Synthesis of Met allopolymers and Direct Visualization of the Single Polymer Chain

Abstract

During the past few decades, the study of the single polymer chain has attracted considerable attention with the goal of exploring the structure–property relationship of polymers. It still, however, remains challenging due to the variability and low atomic resolution of the amorphous single polymer chain. Here, we demonstrated a new strategy to visualize the single metallopolymer chain with a hexameric or trimeric supramolecule as a repeat unit, in which Ru(II) with strong coordination and Fe(II) with weak coordination were combined together in a stepwise manner. With the help of ultrahigh-vacuum, low-temperature scanning tunneling microscopy (UHV-LT-STM) and scanning tunneling spectroscopy (STS), we were able to directly visualize both Ru(II) and Fe(II), which act as staining reagents on the repeat units, thus providing detailed structural information for the single polymer chain. As such, the direct visualization of the single random polymer chain is realized to enhance the characterization of polymers at the single-molecule level.

Graphical Abstract

INTRODUCTION

Over the past decades, to enhance the understanding of the structure–property relationship of polymers, characterization of single polymer chain has attracted widespread attention.^1,2 Microscopy, such as scanning tunneling microscopy (STM) or atomic force microscopy (AFM), allows us to observe the molecular structures directly. However, except for the large biomolecules such as DNA,^3–7 the imaging of structure as well as conformation is mainly limited to crystalline, liquid crystalline, and the aggregation of polymer chains.^8–13 With densely grafted side-chains, the worm-like polymer bottle-brushes can be observed directly,^14–23 yet the dense and long side chains severely affect the polymeric backbone conformation.^24,25 A formidable challenge still remains to address the random coil conformations of amorphous single polymer chain due to the variability and low atomic resolution.

As is known to all, metals or metal ions with high electron density can afford relatively strong signals and enhanced atomic resolution under microscopy like STM. Therefore, if one can introduce these metal components into polymer chains as “staining reagents”, the amorphous single polymer chain could be easily observed and confirmed by STM depending on the position of metal. Generally, metal components are introduced into polymers through three approaches, viz., (1) polymerization of monomers with metal ions in the main chains or pendant groups,^26–30 (2) supramolecular polymerization through coordination of metal ions with organic monomer in the main chains,^31–41 and (3) metalation of pendant groups through postpolymerization modification.^42–53 However, these approaches have certain limitations. In the first approach, the polymerization is typically centered on small and stable metal–organic components, e.g., metallocenes, which are polymerized through living radical polymerization, ring-opening metathesis polymerization (ROMP), and so on.^54–56 In the second approach, the supramolecular polymerization is mainly achieved through dynamic coordination interaction.^{32,36,37,39–41,57} Therefore, precise control of the polymerization is difficult for dynamic supramolecular polymers with labile coordination bonds. In the third approach, metalation of pendant groups after polymerization is also limited to metal with weak coordination to get low-defect polymers.^42,44–48 To date, very few systems were able to harness both strong and weak coordination metal ions together into metallopolymers with discrete supra-molecules as repeat units.

Thanks to the recent flourish of 2D and 3D metallosupramolecular structures with precisely controlled sizes and shapes constructed by coordination-driven self-assembly,^58–82 we introduced two discrete structures, hexameric and trimeric metallosupramolecules, as repeat units to construct metallopolymers due to the high symmetry as well as the appropriate size for STM imaging (Scheme 1). The copper(I)-catalyzed azide–alkyne cycloaddition (CuAAC) “click” reaction, which owns the advantages of fast reaction rates and high yields with mild reaction conditions,^83–86 was selected for the polymerization. In this design, we bridged the 2,2′:6′,2″-terpyridine (tpy) ligand and Ru(II) with strong coordination interaction to create the open-form metal–organic hexameric and trimeric monomers with free tpy.

Scheme 1.

Monomers (A, Hex-M; B, Tri-M and Alkyne-PEG) and Their Corresponding Metallopolymers with 2D Hexameric (C) and Trimeric (D) Supramolecules as Repeat Units

After polymerization by the click reaction, through further metalation of Fe(II) with weak coordination, both strong and weak binding metal ions were harnessed into one polymer system in a stepwise manner. Our initial assumption was that the metalation of Fe(II) of metal–organic monomers would form the ring-closing hexagon or triangle as repeat units, and we could measure the structure and conformation of the amorphous single polymer chain based on the location of Ru(II) and Fe(II). Unexpectedly, according to the results of ultrahigh-vacuum, low-temperature scanning tunneling microscopy (UHV-LT-STM) and scanning tunneling spectroscopy (STS), both the hexagon assembled by intramolecular coordination and linear chains assembled by intermolecular assembly were observed. In contrast, only the intramolecular assembly into the triangle existed in trimeric repeat units.

RESULTS AND DISCUSSION

Scheme 1 depicts the two metal–organic monomers with open-form structures, viz., hexameric monomer (Hex-M) and trimeric monomer (Tri-M), which were prepared according to the procedure as shown in the Supporting Information (Schemes S1–S5 and Figures S1–S76). Briefly, Hex-M was synthesized through a well-established stepwise strategy based on the strong connectivity of 〈tpy-Ru(II)-tpy〉,^87–94 by which six ditopic tpy ligands bearing a 120° angle were bridged linearly with five Ru(II) ions. Similarly, two Ru(II) ions were used to bridge three ditopic tpy ligands with 60° angle together to afford Tri-M. Note that Hex-M and Tri-M are functionalized by dual azide groups, which were designed for the next polymerization by click reaction. The ¹H NMR spectra of Hex-M and Tri-M are shown in Figures 1A and 1E, with proton labeling at the top of Scheme 1. The assignment of peaks associated with the chemical structure was assisted by the two-dimensional (2D) ¹H–¹H gCOSY and NOESY NMR spectra (Figures S41–S44 and S71–S74). After treating the precursors (compounds 8 and 16) bearing chloride as terminal groups with sodium azide, the triplet peak for the methylene unit near chloride at around 3.59 ppm totally disappeared with a new triplet peak at 3.28 ppm appearing, demonstrating that the conversion was completed. The integrals from corresponding peaks fit well with the theoretical value, indicating the high purity of Hex-M and Tri-M.

Figure 1.

¹H NMR spectra of (A) Hex-M (600 MHz), (B) Hex-P (600 MHz), (C) Alkyne-PEG (400 MHz), (D) Tri-P (400 MHz), and (E) Tri-M (400 MHz) in CD₃CN. The asterisk (*) represents the residue solvent CHCl₃.

The Hex-M and Tri-M monomers were further characterized by electrospray ionization–mass spectrometry (ESIMS). Hex-M (Figure 2A) was observed with a dominant set of peaks with continuous charge states from 6+ to 10+ due to successive loss of the counterions, PF ^–₆. The experimental isotope pattern of each charge state agrees well with its theoretical simulation result (inset and Figure S46). The averaged molecular mass is deconvoluted as 7019 Da, which matches the chemical composition [C₃₄₀H₂₉₄N₄₂O₆Ru₅P₁₀F₆₀]. Figure 2C shows the ESI-MS of Tri-M with continuous charge states from 3+ to 4+. The experimental isotope pattern of each charge state is also consistent with the simulated one (inset and Figure S76). The averaged molecular mass is deconvoluted as 3375 Da, which has a good agreement with the chemical composition [C₁₆₈H₁₄₂N₂₄O₆Ru₂P₄F₂₄]. Furthermore, traveling wave ion mobility–mass spectrometry (TWIM-MS) spectra (Figure 2B,D) show a series of narrowly distributed signals with successive charge states for both Hex-M and Tri-M, indicating high isomeric purity of the structures.

Figure 2.

ESI-MS of Hex-M (A), TWIM-MS plot (m/z vs drift time) of Hex-M (B), ESI-MS of Tri-M (C), and TWIM-MS plot of Tri-M (D).

Alkyne-terminated pentaethylene glycol (Alkyne-PEG, Scheme S6 and Figures S77–S79) was selected as another component to polymerize with the monomers (Hex-M and Tri-M) to construct the corresponding metallopolymers (Hex-P and Tri-P) by the click reaction (Scheme S7). During the polymerization process in DMF, a large amount of precipitation was produced, suggesting the formation of metallopolymers (Hex-P and Tri-P). The products were fully characterized by NMR and matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF). Figure 1 shows the ¹H NMR spectra of Hex-P (Figure 1B) and Tri-P (Figure 1D), while 2D ¹H–¹H gCOSY and NOESY NMR spectra are provided in the Supporting Information (Figures S80–S89). The triplet peak at 3.28 ppm corresponding to a methylene unit near the azide group of the monomers (Hex-M and Tri-M) disappeared, indicating monomers were completely reacted. After the formation of triazole by click reaction, this triplet peak was significantly shifted downfield (~1.23 ppm) and the protons of alkynyl group were dramatically shifted downfield (~4.99 ppm).

Because of the strong electrostatic interactions between charged polymer backbones and counterions, the characterization data by size exclusion chromatography, viscometry, and laser light scattering may come from the associations rather than the single chains of the metallopolymers, which can be regarded as a kind of polyelectrolyte. Because all the solution-based techniques have the similar problem, MALDI-TOF mass spectrometry was further used to characterize the metallopolymers with open-form monomers in their solid states. In contrast to losing multiple counterions for ESI-MS of monomers, such metallopolymers are mainly ionized by losing a single counterion, that is, PF₆⁻ in this study. Hex-P (Figure 3A) shows a series of peaks with an interval of ca. 7340 Da, in good agreement with the molar mass of the repeat unit in Hex-P (Hex-M with Alkyne-PEG, C₃₅₆H₃₂₀N₄₂O₁₂Ru₅P₁₀F₆₀; theoretical value: 7334 Da). A high molecular weight signal up to 73400 (m/z) was detected with the degree of polymerization of 10, directly confirming the successful synthesis of the corresponding metallopolymer Hex-P. In terms of polymers with a large polydispersity index value (PDI), it is worth noting that MALDI-TOF generally displays a nonsymmetrical distribution of peaks for polymers with higher abundant signals at low m/z range due to the high ionization efficiency of species with low molecular weight.^95–97 This also suggested that large oligomers with a high polymerization degree might exist in our system; however, they were not ionized due to the limitation of MALDI-TOF. With small monomer Tri-M, a much higher degree of polymerization was detected in Tri-P as shown in Figure 3B. A high molecular weight signal up to 70000 (m/z) was obtained with the degree of polymerization of 19, indicating the successful synthesis of corresponding metallopolymer Tri-P. The average interval between adjacent peaks is about 3684 Da, which agrees with the molar mass of the repeat unit in Tri-P (Tri-M with Alkyne-PEG, C₁₈₄H₁₆₈N₂₄O₁₂Ru₂P₄F₂₄; theoretical value: 3689 Da).

Figure 3.

MALDI-TOF MS of Hex-P (A) and Tri-P (B) with corresponding numbers of repeat units in linear mode.

After polymerization, Fe(II) with weak coordination was introduced for further metalation.^{35,42,98–103} To avoid the undesired cross-link between different polymer chains, a low concentration (0.2 mg mL⁻¹) of metallopolymers (Hex-P and Tri-P) was used in the metalation process. The metalation results were supported by the ultraviolet–visible (UV–vis) absorption spectra as shown in Figure S90. In both metallopolymers, the expected peak at about 496 nm was ascribed to the metal-to-ligand charge-transfer transition (MLCT) of 〈tpy-Ru(II)-tpy〉 connectivity.^31,101,104 After Fe(II) was introduced, a new emerging peak at about 574 nm that was attributed to the characteristic MLCT band of the 〈tpy-Fe(II)-tpy〉 was observed, indicating successful metalation.^{35,98,99,101–104}

Furthermore, ultrahigh-vacuum, low-temperature scanning tunneling microscopy (UHV-LT-STM) was then used in an effort to characterize metallopolymers and address the structure and conformation of polymer chains. Samples were dissolved in acetonitrile and deposited on an Ag (111) surface by drop-casting. It should be noted that to clearly observe a single metallopolymer chain, a low concentration (0.05 mg mL⁻¹) was used to prepare the STM sample. Because of the octahedral coordination geometry and higher electron density around the metal ions, the 〈tpy-Ru(II)-tpy〉 or 〈tpy-Fe(II)-tpy〉 units as the stained spots gave rise to a relatively strong signal in the form of a bright lobe when compared to the organic portions.¹⁰⁵

Figure 4A and Figure S92 show the STM images of Hex-P, in which each of the five bright lobes located closely together correspond to five connected 〈tpy-Ru(II)-tpy〉 units, indicating the successful formation of metallopolymers. Aggregation occurred due to the poor solubility of Hex-P in acetonitrile, making it difficult to observe longer polymer chains, and only large bright spots are observed for aggregates as shown in Figure S92. We also used DMF and DMSO as solvents, which however caused severe background signals during UHV-LTSTM imaging. After self-assembly with Fe(II), six bright lobes including five 〈tpy-Ru(II)-tpy〉 units and one 〈tpy-Fe(II)-tpy〉 unit are observed as a group (Figure 4C and Figure S93). As expected, the repeat units of Hex-P underwent intramolecular coordination into a ring-closed structure, and six bright lobes arranged in a cyclic hexagonal structure (Hex–P-C). The distance between two adjacent bright lobes of the cyclic hexagonal structure is around 2.1 nm, which is well consistent with the calculated size by modeling (Figure S91). Also, because of the poor solubility, we were only able to observe one metallopolymer chain with three cyclic hexagonal repeat units (Figure S93). Besides intramolecular coordination, intermolecular coordination into a linear chain (Hex-P-L) was also observed as shown in Figure 4E and Figure S93. It should originate from the constraint of the repeat units in the metallopolymer chain, where Alkyne-PEG brings two monomers close enough for the occurrence of intermolecular complexation.

Figure 4.

(A) STM image of Hex-P on the supporting Ag(111) surface. (B) STM image of Hex-P with STS data acquisition locations on 〈tpy-Ru(II)-tpy〉 unit shown in red. (C) STM image of Hex–P-C. (D) STS profiling of Hex–P-C with 〈tpy-Ru(II)-tpy〉 unit in red and 〈tpy-Fe(II)-tpy〉 unit in blue. (E) STM image of Hex-P-L. (F) STS profiling of Hex-P-L with 〈tpy-Ru(II)-tpy〉 unit in red and 〈tpy-Fe(II)-tpy〉 unit in blue. The blue line in the inset represents the polymer chain proposed by 〈tpy-metal(II)-tpy〉 units. dI/dV–V results of STS of (G) Ru(II) (red) and (H) Fe(II) (blue) correspond to different electronic features in (D) and (F). Scale bar: 10 nm.

The positions of Fe(II) and Ru(II) ions in the repeat units were identified by the means of dI/dV–V using scanning tunneling spectroscopy (STS) which probed local density of states (LDOS).^106,107 Tunneling spectra were measured by positioning the STM tip above each lobe of the repeat units on the Ag (111) surface at a fixed height with a bias range of ±2V. From dI/dV–V spectroscopic data, the energy gap between the HOMO and LUMO can be directly obtained.¹⁰⁸ Thus, from the STS measurements of each 〈tpy-metal(II)-tpy〉 unit, we were able to differentiate the Fe(II) and Ru(II) centers based on their different energy gaps. Among the six bright lobes of a group, one bright lobe in each group in Figure 4C,E gives the HOMO–LUMO gap of 2.4 eV (Figure 4H), while the remaining five bright lobes provide the gap value of 2.8 eV (Figure 4G). Thus, we can identify the site with a smaller gap, viz. 2.4 eV, as Fe(II) and that of the larger gap, viz. 2.8 eV, as the Ru(II) site. Then, STM imaging was able to directly read the “sequence” of metal ion within the polymer chains (Figure 4D,F), in which red and blue dots represent the 〈tpy-Ru(II)-tpy〉 and 〈tpy-Fe(II)-tpy〉 units, respectively. More importantly, the polymer chain can be profiled depending on the positions of metal ions as shown in Figure 4F, by which we were able to deduce the conformation of the polymer chain.

In the STM image of Tri-P, each of the two bright lobes positioning closely together represents two connected 〈tpy-Ru(II)-tpy〉 within a repeat unit (Figure 5A and Figure S94). With the introduction of Fe(II), all of the repeat units performed intramolecular coordination to form a ring-closing structure, where three bright lobes arranged into a triangle (Tri-P-C) as evidenced by STM imaging (Figure 5C and Figure S95). The distance between two adjacent bright lobes of the cyclic triangular structure is ca. 1.2 nm, which is well consistent with the modeling size (Figure S91). It should be noted that with enhanced solubility up to 12 triangular repeat units were observed in a single polymer chain, much higher than that in the hexameric system. This result is consistent with the MALDI-TOF MS data as shown in Figure 3. The positions of Fe(II) and Ru(II) ions in the repeat unit are also identified by the means of dI/dV–V obtained in STS (Figure 5E,F). We then labeled the 〈tpy-Ru(II)-tpy〉 and 〈tpy-Fe(II)-tpy〉 units with red and blue dots, respectively, on the basis of STS results (Figure 5B,D). The polymer chain is readily profiled by using the blue dots because all 〈tpy-Fe(II)-tpy〉 units are located right in the middle of the chain. Meanwhile, the conformation could be estimated and proposed according to the polymer chain drawn as shown in the inset of Figure 5D. It should be noted that theoretically the molecular weight and polydispersity index for the metallopolymers can be calculated based on the observed polymer single chains from STM images. However, to obtain a statistically average value, hundreds of high-resolution STM images without strong aggregation are needed, which is not practical currently.

Figure 5.

(A) STM image of Tri-P on the supporting Ag(111) surface. (B) STS profiling Tri-P with the 〈tpy-Ru(II)-tpy〉 unit in red. (C) STM image of Tri-P-C. (D) STS profiling of Tri-P-C with the 〈tpy-Ru(II)-tpy〉 unit in red and the 〈tpy-Fe(II)-tpy〉 unit in blue. The blue line in the inset represents the polymer chain proposed by 〈tpy-Fe(II)-tpy〉 units. dI/dV–V results of STS of (E) Ru(II) (red) and (F) Fe(II) (blue) correspond to different electronic features in (D). Scale bar: 5 nm.

CONCLUSIONS

In summary, a novel class of metallopolymers were synthesized by harnessing Ru(II) with strong interaction and Fe(II) with weak interaction into one polymer system in a stepwise manner. First, open-form metal–organic monomers with Ru(II) were polymerized by click reaction to form the corresponding metallopolymers (Hex-P and Tri-P). After further metalation by Fe(II), Hex-P and Tri-P displayed different coordinations or self-assembly behaviors. More importantly, the structures and conformations of single polymer chains can be directly visualized through STM and STS of the 〈tpy-metal(II)-tpy〉 units. The metallopolymers generated by Hex-P include not only hexagons constructed by intramolecular coordination but also hexameric linear chains formed by intermolecular coordination. In a sharp contrast, the further coordination of Tri-P was performed in a more controlled manner with only intramolecular complexation observed to form triangles as the repeat units. The different results are attributed to the size and rigidity of the repeat units. Overall, our endeavors will bring innovative aspects into polymer chemistry field and shed light on the development of innovative approaches to synthesize metallopolymers and study the structure and conformation of polymer chains at the single-molecule level. The characterization with the combination of STM and STS could be expanded into other metallopolymer systems to advance our understanding of the structure–property relationship of polymers.