Hybrid films of chitosan, cellulose nanofibrils and boric acid: Flame retardancy, optical and thermo-mechanical properties

Accepted Manuscript Title: Hybrid films of chitosan, cellulose nanofibrils and boric acid: flame retardancy, optical and thermo-mechanical properties Authors: Khan M.A. Uddin, Mariko Ago, Orlando J. Rojas PII: DOI: Reference:

S0144-8617(17)31000-7 http://dx.doi.org/10.1016/j.carbpol.2017.08.116 CARP 12723

To appear in: Received date: Revised date: Accepted date:

7-5-2017 16-7-2017 27-8-2017

Please cite this article as: Uddin, Khan MA., Ago, Mariko., & Rojas, Orlando J., Hybrid films of chitosan, cellulose nanofibrils and boric acid: flame retardancy, optical and thermo-mechanical properties.Carbohydrate Polymers http://dx.doi.org/10.1016/j.carbpol.2017.08.116 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Hybrid films of chitosan, cellulose nanofibrils and boric acid: flame retardancy, optical and thermo-mechanical properties

Khan M. A. Uddin1, Mariko Ago1, Orlando J. Rojas1,2,*

1

Departments of Bioproducts and Biosystems, School of Chemical Engineering, Aalto University,

FI-00076, Espoo, Finland. 2

Department of Applied Physics, School of Science, Aalto University, FI-00076, Espoo, Finland.

*Corresponding Author: O.J.R., [email protected]

Highlights 

Cellulose nanofibrils (CNF) were used to synthesize films with fire retardancy properties.



The effect of chitosan (CS) boric acid (BA) was demonstrated



The hybrid films displayed optical transparency and strength.



The flammability and the thermal stability were studied with respect to BA loading.



Bicomponent CNF and CS, displayed better fire retardancy than single CS films.

ABSTRACT Chitosan (CS), cellulose nanofibrils (CNF) and boric acid, the latter of which was used as flame retardant, were combined in transparent, hybrid films that were produced by solvent casting. The 1

flammability and the thermal stability of the films were studied with respect to the loading of the inorganic component. Chitosan films displayed fire retardancy properties, which were enhanced in the presence of boric acid. CNF films, in contrast to those from chitosan, were readily flammable; however, when combined with boric acid (30 w%), they became self-extinguishing. Most remarkably, bicomponent films comprising CNF and chitosan, displayed better fire retardancy than that of neat CS films. Moreover, boric acid improved the thermal stability of the bicomponent films. The tensile strength and Young’s modulus of CS, CNF and CS-CNF films improved at intermediate boric acid addition, although a negative effect on elongation was observed.

KEYWORDS: Fire retardancy; boric acid; chitosan; cellulose nanofibrils; transparent films; thermal stability. INTRODUCTION There has been great interest in fire retardancy by adoption of agents that are non-toxic and environmentally safe, including halogen-free compounds. While less studied compared to those based on halogens or phosphor, boric acid has been used as an effective flame-retardant (Lyons, 1970; Baitinger, 1982; LeVan, S. L. 1984). The fire retardant mechanism of boric acid is physical, i.e., achieved by the formation of an impenetrable glass coating or protective layer on the solid surface at high temperatures. This layer traps volatile pyrolysis products, hinders oxygen diffusion, inhibits mass transfer of combustible vapor and prevents the propagation of the exothermal combustion reactions (Kandola, Horrocks, Price, & Coleman, 1996; Lyons, 1970). The effect of the protective layer at an early stage of the combustion stems from the low melting point of boric acid (Tm= 170.9 °C), which suppresses smoldering. The endothermic dehydration of the acid also retards solid decomposition and dilutes the mixture of volatile products. In addition, boric acid acts through other chemical mechanisms, catalyzing dehydration and oxygen-eliminating reactions, for

2

example, in wood, at a relatively low temperature (approximately 100–300 °C); it may also catalyze the formation of aromatics from the subsequently formed polymeric materials (Wang, Li, & Winandy, 2004). These aromatic structures create char and reduce the effective heat of combustion, resulting in lower heat release and slower flame spread. Thus, the acid-catalyzed dehydration reaction increases the amount of char during pyrolysis (Kandola, Horrocks, Price, & Coleman, 1996; Lyons, 1970; Rowell, 1984). Volatile pyrolysis products generated from materials containing boric acid are qualitatively different from those originating by others treatments. After treatment of wood with boric acid, it undergoes decomposition at a temperature similar (Shafizadeh, Bradbury, DeGroot, & Aanerud, 1982), or slightly lower (Wang, Li, & Winandy, 2004) as that for untreated samples, under lower char oxidation rates (Hirata & Werner, 1987). Chitosan is an amine-containing polysaccharide derived from the alkaline deacetylation of chitin, which is soluble in acidic aqueous solutions (Kurita, 2001; Desbrieres, 2004). Chitosan has been used for its oxygen barrier and flame-retardant properties (Butler, Vergano, Testin, Bunn, & Wiles, 1996; Laufer, Kirkland, Cain, & Grunlan, 2012; Hu, Song, Pan, & Hu, 2013). In fact, the presence of multiple hydroxyl and amino groups in chitosan promotes the formation of char during burning (Don, Hsu, & Chiu, 2001). Thus, while both boric acid and chitosan are considered fire retardants (Lyons, 1970; Baitinger, 1982; LeVan, S. L. 1984; Butler, Vergano, Testin, Bunn, & Wiles, 1996; Laufer, Kirkland, Cain, & Grunlan, 2012; Hu, Song, Pan, & Hu, 2013), the question that arises is if their effects can be combined when added to cellulose nanofibrils (CNF). In such systems, intermolecular hydrogen bonding between boric acid, B(OH)4, and hydroxyl groups of CS and CNF can be hypothesized to contribute to fire retardancy effects. Indeed, CNF has drawn widespread attention in the past few years given their availability, biocompatibility, biodegradability and interesting chemical-mechanical properties (Syverud, & Stenius 2009; Yano & Nakahara, 2004; Moon, Martini, Nairn, Simonsen, &Youngblood, 2011; Saito, Kuramae, Wohlert, Berglund, & Isogai, 2013). Among the several potential applications for CNF, transparent and flexible films have 3

been reported. They can offer advantages as far as their strength, low density and gas barrier properties. In addition, they can form bioactive films, inorganic/organic hybrids, filaments and supports for flexible electronic devices (Zhang et al.,2013; Nogi, Iwamoto, Nakagaito, & Yano, 2009; Fang et al., 2014; Huang et al., 2013). However, the highly flammable character of cellulose is a significant drawback for any of these applications and several efforts have been tested in order to endow CNF-based materials with flame retardancy (Horrocks 1996). Although the use of boric acid and chitosan toward fire retardant materials has been explored, so far no attempt has been made to study the effects associated with the combination CNF with boric acid and chitosan for these purposes. Therefore, here we inquire on the interactions of boric acid (added either in the bulk or as a coating layer) in films of CS, CNF or those comprising equal parts of CNF and CS (thereafter termed as “bicomponent films”). As such, the fire retardancy effects as well as physical, optical and mechanical properties of these systems are elucidated in this investigation.

MATERIALS AND METHODS Chitosan (CS) of medium molecular weight (product # 448877, Mw= 300−500 kDa, degree of deacetylation= 80% determined by acid-base titration), boric acid and hydrochloric acid (HCl) were purchased from Sigma-Aldrich and used as delivered. The water used was purified with a Millipore system. The chitosan (CS) suspension was prepared by dissolving 1g in 300 ml (1 mM) HCl solution and stirred overnight with a magnetic stirrer. The suspension was dialyzed against water using a dialysis membrane (MWCO 6-8 kD) until the conductivity dropped below 5 µS/cm. The suspension was then filtered using a glass filter under vacuum. Unmodified cellulose nanofibrils (CNF) were prepared by first diluting never dried wood fibers in deionized water to a solids content of 1.5 w%. The fiber suspension was then passed sequentially (six times) through a M110P fluidizer (Microfluidics Corp., Newton, MA, USA) equipped with a chamber pair (200 and 100 μm) and operated at 2000 bar pressure. Bicomponent chitosan-CNF 4

aqueous suspensions were prepared by mixing equal amounts of both biopolymers and stirring overnight. Preparation of Chitosan, CNF and Chitosan-CNF Films loaded with boric acid. CS, CNF and CS-CNF films loaded with boric acid were prepared via solvent casting and evaporation. The bicomponent (CNF-CS) films were prepared by mixing same amounts of CNF and CS. The CSCNF suspension was prepared by mixing equal volumes of 0.3 w% CS and CNF for 1 hour. 40 ml of the suspension was mixed with a given amount of boric acid under vigorous mixing with a magnetic stirrer. The well mixed suspension was then poured on a polystyrene Petri dish (8.8 cm diameter), followed by evaporation in a fume hood at room temperature. Coating of the films with boric acid. Boric acid solution was used to coat single component (CS or CNF) as well as bicomponent (CS-CNF) films. The boric acid solution of given concentration was spread over the surface drop-wise. Flammability test. The flammability of the prepared films was measured in the presence of air by using a vertical burning test similar to that in standards such as UL94 or ISO 9773 (ISO 9773:1998). More specifically, films were cut (70 mm in length and 10 mm in width) and fixed on a plate and then ignited with a gas flame for two seconds. The total burning time (time it took for the flame to extinguish) and the after-glow time (time during which a red glow of the film persisted after the flame was extinguished) were evaluated from a video recorded during the test. After extinction, the mass of the residue was weighed. Fourier-transform infrared spectroscopy (FTIR) and Thermal Stability. The FTIR spectra of the samples were collected at room temperature using 64 scans in the range of 4000-400 cm-1 with a (Bruker Alpha Eco-ATR). The thermal properties of the films were analyzed by thermogravimetric analyses (TGA). The tests were performed with a TAQ500 thermogravimetric balance operating in nitrogen and oxygen atmospheres, using ~10 mg of samples that were placed on an open pan and 5

heated (constant heating rate of 10 ºC/min) from 40 to 800 °C. Most of the measurements were carried out in duplicate or triplicate and the values reported correspond to the average. Imaging and Optical properties. The CS, CNF and CS-CNF films mixed with examined with a Zeiss Sigma VP scanning electron microscope at 1-2 kV accelerating voltage. The samples were placed on a carbon tape and sputter-coated with gold before SEM imaging to reduce charging. At least five different spots per each sample were observed under the microscope. UV –Vis spectroscopy was used to record the transmission spectra of the prepared films within the visible region. The tests were performed using a PerkinElmer Lambda 950 UV/ Vis/ NIR absorption spectroscopy. Mechanical properties. The films were cut into strips 5 mm wide and more than 4 cm long using a paper-cutting machine. The thickness of the strips was measured with a micrometer. Prior to mechanical characterization, the films were equilibrated for 24 h in a conditioned atmosphere (RH 65 %, 25 °C). The tests were performed using a Deben Microtest with 200N load cell with an elongation rate 0.2 mm/min.

RESULTS AND DISCUSSIONS Films properties. Digital imaging, UV-vis transmission spectra and scanning electron microscopy (SEM) were used to investigate the optical properties, visual appearance and morphology of single (CS or CNF) or bicomponent (CS-CNF) films as well as the respective organic-inorganic hybrids, upon addition of boric acid. Figure 1a-c include the UV-vis spectra of films based on CS, CNF and CS-CNF with boric acid added at different loadings. The CS films showed the highest transparency, over 90% in the range of 400-800 nm. The addition of boric acid introduced a slight improvement in the transparency and density of these films (less void volume), which may be an indication of the interaction between the components (see Figure S1 for film thickness in Supporting Information). 6

Compared to the CS films, those comprising CNF were much less transparent. This is explained by the fact that the extent of fibril deconstruction was limited (given the relatively small number of passes in the microfluidizer) and some residual fibril bundles remained, which restricted the ability for packing into highly densified (and transparent) structures. Certainly, better transparency in CNF films could be attained by increasing the degree of deconstruction or by the suitable choice of isolation and film preparation methods, as we have reported in earlier work (Toivonen et al., 2015). In contrast to the case of CS films, the addition of boric acid slightly reduced the transparency of films consisting of CNF (CNF or CNF-CS). Such negative effect of boric acid was more noticeable at the higher wavelength of the spectrum. Figure 1d includes digital photographs of the films that used a printed-paper sheet as background in order to illustrate their visual appearance and transparency. Scanning electron microscopy (SEM) images of the films (Figures S2 to S4) indicated the surface morphology of the films, which were very smooth. This property was noted to improve with the addition of the inorganic component. In addition to the fact that boric acid may have a space-filling effect, crosslinking of boric acid with available nitrogen and hydroxyl groups are expected to make the films more dense. Flame retardancy. The effect of a flame on the films was evaluated (flammability tests). Figure 2 includes snapshots 1 second after ignition of strips positioned vertically (vertical flammability tests) for CS, CNF and bicomponent (CS-CNF) films with different loadings of boric acid. The results from the flammability tests are included in Table 1 and the residual mass (wt %) as a function of boric acid loading is presented in Figure S5. It was observed that neat CS films (with no added boric acid), did not produce a flame. Only a small combustion of the film was observed but with no signs of afterglow. This confirmed the already reported fire retardant properties of chitosan (Butler, Vergano, Testin, Bunn, & Wiles, 1996; Laufer, Kirkland, Cain, & Grunlan, 2012; Hu, Song, Pan, & Hu, 2013), which acts as an oxygen barrier. The addition of boric acid to chitosan increased the total residue %, as expected from the formation of more char in these films. 7

After ignition, single component CNF films burned readily and displayed a vigorous flame, as expected; the total burning time was only 3 seconds and the afterglow was observed for 2 seconds. This confirmed that CNF films were highly flammable. With increased boric acid loading, the total residue gradually increased from 0% to 86% (Table 1), which resulted from the formation of more char. The CNF film with 30 w% boric acid showed self-extinguishing behavior after ignition and the flame spreading decreased and ceased, leaving a large amount of residue. Upon addition of boric acid, the total burning time of the film increased, due to the reduction of the flame spread rate. We observed a long afterglow, 2 seconds in single component CNF films. The presence of afterglow indicated that the CNF film continued combustion without producing any flame and was completely consumed; no residues were found after the flammability test with single component CNF films. However, the afterglow was gradually suppressed with boric acid addition. A remarkable phenomenon was observed in the case of the bicomponent CS-CNF films loaded with boric acid: an excellent fire retardant effect was determined, similar to that of CS films, with no flame and no afterglow upon ignition. Furthermore, the residue after combustion of bicomponent CS-CNF films at low boric acid contents (< 10%) was higher than that of CS (or CNF) films. In the absence of boric acid and compared to that of the single component CS films, the bicomponent CSCNF films showed a 6% increment of total residue. Overall, the results indicate a synergistic effect of CS and boric acid on fire retardancy of CNF films. The observations can be explained by the binding ability of chitosan with cellulose via interactions such as ionic as well as hydrogen bonding. In addition, physical crosslinking between cellulose and chitosan can make the CS-CNF film more cohesive (Toivonen at al., 2015). Besides, a reduction of gas transport or permeability (better barrier) occurred by the combination of CS and CNF (Österberg et al., 2013; Nair, Zhu, Deng & Ragauskas, 2014; Khan et al., 2012; Azeredo et al., 2010). Simultaneously, the presence of impermeable crystalline components increased the tortuosity to molecular transport, leading to slower diffusion processes and, hence, to lower permeability which makes the CS-CNF film more flame retardant. The thermal stability of CNF was higher than chitosan, which also make the bicomponent CS-CNF film more thermally stable.

8

Thermal and thermo-oxidative stability. The thermal and thermo-oxidative stability of the CS, CNF and CS-CNF films were examined by thermogravimetric analysis in nitrogen and oxygen atmospheres. The purpose of these tests was to assess the degradation behavior of the films in the presence of boric acid. The thermal stability test is similar to that of flammability but without a source of ignition. For CS films in the presence of nitrogen, the first stage of weight loss was around 150 ºC, which corresponded to the loss of adsorbed and bound water (Figure 3a). The second stage of weight loss, in the range of 200-300 °C, was mainly due to depolymerization (decomposition) of chitosan molecules and deacetylation of residual acetyl groups. The third stage, for temperatures higher than 300 ºC, corresponded to the residual decomposition reactions and char formation (De Britto & Campana-Filho, 2004). Initially, the Tonset temperature and the peak height corresponding to maximum weight loss rate of pure CS film (without boric acid) was higher than the CS film loaded with up to 5% with boric acid. This indicated that addition of boric acid caused early start of degradation but decreased the rate of degradation of the CS films. However, at higher boric acid concentrations, 10% or above, and compared to the pure CS film, the Tonset temperature increased. This indicated that boric acid delays the decomposition of chitosan. The FTIR spectra (Figure S6a) include absorption band at 810 cm-1, from to interaction of boron with nitrogen, which is taken as an indication of B-N bonds (Swain, Dash, Behera, Kisku, & Behera, 2013). Also the disappearance of the band at 1245 cm-1, corresponding to the C-N stretching of amine (Peters, 2014), clearly indicated the existence of molecular interactions between boric acid and chitosan. Such interactions may explain the observed improved thermal stability of chitosan. A similar phenomenon was observed with Tmax. On increasing the temperature to 265 ºC the CS films loaded with boric acid showed better thermal stability. Char is important as indicator of the enhanced thermal stability and flammability of the materials. The chitosan films produced significant amounts of char (45%), Table 2, which highlights the known flame retardancy capacity of chitosan. With an increase boric acid concentration, the char 9

production was increased; thus, boric acid had a significant effect on fire retardant effects in chitosan films. For CNF in the nitrogen atmosphere, the first stage around 150 ºC corresponded to the loss of adsorbed and bound water (Figure 3b). The thermal degradation of CNF films was similar to that of pure cellulose, with one-step weight loss (280-390 ºC) due to the depolymerization and decomposition of the glycosyl units to thermally stable aromatic char, which formed the final residue (Horrocks, 2001). The TGA profile of CNF films obtained with different concentrations of boric acid (BA) showed three weight loss stages. The first step (110-210 ºC), was from the evaporation of water from the boric acid. It was observed that the weight loss % increased with increasing boric acid concentration. The second step was similar to the first step of neat CNF films, which indicated that cellulose degradation occurred similarly to pure cellulose. The third step (360410 ºC) was related to formation of complexes between boron and cellulose, as indicated by the FTIR spectra that showed the interaction of boron with hydroxyl groups in cellulose (Figure S6b). In fact, boric acid can undergo an esterification reaction with CNF with increasing temperature up to 450 °C (Dasari, Yu, Cai, & Mai, 2013; Wicklein, Kocjan, Carosio, Camino, & Bergström, 2016). The CNF films containing up to 5% boric acid (BA) showed a slightly lower Tonset; however, at higher boric acid concentrations, a slight increase was observed for Tonset, which indicated a delay in the start of depolymerization and decomposition of cellulose. The degradation rate of CNF films loaded with BA was slower than the pure CNF film, which suggested a slower formation of the volatile and combustible products (levoglucosan) (Horrocks, 2001; Alongi, & Malucelli, 2015). Addition of boric acid, up to 5%, reduced the Tmax (Table 2) of the films, but at 10% or higher boric acid concentration it started to increase again. Both the weight loss and the maximum weight loss rate differential peak height were reduced with increasing boric acid concentration. This may be due to the thermal annealing of boric acid with CNF in the presence of nitrogen, which caused a slower degradation (Kandola, Horrocks, Price, & Coleman, 1996). In addition, this effect may be due to the 10

esterification reaction of CNF in the presence of boron (Dasari, Yu, Cai, and Mai, 2013; Wicklein, Kocjan, Carosio, Camino, & Bergström, 2016). The total residue at 600 °C (the amount of water loss at 100 °C was taken into account) of CNF films with increasing boric acid concentration gradually increased compared to CNF films without boric acid It was observed that bicomponent, CS-CNF film degraded in nitrogen following only one-step profile at 230-280 °C (Figure 3c), which may be due to the depolymerization and degradation of CNF and chitosan and finally, char formation. The CS-CNF films containing boric acid showed a different behavior. Initially, up to 5% loading of boric acid in CS-CNF films, started to degrade at lower temperature (180-220 ºC) than the films with no BA, and degraded in one-step (similarly to CS-CNF film without boric acid). The CS-CNF films with more that 10% boric acid added, displayed an increased degradation temperature and degraded in two steps. The first one, between 220-260 ºC, corresponded to the depolymerization and decomposition of the glycosyl units. The second one, between 260-400 ºC, was due to the complexation, esterification and thermal annealing of CNF with boric acid as we observed earlier with CNF –BA films. The total residue at 600 ºC (note: the amount of water loss at 100 ºC was taken into account) increased in the films prepared at higher boric acid concentrations. In the presence of oxygen, the thermal oxidation of CS and CS-BA films (Figure 4a) took place in two steps. The first step between (230-236 °C) was very rapid, due to the formation of volatile and aliphatic char, and then, the char was oxidized very slowly in the second step, between 236-580 °C. The rate of oxidation decreased by increasing the concentration of boric acid, in a higher temperature range, from 236-800 °C. The CS films loaded with BA showed a slightly higher Tonset (Table 2), which indicated that boric acid increased the degradation temperature of chitosan. At increased concentrations of boric acid, the total residue at 600 °C of the film gradually increased. The thermal oxidation of pure CNF (Figure 4b) in the presence of oxygen, took place in two steps. The first step, between 250-340 OC, was due to the formation of volatile and aliphatic char, then the 11

char was rapidly and completely oxidized in the second oxidation step between 340-430 °C (Alongi, Camino, & Malucelli, 2013). The CNF film obtained with different boric acid loadings showed several weight loss steps. The first step between 90-200 °C was due to evaporation of water, as observed under nitrogen. The second step, between 250-340, was similar to the pure CNF film and, the third step, between 340-500 °C, was due to the oxidation of the char. It was observed that the films with boric acid oxidized at higher temperature compared to the neat CNF film (see Tmax 2 in Table 2) with a slower oxidation rate. A slower oxidation step above 500 °C was observed and the oxidation rate decreased with increasing concentrations of boric acid. The oxidation steps of the CNF films in the presence of BA started with partial oxidation and then continued as an oxidative degradation at a slower rate. These films showed a slightly higher Tonset (Table 2) for cellulose degradation, which indicated that boric acid increased the degradation temperature of cellulose. The total residue at 600 °C of the films increased gradually with increasing boric acid concentration except for 10% BA loading in the films. It can be concluded that boric acid slowed down char oxidation and shifted it to higher temperatures. The reduction of the oxidation rate with boric acid addition indicated that boric acid engaged in complexation, esterification and thermal annealing with cellulose (Kandola, Horrocks, Price, & Coleman,1996; Dasari, Yu, Cai, and Mai, 2013; Peters, 2014; Wicklein, Kocjan, Carosio, Camino, & Bergström, 2016). The TGA and dTGA profiles in oxygen for the CS-CNF films are presented in (Figure 4c). The relevant data are summarized in Table 2. Comparison of the weight loss of bicomponent CS-CNF films in the presence or absence of BA suggested that the presence of boric acid slowed down the char oxidation and shifted to higher temperature. The degradation occurred in multiple steps. This may be due to the esterification and complex formation between boron and nitrogen or hydroxyl groups. In the third step, an even slower oxidative degradation was observed. A higher Tmax, more total residues were determined with higher boric acid content.

12

Thermal and thermo-oxidative stability of the film coated with boric acid. In this section we examine the thermal and thermo-oxidative stability of the CS, CNF and CS-CNF films when coated with boric acid, which is a different procedure as that employed so far for film preparation, which involved simple (bulk) mixing. The corresponding TG and dTG profiles of boric acid-coated CS, CNF and CS-CNF film in nitrogen atmosphere are presented in Figure S7. In the case of boric acidcoated CS films (Table 3, Figure S7a), it was observed the Tonset and Tmax were lower than that for the neat CS film with up to 10 % of boric acid and the maximum weight loss rate at 10% (or higher) boric acid was lower than that of neat CS film. This indicated that compared to uncoated CS films, the CS film coated with boric acid started early degradation but the degradation became slower. At high concentrations of boric acid (30%), used to coat the CS film, the Tonset temperature was higher than for the neat CS films, which indicated an improved thermal stability. As was observed for chitosan, early degradation with decreased degradation rate were noted for CNF and CS-CNF films. The rate of degradation decreased with the boric acid coating. Indeed, except for the CS-CNF-BA 10% sample, the peak height corresponding to maximum weight loss rate decreased with the application of boric acid (see dTG curve Figure S7b and c for CNF and CS-CNF, respectively). This may be due to the same effect of boric acid as we observed earlier for bulk addition. The reason for the exception above is unclear. In all cases, boric acid increased the overall residue production, as expected. In comparison with films prepared by (bulk) mixing, coating with boric acid showed lower Tonset and Tmax (Table 3). This indicates that compared to the case when boric acid was added on the surface, the ‘mixing method’ promoted more extensive interactions between the components. Thus, the improved complexation, esterification and thermal annealing of boric acid with CS and CNF, improved the thermal stability of the films. The overall char production of the mixed and coated films was fairly similar.

13

Overall, the results agree with the fact that boric acid, a weak acid (pKa = 9.24), reacts with diols at low temperature (Köse, & Karan, 2012), resulting in boric acid-polysaccharide complexes. The stable bonds between hydroxyl and boron inhibits the mobility of surface oxides during combustion. Moreover, the trivalent B atom in boric acid has an empty p orbital that is very electrophilic and reacts rapidly with various nucleophiles to form complexes (Cotton & Wilkinson,1980); consequently, crosslinking structures can be formed between boric acid B(OH)3 and hydroxyl groups (-OH) (Ochiai, Shimizu, Tadekoro, & Murakami, 1981; Shibayama, Sato, Kimura, Fujiwara, & Nomura,1988). On the other hand, when combined with boric acid, amino groups react (Lundberg, Tinnis, Selander, & Adolfsson, 2014) and form coordinate bonding (Staubitz, Robertson, Sloan, & Manners, 2010), which influences the properties of chitosan. Mechanical properties. Tensile tests of the single component (CS, CNF) and bicomponent CSCNF films containing boric acid were performed at 65% relative humidity and room temperature (25 °C). Figure S8 shows representative stress−strain curves for the films. It was observed that intermediate additions of boric acid had the most positive effect on the tensile strength of CS, CNF and CS-CNF films (Figure 5a). Single CS and CNF film showed a decreasing trend in elongation upon addition of boric acid. The bicomponent CS-CNF films showed an increase in 30 % in strain with addition of BA up to 3% (Figure 5b). This increase was offset at 5% BA concentration while 10% addition of boric acid drastically decreased the elongation (77% for CS films and 60% for both the CNF and CS-CNF films), which implied that excess (non-interacting) boric acid existed in the system. The excess boric acid probably led to discontinuities in the CS-CNF matrix. The Young’s modulus, increased with the addition of boric acid. This can be due to bond formation between boron and nitrogen present in chitosan, which was clearly observed from the FTIR spectra (Figure S6). Boron also interacted with OH, which also made the film denser, as we observed from the film thickness (Figure S1), up to certain BA concentration level. With increasing concentration of boric 14

acid, the CNF films showed an increased Young’s modulus beyond those for the CS and CS-CNF films. Boric acid increased the intra- and inter- fiber hydrogen bonding and could also increase the adsorption of chitosan on the CNF, which may explain the improved Young’s modulus of the films. With boric acid addition, the rate of stiffening increased slowly in CNF and CS films. Beyond 10% concentration of boric acid, the Young’s modulus started to decrease. It was observed that with the addition of boric acid the Young’s modulus increased about 4.3, 2.8 and 0.65 times for CS, CNF and CS-CNF films, respectively. Bicomponent CS-CNF films displayed a slightly higher Young’s modulus than the neat CNF film but, with the addition of 30% boric acid, the CS-CNF film showed about 60% lower Young’s modulus than that for the neat CNF film, which may be due to the interference in bonding owing to the excess boric acid between chitosan and CNF at such high loading. The films of chitosan and cellulose nanofibrils can be used in many applications, such as flexible displays (Sehaqui, Liu, Zhou, & Berglund, 2010), loudspeaker membranes (Klemm, Heublein, Fink, & Bohn, 2005), flexible electronics (Huang et al., 2013), strain-sensor applications (Yan et al., 2014), light-harvesting platforms (Svagan et al., 2014), solar cells (Fang et al., 2014), batteries (Nystrom et al., 2009; Choi et al., 2014), and many biomedical and optical devices (Czaja, Young, Kawecki, & Brown, 2007; Nogi, Iwamoto, Nakagaito, & Yano, 2009). A high surface smoothness, transparency, mechanical strength and fire retardant capacity are all relevant properties in such applications and therefore the results in this work are expected to broaden the prospects for related hybrid systems.

CONCLUSIONS CS, CNF and CS-CNF films were prepared with different concentrations of boric acid. The fire retardant capacity, thermal stability and mechanical, and optical properties of the films were measured. Chitosan films showed a self-extinguishing fire retardant behavior due to the high 15

oxygen barrier properties. The addition of boric acid to chitosan slightly improved its fire retardancy. For the CNF films, boric acid was shown to be a very effective fire retardant. The flammability of the CNF film gradually decreased with increasing boric acid content. CNF film prepared with 30 w% boric acid showed self- extinguishing behavior during vertical flammability tests. The bicomponent CS-CNF film prepared with equal parts of each component showed very high fire retardancy. A synergistic fire retardancy effect due to addition of chitosan to cellulose was observed. In the absence of boric acid, the CS-CNF films showed self-extinguishing, fire retardant behavior, indicating that potential benefit of adding chitosan. Addition of boric acid increased the thermal stability of CS, CNF, and CS-CNF film and improved char formation. Compared to films coated with boric acid, adding boric acid during preparation produced films with higher thermal stability. The addition of boric acid, up to 5% concentration based on the total mass, was effective to improve the tensile strength of the CS, CNF and CS-CNF films, by crosslinking of boric acid with nitrogen and hydroxyl groups. Boric acid also improved the Young’s modulus of the films. Boric acid, though, had a negative effect on the elongation of the films. Overall, boric acid addition to single or bicomponent films comprising CNF and chitosan improved their fire retardancy, thermal stability, mechanical strength, transparency and smoothness, showing promise for many potential applications.

ACKNOWLEDGMENTS The authors are grateful to the Academy of Finland for funding support under the projects “3DBiomat” (BioFuture2025 program, projects 307485-7) and HYBER project 264677 under the Centres of Excellence Programme (2014–2019).

16

REFERENCES Alongi, J., Camino, G., & Malucelli, G. (2013). Heating Rate Effect on Char Yield from Cotton, Poly(Ethylene Terephthalate) and Blend Fabrics Carbohydrate Polymers, 92, 1327– 1334. Alongi, J., & Malucelli, G. (2015). Cotton flame retardancy: State of the art and future perspectives. RSC Advances, 5, 24239– 24263. Azeredo, H. M. C., Mattoso, L. H. C., Avena-Bustillos, R. J.,Filho, G. C., Munford, M. L., Wood, D., & Mchugh, T. H. J. (2010). Nanocellulose reinforced chitosan composite films as affected by nanofiller loading & plasticizer content. Journal of Food Science, 75, N1N7. Baitinger, W.F. (1982). Fire Retardance Treatment of Fabrics for the 80s: Smolder Resistant Cotton Treated with Boric Acid. Textile Research Journal, 52 (1), 82-86. Butler, B.L., Vergano, P.J., Testin, R.F., Bunn, J.M., & Wiles, J.L. (1996). Mechanical and barrier properties of edible chitosan films as affected by composition and storage. Journal of Food Science, 61(5), 953-956. Choi, K.H., Cho, S.J., Chun, S.J., Yoo, J. T., Lee, C. K., Kim, W., Wu, Q., Park, S.B., Choi, D.H., Lee, S.Y., & Lee, S.Y. (2014). Heterolayered, One-Dimensional Nanobuilding Block Mat Batteries. Nano Letters, 14, 5677–5686. Cotton, F. A., & Wilkinson, G. Advanced Inorganic Chemistry, 4th ed., John Wiley & Sons, New York, Chapter 9, 1980. Czaja, W.K., Young, D.J., Kawecki, M., & Brown, R.M. (2007). The Future Prospects of Microbial Cellulose in Biomedical Applications. Biomacromolecules, 8, 1–12.

17

Dasari, A., Yu, Z. Z., Cai, G. P., & Mai, Y. W. (2013). Recent developments in the fire retardancy of polymeric materials. Progress in Polymer Science, 38, 1357−1387. De Britto, D., & Campana-Filho, S. P. A (2004). Kinetic study on the thermal degradation of N,N,N-trimethylchitosan. Polymer Degradation and Stability, 84(2), 353–361. Desbrieres, J. (2004). Autoassociative natural polymer derivatives: the alkylchitosans. Rheological behaviour and temperature stability. Polymer, 45, 3285– 3295. Don, T.M., Hsu, S.C, & Chiu, W.Y. (2001). Structures and thermal properties of chitosan-modified poly(methyl methacrylate). Journal of Polymer Science Polymer Chemistry, 39, 1646– 1655. Fang, Z., Zhu, H., Yuan, Y., Ha, D., Zhu, S., Preston, C., Chen, Q., Li, Y., Han, X., Lee, S., Chen, G., Li, T., Munday, J., Huang, J., & Hu L. (2014). Novel Nanostructured Paper with Ultrahigh Transparency and Ultrahigh Haze for Solar Cells. Nano Letters, 14, 765– 773. Hirata, T., & Werner. K.E. (1987). Thermal analysis of cellulose treated with boric acid or ammonium phosphate in varied oxygen atmospheres. Journal of Applied Polymer Science, 33, 1533-1556. Horrocks, A. R. In Fire Retardant Materials, Horrocks, A. R., & Price, D., Eds., Woodhead Publishing, Ltd.:Cambridge, England. 2001, 128-181. Horrocks, A. R. (1996) Developments in flame retardants for heat and fire resistant textiles-the role of char formation and intumescence. Polymer Degradation and Stability, 54, 143-154. Hu, S., Song, L., Pan, H., & Hu, Y. (2013). Effect of a novel chitosan-based flame retardant on thermal and flammability properties of polyvinyl alcohol. Journal of Thermal Analysis and Calorimetry, 112, 859–864. 18

Huang, J., Zhu, H., Chen, Y., Preston, C., Rohrbach, K., Cumings, J., & Hu, L. (2013). Highly Transparent and Flexible Nanopaper Transistors. ACS Nano, 7, 2106–2113. ISO 9773:1998, Plastics - Determination of burning behavior of thin flexible vertical specimens in contact with a small-flame ignition source. Kandola, B. K., Horrocks, A. R., Price, D., & Coleman, G. V. (1996). Flame-Retardant Treatments of Cellulose and Their Influence on the Mechanism of Cellulose Pyrolysis. Journal of Macromolecular Science, Part C Polymer Reviews, 36 (4), 721-794. Khan, A., Khan, R. A., Salmieri, S., Le Tien, C., Riedl, B., Bouchard, J., & Lacroix, M. (2012). Mechanical and barrier properties of nanocrystalline cellulose reinforced chitosan based nanocomposite films. Carbohydrate Polymers, 90(4), 1601–1608. Klemm, D., Heublein, B., Fink, H.P., & Bohn, A. (2005). Cellulose: Fascinating Biopolymer and Sustainable Raw Material. Angew. Chem., Int. Ed, 44, 3358–3393. Köse, D., & Karan, B.Z. ( 2012). Mixed-ligand complexes of boric acid with organic biomolecules. Chemical Papers, 66, (1), 54–60. Kurita, K. (2001). Controlled functionalization of the polysaccharide chitin. Progress in Polymere Science, 26, 1921– 1971. Laufer, G., Kirkland, C., Cain, A. A., & Grunlan, J. C. (2012). Clay-chitosan nanobrick walls: completely renewable gas barrier and flame-retardant nanocoatings. ACS Applied Materials & Interfaces, 4, 1643–1649. Lundberg, H., Tinnis, F., Selander, N., & Adolfsson, H. (2014). Catalytic amide formation from non-activated carboxylic acids and amines. Chemical Society Reviews, 43, 2714−2742.

19

LeVan, S. L. (1984). Chemistry of fire retardancy: In The Chemistry of Solid Wood; Advances in Chemistry Series 207; Rowell, R. M., Ed.; American Chemical Society: Washington, DC, USA, Chapter 14, 531-574. Lyons, J.W. The chemistry and uses of fire retardants. New York: Wiley- Interscience, 1970. Moon, R. J., Martini, A., Nairn, J., Simonsen, J., Youngblood, J. (2011). Cellulose nanomaterials review: structure, properties and nanocomposites. Chemical Society Reviews, 40, 39413994. Nair, S.S., Zhu, JY., Deng, Y., & Ragauskas, A.J (2014). High performance green barriers based on nanocellulose. Sustainable Chemical Processes, 2:23. Nogi, M., Iwamoto, S., Nakagaito, A.N., & Yano, H. (2009). Optically Transparent Nanofiber Paper. Advanced Materials, 21, 1595–1598. Nystrom, G., Razaq, A., Stromme, M., Nyholm, L., & Mihranyan, A. (2009). Ultrafast All-polymer Paper-based Batteries. Nano Letters, 9, 3635–3639. Ochiai, H., Shimizu, S., Tadekoro,Y., & Murakami, I. (1981). Complex formation between poly(vinyl alcohol) and borate ion. Polymer, 22, 1456-1458. Österberg, M., Vartiainen, J., Lucenius, J., Hippi, U., Seppala, J., Serimaa, R., & Laine, J. (2013). A fast method to produce strong NFC films as a platform for barrier and functional materials. ACS Applier Materials & Interfaces, 5, 4640–4647. Peters, J.A (2014). Interactions between boric acid derivatives and saccharides in aqueous media: Structures and stabilities if resulting esters. Coordination Chemistry Reviews, 268, 122.

20

Rowell, R.M. The chemistry of solid wood: Advances in chemistry series 207. ACS. 1984. 531574. Saito, T., Kuramae, R., Wohlert, J., Berglund, L. A., & Isogai, A. (2013). An Ultrastrong Nanofibrillar Biomaterial: The Strength of Single Cellulose Nanofibrils Revealed via Sonication-Induced Fragmentation. Biomacromolecules, 14, 248-253. Sehaqui, H., Liu, A., Zhou, Q., & Berglund, L.A. (2010). Fast Preparation Procedure for Large, Flat Cellulose and Cellulose/Inorganic Nanopaper Structures. Biomacromolecules, 11, 2195–2198. Shafizadeh, F., Bradbury, A.G.W., DeGroot, W.F., & Aanerud, T.W. (1982). Role of inorganic additives in the smoldering combustion of cotton cellulose. Industrial and Engineering Chemistry Product Research and Development, 21, 97-101. Shibayama, M., Sato, M., Kimura, Y., Fujiwara, H., & Nomura, S. (1988). 11B n.m.r. study on the reaction of poly(vinyl alcohol) with boric acid. Polymer, 29,336-340. Staubitz, A., Robertson, A. P., Sloan, M. E., & Manners, I. (2010). Amine and Phosphine-Borane Adducts: New Interest in Old Molecules. Chemical Reviews, 110, 4023−4078. Svagan, A. J., Busko, D., Avlasevich, Y., Glasser, G., Baluschev, S., & Landfester, K. (2014). Photon Energy Upconverting Nanopaper: A Bioinspired Oxygen Protection Strategy. ACS Nano, 8, 8198–8207. Syverud, K., & Stenius, P. (2009). Strength and barrier properties of MFC films. Cellulose, 16, 75 85. Swain, S.K., Dash, S., Behera, C., Kisku, S.K., & Behera, L. (2013). Cellulose nanobiocomposites with reinforcement of boron nitride: study of thermal, oxygen barrier and chemical resistant properties. Carbohydrate Polymers, 95,728-732.

21

Toivonen M.S., Kurki-Suonio S., Schacher F.H., Hietala S., Rojas O.J., & Ikkala O. (2015). Waterresistant, transparent hybrid nanopaper by physical crosslinking with chitosan. Biomacromolecules, 16, 1062-1067 Wang, Q., Li, J., & Winandy, J.E. (2004). Chemical mechanism of fire retardance of boric acid on wood. Wood Science and Technology, 38, 375-389. Wicklein, B., Kocjan, D., Carosio, F., Camino, G., & Bergström, L. (2016). Tuning the Nanocellulose–Borate Interaction to Achieve Highly Flame Retardant Hybrid Materials. Chemistry of Materials, 28 (7), 1985–1989. Yan, C., Wang, J., Kang, W., Cui, M., Wang, X., Foo, C. Y., Chee, K. J., & Lee, P. S. (2014). Highly Stretchable Piezoresistive Graphene Nanocellulose Nanopaper for Strain Sensors. Advanced Materials, 26, 2022–2027. Yano, H., & Nakahara, S. (2004). Bio-composites produced from plant microfiber bundles with a nanometer unit web-like network. Journal of Materials Science, 39, 1635-1638. Zhang, Y., Nypelö, T., Salas, C., Arboleda, J., Hoeger, I. C., & Rojas, O. J. (2013). Cellulose Nanofibrils: From Strong Materials to Bioactive Surfaces. Journal of Renewable Materials, 1(3), 195-211.

22

Figure 1. UV-vis spectra corresponding to (a) CS, (b) CNF and (c) CS-CNF films with boric acid added at different concentrations during preparation (the added numerical figures represent the mass percent of boric acid with respect to the total dry mass). Photographs of the strips used in (a-c) are also displayed in (d), under similar illumination.

23

Figure 2. Photo snapshots of the flammability test after 1 sec of ignition (left set of photos), and the residues collected after the test (right set of photos), for films with different boric acid loadings.

Figure 3. TG and dTG thermograms of films of CS (a), CNF (b) and CS-CNF (c) with various boric acid concentration and measured in nitrogen atmosphere. The numbers added by the thermograms indicate the concentration of boric acid. 24

Figure 4. TG and dTG thermograms measured in oxygen for CS (a), CNF (b), and CS-CNF (c) films loaded with different concentration of boric acid. The numbers added by the profiles indicate the initial concentration of boric acid in the films.

25

Figure 5. (a) Tensile strength, (b) Maximum strain % and (C) Young`s modulus of the CS, CSCNF, CNF films with different concentrations of boric acid.

26

Table 1. Flammability test for CS, CNF and CS-CNF films with different content of boric acid (BA). Note: the standard deviation in the measurements of residue % is ~ ± 2% Film

CS

CNF

CS-CNF

BA, % 0 1 3 5 10 20 30 0 1 3 5 10 20 30 0 1 3 5 10 20 30

Burning time, s 3 4 4 4.5 3 3.5 2 -

After-glow time, s 2 1 0.5 0.2 0 0 0 -

Residue, % 79 82 85 87 87 92 95 0 0 19 21 36 59 86 85 86 88 88 85 88 92

The sign “-” is used to indicate that no flame was produced.

27

Table 2. Thermogravimetric data of CS, CNF and CS-CNF films with different concentrations of boric acid measured in nitrogen and oxygen atmosphere.

Film CS

CNF

CS-CNF

BA, % 0 3 5 10 20 30 0 1 3 5 10 20 30 0 1 3 5 10 20 30

Tonset, oC N2 232 221 222 253 260 257 278 277 277 280 295 298 299 248 188 189 189 244 228 231

Tmax, oC O2 219 220 222 226 228 262 266 270 273 271 281 281 182 181 181 187 234 239

N2 254 233 233 263 272 260 358 328 320 314 318 320 328 259 199 200 199 248 234 234

O2 232 233 233 242 235 311 319 314 312 313 315 313 200 200 198 197 246 243

NA NA NA NA NA 414 447 456 459 450 463 462 NA NA NA NA NA NA

Total residue, % at 600 oC N2 O2 45 4 44 8 47 11 48 21 51 33 50 26 3 33 6 36 6 38 11 42 5 45 26 46 32 42 8 36 8 43 9 43 9 42 22 49 50 38

The sign “-” is used to indicate that no test were done at this concentration

28

Table 3. Comparison of thermogravimetric data in nitrogen atmosphere for films prepared by coating and mixing with BA. Films

CS

CNF

CS-CNF

BA content, %

Coating Tonset OC

Tmax C

0 3 10 30 0 3 10 30 0 3 10 30

232 217 225 240 278 272 282 282 248 205 206 205

254 240 245 245 358 321 316 315 259 215 216 216

O

Mix Total residue, % Tonset OC at 600OC 45 232 46 221 48 253 54 257 26 278 29 277 43 295 46 299 42 248 38 189 42 244 48 231

Tmax OC 254 233 263 260 358 320 318 328 259 200 248 234

Total residue, % at 600OC 45 44 48 50 26 36 42 46 42 42 42 50

29