Non-delaminating Polymer Hydrogel Coatings via C,H-Insertion Crosslinking (CHic) – A Case Study of Poly(oxanorbornenes)

Abstract

Robust, non-delaminating polymer coatings and hydrogels are needed for technical and biomedical applications. This study focusses on surface-attached poly(oxanorbornene) hydrogels obtained by simultaneous UV-activated crosslinking and surface-immobilization. The synthesis and copolymerization of two oxanorbornene monomers carrying the UV-crosslinkers malonic acid diazoester or benzophenone, which can both undergo UV-triggered C,H-insertion crosslinking (CHic), is presented. The crosslinking efficiency and network stability of hydrogels made from these self-crosslinkable polymers are studied and compared to the properties of poly(oxanorbornene) networks obtained by UV-triggered thiol-ene-reactions involving a low molecular weight crosslinker. Smooth, defect-free, non-delaminating hydrogel coatings were obtained by CHic, not only on laboratory model surfaces but also on a technical product.

Introduction

Robust adhesion of polymer coatings on substrates is important for many industries. Applications of such coatings include vehicle paints, glues and adhesives, corrosion protection, prevention of biofilm formation, and many others.^[1] In the medical field, functional polymer coatings may, for example, serve as a barrier layer for other materials (e.g. in polymer-coated stents), as a drug reservoir (e.g. in medical products that leach antimicrobials), as a protective layer that prevents biomolecule adhesion (e.g. poly(dimethylsiloxane) or poly(tetrafluoroethylene)-coated devices), or as a layer that ensures biocompatibility.^[2] When designing polymer coatings for medical applications, it is of particular importance that the coatings do not delaminate. Delamination would compromise the function of the device and/or the delaminated films could pose a threat to the patient. In the context of polymer hydrogels, such stability considerations need to include situations where the coatings are swollen in biological fluids where the swelling of the layer adds stress to the interface and strongly facilitates delamination. Therefore, covalently surface-attached and crosslinked polymer layers and hydrogels are important alternatives to merely physisorbed solution-processed or melt-processed polymer coatings.

Many surface-attached polymer architectures are known, including polymer brushes^{[2d, 3]} and surface-attached polymer networks.^[4] Both methods have their merits, but also their problems: while the synthesis of brushes is often complex and multi-step, the formation of surface-attached polymer networks is often based on crosslinking a polymer with added low molecular weight crosslinkers. The latter may lead to phase separation of the crosslinker and the polymer, and thus to coating defects. In the past years, an increasing amount of self-crosslinkable polymers that easily form surface-attached polymer networks have been reported.^[4b] These polymers contain in-built crosslinker groups that are covalently attached to a small fraction of their repeat units. These can either be specific groups that only react with defined other groups (e.g. the azide - alkyne pair, the thiol-ene pair, or the diene-dienophile pair), or groups like benzophenone or diazo-esters that can undergo C,H insertion crosslinking (CHic, Scheme 1).^{[4b, 5]} The advantage of molecules that can undergo CHic is that they can react with most C-H bonds of other molecules. Upon UV irradiation or thermal treatment,^{[4, 6]} they form a covalent bond between the polymer chain containing the CHic crosslinker, and the molecules or substrate containing the former C-H group.^{[4b, 6–7]} Advantageously, when polymers contain CHic crosslinkers, they can attach to substrates containing C-H bonds (i.e. to most polymer substrates) without the need for surface priming.^[8] Additionally, they can be attached to substrates containing oxide or OH functionalities after those have been primed with silanes that contain CHic agents, or alkyl groups.^{[6, 9]} Benzophenon is a well-established CHic molecule that has been used in many contexts,^[10] notably to crosslink polymers.^[8] Recently, Rühe and co-workers have reported CHic agents based on diazoesters.^{[7, 11]} In particular, they found a malonic acid diazo ester that can react both by UV irradiation and thermal treatment.^[7] This malonic acid diazo ester was attached to a methacrylate and could be copolymerized with other (meth)acrylic and styrenic monomers, so that self-crosslinkable polymers were obtained.^[7]

Scheme 1.

Poly(oxanorbornenes) can undergo UV-triggered crosslinking via thiol-ene reactions (P1) or C,H-insertion crosslinking reactions (CHic, P2 and P3). Polymer P2 contains the CHic agent benzophenone, polymer P3 contains the CHic agent malonic acid diazoester. In each system, the polymers were simultaneously crosslinked by their respective crosslinkable groups and surface-attached on a benzophenone-functionalized substrate.

In recent work on surface-attached polymer networks for biomedical applications, we have reported the crosslinking of poly(oxanorbornenes) by thiole-ene reactions of the respective polymers with a low molecular weight tetrathiol crosslinker (using an amine functionalized polymer similar to P1 in Scheme 1a).^[12] These networks were relatively smooth, yet contained coating defects due to phase separation of the polymer and the crosslinker during the coating process. As an alternative, we designed self-crosslinkable poly(alkenylnorbornenes) and poly(oxanorbornenes) that carried the covalently attached CHic crosslinker benzophenone (e.g. P2 in Scheme 1b).^[13] In thes studies, we found that poly(oxanorbornenes) underwent chain scission when irradiated at 254 nm. When UV-irradiated at 365 nm, high irradiation energy doses and BP contents were needed, and the gel content of the coatings remained low.^[13] We therefore adapted the diazoester concept^{[7, 11]} for poly(oxanorbornenes). These polymers are increasingly interesting for applications in biomedical contexts, as they can be easily functionalized with multiple substituents, and their molecular weights can be precisely adjusted.^[14]

Thus, in this report, we present the design and synthesis of a diazo malonic ester-substituted oxanorbornene, from which we synthesized bioactive, self-crosslinkable polymers (P3, Scheme 1c). We describe the copolymerization of this crosslinker with other oxanorbornenes, and study the UV-triggered network formation of the thus obtained polymers. These results are compared to UV-crosslinking of analogous benzophenone polymers P2 (Scheme 1b), and to surface-attached polymer networks made from polymer P1 and a low molecular weight tetrathiol crosslinker in a thiol-ene reaction (Scheme 1a).

Results and Discussion

The aim of this study was to compare the crosslinking efficiency of the diazo ester contained in P3 with the previously reported CHic agent benzophenone, and a low molecular weight thiol-based crosslinker (Scheme 1). To be able to do so, we first had to synthesize polymers P1 to P3 (Scheme 1). These were made from monomer M1 (Scheme 2a). P1 is the homopolymer of M1, which was to be crosslinked with pentaerythritoltetrakis(3-mercaptopropionate). The synthesis of M1 (Scheme 2a) has been reported previously^[15], however the literature procedures were difficult to reproduce and only gave low yields in our hands; we therefore significantly altered the synthesis and work up: M1 was obtained by reaction of the oxanorbornene anhydride R1 with the amine R2. Unfortunately, this reaction stage yielded not only the desired reaction product Z1, but a significant amount of its ring-opened analogue. To induce ring-closure to the desired imide, hexamethyldisilazane was added,^[16] which increased the isolated imide yield to 52%. The N-Boc protective group of the thus obtained product Z1 was near-quantitatively hydrolyzed with hydrochloric acid to give the corresponding ammonium salt Z2. In the third step, this product was deprotonated with triethylamine, and the amine group thus produced was guanidylated using reagent R3 with a yield of 60%. Even though the combined isolated yields of all three reaction steps was only about 28%, the fact that the work-up contained only extraction, filtration and recrystallization steps, and not a single purification by column chromatography, makes this procedure robust, reproducible and transferable to other laboratories. The synthesis of monomer M2 carrying the benzophenone CHic agent (Scheme 2b)^[13] also starts from 7-oxa-norbornene-2,3-dicarboxylic acid anhydride (R1), which was reacted with 2-bromo ethylamine hydrobromide (R3) to yield product Z3. Z3 was ring-closed to the corresponding imide Z4 in acidic medium.^[17] This imide was then reacted under etherification conditions with 4-hydroxybenzophenone to give monomer M2. For the synthesis of the diazo-malonic ester-substituted oxanorbornene M3 (Scheme 2c), the key component of this study, the oxanorbornene anhydride R1 was reacted with 2-aminoethanol (R6) to form N-(2-hydroxyethyl)-3,6-tetrahydrophthalimide (Z5). Z5 could be recrystallized from dichloromethane. This hydrodxy-substituted imide was then esterified with 3-chloro-3-oxapropanoic acid methylester to obtain the desired malonic acid derivate Z6 in good yields (81%) after column chromatography. In the last step, Z6 was reacted with 2-azido-1,3-dimethylimidazolinium hexafluorophosphate (R8), and the desired monomer M3 was obtained after extraction and further purification by column chromatography (yield 86%). The full characterization details of the intermediates and monomers are given in the Experimental Section.

Scheme 2. Synthesis of Monomers M1, M2, M3.

Reaction conditions: I: a) 0 °C, 90 min, MeOH; b) HMDS, 65 °C, 16 h. II: 2M HCl in dioxane:DCM (1:1). III: NEt₃, MeCN:H₂O (1:9), r.t.. IV: NaHCO₃, H₂O, r.t., 2h. V: NaAc, Ac₂O, 80°C, 15 min. VI: K₂CO₃, acetone, 56°C, 72h. VII: 65 °C, 16 h, MeOH. VIII: Py, DCM, 0 °C, 3 h; r.t., 16 h. IX: NEt₃, MeCN/THF, 0 °C, 1 h.

The target polymers P1, P2 and P3 were synthesized by ring-opening metathesis polymerization (ROMP) using Grubbs’ third generation catalyst. Reagent amounts for these syntheses are given in Table S1 in the Supporting Information. To obtain P1, monomer M1 was homopolymerized. P2 was obtained by copolymerization of M2 (carrying the benzophenone crosslinker) and M1. Three different copolymers with a benzophenone content of 5 mol%, 10 mol%, and 15 mol% (labelled as P2-5%, P2-10%, and P2-15% respectively) were synthesized. The repeat unit composition of the thus obtained polymers was confirmed by ¹H-NMR spectrometry (Table S1 and Appendix in the Supporting Information). Similarly, the different versions of P3 with a diazoester content of 5 to 15 mol% (labelled as P3-5%, P3-10%, and P3-15% respectively) were obtained by copolymerization of M3 with M1. Again, the repeat unit composition of the diazoester-bearing copolymers was checked with ¹H-NMR spectrometry (Table S1 and Appendix in the Supporting Information). The molar masses of all seven polymers (number average molecular mass M_n, weight average molar mass M_w and polydispersity index PDI) were determined by gel permeation chromatography and are summarized in Table 1.

Table 1. Molar masses (number average ${\overline{\text{M}}}_{\text{n}}$ and weight average ${\overline{\text{M}}}_{\text{w}}$) and polydispersity indices $(\text{PDI}={\overline{\text{M}}}_{\text{w}}/{\overline{\text{M}}}_{\text{n}})$ of the respective polymers P1-P3.

Polymer	${\overline{\text{M}}}_{\text{n}}$ / kg mol-1	${\overline{\text{M}}}_{\text{w}}$ / kg mol-1	PDI
P1	70	92	1.31
P2-5%	81	110	1.41
P2-10%	83	147	1.76
P2-15%	82	157	1.91
P3-5%	85	123	1.45
P3-10%	83	143	1.72
P3-15%	85	182	2.13

The network formation of the polymers was studied in two states: first, the protected polymers P1 to P3 were crosslinked; second, the crosslinking behaviour of the self-crosslinking polymers without the N-Boc protective groups (P3*) were studied. This comparison was important, because each Boc protective groups contains 9 C-H bonds and thus could be involved in CHic reactions.

However, removal of the Boc group after network formation would cleave these crosslinking points and thus lower the gel content of the network. The surface-attached polymer networks were studied on two kinds of substrates, quartz glass and silicon wafers, which were needed for different kinds of measurements. Each substrate had been pre-functionalized with the CHic agent benzophenone as described in the Experimental Section, so that covalent bonds between the polymer and the substrates formed when the samples were UV irradiated. To obtain the surface-attached polymer networks, each polymer / crosslinker system was spin-coated on a pre-treated silicon wafer or quartz glass, and UV irradiated. The thus obtained networks were then extracted with the appropriate solvent (dichloromethane or water, respectively), and dried. The network layer thickness before and after extraction was measured by ellipsometry. The ability of any polymer / crosslinker system to form a network can be quantified by studying its crosslinking efficiency. This is typically done by determining the gel content of the network under given irradiation conditions. (The gel content is the amount of crosslinked polymer that becomes part of the polymer network, while the non-attached polymer chains that can be removed by solvent extraction are called the sol.) The gel content was obtained from ellipsometry measurements (% gel content = ratio of layer thickness after extraction and layer thickness before extraction, multiplied by 100). When studying the crosslinking behaviour of the tetrathiol system (Figure 1a), preliminary data indicated that substantial amounts of tetrathiol crosslinker (35 mass%, corresponding to a thiol:double bond ratio of 200%) were needed to obtain sufficiently thick surface-attached polymer networks (Table S2 in the Supporting Information). For this thiol:double bond ratio, the gel content as a function of irradiation energy (at a wavelength of 254 nm) was studied. The results are shown in Table S3 and in Figure 1a. The data shows that no crosslinking occurred below an energy dose of 1 J cm^-2, and that the maximum gel content that could be achieved with this system was 60%. This is not satisfactory for many applications, and the fact that the thus obtained polymer layers phase-separated into polymer-rich domains and tetrathiol-rich domains makes this system even less attractive (Figure 1b, atomic force microscopy (AFM) height images indicate that the tetrathiol rich-domains were fully removed during washing and left holes in the network). Crosslinking of polymers P2 with 5 mol% benzophenone content at λ = 254 nm (Figure 1c) occurred at much lower energy does and gave higher gel contents than the P1/tetrathiol system, however the data shows that under these conditions, side reactions occur that lead to bond cleavage at energy doses > 0.5 J cm^-2.^[13] These side reactions are not found when the polymers P2 are crosslinked at λ = 365 nm (Figure 1d), and high gel contents were obtained for these conditions. However, the crosslinking of P2 at λ = 365 nm requires high energy doses (3-6 J cm^-2) until the plateau value of the gel content is reached. AFM images demonstrated that the coatings obtained using P2 carrying the CHic agent benzophenone (Figure 1e) were much more homogeneous than the corresponding P1/tetrathiol system, as no phase separation was observed. Finally, the crosslinking efficiency of the polymers P3 carrying the CHic agent diazoester (Figure 1f) was investigated. An irradiation wavelength of 254 nm was used due to its UV-absorption in this energy range (Figure S1). The data shows that this crosslinker gives networks with gel contents up to 92% (depending on the diazoester content) when using energy doses as low as 0.2 J cm^-2. Using FTIR spectrometry (Figure S2 in the Supporting Information), it could be demonstrated that the diazo group fully vanished upon UV irradiation with an energy dose of 0.2 J cm^-2. Thus, the diazo crosslinker is fully consumed already at that energy dose. In summary, the polymers P3 carrying the diazoester CHic agent had superior crosslinking efficiency. AFM images (Figure 1g) indicated that homogeneous coatings with no phase separation were obtained. Additionally, no tendency towards side reactions was observed, as the energy doses needed for crosslinking were very low.

Figure 1. Gel contents and typical morphology (atomic force microscope height images) of surface-attached polymer networks obtained by UV-induced crosslinking of polymers P1 to P3 (Scheme 1).

a) Gel content of P1 crosslinked with 200 mol% tetrathiol; b) AFM height image of network obtained in a); c) gel content of P2-5% irradiated at λ = 254 nm; d) gel content of P2-5%, P2-10% and P2-15% irradiated at λ = 365 nm; e) AFM height image of network obtained in d); f) gel content of P3-5%, P3-10% and P3-15% irradiated at λ = 254 nm; g) AFM height image of network obtained in f).

As mentioned above, C,H-insertion reactions between the Boc protective groups and CHic agents lead to cleavage of these crosslinks when the Boc group is removed. For the networks made from P2 and P3, this caused a coating thickness loss of up to 50%, depending on the crosslinker content (Figure S3). However, surface-attached polymer networks can be directly obtained by crosslinking the deprotected polymers P2* and P3*. This is shown in Figure 2 for the P3* series. The data shows that the gel contents of polymer P3*-5% remained lower than that of the protected polymer P3-5% in the irradiation energy range investigated. However, gel contents of up to 77% could be obtained when using polymers P3*-10% and P3*-15% (Figure 2a and Table S9 in the Supporting Information). This indicates that the 18 H atoms that were part of the Boc groups were significantly involved in the network formation of the P3 networks. This is plausible both from a statistical and a chemical perspective - first, they were many, and second they were sterically easily accessible thanks to their position in the polymer side chain. The deprotected polymers, on the other hand, had mainly C-H bonds near the backbone, which were more difficult to approach for the CHic agent. Thus, overall, the gel content of the P3* networks remained lower than that of the corresponding P3 networks. However, as indicated by the AFM image in Figure 2b, they were sufficiently high to obtain smooth, homogeneous surface coatings that fully covered the substrate. This was confirmed when the antimicrobial activity of the thus obtained polymer networks were tested. When sprayed with bacteria as described in the supporting information, the surfaces killed 100% of the applied bacteria within 2 hours (Figure 2c).

Figure 2.

Gel content (a.) and typical morphology (atomic force microscopy height image, b.) of surface-attached polymer networks obtained by UV-induced crosslinking of polymer P3*. c.) antimicrobial test results for surfaces coated with P3-15% and P3-15%*, as well as those surfaces after immersion in buffer (pH 7.4, 37 °C)

So far, the here described coatings were immobilized on laboratory surfaces (Si and glass, respectively) as this facilitated the analytical procedures. From an applications point of view, it is important that these polymer coatings can also be applied directly to polymeric substrates. This is illustrated in Figure 3. A PDMS tube was coated with a fluorescently labelled version of polymer P3*. The optical micrograph (Figure 3a) shows a cross-section of the coated tube after irradiation and subsequent rinsing to remove the non-attached polymer. The fluorescence image (Figure 3b) demonstrates the fluorescently coated inner surface of the tube. Figure 3c shows an overlay of the fluorescent image with the optical micrograph. This indicates that coatings fabricated from antimicrobial material can be covalently attached to other polymeric materials. This opens the possibility to coat medical devices like PDMS catheter tubing to reduce the risk of bacterial infections.

Figure 3.

a) Optical micrograph of a PDMS tube with covalently attached fluorescent polymer P3*. b) Fluorescence image of the red fluorescing coating. c) Overlay of the optical micrograph a) and the red channel of image b).

During the processing steps described in this paper, the polymer layers were exposed to different organic and aqueous solvents. Yet in all experiments reported, the layers remained firmly attached to the substrate. Thus, non-delaminating surface- attached polymer networks could be obtained using the diazoester-based CHic agent, both on “real life” technical materials and on laboratory surfaces.

Conclusions

We here reported the synthesis of robust, non-delaminating functional polymer hydrogel coatings that could be obtained both on laboratory and technical surfaces using a CHic-agent - a diazoester-based group that can undergo a C,H-insertion reaction with neighbouring C-H bonds upon UV irradiation. We presented the synthesis of the oxanorbornene monomer M3 carrying the CHic agent, and its copolymerization with functional monomers to obtain the poly(oxanorbornene)-based copolymers P3 and P3*. From these, surface-attached polymer networks were obtained that were much more homogeneous than analogous materials made by the addition of a low molecular weight crosslinker to the poly(oxanorbornene), and had a significantly higher gel content. This is particularly important when planning polymer multilayer systems, for example for the purpose of layer shedding from polymer multi-stacks.^[18] In such system, any layer inhomogeneity will have an impact both on the morphology of the consecutive layers and on the layer interfaces.

In a direct comparison of the polymer system P3 with the polymers P2, which carried the well-known UV-crosslinker benzophenone (another CHic agent), we could show that the diazoester CHic agent is more suitable for poly(oxanorbornenes) (and thus presumably also for poly(norbornenes)), because it can be used under reaction conditions where no side reactions occur, and because much lower irradiation doses are required. The latter is particularly important because the energy dose required to crosslink the benzophenone-based polymers P2 corresponds to an irradiation time of approximately 20 min, which is unfavourably long for continuous industrial productions. For the polymers P3, this time is reduced by a factor of 10. By the use of the functional monomers M1, the coatings obtained had marked antimicrobial activity. Thus, the combined properties of easy processability, applicability to technical systems, and bioactivity make diazoester-containing poly(oxanorbornenes) a promising substance class for biomedical applications, particularly as coatings for medical devices.

Experimental Section

Materials and Methods

The materials and methods used in the following experiments are described in the Supporting Information.

Monomer Syntheses

The synthesis of monomers M1 and M2 were reported previously^[13] and have been included in the Supporting Information.

Synthesis of Monomer M3

Step 1: Synthesis of Z5 (3a,4,7,7a-Tetrahydro-2-(2-hydroxyethyl)-4,7-epoxy-1H-isoindole-1,3(2H)-dione)

The synthesis was performed as reported in the literature.^[19] Exo-3,6-Epoxy-1,2,3,6-tetrahydrophthalic acid anhydride (R1, 8.3 g, 50 mmol, 1 eq.) and 2-aminoethanol (3.5 g, 50 mmol, 1 eq.) were dissolved in methanol and refluxed overnight. The mixture was allowed to cool to room temperature, and the solvent was removed by rotary evaporation. The residue was recrystallized from dichloromethane to yield pure Z5.

¹H NMR (250 MHz, acetone-d₆): δ / ppm = 6.57 (s, 2 H, 1,2), 5.12 (s, 2 H, 3,6), 3.79 (t, J=5.40 Hz, 1 H, 15), 3.65 - 3.39 (m, 4 H, 13,14), 2.91 (s, 2 H, 4,5); ¹³C NMR (62.86 MHz, acetone-d₆): δ / ppm = 177.3 (7,9), 137.4 (1,2), 81.8 (3,6), 59.3 (14), 48.4 (4,5), 41.8 (13); Yield (M = 209.20 g mol^-1): 4.6 g (22 mmol, 44%).

Step 2: Synthesis of Z6 (2-(3a,4,7,7a-Tetrahydro-1,3-dioxo-4,7-epoxy-2H-isoindol-2-yl)ethyl]-methyl propanedioic acid diester)

In a pre-dried Schlenk flask, Z5, the reaction product of step 1 (1.6 g, 7.6 mmol, 1 eq.), was dissolved in dry dichloromethane under nitrogen atmosphere. Dry pyridine (1.2 g, 15.2 mmol, 2 eq.) was added. The reaction mixture was cooled to 0°C using an ice-water-bath. Methyl-3-chloro-3-oxopropionat (R7, 1.15 g, 8.4 mmol, 1.1 eq.) dissolved in dry dichloromethane was added dropwise over three hours. The mixture was kept under nitrogen atmosphere and stirred overnight. It was then washed with 1M HCl (twice), deionized water (once) and aqueous saturated NaCl solution (once). The organic phase was dried with Na2SO4 and the solvent was removed by rotary evaporation. The crude product was purified by column chromatography in a hexane:ethylacetate mixture (1:3 by volume). The retention factor of the product (Z6) was 0.27.

¹H NMR (250 MHz, chloroform-d1): δ / ppm = 6.49 (s, 2 H, 1,2), 5.23 (s, 2 H, 3,6), 4.27 (t, J=5.37 Hz, 2 H, 14), 3.80 - 3.63 (m, 5 H, 13,22), 3.32 (s, 2 H, 17), 2.85 (s, 2 H, 4,5); ¹³C NMR (62.86 MHz, chloroform-d1): δ / ppm = 175.9 (7,9), 166.6 (16), 166.1 (18), 136.4 (1,2), 80.8 (3,6), 61.4 (14), 52.4 (22), 47.4 (4,5), 40.9 (17), 37.4 (13); APCI-MS: calculated for C14H15NO7 [M+H]: 310.09, found: 310.09; EA: calculated: C: 54.37 %, H: 4.89 %, N: 4.53 %; found: C: 54.45 %, H: 4.92 %, N: 4.49 %; Yield (M = 309.27 g mol^-1): 1.9 g (6.1 mmol, 81%).

Step 3: Synthesis of M3 (2-Diazo[2-(3a,4,7,7a-tetrahydro-1,3-dioxo-4,7-epoxy-2H-isoindol-2-yl)ethyl]-methyl propanedioic acid diester)

In a predried Schlenk flask, 2-azido-1,3-dimethylimidazolinium hexafluorophosphate (R8, 955 mg, 3.1 mmol, 1 eq.) was dissolved in dry acetonitrile under nitrogen atmosphere and cooled to 0°C. Triethylamine (607 mg, 6.2 mmol, 2 eq.) and Z6, the reaction product of step 2 (928 mg, 3.1 mmol, 1 eq.) were dissolved in dry tetrahydrofurane and added dropwise. The reaction mixture was stirred for one hour. It was then quenched by adding water, and the product M3 was extracted with dichloromethane. The organic phase was dried with Na2SO4 and the solvent was removed by rotary evaporation. The crude product was purified by column chromatography in a hexane:ethylacetate mixture (1:4 by volume). The retention factor of the product (M3) was 0.38.

¹H NMR (250 MHz, chloroform-d1): δ / ppm = 6.52 (s, 2 H, 1,2), 5.27 (s, 2 H, 3,6), 4.39 (t, J=5.40 Hz, 2 H, 13), 3.88 - 3.68 (m, 5 H, 14,22), 2.88 (s, 2 H, 4,5); ¹³C NMR (62.86 MHz, chloroform-d1): δ / ppm = 176.0 (7,9), 161.3 (16), 160.4 (18), 136.5 (1,2), 80.9 (3,6), 61.6 (14), 52.5 (22), 47.5 (4,5), 37.8 (13); APCI-MS: calculated for C14H13N3O7 [M+H]: 336.08, found: 336.08; EA: calculated: C: 50.15 %, H: 3.91 %, N: 12.53 %; found: C: 46.59 %, H: 3.73 %, N: 11.32 % Yield (M = 335.27 g mol^-1): 890 mg (2.7 mmol, 86%).

Polymer Syntheses

The general polymer synthesis procedure was the same for all polymers. First, a stock solution of Grubbs’ 3rd generation catalyst was prepared in a dry Schlenk flask under nitrogen atmosphere. For this, an appropriate amount of catalyst was dissolved in dry dichloromethane. The appropriate amount of monomer(s) was dissolved in another dry Schlenk tube in dry dichloromethane under nitrogen atmosphere. Both solutions were stirred under nitrogen for 30 min. An appropriate amount of the catalyst solution was removed with a syringe and added to the monomer solution in one shot. The reaction conversion was monitored by ¹H-NMR. The reaction was terminated by addition of ethylvinyl ether (1 mL), and the reaction mixture was stirred for another 30 min. The polymer was precipitated into hexane. A finely dispersed precipitate formed, which was recovered by filtration after 30 min and dried in high vacuum. The reagent amounts for each reaction are summarized in Table S1. The polymers were labelled according to the repeat unit content of crosslinker. For example, P2-5% refers to a polymer made from 95 mol% M1 and 5 mol% M2. The polymer structure was confirmed by ¹H-NMR; the molar masses were determined by gel permeation chromatography (GPC, in chloroform on SDV columns, mass determined against calibration with PMMA standards). The polymers P3* were obtained by dissolving the protected polymers (500 mg) in dry chloroform (2 mL). 4M HCl in 1,4-dioxane (2 mL) was added. The reaction mixture was stirred for 4 hours. The solvents were removed under reduced pressure and the polymer was dried in high vacuum.

Polymer Network Formation

Substrate preparation

Silicon wafers and quartz glass substrates were pre-functionalized with the CHic agent 4-(3-triethoxysilylpropoxy)-benzo-phenone (3EBP) as reported previously and described in the Supporting Information ^[20]

Surface-attached polymer networks - Formation of polymer layers from P1 to P3 by spin-coating

The protected polymers P2 to P3 were dissolved in dry chloroform (at 30 mg mL^-1). Using a syringe with a syringe filter (pore size 0.2 μm, h-PTFE membrane), the solution was applied to the non-moving substrate (either Si or glass, see above), which was then rotated at 3000 rpm (acceleration 1000 rpm/sec) for 10 sec. The substrates were then irradiated with a defined energy dose of UV light having the desired wavelength. After this, the layer thickness obtained (= gel + sol) was determined by ellipsometry. After extraction over night with dichloromethane, they were dried with compressed air. Again, the layer thickness (= gel) was determined by ellipsometry. To activate the polymers, the N-boc protecting group was removed by immersion of the substrates in 2M HCl in dioxane/DCM (1:1 mixture). In the case of P1, equal volumes of appropriately concentrated pentaerythritoltetrakis(3-mercaptopropionate) crosslinker solutions in chloroform were added to the polymer solution before spincoating resulting in a 15 mg mL^-1 polymer solution (see Table S2). Otherwise, the procedure was identical.

Formation of polymer layers from P3* by spin-coating

The deprotected polymers P3* (30 mg) were wetted with deionized water (0.1 mL). To this mixture, methanol (0.2 mL) was added. The mixture was vortexed until the polymer was fully dissolved. Then, methanol (0.7 mL) was added, thus giving a polymer solution of 30 mg mL^-1. With this solution, the substrates were coated as described for P2 and P3.

Scheme 3. Structures of Z5, Z6 and M3. Atom numbers relate to the NMR signal assignments.