Synthesis of Fluorescent, Dumbbell-Shaped Polyurethane Homo- and Heterodendrimers and Their Photophysical Properties

Fluorescent dendrimers have wide applications in biomedical and materials science. Here, we report the synthesis of fluorescent polyurethane homodendrimers and Janus dendrimers, which often pose challenges due to the inherent reactivity of isocyanates. Polyurethane dendrons (G1–G3) were synthesized via a convergent method using a one-pot multicomponent Curtius reaction as a crucial step to establish urethane linkages. The alkyne periphery of the G1–G3 dendrons was modified by a copper catalyzed azide–alkyne click reaction (CuAAC) to form fluorescent dendrons. In the reaction of the surfaces functionalized two different dendrons with a difunctional core, a mixture of three dendrimers consisting of two homodendrimers and a Janus dendrimer were obtained. The Janus dendrimer accounted for a higher proportion in the products’ distribution, being as high as 93% for G3. The photophysical properties of Janus dendrimers showed the fluorescence resonance energy transfer (FRET) from one to the other fluorophore of the dendrimer. The FRET observation accompanied by a large Stokes shift make these dendrimers potential candidates for the detection and tracking of interactions between the biomolecules, as well as potential candidates for fluorescence imaging.


Introduction
Dendrimers, as a pervasive class of macromolecules, have been extensively investigated for over four decades because of their unique monodispersity and multifunctionality. Depending on the types of functionalities present throughout the dendritic wedges, the dendrimers can be classified into two groups. The first group consists of identical functionalities throughout the dendrons, and thus in the dendrimer. These dendrimers are often called 'homodendrimers.' The second group consists of unidentical or dissimilar functions in the dendrons, and the dendrimers are classified as heterodendrimers. Such heterodendrimers composed of two different dendritic wedges of different hydrophobicity or hydrophilicity are called "Janus dendrimers (JDs)" [1][2][3][4]. The presence of dissimilar functions in the dendrons enable JDs to combine different properties associated with different wedges. Compared to amphiphilic copolymers [5,6], their dendritic analogs (JDs) offer more tunability at the molecular level, further affecting their self-assembling nature, where the tunning may be performed for peripheral groups, dendron generation, and the density of branching [3].
In a review, Caminade et al. described three distinct methods for the synthesis of JDs, as shown in Scheme 1 [1]. The first method is the simplest of all three and consists of reacting two dendrons with complementary functionalities. This method simply joins two different dendrons together at their focal points. Dendrons with semifluorinated benzyl ethers moiety connected together to form an amide linkage in the JD [7], and a Pd-catalyzed coupling of two focal points of dendrimeric wedges to create the JD are typical examples of this method [8]. The second method involves the convergent synthesis of dendrons followed by the controlled attachment of one dendron to core (with even number of functionalities, most often a difunctional) via a dendron's focal point. The remaining reactive site of core is then reacted with the other dendron to afford a JD. The very first example of all types of JDs synthesized by Hawker et al. in 1993 (poly(benzyl) ether) stands out as an excellent example of creating JDs using the second method [9]. The third method consists of the convergent synthesis of a dendron. The focal point of the dendron is then utilized to grow new branches using a divergent method [10,11]. Moreover, the syntheses of these dendritic structures via self-assembly caused by noncovalent interactions can also be found in the literature [3,12]. ethers moiety connected together to form an amide linkage in the JD [7], and a Pdcatalyzed coupling of two focal points of dendrimeric wedges to create the JD are typical examples of this method [8]. The second method involves the convergent synthesis of dendrons followed by the controlled attachment of one dendron to core (with even number of functionalities, most often a difunctional) via a dendron's focal point. The remaining reactive site of core is then reacted with the other dendron to afford a JD. The very first example of all types of JDs synthesized by Hawker et al. in 1993 (poly(benzyl) ether) stands out as an excellent example of creating JDs using the second method [9]. The third method consists of the convergent synthesis of a dendron. The focal point of the dendron is then utilized to grow new branches using a divergent method [10,11]. Moreover, the syntheses of these dendritic structures via self-assembly caused by noncovalent interactions can also be found in the literature [3,12]. Employing these methods, several JDs containing various types of linkages have been reported by many research groups since 1993. Percec et al. elegantly described the sequence-defined Janus glycodendrimers that can be designed to mimic the spatial properties of biological membranes, thereby providing a versatile tool in glycobiology [13,14]. In another study, Janus polyethylene glycol (PEG)-based dendrimers have been reported to perform controlled multi-drug loading and sequential release [15]. Buzzacchera et al. recently reported the synthesis of constitutional isomeric libraries of self-assembling dendrons and JDs employing natural and synthetic phenolic acids and PEGs [16]. These dendrimers can be utilized as important tools for nanomedicine and synthetic cell biology. The ongoing research on the potential applications of JDs in the Scheme 1. General methods of synthesis of Janus dendrimers.
Employing these methods, several JDs containing various types of linkages have been reported by many research groups since 1993. Percec et al. elegantly described the sequencedefined Janus glycodendrimers that can be designed to mimic the spatial properties of biological membranes, thereby providing a versatile tool in glycobiology [13,14]. In another study, Janus polyethylene glycol (PEG)-based dendrimers have been reported to perform controlled multi-drug loading and sequential release [15]. Buzzacchera et al. recently reported the synthesis of constitutional isomeric libraries of self-assembling dendrons and JDs employing natural and synthetic phenolic acids and PEGs [16]. These dendrimers can be utilized as important tools for nanomedicine and synthetic cell biology. The ongoing research on the potential applications of JDs in the biomedical field is further supported by the report on amphiphilic spermine-alkyl Janus dendrimers, which self-assemble to form highly ordered crystalline virus assemblies. Such findings can be used for applications in the study of complex biological systems [17]. Moreover, a decade ago, the polymers that are dendronized with self-assembling JDs, possessing one fluorinated and one hydrogenated dendron, were shown to act as reverse thermal actuators [18]. JDs not only have applications

Synthesis of G1 Dendron
The synthesis of generation-one dendron 3 has been previously reported by our group [25]. In brief, a protecting group-free approach using a one-pot multicomponent reaction was employed, where a mixture of 5-hydroxyisophthalic acid 1, diphenylphosphoryl azide (DPPA), and triethylamine was heated to generate an isocyanate in situ, which was then trapped by 4-pentyn-1-ol to form phenolic diurethane 2. The subsequent attachment of 11-bromoundecanol as a spacer group afforded the G1 dendron 3 in a two-step sequence (Schemes 2 and S1). 2. The subsequent attachment of 11-bromoundecanol as a spacer group afforded the G1 dendron 3 in a two-step sequence (Schemes 2 and S1).

Late-Stage Modification (LSM) of G1 Dendron 3
Late-stage modification is one of the most powerful synthetic approaches for the synthesis and functionalization of polymeric or dendritic macromolecules, especially when the functional groups in the molecule are sensitive to at the beginning of the synthetic route. To execute the LSM of dendrons, two different azido-compounds, 4 [34] (Scheme S2) and 5 [35,36], were synthesized using a previously reported procedure (Scheme S3). The LSM of 3 using non-fluorescent 4 under azide-alkyne click conditions yielded blue-fluorescent G1 dendron 7, as shown in Scheme 3. The formation of 1,4substituted-1,2,3-triazole after a click reaction led to a highly fluorescent PU dendron. In a similar approach, the same dendron was clicked with a different azide-clicking partner 5 ( Figure S1 and S2) to furnish a modified dendron 7 (Scheme S4). However, the expected fluorescence of 7 was not observed (under irradiation using 365 nm UV light) after azidealkyne click reaction as reported in the literature [37]. Both of these reactions preceded smoothly at room temperature for 2.5 h, providing excellent product yields. The pure products were obtained by flash chromatography, where the fluorescent dendrons could be easily visualized using a UV lamp during purification. Alternatively, the product can be precipitated in water without requiring chromatography. Scheme 2. Synthesis of first-generation dendron 3 via one-pot multicomponent Curtius reaction.

Late-Stage Modification (LSM) of G1 Dendron 3
Late-stage modification is one of the most powerful synthetic approaches for the synthesis and functionalization of polymeric or dendritic macromolecules, especially when the functional groups in the molecule are sensitive to at the beginning of the synthetic route. To execute the LSM of dendrons, two different azido-compounds, 4 [34] (Scheme S2) and 5 [35,36], were synthesized using a previously reported procedure (Scheme S3). The LSM of 3 using non-fluorescent 4 under azide-alkyne click conditions yielded blue-fluorescent G1 dendron 7, as shown in Scheme 3. The formation of 1,4-substituted-1,2,3-triazole after a click reaction led to a highly fluorescent PU dendron. In a similar approach, the same dendron was clicked with a different azide-clicking partner 5 ( Figure S1 and S2 ) to furnish a modified dendron 7 (Scheme S4). However, the expected fluorescence of 7 was not observed (under irradiation using 365 nm UV light) after azide-alkyne click reaction as reported in the literature [37]. Both of these reactions preceded smoothly at room temperature for 2.5 h, providing excellent product yields. The pure products were obtained by flash chromatography, where the fluorescent dendrons could be easily visualized using a UV lamp during purification. Alternatively, the product can be precipitated in water without requiring chromatography. 5 of 24 Scheme 3. Late-stage modification of G1 dendron 3 forming fluorescent dendrons 6 and 7.

Synthesis of G1 Homo-and Janus Dendrimers
With late-stage-modified dendrons 6 and 7 in hand, G1 homo-and Janus dendrimers were synthesized by allowing these two dendrons (1.2 eq each) to react with hexamethylene-1,6-diisocaynate core 8 at room temperature for 20 h (Schemes 4 and S5). The reaction was catalyzed by dibutyltin dilaurate (DBTDL), which is a workhorse catalyst in urethane chemistry used to convert an isocyanate into a urethane linkage [38]. The crude reaction mixture was purified by flash chromatography to afford a mixture of two homodendrimers, 9 and 11, and a Janus dendrimer 10 as yellow solids. Of these three dendrimers, 9 and 10 were found to exhibit a strong fluorescence; however, dendrimer 11 did not fluoresce when illuminated with a 365 nm UV lamp. Moreover, 77% of the product's total yield was constituted by the blue-fluorescent Janus dendrimer 10. Due to the presence of long non-polar alkyl chains, these solid dendrimers did not have sharp melting points.

Synthesis of G1 Homo-and Janus Dendrimers
With late-stage-modified dendrons 6 and 7 in hand, G1 homo-and Janus dendrimers were synthesized by allowing these two dendrons (1.2 eq each) to react with hexamethylene-1,6-diisocaynate core 8 at room temperature for 20 h (Schemes 4 and S5). The reaction was catalyzed by dibutyltin dilaurate (DBTDL), which is a workhorse catalyst in urethane chemistry used to convert an isocyanate into a urethane linkage [38]. The crude reaction mixture was purified by flash chromatography to afford a mixture of two homodendrimers, 9 and 11, and a Janus dendrimer 10 as yellow solids. Of these three dendrimers, 9 and 10 were found to exhibit a strong fluorescence; however, dendrimer 11 did not fluoresce when illuminated with a 365 nm UV lamp. Moreover, 77% of the product's total yield was constituted by the blue-fluorescent Janus dendrimer 10. Due to the presence of long non-polar alkyl chains, these solid dendrimers did not have sharp melting points.

Growth of Higher-Generation Dendrons
The synthesis of dendrons to the G3 dendrons is shown in Scheme 5 and Scheme S6. The linking group 12 required for the growth of the dendrons was synthesized and reported in previous studies [39]. To prepare G2 dendron 14, a mixture of G1 dendron 3 and linking group 12 was heated in anhydrous DMF under Curtius reaction conditions for three days to obtain a highly viscous dark red oil as a crude product, which yielded phenolic dendron 13 on chromatographic purification. This reaction was carried out in absence of any catalyst, where longer reaction times tend to provide better yields. The reaction set for about 36 h only had a 42% yield. Moreover, only 1.5 eq of 3 was found to afford the highest product yield. This could be due the fact that the phenolic OH of 12 can compete with the alcoholic OH of 3 during a nucleophilic attack on in situ generated isocyanate, leading to polymeric side products. This required us to utilize less than 2 eq of dendron for this specific reaction. The attachment of the undecanol group at the phenolic position in the next step afforded G2 dendron 14 with an excellent yield (90%). Scheme 4. Synthesis of first-generation homo-and Janus dendrimers 9-11.

Growth of Higher-Generation Dendrons
The synthesis of dendrons to the G3 dendrons is shown in Schemes 5 and S6. The linking group 12 required for the growth of the dendrons was synthesized and reported in previous studies [39]. To prepare G2 dendron 14, a mixture of G1 dendron 3 and linking group 12 was heated in anhydrous DMF under Curtius reaction conditions for three days to obtain a highly viscous dark red oil as a crude product, which yielded phenolic dendron 13 on chromatographic purification. This reaction was carried out in absence of any catalyst, where longer reaction times tend to provide better yields. The reaction set for about 36 h only had a 42% yield. Moreover, only 1.5 eq of 3 was found to afford the highest product yield. This could be due the fact that the phenolic OH of 12 can compete with the alcoholic OH of 3 during a nucleophilic attack on in situ generated isocyanate, leading to polymeric side products. This required us to utilize less than 2 eq of dendron for this specific reaction. The attachment of the undecanol group at the phenolic position in the next step afforded G2 dendron 14 with an excellent yield (90%). Next, G2 dendron 14 was heated with same linking group 12 following the same route of reaction to afford G3 phenolic dendron 15. With an increase in the size (branching) of the dendron, the steric effects also increase, thereby decreasing the yield of the reaction. Accordingly, the yield of 15 was slightly lower (62%) than that of G2 phenolic dendron, which in the next step underwent an S N 2 reaction with 11-bromoundecanol in presence of potassium carbonate as a base to afford G3 dendron 16 with an 80% yield. Both G2 and G3 dendrons can be easily purified by rapidly eluting them through a small silica column. Physically, these dendrons are highly viscous at room temperature, and this viscosity decreases when heat is gently applied, emonstrating a sol-gel behavior. This behavior can be attributed to the presence of numerous urethane (NHCOO) linkages, which could be associated with the formation of noncovalent interactions such as hydrogen bonding. Such gel-like compounds could have potential applications in material and biological sciences.

Late-Stage Modification of G2-G3 Dendrons
We utilized LSM approach to further functionalize the periphery of G2-G3 dendrons using CuAAC, as previously shown in G1 dendrons and dendrimers (Schemes 6 and S7). Firstly, the peripheral terminal alkyne groups were modified into 1,2,3-triazoles using azidocoumarin 4 as an azide-clicking partner. The reaction was fast and efficient and was performed at room temperature to produce a highly fluorescent (blue) dendron 18 with a 96% yield. Under similar conditions and using the same set of catalysts (copper sulfate and sodium ascorbate), a different G2 dendron 19 was produced with an 87% yield, employing azidonaphthalimide as the azide-clicking partner in this case. The expected fluorescence, however, was not observed (when illuminated with 365 nm UV lamp) in this molecule, which is contrary to the reported literature.

Synthesis of G2 Homo-and Janus Dendrimers
With surface-modified polyurethane dendrons 18 and 19 in hand, we moved forward to synthesize G2 dendrimers. The synthetic protocol is illustrated in Schemes 7 and S8. As previously mentioned for the synthesis of G1 dendrimers, a tin catalyst (DBTDL, 8.4 eq) was used to convert an isocyanate function to a urethane linkage. To accomplish this, Scheme 6. Late-stage modification of G2 and G3 dendrons 18-21.
After the successful surface functionalization of G2 dendrons using azide-alkyne click chemistry, we began modifying G3 dendron 16 using azides 4 and 5. For this, G3 dendron and azidocoumarin 4 were dissolved in THF, to which a brown solution of copper sulfate and sodium ascorbate in minimum amount of water was added at room temperature. The mixture was then vigorously stirred at rt for 2.5 h in the dark to obtain an intensely blue fluorescing solution. The solution instantly turns blue after stirring the reaction mixture, which is clearly observed with the help of a UV lamp. The crude mixture was then purified by flash chromatography to afford a yellow solid 20 in quantitative yield (97%). Azidealkyne click reaction G3 dendron and azide 5, on the other hand, it afforded a yellow solid as a surface-modified G3 dendron 21 with a 99% yield (Scheme 6).
The incomplete reaction or fragmentation of certain branches are very common drawbacks of a reaction involving a large number of reactions per molecule. This issue becomes serious when the size of the dendron continues to increase in convergent synthesis or during the outward growth of a dendrimer in a divergent method of synthesis. A small amount of heat during synthesis can cause the fragmentation of dendritic branches, leading to imperfections in the dendritic architecture. In our approach, these synthetic limitations were eliminated using click chemistry during the late-stage modification. This is because azide-alkyne click reaction efficiently competes at room temperature, leaving no unreacted sites in the molecule. Additionally, this reaction is compatible with any organic solvents as well as water.

Synthesis of G2 Homo-and Janus Dendrimers
With surface-modified polyurethane dendrons 18 and 19 in hand, we moved forward to synthesize G2 dendrimers. The synthetic protocol is illustrated in Schemes 7 and S8. As previously mentioned for the synthesis of G1 dendrimers, a tin catalyst (DBTDL, 8.4 eq) was used to convert an isocyanate function to a urethane linkage. To accomplish this, blue-fluorescent and non-fluorescent G2 dendrons 18 and 19 (1.2 eq each), respectively were allowed to react with hexamethylene diisocyanate core at room temperature for 20 h. Chromatographic purification of the crude mixture yielded a mixture of three dendrimers, 22-24, as G2 PUDs. Homodendrimer 22 and Janus dendrimer 23 showed an intense fluorescence, but the homodendrimer 24 did not. The distribution of these three products was almost identical to the distribution of G1 dendrimers, with the ratio being slightly different (22:23:24 = 1:4.3:1.3). This strategy of dendrimer synthesis could be promising from a synthetic point of view as it creates asymmetry in a dendritic molecule simply from a reaction of two different dendrons. Additionally, most of the product distribution is occupied by a heterodendrimer (Janus dendrimer). This method can be used to construct synthetically challenging JDs; otherwise, it is impossible to synthesize using the general methods shown in Scheme 1.
The syntheses of G3 homo-and Janus dendrimers, as well as G3 dendrimers, are shown in Scheme 8 and Scheme S9, respectively. Modified G3 dendrons 20 and 21 (1.2 eq each) were reacted with 1 eq of a hexamethylene diisocyanate core in the presence of DBTDL (16.4 eq) using anhydrous DMF as a solvent to afford a mixture of G3 dendrimers 25-27, of which 25 and 27 were homodendrimers and 26 was a Janus dendrimer. As in G2, homodendrimer 25 and Janus dendrimer 26 were blue and fluorescent, whereas the homodendrimer 27 was non-fluorescent. Nonetheless, the statistical distribution of these dendrimers was dramatically different from G1 and G2 dendrimers. The product ratio in this case was found to be 1: 31:1 for 25, 26, and 27, respectively. This showed that most of the product distribution was solely occupied the Janus dendrimer.
In these transformations, DBTDL acts as a catalyst. Sn coordinates with isocyanate thereby make the NCO group more susceptible to a nucleophilic attack via the focal point (OH) of dendron during dendrimer growth. However, a catalytic amount of DBTDL was not sufficient for successfully carrying out these transformations. Instead, a stoichiometric amount of this reagent was required for an excellent yield of the dendrimers because a greater portion of Sn is coordinated in a sphere created by 1,2,3-triazole. The separation of three dendrimers from the crude mixture was accomplished by gradient elution during flash chromatography, where blue-fluorescent G1-G3 homodendrimers were first eluted using acetone-DCM as a mobile phase. The polarity of the mobile phase was then gradually increased to 5% MeOH in DCM to elute blue-fluorescent G1-G3 Janus dendrimers followed by the elution of non-fluorescent homodendrimers. In these transformations, DBTDL acts as a catalyst. Sn coordinates with isocyanate thereby make the NCO group more susceptible to a nucleophilic attack via the focal point (OH) of dendron during dendrimer growth. However, a catalytic amount of DBTDL was not sufficient for successfully carrying out these transformations. Instead, a stoichiometric amount of this reagent was required for an excellent yield of the dendrimers because a greater portion of Sn is coordinated in a sphere created by 1,2,3-triazole. The separation of three dendrimers from the crude mixture was accomplished by gradient elution during flash chromatography, where blue-fluorescent G1-G3 homodendrimers were first eluted using acetone-DCM as a mobile phase. The polarity of the mobile phase was then gradually increased to 5% MeOH in DCM to elute blue-fluorescent G1-G3 Janus dendrimers followed by the elution of non-fluorescent homodendrimers.

Characterization of Dendrimers
Triazole moiety-containing G1-G3 PUDs were first characterized by homonuclear one-dimensional (1D) proton and carbon nuclear magnetic resonance ( 1 H NMR and 13 C) spectroscopy. As shown in the spectra in ESI, the peaks of similar G1-G3 homo-or Janus dendrimers ( Figures S5-S8, S25-S36) or G1-G3 dendrons ( Figures S3, S4 and S9-S24) exhibited similar peak patterns. However, the peak broadening was observed when increasing the generation of dendrimers. This can be associated with the fact that the protons become slightly inequivalent with the increasing size of dendrons or dendrimers. This could partly be caused by a 'tumbling effect.' The slow tumbling of macromolecules in solution leads to faster relaxation of transverse magnetization due to enhanced spinspin interactions leading to sharp peaks in the spectrum. On the other hand, a peak broadening is observed when the relaxation is faster due to an increased viscosity. The increase in the viscosity of these PUDs can be expected with increased generation due to the increased number of polar urethane linkages that can undergo intermolecular hydrogen bonding. For these reasons, peak splitting, and thus the resolution, gradually decreased in the higher-generation PU dendrons and dendrimers.
The representative 1 H NMR spectrum of G2 Janus dendrimer 23 and G3 dendron 20 is depicted in Figure 1. The presence of two different types of NH peaks, one that is more deshielded at 9.50 ppm and the other that is slightly shielded at 6.94 ppm, demonstrate

Characterization of Dendrimers
Triazole moiety-containing G1-G3 PUDs were first characterized by homonuclear one-dimensional (1D) proton and carbon nuclear magnetic resonance ( 1 H NMR and 13 C) spectroscopy. As shown in the spectra in ESI, the peaks of similar G1-G3 homo-or Janus dendrimers ( Figures S5-S8, S25-S36) or G1-G3 dendrons ( Figures S3, S4, S9-S24) exhibited similar peak patterns. However, the peak broadening was observed when increasing the generation of dendrimers. This can be associated with the fact that the protons become slightly inequivalent with the increasing size of dendrons or dendrimers. This could partly be caused by a 'tumbling effect.' The slow tumbling of macromolecules in solution leads to faster relaxation of transverse magnetization due to enhanced spin-spin interactions leading to sharp peaks in the spectrum. On the other hand, a peak broadening is observed when the relaxation is faster due to an increased viscosity. The increase in the viscosity of these PUDs can be expected with increased generation due to the increased number of polar urethane linkages that can undergo intermolecular hydrogen bonding. For these reasons, peak splitting, and thus the resolution, gradually decreased in the higher-generation PU dendrons and dendrimers.
The representative 1 H NMR spectrum of G2 Janus dendrimer 23 and G3 dendron 20 is depicted in Figure 1. The presence of two different types of NH peaks, one that is more deshielded at 9.50 ppm and the other that is slightly shielded at 6.94 ppm, demonstrate that the attachment of the focal point of the dendron (OH) to the hexamethylene diisocyanate core was successful. The more downfield peak was assigned to the NH peak in the vicinity of branching point aromatic rings, whereas the upfield peaks were assigned to the NH in the aliphatic region of the dendrimer. The most deshielded aromatic proton peak at 8.53 ppm was caused by the triazole proton, thereby providing evidence of the successful completion of an azide-alkyne click reaction. Two protons of branching aromatic ring were spotted at 7.55 ppm and 7.20 ppm and labeled as 'a' and 'b', respectively, as shown in Figure 1. G1-G3 dendrons and dendrimers were then investigated by high-resolution electrospray ionization-mass spectrometry (HR ESI-MS) and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). The MALDI-TOF-MS spectra of some of the dendrons and dendrimers are shown in Figure 1. diisocyanate core was successful. The more downfield peak was assigned to the NH peak in the vicinity of branching point aromatic rings, whereas the upfield peaks were assigned to the NH in the aliphatic region of the dendrimer. The most deshielded aromatic proton peak at 8.53 ppm was caused by the triazole proton, thereby providing evidence of the successful completion of an azide-alkyne click reaction. Two protons of branching aromatic ring were spotted at ~7.55 ppm and 7.20 ppm and labeled as 'a' and 'b', respectively, as shown in Figure 1.

Photophysical Properties-Absorption and Emission
Photophysical properties of the G1-G3 dendrimers were investigated using UV-vis and fluorescence spectroscopy ( Figure 2 and Table 1). As shown in Figure 2A, G1 dendrimers (9-11) absorbed at 343 nm and 420 nm. These absorption maxima were assigned to the dendrimers containing naphthalimide and blue fluorophore, respectively. Similarly, G2 dendrimers (22)(23)(24) were found to absorb at 345 nm and 417 nm due to the presence of naphthalimide and coumarin fluorophores, respectively ( Figure 2C). G3 dendrimers (25)(26)(27) exhibited an almost similar pattern of UV absorption showing the absorption maxima at 344 nm and 418 nm ( Figure 2E). In all cases, the Janus dendrimers presented two absorption maxima due to the presence of two fluorophores.   G1-G3 dendrimers did not fluoresce in solid state; however, they showed good fluorescence in a solution. A 1:1 solution of methylene chloride and methanol was chosen as the solvent to study the fluorescence of these dendrimers. The fluorescence emission spectra of the G1 dendrimers are shown in Figure 2B. Homodendrimers 9 and 11 were emitted at a maximum wavelength of 478 nm and 410, respectively, and as expected, Janus dendrimer 10 had two emission peaks at 484 nm and 414 nm. Similar observations were observed for G2 and G3 dendrimers ( Figure 2B,F, respectively), where homodendrimers had one absorption peak, and the Janus dendrimers had two peaks. Moreover, the longer wavelength of each Janus dendrimer had a slight increase in its wavelength compared to the corresponding blue-fluorescent homodendrimer. For example, blue-fluorescent G3 Janus dendrimer 26 was slightly absorbed at a longer wavelength (495 nm) compared to its blue-fluorescent homodendrimer analog 25 (489 nm).

FRET
The bathochromic shift of the blue-fluorescent Janus dendrimer can be explained by a phenomenon called FRET, an acronym for Förster resonance energy transfer or fluorescence resonance energy transfer. For FRET to occur, the fluorescence emission of one fluorophore must overlap with the absorption of the other fluorophore. In our study, the emission caused by naphthalimide fluorophore (~350-450 nm) was found to precisely coincide the absorption of coumarin fluorophore (~350-475 nm). For this reason, the fluorescence energy from the naphthalimide fluorophore (donor) was transferred to the coumarin fluorophore (acceptor), thereby shifting the emission of Janus dendrimers toward the longer wavelength ( Figure 3D). This is evident from the red-shifted emission of Janus dendrimers (Table 1). This shift of emissions toward longer wavelengths was accompanied by an increase in the intensity of emitted light. For instance, G1 Janus dendrimer 10 (shows FRET) exhibited intense fluorescence compared with homodendrimer 9 (no FRET). This red-shifting was found to be somewhat constant (6-7 nm) with the increased generation of dendrimers. This is potentially due to the fact that fluorescence resonance energy transfer is primarily affected by the distance between the donor and acceptor fluorophores, which can become closer together in a solution due to bond rotation. The FRET phenomenon was not observed when G1-G3 Janus dendrimers were irradiated, with the wavelength corresponding to the acceptor (blue) fluorophore simply because there was no possibility of energy transfer. This type of FRET measurement is useful for the detection and tracking the interactions between proteins [40], observing membrane fluidity [41], and sensing [42] applications.

Stokes shift
Unlike homodendrimers, Janus dendrimers (G1-G3) exhibited two Stokes shifts when the compound was irradiated with absorption maxima corresponding to the nathphthalimide fluorophore (343-345 nm). While the first Stokes shift was quite similar to that of homodendrimers, the second was dramatically red-shifted. Janus dendrimers of  Stokes shift is defined as the electronic transition between the excited and ground state of a chemical entity, which is measured as the difference between absorption and emission maxima. It is the outcome of two phenomena-vibrational relaxation and solvent reorganization. When G1-G3 dendrimers are dissolved in polar solvents such as methanol, solvent molecules surround the fluorophore. They can quickly reorientate their dipoles to stabilize the excited state of fluorophore more than the ground state. The difference in energy following solvent reorganization results in Stokes shift [43]. The G1-G3 dendrimers displayed very similar absorption (343-345 nm for naphthalimide fluorophores and 417-420 nm for blue fluorophores) and emission (413-415 nm for naphthalimide fluorophores and 478-489 nm for blue fluorophores) maxima wavelengths. This resulted in the red-shifting (∆λ) of G1-G3 homodendrimers by 68-71 nm, which is a moderately large Stokes shift. Figure 3A,C shows the Stokes shift of G2 homodendrimers.

Stokes shift
Unlike homodendrimers, Janus dendrimers (G1-G3) exhibited two Stokes shifts when the compound was irradiated with absorption maxima corresponding to the nathphthalimide fluorophore (343-345 nm). While the first Stokes shift was quite similar to that of homodendrimers, the second was dramatically red-shifted. Janus dendrimers of all generations exhibited large Stokes shifts (∆λ ca. 141-151 nm). As shown in Figure 3B, the G2 Janus dendrimer revealed both moderately large (4714 cm −1 ) and large (8742 cm −1 ) Stokes shifts (calculated as (1/λ max.abs − 1/λ max.em ) × 10 7 )). A large Stokes shift is an indicative of fast relaxation from the initial state to the final emission state, which is caused by intramolecular energy transfer from one part of the molecule to another part of the same molecule. In our study, naphthalimide fluorophore was found to transfer energy to the blue fluorophores. Compared with other strategies, such as excited state atom transfer [44], and by introducing alternating vibronic structures [45], this approach ensures that large Stokes shifts occur simply by installing two different fluorophores in a dendrimer in one synthetic operation. The large Stokes shift (≥100 nm) of these dendrimers is advantageous for their application in fluorescence imaging, as the wide gap between the absorption and emission causes a decrease in self-absorption.
HRMS spectra of small molecules including dendrons were obtained from ESI-LTQ-Orbitrap. MALDI of larger molecules were recorded with a Bruker Autoflex 3 instrument using ∝ −cyano-4-hydroxycinnamic acid (CCA) as a matrix in positive ion mode.
Purification of compounds was carried out using flash chromatography with irregular silica of 40-60 µm, 60 Å. Small-scale purification was achieved using auto-column flash cartridges packed with 12 g or 40 g silica of 40-75 µm, 60 Å (obtained from Sorbtech and Supelco Technologies). Flow rate was 10-30 mL/min. Mobile phase used in these separations was ethyl acetate, hexane, DCM, or a mixture of these solvents.
3.2. Synthesis of Azide-Alkyne Clicked G1 Dendron, 7 4-Bromo-N-ethyl-1,8-naphthalimide, 5b (Scheme S3). Compound 5b was synthesized by a previously reported procedure [46]. In brief, in a clean RB flask, a mixture of 4-bromo-1,8-naphthalic anhydride 5a (2.5 g, 9.0 mmol, 1.0 eq), 70% ethylamine (3.36 mL, 36.0 mmol, 4.0 eq), and dioxane (90 mL) was refluxed for 7 h taking 2.02 mL of ethylamine as the first aliquot. After 7 h, the second aliquot of triethylamine (1.34 mL) was added and further refluxed for 14 h, cooled to room temperature, and then poured into cold water to collect a yellow solid as product 5b (2.712 g, 99%) on filtration. This product was taken to the next step without further purification. Spectra were similar to the previously reported literature. 1 [36]. Briefly, in a clean RB flask, NaN 3 (2.32 g, 35.03 mmol, 3.6 eq) in NMP (35 mL) was stirred at 50 • C for 24 hr. Then, 5b (2.96 g, 9.73 mmol, 1.0 eq) was added at room temperature and further stirred at room temperature for 24 h. The solution was then diluted with water (100 mL) and extracted with EtOAc, washed with brine, dried over anhydrous MgSO 4 , filtered, and evaporated under reduced pressure. The crude was purified by flash chromatography using 10%-25% EtOAc in hexane to obtain a yellow solid as product 5 (2.51 g, 97%). Spectra were similar with the previously reported literature. 1  Azide-alkyne clicked G1 dendron, 7 (Scheme S4). Dendron 3 (103.0 mg, 0.20 mmol, 1 eq) was dissolved in THF (2.0 mL) in an RB flask to which azidonaphthalimide 5 (160.0 mg, 0.60 mmol, 1.5 eq/triple bond) was added. After adding an aqueous solution of CuSO 4 ·5H 2 O (0.15 eq/N 3 ) and sodium ascorbate (0.30 eq/N 3 ) in a minimum amount of water to the flask, the reaction mixture was stirred vigorously at room temperature under dark conditions. When the alkyne was completely consumed, the reaction mixture was diluted and extracted with DCM, combined organic layers were dried over anhydrous MgSO 4 , and the solvent was evaporated under reduced pressure. The crude was purified by flash chromatography using 24 % acetone in DCM as mobile phase to obtain a yellow solid as product 7 (202.  13  Experimental procedure for the synthesis of G1 dendron 6 (Scheme S1 has been reported previously by our group [25].

Conclusions
In summary, polyurethane dendrons to G3 were synthesized using a convergent method, where the azide-alkyne click reaction is highly effective strategy for preparing dendrons with high yields and easy isolation. When two different dendrons reacted with a difunctional core, Janus dendrimers are favorable. This is demonstrated by the yield of the G3 Janus dendrimer, which accounted for more than 90% of the total product yield, showing the promising synthetic applicability of this approach to introduce unsymmetrical branches in the dendritic system with a high efficiency. Photophysical study of these dendrimers showed the FRET phenomenon from naphthalimide (donor) to blue fluorophore (acceptor) resulting in shift of emission toward the longer wavelengths accompanied by an increased intensity. The FRET measurements in the solution showed that red-shifting is not affected by the generation of dendrimers. Our method can prepare dendrimers with fluorophores to achieve large Stokes shifts. This research establishes the parameters for energy transfer within dendritic molecules and should inform the further development in the preparation of molecular antennas.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author. The data are not publicly available because of the lack of a dedicated server.