Metal-pattern preparation based on selective deposition using soft organofluorine surfaces

Metal patterning is a critical process from academic study to various industrial mass-production fields, including electrode-pattern preparation^1–5 and electromagnetic wave absorbers that use metamaterials.^6–9 As a conventional metal-patterning process, vacuum deposition with a shadow mask has been very widely used. Unfortunately, it suffers from such potential problems as the difficulty of forming a precision shadow mask with a large area and thermal deformation due to high radiant heat from the evaporation source metal with a high melting point, including Ag, Cr, and Ni.¹⁰ 

Selective metal-vapor deposition is the ability to form metal patterns by maskless vacuum deposition. Metal atoms are adsorbed on a patterned layer with a metal-desorption property and desorbed from the surface; the atoms that reach the surfaces without the metal-desorption layer are deposited to form a metal film.^11–13 The core phenomenon of selective metal-vapor deposition is the desorption of metal atoms from organic surfaces. When a pattern with a metal-desorption property is formed on the surface by a certain method and the metal is evaporated on it without a shadow mask, a desired metal pattern can be easily achieved.

Selective metal-vapor deposition was discovered for Mg, Pb, Mn, and Zn using photochromic diarylethenes^11,14,15 and has been extended to Sn, Ca, Al, In, and Ag using oily polydimethylsiloxane (PDMS).^16,17 Although PDMS exhibits desorption of various metals, it requires a thick film of several hundred nanometers or more, and therefore, forming fine patterns is difficult due to its fluidity. Hatton et al. showed that 1 μm-level Ag and Cu patterns can be formed using a contact microprinting method with thin fluorine-based organic materials.^18–21 We also showed that surfaces with a small dispersion component in surface free energy and a flexible surface-molecular structure effectively enhance the desorption of various metal atoms, and we revealed that a material in which an ether moiety is incorporated in a perfluoroalkyl chain, such as perfluoropolyether (PFPE), is more effective than a molecule with simple firm perfluoroalkyl chains.²² PFPE has a flexible structure, and the surface PFPE molecule move easily.^23–28 

Three methods have been proposed for preparing metal-desorption patterns, and the following are the advantages and potential problems of each. The best method is used based on the requirements and cost of the device to be manufactured, compatibility with conventional processes.

1. Photoreaction using laser scanning or a photomask.^11,14,15

   • Advantages: high resolution in the optical diffraction limit; the highest degree of freedom in the pattern shape.

   • Problems: low carrier injection efficiency when forming an electrode pattern for organic electronic devices and weak adhesion due to the presence of an organic metal-desorption layer between the metal layer and the substrate.

2. Printing.^16,18,21

   • Advantages: high degree of freedom in a pattern shape; high resolution when microcontact printing is used.

   • Problems: potential damage underlying an organic layer when the method is applied to organic electronics devices and incompatibility with a continuous vacuum-deposition process in the industry. (The device must be removed from the vacuum chamber and set in the ambient atmosphere during the printing process.)

3. Vacuum deposition using a shadow mask.^29,30

   • Advantage: less heat damage to a shadow mask compared to directly forming a metal pattern through a shadow mask; complementary patterns of direct metal-pattern formation with a shadow mask, such as the formation of isolated metal nondeposition areas; and good consistency in the continuous vacuum-deposition process.

   • Problems: same problems as vacuum deposition with a shadow mask: limited resolution and difficulty preparing a large area precision shadow mask.

In this paper, we demonstrate metal patterning for a variety of metal species based on methods 1 and 2 using PFPE-based materials. We report the extensibility to various metal species and discuss the possibility of fine pattern formation. In particular, we introduce patterning by maskless metal deposition of such metal species as Cr and Ni, which have high melting points and low vapor pressures. Our results will be widely applicable to metal-pattern formation in various fields, including electronics and photonics.

We previously reported that fluorine-based materials in which ether groups are incorporated in alkyl chains, especially commercially available mold-release agents, are effective for exhibiting desorption of various metals.²² However, since these materials are coated on substrates in a solution state, they cannot be used for vacuum deposition. In this study, we used a PFPE-organofluorine material called KY-1901 (a mold-release and antifouling agent manufactured by Shin-Etsu Chemical Co., Ltd.) with a silane coupling site and applicability to the vacuum deposition process and printing. KY-1901 is a chemically very stable material that does not form reactants with any of the metal species used in this study. Since KY-1901 is supplied as a solution in a fluorine-based solvent, we dropped 10–20 μl of it into a molybdenum boat in advance, left it in a room for about 5–10 min to evaporate the solvent, and placed it in a vacuum chamber. We adjusted KY-1901’s evaporation temperature around 250 °C. When a pattern was formed using a stamp, the supplied solution was used without changing it.

The glass substrates used for vacuum deposition were subjected to UV-ozone treatment after being ultrasonically cleaned with acetone for 15 min. Vacuum deposition for the metals and KY1901 was carried out using thermal evaporation under 1 × 10⁻³ Pa at room temperature. The film thickness during deposition was controlled by a quartz crystal thickness-monitor (Inficon XTM/2) and calibrated by interferometrically measured film thickness. The absorption spectra of the metal films were measured with an ultraviolet/visible spectrophotometer MultiSpec-1500 manufactured by Shimadzu Corporation, and the droplet contact angle for the surface-energy evaluation was measured using a device we made.

First, we studied the thickness dependence of the deposited PFPE-organofluorine film for Ag desorption. Figure 1(a) shows the PFPE-organofluorine films after Ag deposition, and Fig. 1(b) shows the corresponding absorption spectra. The numerical value without an index in the graph represents the PFPE thickness. The deposition rate and the film thickness of Ag were 0.16 nm/s and 8.7 nm on the glass substrate (thickness-on-glass) as a control. Figure 1(c) shows the energy dispersive X-ray spectroscopy (EDXS) of the Ag-deposited samples on the glass and PFPE (4.6 nm) surface. With an Ag-film thickness-on-glass of 8.7 nm, when the PFPE-film thickness was 4.6 nm or more, no Ag at 2.125 keV was determined by EDXS, indicating excellent desorption properties.

Next, the PFPE-film thicknesses were set to 11 and 26 nm, Ag was deposited even more thickly, and the desorption property was similarly examined [Figs. 2(a) and 2(b)]. In all the samples, the deposition amount was greatly reduced compared to the glass substrate, and a good desorption property was observed. However, the 11 nm-thick PFPE-film sample was slightly yellow and began to deposit when the Ag evaporation increased to 100 nm thickness-on-glass. On the other hand, the 26 nm-thick film sample showed excellent desorption properties even with an Ag-film thickness-on-glass of 100 nm; no absorption by Ag was observed.

To clarify the surface properties of the PFPE-film that govern the desorption of the Ag atoms, we investigated the dependence of the droplet contact angle on the PFPE-film thickness. Distilled water (H₂O), methylene iodide (MeI), and normal-hexadecane (Hd) were used as droplets [Fig. 3(a)]. As the film thickness increased from 2 nm, the CA of the three droplets increased to a maximum at 10 nm, a result that indicates that the PFPE film does not completely cover the surface when the film thickness is less than 10 nm.

The surface free energy was determined as follows: $γ S D$ = 11.4 mJ/m², $γ S P$ = 0.0 mJ/m², $γ S H$ = 8.1 mJ/m², and γ_S= 19.5 mJ/m². This surface exhibited sufficiently small γ^D for metal desorption. When the film thickness was further increased, a phenomenon was observed in which CA decreased with time immediately after dropping for all the liquids [Fig. 3(a)]. This indicates a change from Young’s condition on the solid surface to Neumann’s condition due to surface fluidity^33–37 [Fig. 3(b)].

Figure 4 shows the molecular model for the changes in CA. The monolayer of the rigid fluorinated alkyl chains is silane-coupled to the substrate [Fig. 4(a), experimental data of this typical hydrophobic coat using silane coupling are not displayed in this paper], the monolayer of the flexible PFPE-type molecules is silane-coupled to the substrate [Fig. 4(b), corresponding to a PFPE thickness of 10 nm], and a PFPE polymer was formed on the silane-coupled molecular layer [Fig. 4(c), corresponding to a film thicker than 10 nm]. The metal is more likely to desorb from the flexible PFPE monolayer than from the rigid perfluoroalkyl chain monolayer, and by forming a soft PFPE polymer film on it, desorption is enhanced based on the active motion of the surface PFPE chains at room temperature.²² The above results show that to obtain enhanced metal-desorption properties for Ag, the PFPE film must be 28-nm thick or more to form a flexible and soft outermost surface. In other words, Ag atoms are desorbed from the PFPE surface under suitable conditions; they neither form a film on the surface nor absorb the atoms into the film.

In the same way as for Ag, the dependence of the desorption properties was investigated for In, Cr, Ni, and Fe on the PFPE-film thickness. Figures 5(a) and 5(b) show the appearance and the absorption spectra of the samples after In deposition. PFPE-film thicknesses of 11 and 29 nm were prepared, and In was vacuum-deposited on them with four thicknesses-on-glass from 11 to 96 nm at a deposition rate of 0.16 nm/s. All the PFPE-films exhibited excellent desorption properties; no In deposition is compared to the glass substrates.

Figures 6(a) and 6(b) show the desorption property for Cr. When the Cr-film thickness-on-glass was thin (9 and 18 nm), no deposition was observed on the two PFPE films; when the thickness-on-glass was 30 nm, and Cr deposition was observed on the thin (11 nm) PFPE film. However, it was not deposited on the thick (29 nm) PFPE film, showing good desorption characteristics. Like Ag deposition, this result also indicates that PFPE thickness should be increased when depositing a thick metal film.

Next, we investigated the desorption properties of Ni and Fe, both of which are typical magnetic metals with high melting points and low intrinsic vapor pressures on the PFPE surface (Figs. 7 and 8). Although these metals have lower intrinsic vapor pressures than Ag, they showed good desorption characteristics regardless of the film thickness, and the light absorption of them is negligible after evaporation. Our previous study showed that metals with low intrinsic vapor pressure are less likely to be desorbed,³⁸ although these results indicate that we must investigate the metal-desorption property and focus on the interaction between the metal and the organic surface itself. Unfortunately, the corresponding physical property values on metal remain unknown.

Finally, we demonstrated metal patterning based on the desorption from the patterned PFPE film using method (3). Figure 9(a) shows examples of metal shadow masks for preparing PFPE-film patterns. The resolution of these shadow masks was at the millimeter level. The PFPE’s film thickness was set to 30–40 nm, which confirms high rate of desorption, and thickness-on-glass of the metals was 15–30 nm.

Figure 9(b) shows a variety of metal patterns obtained by maskless metal deposition on various PFPE patterns. The metal species corresponding to each pattern are indicated by element symbols. The samples were observed by transmitted white light, and therefore, the bright white areas correspond to the no-metal deposited areas. The following are the metal species for which metal-pattern formation was confirmed by this method: Ca, Mg, Yb, Sb, Bi, Pb, Mn, Sn, In, Ag, Al, Cu, Cr, Ge, Ni, Fe, and Co [Ca, Fe, and Co are not shown in Fig. 9(b)]. Metal-pattern formation was found to be applicable to a wide variety of metals.

Figure 10 shows another example for preparing metal patterns based on method (2). The PFPE patterns were made by printing (with a commercially available stamp) on glass substrates, and a variety of metals were evaporated onto the substrate with a PFPE pattern without a shadow mask [Fig. 10(a)]. Figure 10(b) shows a variety of metal patterns obtained by maskless metal deposition. The metal species corresponding to each pattern are indicated by element symbols. Since the samples were observed by transmitted white light, the bright white areas correspond to the no-metal deposited areas.

Next, we deposited PFPE using a fine shadow mask and investigated the resolution of the resulting metallic patterns. A shadow mask was a stainless-steel plate with 100 μm holes. Figure 11(a) shows Cr patterns directly formed using the shadow mask and by selective deposition. The samples were observed by transmitted light, and therefore, the dark areas indicate Cr-deposited regions. Since the Cr pattern obtained by selective deposition was a negative pattern to the direct deposition using a shadow mask, it is a complementary pattern. Such an isolated metal nondeposited pattern cannot be obtained by conventional vapor deposition using a shadow mask.

We next tested how the size of the metal pattern changes based on the size of the holes in the shadow mask [Fig. 11(b)]. After a PFPE-film pattern was formed using a shadow mask, Ag, Cr, and Ni were maskless vacuum-deposited, creating a metal nondeposited pattern larger than the shadow mask’s hole. This result suggests that the deposition of the PFPE material using a shadow mask spreads around the pattern on the substrate surface. Therefore, to obtain a metal pattern of a predetermined size, the shadow mask’s shape must be determined based on the spread. A potential effective alternative is to prepare a high viscous material instead of a liquid by designing the molecular structure of PFPE. A pattern with such an isolated metal nondeposited region is under research and development for applications as a transparent electrode for an organic light-emitting display.^29,30

We demonstrated that metal-pattern formation is possible by maskless deposition for various metal species using a vacuum-depositable and printable PFPE material based on selective metal deposition. This high-efficient metal-atom desorption is based on a low surface-energy dispersive-component and a flexible PFPE structure incorporating ether moieties. The PFPE material chemically bonds to the glass substrate by silane coupling to form a monolayer, but the ability to desorb the metal from the surface, which is a core phenomenon for selective metal deposition, was enhanced with films thicker than the monolayer. We showed that metal patterns can be formed by maskless deposition of various metals by selective deposition on PFPE patterns, obtained by deposition using a shadow mask or printing on a glass substrate. Metals that can be patterned by this method include not only Mg and Pb, but also such metal species with a high melting point as Ag, Cu, Cr, Ge, Ni, Fe, and Co. We also demonstrated maskless deposition formation of isolated 100 μm nonmetal-deposited areas. This method can be applied to such applications as electrode-pattern formation in electronics and electromagnetic wave absorption substances.

This work was partially supported by JSPS KAKENHI (Grant No. 21K05214).

The author has no conflicts to disclose.

Tsuyoshi Tsujioka: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Resources (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal).

The datasets generated during the current study are available from the corresponding author upon reasonable request.