Transparent wood for functional and structural applications

Abstract

Optically transparent wood combines mechanical performance with optical functionalities is an emerging candidate for applications in smart buildings and structural optics and photonics. The present review summarizes transparent wood preparation methods, optical and mechanical performance, and functionalization routes, and discusses potential applications. The various challenges are discussed for the purpose of improved performance, scaled-up production and realization of advanced applications.

This article is part of a discussion meeting issue ‘New horizons for cellulose nanotechnology’.

1. Introduction

Cellulose-based biocomposites are of interest as semi-structural and eco-friendly materials due to the desire to reduce our dependence on synthetic polymers [1]. Low density, favourable mechanical properties, renewable resource origin and low cost are important drivers for using biocomposties in applications ranging from automotive composites to electronic devices [2–5].

Wood derives its mechanical performance primarily from the cellulose component, and is the most widely used bio-based structural material. Wood shows several levels of hierarchical structure [6], and has a characteristic anisotropic organization with open channels for liquid transport [7]. The porous structure is of great complexity combing micro-, meso- and macro-pores [8]. The nanoscale organization of the cell wall as a nanoporous cellulose fibril composite provides opportunity for tailored design of advanced wood-based multi-functional materials with preserved structural performance [9]. Native wood is obviously non-transparent, and this offers an interesting materials science challenge.

Cellulose fibres extracted from wood in the form of pulp fibres have been intensively studied to make transparent paper and films [5]. Wood chips at centimetre scale are subjected to ‘cooking’. The lignin-rich region in the middle lamella and cell wall corner are degraded, so that individual wood tracheid cells in the form of fibres become liberated. These chemical pulp fibres are mainly composed of nanoscale cellulose fibrils and amorphous hemicellulose, and can be used to prepare transparent paper or transparent films. Preparation strategies include cellulose dissolution and regeneration [10], impregnation of porous paper with a polymer [11,12] or decreasing the diameter of the cellulose fibrils to the nanoscale [2,13]. However, the final materials are hard to compete with the original wood tissues in terms of hierarchical structure and oriented cell structure, which may limit their applications where more complex structure is needed. In terms of designing materials based on biological tissues with preserved structure, the generality of the approach was illustrated by Shams et al., who prepared a transparent crab shell [14]. Calcium carbonate and proteins were removed, and the nanoporous chitin fibril network was impregnated with a polymer having similar refractive index as chitin.

Transparent wood was first prepared with preserved structure, for the purpose of wood morphology studies [15]. Li et al. then reported physical properties of transparent wood and pointed out the potential for engineering applications [16]. After this, several publications have discussed various potential applications in smart buildings and structural optics and photonics [15–23]. The present review summarizes recent progress and challenges for transparent wood preparation, improved optical and mechanical performance, new functionalization approaches and applications.

2. Basics of light–wood interaction

When light travels through air and interacts with a solid object, light propagation may continue in the forward direction for refraction and/or absorption, or it may be reflected backwards at the air/solid interface, figure 1a). Some terms are of importance during discussions of optical properties of a transparent material [25]. The total optical transmittance (here we use transmittance for short) of an object is the ratio of transmitted light intensity (both directly transmitted I_T,direct and diffused transmitted I_T,diffuse) to the incident light intensity I_I0. The total transmittance is then (I_T,direct + I_T,diffus)/I_I0. The optical haze is the ratio of diffused transmitted light I_T,diffus to the total transmitted light, I_T,diffus/(I_T,direct + I_T,diffus) depicted in figure 1a. Light is refracted as it transmits through an object due to the difference in the refractive index, and the refraction angle obeys Snell's law, n₁sinθ₁ = n₂sinθ₂ [26]. Several parameters may influence the optical transmittance and haze of an object, including the refractive index, thickness, surface roughness, porosity, pore size distribution, etc. [27–29]. For the case of two-phase materials such as microscale composites, the higher the refractive index contrasts between the two media, the stronger the scattering of light. As a result, a larger percentage of light is reflected, leading to lower transmittance. Light attenuation (I_A in figure 1a) takes place by transferring light to other types of energy, such as heat. With larger thickness or a larger number of solid–solid interfaces in the composite, more light attenuation takes place, and transmittance is lowered. The Beer–Lambert law could be employed to explain this phenomenon [30]. In order to make a composite material transparent, low refractive index contrast between phases and low light attenuation are needed. Obviously, lower thickness results in higher transmittance. It should also be mentioned that for nanocomposites, the small scale of the phase domains makes refractive index differences less critical.

Figure 1.

(a) Sketch illustrating light propagation through a solid object (blue) in a medium (i.e. air), where I_IO, incident light intensity; I_T,directe, intensity of directly transmitted light; I_T,diffuse, intensity of diffusely transmitted light; I_R, intensity of reflected light; I_R,diffuse, intensity of diffusely reflected light; θ₁, incident angle; θ₂, refracted angle, n₁ and n₂ represent the refractive indices of the medium and the solid object. (b) Typical SEM image of wood structure (softwood), showing the aligned porous structure with ‘pipe-like’ wood tracheid fibre cells. Note the abrupt transition between low-density earlywood and higher density latewood regions. Tr, transverse plane; Tα, tangential plane; R, radial plane. The yellow arrow represents the longitudinal direction. Scale bar is 100 µm [24].

Wood is a non-transparent material due to its optical heterogeneous nature, including microscale porous structure, different chemical components in the cell wall with different refractive indices and the presence of strongly light-absorbing chemical entities. Figure 1b is a micrograph of the softwood structure, showing the aligned hollow fibre cells. When light interacts with wood, a combination of reflection, refraction, absorption and transmission occurs. Light scattering takes place at all interfaces between the cell wall (refractive index around 1.56) and air (refractive index of 1.0). Inside the cell wall, mismatch of refractive index of the main chemical components of cellulose (1.53), hemicellulose (1.53) and lignin (1.61) is likely to lead to light scattering [15]. In addition, strong light absorption takes place in wood due to the existence of light-absorbing components in the cell wall, among which lignin accounts for 80–95% of the light absorption [31]. The interaction details depend on the wavelength of the light and the nature of the wood such as density, chemical compositions, wood directions etc. Generally, in order to make wood transparent, light absorption from chemical entities need to be reduced or eliminated. Light scattering at the air/cell wall interfaces and inside the cell walls should be minimized.

3. Transparent wood fabrication

The current preparation approach is mainly based on delignification of the substrate followed by infiltration of a polymer with refractive index matching the wood substrate (figure 2) [16–18]. Delignification decreases light absorption in wood, and the refractive index mismatch in the cell wall. Fink treated wood with a 5% aqueous solution of sodium hypochlorite (recorded as NaClO method) for 1–2 days to remove coloured substances, including lignin [15]. Yano [32] reported lignin removal from wood through sodium chlorite treatment at 45°C for 10 h. Berglund's group obtained delignified wood through modified sodium chlorite method at 80°C for 6 h (recorded as NaClO₂ method). The lignin content was strongly decreased from around 25% to less than 3% [17]. Other researchers have also adapted the NaClO₂ method [21,22,33]. Hu's group removed lignin by cooking in NaOH and Na₂SO₃ solution, followed by hydrogen peroxide (H₂O₂) treatment, resulting in a lignin content lower than 3% (recorded as NaOH/Na₂SO₃ + H₂O₂ method) [34]. Yu et al. [20] also used this method. After delignification processes, the wood colour changes from brownish to white, as shown in figure 2. Berglund's group characterized the change in wood structure upon delignification. The honeycomb-like structure was preserved compared with the original wood (figure 3a,b). The presence of micro-scale and nano-scale voids in the cell wall was attributed to the removal of lignin (figure 3a,b), leading to increased specific surface area. The specific surface areas increase from 1.2 m² g⁻¹ for the original wood to 20 m² g⁻¹ for the delignified wood template (figure 3c) [17]. Preservation of the anisotropic wood structure and increased nanoscale pores was supported by results from small-angle X-ray scattering, [17,35,36] as shown in figure 3d,e. The two-dimensional small-angle X-ray scattering data from original wood and delignified wood both show twofold symmetry anisotropic intensity distribution with a strong streak, indicating the preservation of the anisotropic wood structure. The increased intensity for the delignified wood was confirmed due to the increased pore density by lignin removal. The present authors demonstrated that delignification is not a necessary step for transparent wood preparation [37]. Compared with the delignification alternative, wood template with the same whiteness was obtained by only removing components containing chromophoric structures. Up to 80% of lignin was retained with better-preserved wood structure and higher mechanical strength. The procedure was demonstrated to be more efficient and environmentally benign.

Figure 2.

Sketch showing transparent wood preparation: delignification followed by polymer infiltration. Scale bars in the SEM images are 40 µm. Reproduced with permission from Li et al. [17] (Copyright © 2016 American Chemical Society).

Figure 3.

(a) Images of original wood cross-section showing the cellular structure of wood. (b) Images of delignified wood supporting the presence of a well-preserved wood structure, and the generation of micro-scale and nano-scale voids in the cell wall. (c) Pore size distribution from BET desorption analysis of original wood and delignified wood confirming generation of nanopores due to delignification. Two-dimensional SAXS patterns of original wood (d) and delignified wood (e) [17]. Reproduced with permission from Li et al. [17] (Copyright © 2016 American Chemical Society). (Online version in colour.)

Polymer infiltration of the porous template was then carried out in order to reduce the refractive index contrast and obtain high optical transmittance. Proper selection of transparent polymers with refractive indices matching the cell wall is needed. The refractive index of wood tissue may vary depending on the wood species, delignification efficiency, etc., as pointed out by Fink [15]. The polymers applied to make transparent wood are listed here with refractive index in the bracket: poly(methyl methacrylate) (PMMA, approx. 1.49), epoxy resin (approx. 1.5), polyvinylpyrrolidone (PVP, 1.53), n-butyl methacrylate (1.50), polystyrene (1.59), dibutyl phthalate (1.52), iso-bornyl methacrylate (1.48–1.50), diallyl phthalate and poly vinylcarbazole (1.68) [15–22,34]. Different polymers demonstrate different viscosity, refractive indices and shrinkage properties, which influences the transparent wood preparation and properties. Fink tried with polymer mixtures to minimize the refractive index contrast between wood and infiltrating polymer, and decrease the viscosity and shrinkage of the polymer [15]. After polymerization, transparent wood with an optical transmittance exceeding 80% and haze over 75% can be obtained. The information about transparent wood with various infiltration polymers is summarized in table 1 and in the following text.

Table 1.

Summary of transparent wood preparation and properties. The optical properties are values at the wavelength of 550 nm. T, transverse plane; L, longitudinal direction; t, thickness; Tr, transmittance; E, elastic modulus; σ, tensile strength; ε, strain to failure. All the transparent wood directions are remarked according to our method mentioned above. Mechanical properties from [22] are the data from luminescent transparent wood.

					optical property			mechanical performance
	wood species	template preparation method; temperature; time	polymer	wood thickness direction	t (mm)	Tr (%)	haze (%)	l × w × t (mm3)	E (GPa)	σ (MPa)	ε (%)
Li et al. [17]	balsa	NaClO2; 80°C; 6 h	PMMA	T	0.7	90	approximately 50	60 × 5 × 1.2	3.59	90	approximately 3.5
					3.7	40	approximately 80
Gan et al. [21]	cathay poplar	NaClO2; 80°C; 12 h + H2O2; boiling;4 h	PMMA	—	0.5	86.1	—	50 × 10 × 3	2.66	45.92	approximately 5.5
Yu et al. [38]	beech	NaOH/Na2SO3; boiling; 48 h + H2O2	PMMA	T	5	86	90	60 × 30 × 5	2.67	60.1	approximately 4.8
Yaddanapudi et al. [34]	beech	NaClO2; 95°C; 12 h	PMMA (RI = 1.49)	—	0.1	70	—	50 × 3 × ?	2.5	150	approximately 10
					0.3	30	18
					0.6	15	40
					0.7	10	49
Zhu et al. [18]	bass	NaOH/Na2SO3; boiling; 7 h + H2O2; boiling; 5 h	epoxy (RI = approx. 1.5)	L	2	90	approximately 95	50 × 10 × 3	1.22	23.38	approximately 4
		NaOH/Na2SO3; boiling; 3 h + H2O2; boiling; 3 h		T		approximately 80	approximately 85	50 × 10 × 3	2.37	45.38	approximately 4.3
Zhu et al. [19]	bass	NaOH/Na2SO3; boiling; 12 h + H2O2; boiling	PVP (RI = 1.53)	L	1	90	80	50 × 10 × 3	—	11.7	6
Li et al. [16]	bass	NaOH/Na2SO3; boiling; 3 h + H2O2; boiling; 2–3 h	epoxy (RI = approximately 1.5)	L	5	90	95	—	—	—	—
Fink [15]	broad leaved lime, field elm, ash and champion oak	NaClO; 24–48 h	mixture Aa (RI = 1.55–1.57)	T	up to 4	—	—	—	—	—	—
			mixture Bb (RI = 1.56)

^aPolymers: butyl methacrylate, styrene and dibutyl phthalate.

^bPolymers: iso-bornyl methacrylate, diallyl phthalate and vinyl carbazole.

Despite the success in transparent wood fabrication, challenges exist in order to improve optical and mechanical properties and scale-up the production methods. From the point of rational preparation, the delignification processes are time-consuming and not always eco-friendly. In addition, removal of lignin will weaken the wood structure, so that handling and fabrication of large substrates becomes challenging. Removing only chromophore works for transparent wood preparation; however, the optical performance stability is a potential question for real applications. Polymer properties are important for transparent wood fabrication and performance. Important parameters are refractive index, viscosity, compatibility with cell wall, shrinkage during polymerization, etc. Thus, polymer selection is vital. Furthermore, transparent wood with increased cellulose content is desirable. This can be achieved by compressing the wood structure, although it is dubious if this is feasible for thick structures.

4. Transparent wood optical properties

The current reported transparent wood demonstrates good optical transmittance and high haze (table 1). When light interacts with transparent wood, the following three interactions are considered.

• (i) Reflection at outer gas/transparent wood interface: Transparent wood shows high transmittance and high optical haze, with minor reflection at the outer surface. Li et al. [16] reported about 10% of reflection for transparent wood with thickness in the longitudinal direction.

• (ii) Scattering in the form of refraction and reflection: In transparent wood, light scattering mainly happens at the interface between wood tissue and polymer. The polymers with different refractive indices used for transparent wood preparation are summarized in table 1. The lower the refractive index contrast between the wood template and the polymer, the less scattering will take place at the interfaces. It should be mentioned that with nanoscale phase domain size, the refractive index matching does not need to be as good as is required for micro-scale domain size [38]. The high haze is primarily due to collective scattering inside the composite material. The increased light scattering is apparent when a laser beam is transmitted through wood [17,18]. It is also a challenge to obtain perfect interfaces between the polymer and the wood tissue, because there is substantial risk for interfacial gaps between the polymer in the lumen and the wood cell wall. There are two main reasons: (i) the compatibility between the polymer and the wood template may not be sufficient, and (ii) the polymer will shrink during polymerization.

• (iii) Absorption inside transparent wood: Light absorption is mainly due to the presence of lignin in wood. Delignification and lignin modification were applied by removing components containing chromophores to minimize light absorption. Detailed information is presented in the transparent wood preparation section.

There is interest in considering the effects of material parameters, including wood specimen thickness and cellulose volume fraction on the transparent wood optical properties. As wood thickness increases, the transmittance will decrease and haze will increase. This is caused by the increasing amount of light attenuation with longer light pathway, and the increasing number of polymer/wood interfaces leading to light scattering. Li et al. [17] reported that as the thickness of transparent wood increased from 0.7 to 3.7 mm, the transmittance decreased from 90 to 40%, and the haze increased from 50 to 80% (figure 4a). The cellulose volume fraction is another important parameter. Transparent wood with different cellulose volume fractions but the same thickness can be obtained by compressing wood template before polymer impregnation. Figure 4b shows that the transmittance will decrease and haze will increase, primarily due to the larger number of wood/polymer interfaces [17]. Light scattering of transparent wood based on different wood species may vary substantially due to differences in cellulose volume fraction and cellular morphology.

Figure 4.

Transmittance and haze data for transparent wood as they vary with (a) thickness [17], (b) cellulose volume fraction [17], (c) wood thickness direction (anisotropy) [18] and (d) scattering pattern of transparent wood with different wood thickness direction (anisotropy effects) [18]. Reproduced with permission from Li et al. [17] (Copyright © 2016 American Chemical Society) and Zhu et al. [18] (Copyright © 2016 Wiley).

Optical anisotropy is also of interest. Wood is a highly anisotropic structured material, leading to directional-dependent optical properties. The interface density (number of interfaces per length) is higher when light propagates perpendicular to wood longitudinal direction (or propagates in the transverse plane) compared with propagates along the longitudinal direction. In this paper, transparent wood is termed L-wood when light propagates along the longitudinal and termed T-wood when light propagates in the transverse plane (or perpendicular to wood longitudinal direction). L-wood presents higher transmittance and lower haze than T-wood. This is because L-wood has a lower density of polymer/cellulose interfaces due to the hollow cylinder shape of wood cells (figure 4c) [18]. The light scattering pattern is also influenced by anisotropy. Most of the scattering takes place at the interface between the wood cell wall and the polymer present in the lumen. This interface is located in a more favourable position in the L-wood than in T-wood. Figure 4d illustrates this, in that the scattering is isotropic for L-wood, whereas the scattering effect is highly anisotropic for T-wood. In L-wood, the fibre orientation is parallel to the light direction (z direction), thus the scattering light exhibits a Gaussian-like distribution with similar angular distribution in both the x and y directions [18]. By contrast, the wood fibre in T-wood is aligned in the x direction, influencing scattering in the y direction. Thus, an extremely large scattering angle is observed in the y direction.

5. Transparent wood mechanical properties

The mechanical properties are important for many applications considered. They strongly depend on the wood species (including density, cellulose content, cell structure morphology, annual ring structure etc.), and wood structure anisotropy. From the results in table 1, properties of transparent wood based on balsa, poplar and beech impregnated with the same polymer show substantial differences.

Mechanical property synergies between wood and the polymer are observed for transparent wood tested along the fibre direction (figure 5a) [17,20]. The major load-bearing component in wood is the cellulose nanofibres. They are favourably oriented and can show favourable interaction (bonding) with PMMA (figure 5b,c). Li et al. [17] reported that the tensile properties of transparent wood (V_f = 19%, E = 3.59 GPa and σ = 90 MPa) are much higher than for both the delignified wood template (E = 0.22 GPa and σ = 3 MPa) and for PMMA (E = 1.8 GPa and σ = 44 MPa). Yu et al. [20] observed similar results. Li et al. [17] reported that the tensile properties of transparent wood were improved with increased cellulose volume fraction, and the tensile properties can thus be tailored by compression of the wood template to the desired thickness and cellulose content. Wood is an anisotropic material with higher modulus and strength in the longitudinal direction than in the radial direction [39]. The reason is anisotropic cell wall properties and the geometrical shape of wood fibres. The tensile properties of transparent wood are thus also anisotropic. Zhu et al. [18] reported the mechanical performances of transparent wood along the longitudinal direction are around double that of transparent wood perpendicular to the longitudinal direction.

Figure 5.

(a) Stress–strain curves in uniaxial tension for transparent wood, V_f 0.19 (TW-19) and V_f 0.05 (TW-5), delignified wood (DLW) and PMMA [17]. (b) SEM micrograph of transparent wood showing the nanofibrous nature of the cell wall region and signs of favourable interfacial bonding with PMMA at the PMMA/wood interface [17]. (c) Impact response of transparent wood, sketch (left) and stress–strain curve (right) [16]. (d) Comparison of glass (upper) and transparent wood (bottom) after impact test [16]. Reproduced with permission from Li et al. [17] (Copyright © 2016 American Chemical Society) and Li et al. [16] (Copyright © 2016 Wiley). (Online version in colour.)

One advantage of transparent wood for structural applications is that the fracture toughness is higher than for brittle materials such as glass, so that transparent wood does not shatter in an unfavourable manner [37]. Li et al. [16] reported the impact strength of transparent wood (figure 5d), and noted favourable impact toughness. This shows that transparent wood could potentially be used in applications where impact properties are important.

Incorporation of functional particles into transparent wood introduces multi-functionalities and brings new application directions. The influence of incorporating functional particles on the mechanical properties is also studied. Yu et al. reported that incorporation of functionalized nanoparticles (Cs_xWO₃) could decrease the mechanical properties of transparent wood. A higher content of inorganic particles resulted in reduced strength properties [20,21]. Gan et al. [21] obtained a similar conclusion when magnetic particles (Fe₃O₄) were introduced.

6. Functionalization

Wood shows great potential for multi-functionalities due to the presence of multi-scale pores (from macro-scale lumen channels to nano-scale pores in cell wall), and the compositions of cellulose, hemicellulose and lignin. The main strategies are cell wall modification, cell-wall/lumen interface modification and lumen filling. Involved processes include chemical reactions, physical adsorption or absorption, thermal treatment for phase separation effects, inorganic or metal particle precipitation from salt solution, and combined approaches [9]. Successful examples include magnetic wood [40,41], UV-stabilized wood [42,43], stimuli responsive wood [37] and conductive wood [44–46]. Some technologies can potentially be used to prepare functional wood with optical transparency.

A wood template for transparent wood demonstrates higher porosity compared with the original wood [17]. The pore space is substantial and there are many possibilities to modify the template by adding functional components to the porous regions. One approach for transparent wood modification is to fill the microscale pore space (i.e. lumen space) with polymer liquids (solutions or suspensions) decorated with functional nanoparticles, followed by polymerization/curing [20–23]. Li et al. [23] prepared luminescent transparent wood by impregnation of silicon or CdSe quantum dots (QDs) dispersed in a PMMA monomer/oligomer liquid mixture, and the liquid was then polymerized. The approach is presented in figure 6a. Photoluminescence mapping revealed a uniform distribution of emission centres (figure 6b) with the signal intensity variation less than 10% across the image. The emission and absorption spectra of the QD-modified wood are also provided (figure 6c) because this confirms the successful shift in wavelengths. The quantum yield was 36%, and this as well as observed spectra are similar to that for the liquid suspension of QD. It is therefore concluded that the QDs are well dispersed in lumen space. A later study of luminescent transparent wood incorporated γ-Fe₂O₃@YVO₄ : Eu³⁺ nanoparticles using the same method [21]. Heat-shielding transparent wood [20] was reported by Yu et al., who used Cs_xWO₃ nanoparticles with near-infrared (NIR) absorption properties. The preparation procedure was similar to that in the previous study. Light transmittance curves are presented in figure 6d. With increasing content of nanoparticles, the absorption in the NIR range increases, so that optical transmittance is reduced. The same trend was observed for magnetic transparent wood. Gan et al. [21,22] prepared magnetic wood using dispersed Fe₃O₄ and γ-Fe₂O₃@YVO₄ : Eu³⁺ magnetic particles. SEM and corresponding element mapping, FTIR and X-ray diffraction data confirm the presence of magnetic particles. The saturation magnetization of the transparent magnetic wood increased as the magnetic particles loading increased.

Figure 6.

(a) Sketch of transparent wood modification procedure using QDs in order to achieve luminescence effects [23]. (b) Real colour photoluminescence mapping of Si QDs in transparent wood [23]. (c) Absorption and photoluminescent (PL) emission curves of Si QD wood. Inset image shows the actual samples: (left) transparent wood and (right) Si QD wood prepared for quantum yield (QY) measurements [23]. (d) Transmittance spectra of C_sxWO3/transparent wood [20]. Reproduced with permission from Li et al. [23] (Copyright © 2017 American Chemical Society) and Yu et al. [20] (Copyright © 2017 Royal Society of Chemistry).

Another modification approach is to firstly modify the wood template followed by monomer impregnation and polymerization. Li et al. prepared luminescent wood where lasing emission was observed after impregnation of an organic dye (Rh6G) solution in the wood template. This was then followed by MMA monomer/oligomer impregnation and polymerization [47]. A microscopy image of the specimen and the optical transmission curve are presented to confirm the presence of the dye. Under favourable pump light conditions, cellulose fibres operate as small optical resonators and lead to lasing action so that a wood laser is created by utilization of the microstructural anisotropy in wood. As reported, the nanoparticle or molecular dye content is low (less than 1 wt%). Higher content would strongly reduce optical transmittance. It is a major challenge to increase the content of active component, and yet preserve optical and mechanical properties.

7. Applications

Transparent wood shows low density (approx. 1.2 g cm⁻³), high optical transmittance (over 80%) and haze (over 70%), good mechanical performance, and potential for multi-functional modifications. Interesting application areas include smart buildings and load-bearing structures with photonic function. One may consider building applications where light transmittance can be designed so that artificial light can be partially replaced by sunlight. Li et al. [16] suggested that a transparent wood roof could provide more uniform and consistent illumination during the day (figure 7a). Compared with glass, building panels from transparent wood have lower thermal conductivity, better impact strength and lower density. The high haze with broad scattering angles is an attractive feature for solar cells [48]. Light will travel longer path lengths in the solar cell due to this light scattering phenomenon. The interaction between the light wave and the active medium is improved, so that better solar cell efficiency becomes possible. Zhu et al. [19] demonstrated improved solar cell efficiency of 18.02% by positioning transparent wood on top of solar cells.

Figure 7.

(a) A sketch of a house with a transparent wood rooftop for comfortable lighting conditions [16]. (b) Image of transparent wood with Si QD showing diffused luminescence [23]. (c) Photograph of a luminescent transparent wood stool [37]. (d) PL intensity for QD transparent wood as a function of excitation spot distance from the sample edge: parallel and perpendicular to the fibre direction (red and black), PL of QDs in glass collected as a reference (blue) [23]. (e) Evolution of fluorescence spectrum for the wood laser sample at variable pump pulse energy and fixed pump line length. The inset image is a typical dye-impregnated transparent wood sample illuminated by blue light [47]. (f) Image of model houses showing temperature changes after simulated solar radiation: C_sxWO3/transparent wood house (left) and glass house (right) after 10 min. Note temperature readings below image [20]. Reproduced with permission from Li et al. (Copyright © 2016 Wiley); Li et al. [23] (Copyright © 2017 Wiley); Li et al. [37] (Copyright © 2017 Wiley); Vasileva et al. [47] (Copyright © 2017 Wiley) and Yu et al. [20] (Copyright © 2017 Royal Society of Chemistry).

Obviously, transparent wood does not show an intrinsic optical gain by itself. The addition of active optical media can therefore lead to a broader range of applications. Luminescent transparent wood [21,23] shows diffused luminescence originating from embedded QDs (figure 7b). This may be used in planar light sources, luminescent building elements or design furniture [23]. Li et al. made a luminescent transparent wood model stool by including QDs with different emission colours (figure 7c) [37]. The scattering pattern demonstrates the anisotropic nature of this QD transparent wood nanocomposite (figure 7d). The phenomenon is relevant for wood anatomy studies and wood structure imaging. With luminescent dyes, lasing emission was demonstrated [47], and wooden lasers could potentially be realized. The fibre dimensions and the size distributions influence lasing behaviour (figure 7e). Yu et al. [20] demonstrated transparent wood for heat-shielding buildings/windows by the use of NIR absorbing Cs_xWO₃ nanoparticles. Strong absorption of light was observed at wavelengths from 780 to 2500 nm. The CsxWO3/transparent wood window resulted in lower temperature increase compared with the ordinary glass window reference (figure 7f). Transparent wood may therefore be considered for energy-saving specialty windows. Magnetic nanoparticles have also been used with transparent wood [21,22], and considered applications include magnetic switches and electromagnetic interference shielding.

For future building applications, thick and large-scale transparent wood structures are highly desirable. So far, transparent wood samples reported have been small and of low thickness. That is because with higher thickness, light undergoes longer propagation inside the transparent wood where light attenuation happens (light absorption and scattering), leading to less transmitted light. Therefore, proper polymer selection and chemical modification are vital to minimize the light attenuation. The challenge to increase thickness while managing sufficient transparency also includes the need for scalable processing concepts in order to successfully implement the technologies at large industrial scale. Efforts are needed to improve polymer infiltration for large samples (i.e. metres large), and to minimize polymer shrinkage during polymerization. The long-term stability of optical and mechanical properties will be important for transparent wood in real applications.

8. Summary

Transparent wood was originally described in the early 1990s, with the purpose to facilitate cell structure and morphology studies of wood species. Li et al. then measured physical properties and outlined the possibility to use transparent wood in engineering applications [15]. This has been followed by numerous publications on various applications of the concept. New functionalization procedures in the direction of photonics applications are of great interest, including QD and organic dye additions to the material. IR-adsorbing nanoparticles have also been added to provide heat-shielding function to transparent wood. An important advantage with transparent wood is that it combines functional and structural performance so that load-bearing structures can be prepared with optical, heat-shielding or magnetic functionalities. In order to realize the vision of large-scale wood nanotechnology applications of transparent wood, nanoscale materials science needs to be developed to include scalable processing technologies for industrial implementation.

Data accessibility

This article has no additional data.

Authors' contributions

L.B. conceived the initial idea, and all authors contributed to the final design. Y.L., Q.F. and X.Y. contributed equally to the writing of this paper. All authors gave final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

We acknowledge funding from KTH in the European Research Council project no. 742733, Wood NanoTech.