3920
Langmuir 2009, 25, 3920-3923
Letters Non-Electrostatic Building of Biomimetic Cellulose-Xyloglucan Multilayers Bruno Jean,*,† Laurent Heux,† Fre´de´ric Dubreuil,† Ge´rard Chambat,† and Fabrice Cousin‡ Centre de Recherche sur les Macromole´cules Ve´ge´tales (CERMAV-CNRS), BP 53, 38041 Grenoble Cedex 9, France, and Laboratoire Le´on Brillouin CEA-CNRS, Saclay 91191 Gif-sur-YVette Cedex, France ReceiVed August 27, 2008. ReVised Manuscript ReceiVed October 20, 2008 Layer-by-layer assembly was used to build thin films, consisting of multiple layers alternating cellulose nanocrystals and xyloglucan, benefiting from the strong non-electrostatic cellulose-xyloglucan interaction. Data from atomic force microscopy and neutron reflectivity showed that these well-defined films exhibited a thickness increasing linearly with the number of layers, without increase in surface roughness. These “green” nanocomposite films, reminiscent of plant cell wall, are composed of a regular stack of single layers of cellulose nanocrystals separated by very thin xyloglucan spacers. Such architecture differs from the one formed by cellulose/polycations multilayers, where the cellulose phase itself consists of a double layer.
Introduction Increasing environmental awareness is prompting scientists and industrialists to develop strategies for environmental sustainability that can be in some cases inspired by the living world. Many natural structures are complex assemblies of relatively simple elements, especially designed for specific functions. In this framework, green nanocomposite films made entirely from renewable resources could offer an attractive alternative for many applications, including biomedical or optical devices. Biopolymers and in particular structural polysaccharides may thus emerge as a very interesting class of natural building blocks. Among polysaccharides, cellulose especially possesses outstanding physical properties such as high mechanical resistance and low density, and moreover benefits from tremendous natural abundance and noncompetition regarding food. Cellulose occurs in ViVo in plant cell walls in the form of slender microfibrils, which are combined with other biomacromolecules to form supramolecular assemblies where they act as reinforcer. It is well-accepted that hemicelluloses associate with cellulose to create an interconnected network responsible for load-bearing properties of these natural nanocomposites.1 Xyloglucans (XGs) are the major hemicelluloses in the primary cell wall of dicotyledonous plants. These neutral polysaccharides have a β(1f4) linked D-glucopyranose cellulose-like backbone decorated with dangling R-D-xylosyl and β-D-galactosyl units.2 This substitution is responsible for a certain solubility in water and preferred interaction with the cellulose crystals. Although of crucial importance, the nature of the binding of xyloglucans to cellulose is still uncertain and may occur through concomitant contributions * Corresponding author. E-mail:
[email protected]. † Centre de Recherche sur les Macromole´cules Ve´ge´tales (CERMAVCNRS). Note: CERMAV-CNRS is affiliated with Universite´ Joseph Fourier and member of the Institut de Chimie Mole´culaire de Grenoble. ‡ Laboratoire Le´on Brillouin CEA-CNRS. (1) Carpita, N. C.; Gibeaut, D. M. Plant J. 1993, 3, 1–30. (2) Vincken, J. P.; York, W. S.; Beldman, G.; Voragen, A. G. J. Plant Physiol. 1997, 114, 9–13. (3) Hanus, J.; Mazeau, K. Biopolymers 2006, 82, 59–73.
of van der Waals forces, polar interactions, and hydrogen bonding.3 In this letter, we propose to build new biomimetic cellulose/ hemicellulose nanocomposite films made of alternating layers of cellulose nanocrystals and hemicellulose and investigate their morphology. The strong cellulose-xyloglucan interactions have previously been exploited on the macroscopic scale with applications in the textile, food, and pulp and paper industries.4,5 Our objective here is to focus on the nanoscopic scale to build green nanocomposite films. Cellulose nanocrystals (CNXLs) are obtained from sulfuric acid hydrolysis of microfibrils that yields negatively charged short crystalline rods with a cross section between 3 and 20 nm and a length between 100 nm and several micrometers, depending on the biological origin.6 To prepare such composites, we have used the versatile yet simple layerby-layer (LbL) assembly technique.7,8 This method originally consists of a sequential adsorption on a solid substrate of oppositely charged compounds. It was first developed for polyelectrolytes and then expanded to include various nanobuilding blocks such as proteins or carbon nanotubes.9,10 LbL assembly has also been extended to neutral systems where electrostatic interactions are replaced by different attractive forces such as hydrogen bonding,11 hydrophobic interactions,12 or host-guest interaction.13 The architecture of our CNXLs/XGs multilayer films was characterized by combining neutron reflectivity (NR) measurements and atomic force microscopy (4) Zhou, Q.; Rutland, M. W.; Teeri, T. T.; Brumer, H. Cellulose 2007, 14, 625–641. (5) Kumar, C. S.; Bhattacharya, S. Crit. ReV. Food Sci. Nutr. 2008, 48, 1–20. (6) Revol, J. F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray, D. G. Int. J. Biol. Macromol. 1992, 14, 170–172. (7) Decher, G. Science 1997, 277, 1232–1237. (8) Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210, 831–835. (9) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117–6123. (10) Rouse, J. H.; Lillehei, P. T. Nano Lett. 2003, 3, 59–62. (11) Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30, 2717–2725. (12) Serizawa, T.; Kamimura, S.; Kawanishi, N.; Akashi, M. Langmuir 2002, 18, 8381–8385. (13) Suzuki, I.; Egawa, Y.; Mizukawa, Y.; Hoshi, T.; Anzai, J. Chem. Commun. 2002, n/a, 164–165.
10.1021/la802801q CCC: $40.75 2009 American Chemical Society Published on Web 11/06/2008
Letters
Langmuir, Vol. 25, No. 7, 2009 3921
(AFM) observations. The morphological details of the films are discussed in light of recent studies concerning CNXLs/polycations multilayers.14-20
Experimental Section Materials. The polycations poly(ethylene imine) (PEI, Mw ) ∼25 000) and poly(allylamine hydrochloride) (PAH, Mw ) ∼70 000) and the polyanion poly(sodium 4-styrene sulfonate) (PSS, Mw ) ∼70 000) as well as other chemicals were purchased from SigmaAldrich and used as received. Cotton linters were provided by RhoˆnePoulenc Tubize Plastics (Belgium) and used as the cellulose source without any further purification. Cellulose Nanocrystals. Deuterated cellulose nanocrystals suspensions were prepared in two steps. First, deuterated cotton linters were prepared using the hydrothermal intracrystalline deuteration method described by Nishiyama et al.21 Second, deuterated nanocrystals were generated by acid hydrolysis of deuterated linters following the method by Revol et al.6 Experimental details are given elsewhere.18 The final 7 wt % nanocrystals suspension was obtained by osmotic stress of the initial 3.2 wt % stock suspension using a 35 wt % poly(ethylene glycol) solution. Xyloglucan. Tamarindus xyloglucans were prepared from decorticated seeds of Tamarindus indica L. by extracting the ground seed with ethanol-water solution (8:2, v/v, 6 h, 80 °C), and then with water (1 h, 100 °C). After centrifuging, the residue was extracted again twice with water in the same manner. The water extracts were combined and the resulting xyloglucan recovered by precipitation with ethanol (3 vol). The xyloglucan fraction was then ultrafiltered on an Amicon membrane limit of >104 g mol-1 (Amicon, Beverly, USA). The weight-average and number-average molar masses determined by size exclusion chromatography coupled with light scattering are equal to Mw ) 105 000 and Mn ) 61 000, respectively. Layer-by-Layer Assembly. Multilayers were deposited on polished silicon 〈100〉 wafers, which were first cleaned in the highly corrosive Piranha H2SO4/H2O2 mixture (70:30 v/v) at room temperature for 30 min, followed by intensive rinsing with water, resulting in a negatively charged surface. Film formation was achieved by simple sequential dipping for 20 min in the solution or suspension to be deposited with intermediate rinsing steps with distilled water for 5 min. Prior to the first cellulose nanocrystals deposition, a pure polyelectrolyte multilayer primer, PEI/(PSS/PAH)n with n ) 1 or 3 for AFM or reflectivity experiments, respectively, was adsorbed on the silicon substrate using 4 g/L solutions and NaCl 1 M for PSS and PAH solutions and a 2 g/L solution for PEI. It was previously shown that a first layer of PEI is well-suited to obtain a uniform coating and thus favors subsequent multilayer buildup.22 Furthermore, the primer coating reduces the influence of the substrate on CNXLs adsorption. For AFM measurements, n was equal to 1, whereas for neutron reflectivity measurements, n was chosen equal to 3 to give a thickness that is large enough to fit in the Q-range probed. In consequence, there is an increase in the number of observable oscillations (known as Kiessig fringes) related to the thin film thickness, and thus the data analysis is facilitated and its quality (14) Podsiadlo, P.; Choi, S.-Y.; Shim, B.; Lee, J.; Cuddihy, M.; Kotov Nicholas, A. Biomacromolecules 2005, 6, 2914–2918. (15) Cranston, E. D.; Gray, D. G. Biomacromolecules 2006, 7, 2522–2530. (16) Cranston, E. D.; Gray, D. G. Sci. Technol. AdV. Mater. 2006, 7, 319–321. (17) Podsiadlo, P.; Sui, L.; Elkasabi, Y.; Burgardt, P.; Lee, J.; Miryala, A.; Kusumaatmaja, W.; Carman, M. R.; Shtein, M.; Kieffer, J.; Lahann, J.; Kotov, N. A. Langmuir 2007, 23, 7901–7906. (18) Jean, B.; Dubreuil, F.; Heux, L.; Cousin, F. Langmuir 2008, 24, 3452– 3458. (19) Aulin, C.; Varga, I.; Claessont, P. M.; Wagberg, L.; Lindstrom, T. Langmuir 2008, 24, 2509–2518. (20) Wagberg, L.; Decher, G.; Norgren, M.; Lindstrom, T.; Ankerfors, M.; Axnas, K. Langmuir 2008, 24, 784–795. (21) Nishiyama, Y.; Isogai, A.; Okano, T.; Mueller, M.; Chanzy, H. Macromolecules 1999, 32, 2078–2081. (22) Kolasinska, M.; Warszynski, P. Appl. Surf. Sci. 2005, 252, 759–765.
Figure 1. Neutron reflectivity curves of multilayer samples PEI/(PSS/ PAH)3/(CNXLs/XG)m with m ) 1.5 (9), 2.5 (1), 4 ([), and 5.5 (b). The solid line is the Fresnel reflectivity.
improved. Previous experiments showed that the primer layer had a thickness of 4 nm per PSS/PAH bilayer and a low roughness.18,23 (CNXLs/XG)m multilayers with 0.5 e m e 5.5 were then adsorbed on the primer coating from 7% CNXLs suspension and 0.1 wt % XG aqueous solution. Half-integer values of m thus correspond to CNXLs terminated multilayers, whereas integer values correspond to XG capped films. After the last rinsing step, the samples were dried for 30 min at 60 °C. Neutron Reflectivity. Specular neutron reflectivity experiments were carried out on the time-of-flight reflectometer EROS at the Laboratoire Le´on Brillouin, CEA Saclay, France.24 To access a broad momentum transfer range, data were collected at two different fixed angles, 0.93° and 2°, with a neutron white beam covering wavelengths from 3 to 25 Å. The final accessible Q-range is 0.008-0.1 Å-1. Data were analyzed using a “box” model consisting of dividing the thin film into a series of layers. Each layer was characterized by a finite thickness, scattering length density, and interfacial roughness with the neighboring layer. Reflectivity curves calculated using the optical matrix method take into account the resolution of the spectrometer.25,26 Atomic Force Microscopy. AFM experiments were performed on a Pico plus (Molecular Imaging) commercial instrument. Topography pictures were obtained using contact mode with Mikromash CSC 36 tips. Data treatment (height measurements after baseline correction only) and presentation were realized with the help of Gwyddion Software. Thickness was derived from scratchheight analysis, and rms roughness was evaluated from 5 × 5 µm2 images.
Results and Discussion AFM and NR were used as complementary techniques to characterize the samples prepared by LbL assembly and to get insight into the structure of the multilayers. Figure 1 shows the neutron reflectivity spectra corresponding to PEI/(PSS/PAH)3/ (CNXLs/XG)m samples with m ) 1.5, 2.5, 4, and 5.5. These curves exhibit very pronounced oscillations indicating welldefined total thickness of the multilayers. The spectra can be divided into two regions. For Q < 0.045 Å-1, Kiessig fringes shifting to lower Q values when m increases are observed. This is the signature of an increase of the total thickness with the number of (CNXLs/XG) bilayers. Conversely, for Q > 0.045 Å-1 a Bragg peak with quasi m-independent amplitude and (23) Von Klitzing, R.; Wong, J. E.; Jaeger, W.; Steitz, R. Curr. Opin. Colloid Interface Sci. 2004, 9, 158–162. (24) Ott, F.; Cousin, F.; Menelle, A. J. Alloys Compd. 2004, 382, 29–38. (25) Penfold, J.; Thomas, R. K. J. Phys.: Condens. Matter 1990, 2, 1369– 1412. (26) Higgins, J. S.; Benoıˆt, H. C. Polymers and Neutron Scattering; Clarendon Press: Oxford, 1994.
3922 Langmuir, Vol. 25, No. 7, 2009
Letters
Figure 2. Total thickness of multilayers PEI/PSS/PAH/(CNXLs/XG)m measured by AFM (b) and multilayers PEI/(PSS/PAH)3/(CNXLs/XG)m measured by neutron reflectivity (O) and AFM (×). The solid line is a guide to the eye, whereas dash-dotted and dashed lines are linear fit considering only half-integer values of m.
position is witnessed. Both features unambiguously prove that well-ordered films with repeat units assembled to give a welldefined architecture were produced. To quantitatively characterize the (CNXLs/XG) multilayers, neutron reflectivity data were fitted using a multiple-layers model. This procedure finally gives the total thickness of the multilayered films as well as the composition of each layer. Figure 2 shows the evolution of the total thickness of PEI/(PSS/PAH)3/(CNXLs/XG)m films deduced from NR. The last three points were also cross-checked using AFM. These data are compared to the total thickness of PEI/(PSS/PAH)1/(CNXLs/ XG)m films, which only differ by the thickness of the primer layer, derived from AFM measurements. It can first be noted that both sets of data display a parallel behavior, which evidence the consistency of both techniques. AFM data for n ) 1 display a staircase-like behavior, i.e., they show an increase of the thickness when CNXLs are deposited on the surface and a roughly constant thickness when XGs are added. Despite the limited number of m values, neutron reflectivity data for n ) 3 follow the same trend. Dashed and dash-dotted lines in Figure 2 are linear fits taking into account half-integer values of m, i.e., cellulose nanocrystals-terminated films. The very good agreement between data and fits clearly demonstrate a linear increase of the thickness of the multilayered films with m in both cases. The corresponding slopes are 7, 7.8, and 8.3 nm for AFM, neutron reflectivity, and AFM on the neutron reflectivity samples, respectively. With the thickness of cotton cellulose nanocrystals being between 6 and 7 nm,27 AFM and NR data both show that the adsorption of CNXLs on top of an XG layer obviously occurs as a single layer. Very little variation of the thickness of the film is observed when XG is deposited, suggesting the adsorption of a very thin XG layer on top of the nanocrystals that cannot be detected within the experimental accuracy. This result is consistent with molecular dynamics simulation investigations3 showing that XG macromolecules without long side chains tend to adsorb on cellulose microfibrils with a flat conformation with all the sugar residues interacting with the surface, thus resulting in a thin XG layer. The presence of a Bragg peak in the reflectivity spectra in Figure 1 reflects the existence of a building unit repeated at regular intervals to form the multilayer. Here, this Bragg peak can be observed thanks to the large scattering length density contrast between deuterated CNXLs and protonated XGs. In real (27) Elazzouzi-Hafraoui, S.; Nishiyama, Y.; Putaux, J. L.; Heux, L.; Dubreuil, F.; Rochas, C. Biomacromolecules 2008, 9, 57–65.
Figure 3. 5 × 5 µm2 AFM topography image of sample PEI/PSS/ PAH/(CNXLs/XG)1.5.
Figure 4. Rms roughness of multilayers PEI/PSS/PAH/(CNXLs/XG)m measured by AFM.
space, the position of the Bragg peak at Q ) 0.07 Å-1 corresponds to the thickness of the repeat unit of the multilayer. Even if a precise value is difficult to assess from the experimental curves, a repetition value of about 9 nm can be roughly estimated, which is slightly larger than the lateral size of cotton cellulose nanocrystals. This result is in perfect agreement with a repeat unit of the multilayer consisting of the addition of a single cellulose nanocrystals layer and a thin XG layer. The analysis of the NR data from the shifts of the oscillations as well as from the Bragg peak position are thus fully consistent. Both features show that the (CNXLs/XG) films are composed of stacked CNXLs single layers separated by thin hemicellulose layers. The fitting parameters suggest a surface density for the adsorbed CNXLs layer equal to 40-45%. This close packing can be further evidenced by considering the AFM topography image obtained for sample PEI/(PSS/PAH)3/(CNXLs/XG)1.5 (Figure 3). Figure 3 indeed shows that homogeneous CNXLs layers with high density are obtained. The CNXLs adsorption density is not sensitive to the number of (CNXLs/XG) bilayers (data not shown). CNXLs/polycation multilayers prepared by spin-coating or dipping have recently been investigated using different techniques. With PAH as the polycation, it has been shown by Cranston and Gray using ellipsometry15 and by us using AFM and NR18 that
Letters
the adsorption of CNXLs from cotton occurs as a double layer. These two layers have different packing densities (namely, 50% and 25% for the lower and upper layer, respectively).18 This feature results from the conjunction of enthalpic forces related to complex formation and entropic forces associated with counterion release. Interestingly, the present study shows that, when associated to the neutral XGs, CNXLs adsorption occurs as a single layer. Here, since electrostatic interactions between the two species are absent, there is no entropic gain associated with counterion release, and the only driving force for multilayer buildup is the enthalpic gain linked to the CNXLs/XGs interactions. Thus, the film formation occurs single layer by single layer, resulting in a very well defined architecture, as reflected by the low surface roughness of the samples. As shown in Figure 4, AFM measurements indicate a roughly constant low rms roughness of about 3.5 nm, up to deposition of 5 CNXLs layers. This result again contrasts with the surface roughness increase in the CNXLs/PAH multilayers when the number of layers was increased. This disorganization was attributed to a
Langmuir, Vol. 25, No. 7, 2009 3923
defect addition mechanism associated with the adsorption of CNXLs as a double layer.18
Conclusion The present letter shows, for the first time to our knowledge, that it is possible to prepare cellulose nanocrystals/xyloglucan multilayersusingthestrong,nonelectrostatic,cellulose-hemicellulose interaction. The resulting green composite displays a wellordered architecture that differs from previously studied CNXLs/polycations multilayers. Here, the interaction forces involved result in the periodical repetition of a building block composed of one CNXLs single layer and a thin XG spacer. Remarkably, in such a system, a low surface roughness is kept, despite the increase of the number of layers and the use of polydisperse colloids. Further developments will include evaluation of the mechanical properties of these nanocomposite films based on renewable resources and reminiscent of the natural structures. LA802801Q
