**2.2. Technology of AG production**

The only regular production of AG is realized in the USA where the raw materials are derived from *L. occidentalis* Nutt. [5,6]. The known AG production techniques are all based on extraction of the polysaccharide from larch wood particles (chips, shavings, sawdust) and differ in pre-treatment of the raw material, conditions of extraction and methods of purifying the extracts and the final product [10]. The extraction process is highly dependent on the wood reduction range. Sawdust (specific surface of 164 cm2/g) yields almost 90% of AG during the first 10 min of extraction while chips (specific surface of 17 cm2/g) yield only 10% [68]. To intensify extraction, mechanochemical activation of larch wood and its further treatment with superheated water steam ("autohydrolysis explosion") is proposed [69]. The AG yields increase significantly (more than twice) if larch sawdust extraction is conducted with microwave or pulse ultrasound activation. Microwave or shock-and-acoustic processing of sawdust allows 90% of AG extraction to be completed in 30–60 seconds [70]. The products of larch wood water extraction, together with AG, are various phenolic compounds (low-molecular and oligomeric flavonoids, lignans, lignine substances and tannins), making production of high purity AG a complicated problem [71,72]. Purification of AG water extracts from those impurities is a current issue.

We propose a beneficial, economically and ecologically sound method to obtain dry 95–97% AG [17,73]. There are two main stages: first, extraction and purification of the extract, and second, dry product isolation. Every stage has been carefully studied and theoretically rationalized.

Study of extraction process kinetics is primarily aimed at determining duration of contact between the phases to give the target degree of isolation. Extraction kinetics data determine the geometry of the apparatus. AG water extraction proceeds in two stages, the first of which is fast and the second is slow (Figure 3).

**Figure 3.** AG yield (% of absolutely dry wood, a.d.w.), showing dependence on particle size

The extraction process is determined by AG diffusion and penetration of extracting solution into wood pores as well as by hydrodynamic conditions. Thirty minutes is sufficient to isolate almost the total amount of AG from sawdust, while, in the same conditions, wood chips need several days to yield the major fraction of AG. The quality of AG and other water-soluble components is not affected by wood reduction range.

On the basis of the experimental data, we calculated diffusion constants, mass-transfer coefficients and the Biot diffusion criterion (*Вi*), characterizing the influence of hydrodynamic conditions on AG isolation rate. For the extract (*Yр*), extraction was conducted until the equilibrium of the AG solution concentrations in wood pores reached equilibrium. As , Fourier diffusion criterion *Fog*. Thus, the kinetic equation of the extraction process can be described by the equation [74]:

$$\frac{X - Y^\*}{X\_s - Y\_s} = \sum\_{n=1}^{\infty} A\_n \mathbf{e}^{-\mu\_n^2 Fo\_\mathcal{g}} \tag{1}$$

where *X* is the average AG solution concentration in wood pores, at the moment of time , g/cm3; *Y\** is the equilibrium concentration of the substance isolated in the solution, g/cm3; *Х<sup>s</sup>* is the initial AG solution concentration in the wood pores, g/cm3; and *Ys* - is the initial concentration of the substance isolated in the solution, g/cm3;

$$A\_n = \frac{6Bi^2}{\mu\_n^2 \left(\mu\_n^2 + Bi^2 - Bi\right)}\tag{2}$$

where *<sup>n</sup>* represents characteristic equation roots.

According to substance balance data,

160 The Complex World of Polysaccharides

in veterinary medicine has been proven [64-67].

of AG water extracts from those impurities is a current issue.

which is fast and the second is slow (Figure 3).

**2.2. Technology of AG production** 

rationalized.

additives isolated from Siberian larch, the authors of one study [60] examined the soft wheat flour quality and quantity of gluten, physical properties of the dough, and quality of finished bread, depending on the quantity of the added polysaccharide. The addition of 1% of arabinogalactan to flour causes a significant improvement in the qualitative indices of bread. In this case, AG is totally consumed in the course of bread making because it is utilized by yeast. It is recommended that bread quality can be improved when the flour incorporates 1% mass of AG. When 2–3% of AG is added to flour, the AG content decreases.

The optimum compositions for AG-enriched bakery goods and pastry have been proposed [61,62]. It has been shown that when AG is added to flour in proportions of 1–5% by weight, bakery products are rich in dietary fibre of prebiotic and immunostimulating action, while their energy density is lower due to the decreased amount of sugar in the recipes. The AGenriched bakery and pastry products have a medical effect [59]. Production technology for AG-enriched prebiotic cultured milk products has been developed [63]. The efficiency of AG

The only regular production of AG is realized in the USA where the raw materials are derived from *L. occidentalis* Nutt. [5,6]. The known AG production techniques are all based on extraction of the polysaccharide from larch wood particles (chips, shavings, sawdust) and differ in pre-treatment of the raw material, conditions of extraction and methods of purifying the extracts and the final product [10]. The extraction process is highly dependent on the wood reduction range. Sawdust (specific surface of 164 cm2/g) yields almost 90% of AG during the first 10 min of extraction while chips (specific surface of 17 cm2/g) yield only 10% [68]. To intensify extraction, mechanochemical activation of larch wood and its further treatment with superheated water steam ("autohydrolysis explosion") is proposed [69]. The AG yields increase significantly (more than twice) if larch sawdust extraction is conducted with microwave or pulse ultrasound activation. Microwave or shock-and-acoustic processing of sawdust allows 90% of AG extraction to be completed in 30–60 seconds [70]. The products of larch wood water extraction, together with AG, are various phenolic compounds (low-molecular and oligomeric flavonoids, lignans, lignine substances and tannins), making production of high purity AG a complicated problem [71,72]. Purification

We propose a beneficial, economically and ecologically sound method to obtain dry 95–97% AG [17,73]. There are two main stages: first, extraction and purification of the extract, and second, dry product isolation. Every stage has been carefully studied and theoretically

Study of extraction process kinetics is primarily aimed at determining duration of contact between the phases to give the target degree of isolation. Extraction kinetics data determine the geometry of the apparatus. AG water extraction proceeds in two stages, the first of

An excess of AG inhibits yeast growth, which leads to a decrease in bread quality.

$$Y^\* - Y\_i = b \begin{pmatrix} X \ -Y\_p \end{pmatrix} \tag{3}$$

where *Yi* is the concentration of the substance isolated in the solution at the moment of time , g/cm3; and

$$b = \frac{G\varepsilon}{\rho V} \tag{4}$$

where *G* is the mass of the solid, g; *V* is the volume of the liquid phase, cm3; is the density of the solid, g/cm3; is the specific volume of wood pores occupied by solution, cm3/cm3.

The right side of the equation (3) determines the fraction of the substance transferred from the solid into the solution during the time interval from the moment under consideration until the end of the experiment. The left side determines the increase of the solution concentration during the interval mentioned. Substituting (*X – Y\**) into equation (1) according to (2) leads to:

$$\frac{Y^\*-Y\_i}{X\_s-Y\_s} = \sum\_{n=1}^{\infty} B\_n \mathbf{e}^{-\mu\_n^2 Fo\_g} \tag{5}$$

where Bn = bAn.

For a regular mode of extraction, the first member of the series in equation (5) is sufficient.

Figure 4 graphs the dependence of <sup>i</sup> s s Y\* - Y ln Х -Y on according to equation (5), on the basis of experimental data.

Extrapolating the straight line (Figure 4) i s s Y\* - Y ( ) ln <sup>Х</sup> -Y *<sup>f</sup>* to = 0 gives a diffusion constant, according to the equations:

$$\text{tg}(\alpha) = \cdot \mu\_1^2 \frac{\text{D } \tau}{l^2} \tag{6}$$

and

$$
tg \mu\_1 = \frac{\mu\_1}{b} \tag{7}$$

where *α* is the angle of inclination of the straight line to the time axis.

Solving the characteristic equation (6) regarding 1 gives, according to equation (7), the diffusion constant *D*. On the basis of the experimental data, the diffusion constant at temperatures of 20–25 С is 1.55–2.67·10-10 m2/s. For particles of average size of 5 mm, *Вi>>1*, which demonstrates the insignificant affect of hydrodynamic conditions upon AG isolation rate.

These data on mass transfer were used to calculate the AG extraction process from larch wood particles in similar hydrodynamic conditions. The value of the diffusion constant *D*  was used to calculate the processes conducted at a given temperature.

, g/cm3; and

of the solid, g/cm3;

according to (2) leads to:

where Bn = bAn.

experimental data.

and

rate.

according to the equations:

Figure 4 graphs the dependence of <sup>i</sup>

Extrapolating the straight line (Figure 4) i

1

where *α* is the angle of inclination of the straight line to the time axis.

was used to calculate the processes conducted at a given temperature.

where *Yi* is the concentration of the substance isolated in the solution at the moment of time

*<sup>G</sup> <sup>b</sup> V* 

where *G* is the mass of the solid, g; *V* is the volume of the liquid phase, cm3;

The right side of the equation (3) determines the fraction of the substance transferred from the solid into the solution during the time interval from the moment under consideration until the end of the experiment. The left side determines the increase of the solution concentration during the interval mentioned. Substituting (*X – Y\**) into equation (1)

is the specific volume of wood pores occupied by solution, cm3/cm3.

2

s s

Y\* - Y ( ) ln <sup>Х</sup> -Y *<sup>f</sup>*

2 1 2

n 1 \* <sup>e</sup> *n g Fo <sup>i</sup>*

For a regular mode of extraction, the first member of the series in equation (5) is sufficient.

*s s n Y Y <sup>B</sup> X Y*

s s Y\* - Y ln Х -Y

<sup>D</sup> tg( ) - *<sup>l</sup>*

<sup>1</sup> *tg <sup>b</sup>* 

Solving the characteristic equation (6) regarding 1 gives, according to equation (7), the diffusion constant *D*. On the basis of the experimental data, the diffusion constant at temperatures of 20–25 С is 1.55–2.67·10-10 m2/s. For particles of average size of 5 mm, *Вi>>1*, which demonstrates the insignificant affect of hydrodynamic conditions upon AG isolation

These data on mass transfer were used to calculate the AG extraction process from larch wood particles in similar hydrodynamic conditions. The value of the diffusion constant *D* 

 

(4)

(5)

on according to equation (5), on the basis of

to = 0 gives a diffusion constant,

(6)

(7)

is the density

**Figure 4.** Graph of regular mode of AG extraction. Particle sizes, mm: 1 - 0.056; 2 - 0.86; 3 - 2.61; 4 - 3.24; 5 - 5.37

Based on the experimental data, a mathematical model was developed, material balance was calculated, and the optimal parameters of the extraction process were determined [75]. The model was used to optimize the technological process and production algorithm.

According to the method developed, firstly DHQ and other phenolic extractive substances were isolated from larch wood particles by organic solvent, and then were exposed to extraction by circulating water at 60–80 ºС for 2–3 h. The extract obtained was treated with a cationic flocculant solution to remove mechanical and colloidal impurities. Decolourized extract was ready for being concentrated and further purified by ultrafiltration.

Concentrations of AG water extracts were decolourized by flocculation, which was achieved by ultrafiltration using UAM-150P cellulose acetate membranes (Russia) [76]. Filtration rate decreased with time due to an increase in solution concentrations, viscosities and sedimentation of high-molecular particles on the membrane surface. Initial productivity decreased as initial AG solution concentration increased. However, initial productivity increased with increasing pressure gradient. At the final stage of ultrafiltration, when the process rate approaches a constant level, the higher the pressure gradient the lower the productivity, due to the higher rate of concentration at high pressure during equal time intervals.

Studies have been made of the influence of pressure upon the ultrafiltration process. The maximum degree of concentration is reached at the pressure gradient Р = 0.4 МPа. However, the optimal ratio between productivity and degree of concentration is at Р = 0.2 МPа.

Ultrafiltration results in simultaneous concentration of AG extracts and their purification by almost entirely filtering out low-molecular phenolic impurities. Purification efficiency depends on composition of the extract, membrane characteristics and conditions of filtration as well as a degree of concentration.

According to the IR spectroscopy and HPLC data, filtrate contains, together with phenolic substances, an oligomeric fraction of AG. Atomic absorption and X-ray fluorescence analyses have shown that ultrafiltration purifies AG from metal cations [76]. The total content of dry substances in filtrates is not more than 1–2.5%.

To increase productivity of the ultrafiltration module, we also tested the UAM-500P membrane (Russia). Ultrafiltration dynamics of decolourized AG extracts have shown that filtration rate is in an inverse ratio to initial extract concentration. The use of a macroporous membrane allows ultrafiltration without pre-treatment of AG extracts by a flocculating agent. It has been proven experimentally that productivity of this process is comparable to that of extract decolourized by flocculation. Thus, for the UAM-500P membrane, pore blocking at the initial stage is not a limiting factor, unlike the case for UAM-150P. The optimal conditions of ultrafiltration have been determined to make the technology profitable.

After ultrafiltration, the concentrate was dried in a drying unit. The known methods of dry product isolation, by precipitation in alcohol or acetone [6], are disadvantageous for industrial use from a technological, economical and ecological point of view. The filtrate, without additional treatment, was mixed with fresh water and reused for DHQ extraction.

The method proposed, as compared to known methods, enables the following improvements:


Additional AG purification from high-molecular phenolic impurities was realized by treatment of the water extracts with ecologically harmless oxidant (hydrogen peroxide) [17]. The optimal conditions were found to oxidize impurities without affecting the polysaccharide macromolecule.

For final product isolation from the concentrate, spray drying, lyophilic drying or fluid bed drying can be used.

The experiments showed that spray drying is technically and economically optimal. In a manufacturing pilot, different modes of AG spray drying were tested by varying the starting concentration of AG solution, air temperature at the drier input, air temperature at the drier output and pressure of compressed air at spraying. The temperature of the drying gas (air) was the most technologically relevant. The required humidity of the final product (less than 7%) was reached at an air temperature higher than 100 ºС. The optimal process conditions produced AG of high quality.

On the basis of our study, a technological scheme for isolating high purity AG was developed, involving:

1. Flocculation of AG solution (**FR**)

164 The Complex World of Polysaccharides

profitable.

improvements:

involved

decreased.

drying can be used.

polysaccharide macromolecule.

conditions produced AG of high quality.

analyses have shown that ultrafiltration purifies AG from metal cations [76]. The total

To increase productivity of the ultrafiltration module, we also tested the UAM-500P membrane (Russia). Ultrafiltration dynamics of decolourized AG extracts have shown that filtration rate is in an inverse ratio to initial extract concentration. The use of a macroporous membrane allows ultrafiltration without pre-treatment of AG extracts by a flocculating agent. It has been proven experimentally that productivity of this process is comparable to that of extract decolourized by flocculation. Thus, for the UAM-500P membrane, pore blocking at the initial stage is not a limiting factor, unlike the case for UAM-150P. The optimal conditions of ultrafiltration have been determined to make the technology

After ultrafiltration, the concentrate was dried in a drying unit. The known methods of dry product isolation, by precipitation in alcohol or acetone [6], are disadvantageous for industrial use from a technological, economical and ecological point of view. The filtrate, without additional treatment, was mixed with fresh water and reused for DHQ extraction.

The method proposed, as compared to known methods, enables the following

AG extraction from larch wood is realized after the isolation of DHQ and resin

no expensive sorbents are needed and no toxic or combustible organic solvents are

the closed water cycle allows water consumption and the amount of waste water to be

Additional AG purification from high-molecular phenolic impurities was realized by treatment of the water extracts with ecologically harmless oxidant (hydrogen peroxide) [17]. The optimal conditions were found to oxidize impurities without affecting the

For final product isolation from the concentrate, spray drying, lyophilic drying or fluid bed

The experiments showed that spray drying is technically and economically optimal. In a manufacturing pilot, different modes of AG spray drying were tested by varying the starting concentration of AG solution, air temperature at the drier input, air temperature at the drier output and pressure of compressed air at spraying. The temperature of the drying gas (air) was the most technologically relevant. The required humidity of the final product (less than 7%) was reached at an air temperature higher than 100 ºС. The optimal process

content of dry substances in filtrates is not more than 1–2.5%.

substances, giving a rather high purity of the extract the process is simple, energy-efficient and economically viable

concentrates of dry substance content of up to 40% can be produced


The manufacturing pilot revealed drawbacks to the proposed scheme: low productivity and high time consumption per product unit. Thus, the scheme was optimized.

The most reasonable sequence of steps was determined using a decision tree [77-79] with limited operation sequence combinations: spray drying was always the last step and, thus, was excluded from the decision tree; microfiltration and oxidation always followed flocculation. The decision tree, taking into account the limitations mentioned, is shown in Figure 5.

**Figure 5.** Decision tree

The variants were ranged from 1 to 7 according to these criteria:


The optimal variant is UF-FR-OR-MF (see Table 1), which has a four times higher productivity and a two times lower investment in comparison with the initial scheme.

An optimized technological scheme of arabinogalactan production was implemented at an experimental–industrial scale.


**Table 1.** Criteria of variant estimation
