Optimization of chitosan gated electric double layer transistors by combining nanoparticle incorporation and acid doping

Abstract

In this article, we report on the successful implementation of nanoparticle incorporation and acid doping in tuning the properties of chitosan for proton-based ion gating of metal–oxide electric double layer transistors (EDLTs). SiO₂ nanoparticle (nano-SiO₂) incorporation was initially found to produce rougher surfaces, lower the specific capacitance and postpone the transport of protons, which deteriorated the performance of the transistors. The advantage of nano-SiO₂ incorporation lies in their ability to significantly improve the transistor's negative bias stability due to a proposed proton blocking mechanism. By applying doping of H₃PO₄, the negative effects of nano-SiO₂ incorporation were partially counterbalanced. This is because H₃PO₄ can donate more protons and assist the proton conduction via the donor/acceptor process of amphoteric acids or forming the molecular network of hydrogen bonds. Overall, this work presents a promising strategy to optimize the bio-polyelectrolyte gated EDLTs by combining nanoparticle incorporation and acid doping.

---

1 Introduction

Electrostatic gating with ion based dielectrics, e.g., ionic liquid, ion gel and polyelectrolyte has become a research hotspot in the last few years. Thanks to the strong charge coupling effect of the electric double layer (EDL) formed at the interface of semiconductor and dielectric, EDLTs (also known as ion or electrolyte gated transistors) operate in a small gate voltage (usually <2 V) range. Intriguing field induced surface phenomena, e.g., interfacial superconductivity or insulator-to-metal transition, have been demonstrated.^1,2 Moreover, the ion based charge modulation process has also exhibited great potential in enabling new possibilities in chemical sensing and artificial synapses.^3–7

Natural biomaterials that are biosynthesized, biocompatible and biodegradable are appealing candidates for the use in electronic devices. Chitosan, the deacetylated form of naturally abundant biopolymer chitin, is a water soluble cationic bio-polyelectrolyte when the amine groups on the D-glucosamine units are protonated. Early studies of chitosan as the bio-polyelectrolyte were mainly focused on proton exchange membranes (PEMs) for fuel cells.^8,9 The emergence of proton gated EDLTs has highlighted its potential as a flexible and eco-friendly gate dielectric material in EDLTs.¹⁰ Free standing chitosan membrane based EDLTs were reported recently.^11,12 However, besides low ionic conductivity, neat chitosan hydrogel also exhibits low mechanical strength due to the high water content and relatively loose polysaccharide network.¹³ It is therefore required to look for a rational combination of material composition to tune its properties for the use in EDLTs.

The modification of polymer electrolytes generally involves adding inorganic fillers, doping with acids/ion donors, cross-linking/copolymerization and grafting of functional groups.^14–17 It is well known that polymers can be doped with amphoteric acids to increase their ionic conductivity, in which the acids act both as proton donor and acceptor during the proton migration. H₃PO₄ doped polymers are typical examples of this approach and have been widely studied as high-temperature PEMs.^15,18–20 Meanwhile, inorganic reinforcement fillers, e.g., carbon black, mica and silica, are often added to improve the mechanical strength of polymers.²¹ The presence of an solid component could compensate the loss of mechanical strength after the acid doping.²² Recent studies using nano-particulate fillers also suggested that the high specific surface area of nano-fillers induces the formation of amorphous phase in the biopolymer, which is beneficial to the transport of ions.²³ However, to the best knowledge of the authors, these material modification approaches had not been applied to the bio-polyelectrolytes for EDLTs and their influence on the performance of EDLTs had not been explored.

In this work, chitosan/nano-SiO₂ and chitosan/nano-SiO₂/H₃PO₄ systems were studied as the gate electrolyte of indium gallium zinc oxide (IGZO) EDLTs. The incorporation of nano-SiO₂ was found to produce rougher surface, reduce the specific capacitance and postpone the proton migration, resulting in the deterioration of the transistors gated by chitosan/nano-SiO₂ electrolyte. Such an effect could be partially counterbalanced by further doping of H₃PO₄ because H₃PO₄ can donate more protons and assist their transport. Most interestingly, the advantage of nano-SiO₂ incorporation lies in their ability to significantly improve the negative bias stability due to a proposed mechanism of proton blocking effect. EDLTs with chitosan/nano-SiO₂/H₃PO₄ electrolyte therefore had a better performance in terms of overall considering the bias stability, on/off ratio, etc. This work shows that the combination of nanoparticle incorporation and acid doping is a promising strategy to optimize the bio-polyelectrolytes for the proton based gating of EDLTs.

2 Experimental

The chitosan solution (2.0 wt%) was prepared by dissolving chitosan powders (Sinopham, deacetylate ratio ∼87%) into acetic acid solution of 2.0 wt%. To prepare the composite systems, weighted amount of SiO₂ nanoparticles (Aladdin, 15 nm) were first added into deionized water under sonication. Then acetic acid and chitosan were added stepwise. After the complete dissolving of chitosan by stirring, H₃PO₄ was added and mixed completely. The chitosan : nano-SiO₂ weight ratio was 10 : 1.5 in the chitosan/nano-SiO₂ system and the chitosan : nano-SiO₂ : H₃PO₄ weight ratio was 10 : 1.5 : 1 in the chitosan/nano-SiO₂/H₃PO₄ system. The electrolyte membranes were prepared by drop casting of 0.5 ml corresponding solution onto 2 cm × 2 cm ITO glass and followed by drying at 60 °C for 6 h. The film thicknesses measured with digital thickness gauge (Mitutoyo) were 10.5 ± 2 μm, which was further confirmed by scanning electron microscope (SEM) (Sigma, Zeiss) observation. Surface roughness was determined with an atomic force microscope (AFM) (NTEGRA, NT-MDT).

IZO electrodes were deposited on top of the as-prepared membrane to make the IZO/electrolyte/ITO sandwich structures for the capacitance and proton conductivity characterizations. Frequency dependent capacitance was measured using a LCR meter (HIOKI IM3533-01). Impedance measurements were performed on an impedance analyzer (Solartron 1260). The frequency sweeping range was 1 MHz to 1 Hz and the applied peak-to-peak voltage was 300 mV.

EDLTs with top-contact structure were fabricated on ITO glass with the ITO layer acting as the bottom gate electrodes. The IGZO channel layer (target In : Ga : Zn = 1 : 1 : 1, atomic ratio) of 40 nm thick was deposited on top of the prepared electrolyte by radio-frequency magnetron sputtering. The deposition was performed in Ar at the pressure of 0.5 Pa and power of 80 W. Shadow mask was used to define the area of IGZO deposition. IZO source/drain (S/D) electrodes were then deposited by sputtering with a second shadow mask. The deposition was performed in Ar at the pressure of 0.5 Pa and power of 100 W. The obtained channel has the width of 1000 μm and length of 80 μm (W/L = 12.5). Electrical measurements were performed on an Ever-Being probe station quipped with Source-Meter Units (Keithley 2612B and 2636B).

3 Results

3.1 Characterization of composite bio-polyelectrolytes

Top-view SEM images of the prepared electrolyte films are shown in Fig. 1(a)–(c), and the corresponding AFM images are shown in Fig. 1(d)–(f). The pristine chitosan film displays a rather uniform surface with the root-mean square (rms) roughness estimated to be 0.61 nm. After the incorporation of nano-SiO₂, nanoparticles are visible in the SEM images of chitosan/nano-SiO₂ and chitosan/nano-SiO₂/H₃PO₄ electrolyte films. These nanoparticles distributed quite well in the film, the rms roughness value increased to 10.7 nm and 11.3 nm for the chitosan/nano-SiO₂ and chitosan/nano-SiO₂/H₃PO₄ electrolyte films, respectively.

Fig. 1 Top-view SEM images of (a) chitosan, (b) chitosan/nano-SiO₂ and (c) chitosan/nano-SiO₂/H₃PO₄ films. (d), (e) and (f) are the corresponding AFM images in a range of 5 μm × 5 μm.

Fig. 2(a) shows the specific capacitance versus frequency relation of the prepared electrolyte films. In all cases, the specific capacitance increases by several orders of magnitude as the frequency decreases. The rising gradient is much larger in the high frequency range (100k–2k Hz) than in the low frequency range (2k–1 Hz). Such a strong frequency dependence of specific capacitance is typically found in ion (proton here) based dielectrics. It indicates that the EDL capacitance is dominating in the low frequency range and the protons' contribution to the EDL capacitance is stronger in the at low frequency, as the mobile protons would have enough time to migrate to the interface.^24–26 Starting from about 2k Hz, the quick drop of capacitance indicates the EDL is fading and the dipole polarization of bulk material will eventually dominate.²⁷ Of most interested for EDLTs, is the high specific capacitance (∼μF cm⁻²) in the low frequency range that enables the low voltage operation. In general, one could expect to have higher capacitance with a rougher surface. However, the specific capacitance of chitosan/nano-SiO₂ films was found to be lower than the pristine chitosan. It indicates that despite having a rougher surface, the presence of nano-SiO₂ has inhibited the formation of the EDL, which can be ascribed to two reasons: (I) the blocking effect of nano-SiO₂ in proton migration;⁹ (II) the rigid SiO₂-electrode interface does not support the formation of EDL. On the other hand, the highest specific capacitance was obtained with the chitosan/nano-SiO₂/H₃PO₄ films.

Fig. 2 (a) Frequency dependent specific capacitance curves of the films. (b) Nyquist plots of the films; the arrow indicates where the bulk resistance R was retrieved for the chitosan/nano-SiO₂/H₃PO₄ film. (c) Schematic illustration of the proton transport process with the assistance of H₂O and H₃PO₄.

Typical impedance spectra (Nyquist plots in Fig. 2(b)) of the chitosan based bio-polyelectrolytes consist of a high-frequency distorted semicircle and a low-frequency spike. The former represents bulk relaxation phenomena in these electrolytes whereas the latter is related to the electrode/electrolyte interfacial phenomena.²⁰ The high-frequency end of the semicircle on the real axis was ascribed to the series resistance (R_s) in the electrical connection. The proton conductivity σ was calculated with the equation, σ = d/((R − R_s)A), where d is the thickness of the electrolyte layer, R is the bulk resistance of the electrolyte (obtained as the Z′ value at which −Z′′ goes through a local minimum) and A denotes the electrode surface area (0.20 cm²).^28,29 The obtained σ values are listed in Table 1. The chitosan/nano-SiO₂ electrolyte was found to have the lowest σ, providing proof to the above mentioned proton blocking effect of nano-SiO₂. Further doping with H₃PO₄ has set the σ back to the level of pristine chitosan. In the chitosan, the –NH₂ groups are protonated in the acid assisted dissolution process and the presence of H₂O also provides protons by self-ionization (2H₂O ↔ H₃O⁺ + OH⁻). H₂O molecules form hydrogen bonds with the –OH and –NH₂ groups on chitosan, allowing protons to migrate through the chitosan chain by hopping (the Grotthuss mechanism).^30,31 However, the acetic acid used to dissolve chitosan is known to be volatile and has been reported to evaporate during the drying process.³² The doping of H₃PO₄, which is stable even above 100 °C, increases the concentration of mobile protons.³³ The incorporated H₃PO₄ could bond to the chitosan/SiO₂ via hydrogen bonding and the deprotonated H₂PO₄⁻/HPO₄²⁻ also act as acceptor/donor of proton, providing new pathways to proton transport and leading to faster proton transport.³⁴ Jointing the relative rougher surface, all of these add up to the high specific capacitance and ion conductivity of the chitosan/nano-SiO₂/H₃PO₄ ternary system. A schematic illustration of the proposed proton transfer process is given in Fig. 2(c).

Table 1 Parameters of the electrolyte films and IGZO EDLTs

	IZO/electrolyte/ITO		IGZO EDLTs
	Ci (μF cm^−2) at 1 Hz	σ (S cm^−1)	Ion/Ioff	Vth (V)	μ (cm^2 V^−1 s^−1)
Chitosan	6.7	9.7 × 10^−5	5.1 × 10^5	0.57	3.3
Chitosan/nano-SiO2	5.1	1.1 × 10^−5	1.4 × 10^5	0.60	1.2
Chitosan/nano-SiO2/H3PO4	13.4	6.2 × 10^−5	8.3 × 10^5	0.61	2.0

---

3.2 Transistor performance

With proton gating, the n-type IGZO transistor turns on when protons are driven toward the IGZO/electrolyte interface. The transfer curves are showing in Fig. 3(a)–(c) for the bio-polyelectrolyte gated EDLTs. High on/off ratio in a small V_G range of −0.5 to 1.5 V was produced and the gate current I_G was less than 0.01% of the I_DS. As summarized in Table 1, the field-effect mobility μ was calculated with the equation μ = 2LI_DS/[WC_i(V_G − V_th)²], where C_i is the specific capacitance of the gate dielectric (C_i values at 1 Hz were used), and V_th is the threshold voltage, W and L are the width and length of the IGZO channel, respectively. Comparing to the pristine chitosan, the saturate I_DS, μ and on/off ratio were found to be lower for the chitosan/nano-SiO₂ gated EDLTs. This is reasonable when the rougher surface and lower specific capacitance of the chitosan/nano-SiO₂ gate electrolyte are taken into account. On the other hand, doping of H₃PO₄ helped the EDLT to regain its property in terms of saturate I_DS, μ and on/off ratio. Indicating the high capacitance brought by acid doping is capable of counterbalancing the negative effects caused by nanoparticle incorporation. Moreover, an anti-clockwise hysteresis was observed due to the accumulation and slow relaxation of protons at the interface. EDLTs with composite electrolytes displayed a smaller hysteresis loop than pristine chitosan. For chitosan/nano-SiO₂, it should be related to the lower capacitance and low level of proton accumulation at the interface. For chitosan/nano-SiO₂/H₃PO₄, the small hysteresis can be attributed to the faster migration of protons with the assistance of H₃PO₄.

Fig. 3 Transfer curves of EDLTs gated by (a) chitosan, (b) chitosan/nano-SiO₂ and (c) chitosan/nano-SiO₂/H₃PO₄. (d), (e) and (f) are the corresponding transfer curves measured in negative bias stability (NBS) tests. The dashed line represents the I_DS^1/2 vs. V_G relation that was fitted to obtain the mobility and threshold voltage. The legends in (d), (e) and (f) are the time of the applied negative bias.

3.3 Negative bias stability

Operation stability of the transistor under prolonged bias is a major concern in practical uses. Negative bias is frequently applied to keep the device in off state. For negative bias stability (NBS) tests, V_G was fixed at −0.5 V and the source/drain electrodes were grounded. As shown in Fig. 3(d)–(f), the transfer curves of the chitosan gated EDLT experienced a dramatic shift of V_th in positive direction and a drop of the on state current. Such a positive V_th shift is similar to the previously reported inorganic electrolyte gated EDLT,^35,36 but opposite to metal–oxide transistors gated by conventional type dielectrics.^37–39 In addition, the lowering of the off state current I_DS after negative bias in all the cases is a sign of depleted ion current due to ion accumulation and trapping. Interestingly, there was little shift of transfer curves for the nano-SiO₂ incorporated EDLTs. The extracted ΔV_th and μ change are given in Fig. 4(a)–(b). Both V_th and μ are much more stable with the incorporation of nano-SiO₂.

Fig. 4 (a) The threshold voltage shift ΔV_th and (b) the drift of mobility μ as a function of the bias time in the NBS tests, respectively.

Taking into account the ion based gating mechanism and the fact that nano-SiO₂ incorporation resulted in significant improvement of the NBS, a model is proposed here to explain the positive V_th shift and stabilization effect by nano-SiO₂. As depicted in Fig. 5(a), protons are accumulated at gate/electrolyte interface during the prolonged period of negative gate bias. Considering the slow motion of proton, their arrival to the semiconductor/electrolyte interface will be delayed when switching to positive V_G. This causes the positive shift of V_th and drop of the on state current in pristine chitosan. The high stability of nano-SiO₂ incorporated EDLTs to negative bias is related to the proton blocking effect mentioned previously. During the negative bias stress, some protons will be hold up by the nanoparticles (Fig. 5(b)), these protons can't travel far from their original position due to the blocking effect. However, these protons can go to semiconductor/electrolyte interface when a positive V_G is applied. So the transfer curves of the EDLTs with nano-SiO₂ incorporated chitosan electrolyte were not largely affected.

Fig. 5 Schematic illustration of a proposed mechanism of proton migration and accumulation in (a) pristine chitosan and (b) nano-SiO₂ incorporated chitosan gated EDLTs. In pristine chitosan, mobile protons are readily accumulated at the negatively biased electrode; while in nano-SiO₂ incorporated chitosan, accumulation of photons can be prevented by the solid nanoparticles.

4 Conclusions

In summary, the performance of IGZO EDLTs gated by chitosan bio-polyelectrolyte could be optimized by nanoparticle incorporation and acid doping. The incorporation of SiO₂ nanoparticles was found to increase the surface roughness, reduce the specific capacitance and postpone the proton migration, which resulted in the deterioration of the transistors gated by chitosan/nano-SiO₂ electrolyte. This deterioration effect could be partially counterbalanced by doping H₃PO₄ into the chitosan/nano-SiO₂ electrolyte because H₃PO₄ can donate more protons and assist the proton transport. Most interestingly, due to the proton blocking effect, improved negative bias stability was maintained, stemming from the advantage of nano-SiO₂ incorporation. This work provides a promising strategy to the optimization of bio-polyelectrolyte gated EDLTs by combining nanoparticle incorporation and acid doping. It means that the existing knowledge on polyelectrolyte modification can also be applied to the polyelectrolyte gated EDLTs.

Acknowledgements

This work was supported by National Natural Science Foundation of China (51502131) and China Postdoctoral Science Foundation (2016M591819). The authors thank Dr Xuefei Li for AFM measurements and Prof. Linwei Yu for SEM measurements.

References

1. J. T. Ye, S. Inoue, K. Kobayashi, Y. Kasahara, H. T. Yuan, H. Shimotani and Y. Iwasa, Nat. Mater., 2010, 9, 125–128.
2. H. Shimotani, H. Asanuma, A. Tsukazaki, A. Ohtomo, M. Kawasaki and Y. Iwasa, Appl. Phys. Lett., 2007, 91, 082106.
3. N. Liu, L. Q. Zhu, P. Feng, C. J. Wan, Y. H. Liu, Y. Shi and Q. Wan, Sci. Rep., 2015, 5, 18082.
4. L. Q. Zhu, C. J. Wan, L. Q. Guo, Y. Shi and Q. Wan, Nat. Commun., 2014, 5, 3158.
5. S. H. Jo, T. Chang, I. Ebong, B. B. Bhadviya, P. Mazumder and W. Lu, Nano Lett., 2010, 10, 1297–1301.
6. Y. Yang and W. Lu, Nanoscale, 2013, 5, 10076–10092.
7. C. Wang, W. He, Y. Tong and R. Zhao, Sci. Rep., 2016, 6, 22970.
8. M. Yamada and I. Honma, Electrochim. Acta, 2005, 50, 2837–2841.
9. N. Shaari and S. K. Kamarudin, J. Power Sources, 2015, 289, 71–80.
10. H. Liu and Q. Wan, Nanoscale, 2012, 4, 4481–4484.
11. Y. H. Liu, L. Q. Zhu, P. Feng, Y. Shi and Q. Wan, Adv. Mater., 2015, 27, 5599–5604.
12. G. Wu and H. Xiao, RSC Adv., 2015, 5, 105084–105089.
13. A. S. Hoffman, Adv. Drug Delivery Rev., 2002, 54, 3–12.
14. S. Klongkan and J. Pumchusak, Electrochim. Acta, 2015, 161, 171–176.
15. J. S. Wainright, J. T. Wang, D. Weng, R. F. Savinell and M. Litt, J. Electrochem. Soc., 1995, 142, L121–L123.
16. B. Smitha, S. Sridhar and A. A. Khan, Macromolecules, 2004, 37, 2233–2239.
17. H. W. Zhang, D. Z. Chen, Y. Xianze and S. B. Yin, Fuel Cells, 2015, 15, 761–780.
18. J. A. Asensio, E. M. Sanchez and P. Gomez-Romero, Chem. Soc. Rev., 2010, 39, 3210–3239.
19. Q. Tang, S. Yuan and H. Cai, J. Mater. Chem. A, 2013, 1, 630–636.
20. W. Wieczorek and J. R. Stevens, Polymer, 1997, 38, 2057–2065.
21. B. B. Boonstra, Polymer, 1979, 20, 691–704.
22. D. J. Jones and J. Rozière, J. Membr. Sci., 2001, 185, 41–58.
23. S. Ramesh and C.-W. Liew, J. Non-Cryst. Solids, 2012, 358, 931–940.
24. H. Yuan, H. Shimotani, A. Tsukazaki, A. Ohtomo, M. Kawasaki and Y. Iwasa, J. Am. Chem. Soc., 2010, 132, 6672–6678.
25. S. Ono, K. Miwa, S. Seki and J. Takeya, Appl. Phys. Lett., 2009, 94, 063301.
26. O. Larsson, E. Said, M. Berggren and X. Crispin, Adv. Funct. Mater., 2009, 19, 3334–3341.
27. J. Sun, H. Liu, J. Jiang, A. Lu and Q. Wan, J. Mater. Chem., 2010, 20, 8010–8015.
28. Y. G. Jin, S. Z. Qiao, J. C. D. da Costa, B. J. Wood, B. P. Ladewig and G. Q. Lu, Adv. Funct. Mater., 2007, 17, 3304–3311.
29. B. Kim, Y. S. Cho, J.-G. Lee, K.-H. Joo, K.-O. Jung, J. Oh, B. Park, H.-J. Sohn, T. Kang, J. Cho, Y.-S. Park and J. Y. Oh, J. Power Sources, 2002, 109, 214–219.
30. C. Zhong, Y. Deng, A. F. Roudsari, A. Kapetanovic, M. P. Anantram and M. Rolandi, Nat. Commun., 2011, 2, 476.
31. S. Cukierman, Biochim. Biophys. Acta, Bioenerg., 2006, 1757, 876–885.
32. A. Pawlicka, R. I. Mattos, C. E. Tambelli, I. D. A. Silva, C. J. Magon and J. P. Donoso, J. Membr. Sci., 2013, 429, 190–196.
33. J. R. Stevens, W. Wieczorek, D. Raducha and K. R. Jeffrey, Solid State Ionics, 1997, 97, 347–358.
34. S. Yuan, Q. Tang and B. He, J. Mater. Sci., 2014, 49, 5481–5491.
35. J. Sun, J. Jiang, W. Dou, B. Zhou and Q. Wan, IEEE Electron Device Lett., 2011, 32, 910–912.
36. L. Yanghui, W. Xiang, Z. Li Qiang, S. Yi and W. Qing, IEEE Electron Device Lett., 2014, 35, 1257–1259.
37. J.-S. Seo and B.-S. Bae, ACS Appl. Mater. Interfaces, 2014, 6, 15335–15343.
38. K.-H. Liu, T.-C. Chang, K.-C. Chang, T.-M. Tsai, T.-Y. Hsieh, M.-C. Chen, B.-L. Yeh and W.-C. Chou, Appl. Phys. Lett., 2014, 104, 103501.
39. J. M. Kwon, J. Jung, Y. S. Rim, D. L. Kim and H. J. Kim, ACS Appl. Mater. Interfaces, 2014, 6, 3371–3377.

---