ALD/ALE 2024 Session EM-TuA: Vapor Phase Infiltration

As semiconductor lithography nears sub-10 nm feature sizes with the upcoming adoption of high numerical aperture extreme ultra-violet lithography, patterning materials that enable these small dimensions are not the only challenge. Lithographic materials must facilitate high-precision pattern transfer and ensure uniform feature sizes with minimal placement errors thereby mitigating the stochastic variations that limit density scaling. We have recently introduced a concept of area-selective deposition (ASD) with vapor phase infiltration (VPI) akin to atomic layer deposition, which is a promising tool for high-precision patterning.

Our approach involves utilizing a patterned two-color surface modification monolayer, where regions of the first color serve as non-growth surfaces resembling conventional ASD schemes, while regions of the second color act as growth promoters through enhanced VPI. We engineer the two-color surface modification layers by leveraging the versatility and lithographic compatibility of graft-to-polymer brushes.

This work reports on the selective infiltration of Al₂O₃ using VPI on patterned substrates coated with growth-promoting sequence-defined polypeptoid (PP) polymers and polystyrene brushes (PS) (non-growth surface). We investigate PP side-chain chemistry to determine which monomer combinations are most receptive to ALD precursors and gain insight into how to tune the thickness of the oxide film. Furthermore, we consider how three Al precursors (triethylaluminium, trimethylaluminum (TMA), and dimethyl aluminum isopropoxide (DMAI)), which vary in ligand type and size, interact with and influence growth within the polymers and lead to different oxide film densities.

We confirm successful growth of the oxide film on the PP surface of the substrate using atomic force microscopy (AFM). Additionally, blanket substrates covered with PP or PS polymer brushes show large variations in Al uptake and selectivity, as determined by x-ray photoelectron spectroscopy (XPS) and angle-resolved XPS, depending on the ALD precursor-PP interaction. From XPS we determine that TMA creates a thick oxide film within PP and not in PS. DMAI, the larger molecule of the three and the only one containing an Al-O bond, is shown to not be selective between the two types of polymer brushes. In-vacuo infrared spectroscopy measurements provide mechanistic insight into the reactions with the various precursors. Finally, upon reactive ion etching, we demonstrate successful pattern transfer from the VPI-hardened polymer pattern onto the substrate, providing a bridge for existing patterning technologies to be paired with this selective process.

Sequential infiltration synthesis (SIS) is a vapor phase infiltration technique for the creation of organic-inorganic hybrid materials and/or inorganic nanostructures from a suitable polymer template. So far, the number of inorganic materials that can be grown by SIS is quite limited. To expand the materials library and widen the application fields, studies on kinetics, diffusivity, reaction energy and mechanism are necessary.

In this work, Al₂O₃ is infiltrated in poly (methyl methacrylate) (PMMA) and poly (lactic acid) (PLA) thin films by means of a standard ALD reactor, operating at 70°C in quasi-static mode, using trimethyl aluminium (TMA) and water. Selected polymers are characterized by the presence of ester groups that are expected to act as reactive sites for TMA. In situ spectroscopic ellipsometry reveals that significant swelling and deswelling of PMMA occur during TMA exposure and purging, respectively, due to sorption and subsequent desorption of TMA molecules that are not stably incorporated into the PMMA matrix. PLA exhibits a much larger swelling than PMMA during TMA exposure, but no significant deswelling is observed during purging, suggesting that a large amount of infiltrated TMA molecules is effectively trapped into the polymer matrix by a stable chemical bond. Accordingly, ex-situ XPS analysis demonstrates that much more Al₂O₃ is grown in PLA than in PMMA. Additionally, ex-situ XPS analysis indicates that, in both cases, TMA incorporation occurs through the formation of an Al-O covalent bond at the C-O-C group, as already observed in other biopolymers.

Two density functional theory (DFT) approaches were used to investigate the infiltration of TMA into PMMA and PLA polymers. Binding configurations and energies were modelled using a gas phase model of a ten-unit polymer. For both polymers, TMA was found to form an adduct with the oxygen in the C=O with an exothermic reaction energy, consistent with experiment. Furthermore, TMA was able to exothermically insert into the C-O-C bond of PLA, forming a covalent Al-O bond, aligning with the ex-situ XPS results. Infiltration modelling employed a periodic model with two eight-unit polymer chains. Both PMMA and PLA displayed swelling upon TMA infiltration, saturating with increasing TMA, consistent with experimental findings.

This combined experimental and theoretical study provides insight into SIS of Al₂O₃ in PMMA and PLA. The methodology can be extended to other precursors or polymers, unravelling of the complexities of SIS at the molecular level.

Vapor phase infiltration (VPI), also named sequential infiltration synthesis (SIS), is a technique derived from ALD that enables the incorporation of inorganic species into polymeric matrices. VPI can modify the properties of polymers by creating hybrid organic-inorganic materials with enhanced mechanical, electrical, optical, or catalytic functionalities. In addition, polymer patterns can be used as templates for confined growth, enabling simple conversation of organic nanostructures into inorganic ones, upon polymer removal. With these abilities, VPI is an appealing technique for a wide range of applications- from nanopatterning and sensing to energy storage devices and water filtration membranes. To continue expanding VPI processes, key challenges need to be addressed. These include new VPI chemistries, better understanding and control over growth processes and depth gradients, as well as new polymer chemistries, morphologies, and architecture for VPI.

Here I will describe our efforts in addressing these challenges. To expand VPI chemistries, we demonstrated the use of TDMA ligands in VPI, creating SnO₂ and HfO₂ nanostructures from block copolymer templates. Despite the large volume of these precursors that reduce diffusion through the polymer, exploring the parameter space of VPI enabled us to rationally design these processes. We also explored new ways for fabricating multi-material structures via VPI by controlling either the precursors’ diffusion or their ratios in a three-precursor process, leading to spatially controlled structures and doped metal oxides. We investigated ZnO nucleation and growth in an array of polymers, setting design rules for this important VPI process. Finally, we demonstrate how this knowledge can be harnessed for a new application- protecting degradable polymers from UV radiation. We show how VPI can enhance the growth of ALD on the polymer surface leading to the synergic effect of high durability of these sensitive polymers under destructive UV radiation.

For polymer gas separation membranes that typically rely on the molecular sieving effect, an improved selectivity typically accompanies a lowered permeability, thus limiting a simultaneous improvement of the two parameters (i.e., Robeson upper bound). Here, we discover that polybenzimidazole (PBI), one of the leading membrane materials for H₂/CO₂ separation, can exhibit simultaneously increased H₂ permeability and H₂/CO₂ selectivity, surpassing the Robeson upper bound, once treated by simple, low-temperature (85 °C) 1-cycle alumina atomic layer deposition (ALD).¹ We find that the normal 1-cycle alumina ALD (i.e., not using an intentional exposure-mode protocol), comprising 15 msec sequential dosing of trimethylaluminum (TMA) and water with 10 sec intermediate purging, leads to a surprising vapor-phase infiltration (VPI) of alumina into PBI matrix, penetrating over several micrometers deep into the film. The resulting AlO_x-doped PBI hybrid membrane features remarkably increased H₂ permeability by 120% - 270% and H₂/CO₂ selectivity by 30% at 35 - 200 ℃, compared to the untreated control. This unusual, simultaneously enhanced H₂ permeability and H₂/CO₂ selectivity were attributed to the disrupted PBI polymer chain packing by the AlO_x infiltration, which was found to be mediated by the reaction between infiltrating TMA and amine groups available in PBI. The suggested mechanism was further supported by the TMA-only-infiltrated PBI membrane that displayed a similarly increased H₂ permeability. Given the simple and fast treatment protocol, this demonstration showcases a facile and scalable way of utilizing ALD and VPI for engineering and improving polymeric membranes important for CO₂ capture and future H₂ economy.

¹L. Hu et al., “Few-cycle atomic layer deposition to nanoengineer polybenzimidazole for H₂/CO₂ separation”, Chem. Eng. J. 479, 147401 (2024)