Inkjet-printed thin-film transistors using surfactant-based transition-metal dichalcogenide nanocomposites suspended in polymeric semiconductors

Ink formulations containing a suspension of single-crystalline molybdenum disulfide (MoS2) nanosheets suspended in the polymeric semiconductor poly(3-hexylthiophene-2,5-diyl) (P3HT) were inkjet printed for the fabrication of thin-film transistors (TFT). The MoS2 nanosheets were treated with the surfactant trichloro(dodecyl)silane (DDTS) to functionalize the MoS2 surface and created a more stable suspension, reducing the agglomeration of MoS2 suspended in the P3HT solution. This ink formulation was inkjet printed onto the surface of thermal oxide coated, p+-Si wafers to form common-gate TFT device structures. The printed semiconductor formed the active region of a hybrid MoS2 suspension in P3HT of the TFTs. The field-effect mobility for the hybrid-ink TFTs was found to be three times (3×) higher compared to reference devices using pristine P3HT without the suspension. The functionalized MoS2 suspension was also found to form thinner nanosheet suspensions within the P3HT matrix that resulted in approximately 60% higher field-effect mobility compared to hybrid inks without the surfactant. The enhancement of the electrical properties of the TFTs was determined to be due to a structural change in the thin-film semiconductor. The observed current–voltage (I–V) changes were correlated to measurable structural alterations in the semiconductor thin film characterized by x-ray diffraction, atomic force microscopy, and UV–visible absorption spectroscopy.


Introduction
Inkjet printing technology has received great attention for the fabrication of flexible and wearable electronics, as well as devices for the emerging Internet of Things applications [1][2][3][4][5].The technology enables low-cost processing where drop-on-demand (DOD) techniques provide precise non-contact additive processes that reduce waste in the integration of disparate materials.Solution-processed polymeric semiconductors are one of the best candidates for inkjetprinted thin-film transistors (TFTs).These inks enable thin-film electronics offering low cost, solution processibility, and high mechanical durability compared to inorganic semiconductors [6].However, a major trade-off to this printed microfabrication approach is the degraded electrical performance of the printed device compared to conventional electronic materials.Printed TFTs also have electrical characteristics that are typically poorer compared to spin-coated deposition methods.The centrifugal force created by spin coating results in highly aligned crystalline layers when the film forms.This effect is difficult to achieve by DOD printing methods due to the absence of centrifugal forces in printed droplets [7][8][9] (see figure S1 in the supplementary information).As a result, the high crystalline property in printed polymeric semiconductors is difficult to achieve, hampering the carrier transport properties of TFT devices.
To address these deficiencies for inkjet-printed polymers, formulating organic/inorganic hybrid semiconducting inks may be an effective approach to enhance the performance of inkjet-printed TFTs.
A suspension of single-crystalline nanosheets mixed within the polymeric solution can create electrical pathways for enhancing carrier transport as well as provide surfaces for the polymeric solution to form a more ordered film.Several organic-inorganic hybrid nanocomposite materials have been suggested and reported by mixing polymeric and inorganic semiconductors, such as graphene or two-dimensional (2D) materials [10][11][12][13], TiO 2 nanorods [14], or carbon nanotubes [15] that have higher structural and carrier transport properties compared to the polymeric semiconductor alone.
Previous investigations have shown that deposited MoS 2 layers (from bulk flakes) formed from a drop-cast suspension showed very limited electrical conductivity between the interflakes of the MoS 2 [16].This degraded charge transport may be due to carrier depletion and scattering at the edge of these overlapping MoS 2 flakes [17].Spin-coated hybrid films from MoS 2 suspension in a polymeric semiconductor have been reported where the suspension was found to improve the transport properties of the TFT devices [11][12][13]18], but it is still unclear whether the suspension only affects the electrical characteristics through changes to the electronic structure.Considering the band alignment between the P3HT and MoS 2 , (see figure S2 in the supplemental section, the valence band maximum of the MoS 2 is located 200 meV below the P3HT highest-occupied molecular orbital level.This band offset creates an energy barrier for the hole transport between the MoS 2 and the p-type P3HT semiconductor hampering charge transport from the polymer to the MoS 2 .
In addition, few attempts to fabricate TFTs by inkjet printing using hybrid nanocomposite semiconductor inks have been attempted.Inkjet-printed suspensions of graphene, semiconducting nanowires in insulating polymers, and single-walled carbon nanotubes suspended in a semiconducting solution have been reported [6,19,20] but inkjet-printed active layers for TFT fabrication have not been demonstrated and the role of the suspension is still unknown for devices formed by printing.
In this work, MoS 2 nanosheets suspended in a P3HT solution (using 1,2-dichlorobenzene (DCB, C 6 H 4 Cl 2 ) as the solvent) were used to formulate a hybrid nanocomposite semiconductor ink for inkjetprinted TFTs.Both semiconductor materials have been studied separately and their properties are well known, providing a good reference point for hybrid ink formulations.The electrical properties of the nanocomposite active channel layer were studied as well as the effect of the suspension on the structural characteristics of the thin film.In addition, to effectively form thin layers of suspended MoS 2 nanosheets within the polymer matrix, a surfactant, was also used to create thin layers of MoS 2 nanosheets within the polymer matrix, enhancing the MoS 2 dispersion within the P3HT [21,22].Ultraviolet-visible (UV-vis) spectroscopy was employed to evaluate the density of the exfoliated suspension of the MoS 2 in the ink that was enhanced by the surfactant.The printed devices were comprehensively characterized by current-voltage (I-V) measurements, atomic force microscopy (AFM), UV-vis absorption spectroscopy, and x-ray diffraction (XRD) to analyze and understand the influence of the ink formulation, printing process, and thin-film structure on the electrical characteristics of the printed TFTs.

Experimental
The formulation of the hybrid ink, consisting of MoS 2 suspension in P3HT, started with a MoS 2 nanosheet suspension created from 1 mg ml −1 MoS 2 /DCB mixed in a glass vial.The mixture was sonicated (using an LS-04D Digital Ultrasonic Cleaner Bath, Limplus) for 2 h (figure 1(a)) to help separate the MoS 2 into nanometer-thick sheets.To increases the dispersion of few layered MoS 2 nanosheets a surfactant trichloro(dodecyl)silane (DDTS, C 12 H 25 Cl 3 Si), consisting of 0.5% (volume/volume percent) was added to the MoS 2 /DCB mixture to functionalize the surface of the MoS 2 during the sonication/exfoliation process (figures 1(a) and (b)).The DDTS helped prevent the restacking of MoS 2 nanosheets and created a stable suspension that lasted several days (figure 1(c)).
The polymeric semiconductor was prepared using a P3HT powder mixed into the DCB solvent on a hotplate (10 mg ml −1 P3HT, DCB); the solution was heated for 1 h at 85 • C to form the semiconducting polymeric ink.
The MoS 2 concentration in the P3HT solution was varied by mixing different volumes of MoS 2 suspension into a fixed volume of P3HT solution (figure 1(a)).The volume of the MoS 2 suspension in DCB was varied from 25-200 µl added into 1 ml of P3HT, creating mixtures having MoS 2 concentrations of 0.25, 0.5, 0.75, 1, and 2 wt%, respectively.Prior to printing, all hybrid ink mixtures were sonicated for 30 min.
The fabricated devices have a common bottomgate, bottom-contact TFT structure (see figure 1(a)) with the I-V measurements carried out in a vacuum environment, under dark conditions.Electrical contacts to the semiconductor were created with Cr/Au (3 nm/30 nm) source and drain (S/D) patterns deposited on 100 nm thick thermal SiO 2 on highly doped p + -Si wafers.The contacts were defined using a metal lift-off process and patterned using conventional photolithography.The device geometry was defined by the width-to-channel length ratio (W/L) of ∼1000 µm/10 µm.The contacts were then cleaned in an ozone environment generated by ultraviolet light followed by a solvent-cleaning process using an ultrasonic bath consisting of DI water, acetone, and isopropanol to remove surface residue from the substrate.The substrate surface was next treated with hexamethyldisilazane to create a hydrophobic oxide surface (see figure S3 in the supplemental section).At this point, the hybrid ink (MoS 2 + P3HT) was printed onto the prepared substrate.The inkjet printing was accomplished using a single nozzle printhead (MJ-ATP-01-60 µm, MicroFab) (figure 1(d)).The P3HT and hybrid inks were jetted onto the channel regions that were defined by the photolithographically defined S/D pattern (see figure S4 in the supplemental section).After printing the semiconductor layer, the samples were annealed at 120 • C in a vacuum oven for 1 h and allowed to cool down to room temperature.

Effect of MoS 2 surface functionalization on ink formulation
The prepared MoS 2 dispersion was inkjet-printed onto a SiO 2 surface to determine the wetting properties of the ink and the assembly of the MoS 2 nanosheets on the oxide surface.The assembled MoS 2 structure is shown in the scanning electron microscopy (SEM) micrograph shown in figure 2. Depending on the length of the sonication process, the formed suspensions ranged in size from a few microns (figure 2 The formulation of the hybrid ink requires well-exfoliated MoS 2 nanosheets suspended in the  defects [18,23].Unfortunately, conventional liquidphase exfoliation methods, do not yield suspensions having high density or large lateral size with single or few-layer MoS 2 nanosheets [24][25][26] compared to films formed by chemical vapor deposition or mechanical exfoliation.Thin layers of MoS 2 will disperse more easily in the solution, preventing agglomeration of the nanosheets.The clustering of the suspended MoS 2 was observed to lead to the MoS 2 deteriorating into clusters that settled out of suspension.To overcome this challenge, DDTS was added to the DCB solvent to functionalize the surface of MoS 2 that resulted in a more uniform dispersion of the suspended flakes and enhanced the exfoliation process for creating few-layer MoS 2 structures [21,22]. To determine the effect of the surface functionalization, two suspensions of MoS 2 in DCB were formulated, one ink formulated with and a second ink without the DDTS additive.The suspensions were found to be visually similar in terms of optical transparency when the inks were first formulated but samples without the DDTS treatment were found to visually settle out of the suspension as shown by the increased optical transparency of the ink after 48 h (figure 1(c)).Both suspensions were diluted with 1 mg ml −1 to 0.1 or 0.05 mg ml −1 which are shown in the inset of figure 3. The suspension with DDTS showed a uniform pale yellowish-green color suggesting a well-exfoliated nanosheets suspension.In contrast, the MoS 2 nanosheets without the DDTS treatment have a different color (grey) of suspension suggesting a difference in terms of its size, layer thickness, and distribution in the solvent [27].
Using UV-vis absorption spectroscopy, the optical absorption of the suspended MoS 2 in DCB was compared between formulations with and without the DDTS treatment (figure 3).The A and B absorption peaks, at 625 and 700 nm, respectively, originate from excitonic peaks of few-layer nanosheets and represent the 2H phase of the MoS 2 [16,25,27,28].The C and D peaks, at 500 and 460 nm, respectively, are related to the thickness of the MoS 2 structures where bulk MoS 2 has an indirect bandgap and monolayer structures have a wider direct bandgap [25,29].The absorption was found to increase for the DDTS-treated suspension indicating an increase in few layer MoS 2 nanosheets.The C peak absorption at 500 nm also showed a slight blue shift for the DDTS-treated flakes further showing an increase in bandgap energy of the MoS 2 due to an increase in the creation of few-layer structures from the exfoliation process [25,30].Furthermore, the ratio between the C-D peaks and A-B peaks of the MoS 2 suspension without DDTS was higher at 0.1 mg ml −1 than at 0.05 mg ml −1 indicating an increasing density of few-layer MoS 2 nanosheets in the higher MoS 2 concentration, where the density of few-layer MoS 2 in the suspension increased for DDTS functionalized MoS 2 .

Inkjet-printed hybrid nanocomposite inks for TFT fabrication
To investigate the effect of the suspension on the TFT performance, inks with and without the MoS 2 surface functionalization were used to print hybrid and pristine (P3HT only) active layers.Printed TFTs from the nanocomposite ink were characterized with MoS 2 concentrations from 0.25 wt% to 2 wt%.Generally, a transparent orange tint in the P3HT solution was observed for a fully dissolved (amorphous state) P3HT solution [31,32].This color was observed to darken with increasing concentrations of suspended MoS 2 (inset of figure 4(a)).
Using the current-voltage (I-V) measurements, the saturation field-effect mobility of the inkjetprinted TFTs was extracted as a function of the MoS 2 concentration for the printed hybrid inks (figure 4(a)).The device parameters were extracted using the gradual channel approximation [33]: where I D, sat is the drain to source current in the saturation regime, the channel width (W) to length (L) ratio (W/L) is the normalized channel dimension (1000/10 µm), C ox is a gate oxide capacitance (34.5 nF cm −2 ), V G is the applied gate bias, and the threshold voltage is V th .Reference devices were fabricated using jet-printed P3HT-only inks.A total of eighteen baseline devices were characterized (see figure S5 in the supplementary section) and the average mobility values are shown in figure 4(a).The average extracted saturation mobility of the reference devices was ∼0.011 cm 2 V −1 s −1 .The extracted fieldeffect mobility of the hybrid ink devices increased with increasing MoS 2 concentration, with maximum saturation mobility, µ sat ∼ 0.038 cm 2 V −1 s −1 (at 0.5 wt%) and 0.032 cm 2 V −1 s −1 (at 0.75 wt%) for the TFTs having suspended nanosheets treated with and without DDTS, respectively.In both cases, the obtained mobilities from hybrid inks were enhanced by a factor of three (3×).All devices demonstrated well-defined switching behavior with an on/off current ratio of ∼10 5 and subthreshold swing of ∼3 V/decade with the nanocomposite TFTs having higher on-current levels (regardless of any DDTS surface functionalization) compared to the baseline TFTs.The electrical characteristics of all fabricated devices are summarized in table 1.In all ink formulations, the extracted saturation mobility was found to decrease with increasing MoS 2 concentration after the 0.75 wt% and 0.5 wt% peak for the formulations with and without the surfactant, respectively.
To understand these observations, the DDTS functionalized MoS 2 nanocomposite inks were found to have a higher density of few-layer MoS 2 nanosheets after sonication, based on the absorption measurements shown in figure 3. The higher C-D peaks and the increase in the ratio between the C-D to A-B peaks in the UV-vis spectrum supports the hypothesis that a more effective exfoliation process is obtained with the DDTS surface treatment compared to suspensions created without the surfactant.
The TFT threshold voltages (V th ) were extracted from the linear extrapolation of the √ I D vs. V G plots.In figure 4(d), the pristine P3HT-only ink formulation has a typical V th of −9.4 V while the V th of the TFTs fabricated from the hybrid nanocomposite inks had positively shifted threshold voltages [14,34,35].The V th shift may be due to trapped charges at the interface formed along the MoS 2 nanosheets within the bulk P3HT matrix due to interface states [36].The increasing concentration of MoS 2 nanosheets increases this trap density within the nanocomposite films [22].
To further understand the effect of the trap density on the TFT performance, the electrical stability of the devices was investigated under constant DC gate-bias stress for devices having different ink formulations.Figure 4(e) shows the drain current as a function of time for samples processed using different suspension concentrations operating under V G = −40 V and V D = −10 V.The normalized current was observed to decrease between 33% for P3HTonly films to 70% for the highest concentration of DDTS treated MoS 2 in P3HT.Comparing devices fabricated with inks having 2 wt% MoS 2 in P3HT, the current decay was found to be higher for ink formulations having the surfactant treatment (blue diamonds, in figure 4(e)) compared to suspensions without treatment (green reverse triangles).The results further support the effect of the surfactant to effectively separate layers of MoS 2 , increasing its density in the suspension.This increasing nanosheet density raises the interface density for charge trapping that may be responsible for the observed V th shift in the TFT electrical characteristics.The inclusion of the suspension has improved the transport properties of the printed TFTs but at the expense of the electrical stability of the devices.Continuing investigations are being made to determine this interface effect for other polymeric semiconductors [22].Further investigations of the I-V characteristics suggested changes in the electrical characteristics may not be only due to the electronic structure of the semiconductor.Evidence for a structural dependence on the MoS 2 concentration was observed by AFM (see figure S6 in the supplemental section) with increasing MoS 2 concentration up to 2 wt%, resulting in the decreasing field-effect mobility shown in figure 4(a) for both ink formulations.This effect will be analyzed in the next section.

Dependence of the electrical characteristics on the microstructure in hybrid thin films
As the MoS 2 concentration increases in the hybrid inks, interface states formed at the organic/inorganic boundary increase, imposing a structural influence on the formation of the polymeric thin film in addition to the electrical changes measured in the I-V characteristics.The introduced MoS 2 nanosheets act as surfaces for the P3HT to crystallize during film formation.This influence may affect the longrange ordering of the polythiophene film formation where structural nanofiber-like aggregates have been observed for high-quality materials [31,[37][38][39][40][41].As the concentration of the MoS 2 increases, the higher density of nucleation sites disrupts the long-range ordering through the film, resulting in the introduction of interfacial defects at grain boundaries between the MoS 2 and P3HT that degrades the structural To determine if the structural properties of the film are affected by the MoS 2 nanosheet, AFM scans were performed to characterize the surface morphology of the films.In figure 5, the surface morphology of TFT samples with pristine P3HT (figure 5(a)) Figure 5(b) shows that the formation of nanofiberlike structures [40,41] was observed in the 0.5 wt% MoS 2 on the P3HT surface (highlighted by the white arrows), suggesting an enhancement of the structural quality of the polymeric thin-film matrix.These devices had the highest extracted field-effect mobility.In comparison, the nanocomposite film with 2 wt% MoS 2 showed smaller polymer domains in the P3HT (see figure S6(c) in the supplemental section).
The degree of crystallinity was characterized further between pristine P3HT films and hybrid nanocomposite films by UV-vis spectroscopy, shown in figure 6(a).The UV-vis absorption of pristine P3HT film where the 0-0 (λ ∼ 600 nm), 0-1 (λ ∼ 550 nm), and 0-2 (λ ∼ 520 nm) peaks are clearly observed showing a well-ordered P3HT-based film [42][43][44][45].The absorption spectra of three films having different MoS 2 concentrations are normalized at the 0-2 transition to compare the 0-0 and 0-1 peaks relative to the crystallinity of the film [44,45].The two peaks of the hybrid film have a higher intensity compared to pristine P3HT films, indicating larger crystalline domains [43][44][45].This degree of crystallinity can be quantified and determined by estimating the exciton bandwidth (W ex ) with the following equation [42,43]: where A 0-0 and A 0-1 are peak intensities of the 0-0 and 0-1 vibrational transitions.The main intramolecular vibration peak (E P ) of P3HT is at 0.18 eV [39,42,43].Table 2 shows the extracted exciton bandwidth, W ex , decreasing with 0.5 wt% MoS 2 concentration which indicates an increase in the crystalline order of the thin film [46].A slight increase in W ex is extracted for 2 wt% MoS 2 samples suggesting a lowering in the structural quality as the MoS 2 concentration increases.
To confirm the UV-vis observations, XRD measurements were conducted to further evaluate the degree of crystallinity of the hybrid nanocomposite films [47][48][49].Figures 6(b) and (c) show the out-ofplane and in-plane reflections of the polymeric thin film, respectively.The inset of figure 6(c) shows the orientations of the thin-film microstructure for the alkyl chain lamellar stacking, π-π interchain stacking, and π-conjugation in the polymeric film that are represented by the (h00), (0k0), and (00l) reflections, respectively [49,50].
For the out-of-plane reflections, the (100), (200), and (300) peak intensities of the printed films were found to be higher in the hybrid nanocomposite films compared to a pristine P3HT film.The (100) reflection at an angle 2θ = 5.5 • was found to be strongest for the 0.5 wt% MoS 2 indicating this concentration resulted in a highly ordered alkyl side chain in the P3HT structure [47,51].The (200) and (300) peaks (at 2θ = 10.8 • and 16.3 • , respectively) are observed in the inset of figure 6(b), but the (300) diffraction peak is not clearly defined as this peak overlaps with the MoS 2 (002) reflection 2θ ≈ 15 • [52,53].This (002) MoS 2 peak was only observed in the hybrid thin film made with higher concentrations of MoS 2 suspension (from 0.5 wt% to 2.0 wt%).The in-plane XRD spectroscopy also shows the 0.5 wt% MoS 2 hybrid film had the highest (010) peak intensity due to the stronger π-π interactions in the hybrid films compared to P3HT alone [49].The XRD measurements support the UV-vis results, showing the highest quality films were obtained with a 0.5 wt% MoS 2 concentration in P3HT, correlating to the highest extracted field-effect carrier mobilities in the TFT devices.As the concentration of the MoS 2 increases, the P3HT structural properties were found to degrade and this degradation was correlated to the reduction in the extracted TFT field-effect mobility.

Conclusions
Few-layer MoS 2 nanosheets were incorporated as a suspension into polythiophene-based electronic inks.The hybrid nanocomposite inks were inkjet-printed for the fabrication of TFT devices.For the ink formulation, the exfoliation of the MoS 2 layers was accomplished by sonicating a mixture of DDTS in a DCB solution.UV-vis spectroscopy showed that adding the surfactant increases the creation of few-layer MoS 2 nanosheets, resulting in improved electrical and structural properties of the inkjet-printed P3HT channel TFTs.The printed TFTs from inks formulated with DDTS showed the highest field-effect mobility compared to formulations without the surfactant.The TFT performance was also found to be dependent on the concentration of MoS 2 nanosheets within the nanocomposite ink.The field-effect mobility was observed to increase by as much as three times (3×) with increasing MoS 2 concentration.This effect was found to decrease for TFTs having MoS 2 concentrations >0.5 wt% in P3HT.The dependence of the TFT performance on higher concentrations of MoS 2 additives was attributed to a degradation in the P3HT structural quality, as shown by XRD.The increased MoS 2 concentration increases the density of nucleation sites for nanocrystalline film formation that results in the increased formation of grain boundaries within the thin-film layer.

Figure 1 .
Figure 1.(a) Schematic illustrations of the overall fabrication process for inkjet-printed hybrid channel TFTs.The MoS2 suspension is prepared by adding bulk MoS2 powder in DCB as a solvent and 0.5 vol/vol % DDTS in a glass vial.The suspension is sonicated for two hours in an ultrasonic bath.The well-dispersed MoS2 suspension is mixed with P3HT solution with different volumes (from V1 to V4) to make different concentrations of a suspension in the hybrid ink.The ink is inkjet-printed on an array of S/D electrodes with a single ejector printhead.(b) Schematic illustration showing how DDTS intercalated MoS2 interlayers functionalize the MoS2 surface during sonication.(c) Photograph of vials containing MoS2 suspensions; the suspension with DDTS treatment stayed in suspension while non-DDTS treated suspensions formed a sediment after 48 h, (d) Video capture of the single ejector printhead with a stable ejection of the hybrid ink; droplet characteristics are given in the video captions.
(a)) to well-exfoliated few-layered nanosheets having a lateral size of approximately one micrometer shown in figure 2(b).

Figure 3 .
Figure 3. Normalized UV-vis absorption spectra of two MoS2 suspensions with (blue) or without (green) DDTS treatment.The inset shows a photograph of MoS2 suspensions with (left) and without (right) DDTS treatment.

Figure 4 .
Figure 4. Summary of electrical device characteristics for: (a) mobility changes as a function of MoS2 concentration, (b) transfer characteristics of three TFT devices having different ink formulations; pristine P3HT TFTs and hybrid channel TFTs with and without DDTS, (c) on/off current ratios and (d) threshold voltage changes as a function of MoS2 suspension concentrations.(e) DC gate-bias stress results of devices having different ink formulations.

Figure 6 .
Figure 6.(a) UV-vis spectroscopy results of thin films for the reference and two ink formulations with different MoS2 concentrations.XRD results of all samples with (b) out-of-plane and (c) in-plane P3HT reflections.

Table 1 .
Electrical properties for inkjet printed TFTs with P3HT (reference), P3HT + MoS2 with and without DDTS treated suspensions.

Table 2 .
UV-vis exciton bandwidth for films formed using different concentration of MoS2 suspensions.