THE POTENTIAL OF PLASMA PRINTING
Electrical discharges – or plasmas – have fascinated mankind for centuries, because of their aesthetical beauty and uncontrollable behaviour. It is only for a few decades that plasmas are being used abundantly in industry for a myriad of applications, mostly in the controlled environment of vacuum chambers. In recent years it has become possible to operate plasmas in atmospheric pressure in a controlled manner, opening up the possibility of plasma treatment of surfaces that cannot withstand heat and/or are difficult to use in vacuum. InnoPhysics has developed an atmospheric plasma source that can deliver plasma in a patterned manner, such that the benefits of digital printing now apply to plasma treatment as well. Instead of ink droplets, dots of activated gas are being printed. Presently, the main target market is plastic electronics.
“… this stream often divides itself into a variety of beautiful rivulets, which are continuously changing their course, uniting and dividing again in the most pleasing manner …”
Joseph Priestley, History of Electricity, 1769
Plasma
Plasmas are gases in which a fraction of the atoms or molecules is ionised. One can achieve this by simply heating a gas to a sufficient temperature, but it is more effective to apply electric fields and couple energy directly into the charged species that for a large part define the characteristics of a plasma: electrons and ions. Especially electrons are, due to their low mass, easily accelerated in an electric field. When they collide with neutral atoms or molecules, they transfer some of their energy to these heavier species. This can result in excitation or in ionisation of the species. In the case of excitation, the energy is often released again in the form of light (giving the plasma its appearance) or it stays available for chemical reactions. In the case of ionisation, another electron and an ion are created and thereby the plasma sustains itself [1].
An important parameter in the physics of plasmas is the number of atoms that an electron encounters in its way through an electric field from one electrode to the other; this is easy to understand if one considers that electrons multiply themselves by ionisation. For that they need to pick up sufficient energy between collisions and meet the right number of atoms. The product of pressure and distance is the quantity that describes this. If it is too low, there are not enough atoms on the electron’s path to be ionised. If it is too high, the electron cannot pick up enough energy between collisions to ionise and the plasma will be hard to ignite as well. In the Paschen curve [2] shown in figure 1, the minimal voltage is shown that is required to ignite a plasma between the canonical parallel plate geometry often found in textbooks [3]. This breakdown voltage has a minimum at a specific value of pressure times distance and it is usually around this value that plasma sources are designed. If one considers the numbers, it can be seen that the convenient engineering sizes of some centimetres require sub-atmospheric pressures and that if one wants to operate plasmas at atmospheric pressures, one needs to decrease the distance between electrodes to tens of micrometers.
Figure 1: The Paschen curve describes the breakdown voltage of a discharge between two parallel plates, as a function of the product of the distance between the plates and the pressure of the gas [2,3].
Of course, one can choose to operate a discharge on the right hand side of the minimum of the curve by applying a sufficiently high voltage. This has the downside that the plasma development becomes uncontrollable: at atmospheric pressure small filamentary discharges develop that are reminiscent of lighting, with their destructive effect on contacting surfaces. There is one trick that can mitigate this: introducing a dielectric barrier between the two electrodes. Any current that runs through the discharge from one electrode to the other hits this surface, charges it and thereby counteracts the electric field of the electrodes. While the field diminishes, the discharge dies out before it can become destructive. So called dielectric barrier discharges (DBDs) have become a common design for plasma sources that operate at atmospheric pressure. The materials, manufacturing capabilities and power supply technology of recent years have allowed large steps to be made in the development of well controlled atmospheric pressure gas discharges [4]. As an additional bonus, these plasmas can be chemically active, while remaining relatively cold (i.e. some tens of degrees centigrade).
Applications of plasma treatment
By means of plasmas, we have a way of creating specific molecules efficiently, or to deliver chemical energy to surfaces without heating them. This was realised a long time ago and plasmas found their applications in deposition of layers and in etching. Both are essential steps in the production of semiconductors nowadays. By being able to create controlled plasmas at atmospheric pressure, the versatility of plasmas can be applied to materials that cannot stand vacuum (such as biological tissues) or to materials that are unwieldy for vacuum equipment (such as big rolls of plastic sheets). As such, plasmas are finding applications in the emerging field of plasma medicine [5], in which they are tested for disinfection and stimulation of wound healing. Already more established is their capability of functionalisation of surfaces: the wettability of substrates can be altered substantially and surfaces can for example be prepared before they receive a layer of inks or glue that would otherwise not hold [6]. In the textile industry plasmas are used as a pre-treatment before dyeing, or, using different chemistries, to render textiles water repellent.
Surface modifications can be categorised in three types: activation, etching and deposition (cf. figure 2). The type of modification realised with a plasma depends largely on the gas in which the plasma is ignited. It is therefore possible to make tools that can do any of the three of the operations by changing the gas supply and choosing the right settings for the power supply. In the following we will focus predominantly on activating the surface, although the equipment described is indeed not limited to this application.
Figure 2: Process capabilities divided in three categories.
Changing, activating and functionalising surfaces are terms that are typically used when one talks about changing the surface energy of materials, or adding molecular groups to surfaces in order to change the adhesive properties of specific chemical compounds to the surface. Plasmas achieve this via the reactive particles that are created in the collisions that happen between electrons, ions and neutral atoms and molecules. If oxygen and nitrogen are present, many radical species can be formed, such as O3, OH radicals and ions, H3O+, singlet oxygen, NO species, and many, many more [4]. The chemically energetic particles are capable of opening bonds of surface molecules, establishing new bonds and assembling molecules at the surface of the material.
In general, all these activation effects add to the surface energy of the material. This surface energy is important in the way the surface interacts with other, solid or liquid surfaces. If little energy is available, creating a joint interface with another material is generally not energetically favourable: the individual components are energetically better off on their own and this shows for example by liquids forming beads on the low energy surface, rather than wetting it. If a surface is treated, bonds are broken and the resulting surface energy is larger, it might become favourable for liquids to create a common interface with the treated surface, thereby wetting it. In figure 3 this is illustrated: in the middle one can see a trace where the plastic has been plasma treated and where the fluid spreads. Around it, the liquid forms beads. Surface energy and wettability are quantified by so called water contact angle (WCA) measurements: the angle between the solid surface and the meniscus of the droplet are measured and if one knows the surface energy of the test fluid, one can calculate the surface energy of the solid via Young’s law. High surface energies lead to low contact angles and vice versa.
Figure 3: Illustration of the wetting of a surface. In the middle a
Contact angles do not tell much about the specific surface chemistry though. If one wants to know this, one can use ellipsometry to determine the energy of molecular bonds at the surface and x-ray photoelectron spectroscopy (XPS) for the stoichiometrical composition of the surface. Atomic force microscopy (AFM) can be used to quantify the texture, which is another property that can have its effect on the contact angle. Such more detailed data about the surface can become important if one has certain specific applications in mind. For example, the presence of –NH groups at a surface is very interesting from the viewpoint of living cell adhesion to the surface and therefore their density an important quantity in bio-engineering.
Equipment
Most atmospheric plasma treatment systems that are available on the market are either meant to homogeneously treat surfaces (such as is common for textiles) or generate plasmas of a centimetre in diameter (for example for wound healing). InnoPhysics has developed a plasma printing technique that combines plasma treatment with digital printing. Instead of the droplets that are for example delivered at a specific place on a substrate by an inkjet printer, a plasma is briefly ignited at the desired spot. Referring to the Paschen curve model in figure 1 again, one can see that there are three parameters that determine whether a plasma ignites or not: the gas pressure, the distance between electrodes and the applied voltage. The gas pressure is typically hard to control on small time-scales. The applied voltage is reasonably easy to control on a small time-scale, so this could be a convenient parameter to vary. We chose the unconventional third: the distance between the electrode and the substrate.
The technology behind this can be most easily explained by referring to traditional impact printing techniques in which a needle is mechanically displaced, impacts an ink ribbon and thereby transfers ink to paper. By moving a needle down towards the substrate, the Paschen curve is crossed and a plasma can ignite. This is illustrated in figure 4.
Figure 4: Schematic illustration of the working principle behind plasma printing technology: a plasma is ignited by varying the distance between a needle and the substrate behind which the counter electrode is placed.
Figure 5: Photograph of the plasma print head mounted in Pixdro’s LP50 desktop printer [7].
We chose to make this technology first available in a laboratory set-up. For this we chose OTB-Solar’s Pixdro LP50 inkjet tool [7] as a platform. The modification kit that InnoPhysics provides for this platform consist of a print head assembly, substrate table and high voltage power supply. With these additions, the LP50 delivers the functionality for plasma printing on the scale of R&D level development (see figure 5).
The print head consist of two arrays of twelve needles that can be independently moved towards the substrate. The frequency at which this happens is typically half a kHz per needle. Between the substrate table and the print head an alternating potential is applied with a frequency in the order of tens of kHz. When a needle moves close enough to the substrate table (when d in the Paschen curve decreases), the electric field becomes sufficiently large for the plasma to ignite. Figure 6 shows side and bottom views of the plasmas when all needles are activated. Precise needle control and precise voltage control allow for accurate plasma treatment. In figure 7 the effect of these two parameters on the energy delivered per plasma pulse are plotted.
Since the print head is mounted on an XY-stage, motion of the head with respect to the substrate allows to switch the plasmas at specific locations, thereby creating a pattern area that is plasma treated. While the LP50 system is ideal for research and development work, it does not have the speed required for most production processes. Knowing this, InnoPhysics is working on production scale equipment, which can be integrated in roll-to-roll processing tools.
Figure 6: Side view of a single row (12 needles) and bottom view of the dual row (24 needle) plasma print head simultaneously in action.
Figure 7: The energy dissipated in the plasma as a function of the gap distance.
Process capabilities
In the text above, it was already mentioned that plasmas can be used for a number of surface modifications: activation, etching and deposition. Below we give some examples of processes that can be demonstrated with the plasma printer. The first one is illustrated in figure 8. The substrate was a plastic foil with two layers on top. The first layer was hydrophilic and on top of it was a hydrophobic layer with a thickness of a few nanometres. By plasma printing in an environment of oxygen and nitrogen, the top layer was selectively removed (etched). As the figure shows, this can be made visible with a regular text marker pen: the ink only transfers to the substrate where the hydrophobic layer has been removed. This effect is highly durable and can be made visible time and time again. This process example has been developed in collaboration with the Holst Centre [8]. The idea behind it is to use the large surface energy contrast to selectively deposit functional materials by combining plasma printing with slot-die coating techniques. In this way, patterns of functional inks can be created in a controlled manner. This can be used for customisable printed, organic electronics, such as custom shaped thin film OLED devices. Being selective at the nanometre scale without damaging substrate with a maskless technique is relatively unique. Lasers are the next best alternative: In terms of feature size, lasers are still one step ahead of the plasma print technique. However, operating without damaging the substrate or creating debris in the process, and scaling towards production sizes at an acceptable cost of ownership is difficult to achieve with lasers.
Figure 8: After plasma printing on a substrate, the effect is often not visible, as it would be with inkjet printing. In this case, a large contrast was created between hydrophobic (original surface) and hydrophilic (treated surface), which can be made visible with a marker pen. The ink only sticks on the hydrophilic surface and shows the pattern that was printer.
The second example (cf. figure 9) illustrates the resolution that can presently be obtained with plasma printing. In the left side of the figure one sees two lines made visible with a marker again. These lines were printed on a regular plastic foil (i.e. not having the layer stack of the previous example). One can see that
Figure 9: On the left: different line widths realised by plasma printing. The visualisation is done in the same manner as in figure 5. Shown here are a line of
The third example, visualised in figure 10, shows how different plastics react to plasma treatment as a function of time. On the horizontal axis the treatment time of one location is given and on the vertical axis the surface energy, which has been calculated from contact angle measurement. One can see that surface energy increases of roughly 20 mN/m are obtainable for all the materials that were tried. This change depends strongly on the gas that is administered to the plasma region. For specialty gas mixtures, changes in surface energy of 40-50 mN/m have been demonstrated in the lab, which corresponds to roughly 70o in water contact angle differences between treated and untreated areas. Furthermore, the largest change happens in the first few seconds of treatment. In fact, faster than was anticipated, as otherwise we would have chosen the time intervals differently. Future work will show exactly how steep the change in surface energy is in time. The Functional Polymers group at the Fontys University of Applied Sciences [9] in
Figure 10: The change in contact angle for a number of different plastics as a function of the treatment time. One can see that the largest effect is obtained between the first two data points of each curve.
The fourth example (see figure 11) is a very preliminary result, but one that we would like to show for completeness. In this example, a precursor gas (hexamethyldisiloxane or HMDSO) was fed into the gas nozzle of the print head. This material is often used in the semiconductor industry to create oxide layers. Notice that the contact angle now changes from small to high; i.e. in the opposite direction as it did in the activation examples given before.
Figure 11: By feeding HMDSO in to the plasma region, methyl-rich silicon oxide deposition can be realised. This new top layer has a lower surface energy than the original surface and the droplet does not spread as well anymore.
Markets
As a last topic of this article, we would like to give a small outlook on the market that we see for this new technology. The main market presently targeted is plastic electronics. This is an expanding market with a large variety of different applications that use printing technologies for devices or components on polymer films such as organic LED’s, (organic) solar cells and displays, organic sensors, biomedical chips and RFID tags. For the volume production of many applications in printed electronics, very thin, patterned layers of (semi-)conducting and/or insulating polymers need to be created with high precision and extremely uniform thickness. Printing of such patterned layers can result in significant cost reductions compared to homogeneous deposition techniques that require subsequent patterned removal.
Accurate control of surface energies is required in order to achieve the desired thin film accuracies and uniformities. As described above, plasma printing enables maskless patterning of surfaces or coatings on thin, insulating substrates. It can thus be used for surface energy controlled slot die and/or inkjet printing but also for direct etching of thin organic layers. Many emerging applications demand hybrid manufacturing utilising both slot-die coated patterns and inkjet for which etching and activation in one machine are mandatory
Figure 12: This graph that shows approximately how plasma printing is positioned with respect other printing techniques. Please notice the discrimination between contact and contactless, which is important for fragile substrates, and between with mask and maskless, which has clear implications on the workflow of production.
Acknowledgements
Authors would like to acknowledge the following people for the fruitful and pleasant collaboration. Ronn Andriessen, Juliana Gabel and Jaap Lombaers from the Holst Centre. Jan Bernards, Martijn van Dongen, Renée Verkuijlen, Richard Janssen, Richard van Hout, Kevin van de Wiel en Merel Eland from the Fontys University. Peter Brier, Peter Diepens and Klaus Schiffer from Pixdro/OTB-Solar. Gerrit Kroesen, Guus Pemen and their colleagues from the Eindhoven University of Technology. And finally, all partners in the R&D programs HTT PrintValley, PiD Printing for Inkjet Applications and RAAK Pro Inkjet printing.
References
[1] J.S. Townsend. The Theory of Ionization of Gases by Collision. Constable & Company Ltd.,
[2] F. Paschen. Ueber die zum Funkenübergang in Luft, Wasserstoff und Kohlensäure bei verschiedenen Drucken erforderliche Potentialdifferenz. Wied. Ann., 37:69-96, 1889.
[3] Yu. P. Raizer. Gas Discharge Physics. Springer, Berlin, 1991.
[4] K.H. Becker, U. Kogelschatz, K.H. Schoenbach, and R.J. Barker, ed., Non-equilibrium Air Plasmas at Atmospheric Pressure. Institute of Physics Publishing Ltd., 2005.
[5] K.-D. Weltmann, ed., 3rd International Conference on Plasma Medicine (ICPM-3),
[6] K.L. Mittal, ed., Polymer Surface Modifications. VSP 2000.
[7] www.pixdro.com
[9] www.fontys.edu