Most of this dossier concerns a glass skeleton: DNA origami frames assembled into a lattice and coated with silica (see Silica Nanolattice). This paper asks what else that skeleton can carry. Published in Science Advances (volume 10, eadl0604) on 10 January 2024, with the accepted manuscript deposited by Brookhaven National Laboratory as report BNL-225500-2024-JAAM, it describes using the silica replica as a scaffold for coatings of metals, metal oxides and semiconductors. The eleven authors are led by Aaron Michelson, with Chang-Yong Nam and Oleg Gang as corresponding authors. Their affiliations include Brookhaven's Center for Functional Nanomaterials and National Synchrotron Light Source II, Columbia University and Stony Brook University.
What the paper claims
The abstract presents a "general approach" to making designed, ordered 3D inorganic frameworks. It lists zinc, aluminum, copper, molybdenum, tungsten, indium, tin and platinum, plus composites: aluminum-doped zinc oxide (AZO), indium tin oxide (ITO) and platinum on AZO. The authors end on an aspiration rather than a result, saying the work "paves the road toward establishing a 3D nanoscale lithography." Their argument is that planar lithography struggles to build 3D architecture below about 30 nm, and that self-assembly methods rarely let a designer specify the architecture.
How the scaffold is built
The starting point follows the group's earlier work. DNA origami frames are folded from M13 phage scaffold DNA. Most experiments used octahedra (edge length about 29 nm), with tetrahedra (about 36 nm) in some cases. Binding strands at the vertices let the frames assemble into superlattices during a slow anneal from 50°C to room temperature at −0.2°C per hour, which takes about five days. Sol-gel chemistry then grows silica 4 to 10 nm thick on the DNA bundles. The paper gives pores of roughly 10 to 20 nm for octahedral lattices and 50 to 60 nm for tetrahedral ones. Hydroxyl groups on the silica surface serve as anchoring sites for the next layer.
Two ways in: liquid and vapour
Liquid-phase infiltration (LPI). A drop of metal-salt solution at 20 mM is placed on the silica lattice, left for 5 minutes, spun dry and heated to 250°C in air. Platinum (from sodium tetrachloroplatinate) was used to tune the method. EDS showed metal content levelling off after about 10 minutes of incubation. TEM indicated the coating was self-limiting: longer soaking did not make it thicker. Two minutes left the interior unevenly coated, so the authors used 5 minutes as the standard time. Copper, molybdenum, platinum, tungsten, indium and tin each formed coatings 4 to 5 nm thick, as reported. Gold salt is listed among the chemicals attempted, but gold is not among the elements the abstract names.
The indium–tin case shows that the post-treatment matters. Heating at 300°C in air oxidised the metals only partly and drove off tin, according to XPS. A rapid 5-minute treatment in oxygen at 600°C kept the intended composition, removed chlorine left over from the salts and produced ITO. X-ray diffraction on 95:5 and 50:50 indium:tin samples, together with XPS, supports crystalline ITO, and high-resolution TEM found crystalline domains of about 7 nm.
Vapour-phase infiltration (VPI). This is a variant of atomic layer deposition (ALD). The authors explain that standard ALD uses short pulses separated by a few seconds. In pores below 50 nm, they say, this does not give precursors time to reach the interior, so material builds up near the surface and blocks the pores. In VPI, each precursor sits on the sample under static vacuum for 600 s and is then purged for 600 s. Trimethylaluminum gave alumina (AlOx) and diethylzinc gave zinc oxide (ZnOx), with water as the oxidant.
The two oxides behaved differently. Ten ZnOx cycles filled the pores completely, and cutting to two cycles reduced the total amount without substantially reducing overgrowth. Ten AlOx cycles grew 5 to 7 nm on the struts and left the pores open, uniformly across a sample 5 μm wide. The authors attribute the difference to how strongly each precursor binds to the silica's hydroxyl groups. They also report a limit: where silica growth left pores smaller than 10 nm, VPI clogged much as standard ALD does. Some excess precursor stayed trapped in micrometre-scale domains even with the long purges.
Composites and what was measured about them
One alumina "priming" cycle before six zinc oxide cycles gave a uniform coating with open pores. Cycle ratios of 1:6, 2:5 and 3:4 (alumina:zinc) shifted the composition, with aluminum rising from 10 to more than 30 atomic percent. In a two-probe current–voltage test with patterned platinum contacts on one isolated lattice, the AZO framework carried about four orders of magnitude more current than the bare silica lattice. Photoluminescence of the zinc oxide lattice showed near-band-edge emission at about 375 nm and a broad band from 380 to 500 nm. The authors attribute the broad band to defects, impurities and dopants from the silica, and report that aluminum priming strengthened it.
The most complex structure combined both routes: VPI of alumina-primed zinc oxide, followed by LPI of platinum. The reverse order did not work as well. Cross-sectional EDS showed all three elements coating the interior. SAXS gave a peak at 0.0148 Å⁻¹, which corresponds to the 42-nm simple cubic spacing of the octahedral lattice. WAXS assigned the zinc oxide to space group P63mc and what it calls platinum oxide to Fm3m. Scanning hard X-ray nanotomography at the NSLS-II nanoprobe beamline (12 keV, a beam of about 13 nm, 135 projections) on a domain about 1 μm across showed platinum and zinc distributed through the 3D volume.
What the evidence does and does not show
The paper shows, with imaging, spectroscopy and scattering, that thin conformal coatings of several materials can be placed throughout a DNA-templated silica lattice without destroying its order. It does not test mechanical properties: the strength claims in DNA-Silica Claims vs. Evidence: The Press Multiplier, the Paper Abstract and the Patent List rest on the earlier silica paper, which this one cites as "High-strength, lightweight nano-architected silica" (Cell Rep. Phys. Sci. 4, 101475, 2023). The electrical test is a single two-probe comparison, and the optical data are spectra of a few samples. Applications in catalysis, optics and electronics are possibilities the authors propose, not results. Every structure still contains the silica and, according to the paper's schematic, the DNA underneath, so these are coated composites, not free-standing metal or oxide lattices.
Connection to the patents
The competing-interests statement says that Michelson, Subramanian, Nam and Gang are co-inventors on a pending patent application related to the work. The paper does not give a number. However, the title and claims of US20260158477A1 — 3D Metal, Metal Oxides, and Semiconductor Nanoscale Frameworks Through Templating of DNA-Programmable Lattice Scaffolds (a silicate lattice with pores of about 5 to 100 nm and an inorganic layer on top) match this paper closely. As always, the patent claims what is covered and does not show that it works. The octahedral and tetrahedral frames come from a 2020 Nature Materials paper by Tian et al., which the authors cite. Its title, on DNA-prescribed and valence-controlled material voxels, matches US12548243B2 — 3D-Organized Nanomaterials Through DNA-Prescribed and Valence-Controlled Material. Funding came from the US Department of Energy, the Army Research Office (for the inorganic templating) and the Keck Foundation.
Source notes & attribution
- https://www.osti.gov/pages/servlets/purl/2338127