1306
OPTICS LETTERS / Vol. 33, No. 12 / June 15, 2008
Direct fabrication of physical relief structures via patterned photodeposition of a titanium alkoxide solution 1
J. David Musgraves,1,* B. G. Potter, Jr., and Timothy J. Boyle2 1
Materials Science and Engineering Department, The University of Arizona, 1235 East James E. Rogers Way, Tucson, Arizona 85721, USA 2 Advanced Materials Laboratory, Sandia National Laboratories, 1001 University Boulevard SE, Albuquerque, New Mexico 87106, USA *Corresponding author:
[email protected] Received February 21, 2008; accepted April 25, 2008; posted May 8, 2008 (Doc. ID 93026); published June 6, 2008 Ultraviolet 共 = 248 nm兲 excitation of a photosensitive Ti alkoxide solution was found to generate a metaloxide-based insoluble film on substrates in contact with the solution during illumination. Patterned deposition of 100 m wide lines of material was demonstrated using a slit-shaped aluminum shadow mask during exposure. Stylus profilometry confirmed that the average thickness of the photodeposited film monotonically varied with accumulated UV fluence, exhibiting thicknesses of 10 to 310 nm for fluences of 12 and 192 J / cm2, respectively. Moreover, the surface profile of the film surface at fluences greater than 12 J / cm2 was found to reproduce the near-field Fresnel diffraction pattern anticipated from the slit mask used. © 2008 Optical Society of America OCIS codes: 160.4236, 160.5335.
The formation of complex physical relief structures in optical and dielectric films often relies on postdeposition complex multistep processing strategies, such as photolithographic patterning and chemical etching. These methods often use sacrificial masking materials or require the direct photomodification of the thin-film material to locally alter susceptibility to chemical attack and removal. The latter approach was successfully used to form a variety of patterned relief structures in ceramic oxides formed via the photoexposure of sol–gel processed films produced from photosensitive ligated metal alkoxides 关M共OR兲x兴 [1–3]. The final film structures are again “developed” after photoexposure via a chemical etch. An alternative strategy for the fabrication of thinfilm physical relief structures directly from solution has been developed by using patterned optical irradiation of a photosensitive M共OR兲x solution. Control of solution chemistry and irradiation conditions produce a localized initiation of hydrolysis and condensation reactions that result in the deposition of insoluble oxide-containing material only in the illuminated regions of a substrate in contact with the solution. Thus spatially defined relief structures can be directly formed from solution without any additional processing steps. In this effort, a water-stable heteroleptic titanium alkoxide 共OPy兲2Ti共TAP兲2 [where OPy= pyridine carbinoxide and TAP= 2 , 4 , 6 tris(dimethylamino)phenoxide] precursor forms the basis for the photoactive sol–gel solution. The synthesis of analogous OPy-based Ti alkoxides was previously discussed [4,5]. For this compound, the large steric bulk of the TAP and the bidentate nature of the OPy ligands serve to effectively inhibit the hydrolysis of the 共OPy兲2Ti共TAP兲2 molecule. Previous optical spectroscopic analysis of the electronic and vibrational 0146-9592/08/121306-3/$15.00
structure of these and related precursors coupled with their response to UV-irradiation has indicated that incident energies resonant with cyclic ligand group transitions result in the disruption of these groups and a destabilization of the alkoxide toward hydrolysis [5,6]. In recent work, the irradiation 共 = 248 nm兲 of an 共OPy兲2Ti共TAP兲2 pyridine solution led to the formation of a colloidal suspension in solution and the deposition of an insoluble precipitate on the container wall in contact with the solution [6]. Raman analysis of these films revealed that the 共OPy兲2Ti共TAP兲2 had been selectively hydrolyzed and substantially condensed; spectra exhibited resonances associated both with bridging Ti-O-Ti groups and residual alkoxide ligands. These results were consistent with a partial reaction of the precursor monomers upon optical exposure. By controlling the irradiation conditions of the 共OPy兲2Ti共TAP兲2 solution and by using the appropriate photomask, these results have been extended to photoinduce the formation of physical relief structures directly onto substrates from solution with no additional processing. In this Letter, a 106 mM solution of 共OPy兲2Ti共TAP兲2 in pyridine (Fisher) was synthesized under an inert argon atmosphere. To this solution, 7.59 L of deionized water 共18 M⍀兲 was added to generate a 4:1 ratio of water-to-titanium in solution, which was stirred at 500 rpm. After 30 min, a 10 L aliquot was deposited onto a fused silica substrate and then covered with a second fused silica substrate to confine the solution between the glass slides and form the deposition cell. To measure the effects of total fluence on a thinfilm formation, a series of lines was formed by sequentially irradiating the deposition cell through a 105 m slit (aluminum shadow mask). In between exposures, the cell was horizontally translated by © 2008 Optical Society of America
June 15, 2008 / Vol. 33, No. 12 / OPTICS LETTERS
1 mm increments prior to subsequent irradiation in the new position. Samples were exposed using a KrF excimer laser ( = 248 nm, repetition rate = 5 Hz, pulse length= 10 ns/ pulse) delivering 10 mJ/ cm2 / pulse to the surface of the shadow mask. Lines were patterned using the expose and step process described above for total incident fluences of 12, 24, 48, 96, and 192 J / cm2. After optical processing was completed, the two glass slides comprising the deposition cell were separated in a bath of pyridine and allowed to soak for 5 min to remove any unreacted precursor material. Upon removal from the solvent bath and air drying, a series of lines of material were visually observable on the fused silica slide closest to the photomask. Stylus profilometry using a Dektak 6 M stylus profilometer [Fig. 1(a)] indicated that the average step height for the lines was linear with accumulated fluence [Fig. 1(b)]. No measurable film thickness was observed for fluences ⬍10 J / cm2 using the stylus profilometer. A more complete investigation of microstructure development at the early stages of film formation is now under way. A higher-resolution surface profile obtained from the upper aspect of the photodeposited lines is shown
Fig. 1. Measurement of the photodeposited feature heights as a function of incident laser fluence. (a) Stylus profilometry trace of the deposited material and (b) linear dependence of average step height on laser fluence.
1307
as the solid curve in Fig. 2. Given the irradiation conditions used to create the lines, a near-field Fresnel-type diffraction pattern is expected at the solution-substrate interface (located ⬃1 mm beyond the plane of the single-slit shadow mask). The intensity of the diffracted beam at point P after passing through a rectangular aperture is given by IP =
I0 4
兵关C共u2兲 − C共u1兲兴2 + 关S共u2兲 − S共u1兲兴2其
⫻ 兵关C共v2兲 − C共v1兲兴2 + 关S共v2兲 − S共v1兲兴2其,
共1兲
where C共x兲 and S共x兲 are the Fresnel cosine and sine integrals, C共x兲 =
冉 冊 冕 冉 冊 冕
x
cos
0
x
S共x兲 =
sin
0
x ⬘2 2
x ⬘2 2
dx⬘ ,
dx⬘ ,
共2兲
and the u and v are dimensionless variables describing the illumination and observation conditions used, including the location of the point P in relation to the aperture, the wavelength of the light, and the dimensional details of the aperture [7]. Using the experimentally anticipated image distance 共1 mm兲, the measured slit width 共105 m兲, and the wavelength used 共248 nm兲, the computed intensity profile shows good agreement with the experimentally observed surface profile of the 48 J / cm2 line (see Fig. 2). Thus, the materials deposition technique successfully reproduced complicated intensity patterns in the deposited material. Insight into the evolution of this surface profile during deposition is provided by the stylus profilometer traces (Fig. 3) for each of the lines produced at different UV-fluences, i.e., corresponding to different
Fig. 2. Comparison of the material profile of the upper aspect of a deposited line (solid curve) and the calculated Fresnel diffraction pattern intensity profile at the substrate-solution interface (dotted curve).
1308
OPTICS LETTERS / Vol. 33, No. 12 / June 15, 2008
Fig. 3. Evolution of the upper surface profile of the deposited material with increasing UV fluence. The total incident fluence associated with each trace is provided to the left.
Fig. 4. Surface profile contrast (defined in text) normalized to average feature height for the photodeposited relief structures depicted as a function of total incident UV fluence.
average line thicknesses. It is clear that the beam intensity distribution associated with the Fresnel diffraction pattern is not reproduced by the material in the early stages of film formation. However, under longer irradiation times (greater accumulated fluence), a surface topography consistent with the beam intensity profile becomes increasingly apparent in the material trace with the 48 J / cm2 trace showing the best fit between the computed and experimental profiles. Interestingly, additional irradiation fluence, while continuing to contribute to an increased film thickness, results in the removal of the finer features of the diffraction pattern reproduced in the film surface. This broadening effect is most likely due to incoherent scattering as the writing beam traverses the newly deposited film to the solution interface. Such scattering is likely associated with structural heterogeneities in the material and will result in the disruption of spatial coherence across the beam. The surface profile contrast, defined here as the difference in height between the leftmost maximum and leftmost minimum in the material profile trace for each line, exhibits a monotonic increase with exposure fluence within the range of 24– 96 J / cm2. As shown in Fig. 4, normalizing this contrast to the average height of the film (feature height) demonstrates that the contrast is approximately 50% of the feature height for the lines patterned in the lower fluence range, while there appears to be a saturation in the contrast developed at the largest fluence used. Currently, we are investigating the film formation mechanism that is anticipated to couple photoactivated chemistries associated with the initial hydrolysis and condensation of the alkoxide in solution [5,6] as well as the photoresponse of the newly deposited material to subsequent irradiation as the deposition proceeds. It is anticipated that these two processes contribute to the spatial variation of the effective photodeposition rate that allows the reproduction of a continuously variable (gray scale) intensity pattern
in the resulting film surface profile that is retained as the overall film thickness is increased. In this context, early results indicate a measurable correlation between the nanostructure of the deposited material and such solution chemistry variables as the water content and precursor concentration. In conclusion, solid-state physical relief structures on the micrometer scale have been directly deposited from a photoactive Ti alkoxide solution by UV exposure. The near-field Fresnel diffraction pattern, created by the slit used as the shadow mask, was found to be superimposed on a film whose average thickness was defined by the total UV fluence used. The creation of such physical relief structures presents a host of possible uses from the formation of gratings to the development of cellular scaffolding with a tailored multilength scale structure. Research was supported by the Department of Energy, Office of Basic Energy Sciences, and the State of Arizona Technology and Research Initiative Fund Optics Initiative. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed-Martin Company, for the United States Department of Energy under contract DE-AC0494AL85000. References 1. H. Segawa, S. Adachi, Y. Arai, and K. Yoshida, J. Am. Ceram. Soc. 86, 761 (2003). 2. K. Tadananga, T. Fujii, A. Matsuda, T. Minami, and M. Tatsumisago, J. Sol-Gel Sci. Technol. 31, 299 (2004). 3. N. Tohge, R. Ueno, F. Chiba, K. Kintaka, and J. Nishi, J. Sol-Gel Sci. Technol. 19, 119 (2000). 4. T. J. Boyle, R. M. Sewell, L. A. M. Ottley, H. D. Pratt, C. J. Quintana, and S. D. Bunge, Inorg. Chem. 46, 1825 (2007). 5. J. D. Musgraves, B. G. Potter, Jr., R. M. Sewell, and T. J. Boyle, J. Mater. Res. 22, 1694 (2007). 6. B. G. Potter, Jr., J. D. Musgraves, and T. J. Boyle, J. Non-Cryst. Solids 354, 2017 (2008). 7. E. Hecht, Optics (Addison Wesley, 2002).
