1. Introduction

The transition from incandescent lighting to the use of LEDs is expected to lead to significant energy savings [1–6]. However, despite this, all forms of artificial lighting are major contributors to environmental light pollution due to scattering losses and light not being directed to where it is needed. This in turn leads to the use of higher illumination intensities, thus putting additional strain on energy generation systems. In addition, light pollution has a detrimental impact on the health of humans [7–9] and wildlife [7–14]. As such, several countries and regions have already enacted legislation to address the impact of artificial lighting on the outdoor environment [15]. A potential technological solution to this is to use DOEs, also known as holographic optical elements (HOEs), to redirect the light to only where it is needed. However, these currently do not operate efficiently with non-laser sources such as LEDs. DOEs can be made at relatively low cost, with low intensity exposures, and can achieve nearly 100% diffraction efficiency. DOEs can be created in the form of volume holographic gratings (VHGs). These are DOEs which are recorded as sinusoidal variations in the refractive index of a material upon exposure to an optical interference pattern. In effect, these DOEs operate as Bragg mirrors and thus can be used to redirect light. However, at present they can only work efficiently across narrow angular and spectral ranges thereby limiting their application to single wavelength sources at highly specific angles of incidence, which is desirable in applications such as holographic data storage.

Therefore, in order to make VHGs viable as DOEs that are compatible with broadband LED sources, it is necessary to increase the angular and spectral working ranges. This can be achieved by using thin DOEs with high $\Delta n$. The refractive index modulation, $\Delta n$, is the maximum contrast, or dynamic range, between the peak refractive index at the bright fringes and the average refractive index of the background material. Reduced thickness, T, is beneficial as the angular working range is proportional to the ratio of the grating period, $\Lambda$, to T. However, if $\Delta n$ is sub-optimal for any given material, any reduction in T requires sacrificing diffraction efficiency, $\eta$, i.e. the ratio of the first order diffracted beam intensity to the sum of first and zero order beam intensities. This is the case since the phase accumulation, $\nu$, is proportional to the product of T and $\Delta n$. Only when the phase accumulation for a particular wavelength has a value of $\pi /2$ can the optimal diffraction efficiency be reached. Nonetheless, in order to prevent production of multiple diffraction orders at a specific grating period, care should be taken not to excessively reduce the thickness for materials with higher $\Delta n$. This can be a particular issue with transmission DOEs and longer wavelengths as the grating period tends to be larger.

Another important consideration is the range of grating periods used. These can also be thought of in terms spatial frequencies (i.e. the number of periods (lines) per millimetre). Indeed, there is a range of grating periods within which transmission elements typically operate. In short, the challenge faced is to develop suitable photosensitive materials capable of high $\Delta n$ in thin films that can operate within a relevant range of grating periods without demonstrating more than one diffraction order. For complex DOEs that contain a range of spatial frequencies careful design is needed to ensure that the whole element operates in the volume regime and that the light is diffracted into a single order. This is a potential challenge in transmission-mode VHGs containing very low spatial frequency gratings since these are more likely to produce diffraction into higher orders. Continuing with the theme of sustainability, such a material should also be: relatively low cost, have minimal environmental impact, and robust. For these reasons photosensitive polymers, or photopolymers (PPs), are of interest as they can be made from low cost, environmentally friendly materials, and have well understood chemistry.

The aim of this review is threefold: (i) to provide the scientific community with an update on recent developments in PPs for recording VHGs; (ii) to identify strategies which could potentially be used to generate a novel PP with enhanced $\Delta n$; and (iii) to identify the range of $\Delta n$ needed for transmission DOEs with single diffracted beams close to 100% diffraction efficiency. To the knowledge of the authors, there has been no review published on materials for VHGs since 2018 [16], yet numerous publications have emerged since that time. Prior to that two other reviews were published in 2016 [17,18]. This review has a particular focus on materials with properties suitable for high-efficiency transmission DOEs and it begins with an overview of modelling relevant to the design of materials for recording VHGs which is also used to inform the selection of materials chosen for review in the following section. This is then followed by a discussion which aims to refine the range of possible materials for fabrication of a low-cost material with high $\Delta n$, high $\Delta \theta$, and high $\Delta \lambda$.

2. Theoretical modelling of the specific requirements needed by DOEs for operation with LEDs

For efficient operation of DOEs with LEDs several criteria must be met, the details of which will be laid out in the following sections. These criteria are:

The remainder of this section explains the definition of a volume holographic grating, the advantages of using volume holographic gratings for DOEs, and how to satisfy the criteria listed above, for transmission gratings.

2.1 Volume holographic gratings

Volume holographic gratings, or Bragg gratings, are recordings of optical interference patterns in the form of periodic modulations of the density or refractive index of the transparent recording medium. A characteristic feature of such gratings is that they show only one diffractive order for a given input angle as predicted by the Bragg equation:

Ideal Bragg conditions for maximum energy in the diffracted order occur when eqn. (1) is satisfied. The associated diffraction efficiencies at and close to the ideal Bragg conditions, $\eta$, are described by the Kogelnik Coupled Wave Theory [19] as follows:

The parameter $\Delta \theta$ is the Bragg angle de-tuning and quantifies the angular deviation from the Bragg angle. It should be noted that some authors will refer to the max-min range of the refractive index contrast within a material as $\Delta n$, however, in this paper the refractive index modulation ($\Delta n$) means the peak to mean refractive index contrast. The range of $\Delta \theta$ corresponding to the full width at half maximum (FWHM) of the 1$^{st}$-order diffraction peak is often used to quantify the angular working range at the probing wavelength. For high diffraction efficiency gratings the angular working range can be expressed as:

At the ideal Bragg conditions, $\xi = 0$. Therefore the equation for $\eta$ simplifies to:

Since Eq. (9) defines the diffraction efficiency at the Bragg angle, the phase accumulation parameter can be thought of as a measure of grating strength. The value of $\eta = 1$ is achieved when $\nu = \pi /2$ and, theoretically, for all $\nu (j) = (2j + 1)\pi /2$ where $j = 0,1,2,3\cdots$. In practice however, for phase modulations where $\nu > \pi /2$, overmodulation occurs. This can lead to reduced peak efficiency at the Bragg angle and increased amplitude of side-lobes. It follows that for a Bragg grating with a value of $\nu \leq \pi /2$, any reduction in thickness, $T$, ensuring larger angular working range, without a corresponding increase in $\Delta n$ will lead to a deviation from optimal grating conditions and the peak value of $\eta$ will be less than one.

2.2 Modelling the boundary conditions for DOEs for operation as volume holograms with LEDs and the case for higher $\Delta n$ when designing for LEDs

So far only the definition of Bragg conditions and the diffraction efficiency of transmission volume holograms has been discussed. This section explains the limits of the conditions, i.e.: the conditions for which the DOEs can no longer be considered Bragg gratings and multiple diffraction orders appear (Raman-Nath). Whether a grating is considered a volume grating or a thin grating is determined quantitatively by two factors introduced by Cook-Klein [20] and Moharam-Young [21]. These are referred to as the Q-factor and $\rho$-factor respectively. The Q-factor is defined as:

The Q-factor was originally derived through studies of the diffraction of light by ultrasonic waves [20]. Conventionally it is used used as a convenient mathematical tool for classifying gratings as thin (Raman-Nath) regime gratings or as thick (Bragg or volume) regime gratings. The physical distinction between these regimes being the number of diffraction orders visible for a given angle of incidence. Under the thin regime multiple diffracted beams are visible. Under the thick regime a maximum of one diffracted beam (the 1$^{st}$-order) is available. When only the Q-factor is considered, gratings are typically defined as thick if $Q \geq 10$ or thin if $Q \leq 1$. However, despite being a very practical tool, a fundamental limitation affecting the generality of Q, in its derivation, was the assumption that $\nu < 6$ [21] In fact, it was shown that Q can fail to accurately classify gratings when $\nu$ is as low as 3 [21]. In addition, it was shown theoretically, that for large $\Delta n$, the Raman-Nath regime can exist even when $Q \gg 10$ [21]. The converse is also true. For sufficiently small $\Delta n$ the Bragg regime can occur despite $Q < 10$. As mentioned by Moharam-Young [21], other authors have come to similar conclusions based on theory [22,23]. In addition, there are experimental studies which support this conclusion [23–25]. For example, eight diffraction orders were observed for a hologram for which $Q = 55$ [21]. Therefore, in order reconcile the discrepancy between Q and experimental observations the $\rho$-factor was derived. The advantage of $\rho$ is that it considers $\Delta n$ independently of the physical thickness $T$ of the recording medium. Moharam-Young [21] derived a relation between Q and $\rho$ as follows:

Later, Gaylord et al. [26] provided more rigorous definitions for Raman-Nath (thin) and Bragg (thick) regimes based on a combination of these parameters. In their work a variation of Q, $Q' = Q/\cos {\theta _{B}}$, was combined with $\nu$ to give

Thus $Q'\nu$ along with $\rho$ can be used together to define limits of the Raman-Nath and Bragg regimes. The conditions for both regimes in terms of $Q'\nu$ and $\rho$ are summarised in Table 1. For each regime the corresponding limits must be satisfied simultaneously in order to be either in the Raman-Nath or Bragg regimes. Similar conclusions were also reached in work by Vita et al. [27] which clearly illustrated the parameters which define Raman-Nath and Bragg gratings.

Table 1. Conditions for thin (Raman-Nath), and thick (Bragg or volume) holograms based on Q-factor and $\rho$-factor.

View Table | View all tables in this article

Figure 1 illustrates Eq. (3), diffraction efficiency is plotted against the Bragg de-tuning angle for various values of grating thickness and $\Delta n$. Both plots correspond to theoretical modelling of gratings for which $SF = 800 l/mm$, $\lambda = 633 nm$, and $n = 1.5$. The peak $\eta$ increases with both thickness and $\Delta n$ until the optimal phase parameter is obtained. It can also be seen from Fig. 1(a) that the full width at half maximum of these curves, $\Delta \theta _{FWHM}$, decreases with thickness despite gains in peak diffraction efficiency. The green curve is the same in both (a) and (b). It is the Bragg curve for a grating with thickness, $T = 50\;\mathrm{\mu}\textrm {m}$ and $\Delta n$ corresponds to $\nu = \pi /2$; i.e. $\Delta n = 0.006$. Also shown is the effect of overmodulation. For the red curve the value of $\Delta n$ is sufficiently large that $\nu > \pi /2$.

 

Fig. 1. Illustration of Eq. (3), plotted against the Bragg de-tuning angle for a range of values for thickness (a) and $\Delta n$ (b). The green curve is the same in both (a) and (b). The orange curves in (a) and (b) correspond to a gratings for which, respectively, the thickness and refractive index modulation is half that of the green curve. The red curves in both (a) and (b) show the affects of overmodulation.

Download Full Size | PDF

In Fig. 2 the thicknesses (assuming optimum phase parameter) are plotted as a function of SF for the threshold values of $\rho$ shown in Table 1, using the following equation based on Eq. (11) and Eq. (7):

 

Fig. 2. 3D stem plot of $\Delta n$ vs thickness and grating spatial frequency along with a 2D projection of $\Delta n$ vs SF. The plotted values of $\Delta n$, and their 2D projections, represent conditions where $\rho = 10$ and $\nu = \pi /2$.

Download Full Size | PDF

In Fig. 2, Eq. (7) (or, equivalently, Eq. (6) where $\nu = \pi /2$) is used to plot for the values of $\Delta n$ using the values of thickness corresponding to $\rho = 10$ as input.

As discussed above for any particular grating there is an upper limit on the range of $\Delta n$ for which the grating can be considered a volume hologram. Figure 3 illustrates these limits as a function of $SF$ for a probe wavelength of 633 nm. It can be seen that, for high enough values of $\Delta n$, gratings can be considered to be outside of the Bragg regime and potentially lose light to higher diffraction orders. The refractive index modulations were calculated and plotted using Eq. (15) for the threshold values of $\rho$, below which the grating can no longer be considered thick:

 

Fig. 3. Refractive index modulations, plotted as function of SF using Eq. (15), where $\lambda = 633$ nm, $n = 1.5$ and $\nu = \pi /2$. The Bragg angles were calculated according to Eq. (1). The green and red lines correspond to the limit values of $\rho$. Based on $\rho$, a grating is considered a volume hologram if $\Delta n$ is on or below the green line, a thin hologram if it is on or above the red line, and an intermediate grating if it lies between these two lines.

Download Full Size | PDF

It is clear that $\Delta n$ should not be increased indefinitely in order to maintain being a volume grating which is operational as a DOE for LEDs, particularly at lower spatial frequencies. For the curve $\Delta n( \rho = 2/\pi ^2)$ the value of $\rho$ was derived from the definition of a Raman-Nath (thin) grating: $Q'\nu \leq ~1$. This is equivalent to $2\nu ^2 \rho \leq ~1$, which by solving for $\rho$ becomes:

Thus by setting $\nu =\pi /2$, this inequality could then be used to define the threshold value of $\rho$ as $\rho = 2/\pi ^2$, which was then used with Eq. (7) to generate the curve $\Delta n( \rho = 2/\pi ^2)$ depicted in Fig. 3.

2.3 Material restraints on $\Delta n$, the formula limit concept

In addition to theoretical limits on $\Delta n$, based on Kogelnik’s theory, there are practical limits based on the materials chosen to construct the DOEs. Based on the refractive indices of the material components, the maximum achievable $\Delta n$ of a material can be estimated using the formula limit concept [17]. The formula limit is derived from the Lorentz-Lorenz equation [28,29]. The formula limit, for two component materials, can be expressed as follows:

The first equation, Eq. (17), describes the formula limit in terms of the difference between the refractive index of the mixture of materials ($n_{mix}$) and the refractive index of the, usually inert, background material ($n_{background}$). The background may consist of a matrix material consisting of a thermoset polymer or a refractive index contrasting species such as nanoparticles (NPs). Equation (18) describes the formula limit in terms of the product of the difference between the refractive index of the writing polymer ($n_{polymer}$) and the background ($n_{background}$) and the volume fraction of the writing polymer in the bright fringe. This formula assumes perfect segregation of two incompressible component species: (i) writing polymer and (ii) background components. It is considered accurate when the volume fraction of the writing polymer is low (<10 vol %) and the refractive index contrast is small (<0.1) [17,30]. In practice however, the achieved $\Delta n$ is some fraction of the formula limit, known as the usable fraction [17]. This reflects the fact that not all of the writing monomer is converted into polymer chains that contribute to holographic effects. This imperfect conversion of the recording light pattern into a sinusoidal modulation of refractive index is a result reaction and diffusivity kinetics, which generally occur on overlapping timescales [17]. In addition , the loss of $\Delta n$ due to excess diffusion of photopolymerised components, from bright to dark regions, scales with $1/\Lambda ^{2}$ [17,30,31]. Furthermore, Eq. (18) has been further adapted to include a ‘degree of segregation’ (SD) factor [32]:

Thus for a given volume fraction of writing monomer the maximum $\Delta n$ is determined by (i) the maximum theoretical index contrast between writing monomer(s) and the background and by (ii) the SD of writing monomer(s) and background components. The latter is intimately linked to the reaction/diffusion kinetics of the system during hologram recording [17,30,32–38]. That is, the ratio ($K$) of diffusion rate to the polymerisation (reaction) rate for different ranges of $\Lambda$ [32,38]. For large values of $K$ ($K>1$), effects of diffusion on $\Delta n$ dominate over those due to polymerisation. This tends to be the case when small values of $\Lambda$ or low recording intensities are used. In this scenario, since the distance to be travelled by monomers is small, diffusional blurring occurs. For small values of $K$ ($K<1$), the inverse is true and effects of polymerisation dominate. This tends to be the case when large $\Lambda$ or high intensity recording is used. In this scenario, because of the relatively large distances monomers in dark fringes must have travelled in order to reach reaction sites, a large proportion fail to polymerise by the time polymerisation has terminated (gel point reached); after which point there is a sharp increase in the viscosity of the system [32]. This leads to a large amount of wasted writing polymer that does not contribute to a sinusoidal $\Delta n$. The relative rates of reaction and diffusion are also influenced by the recording intensity, gelation point, binder molecular weight, and the immobilisation rate [38] (which limits diffusional blurring).

A similar formulation of Eq. (18), is provided by [18,39,40]:

The modelling presented so far in this section assumes two-component systems consisting of a writing monomer(s) and background components (matrix materials or nanoparticles). Three-component systems also exist. These usually consist of writing monomers, matrix monomers, and nanoparticles. Detailed modelling for such a three-component system was provided by Smirnova et al. [42]:

3. Review of high $\Delta n$ materials and strategies used to record VHGs

The following sections aim to concisely explore recently published developments not reported by previous reviews [16–18] and provide an analysis in the context of the above requirements and limits for applications in DOEs for LED lighting. The sections have been organised according to the material type. With the exception of sol-gels, these material types are categorised according to the family of the principal photopolymerised monomer. These are: acrylate, including the subcategory of the commercially available Baylfol range of photopolymers; acrylamide; and thiol-ene based photopolymers. In each section there are tables to summarise salient holographic characteristics. In cases where characteristics could not be found from literature, cells are filled with ‘N/A’. Finally, a figure is presented which aims to relate the results of the literature survey to the presented modelling.

3.1 Acrylate monomers

Acrylate monomers here include any photopolymer that use acrylates as a photo-polymerisable monomer. Binders and photoinitiator systems may be different in each case, so the characteristics of the material also vary. Nevertheless, most systems use non-water-soluble components and as such are more robust against environmental changes and moisture. Recent developments show the materials to be readily doped with nanoparticles and have allowed these materials to achieve very high refractive index modulations. Important characteristics of each material are summarised in Table 2.

Table 2. Holographic characteristics of transmission VHGs recorded using acrylate photopolymers.

View Table | View all tables in this article

Shen et al. [43] reported a green sensitive acrylate photopolymer doped with erythrosine B photosensitising dye that achieved high $\Delta n$ while being robust to environmental conditions. The composition consisted of: polyvinyl acetate and N-vinyl carbozole as binder; tetrahydrofurfuryl acrylate and 2-phenoxyethy acrylate as monomers; (2, 2’-bis(2-dichlorophenyl)-4,4’,5, 5’tetraphenyl- 1, 2’-biimidazole as photo-initiator; and erythrosine B as photo-sensitive dye. With this material, the authors recorded a $12\;\mathrm{\mu}\textrm {m}$ thick reflection mode VHG with a $\Delta n$ of 0.065 ($T = 12\;\mathrm{\mu}\textrm {m}$) and diffraction efficiency of 99% after some optimisation of the recording and post treatment processes (dark reaction time, UV curing time, thermal treatment temperature and time). The authors demonstrated in a waveguide application with input and output via a reflective VHG, with angular working range , ‘diagonal FOV’ (field of view), 28$^{\circ }$ and spectral bandwidth of around 22 nm. Shen et al. [44] also reported a similar composition with gold nanoparticles (AuNPs) incorporated (average diameter of 6-8 nm). Other material differences with the previous composition were the use of polycellulose acetate (instead of polyvinyl acetate) and the use of benzene and dimethylformamide as organic solvents. The addition of these nanoparticles showed a $\Delta n$ increase from 0.03 to 0.08 ($T = 15\;\mathrm{\mu}\textrm {m}$), and the waveguide device demostrated had a diagonal FOV of 30$^{\circ }$ and a maximum spectral bandwidth of 30 nm. Reflection holograms were studied in both of these works.

Guo et al. [45] reported a $\Delta n$ of 0.046 ($T = 5\;\mathrm{\mu}\textrm {m}$, $SF= 3250$ l/mm). The composition uses a new 2-((2-(9H-carbazol-9-yl)ethyl)thio)ethyl methacrylate monomer ($n = 1.60$). 2-phenoxyethyl acrylate was used as a solvent. The photoinitiation system consisted of 2,2-bis(2-chlorophenyl)-4,4,5,5-tetraphenyl-1,2-biimida- zole (photoinitiator), 2,5-Bis(2,3,6,7-tetrahydro-1H,5H- pyrido[3,2,1-ij]quinolin-9-ylmethylene)-cyclopentanone (photosensitiser). The matrix polymer consisted of poly(ethylene glycol), triethanolamine (TEA), hexamethylene diisocyanate, 1,4-diazabicyclo[2.2.2]octane, and N,N- dimethylformamide. It is worth noting that no application of post treatment is reported by the authors. In addition, to the high value of $\Delta n$, the material showed high total transmittance (96.62%) across a broad spectral range (400-800 nm), and high peak $\eta$ (95.16%).

Tomita et al. [40] demonstrated an NPC material which demonstrated a $\Delta n$ of 0.045 (0.03) in a transmission VHG recorded at 532 nm at 1000 l/mm (2000 l/mm), thickness used was $5\;\mathrm{\mu}\textrm {m}$. The material which achieved the highest $\Delta n$ consisted of a ‘hyperbranched polymer’ (HBP) and an acrylate based photopolymer blend. The synthesis of HBP involved the polycondensation of a diamine monomer with 2, 4, 6-thrichloro-1, 3, 5,-triazine in N, N-dimethylacetamide. The HBP had a refractive index of 1.82 and host monomers were 4-hydroxybutyl acrylate and tetrahydrofurfuryl acrylate. Importantly, the photoinitiator system here played the additional role of delaying gelation via inhibition by Rose Bengal ester dye, which facilitated improved mutual diffusion of HBPs and monomer. The authors noted that the 5$\mathrm{\mu}$m films demonstrated were in an intermediate regime between Bragg and Raman-Nath.

Alim et al. [46] reported a photopolymer composition which obtained a maximum $\Delta n$ of 0.029 ($T = 11\;\mathrm{\mu}\textrm {m}$) in a volume transmission hologram. These gratings showed a maximum $\eta$ of 100%. Briefly, it consisted of a 1,3-bis(phenylthio)-2-propyl acrylate (BPTPA) as writing polymer and a low refractive index urethane matrix. Compared to 2,4,6-tribromophenyl acrylate, which is a widely used writing polymer, BPTPA showed superior solubility in the urethane matrix by about 50%. A drawback, as with many of the materials reviewed so far, is the need for overnight thermal curing of the polymer matrix prior to holographic recording, providing an obstacle to economical mass production.

Sakhno et al. [47] also proposed the incorporation of photoinsensitive (gold and also silver) nanoparticles to enhance $\Delta n$. The polymer consisted of: Isooctylacrylate (monomer), ethoxylated bisphenol A diacrylate (crosslinking monomer), and Irgacure 1700 260 (photoinitiator). AuNPs were then dispersed in CH$_{2}$Cl$_{2}$ and mixed with the pre-polymer syrup. The maximum $\Delta n$ was measured to be 0.0084 ($T = 20\;\mathrm{\mu}\textrm {m}$) , which is significantly lower than that observed in Shen et al. [44] formulation. The use of AgNPs produced higher $\Delta n$. However, in this case, they were added to the composite ’in situ’. That is, a metal precursor solution is added to the pre- polymer syrup and then the NPs are formed by UV curing or heating after holographic recording. The polymer composition consisted of the following: bifunctional monomers triethylene glycol dimethacrylate, $\alpha$, $\omega$-bis-(metacryloyloxyethylenoxycarbonyloxyethylene)-oxyethylene, and $\alpha$- metacryloyloxy-$\omega$-metacryloyloligo(oxyethylene) as the photopolymerisable matrix; Michler’s ketone and camphorquinone as the photoinitiator system. The metal precursor used was a solution of AgNO$_{3}$ in acetonitrile. The peak $\Delta n$, reported here as that defined by eqn. (6), achieved using these components was 0.021 when the concentration of AgNPs was 1.3 wt. %. This was achieved with a grating period of $0.9\;\mathrm{\mu}\textrm {m}$ and $20\;\mathrm{\mu}\textrm {m}$ thickness.

Zhang et al. [48] studied a series of fluorinated epoxy resins (FTGEs) for use as background matrices in the formulation of holographic recording materials. These formed matrices with refractive indices in the range of 1.44-1.46. N- vinyl carbazole and 2-phenoxyethyl acry- late were used as monomers, 2,2-bis(2-chlorophenyl)- 4,4,5,5-tetraphenyl-1,2-biimidazole as the photoiniator, 2,5-bis(2,3,6,7-tetrahydro- 1H,5H-pyrido[3,2,1-ij]quinolin-9-ylmethylene)- cyclopentanone as photosensitiser, and triethylamine, 1,4-dioxane, triethylenetetramine as amine curing agent. Transmission VHGs were recorded for a range of concentrations of different FTGE species. It was found that the samples containing 23 wt% prop-FTGE achieved the highest diffraction efficiency (91.1%) and sensitivity (0.012 cm$^{2}$ mJ$^{-1}$). The $\Delta n$ obtained was not reported so here it is estimated based on data provided by [48]. Using Eq. (1) and taking n as the average of 1.44 and 1.46 (1.45), the Bragg angle was estimated to be 0.452 radians. Using eqn. (8) with the value for angular working range provided (0.24 $^{\circ }$), T was estimated to be $119\;\mathrm{\mu}\textrm {m}$; this is the estimated effective thickness, the reported physical thickness of the layer is 0.5 mm. Then by using these values along with the maximum reported diffraction efficiency to solve Eq. (9), $\Delta n$ can be estimated as roughly 0.00174.

Gallego et al. [49] investigated the fabrication of VHGs for used as waveguides using three different compositions. The components used in different combinations of concentrations were dipentaerythritol penta/hexaacrylate as monomer, a nematic liquid crystal, N-vinyl-2-pyrrolidone (NVP) as crosslinker, N-phenylglicine (NPG) as radical generator, octanoic acid (OA) as cosolvent, and ethyl eosin (YEt) as dye. In combination with NVP, N-methyl-2-pyrrolidone (NMP) was also used. Gratings were recorded with a 532 nm laser with SF of 1700 l/mm. Thicknesses of the layers were between $20\;\mathrm{\mu}\textrm {m}$ and $30\;\mathrm{\mu}\textrm {m}$. For two of the solutions tested, $\Delta n$’s as high as 0.0099 and 0.0085 were detected. These two solutions consisted of the aforementioned components in the absence of NMP.

Morales-Vidal et al. [50] reported a novel environmentally friendly photopolymer known as Biophotopol. The composition consists of PVA as inert binder, sodium acrylate as polymerisable monomer, TEA as coinitiator and plasticiser and sodium salt 5’-riboflavin monophosphate as photosensitive dye. They achieved a peak $\Delta n$ of 0.0034 ($T = 118.1\;\mathrm{\mu}\textrm {m}$, $SF =$ 1205 l/mm). Later, also including platinum nanoparticles, Garcia-Vacquez et al. [51] recorded reflection holograms in a similar material doped with platinum nanoparticles. Holograms were recorded in $160\;\mathrm{\mu}\textrm {m}$ thick layers with a 460 nm beam. The spatial frequency used was 4988 l/mm. Through optimisation of the recording process and nanoparticle concentration, a peak $\eta$ of 30% was reported; an approximately three-fold enhancement of $\eta$ when compared with nanoparticle-free compositions.

Guo et al. [52] demonstrated AgNP polymer composites that provided improved $\Delta n$ despite low doping with NPs. They achieved a peak $\Delta n$ of 0.0069 ($SF =$ 1000 l/mm). RB and NPG were used as the photoinitiation system. Trimethylolpropane trimethacrylate (TMPTMA) was used as monomer. Low doping was used (1 wt.%) with respect to monomer loading. The increased $\Delta n$, relative to undoped photopolymer, was attributed to the enhanced diffusion as a result of increased macroradical generation rate and oxygen suppression. A periodic redistribution of nanoparticles was also observed.

3.1.1 Bayfol

Bayfol [53] is a well known commercially available acrylate based family of photopolymer materials that have found a wide variety of applications such as recording of full colour holograms [54] and DOEs for augmented reality [55]. The physics and chemistry of the material was well established in a seminal work [53].

In relatively recent work, Riva et al. [56] investigated tuning of $\Delta n$ and scattering in Bayfol films. They reported $\Delta n$’s as high as 0.04 ($SF=$ 2000 l/mm)for transmission VHGs recorded in Bayfol HX photopolymers; in contrast the peak $\Delta n$ for transmission gratings was originally reported approximately 0.020 [53]. The VHGs were recorded using a Lloyd mirror setup. All samples underwent UV curing post recording. This consisted of exposure to UV light at 365 nm and 4.43 mW cm$^{-2}$ for 30 min. Two development grade Bayfol film compositions were used: Bayfol HX TP10 (red, green, blue sensitive, $10 \;\mathrm{\mu}\textrm {m}$ thick) and Bayfol HX TP25 (red and green sensitive, $25\;\mathrm{\mu}\textrm {m}$ thick). Films were laminated to glass substrates after removal of the protective cover. Gratings were recorded with three different intensities: 100 $\mathrm{\mu}$W cm$^{-2}$, 200 $\mathrm{\mu}$W cm$^{-2}$, and 500 $\mathrm{\mu}$W cm$^{-2}$. Using the RGB sensitive material the peak $\Delta n$ , 0.04, was achieved after 28 s exposure to 200 $\mathrm{\mu}$W cm$^{-2}$. Beyond 28 s at this intensity the $\Delta n$ remained the same, however the scattering increased linearly with additional exposure time. Similar results were also found for 100 $\mathrm{\mu}$W cm$^{-2}$, however, with a smaller peak $\Delta n$ occurring at 91 s. In the case of the 500 $\mathrm{\mu}$W cm$^{-2}$ exposure, the corresponding peak $\Delta n$ occurred at 28 s. However, in contrast to the other two intensities used, any additional exposure did not result in increased scattering. In addition, the measured $\Delta n$ was higher when the film was sandwiched between two glass substrates as opposed to the typical one substrate. They also demonstrated precise control of $\Delta n$ by combining laser exposure with either a pre-exposure or co-exposure to an incoherent white light source.

3.2 Acrylamide monomers

Photosensitive materials based on acylamide monomers tend to use water-soluble constituents and are generally easier to prepare but less robust than the acrylates. Again, it will be noticed that the photoinitiating system and binders vary significantly for different materials but even with recent improvements, to date the refractive index modulation tends to be lower than that of the acrylate materials. Important characteristics of each material are summarised in Table 3.

Table 3. Holographic characteristics of transmission VHGs recorded using acrylamide photopolymers.

View Table | View all tables in this article

Rogers et al. [57] reported an improvement of $\Delta n$ in acrylamide-based photopolymer by replacing the TEA photoinitiator with methyldiethanolamine. Values of $\Delta n$ in the range of 0.005 were reported at a thicknesses ranges in the range of 20-47 nm and spatial frequency of 800 l/mm.

Rajesh et al. [58] demonstrated an acrylamide based photopolymer employing a mixture of dyes capable of equalising RGB sensitivity; which is of particular interest for holographic DOEs for LEDs. The corresponding $\Delta n$ was reported as approximately 0.0017. VHGs were then recorded in samples including methylene blue, thereby making the material sensitive to red wavelengths. While the measured $\Delta n$ was low, DEs of over 80% could be achieved for all three wavelengths. The materials were also demonstrated to have a long shelf life.

Pi et al. [59] investigated the effect of using either acrylamide or N-vinyl pyrrolidone (NVP) as monomer in PVA based systems. The base composition, excluding monomers, consisted of PVA, tetraiodofluorescin sodium, TEA, and N,N’-methylenebis(acrylamide) mixed in a mass ratio of 84:10:0.02:5:1. The thicknesses of the films were $140 {\pm } 10\;\mathrm{\mu}\textrm {m}$. The overall maximum $\eta$ was roughly 85%. Below monomer concentrations of 0.10 mol L$^{-1}$ the NVP photopolymers showed a higher $\eta$ growth rate and higher sensitivity relative to acrylamide. For this range of concentrations the activity of either monomer was considered to be the principal factor determining the rate of polymerisation. At higher concentrations, above 0.10 mol L$^{-1}$, the diffusion rate becomes a more influential factor on grating formation. In this scenario the acrylamide-based photopolymer showed better holographic performance. While $\Delta n$’s were not reported this provides a useful insight into variations of photopolymers with PVA binders.

Zhang et al. [60] demonstrated the use of upconversion photoluminescence to enhance diffraction efficiencies. To achieve this near-infrared (NIR) sensitive upconversion nanocrystals (UCNRs) were incorporated into a photopolymer with liquid crystals (LCs). The intention being that, as a result of opposing diffusion rates and directions during recording, the UCNRs would accumulate with the photopolymer in the constructive interference regions, while the LCs would accumulate in the destructive interference regions. The primary monomer used was dimethylacrylamide (DMAA). Hyperbranched acrylate monomer was used as a crosslinker with low viscosity in order to permit the mutual diffusion of DMAA and LCs. The photoinitiator system consisted of 3,3’-carbonylbis(7-diethylami-nocoumarin) and NPG. The three UNCRs tested for holographic characterisation consisted of 60% mol Gd$^{+3}$, along with varying amounts of lanthanide ions: Yb$^{+3}$, Er$^{+3}$, and Tm$^{+3}$. The composition containing 15 wt% UNCR-1 showed a $\Delta n$ of 0.034. This composition consisted of 8.0 mol% Y$^{+3}$, 30.0 mol% Yb$^{+3}$, and 2.0 mol% Er$^{+3}$.

Neipp et al. [61] used an acrylamide-based photopolymer materials in their work on a see-through display based on a holographic waveguide. The photopolymer used consisted of: dipentaerythritol penta/hexaacrylate as monomer, a nematic liquid crystal, NVP as crosslinker, NPG as radical generator, OA as cosolvent, and YEt as dye. The maximum $\eta$ achieved was 90%. In layers with reported thicknesses of ${20} {\pm } {1}\;\mathrm{\mu}\textrm {m}$, a $\Delta n$ of 0.0107 $\pm$ 0.0001 (1690 l/mm) was measured by applying theoretical best-fits according to rigorous coupled-wave theory. Shrinkage of the gratings were also observed after which the thickness was 98% of the original. However the SF increased to 1700 l/mm with a decreased $\Delta n$ of 0.0086 $\pm$ 0.0001.

3.3 Thiol-X monomers

Photopolymers consisting of thiol groups linked to other, ‘X’, monomer groups have received significant attention in recent years. This is due to favorable characteristics they possess such as low shrinkage, reduced oxygen inhibition of the recording process, delayed gelation, and environmental robustness. They are also highly tunable as the molecular weights of the thiol group or the monomer group can be altered in order to change density or refractive index characteristics. Important characteristics of each material are summarised in Table 4.

Table 4. Holographic characteristics of transmission VHGs recorded using thiol-‘X’ photopolymers.

View Table | View all tables in this article

Galli et al. [62,63] recently reported a novel photopolymer specifically designed for holography whose writing chemistry contains a high concentration of sulfur atoms. The writing chemistry of the polymer in question consists of a cyclic allylic sulfide monomer with a tetrafunctional thiol. The maximum reported $\Delta n$ was 0.0346 (1200 l/mm, $15-20\;\mathrm{\mu}\textrm {m}$), which occurred after thermal treatment. Without thermal treatment, addition of the thiol to the photopolymer increased the $\Delta n$ from 35% to 86% of the formula limit. This addition of tetrafunctional thiol enabled the growth of a crosslinked network within which migration of the components could occur. Improved $\Delta n$ was demonstrated with increasing thiol concentration, and ascribed to a combination of increased refractive index and density in exposed regions. and increased mobility. The increase in $\Delta n$ beyond the formula limit was explained by density increases in the bright fringes (attributed to polymer chain redistribution) and density decreases in dark fringes (due to material loss). A 22% mass loss is reported for heating for 2.5 h at 120$^{\circ }$C. This mass loss is understood to be due to a uniform mass loss throughout the material and localised mass losses in dark regions. Despite the high $\Delta n$ achieved, the requirement of thermal treatment for optimisation of $\Delta n$ could introduce significant time and financial cost limitations from the perspective of mass manufacturing of DOEs with this material.

Recently, Hu et al. [64] reported transmission VHGs with a maximum $\Delta n$ of 0.04 in a $5.2\;\mathrm{\mu}\textrm {m}$ thick, thiol-ene ‘click’ chemistry based photopolymer. They increased the immobilisation of diffusing monomers by adding allyl side chains to a linear polyurethane binder ($n<1.48$). The use of linear binder improved orthogonality of the matrix and writing chemistries, while the use of cross-linking side chains served to anchor diffusing writing monomers during recording. The material consisted of 1,3-Bis(2-mercaptoethylio)-2-mercaptopropane (BMEMP, trithiol) monomer, 1,2-ethanedithiol-based diallyl ether (EDTDAE, diene) monomer, a linear polymer binder with allyl side chains as crosslinking agent, and diphenyl(2,4,6-trimethylbenzoyl)phosphine oxide (TPO) as photoinitiator. The linear polymer used as crosslinking agent was synthesised from Polycaprolactone-block-polytetrahydrofuran-block-polycaprolactone (Mn 2000), 1,6-diisocyanatohexane, and 2-(allyloxymethyl)-2-ethyl-1,3-propanediol. Holograms were recorded at three grating periods (0.4, 0.5, and 1) $\mathrm{\mu}\textrm {m}$ using a 405 nm laser at an intensity of 16 mW cm$^{-2}$. The maximum $\Delta n$ of 0.04 was achieved for $\Lambda$ of $0.5\;\mathrm{\mu}\textrm {m}$ when the monomer concentrations were 30 mol % and 43 wt. % thiol-ene. The storage lifetime of unrecorded photopolymer was also investigated, a very sharp drop in recorded $\Delta n$ was observed with 6 day old photopolymer. This has been attributed to haze buildup due to precipitation of monomers from the polyurethane binder.

Later, in a related work, Hu et al. [32] investigated how to manipulate the relative rates of polymerisation reaction and diffusion in order to optimise $\Delta n$ in the absence of crosslinking groups. By varying the molecular weight of the polyurethane binder it was found that the diffusion rate varied over orders of magnitude. Additionally, haze was significantly reduced when high molecular weight binders were used. It was found that the highest $\Delta n$ would occur when a molecular weight of $2.9 {\times } 10^{4}$ Da was used at a spatial frequency of 2500 l/mm. At a lower spatial frequency (1000 l/mm), a lower molecular weight of $0.7 {\times } 10^{4}$ Da was required to achieve the optimal $\Delta n$. Peak $\Delta n$ of 0.023 (1000 l/mm) was found at 7 mW cm$^{-2}$. To further investigate reaction rates, the functionality of a tetrathiol monomer was varied, which in turn varied the gel point of the polymers. The tetrathiol with average functionality of 3.5 proved to have the highest $\Delta n$ of 0.028. They also tested a new binder with norbornene pendant groups, timethylolpropane norbornene ester (TMPNE) instead of the allyl group mentioned in the previous study [64]. Since TMPNE has higher reactivity towards thiol in the thiol-ene reaction it reacts more quickly with the photoinitiator. The results showed that higher $\Delta n$’s were achieved when using TMPNE.

Hu et al. [65] have also realised ‘dynamic binders’ in a thiol-ene based photopolymer. These binders permit a light-regulated viscosity reduction which can greatly improve the diffusion of writing monomers during recording. Such a system takes advantage of dynamic covalent chemistry which enables control of covalent bond formation, thus enabling a system with molecular weights and cross-link densities that are ‘stimulus-tailorable’. The dyanmic binders consisted of polyurethane with allyl sulfide side chains. The writing monomers used were 1,3-Bis(2-mercaptoethylthio)-2-mercaptopropane (BMEMP) and EDTDAE. The binders were dissolved in acetone to make 15 wt.% solutions. The monomers and TPO photoinitiatior were then dissolved in this solution. Holograms with a peak $\Delta n$ of 0.022 ($T= 10\;\mathrm{\mu}\textrm {m}$, $SF = 1000$ l/mm). This was almost twice that seen in formulations without dynamic binder and was attributed to improved diffusion. However, at a higher spatial frequency (2500 l/mm), poorer performance ($\Delta n = 0.011$) was seen with the same composition. This loss in $\Delta n$ was attributed to the higher diffusion present at smaller grating periods being enhanced by the reduction in viscosity, thus increasing diffusional blurring.

Mavila et al. [66], described the scalable synthesis of optically transparent thiol-yne click based photopolymers. They also demonstrated the recording of a transmission VHG with relatively high $\Delta n$ of 0.018. This was achieved using the alkyne group 4, 4‘-Thiobisbenzenethiol-Based Dipropargyl Ether, and the commercially available thiol, TMPTMP, as writing monomers. This achieved a maximum $\eta$ of 80% with 1.24% haze in an $11\;\mathrm{\mu}\textrm {m}$ thick film. In addition, similar to Hu et al. [32,64], a polyurethane binder was used. Again, as in these two studies, thermal treatment was also required.

Guo et al. [67], reported the use of single-wall carbon nanotubes which enhanced the monomer conversion in thiol-ene photopolymer layers. They reported a peak $\Delta n$ of 0.0028 ($T = 81\;\mathrm{\mu}\textrm {m}$, $SF =$ 1000 l/mm) at a concentration of 0.0015 wt.%. In addition to improved $\Delta n$, mechanical properties of the material were improved; there was a three-fold increase in hardness without loss of flexibility.

3.4 Photopolymerisable glasses

In recent years, there has been interest in sol-gel technology for the preparation of photopolymerisable glass. These materials consist of a combination of functional polymers with inorganic nanostructured species. Compared with other photopolymerisable materials, they have improved mechanical robustness, good optical quality, thermal stability, chemical stability, dimensional stability, and minimal shrinkage [68]. Important characteristics of each material are summarised in Table 5.

Table 5. Holographic characteristics of transmission VHGs recorded using photopolymerisable glass.

View Table | View all tables in this article

Mikulchyk et al. [69] demonstrated a novel photopolymerisable hybrid sol-gel (PHSG). The material offers the advantages of mechanical robustness, water-resistance, short curing time, eco-friendliness, and good holographic characteristics. Using two-photon polymerisation (2PP), as opposed to conventional transmission hologram recording, transmission holograms with very high $\Delta n$ were recorded. The value of $\Delta n$ achieved through 2PP (0.0091) was roughly three times the maximum achieved using holographic recording (0.0032); only the holographic value is reported in Table 5. This is largely attributed to the improved uniformity of the $\Delta n$ profile throughout the layer thickness. Remarkably, this material was cured in 45 min at 100 $^{\circ }$C, compared with 5 to 21 days with previously reported materials.

Rogers et al. [70] demonstrated a water resistant sol-gel capable of achieving $\Delta n$s at least as high as ${0.0033} {\pm } {0.0004}$ ($T= 92 {\pm } 11\;\mathrm{\mu}\textrm {m}$, $SF=$500 l/mm). Gratings were also recorded at 1000 l/mm. The effect of heating during dark processes was also studied. The sol-gel consisted of: hybrid precursors trialkoxyorganosilane and a zirconium complex; (3-Aminopropyl) (3-Aminopro-triethoxysilanepyl)triethoxysilane for enhancement of condensation of hybrid nanoparticles; Irgacure 784 as photoinitiator; isopropanol; and deionised waters. The maximum $\Delta n$ (0.0042) was achieved in a relatively thin grating ($T =30\;\mathrm{\mu}\textrm {m}$), which was heated at 90$^{\circ }$C for 30 min after recording.

4. Discussion

Thus far a range of materials with high $\Delta n$ have been presented. These materials each fall into one of four different families of photopolymer materials, each with their own physical advantages and disadvantages. The characteristics which constitute advantages or disadvantages are dependent on the intended application. In this review, the focus is on designing DOEs for operation with broadband LEDs sources. Such materials should therefore: be capable of suitably high $\Delta n$ in the range of SFs of interest, be environmentally robust, have high optical quality, have low cost, and be easily modified. Each material family will now be discussed in terms of its advantages and disadvantages, and foreseen suitability for different applications.

Acrylates have been a popular choice since they offer several advantages over other monomers: strong $\Delta n$, high sensitivity, environmental robustness, water insolubility, good optical quality, and they can be doped with nanoparticles. Some of the highest $\Delta n$ polymer materials belong to this group including one commercially available material. Nonetheless, for the purposes as DOEs for LEDs the main disadvantage is the UV degradation. Some of these materials also required post recording thermal treatment. Acrylamides have also been a relatively popular focus of recent research. This is largely due to several advantages: good film forming ability, easily doped, and high sensitivity. They can be easily doped as a consequence of being water soluble. The water-solubility makes these polymers sensitive to changes in humidity and other environmental changes. Therefore they can be used as stable DOEs if they are sealed from the surrounding environment. In addition, the stability of such gratings can also be improved by using NPG as a photoinitiator [71]. Thiol-X photopolymers have also been discussed. Due to there tunability and stability, they have been of significant interest as materials for DOEs. However, a limitation to their widespread use is their relatively high haze and need for thermal treatment in various stages of the fabrication process. Photopolymerisable glasses, fabricated using sol-gel technology, have also been subject to recent investigations. These materials offer several advantages: low shrinkage, high dimensional stability, high thermal stability, and high chemical stability. However, they are currently limited by having moderate $\Delta n$ (holographically recorded) and UV sensitivity.

Through the process of reviewing recent materials developments several strategies for enhancing $\Delta n$ have been identified. These strategies generally fall into one of two categories: process optimisation, or materials optimisation. Process optimisation includes optimisation of the recording conditions and optimisation of the post recording dark reaction times, thermal curing, or UV curing. However, the focus of this review has most been on optimisation of the material compositions. This usually consists of ways to increase the difference between the refractive index of the writing polymer and the background components. The materials modifications that are made include: the addition of inert nanoparticles that counter diffuse during photo-polymerisation; the use of crosslinkers; addition of new monomer species or combinations of monomers; changes to the binder polymer which is typically polymerised before recording in two-chemistry systems; changes to the photoinitiation system, which affects sensitivity, diffusion rates, and the average refractive index of the non-writing background material; and the addition of liquid crystals.

As mentioned earlier, for a given volume fraction of monomer (assuming incompressible species) the optimal $\Delta n$ is determined by ($n_{polymer}- n_{background}$) and the degree of segregation $SD$ [32]. Following this, most of the material strategies employed by the materials reported here have been employed with the goal of increasing one of or both of these factors. As such, with the goal of increasing maximum potential refractive index contrast, some groups employed strategies of adding high refractive index monomers [43,45,46,62,64], low refractive index binders [62,64], low refractive index photoinitiator system [57], or added high refractive index nanoparticles [18,44,47,51,52,67]. These strategies are summarised in Fig. 4.

 

Fig. 4. Summary of strategies used to increase $n_{polymer}- n_{background}$.

Download Full Size | PDF

There has also been focus on optimising SD. Some of these strategies overlap with the process optimisation methods mentioned above. The materials-based strategies to optimise SD have, however, focused on balancing the ratio of diffusion rate to polymerisation. These include modifying the photoinitiation system in order to reduce the reaction rate [40], increasing miscibility of monomers by adding compatible side groups [46], increasing the polymerisation rate using carbon nanotubes [67], improving miscibility of nanoparticles to prevent aggregation [40,44,47], reducing oxygen inhibition of polymerisation [52], tuning the gel point [32], increasing binder reactivity with the intention of immobilising small mobile chains in bright fringes [64], decreasing viscosity by reducing binder molecular weight [32,65].

Other groups have focused on developing other characteristics of photopolymers by improving their mass production capacity [66], environmental robustness [69,70], and environmental sustainability [50,51]. Strategies to improve these characteristics have been using sol-gel technology [70,72], using commercially available components [66], and biodegradable components [50,51]. Given the favorable qualities of these materials, further work to improve the achievable $\Delta n$ should be conducted.

As outlined in great detail by several authors [32], there remains a key challenge of balancing the ratio of diffusion rate to polymerisation rate across a broad range of spatial frequencies. As highlighted by Fig. 5, the deviation between achievable $\Delta n$ and optimal $\Delta n$ increases with spatial frequency. This is explainable by the issue of diffusional blurring increasing as grating periods get increasingly small [32]. As mentioned, some strategies to reduce diffusional blurring have focused on increasing the binder molecular weight [32]. However, this creates an issue for the design of complex DOEs for LEDs, for which a broad range of SFs may be present, since with high molecular weight binders the issue of insufficient monomer diffusion occurs at the lower range of spatial frequencies.

 

Fig. 5. Graphical summary of the transmission holograms described by Tables 2–5. To calculate the plotted curves Eq. (15) was used with the value of $n$ fixed to 1.5. Actual values of $n$ for each material differ. The data point for Guo et al. [45] is in green as it was probed with a 532 nm laser, hence it is green. Likewise for Rajesh et al. [58], the data point is in blue since the grating was probed with a 488 nm beam.

Download Full Size | PDF

It can be seen that for DOEs for LED applications recent advances have produced a range of materials with sufficient $\Delta n$, at least for non-multiplexed elements, but further work is needed on producing robust materials with low cost, good mass production prospects and a flat $\Delta n$ across the range of spatial frequencies (approximately 500 – 4000 l/mm) of interest for transmission DOEs. There may be potential to utilise understanding recently gained in how diffusion, immobilisation, gelation and material transport can be controlled, in order to improve uniformity of $\Delta n$ across a range of different spatial frequencies.

5. Conclusions

In this review, the optimal requirements for photopolymer materials to function as high efficiency volume transmission holographic DOEs for operation with LED light sources have been determined. It was found that diffraction efficiency, angular working range, and spectral working range, would be optimised when $\rho \geq 10$ and $\nu = \pi /2$ simultaneously. Based on this requirement, it is considered insufficient to arbitrarily maximise the refractive index modulation since for a given spatial frequency as an upper limit on this value exists beyond which the recorded grating would no longer be considered a volume grating based on $\rho$. This is particularly the case at low spatial frequencies since relatively small increases in $\Delta n$ can lead to the formation of intermediate gratings; thereby causing light to be lost to higher diffraction orders. In addition, a brief overview of some high refractive index recording materials published since 2017 was provided and compared quantitatively to the modelled limits for volume gratings. These were split into four material types: acrylate based photopolymer, acrylamide based photopolymer, thiol-ene based photopolymer, and sol-gels/photopolymer hybrids. Some of these materials were also doped with nanoparticles.

It has been demonstrated that the $\Delta n$ of many existing materials is sufficient for efficient operation of DOEs with LEDs; especially at lower spatial frequencies. Nonetheless, there remains scope for future improvements of all material types. These improvements include: improved environmental sustainability, low cost production, reduced haze, improved environmental stability, reducing need for post treatment processes, producing a relatively flat $\Delta n$ response across a range of spatial frequencies. There may be potential to utilise understanding recently gained in how diffusion, gelation and material transport can be controlled, in order to improve uniformity of $\Delta n$ across a range of spatial frequencies.