K for the reason that the ISC is affected by Ne. Note that the above
recorded by the multichannel scaler integrated in the TCSPC system.
III. RESULTS AND DISCUSSION The confocal SOPLIM scheme was first tested. But none of SOP signal was detected, even the pinhole PH1 had been fully opened. The reason should be that the SOP has been completely decayed by the inner optical elements of the scanner. Therefore, the nonconfocal SOPLIM scheme had to be used to avoid the loss of SOP in the scanner. In the nonconfocal SOPLIM scheme, the SOP images were acquired by detector D1, and the C60-PL images longer and shorter than the wavelength of 600 nm were aquired by detectors D3 and D2, respectively. The experimental results are shown in Figure 2. No C60-PL images shorter than the wavelength of 600 nm were detected. For convenience of comparison, the C60-PL images and the SOP images are put together. The pixel number of these images is 64 × 64 and corresponds to a 1150 μm × 1150 μm region of the sample. The resident time of laser beam on a pixel is Tpxl = 28.8000 ms, within which the laser-on time is Ton = 5.0625 ms and the laser-off time is Toff = 23.7375 ms, and 117.965 s is needed to complete a single-frame image. To improve the signal-to-noise ratio, the images in Figure 2 were all generated by the accumulation of five frames, which leads to the acquisition time for an image is 589.824 s or 9.8304 min. From Figure 2, one can see that SOP intensity image (Figure 2A) and SOP lifetime image (Figure 2C) have been successfully obtained by the nonconfocal SOPLIM. Compared with the corresponding C60-PL images (Figure 2B and 2D), the definition of SOP image is obviously lower than that of C60PL image, but the shape of C60 granules still can be recognized in the SOP images. It should be emphasized that the low definition of the SOP image is resulted from the nonconfocal imaging and should be able to be improved by using the following confocal imaging method. That is the laser beam is kept stationary and meanwhile a pinhole is added between lens 1 and detector D1 to block the SOP that does not come from the focus of the objective, and then confocal SOP imaging is performed by using stage-scanning mode instead of laserscanning mode. The essence of this method is that the definition of SOP image is compensated by sacrificing its imaging speed, for stage-scanning is much slower than laserscanning. The SOP intensity distribution in Figure 2A and the C60-PL intensity distribution in Figure 2B approximately complement each other, that is, the relatively bright regions in Figure 2A correspond to the relatively dark regions in Figure 2B and vice versa. This phenomenon can be interpreted by the conservation of energy. According to ref 17, after the singlet excited state 1 C60* is produced by the excitation of C60, the formed 1C60* will release its energy through the following two pathways. The first pathway is to emit fluorescence (C60-PL), and the second is that the 1C60* first forms the triplet excited state 3C60* through intersystem crossing (ISC) and then the 3C60* transfer its energy to O2 to produce O2(a1Δg). These two pathways compete between each other and lead to the complementary relationship between the intensity distributions of SOP and C60-PL. The relatively bright regions in Figure 2A (or the relatively dark region in Figure 2B) are just the regions that have high O2(a1Δg) concentrations. These regions may contain more O2 so that the energy of 3C60* is more conducive to accepted by O2, and meanwhile the fluorescence (C60-PL) emissions are more reduced. The phenomenon of the D
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
infrared-incompatible scanner that is suitable only for UV− visible light. Therefore, it can be reasonable to anticipate that if the optical elements in the scanner are all infrared-compatible, the SOP should be detectable at location 2 and a confocal SOPLIM should be able to be established. Given that the lenses, reflecting mirrors, and galvanometer mirror in the scanner are all more or less infrared-compatible, the key to an infrared-compatible scanner is likely the dichroic mirror DM2. The requirement for this mirror is that it can highly reflect the UV−visible excitation light and meanwhile highly transmit SOP, just like the dichroic mirror that is used by Ogilby et al. in their SOM.2 It needs to be pointed out that even if the confocal SOPLIM scheme can be realized, SOP lifetime imaging is still faced with a great challenge, which is imaging time. As mentioned in the Introduction, SOP is very weak, and ultralong acquisition time has to be consumed in order to get a clear SOP image. For example, the imaging time for the SOP images in Figure 2 is 9.8304 min; for hydrous samples such as cells or other organisms, it is easy to predict that the imaging time will be unendurably long and reach 9830 min based on the fact that the SOP lifetime in H2O is only 3.5 μs,24 over 1000 times shorter than the lifetime of 3.7 ms in C60 solid and resulting in an about 1000 times decrease in the quantum efficiency of SOP. Moreover, given that the concentration and sensitization efficiency of a hydrous-environment-used sensitizer are generally lower than those of C60 solid, the needed imaging time will become even longer. In addition to extraordinarily long imaging time, another disadvantage for SOP imaging is its relatively low image resolution that results from the relatively long wavelength of SOP (about 2 times longer than that of visible light). To avoid long SOP imaging time, a useful tactic is that fluorescence images are first acquired, and then SOP lifetime images are acquired for only a few interesting small domains selected from the fluorescence images, or SOP trace and lifetime are measured for only a few interesting points. For example, obvious difference of SOP lifetime may exist between different areas of a cell such as nucleus and cytoplasm, but this difference may be very slight at a roughly uniform area, so much so that measurements of SOP trace and lifetime for only a few points at different areas may be enough for a specific study. Figure 5 is just a case in point. Viewed from the C60-PL images, the small areas around the points a, b, c, and d look much
Figure 3. SOP spectrum from C60 powder.
lifetime values in ref 22 were all measured in oxygen-free environments, and in any case, it can be sure that our value of 347 ps is much smaller than these values. One explanation for such large lifetime difference, we think, is that the ISC is affected by O2 and thus shortens the fluorescence lifetime of 1 C60*. After all, according to ref 18, the concentration of O2 in C60 solid is as high as 0.233M; O2 should have the ability to affect the ISC just like Ne matrix can affect the ISC. In a word, by affecting the ISC, O2 is able to indirectly affect the fluorescence lifetime of 1C60*. Figure 4 is the C60-PL spectrum acquired by placing the IRCCD spectrograph at location 1. This C60-PL spectrum agrees
Figure 4. C60-PL spectrum from C60 powder.
well with the luminescence spectrum of the C60 powder at room temperature reported in ref 23, corresponding to the broad peak of the luminescence spectrum of C60 in the wavelength range of 750−820 nm (1.65−1.51 eV), with their maximums being both at the position of about 780 nm or 1.59 eV. In addition, ref 23 has shown that at short-wavelength direction the luminescence spectrum of C60 has decayed to zero at the position of 653 nm (1.9 eV), which also explains why no light signal was detected by the detector D2. As shown in Figure 1, when detecting SOP at location 2 in the confocal SOPLIM scheme, SOP needs to propagate in the scanner, and when at location 1 in the nonconfocal SOPLIM scheme, it does not. The fact that SOP can be detected at location 1 and can not be detected at location 2 shows that the scanner has greatly attenuated the SOP and is obviously an
Figure 5. SOP trace and decay lifetime for a single point. Traces a, b, c, and d correspond respectively to the points a, b, c, and d in Figure 2. E
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
■
different from each other but yet roughly uniform within every individual area, and their differences can indeed be seen from the corresponding SOP images; so the SOP dynamics in these areas can be represented by the SOP traces and lifetimes at points a, b, c, and d in Figure 2. These SOP traces, in fact, can also be acquired by using the single-point measurement function in a FLIM; this function allows one to park the laser beam at a specific point to acquire the fluorescence or phosphorescence trace for this point. This tactic in fact has been first used by Ogilby et al. when they measured the SOP trace and lifetime for small subcellular domains (nucleus and cytoplasm) selected from the fluorescence image of a cell.3,13,14 Figure 5 also clearly shows that with the excitation laser being turned on, O2(a1Δg) is accumulated during the time of Ton, which leads to the SOP intensity going up; then with the excitation laser being turned off, the SOP intensity begins to decay during the time of Toff. The lifetime image of SOP in Figure 2 is just built up by using the decay lifetime for every single point. Note that the ISC of 1C60* and the energy transfer from 3C60* to O2 (with the rate constant of 2 × 109 M−1 s−1)17are both fast processes, and O2(a1Δg) can be quickly produced upon the laser excitation of C60. During the laser-on period, O2(a1Δg) is produced at the arrival of every laser pulse and then decays with the lifetime of about 3.70 ms until the next arrival of laser pulse. The pulse interval is 12.5 ns, obviously far shorter than the decay lifetime time of 3.70 ms, so the concentration of O2(a1Δg) can be continually accumulated and the SOP intensity continually rises during the laser-on period.
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: +86-411-84379010; e-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by a grant from the National Natural Science Foundation of China (grant 21173217). REFERENCES
(1) Ogilby, P. R. Singlet oxygen: there is indeed something new under the sun. Chem. Soc. Rev. 2010, 39, 3181−3209. (2) Snyder, J. W.; Zebger, I.; Gao, Z.; Poulsen, L.; Frederiksen, P. K.; Skovsen, E.; Mcilroy, S. P.; Klinger, M.; Andersen, L. K.; Ogilby, P. R. Singlet oxygen microscope: from phase-separated polymers to single biological cells. Acc. Chem. Res. 2004, 37, 894−901. (3) Snyder, J. W.; Skovsen, E.; Lambert, J. D. C.; Poulsen, L.; Ogilby, P. R. Optical detection of singlet oxygen from single cells. Phys. Chem. Chem. Phys. 2006, 8, 4280−4293. (4) Breitenbach, T.; Kuimova, M. K.; Gbur, P.; Hatz, S.; Schack, N. B.; Pedersen, B. W.; Lambert, J. D. C.; Poulsen, L.; Ogilby, P. R. Photosensitized production of singlet oxygen: spatially-resolved optical studies in single cells. Photochem. Photobiol. Sci. 2009, 8, 442−452. (5) Andersen, L. K.; Ogilby, P. R. Time-resolved detection of singlet oxygen in a transmission microscope. Photochem. Photobiol. 2001, 73 (5), 489−492. (6) Poulsen, T. D.; Ogilby, P. R. Solvent effects on the O2(a1Δg)− O2(X3Σg−) radiative transition: Comments regarding charge-transfer interactions. J. Phys. Chem. A 1998, 102, 9829−9832. (7) Miller, H. C.; McCord, J. E.; Choy, J.; Hager, G. D. Measurement of the radiative lifetime of O2(a1Δg) using cavity ring down spectroscopy. J. Quant. Spectrosc. Radiat. Transfer 2001, 69, 305−325. (8) Andersen, L. K.; Gao, Z.; Ogilby, P. R.; Poulsen, L.; Zebger, I. A Singlet oxygen image with 2.5 μm resolution. J. Phys. Chem. A 2002, 106, 8488−8490. (9) Zebger, I.; Poulsen, L.; Gao, Z.; Andersen, L. K.; Ogilby, P. R. Singlet oxygen images of heterogeneous samples: Examining the effect of singlet oxygen diffusion across the interfacial boundary in phaseseparated liquids and polymers. Langmuir 2003, 19, 8927−8933. (10) Zebger, I.; Snyder, J. W.; Andersen, L. K.; Poulsen, L.; Gao, Z.; Lambert, J. D. C.; Kristiansen, U.; Ogilby, P. R. Direct optical detection of singlet oxygen from a single cell. Photochem. Photobiol. 2004, 79, 319−322. (11) Snyder, J. W.; Gao, Z.; Ogilby, P. R. Application of a dithered sampling technique to increase the spatial resolution of singlet oxygen images. Rev. Sci. Instrum. 2005, 76, 013701. (12) Skovsen, E.; Snyder, J. W.; Ogilby, P. R. Two-photon singlet oxygen microscopy: The challenges of working with single cells. Photochem. Photobiol. 2006, 82, 1187−1197. (13) Skovsen, E.; Snyder, J. W.; Lambert, J. D. C.; Ogilby, P. R. Lifetime and diffusion of singlet oxygen in a cell. J. Phys. Chem. B 2005, 109, 8570−8573. (14) Snyder, J. W.; Skovsen, E.; Lambert, J. D. C.; Ogilby, P. R. Subcellular, time-resolved studies of singlet oxygen in single cells. J. Am. Chem. Soc. 2005, 127, 14558−14559. (15) Hatz, S.; Lambertb, J. D. C.; Ogilby, P. R. Measuring the lifetime of singlet oxygen in a single cell: addressing the issue of cell viability. Photochem. Photobiol. Sci. 2007, 6, 1106−1116. (16) Kuimova, M. K.; Yahioglu, G.; Ogilby, P. R. Singlet oxygen in a cell: Spatially dependent lifetimes and quenching rate constants. J. Am. Chem. Soc. 2009, 131, 332−340. (17) Arbogast, J. W.; Darmanyan, A. P.; Foote, C. S.; Rubin, Y.; Diederich, F. N.; Alvarez, M. M.; Anz, S. J.; Whetten, R. L. Photophysical properties of C60. J. Phys. Chem. 1991, 95, 11−12. (18) Howells, S. C.; Black, G.; Schlie, L. A. O2(a1Δg) production and oxygen diffusion in C60 films. Synth. Met. 1994, 62, 1−7.
IV. CONCLUSION Nonconfocal SOP imaging, including intensity and lifetime imaging, is very easy to be realized on a general FLIM by using a sample of C60 powder. Analyses show that the obstacle to confocal SOP imaging is the great attenuation of SOP resulted from the scanner. Therefore, it is reasonable to anticipate that the two following solutions can be adopted to fulfill confocal SOP imaging. The first solution is that the general UV−visible scanner of a FLIM is replaced by an infrared-compatible scanner, and then the SOP detector is placed behind one of the pinholes of the infrared-compatible scanner. The SOPLIM in this solution in fact is no more than an infrared-compatible FLIM, in which both laser-scanning mode and stage-scanning mode can be used to perform confocal SOP imaging. The second solution is that as shown in Figure 1, the SOP returning from the objective is directly introduced into an optical path that is specifically designed for SOP detection by placing a dichroic mirror that reflects SOP and transmits visible light between the objective and the tube lens of a FLIM; meanwhile, a pinhole is added into the optical path and placed behind lens 1 to block the SOP emitted from off-focus positions. Finally, in the condition that the laser beam is kept stationary, confocal SOP imaging is performed by using stage-scanning mode. For the reason that the conjugate relationship between the two focal points of the objective and lens 1 needs to be maintained, only stage-scanning mode can be used for the SOPLIM in the second solution. For hydrous samples, another obstacle to confocal SOP imaging is its ultralong imaging time; and selecting only a few interesting points from fluorescence images to measure their SOP time-dependent traces may be a correct compromise. F
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A (19) Becker & Hickl GmbH. Microsecond decay FLIM: Combined fluorescence and phosphorescence lifetime imaging, Application Note; 2010. www.becker-hickl.com. (20) Nissen, M. K.; Wilson, S. M.; Thewalt, M. L. W. Highly structured singlet oxygen photoluminescence from crystalline C60. Phys. Rev. Lett. 1992, 69, 2423−2426. (21) Wang, J.; Leng, J.; Yang, H.; Sha, G.; Zhang, C. Long-lifetime and asymmetric singlet oxygen photoluminescence from aqueous fullerene suspensions. Langmuir 2013, 29, 9051−9056. (22) Sassara, A.; Zerza, G.; Chergui, M.; Ciulin, V.; Ganière, J. D.; Deveaud, B. Picosecond studies of the intramolecular relaxation processes in isolated C60 and C70 molecules. J. Chem. Phys. 1999, 111, 689−697. (23) Byrne, H. J.; Maser, W.; Rühle, W. W.; Mittelbach, A.; Hönle, W.; Schnering, H. G. V.; Movaghar, B.; Roth, S. Time-resolved photoluminescence of solid state fullerenes. Chem. Phys. Lett. 1993, 204, 461−466. (24) Egorov, S. Y.; Kamalov, V. F.; Koroteev, N. I.; Krasnovsky, A. A., Jr.; Toleutaev, B. N.; Zinukov, S. V. Rise and decay kinetics of photosensitized singlet oxygen luminescence in water. Measurements with nanosecond time-correlated single photon counting technique. Chem. Phys. Lett. 1989, 163, 421−424.
G
DOI: 10.1021/acs.jpca.5b01504 J. Phys. Chem. A XXXX, XXX, XXX−XXX
