Co-assembly of precision polyurethane ionomers reveals role of and interplay between individual components

Industrial and household products, such as paints, inks and cosmetics usually consist of mixtures of macromolecules that are disperse in composition, in size and in monomer sequence. Identifying structure–function relationships for these systems is complicated, as particular macromolecular components cannot be investigated individually. For this study, we have addressed this issue, and have synthesized a series of five sequence-defined polyurethanes (PUs): one neutral-hydrophobic, one single-charged hydrophilic, one single-charged hydrophobic and two double-charged amphiphilic PUs (one symmetric and one asymmetric). These novel precision PUs – that were prepared by using stepwise coupling-deprotection synthetic protocols – have a defined composition, size and monomer sequence, where the chosen sequences were inspired by those that are abundantly formed in the production of industrial waterborne PU dispersions. By performing dynamic light scattering experiments (DLS), self-consistent field (SCF) computations and cryogenic transmission electron microscopy (cryo-TEM), we have elucidated the behavior in aqueous solution of the individual precision PUs, as well as of binary and ternary mixtures of the PU sequences. The double-charged PU sequences (‘hosts’) were sufficiently amphiphilic to yield single-component micellar solutions, whereas the two more hydrophobic sequences did not micellize on their own, and gave precipitates or ill-defined larger aggregates. Both the neutral-hydrophobic PU and the hydrophilic single-charged PU were successfully incorporated in the host micelles as guests, respectively increasing and reducing the micelle radius upon incorporation. SCF computations indicated that double-charged symmetric PUs stretch whilst double-charged asymmetric PUs are expelled from the core to accommodate hydrophobic PU guests within the micelles. For the ternary mixture of the double-charged symmetric and asymmetric hosts and the neutral-hydrophobic guest we have found an improved colloidal stability, as compared to those for binary mixtures of either host and hydrophobic guest. In another ternary mixture of precision PUs, with all three components not capable of forming micelles on their own, we see that the ensemble of molecules produces stable micellar solutions. Taken together, we find that the interplay between PU-molecules in aqueous dispersions promotes the formation of stable micellar hydrocolloids.


Molecular structures of the sequence-defined precision PU(I) materials
Scheme S1 shows the (macro)molecular structures of the PU(I)s central to this study.
Scheme S1: Molecular structures of the PU(I) materials as used in this study. The average molar mass of pTHF is about 2 kDa. Given molecular weight values are for pTHF chains with 28 THFunits. Note that the IPDA building blocks exist in multiple regio-and stereoisomers; this is not shown for brevity. course, all molecules 1 have only one amine and only one Boc-protected amine group, so in this sense amine 1 is defined. Similarly, building block BB (Scheme S3) is molecularly defined as it has only one active carbonate group, only one benzyl-ester group and only one Boc-protected amine group.
We have chosen the isophorone diamine (IPDA) building block for our syntheses, as commercial waterborne polyurethanes are very frequently prepared from isophorone diisocyanate (IPDI). We have considered using hexane-diamine/diisocyanate (HDA/HDI) building blocks instead, which would have resulted in materials with more regular molecular structures and more regularly positioned hydrogen bonds. However, and importantly, this could have easily led to materials that are decidedly more crystalline in nature. Indeed, a comparative study on polyester thermoplastic elastomers, using either HDI or IPDI building blocks, shows clear crystalline phases for only the HDI-material (reference 2). In designing the sequence-defined PU(I)s of this study, we did not want to introduce a factor (i.e. crystallinity) that could dominate the behavior of these materials, given that this factor is not important in industrially employed IPDI-based WPUs. In conclusion, we wanted the designed and prepared sequence-defined PU(I) to reflect the properties of industrial WPUs as close as possible.

Materials
All reagents, chemicals, materials and solvents were obtained from commercial sources, and were used as received: Cambridge Isotope Laboratories for (deuterated) solvents, Aldrich, Acros, ABCR, Merck and Fluka for chemicals, materials and reagents. All solvents were of AR quality. Moisture or oxygen-sensitive reactions were performed under an atmosphere of dry argon. Analytical thin layer chromatography was performed on Kieselgel F-254 precoated silica plates. Column chromatography was carried out on Screening Devices B.V. flashsilica gel (40-63 μm mesh) or normal silica gel (60-200 μm mesh). In the polymer extension reactions, and for scavenging the excess of BB building block, an amine terminated resin was employed (Silicycle Si-Amine catnr. R5203B loading 1.89 mmol/g). For hydrogenations 10% Pd/C Degussa type E101 NE/W (Sigma-Aldrich) was used.

Molecular characterization
NMR spectra were recorded on a 400 MHz Varian Mercury spectrometer at 298 K. Chemical shifts are reported in ppm downfield from TMS at room temperature using deuterated chloroform (CDCl3) as a solvent and internal standard unless otherwise indicated. Abbreviations used for splitting patterns are s = singlet, t = triplet, q = quartet, m = multiplet (or multiple signals), dd = double doublet, and b or br = broad. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) was performed on a PerSeptive Biosystems Voyager-DE PRO spectrometer. As indicated, these measurements were done by using a α-cyano-4-hydroxycinnamic acid (CHCA) matrix, where KAc was optionally added to the matrix to primarily generate K +adducts, and a linear positive mode was employed. Size-exclusion chromatography (SEC) was measured on a Shimadzu LC-10AD VP system with a RID-10A detector and a SPD-M20A diode array detector using a PL gel 5μm mixed-C and a mixed-D column in sequence and using THF as the eluent (flow rate 1 mL min -1 ). The SEC molecular weight and distribution data were recorded relative to polystyrene (PS) standards using refractive index (RI) detection. Attenuated total reflection Fourier-transform infrared (ATR-FT-IR) spectra were recorded on a Shimadzu IRAffinity spectrometer using MIRacle-10 ATR accessory. HPLC-PDA/ESI-MS was performed using a Shimadzu LC-10 AD VP series HPLC coupled to a diode array detector (Finnigan Surveyor PDA Plus detector, Thermo Electron Corporation) and an Ion-Trap (LCQ Fleet, Thermo Scientific). Electrospray ionization (ESI) was used to create charged species for mass detection. Analyses were performed at 298 K using an Alltech Alltima HP C18 3μ column using an injection volume of 1-4 μL, a flow rate of 0.2 mL min -1 and typically a gradient (5% to 100% in 10 min, held at 100% for a further 3 min) of CH3CN in H2O with both these eluents containing 0.1% formic acid.

Mono-Boc isophorone diamine (Boc-IPDA-NH 2 ): isomeric mixture of tert-butyl [3-(aminomethyl)-3,5,5-trimethylcyclohexyl]carbamate and tert-butyl [(5-amino-1,3,3trimethylcyclohexyl)methyl]carbamate (molecule 1)
A solution of isophorone diamine (IPDA; 3 g, 17.6 mmol, 2 molar equivalents) in DCM (20 mL) was stirred at -78 0 C. Boc-anhydride (1.92 g, 8.8 mmol) in DCM (80 mL) was added drop wise. After addition the mixture was allowed to heat up to room temperature, upon which it became hazy. HPLC-MS analysis showed that about 75% of mono-Boc product and 23% of di-Boc product had formed. The reaction mixture was evaporated, a solution of 0.1M formic acid (pH=3) was added, and this solution was washed two times with DCM (2 x 20 mL) to remove di-Boc product. The water layer was then brought to pH=9 with a 0.1M NaOH solution, and the product was extracted using two portions of DCM (20 mL). The combined organic layers were repeatedly washed with a solution of borax buffer (pH 9.3; 9 mL 0.1M NaOH and 91 mL 0.05M sodium tetraborate), until non-functionalized IPDA had disappeared from the organic layer. The presence of IPDA was checked by HPLC-MS analysis on a small sample after reaction with phenyl isocyanate. The organic layer was dried with Na2SO4 and concentrated to obtain the product as a waxy solid. Yield 3.6 grams (76%).
Di-methylol-propionic acid (DMPA, 100 g, 746 mmol) and KOH (48.2 g, 732 mmol) were stirred in DMF (500 mL) for one hour at 100 0 C. Benzyl bromide (153.4 g, 890 mmol, 1.2 molar equivalents) was added dropwise to the hot mixture that was thereafter stirred overnight at 100 0 C under argon. A KBr-suspension formed. The mixture was concentrated to dryness by evaporation of the volatiles. The crude product was dissolved in a 1/1 mixture of ethyl acetate and hexane (1L), and this solution was washed with several portions of water (2L total volume). The organic layer was dried with Na2SO4, filtered and the filtrate was concentrated. The residue was recrystallized from toluene (300 mL) to yield a first crop of 96.6 g (58%) of product. The filtrate was concentrated and once more recrystallized to give a second crop (6.3 grams).

E. The synthesis of PUI-A1 including analytical data
The synthesis of the asymmetrical poly-urethane PUI-A1 is similar to that for the previously published PUI-A2.
Employing extension building block BB, poly-ether mono-amine 5 was converted in 2 steps to PUI-A1 (Scheme S3). In the coupling reaction, BB was used in molar excess. After complete conversion of the coupling reaction, an amine functional scavenger resin was added to the reaction mixture to remove BB, and thus separate it from the desired polyurethane product. Debenzylation was required to arrive at the PUI-A1 end product.
Note that the stepwise syntheses of the PUIs in this paper rely on the use of building block BB. It contains (i) a 4-nitrophenyl-carbonate group that is stable at room temperature and that reliable reacts with amines to produce urethane linked products, (ii) a stable benzyl-ester group that can conveniently be deprotected by mild Pd/C-H2 reduction to give the ionomeric COOH-group, and (iii) a stable Boc-group that allows mild deprotection with TFA to produce a mono-amine reactive group that is suited and ready for the next iterative extension reaction with BB.

MeO-polyTHF2000-IPDA-DMPA(Bn)-IPDA-Boc (polymer 6)
Building block molecule BB (0.31 g, 0.45 mmol, 2 molar equivalents) was dissolved in dioxane (3 mL) together with MeO-polyTHF2000-IPDA-NH2 (polymer 5, 0.7 g, 0.23 mmol) and pyridine (72 µL, 4 equivalents). The mixture was heated to reflux for 16 hours and was kept under an atmosphere of argon. When the reaction was complete ( 1 H-NMR monitoring) a silica gel functionalized with amine groups (Silicycle Si-Amine catnr. R5203B loading 1.89 mmol/g) was added to react with the excess of building block BB. The mixture was stirred for 4 hours at the elevated temperature. After filtration of the reaction mixture to remove the resin, the filtrate was concentrated by evaporation of the solvent. The crude product was dissolved in chloroform, and the organic layer was washed two times with a 0.1M NaOH solution and once with a saturated NaCl solution. The chloroform layer was dried with Na2SO4 and concentrated by evaporation of the solvent to yield a yellowish oil (780 mg). MeO-pTHF2000-IPDA-DMPA(Bn)-IPDA-Boc (polymer 6, 630 mg) and Pd/C Degussa type catalyst (100 mg) were mixed in dioxane (4 mL). The reaction mixture was stirred overnight at room temperature under a hydrogen atmosphere. When the conversion was complete (check with 1 H-NMR for removal of the benzyl ester) the reaction mixture was filtrated over a zeolite plug to remove the Pd-catalyst. After evaporation of the solvent the product was obtain as a yellowish clear oil (491 mg; 80%). The lower yield was due to the presence of remaining solvent in the starting polymer 6.  2941, 2860, 1711, 1524, 1456, 1371, 1304, 1238, 1211, 1169, 1107, 1047, 1011, 995, 961, 901, 866, 831, 810, 772, 746, 727, 696, 667,

(iii) NMR information
Due to the presence of the IPDA units in the prepared PU(I)s, the 1 H-NMR data become complicated as various isomers give resonances at different positions. The below gives an overview of the proton assignments, where the given virtual molecule is a simplified structure that represents the variety of protons that are typically found in the prepared set of PU(I)s.   H-NMR can be used the calculate and estimate the average number molecular weight Mn of the poly-THF segment within the PU(I) sequences, and therefore also the Mn of the PU-material (Table S4 and  Table S5). The actual 1 H-NMR spectra of the PUs are compiled in ESI Section 7.
Briefly, the integral in the 1.5-0.6 ppm region is used to calculate the integral per proton. Next, the integrals of the signals at about 3.4 and 1.6 ppm (of the pTHF units) are used to calculate the average number of pTHF units in the PU-sequence. Note that the integral of the pTHF units is corrected as the signal at 3.4 ppm also accounts for the 3 methoxy-protons (in PUI-A1 and PUI-A2) and does not account for the CH2-OCONH protons (2 or 4 for PUI-A1 and PUI-A2 or PU-S0 and PUI-S2, respectively), while the signal at 1.6 ppm also accounts for 2 protons of every IPDA unit.

Cryo-TEM analyses
Sequence-defined PUs in aqueous 0.1 M TEA solutions were analyzed by cryo-TEM (cryogenic transmission electron microscopy). Dimensions of observed micelle particles were assessed by applying ImageJ software, an open source image processing program, by checking every particle twice, first along the longest axis of the particle and second perpendicular to this axis. Acquired data are collected and shown in Table S7 and Figure S7 to Figure S10. The micelles formed by the combined hosts PUI-A1 and sPUI-S1 and the PU-S0 guest show a broader distribution in radius than the other analyzed micelles.

DSC information
The PU(I) materials were analyzed using a TA Q2000 DSC machine. Data were thermally analyzed using Universal Analysis software.
Samples were heated at 10 0 C/min in the first heating run to 80 0 C or 100 0 C. Thereafter and subsequently, a cyle at a scanning rate of 10 0 C/min and a cycle at a scanning rate of 40 0 C/min. were recorded. Cooling runs went to -85 0 C. In the below, the thermal results are compiled. The DSC traces are shown in ESI Section 7.

PU-S0.
White semi-crystalline material. Melt in the first heating run at Tm = 34.9 0 C (33.5 J/g). Partial crystallization in first cooling run. Then recrystallization at Tcr = -15.8 0 C (31.5 J/g) and subsequent Tm = 21.9 0 C (58.7 J/g) in the second heating run. These transitions represent melting and (re)crystallization of the poly-THF crystalline phase. Possible Tg of amorphous poly-THF phase at about -65 0 C (heating run 40 0 C/min).

PUI-A1.
Thick waxy-oil material. No transition in the first heating run. Partial crystallization in first cooling run at Tcr = -11.4 0 C (54.7 J/g). Then minor recrystallization and melt at Tm = 20.8 0 C (67.2 J/g) in the second heating run. These transitions represent melting and (re)crystallization of the poly-THF crystalline phase. No Tg of the amorphous poly-THF phase is observed.

PUI-A2.
Thick waxy-oil material. No transition in the first heating run. No crystallization in the first cooling run. Then recrystallization at Tcr = -13.3 0 C (35.3 J/g).) and subsequent melt at Tm = 17.0 0 C (42.7 J/g) in the second heating run. These transitions represent melting and (re)crystallization of the poly-THF crystalline phase. No Tg of the amorphous poly-THF phase is observed.
PUI-S2. Thick waxy oil material. No transitions are observed. Fully amorphous material. Apparently, the end groups prevent the poly-THF phase from crystallizing.