Inkjet Printing of Reinforcing Patterns for the Mechanical Stabilization

Article pubs.acs.org/Langmuir

Inkjet Printing of Reinforcing Patterns for the Mechanical Stabilization of Fragile Polymeric Microsieves Jens Hammerschmidt,† Franziska M. Wolf,‡ Werner A. Goedel,*,‡ and Reinhard R. Baumann*,†,§ †

Institute for Print and Media Technology, Chemnitz University of Technology, Reichenhainer Str. 70, 09126 Chemnitz, Germany Physical Chemistry, Chemnitz University of Technology, Strasse der Nationen 62, 09111 Chemnitz, Germany § Printed Functionalities, Fraunhofer ENAS, Technologie-Campus 3, 09126 Chemnitz, Germany ‡

ABSTRACT: Inkjet printing is employed to apply a mechanically stable reinforcing pattern to polymeric microsieves prepared by float casting, where particles are used as molds for the pores. A mixture of silica particles and nonvolatile monomers is cast onto a water surface and subsequently photopolymerized to produce membranes consisting of a polymer film with embedded particles. These composite membranes are transferred onto an aluminum foil. Subsequently, a UV-curable ink is directly inkjet-printed onto the membranes in line patterns of grids or honeycombs and cured by UV radiation to create a mechanically reinforcing pattern. Afterwards, the particles and the aluminum foil are removed by chemical etching. The reinforcing pattern overcasts 40% of the previously manufactured membrane, is mechanically stable, and gives the microsieves such a robustness that they can be handled in further manufacturing processes.

■

particular manufacturing method of the microsieves.1,14,16 For instance, by employing lithographic processes,1,7,8 initially a thin layer of silicon nitride is deposited on a thicker silicon wafer by chemical vapor deposition; after the manufacture of a microsieve out of silicon nitride, the silicon wafer is structurally etched in such a manner that a grid of support bars is generated which functions as a reinforcing structure for the microsieve. In the case of microsieves prepared purely from photoresist, it is as well possible to generate a support structure by repeated application and lithographic structuring of layers of the photoresist.16 All these lithographic methods involve premanufactured masks, high quality optical systems, and subtractive processes with light exposure and development steps to obtain the desired reinforcing patterns. In the approach reported here, we employ instead inkjet printing22−24 to prepare an overlaying macroporous reinforcing pattern directly on float cast polymeric membranes introducing steps IV and V into the process described in Figure 1. This new process provides the following advantages: (a) The method does not apply impact, pressure or extensional stress to the microsieve. (b) The liquid nature of the ink guarantees intimate contact and good adhesion of the reinforcing material to the microsieve. (c) The patterns manufactured by inkjet printing are based upon digital images, which can be adapted to specific requirements of applications. (d) The patterns are deposited in an additive manner only at locations where material is required. Thus, no premanufactured masks and development steps are required.

INTRODUCTION Microsieves are permeable membranes densely interspersed with uniform pores with a thickness smaller than the pore diameter.1 Compared to track etched membranes and tortuous path membranes, these features result in decreased flow resistance and sharp size selectivity for particles in filtration applications.1−6 The early microsieves were manufactured by lithography.1,7,8 In the meantime, various techniques have been developed and optimized to manufacture both inorganic and organic microsieves with pore diameters in the range of nanoand micrometers.9−16 In previous publications we utilized the principle of particle-assisted wetting17−19 to follow an alternative manufacturing strategy based on float casting,20,21 basically pursuing the steps I to III and VI of Figure 1: A dispersion consisting of nonvolatile monomers, particles, and volatile solvents is spread onto a water surface. There the particles arrange in a densely packed layer with a thickness of one particle. The space between the spherical particles is filled with the monomer, which is subsequently cross-linked photochemically to form a solid nonpermeable composite membrane containing the embedded particles. The particles serve as molds for the pores in the manufactured microsieves and have to be removed to achieve permeability. However, because of their thickness in the range of the pore diameter, especially microsieves with submicrometric pores inherently exhibit a very low resistance against mechanical stresses, which can be avoided by mechanically reinforcing patterns. An obvious option is to transfer the microsieves to macroporous grids, but this entails the danger of damage due to mechanical stresses during the transfer. Furthermore, various approaches exist, which are however closely related to the © 2012 American Chemical Society

Received: July 4, 2011 Revised: September 7, 2011 Published: January 31, 2012 3316

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321

Langmuir

Article

Figure 1. Mechanical stabilization of fragile microsieves by applying a reinforcing pattern by inkjet printing.

As we have demonstrated recently,25 inkjet technology can be used for the stabilization of thin nonporous polymer films, utilizing an ink containing a substantial fraction of a volatile solvent requiring the deposition of several layers of ink on top of each other to obtain an appropriate mechanically reinforcing layer thickness. To overcome the drawback of multilayer printing, we substituted in the current approach the previous ink system by a UV-curable ink and integrated the printing procedure directly into the manufacturing process of the microsieves described in Figure 1. One important aspect of the reported approach was the minimization of the area fraction covered by the reinforcing pattern but achieving a maximum of the reinforcing effect. Therefore, we designed the reinforcing patterns as a composite of narrow lines, whose exact printing requirements are described by Soltman26 and Stringer.27

■

of spherical particles of diameter D multiplied by the density of a particle, ρ (for employed silica particles,28 ρ = 1.9 g/cm3), and its volume πD3/6 and divided by the mass fraction, ω, of the particles in the suspension: M = πAρD/(3√3ω). After evaporation of the solvents, the TMPTMA in this layer was cross-linked by exposure to UV−C radiation from a low pressure mercury vapor lamp (Umex, Dresden; radiant power: 3.5 mW/cm2; exposure time: 45 min). Then the Petri dish with the resulting nonpermeable polymer composite membrane was placed in a bigger reservoir of water so that the membrane area was floating on a larger water surface area. An aluminum foil was positioned manually at close distance below the floating membrane in such a way that it was parallel to it and then cautiously lifted out of the water with a tweezers, finally resulting in a transfer of the composite membrane to the foil. To manufacture the reinforcing patterns, the UV-curable ink Crystal UFX 7683 (SunJet) was deposited by using a Dimatix Materials Printer (DMP 2831, Fujifilm Dimatix) with piezoelectric-driven 10 pL printheads (DMC-11610). The jetting parameters of the ink were optimized by adjusting the parameters of the driving voltage pulse (waveform) and by heating the nozzle plate to 35 °C. The drop ejection frequency was set to 1 kHz. For the preparation of the reinforcing patterns, honeycomb patterns were chosen. The drop spacing was varied systematically. After deposition, the ink was cured using the spot lamp BlueWaveTM50 of Dymax emitting the UV-A range, the radiant power was 78 mW/cm2 at 365 nm, and the exposure time was about 9 s. After completing the manufacture of the reinforcing patterns, the aluminum foil was etched off in hydrochloric acid (7 wt %, BASF), and the particles were removed by exposure of the membranes to the vapor emanating from an aqueous solution of hydrofluoric acid (40 wt %, VWR). For analysis, the samples were transferred to silicon wafers or transmission electron microscopy grids and examined using the atomic force microscope Nano Wizzard II (JPK), the scanning electron microscope NanoNovaSEM (Fei), and the light microscope VHX500 K (Keyence).

EXPERIMENTAL SECTION

The manufacture of the microsieves by float casting basically follows the procedures published in ref 21. Hydrophobized monodisperse silica particles with a diameter of 471 nm ± 20 nm were dispersed in a mixture of the nonvolatile monomer 1,1,1-tris(hydroxymethyl)propane trimethacrylate (TMPTMA, Sigma-Aldrich), the photoinitiator benzoin isobutyl ether (Sigma-Aldrich) (mass ratio of particles to monomer to photoinitiator: 3:1:0.06), ethanol (VWR), chloroform (VWR), and ethyl butyl acetate (Merck) (mass ratio of ethanol to chloroform to ethyl butyl acetate: 1:0.99:0.01; mass ratio of particles to solvent mixture: 1:100). This mixture was spread onto a surface of degassed water in a Petri dish under an argon atmosphere. The exposure of TMPTMA to oxygen was avoided to ensure the later cross-linking. The mass of the particle suspension applied to the water surface, M, was such as to yield complete coverage of the surface by a layer of two-dimensionally arranged particles. As detailed in ref 21, it is given by the ratio of the area of the water surface, A, divided by the area per particle, √3D2/2 in a hexagonally closed packed monolayer 3317

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321

Langmuir

■

Article

RESULTS AND DISCUSSION

shows a sample before the application of the reinforcing pattern). However, this uniform order exhibits two kinds of defects, exemplarily shown in Figure 2a: (1) Several pores exceed the regular diameter (the black arrow points to such a defect). Most likely this is caused by an agglomeration of single particles. (2) The space between several pores is completely filled with polymer (the white arrow points to such a defect); i.e., the pore is missing because this location was not occupied by a particle after the particles assembly on the water surface. Defects of type 1 diminish the selectivity, and type 2 defects 163 increase the flow resistance of the microsieves. Additionally, we observed that the diameter of the openings of the pores that faced the oil/water interface (405 nm ± 19 nm, Figure 2b) during the manufacturing process were twice as wide as the openings that faced the oil/air interface (225 nm ± 24 nm, Figure 2c). This result indicates that the vertical position of the particles is not in the middle of the monomer film but protrude to a larger extent out of the oil/water than out of the oil/air interface. This observation, which has already been made previously,21 is due to the fact that the contact angle at the oil/ water interface differs from the contact angle at the oil/air interface. As it appears uniformly on the complete area of the manufactured microsieves, it has no influence on the microsieves performance. The edge of a fractured microsieve is shown in Figure 2d, proving its extreme tenuity. Optimization of Printing Process. As explained above, the aim of our approach is to apply reinforcing patterns onto fragile microsieves by inkjet printing of narrow but thick line patterns. Printing of the reinforcing pattern was done either onto the composite membrane after cross-linking of the monomer, but before removal of the particles, or onto microsieves after the removal of the particles; in both cases the membrane/microsieve was transferred to a solid substrate (silicon wafer, alumina foil) prior to printing. For an initial study single lines composed of three stacked layers were printed at a drop spacing of 60 μm onto a composite membrane that was transferred to a silicon wafer and were cured after deposition of each single layer. During this process, the idle time between the completion of printing of one layer and the start of the curing step was varied. When the curing step is performed 15 s after printing, incident-light microscopy reveals a color change of the membrane on the sides of the printed lines (Figure 3a) which

Characterization of the Microsieves. We manufactured planar microsieves with a thickness of 0.25 μm and covering an area up to 3 cm2 employing the float casting technology described above. These not yet stabilized microsieves have densely packed two-dimensionally arranged open pores and resemble those published recently21 (see Figure 2a; the figure

Figure 3. (a) Light microscopic image of a printed line UV cured 15 s after deposition revealing a color change around the line. (b) Light microscopic image of a printed line UV cured 3 s after deposition revealing the absence of a color change of the membrane.

Figure 2. Results from the analysis of microsieves: (a) microsieve on a transmission electron microscopy grid exposing small defects; (b) opening of the pores which faced the oil/water interface and (c) the oil/air interface during the manufacturing process; (d) microsieve fractured by extensional stress, revealing the tenuity of the microsieve.

might be due to interference colors caused by a thin film of components of the UV curable ink covering of the membrane 3318

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321

Langmuir

Article

Figure 4. Optimization of the drop spacing (DS) and different line morphologies depending on the orientation of the line pattern to the printing direction: (a) Draft of a digital image illustrating the line by line printing and combined line by line UV curing; the edge length of the big squares was set to 1000 μm, drop spacing was varied; (b) 25 drops/1000 μm (DS = 40 μm); (c) 33 drops/1000 μm (DS = 30 μm); (d) 50 drops/1000 μm (DS = 20 μm); (e) 100 drops/1000 μm (DS = 10 μm); the framed arrow points to one bulging line and the black arrow to one line composed of stacked coins (grid patterns were printed on composite membranes with embedded particles that were transferred to a silicon wafer).

composing the reinforcing pattern from lines oriented exclusively nonparallel to the printing direction only.

and posing the danger of blocking the final pores. A similar effect was observed for idle times of up to 3 min between printing and curing. This effect does not occur when the ink is UV cured and therefore solidified immediately after the deposition within 3 s (Figure 3b). Therefore, during subsequent experiments the curing of the ink was carried out immediately after the printing of each single line of the digital image. If the curing is done in such a way, the line shape is not a function of the substrate; if similar lines as shown in Figure 3b are printed onto glas or silicon wafers, one obtains essentially the same results. To obtain defect-free reinforcing patterns, the drop spacing had to be adjusted. By printing the image of a grid, the optimum value was found to be 20−30 μm (Figure 4c,d). If a drop spacing of 30 μm is exceeded, drops do not merge to form continuous lines when printing one layer (Figure 4b). If the drops are placed in closer distance than 20 μm, too much material is applied and the lines are broadening (Figure 4e). On the other hand, a dependence of the morphology of the line patterns on the orientation of the lines regarding the printing direction was observed. In analogy to ref 26, we found at low drop spacing a bulging of the lines oriented parallel to the printing direction and a close to stacked coins arrangement in case of a line orientation perpendicular to the printing direction (see Figure 4d,e). This difference is caused by the line-by-line assembly and curing of the pattern in the single nozzle mode (see in Figure 4a). Narrow line patterns with the morphology of stacked coins appeared by printing wet drops on already dried dropseach drop is deposited and UV cured before the adjacent drop is printed next to it. In contrast, wider bulging line patterns have been observed by printing wet drops on wet drops at low drop spacing. This compares exactly to results we reported previously,25 employing polymer solution inks that solidify upon evaporation of the solvent. In contrast to this technique, the controllability of the solidification can be improved remarkably by utilizing UV curable inks. To achieve a high permeability of the microsieves, the number of covered pores by the reinforcing pattern needs to be minimized. Therefore, we prevented wider bulging lines by

Figure 5. Printing of reinforcing patterns onto microsieves that were transferred to a silicon wafer (top view): (a) digital image of the honeycomb pattern with 20 μm drop spacing; (b) printed pattern on a microsieve; (c) segment of the reinforcing pattern; (d) UV ink (dark area) cured on a microsieve (bright area); (e) measurement of the thickness of the reinforcing pattern by atomic force microscopy; the segment corresponds to the scanning electron microscopy image of (d), and the darker area indicates the surface of the microsieve. 3319

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321

Langmuir

Article

Manufacturing of Reinforcing Patterns. The composite membranes, still containing the pore forming particles, were

transferred from the water surface to solid substrates and subsequently stabilized by printing the reinforcing pattern on top of it. If desired, the particles were removed by exposure to hydrofluoric acid vapor either before or after the printing. Honeycomb patterns have been chosen for reinforcing the microsieve. Applying the optimized printing process (drop spacing of 20 μm), excessive pore covering bulging lines parallel to the printing direction have been prevented by appropriate orientation of the honeycomb pattern with respect to the printing direction (see Figure 5a). This manufacturing process is characterized by a high degree of reproducibility, demonstrated in Figure 5b,c. Although the reinforcing patterns show surface distortions on top of the regular lines (Figure 5c−e), their basic thickness is in the range of several micrometers, i.e., up to 20 times thicker than the membrane itself, which warrants the stability of the reinforced membrane compound and of the final microsieve. The printed pattern covers about 40% of the microsieve area. To obtain freely suspended reinforced microsieves, it turned out to be most practical to transfer the composite membranes before removal of the particles to aluminum foil, apply the reinforcing pattern by inkjet printing, and remove first the alumina foil in a bath of hydrochloric acid. As a result of the etching, hydrogen is released and the ascending bubbles stress the overlaying membranes mechanically. Whereas the virgin membranes are disrupted during this process, the reinforced ones withstand the stress and can be handled with tweezers. In a last step, the pore forming particles were removed in a vapor of hydrofluoric acid, resulting in microsieves as shown in Figure 6. Within the areas of the honeycombs we found no indication of damages or flaws, proving a remarkable stability of the manufactured microsieves, whereas at the rim of the printed pattern damages occur (see the regions in the upper right and lower left corner of Figure 6a).

■

CONCLUSIONS Fragile polymeric/inorganic composite membranes manufactured by float casting and polymeric microsieves derived from these membranes were mechanically stabilized by applying a reinforcing pattern using inkjet printing of a UV curable ink. The printing process was optimized to obtain reinforcing patterns composed of narrow but thick line patterns to cover as few pores as possible and at the same time to achieve sufficient stabilization. The approach finally yields mechanically robust microsieves with still 60% of the virgin microsieve available for filtrations.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected], Ph +49 371 531-31713 (W.A.G.); e-m ail reinhard.baumann@ mb.tu-chemnitz.de, Ph +49 371 531-35843 (R.R.B.).

■

ACKNOWLEDGMENTS The work was done in the project Mikrohips financed by the Bundesministerium für Bildung und Forschung (BMBF). We thank L. Reinhardt for support in atomic force microscopy.

■

Figure 6. Mechanically reinforced freely suspended microsieve (bottom view): (a) areas up to 4 mm2 could be realized; (b) no defects within the honeycombs; (c) magnification showing the good adhesion between microsieve and reinforcing pattern.

REFERENCES

(1) van Rijn, C. J. M.; Elwenspoek, M. C. Micro Electro Mechanical Systems 1995, 83−87. (2) Kuiper, S.; van Rijn, C. J. M.; Nijdam, W.; Krijnen, G. J. M.; Elwenspoek, M. C. J. Membr. Sci. 2000, 180, 15−28.

3320

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321

Langmuir

Article

(3) Brito-de la Fuente, E.; Torrestiana-Sánchez, B.; MartínezGonzález, E.; Mainou-Sierra, J. M. Chem. Eng. Res. Des. 2010, 88, 653−660. (4) Gironés, M.; Lammertink, R. G. H.; Wessling, M. J. Membr. Sci. 2006, 273, 68−76. (5) Gironés, M.; Akbarsyah, I. J.; Bolhuis-Versteeg, L. A. M.; Lammertink, R. G. H.; Wessling, M. J. Membr. Sci. 2006, 285, 323− 333. (6) Kuiper, S.; van Rijn, C. J. M.; Nijdam, W.; Raspe, O.; van Wolferen, H.; Krijnen, G.; Elwenspoek, M. C. J. Membr. Sci. 2002, 196, 159−170. (7) Kuiper, S.; van Rijn, C. J. M.; Nijdam, W.; Elwenspoek, M. C. J. Membr. Sci. 1998, 150, 1−8. (8) van Rijn, C. J. M.; Veldhuis, G. J.; Kuiper, S. Nanotechnology 1998, 9, 343−345. (9) Vogelaar, L.; Barsema, J. N.; van Rijn, C. J. M.; Nijdam, W.; Wessling, M. Adv. Mater. 2003, 15, 1385−1389. (10) Gironés, M.; Akbarsyah, I. J.; Nijdam, W.; van Rijn, C. J. M.; Jansen, H. V.; Lammertink, R. G. H.; Wessling, M. J. Membr. Sci. 2006, 283, 411−424. (11) Bikel, M.; Ç ulfaz, P. Z. C.; Bolhuis-Versteeg, L. A. M.; Garduño Pérez, J.; Lammertink, R. G. H.; Wessling, M. J. Membr. Sci. 2010, 347, 93−100. (12) Prenen, A. M.; Van der Werf, J. C. A. H.; Bastiaansen, C. W. M.; Broer, D. J. Adv. Mater. 2009, 17, 1751−1755. (13) Tong, H. D.; Jansen, H. V.; Gadgil, V. J.; Bostan, C. G.; Berenschot, E.; van Rijn, C. J. M.; Elwenspoek, M. Nano Lett. 2004, 4, 283−287. (14) Greiser, C.; Ebert, S.; Goedel, W. A. Langmuir 2008, 24, 617− 620. (15) Jahn, S. F.; Engisch, L.; Baumann, R. R.; Ebert, S.; Goedel, W. A. Langmuir 2009, 25, 606−610. (16) Warkiani, M. E.; Lou, C.-P.; Gong, H.-Q. J. Micromech. Microeng. 2011, 21, 035002. (17) Xu, H.; Goedel, W. A. Langmuir 2003, 19, 4950−4952. (18) Goedel, W. A. Europhys. Lett. 2003, 62 (4), 607−613. (19) Ding, A.; Goedel, W. A. J. Am. Chem. Soc. 2006, 128, 4930− 4931. (20) Xu, H.; Goedel, W. A. Angew. Chem., Int. Ed 2003, 42, 4694− 4696. (21) Yan, F.; Goedel, W. A. Chem. Mater. 2004, 16, 1622−1626. (22) Singh, M.; Haverinen, H. M.; Dhagat, P.; Jabbour, G. E. Adv. Mater. 2010, 22, 673−685. (23) Yin, Z. P.; Huang, Y. A.; Bu, N. B.; Wang, X. M.; Xiong, Y. L. Chin. Sci. Bull. 2010, 55, 3383−3407. (24) Derby, B. Annu. Rev. Mater. Res. 2010, 40, 395−414. (25) Hammerschmidt, J.; Jahn, S. F.; Wachner, D.; Goedel, W. A.; Baumann, R. R. Proc. Digital Fabr. 2009, 2009, 780−783. (26) Soltman, D.; Subramanian, V. Langmuir 2008, 24, 2224−2231. (27) Stringer, J.; Derby, B. Langmuir 2010, 26, 10365−10372. (28) Iler, R. K. In The Chemistry of Silica; Wiley-Interscience: New York, 1979; pp 387−390.

3321

dx.doi.org/10.1021/la2025236 | Langmuir 2012, 28, 3316−3321