Atomic-Scale Imaging of Polymers and Precision Molecular Weight Analysis Authors: Arkadios Marathianos, Alexandros Magiakos, Yisong Han, Ana Sanchez, Richard Whitfield, Jochen Kammerer, Athina Anastasaki, Paul Wilson, Joseph P Patterson, Christopher Barner-Kowollik, Evelina Liarou Published: 2024 ABSTRACT AbstractPolymer design requires fine control over syntheses and a thorough understanding of their macromolecular structure. Herein, near-atomic level imaging of polymers is achieved, enabling the precise determination of one of the most important macromolecular characteristics: molecular weight. By judiciously designing and synthesizing different linear metal(loid)-rich homopolymers, subnanoscale polymer imaging is achieved through annular dark field-scanning transmission electron microscopy (ADF-STEM), owing to the incorporation of high Z atoms in the side chain of the monomeric units. The molecular weight of these polymers can be precisely determined by detecting and counting their metal(loid) atoms upon ADF-STEM imaging, at sample concentrations as low as 10 μg·mL–1. Notably, a commonly used C, H, and O-containing polymer (i.e., poly(methyl acrylate)) that was thus far inaccessible at the atomic scale is derivatized to allow for subnano-level imaging, thus expanding the scope of our approach toward the atomic-level visualization of commodity polymers. The design of soft matter with predefined properties necessitates the (sub)nanoscale analysis of polymers, tailored with precision for significant performance.1,2 A fundamental characteristic of polymers is their molecular weight (MW). The leading techniques for MW determination are size exclusion chromatography (SEC), 1H nuclear magnetic resonance (1H NMR), and high-resolution mass spectrometry (HR-MS) (Scheme 1). Although well-established, these techniques possess significant limitations when complex systems are targeted, including organometallic3,4 and conjugated polymers,5,6 or complex architectures.7,8 For instance, SEC requires the combination of suitable solvents, columns and MW standards,91H NMR requires distinctive end-groups, while topologically complex polymers, with high dispersity and MW, are not suitable for HR-MS.10,11 Schematic Illustration of the Methods Traditionally Used to Determine the MW of Polymers and Our Approach through ADF-STEM Acknowledging those challenges, Junkers and colleagues developed a universal approach to determine polymer MW through diffusion-ordered NMR spectroscopy (DOSY), overcoming calibration and solvent implications,12−14 while Haddleton and Lester developed a facile strategy for MW online monitoring through DOSY.15,16 However, for polymers with compositional complexity and aggregation behavior in most solvents, solid-state MW analysis is necessary. Costantini and colleagues reported on the MW determination of conjugated polymers, through scanning tunnelling microscopy (STM) and vacuum electrospray deposition (ESD),6 while another powerful example is the work from Matyjaszewski and Sheiko who achieved in-depth analysis of high MW cylindrical brushes using atomic force microscopy (AFM).17−19 Although STM and AFM have provided valuable insights into the understanding of macromolecular characteristics, they are limited to the imaging of conjugated polymers or polymers with very high MW and branching. Within the scope of visual understanding of polymers, electron microscopy (EM) techniques including (cryogenic) transmission electron microscopy, (Cryo-) TEM, and liquid-cell electron microscopy (LC-EM) have revolutionized the field of polymer imaging. Exemplary are the studies from Patterson,1,20−26 Gianneschi,27−33 Sommerdijk,34,35 and de Jonge,36−40 among others. However, the subnano level imaging of nonconjugated synthetic polymers has been largely inaccessible. Apart from their structural complexity, their elemental composition is mostly limited to C, O, H, and N, exhibiting similarity with most TEM support grids. Consequently, the low contrast obtained during conventional TEM does not allow for precise subnano level imaging. Additionally, their light element composition renders them challenging to detect through atomic-level EM methods, such as annular dark field (ADF) scanning transmission electron microscopy (STEM), where contrast depends on the atomic number.1,41−44 Our vision was to overcome those challenges and approach atomic-level analysis of polymers through ADF-STEM, as well as to visualize their MW (Scheme 1), by strategically designing the synthesis of metalloid-rich homopolymers bearing one arsenic (As) atom per monomer unit. For that purpose, free radical polymerization (FRP) and reversible deactivation radical polymerization (i.e., reversible addition–fragmentation chain transfer polymerization, RAFT) were employed to generate polymers with various MW and Đ values. To expand to another polymer family and metal functionality, ferrocenylmethyl methacrylate was used to generate an Fe-rich polymethacrylate. Finally, to render widely used polymers visible on the atomic level, a poly(methyl acrylate) (PMA) was derivatized with ferrocenecarboxylic acid, and its MW and Đ were calculated through imaging. Initially, an As-acrylamide monomer was synthesized according to the literature45,46 and used to generate an As-polyacrylamide (PAsAm) through FRP (PAsAmFRP, Figure S1, SI). A highly dilute (50 μg·mL–1 in 0.1 M NaOH) solution of the purified homopolymer was prepared and placed under vacuum prior to imaging (SI). To gain a first understanding of the As signal, we employed ADF-STEM through a double aberration corrected JEOL ARM200F microscope, operated at 200 kV. At 3 million times magnification (×3M), bright nanoclusters were evident (Figure S2), while at ×8M and ×12M magnification, their structure was elucidated, depicting the randomly coiled polymer chains consisting of As atoms (appearing as bright spots, Figure S2). To enhance sample stability and mitigate contamination, “beam shower” was applied prior to imaging at high magnifications.47 Although the organic content is sensitive and prone to beam damage under the applied conditions,46 the metalloid-rich chains remained intact throughout imaging. An advantage of this approach is that any potential damage to the organic components of the polymers by the electron beam will not affect the results of the MW analysis, as they depend only on the beam-stable metals. Having achieved the detection of single chains and their As atoms, we sought to visualize the MW distribution of the polymers. Three well-defined PAsAms with targeted DPn = 50, 20, and 10 were synthesized through RAFT polymerization (SI, Figures S3–S5), while aqueous-SEC and DOSY NMR were employed to determine the MW of the homopolymers after purification (Table 1).The ADF-STEM of PAsAm50, PAsAm20 (Figure 1b), and PAsAm10 at ×8M magnification revealed polymer chains smaller than in the case of PAsAmFRP (Figures 1a, 2c–f, and S6–S8). Owing to the different average chain length of the imaged polymers, the chain diameter increased with the increase in MW (Figure 1a), while the low Đ polymers exhibited narrow diameter distribution, compared to PAsAmFRP. Importantly, when the As-monomer was imaged under the same EM conditions, only individually scattered single As atoms were detected (Figure S9). Table 1Molecular Weight and Đ Values from SEC, DOSY, and STEM for the Different Metal(loid)-rich HomopolymersPolymerMn,SECaDPn,SECaMw,SECaĐSECaMWDOSYbDPn,DOSYbMWSTEMcDPn,STEMcĐSTEMc,dPAsAm105,700216,3001.102,50092,30081.30PAsAm207,900298,7001.106,900257,500281.20PAsAm5012,1004513,9001.159,6003511,000401.10PAsAmFRP218,000N/A509,0002.30101,700375113,000h417h1.60PFerMMA105,900e21e8,600e1.4e,f3,700g14g3,500131.407,100f25f10,200f aAqueous-SEC, average molecular weight values expressed as MW equivalents relative to PEG/PEO standards,bin D2O/NaOH using an 80 MHz benchtop NMR, calculated through MaDDOSY,15cMWSTEM expressed in g·mol–1, conditions: 200 kV at ×8M magnification (and ×10M for PAsAm10 and PFerMMA10),dcalculated based on the literature,51eCHCl3-SEC, average molecular weight values expressed as MW equivalents to PS orfPMMA standards,gin CDCl3 using an 80 MHz benchtop NMR,haverage of the broad main distribution from 13,000 to 80,000 g·mol–1 (DPn = 48–295) and chains reaching up to 240,000 g·mol–1 (DPn ∼ 885). Molecular Weight and Đ Values from SEC, DOSY, and STEM for the Different Metal(loid)-rich Homopolymers Aqueous-SEC, average molecular weight values expressed as MW equivalents relative to PEG/PEO standards, in D2O/NaOH using an 80 MHz benchtop NMR, calculated through MaDDOSY,15 MWSTEM expressed in g·mol–1, conditions: 200 kV at ×8M magnification (and ×10M for PAsAm10 and PFerMMA10), calculated based on the literature,51 CHCl3-SEC, average molecular weight values expressed as MW equivalents to PS or PMMA standards, in CDCl3 using an 80 MHz benchtop NMR, average of the broad main distribution from 13,000 to 80,000 g·mol–1 (DPn = 48–295) and chains reaching up to 240,000 g·mol–1 (DPn ∼ 885). (a) Histograms showing the distribution of chain diameter for the four polymers and (b) high-resolution ADF-STEM image of PAsAm20 (scale bar: 5 nm). (a) Aqueous-SEC traces for PAsAm10, PAsAm20, PAsAm50 and PAsAmFRP. (b) MWSTEM distributions for the different polymers and ADF-STEM images showing segmented individual polymer chains for (c) PAsAm10, (d) PAsAm20, (e) PAsAm50, and (f) PAsAmFRP (scale bars: 1 nm, images were smoothed post imaging). To determine the polymer MW and Đ, the intensities of single chains were calculated upon subtraction of their background, while the intensity of single As atoms was used as the calibrant, assuming a linear relationship between the integrated ADF intensity of single atoms and very small nanoclusters when kinematic diffraction effects dominate the signals collected by ADF-STEM imaging (Figure S10).48−50 The same process was repeated for each sample individually in the same session. The integrated As atoms’ intensity allowed for determination of the DPn for PAsAm50, PAsAm20, and PAsAm10 through atom counting (Table 1, SI). In other words, the number of As atoms in each chain corresponded to DPn, which was used to calculate the corresponding MWSTEM. The ĐSTEM of the polymers was estimated according to the literature, based on standard deviation (σ), and the relation between Đ and σ (SI).51 For the three well-defined polymers, MWSTEM was comparable with MWSEC, while there was particularly good agreement between STEM and DOSY for PAsAm10 and PAsAm20 (Table 1; Figures 2a–b, S4, S5, S7, and S8). Importantly, in contrast with DOSY, STEM can provide a distribution of MW, representative of the nonidentical chain lengths in a synthetic polymer sample. For PAsAm10, Mn,SEC was significantly higher than MWDOSY because low MWs necessitate better separation for higher accuracy. The ĐSTEM results for PAsAm20 and PAsAm50 exhibited proximity to ĐSEC with both approaches resulting in ĐSTEM ≤ 1.2. To push the limits of our system, we attempted to calculate the MWSTEM of PAsAmFRP. As expected, the STEM results showed a nonsymmetrical distribution of As atoms per chain, with a predominant broad MWSTEM distribution from 13,000 to ∼80,000 g·mol–1, along with the presence of high MW species up to ∼240,000 g·mol–1, with average MWSTEM = 113,000 g·mol–1 (Figure 2b, Table 1). The ĐSTEM was 1.60, and although lower than the corresponding ĐSEC, it illustrated the broad MW distribution of PAsAmFRP. Samples with such high DPn heterogeneity are highly challenging to quantitatively analyze with accuracy from single ADF-STEM images, since the very high MW chains might exhibit similarities with aggregated species; thus, careful interpretation of the images is necessary.44,52 In general, highly pure polymer samples, careful sample preparation (i.e., suitable support grids),53,54 and thorough pretreatment (i.e., vacuum drying, beam shower,47,55SI) are essential requirements, especially when sensitive samples are used.53 To expand the scope of metal functionality and polymer type, we synthesized an Fe-rich polymethacrylate (PFerMMA10) (Figures 3a and S11, SI). As in the case of PAsAms, the MWSTEM (Figure 3b,d,e) was comparable to MWDOSY while lower than Mn,SEC (Table 1, Figure 3b,c). The SEC analysis of PFerMMA10 exhibited distinct deviations when PMMA and PS standards were used (Figure 3c), highlighting the limitations of SEC when samples deviate from the calibrant. The range of MWSTEM (∼1,000–12,000 g·mol–1) with the existence of a second smaller population with MW ∼9,500–12,000 g·mol–1, was depicted in the obtained ĐSTEM = 1.40 (Table 1). Therefore, the calculation of MWSTEM could be successfully achieved both for metalloid- and metal-containing acrylamide and methacrylate homopolymers, while their DPn heterogeneity could be estimated through ĐSTEM calculation. (a) Reaction scheme for the synthesis of PFerMMA10, (b) MWSTEM histogram for PFerMMA10, (c) SEC traces of PFerMMA10, and (d and e) ADF-STEM images for PFerMMA10 (scale bar: 3 nm). Finally, we were interested in applying our approach to widely used C-, H-, and O-containing polymers, without using specially designed monomers. Thus, a PMA20 was synthesized (Figures 4a and S12, SI) and subsequently amidated using 4-amino-1-butanol, according to a literature procedure.56 The amidation of PMA20 to poly(hydroxybutyl acrylamide, PHBAm) was quantitative, with a full shift of the PMA methyl protons as verified by 1H NMR (Figure S14d), and full shift of the 1730 cm–1 peak (C=O, PMA) along with the formation of the 1635 cm–1 (C=O, amide) and 1543 cm–1 (N–H) PHBAm peaks, as verified by FT-IR (Figure S13). The shift toward higher MW was verified by THF-SEC (Figure 4b). The obtained −OH functional polymer was further functionalized through DCC/DMAP coupling with ferrocene (Fc) carboxylic acid (Figure 4a, SI), leading to derivatization of the parent PMA into an Fe-containing polyacrylamide. THF-SEC showed a clear shift toward higher MW (Figures 4b and S14); FT-IR verified the appearance of the C=O band (1700 cm–1) attributed to the Fc-ester (Figure S13), while 1H NMR confirmed the incorporation of the Fc moieties in the polymer (Figure S14d). ADF-STEM (Figure S15) revealed a predominant MWSTEM distribution at 3,000–4,300 g·mol–1 (vs Mn,SEC = 5,000, Figure 4b,c), indicating that on average ∼11 monomer units per chain had been functionalized (vs ∼14 from SEC). Therefore, the achievement of near-atomic level imaging of a commonly used polymer through derivatization has critical potential to serve as a promising strategy to visualize materials that had thus far been unobtainable. To the best of our knowledge, this is the first example of near-atomic level imaging of such a widely used polymer. Owing to the various synthetic tools available, we envisage that the modification of other commonly used polymers (i.e., polystyrene, polyolefins) through different derivatization approaches (i.e., Diels–Alder, click chemistry)57−64 will expand the scope of this approach and establish it as a platform for advanced polymer imaging. (a) Reaction scheme for the derivatization of PMA. (b) SEC traces of the parent PMA, the amidated derivative PHBAm, and the final PBAm-Fc, and (c) MWSTEM histogram of the functional PBAm-Fc. Our work presents the first approach toward atomic level imaging of synthetic polymers and MW determination through atom counting. By combining metal(loid)-containing monomers, different polymerization approaches, and atom counting through ADF-STEM, fundamental polymer characteristics were determined in the subnano scale. Additionally, the subnano level imaging of a widely used polymer (i.e., PMA) was achieved upon derivatization. Our combinatorial approach sets the ground for atomic level analysis of polymer fundamentals that could not be imaged with such precision before and facilitates the profound understanding of their structure–property relationships. Acknowledgments E.L. thanks the Leverhulme Trust, the Institute of Advanced Studies and the Eutopia Science and Innovation Fellowship for funding (ECF-2023-602 and European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 945380), C.B.-K. acknowledges the Australian Research Council for a Laureate Fellowship and QUT’s Centre for Materials Science for continued support; A.M. and P.W. are grateful for funding from Lubrizol and the EPSRC for funding a PhD studentship (Prosperity Partnership, EP/V037943/1); P.W. also thanks the Royal Society and Tata companies (URF\R1\180274). The authors thank the Polymer RTP (UoW) for providing access to SEC and benchtop NMR, the Electron Microscopy RTP (UoW) for providing access to the microscopes, Dr Lijiang Song for access to the MS platform, Dr Ivan Prokes for access to the NMR equipment, and Dr Maria-Nefeli Antonopoulou and Dr Glen Jones for fruitful discussions on RAFT polymerization. Supporting Information Available The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.4c13812. Materials, instrumentation details, methods, synthetic procedures, NMR, SEC, FT-IR, UV–vis spectra, and ADF-STEM images (PDF) Materials, instrumentation details, methods, synthetic procedures, NMR, SEC, FT-IR, UV–vis spectra, and ADF-STEM images (PDF) Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. The authors declare no competing financial interest. Supplementary Material ja4c13812_si_001.pdf References Associated Data Supplementary Materials ja4c13812_si_001.pdf