1. Introduction

Nanostructured materials (NMs), typically one of their physical dimensions in the range of 1–100 nm ^[1,2,3], have been attracting many interests due to their exceptional properties over traditional bulk materials. With material size down to the nanoscale, the percentage of atoms on the surface and the surface-to-volume ratio (or specific surface area) dramatically increase ^[4,5]. More atoms on the surface may have much more dangling bonds to actively form bonds with adjacent molecules or particles, affecting the intrinsic properties of NMs ^[4]. For instance, NMs can exhibit many exotic properties, such as increased chemical activity ^[3], lowered melting point ^[3], large band gap ^[5], fast electrons/ions transportation ^[6], highly efficient photon absorption ^[7], and improved mechanical stress and strain ^[8]. Two general synthesis strategies have been practiced for NMs, i.e., top-down methods (e.g., optical ^[9,10], E-beam ^[11,12,13], nanoimprint ^[14,15,16], and scanning probe ^[17,18,19] lithography methods) and bottom-up methods (i.e., atomic layer deposition ^[20,21,22], solution-based methods ^[23,24,25], and vapor phase deposition ^{[26,27,28,29]}). The bottom-up methods have shown tremendous promises in manipulating atoms and molecules to establish complex nanostructures with dimensions of < 10 nm ^[10], which is difficult to be achieved through the top-down processes ^[30,31].

The classification of nanostructured materials by dimensionality was first proposed by Fiorani and Sberveglieri in 1994 ^[32], suggesting zero-dimension (0D), 1D, 2D, and 3D. The classification later has experienced some modifications. Now, 0D NMs are defined as the materials with all three dimensions in nanoscale and no significant difference in size between the longest and shortest axes ^[33,34], e.g., nanoparticles ^[35], core-shell quantum dots ^[36,37], nano-onions ^[38], and hollow nanospheres ^[39]. 1D NMs are those that two dimensions are within the nanoscale and one dimension beyond nanoscale ^[40], e.g., nanowires ^[41,42], nanorods ^[43], nanotubes ^[41], and nanoribbons ^[42]. 2D NMs are generally considered as materials with a thickness of a few atomic layers and the other two dimensions beyond the nanoscale ^[33], e.g., nanosheets ^[44,45], nanoplates ^[46,47,48], nanodisks ^[49] and nanowalls ^{[50,51,52,53]}. 3D NMs are the materials having all three dimensions beyond nanoscale but exhibiting the merits of nano-size effect ^[33]. Thus, herein the dimension refers to the morphological characteristics of resultant materials, not their inherent structures.

To date, NMs have explored their applications in a large variety of areas, such as agriculture ^{[54,55,56,57]}, food ^[58,59,60], medicine ^{[61,62,63,64,65,66]}, environmental science ^[67], cosmetics ^{[68,69,70,71,72]}, textiles ^{[73,74,75,76,77]}, water purification ^[78,79,80], and sensor technology ^{[81,82,83,84]}. In addition, NMs are very important for developing energy harvest devices (e.g., solar cells) and energy storage devices (e.g., lithium-ion batteries). For example, solar cells have applied semiconductive materials of 0D (nanoparticles) ^{[85,86,87,88,89,90]}, 1D (nanowires, nanotubes) ^{[91,92,93,94,95,96,97,98]}, 2D (nanodisks, nanosheets, and multilayered nanofilms) ^{[99,100,101,102,103,104]}, and 3D (mesoporous structures) ^{[105,106,107,108,109]} nanostructures. Employing nanostructured semiconducting cathodes, the solar cells were found to deliver higher photo-conversion efficiency than bulk semiconducting cathodes ^[110]. In addition, lithium-ion batteries (LIBs) have been dominating portable electronics since commercialization in 1991 due to their high energy density ^[111]. However, many practical problems have been plaguing the development of LIBs ^[112], such as volume expansion of electrode materials, lithium dendritic growth and formation of solid electrolyte interface (SEI). NMs are very effective to resolve these problems via surface protection on electrode, separator modification, and preparation of solid state electrolyte, thus improving the cycling life, electrochemical and mechanical stability of the battery systems ^{[112,113,114,115]}.

Since the very beginning of the 21^st century, atomic layer deposition (ALD) has emerged as a new bottom-up process for various nanostructures ^[116,117]. ALD features its atomic-scale growth controllability ^[118,119], unrivaled conformal coatings in high-aspect-ratio (HAR) substrates up to 100,000:1 ^[120], excellent uniformity over large-scale substrates ^[121,122], and low growth temperature (<400 ℃ and even down to room temperature) ^{[123,124,125]}. As a result, ALD has precisely enabled various nanostructures. To date, ALD has been applied for a large range of inorganic materials, including elements ^[126], oxides ^{[127,128,129]}, nitrides ^{[130,131,132]}, sulfides ^{[133,134,135]}, and ternary or even high-level compounds ^{[136,137,138]}. Recently, the fundamentals of ALD and its applications for nanostructures have been discussed in literatures ^[139,140]. For instance, Marichy et al. ^[139] in 2012 have summarized the applications of novel NMs by ALD in energy conversion and storage.

Analogous to ALD, molecular layer deposition (MLD) as the sister technique is exclusively for depositing purely organic or hybrid organic–inorganic films. MLD to date has synthesized many different polymeric films, e.g., polyamide ^[141], polyurea ^[142], polyurethane ^[143], polyimide ^[144,145], and polyazomethine ^[146]. The organic–inorganic hybrid films via MLD include alucones ^[147], zincones ^{[148,149,150,151,152,153,154,155]}, zircones ^[156,157], titanicones ^{[136,158,159,160]}, hafnicones ^[161], and vanadicones ^[162]. All these have recently been well documented in the review by Meng ^[163].

In this review, we provide an overview on the strategies via ALD or/and MLD for different NMs. Following this introductory section, a brief description on both ALD and MLD techniques is provided. NMs via ALD or/and MLD are discussed in the following parts. Then, we highlight the enhancement of several promising applications using NMs fabricated by ALD or/and MLD, including surface engineering, catalysis, and batteries. The last section summarizes the significance of NMs and outlooks the future development of NMs via ALD/MLD.

---

2. Briefing on ALD/MLD

The ALD technique was emerged in the 1960s in former Soviet Union ^[164,165] and the 1970s in Finland ^[20,166]. A typical ALD cycle of binary compounds is composed of two alternative self-limiting gas–solid reactions induced by two exposures of precursors ^{[167,168,169,170,171,172]}. A representative ALD process is the ALD-Al₂O₃ using trimethylaluminum (TMA) and water as precursors. The first step is the introduction of TMA to react with a substrate surface, as described by Eq 1 (see Figure 1a) ^[163]. Then, a purging step is used to remove excessive TMA and the byproduct CH₄. In a following step, the second precursor of H₂O is dosed and therefore induces another surface reaction, as described by Eq 2. Then, another purging is conducted to evacuate excessive H₂O precursor and the gaseous byproduct CH₄.

$├{\rm{OH + Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{3}}}\left( {\rm{g}} \right) \to ├{\rm{O-Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{2}}}{\rm{ + C}}{{\rm{H}}_{\rm{4}}}\left( {\rm{g}} \right)$

(1)

$├{\rm{O-Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{2}}}{\rm{ + 2}}{{\rm{H}}_{\rm{2}}}{\rm{O}}\left( {\rm{g}} \right) \to ├{\rm{O-Al}}{\left( {{\rm{OH}}} \right)_{\rm{2}}}{\rm{ + 2C}}{{\rm{H}}_{\rm{4}}}\left( {\rm{g}} \right)$

(2)

where "├" refers to the substrate surface and "g" indicates the gaseous phase of precursors or byproducts. Apparently, ALD deposits inorganic thin films in a repeatable layer-by-layer mode, and the film growth is typically influenced by three key parameters: substrate, temperature, and precursor. Different from other source-controlled vapor-phase methods (e.g., chemical and physical vapor deposition), ALD is surface-controlled and this leads to many advantages in film deposition, such as extremely uniformity, unrivaled conformality, and excellent controllability at the atomic level growth.

Similar to ALD, MLD also is a surface-controlled vapor phase process (see Figure 1b). ALD is exclusive for inorganic materials, while MLD features its preciseness for growing purely organic or inorganic–organic hybrid films. A typical MLD process can be exemplified by the deposition of the alucone AlEG, using TMA and ethylene glycol (EG, HOCH₂CH₂OH) as precursors ^[147], as described in Eqs 3 and 4.

$├{\rm{OH + Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{3}}}\left( {\rm{g}} \right) \to ├{\rm{O-Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{2}}}{\rm{ + C}}{{\rm{H}}_{\rm{4}}}\left( {\rm{g}} \right)$

(3)

$├{\rm{O-Al(C}}{{\rm{H}}_{\rm{3}}}{{\rm{)}}_{\rm{2}}}{\rm{ + 2HOC}}{{\rm{H}}_{\rm{2}}}{\rm{C}}{{\rm{H}}_{\rm{2}}}{\rm{OH}}\left( {\rm{g}} \right) \to ├{\rm{O-Al(OC}}{{\rm{H}}_{\rm{2}}}{\rm{C}}{{\rm{H}}_{\rm{2}}}{\rm{OH}}{{\rm{)}}_{\rm{2}}}{\rm{ + 2C}}{{\rm{H}}_{\rm{4}}}\left( {\rm{g}} \right)$

(4)

Apparently, the molecular fragment of –OCH₂CH₂O– attached in the MLD-alucone is far much larger than the atomic part of –O– in the ALD-Al₂O₃. This makes the MLD film of the aluminum alkoxide (i.e., AlEG) grow much faster than the analog of the ALD-Al₂O₃. More details about MLD processes and growth characteristics can be found in the recent review by Meng ^[163].

In spite of these progresses, ALD and MLD are still experiencing their rapid expanding phases, and they are becoming new research thrusts in many applications such as surface engineering, new energies, and catalysis. Along with their widening paces, the costs of ALD and MLD processes are hopefully to reduce in the near future, for they are closely related to the needs and supplies of their precursors. At the current stage, the precursors of ALD and MLD processes in most cases are only needed in small amounts, leading to their high costs.

---

3. Nanostructured materials by ALD

---

3.1. 0D NMs

Due to its surface-controlled nature, ALD exhibits a great promise for fabricating novel 0D nanostructures. The synthesis of copper nanoparticles via ALD was first reported in 2003 by Johansson et al. ^[173]. In the study, a three-step ALD process was conducted: (ⅰ) dosing CuCl, (ⅱ) dosing H₂O to react CuCl with the production of CuO particles, and (ⅲ) dosing H₂ to reduce CuO into Cu particles. SEM images showed that, after 400 cycles of ALD Cu deposition on alumina substrates at 450 ℃, there were well-distributed Cu nanoparticles in average diameter of 60 nm.

ALD also enables to coat a variety of 0D substrates into new 0D nanostructures. ALD was first reported for coating 0D nanoparticles in 2005 by George's group ^[174], when SnO₂ coatings were deposited on Al and ZrO₂ nanoparticles in the range of 250–1000 ℃. The researchers used SnCl₄ and H₂O₂ as precursors for the SnO₂ deposition. The resultant Al@SnO₂ nanoparticles were then tested for thermite applications, which reacted much more quickly and violently than the uncoated Al particles. Subsequently, there have been an increasing number of studies using ALD to coat various thin films (i.e., Al₂O₃ ^{[175,176,177]}, TiO₂ ^[178], and ZnO ^[178]) over nanoparticle substrates (i.e., SiO₂ ^[176,179], ZrO₂ ^[179], Ag ^[178], Ni ^[180], and polyethylene ^[175]) for 0D core–shell nanostructures. For instance, Hakim et al. ^[177] coated 5 nm Al₂O₃ over 26 nm ZrO₂ nanoparticles using TMA and H₂O as ALD precursors. Uniform and conformal Al₂O₃ thin films were confirmed by transmission electron microscopy (TEM) and scanning electron microscope (SEM). Similarly, Hakim et al. ^[176] deposited 5 nm Al₂O₃ on 40 nm SiO₂ nanoparticles using the same ALD procedure (see Figure 2a).

Analogous to the 0D core-shell nanoparticle strategies, the ALD technique has also been used for developing various hollow nanospheres. In 2006, Kei et al. ^[181] conducted the first effort to form a smooth and conformal ALD Al₂O₃ coating on hydrothermally synthesized polystyrene nanospheres. The Al₂O₃ coating was deposited at 200 ℃ for 200 ALD cycles using TMA and H₂O as precursors. The Al₂O₃ film thickness was confirmed being ~24 nm by X-ray reflectivity (XRR). Through heating the resultant core-shell nanostructures in air, Kei et al. ^[181] enabled the hollow nanospheres of Al₂O₃. The similar strategy then has been applied for producing other nanospheres. Ras et al. in 2007 ^[182] deposited a conformal coating of Al₂O₃ at 80 ℃ on another polymer nanospheres of ~40 nm synthesized using polystyrene (PS) and poly-(4vinylpyridine) (P4VP) block copolymers. Then, the polymer cores were thermally decomposed at 300 ℃ in the N₂ atmosphere, yielding hollow Al₂O₃ nanospheres. This work also demonstrated the ALD capability in depositing films over temperature-sensitive polymers. Moreover, TiO₂ and ZnO hollow nanospheres were also reported by Justh et al. ^[183]. TiO₂ and ZnO coatings were first deposited on hydrothermally synthesized carbon nanospheres (CNSs) using the precursors of TiCl₄/H₂O and DEZ/H₂O at 250 ℃, respectively. Then, the CNS-metal oxide composites were heated to 700 ℃ in ambient environment to remove the CNS cores (see Figure 2b, c).

Furthermore, derived from 0D core-shell strategies, a facile and robust approach for fabricating highly ordered TiO₂ nanobowls via ALD was first demonstrated by Park et al. in 2012 ^[184]. The fabrication process of the patterned 0D TiO₂ nanobowls includes a series of procedures: (ⅰ) the self-assembly of PS nanospheres on glass and polyethylene terephthalate (PET) substrates via a scooping transfer technique (see Step A in Figure 2d), (ⅱ) an ALD-deposited 48-nm thick TiO₂ coating over PS nanospheres at 80 ℃ using the precursors of titanium(Ⅳ) iso-propoxide and water (see Step B in Figure 2d), (ⅲ) a CF₄-based reactive ion etching (RIE) process to remove the top of the TiO₂-coated PS nanospheres (see Step C in Figure 2d), and (ⅳ) a annealing process at 450 ℃ for 2 h to remove the PS nanospheres and also transform the amorphous phase of the ALD-deposited TiO₂ to the crystalline anatase phase (see Step D in Figure 2d). The top- and side-view SEM images displayed the eventual 2D TiO₂ nanobowls (see Figure 2e), respectively. The TiO₂ nanobowl structures could be scaled up to various substrates for the potential applications in optical, electronic, and biological devices.

To date, there has been a large number of papers documented in literature for 0D nanostructures via ALD processes. ALD coatings are very feasible in tuning the surface properties of the original nanoparticles, such as serving as protective coatings to prevent particles from oxidation ^[185], reducing electrical conduction ^[186], and improving optical ^[187] or mechanical ^[188] properties of particles. All these are due to the ALD's unique growth mechanism and growth characteristics, i.e., the surface-controlled deposition with the atomic-level growth of high-quality films at low temperatures. Recently, the ALD technique has also been reported to integrate with the traditional fluidized-bed technology for massive production of 0D nanostructured materials ^[189]. This has laid a venue to expand ALD into more applications.

---

3.2. 1D NMs

There are many different 1D nanostructures, such as nanowires, nanotubes, nanorods, nanobelts, and so on. In 2002 Solanki et al. ^[190] reported an ALD process for 1D ZnSe/CdSe superlattice growth. The researchers first pretreated a silicon wafer covered with a 100 nm SiO₂ layer using a 0.1% solution of poly-L-lysine for 1 min, and then coated a layer of ~6 nm Au nanoparticles using gold-nanocluster solution. The coating of Au nanoparticles was examined for adhesion using atomic layer microscope (AFM). A ZnSe/CdSe composite was deposited on the substrate at the temperatures of 400 to 420 ℃, followed by a sequence of 10 ZnSe monolayers and 5 CdSe monolayers. The precursors for ZnSe and CdSe were Zn/Se and Cd/Se, respectively. As a result, a 5 nm thick nanowire was grown (see Figure 3a). High-resolution TEM (HRTEM) analysis confirmed the excellent crystallinity of ZnSe and CdSe (see Figure 3b). This composite nanowire had a photoluminescence effect to excite the 450 and 540 nm wavelength photons.

In addition to the synthesis of 1D nanowires, ALD has also enabled 1D core-shell nanotubular structures. In 2003, Min et al. ^[191] deposited a ~6 nm Ru thin film on carbon nanotubes (CNTs) using Ru(od)₃/n-butylacetate solution and O₂ as ALD precursors at 300 ℃. The resultant CNT@Ru core-shell composite was then oxidized at 500 ℃ with the removal of the inner CNTs, resulting in pure RuO₂ nanotubes. HRTEM confirmed that Ru films grew uniformly and conformally over the inner and outer surfaces of CNTs (see Figure 3c, d). Based on CNTs, there have been a large amount of 1D nanotubular structures reported in literature, such as ntrigen-doped CNTs (NCNTs)@ZrO₂ (see Figure 3e) ^[192], NCNTs@Li₄Ti₅O₁₂ ^[193], NCNTs@SnO₂ ^[194], TiO₂ nanotubes ^[195], single-walled CNTs@Cu₂S ^[196], single-walled CNTs@GaS ^[197], and CNT@Fe₂O₃ ^[198].

Besides CNTs, another popular ALD strategy for nanotubes is based on the use of anodic aluminum oxide (AAO) templates. Typically, AAO templates can be prepared by a two-step anodization procedure ^[199]. First, aluminum sheets were electropolished in a solution of perchloric acid and ethanol, and then anodized in an oxalic acid. Second, the unreacted Al was selectively etched away in a saturated HgCl₂ solution and immersed in a phosphoric acid to etch the barrier layer. Elam et al. in 2006 ^[199] deposited various ALD conformal coatings (including Al₂O₃, TiO₂, V₂O₅, ZnO, and Pd) into nanopores of porous AAO templates. SEM images presented the top view of 15 nm ALD Al₂O₃ coated AAO template (see Figure 3f), showing conformal Al₂O₃ coating grown along high aspect-ratio AAO nanotubular structure. Through further etching AAO templates away using a base solution (e.g., KOH and NaOH), the inner deposited ALD layers could be remained in nanotubular structures, e.g., SnO₂ ^[200] and TiO₂ ^[195] nanotubes. For instance, Meng et al. ^[200] demonstrated a synthesis strategy towards highly structure-tunable SnO₂ nanotube arrays of high aspect ratio (up to 300:1) using ALD. The fabrication procedures of the patterned 1D SnO₂ nanotubes are as follows: (ⅰ) ALD SnO₂ deposition on the inner surfaces of nanopores of AAO templates using SnCl₄ and H₂O, resulting in the formation of tubular SnO₂ nanostructures (see A and B in Figure 3g), (ⅱ) the removal of the top covering of AAO by mechanically polishing to expose an array of SnO₂ nanotubes (see C in Figure 3g), and (ⅲ) a wet-etching of NaOH to dissolve AAO and then harvest an array of separated SnO₂ nanotubes (see D in Figure 3g). Not limited to above cases, there have many more studies using ALD for 1D NMs ^{[201,202,203,204,205,206,207,208]}. All these 1D NMs have taken advantages of ALD's surface-controlled film growth.

---

3.3. 2D NMs

ALD was first performed for 2D nanofilms, and thus 2D nanofilms are the most common format of different nanostructures. Among the 2D nanofilms, ALD particularly enables nanolaminates consisting of multi-layered materials. Each layer in nanolaminates can be precisely controlled with atomic-scale thickness. This is very convenient for developing new functional materials with tunable composition and properties. In this regard, KuKli in 1997 ^[209] reported 2D Al₂O₃/Ta₂O₅ and ZrO₂/Ta₂O₅ nanolaminates (see Figure 4a), achieving low-leakage and high-permittivity dielectric properties. Elam et al. ^[210] later fabricated ZnO/Al₂O₃ nanolaminates in 2002 using TMA/H₂O for amorphous Al₂O₃ and DEZ/H₂O for crystalline ZnO. It was found that the Al₂O₃ layer could suppress the crystal growth of ZnO films and thus the surface roughness of the ZnO/Al₂O₃ nanolaminates could be tuned.

Subsequently, there have been many more studies invested for various nanolaminates, such as HfO₂/Ta₂O₅ ^[211], Al₂O₃/TiO₂ ^[211], AlP/GaP ^[212], HfO₂/SiO₂ ^[213], and Y₂O₃/ZrO₂ ^[214]. These nanolaminates feature their tunable physical properties ^{[211,212,214,215,216]}. For instance, Zaitsu et al. employed ALD Al₂O₃/TiO₂ nanolaminates for tunable optical refractive index over a wide range from 1.870 to 2.318 ^[215]. The researchers also interpreted that film thickness of the alternating Al₂O₃/TiO₂ layers were related to both the number of each sublayer cycle and the certain growth rate of each material (see Figure 4b). Therefore, the volume concentration of the Al₂O₃ coating could be manipulated by adjusting the cycling number of the Al₂O₃ sublayer in the binary ALD system. Ginestra et al. in 2007 ^[214] reported a solid-state electrolyte membrane of Y₂O₃/ZrO₂. The as-deposited nanolaminates were post-annealed for crystallization at 950 ℃ for 2 hours, allowing inter-diffusion between nanolaminate films. A comparison between the as-deposited and annealed samples in the cross-sectional HRTEM images (see Figure 4c, d) showed the inter-diffusion. The annealed films, consisting of the diffused crystal ZrO₂ and Y₂O₃ composite, were investigated for their electrical conductivity. They found that the nanolaminate having 2 mol% Y₂O₃ enabled a conductivity 15 times higher than that of the nanolaminate having 3 mol%. This clearly showed that ALD is a very precise tool for optimizing materials' properties.

---

3.4. 3D NMs

Owing to its conformal coating capability, ALD can also be used for developing 3D novel nanostructures based on complex substrates. Using SiO₂ inverse opal photonic crystals as substrates, for instance, Graugnard et al. in 2006 ^[217] fabricated 3D TiO₂ inverse-shell opals. The fabrication process can be divided into following procedures: (ⅰ) preparation of opal structure using self-assembled and sintered silica nanospheres of ~450 nm (see A in Figure 5a), (ⅱ) infiltration of a template with ZnS or Al₂O₃ sacrificial layer (see B in Figure 5a) and TiO₂ dielectric material (see C in Figure 5a) in silica opal structure via ALD, (ⅲ) inversion of the opal structure by removing both the original template (see D in Figure 5a) and the sacrificial layer (see E in Figure 5a), and (ⅳ) reinfiltration (or backfilling) of the new template with additional dielectric material (see F in Figure 5a). SEM images exhibited the cross section of a doubly infiltrated silica opal template by 15 nm ZnS and 20 nm TiO₂ layers (see in Figure 5b), a thin backbone of TiO₂ after the removal of the silica nanospheres and sacrificial ZnS layer (see in Figure 5c), and backfilling with an additional 20 nm TiO₂ layer on inverse opal structure (see in Figure 5d). The 3D TiO₂ opal structure greatly enhanced photonic properties with a 58% increase of reflective index. Similarily, King et al. ^[218] demonstrated 3D TiO₂/ZnS:Mn/TiO₂ inverse opal structures via ALD. The resultant 3D structure enabled the combination of the strong photoluminescent properties of ZnS:Mn with the full refractive index of TiO₂. These papers successfully demonstrated the unrivaled conformality and atomic-scale growth from the vapor-phase ALD technique.

Another representative 3D nanostructure, gyroid-structured 3D ZnO networks via ALD, was demonstrated by Kim et al. in 2014 ^[219]. They first replicated periodic 3D gyroid networks generated by the self-assembly of polystyrene (PS) in solution-based method. Then, conformal ZnO coating was deposited via ALD on the 3D networks at 80 ℃ using DEZ and H₂O as precursors. The post-process of O₂-plasma treatment at 400 ℃ was conducted for 10 min to remove the PS template, followed by the annealing process at 450 ℃ to transform ZnO to higher crystallinity. Compared with as-synthesized PS networks, the annealed ZnO interconnected skeletons remained 3D gyroid networks with an average pore width of 30 nm (see Figure 5e). Kim et al. ^[219] revealed that the conformal 3D ZnO networks resulted from the ALD principle of surface-limited reactions, promising for photovoltaic devices.

Moreover, 3D graphene nanosheets (GNS) modified by ALD thin films have been successfully demonstrated, such as SnO₂-GNS ^[220], TiO₂-GNS ^[128], and ZrO₂-GNS ^[221] composites. Liu et al. in 2013 ^[221] deposited ZrO₂ films on wrinkled GNS via ALD using Zr(NMe₂)₄ and H₂O. The pristine GNS with wrinkle thickness of ~4 nm (see Figure 5f) was derived from thermal reduction of graphite oxide (GO). SEM image showed the conformal and uniform ZrO₂ films with the wrinkle thickness of ~19 nm grown on GNS wrinkles at 250 ℃ (see Figure 5g). This work demonstrated the excellent conformity and uniformity of ALD thin film over the wrinkled graphene sheets with extremely high surface ratio.

---

4. Nanostructured materials by MLD and both MLD/ALD

---

4.1. 0D NMs

MLD exhibits excellent capabilities of conformally and uniformly coating various 0D substrates to prepare novel 0D nanostructures, owing to its surface saturation growth and precise control at the molecular level. Liang et al in 2009 ^[222] deposited alucone films on silica nanoparticles (250 nm) at 100 ℃ using TMA and EG as precursors in a fluidized bed reactor. Then, the as-prepared samples were soaked in water or oxidized in air to become the porous Al₂O₃ films (see Figure 6a). Yang et al. in 2013 ^[223] deposited organic polyimide films on Au nanoparticles using ethylenediamine (EDA) and 1, 2, 4, 5-benzenetetracarboxylic anhydride (PMDA) as precursors, followed by a calcination at 600 ℃ to receive ~2 nm carbon shell on Au (see Figure 6b). These works indicated the capability of MLD preparing inorganic–organic and organic films over nanoparticles, and these films will lead to carbon coating after thermal treatment. Subsequently, MLD has been employed to deposit various ultrathin films (i.e., alucone ^{[222,224,225,226]}, titanicone ^[227], and polyimide ^[223,228]) over nanoparticle substrates (i.e., silica ^{[222,225,226]}, SiO₂ ^[224], TiO₂ ^[227], titania ^[226], Au ^[223] and Zr ^[228]) for 0D core-shell nanostructures.

The combined ALD/MLD method was also confirmed in preparing 0D nanomaterials. Dameron et al. in 2008 ^[147] deposited a double layer of Al₂O₃-aluone film on ZrO₂ nanoparticles of < 30 nm and BaTiO₃ nanoparticles of < 200 nm (see Figure 6c). Al₂O₃ was first deposited by ALD using precursors of TMA and H₂O. Subsequently, alucone was deposited by MLD in the range of 85–175 ℃ using TMA and EG as precursors. DuMont et al. ^[229] deposited alucone films on ZrO₂ nanoparticles. In this study, three procedures were conducted: (ⅰ) depositing ALD Al₂O₃ on the nanoparticles, (ⅱ) depositing alucone films (AlEG or AlHQ) at 150 ℃, and (ⅲ) post-treating the samples at 400–750 ℃ to transform alucone films to Al₂O₃ and carbon composites. TEM image demonstrated that the pyrolysis of AlEG (see Figure 6d) and AlHQ (see Figure 6e) films led to very conformal Al₂O₃/carbon composite films.

---

4.2. 1D NMs

MLD was employed to deposit organic films (i.e., zincone ^[149], alucone ^{[230,231,232,233]}, polyurea ^[234] and polyimide ^[223]) on different 1D NMs (i.e, Al₂O₃ microtubes ^[149], CNTs ^[223,234], polymer fibers ^[232,233] and nanowires ^[230,231]) for developing novel 1D nanostructures. For examples, Qing et al. in 2009 ^[149] deposited zincone films on Al₂O₃ microtubes by MLD using DEZ and EG as precursors. The microtubes were initially prepared by depositing 500 ALD cycles of Al₂O₃ onto polyvinyl alcohol fibers (PVA), followed by thermal treatment at 400 ℃ to remove PVA. TEM exhibited the comparison of original Al₂O₃ microtubes and zincone films of ~20 nm on the inside and outside of microtubes (see Figure 7a). Subsequently, Yang et al. ^[223] synthesized carbonaceous nanotubes using MLD, following by a series of procedures: (ⅰ) depositing polyimide films on AAO templates by MLD using EDA and PMDA as precursors, (ⅱ) thermally treating at 600 ℃ to decompose the polyimide films under protecting H₂/Ar gas into carbonaceous films inside of AAO tubes, (ⅲ) etching AAO template in NaOH solution to obtain carbonaceous nanotubes. Chen et al. in 2015 ^[234] employed MLD to deposit 11.6 nm polyurea (PU) on MWCNTs using precursors of 1, 4-phenylene diisocyanate (PDIC) and ethylenediamine (ED).

In addition, Gong et al. in 2011 ^[233] deposited alucone films on electro-spun polyvinyl alcohol fiber mats using TMA and GLY as precursors, followed by a thermal treatment to produce hollow Al₂O₃/carbon composite microtubes. This work demonstrated that MLD enabled alucone films with precise control over film thickness, favorable for artificial solid electrolyte interphase (SEI) in lithium-ion batteries.

Nanowires such as CuO ^[231] and silicon nanowires ^[230] were used for fabricating novel 1D nanostructures. Qin et al. in 2011 ^[231] deposited alucone (AlEG) films of ~37 nm on CuO nanowires by MLD using TMA and ethylene glycol (EG) as precursors (see Figure 7b). The resultant alucone/CuO nanowires were transformed to Cu nanoparticles embedded in porous Al₂O₃ nanotubes in H₂ gas at 650 ℃ (see Figure 7c). To clarify, CuO nanowires were chemically reduced into Cu nanoparticles. In addition, the ethyl groups in alucone films thermally decomposed in H₂ gas, resulting in porous Al₂O₃ nanotubes. Following by a calcination at 400 ℃, Cu nanoparticles diffused through porous Al₂O₃ nanotubes to its exterior surface, initially oxidized into Cu₂O and further to CuO (see Figure 7d). Subsequently, Luo et al. in 2015 ^[230] deposited MLD alucone films of ~5.6 nm on silicon nanowires (SiNWs) (see Figure 7e), in comparison with ALD Al₂O₃ films of ~5.3 nm on SiNWs (see Figure 7f). The alucone coating resulted in a "V-shaped" lithiation shape of the SiNWs (see Figure 7g), resulted from the high Li permeability of the alucone layer. In contrast, the Al₂O₃ coated SiNWs exhibited an "H-shaped" lithiation shape (see Figure 7h) because of the low intrinsic Li⁺ conductivity of the Al₂O₃ layer. The researchers demonstrated that MLD was capable of structurally and chemically modifying the surface of SiNWs to improve the performance of Si nanostructured anodes.

---

4.3. 2D NMs

MLD also was used to fabricate 2D nanostructures. For example, Abdulagatov et al. in 2013 ^[227] deposited two kinds of MLD titanicone films (TiGL and TiEG) on Si wafers. The resultant samples were pyrolyzed under argon to transform titanicone into TiO₂/carbon composite (see Figure 8a). Lushington et al. in 2015 ^[235] used the alternated exposures of TMA, EG, and terephthaloyl chloride (TC) by MLD to grow the aluminum alkoxide films. Prior to the MLD process, a 10 nm ALD Al₂O₃ seed layer was deposited on the polished Si (100) substrate and SiO₂ glass substrate. These hybrid films were then thermally treated at 600–900 ℃ in H₂/Ar to transform the alucone films into Al₂O₃/graphitic–carbon composites. The researchers also addressed that MLD was an effective method to tune the composition of the hybrid films by changing the amount of carbon.

In addition to MLD method, the combined ALD/MLD was also reported in preparing novel 2D nanomaterials ^[236,237]. Lee et al. ^[237] applied combined ALD/MLD to prepare ZrO₂-based organic–inorganic nanohybrid thin films for application in organic thin film transistors. First, ZrO₂ was deposited on oxidized Si (100) substrates by ALD at 170 ℃ using Zr(OC(CH₃)₃)₄ (ZTB) and H₂O as precursors. Then, alkene-terminated self-assembled organic layers (SAOLs) were prepared by MLD using a repetition of three precursors of CH₂CH(CH₂)₆SiCl₃ (7-OTS), ozone, and ZTB. Subsequently, Nisula et al. ^[236] prepared a thin-film Li₂Q–Li₂TP cell (Pt/Li₂TP/LiPON/Li₂Q/Cu) via ALD/MLD. First, they deposited 500 ALD cycles (~23 nm) of Pt as current collector on 300 mm Si wafers using (methylcyclopentadienyl)trimethylplatinum and oxygen as precursors. Then, the Li₂Q electrode layer (see Figure 8b) was deposited on Pt at 160 ℃ by MLD using lithium bis-(trimethylsilyl)amide (LiHMDS) and hydroquinone (HQ) as precursors. On the Li₂Q layer, the LiPON layer of ~30 nm as solid-state electrolyte was deposited by MLD at 300 ℃ using precursors of LiHMDS and diethyl phosphoramidate (DEPA). Finally, the Li₂TP electrode layer was deposited by 70 MLD cycles at 200 ℃ using LiHMDS and terephthalic acid (benzene-1, 4-dicarboxylic acid or TPA) as precursors. At 5 mA/cm², the Li₂Q–Li₂TP cell using 15 nm Li₂Q layer could reach 50% of the full capacity within 0.25 s. In addition, this combined ALD/MLD method was utilized to prepare the thin-film Li₂Q/LiPON/Cu cell (see Figure 8c). This study indicated that ALD/MLD thin film techniques enabled the development of advanced thin film electrodes for solid state lithium ion batteries, due to its precise control over thickness and film quality.

Moreover, many studies were devoted to fabricating 2D nanolaminates by combined ALD/MLD method, such as TiO_x-self-assembled multilayers (SAMs)/TTIP-H₂O ^[238], AlO_x-SAMs/Al₂O₃ ^[239], AlO_x-SAMs/ZnO ^[239], AlO_x-SAMs/Al₂O₃/AlO_x-SAMs/ZnO ^[239], ZnO/DEZ-HQ ^[240], ZnO/DEZ-AP ^[240], ZnO/DEZ-ODA ^[240], Al₂O₃/AlEG alloy ^[241,242], 1, 4-phenylene diisocyanate (PDIC)-ethylenediamine (ED)/PDIC-2, 2'-thiobis(ethylamine) (TBEA)/PDIC-ED ^[243], self-assembled organic layers (SAOLs) and Al₂O₃ ^[244], titanium oxide cross-linked polydiacetylene (TiOPDA)-TiO₂ ^[245], hafnicone and HfO₂/Hafnicone ^[161]. For example, Lee et al. in 2007 ^[238] first prepared TiO_x-SAMs/TiO₂ nanolaminate films. ALD TiO₂-coated Si(100) substrate was alternatively exposed in 7-octenyltrichlorosilane (7-OTS) and H₂O to produce an alkene-terminated self-assembled monolayer (SAM). Then, O₃ was introduced to transfer the terminal C = C groups in SAM into carboxylic groups. Finally, titanium hydroxyl groups were generated by sequentially dosing titanium isopropoxide (TTIP) and H₂O. Lee et al. ^[241] fabricated the alucone alloys using different ratios of Al₂O₃:alucone (1:1 and 3:1). Based on different alucone contents, hardness and elastic modulus of the alucone alloys changed from 1.0 ± 0.1 GPa to 13.0 ± 0.2 GPa and 21 ± 8 to 198 ± 8 GPs, respectively. Therefore, mechanical properties of these alloy films such as hardness and elastic modulus can be finely tuned by adjusting the relative number of ALD cycle and MLD cycle in the reaction sequence.

---

4.4. 3D NMs

MLD can be used to prepare organic–inorganic hybrid material on 3D substrate using precursors of organometal and organic alcohol. For example, Hall et al. ^[157] employed MLD to deposit zircone on nanoporous alumina at 150 ℃ using ZTB and EG as precursors. The nanoporous alumina was prepared by a two-step anodization in 0.3 M sulfuric acid solution (0 ℃, 25 V), where an alumina etching process was conducted to separate the first (12 h) and the second anodization step (20 h) (see Figure 9a). SEM showed the cross section of the zircone-coated nanoporous alumina with obliterated pores at the front side (see Figure 9b) and open pores at the back side (see Figure 9c).

Another 3D nanostructure is metal–organic framework (MOF). MOF is composed of inorganic metal-based clusters and organic linkers to generate highly porous crystalline structures ^[246,247]. MOF displays high thermal stability ^[248], high surface area ^[248], and unrivalled versatility in pore composition and geometry ^[249]. The preparation of MOF using atomic and molecular layer deposition has been spurred ^{[246,247,250,251]}, due to the precise control over the thickness at molecular level. For instance, Salmi et al. in 2013 ^[252] first deposited MOF-5 at 225–350 ℃ using zinc acetate (ZnAC₂) and 1, 4-benzenedicarboxylic acid (BDC) as precursors. The as-deposited films were initially amorphous and solid. However, following by exposure in moist air at room temperature and then solvothermal treatment in dimethylformamide (DMF) at 150 ℃, the crystalline and porous MOF-5 phase was obtained. This paper indicated that post-treatment was an effective method for the crystallization of porous MOF. In another work, Ahvenniemi et al. ^[246] fabricated a highly crystalline copper(Ⅱ)terephthalate (Cu-TPA) MOF on p-type Si (100) substrates at 180–280 ℃ by MLD method using Cu(thd)₂ (copper 2, 2, 6, 6-tetramethyl-3, 5-heptanedione) and terephthalic acid (TPA) as precursors. The researchers emphasized that the conformity of MOF films via MLD on different substrates or architectures was owed to its surface saturation growth.

---

5. Promising applications

---

5.1. Surface engineering

ALD/MLD thin film techniques hold a great promise for the surface modification, allowing for new strategies in the modification of chemical and physical properties of nanoscale materials and the synthesis routes to novel nanostructures, due to its capability of precisely controllable thickness and conformal coverage. Different nanostructures (i.e., nanopores ^[253,254], nanowires ^[116,190], nanotubes ^[191,255], nanopatterns ^{[256,257,258]}, and nanolaminates ^[259]) have been modified via ALD thin films for various applications (i.e., optical devices ^{[184,215,218,260,261]}, wettability ^[262,263], and mechanical applications ^[259]).

Highly conformal and uniform ALD coatings can contribute significantly to the photonic band gap properties for attaining high performance optical devices. Park et al. ^[184] investigated the well-ordered TiO₂ nanobowls via ALD for optical properties. This unique nanostructure with 58 nm shell thickness assisted YAG:Ce CPP-capped LED (see Figure 10a) to achieve the maximum luminous efficacy of 119.0 lm/W at 350 mA with an optimized nanobowl structure (see Figure 10b), which is 54% higher than that of a conventional CPP-capped LED (77.2 lm/W). Zaitsu et al. ^[215] demonstrated the controllable refractive index of ALD Al₂O₃ and TiO₂ 2D nanolaminates by tuning the TiO₂ sublayer thickness in the stack. The refractive index of 1.870 to 2.318 was found as a linear proportion to the TiO₂ sublayer thickness varied from 2.0 to 39 Å while Al₂O₃ sublayer remained 5.5 Å thickness. They pointed out that this ALD 2D nanolaminate would be applicable to advanced spectral performance in optical filters with extremely thin layers. Furthermore, King et al. ^[218] developed ALD 3D TiO₂/ZnS:Mn/TiO₂ inverse opal structures through a combination of inverse opal photonic crystals and atomic layer deposition (ALD). The novel 3D nanostructure of highly crystalline ALD thin films attained full refractive index and strong photoluminescence. These above studies have provided the strong confirmation of the enhanced optical properties in ALD nanostructures.

Wetting property describes the ability of a liquid to sustain on a solid surface, resulting from surface chemistry, roughness and geometry. ALD thin film technique can modify the surface chemistry for the sake of tuning wetting property. Halbur ^[262] applied ALD Al₂O₃ coating on flexible polytetrafluoroethylene (PTFE) substrate to control the surface energy and wetting properties for the application of aqueous-based nanoparticle inks. PTFE substrates with 175 cycles of ALD Al₂O₃ coating exhibited uniform Ag nanoparticle distribution when ink dried, while the coffee-ring effect for nanoparticle distribution was found on the bare PTFE substrate. The researchers reported that the wettability of Al₂O₃ coated PTFE could be precisely controlled by tuning the Al₂O₃ layer thickness. Hydrophobicity with over 100° water contact angle (WCA) was observed for less than 100 ALD cycles, while a sudden change to superhydrophilicity of 0° WCA occurred right after 100 ALD cycles. This work addressed that the dramatical improvement of ALD Al₂O₃ coating on uniform distribution of nanoparticle ink casted or printed on PTFE substrate for flexible electronic applications. Dafinone et al. ^[263] demonstrated both superhydrophilicity (<5° water contact angle) of ALD Al₂O₃ coated on inorganic TiO₂/SiO₂ and organic polycarbonate nanoparticle films with a post-treatment of O₂ plasma. The subsequent oxygen plasma treatment essentially removed organic residues from precursors to render the surface superhydrophilicity.

The combined ALD/MLD also demonstrated the modified wetting properties. Gong et al. ^[264] investigated the combined effects of ALD Al₂O₃ and MLD alucone layers on wettability of the polydimethylsiloxane (PDMS) and Si wafer surface. They first found that 100 cycles of ALD Al₂O₃ layer formed a hydrophilic PDMS surface and hydrophilic Si wafer surface, but after storing for 48 h and 24 h in ambient air they lost the hydrophilicity. Whereas, 200 MLD cycles of alucone on Si wafer led to a durable hydrophilic surface with a low contact angle of 10–30° for more than 72 h (see Figure 10c). Then, the researchers combined 100 cycles of Al₂O₃ sublayer on PDMS and 100 cycles of alucone lay on the top, which exhibited the durable hydrophilic surface with a low contact angle of ~20° for 200 h in air (see Figure 10d). Therefore, the combination of sequential vapor infiltration with ALD or MLD on PDMS was effective in engineering and tailoring the chemical properties of surface.

In addition, ALD has been widely recognized as a method to mechanically reinforce nanoparticle thin films on inorganic and organic substrates, due to low-temperature deposition and precisely controllable layer thickness. Dafinone et al. ^[263] investigated the mechanical enhancement of ALD Al₂O₃ coated 0D TiO₂/SiO₂ nanoparticle films. The dramatically increasing elastic modulus and hardness of Al₂O₃-coated TiO₂/SiO₂ films on Si wafers were confirmed with increasing the number of ALD cycle. For example, the use of 15 ALD Al₂O₃ cycles enhanced 4 times for the elastic modulus and 8 times for the hardness of TiO₂/SiO₂ films (see Figure 10e). They also addressed that the ability of low-temperature ALD would be essential to reinforce mechanical properties on organic materials, like polymers and plastics. In addition, Doll et al. ^[265] studied mechanical properties of 140 nm 2D ZnO/ZrO₂ nanolaminate coating on silicon substrates fabricated via ALD. The nanolaminate contained 8 bilayers of 10 nm ZnO and 7nm ZrO₂ sublayers. As-grown (0002) ZnO crystalline textures improved good tribological properties, and the crystalline ZrO₂ sublayers relatively high toughness acted as a load bearing layer to resist high load and as a pathway to dissipate energy to suppress delamination between the bilayers. Romanes ^[259] investigated the ALD films metal oxides, 2D ZnO/Al₂O₃ nanolaminates and ZrO₂ thin films, as the solid lubricants at room temperatures. They revealed that a significant enhancement in friction and wear performance was obtained for ALD-grown thin films. Low friction and wear are significantly demanded for precise devices of high aspect-ratio, like microelectromechanical (MEMS) devices.

Several studies have devoted to enhanced mechanical properties by MLD or combined ALD/MLD prepared films, such as Al₂O₃/alucone/Al₂O₃ ^[266], alucone film (TMA-EA-MA) ^[267], polyurea ^[234], ALGL ^[155], ZnGL ^[155] and TiGL ^[155]. Miller et al. ^[266] studied the mechanical properties of Al₂O₃/AlEG/Al₂O₃ nanolaminates prepared at 155 ℃ on polyethylene naphthalate substrates. The fabricated nanolaminates exhibited different ratios of each layer thickness of 10/3/10 nm, 25/15/25 nm (see Figure 10f), 25/3/25 nm and 25/192/25 nm respectively. Compared with the pure components, the nanolaminates delivered the reduced critical strains at fracture because of the low toughness of the TMA + EG alucone films. Miller et al. ^[267] also investigated the thermomechanical properties of "AB" (TMA-EG) and "ABC" (TMA-EA-MA) alucone films, respectively. The researchers found that the TMA-EG alucone exhibited the elastic modulus of 36.8 ± 5.6 GPa and the Berkovich hardness of 0.47 ± 0.07 GPa, both of which are higher than that of the TMA-EA-MA alucone films (modulus: 13.2 ± 5.0 GPa and hardness: 0.27 ± 0.10 GPa). Besides, AB alucone exhibited thermal expansion coefficients of 12.6 ppm/℃ during heating and 7.1 ppm/℃ during cooling. Therefore, MLD enabled the tailoring of the thermomechanical properties of the coatings. In addition, MLD coatings have enhanced the mechanical properties of 1D MWCNTs. Chen et al. ^[234] improved chemical compatibility between MWCNT and the polyurethane matrix using MLD polyurea coatings, leading to the enhanced mechanical tensile strength of 50% and elastic modulus of 152%. This can be explained by the following two reasons: (ⅰ) better dispersion and interfacial adhesion modified by MLD PU coatings and (ⅱ) urea couplings between MLD polyurea layers and polyurethane polymer matrix. Moreover, Brown et al. ^[155] discovered that the MLD coatings of alucones (precursors: TMA/GL and TMA/EG), titanicone (precursors: TiCl₄/GL), and zincone (precursors: DEZ/GL) could significantly improve the Young's Modulus of MWCNTs. Compared with the pristine CNTs, the MLD coatings exhibited a dramatical decrease in failure strain, a modest improvement in ultimate tensile strength, and a huge increase in Young's Modulus by a factor ranging from 4.3 (for AlGL) to 13.5 (for AlEG).

---

5.2. Catalysis

Catalysis is an important approach to selectively increase the chemical reactivity using catalysts, typically noble metals. ALD has gained many interests as a promising method for synthesis of catalyst materials. Due to its self‐limiting surface reaction, ALD is capable of designing and modifying catalysts at the nanoscale via precisely control composition. Sun et al. ^[268] first reported the single-atom catalysis of isolated Pt atoms anchored to graphene nanosheet via ALD. Pt single atoms and sub-nanometer clusters (see Figure 11a) were ALD-deposited uniformly on graphene at 450 ℃ using MeCpPtMe₃ (held at 65 ℃) and O₂ as precursors. The researchers demonstrated that ALD was advantageous on precisely controlling the morphology, size, density and loading of Pt on graphene by tuning the number of ALD cycles. The catalytic performance examined the higher methanol oxidation reaction (MOR) activity and superior CO tolerance, in comparison with conventional larger-scale Pt/C catalysts (see Figure 11b). Sun et al. ^[268] interpreted that the excellent catalytic performance was attributed to the single Pt atoms. In addition, Christensen et al. ^[269] fabricated Ru–Pt bimetallic nanoparticle catalysts via ALD for fuel cell applications. Ru–Pt nanoparticles of ~1.2 nm diameter was deposited on Al₂O₃ nanospheres by two ALD steps of Ru(DER) (kept at 80 ℃)/O₂ and Pt(MeCp)Me₃ (kept at 50 ℃)/O₂. MOR measurement revealed that an over 90% conversion efficiency was obtained for ALD Ru–Pt nanostructured catalysts at the temperature of 270 ℃, which was almost double that of physical mixture of monometallic Ru and Pt catalysts. The authors indicated that the steep temperature dependence implied high activation energy for the bimetallic catalysts and also supported bimetallic nanoparticles yielded by ALD synthesis method.

For MLD-synthesized catalysts, Ishchuk et al. ^[270] deposited TiEG films on planar SiO₂ membrane (40 nm thick) by MLD using TiCl₄ and EG as precursors at 100–120 ℃. The enhanced photocatalytic activity of annealed TiEG films at 650 ℃ was enabled by the co-existing amorphous-crystalline state. To clarify, amorphous state provided accessible adsorption sites, and crystalline state were photocatalytic active. This work indicated that MLD was a unique method to prepare nanometric hybrid materials for high photocatalytic activity. Similarly, Sarkar et al. ^[271] employed MLD to prepare nonstoichiometric oxides for tailoring the reactivity of metal oxide photocatalysis. Reactivity of TiEG films annealed at 650 ℃ for photodegradation of terephthalic acid (TPA) was highly dependent on the deep valence band. Different from the two-step MLD process in the above discussions, Chen et al. ^[160] employed a four-step MLD process of TiCl₄-ethanolamine (EA)-malonyl chloride (MC)-EA at 100 ℃ on carbon nanocoils (CNCs), followed by a thermal treatment at 600 ℃ in Ar/H₂ to receive porous N-doped TiO₂/amorphous-carbon composite on CNCs. Through annealing at 350 ℃ in air to remove the amorphous carbon and increase the surface area, the porous N-doped TiO₂@CNCs treated at 350 ℃ exhibited the best catalytic efficiency for photodegradation of methylene blue (MB) in air (see Figure 11c) under visible light irradiation (λ ≥ 420 nm).

In addition, catalysts suffered from the irreversible deactivation in catalytic processes, induced by various mechanisms, such as poisoning, coking and fouling, sintering, and attrition ^[272]. ALD has shown its ability to inhibit nanoparticle mobility of metal catalysts by the encapsulation from a conformal ALD coating, which stabilizes catalysts for an extended lifetime. O'Neill et al. ^[273] demonstrated a stabilization approach for copper catalysts via ALD. ALD Al₂O₃ coating wereas deposited onto oxidized copper nanoparticles using the precursors of TMA and H₂O. Then, a heat treatment for as-synthesized nanocomposite was conducted at 973 K to yield porosity in the encapsulation, in order to provide access to the underlying copper particles. The stabilized catalytic behavior indicated that sintering and leaching issues located at edges, corners, and defects of copper atoms were effectively suppressed by the separation from Al₂O₃-coated Cu nanoparticles. Similarly, Feng et al. ^[274] reported the enhanced stability and reactivity of Pd nanoparticle catalysts via ALD Al₂O₃ coating using TMA and H₂O. MOR results indicated that nearly 100% methanol conversion efficiency at 270 ℃ catalytic temperature for 6 h had been obtained for the 1–16 cycles of Al₂O₃ coating on Pd nanoparticle (see Figure 11d). Remarkably, Pd nanoparticle with only 1 cycle of Al₂O₃ coating maintained the nearly same size of fresh Pd nanoparticles after 6 h MOR catalytic process, observed through STEM (see Figure 11e). Feng et al. ^[274] addressed that a single ALD Al₂O₃ cycle (~0.1 nm) was sufficient to suppress sintering issues of Pd nanoparticles during catalytic process. Overall, these useful strategies for catalysis stabilization have been demonstrated to substantially improve the conversion efficiency and reduce the high cost of the use of noble metals in fuel cells and other catalysis applications.

MLD has also shown an ability for the sake of catalysis stabilization. In this regard, Liang et al. ^[275] have protected Pt nanoparticles using a thermally stable, highly dispersed and ultrathin porous alucone shell (1–4 nm). Prior to the deposition of TMA-EG alucone films, Pt nanoparticles were deposited on mesoporous silica particles at 320 ℃ by ALD in a fluidized-bed reactor (FBR) using methylcyclopentadienyl-(trimethyl) platinum(Ⅳ) (MeCpPtMe₃) and O₂ as precursors. After holding the catalysts at 400,600 and 800 ℃ for 4 h in air, the porous alucone layer significantly reduced the sintering rate of Pt nanoparticles, especially for thicker alucone films. The 20 MLD cycles coated catalyst annealed at 800 ℃ exhibited less catalytical activity of oxidizing CO than that annealed at 600 ℃ (see Figure 11f), due to the amorphous-to-γ alumina transition. Subsequently, Gould et al. ^[276] stabilized Ni catalyst using a porous alumina layer, which was prepared by calcinating the TMA-EA-methane (MA) alucone films in O₂ gas. Dry reforming activity of methane and stability at 700 ℃ of the catalysts were evaluated. The 5 MLD-layer modified catalyst reached steady state faster and higher than the 10 MLD-layer modified catalyst (see Figure 11g), resulted from the increased Ni surface area by the alumina pores expansion and NiO reduction under reaction conditions. The 10 MLD-layer modified catalyst was stable even after repeated calcinations and reductions (108 h). Moreover, Zhang et al. ^[277] employed a four-step ABCB MLD method (DEZ/1, 4-Phenylene diisocyanate (PPDI)/ethanediamine (EDA)/PPDI) to deposit Zn-hybrid films on Cu@CNTs composites. Followed by a calcination at 300 ℃ in air and a reduction at 300 ℃ in H₂/N₂ gas, the resultant Cu–ZnO@CNTs composites obtained from 120 MLD cycles delivered the highest efficiency and stability in hydrogeneration catalysis (see Figure 11h), resulted from the Cu–ZnO interaction and the suppressing of Cu NPs sintering.

---

5.3. Batteries

Rechargeable batteries as electrical energy storage (EES) devices play important roles in mitigating the stress of sharply increasing energy demands, deficient energy resources ^{[278,279,280,281]}, environmental pollution ^{[282,283,284]} and global warming issue ^{[285,286,287,288]}. LIBs as the promising EES devices have been widely applied in portable electronics and electric vehicles, owing to their high energy density. Nevertheless, conventional LIBs suffered from the stringent requirements of higher capacity, higher power, and better cyclability. ALD/MLD nanostructures have been targeted as a solution to improve batteries performance since the past few years, due to their conformal and ultrathin films (1–100 nm) in complex structures with precise growth control ^[117]. The importance of ALD ^{[20,166,167,168,169,170,171,172]} and MLD ^{[142,143,146,243,289,290,291,292,293,294]} have been demonstrated in preparing novel electrodes and solid-state electrolytes, protecting the electrode surface, and modifying the separator for LIBs. Many studies related to ALD and MLD have also been devoted to lithium–sulfur batteries ^{[295,296,297]}, sodium-ion batteries ^[114], and lithium air batteries ^[298].

Protection layers are often used at the interface between electrode materials and electrolyte to enhance the battery performance ^[117]. Jung et al. ^[299] demonstrated the surface modification of ALD Al₂O₃ coating on LiCoO₂ cathode to enhance the capacity retention of over 80% after 120 charge/discharge cycles, in comparison with 45% of that on bare electrode (see Figure 12a). The significant improvement was attributed to the protection of LiCoO₂ from Co dissolution or the mitigation of electrode–electrolyte reactions using an ultrathin Al₂O₃ coating of 4~5 Å. This work demonstrated the great promise of ALD thin films used to protect the interface between electrode and electrolyte and further improve the electrochemical performance of LIBs. Similarly, Jung et al. ^[300] deposited ALD Al₂O₃ also on Li[Li_0.20Mn_0.54Ni_0.13Co_0.13]O₂ cathode, followed by a heat-treatment at 300 ℃ in air. The heat treatment induced the inter-diffusion of atoms between the Al₂O₃ coating and the Li[Li_0.20Mn_0.54Ni_0.13Co_0.13]O₂ core, leading to better Li⁺ ion conductivity through the coating. The resultant cathode exhibited a good cyclability of 50 cycles with high capacity retention. Afterwards, there have been a sharp increase in number of ALD protection coating (i.e., ZrO₂ ^[301], TiO₂ ^[302], and SnO₂ ^[303]) on both cathodes (i.e., LiMnO₄ ^[301] and LiNi_0.5Mn_1.5O₄ ^[304]) and anodes (i.e., graphene ^[304], Si ^[305], and V₂O₅ ^[306]) of LIBs. For LIB anodes, ALD coating also have the capability of providing accommodation to the huge volume change (i.e., 400% in Si anode ^[307]) and reducing the formation of solid electrolyte interface ^[304].

Surface modification via MLD is also effective to improve the performance of LIB and Li–S batteries. Piper et al. ^[308] deposited 5 nm AlGL films on 0D Si nanoparticles at 140 ℃. This conformal and flexible inorganic–organic hybrid thin film was believed to infiltrate the porous structure of the electrode and covalently bind to surfaces. AlGL coated Si anode exhibited better cycling behavior than the bare Si (see Figure 12b), due to the fact that AlGL films had the passive properties and excellent stability to help mitigate the secondary reactions between the electrolyte and the Si electrode surface. Chen et al. ^[309] deposited 2D TMA-EG alucone film on lithium metal. This alucone film with high cross-linking structure can effectively mitigate the side interaction between lithium and electrolyte and improve the capacity and Coulombic efficiency of lithium batteries. Besides surface coating on Si nanoparticles and lithium metal, sulfur cathode coated by MLD films was also investigated recently. Li et al. ^[310] deposited 2D TMA-EG alucone films on carbon/sulfur (i.e., C/S) cathodes to improve the performance of the Li–S battery (see Figure 12c). With a 5-cycle alucone coating, the C/S cathode displayed flat potential plateaus, which demonstrated the high reversibility of the Li–S redox reaction.

Furthermore, the ability to ALD conformality in complex structures is extremely helpful for designing novel nanostructures for battery electrode. Liu et al. ^[311] applied ALD to synthesize iron phosphate (FePO₄) conformally coated on nitrogen-doped carbon nanotubes (NCNTs) by using two ALD steps of Fe₂O₃ (ferrocene/ozone as precursors) and PO_x (trimethyl phosphate/water as precursors) at 200~350 ℃. Even though low discharge capacity of 177 mAh/g in the first cycle, the as-synthesized FePO₄@NCNTs exhibited a discharge capacity of 141 mAh/g after 100 cycles and indicated a fairly good cyclability with high capacity retention. Lu et al. ^[312] reported ALD ZnO (DEZ/H₂O as precursors) on carbon black as a nanostructured anode in LIBs. The nanostructured ZnO anode suppressed large volume change (of 228%) during charge/discharge cycles and mitigated the rapid capacity fading, in comparison with bulk ZnO. A durable cyclability with a high reversible capacity of 1026 mAh/g was retained after 500 cycles. The nanocomposites also presented excellent rate capability with a reversible capacity of 1080 mAh/g at 2000 mA/g. The authors revealed that the strong interfacial molecular bindings in the ALD-synthesized nanocomposites provided stable and robust electrical contact during lithiation and delithiation processes. This approach could be extended to the ALD-synthesis of other metal oxide (i.e., TiO₂ ^[313] and SnO₂ ^[303]) based anodes for high power LIBs.

2D nanostructured materials via MLD were also applied in electrode design. Kerckhove et al. ^[162] deposited tetrakisethylmethylaminovanadium (TEMAV)/EG and TEMAV/GL vanadicone films. Then, the as-deposited films were annealed in He gas to receive amorphous V₂O₅ and carbon composites. Cyclic voltammograms results revealed that the as-resultant V₂O₅@carbon composites exhibited better electrochemical performance than the bulk V₂O₅, in terms of capacity, rate performance and cyclability. Kerckhove et al. ^[158] designed another 2D electrode for lithium ion batteries via MLD, using tetrakisdimethylaminotitanium (TDMAT)/GL and TDMAT/EG titanicone films. The as-deposited samples were annealed in air to become anatase TiO₂ phase, while He-annealed samples were amorphous TiO₂ phase. From the electrochemical experiments, anatase and amorphous TiO₂ samples fabricated via this MLD approach both exhibited higher rate capability, compared to their respective anatase and amorphous TiO₂ via ALD processes. Furthermore, Nisula et al. ^[314] prepared a novel lithium ion battery using high rate-capability anode LiTP with a solid-state electrolyte LiPON via the combined ALD/MLD. The as-prepared LiPON thin film as solid-state electrolyte could avoid instability/dissolution issues of liquid organic electrolytes and also improve the electrical conductivity. The novel LiTP–LiPON battery demonstrated a high capacity retention of 97% with a high rate of 3.2 C after 200 discharge/charge cycles (see Figure 12d).

---

6. Conclusions

The miniaturization in nanotechnology requests a facile and high-quality fabrication method for nanostructured materials. To this end, ALD and MLD techniques recently have emerged as the promising fabrication techniques for inorganic, organic, and inorganic–organic hybrid nanostructured materials, due to precise controllability and high conformality at low temperature atomic-scale or molecular-scale deposition. ALD and MLD are both vapor phase methods based on saturating and self-limiting surface reactions using the alternating exposures of different precursors. Our work introduced the necessities of nanostructured materials and the practical capabilities of ALD and MLD, as well as briefly introducing the mechanisms of ALD and MLD. Subsequently, the fabrication strategies via ALD and MLD individually were highlighted to novel nanostructured materials categorized in 0D, 1D, 2D, and 3D structures. Then, we exemplified the applications of ALD and MLD in surface engineering, catalysis, and batteries. With this work, we expect that more solutions will be brought to practical issues by the versatile capabilities of ALD and MLD for nanostructured materials. In this regard, new research efforts have emerged and hopefully will bring forth breakthroughs in many areas. For instance, ALD and MLD have been verified for their versatile capabilities in tailoring interfaces ^[299,308] and designing nanostructured electrodes for energy devices like rechargeable batteries ^[311,312]. Recently, there has been an increase number of solid-state electrolytes ^[236,314] reported by ALD and MLD for these energy systems for improved performance. In addition, the combination of ALD and MLD is becoming a new research thrust for surface engineering, featuring their powerful capabilities for new hybrid materials with desirable properties ^{[259,261,263]}. In catalysis, furthermore, ALD and MLD are becoming new techniques for developing high-efficient catalysts, such as single-atom catalysts ^[268]. Even more, ALD and MLD have been reported for synthesizing structurally 2D materials ^[315] and 3D MOF ^{[246,252,316]}. All these and others not discussed in this work together evidence that ALD and MLD are expanding quickly in their applications with new solutions.

---

Acknowledgements

Funding for this research was provided by the Center for Advanced Surface Engineering, under the National Science Foundation Grant No. OIA-1457888 and the Arkansas EPSCoR Program, ASSET Ⅲ.

---

Conflict of interest

All authors declare no conflicts of interest in this paper.

---