One-shot carboxylation of microcrystalline cellulose in the presence of nitroxyl radicals and sodium periodate

Abstract

Water soluble cellulose derivatives are highly required products for many practical purposes, expanding the limited applications of pure cellulose caused by the highly ordered hydrogen bond network and high crystallinity. In this connection, this paper, presents a new approach to obtain water soluble carboxyl-functionalized cellulosic materials, combining two of the most common selective oxidation protocols for cellulose, i.e. nitroxyl mediated reaction and periodate oxidation, in a one-shot reaction. It was found that, under specific reaction conditions, fully oxidized, 2,3,6-tricarboxy cellulose can be obtained in large amounts. The other valuable oxidized fractions were found to possess large amounts of carboxylic groups, as determined by potentiometric titration. ¹³C-NMR evidenced the presence of three distinctive carboxylic groups in the fully oxidized product, whereas for the partially oxidized samples, ¹³C CP-MAS solid-state NMR did not detect any carbonyl signals. The oxidized products were characterized by means of FTIR and X-ray photoelectron spectroscopy (XPS). Moreover, the changes of the degree of polymerization occurring after oxidative treatments were viscometrically determined.

---

1. Introduction

Cellulose, among the most versatile and widely prevalent biopolymers in nature, has been used for millennia for human basic needs, e.g. as building material, for clothing fabrication and as an energy source, and has found today, through its derivatives, new and exotic applications in the food industry, medicine, cosmetics, flexible display panels, electronic devices and many others.^1–3 Chemical modification of polysaccharides represent the most important route to design new materials with new structures and properties.⁴ Cellulose, having three reactive hydroxyl groups (one primary and the other two, secondary) in its repeating unit, can be easily modified, following typical alcohol group chemistry, the most important transformations being esterification,⁵ etherification,⁵ and oxidation.^6–9 In particular, the oxidation reaction has aroused the interest of many research groups around the world, due to the large variety of products which can be obtained, depending on the reactive site, and employed reagents, offering a broad spectrum of cellulose derivatives for industrial applications.^1,2,4 Many attempts have been made in order to improve the reaction selectivity of the cellulose oxidation, to find benign and cheap reagents, and even to find new paths, able to supply new and highly value-added products. There are hitherto, two main approaches for cellulose selective oxidation: (i) nitroxyl radical-mediated oxidation of the primary OH groups, and (ii) periodate oxidation of the two secondary OH groups, see Fig. 1. These two protocols are considered the most selective processes, in this kind of transformation.

Fig. 1 Possible oxidation routes for cellulose selective oxidation, in the presence of nitroxyl radicals or periodates.

The introduction of the 2,2,6,6-tetramethyl-piperidine-1-oxyl radical (TEMPO), for the selective oxidation of primary hydroxyl groups in cellulose, has been an important step forward in the field of cellulose oxidation, Fig. 1.^10,11 The stable nitroxyl radical TEMPO acts as a mediator in the presence of sodium bromide and sodium hypochlorite, to selectively oxidize the primary hydroxyl groups in cellulose. Recently, an alternative to this protocol has been reported,^7–9 which implies the use of a non-persistent nitroxyl radical, i.e. phthalimide-N-oxyl (PINO) which is generated in situ from its parent hydroxylamine, N-hydroxyphthalimide (NHPI) and an adequate cocatalyst. The use of NHPI for the selective oxidation of cellulose fibers is further evidence of the exceptional catalytic activity of this catalyst proved previously on a wide range of organic compounds,^12–18 allowing the possibility to implement NHPI homogeneous catalysis for industrial applications to be envisaged.^19,20 The NHPI–NaBr–NaClO system also oxidizes the primary hydroxyl groups in cellulose, see Fig. 1;the reported values of the carboxylic groups formed during oxidation being however smaller than in the case of using TEMPO.¹⁰ Nevertheless, the use of NHPI is preferred when products having higher degrees of polymerization are targeted.⁸

The second protocol for the cellulose selective oxidation, presented in Fig. 1, uses periodates, an oxidant able to oxidize the vicinal hydroxyl groups, resulting in two aldehyde groups with a simultaneously breaking of the C₂–C₃ linkage.^21,22 This protocol creates (masked) aldehyde groups which further might serve as anchoring points, suitable for modifying and functionalizing cellulose. Moreover, the dialdehyde cellulose is biodegradable and biocompatible. Also, the aldehyde groups can be further oxidized to carboxylic groups,²³ or to introduce imine bonds between cellulose and amine groups, through a Schiff base reaction with an amine.²⁴ However, the difference in the reactivity of hydroxyl groups in cellulose, as well as accessibility and regioselective control, could offer the circumstances to prepare cellulose products with new and specific features for various applications. Among the huge variety of the cellulosic products, water-soluble derivatives are significant for a large number of applications, such as film forming, emulsion stabilizers, lubricating, gelling agents, in fields ranging from agriculture to food, cosmetics, oil industry, textile, pharmaceutical, etc.²⁵ Most water soluble cellulose derivatives correspond to cellulose ethers, some of them (carboxymethyl, hydroxyethyl, and hydroxypropyl cellulose) being produced industrially in large quantities. The oxidation reaction of cellulose, could be used as an efficient tool to prepare water soluble cellulosic products. However, for this purpose, careful choice of reaction conditions, suitable reagents and the raw cellulose source need to be considered. 6-Carboxy cellulose, or β-1,4-linked glucuronic acid (“cellouronic acid”), has been prepared as a water-soluble fraction during TEMPO-mediated oxidation of cellulose samples pretreated with ammonia.²⁶ The yield of the soluble fraction was only 17% in the case of using Avicel, and the degree of polymerization was found to be 75. Recently two-step oxidations were applied to microcrystalline cellulose to obtain 6-carboxy cellulose.²⁷ The authors found that the water soluble product had a weight-average DP value of only 38, as against 220, the DP of the starting material. To prepare cellouronic acids by TEMPO-mediated oxidation, pretreatment of the native cellulose to convert cellulose I to III with lower crystallinity has been also proposed.²⁸

Another useful oxidizing product which could be prepared by cellulose oxidation is 2,3,6-tricarboxy cellulose (TCC), also named mesotartaric acid/monohydrated glyoxylic acid alternating co-polyacetal. TCC was initially synthesized in a three-step reaction of cellulose with N₂O₄, NaIO₄, and HClO₂.²⁹ A two-step process to prepare TCC was later reported, using N₂O₄ and NaIO₄ for the oxidation of cellulose.³⁰ Besides their multi-step character, these methods use harsh reaction conditions and harmful reagents, like N₂O₄. Very recently, a one-step preparation of TCC was published, which uses 2-azaadamantane-N-oxyl (AZADO) in combination with sodium bromide and a large excess of sodium hypochlorite to oxidize regenerated cellulose.³¹ Although this is a one-step process, the use of very expensive AZADO, hinders its potential industrial implementation. Moreover, the large excess of sodium hypochlorite used, has serious consequences on the depolymerization of the cellulose chain, which dramatically decreased from 1270 in the initial cellulose sample to only 26 in the fully oxidized product!³¹

Taking into account the serious impediments aforementioned, we envisaged that a reconsideration of the oxidizing protocol could be beneficial. Therefore, in this work, we combined the two the most selective protocols for cellulose oxidation in a one shot reaction, to achieve the oxidation of all three hydroxyl groups in the anhydroglucose unit of cellulose. In this way, we report the production of a series of partially oxidized products, and for the specific reaction conditions, even pure 2,3,6-tricarboxy cellulose has been isolated. In addition to the “single reaction” character of the proposed protocol, we avoided the use of very expensive reagents (TEMPO is 250 times cheaper than previously reported AZADO) or a large excess of NaClO as previously used.³¹ Special attention was paid to the depolymerization phenomena, which drastically decreased the molecular weight of the oxidized products. The oxidation reactions were performed at room temperature, in the presence of either TEMPO and periodate, or NHPI and periodate. Two separate control experiments were carried out, in the absence of any nitroxyl radicals. All the oxidized products were rigorously characterized by means of FTIR, NMR (for water soluble compounds) and solid state NMR for the water insoluble products. X-ray photoelectron spectroscopy (XPS) has been further used for the characterization of the oxidized products, whereas the content of the carboxylic (aldehyde) groups were determined by using titration methods. The changes which had occurred on the degree of the polymerization of the oxidized products were determined through viscometric measurements.

2. Materials and methods

2.1. Materials

Avicel® PH 101 microcrystalline cellulose purified, partially depolymerized α-cellulose, with a mean degree of polymerization (DP) of 140, was purchased from Sigma-Aldrich. TEMPO, NHPI, sodium periodate, sodium bromide, 9% (wt) sodium hypochlorite and other chemicals and solvents were of pure grade (Sigma Aldrich), and used without further purification.

2.2. Oxidation reaction protocols

TEMPO-mediated oxidation of cellulose, using TEMPO–periodate–NaBr–NaClO system. TEMPO (0.4 g, 2.5 mmol), sodium periodate (2.7 g, 12.5 mmol) and NaBr (4 g, 40 mmol) were dissolved in 600 mL distilled water under vigorous stirring. Microcrystalline cellulose (5 g) was then suspended in the reaction mixture. The reaction vessel was covered with aluminum foil to prevent the photo-induced decomposition of periodate. 9% NaClO solution (2.97 g, 40 mmol) was added to the cellulose slurry under continuous stirring and the resulting suspension was stirred for a certain time: 4 h and 24 h respectively, at room temperature. The pH of the suspension was carefully maintained at about 10.5 by adding 2 M NaOH solution. After the designated time, the oxidation reaction was stopped by adding 5 mL of ethanol and the oxidized cellulose was filtered and washed several times with deionized water and 0.5 M HCl solution. The obtained water-insoluble fraction was dried by lyophilization, followed by vacuum-drying at 40 °C for 48 h, and weighed to measure the mass recovery ratios. The water-soluble fraction was precipitated with ethanol, the formed precipitate being collected by centrifugation. After centrifugation, the solid fraction was re-dissolved in water, desalted, and oligomers were removed by diafiltration through a Millipore ultrafiltration membrane from polyethersulfone (cut-off: 10 000 g cm⁻¹) in an Amicon cell equipped with a tank filled with pure water (conductivity lower than 3 μS m⁻¹). The diafiltration was stopped when the filtrate conductivity was lower than 10 μS m⁻¹ and the oxidized cellulose was recovered by freeze-drying.

NHPI-mediated oxidation of cellulose, using NHPI–periodate–NaBr–NaClO system. The oxidation protocol is similar to the procedure presented for the TEMPO-mediated oxidation of microcrystalline cellulose using TEMPO–periodate–NaBr–NaClO, except for the disperse medium, which in this case was a mixture of water (500 mL) and acetonitrile (100 mL), to ensure a better solubility of NHPI. All of the other parameters were maintained, including the reaction times: 4 h and 24 hours respectively.

Periodate-mediated oxidation of cellulose. Microcrystalline cellulose (5 g) was immersed for 4 h and 24 h respectively, in water (600 mL) containing sodium periodate (2.7 g, 12.5 mmol). The mixtures were gently stirred at room temperature, in the dark. The pH of the suspension was maintained at about 4 by adjusting the pH with 2 M NaOH or 0.5 M HCl solutions. The oxidation was stopped by adding 5 mL glycerin and the oxidized cellulose was filtered and washed several times with deionized water and 0.5 M HCl, and then dried by lyophilization followed by vacuum-drying at 40 °C for 48 h.

2.3. Characterization methods

UV-vis measurements. The electronic absorption spectra were recorded using a SPECORD 200 Analytik Jena spectrometer.

FT-IR. Approximately 1 mg of dry cellulose sample was pressed into a pellet with 200 mg of potassium bromide and the Fourier transform infrared (FT-IR) spectrum was recorded by a Bruker Vertex 70 with accumulation of 32 scans and a resolution of 2 cm⁻¹, from 4000 to 500 cm⁻¹.

Determination of the aldehyde group content. The dialdehyde group content was determined using a titrimetric method.³² According to this method, each –CHO group reacts with hydroxylamine hydrochloride to form an oxime, while one proton is released. A certain amount of dialdehyde sample was suspended in water, and the pH was adjusted to 3.5 using HCl, then 25 mL hydroxylamine hydrochloride solution (5% w/w) was added to the suspension. The pH of the suspension was carefully maintained around 3.5 by adding 0.1 mol L⁻¹ NaOH, until no change of pH was observed. The cellulose sample was washed several times with water and collected by filtration. The weight of each sample was measured after drying, and the aldehyde content was determined by the consumption of the NaOH solution.

Determination of the carboxylic group content. To a dried sample equivalent to 0.5–1.0 g was added 100 cm³ of 0.5 M NaCl. 10 cm³ of 0.1 M HCl in 0.5 M NaCl was added to the suspension, prior to titration. The titration was carried out by adding 0.1 M NaOH in 0.5 M NaCl from a precision burette. The solution was stirred with a glass propeller and kept in an airtight titration vessel, during titration. All experiments were carried out under thermostatically controlled conditions at 25 °C. An inert atmosphere was maintained by continuous flow of argon. After each addition, the potential was recorded automatically with a Mettler Toledo 70 titrator. It usually took 2.5 h until a stable potential was attained. The stability criterion was a drift of less than 0.5 mV min⁻¹. All presented values are the mean values of three parallel measurements, the standard deviation of measurements being within 4%.

Degree of polymerization (DPv). The degree of polymerization (DP) was determined viscometrically after dissolving the cellulose samples in cuoxam according to the method described by Klemm.³³ The viscosity measurements were performed in a modified Ubbelohde viscometer (capillary length 78 mm, capillary bore width 0.75 mm, volume of the bulb between the marks 7 cm³). The cellulose solution in the viscometer was employed for one measurement only to avoid possible degradation of the sample. The viscosity (η) was calculated from the efflux time of the cellulose solution (t), the blank cuoxam solution (t₀), and from the concentration of cellulose in solution (c). The DP was calculated from the specific viscosity:according to:where: c = cellulose sample concentration in g L⁻¹.

¹³C NMR analyses of water-soluble fractions. ¹H-NMR and ¹³C-NMR spectra were obtained by a Bruker-Avance DRX 400 MHz Spectrometer, equipped with a 5 mm QNP direct detection probe and z-gradients, from solutions in D₂O, using tetramethylsilane TMS (δ = 0.0 ppm) as the internal standard.

¹³C CP-MAS solid-state NMR. ¹³C CP-MAS NMR spectra were recorded with a Bruker-Avance III AQS 400 SB instrument, operating at 9.4 T, fitted with a double air-bearing two-channel probe head. Samples were packed uniformly in a 4 mm zirconium oxide rotor. All measurements were performed at 296 (±1) K. The MAS rate was 10 kHz. Acquisition was performed with a CP pulse sequence using a 2.95 μs proton 90 degree pulse, an 800 μs ramped (100–50%) falling contact pulse and a 2.5 s delay between repetitions. A SPINAL64 pulse sequence was used for ¹H decoupling. The Hartmann–Hahn matching procedure was performed on glycine and the chemical shift scale was calibrated to TMS ((CH₃)₄Si) by assigning the data point of maximum intensity for the alpha-glycine carbonyl signal a chemical shift of 176.03 ppm.

X-ray photoelectron spectroscopy (XPS). The compositional analysis of the studied samples was carried out by X-ray photoelectron spectroscopy (XPS) using a PHI-5000 VersaProbe photoelectron spectrometer (Φ ULVAC-PHI, INC.) with a hemispherical energy analyzer (0.85 eV binding energy resolution for organic materials). The shape of the samples was a “tablet” of dried fibers. A monochromatic Al Kα X-ray radiation (hν = 1486.7 eV) was used as the excitation source. The standard take-off angle used for analysis was 45°, producing a maximum analysis depth in the range of 3–5 nm. Spectra were recorded from at least three different locations on each sample, with a 1 mm × 1 mm area of analysis. Low-resolution survey spectra were recorded in 0.5 eV steps with 117.4 eV analyzer pass energy. In addition, high-resolution carbon (1s) spectra were recorded in 0.1 eV steps with 58.7 eV analyzer pass energy. The XPS data were acquired using the PHI SUMMIT XPS for VersaProbe software.

3. Results and discussion

Nitroxyl radicals, such as stable TEMPO, or non-persistent PINO, are known to be efficient mediators for the C₆ oxidation of cellulose, the other two hydroxyl groups being unaffected. On the other hand, sodium periodate is often used to oxidize C₂ and C₃ atoms in the anhydroglucose unit of cellulose, to form dialdehydes, whereas the C₆ atom remains unoxidized. Based on these findings, we can allege that these two processes are very selective routes to perform cellulose oxidation. However, no information on how the cellulose oxidation will be influenced by the simultaneous presence of the two reagents (nitroxyl radicals and sodium periodate) are reported. The question is whether the selectivity of each reagent will remain, or other some synergistic effect will be prevalent? To answer this question, in our present work, the oxidation of cellulose was accomplished by using simultaneously both oxidizing agents: nitroxyl radicals (either TEMPO, or PINO) and sodium periodate in a one-shot reaction, in the presence of sodium hypochlorite and sodium bromide. Following this procedure, the total oxidation of the three available hydroxyl groups in cellulose becomes feasible. In one parallel run the oxidation was performed in the absence of nitroxyl radicals, only periodate being employed, a classic periodate oxidation of cellulose. Two series of experiments were carried out, varying the reaction time from 4 h to 24 h respectively.

Until a few years ago, TEMPO and its derivatives, were the only nitroxyl radicals able to perform the selective oxidation of C₆ in cellulose. Recently, NHPI was implemented as an efficient mediator for the selective conversion of the primary OH groups in cellulose into carboxylic moieties. However, in this case, the presence of another compound is mandatory to achieve the activation of NHPI by its conversion into the PINO radical. Several compounds are used as co-catalysts for the homolytic bond cleavage between the O and H atoms in NHPI.^12–16 This requirement seems to be a drawback, since compounds such as lead tetraacetate, or cerium ammonium nitrate, co-catalysts for the PINO in situ generation, themselves interfere with the cellulosic substrate, causing side reactions, especially by degrading the macromolecular backbone. In order to avoid this, we tested the ability of sodium periodate, as a possible agent for the in situ formation of the PINO radical. Fortunately, periodate could act as a very efficient co-catalyst for the PINO radical formation, as the UV-vis spectra highlights, Fig. 2.

Fig. 2 UV-vis spectra of a mixture of NHPI (1 mM) and sodium periodate (1 mM) in acetonitrile–water (1 : 1 volume%); inset: UV-vis spectrum of 1 mM NHPI solution in acetonitrile.

The maximum UV-vis adsorption of neat NHPI in acetonitrile, at λ = 294 nm, shifts instantaneously to λ = 422 nm when an equimolar amount of periodate aqueous solution is added. This is a clear evidence of the PINO radical formation, which, further can act as a mediator for the cellulose oxidation. The simultaneous presence of TEMPO (or NHPI) and periodate will fulfil the conditions for the oxidation of all three hydroxyl groups in the anhydroglucose unit, as is suggested in Fig. 3.

Fig. 3 Illustration scheme of the full oxidation of cellulose in the presence of both nitroxyl radical (TEMPO or PINO) and periodate.

In the first stage of the reaction, the three hydroxyl groups, are converted to aldehydes, as follows: the accessible C₆–OH groups on the crystalline surface are converted due to the presence of the nitroxyl radicals (TEMPO or NHPI) in the presence of sodium hypochlorite, whereas the two secondary OH groups, at C₂ and C₃, are being converted into 2,3-dialdehyde cellulose, in a classic periodate oxidation mechanism, concomitantly with the cleavage of the cellulose’s glucopyranose rings between the C₂–C₃ bond. In the next step, due to the presence of nitroxyl radicals and sodium hypochlorite, the aldehyde groups are further oxidized, to form the final oxidation product: the carboxylic groups.

Table 1 presents the carboxylic (aldehyde) group content, yield and degree of polymerization of the oxidized samples. The samples are labeled by letters O, N, T, and P. Sample O is the untreated (original) microcrystalline cellulose, whereas N denotes samples simultaneously oxidized by NHPI and periodate, T those simultaneously oxidized by TEMPO and periodate, and P means the samples oxidized only by periodate. Each label also contains a number (4 or 24) which describes the reaction time (in hours) used to prepare the respective sample. The water soluble samples are identified by a lowercase s, added after the corresponding number.

Table 1 The content of carboxylic (aldehyde) groups, and degree of polymerization of cellulose samples, oxidized at room temperature, in the presence of 2.5 mM periodate per g cellulose

Sample	Reaction time (h)	NHPI (mM per g cellulose)	TEMPO (mM per g cellulose)	Content of carboxylic groups (mM kg^−1)	Content of aldehyde groups (mM kg^−1)	Yield (%)	Degree of polymerization (DP)
O	—	—	—	51	—		140
N4	4	0.5	—	456	—	82	86
N4s	4	0.5	—	—	—	<1	n.d
N24	24	0.5	—	553	—	79	65
N24s	24	0.5	—	—	—	<3	n.d
T4	4	—	0.5	1760	—	52	69
T4s	4	—	0.5	3110	—	31	35
T24	24	—	0.5	1790	—	48	55
T24s	24	—	0.5	3128	—	34	28
P4	4	—	—	—	315	95	107
P24	24	—	—	—	1325	88	75

---

Once the conversion of the hydroxyl groups in the anhydroglucoside unit, to carboxylic groups, reached a certain level, the fully oxidized product tends to become water soluble: see samples N4s, N24s, T4s, and T24s, in Table 1. The amount of carboxylic groups formed upon oxidation for 4 h, is almost four times higher in the case of using TEMPO–periodate, than in the case of using NHPI–periodate (samples T4 and N4, Table 1). When solely periodate was used for the oxidation, no evidence of carboxylic group formation was found, the only hydroxyl groups converted being those on the positions 2 and 3 on anhydroglucose unit, to form dialdehydes. On increasing the oxidation time from 4 h to 24 h, no further improvement of the amount of carboxylic group formation was noticed, samples N24, and T24. However, the prolongation of the reaction, caused supplemental degradation of the cellulose chain, from a degree of polymerization of 86 to 65 for the samples oxidized with NHPI–periodate, and from a degree of polymerization of 69 to 55, for the samples oxidized with TEMPO–periodate. Periodate oxidation of cellulose for a longer time (24 h) instead, leads to an impressive increase on the amount of 2,3-dialdehyde group formation as compared as with the 4 h time reaction, see samples P4, and P24. Moreover, the degradation processes in the case of periodate oxidate are less pronounced as compared with the case of using nitroxyl radical-mediated oxidations. Sample T24s has the highest amount of carboxylic groups, more than 3000 mM per kg cellulose. This remarkable amount induced the solubility of this sample in water, as well as a highly negative value of the ξ-potential: −58 ± 3.85 mV. It has to be pointed out that the water soluble fractions are exclusively obtained when employing the TEMPO–periodate protocol, in the case of the NHPI–periodate system only small amounts of the water-soluble product could be obtained.

3.1. FTIR

The FTIR technique can be used as a straightforward method to evaluate the structural changes which have occurred in the cellulose after oxidation. However, for the periodate oxidized samples, the aldehyde group estimation appears to be difficult, since these groups exist partially or even totally hydrated, and the resulting hemiacetal or hemialdol structures do not exhibit the classical adsorption in FTIR of the carbonyl group. In these conditions, the aldehyde group originating from periodate oxidation appears hardly to be detected and not quantitatively by FTIR.³⁴ Moreover, all the possible oxidized functionalities appearing after the oxidation reaction (aldehyde, keto and carbonyl) absorb in a very narrow region of the spectrum, between 1700 and 1750 cm⁻¹. Fig. 4 shows the FTIR spectra of original and 24 h oxidized samples.

Fig. 4 FTIR spectra of untreated and 24 h oxidized cellulose samples.

The characteristic cellulose peaks for hydrogen bonded O–H stretching, centered at 3420 cm⁻¹ and for sp³ hybridized C–H stretching at 2900 cm⁻¹ are present in all the samples. However, the latter peak, drastically decreased after oxidation (see samples T24 and T24s), due to oxidation at the secondary hydroxyl groups (OH-2 or OH-3 positions). After oxidation, the presence of another peak can be denoted, thus for the samples N24 and T24 a sharp absorption at 1738 cm⁻¹ is attributed to carbonyl groups in the free COOH group, whereas the large and sharp absorption at 1615 cm⁻¹ in sample T24s is due to carboxylate groups in their sodium salt form. The absence of these absorption bands in the P24 sample shows that no carboxylic groups are formed during periodate oxidation (in the absence of any nitroxyl radicals), the absorption peak being localized around 1717 cm⁻¹, which is assigned to C=O stretching vibration of aldehyde groups. The other characteristic peaks of unoxidized cellulose, such as –OH in-plane bending, at 1201 cm⁻¹, C–H deformation stretching vibration, at 1112 cm⁻¹ and asymmetry stretch vibration, around 1165 cm⁻¹, C–O–C stretching vibration of the pyranose ring skeleton at 1059 cm⁻¹, tend to become weaker after oxidation, in all samples, indicating possible decomposition processes during oxidation.^35,36

3.2. XPS

The X-ray photoelectron spectroscopy (XPS) technique has been intensively used in the last decade to study structural changes which occur in polysaccharide backbones during various physical or chemical treatments.^37–39 A low resolution scan of the untreated and 24 h cellulose oxidized samples was performed. These scans were used to determine the O/C ratio, Table 2 and Fig. 5. As expected, the only elements found, were carbon and oxygen, except for the T24s sample, when Na was also detected. This is due to the fact that carboxylic groups in this sample, are in their sodium carboxylate form, see also the FTIR spectrum.

Table 2 The elemental surface composition from XPS survey spectra for original and oxidized cellulose samples

Sample	C1s (%)	O1s (%)	O/C
Avicel	61.30	38.70	0.63
N24	59.46	40.56	0.68
T24	58.74	41.26	0.70
T24s	57.02	42.98	0.75
P24	60.09	39.91	0.66

---

Fig. 5 XPS survey spectra of original and oxidized cellulose samples.

As Table 2 highlights, the atomic ratio O/C gradually increases from 0.63 in the unoxidized sample to 0.66 in the P24 sample, and reached the highest value of 0.75 in the T24s sample, which denote the highest percent of introduced oxygen in the cellulose units.

To examine the types and relative amounts of the different C–O bonds present on the surface of cellulose samples, the high resolution spectra of the C1s region was obtained. The high resolution C1s XPS spectra reveal chemical shifts that could be classified as follows:⁴⁰ unoxidized carbon (C–C, binding energy: 285 eV), carbon with one oxygen bond (C–O– or C–OH, binding energy: 286.5 eV), carbon with two oxygen bonds (O–C–O or C=O, binding energy: 288 eV), and carbon with three oxygen bonds (O=C–O, binding energy: 289.4 eV), Table 3.

Table 3 Relative amounts of differently bound carbons as determined from high resolution carbon C1s determined by XPS

Sample	Relative concentration (%) and binding energy (eV)
	C–C/C–H, C1, 285	C–O–/C–OH, C2, 286.5	O–C–O/C=O, C3, 288	O–C=O, C4, 289.4	C4/C2
O	16.96	64.15	17.39	1.50	0.023
N24	17.77	56.53	21.50	4.20	0.074
T24	27.37	47.08	19.93	5.62	0.120
T24s	23.32	43.51	25.59	7.58	0.170
P14	15.73	59.61	21.24	3.42	0.057

---

From Table 3 one can conclude that after oxidation, the C2 peak area decreases substantially due to the conversion of OH groups into CHO or COOH groups. Another reason for the decrease could be related to the degradation of the D-glucose ring during oxidation. Conversely, the C3 and C4 peaks, consisting of O–C–O/C=O and O=C–O linkage of carbon respectively, significantly increased after cellulose oxidation, which denotes a large amount of carbonylic/carboxylic groups being introduced into the anhydroglucose unit. As a measure of these changes, we used the ratio of C4/C2 area peaks, to evaluate the unoxidized and oxidized cellulose samples. This ratio increased from 0.023 in pure cellulose to 0.074 in the N24 sample and reached the highest value of 0.170 in T24s sample. The ratio C4/C2 differs due to the changes occurring on both carbon type contributions. On the one hand, the C2 peak area became smaller after oxidation (due to the OH groups disappearance), but on the other hand, the C4 peak area largely increased after cellulose oxidation, due to the introduction of a large amount of carboxylic groups.

3.3. ¹³C-NMR of water soluble fractions

Due to the high content of carboxylic groups incorporated, T4s and T24s samples become highly water soluble, allowing therefore the acquisition of the ¹³C NMR spectra in deuterated water. Fig. 6 shows a typical spectrum of such compounds. Since for the original (unoxidized) cellulose sample it is not possible to record the ¹³C NMR in deuterated water its spectrum can be seen later as a solid state NMR spectrum. The T4s sample exhibits in its ¹³C NMR spectrum quite intense peak signals, between 176.87 and 178.35 ppm, which are characteristic for carbons originating from carboxylic groups. The presence of three peaks (see the inset of Fig. 6) proves that there are three distinctive COOH groups inside the anhydroglucose unit.

Fig. 6 ¹³C-NMR spectrum in D₂O of T4s sample. Inset: magnified peaks of “carboxylic” signals.

In the case of the preparation of another water soluble cellulose derivative, i.e. 6-carboxylcellulose, in the ¹³C NMR spectrum there is only one carbon peak around 175 ppm, (corresponding only to one carboxylic group type formed) in accordance with C₆ oxidation performed with nitroxyl radical and sodium hypochlorite (no periodate was present).⁴¹

3.4. ¹³C CP-MAS solid-state NMR

Partially oxidized cellulose products, those which were not water soluble, were analyzed by means of ¹³C CP-MAS solid-state NMR. The spectrum of the unoxidized cellulose, Fig. 7, displays typical cellulose signals in the range of 110–60 ppm. A multiplet around 105 ppm is assigned to the C₁ carbon of cellulose, two signals at 90 and 84 ppm are assigned to the C4 carbons in crystalline and non-crystalline regions, whereas crystalline and non-crystalline C₆ signals are located at 64.4 and 62.5 ppm respectively. Between 76.6 and 70.7 ppm there is a region of overlapping signals, originating from the C₂, C₃, and C₅ carbons. After oxidation some spectral changes occurred. The most noticeable is a new signal peak at 170.5 ppm in the N24 sample, which increased in intensity in the T24 sample. This new signal confirms the introduction of the COOH groups in these samples, as previously evidenced by FTIR. Also, there are changes of the C6 non-crystalline signals, which decrease in intensity from N24 to T24 samples, correlating with the expected conversion of CH₂–OH groups to COOH groups. Chemical modification of a fibril surface can indirectly alter signal intensities at neighboring chemically unmodified surfaces. Also, significant changes occur in the C4 non-crystalline region as the result of the chemical modifications. Further, loss of signal intensity at a particular signal position in CP/MAS ¹³C-NMR spectra as the result of chemical modifications can be the result of shifting signal intensity to another position or loss of signal intensity (discrimination) due to changes in polymer mobility. For all oxidized samples, the signals originating from the C_2,3,5 atoms are lower in intensity than in the original, unoxidized samples, correlating with the conversion of the OH groups linked to C₂ and C₃ to carboxyl (carbonyl) groups. An interesting feature of the sample P24 in ¹³C NMR, is the lack of any resonance in the carbonyl region (210–180 ppm) expected due to the presence of the two carbonyl moieties formed during periodate oxidation. This absence is due to the fast recombination reactions of carbonyl groups to form hemiacetal structures with the remaining hydroxyl groups.^32,42

Fig. 7 ¹³C CP-MAS solid-state NMR spectra of unoxidized and oxidized cellulose samples.

4. Conclusions

The highly water-soluble 2,3,6-tricarboxy cellulose has been prepared by combining the synergic actions of the two cellulose selective oxidants: nitroxyl radicals (TEMPO) and sodium periodate, in a one-shot reaction, at room temperature and pH 10.5, within 4 h. The carboxylic group content as found by potentiometric titration was as high as 3110 mM kg⁻¹. No further significant increase of this amount by prolonging the reaction time to 24 h was observed. In the case of combining another nitroxyl radical (PINO, obtained in situ from its parent hydroxyl amide) with periodate, rather modest amounts of water soluble derivative were formed, and in addition the overall content of carboxylic groups formed was scarce. Notably, the polymerization degree of the synthesized 2,3,6-tricarboxy cellulose was found to be only 4 times lower than that of the starting material, indicating a better preservation of the macromolecular chain than in the case of other reported oxidizing protocols, when the polymerization degree was, impressively 500 times lower as compared with the original cellulose sample.³¹ When periodate alone was employed no carboxylic group formation was detected, 2,3-dialdehyde cellulose products were revealed instead.

References

1. D. Klemm, F. Kramer, S. Moritz, T. Lindstrom, M. Ankerfors, D. Gray and A. Dorris, Angew. Chem., Int. Ed., 2011, 50, 5438–5466.
2. Y. Habibi, L. Lucia and O. J. Rojas, Chem. Rev., 2010, 110, 3479–3500.
3. R. J. Moon, A. Martini, J. Nairn, J. Simonsen and J. Youngblood, Chem. Soc. Rev., 2011, 40, 3941–3994.
4. T. Heinze and T. Liebert, Prog. Polym. Sci., 2001, 26, 1689–1762.
5. S. C. Fox, B. Li, D. Xu and K. J. Edgar, Biomacromolecules, 2011, 12, 1956–1972.
6. S. Coseri, G. Biliuta, B. C. Simionescu, K. Stana-Kleinschek, V. Ribitsch and V. Harabagiu, Carbohydr. Polym., 2013, 93, 207–215.
7. S. Coseri, G. Nistor, L. Fras, S. Strnad, V. Harabagiu and B. C. Simionescu, Biomacromolecules, 2009, 10, 2294–2299.
8. G. Biliuta, L. Fras, S. Strnad, V. Harabagiu and S. Coseri, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 4790–4799.
9. G. Biliuta, L. Fras, V. Harabagiu and S. Coseri, Dig. J. Nanomater. Bios., 2011, 6, 293–299.
10. M. Hirota, K. Furihata, T. Saito, T. Kawada and A. Isogai, Angew. Chem., Int. Ed., 2010, 49, 7670–7672.
11. Y. Okita, T. Saito and A. Isogai, Biomacromolecules, 2010, 11, 1696–1700.
12. S. Coseri, Catal. Rev., 2009, 51, 218–292.
13. S. Coseri, J. Phys. Org. Chem., 2009, 22, 397–402.
14. S. Coseri, G. D. Mendenhall and K. U. Ingold, J. Org. Chem., 2005, 70, 4629–4636.
15. S. Coseri, Eur. J. Org. Chem., 2007, 11, 1725–1729.
16. F. Recupero and C. Punta, Chem. Rev., 2007, 107, 3800–3842.
17. B. Orlinska and J. Zawadiak, React. Kinet., Mech. Catal., 2013, 110, 15–30.
18. L. Melone and C. Punta, J. Org. Chem., 2013, 9, 1296–1310.
19. L. Melone, S. Prosperini, G. Ercole, N. Pastori and C. Punta, J. Chem. Technol. Biotechnol., 2014, 89, 1370–1378.
20. M. Petroselli, P. Franchi, M. Lucarini, C. Punta and L. Melone, ChemSusChem, 2014, 7, 2695–2703.
21. K. A. Kristiansen, A. Potthast and B. E. Christensen, Carbohydr. Res., 2010, 345, 1264–1271.
22. W. Kasai, T. Morooka and M. Ek, Cellulose, 2014, 21, 769–776.
23. U. J. Kim and S. Kuga, J. Chromatogr. A, 2001, 919, 29–37.
24. M. Wu and S. Kuga, J. Appl. Polym. Sci., 2006, 100, 1668–1672.
25. S. Gomez-Bujedo, E. Fleury and M. R. Vignon, Biomacromolecules, 2004, 5, 565–571.
26. D. da Silva Perez, S. Montanari and M. R. Vignon, Biomacromolecules, 2003, 4, 1417–1425.
27. L. Li, S. Zhao, J. Zhang, Z. X. Zhang, H. Hu, Z. Xin and J. K. Kim, Fibers Polym., 2013, 14, 352–357.
28. L. Dantas, A. Heyraud, J. Courtois and M. Milas, Carbohydr. Polym., 1994, 24, 185–191.
29. F. S. H. Head, J. Chem. Soc., 1948, 1135–1137.
30. W. M. Hearon, F. L. Cheng and F. W. John, Appl. Polym. Symp., 1975, 28, 77–84.
31. S. Takaichi, R. Hiraoki, T. Inamochi and A. Isogai, Carbohydr. Polym., 2014, 110, 499–504.
32. U. J. Kim, S. Kuga, M. Wada, T. Okano and T. Kondo, Biomacromolecules, 2000, 1, 488–492.
33. D. Klemm, B. Philipp, T. Heinze, U. Heinze and W. Wagenknecht, Comprehensive Cellulose Chemistry: Fundamentals and Analytical Methods, Wiley-VCH Verlag GmbH, Weinheim, Germany, 1998, vol. 1; pp. 9–21.
34. P. Calvini, G. Conio, M. Lorenzoni and E. Pedemonte, Cellulose, 2004, 11, 99–107.
35. J. Y. Kim and H. M. Choi, Cellul. Chem. Technol., 2014, 48, 25–32.
36. T. Nikolic, M. Kostic, J. Praskalo, B. Pejic, Z. Petronijevic and P. Skundric, Carbohydr. Polym., 2010, 82, 976–981.
37. J. Li, Y. Wan, L. Li, H. Liang and J. Wang, Mater. Sci. Eng., C, 2009, 29, 1635–1642.
38. R. A. N. Pertile, F. K. Andrade, C. Alves and M. Gama, Carbohydr. Polym., 2010, 82, 692–698.
39. T. Topalovic, V. A. Nierstrasz, L. Bautista, D. Jocic, A. Navarro and M. M. C. G. Warmoeskerken, Colloids Surf., A, 2007, 296, 76–85.
40. S. Sun, J. Suna, L. Yao and Y. Qui, Appl. Surf. Sci., 2011, 257, 2377–2382.
41. S. Coseri, A. Doliska and K. Stana-Kleinschek, Ind. Eng. Chem. Res., 2013, 52, 7439–7444.
42. N. Guigo, K. Mazeau, J.-L. Putaux and L. Heux, Cellulose, 2014, 21, 4119–4133.

---