Ion‑Mediated Structural Engineering of Hydrogel Interfaces for Tunable Mechanical and Analyte Diffusion Properties in Electrochemical Biosensors

Background

Hydrogel layers on wearable and implantable electrochemical sensors reduce the mechanical mismatch with tissue and control how analytes reach the electrode. In conventional preparation by chemical crosslinking or sacrificial templating, a denser network improves strength but slows diffusion, and a more porous network speeds diffusion but weakens the layer. Sweat sensors need fast transport of low-concentration analytes under deformation, whereas implanted sensors need restricted transport to avoid saturation at high concentrations, together with insertion strength and resistance to fouling.

Design and approach

Freeze–thaw PAH was immersed in Na₂SO₄, CaCl₂ or both in sequence. Sulfate, a kosmotropic anion, salts out PVA chains and drives phase separation that leaves an interconnected hierarchical pore network while promoting PVA crystallization. Calcium forms egg-box crosslinks with alginate. When calcium is applied first and sulfate second, the calcium–alginate nodes act as nucleation sites for uniform CaSO₄ crystal growth inside the network, producing a dense structure; the reverse order forms pores first and confines CaSO₄ mainly to surfaces. The resulting hydrogels (ISC–PAH) were used as the outer diffusion-control layer over glucose oxidase and a conductive catalytic hydrogel containing nickel-doped Prussian Blue and PEDOT:PSS, so that the layer controls the glucose flux and hence the apparent Michaelis–Menten constant.

How it was evaluated

Four variants (Na₂SO₄–PAH, CaCl₂–PAH, Na₂SO₄/CaCl₂–PAH and CaCl₂/Na₂SO₄–PAH) and pristine PAH were characterized by SEM, porosimetry, FT-IR, XPS, XRD, Raman spectroscopy, contact angle, FRAP with 4, 20 and 40 kDa FITC–dextran, and steady-state currents at Au microelectrodes. Tensile, compressive, cyclic, rheological and swelling tests (artificial sweat and interstitial fluid) and NIH-3T3 viability followed. Glucose sensing was compared on screen-printed carbon electrodes, then in a stretchable patch tested on four healthy participants during cycling with a sugar drink, and in a Pt-wire implantable probe tested by repeated insertion through phantom skin in BSA solution and in four anesthetized mice receiving glucose and insulin injections.

Key results

Average pore size was 1.08 µm for pristine PAH, about 2.50 µm after Na₂SO₄, 1.85 µm after CaCl₂ and about 2.23 µm for Na₂SO₄ followed by CaCl₂, whereas CaCl₂ followed by Na₂SO₄ produced a dense structure with 65 nm pores containing rod-shaped CaSO₄ crystals. In FRAP with 4 kDa FITC–dextran, relative diffusivity was about 7.1 ± 1.4% for Na₂SO₄–PAH and 1.26% ± 0.09% for CaCl₂/Na₂SO₄–PAH, and Na₂SO₄–PAH was the only variant without a diffusion barrier for 40 kDa dextran. Diffusion coefficients from microelectrode currents followed the same order as pore size.

All ion-treated variants were mechanically stronger than pristine PAH. Na₂SO₄–PAH reached a tensile strength of about 237 kPa and an elongation of about 347%, and its moduli together with those of the two sequential variants were 134–143 kPa, compared with 54 ± 6 kPa for pristine PAH. Compressive strength at 90% strain was about 5.17, 3.32 and 3.21 MPa for Na₂SO₄–PAH, CaCl₂/Na₂SO₄–PAH and Na₂SO₄/CaCl₂–PAH, against less than 1 MPa for pristine PAH. The two sequentially treated hydrogels swelled less than 3% over 48 h in artificial interstitial fluid at 37 °C, and fibroblast viability exceeded 90% for all variants.

On electrodes with identical enzyme and catalytic layers, the porous Na₂SO₄–PAH layer gave the highest sensitivity (1.813 µA mM⁻¹), a detection limit of 15.4 µM and a linear range of 0.1–1 mM, whereas the dense CaCl₂/Na₂SO₄–PAH layer gave 0.077 µA mM⁻¹ and extended the linear range to 2–30 mM. The apparent Michaelis–Menten constant rose from 1.49 to 18.7 mM accordingly. The stretchable patch with Na₂SO₄–PAH remained stable over 50 cycles of 10% strain; in four healthy participants, sweat glucose followed blood glucose with a lag of 9–12 min, and all 54 paired points fell in Zone A of the Parkes error grid.

The implantable probe with CaCl₂/Na₂SO₄–PAH was linear from 2 to 30 mM (R² = 0.997), kept its response after three insertions through phantom skin in BSA solution while uncoated and pristine-PAH probes lost signal, and showed coefficients of variation below 3% during glucose cycling between 0 and 20 mM. In four mice, interstitial glucose followed blood glucose with a lag of 9–14 min and an interstitial-to-blood ratio of 0.9531, and 98.6% of 74 paired values fell in Zone A.

Open questions

Fluorescence imaging showed that the CaCl₂/Na₂SO₄–PAH surface began to adsorb some protein after 7 days in protein-containing medium, although diffusivity and signal were maintained for 14 days. The authors plan in situ studies of the ion-sequence-dependent structural dynamics and longer stability evaluations under physiological and in vivo conditions.

How it connects to other research

Fully integrated multiplexed sweat sensor arrays address temperature calibration and on-board signal processing, whereas this study adjusts mass transport and mechanics in the hydrogel layer that analytes cross before reaching the enzyme electrode.