Sensors & Actuators: B.

Chemical 271 (2018) 280–288

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical

journal homepage: www.elsevier.com/locate/snb

Single walled carbon nanohorns composite for neural sensing and T

stimulation
⁎
Stefano Carlia, , Linda Lambertinia, Elena Zucchinia,b, Francesca Ciarpellaa,b, Alice Scarpellinic,
Mirko Pratod, Elisa Castagnolaa, Luciano Fadigaa,b, Davide Riccia
a
Center for Translational Neurophysiology of Speech and Communication, Istituto Italiano di Tecnologia, Via Fossato di Mortara 17/19, 44121 Ferrara, Italy
b
Section of Human Physiology, University of Ferrara, Via Fossato di Mortara 17/19, 44121, Ferrara, Italy
c
Electron Microscopy Facility, Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genova, Italy
d
Materials Characterization Facility, Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genova, Italy

A R T I C LE I N FO A B S T R A C T

Keywords: Oxidized single walled carbon nanohorns (ox-SWCNH) were electrodeposited onto gold microelectrode arrays in
PEDOT conjunction with poly(3,4-ethylenedioxythiophene) (PEDOT) and polystirenesulfonate (PSS), and the properties
Electrodeposition of the new composite material for neural recording and stimulation were assessed. PEDOT/ox-SWCNH com-
Single walled carbon nanohorns posites were compared with films prepared with one of the most notorious carbonaceous material in this field,
Neural sensing and stimulation
the oxidized multi-walled Carbon Nanotubes (ox-MWCNT). The PEDOT/ox-SWCNH exhibited superior charge
Multielectrode arrays
Neural implants
transfer capability, reflecting greater electroactive surface, as confirmed by SEM and EIS characterizations. As a
consequence, a charge injection limit of 11.6 mC/cm2 was observed for the new composite, which is higher than
the one of PEDOT/ox-MWCNT (8.7 mC/cm2). Having confirmed comparable neural recording performance, the
PEDOT/ox-SWCNH composite results very promising for improving therapeutic electrical stimulation in the
central and peripheral nervous systems.

1. Introduction interface during stimulation [3,4]. It has been demonstrated that

An ideal cortical microelectrode array for neural recording and
stimulation needs to be minimally invasive, flexible and robust while
providing low impedance and high charge transfer capability. Typical
neural electrodes are made using noble metals and alloys but recently
glassy carbon based microelectrodes have emerged [1,2]. Moreover, for
neural stimulation, electrodes should be capable of injecting relatively
large currents while minimizing electrode degradation due to faradaic
effects, requirements generally satisfied by increasing the electrode size
and, therefore, negatively affecting biocompatibility. In order to allow
long-term use, the surface of the microelectrodes should be able to fa-
cilitate charge transport and minimize inflammatory reaction and
gliosis.
To improve electrochemical properties and stability of microelec-
trode arrays, conductive polymers (CPs), such as polypyrrole (PPy) and
poly(3,4-ethylenedioxythiophene) (PEDOT), are commonly electro-
deposited onto the metal underlayer in order to significantly reduce the
total impedance as well as to increase the total charge that can be ac-
cumulated. This leads to higher S/N ratios, higher charge injection
limits, smaller voltage excursions and less heat generated at the
PEDOT exhibits higher electrochemical stability than PPy [5], and is
very promising in terms of biocompatibility as well [6,7].
Carbon nanotubes (CNT) have also been widely explored as co-do-
pant and dopant for PEDOT composites, enabling improved neural
signals recording as well as chronic neural stimulation [8–13]. Among
carbon based nanomaterials, high purity grade Single Walled Carbon
Nanohorns (SWCNH) are produced by direct vaporization of pure gra-
phite without the use of metal catalysts, originally by laser ablation
[14], and later by arc discharge [15] or by induction of very intense,
high frequency, eddy currents [16]. Having a basic structure similar to
tiny carbon nanotubes, SWCNH maintain most of the typical properties
of nanotubes including high electrical conductivity, high thermal con-
ductivity and they are easy to be functionalized [17]. Pristine SWCNH
are extremely hydrophobic but they can be readily functionalized with
abundant oxygenated groups, such as carboxylic acid groups, by oxi-
dative treatment in water-acidic media, to obtain hydrophilic oxidized
SWCNH (ox-SWCNH). This treatment promotes the opening of the horn
tip while preserving the conductivity and electrochemical properties.
Here we report, for the first time, the use of ox-SWCNH as co-dopant for
the electrodeposition of PEDOT/PSS onto microelectrode arrays for

⁎
Corresponding author.
E-mail address: stefano.carli@iit.it (S. Carli).

https://doi.org/10.1016/j.snb.2018.05.083
Received 24 January 2018; Received in revised form 14 May 2018; Accepted 15 May 2018
Available online 19 May 2018
0925-4005/ © 2018 Elsevier B.V. All rights reserved.
S. Carli et al.
neural applications comparing their behavior with the similar compo-
site based on oxidized multi walled carbon nanotubes (ox-MWCNT)
[6–11].
2. Experimental
2.1. Materials and equipment
Pristine SWCNH were obtained from Carbonium s.r.l., Padova
(Italy) [16]. ox-MWCNT were from Nanocyl S.A., Belgium (NC
3151, < 4% of COOH). All chemicals were from Sigma Aldrich except
otherwise specified.
2.2. Electrochemical characterization
Cyclic voltammetry (CV), electrochemical impedance spectroscopy
(EIS) and coating electrodepositions were carried out using a Reference
600 potentiostat (Gamry Instruments, USA) connected to a three-elec-
trode electrochemical cell with a Pt wire as a counter electrode and a
Ag/saturated AgCl reference electrode (+0.197 V vs NHE [18]). The
charge storage capability (CSC) was calculated as the time integral of an
entire CV cycle between +0.5 and −0.6 V. EIS were performed in three
electrodes cell configuration by superimposing a voltage sine wave
modulation (10 mV RMS amplitude, 0 V potential applied) within the
frequency interval of 105–1 Hz. The software ZSimpWin V 3.2 (EChem
Software) was used for equivalent circuit modeling of EIS data and χ2
values in the range of 10−4–10−6 were used to estimate the goodness of
the fit. Voltage transients were collected with a Parstat 2273 po-
tentiostat (Princeton Applied Research) in a two electrode configura-
tion by applying cathodic-first balanced current pulses (1 ms single
pulse, charge density 2 mC/cm2). Electrical stimulation consisted of a
series of cathodic-first charge balanced biphasic pulses with 6.37 mC/
cm2 charge density, 500 μs cathodic half-phase period and a frequency
of 500 Hz. The electrochemical characterizations were collected in a
0.9% w/w sodium chloride aqueous solution.
2.3. Gold microelectrode arrays [19]
16-electrode arrays with (1.2 × 1.2) mm total recording area and
60 μm diameter recording sites (Fig. 1a) were used for electrochemical,
morphological and X-ray photoemission spectroscopy (XPS) character-
izations, as well as for neural activity recording. For stability tests under
stimulation electrical pulses, similar arrays have been used, with mi-
croelectrode diameter of 100 μm.
Fig. 1. a) Representative scheme of the whole μEcoG array used for neural signal recording, b) details on the position of NTs and NHs coatings on electrodes and c)
scheme of the NHs modified microelectrode.
281
Sensors & Actuators: B. Chemical 271 (2018) 280–288
2.4. PEDOT/ox-SWCNH and PEDOT/ox-MWCNT electrodeposition
A nanostructured gold layer was electrodeposited following our
previously published work [19]. The composite electrodeposition so-
lution was prepared by suspending ox-SWCNH [20] or ox-MWCNT in
ultrapure water at the concentration of 1 mg/ml (Milli-Q, Millipore,
USA) by horn sonication (Vibra-Cell VCX130, Sonics and Materials,
USA) for thirty minutes (6 s at 66% duty cycle pulses, 4 W/ml). Poly
(sodium 4-styrene sulfonate) (NaPSS, 0.1 M) and 3,4-ethylenediox-
ythiophene (EDOT, 0.05 M) were added to the suspension immediately
after, and the mixture was stirred at room temperature for thirty min-
utes. The electrochemical deposition was carried out in potentiody-
namic mode, in the potential range of 0 V–0.95 V, with a scan rate of
100 mV/s, for a total of 100 cycles. For sake of clarity, PEDOT com-
posites electrodeposited by a deposition mixture containing ox-SWCNH
or ox-MWCNT will be abbreviated with NHs or NTs, respectively. The
deposition followed the detailed scheme reported in Fig. 1b. The same
conditions were used to deposit PEDOT:PSS from a solution of EDOT
(0.05 M) and NaPSS (0.1 M) in water. A schematic representation of the
coated microelectrode is reported in Fig. 1c.
2.5. Optical and surface characterization
High-resolution scanning electron microscopy (SEM) imaging was
carried out using a JEOL JSM 7500FA (Jeol, Tokyo, Japan) equipped
with a cold FEG (field emission gun), operating at 5 kV acceleration
voltage. When needed, the sample was carbon coated with a 10 nm
thick film using an Emitech K950X high vacuum turbo system (Quorum
Technologies Ltd, East Sussex – UK). XPS analyses were carried out
using a Kratos Axis UltraDLD spectrometer (Kratos Analytical Ltd., UK)
on 100 μm microelectrodes. XPS spectra were acquired using a mono-
chromatic Al Κα source operated at 20 mA and 15 kV. High resolution
spectra were collected at pass energy of 20 eV and energy step of 0.1 eV,
and the Kratos charge neutralizer system was used on all specimens.
Spectra have been charge corrected to the main line of the C 1 s spec-
trum set to 284.8 eV and analyzed with CasaXPS software (Casa
Software, Ltd., version 2.3.17). Each sulfur species is represented by a S
2p doublet, due to spin-orbit coupling. In the fitting procedure, each
doublet was fitted to a pair of peaks with the constraints of having the
same full width at half maximum (fwhm), the standard spin–orbit
splitting of 1.2 eV, and a statistical branching ratio of 1/2. The position
of the S 2p doublet was identified by the position of the most intense
component (i.e. S 2p3/2 component).
S. Carli et al.
2.6. Somatosensory evoked potentials (SEPs) and spontaneous activity
recordings
One Long Evans rat (male, 12 weeks, 400 g) was used for recording
neural activity, and was prepared following a published procedure [2]
in compliance with the guidelines established by the European Com-
munities Council (Directive 2010/63/EU, Italian Legislative Decree n.
26, 4/3/2014), and the protocol was approved by the Ethics Committee
for animal research of the University of Ferrara and by the Italian
Ministry of Health (authorization n° 332/2015-PR).
SEPs from four different positions (25 min long sessions) and
spontaneous neural activity from 3 different positions (10 min long
sessions) were acquired using a TDT (Tucker-Davis Technologies, USA)
multi-channel recording system 3 including the ZIF-Clip® headstage
with unity (1×) gain, the RZ2 real-time processor and the PZ2-256
battery-powered preamplifier. Data were digitized at a sample rate of
12207 samples/s at 18-bit resolution and transferred from RZ2 to the
PC by fast fiber optic connection. The experimental setup was placed in
a Faraday cage to reduce electromagnetic noise. In order to mechani-
cally elicit the somatosensory evoked potentials, a vibrating system was
used to produce a multi-whiskers deflection along the horizontal plane.
The whiskers contralateral to the craniotomy were shortened and in-
serted in a Velcro strip attached to a rod moved by a shaker (Type 4810
mini shaker; Bruel & Kjaer) controlled by a National Instruments board
(Austin, TX, USA). The deflection stimulus consisted of 10 truncated
Gaussian trains of 12 ms duration, delivered at 9 Hz. Three different
amplitude stimulations were used in order to obtain multi-whiskers
deflections respectively of 250, 500, and 800 μm. Each train was se-
parated from the others by a 5 s pause and all of them were repeated 20
times with 60 s pause between cycles.
2.7. In-vivo electrophysiological data analysis
Neuroexplorer software (Plexon Inc, USA) was used to compute the
spectral power densities (SPDs) and the spectrograms (data sampling
rate 12207 Hz, segmentation length of 512, Hamming windowing
function, 50% window overlap) of the recorded spontaneous activity
(in the 0–300 Hz frequency range) and of the mechanically evoked
activity (in the 0–600 Hz frequency range; stimulation
amplitude = 250 μm) for electrode performance evaluation. With the
same purpose, the signal-to-noise ratio (SNR) was computed on the LFP
frequency range (0–300 Hz) considering 20 intervals of 1 s after the
stimulus (stimulation amplitude = 250 μm) as signal and 20 intervals of
1 s of pause as pre-stimulus noise. We estimated the signal power by
computing the integral of the spectral power density (SPD) of the signal
and the noise power by computing the integral of the SPD of the noise.
The SNR was finally defined as the ratio between the signal power and
the noise power. SNR data are presented as mean ± standard deviation
calculated for all the recording channels in the four different positions.
Two sample t-test was performed on the resulting SNR mean values of
both NHs and NTs electrodes. To extract SEPs, the acquired traces were
digitally low-pass filtered below 300 Hz (Butterworth 2nd order) to
obtain the local field potential (LFP) and down-sampled to 3051
sample/s using MATLAB software (Mathworks 2016A, USA). The
evoked responses for each amplitude of stimulation were averaged from
5 ms before and 100 ms after the stimulus trigger using EEGLAB
MATLAB toolbox [21].
2.8. Cell culture and in-vitro biocompatibility assay
NHs and NTs have been deposited onto conductive glass (FTO,
fluorine tin oxide, TEC8, 3 mm thick substrates, Pilkington) and the
surfaces were sterilized for thirty minutes in ethanol and then washed
with aqueous phosphate buffer (PBS, pH = 7.4) before using them for
the viability assay. Fibroblast cells, obtained from culture of adult rat
tail specimens [2], were plated over NHs or NTs supports as well as on
282
Sensors & Actuators: B. Chemical 271 (2018) 280–288
control samples (glass coverslips and FTO). The cell viability was
analyzed twenty-four hours after fibroblasts seeding and at three and
seven days of culture by triple staining with propidium iodide (PI; 5 μg/
ml), fluorescein diacetate (FDA; 15 μg/ml) and Hoechst-33342 (3.3 μg/
ml) [2] ImageJ (open source NIH image processing software) was used
to post-processed merged the images. The mean percentages of sur-
viving cells were calculated based on the ratio of total (Hoechst-posi-
tive) nuclei minus PI-positive nuclei (dead cells) divided by the total
nuclei. The viability of NHs and NTs was normalized with respect to the
percentage of surviving cells observed over the uncoated FTO samples.
3. Results and discussion
3.1. Coatings deposition and electrochemical characterization
Chemical vapor deposition has been reported for the deposition of
pristine carbon nanotubes onto metal electrodes with the aim of im-
proving neural recording [9]. Unfortunately, this technique requires
high temperature, in the order of 600 °C–900 °C, that cannot be applied
to an unconstrained variety of materials, including polyimide based
microelectrode arrays. On the contrary, electrochemical deposition of
carbon nanotubes/PEDOT:PSS composites, can be performed at room
temperature and enables an excellent control of the thickness as well as
allowing to localize the coating only at the microelectrode site. In this
study, electrochemical deposition through potentiodynamic mode was
preferred, since it can lead to a precise control of the polymeric film
thickness by varying the number of potential scan cycles [22]. In Fig. 1a
we report a schematic representation of the polyimide based micro-
electrode array used for this study. The single microelectrode is made
from a copper substrate, coated by a thick layer of electroplated gold
and a final layer of electrodeposited nano-porous gold on the top [19].
Subsequently, NHs and NTs were further electrodeposited following
the scheme reported in Fig. 1b. The role of the supplementary porous
gold layer is to increase the roughness in order to improve the adhesion
of electrodeposited NHs and NTs coatings.
Thanks to their extremely high electroactive surface area, both NHs
and NTs coatings provide a dramatic reduction of total impedance with
respect to Au coated microelectrodes that exhibit |Z| values in the order
of 160 KΩ at 1 KHz and up to 1 MΩ at 100 Hz (Fig. S1). Moreover, the
two composites show very similar EIS |Z| values, but NHs yielded
higher CSCs with respect to NTs, as summarized in Table 1. Taking into
account that both NHs and NTs were electrodeposited in similar con-
ditions (see Sections 2.3 and 2.4), the significantly higher charge sto-
rage capacity obtained for NHs suggests a higher electroactive surface
area for this material.
Circuit model analyses of EIS data were performed in order to gain
insight on the charge transfer dynamics at the electrolyte|electrode
interface and on the charge transport within the polymeric film.
In the high frequency region, the Nyquist plot (Fig. 2a) shows a
semicircle that can be ascribed to a fast charge transfer at the interfaces
of electrolyte|electrode or polymer|solution, as well as to rapid charge
transport dynamics in the conductive polymer matrix.
With respect to similar measurements found in literature, a sup-
plemental charge transfer process can be observed in the middle fre-
quency region of the Bode Phase plot (Fig. 2c), with a time constant
tentatively in the range of 103–102 Hz [24–28]. The nature of this
Table 1
Experimental |Z| and CSC values obtained for NHs and NTs coated microelec-
trodes (60 mm diameter) from a set of five different electrodes.
|Z|@ |Z|@1 kHz kΩ |Z|@ |Z|@10 Hz kΩ CSC mC/cm2
10 kHz kΩ 100 Hz kΩ
NHs 3.0 ± 0.2 3.8 ± 0.4 6.0 ± 0.5 12.8 ± 0.9 101.0 ± 18.6
NTs 3.0 ± 0.3 3.5 ± 0.4 6.1 ± 0.6 14 ± 3 43.0 ± 14.9
S. Carli et al.
Fig. 2. Representative EIS Circuit model for NHs (experimental data black circles) and NTs (experimental data red triangles) respectively: a) Nyquist plot b) Bode |Z|
plot, c) Bode phase plot. Fitted data are reported as lines. Inset in 1a represents a magnification in the high frequency domain. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
supplemental charge transfer element is not clear, but it seems to be
strongly related to the roughness of the electrodeposited Au underlayer
[19], in accordance with literature [29], since it is not noticeable for
NHs or NTs electrodeposited onto commercially 3 mm diameter GC
electrodes (see Fig. S2).
The optimized equivalent circuit is reported in Fig. S3. It consists of
the solution resistance (RS), a parallel capacitor (C1) and resistor ele-
ment (R1) to fit the high frequency impedance, a constant phase ele-
ment (CPE1 where 1/ZCPE = Y0(jω)n) in parallel with a resistor (R2) to
fit the middle frequency charge transfer process, and the bulk capaci-
tance of the composite film (represented by the CPE2). Finally, a series
finite length Warburg element (ZD = RDcoth(jωτD)1/2)/(jωτD)½) is re-
quired to fit the low frequency behavior. [23–26,28]
In Table 2 we report characteristic parameters obtained by the fit-
ting of experimental EIS shown in Fig. 2. A significantly lower RS
parameter was obtained for NHs, in the order of 0.8 KΩ, whereas for
NTs a value of 1.2 KΩ was calculated from the fitting. This parameter
depends on the ionic concentration, type of ions, temperature, as well
as on the geometry of the area in which current is carried. It is defined
as RS = ρ (l/A) where ρ is the solution resistivity, l is the length and A is
the area of the electrode. Providing ρ and l are constant during EIS
experiments, and according to the relationship RNHsANHs = RNTsANTs, it
can be deduced that ANHs ≅ 2ANTs accounting for an higher electro-
active surface area for NHs, in accordance with CSC values in Table 1.
Time constants (τ = R1C1) in the order of 0.53 μs and 0.69 μs for
NHs and NTs respectively, were observed for the high frequency charge
transfer process, suggesting comparable electronic interactions at the
PEDOT|Au interface. Moreover, as reported in Fig. 2b and c, the Bode
|Z| and phase plots show a very similar behavior for NHs and NTs in the
frequency range of 105–102 Hz, suggesting that the charge transport
dynamics within the conductive polymer matrix are almost indis-
tinguishable for NHs and NTs.
The low frequency domain is mainly dominated by the diffusional
resistance RD of the Warburg impedance, and its value is up to three
times lower for NHs, suggesting a faster ion transport within the pores.
CPE is used to describe the non-ideal capacitive behavior due, for
instance, to electrode roughness [23]. The physical meaning of the
admittance Q0 can be related to the capacitance of the electrode (at
ω = 1 rad/s and n = 0.95) and can be used to compare the
Table 2
Relevant parameters extracted from circuit modelling of the EIS characterization (from a set of six different NHs and NTs electrodes, respectively).
Rs kΩ R1 kΩ C1 nF R2 kΩ Q0a μS × sn
NHs 0.8 ± 0.1 2.2 ± 0.2 0.24 ± 0.02 12 ± 2 8±1
NTs 1.2 ± 0.1 1.9 ± 0.2 0.38 ± 0.05 6±2 5±1
a
Referred to CPE1.
b
Referred to CPE2 of circuit model reported in Fig. S3.
283
Sensors & Actuators: B. Chemical 271 (2018) 280–288
electroactive surface areas of NHs and NTs composites. Thus, the
higher surface area of NHs films can be predicted also in terms of
higher capacitance, as obtained by EIS fitting, as well as by extra-
polating the low frequency capacitance from experimental EIS
(−Zi = 1/ωCD where CD is the diffusional capacitance), as reported in
Table 2.
3.2. Surface characterization
SEM analysis, as reported in Fig. 3, confirmed a nano-porous
structure for both NHs and NTs. The morphology of the two coatings is
dramatically different and seems to be dominated by the presence of the
carbonaceous co-dopants. In particular, it has been reported that ox-
SWCNH tend to give spherical aggregates [20] that probably affect the
PEDOT growth leading to poly-dispersed nanoaggregates (Fig. 3a–c),
with big pores up to 150 nm in diameter, accounting for the higher
electroactive surface area for NHs, in accordance with the above
mentioned electrochemical characterization. On the contrary, the in-
corporation of ox-MWCNT (average length < 10 μm; average diameter
9.5 nm, according to provider) in NTs film, led to a fibrous network
with pores up to 30 nm in diameter (Figs. 3d–f and S4).
The role of ox-SWCNH, acting as templating-agent during the elec-
trochemically promoted polymer growth, can be confirmed by com-
paring the morphologies of NHs and PEDOT:PSS, as shown in Fig. S5: in
particular the prominent globular shape exhibited by NHs is not ob-
served for PEDOT:PSS.
XPS analysis furnished the S(2p) core-level spectra of NHs and NTs
composites and is shown in Fig. 4. Both spectra can be decomposed
considering the presence of two main sulfur species: the doublet having
S 2p3/2 component at binding energy (164.0 ± 0.2) eV can be assigned
to sulfur atoms of the thiophene moiety in the PEDOT backbone,
whereas the signals at 168 eV correspond to the sulfur atom in the PSS
polymer [30,31]. The relative amounts of the two S species are sum-
marized in Table S6. No significant differences between the two com-
posite coatings were observed, in terms of S(2p) binding energy both
for PEDOT and PSS sulfur signals. It is known that the binding energy is
strictly related to the atomic electronic density thus XPS analysis sug-
gests comparable electronic density around S atoms for both NHs and
NTs composites [32]. The PSS to PEDOT molar ratio can be estimated
na Q0b μS × sn nb RD kΩ CD μF
0.53 ± 0.01 4.1 ± 0.9 0.98 ± 0.01 1.22 ± 0.07 3.1 ± 0.6
0.65 ± 0.04 4±1 0.99 ± ± 0.01 3.8 ± 0.6 1.8 ± 0.3
S. Carli et al. Sensors & Actuators: B. Chemical 271 (2018) 280–288

Fig. 3. SEM image analysis for NHs (a,b,c) and NTs (d,e,f) coated microelectrodes.

by the area ratio between the signal of PSS and PEDOT. This parameter
has been reported in the order of 3.5 for PEDOT:PSS films obtained
from chemically oxidized EDOT, whereas for electrodeposited coatings
is typically expected in the range of 0.68–0.96 [30,31,33]. The PSS to
PEDOT molar ratios obtained for NHs and NTs were 0.66 and 0.92,
respectively: the higher content of sulfonate groups observed for NTs is
likely due to the incorporation of a higher concentration of the in-
sulating polymer NaPSS within the film [33].
3.3. Voltage transients and stability evaluation of microelectrodes during
electrical stimulation
During neural stimulation, microelectrodes must withstand a high
number of electrical stimuli, in the order of millions, and typically in
the form of cathodic-first charge-balanced biphasic pulses. The high
charge density associated to these stimuli and the fast time domain of
these electrical pulses, in the time scale of microseconds, give rise to a
non-ideal behavior at the electrode interface.
In Fig. 5a we report the voltage transients collected by 50 μm mi-
croelectrode arrays (see Fig. 1), corresponding to a stimulating pulse in
the order of 32 μA and a charge injected of 2 mC/cm2 (biphasic
cathodic first, 1 ms pulse) that represents the highest safe current limit
safety to prevent neural tissue damage [34]. As expected, the highest
voltage excursion was observed for Au microelectrode (4.9 V), while it
was extremely reduced for NHs and NTs coated microelectrodes,
thanks to their higher charge transfer capability.
Charge injection limit is defined as the amount of charge that can
move from the electrode to the solution without causing water reduc-
tion and/or oxidation [1]. For example, in order to avoid water re-
duction, the maximum cathodic excursion voltage (Emc) should not rise

Fig. 4. High resolution XPS spectra showing the binding energy values (in eV) collected on S (2p) energy region for NTs (left panel) and NHs (right panel) coatings.

284
S. Carli et al. Sensors & Actuators: B. Chemical 271 (2018) 280–288

Fig. 5. Voltage transient responses during the electrical stimulation.

above 0.8 V [1,12]. Access potential Ea due to iR drop in the solution
were observed for NHs and NTs in the order of 199 ± 11 mV and
243 ± 34 mV, respectively, and this is in accordance with the elec-
trochemical characterization that furnished a higher electroactive sur-
face area (lower Rs) for NHs with respect to NTs: in fact, for the same
injected current Ic, the access potential (Ea = Rs × Ic where Ic is the
cathodic current) is mainly dominated by the solution resistance. Emc
can be obtained by subtracting the voltage access Ea from the voltage
excursion (Fig. 5b) and can be described as Emc = Em − Ea = (Ic × tp)/
Cdl where tp is the pulse duration and Cdl is the electrode capacitance.
The lower Emc value was observed for NHs (Emc = 112 ± 2 mV) when
compared to NTs (Emc = 155 ± 34 mV), and this result reflects an
improved charge injection limit for NHs that has been estimated in the
order of 11.6 mC/cm2 while for NTs was 8.7 mC/cm2, in good ac-
cordance with the value of 7.7 mC/cm2 reported for similar NTs com-
posites (see Table S7) [12]. The reduced voltage excursion exhibited by
NHs during stimulation is helpful to prevent neural tissue damage by
enabling lower heating, as well as limiting the instability of the PEDOT
backbone due to over oxidation [35,36].
The stability of NHs coatings have been investigated under elec-
trical stimulation, according to our previously reported protocol used
for NTs [37], by applying a series of cathodic first balanced pulses at a
charge density of 6.33 mC/cm2. This value corresponds to a very strong
stimulation pattern taking into account that, for instance, pulsing at
charge densities in the order of 5 mC/cm2 induced IrOx delamination
[38]. Both NHs and NTs based microelectrodes can withstand over 5
million pulses with negligible variations in their impedances, as sum-
marized in Table 3. Moreover, SEM analysis of stimulated microelec-
trodes confirmed that the porous structure of pristine NHs and NTs
coatings is preserved, as reported in Fig. S8.
Table 3
Representative magnitude impedance of NHs and NTs based microelectrodes
(100 μm diameter, n = 3) before and after electrical stimulation.
|Z|@1 kHz kΩ |Z|@100 Hz kΩ |Z|@10 Hz kΩ
NHs 1.6 ± 0.5 1.8 ± 0.4 2.6 ± 0.7
NHs stimulated 1.9 ± 0.2 2.1 ± 0.3 3±1
NTs 1.4 ± 0.2 1.6 ± 0.3 3±2
NTs stimulated 1.6 ± 0.5 1.9 ± 0.7 4±3
Au 5.8 ± 0.7 30 ± 4 165 ± 10
3.4. Neural recording and cell viability
It has been widely demonstrated that microelectrodes coated with
NTs provide improved neural recordings when compared to “state of
the art” metal based microelectrodes, as well as to different conductive
polymer coatings [9,13]. For this reason, we have chosen to use NTs
coatings as a benchmark for the performance in neural signal recording
of the NHs coated microelectrodes. For this purpose, both NHs and NTs
were coated on the same neural array (see experimental details and
Fig. 1b).
PSDs and spectrograms of both spontaneous and evoked activity
were calculated for NHs and NTs. When compared to the spontaneous
activity (Fig. S9), PSDs and spectrograms of the evoked activity (Fig.
S10) show an increase of the power content of the signal in the LFP
frequency components and beyond (0–600 Hz). This gain, due to the
stimulation of the responses, is distinguished with both types of coat-
ings essentially in the same manner. Moreover, the evoked response
plot shown in Fig. 6a suggests that both coatings have comparable
capabilities of recording somatosensory evoked activity. Indeed, no
significantly difference in SNR values for the NHs and NTs coated
microelectrodes (2.1 ± 1.2 for the NTs and 2.0 ± 1.5 for NHs, two
sample t-test, α = 0.05, N = 32, p-value = 0.93) were observed. Thus,
NHs and NTs were found to be able to acquire the same neural signal
information from the rat cortex. The multitude of factors that con-
tribute to the total noise level during an in vivo neural recording session
can be defined as biological and non-biological noise sources [39]. Bio-
logical noise is due to the electrical activity of distant neurons/neuronal
processes, additional thermal noise components due to the presence of
brain tissue as well as synaptic noise. Among non-biological sources,
that include the electronic noise due to the amplifier system and the
noise injection at the electrolyte-electrode interface, the thermal noise
is the major contribution due to the electrochemical properties of the
coating. In particular, thermal noise is dominated by the impedance of
the microelectrode, in fact, it can be expressed as (4 Kb T ZR Δf)1/2
where Kb is the Boltzmann constant, T is the absolute temperature, ZR is
the real part of the electrode impedance, and Δf is the noise bandwidth.
The Johnson noise plots, calculated from the real part of the impedance
of electrodes used for neural recordings, are reported in Fig. S11. The
mean body temperature (approximately 310 K) is used for the calcu-
lations. It can be observed that in the analysed frequency domain
(3–300 Hz), there is no significant difference between NHs and NTs in
terms of thermal noises (Fig. S11). Since both composites have been

285
S. Carli et al.
Fig. 6. a) SEPs acquired by both types of electrodes (in blue NHs electrodes and in red NTs electrodes). Data are presented as mean (solid lines) ± SD (dashed lines)
of all channels in 4 different locations (N = 32); b) Quantitative analysis of the viability of fibroblasts cells cultured on glass coverslip as control sample (control,
black bars), over FTO coated with NHs (blue bars) and NTs (red bars) at 1, 3 and 7 days of in vitro culture. The percentages of surviving cells is represented as samples
mean value ± SD. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
tested on the same array and, simultaneously, on the same animal
(according to the deposition scheme shown in Fig. 1b), the biological
noise component and the electronic noise due to the amplifier can be
assumed identical. Therefore, it is not surprising that both NHs and NTs
exhibited a comparable neural recording capability.
In vitro and in vivo studies have demonstrated the low toxicity of
both SWNHs and ox-SWNHs, suggesting how these materials can be
successfully used for biological applications, including neural sensing
and stimulation as well, whereas further studies aimed to the in-
vestigation of the long term biocompatibility of these materials are
currently underway [17,40]. One of the main concerns related to the
biocompatibility of NHs composite could be referred to the possible loss
of carbonaceous nanoparticles by the electrode surface. Despite their
concentration within the PEDOT composite is expected to be dramati-
cally low, this disadvantage cannot be totally avoided. Our group has
proposed an interesting approach that consists of the encapsulation of
NTs nanocomposites with a soft synthetic permanent biocompatible
pHEMA hydrogel [41]. This pHEMA coating preserves the electro-
chemical performance and high quality recording ability of NTs but,
more importantly, acts as a biocompatible protective barrier against
potential nanomaterial detachment. Therefore, the same approach
could be used for NHs coatings in order to improve the biocompatibility
of this composite material. The investigation of the cytocompatibility
using in vitro assays is a simple first-step evaluation of biocompatibility
[42]. For this purpose, we used fibroblast cells cultured over NHs or
NTs coated FTO supports and a cell viability assay was performed at
different time points (1, 3 and 7 days) in order to assess the ability of
the materials to allow cells adhesion and to support their proliferation
(see Fig. S12). As shown in Fig. 6b, after one day of culture both cells
seeded over NHs and NTs adhered less efficiently when compared to
controls, but they were able to grow up and proliferate progressively
from the third to the seventh day of culture. To sum up, no statistically
significant difference in viability was observed for NHs and NTs
(Fig. 6b).
286
Sensors & Actuators: B. Chemical 271 (2018) 280–288
4. Conclusions
Oxidized SWCNH have been used for the first time as co-dopant in
the preparation of PEDOT based microelectrode arrays. With respect to
NTs coatings, NHs based microelectrodes exhibited a higher charge
storage capacity due to a higher electroactive surface leading to a
higher charge injection limit. NHs based microelectrodes proved to
withstand over five million pulses without significant loss in terms of
total impedance and surface damages. A similar performance in the
capability of discriminating between evoked response and basal neural
activity has been found for NHs and NTs. Encouraged by these results,
we believe that the improved ability of NHs to supply fast charge in-
jection pulses, makes this material a good candidate for neural appli-
cations that combine the possibility of recording and stimulating neu-
rons.
Conflict of interest
The authors declare no competing financial interest.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.snb.2018.05.083.
References
[1] S.F. Cogan, Neural stimulation and recording electrodes, Annu. Rev. Biomed. Eng.
10 (2008) 275–309.
[2] M. Vomero, E. Castagnola, F. Ciarpella, E. Maggiolini, N. Goshi, E. Zucchini, et al.,
Highly stable glassy carbon interfaces for long-term neural stimulation and low-
noise recording of brain activity, Sci. Rep. 7 (2017) 40332.
[3] Z. Aqrawe, J. Montgomery, J. Travas-Sejdic, D. Svirskis, Conducting polymers for
neuronal microelectrode array recording and stimulation, Sens. Actuator B: Chem.
257 (2018) 753–765.
[4] R. Green, M.R. Abidian, Conducting polymers for neural prosthetic and neural in-
terface applications, Adv. Mater. 27 (2015) 7620–7637.
S. Carli et al.
[5] K.A. Ludwig, J.D. Uram, J.Y. Yang, D.C. Martin, D.R. Kipke, J. Neural Eng. 3 (2006)
59–70.
[6] A.R. Harris, S.J. Morgan, J. Chen, R.M.I. Kapsa, G.G. Wallace, A.G. Paolini,
Conducting polymer coated neural recording electrodes, J. Neural Eng. 10 (2013)
016004.
[7] N. Chen, L. Tian, A.C. Patil, S. Peng, I.H. Yang, N.V. Thakor, et al., Neural interfaces
engineered via micro- and nanostructured coatings, Nano Today 14 (2017) 59–83.
[8] E.W. Keefer, B.R. Botterman, M.I. Romero, A.F. Rossi, G.W. Gross, Carbon nanotube
coating improves neuronal recordings, Nat. Nanotechnol. 3 (2008) 434–439.
[9] A. Ansaldo, E. Castagnola, E. Maggiolini, L. Fadiga, D. Ricci, Superior electro-
chemical performance of carbon nanotubes directly grown on sharp microelec-
trodes, ACS Nano 5 (2011) 2206–2214.
[10] R. Gerwig, K. Fuchsberger, B. Schroeppel, G.S. Link, G. Heusel, U. Kraushaar, et al.,
PEDOT-CNT composite microelectrodes for recording and electrostimulation ap-
plications: fabrication morphology and electrical properties, Front. Neuroeng. 5
(2012) 1–11 (article8).
[11] H. Zhou, X. Cheng, L. Rao, T. Li, Y.Y. Duan, Poly(34-ethylenedioxythiophene)/
multiwall carbon nanotube composite coatings for improving the stability of mi-
croelectrodes in neural prostheses applications, Acta Biomater. 9 (2013)
6439–6449.
[12] S. Chen, W. Pei, Q. Gui, R. Tang, Y. Chen, S. Zhao, et al., PEDOT/MWCNT com-
posite film coated microelectrode arrays for neural interface improvement, Sens.
Actuator A: Phys. 193 (2013) 141–148.
[13] T.D.Y. Kozai, K. Catt, Z. Du, K. Na, O. Srivannavit, R.M. Haque, et al., Chronic in
vivo evaluation of PEDOT/CNT for stable neural recordings, IEEE Trans. Biomed.
Eng. 63 (2015) 111–119.
[14] S. Iijima, M. Yudasaka, R. Yamada, S. Bandow, K. Suenaga, F. Kokai, et al., Nano-
aggregates of single-walled graphitic carbon nano-horns, Chem. Phys. Lett. 309
(1999) 165–170.
[15] T. Yamaguchi, S. Bandow, S. Iijima, Synthesis of carbon nanohorn particles by
simple pulsed arc discharge ignited between pre-heated carbon rods, Chem. Phys.
Lett. 389 (2004) 181–185.
[16] M. Schiavon, Device and method for production of carbon nanotubes fullerene and
their derivatives, Patent EP1428794 (A2) 2006.
[17] N. Karousis, I. Suarez-Martinez, C.P. Ewels, N. Tagmatarchis, Structure properties
functionalization and applications of carbon nanohorns, Chem. Rev. 116 (2016)
4850–4883.
[18] A.J. Bard, L.R. Faulkner, Electrochemical Methods Fundamentals and Applications,
J. Wiley & Sons, Hoboken, NJ, 2001.
[19] E. Castagnola, A. Ansaldo, E. Maggiolini, T. Ius, M. Skrap, D. Ricci, et al., Smaller
softer lower-impedance electrodes for human neuroprosthesis: a pragmatic ap-
proach, Front. Neuroeng. 7 (2014) 1–8 (article8).
[20] S. Carli, L. Casarin, Z. Syrgiannis, R. Boaretto, E. Benazzi, S. Caramori, et al., Single
walled carbon nanohorns as catalytic counter electrodes for Co(III)/(II) electron
mediators in dye sensitized cells, ACS Appl. Mater. Interfaces 8 (2016)
14604–14612.
[21] A. Delorme, S. Makeig, EEGLAB: an open source toolbox for analysis of single-trial
EEG dynamics including independent component analysis, J. Neurosci. Methods
134 (2004) 9–21.
[22] V. Castagnola, C. Bayon, E. Descamps, C. Bergaud, Morphology and conductivity of
PEDOT layers produced by different electrochemical routes, Synth. Met. 189 (2014)
7–16.
[23] E. Barsoukov, J.R. Macdonald, Impedance Spectroscopy: Theory Experiment and
Applications, 2nd ed., John Wiley & Sons Inc., Hoboken, NJ, 2005.
[24] J. Bobacka, A. Lewenstam, A. Ivaska, Electrochemical impedance spectroscopy of
oxidized poly(34-ethylenedioxythiophene) film electrodes in aqueous solutions, J.
Electroanal. Chem. 489 (2000) 17–27.
[25] M. Asplund, H. von Holst, O. Inganäs, Composite biomolecule//PEDOT materials
for neural electrodes, Biointerphases 3 (2008) 83–93.
[26] P. Danielsson, J. Bobacka, A. Ivaska, Electrochemical synthesis and characterization
of poly(34-ethylenedioxythiophene) in ionic liquids with bulky organic anions, J.
Solid State Electrochem. 8 (2004) 809–817.
[27] S. Carli, E. Busatto, S. Caramori, R. Boaretto, R. Argazzi, C.J. Timpson, et al.,
Comparative evaluation of catalytic counter electrodes for Co(III)/(II) electron
shuttles in regenerative photoelectrochemical cells, J. Phys. Chem. C 117 (2013)
5142–5153.
[28] E. Castagnola, S. Carli, M. Vomero, A. Scarpellini, M. Prato, N. Goshi, et al.,
Multilayer poly(34-ethylenedioxythiophene)-dexamethasone and poly(34-ethyle-
nedioxythiophene)-polystyrene sulfonate-carbon nanotubes coatings on glassy
carbon microelectrode arrays for controlled drug release, Biointerphases 12 (2017)
031002.
[29] X. Cui, D. Martin, Fuzzy gold electrodes for lowering impedance and improving
adhesion with electrodeposited conducting polymer films, Sens. Actuator A: Phys.
103 (2003) 384–394.
[30] G. Greczynski, T. Kugler, W.R. Salaneck, Characterization of the PEDOT-PSS system
by means of X-ray and ultraviolet photoelectron spectroscopy, Thin Solid Films 354
(1999) 129–135.
[31] G. Greczynskia, T. Kuglerb, M. Keila, W. Osikowicza, M. Fahlmanc, W.R. Salanecka,
Photoelectron spectroscopy of thin films of PEDOT–PSS conjugated polymer blend:
a mini-review and some new results, J. Electron Spectrosc. Relat. Phenom. 121
(2001) 1–17.
[32] Z.L. Yang, H.Z. Chen, L. Cao, H.Y. Li, M. Wang, Synthesis and photoconductivity
study of carbon nanotube bonded by tetrasubstituted amino manganese phthalo-
cyanine, Mater. Sci. Eng. B 106 (2004) 73–78.
[33] G. Zotti, S. Zecchin, G. Schiavon, F. Louwet, L. Groenendaal, X. Crispin,
W. Osikowicz, W. Salaneck, M. Fahlman, Electrochemical and XPS studies toward
287
Sensors & Actuators: B. Chemical 271 (2018) 280–288
the role of monomeric and polymeric sulfonate counterions in the synthesis, com-
position, and properties of poly(3,4-ethylenedioxythiophene), Macromolecules 36
(2003) 3337–3344.
[34] R.V. Shannon, A model of safe levels for electrical stimulation, IEEE Trans. Biomed.
Eng. 39 (1992) 424–426.
[35] Y. Takenaka, T. Koike, T. Oka, M. Tanahashi, Degradation of polythiophene film by
doping-undoping cycle, Synth. Met. 18 (1987) 207–212.
[36] H. Harada, T. Fuchigami, T. Nonaka, Degradation and its prevention and the de-
activation and reactivation of electroactive polythiophene films during oxidation/
reduction cycles, J. Electroanal. Chem. 303 (1991) 139–150.
[37] E. Castagnola, A. Ansaldo, E. Maggiolini, G.N. Angotzi, M. Skrap, D. Ricci, et al.,
Biologically compatible neural interface to safely couple nanocoated electrodes to
the surface of the brain, ACS Nano 7 (2013) 3887–3895.
[38] S.F. Cogan, A.A. Guzelian, W.F. Agnew, T.G. Yuen, D.B. mCCreery, Over-pulsing
degrades activated iridium oxide films used for intracortical neural stimulation, J.
Neurosci. Methods 137 (2004) 141–150.
[39] M.E.J. Obien, K. Deligkaris, T. Bullmann, D.J. Bakkum, U. Frey, Revealing neuronal
function through microelectrode array recordings, Front. Neurosci. 8 (2015) 1–30
(Article 423).
[40] P. Araujo, Carbon Nanostructures, Springer International Publishing, Switzerland,
2016.
[41] E. Castagnola, E. Maggiolini, L. Ceseracciu, F. Ciarpella, E. Zucchini, S. Faveri,
L. Fadiga, D. Ricci, pHEMA encapsulated PEDOT-PSS-CNT microsphere micro-
electrodes for recording single unit activity in the brain, Front. Neurosci. 10 (2016)
1–14 (article 151).
[42] W. Li, J. Zhou, Y. Xu, Study of the in vitro cytotoxicity testing of medical devices,
Biomed. Rep. 3 (2015) 617–620.
Dr. Stefano Carli (MSc and PhD in Chemistry, University of Ferrara) has devoted his PhD
investigations to the development of new photochromic dyads. From 2009 to 2016 Dr.
Carli’s research activity was oriented on Dye Sensitized Solar Cells, Perovskite Solar Cells,
as well as on new catalysts for light-induced water splitting. Currently Dr. Carli is a
postdoctoral researcher at the Italian Institute of Technology and his research is focused
on the development of new smart materials for neural sensing and stimulation.
Miss Linda Lambertini is a Master student of the Department of Chemical and
Pharmaceutical Sciences of the University of Ferrara. She devoted her thesis to the de-
velopment of new materials for neural sensing and stimulation at the Italian Institute of
Technology.
Dr. Elena Zucchini received her Master’s Degree in Biomolecular and Cellular Sciences at
the University of Ferrara in 2015 and currently is attending the PhD course in
Translational Neurosciences and Neurotechnologies at IIT@UniFe − Center for
Translational Neurophysiology (University of Ferrara). The main focus of her project is to
improve the neural signal quality of the recordings obtained with novel μECoG arrays in
order to detect brain high frequency activity from the surface of the cortex.
Dr. Francesca Ciarpella received her master’s degree in Biomolecular and Cellular
Sciences at the University of Ferrara and during that time she get expertise in biochem-
ical, cellular and molecular fields. She is now attending her PhD in Biomedical Science
and Biotechnology at the Center for Translational Neurophysiology at IIT@Unife focusing
the attention on the study of long-term biocompatibility of intracortical microelectrodes
and on the development of a seamless device, fully integrated with the brain tissue.
Dr. Alice Scarpellini obtained a bachelor degree in Biomedical Engineering at the
University of Cagliari in 2008. In October, 2010, after moving to Genoa, she received the
master degree in Bioengineering at the University of Genoa with a thesis on chronical
implants of microelectrodes with nanostructured coatings. She currently works as a senior
technician in the Electron Microscopy Facility at the Italian Institute of Technology and
her activities are mainly devoted to the characterization of different materials with var-
ious techniques including HRSEM, HRTEM, HAADF-STEM and EDS analysis.
Dr. Mirko Prato received his Ph.D. in Materials Science and Technology from the
University of Genova, Italy in 2008, discussing a thesis on the surface characterization of
organic self-assembled monolayers. He then moved to the National Institute of Nuclear
Physics (INFN), working in the gravitational waves field. In 2011, he joined as post-
doctoral fellow the Italian Institute of Technology, Nanochemistry Department, working
on X-ray Photoelectron Spectroscopy studies of colloidal nanocrystals and halide per-
ovskites for photovoltaics. Since the beginning of 2017, he is the coordinator of the
Materials Characterization Facility at the same institute, providing support in the scien-
tific activities of IIT research groups dealing with the development, optimization, and
applications of new materials, through a set of state-of-the-art characterization labora-
tories, expert experimental design, and data analysis
Dr. Elisa Castagnola received her MSc in Bioengineering (highest honors) at the
University of Genoa (2007). In 2011, Elisa got her PhD in Robotics, Neurosciences and
Nanotechnologies from the Italian Institute of Technology (IIT), defending a thesis on
“Carbon Nanotube Based Coatings for Low Impedance Neural Microelectrodes”. She has
been a Post Doc at IIT from 2011 to 2016 focusing on the study of innovative neural
microelectrode technology to produce low-noise, long term implantable devices con-
ceived ‘for human use’ by researching new materials for the electrode biological interface.
Currently, Dr. Castagnola is a postdoct researcher at the Department of Mechanical
Engineering San Diego State University.
S. Carli et al.
Prof. Luciano Fadiga (M.D., University of Bologna, Ph.D. in Neuroscience, University of
Parma) is full Professor of Human Physiology at the University of Ferrara and is the
Coordinator of Center IIT@UNIFE. He has a long experience in electrophysiology and
neurophysiology in monkeys (single neurons recordings) and humans (transcranial
magnetic stimulation, study of spinal excitability, brain imaging, recording of single
neurons in awake neurosurgery patients).
288
Sensors & Actuators: B. Chemical 271 (2018) 280–288
Dr. Davide Ricci Davide Ricci received the MSc degree in Physics and the PhD degree in
Electronic Engineering and Computer Science from the University of Genoa, Italy, in 1989
and 1993, respectively. He is currently researcher the IIT Translational Neurophysiology
Center in Ferrara, focusing on the development of devices such as neural electrodes,
flexible nano-actuators and sensors for Brain Machine Interfaces and prosthetics. He has
authored or co-authored more than 100 papers in peer-reviewed international journals,
conference proceedings, and books.