ISSN 1998-0124 CN 11-5974/O4 2 https://doi.org/10.1007/s12274-019-2462-0

Res

earc

h Ar

ticle

In situ fabrication of organic electrochemical transistors on a

microfluidic chip

Jianlong Ji1,2, Mangmang Li1, Zhaowei Chen3, Hongwang Wang1, Xiaoning Jiang2, Kai Zhuo1, Ying Liu1, Xing Yang4, Zhen Gu3, Shengbo Sang1 (), and Yang Shu5 ()

Nano Res., Just Accepted Manuscript • https://doi.org/10.1007/s12274-019-2462-0 http://www.thenanoresearch.com on June. 12, 2019 © Tsinghua University Press 2019 Just Accepted

This is a “Just Accepted” manuscript, which has been examined by the peer-review process and has been accepted for publication. A “Just Accepted” manuscript is published online shortly after its acceptance, which is prior to technical editing and formatting and author proofing. Tsinghua University Press (TUP) provides “Just Accepted” as an optional and free service which allows authors to make their results available to the research community as soon as possible after acceptance. After a manuscript has been technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Please note that technical editing may introduce minor changes to the manuscript text and/or graphics which may affect the content, and all legal disclaimers that apply to the journal pertain. In no event shall TUP be held responsible for errors or consequences arising from the use of any information contained in these “Just Accepted” manuscripts. To cite this manuscript please use its Digital Object Identifier (DOI®), which is identical for all formats of publication.

Address correspondence to First A. Firstauthor, email1; Third C. Thirdauthor, email2

ISSN 1998-0124 CN 11-5974/O4 2 https://doi.org/(automatically inserted by the publisher)

Res

earc

h Ar

ticle

In situ fabrication of organic electrochemical transistors on a microfluidic chip Jianlong Ji1,2, Mangmang Li1, Zhaowei Chen3, Hongwang Wang1, Xiaoning Jiang2, Kai Zhuo1, Ying Liu1, Xing Yang4, Zhen Gu3, Shengbo Sang1 (), Yang Shu5 ()

1 College of Information and Computer, Taiyuan University of Technology, Taiyuan 030024, China 2 Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh 27695, USA 3 Department of Bioengineering, California Nanosystems Institute, University of California, Los Angeles, Los Angeles 90095, USA 4 The State Key Laboratory of Precision Measurement Technology and Instruments, Department of Precision Instrument, Tsinghua University, Beijing 100084, China 5 Department of Chemistry, Colleges of Sciences, Northeastern University, Shenyang 110819, China © Tsinghua University Press and Springer-Verlag GmbH Germany, part of Springer Nature 2018 Received: day month year / Revised: day month year / Accepted: day month year (automatically inserted by the publisher)

ABSTRACT Microfluid chips integrating with organic electrochemical transistors (OECTs) are useful for manufacturing biosensors with high throughput and large-scale analyses. We report here the utilization of alternating current (AC) electrodeposition to fabricate OECTs in situ on a microfluid chip. With this method, the organic semiconductor (OS) layer with a channel length of 8 µm was readily prepared without requiring the post-bonding process in the conventional construction of microfluidic chips. Poly(3,4-ethylenedioxythiophene): poly(4-styrenesulfonate) / graphene quantum dots (PEDOT:PSS/GQDs) composites with different morphologies, such as microfilms, nanodendrites and nanowires were electropolymerized. The mass transfer process of the electropolymerization reaction was evidenced to be diffusion limited. Morphologies, growth directions, and chemical structures of OS layers could be tuned by the amplitude and the frequency of the AC voltage. Transfer and output characteristic curves of OECTs were measured on the microfluidic chip. The maximum transconductance, on/off current ratio and threshold voltage measured in the experiment was 1.58 mS, 246, and 0.120 V, respectively.

KEYWORDS electrodeposition, in situ, microfluidic chip, PEDOT:PSS/GQDs composite, transistor

1 Instruction Organic electrochemical transistor (OECT) has received increasing attention recently because they are mixed electronic and ionic conductive, exhibit intrinsic signal amplification properties and require low operating voltages [1]. Such devices show great potentials in building biosensors, such as sensing biochemical substances [2-4], monitoring cellular activities [5, 6] and interfacing with electrophysiological signals [7, 8]. In particular, microfluidic chips containing OECTs has shown great potential in point-of-care diagnostics [9], wearable biosensors [10], and toxicology testing devices for drug screening [11]. There is great enthusiasm for developing multiplexed microfluidic chips since they can give a fast result and reduce sample usage [12]. Specially, reducing the channel length of OECT leads to enhanced utilization of the substrate space [13] as well as signal amplification capability [14], which is beneficial to achieve high throughput and high sensitivity sensing performance on microfluidic chips.

The most common method for preparing OECT is combing photolithography and spin coating processes [8, 15]. However, the organic semiconductor (OS) is generally incompatible with conventional photolithography processes. Besides, multiple exposures to solvents and chemicals lead to cracking and delamination of OS layers [16]. The additive patterning approach

was preferred where two layers of parylene C (PaC) were deposited and patterned in advance. After the peel-off process, OECTs array was defined by the first layer of PaC [17]. However, conventional bonding techniques like oxygen plasma activation are challenging to bond PaC substrates with polydimethylsiloxane (PDMS) microchannels [18]. Vincenzo reported that the pressure sensitive adhesive was beneficial to solve this problem [11] and an OECT microfluidic chip with a channel length of down to 10 μm was present [19].

Printing techniques, such as screen and inkjet printing, are easy to adapt for massive production. Screen printing patterns OS layers through the mechanical masks onto prefabricated locations. The channel length falls in the range of 50 - 100 μm, which is on the same scale of critical dimensions of traditional microfluidic chips [20]. Inkjet printing utilizes piezoelectric devices to spray OS ink onto selected areas on the substrate. Though the patterning process is eliminated, specially formulated inks and nonstandard equipment are required [16]. Recently, an OECT microfluidic chip integrating floating gates was present to detect DNA hybridization, and the OS footprint was set as 110 μm × 200 μm [9].

Electrodeposition of OS layers has also been extensively studied, which can be used in the field of electrochromic [21, 22], biosensing [23, 24], and lithium batteries [25]. Through the potentiostatic [26] or cyclic voltammetry method [24], OS layers

Address correspondence to S. Sang, sunboa-sang@tyut.edu.cn; Y. Shu, shuyang@mail.neu.edu.cn

were prepared on the electrode surface. However, the connection of the source and the drain electrode with OS layers was still needed to construct OECTs. The problem could be solved by alternating current (AC) electrodeposition. PEDOT fibers connecting the bipolar electrodes by AC-bipolar electrolysis were reported by Yuki [27] recently. Our group also reported several different nanomaterials prepared between planar electrodes [28-30].

Herein, we report that OECTs can be fabricated in situ on a

microfluidic chip by AC electrodeposition. We first characterized the morphology and chemical structures of PEDOT:PSS/GQDs OS layers. In addition, the potential mechanism for regulating the morphology and the growth direction of OS layers by the amplitude and the frequency of the AC voltage was demonstrated. Moreover, the field effect characteristics of OECTs, such as the maximum transconductance, on/off current ratio and threshold voltage were investigated.

Figure 1 (a) Micromachining processes used to obtain the microelectrode chip. (b) Fabrication processes to obtain the microfluidic chip. (c) The experimental setup for AC electrodepositions and electrical characterizations. Red tubes in (b) and (c) represent the wire channel for electrical connections. Blue tubes in (b) and (c) represent the fluid channel for inlets and outlets.

2 Experimental

2.1 Experimental procedure 3,4-ethylenedioxythiophene (EDOT), poly(4-styrene sulfonate sodium) (NaPSS, average Mw ~70,000), and PDMS were purchased from Alfa Aesar (Alfa Aesar Corp., USA). The double distilled water was prepared from a Millipore system (Millipore Q, USA) (resistivity >18 Ω·cm-1). GQDs were provided by Yang Shu at Northeastern University (Shenyang, China), and the preparation method is described in detail elsewhere [31]. The transmission electron microscopy (TEM) image shown in Fig.S1 in the electronic supplementary material (ESM) illustrates that uniform GQDs with a diameter of 2 - 9 nm were prepared. The three peaks at 531 eV, 399 eV, and 285 eV in X-ray photoelectron spectroscopy (XPS) spectrum shown in Fig.S2 are attributed to O 1s, N 1s, and C 1s, respectively. The ratios of nitrogen-to-carbon (N/C) and oxygen-to-carbon (O/C) are 0.3 and 0.23, indicating a relatively high N-doping content and the existence of a large amount of oxygen-containing functional groups on the surface GQDs.

Through the positive lithography and the lift-off processes, microelectrode pairs with 8 μm spacing (Fig.S3 (in the ESM)) were fabricated. Microelectrodes were composed of a 20 nm thick Ti cohesive layer and a 200 nm thick Pt layer. All the micromachining processes were carried out in CapitalBio Corp.

(Beijing, China). Microelectrode pairs before and after AC electrodeposition were illustrated in Fig.S3a and S3b (in the ESM), respectively. Then, microelectrode chips were fixed onto the PCB board by Benzocyclobutene bonding [32]. Gold wires of 200 μm in diameter were used to connect microelectrode pads to the PCB board for the external electrical connection.

To demonstrate OECTs on the microfluidic chip, a 20×15×4 mm copper block and a 5 mm diameter copper rod that were immersed in polydimethylsiloxane were fixed on the glass slide as the master mold. The glass slide was silanized with dimethyldichlorosilane and wrapped in a foil with the shape of a groove. A 10:1 weight mixture of PDMS prepolymer and curing agent (Sylgard 184 silicone elastomer kit, Dow Corning Corp., USA) was then poured into the foil. After keeping at 80 °C for 1 hour in an oven, the PDMS cavity was peeled off from the glass slide. Then, the PDMS cavity, into which both the PCB board and the microelectrode chip could be placed, was bonded with another glass substrate mechanically (Fig.1). Finally, inlets, outlets, and the gate channel were manufactured using metal punches.

In order to investigate the transfer characteristic of OECTs, 100 mM NaCl solutions were utilized as the gate electrolyte medium [33]. An Ag/AgCl electrode was immersed into the electrolyte through the hole pre-punched in the PDMS cavity with its tip at a distance of 200 μm on top of the microelectrodes. The Ag/AgCl electrode was chosen as the gate electrode because the negligible potential dropped at its interface with the electrolyte and

the small cross-sectional area at the tip [34]. It should be noted that the preparation process presented here

is only used for device demonstrations. The cavity was designed to be large enough that the entire PCB board can be integrated into. The microfluidic chip was mechanically bonded to facilitate further characterizations of OS layers (Fig.1). In the future, the micro-flow channel exhibiting a smaller size and higher precision can be obtained by the photoresist development process. Besides, the oxygen plasma activation could be used to increase the bond strength. Moreover, the Ag/AgCl gate electrode could be replaced with the planar electrode [19] to improve devices integration.

The EDOT monomer and NaPSS powder with different molar ratios were dissolved in distilled water (Ⅰ - Ⅳ in Table 1). Besides, electrolytes (Ⅴ, Ⅵ, Ⅶ, and Ⅷ in Table 1) blending with GQDs of mass concentration ranged from 0.2 μg·L-1 to 200 μg·L-1 were prepared to improve OECTs performance. In detail, the mixture was put into an ultrasonic water bath and stirred for 1 hour. Left without disturbance overnight, the mixture was fully exposed to the reaction. Also, the plating solution was added into the microfluidic chip by a syringe pump (Model Pump 11 Elite, Harvard Apparatus, USA) for the in situ AC electrodeposition.

Figure 2 (a) Cyclic voltammetry of electrolyte (Ⅲ) with the scan rate increasing from 10 mV·s-1 to 100 mV·s-1. The range of potential was increased from −0.8 V to +2.0 V vs SCE. The inset illustrates the relationship between the P1 current density and the square root of the scan rate. (b) Cyclic voltammetry of electrolyte (Ⅰ, Ⅱ, Ⅲ, Ⅳ) with the scan rate of 50 mV·s-1.

2.2 Characterization The electrochemical spectra were recorded by an electrochemical workstation (CS2350, CorrTest Corp., China). The microelectrode fabricated by the lift-off process (Fig.1a) with an area of 0.38 mm2 served as the working electrode. A Pt foil was utilized as the counter electrode, and a saturated calomel electrode (SCE) was used as the reference electrode. All solutions were deaerated by bubbling high-pure N2 for 15 min. The working electrode was polished with 0.05 µm alumina powder, scanned in 1 M HNO3 several times and washed thoroughly before characterizations.

The scanning electron microscope (SEM) morphology was visualized by JSM-7100F (JEOL Ltd., Japan) operated at 20 kV. Before SEM analysis, OS layers were coated with 30 nm gold as a conductive layer. The atomic force microscope (AFM) characterizations were carried out under the Bruker Dimension Icon (Bruker, USA) operating in the tapping mode. The thickness of the OS layer was evaluated by the acquisition of topographical images of 35 μm × 35 μm scanning area.

It has been demonstrated that the green line energetically close to the π-π* transition would be more suitable for Raman characterizations of electropolymerized PEDOT:PSS [36] at the neutral state. Thus, vibration spectra were obtained by using a laser Raman spectrophotometer (LabRAM HR Evolution, HORIBA Jobin Yvon, France) with a green laser. In the experiment, the selected wavelength was 532 nm, and the laser power intensity was 13 mW·(cm2)-1. All measurements were conducted within 60 seconds accumulating one time and utilizing 5% of the laser power.

The field effect characteristics of OECTs were measured using a Keithley 2636B source meter (Keithley Instruments, USA). Typical electrical performance parameters such as the threshold

voltage (VT), the on/off current ratio, and the transconductance (gm) were extracted from I-V curves such as the output characteristic curve and the transfer characteristic curve. Moreover, the gate current (IG) was simultaneously acquired during measurements.

3 Results and discussion Although several mechanisms have been reported [37], the electropolymerization of EDOT in NaPSS is considered a complex process [38]. Generally, it begins the initial oxidation of the EDOT monomer. After oxidation, radical cations return to the solution, and the high-density oligomeric region can be built. Then, oligomers are deposited, and different nucleation processes occur depending on the polymerization potential and the monomer concentration [39]. With regards to the chain propagation, mechanisms of the processes are still controversial due to influences of homogeneous or comproportionation reactions [40].

Herein, the cyclic voltammograms with scan rates ranging from 10 mV·s-1 to 100 mV·s-1 were conducted in 104 μM EDOT and 102 μM NaPSS (Ⅲ) between -0.8 V and +2.0 V (vs SCE) (Fig.2a). According to the literature [41], the oxidation peak P1 around 1.17 V (when the scan rate is 10 mV·s-1) is attributed to the diffusion of monomers. It has been verified by the inset of Fig.2a, which shows a linear relationship between the P1 current density and the square root of the scan rate [42]. In addition, P1 locates at a higher potential as the scan rate increases because a stronger concentration polarization could result in a larger overpotential.

Then, cyclic voltammograms with the NaPSS concentration ranging from 1 μM to 103 μM were measured. To compare results with the references [42, 43], the scan rate was set as 50 mV·s-1. It has been demonstrated that [38, 44] the use of NaPSS counterions allowed increasing the EDOT solubility due to the formation of the

polyelectrolyte complex. Besides, the oxidation potential of EDOT monomer decreased with increasing NaPSS concentration due to the stabilization effect of dodecyl sulfate anions to EDOT radicals [45, 46]. As illustrated in Fig.2b, P1 peaks of 103 μM (Ⅳ) and 102

μM (Ⅲ) NaPSS locate at 1.10 V and 1.27 V with a current density of 8.51 and 4.84 mA·cm-2, respectively. It can be explained by the fact that the current density was improved due to that more EDOT participated in the reaction. As the oxidation potential of the EDOT monomer decreased, the mass transfer limitation of the monomer was more likely to occur at lower electrode potential. In terms of 10 μM (Ⅱ) and 1 μM (Ⅰ) NaPSS, nucleation loops substitute the oxidation peaks, indicating that electron-transfer kinetics became

the speed determination step. In this study, the concentration of NaPSS and EDOT used for

OS layers preparation was maintained at 102 μM and 104 μM, respectively, which was the same as that used in the cyclic voltammetry tests shown in Fig.2a. Besides, the sinusoidal AC frequency used for OS layers preparation was between 50 Hz and 500 kHz. The amplitude was set between 4 Vp-p and 8 Vp-p. Consequently, the average scan rate was set between 4×105 and 8×109 mV·s-1, which was at least 4000 times as high as the maximum scan rate (100 mV·s-1) used in the cyclic voltammetric tests. Thus, it is suggested that the AC electrodeposition of OS layers was diffusion limited [47].

Figure 3 (a) The schematic plot of the mass transport mechanism of OS layers preparation. (b) Re[k(ω)] versus AC frequency based on the data in the literature [57]. (c) E versus (VR)equ distributing on the electrode surface. One set of data is taken from the central line of the electrode pair. The other set of data is taken from the electrode edge parallel to the central line. (d) ∇E2 versus (VR)equ along the central line of the electrode pair. For (c) and (d), the position x=0 and y=0 locate at the center of the electrode pair.

According to the electrochemical impedance theory [48], the electric double layers (EDL) only undertake a part of the open-circuit voltage (VO). The un-screened fraction across the electrolyte resistor (VR) could be utilized to drive particles in bulk such as radicals, monomers, dimers, and oligomers [49]. In detail, the overall impedance (Z) is the sum of interfacial impedance resulting from the EDL capacitance (Cdl) in parallel with the faradaic impedance (ZF) plus the electrolyte resistance (R) (Eq. (1)). Concerning the diffusion limited mass transport, ZF is independent of the AC frequency (ω/2π) [50], and hence, VR increases when ω and VO increase (Eq. (2)). Z = R + ZF

1+jωCdlZF Equation 1

VR = R

R+ ZF1+jωCdlZF

× VO Equation 2

In the polymerization reaction, free radicals oxidized from monomers would move along the electric field lines until encountering other radicals and forming dimers as well as subsequent oligomers [51, 52]. Thus, both mass transport of neutral particles and charged free radicals exert significant influences on OS layers morphologies. In detail, neutral particles

such as monomers and oligomers would be subjected to the dielectrophoresis (DEP) force (Eq. (3)) in the non-uniform electric field and be concentrated in the middle of the electrode pair (Fig.3a). DEP assembly would lead to particles structuring when higher concentrations of particles are present [53]. Then, the dipoles induced in the particles would interact with each other through the chaining force (Eq. (4)) if particles are close enough [54]. Besides, radical cations with charge q are subjected to an electrophoretic force FEP (Eq. (5)) [55]. In sum, FDEP, Fchain, and FEP are all related to the electric field strength (E). Therefore, the morphologies of OS layers could be tuned by the AC parameters. FDEP = 2πr3εmRe[k(ω)]∇E2 Equation 3 Fchain = Cπεmr2(k(ω))2E2 Equation 4 FEP = qE Equation 5 where εm is the permittivity of the medium and r is the radius of particles. C depends on geometrical parameters such as the length of particle chains and the distance between the particles. Besides, k(ω) indicates the Clausius-Mossotti factor, and Re[k(ω)] indicates the real component [56].

Figure 4 shows AFM images and AFM profiles of OS layers. The electrolyte used here was Ⅲ. The AC parameters used for (a) were 50 Hz, 4 Vp-p and no DC Offset; for (b) were 1 kHz, 4 Vp-p and 1.0 V DC Offset; for (c) were 10 kHz, 4 Vp-p and 1.0 V DC Offset; for (d) were 20 kHz, 4 Vp-p and 1.0 V DC Offset; for (e) were 500 kHz, 6 Vp-p and 1.0 V DC Offset; for (f) were 500 kHz, 8 Vp-p and no DC Offset. The corresponding SEM images were illustrated in Fig.S4 in the ESM.

As illustrated in Fig.4, OS layers with different morphologies could be obtained. For experiments shown in Fig.4a, 4b, and 4c, 4 Vp-p was used, and the frequency was 50 Hz, 1 kHz, and 10 kHz, respectively. As the frequency increased, the morphology changed from microfilms (Fig.4a) to nanodendrites (Fig.4b), and then to nanowires (Fig.4c). Besides, the frequency exerted significant influences on the nanowire’s diameter. The diameter shown in Fig.4c ranging from 800 nm to 1.6 μm is almost three times larger than that in Fig.4d. However, the frequency used in Fig.4c was only half of that in Fig.4d.

Experimentally, the minimum and maximum VO used were 4 Vp-p and 8 Vp-p; the minimum and maximum frequency used were 50 Hz and 500 kHz, respectively. Hence, the equivalent voltage drops (VR)equ across the electrolyte resistor ranged from the lower limit 0 Vp-p to the upper limit 8 Vp-p. The former one corresponded to the situation that the frequency was small enough and all VO dropped across the interface impedance. Whereas the latter one corresponded to the fact that the frequency was large enough and all VO dropped across R. The finite element method (FEM) simulations (Fig.3c and 3d) demonstrate that both E and ∇E2 increase as (VR)equ increases. Because VR increases as the frequency increases and ZF is independent of the frequency for the diffusion limited mass transport (Eq. (2)), it is reasonable to speculate that both E and ∇E2 would increase as AC frequency increases qualitatively.

As discussed above, Fchain depends on E2 linearly (Eq. (4)). Thus, Fchain increases with the increasing frequency, and oligomers could be more prone to polymerize in a finer nanowire form. When it comes to FDEP, the situation gets a little more complicated. As shown in Eq. (3), FDEP depends on both ∇E2 and Re[k(ω)]. According to the existing data [57], Re[k(ω)] remains to be 1.0 when AC frequency is lower than 10 MHz (Fig.3b) since particles exhibit much higher polarizability than the electrolyte media [32]. Based the fact, it can be considered that the frequency mainly affects FDEP through the ∇E2 term.

Collectively, both Fchain and FDEP increase as the frequency increases. Thus, the diameter of nanowires decreased as AC frequency increased (Fig.4c and 4d). However, the diameter cannot be reduced continuously because the voltage drop (VO-VR)/2 across

the electrode/electrolyte interface decreased as the frequency increased, and electropolymerization reactions cannot proceed spontaneously once (VO-VR)/2 was smaller than the nucleation potential [58]. In addition, polymerization reactions rely on the transport of free radicals. However, radical cations cannot respond to the change of AC frequency if the period of the AC voltage was less than the relaxation time of radical cations, in which case the mass transfer by electrophoresis would expire. In the experiment, the electrode potential was insufficient to drive OS layers to grow when VO was set as 4 Vp-p and the frequency exceeded 500 kHz. If VO was raised to 6 Vp-p (Fig.4e), nanowires with smaller diameters (200 nm) can be obtained again. However, there were bifurcations at the electrode tip. It is similar to the case of metal electrodeposition, where higher overpotentials facilitate the instantaneous nucleation [59]. The phenomenon was further verified in Fig.4f, nanowires with an average diameter of 200 nm grew simultaneously along the electric field lines when VO was raised to 8 Vp-p.

The growth direction of OS layers could also be controlled. As shown in Fig.4b-e, when 1 V DC offset was applied, OS layers started to grow from the end where the positive DC bias was applied. If no DC bias was set, OS layers grew simultaneously at both ends, encountered and connected at the intermediate position (Fig.4a and 4f). Once two parts of OS layers were connected, a current loop formed and OS layers worked as resistors. FDEP , Fchain and FEP that used for the particle manipulation disappeared, and hence the unconnected parts stopped growing as shown in Fig.4f. The typical growth processes with and without a DC bias were demonstrated in Mov.S1 and Mov.S2 (in the ESM).

In order to investigate the influence of AC parameters on the OS layers structures, the Raman spectroscopy characterization was performed. Vibrational bands under 532 nm lasers were selected according to the literature (Fig.5a and 5b) [60]. Taking microfilms (electrolyte Ⅲ, 4 Vp-p, 50 Hz) as an example, the most intense band at 1432 cm-1 or 1430 cm-1 is assigned to the symmetric stretching mode of the aromatic Cα=Cβ band. Less intense bands at 1510 cm-1 and 1560 cm-1 are characteristics of asymmetric Cα=Cβ stretching vibrations. Weaker bands at 1366 cm-1 and 1267 cm-1 are assigned to the Cβ-Cβ stretching vibration and the Cα-Cα inter-ring stretching

Nano Res.

vibration, respectively. Finally, the band at 1105 cm-1 is assigned to the C–O–C deformation.

Figure 5 shows the Raman spectra of OS layers in the range of 1000-1800 cm-1. The electrolyte used was Ⅲ. The AC frequency was set at 50 Hz, and the amplitude was increased from 3.6 Vp-p to 4.0 Vp-p without DC bias for (a, c, e); The AC amplitude was set at 4.0 Vp-p with 1 V DC offset, and the frequency was increased from 20 kHz to 400 kHz for (b, d, f). The relative intensities of bands at 1560, 1510, 1366, 1267 and 1105 cm-1 were normalized to the Cα=Cβ band.

Two AC parameters intervals were investigated. The first one was used for microfilms preparation. The AC frequency was set at 50 Hz, and the amplitude was increased from 3.6 Vp-p to 4.0 Vp-p without DC bias (Fig.5a, 5c, and 5e). The second one was used for nanowire preparation. The AC amplitude was set at 4.0 Vp-p with 1 V DC offset, and the frequency was increased from 20 kHz to 400 kHz (Fig.5b, 5d, and 5f). As shown in Fig.5, All bands were normalized by the Cα=Cβ band. In general, there is no significant change in peak positions. However, both the normalized intensity (Fig.5c and 5d) and the full width at half maximum (FWHM) (Fig.5e and 5f) of bands 1510 cm-1 and 1560 cm-1 increase as the amplitude and the frequency of AC voltage increase.

As noted in the literature [61], no offset of Cα=Cβ band indicates that the prepared OS layers are in the neutral state, which could be a proof of correctness of the laser wavelength selection. Besides, alterations of normalized intensities and FWHM could be ascribed to the change of structural defects and conjugation lengths

of polymer chains [43, 62]. As has been discussed, VR increases as AC frequency and VO increase (Eq. (2)). Thus, higher amplitudes and frequencies of AC voltage lead to the greater FDEP (Eq. (3)) and Fchain (Eq. (4)), which would result in higher internal stress and more backbone structural defects in the polymer chains. Besides, radical cations move faster under the higher electric field strength. Thus, polymerization could occur at a larger spatial scale, and the polymer chains could precipitate with a longer conjugation length.

The transfer characteristic curves were obtained by acquiring the drain current (Id) when fixing the drain voltage (Vd) at −0.6 V and sweeping the gate voltage (Vg) between -0.7 V and 0 V. As illustrated by the Id-Vg curves in Fig.6a-6f, Id is non-zero when no voltage was applied on the gate. Besides, Id increases as the absolute value of Vg increases. It confirms that OECTs are working as p-channel depletion mode as demonstrated in Eq. (6)

PEDOT+: PSS− + Na+ + e− ↔ PEDOT0 + Na+: PSS− Equation 6 Particularly, when a positive Vg is applied, cations in the electrolyte are injected into the OS layer leading to holes depletion as well as conductivity reduction. Thus, OECTs reach the off state when the OS channel is completely de-doped. Conversely, the application of a negative Vg would give rise to cations retraction from the OS layer resulting in holes accumulation as well as conductivity increment. Output characteristic curves have also been acquired by spanning Vd between -0.6 V and 0.6 V and fixing Vg in the range from -0.7 V to 0 V (Fig.S5). Id saturates when a negative Vg and a negative Vd are applied. No saturation is observed when Vd is positive, by which device working mode was verified [63].

According to the“Bernards model” [64], gm of OECTs defining as the slope of the Id-Vg curve is directly proportional to WDL-1·μC*, where μ is the hole mobility; C* is the volumetric

capacitance; W, D and L are channel width, thickness and length, respectively. Experimentally, the OS layers illustrated in Fig.4a-c were grown in the equal electrode spacing and thus have the same L. In addition, the average D and W of the microfilm (Fig.4a) are much larger than that of dendrites (Fig.4b) and nanowires (Fig.4c). Qualitatively, it can be expected that gm of microfilm is higher than that of dendrites and nanowires. This is consistent with the results illustrated in Fig. 6b, 6d, and 6f, where the maximum transconductance (gm,max) is 832 μS, 61 μS, and 49 μS, respectively.

As mentioned above, the steady-state device performance could also be improved by optimizing the product of μC* [65]. It has been demonstrated that PEDOT chains and GQDs could bond tightly together through the strong π-π interactions, which would lower the energy barrier and provide more conductive pathways for the charge transport in PEDOT:PSS [66]. Thus, microfilms were

Nano Res.

prepared by co-electrodeposition of EDOT and GQDs in NaPSS to improve device performance further. The content of the GQDs was varied (Ⅲ, Ⅴ, Ⅵ, Ⅶ, Ⅷ in Table 1) and optimized to give the highest gm (Fig.S6 (in the ESM)). As illustrated in Fig.S7 (in the ESM), there is no significant change of microfilm morphologies when GQDs were co-electrodeposited. But, gm,max is improved to 1.576 mS (Fig.6a) when the mass concentration of GQDs was 2 μg·L-1. The result is comparable to the value reported in the literature [33, 67]. The Raman spectra of PEDOT:PSS/GQDs composites have also been conducted, which indicated that PEDOT:PSS peaks overlapped with the D- and G- bands strongly as the GQDs content used was much less than 1 wt% [68].

Except for gm, electrical performance such as the on/off current ratio, VT and IG were also obtained. As summarized in Table S1 (in the ESM), the on/off current ratios of microfilms are higher than that of nanodendrites and nanowires. Co-electrodeposition of GQDs improves the on/off current ratio significantly. Taking microfilms as an example, GQDs increase the on/off current ratio from 166 to 246. Results of nanowires are in the same order of magnitude as that reported in the literature [63]. VT could be determined by fitting the linear region of Id

1/2-Vg curves. As shown, VT of microfilms is also higher than that of nanodendrites and nanowires indicating that the power consumption of OECTs based on the microfilms would be relatively higher than others. Co-electrodeposition of GQDs led to

a slight increase of VT in the experiment. With regards to IG (the second column in Fig.6), the maximum occurs when no voltage was applied on the gate indicating the contribution of IG to the power consumption [13]. For nanodenrite and nanowire OS layers co-electrodeposted with GQDs, IG is almost one-tenth of Id, which indicates that the contribution of IG to Id modulation is also not negligible [69].

Although in situ fabrication of OECTs on a microfluidic chip was achieved in this work, how to achieve massive production is still a tricky problem as current electrodeposition techniques can only produce one transistor each time. Besides, the reproducibility of the device performance was investigated as shown in Fig.S8 (in the ESM), where ten devices co-electrodeposited with (III) and without (Ⅵ) GQDs were tested separately. Error bars represent the standard deviation of normalized values, and the AC parameters used were 50 Hz, 4 Vp-p and no DC Offset. Generally, the reproducibility is comparable to that of inkjet printing devices in the literature [70] but lower than that of spin coating ones [65]. The phenomenon may be caused primarily by the lower controllability of OS layers shape in the electrodeposition. As the relation between the surface roughness and the conductivity of OS layers has not been evidenced clearly [71], the influence of morphologies also needs to be taken into account in the next step. Related work is in progress in our group.

Figure 6 The transfer characteristic curves (the first column) and IG (the second column) were acquired when Vd was set as -0.6 V. The data set (a) and (b) are for microfilms prepared by electrolyte Ⅵ and Ⅲ with same AC parameters used in Fig.4a; the data set (c) and (d) are for nanodendrites prepared by electrolyte Ⅵ and Ⅲ with same AC parameters used in Fig.4b; the data set (e) and (f) are for nanowires prepared by electrolyte Ⅵ and Ⅲ with same AC parameters used in Fig.4c. Characterizations were operated in a 100 mM NaCl solution with an Ag/AgCl gate electrode.

4 Conclusions Based on the AC electrodeposition, PEDOT:PSS/GQDs OS layers with different morphologies such as microfilms, nanodendrites and nanowires were prepared in situ on a microfluidic chip. The channel length of OS layers was 8 μm defined by the prepared electrode gap. The channel height of OS layers could be reduced to 200 nm by tuning the AC parameters. The growth process of OS layers was found to be diffusion-limited, which was verified by the cyclic voltammetric characterization. In order to elucidate the control mechanism of morphologies and growth directions, mass transport based on the DEP force, the chaining force and the electrophoretic force were discussed. The normalized intensity and

FWHM of Raman spectra demonstrated that both the amplitude and the frequency of the AC voltage exert significant influences on internal stresses and backbone structural defects.

The transfer and output characteristics of OECTs were investigated when the p-channel device was working in the depletion mode. Generally, the field effect performance of OECTs with microfilm OS layers is better than the nanodendrite and nanowire ones. The gm,max, on/off current ratio and VT measured in the experiment was 832 μS, 166 and 0.234 V, respectively. To improve OECTs performance, GQDs were co-electrodeposited with EDOT in NaPSS. The on/off current ratio, gm,max and VT were improved to 246, 1.58 mS, and 0.269 V, respectively

Overall, the use of AC electrodeposition to prepare OS layers in situ could integrate and miniaturize OECTs-based microfluidic

Nano Res.

chips. As a common challenge, broad space exists in achieving high mass production and device reproductivity, especially for sensing applications. In order to further improve OECTs performance, the controllability of the OS layers shapes such as thickness and width need further improvement. Besides, more efforts are required to study the influences of AC parameters on the morphology, which would influence the conductivity, as well as the structure, which would determine the volumetric capacitance of OS layers.

Acknowledgements This work is supported by National Science Foundation of China No. 51705354, 51622507, 61671271; Scientific and Technologial Innovation Programs of Higher Education Institutions in Shanxi No. 183290224-S and 201802029. Electronic Supplementary Material: Supplementary material (TEM, XRD, SEM, reproducibility results, electrical characteristics and video clips) is available in the online version of this article at http://dx.doi.org/10.1007/s12274-***-****-* (automatically inserted by the publisher).

References [1] Rivnay, J.; Inal, S.; Salleo, A.; Owens, R. M.; Berggren, M.; Malliaras, G.

G. Organic electrochemical transistors. Nat. Rev. Mater. 2018, 3, 17086. [2] Tang, H.; Yan, F.; Lin, P.; Xu, J. B.; Chan, H. L. W. Highly sensitive

glucose biosensors based on organic electrochemical transistors using platinum gate electrodes modified with enzyme and nanomaterials. Adv. Funct. Mater. 2011, 21, 2264-2272.

[3] Tang, H.; Lin, P.; Chan, H. L. W.; Yan, F. Highly sensitive dopamine biosensors based on organic electrochemical transistors. Biosens. Bioelectron. 2011, 26, 4559-4563.

[4] Lin, P.; Luo, X.; Hsing, I. M.; Yan, F. Organic electrochemical transistors integrated in flexible microfluidic systems and used for label-free DNA sensing. Adv. Mater. 2011, 23, 4035-+.

[5] Hempel, F.; Law, J. K.-Y.; Thanh Chien, N.; Munief, W.; Lu, X.; Pachauri, V.; Susloparova, A.; Xuan Thang, V.; Ingebrandt, S. Pedot:Pss organic electrochemical transistor arrays for extracellular electrophysiological sensing of cardiac cells. Biosens. Bioelectron. 2017, 93, 132-138.

[6] Lin, P.; Yan, F.; Yu, J.; Chan, H. L. W.; Yang, M. The application of organic electrochemical transistors in cell-based biosensors. Adv. Mater. 2010, 22, 3655-+.

[7] Gu, X.; Yao, C.; Liu, Y.; Hsing, I. M. 16-channel organic electrochemical transistor array for in vitro conduction mapping of cardiac action potential. Adv. Healthc. Mater. 2016, 5, 2345-2351.

[8] Liang, Y.; Ernst, M.; Brings, F.; Kireev, D.; Maybeck, V.; Offenhaeusser, A.; Mayer, D. High performance flexible organic electrochemical transistors for monitoring cardiac action potential. Adv. Healthc. Mater. 2018, 7, 1800304.

[9] White, S. P.; Dorfman, K. D.; C Daniel, F. Label-free DNA sensing platform with low-voltage electrolyte-gated transistors. Anal. Chem. 2015, 87, 1861-1866.

[10] Parlak, O.; Keene, S. T.; Marais, A.; Curto, V. F.; Salleo, A. Molecularly selective nanoporous membrane-based wearable organic electrochemical device for noninvasive cortisol sensing. Sci. Adv. 2018, 4, 2904.

[11] Curto, V. F.; Marchiori, B.; Hama, A.; Pappa, A.-M.; Ferro, M. P.; Braendlein, M.; Rivnay, J.; Fiocchi, M.; Malliaras, G. G.; Ramuz, M. et al. Organic transistor platform with integrated microfluidics for in-line multi-parametric in vitro cell monitoring. Microsyst. Nanoeng. 2017, 3, 17028.

[12] Liao, Z.; Wang, J.; Zhang, P.; Zhang, Y.; Miao, Y.; Gao, S.; Deng, Y.; Geng, L. Recent advances in microfluidic chip integrated electronic biosensors for multiplexed detection. Biosens. Bioelectron. 2018, 121, 272-280.

[13] Kim, S. H.; Hong, K.; Xie, W.; Lee, K. H.; Zhang, S.; Lodge, T. P.; Frisbie, C. D. Electrolyte-gated transistors for organic and printed electronics. Adv.

Mater. 2013, 25, 1822-1846. [14] Friedlein, J. T.; McLeod, R. R.; Rivnay, J. Device physics of organic

electrochemical transistors. Org. Electron. 2018, 63, 398-414. [15] Zhang, M.; Lin, P.; Yang, M.; Yan, F. Fabrication of organic

electrochemical transistor arrays for biosensing. Biochim. Biophys. Acta-Gen. Subj. 2013, 1830, 4402-4406.

[16] DeFranco, J. A.; Schmidt, B. S.; Lipson, M.; Malliaras, G. G. Photolithographic patterning of organic electronic materials. Org. Electron. 2006, 7, 22-28.

[17] 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. 2013, 25, 2135-2139.

[18] Curto, V. F.; Ferro, M. P.; Mariani, F.; Scavetta, E.; Owens, R. M. A planar impedance sensor for 3d spheroids. Lab Chip 2018, 18, 933-943.

[19] Pappa, A. M.; Curto, V. F.; Braendlein, M.; Strakosas, X.; Donahue, M. J.; Fiocchi, M.; Malliaras, G. G.; Owens, R. M. Organic transistor arrays integrated with finger-powered microfluidics for multianalyte saliva testing. Adv. Healthc. Mater. 2016, 5, 2295-2302.

[20] Guo, X.; Liu, J.; Liu, F.; She, F.; Zheng, Q.; Tang, H.; Ma, M.; Yao, S. Label-free and sensitive sialic acid biosensor based on organic electrochemical transistors. Sens. Actuator B-Chem. 2017, 240, 1075-1082.

[21] Fonseca, S. M.; Moreira, T.; Jorge Parola, A.; Pinheiro, C.; Laia, C. A. T. Pedot electrodeposition on oriented mesoporous silica templates for electrochromic devices. Energy Mater. Sol. Cells 2017, 159, 94-101.

[22] Shi, Y.; Zhang, Y.; Tang, K.; Song, Y.; Cui, J.; Shu, X.; Wang, Y.; Liu, J.; Wu, Y. In situ growth of pedot/graphene oxide nanostructures with enhanced electrochromic performance. RSC Adv. 2018, 8, 13679-13685.

[23] Demuru, S.; Deligianni, H. Surface pedot:Nafion coatings for enhanced dopamine, serotonin and adenosine sensing. J. Electrochem. Soc. 2017, 164, G129-G138.

[24] Si, W.; Lei, W.; Han, Z.; Hao, Q.; Zhang, Y.; Xia, M. Selective sensing of catechol and hydroquinone based on poly(3,4-ethylenedioxythiophene)/nitrogen-doped graphene composites. Sens. Actuator B-Chem. 2014, 199, 154-160.

[25] Song, J.; Noh, H.; Lee, J.; Nah, I.-W.; Cho, W.-I.; Kim, H.-T. In situ coating of poly(3,4-ethylenedioxythiophene) on sulfur cathode for high performance lithium-sulfur batteries. J. Power Sources 2016, 332, 72-78.

[26] Si, W.; Lei, W.; Han, Z.; Zhang, Y.; Hao, Q.; Xia, M. Electrochemical sensing of acetaminophen based on poly(3,4-ethylenedioxythiophene)/graphene oxide composites. Sens. Actuator B-Chem. 2014, 193, 823-829.

[27] Koizumi, Y.; Shida, N.; Ohira, M.; Nishiyama, H.; Tomita, I.; Inagi, S. Electropolymerization on wireless electrodes towards conducting polymer microfibre networks. Nat. Commun. 2016, 7, 10404.

[28] Yang, X.; Wang, Q.; Fan, J.; Zhang, M.; Zhou, Z.; Ji, J. Extremely sensitive sers detection through the one-step fabrication of a composite nanostructure of substrate and analyte. ECS Electrochem. Lett. 2015, 4, B17-B20.

[29] Ji, J.; Li, P.; Sang, S.; Zhang, W.; Zhou, Z.; Yang, X.; Dong, H.; Li, G.; Hu, J. Electrodeposition of au/ag bimetallic dendrites assisted by faradaic ac-electroosmosis flow. AIP Adv. 2014, 4, 031329.

[30] Zhang, M.; Yang, X.; Zhou, Z.; Ye, X. Controllable growth of gold nanowires and nanoactuators via high-frequency ac electrodeposition. Electrochem. Commun. 2013, 27, 133-136.

[31] Chen, S.; Hai, X.; Xia, C.; Chen, X. W.; Wang, J. H. Preparation of excitation-independent photoluminescent graphene quantum dots with visible-light excitation/emission for cell imaging. Chem.-Eur. J. 2013, 19, 15918-15923.

[32] Oberhammer, J.; Stemme, G. Bcb contact printing for patterned adhesive full-wafer bonded 0-level packages. J. Microelectromech. Syst. 2005, 14, 419-425.

[33] Khodagholy, D.; Rivnay, J.; Sessolo, M.; Gurfinkel, M.; Leleux, P.; Jimison, L. H.; Stavrinidou, E.; Herve, T.; Sanaur, S.; Owens, R. M. High transconductance organic electrochemical transistors. Nat. Commun. 2013, 4, 2133.

Nano Res.

[34] Stavrinidou, E.; Leleux, P.; Rajaona, H.; Khodagholy, D.; Rivnay, J.; Lindau, M.; Sanaur, S.; Malliaras, G. G. Direct measurement of ion mobility in a conducting polymer. Adv. Mater. 2013, 25, 4488-4493.

[35] Ji, J.; Zhou, Z.; Yang, X.; Zhang, W.; Sang, S.; Li, P. One-dimensional nano-interconnection formation. Small 2014, 9, 3014-3029.

[36] Garreau, S.; Duvail, J. L.; Louarn, G. Spectroelectrochemical studies of poly(3,4-ethylenedioxythiophene) in aqueous medium. Synth. Met. 2001, 125, 325-329.

[37] Heinze, J.; Frontana-Uribe, B. A.; Ludwigs, S. Electrochemistry of conducting polymers-persistent models and new concepts. Chem. Rev. 2010, 110, 4724-4771.

[38] Castagnola, V.; Bayon, C.; Descamps, E.; Bergaud, C. Morphology and conductivity of pedot layers produced by different electrochemical routes. Synth. Met. 2014, 189, 7-16.

[39] Randriamahazaka, H.; Noel, V.; Chevrot, C. Nucleation and growth of poly (3, 4-ethylenedioxythiophene) in acetonitrile on platinum under potentiostatic conditions. J. Electroanal. Chem. 1999, 472, 103-111.

[40] Heinze, J.; Rasche, A.; Pagels, M.; Geschke, B. On the origin of the so-called nucleation loop during electropolymerization of conducting polymers. J. Phys. Chem. B 2007, 111, 989-997.

[41] Du, X.; Wang, Z. Effects of polymerization potential on the properties of electrosynthesized pedot films. Electrochim. Acta 2003, 48, 1713-1717.

[42] Tamburri, E.; Orlanducci, S.; Toschi, F.; Terranova, M. L.; Passeri, D. Growth mechanisms, morphology, and electroactivity of pedot layers produced by electrochemical routes in aqueous medium. Synth. Met. 2009, 159, 406-414.

[43] Tamburri, E.; Guglielmotti, V.; Orlanducci, S.; Terranova, M. L. Structure and i2/i- redox catalytic behaviour of pedot-pss films electropolymerized in aqueous medium: Implications for convenient counter electrodes in dssc. Inorg. Chim. Acta 2011, 377, 170-176.

[44] Youk, J. H.; Locklin, J.; Xia, C.; Park, M.-K.; Advincula, R. Preparation of gold nanoparticles from a polyelectrolyte complex solution of terthiophene amphiphiles. Langmuir 2001, 17, 4681-4683.

[45] Schweiss, R.; Lubben, J. F.; Johannsmann, D.; Knoll, W. Electropolymerization of ethylene dioxythiophene (edot) in micellar aqueous solutions studied by electrochemical quartz crystal microbalance and surface plasmon resonance. Electrochim. Acta 2005, 50, 2849-2856.

[46] Bobacka, J.; Lewenstam, A.; Ivaska, A. Electrochemical impedance spectroscopy of oxidized poly (3, 4-ethylenedioxythiophene) film electrodes in aqueous solutions. J. Electroanal. Chem. 2000, 489, 17-27.

[47] Melato, A. I.; Viana, A. S.; Abrantes, L. M. Different steps in the electrosynthesis of poly(3,4-ethylenedioxythiophene) on platinum. Electrochim. Acta 2008, 54, 590-597.

[48] Orazem, M. E.; Tribollet, B. Electrochemical impedance spectroscopy. John Wiley & Sons: New York, 2011.

[49] Thapa, P. S.; Ackerson, B. J.; Grischkowsky, D. R.; Flanders, B. N. Directional growth of metallic and polymeric nanowires. Nanotechnology 2009, 20, 235307.

[50] Botasini, S.; Mendez, E. Limited diffusion and cell dimensions in a micrometer-layer of solution: An electrochemical impedance spectroscopy study. ChemElectroChem 2017, 4, 1891-1895.

[51] Scharifker, B. R.; García-Pastoriza, E.; Marino, W. The growth of polypyrrole films on electrodes. J. Electroanal. Chem. 1991, 300, 85-98.

[52] Garfias-García, E.; Romero-Romo, M.; Ramirez-Silva, M. T.; Morales, J.; Palomar-Pardavé, M. Mechanism and kinetics of the electrochemical formation of polypyrrole under forced convection conditions. J. Electroanal. Chem. 2008, 613, 67-79.

[53] Velev, O. D.; Gangwal, S.; Petsev, D. N. Particle-localized ac and dc manipulation and electrokinetics. Annual Reports Section "C" (Physical

Chemistry) 2009, 105, 213-246. [54] Velev, O. D.; Bhatt, K. H. On-chip micromanipulation and assembly of

colloidal particles by electric fields. Soft Matter 2006, 2, 738-750. [55] Velev, O. D. Assembly of electrically functional microstructures from

colloidal particles. Colloids and Colloid Assemblies: Synthesis, Modification, Organization and Utilization of Colloid Particles, Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2004.

[56] Jones, T. B.; Jones, T. B. Electromechanics of particles. Cambridge University Press: Cambridge, 2005.

[57] Du, Y.; Cui, X.; Li, L.; Tian, H.; Yu, W. X.; Zhou, Z. X. Dielectric properties of dmso‐doped‐ p edot:Pss at thz frequencies. Phys. Status Solidi B-Basic Solid State Phys. 2017, 255, 1700547

[58] Randriamahazaka, H. Electrodeposition mechanisms and electrochemical behavior of poly(3,4-ethylenedithiathiophene). J. Phys. Chem. C 2007, 111, 4553-4560.

[59] Gamburg, Y. D.; Zangari, G. Theory and practice of metal electrodeposition. Springer-Verlag: New York, 2011.

[60] Sakmeche, N.; Aeiyach, S.; Aaron, J. J.; Jouini, M.; And, J. C. L.; Lacaze, P. C. Improvement of the electrosynthesis and physicochemical properties of poly(3,4-ethylenedioxythiophene) using a sodium dodecyl sulfate micellar aqueous medium. Langmuir 1999, 15, 2566-2574.

[61] Garreau, S.; Louam, G.; Lefrant , S. Optical study and vibrational analysis of the poly (3, 4-ethylenedioxythiophene)(pedt). Synth. Met. 1999, 101, 312-313.

[62] Akimoto, M.; Furukawa, Y.; Takeuchi, H.; Harada, I.; Soma, Y.; Soma, M. Correlation between vibrational spectra and electrical conductivity of polythiophene. Synth. Met. 1986, 15, 353-360.

[63] Alam, M. M.; Jun, W.; Yaoyao, G.; Lee, S. P.; Hsian-Rong, T. Electrolyte-gated transistors based on conducting polymer nanowire junction arrays. J. Phys. Chem. B 2005, 109, 12777-12784

[64] Bernards, D. A.; Malliaras, G. G. Steady-state and transient behavior of organic electrochemical transistors. Adv. Funct. Mater. 2007, 17, 3538-3544.

[65] Inal, S.; Malliaras, G. G.; Rivnay, J. Benchmarking organic mixed conductors for transistors. Nat. Commun. 2017, 8,1767.

[66] Pathak, C. S.; Singh, J. P.; Singh, R. Preparation of novel graphene-pedot: Pss nanocomposite films and fabrication of heterojunction diodes with n-si. Chem. Phys. Lett. 2018, 694, 75-81.

[67] Khodagholy, D.; Doublet, T.; Quilichini, P.; Gurfinkel, M.; Leleux, P.; Ghestem, A.; Ismailova, E.; Hervé, T.; Sanaur, S.; Bernard, C. In vivo recordings of brain activity using organic transistors. Nat. Commun. 2013, 4, 1575.

[68] Kepić, D. P.; Marković, Z. M.; Jovanović, S. P.; Peruško, D. B.; Budimir, M. D.; Holclajtner-Antunović, I. D.; Pavlović, V. B.; Marković, B. M. T. Preparation of pedot:Pss thin films doped with graphene and graphene quantum dots. Synth. Met. 2014, 198, 150-154.

[69] Giuseppe, T.; Anna Giulia, B.; Nicola, C.; Simone, M.; Pasquale, D. A.; Stefano, B.; Matteo, C.; Paolo, C.; Fabio, S.; Roberto, M. Liposome sensing and monitoring by organic electrochemical transistors integrated in microfluidics. Biochim. Biophys. Acta-Gen. Subj. 2013, 1830, 4374-4380.

[70] Kim, S. H.; Hong, K.; Lee, K. H.; Frisbie, C. D. Performance and stability of aerosol-jet-printed electrolyte-gated transistors based on poly(3-hexylthiophene). ACS Appl. Mater. Interfaces 2013, 5, 6580-6585.

[71] Louet, C.; Cantin, S.; Dudon, J.-P.; Aubert, P.-H.; Vidal, F.; Chevrot, C. A comprehensive study of infrared reflectivity of poly(3,4-ethylenedioxythiophene) model layers with different morphologies and conductivities. Sol. Energy Mater. Sol. Cells 2015, 143, 141-151.

Electronic Supplementary Material

In situ fabrication of organic electrochemical transistors on a microfluidic chip

Jianlong Ji1,2, Mangmang Li1, Zhaowei Chen3, Hongwang Wang1, Xiaoning Jiang2, Kai Zhuo1, Ying Liu1, Xing Yang4, Zhen Gu3, Shengbo Sang1 (), Yang Shu5 ()

1 College of Information and Computer, Taiyuan University of Technology, Taiyuan 030024, China 2 Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh 27695, USA 3 Department of Bioengineering, California Nanosystems Institute, University of California, Los Angeles, Los Angeles 90095, USA 4 The State Key Laboratory of Precision Measurement Technology and Instruments, Department of Precision Instrument, Tsinghua University, Beijing 100084, China 5 Department of Chemistry, Colleges of Sciences, Northeastern University, Shenyang 110819, China Supporting information to DOI 10.1007/s12274-****-****-* (automatically inserted by the publisher)

Address correspondence to S. Sang, sunboa-sang@tyut.edu.cn; Y. Shu, shuyang@mail.neu.edu.cn

Nano Res.

Figure S1 shows the morphology of GQDs observed by FEI G20 (Hillsboro, OR, USA). (a) is the obtained TEM image; (b) is the high-resolution TEM image; (c) is the particle size distribution.

Figure S2 ESCALAB 250Xi (Thermo Scientific, America) was utilized to determine the element components and chemical states of the graphene quantum dots (GQDs). The C1s spectrum shows that the carbon element is presented in three different chemical environments (Fig.S2b), which correspond to the sp2 C in graphene (C–C or C=C) at 285.1 eV, sp3 C from C–N and/or C–O at 287.8 eV, and C=N at 287.3 eV, respectively. The high-resolution N 1s spectrum of the GQDs (Fig.S2c) shows two peaks at 399.2 and 400.7 eV, which are attributed to the pyrrolic N (C–N–C) and graphitic N bands, respectively. It is clear that the pyrrolic N is the main N-binding configuration for the N-doped GQDs, which is in good agreement with the previous reports about the pyrrolic N-doped graphene. In fact, the carbons on the graphene edges can be more easily replaced by N element due to their high reaction activity compared with the inner faultless sp2-bonded carbon atoms (such as the graphitic N), and thus benefits to the formation of pyrrolic N. The more pyrrolic N, the better charge carriers transfer property of the GQDs. Besides, the O1s spectrum in Fig.S2d shows two major peaks centered at 530.8 eV and 532.1 eV, which are ascribed to the C=O and C–O groups, respectively.

Nano Res.

Figure S3 shows SEM images of micro-electrodes before (a) and after (b) AC electrodeposition. The electrolyte used here was Ⅲ. The AC parameters were 30 kHz, 4 Vp-p, and 1.0 DC Offset.

Figure S4 shows SEM images of OS layers. The electrolyte used here was Ⅲ. The AC parameters used for (a) were 50 Hz, 4 Vp-p and no DC Offset; for (b) were 1 kHz, 4 Vp-p and 1.0 V DC Offset; for (c) were 10 kHz, 4 Vp-p and 1.0 V DC Offset; for (d) were 20 kHz, 4 Vp-p and 1.0 V DC Offset; for (e) were 500 kHz, 6 Vp-p and 1.0 V DC Offset; for (f) were 500 kHz, 8 Vp-p and no DC Offset.

Figure S5 shows output characteristic curves acquired when Vg ranged from -0.7 V to 0 V with the step of -0.1 V. The data set (a) and (b) are for microfilms prepared by electrolyte Ⅵ and Ⅲ, respectively, with the same AC parameters used in Fig.4a; the data set (c) and (d) are for nanodendrites prepared by electrolyte Ⅵ and Ⅲ, respectively, with the same AC parameters used in Fig.4b; the data set (e) and (f) are for nanowires prepared by electrolyte Ⅵ and Ⅲ, respectively, with the same AC parameters used in Fig.4c. Device characterizations were operated in a 100 mM NaCl solution with an Ag/AgCl gate electrode.

Nano Res.

Figure S6 illustrates gm,max of` OS microfilms versus the mass concentration of GQDs. Error bars represent the standard deviation of normalized values for n= 5 devices, and each device was measured three times.

Figure S7 shows SEM images of OS layers prepared by electrolytes Ⅲ (a), Ⅴ (b), Ⅵ (c), Ⅶ (d), Ⅷ (e) with different mass concentration of GQDs (0 μg·L-1, 0.2 μg·L-1, 2 μg·L-1, 20 μg·L-1, 200 μg·L-1). The AC parameters used were 50 Hz, 4 Vp-p, and no DC Offset.

Figure S8 illustrates gm,max, on/off current ratio and VT for OECTs based on the microfilms. The AC parameters used were 50 Hz, 4 Vp-p, and no DC Offset. The electrolyte used for (a) was Ⅲ; The electrolyte used for (b) was Ⅵ. Error bars represent the standard deviation of normalized values for n= 10 devices, and each device was measured three times. Table S1 illustrates the electrical characteristics of OS layers demonstrated in Fig.6

gm,max (mS) Vg| gm,max (V) Ion/off VT (V)

Fig.6a 1.57604 -0.25 246 0.269

Fig.6b 0.83174 -0.25 166 0.234

Fig.6c 0.14747 -0.3 19 0.135

Fig.6d 0.06112 -0.3 9 0.120

Fig.6e 0.06065 -0.3 13 0.152

Fig.6f 0.04932 -0.25 9 0.132