Langmuir 2004, 20, 7789-7793
7789
Inkjet Printing of Well-Defined Polymer Dots and Arrays Berend-Jan de Gans and Ulrich S. Schubert* Laboratory of Macromolecular Chemistry and Nanoscience, Eindhoven University of Technology and Dutch Polymer Institute (DPI), P.O. Box 513, 5600 MB Eindhoven, The Netherlands
Downloaded via UNIV OF EDINBURGH on January 23, 2019 at 15:21:53 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Received March 1, 2004. In Final Form: June 9, 2004 Inkjet printing represents a highly promising polymer deposition method, which is used for, for example, the fabrication of multicolor polyLED displays and polymer-based electronics parts. The challenge is to print well-defined polymer structures from dilute solution. We have eliminated the formation of ring stains by printing nonvolatile acetophenone-based inks on a perfluorinated substrate using different polymers. (De)pinning of the contact line of the printed droplet, as related to the choice of solvent, is identified as the key factor that determines the shape of the deposit, whereas the choice of polymer is of minor importance. Adding 10 wt % or more of acetophenone to a volatile solvent (ethyl acetate)-based polymer solution changes the shape of the deposit from ring-like to dot-like, which may be due to the establishment of a solvent composition gradient. Arrays of closely spaced dots have also been printed. The size of the dots is considerably smaller than the nozzle diameter. This may prove a potential strategy for the inkjet printing of submicrometer structures.
Introduction Inkjet printing is considered one of the most promising methods for controlled deposition of polymers and functional materials,1,2 especially in relation to the fabrication of multicolor polyLED displays and polymer-based electronics parts. Such devices cannot be prepared by conventional spin coating. Instead, a microscopic patterning technique is needed. Inkjet printing has the advantage of simplicity, low cost, flexibility, and maturity. Outstanding examples include inkjet printed all-polymer transistors3-5 and 130 ppi RGB polyLED displays.6,7 However, the potential of inkjet printing is huge and goes far beyond what has been realized up to now. Future areas of application may include the fabrication of sensors, polymerbased solar cells, or microfluidics devices, all based on polymers as matrix or functional materials. In the field of combinatorial materials science, inkjet printed libraries of polymer dots could bridge the gap between parallel synthesis and property characterization. As inkjet printing requires low viscosities (typically 1-10 mPa s), polymeric materials can only be deposited from dilute solution.8 However, a drying droplet usually deposits its solute as a ring stain that marks the perimeter of the droplet before drying. Ring formation, commonly known as the “coffee drop effect”, is due to the combined * Corresponding author. Fax:
[email protected].
0031-40-247-4186. E-mail:
(1) De Gans, B. J.; Duineveld, P. C.; Schubert, U. S. Adv. Mater. 2004, 16, 203. (2) Calvert, P. Chem. Mater. 2001, 13, 3299. (3) Sirringhaus, H.; Kawase, T.; Friend, R. H.; Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E. P. Science 2000, 290, 2123. (4) Stutzmann, N.; Friend, R. H.; Sirringhaus, H. Science 2003, 299, 1881. (5) Paul, K. E.; Wong, W. S.; Ready, S. E.; Street, R. A. Appl. Phys. Lett. 2003, 83, 2070. (6) Funamoto, T.; Matsueda, Y.; Yokoyama, O.; Tsuda, A.; Takeshita, H.; Miyashita, S. Proc. 22nd Int. Display Res. Conf., Boston 2002; Society for Information Display: San Jose, 2002; p 899. (7) Giraldo, A.; Duineveld, P. C.; Johnson, M. T.; Lifka, H.; Rubingh, J. E. J. M.; Childs, M. J.; Fish, D. A.; George, D. S.; Godfrey, S. D.; McCulloch, D. J.; Steer, W. A.; Trainor, M.; Young, N. D.; Hunter, I. M. Proc. 7th Asian Symp. Information Display (ASID2002), Singapore 2002; Society for Information Display: San Jose, 2002; p 43. (8) De Gans, B. J.; Kazancioglu, E.; Schubert, U. S. Macromol. Rapid Commun. 2004, 25, 292.
action of an increased evaporation rate at the droplet edge, and contact line pinning due to surface irregularities and solute deposition (“self-pinning”).9-11 A capillary-driven flow from the droplet center toward the edge compensates for evaporation losses and transports most of the solute toward the contact line. All previously mentioned applications of inkjet printing require well-defined, often dot-like deposits. Therefore, the elimination of ring formation is of great practical interest. To improve the homogeneity of the deposits, mixtures of low-boiling good solvents and high-boiling bad solvents for the polymer under consideration were used. During evaporation, the solvent quality decreases, and the polymer will precipitate.12 A disadvantage is the poor (mechanical) properties of the deposit formed. It was shown that when a single solvent was used, the homogeneity increases with decreasing rate of evaporation.13 This could be due to the time scale for depinning of the contact line matching the time scale of evaporation. Polyvinylcarbazole dots having a Gaussian shape were inkjet printed, although the optical micrograph provided clearly shows irregularities, probably due to stick-slip (de)pinning of the contact line during evaporation.14 ITO was used as the substrate. The velocity of the print head and the evaporation conditions are also known to affect the shape of the deposit.15 Casting experiments with macroscopic, aqueous drops have shown that ring formation can be (partially) eliminated by the use of a hydrophobic substrate, that is, a film of castor oil on glass.16 In the field of inkjet printing, use (9) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Nature 1997, 389, 827. (10) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Phys. Rev. E 2000, 62, 756. (11) Deegan, R. D. Phys. Rev. E 2000, 61, 475. (12) Lyon, P. J.; Carter, J. C.; Bright, J. C.; Cacheiro, M. WO Patent 02/069119 A1, 2002. (13) Madigan, C. F.; Hebner, T. R.; Sturm, J. C.; Register, R. A.; Troian, S. MRS Symp. Proc. Vol. 625; Materials Research Society: Warrendale, 2000; p 123. (14) Hebner, T. R.; Wu, C. C.; Marcy, D.; Lu, M. H.; Sturm, J. C. Appl. Phys. Lett. 1998, 72, 519. (15) Shimoda, T.; Morii, K.; Seki, S.; Kiguchi, H. MRS Bull. 2003, 28, 821. (16) Takhistov, P.; Chang, H.-C. Ind. Eng. Chem. Res. 2002, 41, 6256.
10.1021/la049469o CCC: $27.50 © 2004 American Chemical Society Published on Web 07/30/2004
7790
Langmuir, Vol. 20, No. 18, 2004
de Gans and Schubert
Table 1. Solvent Properties at Room Temperature liquid
Tb (°C)
Pv (mmHg)
θc (deg)
γLV (mN m-1)
ethyl acetate anisole acetophenone
77.1 153.7 202
76 3.51 0.35
37.4 ( 2.2 57.2 ( 1.3 69.1 ( 1.4
23.39 35.10 39.04
was made of hydrophobic/hydrophilic patterns to control the position of inkjet printed droplets,1,15,17 but not yet for influencing the shape of the deposit. In this contribution, we will demonstrate how homogeneous deposits of polymeric material can be obtained via use of a hydrophobic, perfluorinated surface and a single solvent rather than a mixture of a good and a bad solvent. In addition, we will show that the use of a mixture of two good solvents for the polymer allows even further control over the shape of the deposit and that arrays of dots can be inkjet printed. Experimental Section Materials. Polydisperse polystyrene (N5000, Shell Nederland, Den Haag, The Netherlands; Mn 80 kD, Mw 282 kD) was used. Poly(methyl methacrylate) was purchased from Sigma-Aldrich (Steinheim, Germany; Mn 130 kD, Mw 240 kD). Monodisperse polystyrenes were purchased from Polymer Standards Service (Mainz, Germany; Mn/Mw 17/17.5, 30/34, 60/64, 120/125, 214/ 226 kD). Ethyl acetate (Biosolve, Valkenswaard, The Netherlands), anisole (Acros Organics, Geel, Belgium), and acetophenone (Sigma-Aldrich, Steinheim, Germany) were used as solvents, or mixtures thereof. Solvent properties are listed in Table 1. Solutions used for printing contained 1.0% polymer by weight. Viscosities were measured with an Ubbelohde viscometer at 20.0 °C. Despite the high molecular weight, these solutions could easily be printed without persistent filament formation.8 Substrate Preparation. Glass slides (Marienfeld, LaudaKo¨nigshofen, Germany) were ultrasonicated in acetone for 5 min, rubbed with sodium dodecyl sulfate solution, ultrasonicated in sodium dodecyl sulfate solution for 5 min, flushed with demineralized water to remove the soap, treated with 2-propanol vapor to remove the water in a reflux setup, dried with a flow of nitrogen, and subsequently treated in a UV-ozone photoreactor (PR-100, UVP, Upland, CA) for 30 min to remove any remaining organic contamination. Substrates were used immediately after cleaning. Alternatively, glass slides were coated with (tridecafluoro-1,1,2,2-tetrahydro-octyl)trichlorosilane (ABCR, Karlsruhe, Germany),18 resulting in strongly hydrophobic surfaces. Contact angles were measured with an OCA30 device from Dataphysics (Filderstadt, Germany). The results are listed in Table 1. Inkjet Printer. A microdrop Autodrop device (Norderstedt, Germany) was used, consisting of an automated XYZ-stage in combination with a AD-K-501 micropipet print head.19 The printer offers a workspace of 200 × 200 × 80 mm. The positioning accuracy of the print head is 3 µm. The diameter of the micropipet nozzle is 70 µm. The Autodrop is a piezoelectric inkjet printer. Droplet ejection occurs via contraction of a piezoelectric actuator that generates an acoustic wave. The signal that drives the actuator is a rectangular pulse, with an amplitude of 80 V, and a width of 29 µs. The ejection frequency is 200 Hz, and the print head velocity is 5 mm s-1. Variable amounts of material can be deposited by changing the number of droplets that are printed at a spot. The inkjet printer produces highly uniform droplets with a typical radius error of less than 1%. Unless stated otherwise, all samples were printed as a 2 × 10 array, with 1.00 mm spacing between the dots. Each dot is the result of five drops that coalesce on the surface to form one large drop. Surface Topography. A white light confocal microscope (Nanofocus µSurf, Duisburg, Germany) was used to determine (17) Wang, J. Z.; Zheng, Z. H.; Li, H. W.; Huck, W. T. S.; Sirringhaus, H. Nat. Mater. 2004, 3, 171. (18) Trimbach, D.; Feldman, K.; Spencer, N. D.; Broer, D. J.; Bastiaansen, C. W. M. Langmuir 2003, 19, 10957. (19) microdrop GmbH, Norderstedt, Germany; http:// www.microdrop.de.
Figure 1. Drying pattern of an inkjet printed droplet of a 1 wt % solution of polystyrene in acetophenone on clean glass. the surface topography. The setup uses an external 100 W xenon cold light source and a 20× objective.
Results and Discussion Figure 1 shows a typical drying pattern of a drop (i.e., consisting of 5 droplets printed on top of each other) of a 1.0 wt % solution of polystyrene (Mn 80 kD; Mw 282 kD) in acetophenone on an uncoated, cleaned glass substrate. The viscosity of the solution is 3.5 mPa s. Despite its high boiling point, the droplet dries within a few seconds, due to the high surface-to-volume ratio. The pattern consists of a series of rings that originate from the increase in surface energy when drying with the contact line fixed. Eventually, the droplet will contract in a stick-slip manner, producing the pattern.11 Changing the solvent does not improve homogeneity. Figure 2a shows the deposit that is left by the same solution of polystyrene in acetophenone on a perfluorinated glass slide. After being dried, a regular polystyrene dot is formed. Figure 2b shows cross-sections of the dot in the x- and y-directions. Its diameter is 86 µm, and its height 3.0 µm. The dot is shaped as a volcano; that is, it has a small crater on top. The experiment was repeated using a 1.0 wt % solution of poly(methyl methacrylate) (Mn 130 kD; Mw 240 kD) in acetophenone, which had a viscosity of 3.9 mPa s. In this case, the diameter of the dot is 75 µm, and its height 4.1 µm; that is, the dimensions of the polystyrene and the poly(methyl methacrylate) deposits and their shape are similar. This implies that the dominant factor that determines the deposition process is the solvent and its physical properties, rather than the solute. To check this assumption, polystyrene solutions were printed using anisole as a solvent. Anisole (methoxybenzene) has a boiling point and a vapor pressure that are considerably lower and higher, respectively, than those of acetophenone. The anisole dot is similar to the acetophenone dot, although the crater is more prominent. The shape of the deposit changes completely when using a low-boiling solvent like ethyl acetate. Figure 3a shows the deposit that is left by a drop (i.e., 5 printed droplets) of a 1 wt % solution of polystyrene in ethyl acetate. This solution has a viscosity of 0.68 mPa s. A ring stain that marks the perimeter of the original drop is formed. The cross-section displayed in Figure 3b shows that indeed most of the polymeric material is present at the ring rather than at the center. In principle, this effect can be used to
Inkjet Printing of Well-Defined Polymer Dots
Figure 2. (a) Volcano-shaped polymer dot, formed by an inkjet printed droplet of a 1 wt % solution of polystyrene in acetophenone on perfluorinated glass; (b) cross-sections in the x- and y-directions.
Figure 3. (a) Polymer dot, formed by a droplet of a 1 wt % solution of polystyrene in ethyl acetate on perfluorinated glass; (b) cross-sections in the x- and y-directions.
print thin rings or lines with a resolution better than the limit that is set by the nozzle diameter.20 Drop-casting experiments showed that on the perfluorinated surface the contact line of a droplet of this solution of polystyrene in ethyl acetate is pinned, whereas the contact lines of acetophenone and anisole solution droplets are at least initially unpinned. This may be due to self(20) Cuk, T.; Troian, S. M.; Hong, C. M.; Wagner, S. Appl. Phys. Lett. 2000, 77, 2063.
Langmuir, Vol. 20, No. 18, 2004 7791
Figure 4. (a) Polymer dot, formed by a droplet of a 1 wt % solution of polystyrene in an 80/20 wt % ethyl acetate/ acetophenone mixture on perfluorinated glass; (b) cross-sections in the x- and y-directions.
pinning of the ethyl acetate droplet: The vapor pressure Pv of ethyl acetate, to which the rate of evaporation is proportional, is 1 order of magnitude larger than the vapor pressure of anisole, which is again 1 order of magnitude larger than the vapor pressure of acetophenone (see Table 1). A ring-like deposit that pins the contact line is therefore more easily formed. In addition, the contact angle θc of ethyl acetate is smaller than the contact angle of anisole, which is again smaller than the contact angle of acetophenone. Near the contact line, the evaporation flux J is proportional to (R-r)-λ, with λ ) (π - 2θc)/(2π - 2θc), where r is the distance from the edge toward the center of the droplet and R is its radius. λ increases with decreasing θc, and therefore the flux.9 Viscosity effects cannot explain the observed behavior. The viscosity of the acetophenone solution is considerably higher than the viscosity of the ethyl acetate-based solution. A large acetophenone dot as compared to ethyl acetate is therefore expected, but the opposite is found. The effect of mixing the two solvents is particularly interesting. A typical result, obtained with a mixture of 80% ethyl acetate and 20% acetophenone by weight, is shown in Figure 4a. Figure 4b displays the corresponding cross-sections. Again, a regular polystyrene dot is formed, with a height of 6.2 µm, that is, twice as high as the pure acetophenone dot, and with a diameter of 68 µm. Assuming a cylindrical shape, it follows that the total amount of deposited material must be equal in both cases, as it should, as both concentration and number of deposited droplets are equal in both cases. A striking difference between the pure acetophenone deposit and the deposit left by the mixture is the absence of a crater on top in the case of the latter. Perhaps most surprising is the fact that a small amount of acetophenone already suffices to change the shape of the deposit from ring-like to dot-like, that is, to depin the contact line. Dots, identical to that shown in Figure 4a, were already obtained with mixtures containing 10% acetophenone by weight.
7792
Langmuir, Vol. 20, No. 18, 2004
de Gans and Schubert
Figure 5. Dot height as a function of polymer molecular weight, using monodisperse polystyrene standards. The inset shows the droplet mass (0) and the specific viscosity (2) as a function of molecular weight. A power law was fitted to the viscosity data, the exponent of which is 0.77.
To understand the origin of this phenomenon, the contact angle of the 80/20% w/w ethyl acetate/acetophenone solution mixture on the perfluorinated surface was measured. It was found that the (initial) contact angle θc (t ) 0) is 41.5 ( 1.0°, which is very close to the value of pure ethyl acetate, that is, 37.4°. This proves that the results cannot be explained by a change in contact angle. We then realized that the process of droplet evaporation and solute deposition depends on the composition of the droplet at the contact line rather than the overall composition of the droplet. When using a mixture of a low- and a high-boiling solvent, the composition at the contact line will shift toward a higher fraction of highboiling solvent than in the bulk, due to the increased rate of evaporation at the edge. Therefore, the rate of evaporation at the contact line decreases, and a surface tension gradient is established. A flow will be induced from regions with low to regions with high surface tension when the Marangoni number M ) ∆γL/ηD is sufficiently large.21 Here, ∆γ denotes the surface tension difference, L is the length scale involved, η is the viscosity, and D is the diffusion coefficient. Taking for ∆γ the difference between the bulk values of ethyl acetate and acetophenone, and using typical values for L, η, and D, it follows that M is of the order 106. This result shows that even a very small concentration gradient will be sufficient to cause a Marangoni flow, which will homogenize the liquid droplet and reduce the concentration gradient between the contact line and the bulk.16 To gain insight into the process of dot formation, we studied the height of the dots as a function of molecular weight of the dissolved polymer, using solutions containing 1.0% monodisperse polystyrene by weight in acetophenone. One single substrate was used on which all solutions were printed as 1 × 10 arrays. In addition, the reduced viscosities of the polymer solutions were measured, and their contact angles on the fluorinated substrate were used. The mass of the droplets as generated by the inkjet printer was measured as follows: The total amount of printed material was collected during a certain period of time, weighed, and divided by the number of droplets. The results are shown in Figure 5. It should be emphasized that, although the errors in the dot height are generally small (1-3.5%), the errors between different surfaces are much larger (up to 30%). As for the three highest molecular weight samples, the dot height shows very little variation (21) Pesach, D.; Marmur, A. Langmuir 1987, 3, 519.
Figure 6. (a) Array of polymer dots printed at a mutual distance of 150 µm, formed by droplets of a 1 wt % solution of polystyrene in acetophenone. The typical size of the dots shown here is 29 ( 2 µm, which is smaller than the nozzle diameter; (b) crosssection in the x-direction, corresponding to the line in Figure 6a.
with molecular weight. However, a jump in dot height occurs between 34 and 64 kD. The inset of Figure 5 shows that the mass of the droplets decreases with molecular weight. The specific viscosity, also shown in the inset, increases with molecular weight. A power law, fitted to these data, yields [η]c ∝ Mw0.77, in perfect agreement with the prediction of Zimm for polymer chains in a good solvent.22 Finally, it was found that within the experimental error the presence and the molecular weight of the polymer do not affect the contact angle. It is expected that dot height decreases with viscosity as the dot diameter increases. Furthermore, it is expected that the dot height scales linearly with droplet size, that is, with the cube root of the droplet mass, which also decreases with molecular weight. However, the reason for the discontinuous decrease of droplet height with molecular weight remains unclear. For most applications, the formation of dot arrays will be of interest. Figure 6a shows a dot array that was obtained by printing a rectangular array of single, acetophenone-based droplets at a mutual distance of 150 µm on a perfluorinated surface. Figure 6b displays a corresponding cross-section. Here, the average dot diameter is 29 µm, and no craters can be observed. This could be due to different evaporation conditions when printing an array of closely spaced droplets. Similar results were obtained by Shimoda et al., who studied shape differences of polyLED dots: Crater formation decreased with increasing velocities.15 The effect was attributed to partial drying before the droplet hits the surface. However, Duineveld et al. have shown that droplet mass reduction during flight is negligible, assuming typical values for traveling distance, droplet size, and solvent vapor pressure.23 Both our and Shimoda’s results can be explained (22) Doi, M.; Edwards, S. F. The Theory of Polymer Dynamics; Oxford University Press: Oxford, 1986; p 144.
Inkjet Printing of Well-Defined Polymer Dots
assuming a correlation between rate of evaporation and dot shape: Slow evaporation occurs when the time for printing is small as compared to drying times (i.e., at high printing velocities), as the droplets evaporate in an atmosphere with a high partial solvent pressure, due to evaporation of neighboring droplets. This prevents crater formation. Minimum spacing at which dots can be printed without coalescence of the droplets on the surface is at present about 125 µm. It is seen from Figure 6a that the deposition accuracy is far worse than its theoretical limit, which is set by the positioning accuracy of the print head, that is, 3 µm. This is probably due to random motion of the droplets during evaporation. Confinement of the droplets, either mechanically or by surface energy patterning, will be necessary to produce regular arrays. Finally, it is interesting to note that, by using dilute solutions, structures can be created of size smaller than the diameter of the nozzle (29 vs 70 µm). This may prove to be a potential strategy for the inkjet printing of defined submicrometer structures. Conclusions We have inkjet printed well-defined (arrays of) polymer dots on a hydrophobic, perfluorinated substrate using a nonvolatile solvent such as acetophenone. Different polymers yield qualitatively identical deposits, suggesting that the choice of the solvent may be the key factor determining the deposition process. When a volatile solvent such as ethyl acetate is used, ring-like deposits are formed. Drop-casting experiments show that the (23) Duineveld, P. C.; De Kok, M. M.; Buechel, M.; Sempel, A. H.; Mutsaers, K. A. H.; Van de Weijer, P.; Camps, I. G. J.; Van den Biggelaar, T. J. M.; Rubingh, J.-E. J. M.; Haskal, E. I. Proc. SPIE Vol. 4464; SPIE: Bellingham, 2002; p 59.
Langmuir, Vol. 20, No. 18, 2004 7793
contact line of acetophenone droplets on the surface used is unpinned, whereas the contact line of ethyl acetate droplets is not, which may explain the results. Mixing a small amount of acetophenone, that is, 10 wt % or more, with ethyl acetate changes the shape of the deposit from ring-like to dot-like, probably due to a local increase of the acetophenone concentration at the contact line. Investigation of the dot height as a function of molecular weight using monodisperse polystyrene samples showed a discontinuous decrease with molecular weight. Simultaneously, the mass of the ejected droplets decreases with molecular weight, whereas the solution viscosity increases and the contact angle of the droplets remains constant. Thus, none of these three factors provides an explanation for the observed behavior. Finally, arrays of dots were printed. The size of the dots that are printed is considerably smaller than the nozzle diameter. This may represent a potential strategy for the inkjet printing of submicrometer structures. Acknowledgment. We thank Carlos Sanchez for his help in preparing perfluorinated substrates, Caroline Abeln for the GPC measurements, Prof. Bert de With (Laboratory of Coatings Technology, TU/e) for the use of the confocal scanning microscope, Niek Lousberg and Alper Tiftikci for sharing their expertise, and Wilhelm Meyer from Microdrop for fruitful collaboration. This work is part of the Dutch Polymer Institute (DPI) research program (Project 400). Supporting Information Available: Scanning confocal micrograph of poly(methyl methacrylate) dot and corresponding cross-sections. Scanning confocal micrograph and corresponding cross-sections of polystyrene dot, printed from anisole solution. This material is available free of charge via the Internet at http://pubs.acs.org. LA049469O
