Design and synthesis. Due to the large π−conjugated frameworks, BODIPY dyes tend to aggregate in a face-to-face stacking mode (H-aggregation), and J-aggregation is usually not favored. As a result, most BODIPY dyes usually display ACQ in the aggregation state. To dismiss the strongly π−π interactions between the indacene plane and subsequently achieve the goal of J-aggregation, we introduced a PCP group with a three-dimension (3D) structure, to the meso position of BODIPY core. Moreover, we conjugated the strong electron-donating group, N,N-dimethylaminostyrl, to the 3,5- position of the PCP-BODIPY structure to realize the NIR-I emission (Fig. 1). Take advantage of red-shifted emission, J-aggregates of PCP-BOIDPY dye are expected to be fluorescent in the NIR-II region. The synthesis of PCP-BDP1 and PCP-BDP2 is outlined in Scheme S1. Additionally, two BODIPY dyes, Ph-BDP151 and Ph-BDP2,54 which have the phenyl group on the meso-position of BODIPY core, were synthesized for comparison.
X-ray single crystal structure analysis. To confirm the molecular design concept, we first investigated the molecular packing mode of PCP-BDP1 and PCP-BDP2 via single crystal structure analysis (Table S1). In the single crystal structure, the indacene planes of PCP-BDP1 and PCP-BDP2 are slightly bent due to the weak intermolecular interactions (Fig. 2 and S1). The PCP group is highly twisted to the BODIPY core. The dihedral angles between the phenyl ring that was attached to the BODIPY core and the indacene plane are ranging from 55.5o to 58.5o. In the molecular packing structure of PCP-BDP1, PCP groups are connected to the indacene plane through C-H⋅⋅⋅π (~ 3.8 Å) interactions. Due to the steric effect of the PCP group, the face-to-face molecular packing mode between the indacene plane is disfavored, as a result, PCP-BDP1 molecules are J-aggregated with a slipping angle of 38o.27 The distances between the borondipyrrole plane are ~ 3.7 Å. Ph-BDP1 show different molecular packing mode comparing to PCP-BDP1.51 As shown in Figure S2, Ph-BDP1 molecules are arranged in a zigzag pattern through weak C-H⋅⋅⋅π interactions and no J-dimers were observed in the molecular packing structure. This difference demonstrates that the PCP group plays a key role in tuning the J-aggregation of BODIPY core. In PCP-BDP2, π⋅⋅⋅π interactions between the N,N-dimethylaminphenyl group and the indacene plane, the C-H⋅⋅⋅F hydrogen bond (~ 3.4 Å) between the PCP group and the borondipyrrole plane, and C-H⋅⋅⋅π (~ 3.8 Å) interactions between the PCP group and the N,N-dimethylaminphenyl group dominate the molecular packing structure of PCP-BDP2, which facilitate the J-aggregation packing mode. The slipping angle and the distance between each molecule are determined to be 36o and ~ 3.3 Å, respectively (Fig. 2c).
Photophysical properties. Before investigating the photophysical properties of J-aggregates, firstly, we measured the absorption and emission spectra of PCP-BDP1 and PCP-BDP2 in diluted solutions. In dichloromethane (DCM), PCP-BDP1 displayed the main absorption band (λabs) centered at 523 nm (ε = 41900 M− 1cm− 1), and two emission bands centered at 524 nm and 554 nm, respectively (Figure S3). The sharp and narrow emission band at 524 nm can be assigned as the typical local excite (LE) emission of the BODIPY core, while the broad and weak emission band at 554 nm should be ascribed to the charge transfer (CT) emission band. In most cases, the phenyl ring at the meso-position has a very weak conjugation effect on the BODIPY core because of its free rotation.55 For example, Ph-BDP1 only shows the LE emission at 515 nm in DCM, and the CT band around 550 nm is hardly to be distinguished from the LE emission. However, the free rotation of the PCP group in PCP-BDP1 is inhibited due to the steric effect of the PCP group, leading to the CT from the PCP group to the BODIPY core. This speculation was preliminary proved by measuring the viscosity-dependent emission spectra. As shown in Figure S4, the emission intensity at 554 nm is gradually increased with the viscosity-increasing from 0.6 to 630 cp, while the intensity at 524 nm is almost unchanged. This result suggests that the CT process was greatly enhanced due to the inhibited rotation of the PCP group in the high viscosity media. Moreover, PCP-BDP2 showed the absorption and emission bands centered at 722 nm and 795 nm, respectively, which is red-shifted in comparison with that of Ph-BDP2 (λabs = 700 nm and λem = 750 nm in DCM, Fig. 2d). This result suggests that the CT process is more strengthened in PCP-BDP2 than in Ph-BDP2, which also confirms the electron-conjugating effect of the PCP group on the BODIPY core.
We further carried out the theoretical calculation to understand the different absorption and emission properties between these PCP and phenyl groups substituted BODIPY dyes (Table S2). Time-dependent density functional theory (TDDFT) results show that the main absorption and emission bands of Ph-BDP1 are contributed mainly by the transition from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) (Fig. 2e and S5a). Moreover, the HOMO and LUMO are delocalized on the indacene plane, while the phenyl ring at the meso-position has negligible contribution to the electronic structure of Ph-BDP1 in both ground and excited states. Different from that, the PCP group in PCP-BDP1 is conducive to the electron delocalization in both HOMO and LUMO (Fig. 2e and S5b). This difference makes the main absorption band of PCP-BDP1 is composed of both the first (S1) and second (S2) excited states, which mainly originated from the HOMO→LUMO, HOMO-1→LUMO and HOMO-3→LUMO transitions. The electronic transition of S1→S0 (ground state, f = 0.0900) can be ascribed to the CT emission, while the S2→S0 transition with relative large oscillator strength (f = 0.4344) is a fluorescent LE state. This result is consistent with the fluorescent spectrum of PCP-BDP1 obtained in DCM. The CT process from the PCP group to the indacene plane should be responsible for the different emission behavior of Ph-BDP1 and PCP-BDP1. Interestingly, the main absorption and emission bands of both Ph-BDP2 and PCP-BDP2 are contributed by the HOMO→LUMO transitions (Fig. 2e and S6). The HOMO and LUMO of Ph-BDP1 and the HOMO of PCP-BDP2 are delocalized on the whole molecular skeleton except for the phenyl and PCP groups. However, the LUMO of PCP-BDP2 is localized on the BODIPY core as well as the meso-phenyl ring of the PCP group. This feature favors the strengthened CT process, and therefore, the red-shifted absorption and emission bands of PCP-BDP2 in comparison with that of PCP-BDP1 can be rationalized.
Both the absorption and emission wavelength of PCP-BDP1 remain almost unchanged in different solvents, and only a slight increase of the absorbance and the emission intensity is observed in a polar solvent such as methanol and dimethylsulfoxide (DMSO) (Figure S7). All of these results demonstrate the weak CT of PCP-BDP1. Due to the donor (D)−π−acceptor (A) structure, PCP-BDP2 shows polarity dependent absorption and emission (Figure S8). The absorption and emission bands are red-shifted with the increased solvent polarity. For example, PCP-BDP2 exhibited λem in DMSO at 824 nm, which is 73 nm red-shifted relative to the λem of 751 nm in hexane. Interestingly, the emission intensity increment is also observed in solvents with relatively large viscosity, such as o-dichlorobenzene (1.32 cP at 25°C), 1,4-dioxane (1.20 cP at 25°C) and DMSO (1.98 cP at 25°C). This result suggests that the non-radiative decay process is inhibited due to the restricted intermolecular motion in these solvents with higher viscosity.
J-aggregation behavior in the aggregated state. To further explore the J-aggregation behavior of PCP-BDP1 and PCP-BDP2, we investigated their emission variation in the tetrahydrofuran (THF)-water binary solvents. In THF, PCP-BDP1 and PCP-BDP2 show one sharp absorption band at 523 and 718 nm, respectively. After the addition of water to the solution, the absorbance decreases and red-shifted to 532 and 748 nm when the water volumetric factions (fw) increased to 99%, which indicates the formation of J-aggregates (Figure S9). In the emission spectra, the emission intensity at 520 nm and 560 nm of PCP-BDP1 remain unchanged with fw increasing from 0 to 70%. With fw of 80%, the emission intensity at 520 nm was distinctly increased, accompanied by the slightly red-shifted CT band. This result suggests that PCP-BDP1 begins to aggregate, the restricted molecular motion and enhanced CT process should be responsible for the enhanced LE emission and the red-shifted CT emission. Notably, when fw reaches 90%, the LE emission is distinctly decreased, and the CT emission is almost disappeared. A new broad emission band ranging from 560 nm to 850 nm, which should be assigned as the emission of J-aggregates, is appeared simultaneously (Fig. 3a). Since the emission wavelength of J-aggregates is highly dependent on the size of the aggregates, the coexistence of the J-aggregates with different sizes is the reason for the formation of the broad emission band.56
PCP-BDP2 showed the same aggregation trend as PCP-BDP1. As shown in Fig. 3b, the emission intensity at 790 nm was enhanced and red-shifted to 820 nm with fw increasing from 10–60%. The intramolecular motion restriction and the CT process enhancement induced by the viscosity and polarity increment should be responsible for the enhanced and red-shifted emission, respectively. With fw of 70%, the emission at 820 nm was distinctly quenched, accompanied by the appearance of a new broad peak around 930 nm, suggesting the formation of J-aggregates. When fw reaches 80% and 90%, most of the PCP-BDP2 molecules may be aggregated to J-aggregate, as a result, the emission around 800 nm is almost disappeared, while the emission band of J-aggregates at approximately 900 nm (J1-band) is enhanced. Notably, a weak NIR-II emission band around 1000 nm (J2-band) was also observed as the side peak of the main emission band, which suggests that the emission of PCP-BDP2 J-aggregates is capable of red-shifted to the NIR-II region when the more condensed J-aggregates are formed.
To verify this speculation, we further measured the emission spectra of PCP-BDP1 and PCP-BDP2 in the solid state. The crystalline powder of PCP-BDP1 showed multiple J-aggregates emission bands at 570 nm, 640 nm, and 685 nm, respectively, which is contributed by J-aggregates with different energies (Fig. 3c). Notably, PCP-BDP2 displayed one broad J-aggregates emission centered at 1010 nm, which is consistent with the J2-band observed in THF-water binary solvent with fw of 99% (Fig. 3d). This result confirms that the condensed molecular packing mode should be responsible for the generation of the J2-band of PCP-BDP2. Additionally, we also measured the emission behavior of Ph-BDP1 and Ph-BDP2 in THF-water. With the fw increasing from 0 to 90%, the emission of both two compounds was gradually decreased, indicating the typical ACQ behavior (Figure S10). These results further demonstrated that the PCP group plays a key role in the J-aggregation behavior of PCP-BDP1 and PCP-BDP2.
Encouraged by the NIR-II emission capability of PCP-BDP2 J-aggregates observed in both THF-water and solid-state, we investigated that whether the NIR-II emissive J-aggregates could be stabilized in nanoparticles (NPs) and be employed for the NIR-II imaging. we prepared PCP-BDP2 NPs by encapsulating the PCP-BDP2 aggregates into a Pluronic F-127 matrix. The PCP-BDP2 NPs showed a spherical morphology with a diameter of ~ 70 nm, which was characterized using transmission electron microscopy (TEM) (Fig. 3e). The average diameter of PCP-BDP2 NPs was measured to be ~ 75 nm by a dynamic light scattering (DLS) experiment with a low polydispersity index (PDI) of 0.081 (Fig. 3f). In PBS buffer, the PCP-BDP2 NPs showed λabs and λem around 750 and 1010 nm, respectively, which is comparable with those observed in the aggregation state. This result suggests that the J-aggregates can be efficiently stabilized in the Pluronic F-127 matrix.
The PCP-BDP2 NPs showed relatively high Φf in PBS buffer. The calculated Φf value for PCP-BDP2 NPs was 6.4% (reference dye IR 26, Φf = 0.1%), which is higher than most reported NIR-II emissive BODIPY dyes and J-aggregates such as NJ1060 (1%),21 and NIR-II-WAZABY-01 (0.8%).20 Furthermore, after continuous laser irradiation (808 nm, 100 mW/cm2) for 60 min, the NIR-II emission intensities of PCP-BDP2 NPs remained almost unchanged, however, another famous NIR-II fluorophore, IR1061, showed rapidly decayed emission from 100–75% under the same conditions (Fig. 3h). Moreover, PCP-BDP2 NPs showed no apparent absorption and emission spectral change in PBS in the presence of glutathione, cysteine, and hydrogen peroxide (Figure S11). These results demonstrate the good photo- and chemical-stability of PCP-BDP2 NPs.
In vitro imaging. To verify the biological imaging capability of PCP-BDP2 NPs, we carried out both the in vitro and in vivo NIR-II fluorescence imaging experiments. An 808 laser was used as the excitation resource because of its general availability and reduced biological absorption. A clinically approved NIR-I dye, ICG, was employed as a control. The PL intensity of PCP-BDP2 NPs with different concentrations was detected, and the NIR-II emission signals of PCP-BDP2 NPs showed a linear increase with the concentration, which is consistent with the above results that NIR-II emissive J-aggregates is highly related to the degree of aggregation (Fig. 4a and 4b). Besides, the NIR-II fluorescence signals of PCP-BDP2 NPs and ICG under the different thickness of chicken tissues were collected, the NPs showed good NIR-II imaging penetration depth up to 8 mm, which is higher than the 6 mm of ICG (Fig. 4c). Altogether, these results indicated that PCP-BDP2 NPs hold the promise to be a superb NIR-II fluorescent dye than ICG.
In vivo imaging. We further carried out in vivo NIR-II imaging of cerebral vasculature and hindlimb with the PCP-BDP2 NPs. The NPs in PBS were injected into mice via the tail vein. After the injection for 5 min, the blood vessels in both the brain and hindlimb can be visualized from the surrounding background tissue with high resolution. The brightness and clarity of PCP-BDP2 NPs were higher than that of ICG with the same imaging conditions, implying that the PCP-BDP2 NPs can offer better imaging quality than ICG (Fig. 4d and 4f), which was quantitatively analyzed and compared (Fig. 4e and 4 g). Moreover, the fluorescence signal of PCP-BDP2 NPs decreased gradually with the time increased from 5 min to 24 h, while the fluorescence of ICG is almost undetectable when the time increased to 8 h (Figure S12 and S13). Besides, the PCP-BDP2 NPs were found to accumulate in the liver and spleen, properly due to the uptake by mononuclear phagocytic system-related organs (Figure S14). All the above imaging results demonstrate the promising high-resolution and long-term NIR-II imaging ability of the PCP-BDP2 NPs.
Lymph node imaging. The lymphatic system is important for maintaining fluid homeostasis and immunity, which is increasingly considered as a conduit for the metastasis of a variety of cancers such as breast, melanoma and so on.57 Optical imaging of the lymphatic system can map lymphatic drainage, locate sentinel lymph node (SLN), and visualize multiple lymph nodes.58, 59 In this experiment, PCP-BDP2 NPs were used for in vivo imaging to assess its ability for mapping lymph nodes. The PCP-BDP2 NPs (50 µL,1 mg/mL) was subcutaneously injected into the footpads of nude mice, the lymphatic vasculature and the sentinel lymph node (SLN) was observed immediately (Fig. 5a), and it was still clear after 5 h injection (Fig. 5b). And then the SLN with a diameter of less than 1 mm was removed precisely with the guidance of the fluorescence signal of PCP-BDP2 NPs (Fig. 5c), and the (H&E) histological staining could prove again (Fig. 5d).
Image-guided surgery. Recently, NIR-II fluorescence-guided cancer surgery has been proven feasible clinically, which reduce cancer recurrence and promote the outcomes of cancer surgery.4, 60 To demonstrate that the strong NIR-II fluorescence signal of PCP-BDP2 endow its ability for image-guided cancer surgery, the peritoneal carcinomatosis-bearing mouse model was established, which scatter numerous tumor nodules of various sizes in the peritoneal cavity, especially those with diameters < 1 mm. After the PCP-BDP2 NPs were intravenously injected into the mice for 24 h, the surgery was first performed by a surgeon by opening the mouse abdomen. As we selected luciferase-expressed 4T1 tumors that exhibited bioluminescence after injection with D-luciferin, so the signals of fluorescence of PCP-BDP2 NPs and the bioluminescence of luciferase were well colocalized (Fig. 6a). And then lots of large tumor nodules with diameters > 1 mm were removed by the surgeon’s naked eyes. Then with the guidance of high brightness of PCP-BDP2 NPs, smaller tumor nodules were resected (Fig. 6c). The well-overlapped bioluminescence and fluorescence signals of the removed nodules indicate those were indeed tumors (Fig. 6b), which was also proved by the hematoxylin and eosin (H&E) staining (Fig. 6d), and these results together demonstrate the accuracy of operation.