Modular Synthesis and Patterning of High-Stiffness Networks by Postpolymerization Functionalization with Iron–Catechol Complexes

Bioinspired iron–catechol cross-links have shown remarkable success in increasing the mechanical properties of polymer networks, in part due to clustering of Fe3+–catechol domains which act as secondary network reinforcing sites. We report a versatile synthetic procedure to prepare modular PEG-acrylate networks with independently tunable covalent bis(acrylate) and supramolecular Fe3+–catechol cross-linking. Initial control of network structure is achieved through radical polymerization and cross-linking, followed by postpolymerization incorporation of catechol units via quantitative active ester chemistry and subsequent complexation with iron salts. By tuning the ratio of each building block, dual cross-linked networks reinforced by clustered iron–catechol domains are prepared and exhibit a wide range of properties (Young’s moduli up to ∼245 MPa), well beyond the values achieved through purely covalent cross-linking. This stepwise approach to mixed covalent and metal–ligand cross-linked networks also permits local patterning of PEG-based films through masking techniques forming distinct hard, soft, and gradient regions.


Materials & Instrumentation
All chemicals were purchased and used as received. Poly(ethylene glycol diacrylate) (PEGDA, M n 700, with inhibitor), poly(ethylene glycol methyl ether acrylate) (PEGMEA, M n 480, with inhibitor), pentafluorophenol, acryloyl chloride (with inhibitor), dopamine hydrochloride salt, triethyl amine, and DMPA (2,2-dimethoxy-2-phenylacetophenone), were purchased from Sigma Aldrich and used as received. Bicine was purchased from TCI, and Fe(NO 3 ) 3 (nonahydrate, ACS reagent Grade) was purchased from Fischer Scientific (Acros Organics). Isopropanol, hexanes, methanol, and other solvents were procured from Sigma Aldrich and used as received without further drying or distillation, with the exception of DCM used in PFPA synthesis, which was dried via a PureSolv MD-5 dry solvent system. Fourier Transform Infrared Spectroscopy with Attenuated Total Reflection (FTIR-ATR) analysis of polymer films was conducted on a Thermo Nicolet iS10 spectrometer equipped with a diamond ATR accessory using 64 scans. FTIR measurements were taken at room temperature under ambient atmosphere. Data were corrected using automatic baseline correction.
Raman Spectroscopy was conducted on cut film interior surfaces using a Horiba Jobin Yvon instrument with confocal microscopy and a 632 nm laser. Hole and slit sizes were both set to 500 microns, with 20 accumulations at 2 seconds each.
Optical microscopy images were taken on samples using a Keyence VHX-5000 optical (digital) microscope. Surface illumination was used unless otherwise specified.
PDMS sheeting used in masking procedures was obtained from B & J rubber products with mylar backing and a PDMS sheet thickness of 250 microns. Sheets were cut to the desired geometry using a Trotec Speedy 100 laser cutter. Acrylic panels and holders were laser cut using the same equipment.

Polymer Network Synthesis
In a representative procedure of network preparation, DMPA as photoinitiator, PEGMEA, PFPA, and PEGDA were mixed and sonicated for 2 minutes. Monomer resin was dispensed between two quartz plates, with film thickness controlled by steel spacers placed between the plates. Samples were irradiated on one side with 365 nm UV (3 mW/cm 2 power density) for 2 minutes to crosslink the networks. Samples containing PFPA were then immediately placed into a degassed methanol solution containing excess dopamine hydrochloride salt (3 equivalents relative to PFPA), and Et 3 N (1.5 equivalents relative to dopamine) was added to begin the catechol substitution. The reaction was gently stirred at room temperature under nitrogen sparging for 24 hours, and films were then placed in increasingly dilute HCl washes for 24 hours (0.1 M and 1 mM) to neutralize unreacted starting material and gradually adjust pH. Swollen films were stored in dilute HCl (1 mM).

Deswelling of Large Area Films
To prepare large area films for patterning and mechanical testing, a solvent deswelling procedure was developed. Films swollen in aqueous solutions frequently cracked during direct dehydration under vacuum, so a gentle solvent gradient was developed to slowly remove water from the networks. Water swollen films were placed into a solution with a 3:1 volumetric ratio of DI water and isopropanol (IPA). IPA was gradually added to achieve a 2:1, 1:1, 1:2, and finally 1:3 volumetric ratio of water to IPA with 30 minutes to 1 hour between solvent additions depending on film thickness. Films were then transferred to pure IPA, and hexanes was added stepwise to achieve the same volumetric ratios. Films were then removed from the mixed organic solvent solution and dried directly in vacuum.

Bulk Iron Treatment of Catechol Films
Catechol substituted films were removed from their 1mM HCl storage solutions and blotted dry to remove excess solution. Films were then immediately transferred to a solution of 0.05 M Fe(NO 3 ) 3 and 0.2 M Bicine buffered with KOH to pH 7.5. Solution volume was adjusted to be in significant excess (>5) of the iron content need to fully crosslink films. Unless otherwise noted, bulk samples were left in the iron treatment solution for 3 days. Samples were then removed, blotted dry, and placed into deionized (DI) water to remove residual salts. Films were dialyzed for 3 days with daily exchange of the DI solution, followed by the deswelling procedure described above.

Optical Microscopy Images of Fe 3+ -Catechol Films with High PEGDA Content
Optical microscopy images in the main text ( Fig. 6) were taken on cross-sectioned samples using a Keyence VHX-5000 optical (digital) microscope. Mixed transmission and surface illumination was used to highlight dark, opaque regions that were complexed with iron and more transparent regions that were relatively iron free. Samples were noted to have decreasing iron intrusion depths with increasing covalent crosslinker density at high catechol grafting densities.

Methods for Catechol Film Oxidation
To conduct control studies of intentionally oxidized films, a 0.05 M solution of NaIO 4 was prepared. Catechol containing films were removed from their 1 mM HCl dialysis, blotted dry, and placed in significant excess of the periodate solution (>5 excess mol. catechol groups). Samples were allowed to oxidize for 4 days, removed, and dialyzed in DI water for 24 hours before deswelling and drying. Note, caution should be exercised when handling strong oxidizers. Upon exposure to oxidizing conditions, films were noted to have an immediate color change from transparent to a light orange or yellow hue, deepening to the orange/red hues observed in Figure   S1.

Fe 3+ Patterning of Catechol Films
Patterning of iron-catechol domains could be conducted through a modified literature procedure from a solution-soaked filter paper 2 or directly from solution. For patterning from a soaked filter paper, catechol containing films were removed from a 1 mM HCl storage solution, placed against a glass slide and blotted dry of excess solution. The PDMS mask was then adhered to the film, completely covering the sample, and was then secured to a glass substrate to form an effective seal and prevent iron solution intrusion. The edges of the mask were then secured to the slide using Kapton tape to further prevent leakage. Next, a stack of 4 sheets of filter paper, soaked in the Fe(NO 3 ) 3 solution, was gently pressed against the masked catechol film. The entire assembly was then secured with a top glass slide and fixed in place with binder clips. The assembled patterning holder was then submerged in the iron solution for the desired patterning duration. Figure S2: Graphic depicting the components and setup for patterning catechol films from a soaked filter paper.
To pattern samples directly from solution, a similar PDMS masking approach was used, and films could be patterned from both sides. To pattern films from solution, a porous, rigid sample holder ( Figure S3A) was prepared from laser cut acrylic sheeting to allow the solution to reach the film surface. Two PDMS masks were laser cut for the front and back of the film, and a PDMS "spacer" of comparable thickness to the catechol film (c.a. 250 microns) was cut to act as a gasket to seal the film between the PDMS masks. To prepare the sample for patterning, one PDMS mask was adhered to the porous acrylic backing, then the PDMS spacer was placed on top of the mask forming a border around the edge of the sample holder ( Figure S3B). The catechol film was blotted dry and adhered to the mask on the sample holder, and the top mask was aligned and placed on the surface of the catechol film, sealing with the PDMS spacer below ( Figure S3C). The top porous acrylic holder was placed on the assembly ( Figure S3D), secured with binder clips, and submerged in iron solution for the desired time period. The sample could then be removed ( Figure S3E), dialyzed, and opened ( Figure S3F) to reveal the patterned sample ( Figure S3G).

Results
Rheological testing of Fe 3+ -catechol networks at elevated temperatures revealed minimal stress relaxation during the timescale of beam bending (nominally ~5s, Figure S4). These results indicate that during the timescale of beam bending testing, Fe 3+ -catechol crosslinks behave as quasistatic crosslinks rather than dynamic linkages. To further evaluate the amount of stress relaxation relative to the crosslinked plateau modulus of the iron free network, frequency sweeps at varied temperatures were conducted to generate master curves from the time temperature super positions and the rubbery plateau was identified from (Fig. S5). Even at extended timescales (1000s), samples do not relax to the value of the long timescale crosslinking plateau ( Figure S5).

Beam Bending Experimental Procedure
Beam bending was conducted on sectioned samples. Three-point beam bending was performed on a TA.XT Plus Connect Texture Analyzer with a 50 N load cell. Samples were cut into rectangular strips and deformed at a displacement rate of 0.1 mm/s to a turnaround force between 10 and 30 mN depending on sample stiffness. Beam spans were also altered between 15 and 25 mm to control sample stiffness. Summary data and an example plot showing these tests are shown in Figure S6 and Tables S2, S3, and S4. Moduli were calculated according to equation 1, using the sample stiffness , beam length , beam width , and beam height according to the bending equation ℎ for a center loaded beam with simply supported ends 5 .

SEM Experimental Procedures
Fe 3+ distribution through the film was further characterized by scanning electron microscopy (SEM) using a ThermoFisher Apreo C with accelerating voltages of 5 keV. Flat cross sections were prepared by cutting the samples with a blade after quenching in liquid nitrogen. The samples were sputter-coated with 2-3 nm of platinum before imaging. The chemical compositions across the film were analyzed by line scan of energy-dispersive X-ray spectroscopy (EDS).

SEM EDS Analysis of Iron Diffusion Timescale in Bulk and Patterned Films
Iron diffusion gradients could be designed bulk samples ( Figure S7) or in surface patterned films ( Figure S8) through both temporal and directional control. During iron treatment of bulk catechol films, significant gradients in iron content from the film exterior to interior could be designed through short exposure to the Fe 3+ solution, with increasing iron content until a homogenous film was achieved. Similarly, in a filter paper patterned sample, iron diffusion gradients could be designed to be directional from a single side of the film and intrusion into the film could be controlled through the duration of patterning.

SAXS & WAXS Experimental Methods & Supplementary Data
Small Angle X-ray Scattering Experimental Procedures A standard procedure for scattering was as follows. Silver behenate was loaded into a capillary tube and used as a calibration standard. X-ray scattering was conducted using a 1.54 angstrom Xray beam from a XENOCS Genix 50 W x-ray micro-source, focus size 50 microns, with a XENOCS FOX2D monochromator and Dectris EIGER R 1M detector. X-ray scattering was conducted on each sample for 30 minutes of exposure. Fitting and circular averaging was conducted in Igor pro using the Nika macro program 6 . During SAXS measurements, sections of film in water swollen and dry states were individually placed between protective layers of Kapton tape and affixed to the sample holder. For WAXS measurements, the same instrumentation was used, but no Kapton film was used to cover the material, and dry film sections were suspended in air to reduce noise from Kapton tape layers. All measurements, both SAXS and WAXS, were conducted in triplicate on separate batch of polymer network samples, with no significant deviations in trace morphology observed. Representative data is shown in the main text and in Figures S10, S11, S12, and S13.

SAXS & WAXS Supplementary Data
SAXS measurements of water swollen PFPA or Fe 3+ -catechol films at higher PFPA/catechol content reveal similar morphological features to films discussed in the main text. Notably for water swollen PFPA samples a significant peak is observed ( Figure S9C), corresponding the correlation lengths between PFPA rich regions. Moreover, similar to lower active ester and catechol content networks, upon substitution and Fe 3+ complexation in catechol films, a shoulder is observed at comparable length scale to the PFPA rich domains, indicating the formation of clusters ( Figure   S19D). Scattering of control, catechol networks without iron was alco conducted and is shown in Figure S10.
SAXS and WAXS was also conducted on dried films and demonstrated a decrease in feature size for Fe 3+ -catechol clusters (cf., shift of peak/shoulder features to the right in Figure S11D & S11F) and a lack of crystalline domains ( Figure S12)