Phase Separation of P3HT/PMMA Blend Film for FormingSemiconducting and Dielectric Layers in Organic Thin-FilmTransistors for High-Sensitivity NO2 DetectionSihui Hou,† Junsheng Yu,† Xinming Zhuang,† Dengfeng Li,‡ Yiming Liu,‡ Zhan Gao,‡ Tianying Sun,§

Feng Wang,§ and Xinge Yu*,‡

†State Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Science and Engineering,University of Electronic Science and Technology of China (UESTC), Chengdu 610054, P. R. China‡Department of Biomedical Engineering and §Department of Materials Science and Engineering, City University of Hong Kong,Hong Kong, P. R. China

*S Supporting Information

ABSTRACT: Formation of the semiconductor/dielectricdouble-layered films via vertical phase separations frompolymer blends is an effective method to fabricate organicthin-film transistors (OTFTs). Here, we introduce a simpleone-step processing method for the vertical phase separationof poly(3-hexylthiophene-2,5-diyl) (P3HT) and poly(methylmethacrylate) (PMMA) blends in OTFTs and theirapplications for high-performance nitrogen dioxide (NO2)sensors. Compared to the conventional two-step coatedOTFT sensors, one-step processed devices exhibit a greatenhancement of the responsivity from 116 to 1481% for 30ppm NO2 concentration and a limit of detection of ∼0.7 ppb.Studies of the microstructures of the blend films and the electrical properties of the sensors reveal that the devices formed by theone-step vertical phase separation have better capability for the adsorption of NO2 molecules. Moreover, a careful adjustment ofthe blend ratio between P3HT and PMMA can further improve the performance of the NO2 sensors, ranging from sensitivity toselectivity and to the ability of recovery. This simple one-step processing method demonstrates a potential possibility fordeveloping high-performance, low-cost, and large-area OTFT gas sensors.

KEYWORDS: organic thin-film transistor (OTFT), gas sensor, vertical phase separation, polymer blend, nitrogen dioxide (NO2)

L. INTRODUCTION

With the development of the modern society, a large amountof harmful gases, such as nitrogen dioxide (NO2), hydrogensulfide (H2S), and carbon oxide (CO), have been emitted intothe environment as a result of the combustion of fuels.1 As oneof the major harmful gases, NO2 is hazardous to human healthsince it can cause respiratory problems and increase the risk ofemphysema and bronchitis, even at a low concentration ofppm levels. Exposure to high concentration of NO2 (>100ppm) can led to nose irritation, throat inflammation, and evendeath. In addition, NO2 is also a harmful gas for environment,as it can not only accelerate the formation of microparticles inthe air but also act as the precursor of the acid rain.2 Therefore,effectively monitoring/detecting NO2 by high-performance gassensors is extremely important. To date, various types of gassensors, including resistor type, capacitor type, and transistortype, have been developed for monitoring NO2. Among allthese kinds of sensors, organic thin-film transistor (OTFT)based gas sensors have drawn great attentions due to theirintrinsic advantages such as multiple choices of materials,flexible, and low cost.3−5 Moreover, the sensing signal can be

amplified by the gate input of the OTFTs, thus improving thesignal-to-noise ratio easily by the modulation of gate voltages.6

In the past decade, many works have been performed forachieving high-performance OTFT-based NO2 sensors. Animportant strategy is developing functional organic semi-conductors (OSCs) to improve the sensitivity, selectivity, andstability of the OTFT-based gas sensors.7,8 Another effectiveroute is modifying the microstructures of the OSC films toimprove the capability of the gas adsorption. For instance,introducing tunable nanopores OSC films in OTFT sensorsenables the detection of NH3 at ppb levels.9 Interfaceengineering between OSCs and dielectrics has also beenproved to be a successful route for achieving high-performanceOTFT-based sensors.5 One good method is to introducecrystalline para-sexiphenyl as an interfacial modifying layer forimproving the carrier mobility and concentration, thusimproving the gas response in OTFTs sensors.10 With the

Received: August 30, 2019Accepted: November 4, 2019Published: November 4, 2019

Research Article

www.acsami.orgCite This: ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

© 2019 American Chemical Society 44521 DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

Dow

nloa

ded

via

CIT

Y U

NIV

OF

HO

NG

KO

NG

on

Nov

embe

r 28

, 201

9 at

09:

16:5

7 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

aim of improving the sensitivity, this laboratory developedultrahigh-sensitivity OTFT-based NO2 sensors by the UV−ozone treatment of the polymer dielectrics, and the limit ofdetection (LOD) of ∼400 ppb can be achieved.11 However,these methods either require long duration of materialsynthesis or rely on unique equipment, resulting in the devicefabrication to be complicated and time consuming. Note, oneof the potential advantages in OTFT sensors is low cost, thussimple and effective processing routes are the key to thesuccess of OTFT sensors.Owing to the layer-by-layer layout in the OTFTs, semi-

conducting and dielectric layers typically need a two-stepdeposition. However, the previous mentioned interfaceengineering methods mostly focus on the interface betweenthe semiconducting layers and dielectrics. Recently, the reportsof OTFTs with semiconducting−insulating polymer blendshave shown advantages of great electrical performance, goodenvironmental stability, and remarkable mechanical proper-ties.12−17 In addition, the OSC−polymer blend films assemiconducting layers with great NH3 responsivity of 5 ppmhave also shown the potential application in high-performancegas sensors.18 Furthermore, vertical phase separation betweeninsulating polymers and OSCs can be obtained via carefullycontrolling the blend ratio and the solution preparationprocess. Therein, the one-step formation of both OSC anddielectrics from the polymer/OSC blend solution can beachieved. The resulting films from the vertical phase separationprocesses also exhibit layer-by-layer structures where the OSCsaggregate either at the top or bottom of the dielectric polymerlayer.19−21 Beyond simplifying the fabrication process, OTFTdevices processed by vertical phase separation also exhibitgood electrical performance and enhanced long-term stabilityin some cases owing to the unique interfacial propertiesbetween OSCs and dielectrics.22−24 Since the interfacialproperties play important roles in OTFT gas sensors, thequestion arises as to whether it is possible to use a simplevertical phase separation process to develop high-performanceOTFT-based NO2 sensors.In this work, we introduce a simple one-step spin-coating

method for the fabrication of OTFT-based NO2 sensors. Here,we choose two widely used materials poly(3-hexylthiophene-2,5-diyl) and poly(methyl methacrylate) (P3HT/PMMA) asOSC and dielectric, study the devices with different blendratios on the NO2 sensitivity. During the vertical phaseseparation, substrates with high surface energy typically lead tothe formation of PMMA at bottom.25,26 Thus, a top-contactOTFT structure is adopted and used as the architecture of thegas sensors. Systematic study of the transistor performance andsensing capabilities of the OTFTs with different P3HT/PMMA blend ratios as well as various PMMA molecularweights (MW) is performed. The optimized devices exhibitsuperior NO2 sensitivity, with responsivity of 1481% at 30 ppmand a LOD of ∼0.7 ppb. To the best of our knowledge, theLOD of 0.7 ppb is the record high sensitivity of OTFT-basedNO2 sensors; these devices are even comparable withcommercial metal-oxide sensors.

2. EXPERIMENTAL SECTION2.1. Material Preparation. P3HT (MW > 45 000, RR = 93%) was

purchased from Lumtec Corp., PMMA with different MW (MW = 15k,46k, 120k, and 350k) was purchased from Sigma-Aldrich and solvents(chlorobenzene and anisole) were purchased from Tokyo ChemicalIndustry Company. All the chemicals and solvent were used as

received. PMMA with MW = 120k was dissolved in anisole to form aconcentration of 10 wt % for a two-step spin-coating. For the blendfilms, P3HT and PMMA (MW = 120k) were mixed and dissolved inchlorobenzene accordingly to form a series of weight ratios rangingfrom 1:40 to 1:80, with the total concentration from 4.1 to 8.1 wt %.Here, the concentration of the P3HT in the mixed solution is fixed at1 mg/mL. Finally, the solution was stirred on a magnetic stirring plateovernight to ensure sufficient dissolution and mixing.

2.2. Device Fabrication. Figure 1a,b shows the schematicillustrations of the fabrication process and the device layout, and a

photograph of the devices is shown in Figure S1. The bilayerstructures were formed by vertical phase separation in one-step spin-coating. OTFT gas sensors based on bottom-gate, top-contactarchitectures use the self-stratified P3HT and PMMA as the OSClayer and the dielectric layer, respectively. Figure S2 shows themeasured water contact angles of the pure P3HT and PMMA as 97and 69°, respectively. However, the contact angles of the 1:40, 1:60,and 1:80 P3HT/PMMA blend films were 96, 96, and 95° respectively(Figure S2), indicating that the formation of the layer-by-layerstructure with hydrophobic P3HT on top.27 The fabrication of theseOTFT sensors began on indium tin oxide (ITO) coated glasssubstrates. First, the substrates were cleaned in acetone, deionizedwater, and isopropyl alcohol sequentially (15 min each in anultrasonic bath). Ten minutes of ultraviolet−ozone exposure wasperformed on these substrates after drying in an oven at 80 °C for 1 h.The two-step spin-coated OTFTs are referred as the blend ratioP3HT/PMMA = 1:0, where 400 nm thick PMMA was first spun-coated (1500 rpm for 1 min and baked at 90 °C for 2 h), and P3HTwas spun-coated on top of the PMMA at 1500 rpm for 30 s and thenannealed at 100 °C for 30 min. Here, all the coating processes wereconducted in a glovebox. To study if any damage of PMMA may havehappened during the P3HT spin-coating process, atomic forcemicroscopy (AFM) measurement were conducted for the PMMAfilms before and after the in situ spin-coating of organic solvents (onthe fly dispensing). Figure S3 shows the AFM images of the PMMAfilms after spin-coating chlorobenzene, whose surface is still verysmooth with RMS of 0.313−0.440 nm. These results indicate the insitu spin-coating process of P3HT does not cause any serious damageand obvious dissolution of the beneath PMMA films.28 For those one-step spin-coated OTFTs, solutions with various blend ratios (P3HT/PMMA = 1:40, 1:60, 1:80), were spun-coated at 1500 rpm for 30 s ina nitrogen gas atmosphere. Then, the as-prepared films were annealedat 60 °C for 1 h to remove residual solvents, and the thickness of the

Figure 1. (a) Schematic illustrations of the fabrication process for theOTFT-based NO2 sensors, showing layer-by-layer structures formedby vertical phase separation. (b) Schematic illustration of the workingprocess of the OTFT gas sensors. (c) Transfer curves (VDS = −40 V)of the OTFTs with different blend ratios between P3HT and PMMA.IDS is the drain−source current, IGS is the gate leakage current, andVGS is the gate−source voltage.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44522

blend films with ratios of 1:40, 1:60, and 1:80 are 180, 300, and 440nm, respectively (Figure S4). Finally, source and drain electrodes of40 nm gold were deposited on top of the blend film via thermalevaporation at an evaporation rate of 0.3 Å/s under 3 × 10−3 Pa,where the patterns (channel length = 100 μm, channel width = 10mm) were formed by using a shadow mask.2.3. Measurement and Characterization. An airtight chamber

(∼0.02 L) acting as the testing environment was used for the sensorperformance evaluation, where the OTFT sensors were putted insidethe chamber. Dry air and NO2 were first mixed together inappropriate concentrations. Then, the gas was introduced into thetesting chamber by a mass flow controller at a fixed flow rate of 100standard cm3/min. The electrical properties of the OTFT sensorswere characterized by a semiconductor analyzer (Keithley 4200). Themobility μ of the OTFTs was calculated in the saturation region by eq1

IWC

LV V

2( )DS

iG th

2ikjjj

y{zzzμ= −

(1)

where IDS is the drain−source current and L and W are the channellength and width, respectively. VG is the gate voltage and Vth is thethreshold gate voltage. Ci is the capacitance per unit area of thedielectric layer, which can be calculated by eq 2

C kd0

iε=

(2)

where ε0 is the vacuum permittivity, k is the dielectric constant of thedielectric, and d is the thickness of the gate dielectric.The surface morphologies and the thickness of the bilayer films

were characterized by an atomic force microscope (MFP-3D-BIO,Asylum Research) in a tapping mode and a scanning electronicmicroscope (FEI Inspect F50). The phase separation of blend filmswas evaluated by transmission electron microscopy (TEM, FEI/Philips Tecnai 12 BioTWIN) at an accelerating voltage of 200 kV.The inner molecular packing structures of the blend films werecharacterized by an UV−vis spectrophotometer (SHIMAZU UV-1700) and a Fourier-transform infrared (FT-IR) spectrometer(Thermo Scientific, Nicole-10, Waltham, MA).

3. RESULTS AND DISCUSSIONFigures 1c and S5 show the representative transfer and outputplots of the OTFTs with different P3HT/PMMA blend ratios,where the MW for PMMA is 120k. The electrical properties ofthe OTFTs with different PMMA ratios exhibit obviousvariation, especially the on-current (Ion, where VDS = VGS =−40 V). The Ion is ∼4.65 μA for the reference device (two-stepspin-coated, 1:0), ∼6.49 μA for P3HT/PMMA = 1:40, ∼2.21μA for P3HT/PMMA = 1:60, and ∼1.07 μA for P3HT/PMMA = 1:80. When VDS = VGS = −40 V, the leakage currentfor all devices are at the level of 10−6 A, which is 1 order ofmagnitude smaller than their IDS. Other parameters, includingthe threshold voltage (Vth), subthreshold slope (SS), andcurrent-on/off ratio, of these devices are summarized in TableS1.To evaluate the performance of these sensor devices, OTFTs

were exposed and measured under various concentrations ofNO2, ranging from 0 to 30 ppm. Every cycle of themeasurement was performed after introducing NO2 into thetesting chamber for 2 min, then dry air purging was conductedbefore the next cycle test, as this step allowed the completerecovery of the electrical performance for the sensors.Representative transfer curves (VDS = − 40 V) recorded atdifferent concentrations are shown in Figure 2a,b. Comparedto the 1:0 devices, the 1:40, 1:60, and 1:80 devices exhibitsignificant improvements in sensitivity; devices with P3HT/PMMA of 1:60 show the best sensor performance.

The sensor performance is typically referred to asresponsivity associated with the changes of the key parametersincluding Ion, mobility (μ), and Vth, which are summarized inFigures 2c−e and S6. The responsivity of those parameters isdefined as R = |(YNO2 − Y0)/Y0| × 100%, where YNO2 and Y0are the on current of those parameters in a certainconcentration of NO2 and dry air, respectively. From thedata, we can observe that the responsivity of all the devicesincreases with the increase of NO2 concentration. Interestingly,the one-step-processed self-stratified devices exhibit improvedNO2 sensing performance than those pure 1:0 OTFTs. Forinstance, the responsivity of the 1:0 device is 116% for Ion and35% for μ at 30 ppm NO2, whereas the 1:40, 1:60, and 1:80devices exhibit a significant change of 253, 1481, and 1022%for IDS and 117, 418, and 294% for μ. The NO2 responsivity ofthe best sensors (1:60) is enhanced 12× for Ion and 11× for μcompared with that of the referenced 1:0 device. Owing to theultrasensitive properties, the best sensors (1:60) can effectivelydetect low concentration of NO2, with a responsivity of 487%at 0.5 ppm NO2, while the responsivity of referenced 1:0devices is only 22% at the same NO2 concentration. Theseresults indicate the potential of these sensors for ultralow NO2concentration detection. Vth in OTFT sensors is an importantparameter, which is typically used to evaluate the density ofcharge carriers trapped at the interface between the organicsemiconductors and the dielectrics.29 The trap densitydecreases when NO2 diffuses into the P3HT film, thus causinga decrease of the gate turn-on voltage and a positive shift of theVth.

30 The Vth of OTFTs fabricated by the one-step spin-coating increase 298% (1:40), 486% (1:60), and 364% (1:80)after exposure to 30 ppm NO2. However, the referenced 1:0

Figure 2. Representative transfer plots of the OTFTs with indicatedP3HT/PMMA (PMMA MW = 120k) ratios under different NO2concentrations. (a) Log plots of the drain−source current versus thegate voltage. (b) Plots of the square root of the drain−source currentversus the gate voltage. Variation of Ion (c), μ (d), and Vth (e) of theOTFT sensors with different P3HT/PMMA blend ratios at differentNO2 concentrations.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44523

device only shows an increase of 235% at the same NO2concentration.To understand the reason of the sensing performance

improvement in those one-step-processed sensors. Filmmicrostructures of the blend films were next studied by theUV−vis absorbance spectra (Figure 3a), all absorption spectraof the films with different blend ratios show a peak centered at555 nm, which originated from the broad π−π* electronictransition band. Meanwhile, two additional shoulder peaks at525 and 605 nm can be also observed, which are associatedwith the coupling of the CC double bonds causedsymmetrical stretching and the π−π* electronic transition.24

Furthermore, the intensity of the peak gradually increases asthe proportion of PMMA increases. According to the H-aggregation model,31 the ratio of the absorbance peaksbetween 0-0 and 0-1 (I0‑0/I0‑1) was related to the intra-molecular ordering and conjugation length of P3HT. As shownin Table S2, the increase of the I0‑0/I0‑1 ratio indicates theimprovement of crystallinity, conjugation length, and chainordering in the PMMA-blended P3HT films. More concen-trated solution typically has a higher viscosity and thus couldform a thicker and wetter film by spin-coating, in whichabundant solvent prolongs the solidification time.25,32,33 Filmsformed from solution with a higher blend ratio provide alonger time for chain self-organization in P3HT, whichfacilitates the crystalline structure formation. FT-IR spectrain Figure 3b show that there is no change in the functionalgroups in both blend films and the pure P3HT film. Therefore,the enhancement of the sensing performance is not due to thechemistry-induced microstructural changes. In addition, thethickness of OSC is a critical reason that affects the sensingperformance of devices. Here, selective etching techniqueassociated with transferring P3HT thin film from the blendfilm was conducted to characterize the thickness of P3HT inblend films25,26 (Figure S7). From the AFM cross-sectionalviews, we can observe that P3HT in all blend films is

continuous with thickness of 12.4 nm for the 1:40 film, 9.2 nmfor the 1:60 film, and 5.8 nm for the 1:80 film. The decrease ofthe film thickness was due to the reduction of the P3HTcontent in the films with the increase of the totalconcentration.25,34

We next investigate the morphologies of the pure P3HT filmand the self-stratified P3HT films with different PMMA blendratios. Figures 3c and S8 show the AFM and scanning electronmicroscopy (SEM) images of these P3HT films, respectively; itcan be observed that the pure P3HT film reveals a continuousfilm feature with RMS of 0.588 nm, while the P3HT films inthose self-stratified ones show rugged island-like features andan increase of the roughness with 0.502 nm for the 1:40 film,0.591 nm for the 1:60 film, and 0.675 nm for the 1:80 film.The islandlike morphology in those self-stratified filmsassociated with the kinetic interactions between the solutionmix and the stratification during the solution coating process.The glass transition temperature (Tg) of the blends increases asthe solvent evaporates until room temperature. Then, themotions of polymer chain stop and the final polymer structureis formed. Therefore, the motions of the polymer chain in thePMMA-rich phase also stop and thus form a capping layer.Due to the triple-phase boundaries among P3HT, PMMA, andair, an islandlike surface morphology was formed in the finalfilm structure.35−37 Interestingly, the feature sizes and islanddensities are quite different for those films with different blendratios, where the 1:60 blend film exhibits the maximum densityand minimum feature size (domain diameter of 0.25−0.50μm). These results well-explain the best sensing performancefor those OTFTs with the 1:60 blend ratio, since a higherdomain density and a smaller feature size of P3HT could resultin more interfacial “gaps”. It is well-known that the physicaladsorption of the gas molecules in OSCs films highly dependson the micro/nano-gaps, such as grain boundaries.38 So, wecan summarize the NO2-sensing mechanism of those one-step-processed P3HT/PMMA blend OTFT sensors in the

Figure 3. (a) Normalized UV−vis absorption spectra, (b) FT-IR spectroscopy plots, and (c) AFM images of the films with different P3HT/PMMA(PMMA MW = 120k) blend ratios. (d) Schematic illustration of the interaction of NO2 with sensors from different processes.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44524

following two parts: (1) NO2 diffuses into the active layer andcouples with P3HT, thus forming linkage-type structures; as aresult, hole charge carries are induced in the OSC film andcause Ion to increase;39 (2) NO2 is physically absorbed at themicro/nano-gaps (interfaces between P3HT clusters andPMMA), thus donating hole charge carriers and resulting inthe increase of Ion and μ.40 As illustrated in Figure 3d, inreference devices prepared by two-step spin-coating, part (1)plays the main role of the sensing mechanism, while part (2)contributes very little in this system because the P3HT filmscontinues to exhibit feature with few micro/nano-gaps. Incontrast, those vertical phase separation induced one-step-processed self-stratified P3HT/PMMA films exhibit a largeamount of micro/nano-gaps between P3HT and PMMA,which provides a lot of spaces for NO2 adsorption. As a result,both parts (1) and (2) play significant roles in the sensingmechanism in those one-step-processed OTFT sensors.Optimized self-stratified films with proper blend ratios canleverage the sensitivity of the OTFT and result in a 12×increase in responsivity compared to that of the conventionaldevices.To characterize the repeatability of multiple-cycle measure-

ment, the devices were placed in an airtight dry-air-filledtesting chamber and gate and source/drain bias was applied(VDS = VGS = −40 V) until the output current became stable.Then, various concentrations of NO2, ranging from 0.5 to 30ppm, were introduced into the testing chamber for measuringthe real-time response. The best one-step-processed sensorswith a 1:60 blend ratio and reference devices (1:0) werestudied and compared. As shown in Figure 4a, the sensing

performance of the 1:60 device is greater than that of thereference device. The responsivity of the 1:60 device at 30 ppmNO2 is 1358%, 12× that of the 1:0 device (102%). Here, theincrease of Ion with increasing time of NO2 exposure is due tothe continuous adsorption of NO2 in OSC before saturation.41

Since NO2 is a typical oxidizing gas, some effective chargecarriers are produced in p-type OSCs during NO2 adsorption.Therefore, the carrier concentration increases and causes anenlargement in Ion and μ.30 Note, the responsivity of the 1:60device at low concentration of 0.5 ppm NO2 is still as high as446%, while the responsivity of the 1:0 device is only 28%. To

estimate the LOD of the best one-step-processed sensors viathe linear fitting method, we further measured the 1:60 deviceat lower NO2 concentrations (ranging from 10 to 50 ppb,exposure 5 min for each cycle and 15 min purge for recovery).As shown in Figure 4b, the responsivity to 10, 20, 30, 40, and50 ppb are 96, 210, 297, 360, and 419%, respectively. The real-time responsivities of the sensors were also tested and areshown in Figure 4c. Based on the responsivities at these lowconcentrations in Figure 4b, the estimated LOD for thesesensors can be estimated by eq 3,11,42

Y

a b a b b b a ab b

( )( )

LOD

2 2 2 2 2 2 2 2

2 2 2

η ηη

=

− × + × − − × Δ − × Δ− × Δ

(3)

where a, b, Δa, and Δb are the parameters extracted from thelinear fit of the plot in Figure 4d, and η = 1.96 corresponds to a95% confidence level. The estimated LOD of ∼0.7 ppb for the1:60 device (Figure S9), to the best of our knowledge, is thehighest sensitivity of OTFT-based NO2 sensors in literatures.

43

To further understand this one-step processing route, westudied the sensing properties of the champion device with a1:60 ratio using PMMAs with different MW of 15k, 46k, 120k,and 350k. Figures 5a and S10 show the responsivity of the one-

step-processed devices using PMMA with various MW for 30ppm NO2, where we can find that the responsivity increaseswith the increase of the PMMA MW in the range of 15k to120k, and the resulting responsivity is 254% for 15k, 733% for46k, and 1481% for 120k. While further increase in MW causesa decrease of the responsivity, particularly 340% for 350k. Thequantity and density of the micro/nano-gaps increase in theseblend films with increasing MW of PMMA and reach themaximum at MW of 120k; however, further increase of MW to350k decreases the quantity and density of the gaps (FigureS11). These results further prove the conclusion that theadsorption of NO2 by micro/nano-gaps improves the sensingproperties. TEM was conducted for the 1:60 blend films withPMMAs of different MW to evaluate the phase separation(Figure S12). The dark areas with crystalline features

Figure 4. (a) Percentage change of Ion to multiple cycles of NO2introduction at different concentrations. (b) Typical transfer curves ofthe 1:60 devices with different concentrations of NO2 ranging from 10to 50 ppb. (c) Real-time sensitivities of the 1:60 devices in differentlow concentrations of NO2. (d) Calculation of limit of detection.

Figure 5. (a) Effect of MW of PMMA on the percentage of the outputcurrent changes upon exposure to 30 ppm NO2. (b) Percentage of theoutput current changes of the 1:60 device with various gases. (c)Transfer curves of the 1:60 device (PMMA of MW = 120k) whentested under different environments.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44525

correspond to the P3HT-rich domains since the density ofP3HT is higher than that of PMMA, whereas the light areas areassociated with amorphous PMMA domains.25,44 It can beseen that the blend films with PMMA with MW = 120k exhibitmost homogeneous features and afford most micro/nano-gaps.Either the too low or too high MW of PMMA could causediscontinuity of P3HT, which could either lower the carriermobility and eventually lead to the attenuation of the sensingperformance. The tomography shown here is highly relevantwith the sensing properties and the OTFT electrical perform-ance.Selectivity is another important parameter for practical

sensing applications. To probe the selectivity of these one-step-processed sensors, the responses of the devices with P3HT/PMMA = 1:60 to various common harmful gases wereinvestigated, including an oxidizing gas of SO2 and reducinggases of NH3, H2S, and CO (all gases were mixed with dry airat 30 ppm). As shown in Figures 5b and S13, the response ofthe sensors to NO2 is much greater than that to other fourgases, with a superior 1481% responsivity of NO2 but muchweaker responsivities of 112% for SO2, 25% for NH3, 53% forH2S, and 13% for CO.Finally, we evaluated the possibility of using these one-step-

processed sensors multiple times, since reuse is a critical issueand relevant to the life time. As shown in Figure 5c, when the1:60 device is exposed to 30 ppm NO2 for 10 min, Ion increases8× compared to that of the device before NO2 exposure. TheI−V characteristics of the devices could gradually recover totheir original status after exposure to NO2 was stopped andstored in air for 3 days. Then, the devices further tested in ∼3× 10−3 Pa vacuum still show the same value of the draincurrent to the devices stored in air for 3 days. Such resultssuggest that the NO2 sensing mechanism is more related to thephysical adsorption process but not to a chemical reaction.11

The characteristics of the sensors can recover to their initialstate naturally or by a vacuum method to accelerate thisprocess. So, these NO2 sensors could be reused.

4. CONCLUSIONS

In summary, we developed a simple and effective one-stepspin-coating route for realizing high-performance gas sensors.The one-step-processed sensors using P3HT/PMMA mixedsolution formed a blend film with optimized ratios exhibitgood sensitivity to NO2 with a LOD of 0.7 ppb. Maximizationof the micro/nano-gaps between P3HT and PMMA by verticalphase separation process leverages the sensitivity of NO2. Thisone-step route provides a fundamental understanding of thesensing mechanism and the advantages of high-throughputmethods. Therefore, we believe this one-step coating route willbring possibilities of rapid fabricatication of high-performanceand low-cost gas sensors.

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsami.9b15651.

AFM images; TEM images; cross-sectional SEM images;output curves; parameters of different OTFTs; param-eters variation; limit of detection calculation; represen-tative transfer curves; drain current changes (PDF)

■ AUTHOR INFORMATIONCorresponding Author*E-mail: xingeyu@cityu.edu.hk.ORCIDJunsheng Yu: 0000-0002-7484-8114Feng Wang: 0000-0001-9471-4386Xinge Yu: 0000-0003-0522-1171NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was financially supported by the National Key R&DProgram of China (Grant No. 2018YFB0407102), theFoundation of National Natural Science Foundation ofChina (NSFC) (Grant Nos. 61421002, 61675041, and51703019), the Project of Science and Technology of SichuanProvince (Grant Nos. 2019YFH0005, 2019YFG0121, and2019YJ0178), City University of Hong Kong (Grant No.9610423), and the Research Grants Council of Hong Kong(CityU 11204717). This work was also sponsored by SichuanProvince Key Laboratory of Display Science and Technology.

■ REFERENCES(1) Nejat, P.; Jomehzadeh, F.; Taheri, M. M.; Gohari, M.; Abd.Majid, M. Z. A Global Review of Energy Consumption, CO2Emissions and Policy in the Residential Sector (with an Overviewof the Top Ten CO2 Emitting Countries). Renewable SustainableEnergy Rev. 2015, 43, 843−862.(2) Hamra, G. B.; Laden, F.; Cohen, A. J.; Raaschou-Nielsen, O.;Brauer, M.; Loomis, D. Lung Cancer and Exposure to NitrogenDioxide and Traffic: a Systematic Review and Meta-Analysis. Environ.Health Perspect. 2015, 123, 1107−1112.(3) Zhang, J.; Liu, X.; Neri, G.; Pinna, N. Nanostructured Materialsfor Room-Temperature Gas Sensors. Adv. Mater. 2016, 28, 795−831.(4) Zhuang, X.; Zhang, D.; Wang, X.; Yu, X.; Yu, J. Biocompatibleand Degradable Gelatin Dielectric Based Low-Operating VoltageOrganic Transistors for Ultra-High Sensitivity NH3 Detection. Appl.Phys. Lett. 2018, 113, No. 263301.(5) Wu, X.; Mao, S.; Chen, J.; Huang, J. Strategies for Improving thePerformance of Sensors Based on Organic Field-Effect Transistors.Adv. Mater. 2018, 30, No. 1705642.(6) Li, H.; Shi, W.; Song, J.; Jang, H. J.; Dailey, J.; Yu, J.; Katz, H. E.Chemical and Biomolecule Sensing with Organic Field-EffectTransistors. Chem. Rev. 2018, 119, 3−35.(7) Yang, Y.; Zhang, G.; Luo, H.; Yao, J.; Liu, Z.; Zhang, D. HighlySensitive Thin-Film Field-Effect Transistor Sensor for Ammonia withthe DPP-Bithiophene Conjugated Polymer Entailing ThermallyCleavable Tert-Butoxy Groups in the Side Chains. ACS Appl. Mater.Interfaces 2016, 8, 3635−3643.(8) Wang, B.; Huynh, T.; Wu, W.; Hayek, N.; Do, T. T.; Cancilla, J.C.; Torrecilla, J. S.; Nahid, M. M.; Colwell, J. M.; Gazit, O. M.;Puniredd, S. R.; McNeill, C. R.; Sonar, P.; Haick, H. A HighlySensitive Diketopyrrolopyrrole-Based Ambipolar Transistor forSelective Detection and Discrimination of Xylene Isomers. Adv.Mater. 2016, 28, 4012−4018.(9) Zhang, F.; Qu, G.; Mohammadi, E.; Mei, J.; Diao, Y. Solution-Processed Nanoporous Organic Semiconductor Thin Films: towardHealth and Environmental Monitoring of Volatile Markers. Adv.Funct. Mater. 2017, 27, No. 1701117.(10) Wang, Z.; Huang, L.; Zhu, X.; Zhou, X.; Chi, L. AnUltrasensitive Organic Semiconductor NO2 Sensor Based onCrystalline TIPS-Pentacene Films. Adv. Mater. 2017, 29,No. 1703192.(11) Huang, W.; Zhuang, X.; Melkonyan, F. S.; Wang, B.; Zeng, L.;Wang, G.; Han, S.; Bedzyk, M. J.; Yu, J.; Marks, T. J.; Facchetti, A.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44526

Uv-Ozone Interfacial Modification in Organic Transistors for High-Sensitivity NO2 Detection. Adv. Mater. 2017, 29, No. 1701706.(12) Jiang, C.; Choi, H. W.; Cheng, X.; Ma, H.; Hasko, D.; Nathan,A. Printed Subthreshold Organic Transistors Operating at High Gainand Ultralow Power. Science 2019, 363, 719−723.(13) Gumyusenge, A.; Tran, D. T.; Luo, X.; Pitch, G. M.; Zhao, Y.;Jenkins, K. A.; Dunn, T. J.; Ayzner, A. L.; Savoie, B. M.; Mei, J.Semiconducting Polymer Blends That Exhibit Stable ChargeTransport at High Temperatures. Science 2018, 362, 1131−1134.(14) Jia, X.; Fuentes-Hernandez, C.; Wang, C. Y.; Park, Y.; Kippelen,B. Stable Organic Thin-Film Transistors. Sci. Adv. 2018, 4,No. eaao1705.(15) Yu, X.; Zeng, L.; Zhou, N.; Guo, P.; Shi, F.; Buchholz, D. B.;Ma, Q.; Yu, J.; Dravid, V. P.; Chang, R. P.; Bedzyk, M.; Marks, T. J.;Facchetti, A. Ultra-Flexible, “Invisible” Thin-Film Transistors Enabledby Amorphous Metal Oxide/Polymer Channel Layer Blends. Adv.Mater. 2015, 27, 2390−2399.(16) Zhang, X.; Wang, B.; Huang, W.; Wang, G.; Zhu, W.; Wang, Z.;Zhang, W.; Facchetti, A.; Marks, T. J. Oxide-Polymer HeterojunctionDiodes with a Nanoscopic Phase-Separated Insulating Layer. NanoLett. 2019, 19, 471−476.(17) Wang, G.; Swick, S. M.; Matta, M.; Mukherjee, S.; Strzalka, J.W.; Logsdon, J. L.; Fabiano, S.; Huang, W.; Aldrich, T. J.; Yang, T.;Timalsina, A.; Powers-Riggs, N.; Alzola, J. M.; Young, R. M.;DeLongchamp, D. M.; Wasielewski, M. R.; Kohlstedt, K. L.; Schatz,G. C.; Melkonyan, F. S.; Facchetti, A.; Marks, T. J. Photovoltaic BlendMicrostructure for High Efficiency Post-Fullerene Solar Cells. To Tiltor Not To Tilt? J. Am. Chem. Soc. 2019, 141, 13410−13420.(18) Han, S.; Zhuang, X.; Shi, W.; Yang, X.; Li, L.; Yu, J. Poly(3-Hexylthiophene)/Polystyrene (P3HT/PS) Blends Based OrganicField-Effect Transistor Ammonia Gas Sensor. Sens. Actuators, B2016, 225, 10−15.(19) Lee, J. H.; Lee, Y. H.; Ha, Y. H.; Kwon, J.; Pyo, S.; Kim, Y.-H.;Lee, W. H. Semiconducting/Insulating Polymer Blends with DualPhase Separation for Organic Field-Effect Transistors. RSC Adv. 2017,7, 7526−7530.(20) Kimura, Y.; Nagase, T.; Kobayashi, T.; Hamaguchi, A.; Ikeda,Y.; Shiro, T.; Takimiya, K.; Naito, H. Soluble Organic SemiconductorPrecursor with Specific Phase Separation for High-PerformancePrinted Organic Transistors. Adv. Mater. 2015, 27, 727−732.(21) Ge, F.; Liu, Z.; Lee, S. B.; Wang, X.; Zhang, G.; Lu, H.; Cho,K.; Qiu, L. Bar-Coated Ultrathin Semiconductors from Polymer Blendfor One-Step Organic Field-Effect Transistors. ACS Appl. Mater.Interfaces 2018, 10, 21510−21517.(22) Chu, P. H.; Wang, G.; Fu, B.; Choi, D.; Park, J. O.;Srinivasarao, M.; Reichmanis, E. Synergistic Effect of Regioregular andRegiorandom Poly(3-Hexylthiophene) Blends for High PerformanceFlexible Organic Field Effect Transistors. Adv. Electron. Mater. 2016,2, No. 1500384.(23) Kwak, D.; Choi, H. H.; Kang, B.; Kim, D. H.; Lee, W. H.; Cho,K. Tailoring Morphology and Structure of Inkjet-Printed Liquid-Crystalline Semiconductor/Insulating Polymer Blends for High-Stability Organic Transistors. Adv. Funct. Mater. 2016, 26, 3003−3011.(24) Kang, B.; Ge, F.; Qiu, L.; Cho, K. Effective Use of ElectricallyInsulating Units in Organic Semiconductor Thin Films for High-Performance Organic Transistors. Adv. Electron. Mater. 2017, 3,No. 1600240.(25) Wang, X.; Lee, W. H.; Zhang, G.; Wang, X.; Kang, B.; Lu, H.;Qiu, L.; Cho, K. Self-Stratified Semiconductor/Dielectric PolymerBlends: Vertical Phase Separation for Facile Fabrication of OrganicTransistors. J. Mater. Chem. C 2013, 1, 3989−3998.(26) Ge, F.; Wei, S.; Liu, Z.; Wang, G.; Wang, X.; Zhang, G.; Lu, H.;Cho, K.; Qiu, L. Tailoring Structure and Field-Effect Characteristicsof Ultrathin Conjugated Polymer Films via Phase Separation. ACSAppl. Mater. Interfaces 2018, 10, 9602−9611.(27) Wang, X.; Liu, Z.; Wei, S.; Ge, F.; Liu, L.; Zhang, G.; Ding, Y.;Qiu, L. Ultrathin Semiconductor Films for NH3 Gas Sensors Preparedby Vertical Phase Separation. Synth. Met. 2018, 244, 20−26.

(28) Shao, B.; Liu, Y.; Zhuang, X.; Hou, S.; Han, S.; Yu, X.; Yu, J.Crystallinity and Grain Boundaries Control of TIPS-Pentacene inOrganic Thin-Film Transistors for Ultra-High Sensitivity NO2Detection. J. Mater. Chem. C 2019, 7, 10196−10202.(29) Yu, X.; Marks, T. J.; Facchetti, A. Metal Oxides forOptoelectronic Applications. Nat. Mater. 2016, 15, 383−396.(30) Zhuang, X.; Han, S.; Huai, B.; Shi, W.; Yu, J. Sub-Ppm andHigh Response Organic Thin-Film Transistor NO2 Sensor Based onNanofibrillar Structured TIPS-Pentacene. Sens. Actuators, B 2019,279, 238−244.(31) Spano, F. C.; Silva, C. H- And J-Aggregate Behavior inPolymeric Semiconductors. Annu. Rev. Phys. Chem. 2014, 65, 477−500.(32) Lu, G.; Blakesley, J.; Himmelberger, S.; Pingel, P.; Frisch, J.;Lieberwirth, I.; Salzmann, I.; Oehzelt, M.; Di Pietro, R.; Salleo, A.;Koch, N.; Neher, D. Moderate Doping Leads to High Performance ofSemiconductor/Insulator Polymer Blend Transistors. Nat. Commun.2013, 4, No. 1588.(33) Noriega, R.; Rivnay, J.; Vandewal, K.; Koch, F. P.; Stingelin, N.;Smith, P.; Toney, M. F.; Salleo, A. A General Relationship betweenDisorder, Aggregation and Charge Transport in ConjugatedPolymers. Nat. Mater. 2013, 12, 1038−1044.(34) Rinehart, S. J.; Yuan, G.; Dadmun, M. D. Elucidating theKinetic and Thermodynamic Driving Forces in Polymer Blend FilmSelf-Stratification. Macromolecules 2018, 51, 7836−7844.(35) Han, S.; Yang, Z.; Li, Z.; Zhuang, X.; Akinwande, D.; Yu, J.Improved Room Temperature NO2 Sensing Performance of OrganicField-Effect Transistor by Directly Blending a Hole-Transporting/Electron-Blocking Polymer into the Active Layer. ACS Appl. Mater.Interfaces 2018, 10, 38280−38286.(36) Kim, M.; Lee, J.; Jo, S. B.; Sin, D. H.; Ko, H.; Lee, H.; Lee, S.G.; Cho, K. Critical Factors Governing Vertical Phase Separation inPolymer-Pcbm Blend Films for Organic Solar Cells. J. Mater. Chem. A2016, 4, 15522−15535.(37) Kim, J. S.; Ho, P. K. H.; Murphy, C. E.; Friend, R. H. PhaseSeparation in Polyfluorene-Based Conjugated Polymer Blends: Lateraland Vertical Analysis of Blend Spin-Cast Thin Films. Macromolecules2004, 37, 2861−2871.(38) Seo, Y.; Lee, J. H.; Anthony, J. E.; Nguyen, K. V.; Kim, Y. H.;Jang, H. W.; Ko, S.; Cho, Y.; Lee, W. H. Effects of Grain BoundaryDensity on the Gas Sensing Properties of Triethylsilylethynyl-Anthradithiophene Field-Effect Transistors. Adv. Mater. Interfaces2017, 5, No. 1701399.(39) Yu, X.; Zhou, N.; Han, S.; Lin, H.; Buchholz, D. B.; Yu, J.;Chang, R. P. H.; Marks, T. J.; Facchetti, A. Flexible Spray-CoatedTIPS-Pentacene Organic Thin-Film Transistors as Ammonia GasSensors. J. Mater. Chem. C 2013, 1, 6532−6535.(40) Lee, J. H.; Seo, Y.; Park, Y. D.; Anthony, J. E.; Kwak, D. H.;Lim, J. A.; Ko, S.; Jang, H. W.; Cho, K.; Lee, W. H. Effect ofCrystallization Modes in TIPS-Pentacene/Insulating Polymer Blendson the Gas Sensing Properties of Organic Field-Effect Transistors. Sci.Rep. 2019, 9, No. 21.(41) Song, Z.; Tong, Y.; Zhao, X.; Ren, H.; Tang, Q.; Liu, Y. AFlexible Conformable Artificial Organ-Damage Memory Systemtowards Hazardous Gas Leakage Based on a Single OrganicTransistor. Mater. Horiz. 2019, 6, 717−726.(42) Zheng, H.; Ramalingam, B.; Korampally, V.; Gangopadhyay, S.Large Sensitivity Enhancement in Semiconducting Organic FieldEffect Transistor Sensors through Incorporation of Ultra-FinePlatinum Nanoparticles. Appl. Phys. Lett. 2013, 103, No. 193305.(43) Das, A.; Dost, R.; Richardson, T.; Grell, M.; Morrison, J. J.;Turner, M. L. A Nitrogen Dioxide Sensor Based on an OrganicTransistor Constructed from Amorphous Semiconducting Polymers.Adv. Mater. 2007, 19, 4018−4023.(44) Chen, J.; Chen, Z.; Qu, Y.; Lu, G.; Ye, F.; Wang, S.; Lv, H.;Yang, X. Large Interfacial Area Enhances Electrical Conductivity ofPoly(3-Hexylthiophene)/Insulating Polymer Blends. RSC Adv. 2015,5, 1777−1784.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.9b15651ACS Appl. Mater. Interfaces 2019, 11, 44521−44527

44527