Flexible, all-polymer microelectrode arrays for the ...

Viewer
Transcript

Author's self-archiving copy Biomaterials 32 (2011) 1778e1786

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Flexible, all-polymer microelectrode arrays for the capture of cardiac and neuronal signals Axel Blau a, *, 2, Angelika Murr b,1, 3, Sandra Wolff c, 4, Evelyne Sernagor d, 5, Paolo Medini a, 2, Giuliano Iurilli a, 2, Christiane Ziegler b, 3, Fabio Benfenati a, 2 a

Dept. of Neuroscience and Brain Technologies (NBT), Italian Institute of Technology (IIT), Via Morego 30, 16163 Genoa, Italy Dept. of Physics, University of Kaiserslautern, Erwin-Schroedinger-Strasse, 67663 Kaiserslautern, Germany NanoþBio Center, University of Kaiserslautern, Erwin-Schroedinger-Strasse, 67663 Kaiserslautern, Germany d Institute of Neuroscience, Newcastle University Medical School, Framlington Place, Newcastle upon Tyne NE2 4HH, United Kingdom b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 October 2010 Accepted 6 November 2010 Available online 9 December 2010

Microelectrode electrophysiology has become a widespread technique for the extracellular recording of bioelectrical signals. To date, electrodes are made of metals or inorganic semiconductors, or hybrids thereof. We demonstrate that these traditional conductors can be completely substituted by highly ﬂexible electroconductive polymers. Pursuing a two-level replica-forming strategy, conductive areas for electrodes, leads and contact pads are deﬁned as microchannels in poly(dimethylsiloxane) (PDMS) as a plastic carrier and track insulation material. These channels are coated by ﬁlms of organic conductors such as polystyrenesulfonate-doped poly(3,4-ethylenedioxy-thiophene) (PEDOT:PSS) or ﬁlled with a graphite-PDMS (gPDMS) composite, either alone or in combination. The bendable, somewhat stretchable, non-cytotoxic and biostable all-polymer microelectrode arrays (polyMEAs) with a thickness below 500 mm and up to 60 electrodes reliably capture action potentials (APs) and local ﬁeld potentials (LFPs) from acute preparations of heart muscle cells and retinal whole mounts, in vivo epicortical and epidural recordings as well as during long-term in vitro recordings from cortico-hippocampal cocultures. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Polymer electronics PEDOT:PSS PDMS Replica molding Neuroprosthetics Electrophysiology Neurotechnology

1. Introduction Over the past 40 years, passive microelectrode arrays (MEAs) have been widely applied to the in vitro study of bioelectrical signals from electrogenic cells such as neurons or cardiomyocytes [1]. In vivo, implantable electrodes and MEAs for prosthetics and rehabilitation [2,3] look back at a history of more than 50 years as transducers for cardiac pacemakers [4], neuromuscular electrical stimulation (NMES) [5], electrocorticography (ECoG) [6], deep brain stimulation (DBS) [7], cochlear [8] and retinal implants [9], peripheral nerve [10] and spinal cord [11] interfaces or, more generally, for brainemachine interfaces (BMIs) [12,13].

* Corresponding author. Tel.: þ39 010 71781 729; fax: þ39 010 71781 230. E-mail address: [email protected] (A. Blau). 1 Present address: Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany, www.mpip-mainz.mpg.de. 2 www.iit.it. 3 www.physik.uni-kl.de. 4 www.nbc.uni-kl.de. 5 www.ncl.ac.uk. 0142-9612/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2010.11.014

In all cases, the electrodes and interconnection tracks of MEAs or implantable bioelectronic devices are based on metals or inorganic (semi-)conductors. The main two reasons for this choice are their high conductivity and high-deﬁnition processability with classical microfabrication technologies such as photo or electron beam lithography, evaporation, etching, screen printing, micromachining or electroplating. However, most of these techniques not only require access to special equipment and clean room infrastructure, but also demand for device carrier substrates that withstand the respective processing conditions. These prerequisites impose limits onto device ﬂexibility, biocompatibility, biostability, transparency, or exploitation pathways for their cost-efﬁcient mass fabrication [2]. In particular, partially or fully rigid implants with sharp edges tend to damage the surrounding tissue during insertion and exert chronic stress onto the adjacent biological environment. Throughout their functional lifetime they are also prone to repositioning, thereby causing aversive reactions such as inﬂammation, encapsulation, rejection or break, which may eventually lead to device failure [14]. In consequence, because most biological tissue is soft and curvy, there is a steady trend to create more ﬂexible devices [11,15,16]. Various strategies have been suggested to

Author's self-archiving copy A. Blau et al. / Biomaterials 32 (2011) 1778e1786

overcome some of the above mentioned limitations, particularly in the context of designing cochlear, retinal or deep brain implants [3,17]. Recently, fabrication technologies have emerged which are more compatible with soft and ﬂexible carrier substrates, such as the printing of elastic conductors based on single-walled carbonnanotubes [18] or silver nanoparticles [19], the electroless platinum deposition onto polymers [20], or the contact printing of nanowire ﬁeld-effect transistor (NWFET) arrays onto ﬂexible substrates. While NWFETs have already been exploited as bioelectrical signal transducers, nanowire endings and carrier substrates still need to be metalized for establishing contact to external electronics [21]. Furthermore, most of these composites are mainly non-transparent, and material and fabrication costs are still considerably higher than those for most conductive polymers are. Synthetic metals [22,23] have recently seen a fast pace deveopment in electronics, mainly because they are cheap, ﬂexible and easily processable [24,25]. A large number of applications have been addressed, the most prominent examples being their use in FETs [26], light-weight solar panels [27], ﬂexible displays [28], or common consumer devices such as ﬂexible radiofrequency identity tags or smart labels. They are also becoming promising substitutes for metals in biomedical devices such as artiﬁcial sensors and actuators [29,30], laboratory bench [31] or implantable biosensors [32], bionic interfaces and biofeedback devices, in tissue engineering and as components for enhancing neural probes as covered by various reviews [33e35]. Yet, while insulating polymers are already accepted and widely used as implantable electrode carriers, electroconductive polymers have not yet completely replaced conventional conductor materials in biomedical devices. There is proof that they signiﬁcantly enhance the electrodeetissue interaction [36,37] and the electrical characteristics of conventional metal- [34,38] or semiconductor-based bioelectric signal transducers [39]; recently, their use in neuroprosthetics has been demonstrated [40,41]. However, while all-polymer electrochemical sensors [42] and PEDOT:PSS-based ion-pumps for ion homeostasis control in neurons [30] have already been described, a stand-alone polymer macroelectrode made of polypyrrole-polystyrenesulfonate has been tested for tissue integration [43], and diffuse electrical stimulation through polyaniline-based nanoﬁbrous scaffolds has been reported to have a beneﬁcial effect on cell proliferation and nerve outgrowth [44], there is currently no account on all-polymeric, tissue-like electrode arrays featuring conductive polymers as the only transduction element for the recording of bioelectrical signals in neuroprosthetics. Yet, recent advances in polymer material development and processing technologies propose a variety of new design and fabrication possibilities for generating stiff or ﬂexible, all-polymeric MEAs [42,45,46]. In this context, PEDOT:PSS is a favorite candidate due to its chemical and electrical long-term stability, its relatively low interfacial impedance and its high charge storage capacity for stimulation purposes [47] even under constant polarization conditions [48]. We therefore sought to develop a low-cost, all-polymer MEA fabrication strategy to equally suit diverse types of conductive and insulating polymers. Conceptually, it involves four basic steps as depicted in Fig. 1a. Step 1: The design of a replica-master in photoresist technology (Fig. 1a,) with two-level microstructures, which deﬁne the overall polyMEA geometries and the layout of electrodes, contact pads, and interconnection leads between the two. This master has to be produced only once because it can be replicated by a slightly modiﬁed replica-molding strategy used to generate the polyMEAs (Fig. 1a, 2r-8r, master replication cycle). Step 2: To produce a polyMEA from a master, the master microtopography is imprinted into an insulating polymer sheet of thickness that matches the highest protrusions of the master (Fig. 1a, 2f-3f). These will penetrate and perforate the entire sheet

1779

thereby deﬁning electrodes and pads, while lower microstructures will create buried microcavities that become interconnection tracks. Step 3: After detaching the sheet from its master (Fig. 1a, 4f5f) and placing it with its non-imprinted but partially perforated side onto a temporary handling carrier, all cavities are ﬁlled with a conductive polymer to either coat their surfaces as a thin ﬁlm or, in case of voluminous paste-like conductors, ﬁll the entire cavity volume (Fig. 1a, 6f-7f). Step 4: In a last step, the conducting polymer is covered by an insulating polymer ﬁlm (Fig. 1a and 8f), and the ready-to-use polyMEA is detached from its temporary carrier (Fig. 1a and 9f). 2. Materials and methods All experiments involving animals were carried out in accordance with the guidelines established by the European Communities Council (Directive of November 24th, 1986) and approved by the National Council on Animal Care of the Italian Ministry of Health.

2.1. Master fabrication and replica molding of master replication slabs or microchannel polyMEA scaffold sheets in PDMS The polyMEA fabrication by replica-molding is summarized in the Supplemental Figure S1. polyMEA fabrication was based on a replica-molding strategy from microstructured, two-level (bottom: 100 mm, top: 70 mm) SU-8-100 (MicroChem Corp., USA) master topographies on top of a 4-inch Si wafer with native oxide layer (p-doped, orientation <100>, thickness of 525 25 mm, Si-Mat, Germany), or from epoxy copies thereof. Bottom microstructure geometries were slightly larger in width than top geometries to compensate for mask alignment errors thus preventing undercuts. SU-8 masters were produced using standard microfabrication technologies following common SU-8 processing recipes. An inverse copy of SU-8 microtopographies could be created by placing a 50 50 mm2 plastic frame on the wafer, ﬁlling it with degassed PDMS (Sylgard 184, Dow Corning) to a height of 20 mm (to ensure later ﬁrmness and ﬂatness) and letting it cure (preferably 24 h at room temperature (RT) or alternatively 4 h at 65 C). For all steps involving PDMS, base and catalyst (Sylgard 184, Dow Corning) were mixed thoroughly in a10:1 ratio and degassed for 20 min in vacuum. After its detachment from the SU-8 master, cavities of the PDMS replication slab were ﬁlled with two-component epoxy (Marmo & Ferro, Pattex), avoiding air inclusions, and cured (24 h at RT or 1 h at 65 C). To facilitate polymer demolding from masters, a TeﬂonÔ-like anti-stick coating was deposited. SU-8 microstructures on silicon wafers with native oxide were coated by a ﬂuorineterminated layer generated from C4F8 in a reactive ion etcher (ECR-RIE, 2.46 GHz, MicroSys 350, Roth & Rau, Germany) using the following parameters: 15 sccm C4F8, at 2$103 mbar, 5 C substrate temperature, 550 W microwave power and 200 W RF power for 5 min. In contrast, epoxy master copies were placed in a desiccator and exposed to trichloro(1H,1H,2H,2H-perﬂuorooctyl)silane (Sigma 448931) for 1 h under constant vacuum, then tempered at 110 C for 2 h. Using the SU-8 masters with overall topography heights of mostly 200 mm, PDMS was spin-coated (Delta 80 Gyrset, Süss MicroTec, Germany) onto the wafer for 30 s at 1000 rpm to give layer thicknesses of around 150 mm. For epoxy master copies, degassed PDMS was directly poured onto the epoxy and leveled to the highest topographies by squeezing the PDMS by a stack made of a laser copier transparency, a 5 mm foam mat and a 3 mm polymethyl methacrylate (PMMA) square using a screw clamp or weights. In both scenarios, the PDMS was cured at RT for 24 before peeling it from the SU-8 master or its epoxy copy. If required, ethanol (95%) helped in separating interfacing structures from each other in demolding steps. The PDMS scaffold was placed onto a clean, non-coated laser copier transparency or polytetraﬂuoroethylene (PTFE) tray as a temporary carrier with its imprinted microchannels facing upwards. After its surface hydrophilization by oxygen-plasma treatment (femto, 2.45 GHz, Diener electronics, Germany) at RT, 0.3 mbar and 60 W for 3 min, the channels of the polyMEA scaffold were immediately ﬁlled with organic conductor material. Excess conductor was wiped off from the ﬂat surface on top of the channel-separating sidewalls with e.g., the edge of a microscope slide. To ensure homogeneous PEDOT:PSS ﬁlm formation on the microchannel surfaces of the scaffold [49], the solvent was allowed to evaporate for 2 h on a hotplate being slowly heated up to 50 C (30 min) before placing the conductor-loaded scaffold sheet into an oven, gradually raising the temperature from 50 C up to 125 C and leaving it there for a few minutes. If the solvents were tried to be removed in vacuum, or if the heating slope was too steep, partial dewetting of the narrow channel walls was observed. Thin, ﬂexible ﬁlms resulted from a single ﬁlling/tempering sequence. By repeating this sequence, thicker, yet more brittle, less transparent, and less homogeneous ﬁlms with higher conductivity could be produced. In case of gPDMS, scaffold sheets were kept at 120 C for less than 5 h. After cool-down, the exposed conductor on the backside of the polyMEA scaffold was globally insulated by a second layer of PDMS. For microchannels with their surfaces coated by PEDOT:PSS ﬁlms only, channels were

Author's self-archiving copy 1780

A. Blau et al. / Biomaterials 32 (2011) 1778e1786

Fig. 1. polyMEA design and fabrication concept. (a) Conceptual depiction of master replication (top, index r) and polyMEA fabrication therefrom (bottom, index f). The raised letters ‘p’ and ‘M’ represent electrodes or pad features of the polyMEA. Less elevated letters become the signal leads connecting the electrodes with respective contact pads within the polyMEA. The red lettering represents the original master microstructure in SU-8 or an epoxy copy thereof (1). Upon adding PDMS (2), curing it (3), and then removing the master or its copy (4), either the pattern-imprinted transfer substrate (thick PDMS block, 5r) or the micropatterned polyMEA top side insulation scaffold with through-holes at the electrodes and contact pads (thin PDMS sheet, 5f) are obtained. By ﬁlling the imprinted patterns of the thick PDMS transfer slab with epoxy, placing it onto a carrier tray (6r) and curing it (7r), a 1:1 master copy (1) can be produced. For fabricating polyMEAs, the thin PDMS scaffold (5f) is temporarily placed upside-down onto a non-stick carrier, and its cavities are ﬁlled with an electroconductive polymer (6f) to create electrodes, leads and contact pads (7f). After removal of the non-stick carrier and application of a backside insulation (8f), the polyMEA (9f) is ready for use. (b) Schematic polyMEA cross-section. (b1) polyMEA PDMS scaffold resting on a backside insulation slab. (b2) Zoom onto the basic geometry of a microchannel with its walls coated by a quasi-transparent ﬁlm of conductive polymer (blue). (b3) Cross-section through the polyMEA (b1) along the orange plane with the conductive polymer ﬁlm (blue) covering the microchannel walls. (c) Strategies for ﬁlling microchannels with polymer conductors. (c1) Microchannels can be coated with a thin ﬁlm of conductive PEDOT:PSS only (blue). (c2) To keep the electrical characteristics of PEDOT:PSS at the recording sites, but to increase mechanical stability of parts exposed to stress, critical sections such as pads can be ﬁlled with rubber-like gPDMS (black). (c3) Alternatively, the entire conductor track including the electrodes may be made of gPDMS. ﬁlled with PDMS as well. In a last step, the polyMEA was peeled from its carrier to expose electrodes and contact pads at its top side.

below 10 kU over a distance of 10 mm was achieved. gPDMS required slightly longer curing times than pure PDMS. A too high portion of graphite would impede complete curing of the composite.

2.2. Choice and processing of organic conductor materials The PEDOT:PSS dispersion (CPP105D, H.C. Starck, Germany) with about 1.2% solid content and particle diameters of approx. 80 nm was vacuum ﬁltered through a paper ﬁlter to remove polymer agglomerates. 5% (v/v) ethylene glycol (Sigma) was added as a conductivity enhancer. The gPDMS composite was based on graphite powder being blended into pre-mixed PDMS until a paste with a DC resistivity

2.3. Electrode impedance spectroscopy MEAs were placed in a Faraday box and their pads contacted by a gold-coated spring contact probe to measure electrode impedances in saturated KCl versus an Ag/AgCl reference electrode (Ø 1 mm wire) and a Pt-sheet (w2 3 mm2) as the

Author's self-archiving copy A. Blau et al. / Biomaterials 32 (2011) 1778e1786 counter electrode. Spectra were recorded by a potentiostat (Parstat 2273, Princeton Applied Research, USA). 2.4. Assessment of relative UV/VIS device transparency Overall spectral characteristics of PEDOT:PSS electrodes were determined spectroscopically (HR 2000 þ ES with CUV sample holder and SpectraSuite software, Ocean Optics; FO-6000 ﬁber optic halogen light source, World Precision Instruments). Using a razor blade, rectangles (10 30 mm2) were cut out of a PDMS slab (2 mm thickness) partially coated with PEDOT:PSS and out of a PEDOT:PSS polyMEA (200 mm thickness) consisting of the central electrode matrix or the pad area, respectively. They adhered to the sidewalls of 2.5 ml PMMA cuvettes (7591 05, Plastibrand), which were ﬁlled with deionized water after sample placement. The features of interest were positioned at the height of the passing light beam (Ø 3 mm). With no probe in the light path, light intensity was adjusted just below detector saturation. Sample spectra (3 ms integration, 10averaging) were darkcorrected and subtracted from the reference spectrum of a water-ﬁlled but otherwise empty cuvette. Spectra are depicted in Supplemental Figure S2. 2.5. 60-Channel MEA electrophysiology and imaging setup Extracellular signals were recorded and processed by a commercial 60-channel, 0.01 Hz high-pass ﬁlter-ampliﬁer data acquisition system (25 kHz sampling rate per channel) with a regulated heating platform (MEA60, Multi Channel Systems, Reutlingen, Germany), shielded by a grounded metal bowl with a central hole (Ø 5 mm) for LED illumination. If the on-chip polyMEA polymer conductor ground (GND) electrode was broken or too noisy, an external Pt wire (Ø 0.25 mm, <5 mm immersion length) served as GND. Imaging of the heartbeat was performed through a 5objective (Zeiss) by a webcam (2 megapixel, Trust) installed underneath the MEA. Still images of cells on polyMEAs were taken by a Canon Powershot G2 digital camera (4 megapixels) installed on a Zeiss Axiovert 200 microscope, ﬂuorescence images by an Optronics MicroFire camera (2 megapixels) installed on an Olympus BX 51. For in vivo recordings, a shielded 300 mm ﬂat ribbon cable with a home-built connector clamp based on the contact pads of modiﬁed commercial MEAs (Ayanda Biosystems, Switzerland) connected the pads of a triangular, 16 electrode polyMEA cut out to the ampliﬁer. The grounds of different electrophysiology setups were always separated to reduce noise. 2.6. Preparation and maintenance of retinal whole mounts Retinal recordings were performed using isolated retinal whole mounts from C57bl/6 neonatal mice (P9 and P11). As described elsewhere [50], mouse pups were killed by cervical dislocation and enucleated prior to retinal isolation. The isolated retina was then transferred to the experimental chamber and placed, retinal ganglion cell layer facing down, onto polyMEAs. Better coupling between the tissue and the electrodes was achieved by placing a small piece of polyester membrane ﬁlter (5 mm pores) (Sterlitech, Kent, WA, USA) on the retina followed by a slice anchor holder. Retinas were kept at 32 C and, using a custom-made PDMS perfusion lid with an integrated Pt GND wire, continuously perfused (2e5 ml/min) with artiﬁcial cerebrospinal ﬂuid (ACSF) containing the following (in mM): 118 NaCl, 25 NaHCO3, 1 NaH2PO4, 3 KCl, 1 MgCl2, 2 CaCl2, and 10 glucose, equilibrated with carbogen (95% O2 and 5% CO2). 2.7. Timed illumination of retinal whole mounts For stimulation of retinal whole mounts in vitro, light pulses were applied by LEDs (6000e16000 mcd, Conrad Electronic, Europkauf, Germany) emitting at different wavelengths (402 nm e IR). They were positioned above the center of the array at constant distance to the retina (40 mm) delivering diffuse light through transparent PDMS lids. Pulse duration (w1 s) and interval (w29 s) were controlled by a programmable relay card (C-Control, Conrad Electronic, Germany) as described previously [51]. To align the illumination period with biological activity, the switching voltage was sent through a variable voltage divider into a separate A/D channel of the MCS recording station. 2.8. Stimulus-induced epidural and epicortical recordings In vivo recordings were performed on urethane-anesthetized (0.8 ml/hg i.p., 20% solution) Long Evans rats aged P40. A 3 3 mm2 craniotomy was opened above the primary visual cortex (V1) corresponding to the binocular portion of the primary visual cortex (binocular area Oc1B) contralateral to the stimulus-exposed eye; the dura was left intact or carefully removed depending on experimental purposes. The eyes were protected by artiﬁcial tears (Lacrigel, Farmigea) and the eye bulb was positioned at about 55 from the vertical meridian by using metal rings. Visual stimulation was applied through the contralateral eye using a squared stationary grating alternating in counterphase (0.05 cycles/deg; 90% contrast, reversal rate 1 Hz, mean luminance 40 cd/m2) displayed on a 19” CRT placed about 250 mm in front of the animal. In some experiments, a full-ﬁeld ﬂash (50 ms duration, 1 Hz repetition rate) was used. Animal temperature was held constant (37 C) by

1781

a thermostatic blanket. Heart and breathing rates, as well as corneal and pinching reﬂexes were continuously monitored throughout the experiment. Procedural details have been described before [52]. The square-cut tip of the polyMEA with 16 connected electrodes was manually placed onto the dura of the brain and weighed down by a small PTFE ring (weight: 2 g, OD: 19 mm, height: 6.5 mm). A Pt wire (Ø 0.25 mm) placed subcutaneously served as GND electrode. The weight of the ribbon cable connecting the polyMEA probe to the ampliﬁer was supported by clamping it to a mechanical manipulator afﬁxed to the table. To measure the visually-evoked potential (VEP), up to 59 sweeps of 1990 ms duration were averaged and aligned to a TTL trigger pulse (w4 ms, 5 V) coinciding with grating reversals. As with the retinal whole mount recordings, the trigger signal indicating pattern reversal was fed through a variable voltage divider (w10x attenuation) into a separate A/D channel of the data acquisition system. 2.9. Cortico-hippocampal co-culture preparation and maintenance Mixed neural networks from dissociated cortical and hippocampal brain tissue of embryonic rats (E18) were kept in serum-free culture on polyMEAs and on commercial MEAs (Multi Channel Systems, Germany) as control. Preparation and culture maintenance have been described previously [53] following published protocols. The principle steps are summarized as follows: Before plating cells at a density above 15,000 cells/mm2, polyMEAs and MEAs had been autoclaved and plasma-hydrophilized (3e5 min, 0.3 mbar O2, 40e60 W at 2.46 GHz), been coated with laminin (10 mg/ml) and poly-L-lysine (PLL) (10 mg/ml), been allowed to dry, been washed with sterile ultrapure water to remove soluble PLL residues, and been allowed to dry again. The serum-free cell culture medium was based on NeurobasalÔ medium (NBM) (Gibco) containing B27 supplement (2%, Gibco), alanylglutamine (2 mM, Gibco) and penicillin/streptomycin (both 1 mM, Sigma). A glass ring attached with PDMS on top of each MEA served as a culture medium container and was covered by a transparent, gas-permeable PDMS lid to attenuate evaporation in long-term culture [53]. Cultures were kept in a humidiﬁed (92% RH) incubator at 37 C in a 5% CO2 atmosphere. Half of the medium was exchanged twice a week. During the 2 h recording sessions on the lab bench, the pH was allowed to drift. 2.10. Acute heart preparation Hearts were extracted from embryonic mice (E18) or rats (E18) and had been kept in cold (8 C) NBM for less than 40 min after decapitation. Hearts were placed onto dry, prewarmed MEAs without any adhesion mediator, and been wetted from top by a drop of NBM. A platinum wire (Ø 0.25 mm) was placed into the drop for grounding. While warming up to 36 C, beating set in spontaneously. Extracellular activity was recorded without additional stimulation. Heartbeats were visible to the eye and video-imaged for their later synchronization with the electrical activity by a webcam underneath the MEA. A Supplemental Movie shows heartbeat-correlated activity traces. Supplementary data associated with this article contains a movie showing an exemplary heartbeat-correlated electrical activity trace can be found in the online version, at doi:10.1016/j.biomaterials.2010.11.014.

3. Results and discussion 3.1. Properties of polymer-embedded conductive polymer microelectrodes The 60 electrode polyMEAs were composed of PDMS as a ﬂexible, transparent, water-impermeable scaffold substrate and electrical insulation material. Electrode diameters were below 120 mm (70 mm on the mask design), electrode spacing 400 mm with exception of an 800 mm cross-shaped gap in the center of the polyMEA(Fig. 2, a2; Fig. 4, b & c2). PEDOT:PSS and gPDMS were used as conductors with complementary properties. A commercial PEDOT:PSS formulation (Baytron CPP105D, H.C. Starck), optimized for scratch-resistant adhesion to glass, formed thin, bluish-transparent ﬁlms on PDMS after short thermal treatment (t 5 min, 80 C T 125 C) (Fig. 2, a,c and d; Fig. 4b). As for all PEDOT:PSS formulations, conductivity increased with ﬁlm thickness at the cost of transparency and ﬁlm stability. Conductivity was further enhanced by addition of 5% (v/v) ethylene glycol (EG) [54]. However, PEDOT:PSS ﬁlms on soft PDMS with a Shore A grade of 50 (Sylgard 184, Dow Corning) tended to break upon locally applied pressure e.g., by the ampliﬁer contact pins or when bending the polyMEA beyond 70 . Thick ﬁlms and less homogeneous ﬁlms on non-plasma-treated PDMS cracked more

Author's self-archiving copy 1782

A. Blau et al. / Biomaterials 32 (2011) 1778e1786

Fig. 2. Properties of polyMEAs. Figures (aec) depict some of the basic features of replica-molded polyMEAs. (a1) polyMEA (49 49 mm2) with a multilayer ﬁlm of bluish, yet still transparent, conductive PEDOT:PSS þ 5% EG polymer inlays. A glass ring serves as cell culture container. Bar: 1 mm. Zoom onto the central matrix of the 60 quasi-transparent electrodes and leads (red dotted rectangle; bar: 1 mm) in (a2) and onto ﬁve contact pads (green square; bar: 1 mm) in (a3) of a sibling polyMEA depicted in (a1) with a printed piece of paper in the background to demonstrate transparency. (b) Demonstration of ﬂexibility of a polyMEA, here with non-transparent gPDMS conductors. Bar: 1 mm. (c) The pipe-like cross- section of one electrode and its partial conductor track reveals the concept of coating the walls of microchannels with a thin ﬁlm of the electroconductive PEDOT:PSS polymer. In case of gPDMS, these channels are completely ﬁlled (not shown). Bar: 100 mm. (d) Cortico-hippocampal network (rat) after 38 DIV on a polyMEA with PEDOT:PSS þ 5%EG conductor: The recording electrodes and interconnection tracks are quasi-transparent, thus facilitating the imaging of cells on top of them (red arrows). PDMS and the organic conductor did not have any detectable impact on neural differentiation and electrophysiology. Bar: 100 mm. (e) Characteristic network morphology of a cortico-hippocampal network cultured for 21 DIV on a gPDMS polyMEA demonstrates long-term biological inertness of the gPDMS conductor. Electrode positions are pointed out by small white circles. Merged antibody staining by beta-tubulin III - Alexa 488 (green: dendrites & axons), MAP2 - Alexa 568 (red: dendrites) and Hoechst stain (blue: nuclei). Bar: 100 mm.

easily. In contrast, thin PEDOT:PSS ﬁlms with or without conductivity enhancer seemed irreversibly fused with plasma-treated PDMS surfaces. To further render the contact pads and parts of the leads more resistant to mechanical stress, we ﬁlled the respective cavities with black gPDMS, which had almost the same rubbery consistency and durability as the PDMS backbone after its curing. Thus, the central electrode area and leads, which - probably due to their smaller geometries - turned out to be less vulnerable to mechanical stress, stayed transparent, while the edges became fortiﬁed by gPDMS. To compare the recording performance of the two electroconductive materials, we also produced polyMEAs with gPDMS as the only conductor. While transparency was now partially obstructed by the gPDMS, these polyMEAs had exceptionally high mechanical stability. gPDMS polyMEAs could be stretched by 30e50%, bent by almost 180 or be exposed to punctual pressure without any mechanical or electrical damage. Initially conceptualized as disposable devices, all polyMEA types turned out to be easily recyclable without substantial change in electrical properties and recording functionality. In the case of gPDMS, exposed conductor areas could be easily ‘refreshed’ by superﬁcial abrasion with 600e1000 grit sanding paper. This did not only help in removing any production-related PDMS ﬁlms or biological debris, but also exposed a fresh matrix-embedded graphite surface underneath.

3.2. Performance of polymer microelectrodes All utilized materials tolerated diverse sterilization procedures (e.g., autoclavation, 70% ethanol, oxygen-plasma) without noticeable deterioration. Their disposal is risk-free, not containing any hazardous substances. Lack of long-term cytotoxicity and aversive tissue-reactivity was validated with cortico-hippocampal networks up to 63 days in vitro (DIV). On average, the cultures showed typical network morphologies (Fig. 2, d and e) and survived as long as control cultures on commercial non-polymer MEAs. For PEDOT:PSS electrodes, the electrodeecell interface stayed optically transparent for concurrent morphology studies by light microscopy (Fig. 2d; Fig. 4, b & c2). Transparency dropped almost homogeneously over the entire visible spectrum (400e900 nm) to about 60% for 2 mm thick PDMS slabs, and to about 20% for PDMS coated with a thin PEDOT:PSS þ 5%EG ﬁlm (see Supplemental Figure S2). The thickness of the tested prototype polyMEAs was deﬁned by the overall height of the two-layered master microstructure (here < 200 mm) and the freely selectable thickness of the backside insulation (<500 mm). The absolute impedance of different PEDOT:PSS-based electrodes with diameters below 120 mm varied between 400 kU and 4 MU, being almost constant over the entire physiologically relevant frequency range between 1 Hz and 1 kHz (Fig. 3, a1). At 1 kHz (relevant for capturing action potentials [APs];

Author's self-archiving copy A. Blau et al. / Biomaterials 32 (2011) 1778e1786

1783

Fig. 3. Electrode impedances, electrode noise, and recording characteristics of polyMEAs. (a) MEA impedance characteristics. (a1) Comparison of the absolute impedance as a function of frequency in a Bode magnitude plot for Ø 120 mm polyMEA PEDOT:PSS þ5 %EG electrodes (blue; n ¼ 3 polyMEAs; m ¼ 84 electrodes), PEDOT:PSS þ 5%EG electrodes connected to gPDMS connection pads (green; n ¼ 1; m ¼ 27), metal-like gPDMS electrodes (grey; n ¼ 3; m ¼ 57) and Ø 30 mm or 10 mm TiN-on-Au or ITO electrodes (yellow; n ¼ 3, m ¼ 15) of commercial MEAs (30 or 10/200iR, Multi Channel Systems). Central lines depict the averages of m electrodes of n different MEAs of the same kind, vertical bars the minmax ranges delimited by dotted or dashed border lines. Individual impedance values at 1 kHz are pointed out by colored stars on a vertical grey line: yellow: 35 kU (commercial MEAs); black: 550 kU (gPDMS polyMEAs); green: 1.2 MU (polyMEAs with PEDOT:PSS þ 5% EG electrodes and gPDMS pads); blue: 1.5 MU (PEDOT:PSS þ 5% EG polyMEAs). (a2) Comparison of the phase as a function of frequency (color association as in a1). (b, c) Comparison of signal components and peak-to-peak recording noise in non-averaged in vitro recordings (b) from rat cortico-hippocampal neurons (b1 @ 29 DIV; b2 @ 39 DIV) and epidural in vivo recordings (c) in a 40-days-old rat with PEDOT:PSS þ 5% EG electrodes (b1, b2 and c: orange trace) or gPDMS electrodes (c: green trace) with either external Pt wire GND (b1, c) or on-chip PEDOT:PSS þ 5% EG counter electrodes as GND (b2). Stars in (b) point out individual action potentials, dotted lines local ﬁeld components. ECG-like ﬁeld potential recordings in (c) were digitally low-pass (200 Hz) and band-stop (50 Hz, Q ¼ 10) ﬁltered.

see vertical grey line and colored stars in Fig. 3, a1), their average absolute impedance was about 40 times higher when compared to metal electrodes, but about 15 times lower at 1 Hz (relevant for local ﬁeld potentials [LFPs]). In contrast, graphite-based electrodes showed an exponential increase in impedance towards low frequencies, which is typical for rough metal electrodes. As the phase plots indicate (Fig. 3, a2), the PEDOT:PSS interface behaved mostly resistive while gPDMS was dominated by capacitive elements. Peakto-peak noise depended on both the quality of individual conductor tracks and on the type of ground electrode. For PEDOT:PSS þ 5% EG, the noise stayed below 30 mVpp for external Pt wire grounding (Ø 0.25 mm, 5 mm insertion) (Fig. 3, b1 and Fig. 4, b & c1). It was about three times higher (80 mVpp, Fig. 3, b2) when the on-chip polymer counter electrode was used instead (w100 larger geometrical area compared to recording electrodes). For gPDMS, the peak-to-peak noise was almost identical for both external and internal ground (GND) electrodes and reached down to 15 mVpp. polyMEA recording performance was investigated in various scenarios. Despite the notably lower conductivity of polymer conductors with respect to metals (PEDOT:PSS with conductivity enhancers: <104 S m1; gPDMS: <106 S m1; metals: >106 S m1), which would not make them the ﬁrst choice for capturing tiny bioelectrical signals, signal-to-noise ratios (S/N) exceeded 100:1 for cardiac action potentials (Fig. 4a) and were typically between 2:1 and 5:1 in extracellular in vitro recordings of neuronal signals (Fig. 3b; Fig. 4, b & c1). Cardiac activity with typical signal shapes and components (QRS-complex, T and possibly U waves) was recorded

from acute in vitro preparations from embryonic hearts (mice, E18; rats, E18) (Fig. 4a and Supplemental Movie). Timed light stimulation of acute in vitro preparations of retinal whole mounts (mouse, P9; Fig. 4, c2) induced light-activated on-off responses at 402 nm (Fig. 4, c1) [55]. Neural networks of dissociated cortico-hippocampal tissue (rat, E18) showed spontaneous activity over a period of up to 63 days (38 DIV shown in Fig. 4b) that could not be distinguished from that of control networks cultured on commercial MEAs (Au/TiN, 200/ 30iR, Multi Channel Systems). Individual action potentials and low frequency components could be recorded simultaneously and be separated upon digital ﬁltering. We ﬁnally tested the capability of polyMEAs to capture sensory-driven synaptic activity from neocortical tissue in vivo. LFP responses to patterned visual stimuli were reliably recorded from the primary visual cortex of rats (n ¼ 3, P40) in both epidural and epicortical conﬁgurations (Fig. 4d). Importantly, visual LFP activation latencies were comparable with data available in the literature for this same species (e.g., [56]). 3.3. Outlook on fabrication alternatives and application prospects for polymer microelectrode arrays The chosen fabrication concept allows for variations in the choice of materials (e.g., UV-curable polymers) and in the general production design. For instance, alternative, equally non-serial fabrication routes for the microstructuring of the insulator scaffold such as hotembossing of (metal-)masters into polymers can be pursued. Array properties such as stiffness, transparency, hydrophilicity, and

Author's self-archiving copy 1784

A. Blau et al. / Biomaterials 32 (2011) 1778e1786

Fig. 4. polyMEA in vitro and in vivo recordings. (a) Acute high S/N myocardiograms (mouse, E18) recorded by polymer microelectrodes (blue: PEDOT:PSS þ 5%EG; red: gPDMS) and overlaid with a cardiac action potential from a rat heart (E18) picked up by a TiN-coated Au electrode of a commercial MEA (green): The typical signal structuring (Q, R, S and T) is present in all recordings; the polymer electrodes also picked up an additional low frequency component, which might possibly be identiﬁed as the U wave. (b) Recording characteristics of a PEDOT:PSS þ 5%EG polyMEA in an exemplary in vitro recording from a co-cultured cortico-hippocampal neural network with mixed monolayer and cluster architecture after 38 DIV (rat, E18). Signal traces of 1 s duration were overlaid onto the bright-ﬁeld microscopy image and placed in the vicinity of their respective recording electrodes. The 640 ms zoom-ins onto three signal traces demonstrate how activity travels from the upper right area of the array to the lower central area (with respect to the vertical dotted grey line) and the range of signal types picked up by the polymer electrodes: Action potentials (upper and middle trace) as well as low frequency components can be resolved (middle and lower trace). Colored boxes frame the signal traces, colored circles the respective recording electrodes. Bar: 400 mm. Circle diameters: 150 mm. (c) Exemplary 1 min extract of spontaneous and light-induced activity in (c1) from the ganglion layer of a retinal whole mount (mouse, P9) placed onto a PEDOT:PSS þ 5%EG polyMEA in (c2) with gPDMS contact pads (not visible). Bar: 500 mm. Somewhat delayed retinal on-off bursting response to timed diffuse light (402 nm, w1 s duration) and spontaneous activity during dark phases (w29 s) are distinguishable. Violet bars: light on; grey bars: light off. Square wave: attenuated and recorded relay switching voltage (30 mV, not to scale). Signal shapes and baseline during light-on-bursting response (violet dotted box) vs. spontaneous singleespike activity during dark periods (grey box). (d) Epidural in vivo recordings in (d1) from the visual cortex (d2) of an anesthetized rat (P40). Light-stimulated responses to an alternating (1 Hz) grating stimulus after 59 averaging captured by one out of 16 electrodes of a polyMEA with either PEDOT:PSS þ 5% EG electrodes (upper traces) or gPDMS electrodes (lower traces). The recording electrode for the gPDMS trace is pointed out by a white circle in (d2). Blue traces: ‘raw’ (no ﬁlter); red traces: digitally low-pass (200 Hz, Butterworth 2nd order) and band-stop (50 Hz, Q ¼ 10) ﬁltered signals. Red arrows indicate the 1 Hz pattern switch. Bar: 1 mm.

surface chemistry can be tuned by variations in matrix materials and physical or chemical post-processing steps. Other types of wetting or moldable electrode materials might be chosen including conducting gels or ionic liquids. Electrode properties can be further modiﬁed by electrodeposition of carbon-nanotubes for functional surface nanotexturing or electropolymerization of other, more fractal, conductive polymers. With rather straight-forward design adaptations, several arrays could be stacked to increase the density of signal recording sites, or protruding spike-like 3D electrodes be generated.

Finally, due to its conceptual vicinity to microﬂuidic devices, the integration of chemical stimulation paradigms can be envisioned. Growth-factors may be directly embedded into a modiﬁed, spongelike electrode matrix for their slow release over time to enhance deviceetissue interaction, tissue integration, and tissue regeneration. Alternatively, neuromodulatory compounds or drugs may be passed through the hollow electrode columns for their release at the end of the interconnection channels. For the same reason, these arrays may also ﬁnd application as multisite patch arrays after

Author's self-archiving copy A. Blau et al. / Biomaterials 32 (2011) 1778e1786

further reducing the diameters of the electrode apertures. For optoelectrical stimulation, pixels of optically addressable photovoltaic polymer semiconductors could be incorporated in currently unused areas of the array. 4. Conclusions The presented all-polymer microelectrode arrays feature biophysical and optical properties difﬁcult to realize with classical conductor materials and standard microprobe fabrication technologies. The concept may help in designing a new generation of less invasive, soft and curvature-adaptive in vivo probes (e.g., in the context of cortical surface ﬁeld potential recordings, deep brain implants, pacemakers, and cochlear or retinal implants). It can be foreseen that - similarly to the recent shift from silicon to organic electronics in consumer devices - conductive polymers may complement if not substitute classical conductor materials in future biomedical and neuroprosthetic transducers. Acknowledgements The authors thank Tanja Neumann and Simone Riedel, formerly at the University of Kaiserslautern, Mariateresa Tedesco at the University of Genoa as well as Francesca Succol and Marina Nanni at IIT for their expert advice and assistance in tissue and cell culture preparation. We thank Christian Dauterman and Stefan Trellenkamp at the NBC of the University of Kaiserslautern for their help in the production of the SU-8 masters. Many thanks also to Laura Gasparini at IIT for the antibody staining of the cultures, to Francesco Difato for imaging support and to Mario Cerino for exploring PDMS release strategies. Appendix. Supplementary data Supplementary data associated with this article including a schematic summary of master generation and polyMEA fabrication steps, and UV/VIS spectra, can be found in the online version, at doi:10.1016/j.biomaterials.2010.11.014. References [1] Pine J. A history of MEA development. In: Taketani M, Baudry M, editors. Advances in network electrophysiology using multi-electrode arrays. New York: Springer; 2006. p. 3e23. [2] Stieglitz T. Neural prostheses in clinical practice: biomedical microsystems in neurological rehabilitation. Acta Neurochir Suppl 2007;97(1):411e8. [3] Cheung KC. Implantable microscale neural interfaces. Biomed Microdevices 2007;9(6):923e38. [4] Curtis AB. Fundamentals of cardiac pacing 1st ed. Jones Bartlett Publishers; 2009. [5] Shefﬂer LR, Chae J. Neuromuscular electrical stimulation in neurorehabilitation. Muscle Nerve 2007;35(5):562e90. [6] Kim DH, Viventi J, Amsden JJ, Xiao J, Vigeland L, Kim YS, et al. Dissolvable ﬁlms of silk ﬁbroin for ultrathin conformal bio-integrated electronics. Nat Mater 2010;9:511e7. [7] Coffey RJ. Deep brain stimulation devices: a brief technical history and review. Artif Organs 2009;33(3):208e20. [8] Wilson BS, Dorman MF. Interfacing sensors with the nervous system: Lessons from the development and success of the cochlear implant. IEEE Sensors J 2008;8(1):131e47. [9] Margalit E, Maia M, Weiland JD, Greenberg RJ, Fujii GY, Torres G, et al. Retinal prosthesis for the blind. Surv Ophthalmol 2002;47(4):335e56. [10] Navarro X, Krueger TB, Lago N, Micera S, Stieglitz T, Dario P. A critical review of interfaces with the peripheral nervous system for the control of neuroprostheses and hybrid bionic systems. J Peripher Nerv Syst 2005;10 (3):229e58. [11] Rodger DC, Fong AJ, Li W, Ameri H, Ahuja AK, Gutierrez C, et al. Flexible parylene-based multielectrode array technology for high-density neural stimulation and recording. Sens Actuators B 2008;132(2):449e60. [12] Donoghue JP. Bridging the brain to the world: a perspective on neural interface systems. Neuron 2008;60(3):511e21.

1785

[13] Stieglitz T, Rubehn B, Henle C, Kisban S, Herwik S, Ruther P, et al. Braincomputer interfaces: an overview of the hardware to record neural signals from the cortex. In: Verhaagen J, Hol EM, Huitenga I, Wijnholds J, Bergen AB, Boer GJ, et al., editors. Prog brain res. Elsevier; 2009. p. 297e315. [14] Marin C, Fernandez E. Biocompatibility of intracortical microelectrodes: current status and future prospects. Front Neuroeng 2010;3(8):5. [15] Molina-Luna K, Buitrago MM, Hertler B, Schubring M, Haiss F, Nisch W, et al. Cortical stimulation mapping using epidurally implanted thin-ﬁlm microelectrode arrays. J Neurosci Methods 2007;161(1):118e25. [16] Rubehn B, Bosman C, Oostenveld R, Fries P, Stieglitz T. A MEMS-based ﬂexible multichannel ECoG-electrode array. J Neural Eng 2009;6(3):036003. [17] Adams C, Mathieson K, Gunning D, Cunningham W, Rahman M, Morrison JD, et al. Development of ﬂexible arrays for in vivo neuronal recording and stimulation. Nucl Instrum Methods Phys Res Sect A 2005;546 (1e2):154e9. [18] Sekitani T, Nakajima H, Maeda H, Fukushima T, Aida T, Hata K, et al. Stretchable active-matrix organic light-emitting diode display using printable elastic conductors. Nat Mater 2009;8(6):494e9. [19] Ahn BY, Duoss EB, Motala MJ, Guo X, Park SI, Xiong Y, et al. Omnidirectional printing of ﬂexible, stretchable, and spanning silver microelectrodes. Science 2009;323(5921):1590e3. [20] Dupas-Bruzek C, Drean P, Derozier D. Pt metallization of laser transformed medical grade silicone rubber: last step toward a miniaturized nerve electrode fabrication process. J Appl Phys 2009;106(7):074911e5. 074913. [21] Timko BP, Cohen-Karni T, Qing Q, Tian B, Charles M. Lieber. Design and implementation of functional nanoelectronic interfaces with biomolecules, cells, and tissue using nanowire device arrays. IEEE Trans Nanotech 2010;9 (3):269e80. [22] Hush NS. An overview of the ﬁrst half-century of molecular electronics. Ann N Y Acad Sci 2003;1006:1e20. [23] Inzelt G. Conducting polymers. A new era in electrochemistry. Berlin Heidelberg: Springer; 2008. [24] Malliaras G, Friend R. An organic electronics primer. Phys Today 2005;58 (5):53e8. [25] Forrest SR. The path to ubiquitous and low-cost organic electronic appliances on plastic. Nature 2004;428(6986):911e8. [26] Wen Y, Liu Y. Recent progress in n-channel organic thin-ﬁlm transistors. Adv Mater 2010;22(12):1331e45. [27] Helgesen M, Søndergaard R, Krebs FC. Advanced materials and processes for polymer solar cell devices. J Mater Chem 2010;20(1):36e60. [28] Choi M-C, Kim Y, Ha C- S. Polymers for ﬂexible displays: from material selection to device applications. Prog Polym Sci 2008;33(6):581e630. [29] Brochu P, Pei Q. Advances in dielectric elastomers for actuators and artiﬁcial muscles. Macromol Rapid Commun 2010;31(1):10e36. [30] Isaksson J, Kjall P, Nilsson D, Robinson ND, Berggren M, Richter-Dahlfors A. Electronic control of Ca2þ signalling in neuronal cells using an organic electronic ion pump. Nat Mater 2007;6(9):673e9. [31] Lin P, Yan F, Yu J, Chan HLW, Yang M. The application of organic electrochemical transistors in cell-based biosensors. Adv Mater 2010;22(33):3655e60. [32] Malhotra BD, Chaubey A, Singh SP. Prospects of conducting polymers in biosensors. Anal Chim Acta 2006;578(1):59e74. [33] Rogers JA, Someya T, Huang Y. Materials and mechanics for stretchable electronics. Science 2010;327(5973):1603e7. [34] Guimard NK, Gomez N, Schmidt CE. Conducting polymers in biomedical engineering. Prog Polym Sci 2007;32(8e9):876e921. [35] Berggren M, Richter-Dahlfors A. Organic bioelectronics. Adv Mater 2007;19 (20):3201e13. [36] Green RA, Lovell NH, Poole-Warren LA. Cell attachment functionality of bioactive conducting polymers for neural interfaces. Biomaterials 2009;30 (22):3637e44. [37] Poole-Warren L, Lovell N, Baek S, Green R. Development of bioactive conducting polymers for neural interfaces. Expert Rev Med Devices 2010;7 (1):35e49. [38] Abidian MR, Corey JM, Kipke DR, Martin DC. Conducting-polymer nanotubes improve electrical properties, mechanical adhesion, neural attachment, and neurite outgrowth of neural electrodes. Small 2010;6(3):421e9. [39] Bystrenova E, Jelitai M, Tonazzini I, Lazar AN, Huth M, Stoliar P, et al. Neural networks grown on organic semiconductors. Adv Funct Mater 2008;18 (12):1751e6. [40] Simon DT, Kurup S, Larsson KC, Hori R, Tybrandt K, Goiny M, et al. Organic electronics for precise delivery of neurotransmitters to modulate mammalian sensory function. Nat Mater 2009;8(9):742e6. [41] Guiseppi-Elie A. Electroconductive hydrogels: synthesis, characterization and biomedical applications. Biomaterials 2010;31(10):2701e16. [42] Kafka J, Larsen NB, Skaarup S, Geschke O. Fabrication of an all-polymer electrochemical sensor by using a one-step hot embossing procedure. Microelectron Eng 2010;87(5e8):1239e41. [43] George PM, Lyckman AW, LaVan DA, Hegde A, Leung Y, Avasare R, et al. Fabrication and biocompatibility of polypyrrole implants suitable for neural prosthetics. Biomaterials 2005;26(17):3511e9. [44] Ghasemi-Mobarakeh L, Prabhakaran MP, Morshed M, Nasr-Esfahani MH, Ramakrishna S. Electrical stimulation of nerve cells using conductive nanoﬁbrous scaffolds for nerve tissue engineering. Tissue Eng Part A 2009;15 (11):3605e19.

Author's self-archiving copy 1786

A. Blau et al. / Biomaterials 32 (2011) 1778e1786

[45] Hansen TS, West K, Hassager O, Larsen NB. Direct fast patterning of conductive polymers using agarose stamping. Adv Mater 2007;19(20):3261e5. [46] Choi JS, Cho KY, Yim J-H. Micro-patterning of vapor-phase polymerized poly (3,4-ethylenedioxythiophene) (PEDOT) using ink-jet printing/soft lithography. Eur Polym J 2010;46(3):389e96. [47] Wilks SJ, Richardson-Burns SM, Hendricks JL, Martin DC, Otto KJ. Poly(3,4ethylene dioxythiophene) (PEDOT) as a micro-neural interface material for electrostimulation. Front Neuroeng; 2009:2. [48] Rozlosnik N. New directions in medical biosensors employing poly(3,4-ethylenedioxy thiophene) derivative-based electrodes. Anal Bioanal Chem 2009;395(3):637e45. [49] Okerberg BC, Berry BC, Garvey TR, Douglas JF, Karim A, Soles CL. Competition between crystallization and dewetting fronts in thin polymer ﬁlms. Soft Matter 2009;5(3):562e7. [50] Hennig MH, Adams C, Willshaw D, Sernagor E. Early-stage waves in the retinal network emerge close to a critical state transition between local and global functional connectivity. J Neurosci 2009;29(4):1077e86.

[51] Neumann T, Ziegler C, Blau A. Multielectrode array recordings reveal physiological diversity of intrinsically photosensitive retinal ganglion cells in the chick embryo. Brain Res 2008;1207:120e7. [52] Pizzorusso T, Medini P, Landi S, Baldini S, Berardi N, Maffei L. Structural and functional recovery from early monocular deprivation in adult rats. Proc Natl Acad Sci U S A 2006;103(22):8517e22. [53] Blau A, Neumann T, Ziegler C, Benfenati F. Replica-molded poly(dimethylsiloxane) culture vessel lids attenuate osmotic drift in long-term cell culturing. J Biosci 2009;34(1):59e69. [54] Wang T, Qi Y, Xu J, Hu X, Chen P. Effects of poly(ethylene glycol) on electrical conductivity of poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonic acid) ﬁlm. Appl Surf Sci 2005;250(1e4):188e94. [55] Ekesten B, Gouras P, Yamamoto S. Cone inputs to murine retinal ganglion cells. Vision Res 2000;40(19):2573e7. [56] Pizzorusso T, Fagiolini M, Porciatti V, Maffei L. Temporal aspects of contrast visual evoked potentials in the pigmented rat: effect of dark rearing. Vision Res 1997;37(4):389e95.

Author's self-archiving copy

Supporting Information Flexible, all-polymer microelectrode arrays for the capture of cardiac and neuronal signals Axel Blau a*, Angelika Murr b1, Sandra Wolff c, Evelyne Sernagor d, Paolo Medini a, Giuliano Iurilli a, Christiane Ziegler b, Fabio Benfenati a a

Dept. of Neuroscience and Brain Technologies (NBT), Italian Institute of Technology (IIT), Via Morego 30, 16163 Genoa, Italy, www.iit.it

b

Dept. of Physics, c Nano+Bio Center, University of Kaiserslautern, Erwin-Schroedinger-Strasse, 67663 Kaiserslautern, Germany, www.physik.uni-kl.de, www.nbc.uni-kl.de

d

Institute of Neuroscience, Newcastle University Medical School, Framlington Place, Newcastle upon Tyne NE2 4HH, United Kingdom, www.ncl.ac.uk

*

Corresponding author. Dept. of Neuroscience and Brain Technologies (NBT), The Italian Institute of Technology (IIT), Via Morego 30, 16163 Genoa, Italy. Tel.: +39 010 71781 729; fax.: +39 010 71781 230 E-mail address: [email protected]

1

Present address: Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany, www.mpip-mainz.mpg.de

S1

Author's self-archiving copy

1. Supplemental Figures

Supplemental Figure S1. Summary of master generation and polyMEA fabrication steps.

S2

Author's self-archiving copy

Supplemental Figure S2. Spectral characteristics of pure PDMS, and of PDMS completely or partially coated with PEDOT:PSS. (a) Transparency of cured PDMS (red trace, ≥ 60%) and PDMS coated with PEDOT:PSS (blue trace, ≤ 30%) in water-filled PMMA cuvettes with respect to a water-filled cuvette as the reference (grey trace, set to 100%). The inset photograph depicts the samples placed on a printed piece of paper (font size: 5 pt). The blue and red circles indicate diameter (3 mm) and position of the passing light beam. (b) Transparency of a clear PDMS sample (red trace, ≥ 55%), of the central electrode matrix of a PEDOT:PSS polyMEA (green trace, ≥ 30%) and of the pad area of the same PEDOT:PSS polyMEA (blue trace, >10%), again with respect to a water-filled cuvette as the reference (grey trace, set to 100%). The inset photograph depicts the four sample cuvettes placed on printed paper (font size: 5 pt). The colors of the overlayed lines match the colors of the respective spectral curves. The orange arrow and dashed line indicate the height of the center of the passing light beam. The spectra demonstrate that the spectral trends of PDMS and thin PEDOT:PSS films are complementary. In combination, similar to a neutral density filter, they thus lead to an attenuated, yet almost constant transparency over the entire visible spectrum. While typical PEDOT:PSS films on PDMS notably decrease the overall transparency, it is still sufficiently high for effective illumination of biological specimen in brightfield and fluorescence microscopy. S3

2. Embedded video of heartbeat-correlated cardiac activity (Viewing requires a Flash-enabled PDF reader and Adobe Flash Player)

S4

Carbohydrate arrays as tools for the glycomics revolution