Conformation of Tunable Nanocylinders: Up to Sixth-Generation Dendronized Polymers via Graft-Through Approach by ROMP.

Well-defined dendronized polymers (denpols) bearing high-generation dendron are attractive nano-objects as high persistency provides distinct properties, contrast to the random coiled linear polymers However, their syntheses via graft-through approach have been very challenging due to their structural complexity and steric hindrance retarding polymerization. Here, we report the first example of the synthesis of poly(norbornene) (PNB) containing ester dendrons up to the sixth generation (G6) by ring-opening metathesis polymerization. This is the highest generation ever polymerized among dendronized polymers prepared by graft-through approach, producing denpols with molecular weight up to 1960 kg/mol. Combination of size-exclusion chromatography, light scattering, and neutron scattering allowed a thorough structural study of these large denpols in dilute solution. A semiflexible cylinder model was successfully applied to represent both the static and dynamic experimental quantities yielding persistent length (l p), cross-sectional radius (R cs), and contour length (L). The denpol persistency seemed to increase with generation, with l p reaching 27 nm (Kuhn length 54 nm) for PNB-G6, demonstrating a rod-like conformation. Poly(endo-tricycle[4.2.2.0]deca-3,9-diene) (PTD) denpols exhibited larger persistency than the PNB analogues of the same generation presumably due to the higher grafting density of the PTD denpols. As the dendritic side chains introduce shape anisotropy into the denpol backbone, future work will entail a study of these systems in the concentrated solutions and melts.

Introduction

The ability to systematically synthesize macromolecules with well-defined architecture has long been a goal of polymer chemistry as macromolecules with specific functionalities and complex architectures can possess unique properties. This is, for example, the case of graft or bottlebrush polymers, a class of branched macromolecules consisting of linear polymers as side chains grafted on the main-chain polymers. They exhibited unique properties with potential applications in the field of photonic crystals,[1] photolithograpy,[2] and biological elastomers.[3] However, a precise control on the polymer architecture is still elusive due to polydispersity and defects issues of the polymeric side chains.

Dendronized polymers (denpols) are another class of branched macromolecules bearing unique dendritic side chains. It has attracted great interest stemmed from the expectation that they combine properties of grafted polymers and dendrimers, which have highly dense multifunctional groups at the periphery and intramolecular cavities leading to high solubility and low viscosities.[4] One of the most interesting features of denpols, distinct from conventional linear polymers, is their extended or persistent conformation due to large steric repulsion among the side chains.[5,6] Hence, denpols possess rich perspectives for applications including drug delivery,[7] surface coating,[8] photonic crystals,[9] and molecular assemblies.[10] However, to what extent the polymer chains extend is often overlooked in most applications,[7−10] although chemists would agree that the denpols could transform from Gaussian coil to rod-like structures by increasing the generation of the dendron.[11] The connection between dendron size (cross-sectional radius) and chain rigidity has been rarely investigated.[12,13]

It leads us to study the chain conformation in a systemic manner. Denpols are suitable for the investigation since their excellent control on the side chains at the molecular level can eliminate dispersity and defects issues, which are inevitable for graft polymers. Therefore, it is much easier to modulate their cross-sectional profile and conformation by tuning the sizes of the side-chain dendrons with high fidelity and precision. In this context, recent theoretical work suggests that the main difference between denpols and graft polymers[14] should arise from the higher density in dendritic architectures compared to linear polymer chains.

Persistence length (lp) has been mostly used for quantifying the conformational rigidity of the polymer backbone and can be controlled by grafting density, size of the graft, and length of the main-chain backbone.[15] Although atomic force microscopy (AFM) analysis could determine lp of denpols,[16] one should be aware that AFM can overestimate because adsorption could induce stretching on the substrates.[17] Alternatively, scattering techniques[12] have been used for more reliable estimation of lp in solution, eliminating surface effects. For instance, lp ranges between 1 and 10 nm in polymers containing up to second-generation[18] carbosilane dendrons. On the other hand, lp of 30 nm was reported for the denpol containing fifth-generation aromatic amide dendrons.[12,13] We have reported lp values around 6–8 nm for polynorbornene (PNB) and poly(endo-tricycle[4.2.2.0]deca-3,9-diene) (PTD) containing third-generation ester dendrons in dilute regime by the combination of static light scattering (SLS) and dynamic light scattering (DLS).[19] However, the impact of even higher dendron generation on the denpols conformation still remains largely unexplored for denpols, obtained by graft-through approach, due to the lack of good synthetic access to them.

The main obstacle in the synthesis of large denpols is the precise installation of the bulky side chains with high fidelity. To tackle this problem, many approaches have been investigated for the preparation of well-defined denpols. For example, grafting-to method[20−25] (coupling dendrons to the precursor polymers directly) or most widely used grafting-from approach[26−30] (step by step dendronization of every generation onto the precursor polymer chain) has been successful to prepare denpols carrying up to eighth-generation dendrons.[30,31] However, these methods lead to some defects, and analysis to confirm their structure can be tedious, especially for high generations. The alternative approach is grafting-through or macromonomer method[32−42] via directly polymerizing well-defined macromonomers containing the defect-free dendrons. Although this allows the synthesis of highly pure denpols, it is the most challenging method as the steric hindrance between the bulky dendrons greatly retards the polymerization. Fortunately, recent advances of ring-opening metathesis polymerization (ROMP)[43−45] have provided an excellent solution to overcome this challenge. For example, we reported denpols of poly(endo-tricycle[4.2.2.0]deca-3,9-diene) (PTD) containing up to fourth-generation dendrons[42] and poly(norbornene) (PNB) containing up to fifth-generation dendrons,[35] as well as their block copolymers and gradient copolymers[42] by ROMP. Therefore, we envisioned that the synthetic efforts toward even larger denpols would certainly get a positive feedback from a thorough study of their conformation in solution.

Herein, we report the synthesis of large denpols (up to G6 on PNB and G5 on PTD, Scheme ) having molecular weight up to 1960 kDa along with their unique conformational analysis in solution. The consistent determination of the configurational quantities, obtained from the combination of size-exclusion chromatography multiangle laser light scattering (SEC-MALLS), static and dynamic light scattering (SLS/DLS), and small-angle neutron scattering (SANS) techniques, revealed the effect of backbone structure and generation on the conformation. The high persistence lp up to 27 nm strongly suggests a rod-like conformation of the high-generation denpols.

Scheme 1

Synthesis of the Denpols via ROMP Using Various Monomers and Catalysts

Results

Synthesis

To achieve well-defined denpols by graft-through approach, we selected highly reactive monomers such as NB[35] and TD.[42] The release of their ring strain is essential to overcome the steric hindrance during propagation of a polymer chain. We also introduced the rigid biphenyl spacer between the active monomer and dendron, which is necessary to reduce the steric hindrance during the propagation, compromising slight loss of chain rigidity.[35,42] Recently, we have reported on the conformation of denpols containing G0 and G3 dendron using static and dynamic light scattering[19] (Table , entries 1–5 and 15–16). Since the steric repulsion from the medium-size G3 dendron was moderate, these denpols exhibited semiflexible chain structure. To examine the conformation of larger denpols of higher generation, we prepared PNB with G4 and G5 ester dendrons and PTD with G4 dendron using Grubbs third-generation catalyst (1)[35] (Table , entries 6–10 and 17–18). PTD containing 2-ethyl-1-hexyl side chain[46] was prepared as a reference material for PTD-Gn (entry 14, Table ). We envisioned that comparison of the PNB and PTD chain conformation would lead to a deeper insight into the effect of the backbone structures. The obtained polymers have weight-average molecular weights (Mw) ranging from 292 to 1350 kg/mol and relatively narrow dispersity except entry 10.

Table 1

Characterizations of the Polymers

PNB				PTD
entry	namea	Mwb (kg/mol)	Đb	Entry	namea	Mwb (kg/mol)	Đb
1	PNB-287	79	1.06	14	PTD-931	292	1.16
2	PNB-1587	437	1.19
3	PNB-G3-510	666	1.26	15	PTD-G3-1060	1424	1.25
4	PNB-G3-750	978	1.55	16	PTD-G3-1376	1847	1.20
5	PNB-G3-1257	1640	1.48
6	PNB-G4-301	720	1.25	17	PTD-G4-257	626	1.17
7	PNB-G4-475	1136	1.37	18	PTD-G4-630	1532	1.15
8	PNB-G5-148	678	1.28	19c	PTD-G5-205	938	1.39
9	PNB-G5-208	949	1.45	20d,e	PTD-G5-369	1699	1.41
10	PNB-G5-295	1350	1.76
11c	PNB-G6-78	698	1.45
12c	PNB-G6-128	1145	1.77
13d	PNB-G6-220	1960	1.33

Numbers at the end of the label indicate calculated Nw.

Mw and Đ were determined by SEC-MALLS in chloroform.

Third-generation Grubbs catalyst (2) was used.

The polymer was fractionated using preparative-SEC.

Catalyst 4 was used.

To push the limit of the synthesis beyond the state of the art, we attempted to synthesize PTD containing G5 dendron and even larger PNB containing the largest G6 dendron via ROMP. Synthesis of these extremely large denpols was extremely challenging as such huge denpols have not been prepared by graft-through approach. The synthesis of the monomers NB-G6 and TD-G5 was successfully achieved by following the reported procedure,[35] and their structures were confirmed by NMR and matrix-assisted laser desorption/ionization mass spectroscopy (see Figures S3–S8). However, as anticipated, initial attempts to polymerize TD-G5 and NB-G6 by using 1 did not yield the desired polymers even at 50 °C presumably due to the low reactivity of the congested propagating species. On the other hand, ROMP of NB-G6 using highly active and thermally stable second-generation Grubbs catalyst (2) in toluene at 70 °C afforded the denpols with moderate weight average degree of polymerization (Nw = Mw/M0, where M0 is the molecular weight of a monomer) of 78 and dispersity of 1.45 (Table , entry 11).

Encouraged by this initial result, we further optimized this ROMP in THF at 60 °C, where the faster ROMP was reported compared to toluene[47] and the Mw increased to 1145 kg/mol with Nw of 128 (Table , entry 12). Further trials to obtain higher-molecular-weight polymers by increasing reaction concentration, changing the solvent from THF to 1,2-dichloroethane (DCE) and increasing temperature to 70 °C, were not satisfactory (see Table S1).

As an alternative strategy, we switched the catalyst to the recently developed Ru catalysts containing cyclic (alkyl)(amino)carbine (CAAC), which showed even higher activity toward ring-closing metathesis (RCM)[48] and cross-metathesis (CM)[49] reaction, hoping that it would do the same to ROMP. Indeed, after several trials using several catalysts including 3, we obtained the denpols having larger fraction of high molecular weight despite the broader dispersity (see Table S1), and fortunately, fractionating out their low-molecular-weight portion via preparative-SEC successfully provided the denpol having Mw of 1960 kg/mol with Nw of 220 and narrow dispersity (Table , entry 13) (see Table S1). Notably, this enables us to explore the behavior of high-molecular-weight G6 denpols in the dilute solution for the first time. Likewise, using the new optimized condition to prepare PNB-G6, ROMP of TD-G5 using G2 in THF at 60 °C afforded the PTD-G5 with high Mw of 938 kg/mol, Nw of 205, and moderate dispersity of 1.39 (Table , entry 19). Furthermore, using CAAC catalyst (4) and fractionalization (see Table S2) afforded the PTD-G5 denpols having even higher Mw of 1699 kg/mol corresponding to Nw of 369 with moderate dispersity (Table , entry 20).

Persistence Length from SEC MALLS Rg vs M

To systematically investigate the relationship between dendron generation and conformation, we analyzed the denpols by MALLS, providing the Mw and the radius of gyration (Rg). The relationship between Mw and Rg provides a basis to extract persistence length (lp) using standard wormlike chain models. These quantities are obtained on monodisperse elution by SEC-MALLS so that polydispersity effects can be ignored. We present Rg as a function of effective contour length (L = b0N), where b0 is the length of a repeat unit. The evolution of Rg with L within each fraction of the PNB and PTD denpols is depicted in the double-logarithmic plot (Figure a,b). With increasing generation in both denpols, the observed Rg (L) dependence becomes stronger (steeper slope), implying increase of lp.

Figure 1

Radius of gyration (Rg) as a function of weight-averaged contour length (L), obtained from MALLS-SEC measurement, in THF, for PNB-Gn denpols (a) and PTD-Gn denpols (b). The various dashed lines represent the Benoit–Doty model (eq S1) for different values of persistence length. The red solid lines in (a) and (b) represent the rod scaling behavior. The dependence of the radius of gyration contour length in Benoit–Doty master presentation: Double-logarithmic plot (c) and semilog presentation (d). The plateau region in (d) indicates the flexible coil limit, Rg ∼ (L/lp)1/2..

The experimental data of Figure were well represented (dashed lines) by the Benoit–Doty expression (see Section III.5.1. in the Supporting Information (SI)) using lp as a single adjustable parameter. For all fractions, the experimental Rg values were well superimposed into two master curves representing both limiting rod and coiled behaviors in Figure c,d. For the rod-like regime, the scaling Rg ∼ L1 was tentatively observed in the low L/lp limit, while a flexible coil, Rg ∼ L1/2, was obeyed at high L/lp (>50). The distinct crossover from the rod-like to the flexible coil chain conformation according to the Benoit–Doty master plots (Figure d) occurred at higher L (or Nw) with increasing dendron generation, as revealed in Figure a,b. PNB-G6 and PTD-G5 approached the rod regime (red solid lines), whereas denpols with low generation falls near the flexible-chain regime (dotted black lines in Figure a,b). This general trend is consistently reflected in the increase of lp (from 3.5 to 27 nm) in THF, as listed in Table . The same analysis in chloroform solution gave the similar results (see Table S5), demonstrating their rod-like conformation. Moreover, based on the low but positive values of the second virial coefficient (A2 < 10–4 mol·cm3·g–2), determined by SLS, both CF and THF display similar weak repulsive denpol–denpol interactions, implying similar solvent quality. AFM images showing single chains further support their rigid conformations inferred from the scattering data in solution (see Figures c and S18).

Table 2

Molecular Parameters (Contour Length, L; Radius of Gyration, Rg; Hydrodynamic Radius, Rh; and Persistence Length, lp) of Selected Polymers in Dilute Solutions in THF

	DLS	SANS			GPC-MALLS
code	Rh (nm)	LKa (nm)	lKa (nm)	Rcsa (nm)	Nw	L (nm)	lpb (nm)	Rg (nm)
PNB-287	8.6	93 ± 5	2.0 ± 0.5	N/A	370	185 ± 9	c	c
PNB-G3-1257	32.6	698 ± 35	6.0 ± 0.8	1.8	1182	591 ± 30	9.9	52.1
PNB-G4-301	19.6	240 ± 12	6.4 ± 0.8	1.9	313	157 ± 8	8.0	19.5
PNB-G5-295	24.2	217 ± 11	10.2 ± 1.0	2.3	271	135 ± 7	13.5	26.8
PNB-G6-128	15.4	75 ± 4	15.0 ± 1.5	2.8	134	67 ± 3	27.3	24.9
PTD-G3-1060	17.8	169 ± 9	13.7 ± 1.4	2.2	1444	534 ± 27	11.3	47.7
PTD-G4-630	16.0	113 ± 6	24.2 ± 2.4	2.4	600	222 ± 11	20.1	35.6
PTD-G5-205	16.3	65 ± 3	14.3 ± 1.4	2.8	175	65 ± 3	20.2	18.1

Contour length, LK; cross-sectional radius, Rcs; and persistence length, lK were obtained from the representation of the SANS form factor by Kholodenko model.[50,51]

Obtained from SEC-MALLS data fitting using the Benoit–Doty equation (eq 1 in the SI).

Rg (<10 nm) is too small to be measured by MALLS detector.

Figure 2

(a) Schematic representation of macromononers dimensions, (b) denpol chain within a wormlike conformation. (c) Representative AFM images of PNBG6-220 and (d) PTDG5-369.

SANS Patterns and Internal Structure

The emerged structure of the two denpols is schematically shown in Figure . The cross-sectional radius, Rcs, cannot be obtained by SEC-MALLS enabling access only to the low wave-vector fraction of the form factor sufficient to determine Rg. To obtain the full form factor with access to the local structure of the denpols in dilute solutions, the higher resolution (2 × 10–3 nm–1) offered by the SANS technique is necessary. The normalized intensity patterns for the PNB denpols in dilute THF solutions, from G0 to G6, are shown as log(I) vs log(q) in Figure a and I × q/I0 vs log q in Figure b. Analogous scattering curves recorded for various PTD are available in Figure S15. The scattering profile of PNB-G0 (black squares) conforms to the form factor of a flexible polymer chain with typical scaling I(q) ∼ q–2 at high q’s, as indicated in Figure a. However, for the denpols, the form factor is distinctly different at high q’s (>0.6 nm–1), exhibiting steeper decay, which is a sign of the finite cross-sectional dimension of denpols (Figure a). At intermediate wave-vectors, the characteristic q–1 dependence of semirigid chains for qlp ∼ 1 is observed, whereas for qRg < 1, a low q Guinier plateau is reached.

Figure 3

(a) Scattering intensities from the combined SANS and SLS experiments on PNB-Gn denpols in dilute THF-d8 solution (c = 5 g/l) as a function of wave-vector q. The data along with their representation (dashed lines) have been shifted vertically for clarity. (b) Representation (red dashed lines) of the scattering intensity, normalized by the forward scattering intensity, and the wave-vector (for rods) by the Kholodenko form factor.[50−52]

We used the Kholodenko form factor as it is recognized to be best adapted for the description of semiflexible cylinder (Figure ).[50−52] The representation requires an intensity (I0) prefactor and three size parameters, contour length LK, persistence length lp,K, and cross-sectional radius Rcs (Figure b), as described in SI-II.5 (eqs 3–5). The experimental intensity patterns were well represented by the theoretical form factor (red dashed lines in Figure a). The quality of the Kholodenko model representation is better illustrated in the semilogarithmic plot I.q/I0 vs q (Figure b), and the values of the adjustable three size parameters are reported in Table . Notably, LK and lp,K are in good agreement with the values of the corresponding parameters obtained from SEC-MALLS.

Hydrodynamic Radius

Angular-dependent DLS measurements lead to the straightforward determination of diffusion coefficient D and hydrodynamic radius Rh. The experimental values of Rh for the different samples are reported in Table . The hydrodynamic radius Rh ∼ D–1 is expected to reflect the conformation and the hydrodynamic conditions of the denpols in dilute solutions. It is primarily sensitive to many factors such as the overall size of the denpol, rigidity of a polymer chain, thickness (cross-sectional radius, Rcs), solvent quality, as well as the type of branching.[53,54] However, the contribution of the different parameters is not straightforward to distinguish.[55,56] Different models have been reported for the diffusion of semiflexible chains. We have chosen a model based on an accurate description of the polymer chain motion and standard hydrodynamic assumption (Zimm model) developed by Winkler et al.[55,57] As in the case of the SANS, the representation of D also requires the same three size parameters (contour length, persistence length, and hydrodynamic cross-sectional radius).

We utilized the values of Rcs and lp obtained from MALLS and SANS to compute Rh as a function of L using the Winkler model.[57]Figure compares the theoretical (green and purple dashed lines) with the measured Rh data as a function of the chain contour length. Both axes Rh and L have been rescaled by the persistence length, to adopt a semiflexible chain master plot representation in analogy to Figure c,d for Rg. The evolution of the latter with length, also included (red line, Section II.5 in the SI, eq 2), is systematically smaller than Rh. As expected at high L/lp in the Gaussian coil limit (dashed lines in Figure ), Rh follows a similar scaling to Rg (Rg/Rh = 1.72), i.e., Rh(L) is equivalent to Rg(L). The agreement is moderate at low L/lp, possibly because the model is not well adapted for polymers with low L/Rcs, i.e., for thick and short species. In this region, the simple L/lp scaling does not hold any more and Rg/Rh depends on the system specifics and the precise value of Rcs becomes crucial. Nevertheless, the crossover from rod-like scaling Rh ∼ LW1 and to flexible coil Rh ∼ LW1/2 scaling occurs around LW/lp ∼ 10, consistent with the Rg data (Figure d). Hence, the persistence length can be also estimated from translational diffusion data, as it was reported in our previous work.[19]

Figure 4

Hydrodynamic radius as a function of weight-averaged contour length for the two different backbones, PNB (black spheres) and PTD (red spheres), obtained from SLS/DLS experiments in THF and rescaled by lp from SANS. The red line presents the radius of gyration variation from the Benoit–Doty equation. The dashed lines denote the theoretical prediction for the hydrodynamic radius for the upper (lp= 27.3 nm, Rcs = 2.8 nm, purple) and lower limits of persistence length and cross-sectional radius (lp= 6.8 nm, Rcs = 1.8 nm, green). Theoretical curves for thicknesses and rigidities as the ones measured experimentally fall in the marked area between the two curves. The black dotted lines show the contour length dependence for Gaussian coil (R ∼ Lw0.5) and rigid rod (R ∼ Lw1).

Discussion

Based on the simulation work on bottlebrush polymers,[58,59]lp increases with the degree of polymerization of the side chains (Ns) and counterintuitively with the main chain (N). The dependence lp ∼ Ns leads to deviation from the Kratky–Porod wormlike chain model. However, both predictions are not supported by the experimental evidence that lp is robust over Ns and N variations. For the present denpols, we utilized the wormlike chain model (WLC) consistently to represent form factor and translational diffusion. The former, obtained over broad wave-vector range, combining light and neutron scattering, was found to conform to the model including the denpols cross-sectional radius. The good agreement between GPC-MALLS, SANS and diffusion values supports the applicability of the WLC model in the case of the denpols. As a mesoscopic model cannot provide insight into the origin of the persistency, it cannot resolve the different possible contributions to denpols rigidity. The radial density profile cannot be resolved by SANS, but the appearance of a clear minimum at high q’s (for the large generation) hints the formation of a steep concentration gradient. The consistent analysis of the results neither questions the validity of the WLC model to describe the conformation of the large Gn denpols in solutions nor provides insight into the origin of the measured “apparent” persistence length.

The study on the PNB and PTD systems yields insight into the effect of increasing dendron generation on the conformation of the denpols. Two factors can influence chain conformation: dendron molecular weight (Msc) and branching. The former impacts the conformation since both Rcs and lp increase with generation, as depicted in the log–log presentation of Rcs vs Msc in Figure a. In the case of graft polymers in theta solvent,[60]Rcs is expected to grow as Msc0.5. As to the effect of branching, Borisov et al.[14] reported the scaling, Rcs ∼ Nsc0.75, with Nsc being the number of monomers in the side chain, in the presence of excluded volume interaction (good solvent), while Rcs is rather independent of the branching of the side chain. We note that there is a paucity of experimental results for validation of the theoretical prediction. In the present case, both PTD and PNB denpols show a rather weak increase of Rcs with generation, which conforms to Rcs ∼ Msc0.25 and not to the brushlike Rcs ∼ Msc0.5 dependence. The weak increase with generation indicates that the high-generation denpols are denser than the lower-generation denpols rendering their shape close to colloidal nano-objects. It might also imply dense packing near to the polymer main chain.

Figure 5

(a) Excess persistence length obtained from SANS (filled symbols) and MALLS-GPC analysis (opened symbols). The dashed line represents a linear scaling. lp,0 is the persistence length of the linear chain. (b) A log–log plot of the cross-sectional radius vs dendron generation or equivalently the side-chain molecular weight. The two dashed lines represent different power law dependencies indicated in the plot. (c) Ratio of the persistence length to the cross-sectional radius as a function of generation in log–log presentation.

The persistence lengths obtained from SEC-MALLS (opened symbols) and from SANS (filled symbols) with each generation are reported in Figure b as a function of Msc. The excess persistency (lp,0 = 3.5 nm being the persistence length of the linear chain “precursor”) increases with generation reaching the value of 27 nm for PNB-G6. The observed increase in lp is almost linear to Msc. In fact, lp ∼ Msc was proposed by SCF calculation for bottlebrush polymers.[61]

According to Borisov et al.,[14] this relationship should be robust over the structure of the side chain, implying that a linear side chain would increase the persistence length in the same way as a branched side chain with the same side-chain molecular weight. This conjecture concerning structural similarities between denpols and bottlebrush polymers remains to be examined.

Both lp and Rcs are different for PNB and PTD bearing the same side chains (Figure a). The augmented disparity in the value of lp is already evident in the form factors of PNB and PTD of G4 in the intermediate q-range (Figure S17). The scattering intensity from PTD-G4 with lp = 24 nm is higher than that for PNB-G4 with lp = 6.4 nm. The PTD backbone is more persistent than the PNB backbone of the same generation, and the difference is more marked for high generations.[62] The backbone effect on the persistency could arise from an effectively higher grafting density per unit length in PTD as a result of the compact monomer structure; b0 of TD-Gn macromonomer (b0 = 0.37 nm) is smaller than that of NB-Gn macromonomer (b0 = 0.5 nm) (see Figure or 2). Hence, the grafting density of PTD-Gn is about 25% larger than that for PNB-Gn. According to Borisov and co-workers,[14] the ratio lp/Rcs is an important structural quantity also for nondilute denpols solutions, since it can control the anisotropic excluded volume interactions between denpols,[61] thereby forming liquid crystalline phase.[63] Since lp/Rcs was found to increase with generation for both PTD and PNB backbones, as displayed in Figure c, the persistence length should increase faster than Rcs. However, their values remain below 10, which renders the formation of a liquid crystalline phase

Conclusions

The dendronized PNB containing up to sixth-generation and PTD containing up to fifth-generation dendrons were synthesized by ROMP via grafting-through method for the first time. Successful preparation of unique series of denpols with huge molecular weight up to 1960 kDa enabled us to unravel their structure in solution. The results from SEC-MALLS, DLS, and SANS experiments can be well understood using a semiflexible cylinder model with three characteristic sizes: contour length, persistence length lp, and cross-sectional radius Rcs.

The high-generation denpols display a rod-like conformation, with grafting-induced persistence lp ≃ 27 nm backbone. The increase of lp with generation appears to be governed by the molecular weight of the dendron. Compared to the persistence length, the cross-sectional radius Rsc displays a weaker increase with generation, resulting in an increase of aspect ratio lp/Rcs with generation. Interestingly, the PTD denpols are clearly more persistent than the PNB denpols of the same generation due to the more compact structure and higher grafting density in the PTD. Concerning the variation of the translational diffusion with generation, the shorter G6 (Nw = 78) is probably closer to a spherical/ovalic rigid object than semirigid chains in conformity with AFM images. For the other samples with sufficiently high contour length, the value of Rg/Rh (Figure ) conforms to the wormlike chain models.

High-generation denpols are expected to exhibit excluded volume interactions with strength falling between flexible chains and cylindrical nano-objects. The structure and dynamics of such nano-objects in dense states (semidilute solutions to melt) are anticipated to be distinct[64−66] as estimated from their spatial overlapping and possible formation of liquid crystalline phases.