The Use of Layer‐by‐Layer Self‐Assembly and Nanocellulose to Prepare Advanced Functional Materials

Abstract

The current knowledge about the formation of layer‐by‐layer (LbL) self‐assemblies using combinations of nanocelluloses (NCs) and polyelectrolytes is reviewed. Herein, the fundamentals behind the LbL formation, with a major focus on NCs, are considered. Following this, a special description of the limiting factors for the formation of LbLs of only NCs, both anionic and cationic, and the combination of NCs and polyelectrolytes/nanoparticles is provided. The ability of the NCs and polyelectrolytes to form dense films with excellent mechanical properties and with tailored optical properties is then reviewed. How low‐density, wet stable networks of cellulose nanofibrils can be used as substrates for the preparation of antibacterial, electrically interactive, and fire‐retardant materials by forming well‐defined LbLs inside these networks is then considered. A short outlook of the possible uses of LbLs containing NCs is given to conclude.

The current literature on the use of layer‐by‐layer self‐assembled structures of nanocelluloses and oppositely charged polyelectrolytes for the preparation of strong, adhesive, electrically interactive, and biologically interactive thin films is summarized. By using this renewable and biodegradable nanomaterial, it is shown possible to prepare high‐performance materials in a noncomplicated way.

1. Introduction

Since the introduction of the polyelectrolyte layer‐by‐layer (LbL) self‐assembly^{[ 1} , ^{2 ]} technique there has been a major development in the area of surface functionalization with nanometer precision both regarding chemistry and structure. Due to its inherent simplicity, in combination with the substantial development in the area of nanotechnology, this technique has been extensively used for the development of advanced functional materials^{[ 3 ]} in a wide range of application areas.^{[ 4 ]} Moreover, the recent development of reactive LbLs^{[ 5 ]} has opened further applications of this technique. Of particular interest are bio‐based materials from nanocellulose (NC), not only due to their abundance and sustainability, but additionally because they offer extraordinary chemical and mechanical properties. Specifically, over the last two decades there has been significant development in the area of NC,^{[ 6} , ^{7 ]} and its use in advanced materials. Due to the fact that NCs often contain a high number of charged surface groups (either anionic or cationic), to improve their extraction from different sources and ensure a good colloidal stability in aqueous media, early work demonstrated it was possible to use these materials to form self‐assembled LbL structures together with oppositely charged polyelectrolytes or nanoparticles or by combining anionic and cationic NCs^{[ 8 ]} directly. These self‐assembled LbL structures can naturally be used to create strong and ductile coatings on a variety of 2D and 3D surfaces and can be tuned for specific applications. For example, by using phosphorylated cellulose nanofibrils (CNFs),^{[ 9 ]} with rather low levels of phosphorylation, it was possible to create fire retardant films even though the raw material, cellulose, is inherently flammable. In addition to being used for LbL assembly, NCs can be organized to form a variety of structures including wet stable aerogels or foams and thus serve as substrates for LbL assembly in order to create functional materials and devices. To date there have been significant advances in the LbL self‐assembly of bio‐based materials, both as coating materials and substrates, and in the following we summarize the principles of LbL formation using NCs and also examine the use of wet‐stable NC structures, such as aerogels or foams in combination with the LbL technique to create functional materials and devices.

2. Principles of LbL Growth

The LbL build‐up of polyelectrolytes or nanoparticles on a charged substrate is based on the over‐charging by the polyelectrolyte as they are adsorbed to the solid substrate.^{[ 10 ]} A schematic description of this self‐assembly technique is given in Figure  1 ^{[ 11 ]} and in the original version of the technique a rinsing step is included between each polyelectrolyte treatment to keep the layers as free from polyelectrolyte complexes as possible. If these rinsing steps are not included the layers will be rougher and not as well‐defined. For many applications these rinsing steps might not be necessary as will be discussed below.

Figure 1.

Schematic description of the LbL formation of oppositely charged polyelectrolytes/nanoparticles at the solid liquid interface. Reproduced with permission.^{[ 11 ]} Copyright 2007, The Royal Society of Chemistry.

As indicated in this figure and as shown in earlier investigations the surfaces are recharged in each step which allows for the adsorption in a consecutive layer. The recharging is usually measured as an alternation in the zeta potential between anionic and cationic potential of the treated surfaces.^{[ 10 ]} The adsorption is entropically driven due to the release of counterions from both the substrate and the polyelectrolyte when oppositely charged species interact.^{[ 12} , ^{13 ]} As a result, the adsorbed amount, and hence the layer build‐up can simply be controlled by the charge density of the components and the ionic strength of the surrounding solution. It has been shown that there is a minimum charge density of each component required for LbL formation and that lower charge densities lead to thicker LbL structures.^{[ 14} , ^{15 ]} More specifically, low‐charge‐density polyelectrolytes lead to nonlinear/exponential growth of the layers,^{[ 16 ]} whereas high‐charge‐density polyelectrolytes result in a linear growth after an initial build‐up phase. A full theoretical description of the LbL formation is still not available, despite many attempts,^{[ 17} , ¹⁸ , ^{19 ]} and most likely the LbL formation is strongly dependent on the slow kinetics of polyelectrolyte reconfirmation at interfaces.^{[ 12 ]} As a result, many of the practically investigated LbLs are in a nonequilibrium state. In addition to charge density, the structure (i.e., crosslinking, branching, and molecular weight) of the layer components can have a large effect on the growth of LbL assemblies and the properties of the used polyelectrolytes, in hybrid LbLs of nanoparticles and polyelectrolytes, were demonstrated to have a large effect on the build‐up of the LbLs.^{[ 20 ]} Polyelectrolytes with a crosslinked, 3D structure, such as polyethyleneimine (PEI), resulted in the build‐up in thicker layers as compared to similarly charged linear polyelectrolytes. Further work, to theoretically describe the formation of self‐assembled LbLs, is no doubt necessary in order to be able to specifically tailor the properties of these nanometer thick LbL structures, however tremendous progress has been made in the understanding and the application of LbL assembly.

It was also shown that cellulose nanocrystals (CNCs) and CNFs can be used to form LbLs with a very well‐defined structure together with cationic polyelectrolytes.^{[ 20} , ^{21 ]} It was demonstrated that the formed layers are so well‐defined that they will show different structural colors in diffusely reflected white light to create different structural colors as shown in Figure  2 .^{[ 20 ]}

Figure 2.

Interference colors of films of CNFs and PEI (polyethyleneimine) as a function of the number of layers. For example, 12 refers to a combination of six bilayers. Reproduced with permission.^{[ 20 ]} Copyright 2008, American Chemical Society.

Whereas conventional LbL systems rely entirely on polyelectrolyte assembly, charged nanoparticles offer an exciting way to further control and tailor LbL film growth and properties. This has indeed been shown for charged NC colloids where additional factors, beyond charge density and ionic strength, must be considered when tailoring the thickness of the adsorbed layers. Aulin et al.^{[ 8 ]} produced all‐cellulose LbL assemblies and compared their growth to polyelectrolyte/CNF assemblies. Interestingly, the results showed that the adsorbed amount per bilayer was significantly larger for PEI/CNF layers than for the cationic‐CNF/anionic CNF layers. This highlights the importance of charge density and flexibility of each component to layer growth. The build‐up of the initial 5–7 layers will be dependent on the efficiency in recharging of the supporting surface, which will be dependent on the charge of the components and the 3D structure of the respective components. Figure  3 is also included to demonstrate the differences in growth properties between (PEI/CNF) LbL‐films and (anionic‐CNFs/cationic‐CNFs). Specifically, the fact that the NCs are stiff, highly anisotropic nanocolloids that are not able to conform in the same way as flexible polyelectrolytes in LbL structures must be considered when predicting the growth of the LbL structures. Similar LbL assemblies, prepared exclusively from polyelectrolytes, the detailed mechanisms of LbL growth of NC (and other nanoparticles) containing films is not totally resolved and is naturally dependent, in part, on the properties of the polyelectrolytes and how the NCs have been prepared. Furthermore, due to the high charge density of NCs, additional interactions, such as ion–ion correlation and ion‐dispersion interactions, have to be considered,^{[ 22 ]} which increase the complexity of LbL growth and properties.

Figure 3.

a,b) Estimated adsorbed amount of multilayers from (PEI/CNF) (filled squares) and (cat‐CNF/anionic‐CNF) (open squares) by stagnation point adsorption reflectometry (SPAR) (a) and quartz crystal microbalance with dissipation (QCM‐D) (b). a,b) Reproduced with permission.^{[ 8 ]} Copyright 2010, American Chemical Society.

Regardless of the components of each layer (i.e., polyelectrolyte or nanoparticle) the film growth requires recharging of the surface such that the subsequent oppositely charged layer can be deposited. The efficiency of the recharging of the surfaces by different polyelectrolyte combinations is seldomly quantified and often simple parameters, such as zeta potential, are used to demonstrate the recharging of the surfaces. However, in a recent investigation the recharging of cellulose model surfaces, with different charge densities, by of cationic polyvinylamine (PVAm) and anionic CNF was carefully determined using colloidal probe atomic force microscopy (AFM) measurements.^{[ 23 ]} By fitting the force curves from the AFM colloidal probe measurements to the Derjaguin−Landau−Verwey−Overbeek (DLVO) theory it was possible to determine the surface potential of both the untreated and the LbL‐treated surfaces under different ionic strengths. From these potentials, it was also possible to calculate the surface charge using the Gouy–Chapman model. These investigations showed that with an increasing charge of the cellulose surface it was possible to reach significantly higher surface potentials of the LbL‐treated surfaces compared to the untreated surfaces, see Figure  4 .

Figure 4.

a) Surface potentials of differently charged cellulose surfaces and b) surface potentials of cellulose surfaces modified with (PVAm/CNF) 1.5 bilayers (BLs) with PVAm in the most external layer. The potentials were obtained by fitting AFM‐force data with the DLVO‐theory in an asymmetric model. a,b) Reproduced with permission.^{[ 23 ]} Copyright 2019, American Chemical Society.

For the highest charged surface, a surface potential of 110 mV could be obtained corresponding to a surface charge of 47 mC m⁻².^{[ 23 ]} In order to quantify the colloidal properties of LbL‐treated surfaces it is important to publish these types of quantitative data.

Even though the most common NC containing LbL assemblies rely on ionic interactions between the components there are examples of nonionic systems in which NCs and xyloglucans (XGs) have been used to form LbL films.^{[ 24 ]} The XG and CNCs, containing sulfate esters, were able to form a linear growth of the LbLs as determined by both AFM and neutron scattering. It was further shown that the driving force for the interaction between the cellulose and the XG was of entropic origin, likely due to the release of water during the adsorption process.^{[ 25 ]} These nonionic systems are naturally very interesting as they open up the possibility for creating NC‐based LbL assemblies that are less sensitive to the surface charge and ionic strength of the surrounding solution. However, the colloidal stability might be of concern for low‐charged NC‐systems and higher ionic strengths and the inclusion of highly charged NCs will be difficult due to a too high osmotic pressure in the formed layers.

Generally, it can be concluded that the LbL formation with anionic NCs and cationic polyelectrolytes occurs as expected for an entropically driven process.^{[ 20} , ²¹ , ^{26 ]} In these investigations polydiallyldimethylammonium chloride (DADMAC),^[20,26] cationic starch (CS),^{[ 26 ]} PVAm,^{[ 20 ]} polyallylaminehydrochloride (PAH),^{[ 20 ]} and PEI^{[ 20} , ^{21 ]}were used and they represent a range of strong and weak polyelectrolytes and a range of high‐charge and low‐charge polyelectrolytes. Apart from showing the entropic principle for an LbL formation between anionically charged NCs and cationically charged polyelectrolytes, all these investigations also demonstrated the complexity of forecasting the structure of the formed layers. First of all, it is important to keep in mind that the NCs are anisotropic particles which cannot conform to the same extent as flexible polyelectrolytes. This means that even if the driving force for the formation is of entropic origin the structure of the formed layers will to a large extent depend on the properties of the NCs, such as their water‐binding capacity^{[ 26 ]} and their mechanical properties in the wet state. The water binding capacity of the NCs will also change when they are interfaced with highly charged polyelectrolytes which mean that the properties of the NCs will change as the LbL formation proceeds. It is also well known that the interaction between two oppositely charged weak polyelectrolytes, containing, for example, carboxyl groups and amino groups, will affect (increase) the degree of dissociation/protonation of the polyelectrolytes^{[ 27 ]} meaning that it is difficult to theoretically predict the charge balance between the polyelectrolytes just from the adsorbed amount and the charge of the polyelectrolytes in solution.

Furthermore, it is necessary to consider the influence of properties of the polyelectrolytes on the structure of the formed layers with NCs, both in the wet and in the dry state in order to tailor the final properties of the layers. It was shown that highly charged linear polyelectrolytes form dense structures both in the wet^{[ 26 ]} and dry states,^{[ 20} , ^{21 ]} whereas 3D structured polyelectrolytes, such as PEI, will form thicker layers in the dry state. When an increased ionic strength was used for the adsorption of the linear polyelectrolytes they adopted a shape similar to the PEI and then formed layer structures comparable to the 3D PEI molecules. It was also shown that low‐charged polyelectrolytes, such as CS, formed more extended and water‐rich layers compared with highly charged polyelectrolytes.^{[ 26 ]} All this will naturally be important when tailoring the adsorbed amount and the structure of the formed layers. It is important to stress that in order to at all speculate about the structure and properties of the wet and dry layers it is necessary to use a combination of methods, such as quartz crystal balance with dissipation (QCM‐D) and ellipsometry, or QCM‐D and stagnation point adsorption reflectometry (SPAR) since the layers are water rich and their properties might be important both in the wet state, for example for adhesion performance,^{[ 28 ]} or in the dry state, for example for optical performance.^{[ 20} , ^{21 ]}

This discussion is naturally valid also for the formation of LbLs of oppositely charged NCs. The structure and the properties of the formed layers will depend on the charge density of the NCs, the type of charged groups, i.e., strong or weak groups, and on the dimensions of particles. In order to quantify the mechanisms controlling the interactions between the NCs it is necessary to quantify the size, charge, swelling, and mechanical properties of the NCs and to determine the layer build‐up with a combination of for example QCM‐D and ellipsometry or similar techniques. In a rather early investigation^{[ 8 ]} it was, for example, shown that the water content of the formed NC LbLs was around 65% showing that it will be difficult to discuss the fundamental interaction between the particles by just knowing either the dry adsorbed amount or the wet adsorbed amount where both water and adsorbed NC is combined. As a final remark on the principles of LbL self‐assembly, it is in place to add a few aspects on how prone the LbL self‐assembly is to up‐scaling in comparison to other methods. The basic principle, as shown in Figure 1, is naturally a time‐consuming method that is suitable for preparation of well‐ordered structures for scientific investigations only. A number of different up‐scaling methodologies have been suggested^{[ 4} , ^{29 ]} and among these the use of LbLs to improve adhesion between cellulose‐rich fibers, to prepare strong papers, can be shown as a good example.^{[ 29 ]} This approach showed that it is possible to create LbL structures of different polyelectrolytes on cellulose rich fibers without a rinsing step between the addition of polyelectrolytes, using a pilot scale papermachine, provided that a careful control of the charge of the fibers could be managed. The achieved results also showed the kinetics for the formation of the LbLs were of the order of minutes and could possibly be made even shorter. In a rather recent study, it has also been shown that LbLs of polyelectrolytes could be formed inside wet‐stable aerogels of NC by a fast filtration procedure^{[ 30 ]} showing again the large possibilities for the up‐scaling of the LbL technology. Even though the technique is very interesting, there is naturally a lot of work needed to make it into an industrial method but its robustness paves the way for new processes and materials.

3. Functional Materials Based on LbL

The attractive properties of cellulose nanoparticles, such as their high elastic modulus, large specific surface area, ease of chemical functionalization, and their ability to be self‐assembled using the LbL‐technique offers great opportunities to use them in highly engineered functional materials. In addition, the vast range of substrates and materials available for functionalization with the LbL‐method has resulted in the potential use of NC containing LbL assemblies in numerous application areas.

3.1. Preparation Methods for Thin CNF Films

The preparation of LbL‐films is most commonly performed by the sequential dipping of the intended substrate in aqueous media containing the anionic and cationic components, respectively, often with an intermediate rinsing step. Dipping offers great reproducibility of the film formation and nanometer control over the thickness and the adsorbed amounts by adjustments of parameters, such as pH and ionic strength as discussed above. However, one obvious limitation of the dipping method is the substrate itself which has to be transferred between and submerged in the liquid in the different containers. Naturally, this step‐wise build‐up of the multilayers is not industrially applicable, which has driven the development of continuous LbL‐coating techniques.^{[ 31 ]} Furthermore, the dipping method does not allow for any substantial control over the orientation of the adsorbed components in the multilayer films, which are instead simply trapped on the surface where they arrive. Since highly ordered states rarely are the equilibrium states, any reorganization of the film components occurring on the surface commonly favors relaxations forming more homogeneous films instead of oriented structures.^{[ 32 ]} Orientation is instead induced by applying external fields, such as electric, magnetic, and mechanical.^{[ 33 ]} One technique that offers both the same nanometer control over the film thickness and deposited amount as the dipping technique is the build‐up of LbL‐multilayers using sequential spraying of the components instead of dipping.^{[ 34 ]} In addition to speeding up the LbL‐assembly, while maintaining quality, spraying also offers the possibility of creating in‐plane orientation in multilayer assemblies of high‐aspect‐ratio nanoparticles, such as NCs and polyelectrolytes.^{[ 35 ]} It has also been shown that it is possible to obtain even coatings on different substrates by spraying NC which suggests that oriented LbL‐assemblies of NC are possible with this technique for any substrate.^{[ 36} , ^{37 ]}

3.2. Characteristics of Thin LbL Films Containing Nanocellulose

As stated earlier, the completely individualized and dispersed nanocellulose particles typically have their surfaces chemically modified with carboxyls,^{[ 38 ]} quaternized amines,^{[ 39 ]} phosphonic acid groups,^{[ 40 ]} or sulfate half‐esters^{[ 41 ]} in order to enable their liberation from the parent cellulose source and to render them colloidally stable once dispersed.^{[ 42 ]} This chemical functionalization, with charged groups, also presents the opportunity for using nanocellulose for the formation of multilayers using the LbL‐technique. The feasibility of assembling charged nanocellulose particles with the LbL‐technique to prepare uniform and nanometer‐thin nanocomposite films was demonstrated by Podsiadlo et al.^{[ 43 ]} In this work, CNCs were assembled with pDADMAC by the sequential dip‐coating of glass slides in dispersions of each component.^{[ 43 ]} The assembled films consisted of densely packed randomly oriented CNCs uniformly adsorbed with pDADMAC on the glass surface, see Figure  5a. The film growth, observed by UV–vis absorption, of the CNC/pDADMAC system displayed a linear growth behavior with an increasing number of layers and an average thickness increase of 11 nm per adsorbed bilayer. The thickness of each bilayer is solely dependent on the size of the cellulose nanoparticles as very little thickness increase has been observed when the polyelectrolyte is the capping layer. Thus the effective thickness of a bilayer is equivalent to two overlapping cellulose particles.^{[ 44 ]} The multilayer build‐up of such thin LbL films can also be observed through color changes of the substrates, due to thin‐film interference phenomena. Such color changes were observed for multilayers of (CNC/Xyloglucan)^{[ 45 ]} and (CNC/PAH).^{[ 21 ]} Optically active LbL‐films have also been prepared with cellulose nanofibrils (CNFs), reported by Wågberg et al.,^{[ 20 ]} as mentioned earlier. The dynamics of swelling due to moisture absorption in (PEI/CNF) LbL‐films was studied by Granberg et al. who observed reversible swelling and therefore reversible color changes; making them potentially suitable in moisture sensing and security paper applications.^{[ 46 ]}

Figure 5.

A,B) Height images of 25 bilayers of (PAH/CNC) (A) and 12 bilayers of (PEI/tunicate CNFs) (B) displaying the different packing of the nanoparticles when assembled by LbL self‐assembly. A) Adapted with permission.^{[ 21 ]} Copyright 2006, American Chemical Society. B) Adapted with permission.^{[ 44 ]} Copyright 2007, American Chemical Society.

Conversely, high‐aspect‐ratio tunicate CNFs that were LbL‐assembled with pDADMAC can be used to produce antireflective LbL‐coatings.^{[ 44 ]} The antireflective properties are a result of the formation of a porous LbL‐film of randomly oriented tunicates CNFs instead of a randomly and densely packed film due to the high aspect ratio of the tunicate CNFs, which prevents them packing closely like the CNCs, see Figure 5b. These films have an average refractive index of ≈1.22 and a controlled thickness of ¼ of the wavelength of the incident light, which meets the requirements to be classified as antireflective coverage of a glass substrate. Similar antireflective properties were reported for LbL‐films containing CNFs and chitin nanofibrils. The CNF/chitin LbL‐film also increased the hydrophilicity of the PET substrate.^{[ 47 ]} Figure 5 displays a comparison between the randomly oriented networks of different nanocellulose LbL‐films and the different morphologies obtained when using CNCs and tunicate CNFs in the LbL self‐assembly.

3.3. Mechanical Properties of Nanocellulose Containing LbL‐Films

The ability to uniformly disperse nanoparticles in thin films of nanocelluloses is of great importance when preparing nanocomposites^{[ 48 ]} and the formation of LbL‐films is therefore an excellent method to create mechanically robust coatings. An evaluation of the mechanical properties of cellulose nanofibrils/(poly(ethylene imine) (CNF/PEI) films by buckling mechanics, schematically shown in Figure  6a, where the buckling wavelength is subsequently used to calculate the stiffness of the film. The elastic moduli of (CNF/PEI) films are displayed in Figure 6b. The results showed that the thin films had a Young's modulus (independent of the film thickness between 61 nm and 1.7 µm) of ≈4 GPa at 64% RH, 12 GPa at 40% RH, 16 GPa at 30% RH,^{[ 49 ]} and 17.2 GPa at 0% RH.^{[ 50 ]} A thickness independent modulus was also reported for (CNC/PAH) LbL‐films at 23 °C and 50% RH^{[ 21 ]} using the same buckling evaluation method.

Figure 6.

a) Schematic view of the buckling experiment used for determination of the mechanical properties of LbL‐films, and b) the elastic modulus and thickness of (CNF/PEI) films. a,b) Reproduced with permission.^{[ 50 ]} Copyright 2011, American Chemical Society.

However, the strong adhesion between the substrate and the deposited LbL‐film presents difficulties when free‐standing NC containing LbL‐assembled nanocomposites are desired. Karabulut et al. overcame this obstacle by increasing the number of bilayers, forming micrometer thick LbL‐films in the dry state on a hydrophobized substrate. The adhesion between the hydrophobized substrate and the formed LbL‐film was lowered enough to allow the release of freestanding LbL‐films of (CNF/PEI), which could undergo regular mechanical tensile tests. The Young's modulus of these freestanding LbL‐films was 9.4 GPa and the stress at break was 63 MPa; solvent cast equivalents displayed values of 3.8 GPa and 26.8 MPa, respectively (0% RH).^{[ 51 ]} Free‐standing LbL‐films prepared from cat‐CNFs and phosphorylated CNFs (P‐CNFs) by Ghanadpour et al. displayed superior mechanical properties having a Young's modulus of 9 GPa and a strength of 160 MPa evaluated at 23 °C and 50% RH.^{[ 9 ]}

3.4. Properties and Potential Applications of LbL‐Films Containing Nanocellulose

The robustness of the LbL‐technique and the feasibility to incorporate any charged species in the LbL‐structure together with the nanocellulose has led to the insight of several potential applications for LbL‐coatings and the development of functional LbL‐coatings has been reported. Aulin et al. formed LbL‐multilayers of (CNF/PEI) on poly(lactic acid) (PLA)^{[ 52 ]} substrates and were able to lower the overall oxygen transmission rate of the 133 µm thick PLA substrate from 151.1 (cm³ m⁻² atm day) to 49.3 (cm³ m⁻² atm day) and 9.0 (cm³ m⁻² atm day), with a 0.5 µm thick (20 bilayers) and a 1.8 µm thick (50 bilayers) (CNF/PEI) LbL‐films, respectively. The water vapor transport of the PLA coated with the same (CNF/PEI) coatings decreased from 4.8 (cm³ m⁻² atm day) to 3.3 and 2.3 for the 20 and 50 bilayer coatings, respectively, at 23 °C and 50% RH. The LbL‐coatings were also reported to lower the sensitivity to moisture of the PLA substrate itself.^{[ 52 ]} It is also possible to tune the interaction between CNF and water by adsorbing charged, heat‐responsive block copolymers onto them. For example, Larsson et al. displayed the transition of CNFs from a highly dispersed state in water to a macroscopically aggregated state after block copolymer adsorption and subsequent heating; during which the block copolymer transitioned through a critical solution temperature leading to the aggregation of the modified CNFs.^{[ 53 ]}

It has also been shown to be possible to tune the interactions between two surfaces by coating them with CNFs deposited by the LbL method. The dry strainability of paper sheets increased from 2% to 5% and their tensile strength index went from 25 to 30 Nm g⁻¹ when coated with 10 bilayers of (CNF/PEI). Interestingly, the largest increase in strainability and strength was achieved by having PEI as the capping layer. This was postulated to be due to the greater mobility of the PEI polymer chains which allows them to form entanglements in the joints between adjacent fibers.^{[ 28 ]} By grafting dopamine catechols onto CNFs and subsequently form LbL multilayers with PEI, the wet adhesion between the CNF‐coated surfaces can be significantly improved.^{[ 54 ]} The increased adhesion is created by utilizing the multivalent ion complexation ability of the catechols in addition to interpenetration that occurs between the multilayers. In these tests, the LbL‐multilayers were characterized in the presence of Fe^[3+] and 8 bilayers had a measured thickness of 130 nm. LbL‐assemblies of chemically modified CNFs have also been applied to prepare flame retardant materials. Ghanadpour et al. LbL‐assembled cat‐CNFs with p‐CNFs into µm‐thick free‐standing CNF‐only LbL‐films, which displayed fire self‐extinguishing properties and good char‐forming properties, while still maintaining the attractive properties, such as optical transparency and strength offered by CNFs themselves, see Figure  7a–c.^{[ 9 ]} Figure 7d shows the cross section of the µm‐thick free‐standing LbL‐film and the homogeneous composite formed between the cat‐CNFs and the p‐CNFs.

Figure 7.

a) Freestanding 300 bilayer LbL‐film of (p‐CNF/cat‐CNF). b) The same highly transparent LbL‐film in front of text, and c) its UV–vis transmittance. d,e) Displays the cross‐section of the freestanding film at two different degrees of magnifications. a–e) Reproduced with permission.^{[ 9 ]} Copyright 2017, Elsevier.

The free‐standing LbL‐film was shown to have superior flame‐retardant properties compared to corresponding nanosheets prepared by simply mixing the cat‐ and p‐CNFs. This was attributed to the more intimate contact between the nitrogen‐containing and phosphorous containing CNFs in the free‐standing LbL‐film. Figure 7e shows the cross section of the free‐standing film, its dense homogeneous structure and close contact between the cat‐CNFs and the p‐CNFs in more detail. Improved flame retardancy has also been reported for LbL‐assemblies of p‐CNFs and chitosan on polyurethane foams by Carosio et al.^{[ 55 ]} The (p‐CNF/chitosan) LbL‐multilayers yielded a submicrometer coating on the entire 3D‐structure of the polyurethane foam and prevented melt dripping. During combustion the (p‐CNF/chitosan) coated polyurethane formed thermally stable compounds which acted as a barrier layer, preventing release of flammable volatile compounds which is a prerequisite for propagating the combustion. In addition to supplying flame‐retardancy, LbL‐assemblies of CNFs and graphene oxide (GO) were able to produce highly anisotropic thermal conductors. By utilizingnonelectrostatic interactions, such as van der Waals interactions, Song et al. assembled CNFs and GO using LbL into multilayers where the GO and CNFs were highly oriented within the plane of the LbL‐film.^{[ 56 ]} The GO was subsequently reduced by hydrazine treatment which created large contact areas between the reduced graphene oxide (RGO)‐sheets within the LbL‐film. The increased contact area between the RGO and CNFs reduced the thermal contact resistance in‐plane and thermal conductivities as high as 12 W m⁻¹ K⁻ were reported for 40 bilayers. This was several orders of magnitude higher than the out‐of‐plane thermal conductivity. Table  1 is added to summarize the preparation method, material combination, and the intended application area for some nanocellulose containing LbL‐assemblies.

Table 1.

The preparation method, material combination, and intended application area for some nanocellulose containing LbL‐assemblies

Preparation method	LbL‐combination	Application area	Reference
Dipping	PEI/CNF	Optically active coating	20
Dipping	pDADMAC/tunicate nanowires	Antireflective coating	44
Sequential spin‐coating	chitin nanowires/CNFs	Tuneable surface properties/antireflective coating	47
Dipping	PEI/CNF	Mechanically robust coating	50
Dipping	PEI/CNF	Strong free‐standing LbL nanocomposite	51
Dipping	cat‐CNF/p‐CNF	Flame retardant and strong free‐standing LbL nanocomposite	9
Dipping	PEI/CNF	Gas barrier coatings	52
Dipping	PEI/CNF	Strainable paper	28
Dipping	PEI/catechol‐CNF	Wet‐adhesion	54
Dipping	Chitosan/p‐CNF	Flame retardancy	55
Dipping	GO/CNF	Anisotropic thermal conductors	56

3.5. The Use of Nanocellulose as a Substrate for LbL‐Assembly

Thin and smooth nanosheets of nanocellulose are normally produced by the evaporation or vacuum‐assisted removal of water from an aqueous dispersion of nanocellulose. The final dried nanosheets consist of tightly packed nanocellulose particles and the sheet density is often close to the theoretical value of crystalline cellulose (1500 kg m⁻³).^{[ 7 ]} Within the sheets, the nanocellulose particles are randomly oriented within the plane of the sheet. The sheets themselves are generally highly transparent, the close packing of the particles to the nanoscale dimensions of the components mean that the sheets scatter very little light. Nanocellulose sheets generally display excellent mechanical and barrier properties^{[ 57 ]} and their thickness, typically 20–40 µm, can be easily controlled. These nanosheets inherently have a charged surface, originating from the fibrils from which they were obtained, which makes them excellent candidates for LbL‐functionalization providing potential use in different applications.

CNF nanosheets have been successfully modified using LbL with PVAm and PAA, a combination of polyelectrolytes that has earlier been shown to have antibacterial properties when applied to cellulose‐rich fibers,^{[ 58 ]} to produce nonleaching, contact‐active antibacterial CNF materials. Henschen et al.^{[ 59 ]} studied the antibacterial effect of films of 2,2,6,6‐tetramethylpiperidine‐1‐oxyl (TEMPO)‐oxidized CNFs having different amounts of surface carboxyls, degree of chemical crosslinking, and surface structuring after being LbL‐treated with up to 5 bilayers of PAA/PVAm, with PVAm as the capping layer. The results showed that LbL‐multilayers were formed on all investigated NC substrates with the largest PVAm and bacteria adsorption occurring on the structured film with lowest charge.^{[ 59 ]} This phenomenon is argued to be due to the flatter conformation adopted by the polyelectrolytes when adsorbed onto the highly charged surface,^{[ 60 ]} which prevented efficient recharging of the surface and therefore bacteria adsorption. Conversely, the antibacterial effects of (PVAm/CNF)‐multilayers on cellulose model surfaces were found to be correlated to a higher surface charge of the cellulose model surface and hence an increased PVAm adsorbtion.^{[ 23 ,61]} By comparing different types of cellulose materials it was found that a close packing of the CNFs in the nanosheets limits the adsorption of bacteria. By creating a porous 3D‐structure of CNFs in the form of wet‐stable aerogels it was possible to preserve the high specific surface area of the CNFs to a large extent. The wet‐stability of the formed aerogels in turn allows for LbL‐film formation on the entire aerogel surface. Henchen et al. modified CNF‐aerogels with up to 5 bilayers of the (PAA/PVAm)‐system and reported on the antibacterial properties of the CNF‐aerogel material.^{[ 62 ]}

In the earlier mentioned model study, it was also demonstrated how the properties, such as charge of the model cellulose surface affected the he adsorption and killing of bacteria when being adsorbed onto an LbL‐modified cellulose model surfaces modified with PVAm.^[23]A summary of these results is displayed in Figure  8 .

Figure 8.

The effect of cellulose charge on the adsorption and viability of bacteria. Reproduced with permission.^{[ 23 ]} Copyright 2019, American Chemical Society.

The results showed that a higher charge of the surface, in mC m⁻², resulted in a higher adsorption of bacteria and a better killing efficiency. These results could also explain the results achieved with the LbL‐treated aerogels,^{[ 50 ]} which showed that aerogels act as bacteria “sponges,” efficiently removing bacteria from suspension due to their adsorption to the PVAm‐coated aerogel surfaces. Aerogels having 1 layer of PVAm removed 97% of bacteria from a suspension, while aerogels containing 5 bilayers of (PAA/PVAm) were able to remove 99.9% of bacteria from the suspension.

The large specific surface area of wet stabilized CNF‐based aerogels is beneficial in applications, such as energy storage, particularly in supercapacitors where greater surface area to volume ratios improve performance.^{[ 63 ]} The intimate molecular contact achieved with the LbL‐technique is also beneficial for efficient charge transfer.^{[ 64 ]} Hamedi et al.^{[ 30 ]} utilized LbL‐multilayer formation of different polyelectrolyte combinations and carbon nanotubes (CNTs) on a wet‐stable CNF aerogel. The CNT/PEI system formed highly conductive nanocoatings on the aerogels, forming 3D‐electrodes with a large surface area emanating not only from the aerogel structure but from the CNT/PEI multilayers which are themselves 3D. A symmetric cell of two (CNT/PEI) modified aerogels displayed good capacitive behavior, with the device being able to store up to 11 F g⁻¹.^{[ 30 ]} Similar results were obtained by Erlandsson et al. who prepared miniaturized wet‐stable aerogel beads and subsequently LbL functionalized them with (CNT/PEI), achieving similar conductivity and a capacitance density of 9 F g⁻¹.^{[ 65 ]} The nanometer thin LbL‐films also provide short diffusion lengths for electrolytes inside the LbL‐film which facilitates fast redox‐reactions to take place inside the film on the surface of, for example, redox‐active polymer particles, such as polyaniline (PANI).^{[ 66} , ^{67 ]} Colloidally stable dispersions of PANI‐nanowires are easily prepared by oxidative polymerization of aniline,^{[ 68 ]} these nanowires can in turn be LbL‐assembled with CNTs into (PANI/CNT) multilayers on the surfaces of CNF‐aerogels. Thus could, energy storage devices that benefit from both the large specific surface area, a courtesy of the aerogel scaffold, and the good electrical contact between the PANI‐nanowires and CNTs due to the LbL‐assembly be prepared. These devices benefitted from the redox capacity of PANI and the high electrical conductivity of the CNTs. Luy et al. assembled (PANI/CNT) and (PANI/GO) multilayers on wet‐stable CNF‐aerogels using the fast LbL filtration approach and reported specific capacitance values of 700 F g⁻¹ for the (CNF/PANI) coated aerogel.^{[ 69 ]} Figure  9 displays the sequential assembly of (PANI/CNTs) and (PANI/RGO) in wet‐stable CNF‐based aerogels.

Figure 9.

a) Schematic view of the sequential LbL self‐assembly on the entire surface of a wet‐stable CNF‐aerogel. b–d) The CNF‐aerogel prior to LbL‐functionalization (b), functionalized with (CNT/PANI) (c), and functionalized with (CNT/RGO) (d). a–d) Reproduced with permission.^{[ 69 ]} Copyright 2019, Royal Society of Chemistry.

The sequential nature of the LbL‐assembly technique inherently offers the possibility of depositing several different layer combinations on top of each other. Nyström et al. made use of this and assembled a complete, 3D interdigitated battery by sequentially LbL coating an anode, separator, and cathode on the entire surface of a CNF‐aerogel.^{[ 70 ]} The cathode consisted of 10 bilayers of PEI/CNT and copper hexacyanoferrate, the separator was 30 bilayers of PEI and PAA and finally the anode consisted of 10 bilayers of PEI/CNT. The complete device displayed a total capacity of 5.9 mAh g⁻¹ which is close to the specific capacity of copper hexacyanoferrate. The cycling stability of the interdigitated battery was 67% over 800 cycles, while this is not enough for a commercial application it demonstrates the potential of these types of materials. Figure  10a displays the sequential LbL self‐assembly of the thin‐film device and Figure 10b electron microscopy images of the assembled electrodes and separator on a CNF‐aerogel.

Figure 10.

a) Schematics of the LbL‐process to assemble 3D devices on the entire surface of an aerogel: left: schematic cross sections of the LbL‐functionalized aerogel with one (CNT/PEI) electrode; middle: the (CNT/PEI) electrode and the separator; right: the full device. b) SEM images of the aerogel cross section after the assembly of the (CNT/PEI) electrode (left), electrode with the assembled separator (middle) and the full device with the second electrode (right). c) High‐resolution SEM images after each assembly step showing the substrate, both electrodes and the separator. a–c) Reproduced with permission.^{[ 70 ]} Copyright 2015, Springer Nature.

While the LbL‐technique can be used to create multilayers with different components in each layer, it is also possible to incorporate multiple components into each individual layer. Köklükaya et al. used montmorillonite (MMT), poly(vinyl phosphonic acid) (PVPA), and chitosan (CH) to introduce flame‐retardant properties to CNF‐aerogels.^{[ 71 ]} The aerogels coated with LbL‐quadlayers of the (CH/PVPA/CH/MMT) displayed self‐extinguishing properties immediately after removal of the flame source and the treated aerogels did not ignite when subjected to a large heat flux. The porous structure of the aerogel decreased the thermal conductivity of the material and a temperature difference of 650 °C across a 10 mm thick functionalized aerogel was reported. Figure  11 displays a schematic of the flame penetration test (Figure 11a) and the front and rear sides of a nontreated aerogel and an LbL‐treated aerogel (Figure 1b,c) after the flame penetration test. Figure 11d,e shows the temperatures measured on both sides during the flame test.

Figure 11.

a) Schematic description of flame penetration test. b,c) Photographs of the front and rear sides of reference aerogel (b) and LbL‐treated aerogel (c), 20 s after butane flame application and at the end of the test. d,e) Temperature measured during flame penetration test on the flame‐exposed and unexposed sides of reference aerogel (d) and 5QL‐5 g L⁻¹‐treated aerogel (e). a–e) Reproduced with permission.^{[ 71 ]} Copyright 2017, American Chemical Society.

3.6. Closing Statements and Future Outlook

The combination of the development of NCs and the development of the LbL technique has opened up many possibilities for a bottom up engineering by using renewable nanomaterials. As shown herein, the formation of thin films of NCs, using the LbL technique, can be used to create functional thin films as new interactive materials or to be used to form thin films on or within different substrates. To our knowledge, there are no commercial products available today where NCs are combined with oppositely charged polyelectrolytes or nanoparticles but there are numerous possibilities. The advantage of the technique is its inherent simplicity and its ability to form structurally well‐defined^{[ 20 ]} and oriented structures.^{[ 35 ]} It has also been demonstrated that in some applications,^{[ 29 ]} it is possible to create the LbL structures from polyelectrolytes without an intermediate rinsing step, providing there is a proper control of the charge of the substrate, and that the LbLs are formed within minutes. This type of application would be very suitable also for NCs but a suitable application area is needed in order to propagate the technique to pilot scale and even industrial application. Another future area of large interest is the formation of selective filtration media where the dimensions of the pores can be controlled down to the nanometer and even ångström size^{[ 72 ]} by using polyelectrolyte LbL thin films and it has further been shown that thin films of pure CNF can be used to remove virus particles.^{[ 73 ]} The flux of these materials is naturally very poor but by depositing thin films of CNF in more coarser filtration media it has been possible to combine the separation properties of thin NC films with the higher flux of the supporting membrane.^{[ 74 ]} The LbL technique, using NCs, has not been used in this application area but it represents a very high‐value‐added product application that could defend a higher initial production cost.

The fact that there today is a large‐scale production of NCs, both CNFs and CNCs, means that there will be a supply of NCs in the future. Considering that there is also a huge amount of research in the area, including combinations of NCs and LbLs, means that it is very likely that this technique will be used to prepare new materials in the future. What is needed for the technique to reach commercial scale is a proper upscaling technique since there indeed is a huge market pull for new materials and devices from renewable resources.

Conflict of Interest

The authors declare no conflict of interest.

Biographies

Lars Wågberg has been a professor in fiber technology at KTH Royal Institute of Technology, Stockholm, Sweden, since 2002. The focus of his and his team's efforts has been to quantify the molecular interactions at model surfaces of cellulose, lignin, and hemicellulose, in order to allow for the preparation of new materials from these components using a bottom‐up design with, for example, the layer‐by‐layer self‐assembly. A lot of the efforts have been on characterizing cellulose nanofibrils to incorporate these materials in new, interactive materials where the properties of this renewable and biodegradable nanomaterial are utilized.

Johan Erlandsson received his M.Sc. in chemical engineering specialized in materials chemistry from the Faculty of Engineering LTH at Lund University (Lund, Sweden) in 2013 and his Ph.D. in fiber and polymer science from KTH Royal Institute of Technology (Stockholm, Sweden) in 2019. He is currently a postdoctoral researcher in Prof. Wågberg's lab at KTH Royal Institute of Technology working on the preparation, chemical functionalization, and characterization of cellulose‐nanofibril‐based materials, the main focus being toward cellulose‐based ion‐selective membranes for energy‐storage applications.

Wågberg L., Erlandsson 2001474 * J., The Use of Layer‐by‐Layer Self‐Assembly and Nanocellulose to Prepare Advanced Functional Materials. Adv. Mater. 2021, 33, 2001474. 10.1002/adma.202001474