Prof. P. M. Silin.

Saccharification of Starch in Syrup and Glucose Production by Boiling with Sulfuric Acid at Atmospheric Pressure

At the present time, with the production of the beet-sugar industry considerably reduced, the development of starch-syrup and glucose production is an urgent task. The equipment for these industries is not complicated; the crop feeding them, the potato, is also very simple, while for the benefit of agriculture it is of great importance. All this is especially important for Siberia, where the need for sugar is greater than in European Russia because of the remoteness of the beet-sugar regions, and where agriculture, exhausted by extensive methods of cultivation and by monotonous grain crops, requires the introduction of row crops into the crop rotation¹.

The usual method in Russia for the saccharification of starch is boiling under atmospheric pressure (i.e., at 100° C) with sulfuric acid. Only in the production of glucose at large plants were closed converters and heating under a pressure of 2–3 atm used. The small plants newly arising in the new regions work exclusively in open vats under atmospheric pressure (even in the production of glucose). Given the present high cost of metals and plant equipment, this is quite understandable and inevitable. The aim of the present work is to assist such plants. To elucidate experimentally and theoretically the process of starch saccharification when boiled with sulfuric acid under atmospheric pressure. The work is carried out at the plant, usually by a master according to some tested recipe, which is sometimes regarded as a “secret.” Such secrets are now insufficient, since what is required is the ability to orient oneself more broadly and to produce products of the most varied qualities—from glucose to syrups of low degrees of saccharification. Without guiding data, it is difficult to accomplish such a task, despite the simplicity of the production.

The literature on starch saccharification is quite rich, but works of a purely theoretical physicochemical character predominate²; the conditions under which they were set up differ too much from factory working conditions (for example, starch concentrations do not exceed 10%, whereas in plants one works with 20–35% starch). A considerable part of the work is devoted to determining the influence of elevated temperature and pressure. Still, it would seem possible to use Allihn’s old data: he gives saccharification curves also for a temperature of 100° C; but an attempt to apply these data proves unsuccessful. For example, to the question of how much time is required for saccharification at a starch concentration of 35% with 1¼% \(H_2SO_4\) by weight of starch, so that 50% of the starch is saccharified, Allihn’s diagram gives the answer—after 7–8 hours. In actual fact, at the plant this is accomplished in approximately 3 hours (it may be that this discrepancy is explained by the fact that Allihn did not boil the contents

the starch mass, but merely kept it in a boiling water bath; moreover, he did not carry out the preliminary conversion into soluble starch with vigorous stirring).

Among works that, in their formulation, approach factory conditions, two should be noted—those of Lippmann and Parow¹. Both are devoted to elucidating the influence of elevated temperature and pressure; for work at \(100^\circ\) they provide nothing.

A review of the literature leads to the conviction that, in fact, experimental works are entirely lacking that could serve as guides in conducting the process of saccharification under conditions customary for Russian starch-and-syrup factories.

As is known, the principal processes in the action of \(\mathrm{H_2SO_4}\) on starch are two: 1) gradual saccharification—conversion into glucose (passing through intermediate products—dextrins and maltose), and 2) the process of conversion of glucose into some more complex carbohydrate (or carbohydrates—“gallizin,” “glucosin,” “isomaltose”)—“reversion.”

The products of reversion possess a lower reducing power than glucose. An idea of the degree of saccharification of starch may be formed by determining the accumulated glucose by means of Fehling’s solution and calculating what percentage of the starch taken has been converted into glucose (or into some intermediate products, or products of “reversion,” which also possess, though reduced, reducing power). For technical purposes this method, adopted in the present work, is sufficient. In analyzing the results obtained, it is necessary to take into account the process of “reversion,” which is accelerated especially at high concentrations in the later stages of the experiment, when a sufficient quantity of glucose has accumulated.

Besides “reversion,” another change in glucose is of great importance for technology, connected with the formation of colored substances (“humic” substances)—with an increase in color. It was also proposed to trace the course of this process, and, since the reaction products have not been investigated and are chemically indeterminate, the colorimetric method was chosen—the increase in color was measured.

At the same time, the color reactions with a solution of \(J\) in \(KJ\) and the reaction with alcohol (precipitation of dextrins), which are ordinarily used, were also carried out.

The experimental conditions were chosen to approximate factory conditions. The concentration (absolute dry) of starch was 20%, 30%, and 35% (by weight), and for each of these concentrations three concentrations of \(\mathrm{H_2SO_4}\) were taken—1%, 2%, and 3% by weight of absolutely dry starch². Thus a total of 9 experiments were set up.

Apparatus and conduct of the experiment. The experiments were carried out with commercial potato starch, which had been previously purified: mixed with water, strained through a fine sieve, washed three times with water (by decantation), and dried in air (near a steam-heating radiator—12 hours).

Analysis of the starch gave the following results:

Starch—84.6%
Water—0.6%
N-substances—0.6%.

The sulfuric acid used for saccharification was chemically pure (a solution of exactly determined titer was prepared = \(0.0554\ \dfrac{\mathrm{gr}}{\mathrm{cm.^3}}\), and for each experiment a previously calculated amount was measured from a burette).

The saccharification process was carried out in a flask of 700 cubic cm capacity with a reflux condenser. The condenser tube was lengthened and bent, as shown in the figure, and a rubber tube was fitted onto it. In addition to the condenser, another bent tube was inserted into the stopper, reaching almost to the bottom (shaped as in an ordinary wash bottle); it served for taking samples: in order to take a sample, it was only necessary, by forcing air (blowing by mouth) through the rubber tube fitted onto the condenser, to expel the liquid as from a wash bottle. The first drops were discarded (they rinsed the outlet tube).

The preliminary stage of converting starch into soluble starch, at the high concentrations adopted, could not be carried out in the flask. This was done in a porcelain saucepan: in it, a portion of the water required for the experiment, together with half of the total amount of acid required by calculation for the experiment, was brought to boiling (on an asbestos screen), and then, with stirring, starch milk from the weighed quantity of starch was gradually poured in; the remaining half of the sulfuric acid was added to this starch milk. The pouring-in continued for approximately 40 min; then the weight of the entire mass was brought with hot water to 500 g on a Roberval balance (the saucepan having first been tared), the entire mass was immediately transferred to the flask and brought to gentle boiling on an asbestos screen with the reflux condenser attached. The time count began from the moment the flask boiled. Boiling was not interrupted for a single minute. After 1 hour the first sample was taken. The following samples were taken at intervals of 2 hours (i.e., after 3, 5, 7, and 9 hours from the beginning of the experiment). The experiment lasted 9 hours.

The period of pouring in the starch milk was thus not counted at all. The error resulting from this is not great, since the pouring-in period took place at a lowered temperature (on average about 85—90°); consequently, 40 min of pouring-in could be equivalent to no more than 10 min of boiling (for each 10° C the rate of starch saccharification changes approximately 4-fold); moreover, these 10 min were spent on the first stage—liquefaction—the conversion of starch into soluble starch, and therefore they should not be included in the time of saccharification proper¹. For all experiments this initial period was approximately the same. The condition of comparability was, therefore, observed.

As can be seen, the conditions under which the experiments were set up corresponded in general to the methods of factory operation.

Examination of samples. The sample taken (in a large, wide test tube, tightly closed with a stopper) was immediately cooled under a stream of water (for 1—2 min), and 5 cubic cm were at once taken with a pipette into a volumetric flask (100, 250, or 500 cubic cm) and diluted with water to the mark—for the determination of glucose. With such rapid work it was possible to dispense with neutralization of the acid and not fear its further action. The remaining part of the sample (except for 5 cubic cm) was used for the reaction with I, for the test with alcohol, and for the determination of color.

Glucose was determined by Fehling’s solution by the volumetric method according to Bertrand². For each experiment, the percentage of the starch taken, by weight of the entire mass, was also determined (by inversion with HCl of a separate 5 cubic cm sample, which was taken in the 1st or 3rd hour, when processes of glucose change could not yet be noticeable). The weight of liquid (sample) flowing out of the pipette in 5 cubic cm and the specific gravity of the liquid (pycnometrically) were also determined separately for each experiment. All these data made possible the calculation of the percentage of starch saccharification, the volumetric percentages (g per 100 cm³), and the weight percentages of starch, glucose, and sulfuric acid.

To determine color, the sample was filtered (folded filter, diameter 7 cm). The first 5–6 cu. cm of filtrate were discarded (effect of filter moisture and adsorption). There was concern that the filter paper would nevertheless adsorb the colored substance and reduce the color intensity, but a control determination of the color after a second (repeated) passage through a fresh filter showed that adsorption in this case was quite imperceptible. The color intensity was determined with a Duboscq colorimeter by comparison with a solution of J in KJ.

A solution of J with a concentration of about \(25 \cdot 10^{-4}\ \mathrm{N}\) was prepared. Before each series of color determinations this solution was titrated with sodium thiosulfate solution. This J solution, of thus precisely determined concentration, was, however, too darkly colored, and for determining its color it had to be diluted 10 times with distilled water (10 cu. cm to a 100-cu. cm volumetric flask). The titer of the J solution changes rather quickly. Therefore such a method of work (with preliminary titration each time) is necessary. Other methods for determining color were also tried, but they proved less convenient. For example, Bismarck brown solutions must first be compared in color with some solution adopted as the standard, for instance with the same iodine; moreover, the color stability of Bismarck brown solutions is also low. Solutions of \(K_2Cr_2O_7\) are less suitable in hue than J solutions. Thus, for color comparison there was always used a diluted J solution with a concentration of about \(2.5 \cdot 10^{-4}\ \mathrm{N}\) (the fluctuations were from 2.9 to \(2.3 \cdot 10^{-4}\ \mathrm{N}\)).

The qualitative test with J for starch and the higher dextrins was made as usual: to 1 cu. cm of the cooled sample were added 5–6 drops of diluted J solution in KJ.

The alcohol test (for precipitation of dextrins) was carried out as follows: a sample was taken with a small pipette (capacity \(0.91\ \mathrm{cm^3}\)) into a test tube, and alcohol (93.2 vol. %) was added from a burette dropwise, with shaking, until a dextrin precipitate appeared. The amount of alcohol added was recorded.

Course of the saccharification process. The results of the experiments elucidating the course of saccharification are grouped in Table 1 and presented graphically in diagram 1.

Table 1.

Process of starch saccharification. (Experimental data.)

The weight percent of starch (column 1) did not come out as exactly 20, 30, and 35%, as would have been desired, but only approximately these values. This resulted from the impossibility of accurately taking the total weight of the mass (500 g) when adding water on Roberval balances. It was necessary deliberately to add a small excess of water to compensate for loss by evaporation during transfer from the saucepan to the flask. The percent of sulfuric acid, by weight of starch (column 2), was measured accurately. The acidity of the mass was checked by titration, and it was found that the titration result depends strongly on the indicator used. When titrating with phenolphthalein, results were obtained that confirm the correctness of the ratios taken, i.e. the same figures as in column 2 were obtained. If titration is carried out with methyl orange, a smaller amount of acid is obtained by 0.1–0.2% (column 3). This can be explained by the presence in the starch of salts of phosphoric and weak organic acids, which are displaced by the added sulfuric acid and then, as weak acids, cannot be titrated to completion with methyl orange. Of course, neither titration gives an idea of the concentration of H ions (“active” acid) in the solution; and a difference in titration results of 0.1–0.2% cannot serve as a measure of the amount of weak acids.

Heading 6 (5 columns) gives the degree of saccharification, i.e. the % of starch (by weight of all the starch) converted into glucose (and, in general, into reducing carbohydrates) at one or another moment in time.

These latter data (degree of saccharification over time) are represented graphically in Diagram 1. In this way, saccharification curves were obtained. Along the abscissa axis was plotted the duration of saccharification in hours, and along the ordinate axis—the % of saccharified starch. Each experiment corresponds to its own curve (9 in all).

At first glance at Diagram 1 it is apparent that the maximum saccharification for high concentrations of starch and acid is lower than for low concentrations of starch. For example, at 3% H₂SO₄, for 20% starch the maximum saccharification is 91.9%, whereas for 30% starch it is 89.0%, and for 35% starch only 85.3%. Consequently, despite the rapid course of the saccharification process, much glucose cannot accumulate, since, apparently, the process of its alteration begins to proceed even more rapidly. At high starch concentrations, an increase in the amount of H₂SO₄ likewise changes almost nothing in the attainable maximum of saccharification (cf. Exp. IV with V and VII with VIII), since increasing the concentration of H₂SO₄, while accelerating the saccharification process, simultaneously accelerates the process of alteration of glucose. This process, at high concentrations of starch and H₂SO₄, appears sharply: toward the end of the experiment the saccharification curves begin to bend downward (see the curves of Exp. VII—for 35% starch and 3% H₂SO₄, Exp. VIII—for 35% starch and 2% H₂SO₄, Exp. IV—for 30% starch and 3% H₂SO₄); the amount of reducing sugars does not increase, but even begins to decrease.

An attempt was made to determine the amount of glucose which, by the end of the experiment, had undergone an “irreversible” change: had been converted into substances which, by hydrolysis with acid, can no longer be converted back into glucose. For this purpose, complete inversion with HCl was carried out in a dilute solution of samples taken in the 1st and in the 9th hour of the experiment. It turned out that, within the errors of the determinations, in both cases the amount of glucose was the same. Thus, one cannot speak of decomposition, or in general of an irreversible alteration of glucose (it probably occurs, but to an insignificant degree); the principal process is “reversion”—conversion into some carbohydrate (or carbohydrates) consisting of glucose molecules, whereby the number of carbonyl groups decreases (lower reducing power). This carbohydrate (“glucosine,” “gallisine,” etc.) is capable of being hydrolyzed again in more dilute solutions with a larger amount of acid.

It is very interesting to determine how far the saccharification curves (Diagram 1) agree with one another and whether, knowing the form of the curve ...

for one set of conditions, using the laws of the kinetics of chemical reactions, to determine its form for altered conditions. For low concentrations of starch and sulfuric acid (and also for high ones, but only in the first hours of saccharification, when the reversion process is still insignificant), the reaction should proceed according to the laws of monomolecular reactions, and the rate of reaction should be proportional to the concentration of H₂SO₄. This can serve as a basis for comparing the course of the curves. For example, let us take 2 curves (Exp. IV and VI) for one and the same concentration of starch (30%), but for different amounts of acid—3% and 1%. Theoretically one might expect that the process with 1% H₂SO₄ would proceed exactly as with 3%, only 3 times more slowly; i.e., the degree of saccharification which is attained with 1% H₂SO₄ after 9 hours would be obtained with 3% H₂SO₄ already after 3 hours. In reality, however, with 1% H₂SO₄ after 9 h. 68.2% is saccharified, while with 3% H₂SO₄ after 3 h.—78.3%—much more than might be expected. Thus, with small amounts of H₂SO₄, a disproportionately too small saccharification is obtained, and with large amounts—a disproportionately too large one. The 9 h. of saccharification with 1% H₂SO₄ correspond, it turns out, not to 3 hours for 3% H₂SO₄, but to 2 h. 15 m., when, in fact, saccharification of 68.2% is obtained (see diagram 1). This phenomenon is general: it is observed when comparing any two curves of the diagram. A possible cause of such a phenomenon is the presence in the starch of salts of weak mineral (K₂HPO₄) and organic acids. This is also indicated by the different results obtained in titration with methyl orange and with phenolphthalein as indicators (see above). Part of the H′ ions from H₂SO₄ forms, with the anions of weak acids, molecules of these acids. Thus a certain amount of active H′ disappears and the saccharifying action of H₂SO₄ is reduced. The kinetic data (curves of diagram 1) make it possible approximately to determine the amount of this, as it were, disappeared H₂SO₄.

Let the starch contain so many salts of weak acids that they can bind (consume for the formation of molecules of weak acids) \(x\%\) H₂SO₄ by weight of starch. Thus, if we take 3% H₂SO₄ (also by weight of starch), then only \((3 - x)\%\) acid will act saccharifyingly. With 1% H₂SO₄—only \((1 - x)\%\) of it is active¹). But when we subtract \(x\%\) from 3%, the relative decrease is not as great as when subtracting \(x\%\) from 1%; therefore it is understandable that at low % H₂SO₄ the saccharifying action proves to be disproportionately small. The magnitude \(x\) can easily be determined. Thus, for the example examined, where 9 h. at 1% H₂SO₄ proved equivalent to 2 h. 15 m. at 3%, we may write the equation:

\[ \frac{3 - x}{1 - x} = \frac{9\ \text{hours}}{2\ \text{h.}\ 15\ \text{m.}} \quad \text{or} \quad \frac{3 - x}{1 - x} = \frac{540}{135}. \]

This equation is set up on the basis that the duration of saccharification must be inversely proportional to the amount of active H₂SO₄. From the equation we find \(x = 0.33\). Comparing the other curves with one another for the most varied values of time, we find in an analogous way a series of values of \(x\). They do not go beyond the limits 0.3–0.5, which may be regarded as satisfactory agreement. The mean value is \(x = 0.4\%\).

Thus, applying one or another % H₂SO₄ for the saccharification of starch, we have approximately always 0.4% by weight of starch less of active, acting acid; this loss is felt especially sensitively when small amounts of H₂SO₄ are taken for saccharification²).

¹) This, of course, is not entirely correct, since \(x\) must depend somewhat on the total concentration of H′, with an increase of which the degree of dissociation of weak acids decreases and, consequently, the undissociated part, \(x\), increases.

²) One might think that the magnitude \(x\) varies strongly depending on the kind of starch. Experiments were made with another starch (from another year’s harvest). The value of \(x\) determined from these experiments was also close to 0.4%.

In column 7 of Table 1 are given the amounts of active acid for each experiment, in % by weight of starch (from the figures in column 2, 0.4 has been subtracted); and in column 8 are the volume % (in g per 100 cm³ of solution) of active H₂SO₄ (the figures of column 7 are multiplied by the figures of column 1 and by the specific gravity given in column 4, and the result divided by 100).

The agreement of the curves for different concentrations of starch can be checked on the basis of the concentrations of “active” acid (column 8).

For example, 9 h for 30% starch and 1% H₂SO₄, i.e. for 0.205 vol. % active H₂SO₄ (see the line for experiment VI), corresponds for 20% starch and 3% H₂SO₄, i.e. for 0.567 vol. % active H₂SO₄ (see experiment 1), to

\[ \frac{9 \cdot 0.205}{0.567}=3\ \text{h}\ 15\ \text{min}; \]

the curves of diagram 1 give 3 h 05 min, i.e. almost the same. For 35% starch and 2% H₂SO₄, or 0.645 vol. % active H₂SO₄ (experiment VIII), these same 9 h (experiment VI) correspond to

\[ \frac{9 \cdot 0.205}{0.645}=2\ \text{h}\ 52\ \text{min}. \]

According to the diagram, however, it is 2 h 48 min—also satisfactory agreement.

The law of the saccharification reaction, as a monomolecular one, may be expressed by the usual equation:

\[ \frac{dZ}{dt}=K \cdot C(100-Z) \]

Here \(Z\) is the % saccharification (figures of columns 6); \(t\) is the time in hours; \(K\) is the reaction constant; \(C\) is the volume % of active acid (column 8). In column 9 are given the mean values of the constant \(K\), obtained from each experiment. They do, in fact, fluctuate little. It must be noted that when calculating \(K\) for high concentrations of starch and H₂SO₄, only the first hours of the experiment can be taken, since further on the process of reversion quickly begins to make itself felt and \(K\) decreases. These lowered values were discarded and were not taken into account in calculating \(K\) (thus \(K\) for experiments IV and VII were obtained only from the data for the 1st hour of saccharification and therefore, as less reliable, are placed in parentheses and were not considered in deriving the mean constant from all experiments).

Of course, for high degrees of saccharification and high concentrations of H₂SO₄, no agreement in the sense of the considerations given can be established, since here the process is complicated by the alteration of glucose under the influence of H₂SO₄. Wohl¹) proposed a correction for reversion based on the assumption that the rate of the reversion process is proportional to the concentration of the glucose formed. This correction is expressed by the equation:

\[ \frac{dZ}{dt}=K(100-Z)-K' \cdot Z. \]

An attempt to find the constant \(K'\) in Wohl’s equation proved unsuccessful (the “constant” fluctuates too much). The equation is not applicable under the conditions of our experiments. This is understandable, since this equation is based on incorrect premises. Reversion consists in the formation, from glucose molecules, of some more complex carbohydrate. Consequently, the reversion reaction cannot be monomolecular,²) but is at least (and most probably) bimolecular; consequently,

¹) Wohl. Berichte 1890. 2084.

²) If the reversion reaction were monomolecular, then the form of the saccharification curve would not depend on the concentration of starch. The attainable maximum of saccharification would be the same for any concentrations of starch and acid. All this is a consequence of Wohl’s equation. The data of Table 1 and Diagram 1 show that this is not so. The form of the curve, especially in the final hours of the experiments, when reversion makes itself felt, shows a great dependence on the concentration of starch. This can be the case only if the order of the reversion reaction is higher than first.

the correction term would have to contain \(Z^2\); but this too would be wrong, since the reversion reaction is reversible. Thus, the term \(K'Z\) expresses two erroneous assumptions: monomolecularity and irreversibility of reversion. The first term, \(-K(100-Z)\), also contains an erroneous assumption: the percentage of dextrins remaining at a given moment will not be \((100-Z)\) (i.e., 100% minus the percentage of glucose obtained); in order to obtain the percentage of unchanged dextrins, one must subtract from 100 not only the percentage of glucose, but also the percentage of “gallizin” (“glucosine”), i.e., write \((100-Z-Z')\).

The corrected equation will take such a complicated form that it will prove inconvenient for use.

It is nevertheless interesting to form an idea of the quantity of reversion products accumulating toward the end of the experiment and of the amount of unsaccharified dextrin remaining. Proceeding from the reaction constant for the actual saccharification of starch (0.64) found by us, one can construct a theoretical saccharification curve, for example, for 35% starch and 3% \(\mathrm{H_2SO_4}\) (active acid 1.091% abs.), analogous to the curve of experiment VII¹. Such a curve is given for comparison in diagram I (the uppermost). During the first hour this curve coincides with the experimental curve of experiment VII, but thereafter a divergence begins, reaching 16% by the end of the 9th hour. If the carbohydrate formed as a result of reversion has no reducing power, then 16% will be the amount of this carbohydrate (“gallizin”). If, however, it does reduce, but only more weakly than glucose (as indicated by a number of investigators), its amount must be still greater. It may be asserted that, for determining the amount of dextrins still unsaccharified, the experimental curve gives less correct indications than the theoretical one. For the present example, it turns out that after 5 hours only 3.0% remains, after 7 hours—0.8%, and after 9 hours—0.2% of dextrins.

It is interesting to consider the divergence, the increase in the difference between the course of the theoretical saccharification curve (without reversion) and the experimental one. It is as follows:

The rate of increase in the difference, characterizing the rate of reversion, is, as can be seen, small at first; then, with the increase in the amount of glucose, it increases; and later it rapidly diminishes, despite the almost constant amount of glucose. This can be explained only by the reversibility of the reversion process: this process tends toward a certain equilibrium and therefore begins to slow down. If “reversion” were an irreversible process, the curves of experiments IV, VII, and VIII would not proceed toward the end almost parallel to the abscissa axis, but would be strongly bent downward.

Using the correspondences of the saccharification curves and interpolating for intermediate conditions, Table 2 has been compiled. It gives the degrees of saccharification after intervals of 1 hour for starch concentrations of 20, 25, 30, 35% and for 1.0; 1.5; 2.0; 2.5 and 3.0% sulfuric acid by weight of absolutely dry starch. Tenths of a percent of saccharification have been discarded, since

accuracy of the data in the table is hardly higher than 1%. This table has a purely practical significance. It could serve for quick orientation under factory operating conditions.

Table 2.

Degree (%) of saccharification of starch.

Increase in color. The quantity of colored substances formed during the action of H₂SO₄ on glucose is probably very small. There is no method for isolating them, but it is important to take account of the increase in the quantity of these substances, at least by color, since they lower the quality of the products manufactured at the plant. The method of measuring color has already been described. As the unit in measuring color, the color of an iodine solution (in KJ) of concentration \(\frac{1}{10000}\,N\) is adopted.

The numerical material is grouped in Table 3 and presented graphically in Diagram 2.

Table 3.

Increase in coloration during the saccharification of starch.

In the columns under heading a are given the coloration values obtained from the experiment after 1, 3, 5, 7, and 9 hours. What is characteristic and of interest is only the increase in coloration over time. The coloration value obtained after 1 hour of saccharification is entirely accidental. It occurs in part from colored impurities contained in the starch, and chiefly as a result of scorching of the starch during the period of pouring in the starch milk. Therefore, at higher starch concentrations and smaller quantities of H₂SO₄, when it is more difficult to carry out the process of the initial liquefaction, this “initial” coloration proved greater. It must be noted, however, that in general, in comparison with the subsequent growth of coloration, it is very insignificant.²

In order to eliminate the non-characteristic accidental value of the initial coloration and to make the figures comparable, heading b gives the values of the increase in coloration in comparison with the coloration of the first hour (i.e., for example, for the 7th hour of the 1st experiment, instead of 2.25 we have, after subtracting 0.26—the initial coloration of the 1st experiment—2.25 − 0.26 = 1.99—the increase). Thus we regard the colorations of the 1st hour for all experiments as equal to zero. In Diagram 2 it is precisely these values of the increase in coloration that are plotted (therefore, for the 1st hour, all curves pass through 0).

On reviewing the table and the coloration diagrams, the extremely great diversity and sharp differences between the results of the individual experiments, depending on the conditions under which they were set up, draw attention. Whereas in experiment III, over 9 hours, the coloration increased by only 0.12, in experiment VII during the same time it increased by 23.17, i.e. almost 200 times more. Here the difference is far more significant than the difference in the degree of saccharification (47.7% and 83.0% for the same experiments).

For the coloration curves corresponding to high contents of starch and H₂SO₄, the following form is characteristic: in the first hours—a considerable curvature, corresponding to a transition from a slow to a rapid increase in coloration;

¹ As the unit, the coloration of a solution of J [[unclear: in KI?]] of concentration \(\frac{1}{10000}\,N\) was adopted.

² Only in one unsuccessful experiment, when during the pouring in of the starch milk considerable scorching at the bottom occurred, did the initial coloration reach a value of 2.50.

toward the end of the experiment the increase proceeds at a constant rate, almost along a straight line. Curves of such a form were to be expected. The colored products are apparently obtained only from glucose, not from dextrins. In the first hours there is still little glucose, and therefore the color increases more slowly. As the amount of glucose increases, the process of its alteration also increases. At high starch and acid contents, the amount of glucose quickly reaches a maximum and then almost ceases to change; therefore, corresponding to the almost constant glucose content, there is established a constant, uniform (straight-line) course of the process of increase in color.

The concentration of starch (more precisely, the concentration of the glucose obtained) has a very considerable influence on the increase in color. For example, for experiment VIII and experiment I the concentrations of H₂SO₄ differ little from one another (0.64 and 0.57); meanwhile in experiment VIII the increase in color proved to be 10.48, and in experiment I only 3.06; this occurred because in experiment I we have a considerably less concentrated glucose solution (20% starch) than in experiment VIII (35% starch). It is easy to observe that the rate of increase in color may be proportional to the first, or at least the second, power of the glucose concentration; i.e., here the reaction is also bimolecular (as is the reversion reaction). Then, of course, the rate of increase in color must be proportional to the concentration of the acid (as catalyst). These quantitative relations may be expressed by the equation

\[ \frac{df}{dt}=K' \cdot C \cdot g^{2}. \]

Here \(f\) is color, \(t\) is time (hours), \(K'\) is a constant, \(C\) is the acid concentration (volume %), \(g\) is the glucose concentration (in volume %; it is easily calculated from the data of Tables 1 and 2 by multiplying the values of columns 5 and 6 and dividing by 100).

The results of attempts to determine the constant \(K'\) from the data for the time interval from the 7th to the 9th hour are given in column “c” of Table 3. For \(K'\), values were obtained which, although fluctuating, do so within fairly narrow limits. The fluctuations are explained by the low precision of the color determinations. For the first hours of the experiments these fluctuations in \(K'\) would be still more considerable, since there the increase in color is only slight and errors in its determination are relatively greater. For these same reasons, for the 7th–9th hour the value of \(K'\) from experiments II, III, and VI should be regarded as less reliable (they are placed in parentheses), since the corresponding increase in color is too small (1.33; 0.12; and 0.54). The mean value of the constant \(K'\) from the other experiments is \(K' = 50\).

If one assumes that the increase in color is proportional to the first or third power of the glucose concentration, it is not possible to obtain such an approximately constant value of \(K'\).

Thus, color increases in proportion to the acid concentration and in proportion to the square of the concentration of the accumulated glucose. This also explains why at the beginning it increases so slowly and why it begins to grow so rapidly with the accumulation of glucose as a result of the saccharification process. This also explains why, at greater dilutions, less colored products are obtained: if the starch and acid are diluted, for example, twofold, then, it is true, we slow the saccharification process by a factor of 2 (since the acid is diluted); to obtain the former degree of saccharification it will be necessary to boil twice as long, but the process of increase in color will be slowed by a factor of 8 (owing to the dilution of the acid by a factor of two and owing to the dilution of the starch by \(2^{2}=4\) times); in the end, despite the more prolonged saccharification, the color of the solution will be

\[ \frac{2^{3}}{2}=2^{2} \]

times less, if one takes into account that in the diluted solution

To P. M. Silin’s article “Saccharification of Starch.”

Diagram 1

Course of saccharification

Vertical axis: degree of saccharification (in %)
Horizontal axis: Hours

To P. M. Silin’s article, “Saccharification of Starch”

Diagram 2.

Increase in color intensity

Y-axis: Color intensity (in comparison with an iodine solution, \(I_2\), \(1:100000\))
X-axis: Hours

—less glucose; and, recalculating the color intensity to a more concentrated solution of it, this color will nevertheless be \(\frac{2^n}{2} = 2\) times less than before.

Reaction with J and precipitation with alcohol. The results obtained are given in Table 4. From the table it is seen that the reaction to J ceases, approximately, at the time when saccharification reaches 50% (if compared with Table 1). Alcohol still gives a distinct precipitate with 5–7 volumes of 93° alcohol (by volume) when saccharification has already reached approximately 70%. With further saccharification, precipitation ceases. There is noticeable only a slight turbidity of dextrins when a considerable excess of alcohol is added (8–10 volumes), or a faint cloudy ring when a layer of alcohol is carefully poured over a layer of the solution being tested.

Table 4.

Reactions with J and with alcohol during saccharification of starch.

Conclusions. 1. In its main features, the process occurring under the influence of H₂SO₄ on starch consists of two reactions:

a) a monomolecular reaction of conversion of dextrins into glucose;

b) “reversion”—a bimolecular reaction of formation from glucose of some other, apparently more complex carbohydrate; this reaction is reversible, and therefore can be carried only to a certain limit, depending on the concentration: upon dilution, the carbohydrate obtained as a result of reversion again converts into glucose; at high concentrations and with prolonged action of H₂SO₄, a very considerable quantity of reversion products is formed (for 35% starch—not less than [[unclear: value]] of its weight).

1. The colored substances are formed under the action of H₂SO₄ from glucose according to the type of a bimolecular reaction, i.e., the rate of increase of color intensity is proportional to the square of the glucose concentration. Thus, this process increases extraordinarily rapidly with increasing glucose concentration.

2. Points 1 and 2 of the conclusions show that, in glucose manufacture, it is necessary to keep the concentration low. By lowering the concentration of starch we, it is true, slow down the process of saccharification, but at the same time we slow still more the processes of reversion and of increase in color intensity, since these latter reactions are bimolecular; moreover, upon dilution the limit of the reverse reaction (“reversion” equilibrium) shifts toward obtaining a greater quantity of glucose and a smaller quantity of reversion products.

3. Using Diagram I and Table 2, one may, for the principal conditions of saccharification applied in practice, find the corresponding basis of saccharification.

¹ Designations: b—blue, v—violet, r—red, o—orange, y—yellow.

1. Equation of the curve
   \[ \frac{dz}{dt}=0.64\cdot 1.091\cdot(100-Z); \]
   integrating, we obtain:
   \[ \ln \frac{100}{100-Z}=0.64\cdot 1.091\cdot t. \] ↩↩↩↩

2. Bl. de la Soc. Ch. de Paris XXXV 1285 1906. ↩↩↩