KINETICS AND MECHANISM OF CHEMICAL REACTIONS. …...reaction. The nitronium ion content in the reaction acidic mixture (RAM) is a few tens percent. Thus, in this study, the used RAM

245

ISSN 1990-7931, Russian Journal of Physical Chemistry B, 2016, Vol. 10, No. 2, pp. 245–259. © Pleiades Publishing, Ltd., 2016.Original Russian Text © S.V. Stovbun, S.N. Nikol’skii, V.P. Mel’nikov, M.G. Mikhaleva, Ya.A. Litvin, A.N. Shchegolikhin, D.V. Zlenko, V.A. Tverdislov, D.S. Gerasimov,A.D. Rogozin, 2016, published in Khimicheskaya Fizika, 2016, Vol. 35, No. 4, pp. 20–35.

Chemical Physics of Cellulose NitrationS. V. Stovbuna, *, S. N. Nikol’skiia, V. P. Mel’nikova, M. G. Mikhalevaa, Ya. A. Litvina,

A. N. Shchegolikhina, D. V. Zlenkoa, V. A. Tverdislova, D. S. Gerasimovb, and A. D. Rogozinb

aSemenov Institute of Chemical Physics, Russian Academy of Sciences, Moscow, RussiabAleksin Chemical Plant, Aleksin, Tula oblast, Russia

*e-mail: [email protected] August 10, 2015

Abstract—The physical mechanisms responsible for the kinetics of nitration of cellulose raw materials of dif-ferent origin have been studied. It has been shown that the main nitration rate-limiting factor is the speed ofuntwisting of supercoiled cellulose fibers. This process limits the penetration of nitrating agents into micro-crystalline regions and, thus, the total reaction rate. The constructed physical model provides an adequateexplanation of all the experimentally observed features of the cellulose nitration process, particularly as afunction of cellulose origin (cotton, f lax, wood) and preparation/treatment methods (sulfite, sulfate,bleached, refined with sulfurous or boric acids or acetone). The theoretical results have been tested in prac-tice.

Keywords: cellulose, nitration, physical model, structure formation, hierarchical systemsDOI: 10.1134/S199079311602024X

INTRODUCTIONCellulose, whose annual world production is on

the order of 200 million tons, is the most importantand popular natural biopolymer. The major users ofcellulose are the paper-making and chemical indus-tries, which convert the feedstock mostly to variouscellulose esters, such as nitrates or acetates [1]. Themost technologically convenient and most commonlyused feedstock for chemical production is cotton cel-lulose.

In the Russian Federation, cotton plant is notgrown at all, while commercial cellulose of wood ori-gin is produced from domestic raw materials, particu-larly for export. At the same time, commercial woodcellulose does not meet the requirements of the GOST(State Standard) for esterification; the known esterifi-cation processes do not occur efficiently. Thus, com-mercial softwood cellulose manufactured at theArkhangelsk and Syassky Pulp and Paper Mills (PPMs)is not nitrated to the level of nitration of cotton cellu-lose under comparable conditions; this fact has not yetbeen clearly explained in terms of conventional scien-tific approaches [1].

Typically, the chemistry of cellulose nitration isstudied without taking into account the structural andkinetic characteristics of the reaction. In concentratedacids, the nitration reaction consists in an electro-philic attack of the OH groups of cellulose by thenitronium cation to abstract the proton and form arespective nitrate, or the SE1 electrophilic substitution

reaction. The nitronium ion content in the reactionacidic mixture (RAM) is a few tens percent. Thus, inthis study, the used RAM had the following composi-tion: H2SO4, 67.4%; HNO3, 25.4%; and H2O, 7.2%,which corresponds to a 62.5% conversion of nitric acidto nitronium ions [2]. It is assumed that the nitroniumcation rapidly undergoes diffusion in both the amor-phous and crystalline regions of the cellulose fiber;therefore, the reactions are formally treated as liquid-phase reactions [1]. At the same time, it is clear thatcellulose (both cotton and wood) is a disordered het-erogeneous system characterized by a chaos of aniso-metric structural elements (Fig. 1); the equivalence ofit to a homogeneous reaction system is not obvious.

A fairly large number of studies have addressed thekinetics of cellulose nitration. The kinetics of nitrationof wood cellulose in a mixture of nitric acid anddichloroethane has a pronounced two-stage pattern[3]. The first reaction stage is more rapid; it occursabout 100 times faster and takes only a few minutes; ityields a nitrocellulose (NC) with a degree of substitu-tion of about 2.0–2.2. In the slower stage, which takesa few tens of minutes, the degree of substitutionachieves a value of ~2.5 [3]. Under quasi-homoge-neous conditions, nitration of cotton cellulose withaqueous solutions of nitric acid occurs much morerapidly; it takes about a minute or less and yields anearly pure cellulose trinitrate [4]. The observed dif-ferences are evident from the dependence of thedegree of heterogeneity of the final NC on the realconcentration of the nitrating agent (nitronium ion) in

KINETICS AND MECHANISMOF CHEMICAL REACTIONS. CATALYSIS

246

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY B Vol. 10 No. 2 2016

STOVBUN et al.

the reaction mixture [5, 6]. The higher the concentra-tion, the more heterogeneous the final product; thisrelationship is attributed to the specific features of dif-fusion of active nitronium ions in the cellulose matrix[5] and agrees well with the scheme represented in thisstudy. It was found that the activation energy of nitra-tion of cotton cellulose with vapors of nitric acid [7]and nitrogen dioxide [8] is about 40 kJ/mol of nitrogroups. In addition, for this system, it was shown thatthe effective diffusion coefficient of the nitrating agentin the cellulose matrix is ~10–7 cm2/s [7, 8].

Previously, the physicochemical processes in cellu-lose were described using the averaging scales thatmade it possible to reduce a heterogeneous problem toa phenomenological homogeneous model. Thus,according to the approaches discussed in the 1970sand 1980s, the kinetics of diffusion-controlled reac-tions in wood can be described using the concepts offormal isokinetic zones [9, 10]. Isokinetic zones havedifferent characteristic reaction times, which differ by

orders of magnitude; as a consequence, the totalkinetic dependence has a logarithmic form (poly-chronic kinetics) [11]. It is now becoming clear thatphenomenological approaches have not revealed thestructural features of chemical processes of wood cel-lulose esterification and have not led to a significanttechnological progress.

However, there is every reason to believe that, evenunder conditions of a high hydromodulus (~50) in cel-lulose processing, all the major physicochemical pro-cesses take place inside an isolated fiber which mostlypreserves its morphology and has a well-known struc-tural organization (Fig. 2) characterized by the pres-ence of supercoiling and hierarchical chiral phases[12]. Cellulose is known to comprise compact crystal-line regions—elementary fibrils—composed of com-plementarily packed helical macromolecules, which inturn are helically twisted to form microfibrils. More-over, the structural chiral hierarchy of cellulose obeysthe general laws governing the structure formation that

Fig. 1. Electron micrograph (Phenom G2 Pro) showing the chaos of the structural elements of semibleached sulfite wood cellu-lose: (a) without treatment and (b) after nitration.

300 µm 300 µm

(a) (b)

Fig. 2. Structure of the cellulose fiber. Comparison of the hierarchical structure of the fiber with a supramolecular TFAAS string.

20 µm

Microfibril

Elementary fibril

CrystalliteAmorphous region


CHEMICAL PHYSICS OF CELLULOSE NITRATION 247

were previously determined for chiral biomimetic sys-tems [12, 13].

It has been first shown by the authors elsewherethat the formation of hierarchical chiral phases andsupercoiling are a fundamental principle of structureformation in chiral biomimetic and living systems [13].At the same time, it is quite obvious that it is the mac-roscopic dynamics of winding–untwisting in thesupercoiling of fibrils and macromolecules [14] thatcan limit the access of reactants to the compact phasesof cellulose fibers. In this respect, an adequate physi-cal model of a cellulose fiber is a hierarchical super-coiled supramolecular string in which helical struc-tures are observed on a scale of 2 nm to 10 μm. Thesescales correspond to the scales of helical structuresformed by trif luoroacetylated amino alcohol(TFAAS) biomimetics, the macroscopic dynamics ofwhich and the features of formation of compact phaseswere previously studied by the authors [14, 15].

The aim of this study was to construct a physicalmodel of the nitration process that, taking intoaccount the characteristics of structure formation inchiral systems, would make it possible to explain thespecific features of nitration of wood, f lax, and cottoncellulose and prepare samples with a high nitrogencontent. To this end, the following was studied: (i) themorphology of cellulose and NC; (ii) the kinetics ofexothermic nitration reactions; and (iii) the circulardichroism (CD) of nanocellulose (NNC) gels, themacroscopic dynamics of untwisting of cellulosefibrils, and their van der Waals interactions.

MATERIALS AND EXPERIMENTAL METHODS

The following types of cellulose were used:

—Viscose sulfite cellulose (roll paper), OAOArkhangelsk PPM (RB).

—Cotton Cellulose, UZ Cell.

—Semibleached softwood sulfite cellulose (folder),OAO Syassky PPM (SKhP).

—Bleached softwood sulfite cellulose (liquidflow), OAO Syassky PPM (SKhZh).

—Bleached sulfate cellulose LS (folder), OAOArkhangelsk PPM (ALP).

—Flax cellulose VLV (Vologda).—Bleached hardwood sulfate cellulose (liquid

flow), OAO Arkhangelsk PPM (ALZh).—Bleached softwood sulfate cellulose (liquid

flow), OAO Arkhangelsk PPM (AKhZh).—A gel prepared in the form of a nonsedimenting

suspension of NNC in water using a colloid mill. Thesamples were prepared by mixing an aqueous solutionof NaHSO3 or H3BO3 with NNC in a volume ratio1 : 1 and subsequent gentle shaking of the resultingmixture. The weight concentrations (%) of NaHSO3are given below for the aqueous solution before mixingwith NNC; that is, the NaHSO3 concentration in themixture is 2 times lower.

The alpha-cellulose content in the commercial cellu-loses produced at the PPMs is 81–85%. Since onlybleached brands of softwood and hardwood cellulosesare regarded as original commercial cellulose, the lig-nin content in them is at a level of a few tens of a per-cent and meets the requirements of GOST 595-79.Furthermore, cold refining leads to an additionaldecrease in the content of lignins oxidized duringbleaching owing to their dissolution in an alkalinesolution.

An increase in the α-cellulose content to a levelcorresponding to GOST for cellulose for chemical pro-cessing, i.e., to a value of >95%, was achieved viarefining the commercial cellulose. Cold refining wasconducted in accordance with the procedure specifiedin GOST 6840-78 for the determination of α-cellulosecontent in commercial celluloses by the weightmethod. The process included the following steps:(i) treatment of an air-dry commercial cellulose withan aqueous sodium hydroxide solution with a concen-tration of 17.5% (hydromodulus of 15) for 45 min;(ii) dilution of the resulting mass to an alkali concen-tration of 9.5%; (iii) vacuum filtration and rinsing withthree portions of a 9.5% alkali solution (total volumecorresponds to a hydromodulus of 25); and (iv) wash-ing with water to neutral pH. It was shown that coldrefining of original cellulose in the form of a wet massfor 10–30 min leads to the formation of cellulose witha 98% content of α-cellulose. The measurementresults are shown in Table 1.

Test experiments were aimed at determining thebasic parameter of NC—nitrogen content—and study-ing the kinetics of nitration and obtaining additionaldata on the processes that occur during nitration. Thenitrogen content in NC was determined by severalmethods—elemental analysis, titrimetry, and gravim-etry—that were developed in the course of the study.

Table 1. Content of α-cellulose in commercial cellulosesamples

*A more detailed description of the cellulose types is provided inthe text.

Sample designation*, manufacturer ω, %

RB, OAO Arkhangelsk PPM 85.3UZ Cell 94.1SKhP, OAO Syassky PPM 81.6CKhZh, OAO Syassky PPM 81.6ALP, OAO Arkhangelsk PPM 81.8VLV, Vologda 85.2ALZh, OAO Arkhangelsk PPM 81.0AKhZh, OAO Arkhangelsk PPM 82.5

248


STOVBUN et al.

Elemental analysis was conducted using a CE1106automated CHN analyzer (Italy). The weighed por-tion of the sample was 0.8–1 mg. The burning tem-perature of the sample was 1050°C. The flash point inoxygen for the sample in a tin capsule was 1800°C.Oxygen was dosed using a loop-shaped edosingdevice. Separation of degradation products (nitrogen,carbon dioxide, and water) was provided by gas chro-matography. The CHN content was calculated auto-matically by the software of the instrument. The pro-gram took into account calibration coefficients deter-mined from reference standards, blank experimentresults, and the weighed portion magnitude. Theabsolute error of the method was 0.30–0.50%. Thecomplete analysis cycle time was 10 min.

SOP 2014-03 technique of OOO Biovet-ferment.The titrimetric method for the determination of theweight fraction of nitrate nitrogen consists in thedecomposition of a weighed portion of the test samplewith hydrogen peroxide and the reduction of thenitrate nitrogen to ammonia with a Devarda’s alloy inthe presence of sodium hydroxide. The amount ofammonia is determined by back titration of excess sul-furic acid after the absorption of ammonia. The anal-ysis duration is 4 h.

Rapid gravimetric method for the determination ofthe degree of nitration of cellulose (Semenov Instituteof Chemical Physics, Russian Academy of Sciences).The nitration of cellulose consists in the substitutionof the hydrogen atom of the hydroxyl groups of cellu-lose by NO2; it is accompanied by a significant changein the weight of the elementary unit of the cellulose.This feature makes it possible to calculate the depen-dence of the sample weight change during nitration onthe degree of substitution of hydroxyl groups by nitrogroups. During nitration, the C6H10O5 cellulose ele-mentary unit transits into the C6H10 – 3xN3xO5 + 6x ele-

mentary unit, where 0 < х < 1 is the degree of substitu-tion of H by NO2, i.e., the ratio between the averagenumber of substituted hydroxyls and their total num-ber equal to 3.

Figure 3 shows the plot of the dependence of nitro-gen content (wt %) on the weight change ratio (K),i.e., the weight ratio between NC and original cellu-lose. It is evident from the figure that the weight frac-tion of nitrogen in the NC, to a high precision, linearlydepends on the weight change ratio during nitration:

The rapid gravimetric method is based on theassumption that the substitution of hydroxyl groups bynitro groups is the only process that occurs duringnitration. It is assumed that a possible contribution ofthe formation of acid sulfo esters –COSO2H is negli-gible. This assumption has been confirmed by elemen-tal analysis, which has not revealed the presence ofsulfur in the samples. The results of determination ofthe nitrogen content by the gravimetric and elementalanalysis methods are almost the same.

In the experiments, air-dry cellulose strips with aweight of about 50 mg, which were preweighed on ananalytical balance with a precision of up to ±0.1 mg,were immersed in a nitrating mixture (RAM) placed ina sealed glass container at a predetermined tempera-ture; after certain time intervals, the strips were with-drawn. After that, the sample was immersed in a 60%sulfuric acid solution. The acid of this concentrationwas used as a water-containing medium to provide theinstantaneous cessation of nitration and the interme-diate dilution of concentrated sulfuric acid beforeimmersing the sample into ice water to avoid strongheating during the dilution of sulfuric acid with water.The sample was exposed to 60% sulfuric acid for 30 sand then rapidly transferred into ice water. Next, thesample was immersed into distilled water (100 mL) for10–15 min and then into a sodium bicarbonate solu-tion for 30–40 min. After that, the sample was washedwith water 3 or 4 times with holding in a large volume(~100 mL) for 10–15 min. Next, the wet samples werepressed between two sheets of filter paper and finallydried on the paper at room temperature. Final dryingwas conducted in a drying cabinet at a temperature60°C. The degree of substitution was determined usingthe well-known formula [1]

(1)

Optical microscopy was conducted using the fol-lowing instruments:

—a Mikmed-6 optical microscope (LOMO, Russia);—a Carl Zeiss Jena optical microscope (Germany);—a Nikon optical digital microscope (Japan);—a computerized optical bench (Research system

microscope) based on an Olympus BX51 microscope(Japan);

0.1057 0.0517.ω = −K

3.6 .31.1

N=N

α−

Fig. 3. Percentage of nitrogen in the NC versus celluloseweight change ratio during nitration.

11.0

12.0

13.0

14.0

11.5

12.5

13.5

14.5

1.601.55 1.65 1.70 1.75 1.80 1.85

Nitr

ogen

con

tent

, %

К



—a Veeco VCM-200 confocal optical microscope;—a Leica DMI 6000 inverted optical microscope.Electron microscopy was conducted using the fol-

lowing equipment:—A Phenom G2 pro scanning electron micro-

scope, which allows imaging in a magnification rangeof 80× to 45 000× at a resolution of up to 25 nm.

—A Hitachi700 transmission electron microscope.Atomic force microscopy. All measurements by

scanning probe microscopy were conducted on aSolver HV atomic force microscope (AFM; ZAO NT-MDT, Zelenograd, Russia) operating in the tappingmode at room temperature and atmospheric pressure.A cellulose sample was placed on the surface of a stan-dard silicon wafer used in the microelectronics indus-try, and the wafer with the sample was fixed in theAFM.

AFM measurements of the surface topography ofthe sample were conducted with a gradual scaling-upof the frame in a range of 10 000 to 4 μm2. Standardcantilevers manufactured at ZAO NT-MDT (Zele-nograd, Russia) were used. The natural frequencies ofthe cantilevers were in a range of 110–180 kHz; the tipcurvature radius was 10 nm. Most of the data wereobtained in the constant amplitude mode; the rest, inthe phase contrast distribution mode. In the topo-graphic images recorded in the constant amplitudemode, the height is represented by means of color: thehigher the location of the topography element, thelighter the color.

Dynamic viscosity of refined cellulose. The mostimportant characteristic of cellulose intended forchemical processing is the average degree of polymer-ization of the cellulose macromolecules, which is esti-mated from the dynamic viscosity of cellulose solu-tions according to GOST 14363.2. Table 2 shows theresults of measurement of the dynamic viscosity ofvarious cellulose types after cold refining (α index).These data suggest that the viscosity values depend onthe nature of the feedstock and vary in a range of 35–96 cP, which meets the requirements of GOST. Thelowest viscosity is exhibited by softwood cellulose; thehighest viscosity is found in celluloses based on ahardwood mixture. Flax cellulose occupies an inter-mediate position. The above results suggest that, forthese types of feedstock, the step of controlling thedegree of polymerization can be omitted.

CD coefficients were measured using an SKD-2CD spectra analyzer, which provides the operatingtemperature stabilization in a range of 25–75°C, atwavelengths of 250–750 nm. A 2-mL solution waspipetted into the cuvette of the analyzer. The cuvettewas placed in the working cell of the basic unit of theanalyzer. Measurements were conducted at a relativehumidity of 50–55%.

Raman spectra of the solid products were recordedusing a RamanStation-400 Raman spectrometer

(PerkinElmer). This instrument is equipped with adiode laser with λex = 785 nm (100 mW) for the exci-tation of Raman scattering, an iDus-420 detector(Andor) thermoelectrically cooled to –60°C (1024 ×255 pixel Si-CCD matrix) for Raman spectrumrecording, and a motorized XYZ-positioning stage anda color video camera with a 30× lens for the visualinspection and selection of the sample portion foranalysis (a spot with a diameter of about 5–10 μm) orlaser-beam focusing on the test sample before record-ing the Raman spectrum. During spectra recording,the video camera lens is used for transmitting afocused laser excitation onto the sample and, at thesame time, for collecting the sample-generatedRaman photons and delivering them to the spectro-graph. Owing to the echelle optical design, the instru-ment contains no moving parts, requires no premea-surement calibration, records the total spectrum in asingle scan, and provides a constant optical resolution(1, 1.5, 2, 4, 8, or 16 cm–1) in the entire measuredrange of Raman shifts (95–3400 cm–1). Raman spec-tra of cellulose samples (powdered materials mostly ofwhite color) were recorded using samples with aweight of 5–10 mg placed in 50-μL aluminum cups.To increase the sample density and improve theRaman signal, the powder was manually premolded inan aluminum container using a f lat-faced metal rodwith a suitable diameter. The spectra were recorded ata resolution of 1 or 2 cm–1 in a Raman shift range of100–3200 cm–1 at a laser excitation power of 5–75 mWin different experiments with averaging of 9–16 scansat a CCD-detector exposure time of 5–30 s. Theinstrument management, the selection of experimen-tal parameters, and the spectra processing were con-ducted using the Spectrum v.10.6 software package(PerkinElmer).

X-ray analysis of the cellulose samples was con-ducted using a retrofitted DRON-3 X-ray diffractom-eter with a copper anticathode at a voltage of 30 kV and

Table 2. Dynamic viscosity of the cellulose samples accordingto GOST 14363.2

Sample designationDynamic viscosity,

cP

SZhP (α) after dispersion on an IKA MagicLab 48SZhP (α) 49AKhZh (α) 37AKhZh (α), 1% NaHSO3 solution, 68°C, 60 min 35AKhZh (α) 0.1 M H3BO3, 65°C, 80 min 37ALP (α) 96ALP (α) after grinding 92ALZh (α) 81VLV (α) 70

250


STOVBUN et al.

a current strength of 20 mA (Ni filter). The X-raywavelength was λ ≈ 0.15 nm. Measurements were con-ducted at a temperature of 25°C and a relative humid-ity of 50–55%. The original sample was placed in a f latglass cell with a diameter of 10 mm and a depth of1 mm. In some cases, the sample was held in the air orin a vacuum for a certain time.

Preparation of fibrous NNC semiproducts (woodcellulose, cotton cellulose) was conducted using thefollowing IKA dispersing units (Germany): a EuroStar2000 unit (stirrer with a dissolving nozzle; a rotationspeed of 150–450 rpm) and an IKA MagicLab dis-persing unit (rotor–stator couple of the colloid milltype; a rotation speed of 10000–18000 rpm). Themass concentration during dispersion was 1–3%.

The process time and the rotor–stator gap in thecolloid mill were varied depending on the fibroussemiproduct type. One-percent nonsedimentingnanogels of these NNC semiproducts of wood cellu-lose causing the Tyndall effect were used.

RESULTS AND DISCUSSIONThe elementary structural unit of native cellulose is

an unbranched polymer chain of β-D-glucopyranoseresidues which is bound to other chains, on average,via one or two hydrogen bonds per unit. The macro-molecule length can significantly vary. Thus, thedegree of polymerization is 2500–3100 for wood cellu-lose, 9300–10800 for cotton fibers, and 36000 for f laxfibers [16, 17]. Native cellulose is composed of two—crystalline and amorphous—fractions [1]. The micro-crystalline regions are associated with the elementarycellulose fibrils; the amorphous regions are associatedwith the low-molecular-weight cellulose fractions

located between them [18, 19]. The minimum appar-ent thickness of the elementary cellulose fibrils is 3–5 nm; it is characteristic of wood cellulose [18]. Thethickness of the elementary fibrils of the f lax and cot-ton celluloses is higher; it is 5–6 and 7–9 nm, respec-tively [20].

The elementary fibrils of softwood cellulose, whichare composed of 24 densely and complementarypacked macromolecules, have a helical crystallinestructure [21]. The packing of macromolecules in ele-mentary fibrils is attributed to the specific features oftheir biosynthesis; that is, the crystallization and for-mation of fibrils occur simultaneously during the syn-thesis thereof [22]. Thus, all celluloses isolated from aparticular plant are characterized by a certain cross-sectional dimension of microcrystalline regions.

It is well known that wood cellulose preserves themorphology of the original plant raw material duringthe industrial separation of it from lignin and hemicel-luloses (pulp cooking) [1]. After delignification ofwood cellulose, for example, at the Arkhangelsk PPM,individual cellulose fibers are clearly visible in theresulting mass (Fig. 1). Note that the morphology ofthe prepared cellulose samples does not depend on theisolation method; that is, sulfate or sulfite celluloses ofwood origin exhibit an identical morphology. Themorphology of the cellulose is preserved even undersevere conditions of the nitration reaction (Fig. 1).

An extremely convenient method for studying themorphology and molecular structure of cellulose isRaman spectroscopy (Fig. 4) [1]. It is noteworthy thatthe original wood celluloses from the different suppli-ers are very similar in molecular structure, as evi-denced by the coincidence of Raman spectra in a

Fig. 4. Raman spectra of original commercial celluloses: (1) folder and (2) liquid f low of the Syassky PPM and (3) bleached cel-lulose of the Arkhangelsk PPM, roll paper.

0.5

1.5

2.5

3.5

4.5

1.0

2.0

3.0

4.0

5.0

200 600 1000 1400 1800Raman shift/Wavenumber, cm–1

Ram

an in

tens

ity, a

rb. u

nits

1

2

3



range of 1500–800 cm–1 (Fig. 4). Moreover, the origi-nal samples exhibit almost the same packing, as evi-denced by the identity of the phonon bands at lowRaman shifts in a region of 250–600 cm–1, whichare responsible for the packing of cellulose macromol-ecules in the crystalline regions. The differences inthe Raman spectra observed in a region of 1700–1500 cm–1 are attributed to different lignin content(Fig. 4). The intensity of the Raman band with a max-imum in the vicinity of 1600 cm–1 decreases in propor-tion to the degree of bleaching of cellulose, which cor-responds to the decrease in the lignin content.

The preservation of the hierarchy of linear scales ofnative cellulose during delignification and nitrationindicates the occurrence of effective structure forma-tion and self-organization of the cellulose macromol-ecules [1], which severely limit their self-diffusion andsubsequent swelling/dissolution of the fibers [1]. Thisself-organization is caused by supercoiling [12] andthe presence of a hierarchy of molecular and supramo-lecular helical levels—such that the fibers can be rep-resented like a twisted rope—in the cellulose structure.In addition, in wood cellulose, the additional bindingof fibrils into fibers via hemicelluloses occurs to agreater extent than in f lax or cotton celluloses, whichcontain significantly lower amounts of low-molecu-lar-weight binding components (hemicelluloses andresidual lignin).

Figure 5 shows the experimental kinetic curve ofnitration of softwood cellulose produced at theArkhangelsk PPM (AKhZh) after refining it with aNaOH solution, i.e., after removal of low-molecular-weight hemi-, β-, and γ-celluloses playing the role ofa component that binds the supercoiled fibrils. In thecase of nitration of the refined cellulose, the nitrogencontent achieves 13.5%; however, under the samenitration conditions for the original cellulose, it doesnot exceed 12%. The kinetic curve of nitration exhibitstwo pronounced stages [3], namely, a fast stage, inwhich the nitrogen content achieves ~12.5% within 2–3 min, and a slow stage, in which an additional nitro-gen amount of 0.5–1.0% is “gained” within 25–30 min (see Fig. 5).

The expansion of the kinetics of nitration in twoexponents gives characteristic times of ~0.5 and ~10 min(Fig. 5). In view of the above described features of thehierarchical structure of cellulose, it apparently canhave only two compact organization levels. Therefore,only two effective isokinetic zones should be consid-ered; one of the zones is the region inside the elemen-tary fibrils with diameter X of less than ~10 nm; theother zone is the region inside the microfibrils with adiameter of less than ~100 nm, which seems to be con-sistent with the behavior of the observed two-expo-nential kinetics.

Based on the concept of the two isokinetic zones,the time of molecular diffusion during cellulose nitra-tion for D ~ 10–11 m2/s is determined as follows [23]:

(2)

However, comparison of the found diffusion timewith the characteristic times of nitration to degrees ofsubstitution of 1–2 suggests that nitration is not lim-ited to molecular diffusion. The estimation of the dif-fusion coefficient from characteristic times of the slownitration stage also leads to diffusion coefficient valuesof D ~10–18–10–20 m2/s, which do not have any physi-cal sense for low-molecular-weight compounds [23].Thus, the estimates show that molecular diffusion isnot the rate-limiting step of nitration at any conver-sion value.

Since in the conventional scheme of nitration it isthe nitronium ion that acts as a reaction agent [24], itis necessary to explore the possibility of transfer ofcharged particles inside microcrystalline celluloseregions at high degrees of nitration. The surface ofnanofibrils is much more highly accessible to nitratingagents than their internal volume; therefore, it is thesurface that is first nitrated. Based on the data that acotton cellulose nanofibril consists of 24 cellulosechains [21], it is easy to calculate that the fraction ofOH groups located on the elementary fibril surface is1–1.4 per residue. This result is in good agreementwith the experimentally observed threshold of thebeginning of structural rearrangements of thecellulose matrix during nitration. It is the degree ofsubstitution of ~1.3 that provides the occurrence ofreflections corresponding to crystalline cellulose trini-trate in the X-ray diffraction patterns (Fig. 6).

The cellulose nitration experiments revealed a sig-nificant threshold variation (1.3-fold) in the free vol-

( )− −τ −2

6 40.4 10 10 s.6X= =

D

Fig. 5. Kinetic curve of nitration of softwood cellulose inthe form of a loose folder (ρ ~ 0.2 g/cm3) with a thicknessof about 1 mm at 20°C. The circles and the line show theexperimental data and the two-exponential approxima-tion, respectively, with characteristic times of 36 s and7.6 min, respectively.

0.5

1.5

2.5

1.0

2.0

3.0

0 10 20 305 15 25Time, min

Deg

ree

of n

itrat

ion

252


STOVBUN et al.

ume of the fiber or swelling thereof (Fig. 7) at a degreeof substitution of 1.1–1.3, i.e., at the values providinga change in the X-ray diffraction patterns. It is obviousthat the surface layer of the elementary fibril is the firstto undergo nitration.

As cellulose nitrate accumulates in the surface layerof microfibrils, their internal volume is surrounded bya layer of hydrophobic densely packed NC [5], whichprevents the penetration of nitronium cation inside thefibrils. In fact, the Born energy of transfer of a mon-ovalent ion with a radius of r ~ 1.5 Å [25] from amedium with a dielectric constant of about ε2 ~ 100into a hydrophobic region (ε1 ~ 1–5) is, on the aver-age, as follows [26]:

(3)⎛ ⎞Δ − − −⎜ ⎟πε ε ε⎝ ⎠∼

2

0 1 2

1 1 1 1.5 eV.8

qG =

At this activation energy, the penetration of inside a fibril takes a time longer than 3 × 104 s, whichshould actually stop the nitration process. This effectis fundamental for an increase in the heterogeneity ofthe final NC with increasing actual nitronium ionconcentration in the reaction mixture [5, 6].

Thus, nitration is undoubtedly associated with thestructural rearrangement of elementary fibrils, duringwhich macromolecule fragments become accessible tointeraction with nitronium ions. In addition, thestructural rearrangement of fibrils is dynamic andreversible because the morphology of the original andnitrated celluloses is almost the same (Fig. 1).

The previously observed effects of reversible swell-ing of cellulose fibers in various solvents [1] lead to theformation of microscopically observed blisters, i.e.,bulbs, which in fact are regions of an unwound fibril of

2NO+

Fig. 6. Variation in the X-ray diffraction patterns of cellulose during nitration for degrees of substitution of (1) 2.64, (2) 2.40,(3) 2.12, and (4) 1.3. The peak in the region of 12°–15° corresponds to crystalline cellulose trinitrate.

100

200

300

400

500

600

700

800

5 15 2510 20 30 40 5035 45

1

2

3

4

Angle, deg

Inte

nsity

, arb

. uni

ts

Fig. 7. Untwisting of spruce tracheids in concentrated sulfuric acid. Published by courtesy of E.V. Novozhilov [27].

56 µm 56 µm

(a) (b)



an increased volume. The formation of these bulbsduring the swelling of a cellulose fiber in sulfuric acidis shown in Fig. 7 [27].

According to Euler [28], during the twisting ofstrands, the friction forces exponentially depend onthe number of turns:

(4)

where n is the number of turns and α is the coefficientof friction. Therefore, the swelling of a cellulose fiberwithout untwisting thereof is unrealistic.

In other words, the supercoiled structure of thefiber and the macroscopic dynamics of untwisting ofsupercoiled hierarchical structures [11] are factors thatdetermine the kinetics of chemical transformationsduring nitration. Note that, taking into account theabove Euler’s energy restriction (4), the observedswelling and bulb formation processes (Figs. 7, 8) are,in fact, experimental evidences of the reality ofuntwisting of cellulose fibers.

Consider the kinetics of nitration of the refined(purified from low-molecular-weight impurities)wood cellulose at varying temperatures (Fig. 9). Atpositive temperatures, the initial portion of the kineticcurve of nitration has a pronounced linear pattern(Fig. 9). Moreover, at a temperature of –10°C, thekinetics is linear over the entire range of observation.This form of initial portions corresponding to a zero-order reaction takes place if the reaction rate dependson the OH group concentration only slightly or doesnot depend at all. Formally, this situation is imple-mented if the dimensions and characteristics of theregion in which the reaction occurs do not changewith time. The occurrence and stability of parametersof this region can be attributed to effective nitrationonly in the region of untwisting of elementary micro-fibrils. In this case, at a constant untwisting speed, thereaction will occur only in a cylindrical volume withlength l comparable, for example, with the size of abulb in an untwisting fiber or fibril (Fig. 8). Therefore,the reaction rate will be a constant and the total kinet-ics will be of zero order.

Substitution of the experimental data shows that achange in the nitration rate constant with increasingtemperature in the initial portion is satisfactorilydescribed by the Arrhenius equation with an activationenergy of Ea ≈ 46 kJ/mol and a preexponential factorof K0 ≈ 4.5 × 105 s–1. The respective rate constants wereestimated from the slopes of the kinetic curves in Fig. 9.In this case, total time τtot of two successive pro-cesses—the formation of a reaction site for time τс andthe collision of the nitronium ion with the OH groupfor time τcoll is τtot = τsite + τcoll, while the frequency isνtot = νsiteνcoll/νsite + νcoll. Since the frequency of colli-sions of the nitronium ion with the OH group has avalue of ~1012 s–1 and the relationship νcoll νsite isapparently valid, we can assume that νtot ≈ νsite and

( )exp nπα∼ 2 ,F

@

there is a significant entropy contribution responsiblefor the reaction rate because the formation of a reac-tion site is physically associated with fibril untwistingcooperative processes. According to the literaturedata, the nitration reaction is nearly activationless[24]; therefore, the activation energy value can beattributed to the energy of complementary intermo-lecular interaction of the glucopyranose rings of cellu-lose, which provides a close packing of the rings and

Fig. 8. Dependence of the average thickness of the cellu-lose/NC fiber on the nitrogen content according to elec-tron microscopy. Large measurement errors are associatedwith the initial variability of the fiber thickness. Data onthickness are given for ~102 fibers.

6

8

10

12

14

16

18

20

22

0 0.5 1.0 1.5Degree of nitration

Diameter, μm

Fig. 9. Kinetic curves of nitration of wood cellulose at thedifferent temperatures: triangle, –10°C; square, 2°C; dia-mond, 20°C, and circle, 34°C. The solid lines show thelinear portions of the curves.

0

0.5

1.5

2.5

1.0

2.0

3.0

0 10 205 15 25Time, min

Deg

ree

of su

bstit

utio

n

254


STOVBUN et al.

hinders the penetration of the nitrating agent into thecrystalline regions.

A simple estimate of intermolecular interactionenergy yields a similar value. Cellulose nanofibrilscontain, on the average, two interchain hydrogenbonds per glucopyranose unit [1], which gives a bind-ing energy of chains of about 30 kJ/mol. The helicityof cellulose microfibrils provides the complementarityof interaction of the cellulose chains. A glucose resi-due consists of 21 atoms; assuming that the averageenergy of the van der Waals interaction between twoatoms is ~1 kJ/mol, for the interchain binding energy,we obtain a gain of ~20 kJ/mol. Thus, for the averagebinding energy of cellulose chains, we obtain a value of~50 kJ/mol, which is close to the previously calculatedactivation energy of the nitration process (Ea).

At the same time, it is known that nitration is anexothermic reaction accompanied by a significant—about 130 kJ/mol—heat evolution [29]. Since thekinetic characteristics are markedly sensitive tochanges in temperature by as little as 10°C, the localoverheating in the exothermic reaction has a consider-ably smaller value. Otherwise, the reaction mixturetemperature registered by external devices would besubstantially lower than the actual temperature atwhich the reaction occurs and the temperature depen-dence would be much less pronounced. To explain thiseffect, it is necessary to estimate the size of the regionin which the reaction occurs and the thermalizationtime. Thermalization time τ of the energy released inthe exothermic reaction in the form of nonequilibriumphonons, particularly high-energy phonons, in thethermalization region of size R, is determined by thecooling of it to a thermostat temperature:

(5)

where с is the heat capacity, ρ is the density, and λ isthe heat conductivity. For water, after substitution ofrespective values, at thermalization region size R of 1–10 nm, this estimate yields a value of τ ~ 10–11–10–9 s.In this case, R is an adjustable parameter, which isequal to the distance a high-energy phonon travelsbefore it undergoes thermalization; it was estimatedtaking into account the crystallinity of the lattice of theelementary fibril. At the same time, during nitration,there occur no more than 1021 elementary events ofnitration per cubic centimeter per second. In this case,in the attached volume of radius R of the glucopyra-nose ring (V ~ 4 × 10–18 cm3), there occur no morethan 4 × 103 substitution events per about second orone event per 2.5 × 10–4 s. Further, with allowance forthe τ value, it becomes obvious that each successiveevent of nitration in the thermalization region occursat the thermostat temperature. It is this feature that isresponsible for the sensitivity of the kinetic parametersof the highly exothermic nitration reaction to smallchanges in the thermostat temperature.

2

,c= ρτλR

AFM studies of the cellulose feedstocks revealedthat the untwisting of the cellulose fibers occurs at theend faces, as evidenced by the formation of a dispersedphase observed in the fiber topography profile. Notethat the observed untwisting is apparently geometri-cally similar to the untwisting of supercoiled micro-scopic supramolecular TFAAS strings (Fig. 10c). Letus sequentially consider the dependence of the nitro-gen content on the different factors that reduce thebinding of elementary fibrils and fibrils in a super-coiled state (in “bundles”) using the example of thecellulose manufactured at the Syassky PPM. Thus, inthe case of nitration of the unrefined cellulose, thedegree of nitration achieves 11.46 ± 0.34%. Afterrefining and the removal of low-molecular-weight cel-lulose fractions binding the fibrils, the nitrogen con-tent increases to 12.10 ± 0.36%.

This process is also intensified with the appearanceof charged groups on the lateral surface of the fibrils.For example, the appearance of sulfo groups on thesurface of the cellulose fibers gives rise to a negativecharge on it. In general, the mechanism of this reac-tion can be similar to the SN1 mechanism of the ligninsulfonation reaction [1]. In an acid medium, a carbo-cation is formed; it can be subsequently attacked bythe sulfite present in the medium:

The resulting sulfite is unstable and analyticallyundetectable; however, even a very small excess nega-tive charge is sufficient for the untwisting process.Similar charges on individual cellulose chains in thecomposition of the elementary fibril lead to theirrepulsion; this process should intensify the untwisting.In fact, in a NaНSO3 solution, the amplitude of CDspectra decreases in a wavelength region of 250 nm;this decrease apparently corresponds to the effectiveuntwisting of the elementary fibrils of the NNC tomacromolecules (Fig. 11). Since a strict interpretationof CD spectra is severely hindered, particularly forcomplex systems, such as a NNC gel, here, the inter-pretation of the observed chirooptic effect is also basedon the assumption that the untwisting of supercoiledstrings in the model TFAAS system is also accompa-nied by a similar decrease in the CD intensity in a sim-ilar wavelength range [14].

The available threshold effect consists in a varia-tion in CD, rather than in the solution absorption ingeneral. In fact, the threshold effect corresponds to aNaHSO3 concentration in the final mixture of 2%(0.16 M); in addition, the measured absorption spectrain a range of 220–350 nm for NNC gel samples withthe same NaHSO3 concentrations showed goodagreement for sub- and post-threshold concentrations(2.5%, 0.2 M).

3 2R HSO R O SO H,+ −+ → − −

2 2 2 3R O SO H H O R O SO H O .− +− − + → − − +



Fig. 10. Destruction (untwisting) of structures: (a) a topographic AFM image of a cellulose fiber with a marked region, (b) thesurface profile of the marked region of the fiber, and (c) an optical micrograph showing the untwisting of a supramolecularTFAAS-5 string [11].

20

30

40

50

60

70

0 100 200 300 400 500 600 700

3 µm

02.

7

0.2

0.4

0

20

40

60

80

100

120

140

0.6

0.8

1.0

0 0.60.40.2 0.8 1.0µm

nm

nm

nm

µm(a) (c)

(b)

The treatment of the cellulose produced at theSyassky PPM (SKhZh) with a 1% NaHSO3 solution ata temperature of 60°C for 200 min leads to an increasein the nitrogen content in the final nitration product to12.7 ± 0.4% compared with the nitrogen contentobtained in the nitration of the original cellulose10.8 ± 0.3%. Experimental data on the nitration of therefined cellulose (SKhZh) also show that the treat-ment of the cellulose with 1% NaHSO3 at a tempera-ture of 60°C for 200 min leads to an increase in thedegree of nitration from 12.1 ± 0.4% for the originalrefined cellulose to 12.9 ± 0.4% (Table 3).

Another way to reduce the strength of interactionof elementary fibrils is to provide a decrease in the vander Waals interaction of the fibrils. Thus, during thetreatment of the refined cellulose (SKhZh) sampleswith acetone, the polar OH groups on the surface ofthe elementary fibrils are doped; that is, they attachacetone molecules via the formation of a strong hydro-gen bond. Accordingly, the fraction of OH groups on

the surface decreases, while the fraction of methylgroups with lower polarizability increases. Thestrength of the van der Waals interaction of the neigh-boring functional groups decreases because it is pro-portional to the product of polarizability of the groups.Therefore, the macroscopic van der Waals interactionof the fibrils should decrease, while the degree ofnitration thereof should increase (Table 3).

Treatment with a 0.1 M Н3ВО3 solution at a tem-perature of 60°C for 200 min also leads to an increasein the degree of nitration to 13.07 ± 0.39 apparentlyowing to the plasticization of the cellulose matrix byН3ВО3. Since the treatment with boric acid does notlead to significant changes in the CD spectra of cellu-lose gels, the observed effect is apparently attributedalso to the weakening of the interaction of the cellulosemacromolecules owing to the formation of a compet-ing lattice of hydrogen bonds; it is this feature thatfacilitates untwisting during nitration.

256


STOVBUN et al.

Consider differences in the rates of nitration ofwood, f lax, and cotton celluloses, which can easily berelated to different diameters of the elementary cellu-lose fibrils characteristic of these plants [20]. The dis-tance between the fibers in the contact areas is limitedto the profile of the fibers. On the assumption of a geo-metric similarity between the hierarchical levels of thestructural organization of cellulose, the scale of theirprofile will be proportional to the fiber diameter.Thus, effective distance h between the fibers in thecontact areas is proportional to the diameter of thesefibers. In view of the same similarity, pitch H of thehelix formed via the interweaving of thinner fibers isalso proportional to their diameter:

(6)

where ε and α are the respective proportionality coef-ficients.

The elastic energy due to the helimerization offibers is calculated as follows. Let the fiber be an elas-tic helically curved rod. The curve of the fiber in thehelical twist can be considered weak. Elastic energyWel of the weakly curved rod is defined by the follow-ing relationship [30]:

(7)

where E is the Young’s modulus of the rod material;the z axis is directed along the axis of the undeformedrod; X(z) and Y(z) are the deviations of the rod axisfrom the z axis owing to deformation along the x and yaxes, respectively; and I1 and I2 are the secondmoments of area relative to the x and y axes, respec-

, 1, , 1,h = D H = dε ε α α! !

⎛ ⎞ ⎛ ⎞⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠∫

2 22 2

el 1 22 2 ,2E d Y d XW = I + I dz

dz dx

tively. In the case of a rod with a circular cross section,for the second moments of area, we obtain

(8)

To describe the helical deformation of the rod, the fol-lowing expressions are taken for a uniform twist of therod with pitch H and amplitude A:

(9)

Hence,

(10)

From Eqs. (8) and (10), we find

(11)

that is, the integrand in Eq. (7) is independent of z;therefore, the integration over z is reduced to multiply-ing it by the fiber length. Take into account that elasticenergy is exhibited by both of the interwoven fibers.Thus, their total elastic energy per unit length Wel willbe as follows:

(12)

Twist amplitude A (measured from the deviationfrom the initial position of the geometric center of therod with a circular cross section) can easily beexpressed in terms of fiber diameter d and distance hbetween the fibers:

4

1 2 .64dI = I = π

( ) ( )cos , sin .z zX z = A Y z = AH H

( )2 2

2 2 2 2cos , sin .d X A z d Y A z=H Hdz H dz H

⎛ ⎞− = −⎜ ⎟⎝ ⎠

2 22 2 4

1 22 2 4 ,64

d Y d X d AI + I =dz dx H

⎛ ⎞ ⎛ ⎞ π⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠

π 4 2

el 4 .64Ed AW =

H

Fig. 11. CD spectra of cellulose at the different NaНSO3 concentrations: triangle, 0.5%; diamond, 1%; square, 1.5%; and circle,2%. A variation in the monotonic behavior of the CD dependence on sodium hydrosulfite concentration is clearly seen.

5

6

7

8

9

10

11

12

13

14

220 240 260 280 300Wavelength, nm

CD

inte

nsity

, 10–

3 cm

–1



(13)For distance h between the fibers and helix pitch H,

Eq. (6) can be used. Eventually, taking into accountthat ε 1, Eq. (12) for the specific (per unit length)elastic energy of the interwoven fibers takes the form

Global: ε 1 is incorrect.

(14)

Thus, in terms of the model under discussion, theelastic energy of interwoven fibers (per unit length) isdirectly proportional to the square of the fiber diame-ter; that is, the thinner (wood) fibers undergo repul-sion (owing to a tendency to straighten themselves)weaker than the thicker (cotton) fibers do. Experi-mental value of α is ~102; that is, the helix pitch of thefiber is on the order of a hundred of the fiber diame-ters. However, even in the case of a high Young’s mod-ulus of ~1011 N/m2, which is characteristic of the hard-est crystals, from Eq. (14) in a diameter range of d =3–7 nm, we obtain Wel = 2 × 10–6–10–5 eV/nm, whichis 4–5 orders of magnitude lower than the energy ofthe van der Waals attraction of fibers given in Table 4.Thus, in the case of a supercoiled packing of the fiberwith this pitch, the elastic energy is negligible and willnot be considered in further calculations.

Owing to the partial sulfonation of cellulose fiberssubjected to treatment with a sodium hydrosulfitesolution, similar (negative) elementary charges eappear on the surface of thin fibers and lead to therepulsion of the fibers. Electrostatic repulsion is hin-dered by van der Waals attraction. Elementary fibrilstwisted into a helix with a large pitch can be repre-sented as parallel at a length longer than L at which thehelix portion weakly deviates. In this case, van derWaals attraction energy WvdW of the fibers in a portionwith length L is as follows [14]:

(15)

The van der Waals attraction force is determined bydifferentiation with respect to h:

(16)

The van der Waals force per unit length of the attractedfibers will be as follows:

(17)

Coulomb repulsion force FQ, which provides therepulsion and untwisting of fibers, can be regarded asa mutual repulsion of two similarly charged plates withsurface charge density σ, length L, and width l = νd,where ν ~ 0.1 is the proportionality coefficient. Thisforce is as follows:

.A d h= +

!

!

( )⎡ ⎤π ε π⎢ ⎥

α α⎣ ⎦∼

22 2

el 4 4

1.

64 64+

W = Ed Ed

( )−1 2

vdW 3 .24 2AL dW =

h

( )−1 2

vdW 5 .16 2AL dF =

h

( ) ( )ε ∼

1 2 1 25vdW 5 2 2

12 .16 2 16AL d Af = =

h d d

(18)

where ξ is the dielectric permeability of the medium.The Coulomb force per unit length of the fibers will be

(19)

The threshold linear concentration of elementarycharges along the fiber at which the forces of mutualattraction and repulsion become equal will be as fol-lows:

(20)

Expressions (17) and (19) suggest the following:—In the case of thinner fibers of wood cellulose,

the Coulomb force normalized per unit length thatprovides the untwisting of the fibers is lower than theforce for thicker fibers of cotton and flax cellulosesbecause the force is proportional to the fiber diameter.

—In the case of thinner fibers of wood cellulose,the van der Waals attractive force of fibers normalizedper unit length that provides the winding of the fibersis higher than the force for thicker fibers of cotton cel-

πσξQ

2 ,lLF =

πσ πσ νξ ξ

∼

2 2

Q2 2 .I df = = d

( )1 2

1 25.

32 2L

2

An =e d

εν⎡ ⎤⎢ ⎥ξ π⎣ ⎦

Table 4. Energy (eV/nm) of van der Waals attraction andelectrostatic repulsion of cellulose nanofibrils with differentdiameters

d, nm WvdW Wel W = WvdW – Wel

3 0.185 0.013 0.1724 0.139 0.022 0.1175 0.111 0.036 0.0756 0.092 0.051 0.0417 0.079 0.070 0.009

Table 3. Nitrogen content in NC prepared by physicochemicalmodification of cellulose

Sample designation Nitrogencontent, wt %

SZhP nontreated 10.8 ± 0.3SZhP, 1% NaHSO3 solution, 60°C,200 min

12.67 ± 0.38

SKhP (α) 12.1 ± 0.4SKhZh (α), 1% NaHSO3 solution, 60°C,200 min

12.9 ± 0.4

SKhZh (α), acetone (soaking) 12.96 ± 0.39SKhZh (α), 0.1 М H3BO3, 60°C, 200 min 13.07 ± 0.39AKhZh (α) 13.46 ± 0.40VLV nontreated 11.28 ± 0.37VLV (α) 13.17 ± 0.39

258


STOVBUN et al.

lulose because the force is inversely proportional to thesquare of the diameter. The respective van der Waalsattraction energy values and the linear densities of thecharge sufficient for compensating for it are listed inTables 5 and 6.

Thus, in the presence of a f luctuating charge of thesame magnitude, cotton cellulose fibers must undergomore effective untwisting under the action of repulsingCoulomb forces. Apparently, this feature is anotherfactor leading to a higher degree of nitration of this cel-lulose compared with that of wood and flax celluloses.

CONCLUSIONS

The kinetics of cellulose nitration has been studied;it has been revealed that it is of zero order in the fastreaction stage. It has been found that the activationenergy of the reaction is Ea ~ 46 kJ/mol and the preex-ponential factor of the rate constant is K0 ~ 105 s–1.

It has been theoretically shown that the nitrationprocess is limited to the macroscopic dynamics ofuntwisting of cellulose fibers and fibrils, elementaryfibrils, and macromolecules, rather than to the diffu-sion of the reactants. This assumption has been exper-imentally confirmed by the zero order of the kineticsof nitration, the preexponential factor K0 and activa-tion energy Ea values, AFM micrographs, opticalmicroscopy data, a decrease in the amplitude of theCD spectra of cellulose in NaHSO3 solutions, and a

large body of data on nitration of the samples(Table 3).

In terms of physics, this finding is attributed to theformation of a Born potential barrier of ~1 eV to thepenetration of the nitronium ion into the internal vol-ume of nanofibrils with completely nitrated (hydro-phobic) outer layers at a degree of substitution of ~1.3.This barrier arises because of spatial restrictions in themolecular dynamics of the macromolecules and afree-volume deficit associated with the supercoiling ofthe elementary cellulose fibrils. Phenomenologically,this feature corresponds to the impossibility of transi-tion of the ion from the RAM with a dielectric perme-ability of ε ~ 100 into the nitrated regions of the nano-fibrils with ε ~ 3–5 across a barrier of ~1 eV.

A phenomenological model to describe the elec-trostatic and van der Waals interactions of elementaryfibrils has been constructed on the basis of similarityof scales in the system of supercoiled elementaryfibrils. According to this model, during nitration, thefree-volume deficit is compensated for by the Cou-lomb repulsion of similarly charged elementary fibrilsattracted by van der Waals forces and the subsequentuntwisting of the fibrils.

This process occurs more effectively in cottonfibers; this fact is attributed to the large diameter oftheir elementary fibrils, which have a thickness ofabout 7–9 nm. In thin fibrils of wood cellulosestrongly bound by low-molecular-weight celluloses(3–5 nm), the process is intensified during celluloserefining (Table 3). In weakly bound cotton fibers, theuntwisting of thicker elementary fibrils (7–9 nm) canoccur spontaneously because of the f luctuating sur-face charge that provides repulsion and untwisting ofthe fibrils; this feature leads to a higher nitrogen con-tent after the nitration of cotton cellulose. In the inter-mediate case of f lax fibers containing elementaryfibrils with a diameter of 5–6 nm, it is also necessaryto remove low-molecular-weight celluloses; however,their fraction is most probably lower because a fairlyeffective nitration process can occur without refining(Table 3).

It has been experimentally shown that there arethree alternative paths to select an optimum techno-logical mode. The first path is to reduce the amount ofthe “glue” binding nanofibrils via removing lignin andamorphous low-molecular-weight celluloses duringrefining. Next, it is possible to increase the specificsurface charge of nanofibrils, for example, by a treat-ment with sulfurous acid. Finally, the energy of the vander Waals attraction between the elementary fibrilscan be decreased, for example, by a pretreatment withacetone or boric acid.

Thus, the effect of the helical structural anddynamic factors on the occurrence of the cellulosenitration reaction has been studied for the first time; ithas been found that the chemical reaction includes aset of physical stages that are responsible for the reac-

Table 6. Energy of van der Waals attraction and linearcharge density compensating for it as a function of the linearcharge density versus the fibril diameter

d, nm* WvdW, eV/nm nL

3 0.185 0.234 0.139 0.205 0.111 0.186 0.092 0.167 0.079 0.158 0.069 0.149 0.062 0.13

Table 5. Range of location of charged groups on cellulosenanofibrils that is sufficient for the untwisting of the respec-tive microfibrils as a function of nanofibril thickness

d, nm Lmin, nm Lmax, nm

3 1.3 26.7

4 1.5 29.5

5 1.8 36.8

6 2.8 56.1

7 11.0 219.0



tion rate. The proposed physical mechanism ofdecrease in the cellulose nitration rate explains theentire range of effect observed during the nitration ofraw materials of different origin.

REFERENCES1. N. I. Nikitin, Chemistry of Wood and Cellulose (Akad.

Nauk SSSR, Moscow, Leningrad, 1962) [in Russian].2. I. L. Knunyants and N. S. Zefirov, Chemical Ency-

clopedy (Sov. Entsiklopediya, Moscow, 1988) [in Rus-sian].

3. E. M. Belova, N. G. Vais, V. F. Sopin, A. I. Kazakov,Yu. I. Rubtsov, G. B. Manelis, and G. N. Marchenko,Russ. Chem. Bull. 38, 2244 (1989).

4. V. A. Rafeev, Yu. L. Rubtsov, T. V. Sorokina, andN. V. Chukanov, Russ. Chem. Bull. 48, 66 (1999).

5. V. I. Kovalenko, V. F. Sopin, and G. M. Khrapkovskii,Structural Kinetic Features of Preparation and ThermalDegradation of Cellulose Nitrate (Nauka, Moscow,2005) [in Russian].

6. G. N. Marchenko, V. F. Sopin, et al., Vysokomol. Soe-din., No. 5, 1066 (1989).

7. V. A. Rafeev, Yu. L. Rubtsov, and T. V. Sorokina, Russ.Chem. Bull. 45, 328 (1996).

8. V. A. Rafeev, Yu. L. Rubtsov, and T. V. Sorokina, Russ.Chem. Bull. 45, 1879 (1996).

9. A. I. Mikhailov, L. P. Bel’kova, and V. S. Gromov,Khim. Drev., No. 6, 50 (1980).

10. A. I. Mikhailov, L. P. Bel’kova, and V. S. Gromov,Khim. Drev., No. 6, 59 (1980).

11. E. T. Denisov, O. M. Sarkisov, and G. I. Likhtenshtein,Chemical Kinetics (Khimiya, Moscow, 2000) [in Rus-sian].

12. S. V. Stovbun, A. A. Skoblin, A. M. Zanin, et al., Bull.Exp. Biol. Med. 154, 34 (2012).

13. V. A. Tverdislov, Biophysics 58, 128 (2013).

14. S. V. Stovbun and A. A. Skoblin, Mosc. Univ. Phys.Bull. 67, 317 (2012).

15. D. V. Zlenko and S. V. Stovbun, Russ. J. Phys. Chem. B8, 499 (2014).

16. F. Revol, H. Bradford, J. Giasson, R. H. Marchessault,and D. G. Gray, Int. J. Biol. Macromol. 14, 170 (1992).

17. G. C. Ruben, G. H. Bokelman, and W. Krakow, PlantCell Wall 399, 78 (1989).

18. K. Muhlethaler, Ann. Rev. Plant Phys. 42 (18), 24(1967).

19. A. N. J. Heyn, J. Cell Biol. 9, 181 (1966).20. R. H. Newman, Solid State Nucl. Magn. Reson., 15

(1999).21. A. N. Fernandes, L. H. Thomasb, C. M. Altanerc,

et al., Proc. Natl. Acad. Sci. 108, E1195 (2011).22. C. Somerville, Ann. Rev. Cell Dev. Biol. 22, 53 (2006).23. G. Murch, Diffusion in Crystalline Solids, N.Y., Acad.

Press, 1984.24. C. K. Ingold, Structure and Mechanism in Organic

Chemistry (Cornell Univ., Ithaca, 1969).25. V. I. Gol’danskii, L. I. Trakhtenberg, and V. N. Flerov,

Tunneling Phenomena in Chemical Physics (Nauka,Moscow, 1986) [in Russian].

26. J. N. Israelachvili, Intermolecular and Surface Forces(Academic, New York, 2007).

27. E. V. Novozhilov, D. G. Chukhchin, K. Yu. Terent’ev,and I. A. Khadyko, Khim. Rastit. Syr’ya, No. 2, 15(2012).

28. A. A. Silin, Friction and We (Nauka, Moscow, 1987) [inRussian].

29. G. Considine, Van Norstrand’s Scientific Encyclopedia(Wiley, New York, 2006).

30. L. D. Landau and E. M. Lifshitz, Course of TheoreticalPhysics, Vol. 7: Theory of Elasticity (Nauka, Moscow,1987; Pergamon Press, New York, 1986).

Translated by M. Timoshinina

Documents

KINETICS AND MECHANISM OF CHEMICAL REACTIONS. …...reaction. The nitronium ion content in the reaction acidic mixture (RAM) is a few tens percent. Thus, in this study, the used RAM