A protocol for metal-assisted chemical imprinting of 3D microscale features with sub-20 nm shape accuracy into solid and porous silicon wafers is presented.
Metal-assisted electrochemical imprinting (Mac-Imprint) is a combination of metal-assisted chemical etching (MACE) and nanoimprint lithography that is capable of direct patterning 3D micro- and nanoscale features in monocrystalline group IV (e.g., Si) and III-V (e.g., GaAs) semiconductors without the need of sacrificial templates and lithographical steps. During this process, a reusable stamp coated with a noble metal catalyst is brought in contact with a Si wafer in the presence of a hydrofluoric acid (HF) and hydrogen peroxide (H2O2) mixture, which leads to the selective etching of Si at the metal-semiconductor contact interface. In this protocol, we discuss the stamp and substrate preparation methods applied in two Mac-Imprint configurations: (1) Porous Si Mac-Imprint with a solid catalyst; and (2) Solid Si Mac-Imprint with a porous catalyst. This process is high throughput and is capable of centimeter-scale parallel patterning with sub-20 nm resolution. It also provides low defect density and large area patterning in a single operation and bypasses the need for dry etching such as deep reactive ion etching (DRIE).
Three-dimensional micro- and nanoscale patterning and texturization of semiconductors enables numerous applications in various areas, such as optoelectronics1,2, photonics3, antireflective surfaces4, super hydrophobic, and self-cleaning surfaces5,6 among others. Prototyping and mass-producing 3D and hierarchical patterns has been successfully accomplished for polymeric films by soft lithography and nanoimprinting lithography with sub-20 nm resolution. However, transferring such 3D polymeric patterns into Si requires the etching selectivity of a mask pattern during reactive ion etching and thus limits the aspect ratio, and induces shape distortions and surface roughness due to scalloping effects7,8.
A new method called Mac-Imprint has been achieved for parallel and direct patterning of porous9 and solid Si wafers10,11 as well as solid GaAs wafers12,13,14. Mac-Imprint is a contact-based wet etching technique that requires contact between substrate and a noble metal-coated stamp possessing 3D features in the presence of an etching solution (ES) composed of HF and an oxidant (e.g., H2O2 in the case of Si Mac-Imprint). During the etching, two reactions occur simultaneously15,16: a cathodic reaction (i.e., the H2O2 reduction at the noble metal, during which positive charge carriers [holes] are generated and subsequently injected into Si17) and an anodic reaction (i.e., Si dissolution, during which the holes are consumed). After sufficient time in contact, the stamp's 3D features are etched into the Si wafer. Mac-Imprint has numerous advantages over conventional lithographical methods, such as high throughput, compatibility with roll-to-plate and roll-to-roll platforms, amorphous, mono- and polycrystalline Si and III-V semiconductors. Mac-Imprint stamps can be reused multiple times. Additionally, the method can deliver a sub-20 nm etching resolution that is compatible with contemporary direct writing methods.
The key to attaining high-fidelity imprinting is the diffusion pathway to the etching front (i.e., contact interface between catalyst and substrate). The work of Azeredo et al.9 first demonstrated that ES diffusion is enabled through a porous Si network. Torralba et al.18, reported that in order to realize solid Si Mac-Imprint the ES diffusion is enabled through a porous catalyst. Bastide et al.19 and Sharstniou et al.20 further investigated the catalyst porosity influence on ES diffusion. Thus, the concept of Mac-Imprint has been tested in three configurations with distinct diffusion pathways.
In the first configuration, the catalyst and substrate are solid, providing no initial diffusion pathway. The lack of reactant diffusion leads to a secondary reaction during imprinting that forms a layer of porous Si on the substrate around the edge of the catalyst-Si interface. The reactants are subsequently depleted, and the reaction stops, resulting in no discernable pattern transfer fidelity between the stamp and substrate. In the second and third configurations, the diffusion pathways are enabled through porous networks introduced either in the substrate (i.e., porous Si) or in the catalyst (i.e., porous gold) and high pattern transfer accuracy is attained. Thus, the mass transport through porous materials plays a critical role in enabling the diffusion of reactants and reaction products to and away from the contact interface9,18,19,20. A schematic of all three configurations is shown in Figure 1.
Figure 1: Schematics of Mac-Imprint configurations. This figure highlights the role of porous materials in enabling the diffusion of reacting species through the substrate (i.e., case II: porous Si) or in the stamp (i.e., case III: catalyst thin film made of porous gold). Please click here to view a larger version of this figure.
In this paper, the Mac-Imprint process is thoroughly discussed, including stamp preparation and substrate pretreatment along with Mac-Imprint itself. The substrate pretreatment section within the protocol includes Si wafer cleaning and Si wafer patterning with dry etching and substrate anodization (optional). Further, a stamp preparation section is subdivided into several procedures: 1) PDMS replica molding of Si master mold; 2) UV nanoimprinting of a photoresist layer in order to transfer the PDMS pattern; and 3) catalytic layer deposition via magnetron sputtering followed by dealloying (optional). Finally, in the Mac-Imprint section the Mac-Imprint setup along with the Mac-Imprint results (i.e., Si surface 3D hierarchical patterning) is presented.
CAUTION: Use appropriate safety practices and personal protective equipment (e.g., lab coat, gloves, safety glasses, closed-toe shoes). This procedure utilizes HF acid (48% wt) which is an extremely hazardous chemical and requires additional personal protective equipment (i.e., a face shield, natural rubber apron, and second pair of nitrile gloves that covers the hand, wrists, and forearms).
1. Stamp preparation for Mac-imprint
Figure 2: RCA-1 cleaning process. (a) Solution heating and (b) Si cleaning. Please click here to view a larger version of this figure.
Figure 3: PDMS mold fabrication process. (a) Schematic representation of the process. (b) Photographs of the process steps. Please click here to view a larger version of this figure.
Figure 4: Photoresist UV nanoimprinting process. (a) Photographs of photoresist spin coating. (b) Schematics and photographs of UV nanoimprinting. Please click here to view a larger version of this figure.
Figure 5: Catalytic stamp preparation process. (a) Schematics of the thin film deposition. (b) Photographs of the magnetron sputtering system. (c) Photograph of dealloying process with representative porous gold SEM images. Please click here to view a larger version of this figure.
2. Silicon substrate patterning and cleaning
Figure 6: Si wafer patterning mask layout (A) and single patterned chip (B). Please click here to view a larger version of this figure.
Figure 7: Photographs of substrate porosification procedure (Si anodization). (a) PC-controlled potentiostat connected to two-electrode electrochemical cell. (b) Electrochemical cell with platinum electrode. (c) Si chip with a porous Si layer. Please click here to view a larger version of this figure.
3. Mac-Imprinting setup
Figure 8: Photographs of Mac-Imprint setup (A), stamp before (B) and after (C) contact with Si chip. Please click here to view a larger version of this figure.
Scanning electron microscope (SEM) images, optical microscope scans (Figure 9), and atomic force microscopy (AFM) scans (Figure 10) were obtained in order to study the morphological properties of the Mac-Imprint stamps and imprinted Si surfaces. The cross-sectional profile of the imprinted solid Si was compared to that of the used porous Au stamp (Figure 10). Pattern transfer fidelity and porous Si generation during Mac-Imprint were two major criteria to analyze experimental success. The Mac-Imprint was considered successful if the Mac-Imprint stamp pattern was accurately transferred onto the Si and no porous Si is generated during the Mac-Imprint. The results of a suboptimal experiment (i.e., lack of pattern transfer fidelity along with porous Si generation during Mac-Imprint) are presented in Figure 9a (left).
Figure 9: Representative results: (a) Mac-Imprint of solid Si and porous Si with solid Au film (left and middle, respectively) and solid Si with porous Au film (right). (b) Top-down SEM images of porous Au films with different pore volume fraction (top) and corresponding imprinted Si morphology (bottom). (c) SEM images of various patterns produced by Mac-Imprint. This figure is reprinted with permission9,20. Please click here to view a larger version of this figure.
Figure 10: Representative results of solid Si Mac-Imprint with porous Au stamp: (a) AFM scans of porous Au stamp (left) and imprinted solid Si (right) and (b) overlaid cross-sectional profiles of porous Au stamp (blue) and imprinted solid Si (red). This figure is reprinted with permission20. Please click here to view a larger version of this figure.
Supplemental Figure 1: Photograph of spin coater control display. Please click here to view a larger version of this figure.
Supplemental Figure 2: Magnetron sputter control software screenshots. (a) Evacuation of magnetron sputter chamber. (b) Sputtering control parameters. (c) Ventilation of magnetron sputter chamber. Please click here to view a larger version of this figure.
Supplemental Figure 3: Potentiostat control software screenshot. Please click here to view a larger version of this figure.
Supplemental Figure 4: Linear motorized stage and load cell control software screenshots. (a) Before Mac-Imprint and (b) during Mac-Imprint. Please click here to view a larger version of this figure.
Supplemental Figure 5: Photograph of Mac-Imprint stamp to PTFE rod attachment process. Please click here to view a larger version of this figure.
Mac-Imprint stamps and prepatterned Si chips (p-type, [100] orientation, 1-10 Ohm∙cm) were prepared according to sections 1 and 2 of the protocol, respectively. The Mac-Imprint of prepatterned Si chip with stamps containing 3D hierarchical patterns was performed according to section 3 of the protocol (Figure 9). As shown in Figure 9a, different configurations of Mac-Imprint were applied: solid Si with solid Au (left), porous Si with solid Au (middle)9, and solid Si with porous Au (right)20. The diffusion of the reactants was blocked in the first case, leading to nonlocalized etching and partial porosification of the imprinted Si, which correlates with the same issue in the conventional MACE process22,23. However, when the diffusion was enabled through porous networks (either embedded in Si or Au), high pattern transfer fidelity was observed, which leads to the conclusion that Mac-Imprint is a mass transport dependent process. Also, the imprinted Si surface was roughened after imprinting with porous Au (Figure 9a, right).
It was proposed that surface roughening originates from the porosity of the porous Au used. In order to test the hypothesis, a series of porous Au layers with various controlled pore volume fractions (PVF) was created according to sections 1.4 and 1.5 of the protocol and subsequently implemented for Mac-Imprint (Figure 9b)20. A direct relation between the stamp's PVF and imprinted Si surface roughness was observed, supporting the hypothesis. Additionally, after Mac-Imprint with low PVF stamps, Si was porosified, which was explained by hindered ES diffusion through undeveloped porous Au structure, resulting in delocalization of the etching front20. Thus, a developed and interconnected porous structure is critical for high pattern transfer fidelity during Mac-Imprint. Moreover, imprinted Si porosification was observed at medium PVF when a porous Au layer already had an interconnected porous network. This can be attributed to the high ratio between Au and Si surface areas and subsequent injection of the excessive holes into Si, which also leads to the etching front delocalization and, as a result, porous Si formation20. This process can be controlled through careful adjustment of the HF and H2O2 ratios in the ES.
Implementation of the porous Au stamps along with ES composition variations allows the manufacture of various 3D hierarchical patterns via Mac-Imprint that were previously published in the works of Azeredo et al.9 and Sharstniou et al.20 (Figure 9c).
Further investigations of porous Au/Si interface chemistry, in particular PVF-dependent etch rate and localization, along with imprinting system improvement, will help to make the Mac-Imprint process suitable for industrial scale applications in the future.
The authors have nothing to disclose.
We acknowledge Dr. Keng Hsu (University of Louisville) for insights regarding this work; University of Illinois's Frederick Seitz Laboratory and, in memoriam, staff member Scott Maclaren; Arizona State University's LeRoy Eyring Center for Solid State Science; and the Science Foundation Arizona under the Bis grove Scholars Award.
Acetone, >99.5%, ACS reagent | Sigma-Aldrich | 67-64-1 | CAUTION, chemical |
Ammonium fluoride, >98%, ACS grade | Sigma-Aldrich | 12125-01-8 | CAUTION, hazardous |
Ammonium hydroxide solution, 28-30%, ACS reagent | Sigma-Aldrich | 1336-21-6 | CAUTION, hazardous |
AZ 400K developer | Microchemicals | AZ 400K | CAUTION, chemical |
BenchMark 800 Etch | Axic | BenchMark 800 | Reactive ion etching |
Chromium target, 2" x 0.125", 99.95% purity | ACI alloys | ADM0913 | Magnetron sputter chromium target |
CTF 12 | Carbolite Gero | C12075-700-208SN | Tube furnace |
Desiccator | Fisher scientific Chemglass life sciences | CG122611 | Desiccator |
F6T5/BLB | Eiko | F6T5/BLB 6W | UV bulb |
Gold target, 2" x 0.125", 99.99% purity | ACI alloys | N/A | Magnetron sputter gold target |
Hotplate KW-4AH | Chemat tecnologie | KW-4AH | Leveled hotplate with uniform temperature profile |
Hydrofluoric acid, 48%, ACS reagent | Sigma-Aldrich | 7664-39-3 | CAUTION, extremly hazardous |
Hydrogen peroxide, 30%, ACS reagent | Fisher Chemical | 7722-84-1 | CAUTION, hazardous |
Isopropyl alcohol, >99.5%, ACS reagent | LabChem | 67-63-0 | CAUTION, chemical |
MLP-50 | Transducer Techniques | MLP-50 | Load cell |
Nitric acid, 70%, ACS grade | SAFC | 7697-37-2 | CAUTION, hazardous |
NSC-3000 | Nano-master | NSC-3000 | Magnetron sputter |
Potassium hydroxide, 45%, Certified | Fisher Chemical | 1310-58-3 | CAUTION, chemical |
Rocker 800 vacuum pump, 110V/60Hz | Rocker | 1240043 | Oil-free vacuum pump |
Silicon master mold | NILT | SMLA_V1 | Silicon chip with pattern |
Silicon wafers, prime grade | University wafer | 783 | Si wafer |
Silver target, 2" x 0.125", 99.99% purity | ACI alloys | HER2318 | Magnetron sputter silver target |
SP-300 | BioLogic | SP-300 | Potentiostat |
SPIN 150i | Spincoating | SPIN 150i | Spin coater |
SPR 200-7.0 positive photoresist | Microchem | SPR 220-7.0 | CAUTION, chemical |
Stirring hotplate | Thermo scientific Cimarec+ | SP88857100 | General purpose hotplate |
SU-8 2015 negative photoresist | Microchem | SU-8 2015 | CAUTION, chemical |
SYLGARD 184 Silicone elastomere kit | DOW | 4019862 | CAUTION, chemical |
T-LSR150B | Zaber Technologies | T-LSR150B-KT04U | Motorized linear stage |
Trichloro(1H,1H,2H,2H-perfluorooctyl)silane (PFOCS), 97% | Sigma-Aldrich | 78560-45-9 | CAUTION, hazardous |