Microscopic Structure of Swollen Hydrogels by Scanning Electron and Light Microscopies: Artifacts and Reality.

The exact knowledge of hydrogel microstructure, mainly its pore topology, is a key issue in hydrogel engineering. For visualization of the swollen hydrogels, the cryogenic or high vacuum scanning electron microscopies (cryo-SEM or HVSEM) are frequently used while the possibility of artifact-biased images is frequently underestimated. The major cause of artifacts is the formation of ice crystals upon freezing of the hydrated gel. Some porous hydrogels can be visualized with SEM without the danger of artifacts because the growing crystals are accommodated within already existing primary pores of the gel. In some non-porous hydrogels the secondary pores will also not be formed due to rigid network structure of gels that counteracts the crystal nucleation and growth. We have tested the limits of true reproduction of the hydrogel morphology imposed by the swelling degree and mechanical strength of gels by investigating a series of methacrylate hydrogels made by crosslinking polymerization of glycerol monomethacrylate and 2-hydroxyethyl methacrylate including their interpenetrating networks. The hydrogel morphology was studied using cryo-SEM, HVSEM, environmental scanning electron microscopy (ESEM), laser scanning confocal microscopy (LSCM) and classical wide-field light microscopy (LM). The cryo-SEM and HVSEM yielded artifact-free micrographs for limited range of non-porous hydrogels and for macroporous gels. A true non-porous structure was observed free of artifacts only for hydrogels exhibiting relatively low swelling and high elastic modulus above 0.5 MPa, whereas for highly swollen and/or mechanically weak hydrogels the cryo-SEM/HVSEM experiments resulted in secondary porosity. In this contribution we present several cases of severe artifact formation in PHEMA and PGMA hydrogels during their visualization by cryo-SEM and HVSEM. We also put forward empirical correlation between hydrogel morphological and mechanical parameters and the occurrence and intensity of artifacts.

other hand, after 10 minutes of polymerization, the temperature abruptly dropped down to 40-45°C, which was significantly below the glass transition temperature of the formed polymer 24,25 , the chains can loose their segmental mobility, and the vitrification sets in. In glassy state, some microbubbles can be trapped inside the prepared sample and this could explain the morphology found in the cry-SEM micrographs- (Fig. 3a). Thus, the additional heating to T = 120°C during the reaction was used to keep the forming polymer above its glass transition temperature (that was about 109°C determined by DSC) and thus facilitating the bubbles removal. However, the cryoSEM image obtained for the thus prepared sample for the swollen gel prepared by photopolymerization at T > Tg revealed porous structure as well.

b) Synthesis including freeze-pump-thaw-purge cycles.
The reaction mixture containing GMA monomer, crosslinker (DEDGMA or GDMA) at a level of 0.5 mol-% (in monomers) and 0.3 wt.-% (in monomers) of 2-methylpropiophenone (DAROCUR 1173) photoinitiator went through the three freeze-pump-thaw-purge cycles on the Schlenk line to remove dissolved gases. Each cycle included 5 min of freezing in dry ice/ethanol bath, 15 min of pumping, 15 min of thawing in a tepid water bath, and 15 min of purging with argon. Once the reaction mixture was purified and degassed, it was quickly transferred into the mold under the inert atmosphere and exposed to UV of 10 mW/cm 2 intensity during 90 min. After the photopolymerization, the gel was extensively washed with distilled water.
The reaction mixture was degassed using the Schlenk line prior to polymerization. The idea was as follows: mere purging of the reaction mixture with nitrogen to remove the dissolved oxygen could be accompanied by redissolution of nitrogen enhancing the bubbles formation.
However, degassing on the Schlenk line, due to additional freezing and thawing of the mixture under vacuum, pumps away the non-condensable gases. Efficiency of the freezepump-thaw-purge method is improved when each cycle is repeated several times. However, the PGMA samples prepared in this study via the photopolymerization after three purification cycles, still exhibited pores in the cryoSEM images (Fig. 3b). The monomers were used without purification (a) or were preliminary purified by distillation and degassed (b, c). UV irradiation in the case (a) was accompanied by additional heating (T = 120 °C).

c) Synthesis of hydrogels in the presence of water as diluent
The reaction mixture containing a monomer (HEMA or GMA), 40 vol.-% or 80 vol.-% of water, 1 mol-% (with respect to HEMA or GMA) of DEDGMA crosslinker, 0.75 wt.-% (with respect to both monomer) of APS initiator, and TEMED catalyst (5 μL of 5 wt.-% aqueous solution per 5 g of the mixture) was stirred for 5 min, bubbled with nitrogen, injected between two glass slides separated by a silicone rubber spacer and left overnight at room temperature for complete polymerization. The prepared gels were thoroughly washed with distilled water.
Finally, we changed the crosslinker type: GDMA was used instead of DEGDMA. The idea behind was as follows: GDMA contained an additional hydroxyl group instead of ether group, and hence was more hydrophilic and thus possibly more miscible with GMA monomer. Thus, we expected GDMA units uniformly distributed over the network and not causing assemby of nanodomains during the polymerization. On top of that, similarly to case 2 above, the reaction mixture containing purified GMA monomer, GDMA crosslinker, and DAROCUR 1173 initiator was degassed using freeze-pump-thaw-purge method. However, the improvement in the comonomers compatibility did not lead to the formation of nonporous gel as observed by cryoSEM (Fig. 3c).

Synthesis of IPN hydrogels
Macroporous PHEMA network (Network I) was synthesized by redox-initiated radical polymerization in an aqueous solution containing 80 vol.-% of water, 1 mol-% (with respect to HEMA) of DEGDMA crosslinker, 0.75 wt.-% (with respect to both monomers) of APS initiator and TEMED catalyst (5 μL of 5 wt.-% aqueous solution per 5 g of the mixture).
Polymerization reaction was run overnight at room temperature. The unreacted monomers were removed from the hydrogel by washing. The water-swollen PHEMA gel was immersed into the reaction mixture for the second network (Network II) preparation, containing the monomer (HEMA or GMA), 0.3 mol-% (with respect to HEMA or GMA) of DEGDMA crosslinker and 0.5 wt.-% (with respect to both monomers) of DAROCUR 1173 initiator.
Water in the hydrogel was gradually replaced with the monomers by exchanging the reaction mixture with fresh portion several times after equilibration. The absence of water in the system was confirmed by ATR FTIR method and measuring the refractive index of the liquid.
Once the equilibrium swelling was achieved, the solid PHEMA network swollen in the reaction mixture was irradiated by UV with intensity of 10 mW/cm 2 during 90 min and then thoroughly washed with distilled water to remove the unreacted monomers.

Preparation of fluorescently-labelled hydrogels
A single-network PGMA hydrogel was labelled by incorporating the vinyl-modified fluorescein added directly to the monomer reaction mixture (approx. 0.001 wt.-%). The synthetic procedure of fluorescein modification is described below in Supplementary data. In the case of IPNs, each network was labelled using the modified dyes with different excitation wavelengths in order to distinguish between the two networks. The first (PHEMA) network was labelled with the modified fluorescein, while the second (PGMA) one was labelled with the modified dye DY-677. Concentration of the dyes in both cases was 0.001 wt.-%. After the polymerization of each network, the excess of unreacted dye and monomers was removed by extensive washing of the gel with distilled water until the washings became visually colorless.
The reaction mixture was evaporated under vacuum and purified on a silica gel column with acetone-H2O = 9:1 as the mobile phase. After the purification, the blue-colored product was lyophilized for storage.

Methacryloylation of fluorescein
Fluorescein isothiocyanate ester (0.5 g, 1.3 mmol) with excitation spectrum peak at 495 nm was dissolved in 100 mL of anhydrous tetrahydrofuran, and 0.215 mL (1.5 mmol) of triethylamine was added to the solution. The mixture was cooled to 0°C, and 0.146 mL (1.5 mmol) of methacryloyl chloride in 5 mL of tetrahydrofuran was added dropwise. The mixture temperature was kept around 0°C during the mixing, and then the addition reaction was carried out at r.t. for 24 h. Tetrahydrofuran was evaporated off, and the product was purified by column chromatography with CHCl3-acetone = 9:1 as the mobile phase.
It is interesting to notice that the size of the pores formed H0/1-G0/0.3 sample was the same as for the single G0/0.3 hydrogel (e.g., Fig. 1). Hence, the presence of first PHEMA network virtually did not restrict formation of the porous structure, yet affecting other properties like overall swelling (cf. EWCG0/0.3 = 2.8 g/g vs. EWCH0/1-G0/0.3 = 1.1 g/g).