Direct printed silver nanowire thin film patterns for flexible transparent heaters with temperature gradients

Abstract

In this study, a convenient and facile direct printing method is developed to fabricate patterned silver nanowire (AgNW) thin films. An aqueous silver nanowire (AgNW) ink was first formulated and inkjet printed on flexible substrates with precisely controlled AgNW density per unit area. With a mesh grid structure, the printed heater showed high electrical conductivity and transparency up to 90%. From infrared thermography, a steady-state temperature up to 100 °C was observed uniformly over the printed mesh grid heater with only 4 volts of applied voltage. With precise control on local resistances via AgNW density variation, a gradient heater was designed and printed directly to provide a temperature gradient of up to 10 °C cm⁻¹. When printed on thin polyethylene terephthalate (PET) sheets, the printed heater can have a fast response time within 3 s due to the thermal insulating nature of PET. The printed heaters on PET also showed great mechanical stability and the resistances remained almost the same after 1000 bending cycles.

---

Introduction

Transparent heaters have attracted increasing interest for their wide application in defrosters,¹ displays,² solar panels³ and touch screens.⁴ The principle of electrical heaters is based on Joule heating, which converts electricity to heat through resistive losses. The power dissipated in a resistive conductor can be described by Joule's first law: P = V²/R, where V is the applied voltage and R is the total resistance. A low resistance is regularly necessary to obtain high heating power at low applied voltages, which prevent electrical damage to devices and shocks to the human body. The most commonly-used material for transparent heaters is indium tin oxide (ITO), which possesses high optical transparency (∼90%) and low sheet resistance (<100 Ω sq⁻¹). However, due to the brittleness of ITO, heaters made of ITO have a low tolerance to external stresses and thus little flexibility. To meet the surging demands of flexibility and portability in the next-generation macroelectronic systems, recent study has focussed on the development of new transparent conductive materials with better mechanical strength such as conducting polymers,^5,6 carbon nanotubes (CNTs),^7,8 graphene,^9,10 and metallic nanowires.^11,12 Among them, silver nanowire (AgNW) thin films demonstrate both low sheet resistance and great optical transparency due to its percolating structure.^13,14 Moreover, the network structures of AgNW also allow great flexibility to accommodate external mechanical stresses.¹⁵ Thus, AgNWs have been widely used in transparent heaters.¹⁶

To fabricate flexible AgNW thin film heaters with superior electro-optical properties, it is necessary to develop simple, reliable and cost-effective coating techniques. To fulfil these requirements, solution-process techniques operated at room temperature, such as spray coating,¹⁷ bar coating,¹⁸ or vacuum filtration,^1,12 have been commonly used. These wet coating methods can be easily scalable to produce AgNW thin films with arbitrarily large areas. With these methods, large-area AgNW thin film heaters with uniform temperature profiles and high optical transparency have been reported (Table 1). On the other hand, for biomedical applications, such as PCR & microfluidic devices,^19,20 local temperature gradients are usually needed for proper biochemical reaction conditions. Moreover, transparency is also necessary for these microfluidic applications for cell observation.²¹ To fabricate these patterned thin film heaters with both good local temperature control and transparency and to achieve various temperature distributions, commonly used coating techniques combined with pre-defined masks can be applied²² but with vast amounts of material waste.

Table 1 Comparison of the properties of AgNW-based thin film heaters

Coating method	Materials	Input	Max. temp. (°C)	T%	Substrate	Process temp. (°C)	Rs (Ω sq^−1)	tres (s)
Vacuum filtration^1	GO/AgNW	15 V	210	80	Quartz	700	27	110
Vacuum filtration^12	AgNW	7 V	140	92	PET	R.T.	11	140
Spray^11	AgNW	7 V	55	90	PEN	80	35	100
Spray^30	AgNW	100 mA	48	61	PET	120	15	100
Bar coating^15	AgNW	7 V	105	90	PET	100	10	45
Roll to roll^31	CNT/AgNW	15 V	105	95	PET	100	30	15
This work	AgNW	4–15 V	102	65–90	PET or glass	45	—	3

---

Inkjet printing technology, a direct-writing method, which can precisely control liquid deposition effectively for patterns, is commonly used to fabricate electronic devices and reported in the literature.²³ However, due to the nozzle-clogging problem caused by the large aspect ratio of AgNWs, only a few reports are available in the literature, which apply the inkjet printing process to pristine AgNW pattern fabrication. Wu et al.²⁴ first combined AgNWs with silver nanoparticles for inkjet printing, and the results showed that the addition of AgNWs can effectively decrease the sheet resistance of printed thin films. Later, Lu et al. printed AgNW thin films for organic photovoltaic devices.^25,26 Finn et al.²⁷ further improved the system using the single phase AgNW solution to fabricate a conductive thin film without high temperature sintering. These pioneering studies show the possibility to control the local resistance of 1-D nanomaterials such as AgNW. Based on these studies, a modified aqueous AgNW ink²⁸ was recently formulated with no additive for the inkjet printing process to fabricate transparent optoelectronic devices. In this study, with appropriate pattern designs and precise control of AgNW deposition, transparent thin film heaters have been inkjet-printed to demonstrate efficient heating for either uniform or graded temperature profiles. The characteristics of the printed heaters, such as heating power and response time, will be carefully examined to investigate the performance of the printed heaters.

Experimental

Materials

Ethanol and acetone were purchased from Sigma-Aldrich, USA. Deionized water (Millipore Milli-Q grade) with resistivity of 18.2 MΩ cm⁻¹ was used in this experiment. Silver nanowires (AgNWs, 0.5 wt%) in isopropanol, with diameters of 120 nm and lengths of 3–5 μm, was supplied by Emfutur, Spain. Flexible polyethylene terephthalate (PET) slides of 25 μm thickness were purchased from Nanya plastics, Taiwan. Glass slides of 1 mm thickness were purchased from FEA, Taiwan. Silver paste was purchased from Fujikura Kasei, Japan.

AgNW ink preparation

First, acetone was added to the AgNWs dispersion in a 1 : 1 volume ratio. The mixture was then put in a centrifuge (Z206A, Hermle) at 1500 rpm for 15 minutes for AgNW sedimentation. After centrifugation, the supernatant solvent was removed and the sediment dried slightly at room temperature. The wet AgNW sediment containing a little acetone was then mixed in deionized water to form a 0.8 mg mL⁻¹ aqueous solution. To obtain a stable AgNWs suspension, the AgNW ink was sonicated in an ultrasonic bath (DC300H, DELTA) for 30 minutes at 20 °C. The viscosity of the AgNW ink was 0.92 cP at 20 °C, measured by Brookfield DV-III Ultra Rheometer. Contact angles of the prepared AgNW ink, measured by an in-house goniometer, were 58° and 50° on glass and PET, respectively.

Fabrication and characterization of AgNW thin film patterns

Glass and PET substrates were first rinsed sequentially with water and ethanol and then dried in a vacuum oven at 60 °C. A piezoelectric inkjet printer (JetLab4, MicroFab, USA) was utilized to print the aqueous AgNW ink. Droplets of 70 μm in diameter were ejected consistently from a 50 μm nozzle at a velocity of 5 m s⁻¹. The substrates were placed on a heating stage with an operating temperature of 45 °C. After printing, the printed AgNW patterns, which dried on the heating stage, were connected with copper tapes by gluing with silver paste for electrical connection. Heating performances of the printed AgNW heaters were measured by an IR microscope (Ching Shing Computer-Tech, Taiwan). SEM images were obtained using a Nova NanoSEM 230 (FEI Company, USA). The electrical resistance of the printed patterns was measured by a multimeter (HOLA, DM-2690TU) with two-point mode. The UV spectra of the printed heaters were obtained by a UV spectrophotometer (DH-2000-BAL, USA).

Results and discussion

Resistance of AgNW conductive tracks

According to the Joule heating effect, the electrical resistance of the printed devices can directly determine the heating power. As shown in the literature,²⁹ the major resistance of AgNW thin films comes from the contact resistance between nanowires. Thus, with increasing AgNW densities, the better percolation in AgNW thin films can greatly reduce the electrical resistance. The relationship between the AgNW density and electrical resistance is thus very important to heater design. To precisely control the density of the printed patterns, an inkjet printing approach with multiple passes is used here. First, the resistances of straight lines, printed with a dot-to-dot spacing of 45 μm, were recorded (Fig. 1). Because of the low AgNW density (0.8 mg mL⁻¹), multiple printing layers are needed to create effective AgNW percolation for conductive straight lines. As shown in Fig. 1(a), after depositing 20 layers, the printed line has an average line width of 100 μm. The low concentration leads to a non-uniform thin film, as observed from the white clusters inside the printed thin films and some coffee ring effects around the line edges. Because of the small amount of AgNW, a layer of AgNWs with few contact points (Fig. 1(b)) is formed. Due to less effective contacts, the printed lines, which have a length of 1 cm, have a quite large resistance of around 2000 Ω cm⁻¹. As one increases the number of printing layers, the AgNW density increases and the printed lines show a more solid white colour under the microscope with nearly no coffee ring effects. Higher AgNW densities also lead to better nanowire percolation and thus the resistance decreases quickly to ∼300 Ω cm⁻¹ with 40 printing layers (Fig. 1(c)). As the number of layers keeps increasing, the SEM images (Fig. 1(b)) indicate that dense AgNW layers with large nanowire percolations are formed for those lines made by printing more than 60 layers. Thus, the resistance of the printed lines decreases slowly after 60 printing layers. This experiment shows that the electrical resistance of conductive patterns can be adjusted with the assistance of inkjet printing and one can easily fabricate patterns with different local resistances by controlling the deposited AgNW amounts.

Fig. 1 (a) Optical microscopy images of the printed AgNW straight lines on glass with increasing printing layers. (b) The SEM images of the printed lines in (a). The increase in AgNW density with increasing printing layers shows a better percolating network. (c) Relationship between printing layers and resistance of an AgNW track (1 cm long).

Mesh grid heaters

The inkjet printing process can be applied easily to digitally print conductive mesh grids to provide uniform heating over a certain area. In the literature, most of the AgNW-based thin film heaters are composed of a uniform AgNW layer with a low sheet resistance of less than 100 Ω sq⁻¹.¹⁵ From Joule's law and Ohm's law:one can expect that lower resistance can lead to better electro-thermal power. Moreover, because the AgNW percolating network allows light penetration, AgNW thin film heaters usually have high transparency ranging from 80% to 98% (ref. 1, 11, 12, 15, 30 and 31) (Table 1). However, printing AgNWs over an area with multiple passes can be quite time-consuming. Thus, to create a heater covering large areas with reasonable transparency, here a mesh grid is inkjet-printed to increase the heater transparency and show the capability of inkjet printing technology for well-defined printing patterns. As shown in Fig. 2, a mesh grid composed of 21 × 21 printed conductive tracks is printed over a 1 cm × 1 cm area on a glass slide with both good electrical conductivity and optical transparency. Because the grid structure has a large ratio of open area (about 64%), the transparency of the whole device can reach up to 90% (Fig. 2(b)) over the visible light range (450–750 nm) with low electrical resistance. With more layers printed, the transparency decreases slowly to 67% at 30 layers. This value is quite close to the open area ratio of 64% and shows the strong light shielding effect for higher printed AgNW amounts, as indicated by Finn et al.²⁷

Fig. 2 (a) Optical microscopy images of a printed AgNW mesh (20 layers) on a glass slide. Two sides of the printed mesh are connected to copper tape with silver paste. Variation of (b) optical transmittance and (c) electrical resistance with number of layers of printed AgNW meshes. All the meshes were composed of 21 lines (∼100 μm line width) with line spacings of 500 μm to cover a 1 cm × 1 cm area.

In contrast to single conductive tracks, the overall resistance of the printed mesh grid between the two silver paste electrodes decreases drastically due to the grid structure. Because the tracks parallel to the silver paste electrodes ensure the connection between the vertical tracks, the equivalent resistance R_eq of the printed mesh grid is 15 Ω with 20 layers, much lower than the resistance of a single track (2000 Ω) of 1 cm. The low resistance thus leads to a great heating power (Fig. 3). From the thermographs, the printed mesh heater with 20 layers of the AgNW ink can reach up to 93 °C with an input voltage of only 4 V, which is quite good compared to others (Table 1). The mesh heater also provides uniform heating over a circular area of ∼1 cm diameter, possibly due to the isotropic heat dissipation. The response time (∼200 s) for the heater to reach equilibrium temperature, however, is somewhat lower than those in the literature (Table 1). This slow response time is due to the large thermal capacity of the substrate and will be discussed in a later section.

Fig. 3 (a) Infra-red image of the AgNW mesh on a glass slide under 4 V input voltage for 400 s. The mesh grid heater, indicated by the square of dashed lines, was printed with 20 layers. (b) Transient temperature profiles at the central point (as indicated in (a)) of the AgNW mesh with different input voltages.

Gradient heater

In addition to heating uniformly within an area, one can also fabricate heaters with local or gradient heating effects. Fig. 4(a) shows a simple design that can yield a temperature gradient. The pattern consists of six 5 cm long parallel AgNW tracks with a line spacing of 0.5 mm on glass. In each track, different numbers of printing layers, from 30 to 70, are deposited in each centimetre length to create a gradient (or stepped) decrease in resistance. Because current through each single track remains the same for each zone, the highest temperature occurs at the location with highest resistance, wherein the heating power P = I²R is the highest (Fig. 4(b)). A large gradient of 10 °C cm⁻¹ can be generated with a 15 V input voltage; the temperature decreases quickly from 100 °C to 60 °C within 4 cm (Fig. 4(c)). By employing different voltages, the gradient of the temperature distribution can also be adjusted. At an applied voltage of 5 V, the temperature gradient decreases to 1.25 °C cm⁻¹. Based on these observations, one can also adjust the temperature gradient easily by varying the heating zone lengths (here 1 cm for each zone is used) or by manipulating the AgNW density difference in each zone. Computer simulations can also be used to help design the AgNW patterns for specific thermal gradients.

Fig. 4 (a) Pattern design of a gradient heater on glass: the local resistance was controlled by printing different layers of AgNW lines. (b) IR image of the AgNW gradient heater under 15 V input voltage after 400 s. (c) The temperature distribution along the printed tracks extracted from IR images.

Response time

The other important issue of a heater is the response time (t_res). A good heater should be able to reach a high temperature in a short time after applying electrical power. In the case of heaters on the glass slide, the heat dissipation is huge due to the large heat capacity of 1 mm thick glass. To heat up the glass slide underneath the heater, one can roughly estimate the timescale from energy balance between electrical power and heat loss into the substrate:

Pt_res ∼ ρC_pAtΔT (2)

where ρ is the density, C_p is the specific heat capacity, A is the heating area, t is the thickness of the substrate, and ΔT is the temperature difference at equilibrium. To reduce the heat loss in the substrate and further improve the response time, a thin PET substrate is used here. Using the physical properties given in Table 2 and eqn (2), the response time t_res can be reduced by roughly 28 times with the PET sheet, assuming the heating area A is the same. However, as shown in Fig. 5(d), the response of the same gradient heater on the PET sheet is quite fast and has a t_res ∼ 3 s, about 60 times faster than that on glass. This discrepancy happens simply because the thermal diffusion in glass is larger, and thus leads to a larger heating area. As shown in Fig. 5(b), the heating area on PET is nearly restricted to the area around the printed heater, due to its low thermal conductivity, whereas the heat can diffuse to the surrounding glass substrate within an area about 3 times larger. Therefore, the response time on glass is actually longer than expected using the designed heater area.

Table 2 Physical properties of substrates

	Density (kg m^−3)	Specific heat capacity (J kg^−1 K^−1)	Thickness (mm)	Thermal conductivity (W m^−1 K^−1)
Glass	2400	500	1	1.05
PET	1380	1250	0.025	0.18

---

Fig. 5 (a) and (b) IR image of the AgNW gradient heater on glass and PET less than 15 V input voltage. (c) and (d) P1 to P5 temperature profiles (cross-marked positions in (a) and (b)) of the AgNW gradient heater on glass and PET as a function of time at 15 V input voltage.

Heat dissipation

The diffusional heat transfer around the heaters on various substrates can be roughly estimated by comparing the thermal conductivity k. Because PET has a much lower thermal conductivity, i.e. PET is a better heat insulating material than glass, the heat diffusion is expected to be little cross-sectional directional (the y-direction in Fig. 6(a)). With an applied voltage of 5 V, the steady-state temperature across the heated glass slide decreases slowly from the heater printed at the centre and affects almost all the glass slide area (the slide has a width of 2.5 cm). In contrast, as depicted in Fig. 6(b), the temperature decreases sharply from the along the y-axis on PET and only affects ∼5 mm in width. As the applied voltage increases, the temperature profiles on PET become wider in the y-direction because of higher power input or higher heater temperatures. At the same time, the affected width of the heaters on glass slides remains almost the same regardless of the applied voltage, possibly due to it high thermal conductivity. On the other hand, along the x-axis of the printed heater, the temperature profiles of the heaters are nearly the same for both glass and PET (Fig. 6(c) and (d)), possibly because of the same power input and convective dissipation to the air at the steady-state. These observations shows the importance of the thermal conductivity of the substrates on the temperature distributions. Thus, to apply these heaters for a restricted patterned area, such as a microfluidic channel, one should select the substrate carefully according to its insulating ability.

Fig. 6 The temperature distribution along the y-axis for printed heaters on (a) glass slides and (b) PET sheets. The x-position of the sampling point is the same as point P1 in Fig. 5(a). The steady-state temperature distribution along the x-axis at different input voltages for (c) glass slides and (d) PET sheets. These line plot data are extracted from IR image analyses similar to those shown in Fig. 5(a) or 5(b).

Mechanical stability and critical applied voltage

The printed AgNW tracks on PET show high mechanical stability. As shown in Fig. 7, the resistance of the heater remains nearly the same after bending 500 times, but starts to slightly increase after 1000 times. Nevertheless, the resistance changes only by 8% after 5000 times, and the heating performance of the mesh heater remains functional with roughly the same temperature profiles in Fig. 3. The maximum operating voltage of the heater or the highest heater temperature, however, is determined by the nature of the AgNWs. Because the nanowires have a small radius, the melting temperature is much lower than that of bulk silver.³² As depicted in Fig. 8, when the AgNW thin films are heated, the AgNWs start to melt at around 200 °C, and the printed heaters can fail if the local temperature exceeds this critical value. For the heaters on flexible PET substrates, because of the low glass transition temperature (∼70 °C), this critical temperature is rarely reached. However, if a high temperature, larger than 150 °C, is required, this could be a constraint for AgNW-based heaters. Because the thin films are made of percolating nanowires, through which electric currents pass, the temperature of the AgNW can be higher than the measured or bulk temperature of the substrates. The narrow cross sectional areas of nanowires can yield high local current, and the resulting heat can lead to nanowire fusion. A typical example is demonstrated in Fig. 8(c). A mesh heater made by printing 10 layers is printed on a glass slide to provide uniform heating. The heater works well with an operating voltage of 10 V to provide a steady-state temperature of 70 °C. As one increases the voltage to 15 V for higher temperature requirements, the temperature increases to 120 °C in ∼200 s then suddenly decreases. After the failure, the heater shows a much larger resistance. The damaged heater was examined carefully under SEM and the microstructure shows fused AgNWs over several printed tracks (Fig. 8(d)). This solid evidence indicates that the real temperature of AgNWs can be much higher than the measured temperature from the IR microscope, which only provides a best resolution of ∼10 micrometers.

Fig. 7 (a) The setup of bending tests. (b) Durability test of AgNW mesh heater with different bending radii for 5000 bending cycles.

Fig. 8 SEM images for printed AgNW thin films (40 layers) heated on a hotplate for an hour at (a) 150 °C and (b) 200 °C. (c) Thermal response of a mesh heater (10 layers) on a glass slide at various applied voltages. (d) The SEM image of the damaged heater in (c) after applying 15 V for 500 s.

Conclusions

A simple strategy is developed to directly print conductive silver nanowire patterns for transparent thin film heaters. In contrast to other large-area deposition methods in the literature, the inkjet printing method provides an effective solution for heater designs, including local heating, gradient heating, and uniform heating. The number density of AgNWs per unit area can be precisely controlled, and the printed heaters possess both great optical transparency (up to 90%) and strong mechanical durability on a flexible substrate without post sintering treatment. By varying the printed AgNW density, a large temperature gradient up to 10 °C cm⁻¹ can be obtained. With careful selection of printing substrates, the printed heater can have a fast response (∼3 s) to the input voltages. Due to the nanoscale nature of AgNW, however, this approach can only provide heating temperatures lower than 200 °C. In summary, this study illustrates the feasibility of applying the inkjet printing process to produce arbitrary thin film patterns with 1-D nanowire materials, and the developed inkjet printing technology can be further integrated with other microfluidic applications.

Acknowledgements

The authors are grateful to Ministry of Science and Technology (MOST) of Taiwan for supporting this study.

References

1. X. Zhang, X. Yan, J. Chen and J. Zhao, Carbon, 2014, 69, 437–443.
2. S. Bae, H. Kim, Y. Lee, X. Xu, J.-S. Park, Y. Zheng, J. Balakrishnan, T. Lei, H. Ri Kim, Y. I. Song, Y.-J. Kim, K. S. Kim, B. Ozyilmaz, J.-H. Ahn, B. H. Hong and S. Iijima, Nat. Nanotechnol., 2010, 5(8), 574–578.
3. M. G. Kang, T. Xu, H. J. Park, X. Luo and L. J. Guo, Adv. Mater., 2010, 22(39), 4378–4383.
4. S. Pang, Y. Hernandez, X. Feng and K. Müllen, Adv. Mater., 2011, 23(25), 2779–2795.
5. S. Ji, W. He, K. Wang, Y. Ran and C. Ye, Small, 2014, 10(23), 4951–4960.
6. W. Hong, Y. Xu, G. Lu, C. Li and G. Shi, Electrochem. Commun., 2008, 10(10), 1555–1558.
7. T. J. Kang, T. Kim, S. M. Seo, Y. J. Park and Y. H. Kim, Carbon, 2011, 49(4), 1087–1093.
8. H.-S. Jang, S. K. Jeon and S. H. Nahm, Carbon, 2011, 49(1), 111–116.
9. J. Kang, H. Kim, K. S. Kim, S. K. Lee, S. Bae, J. H. Ahn, Y. J. Kim, J. B. Choi and B. H. Hong, Nano Lett., 2011, 11(12), 5154–5158.
10. H. Woo Kim, S. Jea Park, B.-K. Lee and D. Sung Kim, Appl. Phys. Lett., 2013, 102(10), 101907.
11. C. Celle, C. Mayousse, E. Moreau, H. Basti, A. Carella and J.-P. Simonato, Nano Res., 2012, 5(6), 427–433.
12. S. Wang, X. Zhang and W. Zhao, J. Nanomater., 2013, 2013, 6.
13. S. De, T. M. Higgins, P. E. Lyons, E. M. Doherty, P. N. Nirmalraj, W. J. Blau, J. J. Boland and J. N. Coleman, ACS Nano, 2009, 3(7), 1767–1774.
14. S. De, P. J. King, P. E. Lyons, U. Khan and J. N. Coleman, ACS Nano, 2010, 4(12), 7064–7072.
15. T. Kim, Y. W. Kim, H. S. Lee, H. Kim, W. S. Yang and K. S. Suh, Adv. Funct. Mater., 2013, 23(10), 1250–1255.
16. L. Daniel, G. Gaël, M. Céline, C. Caroline, B. Daniel and S. Jean-Pierre, Nanotechnology, 2013, 24(45), 452001.
17. T. Kim, A. Canlier, G. H. Kim, J. Choi, M. Park and S. M. Han, ACS Appl. Mater. Interfaces, 2013, 5(3), 788–794.
18. C. H. Liu and X. Yu, Nanoscale Res. Lett., 2011, 6(1), 75.
19. K.-C. Li, S.-T. Ding, E.-C. Lin, L. Wang and Y.-W. Lu, Biomicrofluidics, 2014, 8(6), 064109.
20. V. Miralles, A. Huerre, F. Malloggi and M.-C. Jullien, Diagnostics, 2013, 3(1), 33–67.
21. I. F. Yu, Y. H. Yu, L. Y. Chen, S. K. Fan, H. Y. E. Chou and J. T. Yang, Lab Chip, 2014, 14(18), 3621–3628.
22. A. Ohlander, C. Zilio, T. Hammerle, S. Zelenin, G. Klink, M. Chiari, K. Bock and A. Russom, Lab Chip, 2013, 13(11), 2075–2082.
23. S. P. Chen, H. L. Chiu, P. H. Wang and Y. C. Liao, ECS J. Solid State Sci. Technol., 2015, 4(4), P3026–P3033.
24. J.-T. Wu, S. Lien-Chung Hsu, M.-H. Tsai, Y.-F. Liu and W.-S. Hwang, J. Mater. Chem., 2012, 22(31), 15599.
25. H. Lu, J. Lin, N. Wu, S. Nie, Q. Luo, C.-Q. Ma and Z. Cui, Appl. Phys. Lett., 2015, 106(9), 093302.
26. L. Yang, T. Zhang, H. Zhou, S. C. Price, B. J. Wiley and W. You, ACS Appl. Mater. Interfaces, 2011, 3(10), 4075–4084.
27. D. J. Finn, M. Lotya and J. N. Coleman, ACS Appl. Mater. Interfaces, 2015, 7(17), 9254–9261.
28. S.-P. Chen, J. R. Durán Retamal, D.-H. Lien, J.-H. He and Y.-C. Liao, RSC Adv., 2015, 5(87), 70707–70712.
29. Z. Siwei, G. Yuan, H. Bin, L. Jia, S. Jun, F. Zhiyong and Z. Jun, Nanotechnology, 2013, 24(33), 335202.
30. S. Sorel, D. Bellet and J. N. Coleman, ACS Nano, 2014, 8(5), 4805–4814.
31. D. Kim, L. Zhu, D.-J. Jeong, K. Chun, Y.-Y. Bang, S.-R. Kim, J.-H. Kim and S.-K. Oh, Carbon, 2013, 63, 530–536.
32. B.-T. Liu and S.-X. Huang, RSC Adv., 2014, 4(103), 59226–59232.

---