Supporting Information Anisotropic Nanoparticles as Shape-Directing ...
Supporting Information Anisotropic Nanoparticles as Shape-Directing Catalysts for the Chemical Etching of Silicon Guoliang Liu†, Kaylie L. Young†, Xing Liao‡, Michelle L. Personick†, Chad A. Mirkin†‡* †Department of Chemistry and International Institute of Nanotechnology, ‡Department of Materials Science and Engineering, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208, United States
Materials Hydrofluoric acid (HF, 48 wt.% in H2O, 99.99%), tetrachloroauric(III) acid trihydrate (HAuCl4·3H2O, 99.9+%), cetyltrimethylammonium bromide (CTAB, 99%), cetyltrimethylammonium chloride (CTAC, 25 wt.% in H2O), sodium borohydride (NaBH4, 99.99%), L-ascorbic acid (AA, 99+%), trisodium citrate, NaOH, hexamethyldisilazane (HMDS) and hexane were purchased from Sigma-Aldrich, Inc., and used as received. Poly(ethylene oxide)-block-poly(2-vinyl pyridine) (PEO-b-P2VP, Mn=2.8-b-1.5 kg·mol-1, polydispersity index, PDI=1.11) was purchased from Polymer Source, Inc. and used as received. Hydrogen peroxide (H2O2, 30% Solution, AR Select® (ACS)) was purchased from Macron Fine Chemicals. DPN® pen arrays (Type M, no gold-coating) were purchased from Nanoink, Inc. Silicon wafers (100) were purchased from Nova Electronic Materials. Synthesis of spherical nanoparticles PEO-b-P2VP and HAuCl4·3H2O were dissolved separately in water, and then mixed together. The final ink had a PEO-bP2VP concentration of 5 mg·ml-1. The molar ratio between 2-vinylpyridine (2VP) and HAuCl4 in the final mixture ranged from 4:1 to 512:1. After stirring overnight, the solution was dip-coated onto a DPN® pen array. After drying in an N2 stream, the pen array was mounted onto an AFM (NScriptor, NanoInk, Inc.) in a chamber with controlled humidity. The relative humidity was in the range of 75-95% to control the dimensions of patterned polymer features and thus the nanoparticle size. Prior to DPN patterning, to ensure that the silicon substrate was hydrophobic, hexamethyldisilazane (HMDS) was evaporated on the surface by keeping the substrate in a desiccator with two vials of a HMDS and hexane mixture (v:v = 1) for 24 h. The patterned substrate was placed in a tube furnace for thermal annealing. The annealing conditions were programmed as follows: ramp to 150 °C in 1 h, soak at 150 °C for 4 h in Ar, cool down to room temperature in 4 h, ramp to 500 °C in 1 h, soak at 500 oC for 4 h in Ar, cool down to room temperature in 1 h. Afterwards, the substrate was again annealed at 500 °C for 2 h in air to remove any organic residue. By varying the molar ratio between 2VP and HAuCl4 or dwell time during the patterning, the average diameter of the gold nanoparticles was controlled over the 4.0-25.5 nm range. Synthesis of gold nanorods with low aspect ratio Gold nanorods with an aspect ratio of ~5.3 were synthesized as follows. Briefly, gold nanoparticle seeds (~3.5 nm diameter) were first prepared by adding 250 µL of 100 mM HAuCl4 and 600 µL of ice cold 10 mM NaBH4 to a 10 mL aqueous solution of 100 mM CTAB. The solution was stirred for ~ 2 min and then allowed to age for 2 h. The solution appeared brown. To grow gold nanorods, 10 ml of a 100 mM CTAB solution was mixed with 500 µL of 10 mM HAuCl4, 200 µL of 1 M HCl, 100 µL of 10 mM AgNO3, and 100 µL of 100 mM AA. After shaking the solution, 100 µL of the seed solution (100 times diluted) was added. The solution was then stored in a 30 °C water bath overnight. Synthesis of gold nanorods with high aspect ratio Gold nanorods with an aspect ratio of ~25 were synthesized according to methods in the literature (J. Phys. Chem. B 2001, 105, 4065). Briefly, gold nanoparticle seeds (~3.5 nm diameter) were first prepared by adding 600 µL of ice cold 0.1 M NaBH4 to a 20 mL aqueous solution containing 2.5 x 10-4 M HAuCl4 and 2.5 x 10-4 M trisodium citrate. The solution was allowed to age for 2-5 h. A three-step seeding method was used to prepare the nanorods. Three vials (labeled A, B, C), each containing 9 mL of “growth solution” consisting of 0.1 M CTAB and 2.5 x 10-4 M HAuCl4, were mixed with 0.05 mL of 0.1M ascorbic acid. Next 1 mL of the seed solution was added to vial A, swirled, and allowed to sit for 15 s. Then, 1 mL of solution A was transferred to vial B, swirled, and allowed to sit for 30 s. Finally, 1 mL of solution B was transferred to vial C and swirled. Solution C contains the gold rods with an aspect ratio of ~25 (Figure S8).
1
Synthesis of gold nanoplates Gold nanoprisms were synthesized according to the methods in the literature (J. Am. Chem. Soc. 2005, 127, 5312; Nano Lett. 2008, 8, 2526.) with minor differences. Briefly, gold nanoparticle seeds were first prepared by adding 250 µL of 0.01 M HAuCl4, 300 µL of 0.01 M trisodium citrate and 300 µL of ice cold 0.01 M NaBH4 consecutively to 18.95 mL of Nanopure water in a round bottom flask that is vigorously stirred. The solution was stirred for 1 min, and then heated at 40 °C for 15 min to remove any unreacted reducing agent. The resulting gold seeds are approximately 5 nm in diameter and 30 nM in concentration. Next a “growth solution” was prepared, which consisted of 9 mL of an aqueous solution of 0.05 M CTAB with 50 µM NaI, 250µL of 0.01 M HAuCl4, 50 µL of 0.10 M NaOH, and 50 µL of 0.10 M L-ascorbic acid. The growth solution was swirled quickly, after which the solution became colorless. Addition of the aged gold seeds to the clear growth solution resulted in a purple solution containing gold nanoprisms and gold nanospheres after ~30 min. To synthesize ~140 nm edge length nanoprisms, the final seed concentration in the growth solution was 44.5 pM. The thickness of the nanoprisms was ~6.9 nm (Figure S9). To purify the nanoprisms from the spherical nanoparticles that form concomitantly during synthesis, depletion-force induced assembly was utilized (Proc. Natl. Acad. Sci. USA. 2012, 109, 2240). Briefly, the solution containing the gold nanoprisms was aliquoted into 1.5 mL Eppendorf tubes and NaCl was added to a final concentration of 0.1 M. The prisms were allowed to sit with the salt for ~4 h after which they were centrifuged for 15 s using a desktop centrifuge and the supernatant was removed. The assembled prisms form a pellet. The prisms were redispersed in water and centrifuged one more time to remove any excess salt. Finally, the prisms were resuspended in 50 mM CTAB. In the final solution, other nanoplates including trapezoids and hexagons are also present. Synthesis of gold nanocubes Gold nanocubes were synthesized via a previously published method (J. Am. Chem. Soc. 2012, 134, 14542). First, 7 nm diameter Au seeds were prepared by quickly injecting 0.60 mL of ice-cold, freshly prepared 10 mM NaBH4 into a rapidly stirring solution containing 0.25 mL of 10 mM HAuCl4 and 10.0 mL of 100 mM CTAC. The seed solution was stirred for 1 min and then left undisturbed for 2 h. Larger seeds were prepared by consecutively adding 200.0 µL of 10 mM HAuCl4 and 40.0 µL of 100 mM AA to a solution containing 8.0 mL H2O and 1.6 mL 100 mM CTAC. The 7 nm diameter seed particles synthesized above were diluted in 100 mM CTAC to generate a solution which was 1/100 the concentration of the original seed solution. Growth of the larger seed particles was initiated by adding 100.0 µL of the diluted 7 nm diameter seeds. The reaction was swirled immediately after the addition of the seeds and then left undisturbed on the bench until the reaction was complete. Cubes were prepared by consecutively adding 125 µL of 10 mM HAuCl4 and 30 µL of 100 mM AA to 10.00 mL of 10 mM CTAB, swirling the solution after each addition. Particle growth was initiated by adding 0.2 mL of the large seeds to the growth solution, which was then swirled and incubated at 30 °C for 3 h. Anisotropic nanoparticle deposition Au nanoplates (nanoprism, trapezoids, and hexagons) have a high ratio of surface area to volume. If no surfactants are bound to their surfaces to lower the surface energy, these nanoparticles can be easily corroded and reorganize under ambient conditions (Figure S3 right). To avoid the shape change, these nanoplates were stabilized in a concentrated CTAB solution (50 mM). Before deposition, the solution was briefly sonicated for 5-10 s in a warm water bath. After drop-casting the solution on Si substrate and drying, the substrate was quickly immersed in a petri dish filled with warm deionized water to remove excess CTAB. Removing the majority of CTAB is important for the etching. Typically, if not thoroughly rinsed with warm water, a submonolayer of CTAB could still be present on the nanoplates, as shown in AFM measurements (Figure S9). These Au nanoparticles partially covered with CTAB result in white features in the etched trenches (Figure 3). If the nanoparticles are thoroughly rinsed with warm water, the CTAB molecules are predominately removed (Figure S10). These Au nanoparticles yield the features shown in Figure 4. After synthesis, Au nanocubes and nanorods were also stabilized in CTAB solutions (50 mM). However, they are relatively more stable than the nanoplates. For single nanoparticle species etching, the deposition procedure of nanorods and nanocubes was the same as that of the nanoplates. To compare the etching rate of nanoplates and nanocubes, the two nanoparticle species were deposited on the same surface. The deposition process was optimized as follows. The Au nanocubes and nanoplates in CTAB were first centrifuged separately at 6000 rpm for 4 min. After removing the supernatant, the pellet was redispersed in deionized water and centrifuged again. Finally, the nanoparticles were dispersed in deionized water at 4 times the original concentration. The nanoplates were first drop-cast onto a fresh Si substrate. The solution droplet was gently dried with an N2 stream. While drying, a pipette tip was placed horizontally on the surface and moved across the surface to spread the solution over the entire substrate. After drying, the nanocube solution was then drop-casted and dried in similar manner.
2
Chemical etching procedure Hydrofluoric acid and hydrogen peroxide were mixed at a volume ratio of 10:1 (Extreme caution! Highly toxic! ). The high volume ratio between HF and H2O2 is important to ensure that particles etch Si vertically. The Si substrate with deposited nanoparticles was immersed in the solution for a range of etching times from 10 to 60 min. The etched Si substrate was then thoroughly rinsed with deionized water and dried in an N2 stream. After characterization with scanning electron microscopy (SEM) or atomic force microscopy (AFM), the substrate was immersed in aqua regia (HNO3:HCl=3:1 v:v, caution!) to dissolve the nanoparticles. Afterwards, the substrate was again thoroughly rinsed with deionized water and dried in an N2 stream before further characterization with SEM and AFM. SEM, AFM, and TEM Characterization AFM measurements were performed on a Dimension Icon (Bruker, Inc.). SEM analysis of the samples was conducted on a Hitachi S-4800 SEM at an acceleration voltage of 5 kV and a current of 20 µA. The probe current was set to high, and the focus mode was set to ultrahigh resolution (UHR). Only the upper second electron detector was used. For cross-sectional imaging of the etched features, samples were diced with a diamond scribe and mounted on a 45° tilting stub. TEM imaging was performed on a Hitachi STEM HD-2300A in Z-contrast mode at an acceleration voltage of 200 kV and a current of 78 µA.
Figure S1 An array of spherical Au nanoparticles patterned on a silicon substrate with scanning probe block copolymer lithography and the corresponding etched features.
Figure S2. A SEM image of penta-twinned features etched with crystalline Au nanoparticles. As the diameter increases, the Au nanoparticles start to develop facets and the resulting etched silicon features show corresponding facets.
3
Figure S3. Electron microscopy images of Au nanoprisms immediately after synthesis (left) and after sitting on a lab bench for three weeks (right). The sharp tips of the prisms become rounded over time.
4
Figure S4. Cross-sectional SEM images of nanoplates etched into a silicon substrate. A porous silicon layer can be seen only at the top surface that was exposed to HF for a long time. The majority of the silicon in the cross-section is nonporous. The Au nanoparticles used for the chemical etching are highlighted by the white arrows.
Figure S5 Traditional metal-assisted chemical etching using large Au particles for 6 h. (left) Top-down view of a large pore etched into silicon substrate. A porous silicon layer was fully developed at the top surface. (right) Cross-sectional view of a pore etched deep into the silicon substrate. A porous silicon layer was developed at the wall of the pore. The thickness of the porous layer was thinner at the tip of the pore.
Figure S6 Schematics of metal-assisted chemical etching and the formation of porous silicon layers. The two competing reactions, metal-assisted chemical etching and porous silicon formation, affect the final profile of the etched features. The reaction rate of porous formation depends on many parameters including the doping type of Si, level of the dopant, concentration of HF, and the crystal direction of Si. If the reaction rate of metal-assisted chemical etching (rMACE) is faster than that of pore formation (rpore formation), only a thin layer of porous silicon develops at the top surface and the wall of the pore, which are exposed to HF for a relatively long time (Figure S4 and S6 left). If rMACE is slower than rpore formation, the front line of the porous layer will be flat and the formation of pores will dominate over the metal-assisted chemical etching, resulting in a thick porous silicon layer (Figure S5 and S6 right, and ref. J. Phys. Chem. C 2012, 116, 13446). In the former case, there is no porous silicon layer for the diffusion of the reactant and products. If any diffusion does occur, the metalassisted chemical etching can remove the porous layer quickly because of its faster reaction rate. In the latter case, the porous silicon layer is fully developed below the metal catalysts. The porous layer can also serve as the media for the diffusion of reactants/products.
5
Figure S7. Complex features etched into a silicon substrate from an assembly of one short nanorod, one long nanorod, and a nanoprism. The scale bar is 50 nm.
Figure S8. A STEM image of Au nanorods with a high aspect ratio of ~25.
6
Fig S9. An AFM height-image (top) and line scans (bottom) of an as-deposited Au nanoprism on a silicon substrate. The nanoparticle was briefly rinsed with warm water. The CTAB residue can be observed on top of the Au surface. The thickness of the nanoprism is measured to be 6.9 nm.
7
Fig. S10. An AFM height-image (top) and line scans (bottom) of an Au nanoprism on a silicon substrate. The nanoparticle was thoroughly rinsed with warm water. Most CTAB molecules are removed from the Au surface, and thus the effect of CTAB during the etching process is negligible.
8
Materials Hydrofluoric acid (HF, 48 wt.% in H2O, 99.99%), tetrachloroauric(III) acid trihydrate (HAuCl4·3H2O, 99.9+%), cetyltrimethylammonium bromide (CTAB, 99%), cetyltrimethylammonium chloride (CTAC, 25 wt.% in H2O), sodium borohydride (NaBH4, 99.99%), L-ascorbic acid (AA, 99+%), trisodium citrate, NaOH, hexamethyldisilazane (HMDS) and hexane were purchased from Sigma-Aldrich, Inc., and used as received. Poly(ethylene oxide)-block-poly(2-vinyl pyridine) (PEO-b-P2VP, Mn=2.8-b-1.5 kg·mol-1, polydispersity index, PDI=1.11) was purchased from Polymer Source, Inc. and used as received. Hydrogen peroxide (H2O2, 30% Solution, AR Select® (ACS)) was purchased from Macron Fine Chemicals. DPN® pen arrays (Type M, no gold-coating) were purchased from Nanoink, Inc. Silicon wafers (100) were purchased from Nova Electronic Materials. Synthesis of spherical nanoparticles PEO-b-P2VP and HAuCl4·3H2O were dissolved separately in water, and then mixed together. The final ink had a PEO-bP2VP concentration of 5 mg·ml-1. The molar ratio between 2-vinylpyridine (2VP) and HAuCl4 in the final mixture ranged from 4:1 to 512:1. After stirring overnight, the solution was dip-coated onto a DPN® pen array. After drying in an N2 stream, the pen array was mounted onto an AFM (NScriptor, NanoInk, Inc.) in a chamber with controlled humidity. The relative humidity was in the range of 75-95% to control the dimensions of patterned polymer features and thus the nanoparticle size. Prior to DPN patterning, to ensure that the silicon substrate was hydrophobic, hexamethyldisilazane (HMDS) was evaporated on the surface by keeping the substrate in a desiccator with two vials of a HMDS and hexane mixture (v:v = 1) for 24 h. The patterned substrate was placed in a tube furnace for thermal annealing. The annealing conditions were programmed as follows: ramp to 150 °C in 1 h, soak at 150 °C for 4 h in Ar, cool down to room temperature in 4 h, ramp to 500 °C in 1 h, soak at 500 oC for 4 h in Ar, cool down to room temperature in 1 h. Afterwards, the substrate was again annealed at 500 °C for 2 h in air to remove any organic residue. By varying the molar ratio between 2VP and HAuCl4 or dwell time during the patterning, the average diameter of the gold nanoparticles was controlled over the 4.0-25.5 nm range. Synthesis of gold nanorods with low aspect ratio Gold nanorods with an aspect ratio of ~5.3 were synthesized as follows. Briefly, gold nanoparticle seeds (~3.5 nm diameter) were first prepared by adding 250 µL of 100 mM HAuCl4 and 600 µL of ice cold 10 mM NaBH4 to a 10 mL aqueous solution of 100 mM CTAB. The solution was stirred for ~ 2 min and then allowed to age for 2 h. The solution appeared brown. To grow gold nanorods, 10 ml of a 100 mM CTAB solution was mixed with 500 µL of 10 mM HAuCl4, 200 µL of 1 M HCl, 100 µL of 10 mM AgNO3, and 100 µL of 100 mM AA. After shaking the solution, 100 µL of the seed solution (100 times diluted) was added. The solution was then stored in a 30 °C water bath overnight. Synthesis of gold nanorods with high aspect ratio Gold nanorods with an aspect ratio of ~25 were synthesized according to methods in the literature (J. Phys. Chem. B 2001, 105, 4065). Briefly, gold nanoparticle seeds (~3.5 nm diameter) were first prepared by adding 600 µL of ice cold 0.1 M NaBH4 to a 20 mL aqueous solution containing 2.5 x 10-4 M HAuCl4 and 2.5 x 10-4 M trisodium citrate. The solution was allowed to age for 2-5 h. A three-step seeding method was used to prepare the nanorods. Three vials (labeled A, B, C), each containing 9 mL of “growth solution” consisting of 0.1 M CTAB and 2.5 x 10-4 M HAuCl4, were mixed with 0.05 mL of 0.1M ascorbic acid. Next 1 mL of the seed solution was added to vial A, swirled, and allowed to sit for 15 s. Then, 1 mL of solution A was transferred to vial B, swirled, and allowed to sit for 30 s. Finally, 1 mL of solution B was transferred to vial C and swirled. Solution C contains the gold rods with an aspect ratio of ~25 (Figure S8).
1
Synthesis of gold nanoplates Gold nanoprisms were synthesized according to the methods in the literature (J. Am. Chem. Soc. 2005, 127, 5312; Nano Lett. 2008, 8, 2526.) with minor differences. Briefly, gold nanoparticle seeds were first prepared by adding 250 µL of 0.01 M HAuCl4, 300 µL of 0.01 M trisodium citrate and 300 µL of ice cold 0.01 M NaBH4 consecutively to 18.95 mL of Nanopure water in a round bottom flask that is vigorously stirred. The solution was stirred for 1 min, and then heated at 40 °C for 15 min to remove any unreacted reducing agent. The resulting gold seeds are approximately 5 nm in diameter and 30 nM in concentration. Next a “growth solution” was prepared, which consisted of 9 mL of an aqueous solution of 0.05 M CTAB with 50 µM NaI, 250µL of 0.01 M HAuCl4, 50 µL of 0.10 M NaOH, and 50 µL of 0.10 M L-ascorbic acid. The growth solution was swirled quickly, after which the solution became colorless. Addition of the aged gold seeds to the clear growth solution resulted in a purple solution containing gold nanoprisms and gold nanospheres after ~30 min. To synthesize ~140 nm edge length nanoprisms, the final seed concentration in the growth solution was 44.5 pM. The thickness of the nanoprisms was ~6.9 nm (Figure S9). To purify the nanoprisms from the spherical nanoparticles that form concomitantly during synthesis, depletion-force induced assembly was utilized (Proc. Natl. Acad. Sci. USA. 2012, 109, 2240). Briefly, the solution containing the gold nanoprisms was aliquoted into 1.5 mL Eppendorf tubes and NaCl was added to a final concentration of 0.1 M. The prisms were allowed to sit with the salt for ~4 h after which they were centrifuged for 15 s using a desktop centrifuge and the supernatant was removed. The assembled prisms form a pellet. The prisms were redispersed in water and centrifuged one more time to remove any excess salt. Finally, the prisms were resuspended in 50 mM CTAB. In the final solution, other nanoplates including trapezoids and hexagons are also present. Synthesis of gold nanocubes Gold nanocubes were synthesized via a previously published method (J. Am. Chem. Soc. 2012, 134, 14542). First, 7 nm diameter Au seeds were prepared by quickly injecting 0.60 mL of ice-cold, freshly prepared 10 mM NaBH4 into a rapidly stirring solution containing 0.25 mL of 10 mM HAuCl4 and 10.0 mL of 100 mM CTAC. The seed solution was stirred for 1 min and then left undisturbed for 2 h. Larger seeds were prepared by consecutively adding 200.0 µL of 10 mM HAuCl4 and 40.0 µL of 100 mM AA to a solution containing 8.0 mL H2O and 1.6 mL 100 mM CTAC. The 7 nm diameter seed particles synthesized above were diluted in 100 mM CTAC to generate a solution which was 1/100 the concentration of the original seed solution. Growth of the larger seed particles was initiated by adding 100.0 µL of the diluted 7 nm diameter seeds. The reaction was swirled immediately after the addition of the seeds and then left undisturbed on the bench until the reaction was complete. Cubes were prepared by consecutively adding 125 µL of 10 mM HAuCl4 and 30 µL of 100 mM AA to 10.00 mL of 10 mM CTAB, swirling the solution after each addition. Particle growth was initiated by adding 0.2 mL of the large seeds to the growth solution, which was then swirled and incubated at 30 °C for 3 h. Anisotropic nanoparticle deposition Au nanoplates (nanoprism, trapezoids, and hexagons) have a high ratio of surface area to volume. If no surfactants are bound to their surfaces to lower the surface energy, these nanoparticles can be easily corroded and reorganize under ambient conditions (Figure S3 right). To avoid the shape change, these nanoplates were stabilized in a concentrated CTAB solution (50 mM). Before deposition, the solution was briefly sonicated for 5-10 s in a warm water bath. After drop-casting the solution on Si substrate and drying, the substrate was quickly immersed in a petri dish filled with warm deionized water to remove excess CTAB. Removing the majority of CTAB is important for the etching. Typically, if not thoroughly rinsed with warm water, a submonolayer of CTAB could still be present on the nanoplates, as shown in AFM measurements (Figure S9). These Au nanoparticles partially covered with CTAB result in white features in the etched trenches (Figure 3). If the nanoparticles are thoroughly rinsed with warm water, the CTAB molecules are predominately removed (Figure S10). These Au nanoparticles yield the features shown in Figure 4. After synthesis, Au nanocubes and nanorods were also stabilized in CTAB solutions (50 mM). However, they are relatively more stable than the nanoplates. For single nanoparticle species etching, the deposition procedure of nanorods and nanocubes was the same as that of the nanoplates. To compare the etching rate of nanoplates and nanocubes, the two nanoparticle species were deposited on the same surface. The deposition process was optimized as follows. The Au nanocubes and nanoplates in CTAB were first centrifuged separately at 6000 rpm for 4 min. After removing the supernatant, the pellet was redispersed in deionized water and centrifuged again. Finally, the nanoparticles were dispersed in deionized water at 4 times the original concentration. The nanoplates were first drop-cast onto a fresh Si substrate. The solution droplet was gently dried with an N2 stream. While drying, a pipette tip was placed horizontally on the surface and moved across the surface to spread the solution over the entire substrate. After drying, the nanocube solution was then drop-casted and dried in similar manner.
2
Chemical etching procedure Hydrofluoric acid and hydrogen peroxide were mixed at a volume ratio of 10:1 (Extreme caution! Highly toxic! ). The high volume ratio between HF and H2O2 is important to ensure that particles etch Si vertically. The Si substrate with deposited nanoparticles was immersed in the solution for a range of etching times from 10 to 60 min. The etched Si substrate was then thoroughly rinsed with deionized water and dried in an N2 stream. After characterization with scanning electron microscopy (SEM) or atomic force microscopy (AFM), the substrate was immersed in aqua regia (HNO3:HCl=3:1 v:v, caution!) to dissolve the nanoparticles. Afterwards, the substrate was again thoroughly rinsed with deionized water and dried in an N2 stream before further characterization with SEM and AFM. SEM, AFM, and TEM Characterization AFM measurements were performed on a Dimension Icon (Bruker, Inc.). SEM analysis of the samples was conducted on a Hitachi S-4800 SEM at an acceleration voltage of 5 kV and a current of 20 µA. The probe current was set to high, and the focus mode was set to ultrahigh resolution (UHR). Only the upper second electron detector was used. For cross-sectional imaging of the etched features, samples were diced with a diamond scribe and mounted on a 45° tilting stub. TEM imaging was performed on a Hitachi STEM HD-2300A in Z-contrast mode at an acceleration voltage of 200 kV and a current of 78 µA.
Figure S1 An array of spherical Au nanoparticles patterned on a silicon substrate with scanning probe block copolymer lithography and the corresponding etched features.
Figure S2. A SEM image of penta-twinned features etched with crystalline Au nanoparticles. As the diameter increases, the Au nanoparticles start to develop facets and the resulting etched silicon features show corresponding facets.
3
Figure S3. Electron microscopy images of Au nanoprisms immediately after synthesis (left) and after sitting on a lab bench for three weeks (right). The sharp tips of the prisms become rounded over time.
4
Figure S4. Cross-sectional SEM images of nanoplates etched into a silicon substrate. A porous silicon layer can be seen only at the top surface that was exposed to HF for a long time. The majority of the silicon in the cross-section is nonporous. The Au nanoparticles used for the chemical etching are highlighted by the white arrows.
Figure S5 Traditional metal-assisted chemical etching using large Au particles for 6 h. (left) Top-down view of a large pore etched into silicon substrate. A porous silicon layer was fully developed at the top surface. (right) Cross-sectional view of a pore etched deep into the silicon substrate. A porous silicon layer was developed at the wall of the pore. The thickness of the porous layer was thinner at the tip of the pore.
Figure S6 Schematics of metal-assisted chemical etching and the formation of porous silicon layers. The two competing reactions, metal-assisted chemical etching and porous silicon formation, affect the final profile of the etched features. The reaction rate of porous formation depends on many parameters including the doping type of Si, level of the dopant, concentration of HF, and the crystal direction of Si. If the reaction rate of metal-assisted chemical etching (rMACE) is faster than that of pore formation (rpore formation), only a thin layer of porous silicon develops at the top surface and the wall of the pore, which are exposed to HF for a relatively long time (Figure S4 and S6 left). If rMACE is slower than rpore formation, the front line of the porous layer will be flat and the formation of pores will dominate over the metal-assisted chemical etching, resulting in a thick porous silicon layer (Figure S5 and S6 right, and ref. J. Phys. Chem. C 2012, 116, 13446). In the former case, there is no porous silicon layer for the diffusion of the reactant and products. If any diffusion does occur, the metalassisted chemical etching can remove the porous layer quickly because of its faster reaction rate. In the latter case, the porous silicon layer is fully developed below the metal catalysts. The porous layer can also serve as the media for the diffusion of reactants/products.
5
Figure S7. Complex features etched into a silicon substrate from an assembly of one short nanorod, one long nanorod, and a nanoprism. The scale bar is 50 nm.
Figure S8. A STEM image of Au nanorods with a high aspect ratio of ~25.
6
Fig S9. An AFM height-image (top) and line scans (bottom) of an as-deposited Au nanoprism on a silicon substrate. The nanoparticle was briefly rinsed with warm water. The CTAB residue can be observed on top of the Au surface. The thickness of the nanoprism is measured to be 6.9 nm.
7
Fig. S10. An AFM height-image (top) and line scans (bottom) of an Au nanoprism on a silicon substrate. The nanoparticle was thoroughly rinsed with warm water. Most CTAB molecules are removed from the Au surface, and thus the effect of CTAB during the etching process is negligible.
8
