Conducting Polymer Nanoparticles with Intrinsic Aqueous Dispersibility for Conductive Hydrogels

Abstract

Conductive hydrogels are promising materials with mixed ionic-electronic conduction to interface living tissue (ionic signal transmission) with medical devices (electronic signal transmission). The hydrogel form factor also uniquely bridges the wet/soft biological environment with the dry/hard environment of electronics. The synthesis of hydrogels for bioelectronics requires scalable, biocompatible fillers with high electronic conductivity and compatibility with common aqueous hydrogel formulations/resins. Despite significant advances in the processing of carbon nanomaterials, fillers that satisfy all these requirements are lacking. Herein, intrinsically dispersible acid-crystalized PEDOT:PSS nanoparticles (ncrys-PEDOT_X) are reported which are processed through a facile and scalable nonsolvent induced phase separation method from commercial PEDOT:PSS without complex instrumentation. The particles feature conductivities of up to 410 S cm⁻¹, and when compared to other common conductive fillers, display remarkable dispersibility, enabling homogeneous incorporation at relatively high loadings within diverse aqueous biomaterial solutions without additives or surfactants. The aqueous dispersibility of the ncrys-PEDOT_X particles also allows simple incorporation into resins designed for microstereolithography without sonication or surfactant optimization; complex biomedical structures with fine features (< 150 μm) are printed with up to 10% particle loading . The ncrys-PEDOT_X particles overcome the challenges of traditional conductive fillers, providing a scalable, biocompatible, plug-and-play platform for soft organic bioelectronic materials.

1. Introduction

Bioelectricity is vital to many physiological processes in living organisms, such as the contraction of muscles, transmission of nerve impulses, regulation of heartbeats, and tissue regeneration, among others. The desire to implement bioelectricity to monitor or influence biological events has led to the emergence of bioelectronics, a discipline which aims to merge biology and microelectronics–typically through implantable or wearable devices.^[1] These medical technologies require a high-quality interface between tissue and electronics that enables both ionic and electronic communication. The first-generation of bioelectronic devices relied on inorganic interfaces which pose several disadvantages including 1) a strain mismatch which can lead to cell death and fibrosis at the tissue-device interface, 2) poor compatibility with magnetic fields used during magnetic resonance imaging (MRI), 3) challenges associated with translating both ionic and electronic signals, and 4) limited biodegradability. To overcome these challenges, researchers have been developing soft organic materials with mixed ionic and electronic conduction that can seamlessly integrate with cells and living tissues.^[1–4] Conductive hydrogels have gained significant attention as their structure mimics the extracellular matrix, the natural microenvironment of the cells—as such, conductive hydrogels have been aggressively explored for tissue engineering and regenerative medicine (TERM). The biological basis of such TERM applications is the facilitated transmission of endogenous or exogenous bioelectricity throughout the hydrogel, requiring charge percolation throughout the entire material.

Conductive hydrogels consist of hydrophilic polymeric networks with incorporated electroactive fillers, the most common of which include carbon nanotubes, graphene nanomaterials, conjugated polymers, and inorganics (gold, silver, metal oxides, etc.).^[5] Rationally designing hydrogels to promote homogeneous, 3D charge percolation for TERM and other bioelectronic applications remains a challenge. A significant limitation is the surface energy mismatch between the hydrophilic hydrogel precursor and the hydrophobic, typically carbon-based nanomaterial filler; this incompatibility requires laborious, multistep procedures to achieve sufficient loading for charge percolation. Of the available conductive fillers, conjugated polymers (CPs) have distinct advantages due to their enhanced chemical tunability, flexibility, biocompatibility, and solution processability.^[6] To achieve the necessary intrachain and interchain charge transport across long length scales for sufficient conductivity in thin-films, there are many processing methods such as spin-coating, blade-coating, or other facile techniques to obtain favorable packing with minimal polymer.^{[7, 8]} However, as these processing methods cannot be translated to the tortuous, porous form factor of a hydrogel, achieving this same favorable packing is challenging, and has been traditionally limited to incorporating high loadings of CP filler.^[9]

Alternatively, CPs can be preorganized into colloidal dispersions which are stabilized by polyelectrolytes surfactants—enabling both solution processibility and high electrical conductivity. The most prominent CP dispersion is poly(3,4-ethylenedioxythiophene) doped with poly(styrenesulfonate) (PEDOT:PSS) which can feature conductivities beyond 4000 S cm⁻¹ when processed appropriately,^[10] leading to its widespread adoption within thin film applications such as chemical sensing, energy storage, and antistatic coatings. The commercial presence and biocompatibility of PEDOT:PSS has also inspired its use within conductive hydrogels for biomedical applications;^[11] however, as the ink is designed for 2D coatings, its direct incorporation within 3D hydrogels typically affords low conductivities between 10⁻³ and 10⁻¹ S cm⁻¹ (Table S2, Supporting Information). While several groups have formulated conductive hydrogels directly out of PEDOT:PSS dispersions,^[12–14] many applications require the chemical and mechanical features of commonplace biomaterials to facilitate biocompatibility and biological outcomes such as cellular differentiation,^[2] necessitating a method to endow natural and synthetic biomaterials with conductivity. Here, we report intrinsically dispersible acid-crystalized PEDOT:PSS nanoparticles (ncrys-PEDOT_X) to enable 3D charge percolation within hydrogels for organic bioelectronic applications. Using an acid-based nonsolvent induced phase separation (NIPS) method, commercial PEDOT:PSS inks were coagulated into stable aggregates of concentrated PEDOT with tuneable PSS surfactant to enable dispersion within prototypical hydrogels. Compared to other common conductive fillers ncrys-PEDOT_X displays enhanced dispersibility, enabling homogeneous incorporation at relatively high loadings within diverse aqueous biomaterial solutions without additives or surfactants. To demonstrate the plug and play nature of the particles ncrys-PEDOT₂₀ was directly incorporated within both natural biomaterial derived hydrogels and an aqueous resin designed for projection microstereolithography without sonication or surfactant optimization. With only a single step for conductive filler incorporation, hydrogels were endowed with significant conductivity (> 1 S cm⁻¹) rivaling other reported systems in performance and simplicity. Furthermore, high fidelity complex biomedical structures were printed with up to 10% loading of conductive particles with fine features (<150 μm), biocompatibility, and hemocompatibility. The ncrys-PEDOT_X particles overcome the challenges of traditional conductive fillers, offering a scalable, conductive, biocompatible, and modular platform for soft organic bioelectronic materials.

2. Results and Discussion

2.1. Dispersible and Tunable ncrys-PEDOT_X: Preparation and Characterization

The synthesis of PEDOT nanoparticles is typically achieved through bottom-up, template-, or emulsion-based methods which suffer from limited tunability, low yields, and particles that lack aqueous processability—the latter of which is a critical requirement when working with biomaterial dispersions. While reactors have been reported to afford submicron PEDOT particles via aerosol vapor polymerization and electrochemistry,^[15–17] such methods require sophisticated instrumentation and, as the afforded particles lack a polyanionic surfactant such as PSS, have both limited aqueous dispersibility and require external dopants (Table S1, Supporting Information). Alternatively, commercial PEDOT:PSS can be concentrated via lyophilization and redispersed within a biomaterial;^[18–20] while this top-down approach is scalable and accessible due to the commercial availability of PEDOT:PSS, the afforded composites demonstrate limited conductivity as well as leeching over time due to the excess insulating PSS dispersant. Recently an acid-based nonsolvent induced phase separation (NIPS) approach was applied to coagulate PEDOT:PSS into stable, microparticles to both partially remove insulating PSS and enhance the crystallinity of the PEDOT assemblies—however this method afforded porous systems with size features too large for dispersion within conductive hydrogels.^[21] Acid treatment has also been recently implemented to improve and investigate mixed conduction within thin film devices.^{[22, 23]} In this work, we leverage a top-down NIPS-based processing approach to commercial PEDOT:PSS (Heraeus Electronic Materials, 1.1—1.3% solid content, Clevios PH1000) to prepare tuneable PEDOT:PSS nanoparticles with intrinsic dispersibility for conductive hydrogel applications (Figure 1a). The reported method was designed to be simple for adoption by nonchemists and affords particles on the multigram scale within an academic laboratory setting (Figure 1b).

Figure 1.

Design of dispersible nanoparticles of crys-PEDOT_X from commercial PEDOT:PSS. Commercial PEDOT:PSS solution a) can be converted into dispersible acid-crystallized nanoparticles (ncrys-PEDOT_X) by b) nonsolvent induced phase separation (NIPS). c) SEM image of particles coagulated in an acid bath with 20% sulfuric acid (ncrys-PEDOT₂₀) with a corresponding histogram displaying particle dimension statistics. d) S 2p high resolution XPS spectra including fits for PEDOT (orange) and PSS (gray). PEDOT/PSS ratio was calculated from an area comparison of the S 2p fits. e) X-ray diffraction (XRD) patterns of ncrys-PEDOT₅ and ncrys-PEDOT₂₀, focused on the (010) reflections corresponding to π–π stacking between adjacent PEDOT chains. f) Current–voltage curves of pelletized samples of ncrys-PEDOT_X particles are characterized by ohmic behavior; the inset shows a representative disc-shaped ncrys-PEDOT₂₀ pellet. g) Conductivity comparison of ncrys-PEDOT_X with other reported conjugated polymer nanoparticles from Table S1 (Supporting Information).

Commercial PEDOT:PSS (1—1.3 wt%) was added dropwise into a coagulation solution consisting of both isopropanol (IPA) and sulfuric acid. Collected particles were subsequently comminuted into a fine powder using a cryogenic grinder to afford nanoparticles with an average diameter of 81.5 nm (Figure 1c). As a nonsolvent, IPA dehydrates the colloidal dispersion, enabling the NIPS phenomenon. The autoprotolysis of sulfuric acid is commonly used to stabilize the segregation of insulating PSS from the conductive PEDOT; therefore, excess PSS is selectively removed the ionomer complex, while PEDOT forms crystalline domains via π–π stacking.^[24] The degree to which PSS is lost and PEDOT crystallizes within the particles is directly related to the concentration of sulfuric acid within the coagulation bath. XPS was used to investigate the impact of the coagulation bath composition on PSS extraction. All spectra were referenced to the C 1s level set to 284.8 eV—analysis of the S 2p spectral region reveals the presence of PSS and PEDOT at binding energies of ≈168 and ≈164 eV, respectively (Figure 1d).^{[25, 26]} A comparison of the relative intensities of these spectral components demonstrates that increasing concentrations of the acid within the coagulation bath correlates with increased PEDOT/PSS ratios; this is also demonstrated via UV/Vis analysis (Figure S1, Supporting Information). The XRD spectrum of the dispersible ncrys-PEDOT₅ and ncrys-PEDOT₂₀ particles represent conditions with low and high concentrations of sulfuric acid in the coagulation bath, respectively. The relative decrease in the intensity of the amorphous halo of PSS (≈2θ = 16–17) further supports the loss of PSS at greater sulfuric acid concentrations. Furthermore, at the elevated acid concentration the d₍₀₁₀₎ at ≈2θ = 25 shifts to a higher angle (Figure 1e)—this reflection is broad and diffuse in pristine PEDOT:PSS.^[27] This reduced d₍₀₁₀₎ spacing demonstrates a decrease in interchain π–π stacking distance of PEDOT, suggesting crystallization enhancement when coagulated with increasing sulfuric acid concentration, consistent with similar investigations of acid-crystallized PEDOT:PSS thin films.^[27] The experimental peaks were fitted to a Gaussian function and the full-width at half-maximum (FWHM) was obtained; the crystallite size was calculated using the Debye–Scherrer equation with a K-constant of 0.9. The calculations revealed that the crystallite size of the particles increases from 1.26 to 1.58 nm upon treatment with more concentrated sulfuric acid.

PEDOT:PSS films processed via acid crystallization typically display enhancements in conductivity—to investigate the conductivity of ncrys-PEDOT_X, disc-shaped pellets were formed by applying hydrostatic pressure to the particles and were subsequently analyzed via four-point probe measurements (Figure 1f, see the Experimental Section), a typical approach to quantify conductivity of conjugated polymer-based particles.^[15] As a control, we also assessed the conductivity of lyophilized PEDOT:PSS as it is commonly used as a conductive filler for hydrogels and other composites.^[18–20] Both ncrys-PEDOT₅ and ncrys-PEDOT₂₀ demonstrated remarkable conductivities of 87 and 410 S cm⁻¹, values significantly larger than lyophilized PEDOT:PSS (4 S cm⁻¹). To the best of our knowledge, this is the highest reported electrical conductivity for a solid-state conducting polymer powder (Figure 1g; and Table S1, Supporting Information).^{[15, 28–32]} While the exact mechanism of conductivity enhancement from acid crystallization is currently under investigation, it is generally accepted that the phenomena involves the removal of PSS and enhanced crystallinity;^{[10, 24]} which is consistent with the XRD, UV/Vis, XPS, and conductivity analysis of ncrys-PEDOT_X.

2.2. Conductive ncrys-PEDOT_X Incorporated Hydrogels

Conductive hydrogels are typically synthesized directly from aqueous solutions of hydrophilic natural biopolymers (alginate, gelatin, collagen, chitosan) or synthetic polymer formulations (pHEMA, PEO, PEGDA).^{[5, 6, 33]} Hydrophobic carbon-based nanomaterials such as carbon nanotubes (CNTs) and graphene (G) are common fillers for conductive hydrogels but lack stability in water, requiring complex surfactants and/or harsh oxidation to disperse,^{[34, 35]} thus complicating their application within conductive hydrogels. The residual PSS within ncrys-PEDOT_X can act as an internal surfactant to stabilize their dispersion within aqueous biomaterial solutions, overcoming this limitation. The tunability with which this internal surfactant remains within the particle can therefore act as a handle to promote aqueous dispersibility. The kinetic stability of ncrys-PEDOT_X was investigated by monitoring the absorbance of 1 wt% (10 mg mL⁻¹) ncrys-PEDOT_X aqueous dispersions over time (see the Experimental Section)—a common method to profile the sedimentation kinetics of nanomaterial dispersions.^{[36, 37]} Both ncrys-PEDOT₅ and ncrys-PEDOT₂₀ were first dispersed via stirring in either DI H₂O or 2% aqueous sodium alginate, the latter of which is a common precursor for conductive biomaterials (Figure S2, Supporting Information);^[38–40] such high filler concentrations are similar to those used with graphene oxide, and are orders of magnitude more concentrated than what can be achieved with untreated CNTs or G. For example, graphene nanoplatelets can only be dispersed in water without oxidation or complex surfactant optimization at concentrations less than 0.01 mg mL⁻¹.^[41] In both DI and aqueous sodium alginate solutions the sedimentation of ncrys-PEDOT₅ and ncrys-PEDOT₂₀ were similar, however ncrys-PEDOT₅ had a relatively slower rate, potentially due to the elevated amount of PSS internal surfactant. As ncrys-PEDOT₂₀ displays significantly enhanced conductivity, but with a similar aqueous sedimentation profile to ncrys-PEDOT₅, ncrys-PEDOT₂₀ was further explored in aqueous dispersions of chitosan and collagen due to their widespread use within the synthesis and design of hydrogels for biomedical applications.^[33] The particles had remarkable kinetic stability in each of the aqueous biomaterial-based hydrogel precursors—particularly considering sonication and external surfactants were not utilized (Figure 2a; and Figure S3, Supporting Information). Importantly the sedimentation rate is significantly slower than the rate of hydrogel crosslinking (≈0.5–1 h), allowing these particles to be directly added to aqueous hydrogel precursor solutions for conductive hydrogel formulations.

Figure 2.

Dispersibility of ncrys-PEDOT₂₀ within conductive hydrogels. a) Sedimentation kinetics of 10 mg mL⁻¹ ncrys-PEDOT₂₀ in DI H₂O, 2% aqueous sodium alginate, and 2% aqueous chitosan solutions were monitored by the change in absorbance at the meniscus of the particle-incorporated solutions. b) Conductive hydrogels prepared by simply adding the ncrys-PEDOT₂₀ to aqueous biomaterial dispersions, stirring the mixture, and curing the hydrogel via thermal, covalent, ionic, or photoinitiated crosslinking methodologies. Images display the particle loaded dispersions before (left) and after (right) crosslinking. For the particle loaded alginate dispersion, crosslinking was demonstrated by the “alginate-worm experiment” within a solution of 1 m CaCl₂.

We then investigated if ncrys-PEDOT₂₀ impacted the prototypical chemistries that are utilized to crosslink biomaterials into hydrogels. Many natural aqueous biomaterial dispersions are crosslinked thermally, covalently, ionically, or via pH. Representative biomaterial dispersions of polysaccharides (agarose, chitosan), protein (gelatin), and synthetic resin (PEGDA) were loaded with 1 wt% (10 mg mL⁻¹) of ncrys-PEDOT₂₀, stirred and subsequently crosslinked (Figure 2b, see the Experimental Section). Each dispersion successfully gelled demonstrating that the particles do not interfere with hydrogel-bonding networks (gelatin), Schiff base chemistry (chitosan), ionic crosslinking (alginate), or radical photopolymerization (PEGDA). Furthermore, each hydrogel crosslinked rapidly and demonstrated no evidence of sedimentation—this suggests compatibility of ncrys-PEDOT₂₀ with the common modalities used throughout biomaterial/hydrogel chemistry. Analysis of the swelling and conductivity of the alginate hydrogel with 1 wt% ncrys-PEDOT₂₀ suggests the particles do not have a deleterious impact on hydrogel stability (Figure S4, Supporting Information), however such affects are anticipated to be highly dependent on the hydrogel composition and are a topic for future investigation. As the loading of conductive fillers can increase the modulus (i.e., stiffness) of the hydrogel, we screened several hydrogels to benchmark the modulus reinforcement of ncrys-PEDOT₂₀ compared to other common fillers. The modulus reinforcement is dependent on polymer–particle interactions, therefore the impact ncrys-PEDOT_X has on a hydrogel will be dependent on several factors (polymer structure, polymer loading, crosslinker structure, crosslinker loading, additives, type of ncrys-PEDOT_X, etc.), as such it is challenging to make generalizable claims. As a representative example, mechanical analysis (DMA) of the alginate hydrogel demonstrates a slight decrease in the Young’s modulus of the hydrogel with ncrys-PEDOT₂₀ incorporation—however both hydrogels feature similar stiffness (i.e., soft tissue) to the prior report of the alginate control (Figure S5, Supporting Information).^[38] The slight decrease in the modulus of the alginate hydrogel may suggest the particles reduce the crosslink density as the ionic nature of the calcium crosslinking is weak. This is supported by the slight increase in the modulus of the chitosan hydrogels with particle incorporation—these hydrogels are covalently crosslinked. At 1 wt% loading of ncrys-PEDOT₂₀, each of the tested gels demonstrates substantially less modulus reinforcement than what is seen on average with CNTs (130%) or reduced graphene oxide (170%).^[5] Recent work with hydrophilic CPs has suggested that hygroscopic side chains can reduce modulus reinforcement;^{[42, 43]} the internal PSS surfactant may behave similarly, preventing detrimental stiffening upon loading. As the impact of ncrys-PEDOT_X on stiffness is hydrogel dependent, future application of the particles should also investigate reinforcement behavior.

For bioelectronic applications conductive hydrogels require sufficient electronic conductivity to facilitate the transmission of endogenous or exogenous bioelectricity throughout the material, requiring 3D charge percolation. Prior experiments have suggested that as the loading of conducting filler is increased, a dramatic enhancement in conductivity is observed which is attributed to the formation of percolation pathways for charge carriers.^[44–46] The filler volume required to achieve percolation is related to the aspect ratio; for fillers of roughly spherical shape, similar size, and random orientation, a loading of 16% has been found—known as the Sher–Zallen invariant.^[47] While 1D and 2D materials such as CNTs or G would require reduced loadings to achieve percolation, these materials have limited aqueous dispersibility and the relatively high surface area of their form factor compared to spherical particles promotes aggregation. As high loadings of conductive filler are typically incompatible with hydrophilic hydrogel precursors, this challenge has been circumvented by processing hydrogels directly out of conjugated polymers—typically PEDOT:PSS.^[12–14] However, many applications require the chemical and mechanical features of hydrogels with a natural or synthetic biomaterial base, rather than a CP base to facilitate biocompatibility and biological outcomes such as cellular differentiation. To observe the endowed percolation behavior of ncrys-PEDOT₂₀ within a prototypical hydrogel, particles were systematically loaded within the aqueous PEGDA resin (see the Experimental Section)—this resin was chosen due to its widespread use for 3D tissue engineering constructs, biosensing media, and drug-controlled release matrices. As anticipated, a significant increase in conductivity was observed between 15% and 20% incorporation, with the highest loading displaying a remarkable conductivity of 1.1 S cm⁻¹ (Figure 3a). As a loading of 20% or more of acid-crystallized CP particles may negate the desired properties of the composite hydrogel, we also explored a postpolymerization strategy to achieve percolative electronic transport with significantly lower particle loading. PEGDA hydrogels with 1 and 5 wt% loadings show significant enhancements in conductivity upon subsequent in situ polymerization of EDOT. With this approach, we achieve 8.2 S cm⁻¹ with 5% particle loading, which is notably > 400 fold higher than an EDOT postpolymerization in PEDGA alone (0.02 S cm⁻¹). It should be noted that the use of EDOT (as opposed to derivatives thereof), and its particular polymerization and loading conditions is shown as a proof of concept and is the subject of future investigation. Furthermore, extrinsic factors such as applied stress can impact the conductivity of the hydrogels due to changes in interparticle distance (Figure S6, Supporting Information). A survey of literature examples reporting conductive hydrogels derived from 1) a CP base or 2) a biomaterial base with loaded conductive filler demonstrates that the afforded hydrogel has remarkable conductivity, well beyond the typical range of 10⁻³–10⁻¹ S cm⁻¹ (Figure 3b; and Table S2, Supporting Information).

Figure 3.

Conductivity characterization of conductive hydrogels with incorporated ncrys-PEDOT₂₀. a) Conductivity of photoinitiated hydrogels with square form factors via four-point probe analysis. Conductivity was measured on hydrogels with increasing loading of ncrys-PEDOT₂₀ with (dark blue outline) and without subsequent in situ polymerization of EDOT. The cartoons represent percolation being achieved through either (left) connection of particles via PEDOT from in situ post-polymerization or (right) particle contact through elevated concentration. The dotted lines represent conductivity at 0 wt% particle loading. b) Conductivity comparison of ncrys-PEDOT₂₀ loaded PEGDA hydrogel from (a) with other reported conductive hydrogels from Table S2 (Supporting Information). Each example is categorized by the order of magnitude of the reported conductivity (x-axis), and by the base composition of the hydrogel.

2.3. 3D Printing of ncrys-PEDOT_X Incorporated Resins

3D printing has matured into a promising technique for rapidly generating soft matter with complex form factors for biomedical applications, such as stents, scaffolds for tissue regeneration, implants, and soft robotics, among others.^{[42, 48, 49]} Printing conductive hydrogels while maintaining structural control of fine (<200 μm) features remains a challenge,^[42] as the incorporation of electroactive fillers into printable resins can dramatically impact the rate of photopolymerization or clog the printing nozzle due to particle aggregation during the printing process—arising from poor particle dispersibility.^[50] Alternatively, 3D printed insulating hydrogels can be endowed with electrical conductivity through 1) soaking the part in monomer solution, and 2) subsequently polymerizing the monomer in situ to generate a double network hydrogel—however such systems have limited control of the final CP loading, and can lack homogenous distribution of incorporated CP.^[9] While a growing literature has adopted making 3D printable resins directly from PEDOT:PSS,^{[13, 14, 51–53]} many biomedical applications require the high water content, mechanical properties, and biocompatibility of widely utilized synthetic biomaterials which mimic the native extracellular matrix microenvironments.^[54] It would be attractive to simply add a dispersible conductive filler to a pre-existing aqueous resin to directly print conductive hydrogels.

An aqueous photoprintable conductive resin was formulated through the addition of ncrys-PEDOT₂₀ into a mixture of poly(ethylene glycol)diacrylate (PEGDA) and lithium phenyl (2,4,6-trimethylbenzoyl) phosphinate (LAP) photoinitiator in water (Figure 4a, see the Experimental Section). The aqueous PEGDA resin was utilized to demonstrate the printability of ncrys-PEDOT_X due to its widespread adoption for 3D printing and its noncytotoxic nature;^{[54, 55]} ncrys-PEDOT₂₀ was used as the filler as it displayed the highest conductivity of the particles investigated while maintaining dispersibility. Microcontinuous liquid interface production (μCLIP) projection stereolithography, was used to rapidly print conductive hydrogels with the ncrys-PEDOT₂₀ incorporated at loadings of 1, 5, 10, and 15 wt% (Figure 4b).^{[56, 57]} The kinetic stability of ncrys-PEDOT₂₀ within the resin was investigated via sedimentation; a high loading of 10 wt% ncrys-PEDOT₂₀ was loaded into the aqueous resin simply through vortexing. A remarkable kinetic stability was observed without any sonication or added surfactants (Figure 4c); potentially due to the enhanced viscosity of the resin compared to the biomaterial dispersions (Figure 2a). As the average print time of a part is 15 min, and the part is printed from the bottom of the bath, the particles remain dispersed throughout the μCLIP 3D printing process.

Figure 4.

μCLIP 3D printing of ncrys-PEDOT_X. a) ncrys-PEDOT₂₀ can be dispersed within aqueous resins for 3D printing conducting hydrogels. Particles were added to a photocurable resin which consisted of poly(ethylene glycol)diacrylate (PEGDA), lithium phenyl (2,4,6-trimethylbenzoyl) phosphinate (LAP), and water. b) Diagram of the microcontinuous liquid interface production (μCLIP) stereolithography process to generate hydrogels. (Right) Image of the homemade μCLIP printer used to print all structures. c) Sedimentation kinetics of the PEGDA resin with 10 wt% loading of ncrys-PEDOT₂₀ (100 mg mL⁻¹) demonstrating stable dispersions within the timeframe of a print. (Inset) Image of the particle-loaded resin after sitting for 1 h. d) Cell viability of L929 cells treated with either media alone or extracts of 3D printed ncrys-PEDOT₂₀ incorporated hydrogels after incubation for 48 h. e) Quantitative measurement of cell viability using live/dead assay after incubation for 48 h. f–i) Fluorescent images of cells taken after live/dead assay, where calcein AM (green) represents live cells and ethidium homodimer (red) represents dead cells in contact with each ncrys-PEDOT₂₀ incorporated hydrogel. Scale bars for the images are 100 μm. (inset) Image of particle loaded hydrogel with the square form factor used for both four-point probe and biocompatibility analysis. j) Renderings of a complex tubular scaffold with a form factor consistent for an esophageal stent with 133.33 μm diameter struts. Photography (left) and SEM imaging (right) of PEGDA stents with k) 5 wt% and l) 10 wt% loading of ncrys-PEDOT₂₀. Scale bars for the photographs and SEM images are 2 mm.

Biocompatibility is an important and often overlooked property of materials intended for bioelectronic applications.^[6] The biocompatibility of ncrys-PEDOT₂₀ was assessed by proxy of particle-loaded 3D printed PEGDA gels. Following a modified ISO 10993-5 protocol, in vitro cytotoxicity tests were performed both on extracts from the particle-loaded hydrogels and through direct contact via alamarBlue (Figure 4d) and live/dead (Figure 4e) assays, respectively (see the Experimental Section). Both assays confirmed high cell viability (>95%) at all loadings up to 15% of ncrys-PEDOT₂₀. Significant differences in viability were not observed between the PEGDA hydrogels with and without particle incorporation, therefore the particles appear to be biocompatible. Imaging from the live/dead analysis (Figure 4f–i) demonstrates slight enhancements in cell spreading/adhesion at increased loadings of the ncrys-PEDOT₂₀ particles. Further, hemolysis testing on isolated ncrys-PEDOT₂₀ particles suggests high blood compatibility (5.77% hemolysis), an important material requirement for bioelectronic applications such as regenerative engineering or healthcare monitoring. The high compatibility and positive cell outcomes in the presence of the acid-crystallized PEDOT particles are consistent with other observations of PEDOT:PSS incorporated systems.^{[19, 21, 58]}

To demonstrate the capability to print complex, biomedically relevant geometries, particle incorporated aqueous PEGDA resins were printed (see the Experimental Section). Complex tubular scaffolds resembling esophageal stents with 133.33 μm diameter struts were printed (Figure 4j). Stents with both 5% (Figure 4k) and 10% particle loading (Figure 4l) demonstrated high structural fidelity with complex, small (< 150 μm) features. Such fine features are extraordinary considering the relatively high loading of conductive filler compared to similar formulations,^{[5, 19]} and that both sonication and surfactants were not required for resin preparation. As the dark ncrys-PEDOT₂₀ particles did not interfere with the photopolymerization process, acid-crystallized PEDOT particles can be incorporated within other biomedical resin formations and printed for a wide variety of applications such as biological signal recording, strain sensing, stimulation electrodes, drug delivery, and scaffolds for tissue regeneration.

3. Conclusion

A simple, scalable acid-based nonsolvent induced phase separation (NIPS) approach was leveraged to afford acid crystalized PEDOT nanoparticles (ncrys-PEDOT_X) from commercial PEDOT:PSS with dispersibility arising from an internal surfactant. By tuning the volume ratio of sulfuric acid within the coagulation bath, the PEDOT/PSS ratio was optimized to afford particles with conductivity rivaling all reports of conjugated polymer particles to date (σ_{ncrys-PEDOT20} = 410 S cm⁻¹). Evidence from spectroscopy and scattering suggest that both the removal of insulating PSS and a decrease in the interchain π–π stacking distance of PEDOT lead to enhanced conductivity within the crystallized particles, similar to reports of acid crystalized thin films of PEDOT:PSS, but in solution. Sedimentation and conductivity analyses suggests there is a fine balance when removing PSS via acid crystallization to achieve enhanced conductivity, while maintaining dispersibility.

When incorporated within aqueous dispersions of both natural (polysaccharides and proteins) and synthetic biomaterials, ncrys-PEDOT₂₀ demonstrated high compatibility without the need of external surfactants or sonication—this was observed at concentrations orders of magnitude higher than what is possible with pure graphene or carbon nanotubes. Furthermore, the particles did not interfere with the toolbox of chemistries typically used to crosslink hydrogels (hydrogen bonding, ionic bonding, Schiff-base chemistry, and radical photopolymerization). By directly adding ncrys-PEDOT₂₀ to common hydrogel formulations, highly conductive composites were achieved with a percolation threshold between 15 and 20 wt% loading. The observed conductivity (> 1 S cm⁻¹) is remarkable considering the base of the hydrogel was not derived from a CP. Given the aqueous dispersibility and kinetic stability at high loadings (100 mg mL⁻¹), the ncrys-PEDOT₂₀ incorporated PEGDA resin was utilized for microstereolithography. The afforded hydrogels demonstrated biocompatibility and structural fidelity with complex, small (< 150 μm) features. Such fine features are incredible considering the high loading of conductive filler, and that both sonication and surfactants were not required for resin preparation. As the reported particles display enhanced aqueous dispersibility, high conductivity, biocompatibility, and scalable access from a common commercial ink (PEDOT:PSS), we anticipate ncrys-PEDOT_X offers a robust platform toward conductive hydrogels for TERM and other bioelectronic applications. Unlike other conductive hydrogel formulations which require complex multistep procedures and additives, ncrys-PEDOT_X can be directly added to endow already existing and commonplace hydrogel formulations for biomedical applications with conductivity.

4. Experimental Section

Materials

PEDOT:PSS aqueous solution (Clevios PH1000) was purchased from Heraeus Electronic Materials. PEDOT:PSS dry redispersible pellets, sulfuric acid (ACS reagent, 95.0–98.0%), gelatin from porcine skin (gel strength 300, Type A), chitosan (medium molecular weight), agarose (Type I, low EEO), sodium alginate, and poly(ethylene glycol) diacrylate (PEGDA, M_n = 575 g mol⁻¹) were purchased from Sigma-Aldrich and used as received. Rat tail collagen (Type I, 2.05 mg mL⁻¹ in 0.6% acetic acid) was purchased from First Link (UK) Ltd. Glutaraldehyde (≈50% in water, ≈5.6 mol L⁻¹) and lithium phenyl(2,4,6-trimethylbenzoyl)phosphinate (LAP) were purchased from TCI.

Preparation of ncrys-PEDOT_X Nanoparticles

Acid crystallized PEDOT:PSS nanoparticles (ncrys-PEDOT_X) were synthesized by dropwise addition of PEDOT:PSS aqueous solution into a coagulation bath consisting of sulfuric acid in isopropanol under stirring for 12 h. Subsequently the particles were filtered, washed with isopropanol, deionized water, and acetone. The particles were then dried in vacuo and milled (Freezer/Mill SPEX Cryogenic Grinder) for three cycles where one cycle included 5 min of precool time, 15 min of grinding, and 2 min of rest. The speed was maintained at a rate of 15 counts per second (cps). Crystallization of 100 mL of ink within a 20% acid bath yielded 802 mg of ncrys-PEDOT₂₀.

Size Analysis

To probe the size distribution and morphology of the PEDOT nanoparticles, scanning electron microscopy was conducted using a Hitachi S-4800 SEM operated at 5 kV. Nanoparticles were dispersed in methanol and were drop cast and air-dried on carbon-coated tapes prior to imaging. The high conductivity of the particles precluded the need for a secondary conductive coating. The resulting images were analyzed using ImageJ, data were collated from over 200 individual particles.

X-Ray Photoelectron Spectroscopy

Measurements were performed with a NEXSA G2 XPS instrument, using a 15 keV Al-Kα X-ray source. Survey and high-resolution spectra of C 1s and S 2p core levels were recorded. All spectra were adjusted to the C 1s level of 284.8 eV. Data analysis was performed using the Thermo Fisher Avantage software.

X-Ray Diffraction

PXRD data were collected at room temperature on a STOE-STADI-P powder diffractometer equipped with an asymmetric curved Germanium monochromator (Cu Kα1 radiation, λ = 1.54056 Å) and 1D silicon strip detector (MYTHEN2 1K from DECTRIS). The line focused Cu X-ray tube was operated at 40 kV and 40 mA. Powder was packed in an 8 mm metallic mask and sandwiched between two polyimide layers of tape. Intensity data from 1.5° to 70° two theta were collected over a period of 120 min. Instrument was calibrated against a NIST Silicon standard (640d) prior the measurement.

Conductivity Measurements

For particle conductivity measurements 50 mg of ncrys-PEDOT_X or commercial lyophilized PEDOT:PSS was pressed into a pellet with an 8 mm diameter by pressing under a hydrostatic pressure of 3500 kg cm⁻² for 2 min. Four-point conductivity measurements were performed using a Lucas Labs Pro4 four-point probe head with a Keithley 9 2400 source meter. Samples were measured in triplicate and the average pellet or gel thickness was determined via caliper.

Sedimentation Analysis

Dispersions were prepared using 10 mg mL⁻¹ concentrations of either ncrys-PEDOT₅ or ncrys-PEDOT₂₀ in either DI H₂O and 2 wt% sodium alginate in DI H₂O. Dispersions (10 mg mL⁻¹) of ncrys-PEDOT₂₀ were also prepared in 2 wt% chitosan in 1% acetic acid and 0.2 wt% collagen in 0.6% acetic acid. At predetermined time intervals, aliquots of the supernatants were carefully collected from just below the surface of the meniscus and analyzed via UV–Vis at 633 nm. The ratio of the absorbance at a given time point compared to that of the initial time point was used to measure the stability of the suspensions.

Hydrogel Preparation

Gelatin: 400 mg of gelatin was added to 20 mL of DI H₂O and stirred at 45 °C for 1 h. A solution of ncrys-PEDOT₂₀ (10 mg mL⁻¹) was prepared using the gelatin stock solution and stirred at 45 °C for 1 h. For the vial tilt experiment, 4 mL of particle-loaded gelatin dispersion was slowly cooled to room temperature. For the reinforcement experiment, 6 mL of the particle-loaded gelatin dispersion was poured into a polystyrene dish and slowly cooled to room temperature. Discs were punched using an 8 mm biopsy punch and washed with DI H₂O. Alginate: 400 mg of sodium alginate was added to 20 mL of DI H₂O and stirred at room temperature overnight. A solution of ncrys-PEDOT₂₀ (10 mg mL⁻¹) was prepared using the alginate stock solution and stirred at room temperature for 1 h. For the alginate worm experiment, particle-loaded alginate dispersion was injected into an aqueous solution of 1 m CaCl₂. For the reinforcement experiment 6 mL of the particle-loaded alginate dispersion was poured into a polystyrene dish lined with a filter paper soaked with 0.05 m CaCl₂ and incubated for 24 h at room temperature according to previously reported procedures.^[38] Discs were punched using an 8 mm biopsy punch and washed with DI H₂O. Chitosan: 2 g of chitosan was added to 100 mL of aqueous acetic acid (1% v/v) and stirred at room temperature until homogeneous. A solution of ncrys-PEDOT₂₀ (10 mg mL⁻¹) was prepared using the chitosan stock solution and stirred at room temperature for 1 h. For the vial tilt experiment 1 mL of aqueous glutaraldehyde (50%) was added to 4 mL of particle loaded chitosan dispersion at room temperature. For the reinforcement experiment 6 mL of the particle loaded chitosan dispersion was poured into a polystyrene dish lined with a filter paper soaked in 25% aqueous glutaraldehyde and incubated for 24 h at room temperature. Discs were punched using an 8 mm biopsy punch and washed with DI H₂O.

Mechanical Characterization

The mechanical properties of the hydrogels were investigated in the compression mode using a DMA (850, TA Instruments) with a 0.05 N preload force, at 25 °C under immersion in DI H₂O. The elastic modulus was determined from slope of the stress–strain curve within the elastic region. The elastic modulus, compressive strength, and compressive strain were calculated from five hydrogels discs.

Printing of ncrys-PEDOT₂₀ Incorporated Resins

A home-made μCLIP printer was used to print all structures utilized throughout the study. Conditions used for printer setup and operation followed previously reported procedures.^{[56, 57]} Focal plane pixel resolution was set at 3.98 × 3.98 μm, and 5 μm layer thickness was used. The system is similar to those previously described, with updates to the light projector and light path. A hydrogel resin consisting of PEGDA (39.5 wt%), LAP (0.5 wt%), and water (60 wt%) was used with varying loading of ncrys-PEDOT₂₀. Particles were loaded into the resin at 1, 5, or 10 wt%, initially vortexed for 2 min, and vortexed for 1 min immediately both before use and between each print. Using the μCLIP printer, esophageal stents and square form factors were printed.^[59] Esophageal stents were designed to be 10.93 mm long with supports and have struts 133 μm in diameter. Squares were 8 mm in length and height and were 1 mm in thickness, with 1 mm tall supports on one side to adhere prints to the platform. Inks with 1 wt% ncrys-PEDOT₂₀ were printed with 0.0008 mJ mm⁻² power, 5 wt% loading with 0.2188 mJ mm⁻² power, and 10 wt% loading with 0.3063 mJ mm⁻² power. In situ polymerization of particle loaded prints with EDOT followed previously reported procedures.^[60]

Biocompatibility Analysis

In vitro cytotoxicity tests were performed on particle-loaded PEGDA hydrogels based on the ISO 10993-5 protocol.^{[6, 61]} 3D printed hydrogels with square (8 mm × 8 mm × 1 mm) form factors were assessed with either 1, 5, 10, or 15 wt% ncrys-PEDOT₂₀. Indirect Extract Tests: Before performing cytotoxicity experiments, each hydrogel was washed with ethanol, DPBS, and Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) media (Gibco; Cat no. 11 320 033). DMEM/F-12 media supplemented with both 10% fetal bovine serum (FBS) and 1% antibiotic–antimycotic (2 mL) was added directly onto 0.15 g of hydrogel and incubated at 37 °C with 5% CO₂ in a humidified incubator for 24 h to create extracts. Separately, a suspension of L929 cells was prepared at a concentration of 5 × 10⁴ cell mL⁻¹ and 100 μL of the suspension was dispensed into each well of a 96-well plate. The plate was incubated at 37 °C with 5% CO₂ in a humidified incubator. After 24 h, culture media was removed from the wells and replaced by 100 μL of extract media from the hydrogels. Following 48 h, the cell viability was assessed via alamarBlue assay (Invitrogen; Cat no. A50101). Samples were repeated in quintuplicate. Direct Tests: L929 cells were prepared at a concentration of 5 × 10⁴ cell mL⁻¹ and 200 μL of this suspension was added directly on the particle-incorporated hydrogels within a 24-well polystyrene tissue culture plate. The plate was incubated at 37 °C with 5% CO₂ in a humidified incubator for 4 h for cell attachment, and 300 μL of DMEM/F-12 media supplemented with both 10% fetal bovine serum (FBS) and 1% antibiotic–antimycotic was carefully added into each well. Following 48 h, a live/dead solution was prepared from the LIVE/DEAD Viability/Cytotoxicity Kit (Invitrogen; Cat no. L3224) with 0.5 μL of calcein AM and 2.0 μL of ethidium homodimer-1 (EthD-1) in 1 mL of DPBS. Each hydrogel was stained with live/dead solution (200 μL), incubated for 30 min, and imaged with a Nikon confocal microscope. The cell viability percentage was calculated using Fiji (ImageJ).

Hemocompatibility Analysis

12 mL of the blood samples were collected from three female rabbits and pooled together in anticoagulant vials. The hemocompatibility of ncrys-PEDOT₂₀ was assessed using a 50 mg mL⁻¹ stock solution in DI water. The hemolysis assay protocol was adapted from prior literature procedures.^{[62, 63]} Briefly, 300 μL of particle solution was added to 4 mL of 0.9% saline solution and equilibrated for 30 min at 37 °C. 200 μL of diluted blood (4 mL of blood diluted in 5 mL of 0.9% saline solution) was added to the sample. A negative control was prepared by adding 200 μL of diluted blood to 4 mL of 0.9% saline solution (0% hemolysis) and a positive control was prepared by adding 200 μL of diluted blood to 4 mL of DI water (100% hemolysis). The sample was then incubated at 37 °C for 1 h, centrifuged at 1000 rpm for 5 min, and the absorbance of the supernatant was measured at 545 nm. The hemolysis experiment was performed in triplicate.

Data Availability

The authors declare that the data supporting the findings of this study are available within the paper and its Supporting Information files.