Antifouling strategies for protecting bioelectronic devices

Bioelectronic devices are pivotal to a number of areas ranging from biomarker sensing, to cancer and epilepsy diagnostics, to blood sugar monitoring and brain activity recording.^1–3 However, many if not most of those use cases, especially those that have to occur in the complex fluid environments that are inevitably encountered in living systems, that are negatively affected by fouling that occurs as a result of non-specific adsorption of chemical and biological material on the device surfaces. As a result, a remarkable amount of effort⁴ has gone into the development of antifouling strategies that protect the long-term functionality of bioelectronic devices (Fig. 1). In this Research Update, we discuss general antifouling principles and major fouling mechanisms and give some examples of the strategies that are currently used to prevent foulant binding to device surfaces and remove them after fouling has occurred.

Fouling, and in particular biofouling, proceeds through four distinct stages.⁵ First, pristine surfaces get covered with a conditioning layer of small molecules; in the second stage, the conditioning layer gets covered with the main foulant layer; in the third stage, the fouled surface is subject to a robust biofilm growth; and finally, this biofilm progresses further to macrofouling.⁶ Adsorption of the conditioning layer of sugars or other small molecules usually happens within the first several minutes of immersion in solution.⁷ At this stage, the film is not yet continuous and its morphology often resembles islands. Protein adsorption follows during the second stage, which can last hours and produces a more continuous film. Bacterial film growth and surface colonization occur during the third stage. Bacteria secrete a variety of extracellular polymeric substances (EPSs), which are composed mainly of proteins and polysaccharides, e.g., hexose, hexosamine, and ketose,^8–10 and can also trap a variety of other species. In the final stage, larger foulants will attach to and grow on the surface. This process is termed macrofouling and usually takes days or weeks.¹¹ Once a biofilm has appeared on the sensor surface, it is exceedingly difficult to remove unless harsh and destructive chemical or physical treatments are used; therefore, it is much more practical to focus on preventing fouling before the biofilm has formed. Thus, for this update, we will concentrate on recent advances developed by the researchers to suppress the initial surface attachment and reverse subsequent protein layer growth.

Most antifouling strategies can be grouped into several somewhat overlapping classes (Fig. 2). The broadest distinction could be made between the approaches that try to prevent adsorption to the device surfaces and those that aim to remove the films after they have been formed. Prevention and removal strategies can also be classified further based on their general working principles.⁴ 

Physical strategies rely on direct manipulation of the sensor interface to prevent or slow down the foulant attachment to those surfaces or to facilitate removal of foulants once they have attached. This group also includes a number of approaches for physical cleaning of the surface that ranges from a simple mechanical cleaning⁵ to a more sophisticated and gentle stimuli-responsive removal.¹² A large number of these approaches rely on porous membranes and filters^13–16 or porous electrodes^17–20 to restrict or exclude undesired species from reaching the surface of the device active components. Fine-tuning the pore size may also allow only molecules of specific sizes to access the sensor surface, enabling the design of highly specific and sensitive detectors that work in biological fluids.^21–23 An alternative approach involves engineering device surface roughness or its wetting properties to impede foulant adsorption. Other variations use directed flows to actively prevent foulants from reaching the device surface.²⁴ These approaches provide some of the most robust antifouling strategies, although they often suffer from slow response times, limited mass transport kinetics, adverse effects, fouling of the protective membrane, or degradation accompanying the physical removal of the foulant.

Chemical strategies, which rely on functionalization of the interface with moieties that have antifouling properties, cover a rather large group of approaches. They often rely on chemical modifiers that increase the hydrophilicity by using polar chemical groups, hydrophilic polymers, and self-assembled layers. Typical examples include polyethylene glycol (PEG) and its derivatives, zwitterionic species [phosphorylcholine (PC)], peptides with alternating or random mixed-charge [glutamic or aspartic acid (E/D) and lysine (K) based, e.g., EKEKEKEC],²⁵ and other polymers such as polysaccharides, polyoxazolines, poly(hydroxy acrylates), and hyperbranched polyglycerol (HPG) or hybrid materials such as PEG-SB, which all have antifouling properties that can be traced to their highly hydrated states.^26–29 

The spectrum of the strategies that researchers use to immobilize these and other coatings on the device surface is equally vast. Common approaches include self-assembly,^30–32 electro-polymerization,^33–36 atom transfer radical polymerization (ATRP),^37–41 and reversible addition–fragmentation chain transfer (RAFT).^42,43 Self-assembly offers easy access to well-defined, monolayer structures, but this method is applicable only to a small subset of smooth (and typically metal) substrates.^31,32,44,45 Antifouling polymer films offer a much more versatile antifouling modification approach and can be deposited using graft-to and graft-from strategies^46–49 or electrografting.^50–53 

Some of the general drawbacks of these approaches include limited compatibility with some biological species critical for device functionality, the potential to block the device surface from access by the species of interest, low yield of some of these functionalization techniques, and the potential for generating reactive species during chemical modification steps or triggering an adverse immune response.

Biological strategies comprise a relatively limited subset of approaches that rely on affinity binding to capture and deplete foulants before they reach the device surface or use enzymatic catalysis to destroy them once they reach the device. These approaches are strongly selective and can be applied in a highly targeted mode; however, they are mostly suitable for small input volumes and are generally not applicable for in vivo sensing and high-throughput operation.

Complicated fouling mechanisms and a wide variety of existing antifouling strategies make it difficult to come up with a universal approach for assessing antifouling performance. The most commonly used metrics are (1) mass of the foulant adsorbed on the surface, (2) durability and lifetime of the antifouling coating under a fouling challenge, and (3) degradation or loss of quality in a signal detected by using a sensor (the indirect signal change method). The adsorption of foulants on surfaces method is widely utilized in membrane-based antifouling strategies. The adsorption of single protein or a mixture of different proteins from single-protein solutions, diluted serum, or blood plasma onto surfaces is used as a quantitative metric for antifouling performance. A surface is considered “antifouling” when it allows fouling levels of less than 1000 pg/mm², “low-fouling” when it achieves levels of less than <100 pg/mm², “ultra-low fouling” when it achieves levels of less than 50 pg/mm², and “non-fouling” when the protein adsorption levels fall below the limit of detection of the technique used to quantify the protein adsorption on the sensor surface.⁵⁴ 

Even though the lifetime-based method is very intuitive, it is still difficult to make an apples-to-apples comparison for different antifouling strategies from the literature since the foulant composition and concentration vary so much. The indirect signal change method is mostly used in cases where implantable or circulating sensors are present. The signal changes from fluorescence, phosphorescence, Electrochemical Impedance Spectroscopy (EIS), fluorescence resonance energy transfer (FRET), voltammetry, amperometry, colorimetry, or potentiometry are used to quantify specific antifouling performance.

We now survey some of the common antifouling strategies that researchers use to protect bioelectronic devices. As we mentioned before (Fig. 2), we can group them into two very broad categories based on whether a particular strategy aims to remove the foulants from the surface or prevent them from being adsorbed on the surface at the first place.^55,56

Mechanical removal of the fouling layer from a device surface is probably the most straightforward antifouling strategy; unfortunately, it is not always available and often can damage the sensor surface.⁵ These drawbacks can be mitigated by more sophisticated mechanical actuation approaches, for example, by using piezoelectric elements. The U.S. Navy has patented a sensor, which vibrates upon excitation by electric pulses, thus removing fouling material from the surface.⁵⁷ Similarly, Rahmoune and Latour⁵⁸ reported that coating the sensor surface with electrically actuated piezoelectric material helps combat biofouling. Lizotte et al. utilized the Clean Sweep wiper system for the cleaning of their multi-parameter probe.⁵⁹ Alternatively, mechanical devices can be used to expose the sensor for a minimum time required to sample and then remove it, as was demonstrated by Baxter with a marine sensor mounted in a protective housing above the water level and immersed in the seawater only for the brief time needed for data collection.⁶⁰ In a variation of this approach, Garner developed an apparatus with a protective shutter that could be opened to expose the sensor for measurements and then closed to prevent fouling.⁶¹ We note that this approach does not eliminate the potential for fouling but does limit the amount of the foulant that can accumulate on the device surface.

Mott et al. reported using ultrasound to prevent fouling on glass tubing.⁶² This approach will promote formation of biocides and/or antibiotics on surfaces and enhance biocide efficiency. The combination of ultrasound and low concentration of gluteraldehyde, gentamicin, or other biocides will decrease the bacterial concentration significantly.^63–66 Bott⁶⁷ and Mermillod-Blondin et al.⁶⁸ reported that ultrasonic irradiation also helped to detach foulants adsorbed on surfaces. Piedrahita and Wong reported direct integration of ultrasonic emitters with sensors that helped to prevent fouling.⁶⁹ Overall, the effectiveness of these approaches ranges from hours to months of protection depending on the detailed types of mechanical cleaning and fouling layers.

Nature provides a number of examples of surfaces that respond to environmental stimuli. Pilot whales have skins that can remove small patches of skin cells by enzymatic digestion.⁷⁰ Fish have flexible oil-repellent skins that exhibit excellent oleophobicity even when they are subjected to strong mechanical deformation or stimulation.⁷¹ These systems have provided plenty of inspiration for the development of artificial stimuli-responsive antifouling device coatings. The antifouling effectiveness for these systems may range from hours to several days.

Dam-Johansen and co-workers investigated methacrylate copolymer-based self-polishing coatings, which were inspired by pilot whale skin.⁷² Polymers in these coatings will hydrolyze and dissolve in seawater or similar environments, regenerating a foulant-free surface. Ganguli et al. used an underwater circulatory system to form protective coatings based on an anionic surfactant that is able to respond to salinity of seawater¹² and dissolve after a programmable period of time. The authors reported that this self-polishing coating was able to dramatically reduce surface fouling from Pseudoalteromonas carrageenovora, a typical biofilm forming marine bacterium.¹² 

Researchers have reported using poly(N-isopropylacryl-amide) (pNIPAm) brushes on different substrates to control hydrogen bonding and polymer swelling in order to drive off foulants from substrate surfaces.^73–75 pNIPAm brushes have a lower critical solution temperature (LCST) of ∼33 °C in water, and these polymers will show different surface energy and wettability at temperatures higher or lower than LCST.⁷⁴ This LCST transition will switch the polymers between hydrophobic and hydrophilic states, which could lead to the removal of foulants based on their preference for hydrophobic surfaces. For example, more than 90% of Staphylococcus epidermidis and H. marina cells were still released from pNIPAm-grafted poly(styrene) surfaces even after 72 h incubation after pNIPAm brushes had been switched to a more hydrophilic, swollen state.⁷⁵ The same coating can also be grafted to initiator-modified self-assembled monolayers (SAMs)⁷³ and also trigger efficient bacterial film removal upon wettability switching.⁷⁶ 

Zarzar et al. reported a pH actuated hydrogel polymer coating,⁷⁷ which consisted of silicon nanocolumns in polyacrylamide layer and swelled and contracted in response to pH, mimicking dermal denticles embedded under bristling skin from sharks. In another example, Aizenberg and co-workers reported a reversible system to control the surface energy and microtopography of the substrate in order to remove aggregated foulants.⁷⁸ 

Physisorbed foulants can also be degraded by enzymes.⁷⁹ Bioelectronics researchers can borrow these approaches from the water treatment industry, where enzymatic digestion is commonly applied to help with fouling control.^80,81 For example, Shi et al. attached trypsin to the poly(methacrylic acid)-graft-poly(ether sulfone) (PMAA-g-PES) membrane⁸² and reported robust fouling resistance against multiple cycles of BSA filtration over 15 days. Koseoglu-Imer et al. also reported a cellulose acetate membrane modified by serine protease enzyme (Savinase) achieving high protein rejection during membrane filtration.⁸³ These approaches can be easily adapted to various biosensors as long as these enzymes do not interfere with the sensing mechanism.

Another approach, demonstrated by Bar et al.,^84,85 used phosphotriesterase (PTE)-like lactonases, which can hydrolyze lactones including the quorum sensing (QS) lactones to build an active substrate that can resist foulant aggregation. Phosphonium salts have broad antibacterial activity, thermal stability, and low toxicity,⁸⁶ and their positive charge attracts to negatively charged cell walls, causing interruptions in cell membranes and lysing the cells.⁸⁶ Martín-Rodríguez et al. studied the antifouling activity of organic phosphonium salts with different alkyl chains⁸⁷ and reported that molecules with longer than C₇ alkyl chains exhibit biocidal activities and that long saturated alkyl chains (C₁₆–C₁₈) show high tyrosinase inhibition activities.

Other approaches use surface coatings that release biocidal chemicals;^56,88–90 Ag or Cu nanoparticles, which also have biocidal properties;⁹¹ and antibiotics.⁹² Even though these approaches are effective, the chemicals released by the coatings can be detrimental to the environment or they can lead to the emergence of antibiotic-resistant bacterial strains.⁷⁹ Typical challenges associated with these approaches include low enzyme stability and strong potential for degradation of enzymes by proteases, which may ultimately limit practical applications of this strategy.

A common approach of using electrolysis of saltwater to produce biocidal chlorine and hypochlorous acid⁹³ is not viable for most biosensors as it will rapidly degrade the devices. An alternative approach utilizes local electric fields to kill fouling organisms directly, instead of forming a biocidal intermediate. Nakayama et al. reported biocidal activity for TiN electrodes,⁹⁴ and Nakasono reported similar activity for graphite–silicone electrodes as well.⁹⁵ Amr and Schoenbach also reported a strategy that degrades foulants using short duration electrical pulses, which cause membrane breakdown.⁹⁶ 

Direct ultraviolet irradiation of surfaces can prevent biofouling and has often been utilized in marine sensors and wastewater treatment;^97–99 however, it is clearly not ideal for biosensor use due to high power requirements and potential for indiscriminate surface damage. Alternative approaches explore coating of surfaces with photocatalytic materials, which require much smaller irradiation doses. Linkous et al. reported application of TiO₂ and WO₃ coatings for photocatalytic inhibition of foulants.¹⁰⁰ Morris and Walsh reported zinc oxide based photoactive coating comprised of 20% photoactive zinc oxide and about 5% of photosensitizers to protect and preserve surfaces submerged under water that are subject to fouling.¹⁰¹ 

These approaches, which are closely related to mechanical removal strategies described earlier, use induced fluid flow over the device interface to prevent the foulant from reaching and sticking to the device surface.^102–112 As reviewed by Wisniewski and Reichert,²⁴ there are mainly four types of flow-based systems: Manchester open flow microperfusion sensor, microdialysis sensors, Graz open flow microperfusion sensor, and ultrafiltration sensors (see Fig. 3). Even though these approaches have been realized in tissue-deployed sensing systems, they require pumps, vacuums, and fluid reservoirs that introduce significant complexity and energy consumption. In an interesting development of this concept, Towe et al. reported a fully implantable microdialysis-based sensor that will regenerate every other week, which only requires a small volume (picoliter to nanoliter) perfusion system.^24,113

Hydrogels have been widely used in applications ranging from hygienic products to drug delivery to pharmaceuticals and tissue engineering.¹¹⁴ The most popular hydrogels are probably the biocompatible poly(hydroxyethyl methacrylate) (PHEMA)^115–117 and poly(ethylene glycol) (PEG).^118,119 Multiple reports have shown the antifouling efficiency of hydrogels.^120–122 Hydrogels coated on surfaces create a hydrophilic semi-permeable interface that still allows water soluble analytes to translocate through and reach the device surface. Changing crosslinking density will lead to a change in water content in the gel and influence the analyte diffusion. Brinkman et al. reported using chemical modification to enhance the biocompatibility and antifouling activity of heparin-modified polyvinyl alcohol (PVA) hydrogels,¹²³ showing that it led to a significant increase in recalcification times (RCT, a measure of the time taken for clot formation in recalcified blood) of plasma in contact with the hydrogel-modified surfaces. Hydrogels can also play a more active role in bioelectronic devices and serve as a matrix to incorporate enzymes, which can enhance antifouling performance.^{117,124–128} Several groups reported using PHEMA coating to achieve enhanced antifouling performance of sensors such as a calcium ion sensitive field effect transistor (ISFET), Ag/AgCl reference electrode catheter, and catheter-tip pH/pCO₂ sensor.^129–131 

Hydrogel-based approaches always need to overcome several challenges: (1) weak interaction between the hydrogel and the underlying substrate; (2) low mechanical stability that leads to the coating damage by implantation forces; (3) potential incompatibility between the sensor components and the hydrogel monomer, solvent, and crosslinking agent; and (4) additional diffusion barrier that the hydrogel structure poses for the analyte trying to reach the sensor surface.

Phospholipid coatings mimic the structure and composition of cell membranes and thus provide a naturally fouling-often capable of self-repair. Lewisresistant surface coating that is often capable of self-repair. Lewis reported using phospholipids and phospholipid-based biomaterials such as dipalmitoylphosphatidylcholine (DPPC), diacetylenic phosphatidylcholine (DAPC), 2-methacryloyl-oxyethylphosphorylcholine (MPC), segmented polyurethanes (SPUs), and their copolymers to enhance the biocompatibility of bioelectronic devices.¹³² Ishihara et al. examined plasma adsorption onto different phospholipid polymer coatings and reported that protein adsorption was inversely correlated with the free water fraction in the coating.¹³³ A further analysis showed that proteins adsorbed onto these coatings kept their native structure, indicating that protein–surface interactions were minimized. Phospholipid-based coatings are inherently fragile, cannot be dried, only function in a limited pH range, and are susceptible to electroporation. The potential for the oxidative damage in extreme environments could further compromise their antifouling performance.

Nafion, a biocompatible sulfonated tetrafluoroethylene based fluoropolymer-copolymer, which is widely available commercially, is relatively easy to deposit on surfaces using dip-coating,¹³⁴ and it has been used as membrane coatings on biosensors for more than a decade.^{126,135–139} The strong anionic nature of Nafion membranes on sensor surfaces retards foulant adsorption¹⁴⁰ and enhances the sensor lifetime.¹⁴¹ For example, Moussy et al. studied the performance of the Nafion coating on devices during in vivo long-term applications^{138,141–145} and reported stable responses and lower instances of inflammatory reactions.

Surfactants represent another broad class of chemicals that are widely used to form antifouling coatings. Multiple groups reported treating device surfaces with Pluronic^TM surfactants (PEO–PPO–PEO triblock copolymers) to enhance device resistance to protein adsorption and cell adhesion.^146–148 Bakker et al. reported using surfactants on electrochemical sensors to improve mobility and selectivity for Ca²⁺, SCN⁻, Br⁻, NO³⁻, I⁻, and $SO42−$ ions.^149,150 Vadgama et al. reported that surfactant coatings such as Tween 80 or Triton X-100 and bis(2-ethylhexyl) hydrogenphosphate reduced electrode fouling by enzymes in blood.^151–153 Reddy et al. reported that anionic surfactants show stronger antifouling performance than non-ionic ones, which were, in turn, better than cationic surfactants.^151,152 Electrostatic repulsion from anionic surfactants helps reduce adhesion of negatively charged blood-borne cells and increase the fluidity of the membranes.^151,152

However, long-term in vivo applications for surfactant-plasticized membranes are not recommended. Lindner et al. reported that plasticizers leaching out from implanted sensor surfaces will lead to enhanced inflammation.¹⁵⁴ 

Covalent organic frameworks (COFs) are a group of mesoporous polymers that form defined skeletons with regular nanoscale pores.^155,156 COFs exhibit excellent structural stability and their regular porosity, which essentially acts as a precision membrane, could allow species of interest to pass through the coatings and reach the protected sensing surfaces.^157,158 Yang et al. reported using COFs (COF_{TAPB−BMTTPA}) as membrane coatings on graphene FET sensors to reduce fouling,¹⁵⁹ as shown in Fig. 4. The authors demonstrated Hg²⁺ sensing at the ppm level in the presence of poly-(3,4-ethylenedioxythiophene):poly-(styrenesulfonate) (PEDOT:PSS) fouling solution at concentrations ranging from 10⁻¹⁰ M to 10⁻⁴ M. The COF coating can even enhance device performance in neural activity monitoring in a complex fouling environment.¹⁵⁹ Engineering the building blocks that form the crystalline mesoporous structure of COFs gives many possibilities for fine-tuning the network porosity and functionalizing the network walls with different receptors, opening up a large design space for constructing sensors that can operate in complex fouling conditions.

Carbon-based coatings are another common approach to improve material biocompatibility.¹⁶⁰ Diamond-like carbons (DLCs) are chemically inert ∼10 nm thick hydrocarbon coatings that can be deposited on substrates by bombarding hydrocarbon vapors with argon beams.^161,162 Critically for bioelectronic applications, this process occurs under room temperature and is non-destructive to the devices and substrates. Macrophage and fibroblast cells grown on DLC coatings showed no adverse morphological and functional changes in cell culture and in in vivo tests.¹⁶³ Higson et al. reported increased lifetime for sensors with DLC coatings in whole blood testing for less than 1 h.^161,164,165 Disadvantages of the strategies based on DLC coatings include the need for specialized atom bombardment equipment and the also somewhat limited ability to control these coatings on device surfaces.

Surface topography and the accompanying variations in surface chemistry are surprisingly powerful and complex mediators of foulant adhesion.^166,167 Plants, marine organisms, and other living species use unique nanoscale topography to protect themselves from fouling,^90,168 and these topographies have been inspiring researchers to develop new types of antifouling surfaces^55,56 based on similar physical principles (Fig. 5). For example, shark skin has unique morphology consisting of patterned dermal denticles covered by riblets, which help form vortices along the shark’s body, reducing the drag flow and preventing foulant aggregation on the skin.¹⁶⁹ Brenan et al. reported a micro-patterned antifouling surface, fabricated using polydimethylsiloxane (PDMS) soft lithography,¹⁷⁰ which mimics shark skin and shows excellent antifouling activities.¹⁷¹ Kirschner and Brennan also reported antifouling behavior in another surface that mimics shark skin topography.⁵⁶ In another example, the skin of a pilot whale contains nanopores surrounded by nanoridges that also help reduce fouling.¹⁷² Cao et al. reported a similarly designed surface topography based on poly(acrylic acid) (PAA) and polyethyleneimine (PEI) multilayers fabricated by polyelectrolyte layer-by-layer spray coating to alter organisms’ adhesion on surfaces.¹⁷³ 

Perhaps, the best-known example of antifouling surface topography is the lotus leaf, whose superhydrophobic surface shows remarkable self-cleaning characteristics.^169,174 The lotus leaf surface consists of dense hydrophobic bumps that minimize the surface energy, resulting in minimal adhesion on the surface.¹⁶⁹ Pechook et al. deposited paraffin and fluorinated waxes on different substrates (cooper, glass, and silicon) to form nanotopographies with superhydrophobic and self-cleaning characteristics similar to the lotus leaf and showed that they inhibit biofilm formation and maturation on modified surfaces.^174–176 

Aizenberg et al. reported a slippery liquid infused porous surface (SLIPS) inspired by a different plant species, a pitcher plant, Nepenthes.¹⁷⁷ A pitcher plant allows water to penetrate through its leaf surface and form a wetting film that repels oil (in contrast to the lotus leaf case where the water droplets are repelled by the surfaces and air trapped in between textures). The authors demonstrated that by modifying the composition of the lubricants, they can achieve strong antifouling behavior in oil, water, and even blood. SLIPS also inhibits the accumulation of bacteria such as Pseudomonas aeruginosa, Escherichia coli, and Staphylococcus aureus. In addition to being tailorable to protect from a wide range of foulants, SLIPS can withstand extreme conditions such as high temperatures, extreme pH, and UV irradiation.¹⁷⁸ 

Rice leaves and butterfly wings inspired Bixler and Bhushan to design a surface topography that combines hydrophobic surface roughness with the anisotropic surface flow mechanism to create low adhesion self-cleaning surfaces.¹⁷⁹ The rice leaves have parallel grooves with a sinusoidal pattern, which promotes anisotropic flow. In addition, these leaves are covered by micropapillae of epicuticular wax, which creates superhydrophobicity and results in ultra-low adhesion of contaminants.^169,179 Butterfly wings also exhibit similar long microgroove topographies. Bixler and Bhushan used lithography to fabricate substrates mimicking these topographies and then coated them with hydrophobized silica nanoparticles. These surfaces were able to achieve low drag flow and self-cleaning behavior that resulted in excellent antifouling characteristics.¹⁷⁹ Of all antifouling strategies discussed here, surface topography-based systems are perhaps the hardest to compare. Some reports indicate very good antifouling effectiveness of these surfaces over several hours; however, the variation in the surface topography and surface structure and scarcity of long-term effectiveness studies limit our ability to perform such a quantitative comparison.

Continuous diffusion filtering (CDF) is a somewhat unusual approach, usually realized in microfluidic chips, that relies on fluid flow to keep foulants from reaching the sensor surface. In this approach, the sensor surface is separated from the sampling stream (such as blood) by a laminar flow of buffer solution. Small molecules diffuse quickly through this buffer layer and can reach the sensor surface, while most background proteins such as HAS, IgG, and Fib, which have two orders of magnitude lower diffusion coefficients, are swept away before they can attach to the sensor surface. Plaxco, Soh, and co-workers reported continuous monitoring of DOX levels in blood using a custom-designed chip (Fig. 6) that combined CDF and a DNA aptamer-based biosensor to enable continuous doxorubicin (DOX) monitoring in live rat bloodstream for 4.5 h.²³ 

In another example, the same group also covered electrode surfaces with biocompatible polysulfone membranes (∼0.2 μm in pore sizes) to monitor blood flow in the vein and demonstrated that these devices can monitor drug pharmacokinetics in rats.¹⁸⁰ These in vivo sensing platforms are also capable of small molecule sensing. Feng et al. also reported a similar approach for in vivo electrochemical dopamine sensing in the mouse brain by using a polytannic acid-doped nanoporous PANI membrane to protect carbon fiber electrodes.¹⁸¹ 

Self-assembled monolayers are highly ordered monolayers of typically amphiphilic molecules, which normally have the molecular structure of R₁–(CH₂)_n–R₂, where R₁ is a headgroup, (CH₂)_n is a nonpolar alkane chain, and R₂ is an anchoring group that is chemically attached to the substrate. Unlike some other surface coatings, the majority of the SAM stability comes from the hydrophobic interactions of the long alkane chains that form its bulk, and longer alkane chains (with n > 11) tend to form more stable SAMs. The anchoring group R₂ directs the amphiphile to the surface but usually does not form a strong bond; in fact, some mobility of the amphiphile molecule on the substrate is beneficial for the ability of the SAM to complete high-quality assembly. Due to the tight packing of the SAM molecules in the monolayer, the headgroup, R₁, fully determines the chemical nature of the SAM surface.

For different substrates, the selection of the anchoring group, R₂, will be different: thiol (–SH) is used on gold or silver; silane (–SiCl₃) is used on glass and silicone; and silane or phosphate is used on metal oxides. When hydrophilic headgroups are used, they significantly decrease the free energy between the SAM surface and solution¹⁸² and as a result would reduce the fouling propensity of these surfaces. The dense hydrophobic interlayer of alkane chains also makes SAM-based coatings good candidates for applications where direct contact of the analyte molecules with the sensor surfaces is not critical.^183–186 Chapman et al. studied antifouling activities of more than fifty ω-functional SAM coatings terminating in different groups from Fbg and Lys solutions and reported that surfaces with hydrophilic, electro-neutral, or hydrogen-bond accepting functionality all contribute to the excellent resistance to protein adsorption.¹⁸⁷ 

Oligo(ethylene glycol)-terminated SAMs (OEG_n SAM, where n denotes the number of ethylene glycol units) are also often used as antifouling surfaces.^{182,188–195} Prime et al. reported using oligo(ethylene glycol) (OEG) terminated SAMs to prevent device fouling from single-protein solutions and reported that they can effectively reduce adsorption from single-protein solutions.^196,197 They also reported that fouling from blood plasma decreased with increasing number of ethylene glycol units, n. Boozer et al. used a mixture of R₂–(CH₂)₁₁–OEG_n and R₂–(CH₂)₁₁–OEG_n–F (F denotes COOH, NH₃, and biotin) to create SAMs with receptors attached to F functional groups, preparing versatile multichannel biosensor surfaces capable of probing sequence-specific DNA conjugates and directing targets to the appropriate spots on the surface.^198–201 

Zwitterionic SAM surfaces, which typically present an equal amount of positive and negative charges, are usually strongly hydrated, which gives them some antifouling activity.^202–204 Holmlin et al. reported using zwitterionic SAMs to reduce adsorption from Fbg and Lys solutions.²⁰³ Those SAMs were prepared with an equal amount of alkanethiols terminating in a positively charged tri(methyl)ammonium (TMA) functional group and a negatively charged sulfonate group (SO). SAMs consisting of alkanethiols terminating in sulfobetaine (SB) or phosphorycholine (PC) zwitterion groups also demonstrated good antifouling performances.^203–206 SB SAMs were also more effective in resisting Fbg and Lys adsorption than PC SAMs.^203,205,206

Researchers have reported antifouling SAMs terminating in different hydrophilic groups including tri(propylene sulfoxide) (TPS),²⁰⁷ mannitol,²⁰⁸ gulitolor mannonamide,²⁰⁹ hyperbranched or dendritic polyglycerol,²¹⁰ and oligopeptides (Fig. 7).²¹¹ Overall, SAMs with densely packed helical OEG_n are considered to be better antifouling coatings in terms of both preventing protein adsorption and fouling resistance from blood plasma than SAMs with other hydrophilic moieties.⁵⁴ Typical protein adsorption from single-protein solutions and blood plasma onto SAM surfaces is in the range of 100 pg/mm²–3000 pg/mm², as measured by surface plasmon resonance (SPR) and ellipsometry.⁵⁴ 

Polymers represent one of the largest and the most diverse families in antifouling surface modifications.⁵⁴ Polyethylene glycol (PEG) has been a staple of the antifouling coating strategies. However, due to high solubility and strong interactions of linear PEG chains with water, direct adsorption of PEG alone to a surface is challenging and is not recommended, even though several groups reported using electrostatic or hydrophobic physisorption to immobilize PEG structural motifs on substrates.^212–215 Instead, most of the common modification strategies use covalent bonding to attach PEG-bearing moieties to surfaces (Fig. 8). Blättler et al. reported using plasma polymer deposition to graft comb copolymers comprising a polyelectrolyte backbone and PEG side chains such as poly(l-lysine)-g-poly(ethylene glycol) (PLL-g-PEG) to aldehyde plasma-modified substrates.²¹⁶ A triblock copolymer of PEG with poly(propylene oxide) (PPO)²¹⁷ or with poly(butadiene) (PB)²¹⁸ can be adsorbed on hydrophobic surfaces and chemically grafted by γ-irradiation. Zoulalian et al. reported grafting PEG–poly(alkyl phosphonate)–PEG (PEG–PAP–PEG) to TiO₂ and Nb₂O₅ surfaces via the PAP block and showed that these coatings reduced fouling in blood plasma.^219,220 Pop-Georgievski et al. prepared stable antifouling coating with terminal PEG groups that were bonded to reactive groups on substrate surfaces.²²¹ Reactive ω-trichloro silane-PEG (PEG-OSiCl₃)²²² and ω-mercapto-PEG (PEG-SH)^223–227 can be covalently bonded to silicon and gold surfaces, respectively. Similarly, PEG with a catecholic chelating group, such as dopamine (PEG–DOPA)^228–230 or anachelin (PEG–ANACH),²³¹ can bond to metal oxide surfaces by coordination bonds.^182,192,232 Carboxylic acid groups on alkanethiols terminating in either OEG or COOH groups can covalently bond to ω-amino-PEG.²³³ Note that all these PEG-SAM coatings inhibit blood plasma adsorption on the surfaces. PEG-functionalized surfaces prepared by plasma polymerization also demonstrated excellent antifouling activities, inhibiting the adsorption of BSA, IgG, and Fbg proteins.^234–236 

Multiple groups reported enhanced antifouling performance with increasing grafting density σ (number of polymer chains per nm²) and increasing length of the PEG chains.^{223,225,227,237–240} Increased PEG group density triggers a structural transformation from a mushroom regime characterized by isolated individual grafted chains to a brush regime with packed deformed polymer coils. Such polymer brushes form when the average polymer-chain anchoring point S (S⁻² = σ) distance approaches the size of the polymer in solution, given by the double Flory radius R_F of the polymer random coil (S = 2R_F).²⁴¹ At even higher grafting densities, the polymer coils will deform further.^241–243 We will discuss this behavior in more detail in Subsection II E 2. We note that not all PEG polymers can achieve high grafting density because of the strong excluded volume interactions between the chains. Even when the polymers can achieve high grafting density, such as PEG-SH, PEG–DOPA, and PEG–ANACH,^225,226,244 high osmotic pressure in aqueous solution will cleave the coordinate bonds between the coatings and the substrate, requiring additional chemical stabilization.²²⁹ 

Unsworth et al. reported an extensive investigation of gold PEGylation with R–PEG_x-SH compounds of different molecular weights and terminal groups R.^{223–226,245} Researchers reported that increasing PEG grafting density decreased fibrinogen adsorption from blood plasma.^225,226 Pop-Georgievski et al. reported coating gold and other inert substrates with poly(dopamine) and ω-amino-PEG to resist protein adsorption.²²¹ 

Polymer brushes are thick polymeric films made from closely packed polymer chains on the surface.^241,242,246 When those chains are grafted at a high density, proximity will force those chains into conformations different from the typical random coil conformations that polymer chains adopt in dilute solutions.^242,243 Polymer brushes are synthesized by “grafting-to” or “grafting-from” approaches. The “grafting-to” approach simply attaches polymer chains to the surface by physisorption,^202,247 chemisorption,^229,248 or covalent bonding.^223–227 The final structure can be characterized by grafting density ω, the number of polymer chains per nm² area.^241,242 As a general rule, higher polymer densities and thicker layers lead to better antifouling behavior: protein adsorption will decrease with increasing grafting density and increasing length of the polymer chains.^{223,225,227,237–239,249}

The “grafting-from” approach uses controlled polymerization to grow each polymer chain from a seed attached to the surface to produce dense polymer brushes. Common synthetic strategies include anionic and cationic polymerization,^250,251 nitroxide-mediated polymerization (NMP),²⁵² ring-opening polymerization,²⁵³ reversible addition–fragmentation transfer polymerization (RAFT),²⁵⁴ single-electron-transfer living radical polymerization (SET-LRP),²⁵⁵ and atom-transfer radical polymerization (ATRP).²⁵⁶ These reactions occur when the initiators or chain-transfer agents (CTAs) are exposed to monomer solutions.^242,246 Initiators, CTA, or precursor polymers can be attached to the substrate surface via specific chemical reactions.^257–263 Typically, additional chemistry steps are necessary to attach such seeds to chemically inert inactive substrates.^246,260,264 The surface-initiated ATRP method (SI-ATRP), a method where initiator SAMs are deposited by micro-contact printing (μCP) or dip-pen nanolithography, will coat surfaces with a designed pattern of initiator.^242,265 The length of the “grafted-from” polymer chain can then be controlled by controlling the polymerization time for SI-ATRP polymerization and can be used to fine-tune the brush thickness with nanometer precision.^266–269 

Ma et al. reported low-fouling surfaces of oligo(ethylene glycol) methacrylate (HOEGMA) polymer brushes with methacrylate backbone and OEG side chains prepared by SI-ATRP.^270,271 This type of antifouling polymer brushes can be grown on gold,^{199,266,270–273} silica/silicon,^260,274,275 TiO₂,²⁷⁶ polymeric nanocapsules,²⁶³ and various polymer surfaces.^264,269 Brown et al. reported a well-controlled SI-ATRP of pHOEGMA and pMeOEGMA brushes by optimizing the polymerization parameters.²⁷⁴ Rodriguez-Emmenegger et al. reported a 90% decrease in fouling from blood plasma and different body fluids on 30 nm-thick pHOEGMA and pMeOEGMA brushes prepared under the optimized conditions.^{268,269,277–279} Kizhakkedathu et al. reported using PEG-based N-substituted acrylamide from polystyrene and oligo(ethylene glycol) methyl ether acrylamide (pMeOEGA) from latex as an antifouling coating that was protective in undiluted blood plasma.^280,281

Methacryloyloxyethyl phosphorylcholine (MPC) polymer brushes, which form one of the most hydrophilic surfaces with a water contact angle of θ_adv = 18° and θ_rec = 7°,^233,282 can be prepared by SI-ATRP polymerization.^282–284 Rodriguez Emmenegger et al. reported using such pMPC brushes as antifouling coating in blood plasma.²³³ N-(3-sulfopropyl)-N-(methacryloxyethyl)-N,N-dimethylammonium sulfobetaine methacrylate (SBMA) based polymer brush can be used to create hydrophilic surfaces^285–288 that strongly repel foulant adsorption in single-protein solutions^233,289,290 but showed only limited antifouling performance in blood plasma challenge tests.^233,291

Zhang et al. reported using poly(carboxybetaine methacrylate) (pCBMA) with covalently attached receptors as antifouling coatings to protect from Fbg, human choriogonadotropin (hCG), and Lys protein adsorption.²⁵⁷ A follow-up work reported using carboxybetaine acrylamides (pCBAA) with a methylene (pCBAA-1), ethylene (pCBAA-2), propylene (pCBAA-3), and pentylene (pCBAA-5) spacer between the charged groups and carboxybetaine methacrylate with an ethylene spacer (CBMA-2) as an antifouling coating from Fbg adsorption in various ionic strengths and pH.²⁹² The main limiting factor for using pCBMA and pCBAA films remains the lack of control in polymer brush thickness.^269,278 Poly(CBMA-2) brushes were also prepared by the “grafting-to” approach to create an antifouling coating effective against undiluted blood plasma. Dopamine at one end of the polymer chain (DOPA-pCBMA) can be grafted to the SiO₂ film deposited on a gold layer or silicon microcantilevers on a suspended microchannel resonator.^248,261

Another group of studies created poly(ampholyte) brushes with alternating positively and negatively charged monomers.^{291,293–295} McCormick and Johnson reported preparation using charged monomers,²⁹⁵ copolymerized as ion pairs.^296,297 Bernards et al. reported the preparation of polymer brushes consisting of [2-(methacryloyloxy)ethyl] trimethylammonium chloride (pTMA) and 3-sulfopropyl methacrylate potassium salt (pSA) by ATRP.²⁹⁴ These surfaces exhibit antifouling activity against Fbg and Lys protein adsorption.^293,297,298

The SI-ATRP approach can also form hydrophilic polymer brushes based on hydroxy-contained monomers such as 2-hydroxyethyl methacrylate (HEMA),^{266–268,277,299,300} 3-hydroxypropylmethacrylate (HPM),^278,301 and N-(2-hydroxypropy) methacrylamide (HPMA)^{267,277,278,300,302} to create antifouling coatings. These brushes prevented fouling completely in single-protein solution challenge tests^{261,268,277,278,300,301} and also showed negligible fouling rates (below the SPR detection limit) for various biological fluids including cerebrospinal fluid, urine, saliva, fetal bovine and calf sera, eggs, and milk.²⁷⁷ The advantages of pHPMA brushes are their total resistance to undiluted blood plasma and long-term stability, while their downside is the lack of reliable synthetic control over their thickness. The hydroxyl groups in these brushes cannot be activated in a fully reversible manner, which also limits their practical applications in bioelectronic devices.

Researchers also grafted different polymer brushes such as peptoids,³⁰³ poly(β-peptoids),³⁰⁰ and polyoxazolines³⁰⁴ to surfaces to protect them from single-protein foulant solutions and undiluted blood plasma for a short contact duration. Saccharide polymer brushes³⁰⁵ and brushes of N-substituted acrylamide containing different carbohydrates³⁰⁶ also showed good antifouling behavior against single-protein solutions; however, their protection against blood plasma was not very good.

As a reminder, a surface is considered “antifouling” when it allows fouling levels of less than 1000 pg/mm².⁵⁴ PEG surfaces can achieve fouling levels as low as 20 pg/mm²,²²¹ whereas polymer brush surfaces, such as pHPMA, achieve blood plasma fouling levels as low as 0 pg/mm²–10 pg/mm².⁵⁴ 

The hybrid membranes, which combine several different approaches, represent a further evolution of antifouling strategy development efforts. Kwak et al. synthesized hybrid organic/inorganic reverse osmosis membrane with aromatic polyamide coating and titanium dioxide nanoparticles and showed that they result in antifouling activity.³⁰⁷ This system combined physical protection inherent to the membrane with the photocatalytic activity of TiO₂ nanoparticles, which produced reactive oxygen species such as hydroxyl radicals and hydrogen peroxide³⁰⁸ that attack bacterial membranes. The antifouling performance of this hybrid membrane was evaluated by the survival ratio of E. coli, with and without UV light exposure. The experiments showed that 3 hours of UV exposure led to only ∼10% of the E. coli cells surviving on the membrane, and 4 h of UV exposure led to complete sterilization.

Noy and co-workers demonstrated another example of hybrid membrane by constructing a synthetic membrane with carbon nanotube porins (CNTPs).^309,310 This system combined antifouling activity of phospholipid bilayers with the size selectivity and transport efficiency of carbon nanotube pores. CNTPs, 10 nm–20 nm CNT segments that can form transmembrane channels in biological membranes, mimics biological ion channels and can support efficient transport of water, protons, and small ions (Fig. 9). Chen et al.³¹¹ exploited highly efficient proton transport in small diameter carbon nanotubes to develop integrated silicon nanoribbon transistor pH sensors that were protected by the hybrid membrane that contained CNTP channels. Experiments demonstrated that these sensors were able to function in a variety of complex biological fluids such as BSA, simulated milk, and bovine plasma without performance degradation even after 72 h of exposure to foulants.³¹¹ 

Unlike most of the chemical and physical strategies described in Secs. II B 2–II B 7 and II C–II E, which mostly built protective barriers at the sensor surfaces, this approach mitigates fouling by reducing the concentration of interfering molecules in solution. These approaches go beyond “conventional” sample pre-purification techniques, such as liquid chromatography, which are time-consuming and equipment-intensive and often result in significant dilution of samples. One such approach relies on antibody-modified magnetic beads that capture and remove redundant proteins such as HSA and IgG in biological samples, resulting in reduced fouling propensity. Kongsuphol et al. reported the application of two types of magnetic beads (MBs) decorated with anti-HSA/anti-IgG and anti-TNF-α.³¹² The first kind of beads depleted HSA and IgG in serum, while the second beads captured TNF-α. Mei et al. developed a microfluidic platform that will allow mixing of low volume samples.³¹³ They used three types of depletion beads (anti-HSA, protein A, and protein G) to purify four human serum samples simultaneously and deplete up to 95% of IgG and HSA proteins in less than 10 min. While in some instances these approaches provide an elegant solution to the problem of surface fouling, they are by necessity restricted to small sample volumes and short sensing duration as longer exposures and larger volumes would overwhelm even the most efficient capture system.

This Research Update has focused on the antifouling strategies that largely emphasize surface modifications, which still are the most widely used and the most versatile approaches. However, these approaches are starting to run out of room in the quest to increase performance. Zwitterionic coatings are sensitive to pH and applied fields that can modulate the surface charge enough to encourage fouling. Peptide-based coatings are subject to hydrolysis and enzymatic degradation. PEG-based coatings may be vulnerable to oxidative damages in the presence of oxygen. The situation only gets worse when we use these systems in long-term, continuous fouling conditions, which, unfortunately, are the regimes closest to practical applications.

As we emphasized in Sec. II F, hybrid approaches, which combine surface modifications with complementary antifouling strategies, could provide significant antifouling benefits and enable the sensors to reach higher specificity and selectivity. Recent advances in materials science, such as the development of several broad classes of nanomaterials, controlled nanoporous solids (such as MOFs ad COFs), and even exotic designer materials, such as DNA origami, could provide new avenues for constructing efficient barriers for fouling and biofouling. Bio-inspired materials also show great potential in antifouling strategies. Biopolymers could be the next best antifouling material with exceptional stability, bio-compatibility, and environmental friendliness. Finally, as nanofabrication techniques continue to shrink the characteristic sizes of devices, we need to explore how these changes affect fouling and how we can exploit size reduction to minimize it.

X.C. acknowledges the fellowship support from the UC Office of the President through the UC Laboratory Fees Research Program under Grant No. LGF-18-545710. This work was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering. The work at LLNL was performed under the auspices of the U.S. Department of Energy under Contract No. DE-AC52-07NA27344. The work at the Molecular Foundry was supported by the Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

Data sharing is not applicable to this article as no new data were created or analyzed in this study.