LETTER

Inkjet printing of single-crystal films

doi:10.1038/nature10313

Hiromi Minemawari1, Toshikazu Yamada1, Hiroyuki Matsui1, Junya Tsutsumi1, Simon Haas1, Ryosuke Chiba1,2, Reiji Kumai1,3 & Tatsuo Hasegawa1

The use of single crystals has been fundamental to the development of semiconductor microelectronics and solid-state science1. Whether based on inorganic25 or organic68 materials, the devices that show the highest performance rely on single-crystal interfaces, with their nearly perfect translational symmetry and exceptionally high chemical purity. Attention has recently been focused on developing simple ways of producing electronic devices by means of printing technologies. Printed electronics is being explored for the manufacture of large-area and flexible electronic devices by the patterned application of functional inks containing soluble or dispersed semiconducting materials911. However, because of the strong self-organizing tendency of the deposited materials12,13, the production of semiconducting thin films of high crystallinity (indispensable for realizing high carrier mobility) may be incompatible with conventional printing processes. Here we develop a method that combines the technique of antisolvent crystallization14 with inkjet printing to produce organic semiconducting thin films of high crystallinity. Specifically, we show that mixing fine droplets of an antisolvent and a solution of an active semiconducting component within a confined area on an amorphous substrate can trigger the controlled formation of exceptionally uniform single-crystal or polycrystalline thin films that grow at the liquid air interfaces. Using this approach, we have printed single crystals of the organic semiconductor 2,7-dioctyl[1]benzothieno[3,2-b][1] benzothiophene (C8-BTBT) (ref. 15), yielding thin-film transistors with average carrier mobilities as high as 16.4 cm2 V21 s21. This printing technique constitutes a major step towards the use of high-performance single-crystal semiconductor devices for largearea and flexible electronics applications. Antisolvent crystallization is recognized as the best method of achieving controlled and scalable solidification, which is useful in pharmaceutical science14, for example. To achieve this, an antisolvent (a liquid in which a substance is insoluble) is added to the solution of the substance in a solvent that is miscible with the antisolvent. Here we make use of this concept in microliquid inkjet printing processes. A solution of a semiconductor and an antisolvent for the semiconductor are used as the two kinds of ink; the inks are individually printed at arbitrary positions to form a microliquid intermixture between the inks on the top of substrates. We found that optimized printing conditions enable controlled formation of patterned singlecrystal thin films having molecularly flat surfaces, in contrast to conventional inkjet printing processes that produce films with a non-uniform thickness distribution. This is a conceptual extension of the doubleshot inkjet printing process that was developed to produce films of charge-transfer compounds that are hardly soluble16,17. We used 1,2dichlorobenzene (DCB) as the solvent and N,N-dimethylformamide (DMF) as the antisolvent for the semiconductor C8-BTBT. These organic liquids show very different solubilities for C8-BTBT (the solubility at 20 uC is 400 times higher in DCB than in DMF), but have similar boiling points and are miscible with one another. A schematic representation of this printing process is shown in Fig. 1a. We used silicon wafers with 100-nm-thick silicon dioxide layers
1

as substrates. We produced the wetting/non-wetting surface patterning on the silicon dioxide layers by using a combination of ultraviolet/ozone treatment, hexamethyldisilazane treatment, and photoresist patterning18. We used a piezoelectric inkjet printing apparatus with double inkjet printing heads, from which a droplet of 60 picolitres is ejected at a repetition frequency of 500 Hz. In the process, the antisolvent ink (pure anhydrous DMF) is printed first and then overprinted with the solution ink (a 28 mM solution of C8-BTBT in DCB). In the formation of all the pieces of film shown in Fig. 1b, 42 shots of antisolvent ink were printed first and then 6 shots of solution ink were overprinted, all within a second. The deposited droplets are confined and intermixed in a predefined hydrophilic area on the upper surface of the substrate. During the initial stages of film formation, tiny floating bodies begin to form at the surface of the liquid and can be seen in microscope images (Supplementary Movie). Each floating body acts as a nucleus for further crystallization and undergoes subsequent growth to form a larger floating body. These bodies eventually cover the entire surface of the droplet (step 3 in Fig. 1a). A few creases can be seen on the surfaces of the droplets during liquid evaporation, indicating the solid nature of the films (step 4 and Supplementary Fig. 1)19. Although nuclei are generated randomly, mostly at the perimeters of the deposited droplet (solidliquidair interfaces), we found that nucleation can be controlled through appropriate design of the droplet configuration, which is shaped by the predefined hydrophilic area as well as by the ink volume. For example, a hydrophilic area containing a protuberance, as presented in Fig. 1b, was quite effective in causing local seeding of floating bodies in the protrusive area. We propose that local seeding is associated with the comparatively higher rate of solvent evaporation in areas with a high surface area-to-volume ratio. After seeding, the growing front moves slowly to the other end of the droplet until the large single-domain floating body covers the entire liquidair interface (see Supplementary Movie). The solvent then evaporates very slowly, taking about 1050 times longer than is the case without the solute, most probably because the droplet is completely covered by the solid film. During this slow evaporation, the creases in the films become smoothed out, and films with thickness of about 30200 nm are eventually obtained on the amorphous substrate. The film adheres tightly to the substrate. The morphology of the films as well as their single-domain nature depends on a variety of printing conditions, such as substrate temperature, the concentration and volume of the solution, the solutionantisolvent ratio and the shape of the hydrophilic area on which the droplet is deposited. The thickness profile of the film differs markedly from that of conventional inkjet printing deposits. Conventional inkjet printing is known to produce a characteristic thickness distribution in which both ends of the deposit are considerably thicker than its centre, known as the coffee-ring effect (see Supplementary Fig. 2)20. The uniform nature of the deposits produced by our process can be ascribed to temporal discrimination between solute crystallization and solvent evaporation within the deposited droplet (see Supplementary Fig. 1)16. The occurrence of supersaturation in the intermixed microliquid droplet results

National Institute of Advanced Industrial Science and Technology (AIST), AIST Tsukuba Central 4, Tsukuba 305-8562, Japan. 2Department of Applied Physics, The University of Tokyo, Hongo 113-8656, Japan. 3CMRC, Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), Tsukuba 305-0801, Japan. 0 0 M O N T H 2 0 1 1 | VO L 0 0 0 | N AT U R E | 1

2011 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER
a
b

8 mm

500 m 1 2 3 4 8 mm

c
b

[1 1 0]

Polarizer Analyser 100 m

20

40

15 10 5 0 60

Distance (m)

Figure 1 | Inkjet printing of organic single-crystal thin films. a, Schematic of the process. Antisolvent ink (A) is first inkjet-printed (step 1), and then solution ink (B) is overprinted sequentially to form intermixed droplets confined to a predefined area (step 2). Semiconducting thin films grow at liquidair interfaces of the droplet (step 3), before the solvent fully evaporates (step 4).

b, Micrographs of a 20 3 7 array of inkjet-printed C8-BTBT single-crystal thinfilms. c, Crossed Nicols polarized micrographs of the film. d, Expanded micrograph of the film, showing stripes caused by molecular-layer steps. e, Atomic-force microscopy image and the height profile (below) showing the step-and-terrace structure on the film surfaces.

in solute crystallization before solvent evaporation. In the microscope images of the films shown in Fig. 1d, we can see stripe-like features with intervals of several micrometres to several tens of micrometres. Atomic-force microscopy showed that the stripes are associated with the height of the molecular step, which is estimated to be about 2.6 2.8 nm (Fig. 1e)21. This value is consistent with the thickness of one molecular layer of C8-BTBT (csinb 5 2.92 nm, where c and b are the unit cell parameters) (ref. 22). We conclude that the stripe-like features are associated with the step-and-terrace structure of C8-BTBT. In images recorded through crossed Nicol prisms, the colour of almost the entire film changes from bright to dark, simultaneously, on rotating the film about an axis perpendicular to the substrate (Fig. 1c). In addition, when we use hydrophilic areas with different configurations such as a simple square, rectangle or circle, we obtained polycrystalline films composed of some crystal domains (see Supplementary Fig. 3). From these observations, we conclude that with appropriate design of both the droplet shape and printing conditions, single-domain crystal films that cover nearly the whole region of the printed deposits could be produced with high probability (Supplementary Fig. 4). We also noticed that the step-and-terrace structures in Fig. 1d form concentric ellipses, and propose that this feature is formed by epitaxial growth on top of thinner single-domain crystal films at a later stage (see Supplementary Fig. 5). X-ray diffraction data for the films are shown in Fig. 2a and b. The observed out-of-plane diffraction spots are consistent with a molecular layer structure that is parallel to the a and b axes. The observation of Bragg reflections up to 14th order indicates that the films have a highly crystalline nature. At high incident angles of the X-rays, we observed 16 diffraction spots that could be ascribed to Bragg reflections with indices that include an in-plane component (Fig. 2b), where the refined unit cellmonoclinic P21/a, a 5 5.91(15) A, b 5 7.88(1) A, c 5 29.12(19) A, b 5 91.0(8)u, V 5 1357(4) A3is consistent with that of the bulk crystal22. These results provide unambiguous evidence that the films are single-crystalline with a long-range translational symmetry. The data show that the growth direction is parallel to [1 21 0] in many (about 6070%) of the deposited films (Fig. 1c). On the other
2 | N AT U R E | VO L 0 0 0 | 0 0 M O N T H 2 0 1 1

X-ray c*
(0 0 2) (0 0 3)

= 5
(0 0 5) (0 0 7)

= 10
(0 0 12) (0 0 14)

= 80
(1 2 0) (1 1 1) (0 2 1) (1 1 2) (1 2 2)

= 85
(1 2 1) (1 1 0) (0 2 1) (1 1 1) (1 1 1) (1 2 1) (1 1 0) (1 3 1)

= 90
(1 1 1)

Height (nm)
(0 2 0)

45
10 m

(1 1 0)

c
LUMO

d
2 E//a 1 E//b 0 3 3.5 4 Photon energy (eV) a b b
HOMO

(105 cm1)

Figure 2 | Synchrotron-radiated single-crystal X-ray diffraction and polarized absorption spectra. Oscillation photographs for out-of-plane diffraction (a) and for high-incident-angle diffraction (b) of inkjet-printed C8BTBT single-crystal thin films, where v is the incident angle. The Bragg reflections observed in b correspond to the indices, which contain in-plane components. The refined unit cell obtained from the reflections is consistent with that of the bulk crystal. c, Polarized optical absorption spectra with coefficient a and with polarization parallel to the a and b axes in the singlecrystal film, demonstrating optical anisotropy with regard to these principal axes. d, View of the molecular arrangement of C8-BTBT in the crystal21.

2011 Macmillan Publishers Limited. All rights reserved

LETTER RESEARCH
hand, the bright-to-dark images observed through the crossed Nicol prisms originate from in-plane optical anisotropy of the single-crystal films. Figure 2c shows the polarized optical absorption spectra of our single-crystal films with the electric field of the light parallel to the a or the b axis. The spectra show a clear anisotropy in their absorption intensity and peak energy. The absorption intensity is much higher along the a axis and peaks at 3.43 eV, whereas the absorption intensity along the b axis is comparatively weak and peaks at a higher energy of 3.47 eV. We note that the transition dipole for the lowest electronic excitation between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) is polarized parallel to the molecular plane (Fig. 2c). The difference in the absorption intensity can be clearly ascribed to the orientation of the molecular planes within the ab plane. In contrast, the difference in peak energies is due to Davydov splitting along the a and b axes; this is characteristic of herringbone-type molecular arrangements within single-crystal films, as observed in anthracene23 or pentacene24. Field-effect devices were fabricated for the single-crystal films with a top-contact/top-gate geometry, composed of 30-nm Au films as the source and drain electrodes, and films of parylene C (capacitance per unit area of C 5 4.2 nF cm22) as the gate dielectric layers. The typical channel width and length were 145 mm and 100 mm, respectively. The direct-current field-effect characteristics at room temperature (300 K) were measured in an argon-filled glove box. The transfer and output characteristics of this device are shown in Fig. 3. The mobility in the saturation regime reaches 16.4 cm2 V21 s21 on average, and the maximum value is as high as 31.3 cm2 V21 s21. The on/off current ratio is 105107, and the subthreshold slope was about 2 V per decade with a threshold voltage of about 210 V. Injection barriers at the source/drain contacts may have remained, as manifested by the slightly nonlinear source/drain currentvoltage (Isd2Vsd) dependence at low voltages. Hardly any current hysteresis was observed in the transfer and output characteristics, where the shift in the threshold voltage from forward to reverse sweeps was less than 0.1 V. This
a
Source Drain Gate Parylene C

feature is probably associated with the negligible charge-trapping effects between the single-crystal surface of C8-BTBT and the parylene gate dielectric layer. The slope of the transfer curve (Fig. 3c) presents a distinct kink feature, as reported in other organic singlecrystal devices8, which clearly demonstrates the high quality of the semiconductorinsulator interface. We also found that the characteristics were not influenced by the existence of a few domain boundaries and were not degraded by more than 10% after the films were kept in air for 8 months. This device performance is much higher than the previous report for C8-BTBT21 and is comparable to the highest performance obtained (for a rubrene single-crystal device8). We consider that the following characteristics of the film formation process are important to achieve high-quality single-crystal film: (1) the liquidair interfaces need to be ideal locations for diffusion and self-organization of organic molecules (as for LangmuirBlodgett films25) and (2) the gradual growth of single-crystal films is only possible because of the fluidic nature of the microliquid droplet in which laminar flow dominates over turbulent flow26. The technique should be applicable to a broad class of functional soluble materials. The rather broad distribution of apparent mobility (Fig. 3b) indicates that further improvements of our technique should be possible, in areas such as ink composition, the optimization of equipment and the environment, and also subsequent device processing. For example, there is plenty of scope for improving the source/drain contacts. Nonetheless, we believe that this drop-on-demand, non-vacuum and room-temperature printing process of patterned single-crystal semiconductor films is in principle a useful new way of producing transistor arrays on top of plastic substrates, which is indispensable for realizing large-area, light-weight and high-speed electronic products.
Received 20 February; accepted 14 June 2011. Published online 13 July 2011. 1. 2. Teal, G. K. Single crystals of germanium and siliconbasic to the transistor and integrated circuit. IEEE Trans. Electron. Dev. 23, 621639 (1976). Tsui, D. C., Stormer, H. L. & Gossard, A. C. Two-dimensional magnetotransport in the extreme quantum limit. Phys. Rev. Lett. 48, 15591562 (1982). De Poortere, E. P. et al. Enhanced electron mobility and high order fractional quantum Hall states in AlAs quantum wells. Appl. Phys. Lett. 80, 15831585 (2002). Ohtomo, A. & Hwang, H. Y. A high-mobility electron gas at the LaAlO3/SrTiO3 heterointerface. Nature 427, 423426 (2004). Tsukazaki, A. et al. Observation of the fractional quantum Hall effect in an oxide. Nature Mater. 9, 889893 (2010). Sundar, V. C. et al. Elastomeric transistor stamps: reversible probing of charge transport in organic crystals. Science 303, 16441646 (2004). Briseno, A. L. et al. Patterning organic single-crystal transistor arrays. Nature 444, 913917 (2006). Takeya, J. et al. Very high-mobility organic single-crystal transistors with in-crystal conduction channel. Appl. Phys. Lett. 90, 102120 (2007). Yan, H. et al. A high-mobility electron-transporting polymer for printed transistors. Nature 457, 679686 (2009). Rivnay, J. et al. Large modulation of carrier transport by grain-boundary molecular packing and microstructure in organic thin films. Nature Mater. 8, 952958 (2009). Noh, Y.-Y., Zhao, N., Caironi, M. & Sirringhaus, H. Downscaling of self-aligned, allprinted polymer thin-film transistors. Nature Nanotechnol. 2, 784789 (2007). Whitesides, G. M. & Grzybowski, B. Self-assembly at all scales. Science 295, 24182421 (2002). Lim, J. A., Lee, H. S., Lee, W. H. & Cho, K. Control of the morphology and structural development of solution-processed functionalized acenes for high-performance organic transistors. Adv. Funct. Mater. 19, 15151525 (2009). Tung, H.-H., Paul, E. L., Midler, M. & McCauley, J. A. Crystallization of Organic Compounds: An Industrial Perspective 179205 (Wiley, 2009). Ebata, H. et al. Highly soluble [1]benzothieno[3,2-b]benzothiophene (BTBT) derivatives for high-performance, solution-processed organic field-effect transistors. J. Am. Chem. Soc. 129, 1573215733 (2007). Hiraoka, M. et al. On-substrate synthesis of molecular conductor films and circuits. Adv. Mater. 19, 32483251 (2007). Hasegawa, T., Hiraoka, M. & Yamada, T. Double-shot inkjet printing of donor acceptor-type organic charge-transfer complexes: wet/nonwet definition and its use for contact engineering. Thin Solid Films 518, 39883991 (2010). Sirringhaus, H. et al. High-resolution inkjet printing of all-polymer transistor circuits. Science 290, 21232126 (2000). Chen, Z., Liu, M., Liu, G.-y. & Tan, L. Evaporation induced two-dimensional buckling within liquid droplet. Appl. Phys. Lett. 95, 223104 (2009).
0 0 M O N T H 2 0 1 1 | VO L 0 0 0 | N AT U R E | 3

b
Counts

16 12 8 4 0 0 10 20 30 Mobility (cm2 V1 s1)

3. 4. 5. 6. 7. 8. 9.

20 Source

Counts

Si/SiO2

C8-BTBT film

15 10 5 0 104 105 106 107 On/off ratio 108

Drain

10. 11.

c
[Isd]1/2 (102 A1/2)

d
Vsd = 60 V
1 102 Isd (104 A) Isd (mA) 104 106 108 0 60 40 20 Vg (V) 0 20

1.5

Vg = 60 V
1 50 V 0.5 40 V 30 V 0 10 20 30 40 50 Vsd (V)

12. 13. 14. 15. 16. 17. 18. 19.

0.5

Figure 3 | Transistor characteristics for the inkjet-printed C8-BTBT singlecrystal thin films. a, Schematic of the device structure and micrograph of the thin-film transistors. b, Distribution of mobility and on/off ratio measured over 54 transistors. Average mobility is 16.4 6 6.1 cm2 V21 s21. c, Transfer characteristics at Vsd 5 260 V. d, Output characteristics at various gate voltages Vg.

2011 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER
20. Deegan, R. D. et al. Capillary flow as the cause of ring stains from dried liquid drops. Nature 389, 827829 (1997). 21. Uemura, T., Hirose, Y., Uno, M., Takimiya, K. & Takeya, J. Very high mobility in solution-processed organic thin-film transistors of highly ordered [1]benzothieno[3,2-b]benzothiophene derivatives. Appl. Phys. Expr. 2, 111501 (2009). 22. Izawa, T., Miyazaki, E. & Takimiya, K. Molecular ordering of high-performance soluble molecular semiconductors and re-evaluation of their field-effect transistor characteristics. Adv. Mater. 20, 33883392 (2008). 23. Pope, M. & Swenberg, C. E. Electronic Processes in Organic Crystals and Polymers 2nd edn, 5966 (Oxford Science, 1999). 24. Faltermeier, D., Gompf, B., Dressel, M., Tripathi, A. K. & Pflaum, J. Optical properties of pentacene thin films and single crystals. Phys. Rev. B 74, 125416 (2006). 25. Ulman, A. An Introduction to Ultrathin Organic Films, from Langmuir-Blodgett to SelfAssembly 1st edn (Academic, 1991). 26. deMello, A. J. Control and detection of chemical reactions in microfluidic systems. Nature 442, 394402 (2006). Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Acknowledgements We are grateful to Nippon Kayaku for providing C8-BTBT. We thank K. Takimiya and S. Horiuchi for discussions, H. Okamoto and H. Matsuzaki for help with optical measurements, K. Kobayashi for help with the X-ray measurements, and T. Iwadate for help with atomic-force microscopy and measurements of device characteristics. The synchrotron X-ray study was performed with the approval of the Photon Factory Program Advisory Committee (no. 2009S2-003). This work was supported by the New Energy and Industrial Technology Development Organization (NEDO) through a Grant for Industrial Technology Research and also by the Japan Society for the Promotion of Science (JSPS) through its Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST Program). Author Contributions H. Minemawari was responsible for ink fabrication, inkjet printing, microscopic observations, X-ray diffraction measurements, and measurements of the device characteristics of all the films. T.Y. prepared substrates with the wet/non-wet surface patterning, assisted in inkjet printing and X-ray diffraction measurements, and performed atomic-force microscopy and device characteristics measurements. H. Matsui guided sample preparation and inkjet printing, and conducted DFT molecular orbital calculations. J.T. assisted with X-ray diffraction measurements and performed optical anisotropic absorption measurements. S.H. assisted with optical anisotropic absorption measurements. R.C. assisted in the ink fabrication. R.K. assisted with X-ray diffraction measurements. T.H. conceptualized and directed the research project, and wrote most of the manuscript. All the authors discussed the results and commented on the manuscript. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of this article at www.nature.com/nature. Correspondence and requests for materials should be addressed to T.H. (t-hasegawa@aist.go.jp).

4 | N AT U R E | VO L 0 0 0 | 0 0 M O N T H 2 0 1 1

2011 Macmillan Publishers Limited. All rights reserved