Metal Oxide Thin Films for Advanced Photonic Applications

Abstract

Oxide materials represent a versatile and fundamental platform for photonics, allowing the manipulation of light through optical property engineering. This review focuses mainly on the physics and applications of simple oxides, analysing their use in the realisation of dielectric mirrors, in particular of distributed Bragg reflectors, and planar microcavities. Critical aspects regarding the design of multilayer structures, the control of optical confinement and the improvement of the quality factor in passive devices are discussed. However, to provide a complete picture of the evolution of the field, the section dedicated to oxide materials anticipates future directions dominated by complex oxides such as lithium niobate, lithium tantalate and barium titanate required for active photonics. In this context, a necessary technological paradigm shift is highlighted: the transition from the current use of film-on-insulator platforms to the direct epitaxial growth of these functional materials, an essential step for the scalability and monolithic integration of future photonic devices.

1. Introduction

Despite Fraunhofer’s observations on the antireflective effect of a thin film deposited on glass in 1817 and Fresnel’s later complete description of optical interference, it was not until the 20th century that this knowledge was applied for technological purposes (i.e., optical coatings) due to the development of thin-film deposition techniques [1,2]. Optical coatings are systems composed of one or more thin films of material deposited on a substrate that modify their interaction with light, usually by controlling reflection, transmission, absorption and polarisation [3]. Thus, their importance in modern society is understood, becoming indispensable in fields such as energy [4,5,6], medicine [7], communications [8,9], research [10] and others [11]. Broadly speaking, these optical films can be classified according to the type of material used in metallic coatings, dielectric coatings, and hybrid coatings (combining both kinds of material), respectively [12]. Although metallic coatings have some disadvantages, including an intrinsic absorption that restricts maximum reflectivity (90–98%), a low threshold for laser-induced damage due to heating, degradation in air and moisture (e.g., metals such as aluminium and silver undergo degradation through oxidation or tarnishing), and the inability to tailor their performance in terms of spectral and wavelength selectivity, these coatings are predominantly used in broadband reflective optics [13]. Nevertheless, it is pertinent to mention that, although they are categorised as optical coatings, the mechanism by which light is reflected is not based on the interference phenomenon but rather on the interaction of the incident electromagnetic field with the free electrons in the metal.

However, a robust way to overcome these limitations is to use dielectric thin films. Known also as optical interference coatings [11], dielectric ones are better suited for applications where metallic counterparts fail. This advantage comes from their specific properties, such as low optical losses that result in a reflectivity of over 99.9%, angular and spectral engineering through alternating low- and high-refractive-index materials and adjusting the thickness of each thin film [14], a superior laser-induced damage threshold (LIDT) [15,16], and a superior mechanical and chemical stability [17].

The refractive index, transparency window, stability in the working environment, and compatibility with the deposition technology are the main key factors that determine which dielectric materials can be employed to obtain optical coatings. A detailed description of the dielectric materials (especially oxides) used for the manufacturing of reflective structures (i.e., dielectric mirrors) will be presented in the relevant subsequent section. Stoichiometric complex oxides will also be briefly reviewed, as they represent a class of materials that redefines technological boundaries, facilitating the transition from bulky optical components to high-performance photonic integrated circuits (PICs).

As previously stated, dielectric materials are a class of materials whose properties (low absorption, tuneable refractive index, etc.) make them ideal candidates for optical coatings with various functionalities, such as antireflective or high-reflective coatings, filters, and others [18]. Dielectric materials can be classified into inorganic (e.g., metal oxides, nitrites, fluorides, etc.) and organic types [19]. Among the inorganic dielectrics, materials are commonly categorised according to the dominant anion. These include: oxides (e.g., SiO₂, Al₂O₃, HfO₂, Ta₂O₅, TiO₂, etc.), which cover the UV–Vis–NIR spectral domains and exhibit a wide range of refractive indices; fluorides (e.g., MgF₂, CaF₂, etc.), employed as low-index materials [20,21,22]; nitrides (e.g., SiN, AlN), valued for their high refractive indices and barrier properties [11,23]; and sulfides/selenides (e.g., ZnS, ZnSe), preferred for their applicability in the infrared region [24,25,26]. In contrast, organic dielectrics, such as polymer-based materials, are advantageous when low-temperature processing or flexible substrates are required [18].

The performance of an optical interference thin film is influenced by both the choice of material and deposition technique, which governs its macroscopic (refractive index, density, absorption, thickness, etc.) and mesoscopic/microscopic properties (e.g., roughness, interface formation, structure, composition, stoichiometry, etc.). Moreover, these parameters are directly linked to the functionality of these coatings, which encompasses a wide range, including antireflective coatings (AR), dielectric mirrors, various types of filters (e.g., bandpass, longpass, shortpass, notch, etc.), dichroic mirrors, beam splitters, and others [19].

Despite the versatility of optical coatings, this review will focus mainly on dielectric mirrors, especially Bragg structures, with an emphasis on their integration into planar optical microcavities. Deposition strategies will also be discussed, which will be divided into physical vapour methods (PVD), chemical vapour methods (CVD) and solution routes, the latter being a suitable candidate for industrial manufacturing due to their low cost and scalability. Furthermore, a comparison between the advantages/limitations of each thin film process category will be provided. In the following, we briefly present dielectric coatings with reflective architecture and the basic principles of optical microcavities based on a planar configuration.

2. Oxide Materials and Their Subsequent Functionalization for Dielectric Coatings

Transparency over almost the entire UV-Vis-NIR spectrum and a wide range of refractive indices are just a few of the attractive features of dielectric oxides (such as SiO₂, Al₂O₃, TiO₂, ZrO₂, HfO₂, ZnO, etc.), making them preferred for the manufacturing of functional optical coatings, such as dielectric mirrors, filters, antireflective coatings, and others. Along with the latter-mentioned characteristics, chemical inertness and thermal stability are also important factors, especially when they are employed for industrial purposes in hazardous environments (e.g., high temperature, corrosion, and radiation). In the following subsections, specific material properties are detailed. But an overall trend is observed across all oxide pairs: PVD techniques yield the best packing density and refractive indices, whereas solution-based methods have better scalability and are more cost-effective at the expense of porosity.

The most commonly used material to fabricate dielectric mirrors is silicon dioxide (SiO₂). Known for its high transparency and low refractive index (tuneable through the underlying deposition method and parameters—will be subsequently discussed), it provides sufficient optical contrast when paired with materials such as TiO₂, ZrO₂, HfO₂, Ta₂O₅, Nb₂O₅ and even Al₂O₃, among others [27,28,29,30,31,32,33,34]. However, the pair SiO₂/TiO₂ offers the largest optical contrast, rendering it perfect for high-reflectance dielectric mirrors with a smaller number of pairs and sharper stopbands. Koohee Han and Jaun Hyeun Kim demonstrated the functionality of a SiO₂/TiO₂-based multilayer structure for energy-efficient window applications fabricated through a sol–gel route [35]. Considering economic and industrial perspectives, the structure designed for the NIR region (at a target wavelength -λ- of 1000 nm) consists of only three thin films: high-, low-, and high-refractive-index materials (i.e., TiO₂/SiO₂/TiO₂). They determined the refractive index values for TiO₂ (n = 2.14) and SiO₂ (n = 1.44) through ellipsometry, which are satisfactory values for obtaining a highly reflective mirror. The same SiO₂/TiO₂ pair was chosen to obtain a distributed Bragg reflector through e-beam evaporation, which acts as an optical filter with both wavelength and angular dependency [19]. Devising this type of heterostructure ensured a minimization of the wavelength-converting system etendue, which emits in the green wavelength while absorbing in the blue one. To ensure this, the multilayer structure was deposited at room temperature, followed by thermal annealing after each TiO₂ deposition [36,37], thus ensuring an increased refractive index of n = 2.5 for titania, whilst for silicon dioxide it is n = 1.5 at λ = 550 nm. Boucher et al. used SiO₂/TiO₂-based Bragg reflectors to fabricate an erbium-doped cavity through rf sputtering [38]. They produced much denser films compared to those obtained through solution-based methods (such as sol–gel), resulting in higher refractive indices, which were determined using m-line spectroscopy ($n_{{\mathrm{T}\mathrm{i}\mathrm{O}}_2}=2.3$ and $n_{{\mathrm{S}\mathrm{i}\mathrm{O}}_2}=1.44$ at λ = 1542 nm).

Another oxide material is niobium pentoxide (Nb₂O₅), which represents a viable alternative to SiO₂ in the category of materials with a high refractive index. Along with a desirable low absorption coefficient and a wide transparency range, its water/air stability and resistance to acidic environments make it the perfect candidate for both optical and electronic applications [39,40]. In this context, a SiO₂/Nb₂O₅-based multilayer refractive structure was fabricated for blue light by Yuan et al. using e-beam deposition [41]. The optical properties of the individual films grown at a low rate (0.1 nm/s) were investigated through reflectometry, and the refractive index values of SiO₂ and Nb₂O₅ (at a central wavelength of λ = 450 nm) were 1.462 and 2.294, respectively, indicating sufficient contrast to obtain a reflecting surface. Within the framework of micromechanical elements, Richter et al. [42] demonstrate a multilayer structure with stress-controlled reflection based on the SiO₂/Nb₂O₅ pair using rf magnetron sputtering for the low-refractive target material and reactive sputtering in an oxygen pressure for the high-refractive target material. Prior to fabricating the multilayer structure, they investigated various oxide materials and decided to use SiO₂ (n = 1.48) and Nb₂O₅ (n = 2.33). Compared to the e-beam deposition technique, which is known to produce porous coatings [43], sputtering methods offer a fine control of stoichiometry and denser films with low optical losses [44].

Owing to a broad-spectrum transparency and high thermal and chemical stability, tantalum pentoxide (Ta₂O₅) is among the high-refractive-index materials frequently used in optical coatings, from antireflective to guiding thin films (cores) in waveguides operating at high temperatures [45,46,47,48]. In addition, this oxide exhibits relatively low optical and mechanical losses at room temperature, making it favourable for applications in ultra-high-sensitivity optical interferometry for current gravitational wave detection (GWD) [49]. However, both SiO₂ and Ta₂O₅ exhibit a peak in the mechanical loss angle at cryogenic temperatures, which is detrimental for noise-sensitive applications. A strategy for obtaining high-quality optical coatings with reduced thermal noise consists of preparing oxide coatings with multiple cation species, ensuring a higher refractive index and lowering the energy dissipation, as well as increasing crystallisation temperatures, all of which directly affect the sensitivity of GWDs [50]. In this context, Amato et al. [51] investigate the influence of the cation ratio Ti/(Ta + Ti) on the performance of the Bragg reflector, which was deposited through ion beam sputtering (IBS). The choice of using this technique is not accidental. The ion beam sputtering technique is recognised as a state-of-the-art deposition method for fabricating high-quality optical thin films with a high density, low absorption, and scattering losses [52,53]. Moreover, this method offers thickness uniformity over a large area and smooth interfaces, which are critical factors for multilayer structures, such as dielectric mirrors [54]. After the multilayer structure deposition, an annealing treatment was performed at 500 °C for 10 h in a normal atmosphere, which is a standard post-deposition step for dielectric mirrors in the GWD field, diminishing both optical and mechanical losses. The ellipsometry analysis conducted on the amorphous oxidic thin films concluded that the optimal cation ratio is approximately 21%. However, the authors reinforced the necessity for further studies on the impact of annealing conditions, as well as the cation ratio on the optical and mechanical properties of these oxide coatings. The values of the refractive index and extinction coefficient of Ta₂O₅ thin films processed using different deposition techniques are presented in the table below (

Table 1. Optical properties of Ta₂O₅ thin films deposited through different techniques.

Process	Technique	n@550 nm	k@550 nm	References
PVD	Ion beam sputtering (IBS)	2.12	≤2 × 10−4	[54]
	Electron beam evaporation (e-beam)	2.0419	-	[55]
	Radio frequency magnetron sputtering (RF magnetron)	<2.2	-	[56]
	Pulsed laser deposition (PLD)	>2.2	>5 × 10−4	[57]
CVD	Atomic layer deposition (ALD)	2.17	-	[58]
	Plasma-enhanced chemical vapour deposition (PECVD)	2.13	<1 × 10−4	[59]
Solution-based	Sol–gel/dip coating	1.70–1.72	<1 × 10−4	[60]

).

Zirconia, or zirconium dioxide (ZrO₂), is a transition metal oxide that boasts a high refractive index and a band gap greater than 5 eV [55], minimal optical losses, and a wide transparency window spanning the visible to near-infrared region, making it ideal for combining with SiO₂ to obtain optical coatings [56]. A SiO₂/ZrO₂ double-layer coating was fabricated through the sol–gel technique for antireflective coating in solar cells [57]. Moreover, the same pair of oxide materials can be used as a filter in thermophotovoltaic applications. Mansoor et al. [58] numerically investigated the influence of the architecture of a ZrO₂/SiO₂-based filter on the system efficiency. The modified double-quarter-wave stack configuration shows promising results, leading to an efficiency of approximately 29% compared to the case without, where the efficiency was only 16%. Two independent scientific papers [59,60], provided the optical property data of ZrO₂ and SiO₂ for their study, while considering that the solution-based methods present a higher porosity [61], leading to a decreasing refractive index. A detailed comparison based on the fabrication methods of ZrO₂ thin films is presented in

Table 2. Optical properties of ZrO₂ thin films deposited through different techniques.

Process	Technique	n@550 nm	k@550 nm	References
PVD	Electron beam evaporation (e-beam)	>1.76–2.02	-	[62]
	DC magnetron sputtering	2.13–2.05	-	[61]
	Pulsed laser deposition (PLD)	>2.2	>5 × 10−4	[57]
CVD	Atomic layer deposition (ALD)	1.8–2.10	-	[63]
Solution-based	Sol–gel/dip coating	1.70–1.72	<1 × 10−4	[60]

.

For optical coating, silicon dioxide can be used alongside hafnia or hafnium dioxide (HfO₂). Its high refractive index, wide transparency from UV to mid-IR region (due to its bandgap, which is larger than 5.5 eV [64]) and significant thermal and chemical stability [65] propelled HfO₂ for the manufacture of antireflective coatings [66] and highly reflective mirrors [67], but also in other applications, such as memory devices [68], among others [69,70]. Moreover, this oxide is known to have a high laser-induced damage threshold (LIDT); thus, Gong et al. [71] fabricated a highly reflective mirror based on SiO₂/HfO₂ by e-beam evaporation, showing superior resistance induced by a picosecond laser compared to a SiO₂/Ta₂O₅ mirror. As previously mentioned, the optical properties are closely related to the deposition technique used to obtain the HfO₂ thin films, and more precisely, to the deposition parameters. The table below presents the values of the refractive index, as well as the extinction coefficient at λ = 550 nm, depending on the deposition method used (

Table 3. Optical properties of HfO₂ thin films deposited through different techniques.

Process	Technique	n@550 nm	k@550 nm	References
PVD	Electron cyclotron resonance ion beam deposition	1.74–1.95		[72]
	Electron beam evaporation (e-beam)	2.0419	-	[73]
	Radio frequency magnetron sputtering (RF magnetron)	<2.2	-	[74]
	Pulsed laser deposition (PLD)	>2.2	>5 × 10−4	[75]
CVD	Atomic layer deposition (ALD)	2.17	-	[76]
	Plasma-enhanced chemical vapour deposition (PECVD)	2.13	<1 × 10−4	[77]
Solution-based	Sol–gel/dip coating	1.70–1.72	<1 × 10−4	[78]

).

Aluminium oxide or alumina (Al₂O₃) is used in electronic and optoelectronic devices due to its dielectric and protective characteristics, while its high transparency in a wide wavelength range makes it a suitable candidate for optical coatings [79,80]. For these applications, Al₂O₃ can be manufactured into thin films, through various physical, chemical and solutions-based methods [81,82,83,84,85]. In the context of optical coatings, Al₂O₃ thin films can be used in conjunction with SiO₂ as a multilayer structure, which can act as a reflective filter [86]. The same material pair was integrated as an antireflective coating in GaAs-based photovoltaic cells [87]. In the table below, the influence of the deposition method used on the optical properties of Al₂O₃ thin films is presented (

Table 4. Optical properties of Al₂O₃ thin films deposited through different techniques.

Process	Technique	n@550 nm	k@550 nm	References
PVD	Electron beam evaporation (e-beam)	<1.61	<0.003	[88]
	Radio frequency magnetron sputtering (RF magnetron)	1.634–1.667	0.0004–1.6651	[89]
	Pulsed laser deposition (PLD)	>1.732–1.805	<0.01	[90]
CVD	Atomic layer deposition (ALD)	>1.64–1.67	-	[82]
	Plasma-enhanced chemical vapour deposition (PECVD)	>1.65–1.69	<1 × 10−4	[91]
Solution-based	Sol–gel/dip coating	1.51/1.52	<1 × 10−4	[92]

).

Zinc oxide (ZnO) is another oxide that targets standalone optical coatings (antireflective coatings [93], UV filters [94,95], etc.), along with other applications in various fields, due to its optical, piezoelectric, electrical, and biological properties [96,97,98]. In addition to a wide band gap (3.3 eV) and optical transparency in the visible range, this semiconductor oxide also benefits from a high exciton binding energy at room temperature (60 meV), making it suitable for the development of optical planar microcavities due to the strong coupling, which generates exciton–polaritons at room temperature [99,100,101,102]. Regarding the methods employed to grow thin films, a variety of deposition techniques can be used, from physical vapour (such as sputtering-based deposition [103,104,105], e-beam evaporation [106,107], pulsed laser deposition [108,109]) to chemical [110] and solution-based solutions [111], depending on the intended purpose. A detailed presentation of the deposition techniques used to obtain ZnO thin films is available elsewhere [96,112,113,114]. Since 2008, when ZnO was introduced and experimentally validated as a cavity material due to its attractive properties [102], several reports on strongly coupled ZnO-based microcavities have emerged [100,115,116,117,118,119,120,121,122]. However, the first mention of ZnO in the context of optical resonators is much earlier, namely when Zamfirescu et al. modelled a polariton laser, i.e., an optical microcavity comprising a ZnO layer between two Mg_0.3Zn_0.7O/ZnO Bragg mirrors, and calculated a high value for the Rabi splitting in a vacuum (around 120 meV), stating that this could lead to the opening of a new generation of polariton lasers that can operate at a threshold power of only 2 mW at room temperature [123].

2.1. Thin-Film Processing of Oxide Materials for Dielectric Mirrors

The possibilities of combining oxide materials to obtain functional coatings, more precisely dielectric mirrors, do not stop at the pairs formed between SiO₂ and the other previously mentioned materials. The table below contains several pairs of oxide materials (

Table 5. Several examples of material pairs used in fabricating dielectric mirrors.

Pair	DBR Capability	Substrate Dimensions	Technique	References
SiO2/ZrO2	R ≥ 95% (at 366 nm) & PGB* = 82 nm for eight-pair structure	Si wafer (its size is omitted)	Reactive helicon wave-excited plasma sputtering method	[28]
SiO2/Al2O3	R ≥ 97% at 290 nm	The dimensions of the substrate are not mentioned	Radio frequency sputtering method	[32]
HfO2/TiO2	R ~ 80% at 385 nm	-	Radio frequency sputtering method	[124]
Al2O3/ZrO2	R = 99.8% at 377 nm for a 12.5-thin film pair structure	Si wafer (its size is omitted)	Pulsed laser deposition	[125]
Al2O3/YSZ	R = 99.7% for 10-period bilayer films	3-inch Si substrate	Pulsed laser deposition	[126]
Al2O3/TiO2	R = 98.1% (at 471 nm) & PBG* of 79 nm for six-period bilayer films	2 × 2 cm2	E-beam method	[127]
Al2O3/HfO2	85.7% at 541.7 nm	At least 10 cm of silicon substrate	ALD	[128]
SiO2/TiO2	99.99% & tuneable PBG* from visible to NIR due to the solution ageing	Glass substrates	Sol–gel—spin coating	[129]

*PBG stands for photonic bandgap, and R stands for reflectance (will be discussed in the relevant section of this review).

) used in the fabrication of dielectric mirrors, particularly of Bragg reflectors, along with their performance, substrate dimensions, and fabrication methods employed.

In the context of this review, which is focused on metal oxide-based Bragg structures and microcavities, the main criterion for selecting a material depends on the target application. This factor is, as discussed previously, the transparency window, which is preferred to be as wide as possible, spanning the UV–NIR range. The second factor is constrained by the technological maturity of deposition processes, which can ensure precise dimensional control, thickness uniformity, low scattering losses, and material pair compatibility for multilayer architectures. Ever since Fabry–Pérot microcavities (or planar microcavities) became key tools in quantum electrodynamics studies, inorganic materials have dominated the scene, benefiting from a wide range of compositions capable of providing substantial dielectric contrast and a wide range of mature thin-film deposition technologies [130]. Furthermore, outside this context of optical microcavities, inorganic materials have been the main players in the optical coating industry [13,33,131].

Besides the classical metal oxides, materials such as monocrystalline semiconductors [132,133], superlattices [134], nitrides [135], fluorides [136], oxide nanoparticles [137], and also polymers [138,139] have been chosen to manufacture DBRs through different processes (PVD, CVD, solution-based processing and even hybrid methods). As previously mentioned, the choice of thin-film deposition technique is critical, as the properties of the materials depend intrinsically on the deposition method and the parameters used. Although this process is not profitable for dielectrics, it is worth mentioning that epitaxy-based techniques (such as molecular beam epitaxy (MBE)) remain the standard for semiconductor microcavities (and their DBRs), offering several attractive features such as a high crystal quality, ultra-smooth interfaces, fine control of thickness/composition, wafer-scale uniformity, monolithic integration, and in situ optical monitoring. However, in the context of dielectric coatings, their use remains limited due to infrastructure costs, high process temperatures and, especially for MBE, high-vacuum requirements, as well as lower growth rates compared to PVD/CVD (Table 6).

In contrast to MBE, PVD and CVD are less demanding process-based. Thin-film processing through CVD is favoured because it produces uniform, high-purity and quality coatings. Through its surface-reactive mechanisms, CVD offers a series of appealing features, including an excellent coverage, even on irregular shapes, high density, and good substrate adhesion to a variety of substrates. Nevertheless, high operating temperatures and expensive equipment are among the disadvantages of this technique, which reduce its compatibility with materials that are sensitive to high temperatures and increase manufacturing costs [140]. PVD is the solution to overcome the limitations imposed by CVD [141]. Although some PVD methods allow the film growth even at room temperature, PVD generates dense and well-adhered films, allowing the coating of a wide range of substrates and, in many cases, offering cost advantages over alternative routes.

On the other hand, solution-based processes (such as spin coating and dip coating) are most suitable when a low-cost, no-vacuum, and large-surface-area deposition is required [142]. A wide range of materials is used, ranging from polymers and inorganics to hybrid organic/inorganic combinations. However, in the case of multilayer structures, such as Bragg structures, where a repeated deposition of the constituent thin films is required, solution processing becomes difficult. It is necessary to use materials soluble in orthogonal solvents so that the newly deposited thin film does not compromise the one underneath. Additionally, the interface quality in the case of solution-based processes is poor, which is detrimental to the performance of the photonic device.

Table 6. Thin-film deposition processes, their characteristics and limitations.

Deposition Process	Advantages	Limitations/ Disadvantages	Scalability	Optical Performance	References
PVD	High temperatures are not necessary; high-density coatings; good adhesion to various substrates	Poor coverage on complicated geometric shapes; low deposition rate; vacuum equipment	Suitable for large area coatings, but requires high-energy sources and high-vacuum pumping systems, which can limit the throughput	Produces high-density, stoichiometric films with superior crystallinity and high refractive indices; essential for high-Q microcavities	[143,144,145,146,147]
CVD	Uniform, high-purity and high-density coatings; excellent coverage, even on irregular shapes; good substrate adhesion to various substrates	High operating temperatures; expensive equipment and manufacturing costs; incompatible with high-temperature-sensitive materials	A standard industrial process capable of large-scale production with excellent throughput for semiconductor and coating industries	Offers exceptional conformal coverage and precise atomic-level thickness control, resulting in high-quality films with low scattering
Solution-based	Cheap; no vacuum or sophisticated equipment required; large area coverage; decent substrate adhesion on any type of substrates, including flexible ones—can be improved	Poor quality interface; material limitation due to the orthogonal solvents; defects induced by solvent	Cost-effective and compatible with high-throughput methods like roll-to-roll or dip-coating for large-area applications	Allows for easy tuning of optical constants through chemistry, though films often exhibit higher porosity and lower refractive indices than PVD methods

Table 6. Thin-film deposition processes, their characteristics and limitations.

Deposition Process	Advantages	Limitations/ Disadvantages	Scalability	Optical Performance	References
PVD	High temperatures are not necessary; high-density coatings; good adhesion to various substrates	Poor coverage on complicated geometric shapes; low deposition rate; vacuum equipment	Suitable for large area coatings, but requires high-energy sources and high-vacuum pumping systems, which can limit the throughput	Produces high-density, stoichiometric films with superior crystallinity and high refractive indices; essential for high-Q microcavities	[143,144,145,146,147]
CVD	Uniform, high-purity and high-density coatings; excellent coverage, even on irregular shapes; good substrate adhesion to various substrates	High operating temperatures; expensive equipment and manufacturing costs; incompatible with high-temperature-sensitive materials	A standard industrial process capable of large-scale production with excellent throughput for semiconductor and coating industries	Offers exceptional conformal coverage and precise atomic-level thickness control, resulting in high-quality films with low scattering
Solution-based	Cheap; no vacuum or sophisticated equipment required; large area coverage; decent substrate adhesion on any type of substrates, including flexible ones—can be improved	Poor quality interface; material limitation due to the orthogonal solvents; defects induced by solvent	Cost-effective and compatible with high-throughput methods like roll-to-roll or dip-coating for large-area applications	Allows for easy tuning of optical constants through chemistry, though films often exhibit higher porosity and lower refractive indices than PVD methods

2.2. Thin Film Properties for Dielectric Mirrors

Regarding thin-film preparation techniques, as previously mentioned, the properties of the materials intrinsically depend on the deposition method used. The table below (Table 7) summarises representative values of the refractive indices determined for SiO₂ and TiO₂ thin films obtained through PVD, namely PLD, magnetron sputtering and e-beam evaporation techniques. It appears that, without ion assistance (IAD/PIAD) or an annealing step [148], electron beam deposition tends to produce less dense films and thus lower refractive indices [149]. However, the deposition of metal oxide films through the e-beam technique presents several important advantages: high deposition rates, which reduce the fabrication time and, implicitly, the manufacturing costs; a good control of thickness and scalability; and a high reproducibility of the process [41].

Classical studies show that PIAD leads to a higher n and lower k than conventional evaporation, confirming the densification effect of ion bombardment [62]. In contrast to the e-beam technique, magnetron sputtering can provide values higher than n = 2.24 at λ = 550 nm for TiO₂ films, even at room temperature [150]. Also, high refractive index values are an excellent indication of high packing density, and this technique is also known for large-area uniformity. Another technique, relevant from an industrial point of view, is ion beam sputtering (IBS), due to its high-quality and homogeneous thin films, even when deposited on large surfaces. This deposition method is valued for its ability to produce thin films with low scattering and absorption and a high density (and thus a high refractive index) [50,72].

And, last but not least, PLD is an energetic technique that frequently produces very dense films with excellent stoichiometry but comes with the risk of particle/droplet formation. The latter-mentioned issue can be addressed by using industrial PLD equipment. For example, Solmates PLD systems are equipped with a disc having four slits arranged at 90°, positioned between the target and the substrate, which rotates and removes slower ablated droplets from the plasma.

Another key parameter which is dependent on the conditions and deposition method is the mechanical stress within the thin films, which in turn will determine the performance as well as the reliability of photonic devices. While the subject of “stress in thin films” is an entire review article on its own, we will address it summarily in this paper. The stress can be intrinsic, thermal or structural [151]. The thermal stress may arise from the mismatch between the thermal expansion coefficients (CTE) of the deposited film and the substrate, while the intrinsic one is determined by the kinetics of the deposition [152,153]. One such example is SiO₂ obtained through PECVD, which has an intrinsic stress highly dependent on the choice of precursor, RF power, substrate temperature, and gas flow rate [154].

A well-known method to overcome the residual stress is increasing the substrate temperature; at the same time, increasing the deposition rate can further increase the tensile intrinsic stress due to local bond disorder and the appearance of voids [155]. Recent studies [156] have also shown that the gas flow during ICPVD deposition is the primary parameter that can adjust the stress in SiO₂ thin films between tensile and compressive states, whilst not significantly altering the refractive index of the material. By doing so, the obtained structure can maintain its structural integrity as a suspended photonic structure.

The kinetic models also suggest that the stress in thin films is formed between the tensile and compressive states. The latter is formed when atoms are inserted at the boundaries, while tensile stress is generated through grain boundary formation [152]. For example, in the case of PLD, “atomic peening” and collision-induced densification result in compressive stress [153]. Another key feature to take into account is that, in the case of manufacturing DBRs on large wafers, the stress appears due to the interactions of the thin film and the substrate curvature, resulting in a nonlinear stress distribution across the wafer [151].

The main problems that arise from stress-related issues in thin films are cracking and delamination. At the same time, stress-induced birefringence, which appears due to the elastic–optic effect, can alter the polarisation and distort the device temperature sensitivity [157].

Overall, the results confirm that the choice of preparation method must be made with caution, as optical performance is closely correlated with the process used.

Table 7. Examples of refractive index variations depending on the deposition method.

Material	Refractive Index n @ Wavelength λ	Technique	References
Silicon dioxide (SiO2)	n = 1.46–1.52 @ λ = 633 nm	PLD	[158]
	n = 1.41–1.52 @ λ = 633 nm	RF magnetron sputtering	[159]
	n = 1.43–1.49 @ λ = 633 nm	E-beam evaporation	[149]
Titanium dioxide (TiO2)	n = 2.6 @ λ = 550 nm	PLD	[160]
	n = 2.57–2.74 @ λ = 550 nm	RF magnetron sputtering	[161]
	n = 2.06–2.08 @ λ = 550 nm	E-beam evaporation	[162]

Table 7. Examples of refractive index variations depending on the deposition method.

Material	Refractive Index n @ Wavelength λ	Technique	References
Silicon dioxide (SiO2)	n = 1.46–1.52 @ λ = 633 nm	PLD	[158]
	n = 1.41–1.52 @ λ = 633 nm	RF magnetron sputtering	[159]
	n = 1.43–1.49 @ λ = 633 nm	E-beam evaporation	[149]
Titanium dioxide (TiO2)	n = 2.6 @ λ = 550 nm	PLD	[160]
	n = 2.57–2.74 @ λ = 550 nm	RF magnetron sputtering	[161]
	n = 2.06–2.08 @ λ = 550 nm	E-beam evaporation	[162]

2.3. Stoichiometrically Complex Oxides with Optically Tunable Properties Under Electrical, Mechanical or Acoustic Excitation

Current trends in the field of photonics regarding the transition from bulk crystal-based devices to photonic integrated circuits (PICs), and, more specifically, to nonlinear integrated photonics, propel the utilisation of thin films based on stoichiometric complex oxides (such as lithium niobate, lithium tantalate, and barium titanate) since they simultaneously promise both ultrafast light control and massive optical parallelism, something that classical electronic systems cannot achieve [163,164,165]. Although not the primary objective of this work, it is worth mentioning that stoichiometric complex oxides with electro-optical properties constitute the link between electronics and photonics and underlie a wide range of applications, from communications and computing to sensors and quantum information processing.

The history of lithium niobate (LiNbO₃ or Li₂O × Nb₂O₅ or LN) dates back to 1949, when Matthias and Remeika first synthesised it at Bell Laboratories [166], and, sixteen years later, the first single crystal was obtained using the Czochralski technique [167]. High electro-optic (${\mathrm{r}}_{33}=30.9 \mathrm{p}\mathrm{m}/\mathrm{V}$) and second-order nonlinear coefficients (${\mathrm{d}}_{33}=27 \mathrm{p}\mathrm{m}/\mathrm{V}$), as well as a high value for the refractive index (around n = 2.2 for ordinary axis and n = 2.1 for extraordinary one [168]) and transparency from 0.4 to 5.5 µm, are some attractive features of this material (monocrystal), which is particularly suitable for applications where electro-optic modulation is desirable, such as communication technology [163,169,170]. Nevertheless, the existing technology based on the smart cut method is expensive and time-consuming (Figure 1), and the thin films (TFLN) obtained are not epitaxial, which alters their outstanding properties. Therefore, a viable and cheaper alternative would be to grow this material through a classical thin film deposition technique, which requires a compatibility with standard semiconductor technology processing on an industrial-grade level substrate, such as PVD (PLD [171], sputtering-based methods [172]) and solution-based routes (sol–gel [173], etc.). For a detailed presentation of TFLN-based photonics, there are several publications in the literature dedicated to this topic in recent years [174,175,176,177,178,179].

Lithium tantalate (LiTaO₃ or Li₂O× Ta₂O₅ or LT) is another ferroelectric oxide that emerged as a cheaper alternative to LN in the context of this new trend concerning the increased demand for ultra-high-speed PICs, imperative for energy-efficient data centres and high-performance computing for AI [165,181]. Compared to the economic limitations caused by the lack of a large consumption market for LN in insulator structures (LNOI), LT has already entered mass production due to its use in 5G radio frequency filters [182,183]. This material shows an enhanced chemical stability by simply replacing the Nb atoms with Ta atoms [184], as well as a wider bandgap (3.93 eV [185]), facilitating nonlinear optical conversion in the Vis and UV regions [186]. Although the Pockels coefficient is similar to that of LN (${\mathrm{r}}_{33}=30.5 \mathrm{p}\mathrm{m}/\mathrm{V}$), LT exhibits advantages such as: a birefringence reduced by more than 10 times (which suppresses mode mixing in the guides); a higher optical destruction threshold; and 10-times-lower microwave losses [187,188], essential for quantum transduction [189,190]. Although, historically, the low Curie temperature (610–700 °C) has prevented the fabrication of waveguides with traditional ion diffusion methods [191], recently, Li et al. [192] fabricated the first PIC platform on LNOI through direct etching using diamond-like carbon (DLC) masks, thus validating and expanding other materials’, such as BaTiO₃ and LT, ability to support advanced applications, such as soliton micro-comb generation and industrial-scale electro-optical switching. Regarding the preparation of LT thin films through classical deposition methods, these are largely divided: sol–gel [193,194], magnetron sputtering [195,196] and MBE methods [197,198]. Additional methods have also been employed for manufacturing these thin films, such as metal–organic chemical vapour deposition (MOCVD), liquid-phase epitaxy (LPE) and PLD. Regardless, epitaxial thin films have been obtained only on specific substrates, such as LN or sapphire [199,200,201].

Barium titanate (BaTiO₃ or BaO × TiO₂ or BT) is a lead-free ferroelectric material, which in recent years has had a growing range of applications, especially in the photonic field [180,202], and more specifically in nonlinear optics due to its Pockels coefficient being larger than 1000 pm/V [203], the highest in comparison to other stoichiometric complex oxides [204,205]. Other applications include capacitors, sensors, transducers and optical devices due to its piezoelectric, dielectric and optical features [206]. Similarly to LN, due to its transparency window that is wide, from 460 to 6300 nm, BTO becomes another candidate for communication technology. Waveguides are a possible application of this ferroelectric material, due to its high refractive index value as well as its epitaxial integration on silicon-on-insulator substrates [204,205,206,207,208]. The development of BTO epitaxial films for photonic devices is predominantly based on ultra-high-vacuum techniques, such as (MBE) [209,210] and ALD, but also on other techniques (e.g., PLD [211,212,213], sputtering [214]), where the crystalline quality and optical properties critically depend on lattice–substrate matching and the use of buffer thin films to avoid the formation of polycrystalline phases [213,215]. Although these methods offer a superior performance, the high costs and the difficulty of “top-down” processing—caused by the chemical resistance of complex stoichiometric oxides to etching—have required the adoption of strategies for hybridising the films with CMOS-compatible materials, such as silicon or silicon nitride [216,217,218]. Alternatively, “bottom-up” approaches based on wet chemistry (sol–gel, nanoparticles) offer significant advantages in terms of cost and large surface coverage, but their applicability remains limited by the need for high-temperature thermal treatments, often incompatible with conventional integrated circuits, or by the difficulty of controlling the thickness and structural homogeneity [219,220].

Moreover, strain engineering has become a key strategy in improving the performance of complex non-centrosymmetric thin films (piezo-, pyro-, and ferroelectrics). This can result in enhancing functional properties or even unlocking entirely new ones. The aim of using the strain engineering approach is to reveal the changes in optical properties (band gap value, refractive index, etc), as well as in the electrical ones (polarisation and dielectric behaviour), to enhance the efficiency of these materials in optical devices such as MOEMS [221] or, as stated, pave the way for multifunctional devices. For example, different from the conventional epitaxial misfit homogenous strain induced in ferroelectric or multiferroic thin films, strain gradients can be induced into the film thickness by growing compositionally graded heterostructures through dopants [222,223], where the evolution of the lattice parameter with composition variation will impose a “chemical pressure” as a driving force for the strain gradient.

3. Oxide Dielectric Coatings Applications: Mirrors and Their Role in Planar Optical Cavities

Dielectric mirrors are optical coatings recognised as one-dimensional planar photonic crystals (1D) that operate by exploiting the constructive interference of the successive reflections at the dielectric interfaces (Figure 2). Unlike their metallic counterparts, dielectric mirrors are known for their ability to produce a high reflection band in which light is prohibited from propagating due to the back diffraction or reflection (this spectral region is called photonic band gap or, in short, PBG [224]), and thus their technological applicability is widespread in various fields, being used as optical structures in sensors [225], filters [226], optoelectronic and photonic devices [227,228,229,230,231,232], as well as waveguide structures in optical fibres [41].

Additionally, dielectric mirrors present reduced absorption losses, which make them excellent candidates for GWDs in order to reduce thermal noise and enhance sensitivity [50]. When used in cryogenic-specific applications such as ultra-sensitive detectors, mechanical losses are no longer neglected due to their substantial number of pairs in their structure.

A particular case of dielectric mirrors is represented by Bragg mirrors, whose architecture relies on a multilayer structure comprising quarter-wavelength thin films of two dielectric materials with different refractive indices ($n_{\mathrm{l}\mathrm{o}\mathrm{w}}$ and $n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}$). Bragg structures can also be fabricated using a single material in which the pore gradient varies periodically [233,234,235]. In the quarter-wavelength configuration, the thickness ($d_i$) of each thin film “i” is equal to

$$d_i={\displaystyle \frac{{\lambda }_{\mathrm{C}}}{4{\mathrm{n}}_{\mathrm{i}}}}$$

(1)

where $n_{\mathrm{i}}$ stands for $n_{\mathrm{l}\mathrm{o}\mathrm{w}}$ and $n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}$, and ${\lambda }_{\mathrm{C}}$ is the central wavelength. As a consequence of the interaction of light with the 1D photonic crystal (i.e., coherent light diffraction), the reflectivity spectrum of a DBR exhibits a peak. The material selection is crucial, as the reflectance of the Bragg structure (${\mathrm{R}}_{\mathrm{D}\mathrm{B}\mathrm{R}}$) depends on both the number of pairs ($\mathrm{N}$) and the optical contrast between them ($\mathsf{\Delta }n=n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}-n_{\mathrm{l}\mathrm{o}\mathrm{w}}$) [41]. Thus, as the ratio of $n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}$ to $n_{\mathrm{l}\mathrm{o}\mathrm{w}}$ increases, both the spectral width and reflectivity follow this trend.

Therefore, to maintain a high reflectivity with a minimum number of thin-film pairs, the optical contrast between the materials must be quite large [236]. The maximum value of reflectivity can be calculated using the following equation that takes into account the refractive index of the substrate $n_{\mathrm{s}}$ [237]:

$${\mathrm{R}}_{\mathrm{D}\mathrm{B}\mathrm{R}}={\left[{\displaystyle \frac{1-n_{\mathrm{s}}{\left(\frac{n_{\mathrm{l}\mathrm{o}\mathrm{w}}}{n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}}\right)}^{2\mathrm{N}}}{1+n_{\mathrm{s}}{\left(\frac{n_{\mathrm{l}\mathrm{o}\mathrm{w}}}{n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}}\right)}^{2\mathrm{N}}}}\right]}^2$$

(2)

As previously stated, the most common pair of materials in dielectric mirrors is SiO₂/TiO₂, owing to their large refractive index contrast, transparency over the visible and near-infrared regions, and exceptional chemical and thermal stability. Feng et al. [238] employed a deposition technique more compatible with the fabrication processes of optoelectronic devices based on erbium-doped semiconductors to obtain a DBR. Combining the e-beam evaporation technique with a post-annealing treatment of TiO₂ films, they successfully fabricate SiO₂/TiO₂-based DBR on Al₂O₃ substrates with impressive results, achieving a reflectivity of over 95% at 1.5 µm. They emphasise the importance of integrating this dielectric DBR with III-nitride optoelectronic devices for optical communications (IR emitters, erbium-doped fibre amplifiers (EDFAs) and other photonic devices).

The same material pair was used to obtain DBRs in InGaN/GaN flip-chip light-emitting diodes (FC-LEDs) through radio frequency magnetron reactive sputtering [239]. This approach aims to address the issues of the reflective p-electrodes, which restrict the light extraction efficiency (LEE) through light absorption and current crowding. As is well known, the width of the stopband (PBG) increases with optical contrast, while the number of pairs dictates how steep the edge is [240], and it is determined by the following equation:

$${\mathsf{\Delta }\lambda }_{\mathrm{P}\mathrm{B}\mathrm{G}}={\displaystyle \frac{{4\lambda }_{\mathrm{C}}}{\mathsf{\pi }}}\arcsin \left({\displaystyle \frac{\mathsf{\Delta }n}{n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}+n_{\mathrm{l}\mathrm{o}\mathrm{w}}}}\right)$$

(3)

By varying the number of pairs, they prepared Bragg structures with an over 95% reflectance at a normal incidence angle in the ranges of 400–530 nm (using 12 sequential pairs) and 385–720 nm (using 20 sequential pairs).

Given the dielectric nature of the DBRs, these multilayer structures require transparent conductive coatings (TCO, such as ITO) to be used in optoelectronic devices (photovoltaic cells, LEDs, etc.). Another important aspect of a DBR for these types of applications is optimising the morphology of each dielectric thin film, as any defect (such as pores) is detrimental to the conductivity of the TCO. Additionally, the interfaces between these metal oxide thin films significantly influence the efficiency and optical performance of the Bragg mirror, making well-defined thin films desirable. In this regard, Tahereh et al. [241] studied the influence of the annealing temperature on the morphological, optical and electrical properties of the SiO₂/TiO₂/ITO structure deposited on glass substrates through thermal evaporation. Therefore, an improvement was noted by increasing the annealing temperature from 250 °C to 550 °C: the surfaces became much smoother, more homogeneous and uniform, and the resistivity of the structures diminished, while the reflectance reached a maximum of 94.4%. A smooth surface minimises light scattering and maintains the specular component of the reflection; in contrast, a rougher surface introduces significant diffuse scattering, diminishing the effective reflectance, especially when the roughness becomes comparable to a relevant fraction of the wavelength [242,243]. The relation between the scattering parameter, denoted $\mathrm{s}$, and the roughness of a surface thin film ($\mathrm{r}$) is given by the following [244]:

$$\mathrm{s}={\left({\displaystyle \frac{4\mathsf{\pi }\mathrm{r}}{\lambda }}\right)}^2$$

(4)

In summary, the optoelectronic properties can be controlled by introducing an additional step, namely an annealing treatment.

Scientific articles on the large-scale deposition of these optical thin films are scarce in the literature, a phenomenon explainable by a series of technical and economic factors, such as the challenge of obtaining uniform films on large surfaces, defects induced by mechanical stress, cost in terms of time and manufacturing, and difficult metrology on large surfaces (uniformity maps and statistics), which means that few works explicitly treat the subject as such. However, Christidis and his colleagues [245] prepared SiO₂/TiO₂ nanomultilayers to photonically enhance a thermal protection system (TPS) against the thermal radiation encountered by a spacecraft during space missions. The oxide thin films were prepared through e-beam evaporation on 4″ silicon substrates. The photonic structure consists of 18 thin films of TiO₂/SiO₂ and was designed to provide a high reflectivity spanning the spectral domain of 700–1600 nm (Figure 3).

Another combination of oxide materials used to fabricate DBRs is SiO₂/HfO₂, especially for visible and near-UV applications. Réveret et al. [246] employed a SiO₂/HfO₂-based Bragg stack in a planar microcavity using ion beam-assisted electron-gun vacuum evaporation. They obtained reflective coatings with a small number of pairs over a 60 × 60 mm² surface with a uniformity smaller than 4%, a high reflectivity, a broader PBG, and a lower penetration depth (LDBR), which are beneficial for achieving an efficient planar microcavity-based device. The penetration depth of a DBR ($L_{DBR}$) is a crucial parameter in the fabrication of optical microcavities, since it directly influences the quality factor ($\mathrm{Q}$), mode volume (${\mathrm{V}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) and resonance. More will be discussed in the section dedicated to optical microcavities.

DBR-Based Optical Microcavities

Charles Fabry and Theodore Pérot first proposed the concept of an optical microcavity in the late 19th century, having the ability to confine light within a reduced volume for a relatively long period of time [247]. These microstructures can alter the local photonic density of states (LPDOS) owing to their inherent resonance modes, which can be designed by choosing the component materials, geometry and dimensions [248,249]. Consequently, these microresonators have gained increased attention from researchers over the years due to their light coupling, enabling their utilisation in many applications, such as filters [250,251], lasers [252], sensors [253,254] and optoelectronic devices such as organic solar cells [232,255].

This review is mainly focused on metal oxide-based photonic applications fabricated through PVD, particularly devices such as Fabry–Pérot. A comprehensive discussion on the different types of microcavities depending on the structure employed can be found elsewhere [232]. Having stated that, a Fabry–Pérot cavity consists of two highly refractive mirrors, placed parallel to each other, sandwiching a thin film, which is known as the cavity layer (Figure 4). Depending on the material placed in the cavity layer (spacer), the optical microcavities are known as being passive when this material is transparent (non-emitting) for cavity photons, i.e., it does not introduce optical gain and has negligible absorption/scattering.

The nature of the mirrors used in the fabrication of a microcavity could be metallic, dielectric or both. In the latter situation, the optical microstructures are referred to as hybrid and plasmonic Fabry–Pérot cavities [256,257,258,259]. However, as previously mentioned, the metallic ones are not favourable due to their relatively low reflectivity and high absorption, which limit their efficiency [260]. By replacing metallic mirrors with dielectric counterparts, particularly Bragg mirrors (discussed earlier), the related issues can be addressed. This strategy has been employed in vertical cavity surface-emitting lasers (VCSELs) since the 1990s [261]. The working principle of a Fabry–Pérot microcavity is based on stationary electromagnetic waves that are formed by the interference of waves reflected between the two mirrors. The resonance phenomenon occurs when the distance between the mirror surfaces (i.e., optical path length) is a multiple of ${\lambda }_c/2$, permitting light to be confined within the cavity layer [262].

Intensive studies have been carried out in recent years due to their capability of allowing precise control of the optical response in the regime of strong exciton–photon coupling, leading to the formation of a composite quasiparticle, namely exciton–polaritons, in the cavity [263]. The possibilities of achieving Bose–Einstein condensation (BEC) [264,265,266,267] and polariton lasing [137,268,269,270], even at room temperature, are some of the current topics in the field of microcavity research. The generation of entangled photons is a cutting-edge topic, driven by the move towards integrated on-chip sources (SiN/Si/SiC) for scalable and telecom-compatible quantum communications [271,272,273]; by the optimisation of nonlinear optical schemes in microresonators, such as spontaneous parametric down-conversion (SPDC) and spontaneous four-wave mixing (SFWM) [271,273]; and by deterministic quantum dot sources that simultaneously pursue high brightness and high entanglement fidelity [274,275]. In parallel, resonator schemes exploiting biexciton resonant hyperparametric scattering (BRHPS) in microcavities offer efficient and spectrally selectable pathways for entangled pairs, where the control of exciton–photon coupling (Rabi splitting) is essential [276,277].

Compared to other microcavities, such as whispering gallery resonators and complex photonic structures, the Fabry–Pérot cavity is easier to design and fabricate, although it generally offers less spatial confinement (larger ${\mathrm{V}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) and a more modest temporal confinement (lower $\mathrm{Q}$) of the electric field. Therefore, the two previously mentioned quantities related to the confinement of the electric field describe the following: (a) the mode volume (${\mathrm{V}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) quantifies the effective volume in which the light is concentrated and is directly linked to the maximum value of the electric field intensity within the cavity (spatial confinement); and (b) the quality factor ($\mathrm{Q}$) expresses the ratio of the energy stored in the cavity to the power lost per cycle or, equivalently, the lifetime of the photons in the cavity (temporal confinement) [278,279]. In the following, we will provide quantitative definitions of the mode volume (${\mathrm{V}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) and quality factor ($\mathrm{Q}$) to shed light on the previous statements. Since the effective volume (${\mathrm{V}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) depends on the electric field distribution ($\mathrm{E}(\overrightarrow{\mathrm{r}})$) and the relative permittivity of the medium ($\mathsf{\epsilon }(\overrightarrow{\mathrm{r}})$), in conventional dielectric cavities, it cannot fall below the order $~{{(\lambda }_{\mathrm{c}}/2n)}^3$. However, the effective volume can be approximated by the effective cavity length (${\mathrm{L}}_{\mathrm{e}\mathrm{f}\mathrm{f}}$) in the case of a planar configuration of the microcavities [280,281,282]:

$${\mathrm{L}}_{\mathrm{e}\mathrm{f}\mathrm{f}}={\mathrm{L}}_{\mathrm{M}\mathrm{C}}+2{{\mathrm{L}}_{\mathrm{D}\mathrm{B}\mathrm{R}}}^{\prime }\approx {\mathrm{L}}_{\mathrm{M}\mathrm{C}}+4({\mathrm{d}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}+{\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{w}}){\displaystyle \frac{n_{\mathrm{e}\mathrm{f}\mathrm{f}}}{\left|\mathsf{\Delta }n\right|}},$$

(5)

where ${\mathrm{L}}_{\mathrm{M}\mathrm{C}}$ is the length of the planar microcavity, ${{\mathrm{L}}_{\mathrm{D}\mathrm{B}\mathrm{R}}}^{\prime }$ is the length of the DBR, and ${\mathrm{d}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}$ and ${\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{w}}$ are the thicknesses of each layer of low- and high-index material, respectively. Additionally, $n_{\mathrm{e}\mathrm{f}\mathrm{f}}$ represents the effective refractive index of the DBR structure and can be expressed as

$$n_{\mathrm{e}\mathrm{f}\mathrm{f}}=\sqrt{{\displaystyle \frac{{\mathrm{d}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}{\left(n_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}\right)}^2}{{\mathrm{d}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}+{\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{w}}}}+{\displaystyle \frac{{\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{w}}{\left(n_{\mathrm{l}\mathrm{o}\mathrm{w}}\right)}^2}{{\mathrm{d}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}+{\mathrm{d}}_{\mathrm{l}\mathrm{o}\mathrm{w}}}}}$$

(6)

In this context, it is emphasised again that the use of high-dielectric-contrast material pairs in DBR mirrors is essential. By using high optical contrast materials, the reflectivity and bandwidth of mirrors are maximised, thus decreasing the field penetration length in the reflector stacks and resulting in a reduced effective optical length of the cavity. The result is a smaller mode volume and stronger field confinement (and, frequently, a higher $\mathrm{Q}$ factor). On the other hand, the quality factor ($\mathrm{Q}$) is a dimensionless parameter that can be calculated from the emission, transmission, or reflectance spectra of the cavity mode, which is characterised by the wavelength (${\lambda }_{\mathrm{c}}$) (or frequency ${\mathsf{\omega }}_{\mathrm{c}}$). That being said, the quality factor ($\mathrm{Q}$) can be determined from its spectral width (FWHM):

$$\mathrm{Q}={\displaystyle \frac{{\lambda }_{\mathrm{c}}}{\mathsf{\Delta }\lambda }}={\displaystyle \frac{{\mathsf{\omega }}_{\mathrm{c}}}{\mathsf{\Delta }\mathsf{\omega }}}={\mathsf{\omega }}_{\mathrm{c}}{\mathsf{\tau }}_{\mathrm{c}}$$

(7)

where ${\mathsf{\tau }}_{\mathrm{c}}$ is the photon lifetime in the cavity. Moreover, in the case of a DBR-based Fabry–Pérot structure, the quality factor can be expressed as follows [283,284,285]:

$$\mathrm{Q}={\mathrm{m}}_{\mathrm{e}\mathrm{f}\mathrm{f}}\left[{\displaystyle \frac{\mathsf{\pi }{\left({\mathrm{R}}_{\mathrm{D}\mathrm{R}\mathrm{B}1}{\mathrm{R}}_{\mathrm{D}\mathrm{B}\mathrm{R}2}\right)}^{\frac{1}{4}}}{1-\sqrt{{\mathrm{R}}_{\mathrm{D}\mathrm{R}\mathrm{B}1}{\mathrm{R}}_{\mathrm{D}\mathrm{B}\mathrm{R}2}}}}\right],$$

(8)

where ${\mathrm{m}}_{\mathrm{e}\mathrm{f}\mathrm{f}}={\mathrm{m}}_{\mathrm{c}}+{\displaystyle \frac{n_{\mathrm{l}\mathrm{o}\mathrm{w}}}{\mathsf{\Delta }n}}$ is the effective microcavity mode order. The last relation that links the quality factor to the effective cavity mode order was introduced to highlight the influence of the penetration depth (${\mathrm{L}}_{\mathrm{D}\mathrm{B}\mathrm{R}}$) of each Bragg mirror on $\mathrm{Q}$. In a realistic scenario, where absorption, scattering and coupling losses degrade the quality factor of the microcavity [286,287], both the materials and the preparation techniques used for Bragg stacks become decisive and strongly influence the value of $\mathrm{Q}$.

Considering the previous statements, below, we provide a table of different planar microcavities, especially based on oxide material, along with the quality factor and manufacturing method (

Table 8. Examples of planar microcavities for UV, visible and near-infrared emission.

Bottom DBR	Active/Passive Layer	Top DBR	Deposition Technique	$\mathbf{Q}$	References
SiO2/HfO2/Al	ZnO	SiO2/HfO2	Ion beam-assisted electron beam vacuum evaporation/polished substrate/ion beam-assisted electron beam vacuum evaporation	4250	[288]
SiO2/HfO2	HfO2	SiO2/HfO2	Ion beam-assisted electron-gun vacuum evaporation	3700	[246]
Al2O3/YSZ	ZnO	Al2O3/YSZ	Pulsed laser deposition—monolithic integration	1000	[120]
SiO2/Si3N4	ZnO	SiO2/Si3N4/AlN/AlGaN	Molecular beam epitaxy/molecular beam epitaxy/radio frequency plasma-enhanced chemical vapour deposition (RF-PECVD)	675	[118]
ZnO/ZnMgO	ZnO	ZnO/ZnMgO	Oxygen plasma-assisted molecular beam epitaxy	670	[289]
AlN/AlGaN	ZnO	SiO2/HfO2	Molecular beam epitaxy/molecular beam epitaxy/ion beam-assisted electron beam vacuum evaporation	650	[290]
SiO2/HfO2	ZnO	SiO2/HfO2	RF magnetron sputtering/pulsed laser deposition/rf magnetron sputtering	500	[291]
AlN/AlGaN	ZnO	SiO2/HfO2	Low-pressure metal–organic chemical vapour deposition/plasma assisted molecular beam epitaxy/electron beam evaporation	100	[115]
AlGaN//GaN	ZnO	SiO2/Si3N4	Low-pressure metal–organic chemical vapour deposition/plasma assisted molecular beam epitaxy/remote plasma-enhanced ultra-high vacuum chemical vapour deposition	100	[102]
SiO2/Ta2O5	Organics (CBP:BSBCz)	Ta2O5/SiO2	Radio frequency magnetron sputtering/vacuum sublimation/ radio frequency magnetron sputtering	1000	[292]
SiO2/TiO2	TiO2	SiO2/TiO2	Spin coating—monolithic Integration	35	[293]
SiO2/TiO2	Er3+dopedSiO2	SiO2/TiO2	RF sputtering technique—monolithic integration	890	[294]
SiO2/TiO2	Er3+dopedSiO2	SiO2/TiO2	RF sputtering technique—monolithic integration	171	[295]

). A passive layer is described as a transparent, non-emissive, gainless layer whose function is to control propagation (phase/amplitude) and reflectance with minimal loss.

4. Conclusions and Perspectives

In this review, the features and characteristics of highly reflective coatings, particularly Bragg mirrors, have been highlighted, along with the planar microcavities. The importance of planar cavities in future photonic devices (i.e., quantum sources of entangled photons) has also been addressed, although their architecture limits confinement effects. However, improvements in planar structures can be achieved by etching different micropillars, achieving 3D confinement and improving emitting mode coupling by suppressing leaky modes. From the perspective of materials, optical coatings cover a wide spectrum: from inorganic classes (oxides) to organic counterparts. As we have observed, the parameters of the mirror (reflectivity) and the cavity (i.e., quality factor, mode volume) are influenced by materials, techniques and ultimately architecture (in the case of the microresonator). This field of microcavities based on dielectric mirrors is evolving rapidly, and, to advance it, additional developments oriented towards scalable large-area processes, with high uniformity, minimal optical losses, and manufacturing costs compatible with industrial-scale integration, are essential.

In selecting the performance and industrial viability of oxide-based photonic devices, the manufacturing methods play a crucial role. Through PVD methods, the thin films obtained have a higher density with minimal scattering and can be obtained in a single manufacturing step, but at a higher cost. CVD is a robust alternative, especially for complex surfaces/substrates, but it is not suitable for high-temperature-sensitive materials. The solution-based methods are the “go to” methods at the moment, due to their cost efficiency for large areas at the expense of the poorer adhesion of thin films to the substrates or defect-induced solvents. Moreover, the importance of complex stoichiometric oxides is colossal, as they seem to be fundamental pillars for the development of modern photonics due to their superior electro-optical and nonlinear responses. Although the current state of the art is predominantly based on thin film-on-insulator platforms, the long-term perspective indicates a necessary paradigm shift towards the direct epitaxial growth of these films through deposition techniques. This transition from hybrid assembly to monolithic integration is imperative to ensure industrial scalability, cost reduction and compatibility with CMOS production flows on large wafers.

The presented materials throughout this article have a multifunctionality which is considered a significant “design bonus” in terms of developing devices that can act as a functional optical coating, as well as an active sensor. This can be seen in Micro-Opto-Electro-Mechanical Systems (MOEMS), which represent a complex miniaturised device that uses mechanical movement to manipulate light. By embedding lenses, mirrors or gratings with micro-actuators and electrical systems, MOEMS can switch, deflect or filter light beams with extreme precision and speed.

In this context, the critical challenge for future research is no longer just the deposition itself, but achieving a rigorous stoichiometric control and impeccable in situ crystalline quality, which would allow epitaxially deposited films to rival the optical performance of bulk monocrystalline substrates.