1D FT-IR and dynamic 2D FT-IR spectroscopy

1D FT-IR and dynamic 2D FT-IR spectroscopy

(Parte 1 de 2)

Molecular interactions in bacterial cellulose composites studied by 1D FT-IR and dynamic 2D FT-IR spectroscopy

Marta Kacurakova,a Andrew C. Smith,a Michael J. Gidley,b Reginald H. Wilsona,* aNorwich Laboratory, Institute of Food Research, Norwich Research Park, Colney Lane, Norwich NR47 UA, UK bUnile er Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK

Received 19 July 2001; received in revised form 21 March 2002; accepted 10 April 2002


Specific strain-induced orientation and interactions in three Acetobacter cellulose composites: cellulose (C), cellulose/pectin (CP) and cellulose/xyloglucan (CXG) were characterized by FT-IR and dynamic 2D FT-IR spectroscopies. On the molecular level, the reorientation of the cellulose fibrils occurred in the direction of the applied mechanical strain. The cellulose-network reorientation depends on the composition of the matrix, including the water content, which lubricates the motion of macromolecules in the network. At the submolecular level, dynamic 2D FT-IR data suggested that there was no interaction between cellulose and pectin in CP and that they responded independently to a small amplitude strain, while in CXG, cellulose and xyloglucan were uniformly strained along the sample length. © 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Cellulose composites; FT-IR spectroscopy; Dynamic 2D FT-IR spectroscopy; Linear stretching

1. Introduction

Cellulose is composed of -D-glucopyranose units joined by (1 4)-glycosidic links, and is the primary structural element of the cell wall: it has a high molecular weight and crystallinity.1 Xyloglucan is the major hemicellulose component in primary cell walls, with chains of (1 4)- -D-glucan with xylosyl units linked to the glucosyl units in the C-6 position. Pectin is a term for a group of heterogeneous polysaccharides whose backbone consists of (1 4)-linked -D-galacturonic acid repeating-units. Bacterial (Acetobacter xylinum) cellulose-based composites containing xyloglucan or pectin have been shown to possess organizational features similar to those observed in primary plant cell walls.2–4

Infrared (IR) spectroscopy has been extensively used in cellulose research and IR band assignment, orientation data,5,6 and structural details7,8 have been produced. Polymer stiffness is considerably enhanced by molecular orientation, and polarized IR spectroscopy can be used to study orientation induced by mechanical strain and to characterize the segmental mobility of polymers under the influence of an external perturbation.9 Dynamic 2D FT-IR spectroscopy has been used to unravel the IR bands affected by deformation and the orientation of submolecular groups in cellulose I in order to probe the relationship between hydrogen bonding and cellulose structure10–12 in dry spruce-pulp samples. We have recently reported the first application of linear stretching FT-IR microscopy and 2D FT-IR spectroscopy13 to functional cell walls in onion epidermis. However, it is necessary to determine the 2D cross-correlation peak frequencies for pure cellulose and its composites with other cell-wall biopolymers in order to establish an interpretation of cross peaks in hydrated systems.

Here we report the results of an FT-IR study of mechanically strained hydrated cellulose composites and interpret the data in terms of molecular mobility and interactions. Well-characterized model systems were used for detailed studies to allow interpretations to be applied to polymers of intact plant cell walls. Polarized IR spectra allow the determination of molecular alignment and contribute towards an understand- ing of the structure–function relationship in biopoly-

* Corresponding author. Tel.: +4-1603-255000. E-mail address: reg.wilson@bbsrc.ac.uk (R.H. Wilson).

mer mixtures at different water contents. The analysis of the dynamic 2D FT-IR spectra and two-dimensional correlations are shown for samples measured with a polarized IR source.

2. Experimental

Materials.—The Acetobacter cellulose samples cellulose (C), cellulose/pectin (CP, containing 20% apple pectin of 67% degree of methyl esterification), and cellulose/xyloglucan (CXG, containing 35% of tamarind xyloglucan) were prepared as described previously.2–4 Fermentation duration of 24 h gave 10–20 m thick films, which were still measurable with IR absorbance up to about 2 units in transmission, yet strong enough to be mounted in the clamps of the stretching device. The weight fraction of different cellulose morphologies in the C and CP samples has been reported as 57% I , 25% I crystalline, and 18% non-crystalline,4 while in CXG the corresponding values were 21% I , 32% I crystalline, and 47% noncrystalline.2,3

Water content.—The water content was gravimetrically determined after the cellulose composites were equilibrated at discrete relative humidity (RH) values.13 The samples at 100% relative humidity (RH) showed (Table 1) water contents (w/w) of: 42% for C, 61% for CP, and 94% for CXG. Samples used for dynamic 2D and linear stretching experiments were measured at 97% RH (defined as wet samples, w-samples) and certain samples were measured after partial drying at 84% RH (defined as moist, m-samples) as shown in Table 1. Samples dried below 76% RH were effectively dry, brittle, and not suitable for FT-IR mechanical investigation.

FT-IR spectroscopy.—A Bio-Rad FTS 6000 (Cambridge, MA) spectrometer, equipped with a liquid nitrogen cooled MCT (HgCdTe) detector, was used to collect transmission spectra. In linear stretching exper- iments, 128 scans were co-added before Fourier transformation with the IR beam polarized parallel or perpendicular to the stretching direction using a KRS 5 (TlBr and TlI mixed crystal) wire-grid polarizer (Graseby Specac, UK). The resolution was 8 cm−1 and three to five replicates were measured in order to ensure reproducibility of measurement. The spectrometer was controlled by a computer running WIN-IR PRO (Bio-Rad) version 2.9.

Linear stretching.—Polarized FT-IR spectra were collected from oriented samples. The composite samples were washed in water, cut to 15×10 m and mounted between the jaws of a Bio-Rad polymer stretcher. The sample was stretched ( l is the change of length) in steps of 3% strain, ,( l/l) from 0 to 3%.

Infrared dichroism.—Infrared dichroism reflects the mean orientation of the transition moments of the corresponding vibrational modes. Anisotropy following an applied deformation is characterized by the dicular polarization.14–16 The cellulose composites were studied as a function of uniaxial strain. For a polymer network, the segmental orientation (F) detected by IR spectroscopy17 expressed in terms of the dichroic ratio is: F=C[(R−1)/(R+2)], where C= (2 cot2 +2)/(2 cot2 −1). For a given absorption band, is the angle between the transition moment vector of the vibrational mode and a directional vector characteristic of the chain segment. The glycosidic (C O C) link (band at 1162 cm−1) is approximately co-aligned with the molecular long axis1,6 and the stretching direction. In this case, =0° and the equation reduces to: F=(R−1)/(R+2).

Data analysis.—In order to analyze the strain-induced IR spectral changes, A and F of selected bands were used to characterize the degree of orientation of cellulose in the individual composites9,14,17 in correlation with uniaxial strain. One of the effects of stretching is a change in the thickness of the polymer film. In order to minimize resultant error in A, each polarized spectrum was normalized to the corresponding unstretched IR absorption in the (CH) region at (1400–1300 cm−1) where intensity is not orientation dependent.18 For molecular orientation analysis A was used, in which case the bands from vibrational dipoles parallel to the polymer-chain axis appeared with positive intensity, whereas those that are perpendicular were of negative intensity.15

Dynamic 2DF T-IR spectroscopy.—In dynamic 2D

FT-IR, a small-amplitude oscillatory strain is applied to a sample and the resulting spectral changes are measured as a function of time.13 The dynamic perturbation generates directional changes in the transi-

Table 1 Sorption isotherm data

CXGRelative humidity (%) SamplesCC P

Water content (w/w%) in the cellulose composite samples: cellulose (C), cellulose/pectin (CP) and cellulose/xyloglucan (CXG).

Table 2 Assignment and orientation of cellulose, pectin and xyloglucan FT-IR absorption bands a as they were observed in the cellulose composites

Bond, orientationFrequency (cm−1) OriginAssignment

Key: IR vibrations: , stretching; , bending; w, wagging; s symmetric; as asymmetric; C, cellulose; P, pectin; XG, xyloglucan; , parallel; , perpendicular orientation.

a IR band assignment is based on literature using Refs. 5,6,13,21,2.

tion moments of functional groups whose relaxations then depend upon inter- and intramolecular couplings. The resulting dynamic absorbance spectra, which vary sinusoidally with the stretching frequency, are deconvoluted into two separate spectra: the in-phase (IP) spectrum reveals the re-orientational motions of electrical dipole moments occurring simultaneously, and the quadrature (Q) spectrum those which are /2 out-ofphase with respect to the applied strain. From the IP and Q spectra, synchronous 2D correlation spectra can be obtained with correlation peaks at wavenumber coordinates where the IR signal responses are in-phase with each other.16,19,20

Dynamic 2D FT-IR spectra were measured on samples ( 15×10 m) mounted in a Bio-Rad polymer stretcher and pre-stretched to 21% strain. The spectra were acquired on a BioRad FTS 6000 equipped with an MCT detector. The IR beam was polarized parallel or perpendicular to the direction of stretch using a KRS 5 wire-grid polarizer. Sample modulation frequencies (SMF) of 5, 10, 16, 20, 25, and 50 Hz, were used with an amplitude of 100 m. Two scans at resolution 8 cm−1 were co-added and the IP and Q spectra were ratioed to the static spectra. 2D correlation spectra were constructed using MATLAB software, version 6.0 from the IP and Q spectra using cross-correlation analysis.20

3. Results and discussion

Linear stretching experiments were performed to examine the changes in infrared absorption spectra of samples with applied uniaxial strain. From the polarized FT-IR spectra, dichroic difference, A, was calculated to evaluate the fibril reorientation, and dichroic ratio, R, was used to calculate the segmental orientation, F, of cellulose in order to compare the cellulose orientation as a function of strain in the different composites. Dynamic 2D FT-IR small-amplitude oscillatory strain was applied to the samples to study the inter- and intramolecular couplings at the sub-molecular level. 1DF T-IR, transmission spectroscopy.—The FT-IR spectral band assignments based on literature data are listed in Table 2. The absorption spectra of the cellulose composites C, CP, and CXG showed sharp IR bands across the whole 4000–800 cm−1 region (Fig. 1). The intense absorption and high cellulose content compared to that of xyloglucan or pectin caused the CP and CXG spectra to be dominated by cellulose bands. In CP and CXG spectra the non-cellulose polymers could not easily be distinguished, but in CP, pectin bands were observed at 1738 and 833 cm−1 and in CXG xyloglucan bands were seen at 1371 and 1317 cm−1 with the anomeric band at 897 cm−1 more intense than in pure

C due to the contribution from -linked xyloglucan (Fig. 1). In the CP spectrum after subtraction of C, bands at 1738, 1243, 1146, 1097, 1019, 960, and 833 cm−1 were seen, consistent with pectin21 andi nt he CXG spectrum after subtraction of C, shoulders at 1130, 1079, 1045, and 944 cm−1 corresponding to xyloglucan bands22 (Table 2) were observed.

Linear stretching and dichroic spectra. Wet samples.—Linear stretching was first performed at the highest measurable water content (about 97% RH, cellulose/xyloglucan (CXGw) samples (where subscript w denotes wet samples). The unstretched samples showed very small or zero A across the entire spec- trum. Dichroism of the cellulosic as(C-1 O C-4) glycosidic link mode at 1162 cm−1 was used as a measure of orientation.6,23 Increased dichroism in the parallel direction was observed with increasing strain (Fig. 2(a)), indicating that the polymer aligned along the direction

H2 O 6-H cellulose groups and pyranosyl-ring vibrations,13 increased in intensity without significant frequency shift. In CPw (Fig. 2(b)), the intensity at 1162 cm−1 was much higher relative to the two-ring vibra- tions, which broadened, reflecting a wider distribution of the intra- and intermolecular hydrogen bond-ener- gies compared to Cw.I n CXGw, the glycosidic 1062 cm−1 band was the only one to show a significantly positive A, but relatively weak xyloglucan bands at 1075 and 1045 cm−1 (Table 2) and also, frequency shifts (Fig. 2(c)) in the 1070–1020 cm−1 region were observed attributable to molecular deformation.16 and CXGm were measured at 84% RH, where the subscript m denotes moist samples which contained only about 12–2% water (Table 1). In the m samples, the spectral features were similar to those for the w-series, but failure typically occurred at lower strain (18– 21%). Therefore, water plays an important role in the deformation of the sample, consistent with the comment that the mechanical properties of these systems depend on hydration level.24

Pre-stretched sample.—An oriented (30% strain) Cw sample was linearly stretched into a second dimension

(Cws), perpendicular to the initial sample orientation. At the starting point of stretching the molecular orien- tation ( A) was observed to be perpendicular (Fig. 2(d)), but with increasing strain the negative bands at 1064, 1057, and 1037 cm−1 tended to zero showing reorientation into the new stretching direction. This experiment demonstrated that the reorientation into the stretching direction occurs regardless of whether the initial orientation is random or partially orientated. The orientation is not perfect but it was certainly preferred.

Orientation function.—The orientation function (F) of the cellulose glycosidic bond, a measure of distribution of cellulose chains segments, is shown as a function of strain in Fig. 3. Like A, F increased17,25 in all the samples but was generally larger for w-samples than for m-samples (Fig. 3). The enhanced ability of cellulose to orient in a wetter environment can be interpreted in

Fig. 1. FT-IR spectra of dry cellulose composites C (upmost), CP (centre), CXG.

Fig. 2. Dichroic difference spectra ( A=A −A ) taken from transmission FT-IR spectra, of wet, Cw (a), CPw (b), CXGw (c)

Fig. 3. Orientation function F of wet, w ( ) and moist, m ( ) composites: Cw, Cm (a), CPw, CPm (b), and CXGw, CXGm (c), as a function of strain (%). Duplicates are shown.

terms of water lubricating the molecular motion within the fibrillar layers. In both the w- and m-samples, F for CXG was lower than for C and CP. In the original studies of these systems, the individual macromolecular networks may orient independently as a result of fewer fibril contacts4 in CP, whereas in CXG the greater extensibility is related to alignment of cellulose in crosslinked cellulose–xyloglucan domains.3 The orientation function (Fig. 3) showed that this network provides somewhat lower reorientation ability compared to C or CP. Whitney et al.3 suggested that C contained a greater number of mechanically relevant interactions per unit volume, which are strained or broken upon elongation of the bulk material compared to CXG.

(Parte 1 de 2)