In the late 19^th century, Gabriel Lippmann, the 1908 Nobel laureate in physics, succeeded in recording volume holograms for the first time in human history by illuminating 1D interference on silver halide-sensitized gelatin film [1, 2]. Subsequently, in 1948 and 1962, Gabor and Denisyuk [3–5], respectively, advanced the optical systems for efficient recording and playback of volume holograms. Holograms, practically used in our daily lives, for example, in the anticounterfeiting film in banknotes, are based on the holograms of Gabor and Denisyuk. Traditional research on volume holograms focused on the efficient recording and reconstructing of 3D images. To this end, 1D volume gratings rather than 2D and 3D were primarily used in most cases, and the recordable materials were limited to silver halide-sensitized gelatin film. This research stream persisted for a time until there was another phase of research transformation from the 1980s to the 2010s.

With the discovery of the potential for volumetric holograms to store massive amounts of data [6, 7], optical systems for recording and reading more complex volume holograms (e.g. multiplexing of volume holograms in the same volume) were actively investigated. While advancements in modern semiconductor manufacturing technology (lithography) have made it common to record several terabytes (TB) within square centimeters of space, recording more than 1 gigabyte (GB), even with advanced data storage media such as compact disks (CDs), seemed like a pipe dream during the early stages of holographic data storage research. Furthermore, the materials for volume holographic recording, once limited to silver halide-sensitized gelatin film, have undergone extensive diversification into photorefractive polymers [8, 9], photopolymers [10–20], polymer-dispersed liquid crystals (PDLC) [21–23], and photoaddressable polymers [24–26]. These materials are portable, easy to manufacture, and capable of mass production; More importantly, they exhibit excellent diffraction efficiency (DE). At this point, it is important to note that the term recording is predominantly used for inscribing volume holograms. These recording mediums for volume holograms can reversibly or irreversibly reconfigure their molecular orientations and electron density with respect to the intensities and polarizations sinusoidally distributed along the grating vector of the incident interference pattern. Thus, the refractive index (RI) can be volumetrically modulated merely by the illumination of an interference pattern without a solvent development process (i.e. lithographic etching), and volume holograms can be inscribed even in a thick film (even up to a few millimeters). Thus, holographic recording of volume gratings has been used more prevalently than holographic lithography.

Has holography been limited to the recording and playback of volumetric gratings and 3D images? Not at all. Holography has had broader applications in terms of micro/nanomanufacturing. It has been widely used as an efficient method to reduce the period and half-pitch of line patterns in conventional photolithography, which has been a gold standard in conventional semiconductor manufacturing. This holographic line patterning has even found utility in cutting-edge extreme ultraviolet (EUV) lithography, considered state-of-the-art in photolithography. However, in this context, it is noteworthy that holography in conventional photolithography typically involves exposing thin layers of photoresist (PR) to the 1D interference patterns and subsequent solvent development to etch out selectively exposed or non-exposed areas [27–30]. Since it relies on solvent development, this holographic lithography may not qualify for generating volumetric gratings. This is due to the fact that the evaporation of solvent likely causes capillary-driven collapses of the patterned structure, which would be promoted particularly for a higher aspect ratio of structure. Also, volume gratings are not necessarily required in the semiconductor manufacturing process.

Beyond a thin PR layer, holographic lithography can be implemented as relatively thick PR films with thicknesses of at least several micrometers to form 3D volume gratings. In particular, the experimental realization of photonic crystals (PhCs) has been a significant topic in optics and photonics since the 1990s [31–44]. Holographic lithography was introduced as an efficient approach to creating 3D PhCs operating in the visible spectrum in 2000, by Turberfield and his colleagues [45]. This involved using an interference pattern resulting from the mixing of at least four beams on about 20 μm thick PR film, followed by a solvent development process [29]. Prior to this work, holographic lithographic fabrication of PhCs had largely remained limited to 1D or 2D structures [27, 28]. Subsequently, holographic-creating 3D PhCs became very active [29, 46–49]. In this context, it is important to emphasize that the method described above for holographic patterning of PR into distinctly binary structures (all or nothing according to the intensities of interference pattern) is referred to as lithography rather than recording.

What is the difference between holographic recording and holographic lithography? It would be difficult to distinguish them based solely on the fabrications of micro or mesoscale gratings. In particular, the possible resolution or critical dimension (CD) of gratings does not matter for this, because the CD of both holographic fabrications is generally defined by the half-pitch of the interference pattern. Generally, visible and ultraviolet lasers have been used as source beams of holographic fabrications. Therefore, the possible CD for holographic recording and lithography has ranged from 300 nm to a few micrometers. Despite this, these two terms have been used differently over the past half-century, as mentioned above. In this review, we aim to quantitatively compare the structural differences of surface or volumetric gratings that can be developed by holographic recording and holographic lithography, as well as the diffractive characteristics of these gratings achievable with each method. Then we discuss future research directions, including the transformative applications of both methods.

The prefix litho- refers to the act of carving stone [like silicon (Si) lumps] to create specific patterns or structures (i.e. topological features). Graphy refers to the act of drawing or depicting. Thus, lithography refers to the process of designing patterned images and using them as blueprints to carve out specific structures. Therefore, holographic lithography refers to creating microstructures by selectively leaving behind or carving out PR according to the bright or dark portions of holographic interference patterns. As a result, surface or volume gratings produced by holographic lithography exhibit binary structural features, which are similar to binary digital signals based on 0 and 1 (repeating the presence or absence of PR), as shown in Figs. 1(a)–1(c) [28, 45, 50]. Overall, holographic lithography can translate only the intensity information of a holographic pattern [i.e. intensity interference pattern (IIP)] [51, 52] into the photosensitive film (i.e. PR).

Figure 1. Surface and volume gratings fabricated by holographic lithography and holographic recording. (a) 1D line and (b) 2D quasi-crystalline surface gratings, which are respectively developed by 1D and 2D holographic lithography. (a) and (b) Reproduced with permission [28, 50] Copyright © 1999, American Institute of Physics and Copyright © 2007, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. (c) 3D volume gratings or photonic crystals (PhCs), developed by 3D holographic lithography. Reproduced with permission [45] Copyright © 2000, Macmillan Magazines Ltd. (d) Cross-sectional and (e) top-view of a transmissive 1D volume grating, developed by holographic recording on photopolymer (i.e. zirconium-based glass). Reproduced with permission [14] Copyright © 2006, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. (f) A reflective 1D volume grating, developed by holographic recording on photopolymer. Reproduced from [56] Copyright © 2016, Y. Montelongo et al.

In contrast, as mentioned earlier, holographic recording does not involve an etching or solvent development process [10–26]. Both intensity and polarization information that is sinusoidally varied along the grating vector of the holographic interference pattern can be directly encoded in the aforementioned holographic media as a sinusoidally modulated RI. When two beams with the same intensity and polarization interfere, the intensity is sinusoidally modulated along the grating vector (i.e. IIP). When this IIP is exposed to photopolymers, holographic photopolymerization (diffusion type) [10–15, 17, 18] or photorearrangement of molecular structure (non-diffusion type) [16, 19, 53] can be induced in proportion to the intensity, resulting in a sinusoidal variation of molecular density and consequently leading to a sinusoidal modulation of the RI. This holographic recording using IIP can also be achieved in photoaddressable polymers [24–26]. Photoaddressable polymers are typically composed of azobenzene molecules that can undergo photoisomerization (from trans- to cis-form) [24–26, 54, 55]. In particular, the long axis of azobenzene molecules can become aligned perpendicular to the polarization of the incident light; The degree of this alignment is correlated with the incident beam intensity. Thus, the polarization information of the interference pattern can be translated into photoaddressable polymers. With interfering beams with the same intensity but different polarizations, it is possible to create polarization interference patterns (PIP) [51, 52]. In PIP, the polarization is sinusoidally modulated, whereas the intensity is consistent. The illumination of PIP in photoaddressable polymers can drive the sinusoidally distributed orientation of azobenzene molecules and the resultant RI [24–26]. Thus, the gratings formed by holographic recording can exhibit distinctly different characteristics as compared to binary gratings formed by holographic lithography. In general, the term pattern implies the presence or absence of a unit structure regularly arrayed over a large area, whereas holographic recording creates unusual patterns with continuously changing material density or molecular alignment, in a sinusoidal manner [Figs. 1(d)–1(f)] [14, 56].

In this context, the most significant difference between holographic recording and holographic lithography in terms of a fabrication process lies in the presence or absence of a solvent development step. However, this difference does not only lead to structural disparities between binary and sinusoidal gratings as follows. When fabricating surface gratings using holographic lithography, such as 1D and 2D gratings, positive PR (like the AZ series) has mainly been employed. The molecular changes in the AZ series PR can be induced selectively in the bright areas of the interference pattern, resulting in a selective increase in the solubility of the developer [29]. However, this positive PR is likely coated into a thin film with a tens of nanometer thickness rather than thick films of several micrometers or more. Therefore, to produce volume gratings with holographic lithography, high-viscosity PR such as SU8, which are better suited for generating thick films, have been typically used [29]. The high viscosity of SU8 originates from its bulky molecular architectures consisting of eight to 12 epoxy groups and multiple benzene rings per molecule [29]. Selective proton release and induction of epoxy crosslinking can occur selectively in the bright regions of the interference pattern, preventing them from dissolving in the developer. Thereby, SU8 can serve as a negative PR. Mixing of more than three beams, steering it to a thick SU8 PR, and the subsequent solvent development process allows us to create 3D volume gratings (3D PhCs).

Unfortunately, however, several challenges lie in the holographic lithographic definition of SU8, which can compromise the structural fidelity of 3D volume gratings. Due to the relatively high epoxy content per molecule in SU8 (eight to 12 epoxy groups per molecule), a high cross-linking density is inevitable [29]. Generally, a higher cross-linking density leads to higher volume shrinkage, which exerts stress throughout the film and causes structural deformation [57–60]. It is known to result in volume shrinkage of up to 40% for SU8. When continuous channels such as 1D or 2D are produced, the issue of structural deformation due to volume shrinkage is relatively less pronounced [Fig. 1(a)]. However, when creating 3D cellular solids, this becomes a significant problem [Figs. 2(a)–2(c)] and severely restricts the allowable thickness of SU8. In fact, most holographic lithography has been performed on thicknesses of 15–20 μm or less [45–48, 50, 57–60]. Although some exceptional results from 3D holographic lithography on 30–40 μm thick SU8 have been reported recently [60], the structures experienced significant structural deformation, as evidenced by the extent of turning white to the naked eye [(Fig. 2(a)] [60]. While efforts have been made to improve the mechanical properties of thick PR by introducing rigid alternatives (e.g. organic-inorganic hybrid such as POSS) [17, 61] or enhancing solvent development (using supercritical fluids or low-surface-tension solvents such as isopropanol) [59, 62], the issue of volume shrinkage still remains unresolved. Although epoxy is less prone to volume shrinkage compared to acrylates (i.e. another commonly used molecular motif for PR), it still does not entirely mitigate volume shrinkage-driven structural deformation.

Figure 2. 3D volume gratings developed by 3D holographic lithography. (a) Macroscopic and (b) and (c) microscopic cross-sectional views of the holographically etched 3D volume gratings [SU8 photoresist (PR)]. Reproduced with permission [60] Copyright © 2023, Wiley-VCH. In particular, these structures are developed by metasurface-based raster and scanning illumination of a single beam. (d) Top, (e) and (f) cross-sectional microscale, and (g) top-macroscale views of the holographically etched 3D volume gratings (SU8 PR, which is structured into a woodpile lattice). Reproduced with permission [64] Copyright © 2019, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim.

Another practical issue with 3D holographic lithography is the amplification of structural heterogeneity due to thermal heterogeneity [57, 60, 63]. To promote the crosslinking of SU8, the diffusion of protons generated by light exposure needs to be accelerated [29]. Therefore, holographic lithography requires a post-baking of PR after exposing the holographic light pattern. Typically, post-baking is performed on a hotplate in most labs at university levels, leading to a temperature gradient from the substrate to the SU8 surface. This thermal gradient can spatially facilitate chemical heterogeneity (i.e. proton density), leading to the non-uniformity of the developed 3D patterns along the SU8 thickness.

An as-received SU8 from a commercial company is a form of solid dissolved in a solvent; thus, the solvent needs to be totally removed through a soft-baking process before the illumination of the holographic light pattern [29]. This is also typically carried out on a hotplate, making it prone to inducing residual solvent heterogeneity. Consequently, this can further amplify the non-uniformity of patterns formed by holographic lithography.

Due to the aforementioned practical issues, macroscopic images of 3D patterns produced by holographic lithography generally do not exhibit uniformity of the diffractive structural colors. This issue did not gain prominence until the mid-2010s. Research papers published up until the mid-2010s, following the initial reports of 3D holographic lithography PhC fabrication in 2000, typically did not include macroscopic images of the formed patterns [29, 45–48, 50, 57–63]. Surprisingly, the merely reflective spectral results, which seemed to be measured from the well-developed local area of holographically formed 3D volume gratings, had been included in the literature. In other words, the pattern uniformity of these 3D volume gratings was not quantitatively analyzed over a large area. As summarized in Figs. 2(d)–2(g) [64], it was only after the mid-2010s that macroscopic images of 3D patterns produced by holographic lithography began to appear in the literature, often showing mottled rather than uniform structural colors [64], or even appearing white, as mentioned earlier [60]. This indicates that the non-uniformity of patterns produced by holographic lithography is quite severe on a macroscopic scale.

What about a holographic recording? As mentioned previously, holographic recording eliminates the need for a solvent development process, thus preventing structural deformation caused by capillary effects. Furthermore, both photopolymers and photoaddressable polymers avoid thermal processes such as baking during grating formation, which reduces structural heterogeneity related to thermal gradients. It is worth noting that some volume holographic materials required solvent development and thermal treatment for hologram recording in exceptional cases [14, 17, 65–67]. However, these less-controllable processes adversely affected the reproducibility of holographic recording on such materials. Consequently, photopolymers and photoaddressable polymers, which do not necessitate solvent development and thermal treatment, are now widely employed for recording volume gratings [11, 16–18, 24–26]. Importantly, since both photopolymers and photoaddressable polymers typically use a polymeric matrix, they exhibit considerable resistance to volume shrinkage. It is well known that a diffusion-type photopolymer is more prone to volume shrinkage than a non-diffusion-counterpart [10–20]. Nevertheless, the volume shrinkage even for a diffusion-type holographic recording is far below (less than 1%) that for holographic lithography (~40%). As a result, holographic recording can be successfully performed on very thick films exceeding 100 µm in thickness [see Fig. 1(d)] [14, 16].

However, there have been relatively few reports of 3D volumetric gratings created with holographic recording in the last three decades. Instead, the predominant practice has been to record 1D interference patterns in a transmissive or reflective manner [see Figs. 1(d)–1(f)]. This tendency may stem from the fact that volume gratings have mainly been used for recording 3D holographic images or data storage applications, rather than applications related to PhCs and the related structural colorization. This distinction is one reason why holographically recorded volume gratings are sometimes referred to separately as holographic optical elements (HOEs). In general, volume gratings created with holographic lithography are not typically labeled as HOEs. Nevertheless, both the holographically etched 3D PhCs and HOEs can be simultaneously viewed as volume gratings for unidirectional diffractive light bending. As a result, we emphasize the need to broaden our perspective and unify both approaches.

Despite their distinct structural motifs, both the holographically etched and recorded volume gratings function as diffractive optical components, as mentioned earlier (binary profile versus sinusoidally modulated profile). This leads to the question: What are the differences in their diffractive optical properties?

Diffraction from a grating can be intuitively quantitated by Fourier optics [26–49]. From the point of view of Fourier optics, diffraction originates from momentum kicks stemming from the waves of matter (i.e. gratings). The wave nature of a grating is represented by the grating vector, which defines the possible momentum kick for incoming light and the resultant diffraction. To form a binarily modulated grating, several sinusoidal waves with varying wavelengths (i.e. varying grating vectors) need to be mixed accordingly [see Fig. 3(a)] [26]. Thus, these mixed waves of matter in a binary grating should each put a different momentum on the incoming light independently, which consequently leads to several different diffractions [Fig. 3(b)] [26]. In other words, unwanted frequency mixing of diffraction is inevitable for binary gratings. By contrast, a sinusoidally modulated grating can drive a single momentum kick (singular grating vector) of the incoming light [Fig. 3(a)], resulting in diffraction with only the desired frequency [Fig. 3(c)]. The importance of sinusoidally modulated RI in surface and volume gratings was recently elucidated by the concepts optical Fourier surface (OFS) [52, 68–79] and optical Fourier volume (OFV) [26].

Figure 3. Fourier optics for diffractive volume gratings (1D). (a) Fourier synthesis of a binary and sinusoidal grating. Dispersion relation of (b) binary and (c) sinusoidal 1D volume gratings. These results are based on numerical calculations. (d) and (e) The numerically predicted effect of the thickness of volume gratings on a diffraction (top panel): Top is schematic for strengthened and weaken k-vector clouding, for a relative thin and thick volume grating. The experimentally exploited diffractive dispersion as a function of thickness of 1D volume gratings (bottom panel), holographically recorded on photopolymers (bottom left) and photoaddressable polymers (bottom right). All above contents are reproduced with permission [26] Copyright © 2022, Wiley-VCH.

In the same manner, not only the sinusoidal variation of the IR within a volumetric space but also the thickness of volume gratings is important in Fourier optics [Figs. 3(d) and 3(e)] [26, 49]. The significance of this thickness has not been emphasized in the previously reported manufacturing of 3D PhCs. However, with the recent establishment of the OFV concept, it has been confirmed that even when the RI of the volume grating varies sinusoidally, the loss of Fourier optical characteristics can occur if the thickness of the volume grating is relatively thin. To maintain the characteristics of OFV, the sinusoidally modulated RI variation needs to persist along the vertical direction [i.e. along the thickness of PhCs (z-axis)] as well as the lateral directions (x- and y-axes). In other words, a grating vector along the z-axis can converge into a singular point for a thicker volume grating. The limited thickness of the volume grating leads to a distributed grating vector and the resultant k-vector clouding effects for volume diffraction. Note that the importance of thickness is valid only for volume gratings rather than surface gratings.

In addition to the spatial profile of RI, the RI contrast across the grating vector is another key to the activation of a desired frequency diffraction [26]. In general, the achievement of higher contrast of RI along the grating vector (also called dynamic range in Kogelnik’s coupled wave theory)[80] has been believed to be a the grail in developing PhCs since it can widen the photonic bandgap of PhCs [81–83]. This is the reason why a diamond lattice has been considered a champion PhC. A diamond lattice can be built with a lower volume fraction than others, which in turn can render it with a higher RI. The lower limit in the required RI contrast to achieve a complete photonic bandgap is the lowest for diamond lattices among 3D PhCs, which can lower the bar for the manufacturing process and material palette. However, as shown in Figs. 4(a) and 4(b), an increase in RI contrast beyond a critical regime (i.e. order of 10⁻¹) could give rise to the degeneracy of diffraction resulting from the over-accumulated phase in the volume grating [26]. This range of RI contrast is generally attained from holographically etched PhCs (across air and material). Also, the over-accumulated phase can activate the higher-order diffraction. Overall, even with a sinusoidally modulated grating, RI contrast across the grating vector should be not too high or too little for a desired single-frequency diffraction (i.e. order of 10⁻⁴). Otherwise, deteriorated k-vector clouding is inevitable.

Figure 4. Fourier optics for exploiting the effect of refractive index contrast (dynamic range, ∆n) on the dispersive behavior of diffraction (angular selectivity of diffraction). (a) Achievable diffraction efficiency of 1D volume gratings with varying ∆n and thickness (d). (b) Angular selectivity of diffraction efficiency of 1D volume gratings with varying ∆n and thickness. All above contents are reproduced with permission [26] Copyright © 2022, Wiley-VCH.

In summary, the following conditionals are required for gratings from a Fourier optical perspective: Firstly, the variation of the RI must be induced to be sinusoidal within a thick volume of the gratings, along the grating vector, in a highly uniform manner. Simultaneously, the RI contrast should be at a moderate level, typically in the order of 10⁻⁴. The only known method that can satisfy these criteria to produce such OFVs is holographic recording [26], and it is not achievable with holographic lithography, as mentioned above.

However, from a scalable manufacturing perspective, traditional holographic recording can have the following drawbacks. Holographic recording typically involves the interfering of two laser beams and steering the resultant interference pattern onto a photopolymer or photoaddressable polymer. Expanding an as-launched coherent laser beam from a few millimeters to even a centimeter scale is practically challenging and dependent on highly trained personnel (expanding a laser beam by using a spatial filter is highly sensitive for the personnel). Additionally, the consistent operation and maintenance of the coherency of the laser beam can be error-prone in ambient experimental conditions such as mechanical vibration and humidity.

In 2022, Mathias Kolle et al. [84] reported that this practical challenge of holographic recording can be effectively addressed by a standard desktop projector-assisted recording of the Lippman hologram (see Fig. 5). A merely projector-sourced illumination of a scalable light pattern onto the photopolymer-mirror stacks can record an OFV along about 16 μm thick volumetric film (i.e. sinusoidally modulated 1D Bragg stacks). As shown in Fig. 6 [84, 85], inch-scale diffractive color graphics, encoded by scalable holographic recording, were obviously vivid without mottled colors, common to the holographically etched [64] or self-assembled volume gratings (e.g. colloidal opals [86, 87] and block copolymer stacks [88, 89]).

Figure 5. A standard desktop projector-assisted scalable recording of 1D optical Fourier volume (i.e. holographically recorded reflective 1D volume gratings). Reproduced from [84] Copyright © 2022, B. H. Miller et al., under exclusive licence to Springer Nature Limited.

Figure 6. Inch-scale diffractive color graphics, encoded by scalable holographic recording. (a)–(g) Representative macroscopic views of holographically recorded reflective 1D volume gratings [84, 85]. (a)–(d) Reproduced with permission [85] Copyright © 2022, American Institute of Physics. (e)–(g) Reproduced from [84] Copyright © 2022, B. H. Miller et al., under exclusive licence to Springer Nature Limited.

Of course, in the case of holographically etched volume gratings, macroscopic colorization has mainly been reported for 3D lattice (e.g. woodpile) [14, 17, 64–67], whereas the scalable holographic recording-counterpart mentioned above has only been reported for 1D volume gratings [10–20]. Moreover, when compared to 2D or 3D volume gratings, 1D volume gratings inherently exhibit a more uniform and robust Fourier potential along a specific axis, resulting in more vivid colors at a fixed view angle. Therefore, it can be challenging to directly compare the structural colorizations between holographically etched and holographically recorded volume gratings. Nevertheless, from the aforementioned comparison [between Figs. 2(g) and 6], it can still be concluded that the uniformity of the color graphics remains better in holographic recording compared to holographic lithography because the holographic recording does not involve processes that compromise structural fidelity, such as solvent development and thermal treatment. The commercially available photopolymer, mainly from Bayer Material Science LLC (currently, Covestro AG with the product name Bayfol HX) [84, 90], is limited to a thickness of 16 μm; The reported experiments on holographic recording have mainly been conducted at this thickness. This thickness is similar to that of 3D volume gratings that can be produced by holographic lithographic etching of SU8 [45, 57–59]. However, custom photopolymers can be readily manufactured with thicknesses exceeding 100 μm [16, 26, 91], and they have consistently recorded volume gratings along the film thickness. Indeed, it turned out that a thicker volume grating (1D) enables a narrow distribution of k-vector at the far-field diffraction point [see Figs. 3(d) and 3(e) and Fig. 4]. However, this systematic analysis of the effect of thickness on the volumetric diffraction was restricted to the transmissive grating, and not appropriate for the application of a far-field structural colorant (only iridescent colors are visible). Even though Kolle et al. [84] theoretically elucidated that the thicker reflective-type 1D OFVs is better suited for encoding more color graphics, its experimental validation is yet to be exploited.

In the last part of this section, we would like to mention that the manufacturing requirements for OFSs are less stringent compared to OFVs because they only need to satisfy the condition of sinusoidal variation in topological RI. Several manufacturing processes have already been proposed for OFS production, such as scanning probe lithography [76] and holographic inscription [68–75, 77–79] (i.e. a method using holographically guided, directional mass migration in azobenzene materials (called directional photofluidization) [92–96]. The latter is also referred to as directional photofluidization lithography [54, 97–100]. Additionally, although not yet reported, options such as over-dose photolithography or gray-scale electron beam lithography can potentially provide a sinusoidally modulated topological profile of a surface pattern.

Until now, the discussion of volume gratings has mainly been from the perspective of diffractive optical materials. In particular, we focused on the fact that holographic recording from a Fourier optics-based diffractive perspective can give exotic optical properties to volume grating that holographic lithography cannot provide. However, periodic patterns based on volume grating can find applications in a wide range of fields beyond diffractive optical materials.

For instance, precisely controlled 3D porous structures can potentially revolutionize fields spanning battery technology (solid-state electrolytes) [101, 102], photocatalytic materials [103], phononic materials (controlling sound and heat propagations) [29, 104, 105], and mechanical metamaterials [60]. From an application perspective, holographic recording may have limited potential compared to holographic lithography, stemming from the fact that holographic recording is not well-suited for fabricating regularly porous structures, as previously mentioned. Indeed, the immediate practical applications of holographic lithography are rapidly growing from initially limited PhCs to battery electrodes [101, 102] and photocatalytic materials [103]. This considerable progress has benefitted from recent advances in converting holographically etched 3D PR structures into a vast variety of material counterparts, including metal oxides, carbon, and metals, as summarized in Fig. 7.

Figure 7. Metallic 3D volume gratings templated by holographically etched 3D volume gratings. (a) Microscopic perspective views of the MnO₂ and Mi-Sn 3D structures. Reproduced with permission [101] Copyright © 2015, National Academy of Sciences of the United States of America. (b) Templating of 3D porous gold (Au) structures from holographically etched 3D volume gratings. Reproduced from [103] Copyright © 2020, National Academy of Sciences of the United States of America.

What we want to emphasize here is that like holographic recording, holographic lithography has also evolved to be more scalable and robust in its processing. Coherent and long-distance steering of more than three beams directly launched from lasers to form a 3D interference pattern and their stable/reliable operation is practically challenging [60, 106, 107]. More critically, this direct mixing of multiple beams is not compatible with large-area manufacturing. To overcome technical restrictions, illuminating a single beam on a large-area phase mask (i.e. metasurface and diffractive surface gratings) for holographic lithography has been proposed (Fig. 8) [60, 106, 107], which greatly improves the scalability and reliability of holographic lithography. This phase mask-based holographic lithography has furthered the expansion of its application and design spaces.

Figure 8. Representative strategies for scalable holographic lithography. (a) A schematic for raster and scanning holographic lithography with the assistance of a meta-mask. Reproduced with permission [107] Copyright © 2019, S. M. Kamali, et al. (b) A schematic for proximity nanopatterning (PnP) with the assistance of a diffractive phase mask. Reproduced with permission [106] Copyright © 2004, National Academy of Sciences.

In contrast, the materials and optical processing of holographic recording are still limited to the creation of diffractive volume gratings. Although the projector-assisted scalable recording of volume holograms was recently proposed [84], the substrate material for holographic recording is much narrower than that of holographic lithography. Epoxy or acrylate-based organic materials [10, 11, 18, 91], or an organic material-dispersed ceramic matrix such as zirconium oxide and silica [14, 67], can be considered as typical holographic photopolymers. Photoaddressable polymers [24–26], as mentioned earlier, are mainly based on photochromic molecules such as azobenzene; Liquid crystal (LC) molecules or polymers can be co-mixed with azobenzene molecules. There is only one commercially available holographic recording material (the photopolymer product Bayfol HX) [84, 90], as mentioned above. These relatively limited substrates for holographic recording medium originate from the restricted purpose of holographic recording (i.e. HOEs).

Several researchers have attempted to induce counter-diffusion during holographic recording to rearrange the homogeneously dispersed nanoparticles (titanium dioxide, ZrO, and gold) in photopolymer into a regularly distributed composite form to contribute to the diversity of structures achievable with holographic recording [14, 15, 108]. During the holographic recording on photopolymer, these nanoparticles which lack photosensitivity, can inversely diffuse with respect to the diffusion direction of photosensitive monomers (e.g. acrylates and epoxy) [14, 15, 108, 109], leading to directional phase separation from bright regions to dark regions of IIP, as shown in Figs. 9(a)–9(c). However, this method was oriented to enlarge the dynamic range of HOEs rather than develop 3D functional structures. Using this counter diffusion of holographic recording, PDLC has been developed where LC molecules are accumulated selectively within the dark regions of IIP to provide an electrically tunable HOE [21–23]. The orientation of LCs within dark regions of IIP is tuned according to the external voltage; Thus, the dynamic range of HOEs can be electrically reconfigured. Also, after the selective dissolution of LC molecules from holographically photopolymerized volume gratings, a regularly ordered porous structure can be attained [Figs. 9(d) and 9(e)] [21, 22]. To the best of our knowledge, this is the only example of the holographic recording of 3D volume gratings. Even this example was published a long time ago, in 2002–2003, and there have been no other examples since then.

Figure 9. 3D composite structures developed by holographic recording. (a)–(c) Macroscopic and microscopic views of volumetrically assembled titanium nanoparticles fabricated by 1D holographic recording-induced counter diffusion. Reproduced with permission [13] Copyright © 2005, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. (d) and (e) Microscopic views of 3D porous structures (binary volume gratings) developed by 3D holographic recording on polymer-dispersed liquid crystals (PDLC) and subsequent selective removal of the counter-diffused LC molecules. Reproduced with permission [21] Copyright © 2002, WILEY‐VCH Verlag GmbH, Weinheim, Fed. Rep. of Germany.

To widen the scope of holographic recording, particularly toward commodity manufacturing of 3D structures, counter diffusion-based 3D structuring needs to be more generalized in terms of optical processing and available material libraries. Except for the recent suggestion by Kolle et al. [84] (using a desktop projector), holographic recording using two-beam mixing is still prevalent and profoundly suffers from limitations such as low scalability and limited lattice and dimension. Inspired by the recent progress of a scalable holographic lithography, there is a need for the development of single beam-accessible phase mask holographic recording to make it more compatible with ultra-scale 3D lattice engineering of super-depth patterns, which could be widely used for imaging models of scattering medium [110], optical metamaterials [111–115], and PhCs. Along with this direction, the photosensitive inks for on-demand holographic recordings are yet to be exploited.

National Research Foundation (NRF) of Korea grant (NRF-2022M3H4A1A02074314 and NRF-RS-2023-00272363); Samsung Research Funding & Incubation Center for Future Technology grant (SRFC-MA2301-02); The KIST Institutional Program (Project No.: 2V09840-23-P023); Korea University grant.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

No data were generated or analyzed in this study.