Atomic layer deposition of biobased nanostructured interfaces for energy, environmental and health applications

Introduction

Nanostructure science and technology are a broad and interdisciplinary area of research and development that has been exponentially growing in the past few years. They are already having a significant commercial impact, which will assuredly increase in the near future. Engineered nanomaterials are resources designed at the molecular (nanometer) scale to take advantage of their small size and novel properties which are generally not seen in their conventional, bulk counterparts. The two main reasons why nanomaterials can have different properties are: (i) the increase of relative surface area and (ii) the new quantum confinement effects. Nanomaterials have a much greater surface area to volume ratio than their conventional forms, which can lead to greater chemical reactivity and affect their strength [1]. Also at the nanoscale, quantum confinement effects can become much more important in determining the materials properties and characteristics, leading to novel optical, electrical and magnetic behaviors. These properties allow great impacts in electronics, medicine, energy harvesting [2–4], sensors [5] and other fields [6–8].

Nanomaterials designing can be basically described by two approaches: (i) the top down: one refers to slicing or successive cutting of a bulk material to get nanosized particle and (ii) the bottom up: one refers to the approach to build a material up from the bottom: atom-by-atom, molecular-by-molecular or cluster-by-cluster. Using these two approaches, many different methods have been used to design nanostructures such as wet chemical synthesis of nanomaterials, mechanical grinding, heating and gas phase synthesis of nanomaterials (such as chemical vapor deposition [CVD], physical vapor deposition [PVD], atomic layer deposition [ALD] and laser ablation). The gas-phase synthesis methods are of increasing interest because they allow controlling process parameters in order to be able to design the size, the shape and the chemical composition of the obtained nanostructures. Among the gas phase synthesis methods, the atomic layer deposition (ALD) shows several advantages in the synthesis of nanostructured materials such as [8–10]: (i) a wide range of deposition materials: oxide, nitride, carbide and others, (ii) a conformal coating ability on high aspect ratio templates, (iii) a thickness control on the angstrom range, (iv) a high chemical purity of the deposited films and (v) a high chemical composition control of the deposited layer.

In this context, the aim of this paper is to give a brief review on the synthesis of different nanostructured materials based on the ALD. We will show mainly examples of how this method can be used to design single nanopores for sensing applications and membrane for gas separation or water treatment in which the performance depends on the nanostructure morphologies/interfaces and the immobilization conditions of biomolecules.

Atomic layer deposition (ALD)

ALD is a vapor deposition method of ultrathin layers. It was known for his debut under the name of atomic layer epitaxy (ALE) [11]. Some sources give the origins of ALD to Professor Aleskovskii and his team in 1960 that realized the deposition of TiO₂ from TiCl₄ and H₂O, as well as the deposition of GeO₂ from GeCl₄ and H₂O [12]. ALD is a deposition technique derivate from the CVD technique. It based on two self-limiting reactions absolutely separated in gas phases [9]. When two precursors react on gas phases during the CVD deposition to produce a thin film on the surface of the substrate, the same precursors react separately in ALD with the substrate surface to produce a uniform coating. No other thin film technique can approach the conformity achieved by the ALD on high aspect ratio structures.

ALD processes permit the deposition of a large variety of materials classes, such as metals, oxides, nitrides, sulfides and phosphates [9]. The chemistry taking place in ALD is rather rich as well. ALD can be used for the deposition of thin films onto various supports [1, 2, 4, 13–17]. It allows the coating of flat surfaces and complex structures (Fig. 1) in a conformal and homogeneous manner with a precise control of the thickness of the deposited film in the range of a few angstroms [1, 8]. In addition the ALD technique (mainly of thin oxide films) leads to the formation of –OH terminal on the surface of the ALD layer. The OH group could be grafted with different chemical functions including self-assembled monolayers to enhance the biocompability of these layers. These characteristics, combined to the versatility in term of materials that can be deposited, make ALD a technique of choice for the fabrications of complex membranes for the separation and molecular recognition at the nanoscale.

Fig. 1:

SEM images of (a) ALD nanolaminates of ZnO and Al₂O₃ cross-sections with bilayer thickness of 100 nm, (b) Urchin-like substrate architecture designed by ALD [2, 4], (c) ZnAl₂O₄ multi-concentric nanotubes enabled by ALD and (d) ALD ZnO thin film on Si nanowires [18] for photovoltaic applications.

ALD for nanopore technology

Sensors based on single nanopore technology aims to develop tools for the detection of biomacromolecules and nanoparticles [19, 20]. Among all the potential applications, DNA low cost and fast sequencing is the challenge of the 21st century which academic researchers and companies have been trying to achieve since the past two decades. This field started in 1994 when Kasianowitcz et al. [21] showed the possibility to discriminate polynucleotides using a α-hemolysin inserted in a lipidic bilayer. For the last 10 years, with the development of nanotechnology, several teams were interesting by solid-state nanopore. The latters are constituted either by polymer track-etched membrane, silicon based membrane or graphene nanosheet [22]. Whatever the strategy or the material used to tailor the nanopore, the limitations are the difficulty to define a method able to achieve a good control of the diameter size and permits the functionalization of nanopore surface wall. Atomic layer deposition is a method rather neglected to design single nanopore. Nevertheless, its potentiality was clearly reported in the case of multipore membrane since it permits achieving a homogeneous coating on ultra-high aspect ratio nanopore of alumina membrane L/D ∼5000 [23, 24].

Since 2004, atomic layer deposition has been used to design low aspect ratio nanopores (Fig. 2). Chen et al. tuned a SiN nanopore by Al₂O₃ deposition to produce nanopore around 2 nm [25]. Inside this nanopore they studied the translocation of DNA [26]. They demonstrated that this strategy can passivate the non-ideal surface of SiN and obtain a better signal to noise ratio. This method allows fashioning the nanopore as required (length and diameter). To upgrade to nanopore arrays, the main limitation is the fabrication of similar nanopores before the ALD deposition. Torre et al. used TiO₂ to design arrays containing 64 nanopores with homogeneous diameter (between 4 and 10 nm) [27]. These arrays were tested to detect DNA linked with a fluorescent quantum dot using both electrical and optical signals. Another example was demonstrated by Chen et al. [28]. The latter group have used ALD of TiO₂ to reduce nanopore from 2.6 nm to 2 nm and to study the influence of the surface on DNA translocation by voltage clamp method.

Fig. 2:

Illustration of nanopore design by ALD and typical current trace record during translocation event (here a polystyrene microsphere with a diameter of 100 nm through a nanopore with a diameter of 104 nm and a length of 13 μm).

Even if the nanopore array presents potentiality for sensing application, more fundamental work on macromolecules and nanoparticles are essential. In this domain, ALD is still a method of choice. Indeed, the utilization of oxide material creates functional groups on the surface, typically –OH groups. The latter can be modified by more and less complex silane group. This step can prove to be indispensable if material is sensible to corrosion induced by saline solution under an electric flied such as ZnO [29]. Typically the surface modification of oxide deposited by ALD was used to demonstrate the influence of nanopore surface charge on macromolecules both mobility and the energy barrier of entrance of the macromolecules inside the nanopore. Starting from SiN nanopore coated by Al₂O₃, Kim and coworkers [30] have grafted aminopropyle silane. They demonstrate that a nanopore with an opposite charge than DNA induces an increase of both translocation time and energy barrier of DNA entrance. These counterintuitive results were also reported by Cabello-Aguilar et al. [31] in the case of carboxylate modified nanoparticles. Authors have reported that if the nanopore exhibits a negative surface charge, the nanoparticle mobility increases and the global energy barrier of entrance inside the nanopore decreases with its diameter, converse to the uncharged nanopore. Similar results were reported more recently by Lepoitevin et al. on translocation of poly adenosine [32]. Another interesting point of this study is the application of ALD to tune a single nanopore with high aspect ratio by successive layer of Al₂O₃/ZnO. The same strategy was used to tune hydrophobic nanopores with a diameter of 5 nm for discriminate small carboxylate PEG [33]. Based on cylindrical nanopore designing by ALD, Lepoitevin et al. have proposed a pH responsive nanopore by functionalization with PEG-biotin-avidin. The latter nanopore is closed at pH below 4 due to PEG/protein interaction at the nanopore entrance. Additionally this nanopore permits to detect a protein labeled with a biotin group [34].

Besides artificial solid-state nanopores, another strategy attempt to emerge is based on the direct confinement of biological channel inside a solid state nanopore. This strategy aims to combine the advantages of both solid-state nanopores (robustness) and biological channels (selectivity) [35, 36]. In a different first attempt to confine a biological channel in the nanopore, the biological properties were not recovered even if interesting selectivity was demonstrated [37, 38]. Conceptually to recover properties of a confined biological channel inside a nanopore, the latter must own: (i) a diameter close to the hydrophobic part of the channel, (ii) surface functions which allow mimicking of phospholipid chain/channel interactions and (iii) a geometry which allows channel relaxation during transportation process [17]. This concept was demonstrated by Abou-Chaaya et al. [39] in the case of simple ionic channel: gramicidin A. Authors have fashioned hydrophobic nanopore from 10 to 2 nm by ALD. The conductance measurements reveal for the first time an ionic transport mechanism through gA inside a polymeric solid state nanopore similar to that in biological membranes when the nanopore diameter become close to the gramicidin one. In another study, Cabello-Aguilar et al. [17] inserted a α-hemolysin inside a hydrophobic nanopore tuned by ALD. This hybrid system exhibits the same polynucleotide discrimination properties than α-hemolysin in the biological membrane.

As previously mentioned the example of using ALD to tune a single nanopore are rare but present several potentialities. At first the tuning of the solid-state nanopore will find applications in sensors as well as in fundamental understanding of the ionic transport or the macromolecules translocation phenomena (Fig. 3). For hybrid biological/artificial nanopore, the design of the nanopore by ALD is essential. The recent proofs of feasibility have opened new perspectives. Indeed one can imagine confining aquaporin to create selective filters for water and ligand gate channels to detect the presence of toxins.

Fig. 3:

Examples of applications which should be achieved using a nanopore tailored by atomic layer deposition.

ALD for membrane applications

The high efficiency and the low cost of membrane technologies make it a promising technology for different applications such as gas purification, water desalination, catalysis and environmental issues [40]. One of the separation processes is the molecular sieve mechanism. It is a size based mechanism that depends on the particles size and the pore diameter in the membrane. Most industrial membranes have a sub-nanometer up to micrometers pores diameter. However, some separation processes especially in gas separation process required an angstrom pore size range. From the different methods to control the pore diameter, ALD has attracted attention due to the conformal coating and the high thickness control offered by this technique.

Track etched polycarbonate membranes with diameter and thickness of 25 mm and 6 mm, respectively and with 30 nm pores size have been reduced by Al₂O₃ ALD. The scanning electron microscopy (SEM) image shows the continuous pore reduction when we increase the number of ALD cycles from 10 to 300 cycles. The TEM image after dissolving the polymers shows a uniform coating along the pores. The Al₂O₃ ALD inorganic coating improves the membrane hydrophilicity and enhances the chemical stability to acids and organic solvents. A decrease in water flux and more protein retention have been detected when the pores size decreases [41]. In-situ N₂ and Ar permeance measurements have been performed in alumina tubular membranes during the Al₂O₃ ALD cycles demonstrated that the ALD can reduce the pore diameter to molecular size [42]. The same experiment has also been performed on alumina tubular membranes during SiO₂ and TiO₂ ALD cycles [43]. The N₂ permeance measurements show a progressive reduction in pores diameter during the SiO₂ and TiO₂ ALD cycles corresponding to 1.3 ± 0.1 Å per SiO₂ cycle and 3.1 ± 0.9 Å per TiO₂ cycle. Cross-sectional SEM and electron probe microanalysis (EPMA) have been performed on anodic alumina membranes with Al₂O₃ ALD and shows a uniform coating when using sufficient reactant exposure times [23]. EPMA measurement following ZnO ALD of anodic alumina membranes shows the infiltration of Zn into the nanopores when the exposure time is increased. Mesoporous silica membranes have been tuned using catalyzed atomic layer deposition (C-ALD) of silicon dioxide. Pyridine was used as a catalyst to reduce the deposition temperature [44]. Gas separation measurement shows a transition from the Knudsen diffusion mechanism to a molecular sieving mechanism. A selectivity of 8.6 between CH₄ and H₂ has been measured at 473 K after the SiO₂ ALD deposition.

In order to increase membrane selectivity, ALD has been used to enhance surface reaction between the membrane surface and the gas phases by activating the membrane surface with different functions. ALD amino functionalization has been performed on silica membrane in order to enhance CO₂ transport. It is found that high loadings of amino groups, in which interaction with the silica surface is minimized, promote the highest CO₂ transport [45]. Track etched polycarbonate membranes have been also tuned by TiO₂-ALD and then gas permeance measurements were performed to assure the tailoring of pore sizes in PC membranes using the TiO₂-ALD [46]. Nano-porous supported QI-phase LLC polymer membranes has been modified by Al₂O₃ ALD [47] to reduce the pore diameter to the angstrom range for gas separation application. Ten cycles of Al₂O₃ ALD have been enough to enhance the selectivity of H₂/N₂ from 12 to 65 with a decrease of 40 % in the H₂ permeability. Diatom species have been coated by TiO₂-ALD [48] to reduce the pores size from 40 nm to less than 5 nm in order to enhance the filtration properties of the membrane. High-density polyethylene (HDPE) particles (16 and 60 nm) have been coated with a thin layer of Al₂O₃ using a fluidized bed reactors [49] then a polymer/ceramic nanocomposite membranes were fabricated by extruding alumina coated HDPE particles. An inclusion of 7.29 vol.% alumina flakes led to reduce of the diffusion coefficient to the half compared to the uncoated particles. Due to the voids formed during the extrusion process at the polymer/ceramic interface, an increase in the permeability of the membrane has been detected.

Drobek et al. [40] reported an innovative approach based on the ALD of ZnO thin films deposited on the grains of a macroporous ceramic support and their subsequent conversion to ZIF-8 using a 2-methylimidazole/methanol solution and under solvothermal conditions. The physico-chemical characterization of the ZIF-8/ZnO/α-Al₂O₃ nanocomposite membranes was completed by a study of their gas transport properties. Reproducible ZIF-8/ZnO/α-Al₂O₃ nanocomposite membranes were produced and tested in the separation of binary gas mixtures: membranes were found to extract H₂ from H₂/CO₂ and H₂/CH₄ equimolar gas mixtures with selectivities of about 7.8 and 12.5, respectively, measured at 100 °C.

Electrofluidic applications or electrical manipulation of charged species like ions, DNA, proteins, and nanoparticles have a high importance in solid state membrane fabrication. In such systems, the pore diameter and the surface chemistry (surface charge, hydrophobic, hydrophilic, etc.) have attracted attention. Controlling such parameters can lead to the fabrication of electric circuit elements, such as diodes and transistors or to biological elements for DNA or proteins sieving and sensing. ALD of TiO₂ has been performed on multipores of Si₃N₄ sandwiched between 2 TiN layers (TiN 30 nm/Si₃N₄ 20 nm/TiN 30 nm) membrane after an E-beam lithography and reactive ion etching (RIE) processes to make the pores [50] and to reduce the pore diameter to sub 10 nm. This membrane has been used as ionic field effect transistor (IFET) which is the electrofluidic version of the semiconductor field effect transistor (FET) that works on ions instead of electrons.

The surface of nanoporous alumina membranes has been modified by ZnO-ALD for biological applications [51]. ZnO coating enhances the antimicrobial activity of the alumina membrane against Escherichia coli and Staphylococcus aureus bacteria. A volatile organic compounds (VOC) filter has been elaborated with two different structures: the first one is the TiO₂ ALD modification of an anodic aluminum oxide (AAO) membrane, and the second one is the TiO₂ ALD modification of a nanodiamond substrate [52]. The adsorption capacities test for toluene such as VOC shows that the TiO₂ surface can allow adsorption of toluene. Ceramic microfiltration membranes with an average pore size of 50 nm have been tuned using Al₂O₃ ALD [53]. After the Al₂O₃ ALD deposition, the membranes show a decrease in water flux, and an increase in the retention of bovine serum albumin. After 600 ALD cycles water flux decreases from 1698 L (m² h bar)⁻¹ to 118 L (m² h bar)⁻¹, and BSA retention increases from 2.9 % to 97.1 %.

Conclusions

In conclusion, the precise control of the thickness of the ALD layers combined to the high step coverage permitted to introduce new approaches for nanopore and membrane design. Tuning the pore diameter, geometry and chemical interactions with the surface species may lead to optimize separation membranes to meet the requirements for many different applications.

The example of using ALD to tailor single nanopores are rare but present several potentialities. The studies have shown that tuning both geometry and size of the nanopore will be essential to further both applications and fundamental investigations. Currently, nanopore fashioned by ALD have permit reporting interesting results about hybrid biological/artificial nanopore. The latter is the most promising strategy to combine the advantage of biology (selectivity) and material (robustness), but requires well-defined solid-state nanopore.

Other applications relying on the nanostructured of interfaces by ALD such as biofuel cells and optical and electrochemical biosensors are very interested in these processes. Although the nanostructured interfaces designed by ALD for biofuel cells applications have not been reported yet, work in progress in our groups show that the TiO₂ deposited by ALD can efficiently immobilize enzymes that allow potential applications in biofuel cells. Recently we reported the tuning of optical and electrical properties of ZnO thin film deposited by ALD using different approaches: (i) control of the ZnO thickness [15], (ii) deposited of Al₂O₃/ZnO nanolaminates [14], (iii) using electrospun nanofibers as a template [5, 54, 55], and (iv) using ZnO as an active medium in a fiber-optic Fabry–Perot interferometer [56]. The improved optical and electrical properties of ZnO deposited by ALD makes them prominent materials for optical and electrical biosensors applications.