Microfabrication of Ni-Fe Mold Insert via Hard X-ray Lithography and Electroforming Process

Abstract

In this research, a Ni-Fe mold insert for the efficient replication of high aspect-ratio microstructure arrays was fabricated via hard X-ray lithography and an electroforming process. For the X-ray exposure on a photoresist, a gold-based X-ray mask was prepared with conventional UV photolithography. The gold thickness was designed to be over 15 μm to prevent development underneath the absorber and to enhance the adhesion strength between the photoresist and substrate. By using the X-ray mask, a positive-type photoresist was selectively exposed to X-ray under an exposure energy of 4 kJ/cm³. Thereafter, the exposed region was developed in a downward direction to effectively remove the residual photoresist from the substrate. During the evaporation process, deionized water mixed with a surface additive prevented the bending and clustering of the photoresist microstructure arrays by lowering the capillary force, resulting in a defect-free mother structure for electroforming. Lastly, the mother structure was uniformly Ni-Fe electroformed on a conductive substrate without the formation of any pores or detachment from the substrate. Based on the proposed microfabrication process, a Ni-Fe mold insert with a 183 μm pattern size, 68 μm gap size, 550 μm height, 2116 microcavities and a hardness of 585 Hv was precisely manufactured. It can be utilized for the mass production of high aspect ratio metal and ceramic microstructure arrays in micro molding technologies.

1. Introduction

Recently, several studies have established and developed micro molding technologies for the mass production of metal or ceramic microstructure arrays with high aspect ratios. Micro-powder injection molding, hot embossing and UV embossing are broadly used to replicate metal- or ceramic-based microstructure arrays [1,2,3]. Many microstructure array components with highly specified dimensions and tolerance have been utilized for a wide range of applications, including piezoelectric composites [4], heat sinks [5] and microfluidic systems [6]. For the mass production process, the reverse-shaped mold insert should be precisely fabricated with micrometer-scale accuracy because it directly determines the pattern dimension and aspect ratio. Conventional micromachining processes including milling, drilling, electro-discharge machining, laser ablation and silicon etching have been introduced to prepare micro-scale mold inserts [7]. However, as the pattern and gap dimensions become smaller with higher aspect ratios, these methods encounter technical difficulties in manufacturing highly accurate dimensions due to the resolution limit of the mechanical tools involved [8]. In addition, it may consume a great deal of time as the number of potential patterns has increased to over 1000. Regarding accuracy and efficiency, new technology is required to prepare micro mold inserts with a target dimension matching microstructure arrays.

As an alternative method, hard X-ray lithography and electroforming was employed for the efficient fabrication of a micro mold insert. This novel process has been verified to obtain a micro mold insert owing to high accuracy within 0.2 μm and excellent sidewall roughness under 0.06 μm. The hard X-ray lithography utilizes an irradiated X-ray as a beam source to fabricate a base mother structure for the subsequent electroforming process. After X-ray exposure, the irradiated region is etched out to form a space for infiltrating the electroforming materials. However, the microstructure arrays of the mother structure may be bent and clustered due to the presence of a capillary force during evaporation. This phenomenon easily occurs when smaller micro-scale patterns and gap dimensions and higher aspect ratios are involved [9]. Next, the electroforming process is performed to produce a final mold insert. Based on the combined technologies, research on the fabrication of micro mold inserts has been examined. Galhotra et al. fabricated the electroplated mold insert using pure nickel material [10]. Bacher et al. also conducted the fabrication process of mold insert with nickel material [11]. Cao et al. used the pure nickel to obtain a mold insert with a diameter of 200 μm and hole center to center spacing 700 μm [12]. Kim et al. electroplated the nickel mold insert with a height of 300 μm and aspect ratio of 15:1 in various shapes [13]. Kim et al. manufactured the nickel mold insert of the microlens arrays showing various diameter 200 μm, 300 μm and 500 μm [14]. Park et al. utilized the Ni material for fabricating a separated mold insert with a diameter of 200 μm and 5 microcavities [15]. Marques et al. shaped the reverse shaped Ni mold insert with 500 μm in height [16]. Kim et al. fabricated a sectioned micromold system with a diameter ranging from 0.2 to 1.5 mm [17]. Lee et al. developed a separated mold system for the replication of five polymer micropillar arrays with a 200 μm dimension and an aspect ratio over 5 [18]. Katoh et al. made a nickel mold insert with a 5 μm dimension and structure height of 15 μm [19]. Most of these studies achieved a pattern dimension in micro-scale and an aspect-ratio over 3 using pure nickel material and other materials. However, the gap sizes among the microstructure arrays were much larger than the pattern sizes, resulting in low structural density. As previously mentioned, the collapse of mother microstructure arrays easily occurs as a result of capillary forces. To enhance the material properties in ceramics and metals, a gap size smaller than the pattern size is in high demand for applications in the medical field and defence products.

In this study, we proposed and developed a method for fabricating a Ni-Fe mold insert with highly aligned microcavities for the mass production of metal- or ceramic-based high aspect ratio microstructure arrays. The X-ray exposure condition was analyzed and optimized based on the amount of dose on the absorber and photoresist. With the help of a surface additive, the photoresist microstructure arrays did not display bending and clustering with the help of surface additive, and the residual photoresist was effectively removed by downward direction development. On the mother platform, the Ni-Fe material was gently electroformed without any formation of pores or detachment from a substrate. As a result, a Ni-Fe mold insert with highly aligned microcavities was successfully produced with a 183 μm pattern size, 68 μm gap size, 550 μm thickness, 2116 holes and a hardness of 585 Hv. It can be applied and utilized in micro molding technology fields that require metal and ceramic microstructure arrays with high aspect ratios.

2. Materials and Methods

2.1. Hard X-Ray Lithography Process

Before the electroforming process, a precise mother structure of the final Ni-Fe mold insert was prepared by hard X-ray lithography. First, an X-ray mask was delicately fabricated by conventional UV photolithography and subsequent gold electroplating for selective X-ray exposure. A 200 μm-thick polyimide film on a silicon wafer was used as a basic membrane, and it was deposited by a thin seed layer of 20 nm chromium and 100 nm gold using an electron beam evaporator. For micropatterning, a negative-type photoresist (SU-8 3010, MicroChem, Westborough, MA, USA) was uniformly spin-coated on the seed layer at 700 rpm for 30 s, resulting in a 23 μm thickness. The seed layer was then exposed to a UV source for 17 s and followed by the development process to etch out the unexposed area. Next, the SU-8 microstructure was gold electroplated under a current density of 1 mA/cm³ for 5 h using a customized electrochemical cell which was included in the mixture of deionized water and gold electrolyte (KAu(CN)₂; LT Metal, Seoul, Korea). As a result, the 15 μm-thick gold layer was electroplated as X-ray absorber. For hard X-ray lithography, poly (methyl methacrylate) (PMMA; Goodfellow, Cambridge, UK) was selected as a positive-type photoresist because of its excellent contrast and high stability. The PMMA with thickness of 1100 μm was sliced and annealed for dehydration and release of residual stress. As the base substrate, a polished graphite sheet with a thickness of 2000 μm was spin-coated four times using a liquid-type PMMA C9 before bonding with the PMMA. The graphite sheet exhibits a high adhesion ability with PMMA as well as a low back-scattering effect during X-ray exposure [20]. With these advantages, the graphite was adhered to the PMMA photoresist using methyl methacrylate (MMA; Sigma Aldrich, St. Louis, MO, USA).

By using the prepared X-ray mask, the top surface of the PMMA on the graphite was exposed to synchrotron X-ray irradiation. To remove any air gaps, the mask was attached securely to the sample using a vacuum pump. The exposed dose on the bottom surface was determined to be 4 kJ/cm³, which is sufficient to fully eliminate the exposed region of PMMA [21]. A double-mirror system with an angle of 0.4° and an 18 μm aluminum filter was used to cut off an excessive high-energy X-ray and to attenuate a low-energy X-ray.

Table 1. Experimental setup condition of the 9D beamline.

Setup Condition	9D Beamline
Electron energy (GeV)	3.0
Beam current (mA)	360
Gold thickness (μm)	15
Exposure energy (kJ/cm3)	4.0
Mirror angle (°)	0.4
Aluminum filter thickness (μm)	18

summarizes the experimental setup conditions of the 9D beamline in PLS (Pohang Light Source). Thereafter, the selectively exposed area was dissolved using a GG-developer, which was a mixture of water and three different organic solvents (15 vol.% deionized water, 60 vol.% 2-(2-butoxyethoxy) ethanol, 20 vol.% tetrahydro-1, 4-oxazine, and 5 vol.% 2-aminoethanol). The exposed top surface was laid downward to improve the development rate, and the developer was gently agitated with a magnetic stirrer for 6 h [22]. The rinsing solution consisting of water and an organic solvent (20 vol.% water and 80 vol.% butoxy-ethoxyethanol) was used for 3 h to clean the residual GG-developer as well as the residual PMMA resist in the gaps of the microstructures. Finally, the dissolved region was rinsed with mixture of deionized water and surface additive (BYK 3440; Uni Trading Corporation, Gwangju, Korea) for 12 h. For the microstructure arrays, the dimensions including pattern size, gap and thickness were characterized using an optical microscope.

2.2. Metal Electroforming

The metal mold insert was fabricated by Ni-Fe electroforming onto the PMMA microstructure arrays on the graphite substrate. In the present study, a Ni-Fe composite was chosen as a mold insert material due to its excellent hardness and wear resistance [23]. With these advantages, a Ni-Fe mold insert could be utilized for the mass production of microstructures, which is one of the important aspects of bottom-up manufacturing processes. An electroforming process using an Ni-Fe solution in a sulphate bath was conducted on the opened area of bonded PMMA microstructures at 40 °C. In the early stage, a low current density of 1 A/dm² was applied to maintain a stable state. Thereafter, the current density was gradually increased to 3 A/dm². The electroforming process was continued until the Ni-Fe layer perfectly wrapped up the PMMA microstructure arrays.

Table 2. Composition and conditions of Ni-Fe alloy electroforming solution.

Description	Concentration
Nickel sulfate (M)	0.6
Iron sulfate (M)	0.01
Boric acid as a buffer solution (M)	0.5
Complex agent (M)	0.1
Brightener (g/L)	6
Sodium lauryl sulfate as a surfactant (g/L)	1.5
Direct current density (A/dm2)	3
Temperature (°C)	40

provides the composition and conditions of the Ni-Fe alloy electroforming solution. After forming a 1100 μm-thick Ni-Fe layer, a polishing process was performed to ensure a flat surface on the top and bottom sides. Finally, the PMMA microstructure arrays in the mold insert were dissolved by the cleaning process with acetone. The hardness of the Ni-Fe mold insert was measured by a Micro Vickers Hardness tester (FM-700, FUTURE-TECH, Kawasaki, Japan) under the standard test method (KS B 0811). In total, nine values were measured at random spots using the load of 19.61 N during 10 s and only five values except for two values from maximum and minimum were calculated to obtain an average value. The dimensions of the mold insert including pattern size, gap and thickness were characterized by means of optical microscopy.

3. Results and Discussion

3.1. Analysis of the X-Ray Exposure Condition

Figure 1a shows the schematic process of X-ray irradiation into PMMA through the X-ray mask. A SU-8 micropattern layer should exhibit good X-ray transmittance to ensure the radiation reaches the bottom surface of the PMMA. Conversely, a gold micropattern layer is responsible for absorption of the exposed X-ray radiation to prevent the development of an unexposed PMMA area [24]. Figure 1b compares the X-ray transmittance of each material used in this study. The 23 μm-thick SU-8 layer displayed high X-ray transmittance from the low to high photon energy range. Compared to the SU-8 material, the gold layer displayed low X-ray transmittance due to its excellent X-ray absorption, a quality enhanced as the gold thickness increased. Although the prepared X-ray mask can efficiently block X-ray radiation, a high-energy photon range exceeding 10,000 eV was transmitted through the gold layer. This could result in development on the top surface as well as a detrimental back-scattering effect on the bottom surface beneath the absorber. Thus, a white beam with an electron energy of 3.0 GeV and current of 360 mA was filtered using two Beryllium windows in a double mirror system and an 18 μm aluminum filter to cut off the high-energy photons, as shown in Figure 1c. When the X-ray white beam passed through the Be windows, the photon flux was dropped to about 66% of the original beam source but was still within the high-energy photon range. The effect of the double-mirror system was clearly recognized by eliminating the photon flux at the high photon energy range. The photon fluxes on the PMMA top and bottom were properly designed to remove the exposed area during development. The photon flux on the absorber top did not transmit through the 15 μm gold layer and, therefore, represented the zero value. Judging from the results, it is expected that the gold layer can perfectly block the X-ray during the exposure process. In general, dose states on both the absorber and PMMA are recommended to form a defect-free mother structure under four conditions [25]. (1) The dose on the absorber top should be lower than 100 J/cm³ to avoid development. (2) Ideally, the dose on the absorber bottom has a value under 10 J/cm³ to ensure a higher adhesion strength between the PMMA and substrate. (3) The dose on the PMMA top should not exceed 20 kJ/cm³. The dissoluble dose range during the development process is generally from 4 to 20 kJ/cm³. When the top surface of the PMMA is exposed to a dose greater than 20 kJ/cm³, it can be destroyed by thermal swelling and result in structural distortion. (4) The dose on the PMMA bottom should be higher than 4 kJ/cm³ to clearly etch out the exposed region.

Table 3. The dose conditions used in this study.

Dose	Requirement	Condition
Absorber top	<100 J/cm3	8.3 J/cm3
Absorber bottom	<10 J/cm3	4.4 J/cm3
PMMA top	<20 kJ/cm3	14.6 kJ/cm3
PMMA bottom	≥4 kJ/cm3	4.0 kJ/cm3

summarizes the dose conditions and requirements, all of which were satisfied in this study. With the prepared X-ray mask and optimized exposure conditions, the PMMA on the graphite sheet was X-ray irradiated. The development process was carried out to acquire a mother structure for electroforming as a next step.

3.2. Stability of Micropillar Arrays during Development Process

To acquire a mother structure for electroforming, the irradiated zone of the PMMA resist was dissolved using a GG-developer solution. The sample was initially placed in the upward direction, which is a conventional method of development, and the exposed region was etched out by the solution. Thereafter, the residual PMMA was rinsed with a rinsing solution and then DI water; however, the microstructure arrays became bent and clustered due to a capillary force interaction when the DI water was evaporated off their surface, as shown in Figure 2a. Generally, as the lateral dimension of a microstructure decreases and its aspect ratio increases, the surface area to volume ratio largely increases [26]. Thus, microstructure arrays are significantly susceptible to capillary forces during the evaporation process. To minimize the capillary force, DI water was mixed with a surface additive, the result of which is displayed in Figure 2b. The mixture had a smaller contact angle on the PMMA, leading to higher stability [9]. In general, the capillary force is proportional to the liquid and vapor surface energy. By adding the additive into the DI water, the value of surface energy decreased from 71.58 into 31.99 mN/m, resulting in lower capillary force during the evaporation. As a result, the microstructure arrays did not bend or cluster after the evaporation process, although a gel-like layer remained on the bottom surface of the substrate, as illustrated in Figure 2c. This critical layer exhibited a non-conductive property and could hinder the formation of a uniform Ni-Fe layer. To avoid this defect, the product was mounted in the downward direction to utilize a gravity-induced pulling force and better convective transport of the GG solution [22]. Taking advantage of downward development, the gel-like layer on the bottom side was clearly removed, as shown in Figure 2d. Based on the optimized development and evaporation process, a satisfactory microstructure quality was achieved with a pattern size of 183 μm, gap size of 68 μm and thickness of 1100 μm, as shown in Figure 3.

3.3. Electroforming Process and Hardness Property

The micro mold insert was fabricated by electroforming a Ni-Fe material onto the PMMA mother structure. Typically, the Ni-Fe alloy exhibits greater hardness than a pure Ni composition, leading to higher resistance against severe pressure and a greater number of uses during the molding process. In this study, the composition of Ni and Fe containing 0.6 M Ni and 0.01 M Fe was only used because a higher amount of Fe might induce the internal stress, resulting in bending of electroformed materials. For the electroforming process over 1100 μm, it is important to ensure a clean surface on the bottom without showing any residue. If the PMMA residue remains on the bottom, the area will exhibit a non-conductive character by interrupting a uniform Ni-Fe layer, as shown in Figure 4a. Although the residual PMMA was clearly dissolved, the Ni-Fe mold insert could be detached from the substrate before covering the total thickness, as shown in Figure 4b. This is caused by the weak adhesion strength between the PMMA and substrate. Even if the PMMA microstructure arrays withstand the capillary force during evaporation, greater adhesion strength may be required because higher internal stress occurs during the electroforming process. Therefore, to enhance the adhesion strength, the mirror angle and gold thickness of the X-ray mask were changed to 0.45° and 21 μm, respectively, by lowering the dose on the absorber bottom from 4.4 to 0.4 J/cm³. The lower dose on the absorber bottom means a weaker back-scattering effect on the adhesion layer [20].

Based on the variation of the beam condition, the Ni-Fe alloy completely wrapped the mother structure without any detachment, as shown in Figure 4c. After the polishing process on both sides, a Ni-Fe mold insert with a hardness of 585 Hv was acquired, as shown in Figure 5, displaying a pattern size of 183 μm, gap size of 68 μm, thickness of 550 μm and 2116 holes. Based on the optimization process, this Ni-Fe mold insert could be utilized in industrial fields for injection molding and embossing processes in which the fabrication of microstructure arrays with high aspect ratios is required during mass production.

4. Conclusions

In the present study, a microfabrication process of a Ni-Fe mold insert for micro molding technologies was proposed and developed using hard X-ray lithography and an electroforming process. The doses of both the absorber and PMMA were analyzed to prevent development underneath the absorber and to improve the adhesion strength between the PMMA and graphite substrate. Based on the evaluation of the beam source, the key conditions were designed, including a 15 μm-thick X-ray mask, mirror angle of 0.4° and exposure energy of 4 kJ/cm³. By utilizing the initial beam conditions, the mother structure for electroforming was successfully prepared without any bending, clustering or residual PMMA on the bottom area. However, during the electroforming process, the PMMA microstructure arrays were detached from the substrate before completely wrapping the entire thickness. To enhance the adhesion strength, the X-ray mask thickness and mirror angle were changed to 21 μm 0.45°, respectively, by lowering the dose on the absorber bottom to 0.4 J/cm³ while maintaining an exposure energy of 4 kJ/cm³. As a result, the Ni-Fe material successfully covered the PMMA microstructure arrays without any detachment from the substrate. Based on this optimized microfabrication process, a Ni-Fe mold insert with a pattern size of 183 μm, gap size of 68 μm, height of 550 μm, 2116 microcavities and a hardness of 585 Hv was delicately replicated. From the research, we confirmed the new insights to obtain a defect-free Ni-Fe mold during X-ray lithography and electroforming process as follows.

• During the evaporation process, a rinsing solution with surface additive provided the higher stability of micropillar arrays by lowering the interfacial energy between the PMMA and liquid.
• The downward direction method had the advantage of clear removal of PMMA on the bottom side owing to a gravity-induced pulling force and better convective transport of the etching solution.
• During the electroforming, a lower contact angle and higher gold thickness resulted in greater adhesion between the PMMA and substrate by lowering the dose on the absorber bottom.