Pico- to nanosecond pulsed laser-induced forward transfer (LIFT) of silver nanoparticle inks: a comparative study
- 159 Downloads
Silver nanoparticle inks are among the key functional materials used in printed electronics. Depositing it by laser-induced forward transfer remains a challenging task because the non-linear rheological nature of these inks narrows the range of the laser processing parameters. Understanding, therefore, the influence of the laser parameters on the ejection dynamics and deposition quality is of critical importance. The influence of the laser pulse duration from pico- to nanosecond-laser-induced jet dynamics was investigated using time-resolved shadowgraphy imaging. Jet speed and surface area analyses showed that in the lower laser fluence level range, picosecond pulses induce higher surface area ejections which propagate at higher velocities. As the laser fluence levels were increased, the difference in jet velocity and surface area evolutions narrows. Deposition analysis showed a similar behavior with lower transfer thresholds and larger depositions at lower fluence range when picosecond-laser pulses were used.
The field of printed electronics has been actively explored for the last couple of decades [1-4]. During this period, a number of printing techniques, such as screen printing , inkjet printing  and jet printing , were optimized for application in display technology [8, 9], RFID tags [10, 11], bio-engineering  and sensing [13, 14]. While these established techniques have been pushed to their technical limits, there is an industrial need for higher printing resolution and increased printing speed. As well, current printing techniques lack flexibility and/or versatility, e.g. screen printing is characterized by high yield, but it is a mask-based approach [5, 15]; inkjet printing is digital, but cannot be applied to highly viscous materials ; jet printing can operate with viscous materials, but is relatively slow . The increasing needs for flexibility and versatility are driving research to develop a printing technology which is digital, allowing for printing of micro-deposits with high yield and is applicable to the wide range of materials. Laser-based printing techniques like laser-induced forward transfer (LIFT) provide a potential solution to these demands [18, 19].
Next to the laser fluence, one of the most dominant laser parameters, in any process that involves laser–material interaction, is the duration of the laser pulse. Prior works show that LIFT is commonly executed using laser pulse durations ranging from femtoseconds (fs) to nanoseconds (ns) [42-44]. It was demonstrated that even a CW (continuous wave) laser can be successfully applied to LIFT printing . Prior works studied the LIFT ejection dynamics of gold using laser pulses in the nanosecond and picosecond (ps) range [22, 46]. It was found that for metal films, the pulse duration played a critical role in the transfer regime. A solid transfer regime was observed for picosecond pulses where nanosecond pulses exhibited melted droplets. Previously, we investigated laser fluence threshold levels at which donor material is ejected from the carrier, when using ps- and ns-laser pulses, by analyzing LIFT-printed silver nanoparticle ink depositions . We observed that the fluence threshold is lower when using the ps-laser pulses, compared to ns pulses. This difference was attributed to the additional heat diffusion energy loss to the carrier substrate and the surrounding donor material in case of ns LIFT.
However, no comparative study between ps and ns LIFT was carried with respect to the ejection dynamics—i.e. size and velocity of ejections—of shear thinning, silver nanoparticle based, inks. This paper compares the ejection dynamics of picosecond versus nanosecond LIFT, focusing on the velocity and surface area evolutions in time. The LIFT dynamics study is complemented by the deposition post-analysis, comparing volumes and the diameters of ps and ns LIFT-printed deposits.
2 Theoretical aspects of ejection dynamics
3 Materials and methods
Donor layer thickness measurements
3.2 Experimental setup
Two Nd:YAG-based, frequency-tripled laser sources were used. Both lasers emit a wavelength of 355 nm. A Coherent AVIA-355–4500 laser source was used, emitting laser pulses of tp = 30 ns, while a Coherent Talisker 355–4 produces pulses of tp = 15 ps. The intensity profiles of both laser sources are Gaussian.
A piccolo AOT-YVO-1 SP (InnoLas, Germany) emits 80-ps-laser pulses at 532 nm (green) wavelength and was used to excite fluorescent rhodamine 6G dye. The latter then emits broad bandwidth light, used to illuminate the LIFT jet. A light homogenizer, composed of a condensing and collimating lens pair (ACL50832U-A, LA1401-A, by Thorlabs GmbH, Germany) with the light diffuser (#47–995, by Edmund Optics, Germany) in between, evenly distributes emitted light and focuses on the plane where LIFT jet is induced. The emitted light, which is not “blocked” by the jet, is captured by the microscope objective and directed to the CCD camera (Prosilica GC1380, Allied Vision, Germany).
A pulse generator is used to trigger single UV pulses of the LIFT laser source, while the delay generator controls the delay time between the latter and the green illumination laser pulse. By scanning the delay, time-resolved evolution of LIFT jet formation can be recorded. The shadowgraphy imaging setup is capable of capturing one image per LIFT event. Therefore, after the image is captured, the carrier substrate is moved, and a next laser pulse is focused on a fresh section of the donor layer position.
4 Results and discussion
In this section, the comparative results of the LIFT ejection dynamics and deposition dimensions are presented. The velocities as well as the surface area evolutions of ps- and ns-laser-induced jets are compared. Depending on the main LIFT parameters (laser fluence, donor thickness and viscosity of the donor material), the time interval when the jet is in contact with the receiver substrate can range from hundreds of microseconds up to milliseconds. Compared to this time range, LIFT jet formation (bubble expansion, bubble collapse and jet elongation) takes place during a relatively short period, and ranges up to tens of microseconds. Since we are only interested in the LIFT ejection dynamics, no receiver substrate was used focusing only on the high-pressure bubble expansion and collapse.
After the time-resolved imaging of the LIFT dynamics, a receiver was placed at HP = 500 µm to transfer deposits and study their dimensions. The volumes and the diameters of the deposits transferred using ps- as well as ns-laser sources are compared.
4.1 Ejection front velocity analysis
Figure 5 shows that ps-laser-induced bubbles develop faster than those induced by the nanosecond pulses. This difference is significant at low pulse energies (1–5 µJ), but negligible for jets induced by pulse energies over 8 µJ. This is attributed to the combination of two effects—heat diffusion during the absorption of the laser pulse and the shear thinning effect of the silver ink, that is, during the bubble expansion, the governing force originates from the difference between the (atmospheric) pressure outside the donor material P0 and the gas pressure inside the bubble PH. The absorption volume—i.e. the volume where the laser energy is absorbed—is defined by the diameter of the laser beam focused on the donor surface and the optical penetration depth of the laser radiation, which depends on the material and the laser wavelength.
4.2 Ejection surface area analysis
For visual convenience, the horizontal axis representing the ps-laser-induced ejections on the left side of the graph is mirrored with the respect to the ns-laser data on the right. It is observed from Fig. 8 that, in the case of ps-LIFT, the spread in the data is larger than in the case of ns-LIFT. This is attributed to limitations in the experimental setup. It was found empirically that, during the laser-induced bubble growth phase and the beginning of the collapse phase, the temporal surface area evolution follows a second-order polynomial function. However, as the bubble collapse slows down and the bubble turns into a jet (around 8 µs in Fig. 8), the jet surface area evolution deviates from this quadratic function. Therefore, second-order polynomial fitting was applied up to this time instance.
From Fig. 8 it is evident that when lower pulse energies are applied, the maximum surface area is higher in case of ps-laser pulses, than in case of ns pulses. This means that in the lower pulse energy range, a picosecond pulse is able to eject more material than a ns pulse. However, as higher pulse energies are applied, i.e. higher than 5 μJ, the difference of the maximum surface area of the ejection induced by picosecond and nanosecond pulses becomes negligible. This result coincides with that presented in 4.1 and could also be reasoned by the heat diffusion analysis and the shear thinning of silver nanoparticle ink.
4.3 Deposition morphology, diameter and volume
This section compares the diameter and volume of silver nanoparticle ink depositions printed using ps- and ns-laser pulses. The deposits were printed onto the receiver from two donor layers having different thicknesses HD, namely 7 µm and 35 µm. A printing distance HP of 500 µm was selected. A laser spot diameter of 60 µm was used, allowing a fluence level scan up to 0.8 J/cm2. For statistical analysis of the reproducibility, LIFT deposition experiments were repeated ten times for each set of the parameter combination. Only cases with all ten successful depositions were analyzed. The lower laser fluence level cases, where not all jets reached the receiver substrate, were disregarded.
Compared to a donor layer thickness of 35 µm, when a thinner donor layer (here HD = 7 μm) is used, the range of the laser fluence levels, where clean depositions are observed, is smaller, see Fig. 11. There are three main reasons, why thinner donor layers exhibit unclean depositions. First, if jet is propagating at a velocity higher than the thin(er) donor layer can “feed”, the jet thickness reduces too rapidly, breaking up before reaching the receiver substrate. As a result, a droplet (or multiple droplets) of the donor material falls on the receiver substrate. Second, if the laser fluence level is further increased, the jet becomes turbulent, resulting in even more debris. Turbulent jetting regime is a known phenomenon and was reported by Boutopoulos et al. . Last, further increasing the laser fluence level results in a pressure PB rupturing the bubble, spraying many small droplets onto the receiver substrate.
Figure 10 shows that the fluence threshold at which successful depositions occur is lower in case of ps-LIFT than for ns-LIFT. In addition, for ps-LIFT, the diameter and volume of depositions at low-fluence levels are lager. However, when the laser fluence level reaches 0.5 J/cm2, differences in deposition dimensions between ps-LIFT and ns-LIFT become negligible, see Fig. 11. This 0.5 J/cm2 fluence level matches with the one where the difference between the speed of the ps-induced and ns-induced jets diminishes as shown in Fig. 6. For thinner donor layer (7 µm), however, the effect could not be seen because higher fluence levels induced satellite droplets, turbulent jets and material splashing, resulting in debris on the receiver.
In this work, we used time-resolved shadowgraphy imaging to compare the dynamics of silver nanoparticle ink ejections, induced by ps- and ns-laser pulses. We found that at relatively low-fluence levels (< 0.5 J/cm2), the velocity of jets induced by ps-LIFT is higher than those induced by ns-LIFT. Also, the maximum surface area of LIFT ejections induced by ns pulses is smaller than the ones induced by ps-laser pulses. This was attributed to the lower thermal diffusion length in case of ps pulse, leading to a more confined initial bubble and higher internal pressure. In the case of higher fluence levels (> 0.5 J/cm2), the difference in jet velocity and surface area diminished. This observation coincides with the reduced impact of thermal diffusion on the initial bubble pressure and the drop in viscosity of the silver nanoparticle ink at higher fluence levels. Analysis of the deposition diameter and volume showed a clear correlation with the time-resolved dynamics analysis, where ps-LIFT jets showed lower deposition threshold and larger depositions at lower laser fluence levels compared to ns-LIFT.
With this shadowgraphy imaging study, we provide new insights into understanding the role of the laser pulse duration from ps to ns on the LIFT process of silver nanoparticle ink. Whereas pulse duration is a major parameter in the laser–material interaction for solid film-based LIFT, in the case of highly viscous inks, where a gas bubble is generated, pulse duration does not play a significant role. These results confirm that deposits transferred by picosecond-laser system can also be obtained using more cost-effective nanosecond-laser sources.
This work is part of the research project, titled “High-resolution deposition of high-viscosity materials using Laser Induced Forward Transfer” (Hi-LIFT, Project Number 14641), which is (partly) financed by the Dutch Research Council (NWO). The authors would like to thank Merijn Giesbers and Edsger Smits for inspiring discussions.
- 3.V. Subramanian, J. B. Chang, A. De La Fuente Vornbrock, D. C. Huang, L. Jagannathan, F. Liao, B. Mattis, S. Molesa, D. R. Redinger, D. Soltman, S. K. Volkman, and Q. Zhang, ESSCIRC 2008 - Proc. 34th Eur. Solid-State Circuits Conf. 17 (2008).Google Scholar
- 7.S. Sanaur, A. Whalley, B. Alameddine, M. Carnes, and C. Nuckolls, Org. Electron. Physics, Mater. Appl. 7:423 (2006).Google Scholar
- 11.A. Blayo and B. Pineaux, 27 (2006).Google Scholar
- 18.M. Morales, D. Munoz-Martin, A. Marquez, S. Lauzurica, and C. Molpeceres, Laser-Induced Forward Transfer Techniques and Applications, Second Edn (Elsevier Ltd., 2017).Google Scholar
- 21.M. Zenou, A. Sa’Ar, and Z. Kotler, Small 11, 4082 (2015).Google Scholar
- 22.R. Pohl, C. W. Visser, G. W. Römer, D. Lohse, C. Sun, and B. Huis in ’T Veld, Phys. Rev. Appl. 3, 1 (2015).Google Scholar
- 31.K. S. Kaur, R. Fardel, T. C. May-Smith, M. Nagel, D. P. Banks, C. Grivas, T. Lippert, and R. W. Eason, J. Appl. Phys. 105, (2009).Google Scholar
- 37.A. Kamyshny, J. Steinke, and S. Magdassi, 1 (2011).Google Scholar
- 46.R. Pohl, C. W. Visser, G. R. B. E. Römer, C. Sun, A. J. Huis In’T Veld, and D. Lohse, J. Laser Micro Nanoeng. 10:154 (2015).Google Scholar
- 47.J. Miksys, G. Arutnov, M. Giesbers, and G. Römer, in Proc. LPM2018 (2018), pp. 1–4.Google Scholar
- 49.M. Duocastella, J. M. Fernández-Pradas, J. L. Morenza, and P. Serra, J. Appl. Phys. 106, (2009).Google Scholar
- 50.S. A. Mathews, R. C. Y. Auyeung, H. Kim, N. A. Charipar, and A. Piqué, J. Appl. Phys. 114, (2013).Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.