NMR Study of the Dissolution of Core-Crystalline Micelles

9 hours ago - Short fragments of the core-crystalline micelles formed by a sample of poly(ferrocenyldimethylsilane)-block-poly(isoprene) (PFS-b-PI) bl...
3 downloads 0 Views 2MB Size
Download PDF
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

NMR Study of the Dissolution of Core-Crystalline Micelles Qing Yu,† Dmitry Pichugin,† Menandro Cruz,† Gerald Guerin,*,† Ian Manners,‡ and Mitchell A. Winnik*,† †

Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON M5S 1H6, Canada School of Chemistry, University of Bristol, Bristol, U.K. BS8 1TS

‡

S Supporting Information *

ABSTRACT: Short fragments of the core-crystalline micelles formed by a sample of poly(ferrocenyldimethylsilane)-blockpoly(isoprene) (PFS-b-PI) block copolymer (BCP) underwent self-seeding in decane when heated above its dissolution temperature. Variable temperature (VT) 1H NMR and diffusion-ordered pulsed-gradient spin−echo (DOSY) NMR were used to monitor the behavior of micelles that dissolved as a function of increasing temperature. We examined a sample of micelle fragments of PFS65-b-PI637 characterized by Ln = 39 nm and Lw/Ln = 1.13. The PI corona had high mobility and gave a 1 H NMR signal in both micellar and unimer forms. In contrast, the PFS component could only be detected for the dissolved unimer. We found from 1H NMR that essentially all the BCP molecules were incorporated into the micelles at temperatures up to and including 50 °C, at the limit of NMR detection. Both PFS and PI resonances could be detected between 70 and 100 °C, and the integration ratio of the PFS-to-PI peaks increased with temperature. DOSY NMR measured the self-diffusion coefficients (Ds) of the micelle fragments and unimer at these temperatures. The hydrodynamic radii (Rh) for these species were calculated from Ds using the Stokes−Einstein equation. The PFS signals gave Rh values in the range of 5−6 nm at temperatures between 80 and 100 °C, consistent with unimer diffusion. PI signals were fitted by an exponential decay at 25 °C with Rh = 38 nm characteristic of the micelle fragments and at 90, 95, and 100 °C with Rh ≈ 6 nm, corresponding to unimer. At intermediate temperatures (70−85 °C), PI signals were fitted to a sum of two exponential terms, consistent with a fast diffusing species and a slow diffusing species. Interestingly, we noticed that the size of the micelle fragments at elevated temperatures (80 and 85 °C) was sensitive to sample history; samples heated directly to the elevated temperatures were found to be shorter than those heated stepwise.

■

INTRODUCTION While studies of how polymer molecules crystallize in solution go back more than half a century, there have been many fewer studies about how polymer crystals dissolve. Polymer crystals are susceptible to “self-seeding”,1,2 and these experiments provide important information about how polymer crystals melt or dissolve.3,4 In a self-seeding experiment, a suspension of polymer crystals is heated to a temperature at which the crystals seem to disappear. Upon cooling, one obtains a suspension of uniform single crystals, whose size depends sensitively upon the annealing temperature. A similar phenomenon occurs in the bulk state, where samples heated to the point of melt clarity form uniform single crystals upon cooling.5 The common explanation for self-seeding is that when polymers crystallize, kinetic factors lead to a distribution of regions of crystal perfection. Upon heating, the less crystalline regions dissolve or melt first, followed by the somewhat more crystalline regions. There is a temperature region prior to the dissolution or melting of the entire sample, where microscopic crystallites persist. These crystallites serve as seeds to nucleate crystallization upon cooling, so that all of the dissolved or © XXXX American Chemical Society

melted polymer molecules (unimer) grow onto the surviving seeds. Reiter and co-workers5 have shown that in the selfseeding range of annealing temperatures the fraction of surviving seeds appears to decrease exponentially with an increase in annealing temperature. They also showed that this fraction of surviving seeds persists as the annealing time at a given temperature is increased, suggesting that the surviving seeds are at thermodynamic equilibrium with the surrounding medium. Core-crystalline block copolymer (BCP) micelles are low curvature structures, e.g., 2D platelets and 1D elongated fibers that exhibit many of the properties of polymer crystals. Many examples have been studied.3,6−16 Some compositions can undergo seeded epitaxial growth, in which 1D or 2D micelle fragments suspended in a selective solvent are treated with a solution of BCP molecules (unimer) dissolved in a common good solvent for both blocks.17−19 In the case of fiber-forming Received: January 15, 2018 Revised: April 2, 2018

A

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Scheme 1. Structure of PFS65-b-PI637a

a

The microstructure of the PI block (x = 0.49, y = 0.37, z = 0.14) was determined by 1H NMR for a solution of the BCP in benzene-d6.

b-PI BCP micelles in decane are sufficiently mobile to yield sharp peaks in 1H NMR spectra. Diffusion-ordered spectroscopy (DOSY) NMR is a powerful tool for the analysis of mixtures.34 It is based on pulsed field gradient-spin echo (PFG-SE) NMR methodology that itself is commonly used to measure self-diffusion coefficients (Ds) in a variety of different systems such as colloids and polymers.35 When one encounters a mixture of species of different diffusivities, the overlapping signals from these species can often be distinguished by PFG-SE NMR. For example, Muller et al.36 demonstrated the use of DOSY NMR to investigate polymer mixtures and molecular weight distribution. Dobson and co-workers examined solutions of micrometer long amyloid fibrils using PFG-SE NMR and showed that rotational diffusion dominated the relaxation measurements. They showed how these data could be used to calculate the distribution of rod lengths for this sample.33 Bertin et al. performed a compositional analysis on mixture of homopolymer and block copolymer, demonstrating that DOSY can be an effective method to monitor the polymerization and purification procedures.37 DOSY can also be employed to study selfassembly processes38 and to determine the critical micelle concentration (CMC).39 In previous studies reported by our group,31 we used PFG NMR to examine the ligand exchange process in which a linear polymer, poly(2-(N,N-dimethylamino)ethyl methacrylate), displaced TOPO from the surface of TOPO-coated CdSe quantum dots. In this paper, we examine the behavior of PFS-b-PI micelles in decane subjected to self-seeding conditions. We test the effectiveness of variable temperature 1H NMR (VT-NMR) and DOSY NMR measurements to monitor micelle fragment samples of a PFS-b-PI sample that we have examined previously. Thus, we have a good understanding of the micelles formed by this BCP under a variety of different sample preparation conditions. The structure of this PFS65-b-PI637 sample is presented in Scheme 1. At room temperature, all of the polymer molecules are incorporated into the micelles. In 1H NMR spectra, after suppression of the decane peaks, only signals from the PI corona chains can be detected, and the DOSY signal gives a Ds value consistent with the diffusion of short micelles. At 100 °C, all the polymer molecules dissolve, and signals from the ferrocene protons of the PFS and those of the PI protons are prominent. Here the DOSY signal gives a Ds value consistent with the diffusion of molecularly dissolved unimer. These experiments allow us to monitor directly the fraction of polymer in the micelles that has dissolved at each annealing temperature and also provide insights into changes in size of the surviving micelle fragments. Over a small range of temperatures from 70 to 85 °C, it is possible to detect simultaneously the diffusion of the micelles and unimer, but the most precise values of the fraction of polymers in the micelles

BCPs, typically characterized by a crystalline core-forming block shorter than the soluble corona-forming block, the added unimer deposits on the ends of the growing micelles. The best examples are those in which poly(ferrocenyldimethylsilane) (PFS) is the core-forming block. Here the micelles grow at a common rate, leading in the end to micelles with a very narrow length distribution. This initiated growth process has many features in common with the living polymerization of molecular monomers, and we refer to this process as living crystallizationdriven self-assembly (CDSA).20,21 Some core-crystalline rod-like BCP micelles undergo selfseeding when a solution of the micelle fragments in a selective solvent is heated, forming longer micelles of uniform length upon cooling.15,22 Here, tthe most intensively studied system involves PFS BCPs.23−25 As in the case of homopolymer crystals, the length of the micelles formed upon cooling increases sensitively with small increases in the annealing temperature. Since the length and the number of micelles formed after annealing remain constant over a wide range of heating times, ranging from 30 min to 24 h,26−28 we inferred that the surviving seeds in the hot solution are at thermodynamic equilibrium. In a recent publication, we explained the strong decrease in the fraction of surviving seeds upon heating in terms of a Gaussian distribution of dissolution temperatures for the micelle fragments.29 We also understand that the degree of crystallinity of a polymer sample depends on many factors such as the crystal growth rate, the age of the crystal, and its thermal history.30 Thus, these factors should have an important effect on the fraction of crystallites that survive heating in a solvent to a given temperature. Most of our ideas about the mechanism of self-seeding are based on observations of the size of crystals obtained following a heating and cooling cycle. This model anticipates that during the heating cycle the solution consists of a mixture of molecularly dissolved polymer (unimer) and intact seed crystallites. We are particularly interested in developing a deeper understanding of how self-seeding operates on solutions of core-crystalline BCP micelles. Here at elevated temperatures, the solution should consist of a mixture of unimer and colloidally stable micelles. We are unaware of previous experiments that attempt to monitor directly the fraction of polymer that has dissolved for samples annealed at different temperatures. Here we explore the idea that variable temperature 1H NMR and DOSY NMR measurements can provide important new insights into the polymer dissolution process. Colloidal objects of nanometer dimensions are normally too large for solution-state NMR measurements.31,32 However, Dobson and co-workers33 cite some examples of large biomolecular systems in which the flexible regions have been characterized by solution state NMR spectroscopy. As we will see below, the polyisoprene (PI) corona chains of rodlike PFSB

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

time. A value of τ2 = 5 ms was used in all measurements, and the value of τ1 changed accordingly with diffusion delay. For a mixture that consists of two independently diffusing species with overlapping resonances, such as the PFS-b-PI unimer and PFS− PI micelle in a high-temperature decane solution, the diffusion decay is the sum of two exponential terms

that has dissolved come from the integration of the PFS and PI signals in the 1H NMR spectra.

■

EXPERIMENTAL SECTION

Purification of Block Copolymer PFS65-b-PI637. The PFS65-bPI637 sample is the same polymer originally described in ref 40. It was originally reported to have a composition PFS53-b-PI637. This composition was determined in two steps. First, an aliquot of the PI block was taken from the synthesis and analyzed by gel permeation chromatography. Its molecular weight was reported as Mn = 43 300 g mol−1, giving a number-average degree of polymerization (DPn) of 637.40 The corresponding length of the PFS block was determined by comparing integration values of the PFS and PI signals in the 1H NMR. At the beginning of the experiments reported here, DOSY measurements on the micelles formed by this BCP showed the presence of a small amount of PI homopolymer in the sample. This impurity was removed by fractional precipitation. The polymer was dissolved in minimum amount of THF and sonicated briefly to promote complete dissolution. The polymer solution was then transferred into hexane in a vial and allow to stand for an hour. This solution was then diluted with more hexane, transferred into 4 7 mL vials, and allowed to stand overnight. This process led to the formation of long fiberlike micelles. The resulting solution was centrifuged three times at 20 000 rpm for 20 min, followed by removal of the supernatant. Here, comparison of the 1H NMR integration values of the PFS and PI signals led to a composition of PFS65-b-PI637 for the purified polymer. Preparation of PFS65-b-PI637 Micelle Seed Fragments. PFS65b-PI637 (30 mg) was added to decane (5 mL) and then heated at 100 °C for 30 min to dissolve the sample. Upon slow cooling to room temperature (23 °C), long micelles (>1 μm) of uniform width were obtained. This micelle solution was then subjected to sonication in a 60 W sonication bath for 30 min (3 × 10 min intervals) at room temperature. Short micelle fragments were obtained with number- and weight-average lengths of Ln = 39 nm and Lw = 44 nm, respectively. NMR Spectroscopy Experiments. All NMR experiments were performed on an Agilent DD2 600 spectrometer. The concentration of the PFS65-b-PI637 unimer solution in benzene-d6 and micelle fragment solutions in decane were 6 mg/mL. All experiments performed in decane were run with a capillary insert containing DMSO-d6 as a deuterium lock. Variable Temperature (VT) 1H NMR. Proton NMR spectra of PFS65-b-PI637 micelle fragment solutions in decane were obtained at various temperatures that ranged from 25 to 100 °C. Two sets of VT 1 H NMR experiments were performed. In the first set of experiments (VT.01 1H NMR), one sample was heated stepwise from 25 to 100 °C, and two spectra were taken at each temperature examined: the first was recorded when the sample just reached the desired temperature, and the second one was recorded after equilibrating for 15 min. In the second set of experiments (VT.02 1H NMR), multiple samples were used to minimize issues related to sample history. One freshly prepared sample was used at each temperature, and NMR spectra were taken after equilibrating for 15 min at the target temperature. The acquisition time for each proton NMR spectrum is around 5 min. DOSY NMR. DOSY experiments employed the pulse field gradient spin echo NMR technique using a pulse sequence [Pulse Field Gradient Double STimulated Echo (Dpfgdste)] proposed by Jerschow and Muller36 to compensate for complications caused by convection in liquid samples at elevated temperatures. This pulse sequence is presented in Figure S1 of the Supporting Information. The diffusion coefficient of a single diffusing species41 in a sample can be calculated from the expression

I / I0 =

∑ fi exp(−2τ2/T2i) exp(−τ1/T1i) × exp[− (γgδ)2 (Δ − δ /3)Di]

(2)

where f i is the fractional intensity contributed by species i (∑f i = 1) with self-diffusion coefficient Di and longitudinal and transverse relaxation times of T1i and T2i, respectively. To extract diffusion coefficients from the experimental data, the resonance intensities I were plotted versus [(γgδ)2(Δ − δ/3)], from which D values could be calculated either from eq 3 for single species or from eq 4 for two species.

y = B + A exp(− xD)

(3)

y = B + A uni exp(− xDuni) + A mic exp(− xDmic)

(4)

where Duni and Dmic are the diffusion coefficients of the unimer and the micelles, respectively. The fractions of each independently diffusing species, f uni and f mic in the case of PFS−PI in high-temperature decane solution, can be estimated from the combination of eqs 5 and 6

A uni /A mic = [funi exp(− 2τ2/T2uni) exp(− τ1/T1uni)] /[fmic exp(− 2τ2/T2mic) exp(− τ1/T1mic)] funi + fmic = 1

(5) (6)

where Auni and Amic are obtained from exponential fitting of experimental data, and the relaxation times T1 and T2 for unimer and micelles should be determined at each temperature. The details of the experimental setup are shown in Tables S1 and S2 for measuring diffusion coefficients of solvent decane and sample PFS65-b-PI637 (6 mg/mL), respectively. The diffusion coefficients of solvent decane at different temperatures were employed to calculate the solution viscosity at each temperature, so that the hydrodynamic radii of the PFS65-b-PI637 micelles seeds could be calculated from measured D values using the Stokes−Einstein equation

R h = kBT /(6πηD)

(7)

VT-DOSY spectra of PFS65-b-PI637 micelle fragment solutions in decane were obtained at various temperatures ranging from 25 to 100 °C. Two sets of VT-DOSY measurements were carried out using micelle solutions with different sample histories. In the first set of experiments (denoted DOSY.01), only one micelle solution (6 mg/ mL in decane) was used. This solution was heated to elevated temperatures stepwise from 25 to 100 °C, with DOSY measurements recorded at discrete temperature intervals. In the second set of experiments (DOSY.02), a series of identical samples were prepared. Here a sample was used for a DOSY measurement at only one temperature, so that results for experiments at different temperatures refer to separate samples. VT-DOSY measurements on neat decane were carried out as a means of calculating the viscosity of decane at each measurement temperature. Details are provided in the Supporting Information, including Figure S2 and Table S3.

■

RESULTS AND DISCUSSION Polymer Purification. The PFS65-b-PI637 polymer used here is the same sample denoted “PFS53-b-PI637” reported in several previous publications.25,40,42,43 The PI block was synthesized first and was found to have DPn ≈ 637 by GPC. The block ratio was determined by comparing 1H NMR integration values of the PFS and PI signals for samples in a ccommon good solvent, benzene-d6 (Figure S3a). We found

I = I0 exp(− 2τ2/T2) exp(− τ1/T1) exp[− (γgδ)2 (Δ − δ /3)Ds] (1) where Ds is the self-diffusion coefficient and T1 and T2 are the longitudinal and transverse relaxation times, respectively. δ is the diffusion gradient length, γ is the particular magnetogyric ratio, g is the magnitude of the gradient pulse, and Δ = τ2 + τ1 is the diffusion delay C

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules that the micelle solution in decane was not pure by DOSY NMR (Figure S3b), as the stimulated echo intensity attenuation as a function of the gradient strength fit much better to the two-exponential function (red solid line) than a single exponential (blue solid line). This was an indication of the presence of two species in the sample solution with different diffusion rates. The slow diffusing species was characterized by Ds = 6.1 × 10−8 cm2/s, and its hydrodynamic radius (Rh) was calculated to be 40 nm. The fast diffusing species, with Ds = 7.3 × 10−7 cm2/s (Rh = 3.5 nm), was consistent with the presence of PI637 homopolymer.44 For the experiments reported previously with this BCP,25,40,42,43 the presence of PI homopolymer does not affect the interpretation of the self-assembly experiments. Here, however, the homopolymer would interfere with the interpretation of the NMR measurements. Thus, the impurity was removed by fractional precipitation. Reanalysis of the polymer by 1H NMR (Figure S3c) gave a higher ratio of PFS to PI, and we assign the new composition as PFS65-b-PI637. In Figure S3d, the DOSY NMR of micelle fragments in decane of the purified sample showed that only one species was present, with Rh = 38 nm. In Figure S4 we compare GPC traces in THF of the original “PFS53-b-PI637” sample and the sample denoted PFS65-b-PI637 after purification. Both peaks are monomodal, but after purification, there was a decrease in Đ from 1.07 to 1.04. The microstructure of the PI block was unchanged before and after the purification. From 1H NMR spectra run in benzene-d6, we found the composition to be 49% 3,4-repeats, 37% 1,2repeats (y), and 14% 1,4-repeats. Diffusive Motion for Polymer Samples at Room Temperature. Sample preparation began with the formation of long fiberlike micelles using the purified sample of PFS65-bPI637. This polymer in decane (c = 6.0 mg/mL) was heated at 100 °C for 30 min and then allowed to cool slowly to room temperature. After aging at room temperature for a day the sample was subjected to mild sonication to produce micelles fragments, which we sometimes refer to as seeds. These micelle fragments are relatively uniform in size as shown in the TEM image in Figure 1, characterized by Ln = 39 nm and Lw = 44 nm. As mentioned above, the 1H NMR of PFS65-b-PI637 in benzene-d6 (Figure S3c), a common good solvent, showed peaks for both the PI and PFS components. In contrast, at 25 °C, the 1H NMR of PFS65-b-PI637 micelle fragments in decane (Figure 1c) gave peaks at 4.6 and 5.7 ppm associated with the PI corona, but no signals from PFS. We know that under these conditions all of the PFS chains are confined to the micelle core. Not surprisingly, the mobility of these chains is too restricted to contribute to the high-resolution NMR spectrum in Figure 1c. We examined the diffusive motion of the PFS65-b-PI637 unimer in benzene-d6 and the PFS65-b-PI637 micelle fragments in decane at room temperature (23 °C) by DOSY NMR. A typical PFG-SE 1H NMR spectral data set showing the intensity decay for the sample in benzene-d6 is shown in Figure 2. From bottom to top, the intensities of the peaks decay with increasing magnetic field gradient strength. The choices of the lowest and highest gradients were determined such that the peak intensity decayed to 10% of its original value in the last increment. A total of 31 increments were used in each DOSY experiment to ensure that there would be enough valid data points to calculate diffusion coefficients. For this purified PFS65-b-PI637 sample, the stimulated echo intensity decays fit well to a single-exponential

Figure 1. (a) TEM image (scale bar = 100 nm) and (b) histogram of the length distribution of PFS65-b-PI637 micelle fragments prepared in decane. From the histogram, we calculate Ln = 39 nm and Lw = 44 nm. (c) Proton NMR spectrum of PFS65-b-PI637 micelle seeds in decane (6 mg/mL) at 25 °C. The small shift in the peak positions of the PI protons here compared to those in Figure S3 is due to the difference in solvent.

Figure 2. Typical PFG 1H NMR spectra for PFS65-b-PI637 polymers (6 mg/mL) in benzene-d6 at 25 °C as a function of magnetic gradient amplitude (g) from low (bottom) to high (top).

D

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 3. Stimulated echo intensity attenuation intensity of the PFS65-b-PI637 NMR signal at (a) 4.5−5.4 ppm (PI, top) and (b) 4.1−4.3 ppm (PFS, bottom) in C6D6 at room temperature as a function of the gradient strength g. From the decay curves obtained at room temperature, diffusion coefficients of 5.3 × 10−7 and 4.9 × 10−7 cm2/s were calculated from the PI and PFS signals, respectively. From the decay curves obtained for the micelle fragments in decane at 85 °C, the results from DOSY.01 showed that the intensity decay of (c) the PI signal fits a two-exponential function, and (d) the PFS signal fits a mono-exponential function. The red solid lines are the curves of the best fit. Decays were monitored to 10% of the initial value.

g/mol for these micelle fragments, a radius of gyration Rg = 41 nm, and a mass per unit length (linear aggregation number) of Nagg,L = 2.2 chains/nm. The value of Lw obtained by fitting the SLS data to a form factor analysis for a thick rigid rod (118 nm) was significantly larger that the corresponding length (70 nm) obtained by TEM. This difference is to be expected because the PI chains extend into the solution beyond the ends of the core and contribute to the scattering signal. In the dry state on the TEM grids, PI does not provide sufficient contrast to be detectable in the TEM images. From the rigid rod model used to interpret the scattering results, we can calculate a rotational diffusion coefficient Dr ≈ 1700 s−1. Multiangle dynamic light scattering (DLS) measurements on this sample gave apparent diffusion coefficients that were very similar at all scattering angles. The value of the hydrodynamic radius calculated from the Stokes−Einstein equation (Rh = 40 nm) is approximately half the value of Lw determined by TEM. The revised composition of this BCP, as described above, can in principle affect the interpretation of the SLS data. We used the Dale−Gladstone relationship45 to calculate values of dn/dc, and this calculation weights the dn/dc values of the two blocks in terms of their weight fraction in the polymer. The data analysis in ref 43 used a value of dn/dc = 0.141 mL/g for PFS53b-PI637 in decane. We now calculate a value of dn/dc = 0.145 mL/g for PFS65-b-PI637 in decane. This correction changes the value of (dn/dc)2 by about 6% and thus makes only small changes in the values of Mw and Nagg,L determined from the SLS measurements. The micelle fragments examined here were characterized by a TEM length of Lw = 44 nm. Both sets of samples are likely to be characterized by similar values of the linear aggregation number (Nagg,L = 2.2 chains/nm). Because of their small size

term (eq 1) for the PI signal intensities measured at 4.5−5.4 ppm and for the PFS signal intensities measured at 4.1−4.3 ppm (Figure 3a,b). Data analysis of the PI signal led to a Ds = 5.3 × 10−7 cm2/s, whereas analysis of the PFS signal gave Ds = 4.9 × 10−7 cm2/s. As the two blocks are covalently connected, the block copolymer must diffuse as a single entity; thus, the 7% difference between the two fitted Ds values is an indication of the accuracy of the measurement. From the Stokes−Einstein equation, the average hydrodynamic radius of the unimer was calculated to be 7.1 nm. DOSY NMR measurements at 25 °C were used to examine the diffusive motion of the PFS65-b-PI637 micelle seed sample. As shown in Figure S3d, the intensity decay also fits well to a single-exponential term, characterized by Ds = 6.7 × 10−8 cm2/ s, corresponding to Rh = 38 nm. This value is close to the number-average length of these micelle fragments (39 nm) determined by TEM. These DOSY NMR experiments show that the PFS65-b-PI637 unimer and the micelle seeds formed by this block copolymer have very different diffusion rates that can readily be distinguished. This finding is important as it suggests that DOSY could be a promising tool to study the micelle dissolution at elevated temperatures, where both unimer and micelles will be present at the same time. Further information about the structure of these micelle fragments is available from previous work in our laboratory, reported in ref 43, in which we prepared a micelle fragment solution in decane of this BCP using the same protocol described above. The length of the micelle fragments depends sensitively on the sonication conditions, and the sample examined previously was characterized by TEM to have Ln = 50 nm and Lw = 70 nm. A Zimm analysis of the multiangle static light scattering (SLS) data gave values of Mw = 1.5 × 107 E

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 4. (a) 1H NMR spectra of PFS65-b-PI637 micelle fragments in decane at various temperatures from 50 to 100 °C. These spectra were taken from samples immediately after they reached the desired temperature. The percentage of PFS65-b-PI637 present as dissolved unimers was calculated. This percentage was calculated from the ratio of PFS to PI signal integrations, assuming that the intensity of the PI signal remains constant at each temperature. In order to test whether kinetic factors affected the fraction of BCP that dissolved at each temperature, NMR spectra (b) were measured at 0 and 15 min after the probe reached the desired temperature. Note that at various elevated temperatures annealing for 15 min led to less unimer present (denoted by vertical arrows) than at t = 0. To minimize the effect of sample history, (c) multiple identical samples were heated for 15 min at each temperature before taking the proton NMR. These data points (△) are denoted “heated directly” in the plot. The data points (○) are for samples heated stepwise and annealed 15 min at each temperature.

compared to that of the ferrocene signal did not change between 25 °C, wwhere all of the polymer is incorporated into micelle fragments, and 100 °C, where all the polymer is present as dissolved unimer. The DOSY experiments described below provide a test of the second assumption. The DOSY signal of the PFS component always appears as a single-exponential decay curve with a Ds value characteristic of the unimer. There is no long tail in the DOSY decay characteristic of a slowly diffusing PFS component. In the first set of experiments, a single solution of PFS65-bPI637 micelle fragments (Ln = 39 nm, Lw = 44 nm) in decane (6 mg/mL) was heated stepwise through a series of temperatures. The 1H NMR spectrum was recorded at two different time points at each temperature. One spectrum was taken as soon as the instrument indicated that the temperature had been reached. A second spectrum was taken 15 min later. Then the temperature of the sample was increased to the next step. In the second set of experiments, multiple identical micelle fragment solutions were prepared, and one sample was used for each temperature. These samples were equilibrated for 15 min

and morphology, rotational diffusion should make very little contribution to the overall diffusion measured by DOSY NMR.46 They are small enough that their diffusion can be modeled as that of a hard sphere. The value Rh = 38 nm calculated with the Stokes−Einstein equation is similar in size to the TEM length. Monitoring Micelles Dissolution at Elevated Temperatures. Variable Temperature NMR Measurements. The PFS65-b-PI637 micelle fragments in decane dissolve at high temperature and appear as unimers in the hot solution. To determine the fraction of unimer in the hot solution, we needed to make two assumptions. First, we assumed that the intensity of the PI signal is not affected by the change in temperature and that all of the units of PI corona chains are mobile, including those located near the surface of the core at room temperature. Second, we assumed that the PFS signal observed by 1H NMR is only due to dissolved PFS unimer, not amorphous but possibly mobile PFS in the core. We tested the first assumption by running 1H NMR spectra in the presence of ferrocene as an internal standard. As shown in Figure S8, the PI integration F

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Annealing at elevated temperatures reduces the fraction of BCP that dissolves as unimer at each temperature. Information about how the micelle fragments dissolve requires a different kind of experiment. TEM measurements of the micelles obtained after the solution has cooled to room temperature (for example, Figures S5−S7) show that annealing at higher temperatures leads to a smaller number of longer micelles. This is a typical self-seeding result23,25 and indicates that some of the micelle fragments dissolve completely when the temperature of the solutions is increased. DOSY NMR Measurements. The experiments described in the previous section allowed us to quantify the fraction of polymer in the core-crystalline micelle fragments that dissolved when their solutions were heated to typical self-seeding temperatures. Here we explore whether DOSY measurements at elevated temperatures can complement these experiments by determining the corresponding diffusion coefficients of the mobile species in solution. Initial high-temperature experiments, carried out with a traditional PFG-SE pulse sequence, gave results that appeared to be distorted because of convection in the sample tube. Therefore, we turned to the Pulsed Field Gradient Double STimulated Echo (Dpfgdste) pulse sequence developed by Jerschow and Muller36 to compensate for these complications. In the first set of experiments (DOSY.01), a single sample of PFS65-b-PI637 micelle fragments (6.0 mg/mL in decane) was heated stepwise to various temperatures from 25 to 100 °C. Data acquisition was relatively slow, and each DOSY measurement at each temperature required 2 h to complete. Intensity decay profiles were measured for both the PI signal at 4.4−5.1 ppm and the PFS signal at 3.8−4.2 ppm. These experimental data were then fitted using a monoexponential function eq 1. When the quality of this fit was poor, suggesting the presence of two independent diffusing species (unimer and micelle), the data were then fitted to a sum of two exponential terms, eq 2. In Table 1, we present the values of the diffusion coefficients Ds determined in this way, along with Rh values calculated using the appropriate viscosity of decane at that temperature. At 25 °C, we found only a single (slow) diffusing species, the intact micelle fragment, with Rh = 38 nm. At the two highest temperatures (95 and 100 °C), we found only a single (fast) diffusing species with Rh ≈ 5.5 nm. We found similar values of Ds and Rh from measurements of both the PI signals and the PFS signals in the NMR. On the basis of measurements at 25 °C in benzene-d6, where we found Rh ≈ 7 nm, we can assign the fast diffusing species to the PFS65-b-PI637 unimer. At 80 and 85 °C, the decay of the PFS signal followed a single-exponential profile, which we associate with unimer. At 70 and 75 °C, the PFS signal was too weak to obtain a meaningful fit. In contrast, fitting the PI decays at 70, 75, 80, and 85 °C required two exponential terms, indicating that micelle fragments and unimers were both present. Figure 3c,d shows that at 85 °C the decay of the PI signal fits better to a two-exponential function than a monoexponential function. In contrast, at this temperature, the PFS signal is well fitted by a monoexponential function. The fast decays determined from both the PFS and PI signals were consistent with Rh ≈ 5−7 nm. The slow decays from the PI signal gave a value of Rh = 45 nm at 70 °C, decreasing to ca. 36 nm at 80 and 85 °C. One of the problems with the data collected for the DOSY.01 experiments is that the samples were annealed at successively increasing temperatures over many hours. In order

at the corresponding temperatures before measuring the proton NMR. Spectra were taken with a pulse sequence that suppressed signals from the solvent decane.47 A series of spectra, covering the range of 3.5−6.5 ppm for temperatures between 50 and 100 °C, are presented in Figure 4a. The peaks due to PI protons do not change in shape significantly over this range of temperatures. The spectra were normalized so that the integration of PI peaks from 4.4 to 5.1 ppm remain constant. Signals from PFS block started to appear at 70 °C, and the intensity became stronger with increasing temperature. At 100 °C, the ratio of intensities for the PFS peaks to the PI peaks was essentially the same as detected in the room temperature spectrum in benzene-d6. This result indicates that all of the BCP in the micelles had dissolved at this temperature. In Figure 4b, we plot the mole fraction (percentage) of the BCP that had dissolved at each temperature. This fraction was calculated by comparing the integrated intensities of the two PFS peaks at ca. 4 ppm to the integrated intensities of the PI peaks at 4.4−5.1 ppm. This analysis assumes that the signal from the mobile PI corona chains remains constant at all temperatures. Note that in this experiment we used a single micelle fragment sample. It was taken stepwise through each of the annealing temperatures and, as a consequence, was subjected to long-term thermal annealing by the time the highest sample temperatures were examined. In order to test whether there was also a kinetic component to the dissolution of BCP molecules from the micelle fragments, we carried out a second NMR measurement at each temperature. This measurement was made 15 min after the initial measurement. Figure 4b shows that at elevated temperatures there was a small decrease in the amount of dissolved unimer in the solution in the second measurement. It is known that semicrystalline polymers remodel when annealed at temperatures below their melting point. For typical platelet homopolymer crystals, the platelets become thicker and the degree of crystallinity increases.48,49 Less is known about core-crystalline micelles, but we recently examined one sample of PFS65-b-PI637 micelle fragments annealed in decane at 75 °C.43 Here we found an increase in the number of BCP molecules per nanometer of micelle length, and electron tomography measurements showed that the micelle core became wider and more rectangular. This result suggests that at 75 °C some of the BCP molecules that initially dissolve recrystallize onto the micelle fragment in solution. We have also noted that annealing or even prolonged aging at room temperature decreases the fraction of the micelles that dissolves at a given temperature. These results are consistent with an increase in crystal perfection upon aging or annealing. There are some very interesting results in Figure 4b,c. The data points in Figure 4b presented as diamonds refer to a sample heated stepwise to each temperature, with the spectrum taken when the temperature was reached. The circles refer to the same sample rerun after aging for 15 min after the end of the first spectrum. At temperatures above 75 °C, we see a decrease in the PFS signal intensity, suggesting that some of the PFS has been removed from solution, likely by deposition on surviving micelle fragments. Another way of observing a sample history effect is to compare the data reported as circles with data obtained from samples heated directly to the measurement temperature. These data are reported as triangles in Figure 4c. Here we see that a substantially larger fraction of the micelles dissolve upon heating directly to the measurement temperature. G

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

close to those determined from the PFS signal. While the VTNMR experiments suggested that not all micelle fragments dissolved below 100 °C, the fragments’ contribution to the PI signal decay in the DOSY experiment was not significant. At 70, 75, 80, and 85 °C, the fits to the PI decay required two exponential terms. The fast PI decays gave Rh values somewhat smaller than those found from the PFS decays. These differences are likely a consequence of fitting three independent parameters (eq 2) instead of one (eq 1). Another interesting feature of these experiments is that the size of the surviving micelle fragments appears to decrease, from Rh = 39 nm at 70 °C to 21 nm at 85 °C. We plot all of the Rh values obtained through DOSY measurements in Figure 5. This plot

Table 1. DOSY.01 Experimental Diffusion Coefficients and Calculated Rh Values of PFS65-b-PI637 Unimer and Micelle Fragments in Decane at Various Temperaturesa DOSY.01b T (°C) 25 70 75 80 85 90 95 100

108Ds (cm2/s) PI signal 6.7 11 72 12 86 15 97 16 93 74 120 120

PFS signal

130 87 89 120 130

Rh (nm) PI signal 38 45 6.9 43 6.1 36 5.7 37 6.3 8.1 5.5 5.6

PFS signal

R2 values of the fits PI signal

PFS signal

0.9991 0.9998 0.9998 0.9998

0.9673

0.9997

0.9958

0.9982 0.9992 0.9997

0.9953 0.9975 0.9979

4.3 6.7 6.8 5.3 5.5

a

A capillary tube insert containing DMSO-d6 was used as a deuterium lock. bIn DOSY.01 experiments, a single sample of micelle fragments in decane was examined stepwise at each increasing temperature. The measurement at each temperature took ca. 2 h.

to minimize sample history effects on our PFS65-b-PI637 micelle fragment sample, we carried out a second set of DOSY experiments (DOSY.02). Here we prepared multiple identical samples, each at 6 mg/mL. We used one sample for each temperature. Each sample was held at the measurement temperature for about 2 h, but the extent of annealing for the samples measured at the highest temperatures was reduced. The results are summarized in Table 2. The PFS signal could be detected starting at 75 °C, and these intensity decays gave good fits to a single-exponential function. Values of Rh were in the range of 5.1−5.9 nm at all temperatures, indicating that the unimer does not change its dimensions significantly in this temperature range. At 90, 95, and 100 °C, the PI signal decay also fit well to one exponential. Calculated values of Rh were

Figure 5. Hydrodynamic radii of all the species detected by DOSY NMR at temperatures from 25 to 100 °C in decane.

emphasizes that the size of the unimer does not change over the range of temperatures examined. In contrast, the two sets of DOSY experiments gave somewhat different results for the size of the surviving seeds at temperatures between 70 and 85 °C. The first set of experiments suggests a small decrease in Rh from 45 to 36 nm with increasing temperature. The second set of experiments (DOSY.02) suggested a more striking decrease in size, from 40 to 21 nm over this temperature range. Lastly, we examined whether we could fit the twocomponent DOSY data to eq 5 as an independent measurement of the fraction of material present as micelle fragments and as unimer. This turned out not to be possible without further strong assumptions about the data, since this analysis requires experimental values of T1 and T2 of each species at each temperature. Further comments about this analysis and the assumptions that need to be made are provided in the Supporting Information. TEM Measurements. As a control to monitor how the extent of annealing affect micelle morphology, we chose the samples examined by DOSY.02 at 70, 80, 90, and 100 °C, each measurement taking 2 h. After each sample cooled to room temperature (typical self-seeding conditions), aliquots were taken for TEM analysis. These TEM images are presented in Figures S5−S7. Micelle elongation compared to the original seeds was found in all of the samples, but the extent of elongation for the samples heated at 70 and 80 °C was rather small. Here the Ln values increased to 48 and 54 nm, respectively. At higher temperatures, the resulting micelles showed significant increases in length and had an average length of 392 nm at 90 °C and 1264 nm at 100 °C. The Lw/Ln value of 70 °C sample was 1.14, which was similar to that of the initial fragments, whereas at higher temperatures, narrower

Table 2. DOSY.02 Experimental Diffusion Coefficients and Calculated Rh Values of PFS65-b-PI637 Unimer and Micelle Fragments in Decane at Various Temperaturesa DOSY.02b T (°C) 25 70 75 80 85 90 95 100

108×Ds (cm2/s) PI signal 6.7 13 110 17 120 21 120 27 110 99 110 120

PFS signal

100 94 99 110 110 120

Rh (nm) PI signal 38 39 4.5 32 4.5 26 4.5 21 5.2 6.1 5.5 5.8

PFS signal

R2 values of the fits PI signal

PFS signal

0.9991 0.9994 0.9998

0.9876

0.9999

0.9982

0.9996

0.9986

0.9994 0.9987 0.9999

0.9970 0.9975 0.9997

5.1 5.9 5.9 5.6 5.8 5.9

a

A capillary tube insert containing DMSO-d6 was used as a deuterium lock. bIn DOSY.02 experiments, separate samples of micelle fragments in decane were prepared for each temperature. The samples were heated rapidly to the measurement temperature. The measurement at each temperature took ca. 2 h. H

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules distributions of 1.04, 1.06, and 1.02 were found for 80, 90, and 100 °C samples, respectively. These results are different from what we reported in a previous paper using the same block copolymer “PFS53-b-PI637”.25 Those experiments were carried out at a much lower micelle fragment concentration (0.02 mg/ mL in decane). Under self-seeding conditions, this sample yielded elongated micelles with Ln = 2500 nm after heating at 70 °C for 30 min. In the VT-NMR and DOSY NMR experiments reported here, the micelle fragment concentration was 6 mg/mL. This result points to an important concentration effect in self-seeding experiments that we are currently investigating. It will be the subject of a future publication.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (G.G.). *E-mail [email protected] (M.A.W.).

■

ORCID

SUMMARY We used variable temperature (VT) 1H and DOSY NMR to examine the behavior of short core-crystalline micelle fragments of PFS65-b-PI637 BCP in decane subjected to self-seeding conditions. We examined a sample of micelle fragments of PFS65-b-PI637 characterized by Ln = 39 nm and Lw = 44 nm, and monitored the fraction of BCP that dissolved as a function of temperature. Our experiments took advantage of the difference in mobility of the PI corona chains and PFS in the micelle core. In 1H NMR measurements, the signal from the PFS component could be detected only for the dissolved unimer. In this way, we found that all of the BCP molecules were incorporated into the micelles at temperatures up to and including 50 °C. Both PFS and PI resonances could be detected between 70 and 100 °C, and at 100 °C, the integration ratio of the PFS-to-PI peaks was essentially the same as for a solution of the BCP in benzene-d6, a common good solvent. Thus, we infer that essentially all of the polymer had dissolved at 100 °C. At intermediate temperatures in decane, the fraction of unimers in solution slightly decreased over time, consistent with some regrowth of the micelle fragments as reported previously in ref 43. We also examined pulsed-gradient spin echo (DOSY NMR) experiments at these temperatures with the goal of measuring self-diffusion coefficients (Ds) and calculating hydrodynamic radii (Rh) for the micelle fragments and unimers. Another interesting aspect of the DOSY results is that the size of the micelle fragments at elevated temperatures (80 and 85 °C) was sensitive to sample history, with smaller fragments (Rh = 26 and 21 nm) obtained for samples heated quickly to the measurement temperature. Samples subjected to prolonged annealing at lower temperatures showed a much smaller reduction in Rh (36 and 37 nm) when heated to 80 or 85 °C. DOSY NMR is a very interesting method to investigate the dissolution of core-crystalline micelles upon heating and the evolution of micelle size. Using the combination of 1H and DOSY NMR, we could quantify the fraction of free unimer and micelles in solution at different temperatures. Despite the detection limit of DOSY NMR and the fact that we need samples with higher than usual concentrations for this type of study reported here, our experiments provide useful insights into the strengths and limitations of this methodology.

■

(Figure S1), experimental decane viscosities (Figure S2), NMR purification results (Figure S3), GPC traces of polymer purification (Figure S4), characterizations of micelle fragments from DOSY (Figures S5−S7); tables summarizing the data obtained by DOSY (Tables S1− S4) (PDF)

Gerald Guerin: 0000-0003-4997-0561 Ian Manners: 0000-0002-3794-967X Mitchell A. Winnik: 0000-0002-2673-2141 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The Toronto authors thank NSERC Canada for their support of this research and the CSB Imaging Facility at the University of Toronto.

■

REFERENCES

(1) Blundell, D. J.; Keller, A.; Kovacs, A. J. A New Self-nucleation Phenomenon And Its Application to The Growing of Polymer Crystals From Solution. J. Polym. Sci., Part B: Polym. Lett. 1966, 4 (7), 481− 486. (2) Blundell, D. J.; Keller, A. Nature of self-seeding polyethylene crystal nuclei. J. Macromol. Sci., Part B: Phys. 1968, 2 (2), 301−336. (3) Zheng, J. X.; Xiong, H.; Chen, W. Y.; Lee, K.; Van Horn, R. M.; Quirk, R. P.; Lotz, B.; Thomas, E. L.; Shi, A.-C.; Cheng, S. Z. D. Onsets of Tethered Chain Overcrowding and Highly Stretched Brush Regime via Crystalline−Amorphous Diblock Copolymers. Macromolecules 2006, 39 (2), 641−650. (4) Li, B.; Li, C. Y. Immobilizing Au nanoparticles with polymer single crystals, patterning and asymmetric functionalization. J. Am. Chem. Soc. 2007, 129 (1), 12−13. (5) Xu, J.; Ma, Y.; Hu, W.; Rehahn, M.; Reiter, G. Cloning polymer single crystals through self-seeding. Nat. Mater. 2009, 8 (4), 348−353. (6) Cao, L.; Manners, I.; Winnik, M. A. Influence of the Interplay of Crystallization and Chain Stretching on Micellar Morphologies: Solution Self-Assembly of Coil−Crystalline Poly(isoprene-blockferrocenylsilane). Macromolecules 2002, 35 (22), 8258−8260. (7) Massey, J. A.; Temple, K.; Cao, L.; Rharbi, Y.; Raez, J.; Winnik, M. A.; Manners, I. Self-Assembly of Organometallic Block Copolymers: The Role of Crystallinity of the Core-Forming Polyferrocene Block in the Micellar Morphologies Formed by Poly(ferrocenylsilane-b-dimethylsiloxane) inn-Alkane Solvents. J. Am. Chem. Soc. 2000, 122 (47), 11577−11584. (8) Fairley, N.; Hoang, B.; Allen, C. Morphological control of poly(ethylene glycol)-block-poly(epsilon-caprolactone) copolymer aggregates in aqueous solution. Biomacromolecules 2008, 9 (9), 2283− 2291. (9) Du, Z.-X.; Xu, J.-T.; Fan, Z.-Q. Micellar Morphologies of Poly(εcaprolactone)-b-poly(ethylene oxide) Block Copolymers in Water with a Crystalline Core. Macromolecules 2007, 40 (21), 7633−7637. (10) Fujiwara, T.; Miyamoto, M.; Kimura, Y.; Iwata, T.; Doi, Y. SelfOrganization of Diblock and Triblock Copolymers of Poly(l-lactide) and Poly(oxyethylene) into Nanostructured Bands and Their Network System. Proposition of a Doubly Twisted Chain Conformation of Poly(l-lactide). Macromolecules 2001, 34 (12), 4043−4050. (11) Zhang, J.; Wang, L. Q.; Wang, H.; Tu, K. Micellization phenomena of amphiphilic block copolymers based on methoxy

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.8b00098. Additional experimental details and analysis of DOSY NMR decays; the pulsed field gradient NMR sequence I

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

polymers. J. Polym. Sci., Part B: Polym. Phys. 2006, 44 (12), 1738− 1750. (28) Maus, A.; Hempel, E.; Thurn-Albrecht, T.; Saalwachter, K. Memory effect in isothermal crystallization of syndiotactic polypropylene - role of melt structure and dynamics? Eur. Phys. J. E: Soft Matter Biol. Phys. 2007, 23 (1), 91−101. (29) Guerin, G.; Rupar, P. A.; Manners, I.; Winnik, M. A. Explosive dissolution and trapping of block copolymer seed crystallites. Nat. Commun. 2018, 9 (1), 1158. (30) Strobl, G. Crystallization and melting of bulk polymers: New observations, conclusions and a thermodynamic scheme. Prog. Polym. Sci. 2006, 31 (4), 398−442. (31) Shen, L.; Soong, R.; Wang, M.; Lee, A.; Wu, C.; Scholes, G. D.; Macdonald, P. M.; Winnik, M. A. Pulsed field gradient NMR studies of polymer adsorption on colloidal CdSe quantum dots. J. Phys. Chem. B 2008, 112 (6), 1626−1633. (32) Tong, L.; Lu, E.; Pichaandi, J.; Cao, P.; Nitz, M.; Winnik, M. A. Quantification of Surface Ligands on NaYF4 Nanoparticles by Three Independent Analytical Techniques. Chem. Mater. 2015, 27 (13), 4899−4910. (33) Baldwin, A. J.; Anthony-Cahill, S. J.; Knowles, T. P.; Lippens, G.; Christodoulou, J.; Barker, P. D.; Dobson, C. M. Measurement of amyloid fibril length distributions by inclusion of rotational motion in solution NMR diffusion measurements. Angew. Chem., Int. Ed. 2008, 47 (18), 3385−3387. (34) Johnson, C. S. Diffusion ordered nuclear magnetic resonance spectroscopy: principles and applications. Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34 (3−4), 203−256. (35) Morris, K. F.; Johnson, C. S. Resolution of Discrete and Continuous Molecular-Size Distributions by Means of DiffusionOrdered 2d Nmr-Spectroscopy. J. Am. Chem. Soc. 1993, 115 (10), 4291−4299. (36) Jerschow, A.; Muller, N. Diffusion-separated nuclear magnetic resonance spectroscopy of polymer mixtures. Macromolecules 1998, 31 (19), 6573−6578. (37) Viel, S.; Mazarin, M.; Giordanengo, R.; Phan, T. N.; Charles, L.; Caldarelli, S.; Bertin, D. Improved compositional analysis of block copolymers using diffusion ordered NMR spectroscopy. Anal. Chim. Acta 2009, 654 (1), 45−48. (38) Cain, J. B.; Zhang, K.; Betts, D. E.; DeSimone, J. M.; Johnson, C. S. Diffusion of block copolymers in liquid CO2: Evidence of selfassembly from pulsed field gradient NMR. J. Am. Chem. Soc. 1998, 120 (36), 9390−9391. (39) Bakkour, Y.; Darcos, V.; Li, S.; Coudane, J. Diffusion ordered spectroscopy (DOSY) as a powerful tool for amphiphilic block copolymer characterization and for critical micelle concentration (CMC) determination. Polym. Chem. 2012, 3 (8), 2006−2010. (40) Gilroy, J. B.; Rupar, P. A.; Whittell, G. R.; Chabanne, L.; Terrill, N. J.; Winnik, M. A.; Manners, I.; Richardson, R. M. Probing the structure of the crystalline core of field-aligned, monodisperse, cylindrical polyisoprene-block-polyferrocenylsilane micelles in solution using synchrotron small- and wide-angle X-ray scattering. J. Am. Chem. Soc. 2011, 133 (42), 17056−17062. (41) Stejskal, E. O.; Tanner, J. E. Spin Diffusion Measurements: Spin Echoes in the Presence of a Time-Dependent Field Gradient. J. Chem. Phys. 1965, 42 (1), 288−292. (42) Rupar, P. A.; Cambridge, G.; Winnik, M. A.; Manners, I. Reversible cross-linking of polyisoprene coronas in micelles, block comicelles, and hierarchical micelle architectures using Pt(0)-olefin coordination. J. Am. Chem. Soc. 2011, 133 (42), 16947−16957. (43) Guerin, G.; Rupar, P.; Molev, G.; Manners, I.; Jinnai, H.; Winnik, M. A. Lateral Growth of 1D Core-Crystalline Micelles upon Annealing in Solution. Macromolecules 2016, 49 (18), 7004−7014. (44) Assuming that Rg/Rh = 1.78 for monodisperse chains in good solvent, we can calculate a value of Rg for this homopolymer (Rg = 6.2 nm). This value is consistent with the value evaluated from data reported for 3,4-rich polyisoprene by Fetters. Burchard, W. Solution Properties of Branched Macromolecules; Springer: Berlin, 1999; Vol. 143, pp 113−194. Fetters, L. J.; Lohse, D. J.; Colby, R. H. Chain

poly(ethylene glycol) and either crystalline or amorphous poly(caprolactone-b-lactide). Biomacromolecules 2006, 7 (9), 2492−2500. (12) Schmalz, H.; Schmelz, J.; Drechsler, M.; Yuan, J.; Walther, A.; Schweimer, K.; Mihut, A. M. Thermoreversible Formation of Wormlike Micelles with a Microphase-Separated Corona from a Semicrystalline Triblock Terpolymer. Macromolecules 2008, 41 (9), 3235−3242. (13) Lazzari, M.; Scalarone, D.; Hoppe, C. E.; Vazquez-Vazquez, C.; Lòpez-Quintela, M. A. Tunable Polyacrylonitrile-Based Micellar Aggregates as a Potential Tool for the Fabrication of Carbon Nanofibers. Chem. Mater. 2007, 19 (24), 5818−5820. (14) Lee, E.; Hammer, B.; Kim, J. K.; Page, Z.; Emrick, T.; Hayward, R. C. Hierarchical helical assembly of conjugated poly(3-hexylthiophene)-block-poly(3-triethylene glycol thiophene) diblock copolymers. J. Am. Chem. Soc. 2011, 133 (27), 10390−10393. (15) Qian, J.; Li, X.; Lunn, D. J.; Gwyther, J.; Hudson, Z. M.; Kynaston, E.; Rupar, P. A.; Winnik, M. A.; Manners, I. Uniform, high aspect ratio fiber-like micelles and block co-micelles with a crystalline pi-conjugated polythiophene core by self-seeding. J. Am. Chem. Soc. 2014, 136 (11), 4121−4124. (16) Kynaston, E. L.; Gould, O. E. C.; Gwyther, J.; Whittell, G. R.; Winnik, M. A.; Manners, I. Fiber-Like Micelles from the Crystallization-Driven Self-Assembly of Poly(3-heptylselenophene)block-Polystyrene. Macromol. Chem. Phys. 2015, 216 (6), 685−695. (17) Arno, M. C.; Inam, M.; Coe, Z.; Cambridge, G.; Macdougall, L. J.; Keogh, R.; Dove, A. P.; O’Reilly, R. K. Precision Epitaxy for Aqueous 1D and 2D Poly(epsilon-caprolactone) Assemblies. J. Am. Chem. Soc. 2017, 139 (46), 16980−16985. (18) Qian, J.; Guerin, G.; Cambridge, G.; Manners, I.; Winnik, M. A. Seeded growth and solvent-induced fragmentation of fiberlike polyferrocenylsilane-polyisoprene block copolymer micelles. Macromol. Rapid Commun. 2010, 31 (9−10), 928−933. (19) Gädt, T.; Ieong, N. S.; Cambridge, G.; Winnik, M. A.; Manners, I. Complex and hierarchical micelle architectures from diblock copolymers using living, crystallization-driven polymerizations. Nat. Mater. 2009, 8 (2), 144−150. (20) Wang, X.; Guerin, G.; Wang, H.; Wang, Y.; Manners, I.; Winnik, M. A. Cylindrical block copolymer micelles and co-micelles of controlled length and architecture. Science 2007, 317 (5838), 644− 647. (21) Gilroy, J. B.; Gädt, T.; Whittell, G. R.; Chabanne, L.; Mitchels, J. M.; Richardson, R. M.; Winnik, M. A.; Manners, I. Monodisperse cylindrical micelles by crystallization-driven living self-assembly. Nat. Chem. 2010, 2 (7), 566−570. (22) Tao, D.; Feng, C.; Cui, Y.; Yang, X.; Manners, I.; Winnik, M. A.; Huang, X. Monodisperse Fiber-like Micelles of Controlled Length and Composition with an Oligo(p-phenylenevinylene) Core via “Living” Crystallization-Driven Self-Assembly. J. Am. Chem. Soc. 2017, 139 (21), 7136−7139. (23) Qian, J.; Guerin, G.; Lu, Y.; Cambridge, G.; Manners, I.; Winnik, M. A. Self-seeding in one dimension: an approach to control the length of fiberlike polyisoprene-polyferrocenylsilane block copolymer micelles. Angew. Chem., Int. Ed. 2011, 50 (7), 1622−1625. (24) Qian, J.; Lu, Y.; Cambridge, G.; Guerin, G.; Manners, I.; Winnik, M. A. Polyferrocenylsilane Crystals in Nanoconfinement: Fragmentation, Dissolution, and Regrowth of Cylindrical Block Copolymer Micelles with a Crystalline Core. Macromolecules 2012, 45 (20), 8363− 8372. (25) Qian, J.; Lu, Y.; Chia, A.; Zhang, M.; Rupar, P. A.; Gunari, N.; Walker, G. C.; Cambridge, G.; He, F.; Guerin, G.; Manners, I.; Winnik, M. A. Self-seeding in one dimension: a route to uniform fiber-like nanostructures from block copolymers with a crystallizable coreforming block. ACS Nano 2013, 7 (5), 3754−3766. (26) Massa, M. V.; Lee, M. S. M.; Dalnoki-Veress, K. Crystal nucleation of polymers confined to droplets: Memory effects. J. Polym. Sci., Part B: Polym. Phys. 2005, 43 (23), 3438−3443. (27) Lorenzo, A. T.; Arnal, M. L.; Sánchez, J. J.; Müller, A. J. Effect of annealing time on the self-nucleation behavior of semicrystalline J

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Dimensions and Entanglement Spacings. In Physical Properties of Polymers Handbook; Springer: New York, 2007; pp 447−454. (45) Gladstone, J. H.; Dale, T. P. Researches on the Refraction, Dispersion, and Sensitiveness of Liquids. Philosophical Transactions of the Royal Society of London 1863, 153, 317−343. (46) Baldwin, A. J.; Christodoulou, J.; Barker, P. D.; Dobson, C. M.; Lippens, G. Contribution of rotational diffusion to pulsed field gradient diffusion measurements. J. Chem. Phys. 2007, 127 (11), 114505. (47) Smallcombe, S. H.; Patt, S. L.; Keifer, P. A. WET solvent suppression and its applications to LC NMR and high-resolution NMR spectroscopy. J. Magn. Reson., Ser. A 1995, 117 (2), 295−303. (48) Rochette, C. N.; Rosenfeldt, S.; Henzler, K.; Polzer, F.; Ballauff, M.; Tong, Q.; Mecking, S.; Drechsler, M.; Narayanan, T.; Harnau, L. Annealing of Single Lamella Nanoparticles of Polyethylene. Macromolecules 2011, 44 (12), 4845−4851. (49) Osichow, A.; Rabe, C.; Vogtt, K.; Narayanan, T.; Harnau, L.; Drechsler, M.; Ballauff, M.; Mecking, S. Ideal polyethylene nanocrystals. J. Am. Chem. Soc. 2013, 135 (31), 11645−11650.

K

DOI: 10.1021/acs.macromol.8b00098 Macromolecules XXXX, XXX, XXX−XXX