[9] This paper appears during the printing of the author's original MS.

SECTION VI. THE LIGNOCELLULOSES

(p. 131) ~Lignocellulose Esters.~--By a fuller study of the ester reactions of the normal celluloses we have been able to throw some light on the const.i.tutional problems involved; and we have extended the investigations to the jute fibre as a type of the lignocelluloses, from the results of which we get a clearer idea of the relationships of the const.i.tuent groups.

Taking the empirical expression for the complex, i.e. the entire lignocellulose, the formula C_{12}H_{18}O_{9}, we shall be able to compare the ester derivatives with those of the celluloses, which we have also referred to a C_{12} unit. But we shall require also to deal with the const.i.tuent groups of the complex, which for the purposes of this discussion may be regarded as (a) a cellulose of normal characteristics--cellulose alpha; (b) a cellulose yielding furfural on boiling with condensing acids--cellulose beta; and (c) a much condensed, and in part benzenoid, group which we may continue to term the lig_none_ group.

The latter has been specially examined with regard to its proportion of OH groups, as a necessary preliminary to the investigation of esters, in producing which the entire complex is employed. It will be shown that the ester groups can be actually localised in various ways, as in the main entering the cellulose residues alpha and beta. But that the lignone group takes little part in the reactions may be generally concluded on the evidence of its non-reactivity as an isolated derivative, (1) By chlorination, &c. it is isolated in the form of an amorphous body, but of constant composition, represented by the formula C_{19}H_{18}Cl_{4}O_{9}. This compound, soluble in acetic anhydride, was boiled with it for six hours after adding fused sodium acetate, and the product separated by pouring into water. The dilute acid filtered from the product contained no hydrochloric acid nor by-products of action.

The product showed an increase of weight of 7.5 p.ct. For one acetyl per 1 mol. C_{19}H_{18}Cl_{4}O the calculated increase is 8.0 p.ct. It is evident from the nature of the derivative that this result cannot be further verified by the usual a.n.a.lytical methods. (2) The chlorinated derivative is entirely soluble in sodium sulphite solution. This solution, shaken with benzoyl chloride, with addition of sodium hydrate in successive portions, shows only a small formation of insoluble benzoate, which separates as a tarry precipitate. (3) The empirical formula of the lignone complex in its isolated forms indicates that very little hydrolysis occurs in the processes of isolation. Thus the chlorinated product we may a.s.sume to be derived from the complex C_{19}H_{22}O_{9}. In the soluble by-products from the bisulphite processes of pulping wood the lignone exists as a sulphonated derivative, C_{24}H_{23}(OCH_{3})_{2}.(SO_{3}H).O_{7}. The original lignone may be regarded as pa.s.sing into solution as a still condensed complex derived from C_{24}H_{26}O_{12} (Tollens). There is evidently little attendant hydroxylation, and another essential feature is the small molecular proportion of groups showing the typical sulphonation.

It appears that in the lignone the elements are approximately in the relation C_{6} : H_{6} : O_{3}, and it may a.s.sist this discussion to formulate the main const.i.tutional types consistent with this ratio, viz.:

(1) The trihydroxybenzenes C_{6}H_{3}(OH)_{3}.

(2) Methylhydroxyfurfural C_{5}H_{2}O.(OH)(CH_{3}).

(CH_{3}) / (3) Methylhydroxypyrone CO

< p="">

__________________

(4) Trioxycyclohexane CH--CH--CH--CH--CH--CH / / / O O O

It is probable that all these types of condensation are represented in the lignone molecules, since the derivatives yielded in decompositions of more or less regulated character are either directly derived from or related to such groups. For the moment we pa.s.s over all but the general fact of complexity and the marked paucity of OH-groups. It would be of importance to be able to formulate the exact mode of union of the lignone with the cellulose residues to const.i.tute the lignocellulose.

The evidence, however, does not carry us farther than the probability of union by complicated groups and of large dimensions; for not only is the lignone isolated in condensed and non-hydroxylated forms, but the cellulose also is not hydrated or hydrolysed further than in the ratio 3C_{6}H_{10}O_{5}.H_{2}O. It is probable, therefore, that the water combining with the residues at the moment of their resolution is relatively small.

Lastly, we have to remember, when dealing with the statistical results of the reactions to be described, that the approximate proportions per cent. of the const.i.tuent groups are:

Cellulose alpha 65 " beta 15 = 100 lignocellulose.

Lignone 20

~Jute Benzoates.~--In preparing the jute for treatment it was boiled in alkaline solution (1 per cent. NaOH), washed with water and dilute acid, again washed, dried, and weighed.

In the ester reaction the reagents were employed in the proportion C_{12}H_{18}O_{9} : 3NaOH : 2C_{6}H_{5}COCl. A series of quant.i.tative experiments gave yields of 126-130 p.ct. of benzoate [calculated for mon.o.benzoate 134 p.ct.].

The results were confirmed by ultimate a.n.a.lysis. The mon.o.benzoate therefore represents a maximum, and this molecular proportion is one-half of that observed with the normal cellulose, calculated to the same unit.

_Localisation of Benzoyl Group._--The entrance of the ester group affects the typical colour reactions of the lignocellulose, which are fainter. The ferric ferricyanide reaction almost disappears. The lignone group is unaffected, and combines with chlorine as in the original. The lignone chloride is removed by sodium sulphite solution, and the residue is a _cellulose benzoate_. The loss of weight due to the elimination of the lignone was 12.7 p.ct. Calculating per 100 of the original lignocellulose this becomes 16. These statistics further confirm the localisation of the benzoyl group in the cellulose residue. It is to be noted that the presence of the benzoyl group renders the cellulose more resistant to hydrolytic actions. Thus, to bring out this fact more prominently, we may calculate the yield of residual cellulose benzoate p.ct. of original jute, and we find it 109 p.ct. Taking a maximum proportion for original cellulose--viz. 85--this benzoate represents a yield of 129 p.ct., as against the theoretical for a mon.o.benzoate, 132 p.ct.

_Furfural Numbers._--The percentage of furfural obtained by boiling with HCl of 1.06 sp.gr. was 3.02 and 3.29 in separate determinations.

Calculating to the original lignocellulose, the percentage, 4.21, indicates a considerable loss of the furfural-yielding const.i.tuent. The effect was also apparent in the cellulose (benzoate) isolated by chlorination &c., the percentage being 1.39 p.ct., and calculated to the original jute benzoate 1.59 p.ct. Under the conditions adopted in dissolving away the chlorinated lignone the original non-benzoated lignocellulose would have yielded a cellulose giving 6 to 7 p.ct.

furfural.

Since the benzoyl group is hardly calculated to produce a const.i.tutional change affecting the furfural constants, it was necessary to examine the effect of the preliminary alkaline treatment, and the change in the furfuroid group was in fact localised in this reaction. It was found that, on washing the alkali from the mercerised jute, and further purifying the residue, this latter yielded only 4.2 p.ct. furfural [3.4 p.ct. on original fibre]. The alkaline solution and washings were acidified and distilled from 10 p.ct. HCl, yielding an additional 3.6 p.ct. calculated to the original lignocellulose. By treatment with the concentrated alkali, therefore, the furfuroid of the original lignocellulose undergoes little change, but is selectively dissolved.

This point is under further investigation.

(p. 132) ~Acetylation of Lignocelluloses.~--Acetates are readily formed by boiling the lignocelluloses with acetic anhydride. The derivatives obtained from jute are only generally mentioned in the 1st edition (p.

132). A further study of the reactions in regard to special points has led to some more definite results. The _yields_ of product by the ordinary and simple process are 114-115 p.ct. But on a.n.a.lysing the product an important discrepancy is revealed.

For the saponification we employ a solution of sodium ethylate in the cold. The following numbers were obtained:

Acetic acid Hydrocellulose residue 27.2 77.8 Calc. for diacetate on C_{12}H_{18}O_{9} 30.8 78.4

The derivative is approximately a diacetate, and on the a.s.sumption of a simple ester reaction the yield should be 127 p.ct. a.s.suming that the difference of 13 p.ct. is due to loss of water by internal condensation, it appears that for each acetyl group entering, 2 mol. H_{2}O are split off.

The jute acetate showed the normal reaction with chlorine, and the lignone chloride was dissolved by treatment with sodium sulphite solution. The fibrous residue was colourless. It proved to be a cellulose acetate. The following numbers were obtained on saponification:

Acetic acid Cellulose 31.6 70.0 30.9 68.8 Calc. for diacetate on C_{12}H_{20}O_{10} 29.4 79.9

The interpretation of these numbers appears to be this: in the original reaction with the lignocellulose it is the cellulose residue which is acetylated, and at the same time condensed. The cellulose residue which undergoes condensation is not of the normal const.i.tution, since the normal cellulose is acetylated without condensation (see p. 41). On saponification a portion of the cellulose, in again combining with water, is hydrolysed to soluble products. The lignone group as it exists in the lignocellulose has no free OH groups, and probably no free aldehydic groups such as would react with the anhydride. Such groups may, however, be originally present, and may take part in the internal condensations which have been shown to occur. The furfural constants of the lignocellulose are unaffected by the acetylation and condensation.

The hygroscopic moisture of the product is lowered from 10-11 p.ct. in the original to 4.5 p.ct. The ferric ferricyanide reaction is inhibited by the disappearance of the reactive groups, upon which this curious and characteristic phenomenon depends (1st ed.).

~Acetylation of Benzoates.~--The cellulose dibenzoate (C_{12} basis) and the jute mon.o.benzoate were acetylated under comparative conditions The results were as follows:

C_{12} basis Cellulose dibenzoate Jute mon.o.benzoate Calc. for Calc. for Found diacetate on Found diacetate on Ester reaction dibenzoate mon.o.benzoate Yield 111 p.ct. 115 p.ct. 124 p.ct. 120 p.ct.

Saponification {Cellulose} {Lignocellulose} 53.5 52.6 59.8 61.9 NaOH combining 21.3 23.9 28.4 24.3

From these results it would appear that the number of acetyl groups entering the benzoates is the same as with the unbenzoylated fibres, the benzoyl has no influence upon the hydroxyls as against the acetyl. At the same time the internal condensation noticed in the acetylation of the jute appears not to occur in the case of the benzoate.

~Nitric Esters.~--The numbers resulting from the quant.i.tative study of the ester reaction and product (1st ed. p. 133) show a very large divergence of the yield of product from that which would be calculated from its composition (N p.ct.) on the a.s.sumption that the ester reaction is simple. We have repeated the results, and find with a yield of 145 p.ct.

that the product contains 11.8 p.ct. N.

The reaction

C_{12}H_{18}O_{9} + 4HNO_{3} - 4H_{2}O

gives a tetranitrate with 11.5 p.ct. N and a yield of 159 p.ct. The ester reaction, therefore, is not simple. There are two sources of the loss of weight. The first of these is evident from the occurrence of certain secondary reactions which result in the solution of a certain proportion of the fibre substance in the acid mixture. To determine this quant.i.tatively we have devised a suitable variation of the method of combustion with chromic acid (1st ed.).

The variation is required to meet the difficulty occasioned by the tension of the nitric acid and products of deoxidation. The mixed acids (10 c.c.), containing the organic by-products in solution, are carefully diluted in a small flask with an equal volume of water, preventing rise of temperature. Nitrous fumes are evolved during the dilution. Strong sulphuric acid (15 c.c.) is now added, and the residue of nitrous fumes expelled by a current of air, agitating the contents of the flask from time to time. The combustion with CrO_{3} is then proceeded with in the ordinary way. The gases evolved are measured (total volume) and calculated to C present in the form of products derived from the lignocellulose; and, a.s.suming that this contains 47 p.ct. C, we may express the result approximately in terms of the fibre substance. The method was controlled by blank experiments, in which citric acid was taken as a convenient carbon compound for combustion.

The C found was 34.9 p.ct. as against 34.3 p.ct. calculated. By this method we find that with maximum yields of nitrate at 143-145 p.ct. the organic matter in solution in the acid mixture amounted to 4.9 to 5.3 p.ct. of the original lignocellulose.

Introducing this quant.i.ty as a correction of the yield of nitrate in the original reaction, we must express the 143 parts as obtained from 95 of fibre substance instead of 100.

The yield per molecule C_{12}H_{18}O_{9} (= 306) is therefore 462, whereas for a tetranitrate formed by a simple ester reaction the yield should be 486. The difference (24) represents 1.5 mol. H_{2}O split off by internal condensation.

The correction for total N is relatively small, raising it from 11.5 to 12.2, which remains in close agreement with the experimental numbers.

_Mon.o.benzoate._--Treated with the acid mixture yields a mixed nitrate.

The yield is 130 p.ct., and the product contains 7.6 p.ct. O.NO_{2} nitrogen. These numbers approximate to those required for reaction with 4HNO_{3} groups, three of the residues entering the cellulose, and one (as NO_{2}) the benzene ring of the subst.i.tuting group. For such a reaction the calculated numbers are: Yield 144 p.ct.; O.NO_{2} nitrogen 7.1 p.ct.

The experimental numbers require correcting for the amount of loss in the form of products soluble in the acid mixture, viz. 7.6 p.ct.; but they remain within the range of the experimental errors sufficiently to show that the benzoyl group limits the number of OH groups taking part in the ester reaction to three. The corrected yield per 1 mol. of jute benzoate (410) is 576, as against the calculated 590 for 4HNO_{3} reacting. A loss of 1H_{2}O per molecule by internal condensation is therefore indicated.

~Denitration.~--The removal of the nitric groups from the esters is effected by digestion with ammonium sulphide. But the reactions are by no means simple. There is considerable hydrolysis of the lignocellulose to soluble products. Thus the _tetranitrate_ yields only 46.4 of denitrated fibre in place of the calculated 66. The product is a cellulose, yielding only 0.5 per cent. furfural. The hydrolysed by-products, moreover, when freed from sulphur and distilled from hydrochloric acid, yielded only an additional 2.5 p.ct. furfural, calculated to the original lignocellulose.

These statistics confirm the evidence that the ester reaction is not simple. Such changes take place in the lignone-beta-cellulose complex that they revert, not to their original form, but to soluble derivatives of different const.i.tution. The mixed nitrate from the benzoate is denitrated to a cellulose amidobenzoate, which confirms the localisation of a nitro-group in the benzoyl residue.

(p. 157) ~General Characteristics of the Lignocelluloses.~--Later investigations have somewhat modified and simplified our views of the const.i.tution of the typical lignocellulose (jute), so far as this can be dealt with by the statistics of its more important decompositions (original, pp. 157-161).