Performance assessment of polymer based electrodes for in vitro
electrophysiological sensing: the role of the electrode impedance
Maria C. R. Medeiros*
a, Ana L. G. Mestre
b,c, Pedro M. C. Inácio
b,c, José M.L. Santos
d,e, Inês M.
Araujo
d,e, José Bragança
d,e, Fabio Biscarini
fand Henrique L. Gomes
b,ca
IT-Instituto de Telecomunicações, Departamento de Engenharia Eletrotécnica e de Computadores,
Universidade de Coimbra, 3030-290 Coimbra, Portugal.
b
Universidade do Algarve, FCT, Faro, Portugal.
c
IT-Instituto de Telecomunicações, Av. Rovisco, Pais, 1, Lisboa, Portugal.
d
Department of Biomedical Sciences and Medicine, University of Algarve, 8005-139 Faro,
Portugal.
e
Centre for Biomedical Research, CBMR, University of Algarve, 8005-139 Faro
f
Life Science Dept. - University of Modena and Reggio Emilia ,Via Campi 103, I-41125 Modena,
Italy.
ABSTRACT
Conducting polymer electrodes based on poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) are used to record extracellular signals from autonomous cardiac contractile cells and glioma cell cultures. The performance of these conducting polymer electrodes is compared with Au electrodes. A small-signal impedance analysis shows that in the presence of an electrolyte, both Au and polymer electrodes establish high capacitive double-layers. However, the polymer/electrolyte interfacial resistance is 3 orders of magnitude lower than the resistance of the metal/electrolyte interface. The polymer low interfacial resistance minimizes the intrinsic thermal noise and increases the system sensitivity. However, when measurements are carried out in current mode a low interfacial resistance partially acts as a short circuit of the interfacial capacitance, this affects the signal shape.
Keywords: Organic bioelectronics, polymer electrodes, extracellular-signals, PEDOT:PSS * [email protected]; phone +351 239 796 200; fax +351 239 796 247
1. INTRODUCTION
Microelectrode arrays (MEAs) are considered to be a basic platform for the development of cell-based sensors. These are substrate-integrated extracellular electrode matrices kept in contact with cells in culture1,2. MEA-based neuronal-electronic interfaces have been shown to facilitate the study of neuronal network processes, effects of pharmacological drugs and mechanisms underlying pathological conditions. Recently, there is a significant effort in improving the MEA technology. These efforts have been directed at two major challenges (i) decreasing the device dimensions and (ii) improving the electrical coupling between the cell and the sensing device. Smaller devices are better for parallel activity mapping, hence the application of nanofabrication technologies in this field. Devices incorporating nanostructures have been used to record electrical activity from neurons and cardiomyocytes3–7. Often in these devices the cells spontaneously engulf protruding structures. This intimate contact between the cell and electrode improves
signal-to-noise ratio as compared to planar electrode systems. Another goal being pursued is the lowering of electrode impedance. The most successful strategy is the use of conducting polymer surfaces5,8–11. It has been shown that polymers offer a low impedance that facilitates signal transduction from the cell to the recording electrode12–16.
This contribution compares the performance of conventional Au electrodes with ink-jet printed polymer electrode to record extracellular signals from cells in vitro. The objective is to gain insight into the impedance parameters that control the electrical coupling between the cells and planar extracellular electrodes. The performance of the electrodes is demonstrated using signals recorded from glioma cells cultures and from clusters of autonomous cardiac contractile cells. The paper is structured as follows: First, the basic measurement system is presented in section 2. Section 3 introduces the equivalent circuit model that describes the displacement current method and how the voltage signal is related with the signal in current. Next, the impedance of metallic and polymer based electrodes is presented and discussed. Finally, the conclusions highlight the role of the individual impedance parameters in shaping the signal in current.
2. ELECTRODES CHARACTERISTICS AND EXPERIMENTAL SETUP
Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate PEDOT:PSS electrodes were ink-jet printed on glass substrates. Printing was performed in air using a Fujifilm Dimatix Material Printer (DMP) 2831, with a DMC-11610 cartridge. Samples were annealed on a hot plate at 60˚C for 8h. After annealing, ethyleneglycol (EG) was deposited on the structure by immersion and then the devices were dried in a vacuum oven at 60˚C for 12h. Fig. 1 (a) shows a section of two parallel PEDOT:PSS electrodes printed on glass substrates. As test cells, we used cardiomyocytes differentiated from mouse embryonic stem cells (ESC) described elsewhere17,18 and Rat glioma C6 cells (American Type Culture Collection, ATCC). A photograph of a confluent C6 cell population on top of gold electrodes is show in Fig. 1(b). An EB with multiple foci of contractile activity is shown in Fig. 1(c). This EB has approximately 1000 cells. The sensing electrodes with cells were maintained at 37 ºC in an incubator (HERACell®150) with a humidified atmosphere with 5% of CO2.
All electrical measurements were performed with a Stanford low-noise current amplifier (SRS 570), or alternatively in voltage mode using the voltage amplifier (SRS 560), connected to a dynamic signal analyzer (Agilent 35670A). The low noise pre-amplifiers operated with internal batteries. Small-signal-impedance measurements were carried out by a Fluke PM 6306 impedance meter. All electrical measurements were carried out inside of a thick iron based Faraday cage to shield low frequency interferences and the entire system is mechanically decoupled from external vibrations. Fig. 1(d) shows a schematic diagram of the measuring set-up.
Figure 1. (b) photo of cardiac
Fig. 2 sh data. The bas distance L. O the electrode a double-laye as depicted i capacitance ( the polarizati into account electrode, the relatively at electrode has equivalent ci
Cells and electr ograph of a mono c cells and (d) sc
hows the schem sic sensing sys One electrode es are immersed er at the electro in Fig. 2. The (CD). RD takes ion resistance o the bulk elect e resistance RC
a long distanc s high impedan ircuit used in li
rodes used. (a) s olayer of adhere chematic diagram
3. E
matic layout o tem is compris acts as a sensi d in an electro olyte/electrode double-layer i into account a of the interface trolyte conduct C models the si ce from the se nce and it is na iterature19,20. section of two pa nt C6 cells on to m of the measuriXPERIMEN
of the electrode sed of two para ing electrode, lyte solution th e interface. Thiis described by all the possible e. The bulk ele tivity (RS) and
ignal loss betw ensing electrod amed seal impe
arallel polymer e op of a PET sub
ing set-up.
NTAL SET
es and the cor allel co-planar the other is gr he ions screen is interfacial re y an interfacia e contributions ectrolyte solutio capacitance (C ween cell and t
de. Therefore, edance (Zseal).
electrodes ink-je strate with gold
TUP MODE
rresponding eq electrodes of w rounded and it
the electrode p egion has been
al resistance (R s to the resistan on is also desc CS). When cel the measuring the electrical This simple m et printed on top electrodes. (c) p
EL
quivalent circui width D and le used as a refe potentials by fo described usin RD) in parallel nce including t cribed by a para lls are in conta electrode. The path from the model has beenof a glass subst photograph of an it used to inter ngth W, separa erence electrod forming a Helm ng an equivalen with the doub the charge tran allel RC netwo act with the m e reference elec e cell to the r adapted from trate. n EB rpret the ated by a de. When mholtz or nt circuit ble-layer nsfer and ork takes measuring ctrode is reference standard
ZSeal Counter electrode
ZT
J
R1 prPri L..../
D Measi electrAhab
Rvs(
t) L. is (t) D n CD uring ode louble-layer RF Figure 2. impedanc The imp thermal nois better electri understand th and the bulk These are th behave as a t At low frequ short-circuite of the interfa such as Au i and loss (1/ω equivalent ci dispersion h equivalent ci polymer elec in contrast w interface has the interfacia does not beh gold/electroly electrodes. T comparison, It is now electrode sen signals are PEDOT:PSS good electric low interfaci . Schematic dia ce amplifier is us pedance plays a e and basically ical coupling b heir frequency k electrolyte so he parameters two-layer syste uencies the hig ed giving rise t acial and bulk immersed in ce ωRP) with ω=2. ircuit network as not been ta ircuit network ctrode immerse with the gold e s a very low re al double-layer have as a two-l yte system. The impedance the impedance w important to u nsitivity for bi detected and S/electrolyte in cal coupling be al resistance w agram representi sed. The amplifia crucial role o y defines the n between cell an y dependence. olution can be measured ext em, with one in gh capacitive i to a transition f capacitance. T ell culture med .π.f for an Au e k depicted in t
aken into acco to fit the disp ed in a cell cult electrodes desc esistance. The r resistance (RD layer system a Fig. 3 (c) co e of the gold i e of the PEDOT understand how oelectronics si amplified as nterface becaus etween the cell whereas the high
ing the electrica ied signal is
4.
of the transduc noise floor. Hi nd electrode. T When measur equivalent to ternally by an nterfacial layer interfacial laye from a high ca This is known dium exhibit th electrode is sh the inset of Fi ount in our m persion at low ture medium. B cribed above, t bulk electrolyt D) is estimated and therefore d mpares the fr s several orde T electrodes is w these differe ignal detection s voltage, the se of its high s and the elect h interfacial ca al coupling betw . 3900RESULTS
cer ability to de igher values of Therefore, it is ing the impeda a frequency-d n impedance a r having a high er is probed. A apacitance to a as a Maxwell his behavior. T hown in Fig. 3ig. 3 (a). The model. Usually
frequencies. F Both the capac
the loss rises r te resistance (R d to be around does not show requency depe
rs of magnitud relatively flat ences in the imp n. The first ob en the gold/el
interfacial res trodes. At low apacitance prov
ween the cells a
, 375
etect bioelectri f capacitance a s important to ance the comp dependent para analyzer. Typ h capacitance in As the frequen low capacitan l-Wagner relax The frequency (a). The contin
fitting is not a constant ph Fig. 3 (b) show citance and loss
rapidly for low RS) is estimate 300 Ω. Interes a Maxwell-W endence of the de higher and upon increasin pedance param vious implicat lectrolyte inte sistance. PEDO frequencies th vides a low imp
and the measuri
, and 550
ical signals. Th and low values measure the s ponents that de allel resistor (R pically, electro n series with a ncy increases th nce that corresp
xation. In gene y dependence o nuous lines rep perfect becau hase element ws the frequen s are frequency w frequencies, ed to be approx sting, the polym Wagner relaxatio e impedance is strongly fre ng frequencies meters can be ex tion is related erface is subs OT:PSS/electro he good couplin pedance path f ing circuit. A tr . he resistance g s of resistance system impeda escribed the in RP) and capacit ode/electrolyte low capacitan he high capac ponds to the se eral metallic el of the capacitan present the fit u use the low fr
must be adde ncy dependenc y dependent. H which shows ximately 1.3 k mer/electrolyte on as observed for gold and equency depen up to 1 MHz. xplored to imp with thermal stantially nois olyte interface ng is mainly du for high freque
rans-generates provide ance and nterfacial tor (CP). systems ce layer. itance is ries sum lectrodes nce (CP) using the requency d to the ce of the However, that the kΩ while e system d for the PEDOT ndent. In prove the noise. If ier than provide ue to the encies.
. Figure 3. frequency of the fre experime frequency impedanc In our me where RF is understand h in our previo The current Comparison be y dependence of equency has a dis ental data (thick
y dependence o ce for Au and for
easurement setu
the feedback r how this curren ous work 21.
flowing tr
etween the imped f a gold electrod spersion know a k lines). Fitting of the capacitan r PEDOT:PSS e up we use a tra resistance and nt is related wit rough the inter
dance of a gold de system immer as Maxwell-Wag parameters are nce and loss for electrodes immer ans-impedance iS (t) the curre th vS (t), the vo rfacial double-l ≅ electrode system rsed in a cell cu gner relaxation. T RD =1.4 MΩ, r a PEDOT:PSS rsed in a cell cul
e amplifier, wh
ent flowing trou oltage signal ge layer can be ex
1
m with a printed lture medium. T The inset shows CD =55 nF, RS S electrode. (c) lture medium.
ose output volt
ugh the double enerated by the xpressed as d PEDOT:PSS el The capacitance the equivalent c =1.13 kΩ, CS The frequency tage is given b e-layer impeda e cell. This rel
lectrode. (a) Typ and loss as func circuit used to fi =0.36 nF. (b) dependence of
by
ance. It is impo lation has been
pical ction t the The f the (1) ortant to n derived (2)
where vS (t) i τ is the time iS (t) sign peak amplitu multiplying f current acros derivative of To under by Eq. (2) to edges of the upward and t DC value wh The beha cooperative p of the electro using gold el the PEDOT: capacitance b spike record approximatel Figure 4. between to the PE electrode s a voltage ram constant for th al is the sum o ude proportion factor for the ss the capacito f the original si
stand how the o a square volt voltage pulse to a downward hen RD is small avior described phenomena an ode. Further de lectrode with th :PSS electrode but the PEDO ded using the
ly 5 pA. . Signal shapes a signal recorde EDOT:PSS elect causes a pulse d mp rising at con he device to be of two independ
nal to the prod current. Basic or. Under these
gnal vs (t).
signal shape is tage pulse. Th
the time deriv d current spike l.
d above was ex d generate squ etails can be fou
he signals reco e was shifted
T:PSS electrod PEDOT based
. (a) the displac ed using a Au el
trode was shifte decaying to a ste nstant rate m=d charged or dis dent terms, a c duct mCD that cally, a rapidly e conditions, t s affect by the he simulation is vative, dνS/dt, f . The current s xperimentally uare like voltag
und in a previo orded using a P along the yy de has a much d electrode do cement current ectrode with a s ed along the yy eady state DC cu dvS (t)/dt. // scharged. component prop t decays with y varying volta the measured c circuit compon s represented i forces a large s spike should fa observed usin ge signals with ous publication PEDOT:PSS el axis by 65 p h smaller interf
oes not return
signal in respo signal using a PE
axis by 65 pA. urrent. These sign
portional to a time consta age signal prod
current signal nents lets simu in Fig. 4 (a). A surge current t all down to zer ng glioma cells h a time length n21. Fig. 4 (b) c lectrode. For cl pA. Both elec facial resistanc n to zero, but
onse to rectangu EDOT:PSS elect The low interfa nals have been r
and ant ≅ .
duces a large shape is also ulate the curren As expected at through the cap ro when RD is
s cultures. The h that is propor
compares the cu larity the time ctrodes have s ce RD. This exp
instead to a
ular voltage pul trode. The time facial resistance recorded using g a transient term . Hence, CD a transient displ proportional to nt response as p
t the rising and pacitor giving large and to a ese cells engag rtional to the d
urrent signals r trace correspo similar low fr plains why the steady state v
lse. (b) Compar trace correspon of the PEDOT: glioma cells cultu
(3) m with a acts as a lacement o m, the predicted d falling rise to a constant ge into a depth (D) recorded onding to requency e current value of rison ding :PSS ures.
Next, we electrodes w , were use higher than t capacitance e approximatel amplitude do transfer proc couples exac Figure 5. using PED is the sign nF. For c PEDOT cultures and Au electrode difference be significant d several order the intrinsic t parallel with the sensing c e discuss how ith the same p ed to record sig the other. The electrode is art ly 2.5 times h oes not exactly cess of the EB ctly in the same
Comparison be DOT:SS electro nal recorded usi larity, the upper
:PSS electrode in cluster of c es. The study etween the two difference on t rs of magnitud thermal noise a the interfacial capacitance. It i the electrode physical dimen gnals from the signals recorde tificially shifted higher than th
scale with the B from one ele e way to the sen
etween two time odes with identic ng an electrode time trace is art
es produced by contractile cells of the frequen o electrodes is the interfacial de lower than t
and can be use l capacitance. W
is shown that t
capacitance a nsions but with same cluster o ed in the two e d by 80 pA. Th he signal recor e increase in ca ectrode to the nsing surface. e traces of bioel cal dimensions b with 48 nF. The tificially shifted
5. CO
y ink-jet printin s. Their electri ncy dependenc s not only relaresistance. Th the Au/electrol ed to lower the When the resis this effect disto
affects the disp h different poly of contractile ce electrode syste he peak of the rded with the apacitance. Am
other. After th
lectrical signals ut with very diff e upper time trac
by 80 pA.
ONCLUSIO
ng were used t ical performan ce of the elect
ted with differ he PEDOT/PS lyte interface. A
system noise f stance is too sm orts the shape o
placement curr ymer roughnes ells. The capac ems are shown signal recorde
low capacitan mong the possib he transfer it i recorded from t fferent interfacia ce is the signal re
ONS
o measure extr nce was evaluatrode/electroly rences on the SS/electrolyte s
A low resistan floor. However mall it can act of the signal wh rent signal am ss and consequ citance of one e in Fig. 5. The d in the high c nce electrode. ble reasons for
is difficult to
the same cluster l capacitances. T ecorded using an
racellular bioe ated and compa yte impedance interfacial cap system has an nce is desirable r, the interfacia as a leakage cu hen measured mplitude. Two uent different v electrode syste e time trace of capacitance ele The measure this discrepan guarantee that r of contractile c The lower time t
n electrode with
lectrical signal ared with conv shows that th pacitance but a n interfacial re
e because it m al resistance ap urrent path tha in current. PEDOT values of em is 3.5 the high ctrode is ed signal ncy is the t the EB cells trace h 148 ls in cell ventional he major also by a esistance minimizes ppears in at bypass
REFERENCES
[1] Spira, M. E.., Hai, A., “Multi-electrode array technologies for neuroscience and cardiology,” Nat. Nanotechnol.
8(2), 83–94 (2013).
[2] Obien, M. E. J., Deligkaris, K., Bullmann, T., Bakkum, D. J.., Frey, U., “Revealing neuronal function through microelectrode array recordings,” Front. Neurosci. 9(JAN), 423 (2015).
[3] Fejtl, M., Stett, A., Nisch, W., Boven, K. H.., M??ller, A., “On micro-electrode array revival: Its development, sophistication of recording, and stimulation,” Adv. Netw. Electrophysiol. Using Multi-Electrode Arrays, 24–37 (2006).
[4] Blau, A., Murr, A., Wolff, S., Sernagor, E., Medini, P., Iurilli, G., Ziegler, C.., Benfenati, F., “Flexible, all-polymer microelectrode arrays for the capture of cardiac and neuronal signals,” Biomaterials 32(7), 1778–1786 (2011).
[5] Furukawa, Y., Shimada, A., Kato, K., Iwata, H.., Torimitsu, K., “Monitoring neural stem cell differentiation using PEDOT-PSS based MEA,” Biochim. Biophys. Acta - Gen. Subj. 1830(9), 4329–4333, Elsevier B.V. (2013).
[6] Hierlemann, A., Frey, U., Hafizovic, S.., Heer, F., “Growing Cells Atop Microelectronic Chips: Interfacing Electrogenic Cells In Vitro With CMOS-Based Microelectrode Arrays,” Proc. IEEE 99(2), 252–284 (2011). [7] Morin, F. O., Takamura, Y.., Tamiya, E., “Investigating neuronal activity with planar microelectrode arrays:
achievements and new perspectives.,” J. Biosci. Bioeng. 100(2), 131–143 (2005).
[8] Fang, Y., Li, X.., Fang, Y., “Organic bioelectronics for neural interfaces,” J. Mater. Chem. C 3(25), 6424–6430, Royal Society of Chemistry (2015).
[9] Nyberg, T., Shimada, A.., Torimitsu, K., “Ion conducting polymer microelectrodes for interfacing with neural networks,” J. Neurosci. Methods 160(1), 16–25 (2007).
[10] Richardson-Burns, S. M., Hendricks, J. L.., Martin, D. C., “Electrochemical polymerization of conducting polymers in living neural tissue.,” J. Neural Eng. 4(2), L6–L13 (2007).
[11] Svennersten, K., Larsson, K. C., Berggren, M.., Richter-Dahlfors, A., “Organic bioelectronics in nanomedicine,” Biochim. Biophys. Acta - Gen. Subj. 1810(3), 276–285, Elsevier B.V. (2011).
[12] Venkatraman, S., Hendricks, J., King, Z. A., Sereno, A. J., Richardson-Burns, S., Martin, D.., Carmena, J. M., “In vitro and in vivo evaluation of PEDOT microelectrodes for neural stimulation and recording,” IEEE Trans. Neural Syst. Rehabil. Eng. 19(3), 307–316 (2011).
[13] Leleux, P., Badier, J. M., Rivnay, J., Bénar, C., Hervé, T., Chauvel, P.., Malliaras, G. G., “Conducting Polymer Electrodes for Electroencephalography,” Adv. Healthc. Mater. 3(4), 490–493 (2014).
[14] Larsson, K. C., Kjäll, P.., Richter-Dahlfors, A., “Organic bioelectronics for electronic-to-chemical translation in modulation of neuronal signaling and machine-to-brain interfacing,” Biochim. Biophys. Acta - Gen. Subj.
1830(9), 4334–4344, Elsevier B.V. (2013).
[15] Sessolo, M., Khodagholy, D., Rivnay, J., Maddalena, F., Gleyzes, M., Steidl, E., Buisson, B.., Malliaras, G. G., “Easy-to-fabricate conducting polymer microelectrode arrays,” Adv. Mater. 25(15), 2135–2139 (2013). [16] Khodagholy, D., Doublet, T., Quilichini, P., Gurfinkel, M., Leleux, P., Ghestem, A., Ismailova, E., Hervé, T.,
Sanaur, S., et al., “In vivo recordings of brain activity using organic transistors,” Nat. Commun. 4, 1575 (2013). [17] Cells, E. S., Current Protocols in Stem Cell Biology, Stem Cells 44(201) (2007).
[18] Fuegemann, C. J., Samraj, A. K., Walsh, S., Fleischmann, B. K., Jovinge, S.., Breitbach, M., “Differentiation of mouse embryonic stem cells into cardiomyocytes via the hanging-drop and mass culture methods,” Curr. Protoc. Stem Cell Biol.(SUPPL.15), 1–13 (2010).
[19] Fromherz, P., “Semiconductor chips with ion channels, nerve cells and brain,” Phys. E Low-Dimensional Syst. Nanostructures 16(1), 24–34 (2003).
[20] Grattarola, M.., Martinoia, S., “Modeling the Neuron-Microtransducer Junction: From Extracellular to Patch Recording,” IEEE Trans. Biomed. Eng. 40(1), 35–41 (1993).
[21] Medeiros, M. C. R., Mestre, A., Inácio, P., Asgarif, S., Araújo, I. M., Hubbard, P. C., Velez, Z., Cancela, M. L., Rocha, P. R. F., et al., “An electrical method to measure low-frequency collective and synchronized cell activity using extracellular electrodes,” Sens. Bio-Sensing Res. 10, 1–8, Elsevier B.V. (2016).
