A fast growing research area, single-chain nanoparticles (SCNP) have demonstrated promise for use in catalysis [1], nanoreactors [2], controlled release [3], chemical sensors [4], enzyme mimetics [57], diagnostics [8] and optics [9]. This breadth of application is attributed to the precise control over the composition, functionality, geometry and topology this technology permits for accessing architectures in the sub-20 nm size regime. These features are programmed into SCNP a priori by means of robust living polymerization techniques including ring-opening metathesis polymerization [10] (ROMP), reversible addition-fragmentation chain-transfer [11] (RAFT), and atom transfer radical polymerization [12] (ATRP). Such techniques afford parent polymers with narrow molecular weight distributions (Đ), targeted molecular weights, tunable functionality and diverse architectures [1317]. These parent chains are then subjected to intra-molecular cross-linking under ultra-dilute conditions, thus folding the chains into architecturally defined nanoparticles with dimensions smaller than the original dimensions of the solvated coil (Fig. 1) [18].

Fig. 1
figure 1

Schematic representation of SCNP synthesis

Although young, this field is developing rapidly; a number of important contributions have appeared in the recent literature. [1, 12, 1921] The Meijer group showed that the molecular weight and backbone rigidity of the parent polymer does not significantly influence the SCNP fabrication induced by supramolecular interactions. Instead, the solvent plays a key role in mediating intra-molecular folding [21]. Meijer and coworkers also expanded the scope of possible architectures by successfully synthesizing SCNP from cylindrical brush block polymers as well as an orthogonally collapsible ABA triblock copolymer [6, 22]. Additionally, this group studied temperature and co-solvent induced changes in the secondary structures of benzene-1,3,5-tricarboxamides (BTA)-based SCNPs, and developed a compartmentalized metal ion SCNP sensor that shows impressive binding affinities [4, 23]. The Barner-Kowollik group demonstrated a photo-induced Diels–Alder ligation to create SCNP in addition to using metal–ligand complexation to induce controlled folding by linking chain ends [1, 24]. Other work has demonstrated that SCNP can be formed from a host of reactions including benzocylcobutene coupling [25], olefin cross-metathesis [26], azide–alkyne click chemistry [19, 2730], thiol-ene click chemistry [31, 32], Bergman cyclization [3335], Curtius rearrangement [36], Glasser–Hay coupling [37], nitrene-mediated C-alkylation (sulfonyl azide) [38], oxidative polymerization through pendant moieties [12], benzosulfone chemistry [39], carbodiimide coupling [40], benzoxazine chemistry [11], amide bond formation [10, 31], cinnamoyl [40] and coumarin [2] photodimerization, guest–host interactions [41], hydrogen bonding [20, 4244], and vulcanization [45].

These methods draw synthetic parallel to nature’s elegant biomacromolecules, albeit in a crude and simplistic fashion [3, 5]. Given that synthetic polymers are always subject to distributions in molecular weights and heterogeneities in microstructure [46, 47], it is important to develop convenient and consistent methods for characterizing SCNP fabrication. We discuss here the analysis of SCNPs using size exclusion chromatography (SEC), specifically the application of multi-detection SEC. We have recently shown that when applied in conjunction with microscopy and dynamic light scattering (DLS), multi-detection SEC is a powerful tool for studying the folding of individual polymer chains into architectural defined nanostructures.

Polymer design and SCNP synthesis

There are a variety of synthetics strategies that can be used to fabricate SCNP [48]. The first methodology successfully employed in our labs involves the addition of an external, bifunctional cross-linker to a chain decorated with complementary reactive handles (Scheme 1) [10]. Polymer chains decorated with cyclic anhydride units were folded by the introduction of p-aminophenyl disulfide as the bifunctional cross-linker. We chose to work with poly(norbornene-exo-anhydride), as the requisite reactive monomer is commercially available and polymerization (or copolymerization with cyclooctene) via ROMP proceeds rapidly with excellent control. In series 1a, the number of intra-chain linkages can be controlled by modulating the amount of bifunctional diamine cross-linker added to a solution of the polyanhydride. In series 1b, the extent of intra-chain cross-links is controlled by changing the feed ratio of anhydride co-monomer relative to cyclooctene, followed by the addition of stoichiometric amounts of diamine. Using a cross-linker with a reversible covalent bond permits the unfolding and refolding of the chain and allows us to better characterize this coil to globule transition [10].

Scheme 1
scheme 1

Synthetic approach to series 1 via addition of an external bifunctional cross-linker

The second route we are using to fabricate SCNPs involves the homocoupling of a built in functional unit [49]. For series 2, foldable methacrylic chains bearing pendant anthracene units were synthesized by RAFT polymerization (Scheme 2). Anthracene is well known for its ability to undergo a photo-induced 4 + 4 cycloaddition when irradiated with 350 nm-centered UV light, resulting in photodimers that can be reversed via irradiation with wavelengths <300 nm or thermally at temperatures >180 °C [50], providing an attractive route to synthesizing thermally responsive and photosensitive nanomaterials. In addition, using light as stimulus for SCNP fabrication offers additional functional group tolerance that other cross-linking strategies lack [49].

Scheme 2
scheme 2

Synthetic approach to series 2 using photo cross-linkable pendant groups

A third method for SCNP fabrication takes advantage of multiple orthogonal chemistries triggered sequentially to fold synthetic chains into particles [31]. Scheme 3 shows a series of poly(oxanorbornene anhydride-co-cyclooctadiene) materials synthesized using ROMP. Analogous to series 1, the polymer structures contained in series 3 feature anhydride groups as reactive handles that can be easily modified with a variety of nucleophiles, simultaneously inducing folding and installing functionality. In this system, aniline tetramer (AT) was used to fold the chains via supramolecular interactions (ππ and hydrogen bonding) between these pendant groups. The chains were then reacted with p-aminoaniline, a bifunctional covalent cross-linker that pulls in the remaining anhydride units and further collapses the chain. Finally, internal olefin units installed within the backbones during ROMP were connected using photoinitiated thiol-ene click chemistry with 1,6-hexanedithiol as a bifunctional cross-linker. This sequential folding approach demonstrates that the application of orthogonal chemistries in a synergistic fashion results in more tightly collapsed structures with the goal of more closely mimicking natural folded polymeric constructs. [31]

Scheme 3
scheme 3

Synthetic approach to series 3 using sequentially activated orthogonal intra-chain interactions

SCNP characterization

Conventional SEC

SEC with conventional calibration is the standard method used to characterize SCNPs formation. Comparing the retention time of the parent polymer chains to the retention time of the folded nanoparticles after intramolecular cross-linking is induced typically reveals shifts to longer retention times. This indicates a decrease in hydrodynamic volume based on the inherent principles of size exclusion chromatography [5153]. This is clearly demonstrated in all of the polymer series discussed here [10, 31, 49].

SEC traces for series 1a below show the effect of adding increasing amounts of bifunctional cross-linker: as more intra-chain linkages are formed the SEC traces shift to longer retention times as expected (Fig. 2a). To ensure this is truly an effect of conformational changes and not due to other factors, such as interactions between the SEC columns and the analyte, we take advantage of the cleavable disulfide linkage that was installed during single-chain nanoparticles formation. Cleaving this linkage after the initial folding reaction takes place allows us to determine that the shifts in the retention time is actually due to the folding of the chain and not other behavior that might affect retention time. This is indeed the case, as shown in Fig. 2b. The reduction of the disulfide bridges via dithiothreitol (DTT) results in the unfolding of the SCNPs, confirmed by SEC traces showing that retention times (and therefore hydrodynamic volume) revert to those of the parent polymer chain, regardless of the amount of cross-linker added. Furthermore, refolding of the unfolded chains via thiol oxidation in the presence of catalytic FeCl3 revealed shifts to longer retention times, indicating the reduction in the solvent dynamic volume (refolding), confirming the utility of this characterization method as well as demonstrating this chemistry as a route to responsive nanostructures (Fig. 2c).

Fig. 2
figure 2

SEC traces for series 1a. a Addition of varying amounts of cross-linker, b after unfolding via disulfide reduction, c reversible folding and unfolding via redox chemistry (Tuten et al. [10]—adapted with permission of The Royal Society of Chemistry)

Analyzing the SEC traces from series 2 shows that the extent of intra-molecular folding or collapse can be controlled very easily using built-in reactive co-monomer. Both reaction time and the amount of co-monomer incorporation can be used to control the extent of intra-chain cross-links when employing this method. Figure 3a–c highlights this behavior. All samples show steady shifts to longer retention times with increasing irradiation times, an indication that folding is occurring with the formation of anthracene photodimers. Anthracene dimerization was confirmed by the disappearance of the characteristic anthracene absorption peak in the UV–Vis spectra (insets in Fig. 3). The effect of reactive co-monomer incorporation is clearly observed here: samples with 10 % (2a-10) anthracene co-monomer showed the least pronounced changes in retention time when compared to the samples with 20 % (2a-20) and 46 % (2a-46) anthracene co-monomer (Fig. 3).

Fig. 3
figure 3

SEC traces and corresponding UV–Vis spectra for series 2 photodimerization studies (Frank et al. [49]—adapted with permission of John Wiley And Sons Inc)

Conventional SEC reveals the same behavior even when multiple orthogonal intra-chain cross-linking chemistries are used. Figure 4 shows that for polymers in series 3, the expected shifts in retention time confirm the efficacy of each folding reaction when engaged sequentially. As expected, the SEC trace of the final nanoparticle shifted to the longest retention time indicating a more tightly wrapped structure consistent with theory and experiment [6].

Fig. 4
figure 4

SEC traces displaying the decrease in hydrodynamic volume with each sequential folding reaction for series 3 (Chao et al. [31]—adapted with permission of The Royal Society of Chemistry)

Triple detection SEC

The application of standard SEC as described above provides useful qualitative information about changes in the solution conformation of polymer chains during SCNP fabrication. Still this method is not without ambiguities, especially with regards to actual changes in molecular weight (or lack thereof) when triggering intra-chain cross-linking. Since relative molecular weight and retention time are linked in traditional SEC, it is impossible to formally characterize SCNP molecular weight with this technique. Furthermore, while shifts in retention time certainly signify changes in hydrodynamic volume it is impossible to quantify these changes with standard SEC measurements alone. Dynamic light scattering (DLS) in conjunction with SEC can be a very useful tool for SCNP characterization, but lacks the chromatographic separation provided by SEC.

To alleviate some of these issues our group implemented SEC with multiple modes of detection for SCNP characterization. The use of multi-angle light scattering (MALS) detection with SEC allows the determination of polymer and SCNP molecular weight independent of retention time [54]. The addition of an inline differential viscometer permits characterization of molecular conformation, including hydrodynamic radius, for each thin chromatographic slice as the sample elutes from the SEC column. With this setup, the absolute molecular weight of the parent polymers can be easily measured and used to calculate the expected SCNP molecular weight. Comparing this value with the experimental absolute SCNP molecular weight can provide useful quantitative information about the chemistry involved in this process. Changes in retention time, which are assumed to correlate with changes in solvated volume, can likewise be readily quantified by this method. Recent work from our lab demonstrating the utility of this technique is highlighted below.

Table 1 summarizes the triple detection data for series 1b. As expected, the MALS detector trace shows a shift to longer retention time after addition of cross-linker similar to what we observed using standard SEC. Examining the absolute molecular weight reveals that although the SCNP curve shifted to a longer retention time the molecular weight actually increased by the amount predicted with the addition of stoichiometric cross-linker. In addition, the viscometric data confirm and quantify the decrease in hydrodynamic radius. It further reveals a decrease in intrinsic viscosity after SCNP formation consistent with theory and experiment [55].

Table 1 SEC data for series 1b

The SEC data for series 2 showed similar behavior where shifts to longer retention times occur after folding as expected. The viscometric data reveals a decrease in both intrinsic viscosity and hydrodynamic radius. The absolute molecular weight, predicted to remain constant, changes only slightly and is consistent with a primarily unimolecular phenomenon (Table 2). This series also displays the trend, although unsurprising, that the extent of reactive co-monomer incorporation directly dictates how much folding will occur.

Table 2 SEC data for series 2 before and after complete photo-induced folding

SEC MALS data from series 3, shown in Table 3, again highlight how useful this technique can be as a quantitative tool. For these polymers, the comparison between the predicted and the measured absolute molecular weights demonstrates that the folding chemistry proceeded as designed. In addition, DLS measurements confirmed that retention time shifts witnessed in SEC measurements (Table 3) are indeed due to folding.

Table 3 Comparison of predicted vs. actual absolute Mw for series 3

Intra-chain versus inter-chain reactions

A crucial bit of qualitative data provided by SEC MALS is the distinction between intra-chain folding and inter-chain coupling. In our experience, MALS is capable of detecting larger nanoaggregates in very minute concentrations while traditional SEC detectors such as UV and RI do not show the characteristic aggregate peak [10, 49]. For series 1, multi-chain aggregates were detectable via MALS while virtually unseen in RI traces (Fig. 5) [10].

Fig. 5
figure 5

MALS and RI detection traces for SEC of 1b highlighting the ability of MALS to detect large, multi-chain aggregates (Tuten et al. [10]—Reproduced by permission of The Royal Society of Chemistry)

Likewise for series 2, the UV detector fails to detect inter-chain aggregates that are readily observed by MALS, evident in the emergence of peak shoulders in MALS traces. These data also suggest that at a certain level of reactive co-monomer incorporation inter-chain coupling becomes unavoidable (Fig. 6) [49]. Thus, triple detection SEC provides a robust method for analyzing polymer folding chemistry and confirming single-chain behavior.

Fig. 6
figure 6

SEC traces for 2c-46 photodimerization studies showing the appearance of multi-chain aggregates. a UV detector traces, b MALS detector traces (Frank et al. [49]—Reproduced by permission of John Wiley And Sons Inc)

Correlating solution measurements with other characterization tools

We highlighted in the previous sections how triple detection SEC can be applied to effectively characterize single-chain folding. Additional secondary characterization techniques can offer invaluable information about SCNP size and shape, including small angle neutron scattering (SANS) [3, 23], thorough DLS measurements [4, 23], atomic force microscopy (AFM) [11, 19, 20, 26, 29, 30, 33, 39, 42, 43, 56, 57], and transmission electron microscopy (TEM) [2, 7, 10, 11, 27, 38, 49]. We found TEM particularly useful as a tool to provide visual evidence for SCNP formation. Figure 7 shows representative images from series 1 (1b) and series 2 (2a-10). In each case, the images reveal nanostructures with dimensions on the same order of magnitude as numbers obtained from solution measurements. Interestingly, series 2 SCNP adopt an oblong ellipsoid geometry, rather than a spherical conformation. We are still investigating the cause of this behavior, which is consistent with form factor measurements made by others [3, 23].

Fig. 7
figure 7

TEM images of SCNP obtained from 1b (a) and 2a-10 (b)


In summary, conventionally calibrated SEC is a satisfactory method to characterize SCNP via interpretation of comparative shifts in retention time. Evolving this tool by marrying SEC with MALS and viscometric detection affords a precise and reliable method for the qualitative and quantitative study of SCNP fabrication. This methodology facilitates the detailed investigation of structure property relationships using absolute molecular weight, intrinsic viscosity, and hydrodynamic radius of SCNP systems. During SCNP formation, the decrease in both the intrinsic viscosity and hydrodynamic radius is observed with, or in the absence of, concurrent changes in the absolute molecular weight depending on the folding technique employed. Triple detection SEC in conjunction with additional characterization tools including TEM and DLS provides a variety of useful information for studying the fascinating and rapidly expanding research area of single-chain folding processes and single-chain nanoparticles.