LETTER pubs.acs.org/Langmuir
Patterned Removal of Molecular Organic Films by Diffusion Corinne E. Packard,* Katherine E. Aidala,* Sulinya Ramanan, and Vladimir Bulovic Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
bS Supporting Information ABSTRACT: We demonstrate that “contact patterning” subtractively patterns a wide range of molecular organic films of nanoscale thickness with nanometer-scale accuracy. In “contact patterning”, an elastomeric stamp with raised features is brought into contact with the organic film and subsequently removed, generating patterns by the diffusion of the film molecules into the stamp. The mechanism of material removal via diffusion is documented over spans of minutes, hours, and days and is shown to be consistently repeatable. Contact patterning provides a photolithography-free, potentially scalable approach to subtractive patterning of a wide range of molecular organic films.
M
icrocontact printing-based methods offer viable approaches for achieving microscale feature definition for a wide range of materials, including vapor-deposited molecular organic solids.17 We recently demonstrated that a variety of molecular organic thin films can be patterned by using relief-patterned polydimethylsiloxane (PDMS) stamps to remove a controlled thickness of vapor-deposited material.8 In contrast to earlier studies, our “contact-patterning” process eliminates the need for elevated processing temperatures, partially cured polymer stamps, increased applied pressures, or combinations of these.1,4,7 The present study investigates the mechanism underlying the contact-patterning method. We find that the patterning process is time dependent, with more material being removed for longer contact times between the stamp and deposited film. Data from this study show that patterning occurs by diffusion of the organic chromophores into the PDMS stamp during contact. The diffusive nature of the patterning process makes possible repeated patterning with a single stamp and nanometer-scale control over the removed film thickness. Figure 1a shows a schematic of the contact-patterning technique. A PDMS stamp is brought into contact with a thermally evaporated molecular organic film, left in contact for a predetermined amount of time, and then removed (experimental details are included in Supporting Information). As detailed in ref 8, patterning may completely remove the organic thin film, revealing the substrate, or it may remove a finite thickness of the film, depending on the material. Figure 1b shows that the amount of material removed exhibits a time dependence for the five different organic materials studied here: 3-(4-biphenylyl)-4-phenyl-5-tert-butylphenyl-1,2,4-triazole (TAZ), N,N0 -bis(3-methylphenyl)-N,N0 -bis(phenyl)-benzidine (TPD), N,N0 -bis(3-methylphenyl)-N,N0 -bis-(phenyl)-9,9-spiro-bifluorene (spiro-TPD), 2,20 ,200 -(1,3,5-benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi), and tris(8-hydroxy-quinolinato)aluminum (Alq3). Each point in the data series is the average thickness removed, as calculated from six ellipsometry measurements from two different contact-patterned areas on the same organic film and three measurements of the as-deposited thickness. The error bars reflect the standard r 2011 American Chemical Society
deviation and include the propagated error from the as-deposited and the postpatterning measurements, neglecting covariance between the two. Additional data (not shown) reveals that while the trend remains consistent, the quantitative amount of material removed was dependent on the age of the stamp (which is defined as time between the casting of the PDMS and use of the stamp for contact-print patterning), with stamps removing less organic material with age. The present study controls for this aging effect by using stamps prepared in a single batch and by completing each experiment over a comparatively short period of time. Figure 1b shows a large disparity in the amount of material removed by the contact-patterning of different thin films. After 60 min of stampfilm contact time, (229 ( 25) nm of TAZ and (94 ( 6) nm of TPD are removed, while only (39 ( 1) nm of spiro-TPD, (12 ( 1) nm of Alq3, and (19 ( 1) nm of TPBi are removed. For all of the films, however, patterning occurs to greater depths the longer the PDMS stamp remains in contact with the organic film. In an earlier study by Choi et al. employing polymeric stamps to pattern N,N0 -bis(naphthalen-1-yl)-N,N0 -bis(phenyl)-benzidine (NPB) and Alq3,4 the proposed mechanism for the organic thin-film removal was based on adhesion. There, a favorable static work of adhesion between the surfaces of the polymer stamp and organic film, calculated based on contact angle measurements, was presented as evidence of adhesion-based lift-off. In our study, the time-dependence of the material removal suggests that the patterning mechanism is not strictly due to adhesion of the organic film molecules to the PDMS stamp. An alternative explanation is that organic thin film removal is controlled by diffusion of molecules from the organic thin film into the PDMS matrix during the PDMSstamp/film contact. The driving force for diffusion of solutes into a matrix is proportional to the concentration gradient that develops within Received: May 6, 2011 Revised: June 23, 2011 Published: June 23, 2011 9073
dx.doi.org/10.1021/la2016929 | Langmuir 2011, 27, 9073–9076
Langmuir
Figure 1. (a) Schematic of the contact-patterning process. A PDMS stamp is brought into contact with a molecular organic film, left in contact for a prescribed amount of time, and delaminated to subtractively pattern the film. (b) Prolonged contact with the PDMS stamp results in a greater thickness of material removed by stamping for each of the five organic materials studied here. Error bars represent the standard deviation of thickness for multiple sample measurements.
the PDMS,9,10 which is consistent with the trend observed in Figure 1b. The rate of removal declines over time, presumably due to the decreasing concentration gradient as the organic dye molecules continue to diffuse into the PDMS polymer network with time. Differences in material removal rates and final saturation values are likely linked to differences in diffusivity between various organic molecules. To test whether the organic dye molecules diffuse into the PDMS stamp, we prepared a series of samples with thin films of organic molecules on PDMS, glass, and silicon substrates. All samples underwent exposure to oxygen plasma for a period of 30 s, after which the photoluminescent intensity of their chromophores was tested. While oxygen plasma is known to cause removal or photooxidation of organic dye molecules on exposed surfaces,1114 a previous study15 that employed structurally similar organic dyes showed that these molecules remained protected from oxidation when embedded within a polymer matrix. Similarly, if the organic dye molecules in our experiments diffuse into the bulk of a PDMS stamp, they could become protected from the caustic effects of the oxygen plasma exposure, remaining luminescent. Two sets of samples were prepared, with each set including a glass substrate, a flat piece of PDMS, and a silicon substrate. All glass and flat PDMS substrates were first covered with shadow masks to define a 5 mm 5 mm window in the center of each. We evaporated a 20 nm thick film of spiro-TPD on one set of samples and a 7.5 nm thick film of Alq3 on the other. The organic thin films on the silicon substrates are patterned with a set of PDMS stamps molded with a pattern consisting of interpenetrating arrays of square features 100 μm, 500 μm, and 1 mm in size. These molded PDMS stamps are placed in contact with the organic thin films on silicon for a period of 20 min, which is sufficient time to pattern through the deposited organic film thickness for each material. The molded PDMS stamp is then delaminated from the substrates and retained for this study.
LETTER
Figure 2. Photographs of the effect of oxygen plasma exposure for spiro-TPD (ac) and Alq3 (df) organic dyes shown on various substrates under UV illumination. See text for full description of samples. (a, d) Photoluminescence of the organic dye molecules prior to the oxygen plasma treatment. (b, e) Photoluminescence from the organics ceases on the glass substrates (left-most in each image), but persists in the PDMS samples (middle and right) after 30 s long oxygen plasma exposure. (c, f) Photographs of the PDMS samples in (b) and (e) after 3 weeks of storage in a low vacuum (>0.1 Torr) environment. All samples are approximately 1 cm 1 cm in size.
Photographs in Figure 2 show the glass substrate with evaporated organic thin film, the molded-PDMS stamp used for patterning of films on silicon, and the flat-PDMS substrate with evaporated organic film, arranged from left to right in each photograph. In all photographs, the samples are illuminated by a hand-held ultraviolet lamp (emitting λ = 365 nm wavelength light) such that photoluminescence from the organic material is apparent. The exact positioning of the samples and lighting was not maintained for all of the photographs, but each image clearly shows the luminescence of the samples, with Figure 2ac and df pertaining to spiro-TPD and Alq3, respectively. Before oxygen plasma treatment, the photoluminescence indicates the presence of the organic dye molecules in all three samples for each dye material. In the case of the molded-PDMS stamps (center samples in the photographs), the photoluminescent organic dye has transferred to the raised PDMS features that were in contact with the organic films on silicon substrates. After exposure to oxygen plasma, the glass substrates no longer show photoluminescence, a result consistent with reports in the literature that intentionally employ oxygen plasma treatments to remove organic species from surfaces. In both flat and molded PDMS samples, however, the photoluminescence persists despite the plasma treatment, indicating that organic dye molecules remain and have not become oxidized. This result provides evidence that the dye molecules are not located on the surface of the PDMS, but that they diffuse into the bulk of the PDMS stamp where they are protected from the reactive plasma species. Molecular diffusion is also evident in the Figure 2c photograph of the spiro-TPD film on PDMS which shows that, after 3 weeks of storage in a dry vacuum environment, the patterned squares of spiro-TPD molecules on the molded-PDMS stamp are no longer discernible. Instead, the entire stamp faintly glows the bluish 9074
dx.doi.org/10.1021/la2016929 |Langmuir 2011, 27, 9073–9076
Langmuir color of spiro-TPD. Similarly, the shadow-masked square pattern on the flat-PDMS has lost its well-defined boundaries, revealing diffusion of molecules away from their original positions. For the same time period, Figure 2f shows that Alq3 molecules do not diffuse far enough to be visibly different from the photograph of Figure 2e, consistent with the slower diffusion of Alq3 in comparison to spiro-TPD (cf. the shallower contact-patterning depth of Alq3 in Figure 1b). In rubbery polymers, such as the PDMS used in this study, a prevailing theory for diffusion of molecules in the polymer matrix is based on the concept of free volume.9,10 Free volume in the network of the polymer chains consists of permanent void spaces between the chains of the polymer backbone as well as temporary openings that arise from chain movement above the glass transition temperature.9,10 Since uptake of molecules by diffusion relies on the movement of molecules through the free volume in the polymer, diffusion constants are generally smaller for bigger molecules and for stronger interactions of the molecules with the polymer chains.9 Lacking known diffusion constants for these organic materials in PDMS, we can instead look at sublimation temperature (data in Table S1 of the Supporting Information), which reflects a combination of the physical properties that are used as inputs in polymer diffusion models, with larger and more strongly interacting molecules having a higher sublimation temperature.16 The connection between sublimation temperature and diffusion rate, however, is not a direct correlation; the results shown in Figure 1 are consistent with trends for the organic film removal in ref 8, but we find that trends in sublimation temperature alone are insufficient in capturing the subtleties of organic material removal by diffusion into PDMS. The results of an earlier study by Lange et al.17 provide further evidence that diffusion of organic small molecules in PDMS is influenced by free volume within the polymer matrix, which we tested through modification of the structure of our PDMS stamps by extracting residual low molecular weight oligomers. The oligomer extraction was accomplished by exposing the cured PDMS stamps to a series of solvents that swell the polymer chain network and allow oligomers to diffuse out of the PDMS matrix. Lange et al. find that diffusion of organic dye molecules is enhanced in the oligomer-extracted PDMS and attribute this result to a greater degree of free volume in the extracted polymer.17 Similarly, our results on contact-patterned spiroTPD thin films show that the amount of material removed by stamping with a PDMS stamp that has been cured and extracted according to the procedures in ref 18 results in a 2-fold enhancement in material removal after a 60 min stamp application ((45 ( 3) nm for extracted stamps versus (19 ( 10) nm for as-cast PDMS). It is instructive that these early results show that control over material removal rates can be tailored by manipulating the polymer network and by material selection based on the sublimation temperature. Further study is necessary to determine the precise dynamics at the interface of the PDMS and the film, the exact role of the oligomers, and to gain predictive power over diffusion rates of organic small molecules. We note that all of our contact-patterning results have been performed on smooth, vapor deposited surfaces (cf. Figure S1 of the Supporting Information). We expect, however, that this patterning technique may be extended to spun-cast materials, but additional study is required to confirm this point, especially regarding the effect of spun-cast film roughness on the fidelity of contact-patterning. Though the industrial feasibility of patterning over large areas via microcontact printing methods remains to be demonstrated,
LETTER
Figure 3. Fluorescence micrograph of repeated stamping trials on three separate areas of a silicon substrate coated with a 20 nm thick film of spiro-TPD. Dark areas indicate where material has been removed after the first (a), second (b), and third (c) patterning attempts with a single PDMS stamp. Scale bars represent a distance of 500 μm.
several schemes for scaling up have been suggested, including roll-toroll and wave printing.19,20 In these processes, the issue of stamp reusability is critical. In adhesion-based printing methods, a cleaning step must be included between each patterning attempt to refresh the surface of the stamp. Given the diffusive nature of the patterning process described in this paper, reusing the same stamp for repeated patterning is a possibility. Figure 3 shows the results of repeated patterning attempts using a single stamp that was placed in contact for 20 min sequentially on three different regions on a silicon substrate with an initial layer thickness of 20 nm of deposited spiro-TPD. Minimal loss in pattern fidelity is observed over the course of three patterning attempts with the same stamp, suggesting that repeated contact-patterning is possible, although the stamp may gradually saturate. Optimizing the process by altering the stamp, time, temperature, or thickness of the films could result in complete material removal across repeated uses. In summary, we investigated the mechanism for subtractive contact-patterning of a range of molecular organic thin films. Material removal by PDMS stamps is found to be a time-dependent process that occurs due to diffusion of organic small molecules from the organic thin film into the polymer stamp. Results concerning material removal rates and oligomer extraction are consistent with existing free-volume models for diffusion of molecules into rubbery polymers such as the PDMS we employ as patterning stamps. We expect that the contact-patterning process should be scalable to patterning over large areas by adopting roll-to-roll processing geometries.
’ ASSOCIATED CONTENT
bS
Supporting Information. Table of properties, AFM images, and experimental details for PDMS, stamp, and film preparation, characterization techniques, and oligomer extraction procedure. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*(C.E.P.) Mailing address: Department of Metallurgical and Materials Engineering, Colorado School of Mines, Golden, Colorado 80401, United States. E-mail:
[email protected]. (K.E.A.) Mailing address: Department of Physics, Mount Holyoke College, South Hadley, Massachusetts 01075, United States. E-mail:
[email protected].
’ ACKNOWLEDGMENT The authors gratefully acknowledge support from the Center for Excitonics, an Energy Frontier Research Center funded by the U. S. Department of Energy under Grant Number DE-SC0001088, 9075
dx.doi.org/10.1021/la2016929 |Langmuir 2011, 27, 9073–9076
Langmuir
LETTER
Office of Science and Office of Basic Energy Sciences (C.E.P.), the Clare Boothe Luce Foundation (K.E.A.), and the NSF Electronics, Photonics, and Magnetic Devices program (Grant Number 1001994). The authors also acknowledge the use of the facilities of the NSF MRSEC Center for Materials Science and Engineering.
’ REFERENCES (1) Kim, J. K.; Suh, K. Y. Appl. Phys. Lett. 2008, 92, 223107. (2) Takakuwa, A.; Misaki, M.; Yoshida, Y.; Yase, K. Thin Solid Films 2009, 518, 555–558. (3) Bolognesi, A.; Botta, C.; Yunus, S. Thin Solid Films 2005, 492, 307–312. (4) Choi, J. H.; Kim, D.; Yoo, P. J.; Lee, H. H. Adv. Mater. 2005, 17, 166–171. (5) Kim, C.; Cao, Y.; Soboyejo, W. O.; Forrest, S. R. J. Appl. Phys. 2005, 97, 113512. (6) Kim, K. H.; Bong, K. W.; Lee, H. H. Appl. Phys. Lett. 2007, 90, 093505. (7) Wang, Z.; Zhang, J.; Xing, R.; Yuan, J. F.; Yan, D. H.; Han, Y. C. J. Am. Chem. Soc. 2003, 125, 15278–15279. (8) Yu, J. Improving OLED Technology for Displays. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, 2008. (9) Crank, J.; Park, G. S. Diffusion in polymers; Academic Press: London, New York; 1968; p xii, 452 pp. (10) Neogi, P. Diffusion in polymers; Marcel Dekker: New York, 1996; p ix, 309 pp. (11) Kukhto, A. V. J. Appl. Spectrosc. 2003, 70, 165–194. (12) Lu, D.; Wu, Y.; Guo, J. H.; Lu, G.; Wang, Y.; Shen, J. C. Mater. Sci. Eng., B 2003, 97, 141–144. (13) Ramamurthy, V.; Schanze, K. S. Organic, physical, and materials photochemistry; Marcel Dekker: New York, 2000. (14) Gan, F.; Xu, L. Photonic glasses; World Scientific: Singapore; Hackensack, NJ, 2006. (15) Chowdhury, S.; Bhethanabotla, V. R.; Sen, R. Measurement of Diffusion Coefficient of Oxygen in FRP Using Luminescence Quenching; Composites & Polycon 2007, Tampa, FL, Oct. 1719, 2007; http:// www.acmanet.org/resources/07papers.cfm. (16) Schwoerer, M.; Wolf, H. C. Organic molecular solids; Wiley-VCH: Weinheim, 2007. (17) Lange, J. J.; Collinson, M. M.; Culbertson, C. T.; Higgins, D. A. Anal. Chem. 2009, 81, 10089–10096. (18) Herrmann, K.; Jennett, N. M.; Kuypers, S.; McEntegaart, I.; Ingelbrecht, C.; Hangen, U.; Chudoba, T.; Pohlenz, F.; Menelao, F. Z. Metallkd. 2003, 94, 802–806. (19) Decre, M. M. J.; Schneider, R.; Burdinski, D.; Schellekens, J.; Saalmink, M.; Dona, R. Wave printing (I): towards large-area, multilayer microcontact printing; In Nontraditional Approaches to Patterning; Yang, S., Xia, Y., Liu, J., Lakeman, C., Eds.; Mat. Res. Soc. Symp. Proc. 2004, EXS-2, 5961. (20) Rogers, J. A.; Bao, Z. N.; Makhija, A.; Braun, P. Adv. Mater. 1999, 11, 741–745.
9076
dx.doi.org/10.1021/la2016929 |Langmuir 2011, 27, 9073–9076
