Supporting Information Shape-controlled metal-metal and metal ...
Supporting Information
Shape-controlled metal-metal and metal-polymer Janus structures by thermoplastic embossing Molla Hasan, Niloofar Kahler, and Golden Kumar* Department of Mechanical Engineering, Texas Tech University, Lubbock 79409, TX
*Corresponding author: Golden Kumar ([email protected])
S-1
Experimental details Thermoplastic embossing experiments were performed in air using a custom-built parallel plate setup illustrated in Figure S1. Two steel plates (4×4×1 inches3) equipped with cartridge heaters and thermocouples were installed on Instron 5966 with a load cell of 10 kN. The plates were lapped and mirror polished to ensure parallelism. For core-shell Janus structures, a silicon template (closed or through-etched) was placed on a lower heating plate. Two thermoplastic materials with an overlapping supercooled liquid temperature range were stacked on top of the template (Figure S1-a).
Figure S1: Schematic of co-embossing and additive embossing fabrication of Janus structures from thermoplastic materials. Core-shell Janus structure can only be produced using coembossing (a) whereas rod shaped Janus structures can be fabricated using both co-embossing (b) and additive embossing (c). S-2
The processing temperature and load were selected based on previous knowledge of thermoplastic forming of individual materials. The processing temperature was selected such that the viscosity of both thermoplastic materials dropped below 1010 Pa.s. At these viscosities, a pressure in the range of 10-50 MPa was sufficient for template filling and joining of metallic glasses (MGs). Silicon templates were fabricated by photolithography and deep-reactive-ionetching (DRIE). The silicon templates were dissolved in 40% KOH solution to release the free Janus structures. Rod shaped Janus structures were fabricated by both co-embossing (Figure 1S-b) and additive embossing (Figure 1S-c). In co-embossing, a through-etched silicon template was sandwiched between two thermoplastic materials. In additive embossing, the materials with a higher Tg was embossed against a through-etched template. The filling length can be controlled through processing temperature and pressure values. Subsequently, the second thermoplastic materials with a lower Tg was embossed from the other end of the template. This provides an independent control over the fractions of two thermoplastic materials in Janus rods. Additive embossing was also used to produce Janus structures from thermoplastic materials that have non-overlapping supercooled liquid temperature ranges. To achieve this, the material with higher Tg was patterned with nanorods using porous alumina templates. After etching the alumina in KOH, the patterned sample was embossed on through-etched silicon template. The silicon was cooled to the processing temperature of low-Tg material which was then embossed from the opposite side of the partially filled template. During final embossing, the high-Tg nanorods served as anchoring sites for the low-Tg material. Both co-embossing and additive embossing produced durable joints among MGs and thermoplastic polymers. Silicon based templates were handled with extreme care due to their fragile nature. Any other template material can be used as long as it can withstand the temperature and processing conditions used in thermoplastic embossing of metallic glasses. List of suitable template materials and their desirable properties for thermoplastic forming of MGs are published elsewhere1-4. Durability of joints Durability of joints formed by co-embossing and additive embossing of thermoplastic materials was tested by room temperature tensile tests. As described in the manuscript, a strong joining was obtained in both co-embossing and additive embossing of dissimilar MGs. Similar strategy was applied for joining of MGs and polymers. The bonding between MGs and polymers was less strong compared to MG-MG joints but stable Janus structures could still be formed. Figure S2 shows the SEM image of mechanical joint between nano-patterned Pt-based MG and PMMA. After separation from the MG, PMMA surface reveals the formation of deep holes and stretched ligaments. MG nanostructures did not fracture during separation as observed in pulling of MGMG joints. Despite weaker adhesion, the PMMA-Pt-based MG Janus pillars remained intact after all the processing and handling steps. Therefore, Janus structures from MGs and polymers can be fabricated by mechanical interlocking of surface nanostructures.
S-3
Figure S2: Fracture of mechanical joint formed between nano-patterned Pt-based MG and PMMA. After separation, the PMMA surface shows the negative replica of nano-patterned Ptbased MG indicating the formation of intimate contact between two materials. Fabrication of core-shell Janus structures Core-shell structures were fabricated using the concept of blow-molding of MGs5-6. A thin sheet of MG can thermoplastically inflate into a hollow shell when subjected to an air pressure difference (Figure S3-a). The final thickness and height of shell can be calculated from the applied pressure, the processing time, the viscosity, and the strain-rate4. The stress developed in hemispherical MG shell during deformation is given by: Pd Stress = , (S1) 4δ o where P is the pressure difference across the MG1, d is the diameter of shell, and δo is the thickness of MG1 sheet. Thermoplastic blowing continues as long as the applied stress is higher than the Newtonian flow-stress of MG sheet: •
Flow − stress = 3η1 ε ,
(S2)
Figure S3: Schematic of thermoplastic blow-molding of MGs (a) and fabrication of core-shell Janus structures using the blow-molding concept (b). MG1 and MG2 correspond to two different MGs. S-4
•
where η1 is the viscosity of MG1 supercooled liquid and ε is the strain-rate. The strain ε in a hemispherical shell is given by7:
ε = ln(δ f / δ o ) ,
(S3)
where δf is the final thickness of hemispherical shell. The pressure and processing time (t) required to achieve the final thickness δf can be calculated by combining Equations (S1) to (S3):
Pd 3η1ε 3η1 > = ln(δ f / δ o ) , 4δ o t t
12η1δ o ln(δ f / δ o ) , (S4) d The shell height (h) can be calculated from the thickness values using conservation of volume. Equation (S4) can be used to control the thermoplastic processing conditions for desirable dimensions of hollow MG shells. To fabricate core-shell Janus structures, we replace air with a thermoplastic material, MG2 (Figure S3-b). The additional requirement is that the applied pressure should be sufficient to fill the core with MG2. The required pressure for filling cylindrical feature of height h with MG2 can be estimated using Hagen-Poiseuille law: ⇒ Pt >
32η 2 P= t
2
h , d − 2δ f
(S5)
where η2 is the viscosity of MG2 supercooled liquid at the co-embossing temperature. Therefore, by using Equations (S4) and (S5) one can precisely control the dimensions of core-shell Janus structures. The initial thickness of MG1 has profound effect on the outcome of co-embossing process but MG2 can be of any thickness as long as sufficient amount of material is supplied to fill the core. Figure S4 shows examples of large number of core-shell Janus structures with varying sizes and shapes fabricated by co-embossing of MGs.
Figure S4: Examples of core-shell Janus structure in large quantity (a), circular cross-section (b), and square cross-section (c). In images a & b, the cores and shells are made of Ni-based MG and Pd-based MG, respectively. In image c, the core is Pd-based MG whilst the shell is Ni-based MG. S-5
Fabrication of rod shaped Janus structures by co-embossing The rod shaped Janus structures were produced by simultaneous filling of through-etched templates with two different thermoplastic materials (Figure S5). The fraction of two materials can be estimated using Hagen-Poiseuille law. The pressure required to fill a cylindrical cavity with a Newtonian fluid is given by:
P=
32ηh 2 , td 2
(S6)
where P is the applied pressure, t is the embossing time, η is the viscosity of thermoplastic material, h is the height and d is the diameter of cylindrical cavity. The embossing time to simultaneously fill the cavity with two different liquids A and B of viscosities ηA and ηB is (Figure S5):
32h 2 t= Pd 2
−2
1 1 + . η η B A
(S7)
The fraction of filling lengths of two liquids is related to their viscosity values as:
ηB hA = and h = h A + hB . hB ηA
(S8)
Figure S5: Schematic of co-filling process using two different thermoplastic materials. Two thermoplastic materials, A and B, are simultaneously embossed from the opposite ends of a through-etched template.
According to these equations under any embossing conditions the liquid with lower a viscosity will constitute the large fraction of the Janus rod. This limitation can be alleviated by continuing the embossing operation beyond t to displace the low viscosity fluid with the high viscosity liquid. It allows controlling the lengths of materials in Janus structures despite the fixed ratio of their viscosity values. In addition, processing temperature can also be used to alter the viscosity ratio of co-flowing supercooled liquids owing to their different fragility values. Effect of trapped air One of the factors that may affect the joining of thermoplastic materials in through-etched templates is the presence of trapped air. Here, we consider the effect of trapped air theoretically and experimentally. Initially the template cavities are filled with an air at atmospheric pressure S-6
(Pi ~ 0.1 MPa) which remains trapped during the co-embossing of thermoplastic materials A and B (Figure S6). With the increase in embossing pressure, the volume (Vf) of trapped air decreases. For example, the volume of trapped air will decrease to ~1/100 of its starting value (Vi) when the embossing pressure reaches 10 MPa (typical embossing pressure). Moreover, the trapped air is displaced towards the sidewalls of the template because of the parabolic flow profiles of thermoplastic materials. Therefore, the presence of trapped air does not severely affect the interface of thermoplastic materials during co-embossing using a through-etched template. Figure S6: Effect of trapped air on the interface during coembossing of two thermoplastic materials from two opposite ends of a through-etched template. The trapped air is squeezed towards the sidewalls because of parabolic flow profiles of thermoplastic materials. The volume of trapped air decreases to 1% of its initial volume at an applied pressure of 10 MPa. To verify this hypothesis, Pt-based metallic glass was thermoplastically embossed against a template cavity closed at one end. Figure S7 illustrates the process and the SEM image of metallic glass after removing from the template. Despite the presence of trapped air, the tip of metallic glass faithfully replicates the bottom of the template except the edges. This suggests that the trapped air is pushed towards the corners of the template cavity during embossing. Because of parabolic profile of flowing metallic glass supercooled liquid, it touches the template bottom at the center and then expands outwards. At the final stage, majority of the trapped air is squeezed in the corners. These theoretical and experimental demonstrations indicate that the effect of trapped air is very small. Figure S7: An example of Pt-based metallic glass thermoplastically embossed against a closed template cavity. The SEM image shows that the metallic glass replicated the bottom of the template except edges, suggesting that the trapped air is squeezed towards the edges.
S-7
Co-filling of templates from the same end Filling of templates can also be achieved by co-embossing of two thermoplastic materials flowing side-by-side (Figure S8). This can produce Janus structures with vertical interface as shown by the examples. When two supercooled liquids of different viscosities co-flow in a template cavity, the more viscous fluid has a lower velocity and fills to lower depth. However, the velocity profile at the interface remains continuous due to balancing of the shear stress. Figure S8: Variant of coembossing when two thermoplastic materials flow side-by-side into a template cavity. The resulting Janus structures have vertical interface as shown by example of pillars made from Ni-based and Pd-based MGs. Multiphasic structures Co-embossing and additive embossing can be combined to generate multiphasic structures as shown in Figure S9. The materials with overlapping supercooled temperature range are coembossed from one side of the through-etched template resulting in formation of core-shell morphology. Subsequently, the material with a lower processing temperature is filled from the opposite side of the template. Examples show pillars consisting of three MGs organised in different arrangements (Figure S9).
Figure S9: Fabrication of structures comprised of three different thermoplastic materials. Initially, materials A and B are co-embossed to form core-shell structures. Subsequently, third material C is added to make three-phase structures. Examples show images of pillars made from Pt-based, Pd-based, and Ni-based MGs.
S-8
References (1) (2)
(3) (4)
(5) (6) (7)
Kumar, G.; Desai, A.; Schroers, J., Bulk Metallic Glass: The Smaller The Better. Adv. Mater. 2011, 23, 461-476. Schroers, J.; Pham, Q.; Desai, A., Thermoplastic Forming of Bulk Metallic Glass - A Technology for MEMS and Microstructure Fabrication. J. Microelectromech. Syst. 2007, 16, 240-247. Schroers, J., Processing of Bulk Metallic Glass. Adv. Mater. 2010, 22, 1566-1597. Sarac, B.; Kumar, G.; Hodges, T.; Ding, S. Y.; Desai, A.; Schroers, J., Three-Dimensional Shell Fabrication Using Blow Molding of Bulk Metallic Glass. J Microelectromech. Syst. 2011, 20, 28-36. Schroers, J.; Pham, Q.; Peker, A.; Paton, N.; Curtis, R. V., Blow Molding of Bulk Metallic Glass. Scr. Mater. 2007, 57, 341-344. Schroers, J.; Hodges, T. M.; Kumar, G.; Raman, H.; Barnes, A. J.; Quoc, P.; Waniuk, T. A., Thermoplastic Blow Molding of Metals. Mater. Today 2011, 14, 14-19. Dealy, J. M., Official Nomenclature for Material Functions Describing the Response of a Viscoelastic Fluid to Various Shearing and Extensional Deformations. J. Rheol. 1995, 39, 253-265.
S-9
Shape-controlled metal-metal and metal-polymer Janus structures by thermoplastic embossing Molla Hasan, Niloofar Kahler, and Golden Kumar* Department of Mechanical Engineering, Texas Tech University, Lubbock 79409, TX
*Corresponding author: Golden Kumar ([email protected])
S-1
Experimental details Thermoplastic embossing experiments were performed in air using a custom-built parallel plate setup illustrated in Figure S1. Two steel plates (4×4×1 inches3) equipped with cartridge heaters and thermocouples were installed on Instron 5966 with a load cell of 10 kN. The plates were lapped and mirror polished to ensure parallelism. For core-shell Janus structures, a silicon template (closed or through-etched) was placed on a lower heating plate. Two thermoplastic materials with an overlapping supercooled liquid temperature range were stacked on top of the template (Figure S1-a).
Figure S1: Schematic of co-embossing and additive embossing fabrication of Janus structures from thermoplastic materials. Core-shell Janus structure can only be produced using coembossing (a) whereas rod shaped Janus structures can be fabricated using both co-embossing (b) and additive embossing (c). S-2
The processing temperature and load were selected based on previous knowledge of thermoplastic forming of individual materials. The processing temperature was selected such that the viscosity of both thermoplastic materials dropped below 1010 Pa.s. At these viscosities, a pressure in the range of 10-50 MPa was sufficient for template filling and joining of metallic glasses (MGs). Silicon templates were fabricated by photolithography and deep-reactive-ionetching (DRIE). The silicon templates were dissolved in 40% KOH solution to release the free Janus structures. Rod shaped Janus structures were fabricated by both co-embossing (Figure 1S-b) and additive embossing (Figure 1S-c). In co-embossing, a through-etched silicon template was sandwiched between two thermoplastic materials. In additive embossing, the materials with a higher Tg was embossed against a through-etched template. The filling length can be controlled through processing temperature and pressure values. Subsequently, the second thermoplastic materials with a lower Tg was embossed from the other end of the template. This provides an independent control over the fractions of two thermoplastic materials in Janus rods. Additive embossing was also used to produce Janus structures from thermoplastic materials that have non-overlapping supercooled liquid temperature ranges. To achieve this, the material with higher Tg was patterned with nanorods using porous alumina templates. After etching the alumina in KOH, the patterned sample was embossed on through-etched silicon template. The silicon was cooled to the processing temperature of low-Tg material which was then embossed from the opposite side of the partially filled template. During final embossing, the high-Tg nanorods served as anchoring sites for the low-Tg material. Both co-embossing and additive embossing produced durable joints among MGs and thermoplastic polymers. Silicon based templates were handled with extreme care due to their fragile nature. Any other template material can be used as long as it can withstand the temperature and processing conditions used in thermoplastic embossing of metallic glasses. List of suitable template materials and their desirable properties for thermoplastic forming of MGs are published elsewhere1-4. Durability of joints Durability of joints formed by co-embossing and additive embossing of thermoplastic materials was tested by room temperature tensile tests. As described in the manuscript, a strong joining was obtained in both co-embossing and additive embossing of dissimilar MGs. Similar strategy was applied for joining of MGs and polymers. The bonding between MGs and polymers was less strong compared to MG-MG joints but stable Janus structures could still be formed. Figure S2 shows the SEM image of mechanical joint between nano-patterned Pt-based MG and PMMA. After separation from the MG, PMMA surface reveals the formation of deep holes and stretched ligaments. MG nanostructures did not fracture during separation as observed in pulling of MGMG joints. Despite weaker adhesion, the PMMA-Pt-based MG Janus pillars remained intact after all the processing and handling steps. Therefore, Janus structures from MGs and polymers can be fabricated by mechanical interlocking of surface nanostructures.
S-3
Figure S2: Fracture of mechanical joint formed between nano-patterned Pt-based MG and PMMA. After separation, the PMMA surface shows the negative replica of nano-patterned Ptbased MG indicating the formation of intimate contact between two materials. Fabrication of core-shell Janus structures Core-shell structures were fabricated using the concept of blow-molding of MGs5-6. A thin sheet of MG can thermoplastically inflate into a hollow shell when subjected to an air pressure difference (Figure S3-a). The final thickness and height of shell can be calculated from the applied pressure, the processing time, the viscosity, and the strain-rate4. The stress developed in hemispherical MG shell during deformation is given by: Pd Stress = , (S1) 4δ o where P is the pressure difference across the MG1, d is the diameter of shell, and δo is the thickness of MG1 sheet. Thermoplastic blowing continues as long as the applied stress is higher than the Newtonian flow-stress of MG sheet: •
Flow − stress = 3η1 ε ,
(S2)
Figure S3: Schematic of thermoplastic blow-molding of MGs (a) and fabrication of core-shell Janus structures using the blow-molding concept (b). MG1 and MG2 correspond to two different MGs. S-4
•
where η1 is the viscosity of MG1 supercooled liquid and ε is the strain-rate. The strain ε in a hemispherical shell is given by7:
ε = ln(δ f / δ o ) ,
(S3)
where δf is the final thickness of hemispherical shell. The pressure and processing time (t) required to achieve the final thickness δf can be calculated by combining Equations (S1) to (S3):
Pd 3η1ε 3η1 > = ln(δ f / δ o ) , 4δ o t t
12η1δ o ln(δ f / δ o ) , (S4) d The shell height (h) can be calculated from the thickness values using conservation of volume. Equation (S4) can be used to control the thermoplastic processing conditions for desirable dimensions of hollow MG shells. To fabricate core-shell Janus structures, we replace air with a thermoplastic material, MG2 (Figure S3-b). The additional requirement is that the applied pressure should be sufficient to fill the core with MG2. The required pressure for filling cylindrical feature of height h with MG2 can be estimated using Hagen-Poiseuille law: ⇒ Pt >
32η 2 P= t
2
h , d − 2δ f
(S5)
where η2 is the viscosity of MG2 supercooled liquid at the co-embossing temperature. Therefore, by using Equations (S4) and (S5) one can precisely control the dimensions of core-shell Janus structures. The initial thickness of MG1 has profound effect on the outcome of co-embossing process but MG2 can be of any thickness as long as sufficient amount of material is supplied to fill the core. Figure S4 shows examples of large number of core-shell Janus structures with varying sizes and shapes fabricated by co-embossing of MGs.
Figure S4: Examples of core-shell Janus structure in large quantity (a), circular cross-section (b), and square cross-section (c). In images a & b, the cores and shells are made of Ni-based MG and Pd-based MG, respectively. In image c, the core is Pd-based MG whilst the shell is Ni-based MG. S-5
Fabrication of rod shaped Janus structures by co-embossing The rod shaped Janus structures were produced by simultaneous filling of through-etched templates with two different thermoplastic materials (Figure S5). The fraction of two materials can be estimated using Hagen-Poiseuille law. The pressure required to fill a cylindrical cavity with a Newtonian fluid is given by:
P=
32ηh 2 , td 2
(S6)
where P is the applied pressure, t is the embossing time, η is the viscosity of thermoplastic material, h is the height and d is the diameter of cylindrical cavity. The embossing time to simultaneously fill the cavity with two different liquids A and B of viscosities ηA and ηB is (Figure S5):
32h 2 t= Pd 2
−2
1 1 + . η η B A
(S7)
The fraction of filling lengths of two liquids is related to their viscosity values as:
ηB hA = and h = h A + hB . hB ηA
(S8)
Figure S5: Schematic of co-filling process using two different thermoplastic materials. Two thermoplastic materials, A and B, are simultaneously embossed from the opposite ends of a through-etched template.
According to these equations under any embossing conditions the liquid with lower a viscosity will constitute the large fraction of the Janus rod. This limitation can be alleviated by continuing the embossing operation beyond t to displace the low viscosity fluid with the high viscosity liquid. It allows controlling the lengths of materials in Janus structures despite the fixed ratio of their viscosity values. In addition, processing temperature can also be used to alter the viscosity ratio of co-flowing supercooled liquids owing to their different fragility values. Effect of trapped air One of the factors that may affect the joining of thermoplastic materials in through-etched templates is the presence of trapped air. Here, we consider the effect of trapped air theoretically and experimentally. Initially the template cavities are filled with an air at atmospheric pressure S-6
(Pi ~ 0.1 MPa) which remains trapped during the co-embossing of thermoplastic materials A and B (Figure S6). With the increase in embossing pressure, the volume (Vf) of trapped air decreases. For example, the volume of trapped air will decrease to ~1/100 of its starting value (Vi) when the embossing pressure reaches 10 MPa (typical embossing pressure). Moreover, the trapped air is displaced towards the sidewalls of the template because of the parabolic flow profiles of thermoplastic materials. Therefore, the presence of trapped air does not severely affect the interface of thermoplastic materials during co-embossing using a through-etched template. Figure S6: Effect of trapped air on the interface during coembossing of two thermoplastic materials from two opposite ends of a through-etched template. The trapped air is squeezed towards the sidewalls because of parabolic flow profiles of thermoplastic materials. The volume of trapped air decreases to 1% of its initial volume at an applied pressure of 10 MPa. To verify this hypothesis, Pt-based metallic glass was thermoplastically embossed against a template cavity closed at one end. Figure S7 illustrates the process and the SEM image of metallic glass after removing from the template. Despite the presence of trapped air, the tip of metallic glass faithfully replicates the bottom of the template except the edges. This suggests that the trapped air is pushed towards the corners of the template cavity during embossing. Because of parabolic profile of flowing metallic glass supercooled liquid, it touches the template bottom at the center and then expands outwards. At the final stage, majority of the trapped air is squeezed in the corners. These theoretical and experimental demonstrations indicate that the effect of trapped air is very small. Figure S7: An example of Pt-based metallic glass thermoplastically embossed against a closed template cavity. The SEM image shows that the metallic glass replicated the bottom of the template except edges, suggesting that the trapped air is squeezed towards the edges.
S-7
Co-filling of templates from the same end Filling of templates can also be achieved by co-embossing of two thermoplastic materials flowing side-by-side (Figure S8). This can produce Janus structures with vertical interface as shown by the examples. When two supercooled liquids of different viscosities co-flow in a template cavity, the more viscous fluid has a lower velocity and fills to lower depth. However, the velocity profile at the interface remains continuous due to balancing of the shear stress. Figure S8: Variant of coembossing when two thermoplastic materials flow side-by-side into a template cavity. The resulting Janus structures have vertical interface as shown by example of pillars made from Ni-based and Pd-based MGs. Multiphasic structures Co-embossing and additive embossing can be combined to generate multiphasic structures as shown in Figure S9. The materials with overlapping supercooled temperature range are coembossed from one side of the through-etched template resulting in formation of core-shell morphology. Subsequently, the material with a lower processing temperature is filled from the opposite side of the template. Examples show pillars consisting of three MGs organised in different arrangements (Figure S9).
Figure S9: Fabrication of structures comprised of three different thermoplastic materials. Initially, materials A and B are co-embossed to form core-shell structures. Subsequently, third material C is added to make three-phase structures. Examples show images of pillars made from Pt-based, Pd-based, and Ni-based MGs.
S-8
References (1) (2)
(3) (4)
(5) (6) (7)
Kumar, G.; Desai, A.; Schroers, J., Bulk Metallic Glass: The Smaller The Better. Adv. Mater. 2011, 23, 461-476. Schroers, J.; Pham, Q.; Desai, A., Thermoplastic Forming of Bulk Metallic Glass - A Technology for MEMS and Microstructure Fabrication. J. Microelectromech. Syst. 2007, 16, 240-247. Schroers, J., Processing of Bulk Metallic Glass. Adv. Mater. 2010, 22, 1566-1597. Sarac, B.; Kumar, G.; Hodges, T.; Ding, S. Y.; Desai, A.; Schroers, J., Three-Dimensional Shell Fabrication Using Blow Molding of Bulk Metallic Glass. J Microelectromech. Syst. 2011, 20, 28-36. Schroers, J.; Pham, Q.; Peker, A.; Paton, N.; Curtis, R. V., Blow Molding of Bulk Metallic Glass. Scr. Mater. 2007, 57, 341-344. Schroers, J.; Hodges, T. M.; Kumar, G.; Raman, H.; Barnes, A. J.; Quoc, P.; Waniuk, T. A., Thermoplastic Blow Molding of Metals. Mater. Today 2011, 14, 14-19. Dealy, J. M., Official Nomenclature for Material Functions Describing the Response of a Viscoelastic Fluid to Various Shearing and Extensional Deformations. J. Rheol. 1995, 39, 253-265.
S-9
