1

Crystallinities of Nanocrystalline and Nanofibrillated Celluloses by FT- Raman Spectroscopy

Umesh P. Agarwala, Richard S. Reinera, Ilari Filpponenb, Akira Isogaic, Dimitris S. Argyropoulosd aUSDA Forest Service, Forest Products Laboratory, Madison, WI E-mail: uagarwal@fs.fed.us bTKK, Espoo, Finland cGraduate School of Agricultural & Life Sciences The University of Tokyo Tokyo,Japan dDepartment of Forest Biomaterials North Carolina State University Raleigh, NC ABSTRACT Recently, a new method based on FT-Raman spectroscopy was proposed to determine cellulose I crystallinity. It was reported that in the Raman spectrum of cellulosic materials the peak intensity ratio of 380 and 1,096 cm-1 bands can be used to determine cellulose crystallinity. In that study, Raman crystallinities of the calibration set Whatman CC31 samples were found to be greatly correlated with the WAXS data (Segal-WAXS-21°; coefficient of determination R2 = 0.98). Average standard error calculated from replicate Raman acquisitions indicated that the Raman crystallinity model was highly reliable. In the present investigation, the Raman and WAXS methods were applied to nanocrystalline and nanofibrillated celluloses. Calculated crystallinities by the two methods were found to be significantly different. INTRODUCTION

Cellulose nanocrystals (CNCs) and cellulose nanofibers (CNFs) are important new materials in field of cellulose science and technology. While CNCs are elongated crystalline rod-like nanoparticles NFCs consist of alternating crystalline and amorphous domains. These materials have generated remarkable excitement in the scientific community. The materials have unique physicochemical properties and have been used as reinforcing components in nanocomposites. Additional benefits arise from the fact that they are derived from the biomass that is both renewable and sustainable. Among other things, crystallinity has an important effect on the physical, mechanical, and chemical properties of cellulose based materials. For example, with increasing crystallinity, tensile strength, dimensional stability, and density increase, while properties such as chemical reactivity and swelling decrease. Cellulose crystallinity is defined as the mass fraction of crystalline domains in cellulose materials.

2

Some of the frequently used techniques for estimating cellulose crystallinity are wide-angle X-ray scattering (WAXS) [1-3], solid state 13C cross-polarization/magic-angle spinning (CP/MAS) nuclear magnetic resonance (NMR) spectroscopy [4-6], and Fourier transform– infrared (FT-IR) spectroscopy [7, 8]. Although WAXS is the most extensively used technique, the X-ray method involves isolation of amorphous background from the diffraction pattern, which in the case of cellulose crystallites is not always easy for reasons that have to do with cellulose crystallites being small and, at lower crystallinities, the diffraction pattern being poorly defined (weak broad features). Similarly, other techniques have limitations as well and these have been discussed in the literature [e.g., 9 and references cited therein]. For the widely used simple Segal-WAXS method [1], which involves measurement of 2 heights, the crystallinity values were significantly higher than values from the other methods [10]. Raman spectroscopy has proven to be a useful technique in the cellulose and lignocellulose field with numerous applications [11, 12]. It has become an important analytical technique for nondestructive, qualitative, and quantitative analysis of cellulose-containing materials. In particular, the FT–Raman technique has an added advantage due to its ability to successfully analyze most materials that are fluorescent in conventional Raman (visible laser excitation). Moreover, in the macro mode of sampling with 1,064-nm excitation, any anisotropy associated with cellulose fiber materials is minimized because the scattering arises from the bulk of the sample and is not limited to the surface. Crystallinity measurements using Raman on semicrystalline polymers including cellulose [9, 13, 14, 15] have been performed. Of the two Raman methods proposed for cellulose crystallinity estimation [9, 15], the method of Agarwal et al. [9] is used here to calculate the crystallinities of CNCs and NFCs. The latter method produced an improved correlation between Raman and WAXS crystallinities [9, 15]. EXPERIMENTAL Materials

The nanocrystalline cellulose sample CNC-1 was prepared by acidic hydrolysis (similar to the procedure used by Araki et al. [16]). A typical procedure was as follows: 2.0 g of cellulose pulp obtained from Whatman #1 filter paper was blended by an Osterizer Blender. Resulting pulp was hydrolyzed with 100 mL of 2.5 M HBr at 100°C for 3 hours. The ultrasonication was applied during (every 60 minutes) the reaction (Omni-Ruptor 250W ultrasonic homogenizer, 50% power, 5 min). The resulting mixture was diluted with de-ionized (D.I) water followed by five cycles of centrifugation at 1500g for 10 min. (IEC Centra-CL3 Series) to remove excess acid and water soluble fragments. The fine cellulose particles became dispersed in the aqueous solution approximately at pH 4. The turbid supernatant containing the polydisperse cellulose particles was then collected for further centrifugation at 15000 g for 45 min (Automatic Servall Superspeed Centrifuge) to remove ultra-fine particles. Ultra-fine particles with small aspect ratio were removed from the upper layer, and the precipitation (after the high-speed centrifugation) was dried using a lyophilizing system (Labconco, Kansas City, MU). CNC-2 sample was produced at Forest Products Laboratory by sulfuric acid hydrolysis of wood pulp [17] using 64% sulfuric acid (acid:cellulose = 8:1). For NFCs, bleached softwood kraft pulp sample was TEMPO oxidized (1.3 mmol/g of COONa)

3

at the Department of Biomaterial Sciences (The University of Tokyo). The method used was similar to that of Saito and Isogai [18]. The TEMPO treated sample was disintegrated in water to obtain NFCs. FT–Raman and crystallinity

CNCs, NFCs, and other cellulose samples were analyzed with a Bruker RFS 100 and MultiRam spectrometers (Bruker Instruments Inc., Billerica, Massachusetts). Both the Raman systems are equipped with a 1,064-nm 1,000-mW continuous wave (CW) diode pumped Nd:YAG laser. About 20-30 mg of dry CNCs and NFCs were sampled in “Aluminum well” (a Raman sampling accessory) or a pellet was made with ~ 100 mg of the sample. For rest of the cellulose samples, approximately 250 mg of each sample was pressed into a pellet with the help of a hydraulic press. To make a pellet, 276 9 106 dyn/cm2 compressive pressure was applied. The laser power used for sample excitation was 600 mW, and 1,024 scans were accumulated. Bruker’s OPUS software program was used to find peak positions and process the spectral data. From the Raman spectra, amorphous contributions in the frequency region 250–700 cm-1 were removed by first normalizing (making band intensity equal) the spectra on 897 cm-1 band and then subtracting the corresponding spectrum of 120-min milled sample [9]. This was necessary because Raman spectroscopy is a semi-quantitative technique. 897 cm-1 band in the spectrum was chosen because its peak height was minimally impacted by change in crystallinity. For plotting purposes, data were converted to ASCII format and then exported to Excel (Microsoft Corp., Redmond, Washington). Raman crystallinity was determined using the univariate method of Agarwal et al. [9] and consisted of calculating the peak height ratio I380/I1096 from the spectra of samples from which spectrum of amorphous cellulose was subtracted. Cellulose crystallinities in the samples were estimated by using the following correlation equation

( )0065.0

0286.0096,1380Raman

−=

IIX

X-ray and crystallinity

With the exception of CNCs-1 (see Table 1 later), wide-angle X-ray diffraction profiles were recorded on a Bruker X-ray diffractometer with a Hi-Star 2-D area detector at the Materials Research Science and Engineering Center, University of Wisconsin, Madison. Diffractograms were obtained on the same sample pellets (or powder in the case of CNCs-2) that were analyzed in FT–Raman. For CNCs-1, Omni Instruments Wide Angle XRD was used and the diffractogram was obtained on the powdered sample. For calculating the crystallinity index, the ratio of the crystallinity part of the 002 peak to the total absolute peak height was used. The peak present at about 22.5° (2θ) corresponds to the [002] crystal planes. The crystalline portion of the total contribution at 22.5° was calculated by Segal method (Segal et al. 1959) and involved subtracting out the amorphous contribution at 21° (2θ). In the following, this crystallinity is designated as Segal-WAXS-21°. The latter was the peak position in the diffractrogram of 120-min milled amorphous sample. The method is same as used in reference [9]. WAXS crystallinity was also calculated in the conventional manner by subtracting out the amorphous contribution at 18° (2θ). Here, the latter is specified as Segal-WAXS-18°.

4

RESULTS AND DISCUSSION As reported previously [9] and shown in Fig.1, for the calibration set cellulose samples, the

Raman ratio plot for 380/1,096 (Fig. 1, Raman-Univariate) generated excellent regression (R2 = 0.992) and showed good sensitivity to change in crystallinity. Correlation of the calibration set crystallinities with Segal-WAXS-21° is also shown in Fig. 1 (WAXS) and evidently, compared to Raman-Univariate is somewhat inferior. The calibration set consisted of control (80.5% crystalline) and 120-min milled (0% crystalline) Whatman CC31 and six cellulose mixtures produced with crystallinities in the range 10.9–64% [9].

Fig. 1 Univariate analysis of the calibration set samples showing the correlation between calibration set crystallinity versus 380/1,096 Raman intensity ratio or WAXS crystallinity [9].

Univariate-Raman, Segal-WAXS-18°, and Segal-WAXS-21° crystallinities of CNCs, NFCs,

and selected cellulose samples [9] are reported in Table 1. Clearly, the Raman crystallinities of the CNCs-1 and CNCs-2 were identical (Table 1). However, although obtained on two different X-ray instruments, that was not the case for either sets of Segal-WAXS crystallinity data. The reasons for different Segal-WAXS-18° (or Segal-WAXS-21°) crystallinity values are not clear. In any case, Segal-WAXS-18° crystallinity for CNC-1 in Table 1 was similar to that reported in literature for other nanocelluloses [19].

5

Table 1: Crystallinities of CNCs, CNFs, and other cellulose products Sample Univariate-Raman Segal-WAXS (18°) Segal-WAXS (21°)CNCs-1a 66.0 91.0 b 80.0 b Control-1 70.7 80.0 b 75.0b CNCs-2 66.0 51.3 (67.3b) 43.7 (57.6b) Control-2 69.6 77.0b NFCs-1c 40.4 NAd TEMPO-NFCs-1 40.5 NAd NFCs-2e 33.0 NAd TEMPO-NFCs-2 54.4 NAd Starting Pulp-NFCs-2 62.3 NAd Other cellulose materials Whatman CC31 78.0f 90.3 80.5f Whatman CC41 77.9f 87.8 74.6f CF-11 72.0f 84.0 74.4f Avicel PH-101 59.5f 75.6 60.9f Whaman 541 paper 56.2f 57.1 50.0f aFrom Whatman # 1 paper bAfter a baseline is drawn under the diffractogram cFrom bleached softwood kraft pulp dNA = not yet available e From eucalyptus pulp fRef. [9]

For CNCs-1, the difference of 11% between the two Segal-WAXS crystallinities (Table 1) is understandable because it is well known that compared to other methods Segal-WAXS-18° gives higher values [10, 20]. Actually, WAXS itself gives different values of crystallinity depending upon how the crystallinity is calculated from the X-ray intensity data [10, 20]. The issue of cellulose crystallinity index vis-à-vis measurement techniques was discussed by Park et al. [20] where four different techniques incorporating X-ray diffraction and solid-state 13C nuclear magnetic resonance (NMR) were compared. It was concluded that the Segal-WAXS-18° produced significantly higher crystallinity values and did not provide an accurate measure of the crystallinity of cellulose. Additionally, earlier Park et al. [10] used three X-ray methods (peak height, peak deconvolution, and amorphous contribution subtraction) to calculate the crystallinity index of 11 cellulose samples. The three methods produced significantly different crystallinity indexes. Compared to the deconvolution approach, peak height method gave values that were significantly higher – 20 to 30 points higher. Moreover, worth noting was also that the differences depended upon the nature of the samples. For the 11 samples analyzed, the average crystallinity indexes for the peak height, peak deconvolution, and amorphous contribution subtraction methods were 86, 60, 68, respectively. This indicated that the index varied significantly and depended upon how the analysis was performed. Agarwal et al. [9] have also discussed this issue in their Raman crystallinity publication [9] and suggested that instead of

6

measuring the amorphous contribution at 18°, as is done in the traditional Segal method, the measurement be carried out at 21° (called here Segal-WAXS-21°). Indeed, these researchers have done that [9] for the cellulose materials investigated in that study and correlated thus calculated WAXS crystallinities (Segal-WAXS-21°) to the ones obtained by Raman methods (e.g., Fig. 1, univariate-Raman). For a number of cellulose products –Whatman CC31, CF-11, Avicel PH-101, Whaman 541 paper, and Whatman CC41, both the Segal-WAXS-18° and Segal-WAXS-21° crystallinities are listed in Table 1. For these samples, Segal-WAXS-21° crystallinities were significantly lower compared to the Segal-WAXS-18° data but were similar to univariate-Raman values (Table 1). Notwithstanding the discrepancy between Segal-WAXS-21° and Raman methods, comparing exclusively the Raman crystallinity data for nanocellulose samples (both CNCs and CNFs), it can be noted that not only both the CNC samples gave same crystallinity (66%) the value was significantly higher (≥26%) compared to the NFCs (Table 1). Although CNCs are supposedly 100% crystalline, it is not clear why a reduced value of 66% is generated. Perhaps, it has something to do with either the small size of the CNCs or the increased number of molecules on the crystallites’ surface. The reasons for lower CNCs crystallinity presently remain under investigation. Similarly, Raman crystallinity data on NFCs (Table 1) indicated that the crystallinity was lower, compared to CNCs, significantly (≥26) and for NFCs-1 remained same as that of the control. As of now, the reasons behind the crystallinity drop for the other NFCs sample (NFCs-2 in Table 1) remain to be investigated. CONCLUSION For the two CNC samples the Raman crystallinity of 66% was obtained. On the other hand, values of 40 and 33% were generated for the two NFCs which were more than ≥26% lower compared to the crystallinity of the CNCs. Between the NFCs, compared to control the crystallinity of NFCs-2 dropped by about 21 points. However, the NFCs-1 sample showed no such reduction. In contrast to Raman, the WAXS methods gave mixed results – higher crystallinity for CNCs-1 and similar or lower crystallinity for CNCs-2. ACKNOWLEDGEMENTS: The authors would like to thank Sally Ralph for the help received during freeze drying and Raman spectra acquisitions. REFERENCES 1. Segal L, Creely JJ, Martin AE, Conrad CM (1959) An empirical method for estimating the

degree of crystallinity of native cellulose using the X-ray diffractometer. Textile Res J 29:786–794.

2. Jayme G, Knolle H (1964) The empirical X-ray determination of the degree of crystallinity of cellulosic material. Papier 18:249–255.

3. Leppanen K, Andersson S, Torkkeli M, Knaapila M, Kotelnikova N, Serimaa R (2009) Structure of cellulose and microcrystalline cellulose from various wood species, cotton and flax studied by X-ray scattering. Cellulose 16:999–1015.

4. Horii F, Hirai A, Kitamaru R (1987) CP/MAS 13C NMR spectra of the crystalline components of native celluloses. Macromolecules 20:2117–2120.

5. Newman RH, Hemmingson JA (1990) Determination of the degree of cellulose crystallinity in wood by carbon-13 nuclear magnetic resonance spectroscopy. Holzforschung 44:351–355

7

6. Newman RH (1999) Estimation of the lateral dimensions of cellulose crystallites using 13C NMR signal strengths. Solid State NMR 15:21–29.

7. Richter U, Krause T, Schempp W (1991) Untersuchungen zur Alkalibehandlung von Cellulosefasern. Teil 1. Infrarotspektroskopische und Ro¨ntgenographische Beurteilung der Anderung des Ordnungszustandes. Angew Makromol. Chem. 185(186):155–167.

8. Hulleman SHD, Van Hazendonk JM, Van Dam JEG (1994) Determination of crystallinity in native cellulose from higher plants with diffuse reflectance Fourier transform infrared spectroscopy. Carbohydrate. Res. 261:163–172.

9. Agarwal UP, Reiner RS, Ralph SA (2010) Cellulose I crystallinity determination using FT-Raman spectroscopy: univariate and multivariate methods. Cellulose 17, 721-733.

10.Park S, Johnson DK, Ishizawa CI, Parilla PA, Davis MF (2009), Measuring the crystallinity index of cellulose by solid state, 13C nuclear magnetic resonance. Cellulose 16:641–647.

11.Atalla RH, Isogai A (1998) Recent developments in spectroscopic and chemical characterization of cellulose. In: Dumitriu S (Ed.) Polysaccharides—Structural diversity and functional versatility, 2nd edn. Marcel Dekker, New York, pp 123–157.

12.Agarwal UP (2008) Raman spectroscopic characterization of wood and pulp fibers. In: Hu TQ (Ed.) Characterization of lignocellulose materials. Blackwell, Oxford, pp 17–35.

13.Stuart BH (1995) Polymer crystallinity studied using Raman spectroscopy. Vib. Spectrosc. 10:79–87.

14.Paradkar RP, Sakhalkar SS, He X, EllisonMS (2003) Estimating crystallinity in high density polyethylene fibers using online Raman spectroscopy. J. App. Poly. Sci. 88:545–549.

15.Schenzel K, Fischer S, Brendler E (2005) New method for determining the degree of cellulose I crystallinity by means of FT Raman spectroscopy. Cellulose 12:223–231.

16.Araki J, Wada M, Kuga S, Okano T (1999) Influence of surface charge on viscosity behavior of cellulose microcrystal suspension. J. Wood Sci. 45, 258–261.

17.Dong XM, Kimura T, Rovol J-F, Gray DG (1996) Effects of ionic strength on the isotropic-chiral nematic phase transition of suspensions of cellulose crystallites. Langmuir 12, 2076-82.

18.Saito T, Isogai A (2004) TEMPO-mediated oxidation of native cellulose. The effect of oxidation conditions on chemical and crystal structures of the water-insoluble fractions. Biomacromolecules 5:1983–1989.

19.Teixeira EM, Correa AC, Manzoli A, Leite FL, Oliveira CR, Mattoso LHC (2010) Cellulose nanofibers from white and naturally colored cotton fibers. Cellulose 17:595–606.

20.Park S, Baker JO, Himmel ME, Parilla PA, Johnson DK (2010) Cellulose crystallinity index: measurement techniques and their impact on interpreting cellulase performance, Biotechnology for Biofuels 3:10.

Presented by: Umesh AgarwalInstitution: USDA FS FPL

Crystallinities of Nanocrystalline and Nanofibrillated Celluloses by FT-

Raman Spectroscopy

Outline

• Various methods for cellulose I crystallinity determination

• Crystallinity by FT-Raman• Crystallinity of CNCs and CNFs

How did we get into this?

Sour Orange: Control -1, 2, 7, 8; elevated CO2 – 3-6

Characterization of Lignocellulosic Materials – Ed.,Hu (2008), pp. 30

Park et al. Biotechnology for Biofuels, 3, 10(2010)

Various Methods

CI of Avicel-PH-101

Park et al. Biotechnology for Biofuels, 3, 10(2010)

Crystallinity from a Diffraction Pattern

Andersson et al.,Trees 18, 346(2004)

Cellulose I crystallinity determination by FT-Raman spectroscopy

Agarwal et al., Cellulose 17, 721(2010)

Calibration Set Raman spectra

Univariate-Raman and WAXS Correlations

Correlation between M-Segal-21°WAXS and Raman

( )0065.0

0286.0096,1380Raman

−=

IIX

Raman Crystallinity

Table 1: Raman vs. WAXS

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)Whatman CC31 78.0 95.3 80.5

Whatman CC41 77.9 94.7 74.6

CF-11 72.0 89.6 74.4

Avicel PH-101 59.5 82.4 60.9Whatman 541 paper 56.2 67.5 50.0

Segal et al., Textile Res. J. 29, 786(1959)

CNCs-1• Source – Whatman FP #1• Preparation – HBr• Appearance – rod like; 7-

8x100-400 nm

CNCs-2• Source – Whatman CF1• Preparation – H2SO4

• Appearance – rod like;

Comparison: CNCs-1 vs. CNCs-2

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)CNCs-1b 66.0 92.3 68.4

Control-1* 70.7 94.7 77.3

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)CNCs-2 66.0 92.6 70.8

Control-2 69.6 94.0 75.4

CNFs-1• Bleached softwood kraft• Aqueous media, pH 10• Regioselective oxidation

@ C6-OH• Disintegration – blender• Carboxylate content – 1.3

mmol/g

CNFs-2• Bleached Eucalyptus

kraft• Aqueous media, pH 10• Regioselective oxidation

@ C6-OH• Disintegration –

Centrifugal pump and microfluidizer

• Carboxylate content – 1.3 mmol/g

Comparison: CNFs-1 vs. CNFs-2

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)CNFs-1d 40.4 66.0 31.8

TEMPO-CNFs-1 58.5 91.4 69.2

Further support for CNFs-1 being more amorphous

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)CNFs-2e 34.5 73.1 39.1

TEMPO-CNFs-2 54.4 83.2 52.2Starting Pulp-CNFs-2 60.5 79.6 50.0

Table 2: CNCs and CNFs

Sample Univariate-Raman

Segal-WAXSa

(18°)M-Segal-WAXS

(21°)CNCs-1b 66.0 92.8 c 68.4

Control-1* 70.7 94.7 77.3

CNCs-2 66.0 92.6 70.8

Control-2 69.6 94.0 75.4

CNFs-1d 40.4 66.0 31.8

TEMPO-CNFs-1 58.5 91.4 69.2

CNFs-2e 34.5 73.1 39.1

TEMPO-CNFs-2 54.4 83.2 52.2Starting Pulp-CNFs-2 60.5 79.6 50.0

• For CNCs, compared to M-Segal-WAXS-21 Raman gave some what lower crystallinities (4-7% lower)

• For CNFs:(a) In both Raman and WAXS the crystallinity was lower compared to TEMPO tr. control(b) The avg. values were similar between the methods (37.5 in Raman to 35.5 in WAXS)

• Raman method can reliably estimate the crystallinity of cellulose nanomaterials

Conclusions

Collaborators

• Rick Reiner – FPL• Sally Ralph – FPL• Ilari Filpponen – TKK, Espoo• Akira Isogai – Univ. Tokyo• Dimitris Argyropoulos – NCSU