Matrix Assisted Laser Transfer of Electronic Materials for Direct Write Applications
R.C.Y. AUYEUNG*, H.D. WU*, R. MODI**, A. PIQUÉ, J.M. FITZ-GERALD, H.D. YOUNG*, S. LAKEOU*, R. CHUNG† and D.B. CHRISEY
Naval Research Laboratory, Code 6372, Washington DC 20375
*SFA Inc., 9315 Largo Drive West, Largo, MD 20774
**George Washington University, Department of Mechanical and Aerospace Engineering, Washington DC 20052
†Geo-Centers, Inc., 1801 Rockville Pike, Rockville, MD 20852
A novel laser-based direct-write technique, called Matrix Assisted Pulsed Laser Evaporation Direct Write (MAPLE-DW), has been developed for the rapid prototyping of electronic devices. MAPLE-DW is a maskless deposition process operating under ambient conditions which allows for the rapid fabrication of complex patterns of electronic materials. The technique utilizes a laser transparent substrate with one side coated with a matrix of the materials of interest mixed with an organic vehicle. The laser is focussed through the transparent substrate onto the matrix coating which aids in transferring the materials of interest to an acceptor substrate placed parallel to the matrix surface. With MAPLE-DW, diverse materials including metals, dielectrics, ferroelectrics, ferrites and polymers have been transferred onto various acceptor substrates. The capability for laser-modifying the surface of the acceptor substrate and laser-post-processing the transferred material has been demonstrated as well. This simple yet powerful technique has been used to fabricate passive thin film electronic components such as resistors, capacitors and metal lines with good functional properties. An overview of these key results along with a discussion of their materials and properties characterization will be presented.
Keywords: direct write, laser forward transfer, matrix assisted pulsed laser evaporation (MAPLE), matrix assisted pulsed laser evaporation direct write (MAPLE-DW), rapid prototyping, electronics devices.
5
1. Introduction
There is a strong need in industry for rapid prototyping and manufacturing of passive electronic components on various substrates in the mescoscopic regime (micron to mm range). This capability is required in order to fabricate smaller and more versatile electronics devices, to iteratively measure the performance of circuits too difficult to model and to quickly design and test circuits without a time-consuming photolithographic mask process.
Direct-write technologies provide a flexible, maskless and efficient technique for depositing a wide variety of materials under ambient conditions which can be easily integrated into a CAD/CAM system. Direct-write methods do not compete directly with photolithography in manufacturing sub-micron size components, but are intended to supplant current surface mount technologies in the mesoscopic regime.
Various direct-write technologies have been developed in the last few years. Some of these techniques include micropen, inkjet, plasma spray, focused ion beam, e-beam and other liquid microdispensing approaches. Laser-based direct-write techniques[1] include Laser Induced Forward Transfer (LIFT), laser CVD, laser particle guidance and Matrix Assisted Pulsed Laser Evaporation Direct Write (MAPLE-DW).
In this paper, the principles of the MAPLE-DW technique will be described and results from some of the fabricated structures will be presented.
2. Background
Ever since the invention of the laser in 1960, much research has been directed towards the interaction of laser radiation with materials. Its high coherence, directionality and brightness properties make laser radiation ideal for materials processing applications. These same properties also make it ideal for direct-write applications where a highly controlled source of energy with well-defined spatial qualities can be used to process materials in a specific location and time interval. By choosing the correct laser wavelength, fluence and optical beam delivery system, many types of materials can be deposited at room temperature under ambient conditions with feature sizes in the micron range. Various laser-based direct-write techniques have been used to deposit metals[2] and dielectrics[3] onto various substrates with varying degrees of success for each.
A promising new direct-write technique that allows the deposition of almost any kind of material onto any substrate type and shape is the MAPLE-DW process[4]. This technique combines aspects of both the MAPLE[5] and LIFT[6] processes into a simple but versatile direct-write technology. In the MAPLE process, large and fragile organic molecules to be deposited into thin film form are first dissolved in a dilute (usually frozen) matrix target. Then, a low fluence UV laser beam strikes the target to gently warm the solvent matrix to gently release the organic molecule while the now volatile solvent molecules are removed by vacuum pumping. This pulsed thermal process heats the surface to below the decomposition temperature of the organic molecules while the solvent matrix is heated to its evaporation point. The combined action of the evaporating matrix desorbs the large species intact and deposits them as uniform thin films.
LIFT is a simple pyrolytic direct-write technique where focused laser radiation vaporizes a thin film coating (<2000Å) on one side of an optically transparent support. The vaporized material is transferred onto a substrate placed adjacent to this coating. Compared to MAPLE, LIFT requires a higher laser fluence because it must remove the thin film from its support by physical vapor deposition. For this reason and to maintain good lateral resolution, the thin film coating on the support is usually less than 1000 Å thick. Thicker depositions would require time-consuming iterative steps. Because the laser radiation tends to atomize the thin coating, LIFT is best suited for metals since multicomponent crystalline materials would not be able to be transferred intact. Other disadvantages of LIFT include poor spatial resolution, adhesion and morphology.
MAPLE-DW takes the experimental support of LIFT and combines it with the matrix desorption mechanism of MAPLE into a versatile direct-write process that takes place at room temperature. The support is transparent at the laser wavelength and consists of a coating of the materials of interest dispersed in a matrix. The matrix can consist of organic solvents, binders, dyes or other components, which aid in the transfer and particle bonding process. Ideally, when the laser beam strikes this coating and transfers it to the substrate, most of the matrix components should be removed, easily decomposed or evaporated after the transfer. An important benefit of MAPLE-DW is that unlike other approaches, this technique does not change the properties of the material of interest after the transfer.[7]
Another advantage of MAPLE-DW, which distinguishes it from other techniques, is that it can operate additively or subtractively. The presence of the laser beam allows in-situ, pre- or post-processing of the transferred material and substrate. For example, the laser can be used to pre-clean the substrate surface, micromachine vias or channels, trim circuit elements, or sinter the deposited material.
The MAPLE-DW process has been used in this work to successfully deposit metal lines, resistors and capacitor structures with good properties on various substrates.
3. Experimental Procedure
The experimental setup used in MAPLE-DW is very similar to that used in micromachining or the LIFT process. A UV laser operating at 355 nm with a 5 ns pulsewidth is focused onto a UV fused silica disk which acts as the optically transparent support. This disk is coated on one side with the material of interest and its corresponding matrix. The disk is placed with its coating side adjacent to the receiving substrate to within tens of microns gap separation. The entire disk and substrate is placed on a computer-controlled X-Y table which is synchronized to the output of the laser. The optical delivery system can generate UV focal spots from 10 to 300 micron diameter and fluences over 5 J/cm2. The UV beam energies were monitored continuously during each deposition by an in-line beam-splitter and focal burn patterns were recorded on polyimide or UV photosensitive film. A secondary IR Nd:YAG laser operating in the free-running mode provides 1.06 µm pulses at 10-40 Hz with a 50 µs width for post-processing the transferred material. The table speed can be adjusted so that the IR focal spots overlap accordingly to give the required sintering dwell times.
Materials were transferred onto various substrates such as glass, alumina and polyimide. By optimizing the deposition parameters, deposited materials showed good morphology, linewidth and functional properties. Metal lines were transferred by MAPLE-DW between two photolithographically patterned Au electrode pads. An e-beam deposited gold line served as a reference on each sample. A standard 4—point probe DC measurement was used to characterize the resistivity of these lines. Resistor lines were transferred by direct-write between two Au electrode pads separated by 700 µm and its electrical properties characterized by both a 4-point probe measurement and an HP4291B impedance analyzer. Dielectrics were evaluated by transferring a (~0.7x0.8 mm) pad of material over a previously patterned interdigitated capacitor structure (IDC). This structure allowed quick characterization of dielectric material properties and avoids the complication of depositing extra metal pad layers as would be found in a parallel-plate capacitor structure. The capacitance and loss tangent values of the capacitors were measured on an HP4284A LCR meter up to 1 MHz and on the impedance analyzer from 1 MHz to 1.8 GHz. Values measured at 1 MHz agreed within error on both systems and provide confidence in our measurement technique.
All samples were evaluated by optical and scanning electron microscopy (SEM). Both surface and fracture cross-sectional SEM’s were performed on the different substrates.
4. Results and Discussion
Silver lines were transferred onto polyimide substrates and after a furnace anneal at 300 °C to react the metal precursor, the resistivities were measured to be 1.1 to 1.6X bulk Ag. A photograph of the line as well as profilometer scans across 6 different locations along a 7 mm-long segment are shown in Fig. 1. The lineshape definition is good with minimal debris as deposited. Only a single laser pass was used to deposit these 2.5 µm (average) thick lines and the 40-µm linewidth was comparable to the laser spot size. Silver lines were also deposited onto alumina and glass substrates with somewhat higher resistivities. A scotch tape test of the Ag line on glass left it intact which demonstrates the excellent adhesion of the transferred material.
Laser sintering was attempted on some of the higher resistivity silver lines with mixed results. At low levels of IR power absorption, only the surface silver layer showed signs of annealing or melting as shown in Fig. 2. When the IR laser power was increased, the silver line tended to “bead” or eventually delaminate from the substrate surface. The thermal diffusion depth for silver is over 10 mm for a 1ms laser pulse so it is possible that sufficient porosity remains in the cross-section preventing efficient thermal conduction of the heat from the surface layer to the entire volume.
Polymer resistor lines were transferred between two Au pads on alumina substrates. By changing the formulations of the starting resistor material, nearly 4 orders of magnitude change in the sheet resistance was obtained in the transferred lines as shown in Fig. 3. The drop-off at higher frequencies is not unique to polymer resistors as it is also observed in cermet resistor material formulations[8]. The oscillations at the highest frequencies are due to resonance effects from the test leads and sample during the experiment. Note the good agreement between the measured sheet resistance values at 1 MHz and those actually specified for the starting material. This agreement demonstrates again that the MAPLE-DW process does not significantly alter the intrinsic properties of the material after transfer.
The dielectric material barium titanate (BTO) was transferred onto pre-fabricated IDC’s on polyimide and alumina substrates. As shown in Fig. 4, an SEM cross-section of the transfer shows that a high packing density was achieved.
The measured capacitance and loss tangent of a BTO IDC from 1 MHz to 1.8 GHz are shown in Fig. 5. The measured dielectric constant er of 70 and loss tangent of 4 % (after correcting for the substrate) at 1 MHz is significantly below the corresponding values for the powder starting material. Due to the high dependence of the dielectric constant on the porosity of the dielectric material[9], it is possible that the air volume fraction in the BTO transfer is still sufficient to dramatically lower the er value.
5. Summary
In conclusion, we have demonstrated the versatility and strength of the MAPLE-DW technique as a rapid direct-write technology in depositing good-quality electronics materials under ambient environmental conditions. Silver lines with bulk-like conductivities have been deposited onto polyimide with a short furnace treatment as the only post-processing step. Resistor lines with sheet resistances of 4 decades have been obtained by only changing the composition of the starting material. Barium titanate compositions have been transferred as pads and their measured dielectric constants and loss show good promise as useful capacitor structures.
In addition, we have shown that MAPLE-DW can deposit different types of materials onto various substrates such as polyimide, metals, glass and alumina. The usefulness of this technique has been strongly demonstrated in this paper with high-quality material properties obtained directly with a simple transfer and furnace treatment step. No other pre- or post-processing procedures were used. Improvements in the materials formulations as well as a deeper understanding of the mechanism of the MAPLE-DW process should bring further success of this technology as a rapid prototyping tool.
6.
Acknowledgements
We gratefully acknowledge support for this work by the DARPA-MICE program and the Office of Naval Research.
References
[1] J.M. Fitz-Gerald, D.B. Chrisey, A. Piqué, R.C.Y. Auyeung, R. Modi, H.D. Young, H.D. Wu, S. Lakeou and R. Chung: Proc. Spring MRS, Symp. V, San Francisco, CA, 2000.
[2] M.J. Renn, M. Essien, B.H. King and W.D. Miller: ibid.
[3] R. Pastel, P. Geiser, E. Nadgorny and A. Struthers: ibid.
[4] D.B. Chrisey, A. Piqué, R.C.Y. Auyeung, R.A. McGill, R. Chung, S. Lakeou, P. Wu, J. Fitz-Gerald, H.D. Wu and M. Duignan: Surface Engineering, “Science and Tech. I”, ed. By A. Kumar, Y.-W. Chung, J.J. Moore and J.E. Smugeresky, (The Minerals, Metals & Materials Soc., 1999) pp.143.
[5] R.A. McGill, R. Chung, D.B. Chrisey, P.C. Dorsey, P. Matthews, A. Piqué, T.E. Mlsna, J.L. Stepnowski: IEEE Trans. On Ultrasonics, Ferroelectrics and Frequency Control 45, (1998) 1370.