SECTION II. SYNTHETICAL DERIVATIVES—SULPHOCARBONATES AND ESTERS
(p. 25) Cellulose sulphocarbonate.—Further investigations of the reaction of formation as well as the various reactions of decomposition of the compound, have not contributed any essential modification or development of the subject as originally described in the author's first communications. A large amount of experimental matter has been accumulated in view of the ultimate contribution of the results to the general theory of colloidal solutions. But viscose is a complex product and essentially variable, through its pronounced tendency to progressive decomposition with reversion of the cellulose to its insoluble and uncombined condition. The solution for this reason does not lend itself to exact measurement of its physical constants such as might elucidate in some measure the progressive molecular aggregation of the cellulose in assuming spontaneously the solid (hydrate) form. Reserving the discussion of these points, therefore, we confine ourselves to recording results which further elucidate special points.
Normal and other celluloses.—We may certainly use the sulphocarbonate reaction as a means of defining a normal cellulose. As already pointed out, cotton cellulose passes quantitatively through the cycle of treatments involved in solution as sulphocarbonate and decomposition of the solution with regeneration as structureless or amorphous cellulose (hydrate).
Analysis of this cellulose shows a fall of carbon percentage from 44.4 to 43.3, corresponding with a change in composition from C6H10O5 to 4C6H10O5.H2O. The partial hydrolysis affects the whole molecule, and is limited to this effect, whereas, in the case of celluloses of other types, there is a fractionation of the mass, a portion undergoing a further hydrolysis to compounds of lower molecular weight and permanently soluble. Thus in the case of the wood celluloses the percentage recovered from solution as viscose is from 93 to 95 p.ct. It is evident that these celluloses are not homogeneous. A similar conclusion results from the presence of furfural-yielding compounds with the observation that the hydrolysis to soluble derivatives mainly affects these derivatives. In the empirical characterisation of a normal cellulose, therefore, we may include the property of quantitative regeneration or recovery from its solution as sulphocarbonate.
In the use of the word 'normal' as applied to a 'bleached' cotton, we have further to show in what respects the sulphocarbonate reaction differentiates the bleached or purified cotton cellulose from the raw product. The following experiments may be cited: Specimens of American and Egyptian cottons in the raw state, freed from mechanical, i.e. non-fibrous, impurities, were treated with a mercerising alkali, and the alkali-cotton subsequently exposed to carbon disulphide. The product of reaction was further treated as in the preparation of the ordinary solution; but in place of the usual solution, structureless and homogeneous, it was observed to retain a fibrous character, and the fibres, though enormously swollen, were not broken down by continued vigorous stirring. After large dilution the solutions were filtered, and the fibres then formed a gelatinous mass on the filters. After purification, the residue was dried and weighed. The American cotton yielded 90.0 p.ct., and the Egyptian 92.0 p.ct. of its substance in the form of this peculiar modification. The experiment was repeated, allowing an interval of 24 hours to elapse between the conversion into alkali-cotton and exposure of this to the carbon disulphide. The quantitative results were identical.
There are many observations incidental to chemical treatments of cotton fabrics which tend to show that the bleaching process produces other effects than the mere removal of mechanical impurities. In the sulphocarbonate reaction the raw cotton, in fact, behaves exactly as a compound cellulose. Whether the constitutional difference between raw and bleached cotton, thus emphasised, is due to the group of components of the raw cotton, which are removed in the bleaching process, or to internal constitutional changes determined by the bleaching treatments, is a question which future investigation must decide.
The normal sulphocarbonate (viscose).—In the industrial applications of viscose it is important to maintain a certain standard of composition as of the essential physical properties of the solution, notably viscosity. It may be noted first that, with the above-mentioned exception, the various fibrous celluloses show but slight differences in regard to all the essential features of the reactions involved. In the mercerising reaction, or alkali-cellulose stage, it is true the differences are considerable. With celluloses of the wood and straw classes there is a considerable conversion into soluble alkali-celluloses. If treated with water these are dissolved, and on weighing back the cellulose, after thorough washing, treatment with acid, and finally washing and drying, it will be found to have lost from 15 to 20 p.ct. in weight. The lower grade of celluloses thus dissolved are only in part precipitated in acidifying the alkaline solution. On the other hand, after conversion into viscose, the cellulose when regenerated re-aggregates a large proportion of these lower grade celluloses, and the final loss is as stated above, from 5 to 7 p.ct. only.
Secondly, it is found that all the conditions obtaining in the alkali-cellulose stage affect the subsequent viscose reaction and the properties of the final solution. The most important are obviously the proportion of alkali to cellulose and the length of time they are in contact before being treated with carbon disulphide. An excess of alkali beyond the 'normal' proportion—viz. 2NaOH per 1 mol. C6H10O5—has little influence upon the viscose reaction, but lowers the viscosity of the solution of the sulphocarbonate prepared from it. But this effect equally follows from addition of alkali to the viscose itself. The alkali-cellulose changes with age; there is a gradual alteration of the molecular structure of the cellulose, of which the properties of the viscose when prepared are the best indication. There is a progressive loss of viscosity of the solution, and a corresponding deterioration in the structural properties of the cellulose when regenerated from it—especially marked in the film form. In regard to viscosity the following observations are typical:—
(a) A viscose of 1.8 p.ct. cellulose prepared from an alkali-cellulose (cotton) fourteen days old.
(b) Viscose of 1.8 p.ct. cellulose from an alkali-cellulose (cotton) three days old.
(c) Glycerin diluted with 1/3 vol. water.
| a | b | b | c | |
| Diluted with equal vol. water | ||||
| Times of flow of equal volumes from narrow orifice in seconds | 112 | 321 | 103 | 170 |
Similarly the cellulose in reverting to the solid form from these 'degraded' solutions presents a proportionate loss of cohesion and aggregating power expressed by the inferior strength and elasticity of the products. Hence, in the practical applications of the product where the latter properties are of first importance, it is necessary to adopt normal standards, such as above indicated, and to carefully regulate all the conditions of treatment in each of the two main stages of reaction, so that a product of any desired character may be invariably obtained.
Incidentally to these investigations a number of observations have been made on the alkali-cellulose (cotton) after prolonged storage in closed vessels. It is well known that starch undergoes hydrolysis in contact with aqueous alkalis of a similar character to that determined by acids [Béchamp, Annalen, 100, 365]. The recent researches of Lobry de Bruyn [Rec. Trav. Chim. 14, 156] upon the action of alkaline hydrates in aqueous solution on the hexoses have established the important fact of the resulting mobility of the CO group, and the interchangeable relationships of typical aldoses and ketoses. It was, therefore, not improbable that profound hydrolytic changes should occur in the cellulose molecule when kept for prolonged periods as alkali-cellulose.
We may cite an extreme case. A series of products were examined after 12-18 months' storage. They were found to contain only 3-5 p.ct. 'soluble carbohydrates'; these were precipitated by Fehling's solution but without reduction on boiling. They were, therefore, of the cellulose type. On acidifying with sulphuric acid and distilling, traces only of volatile acid were produced. It is clear, therefore, that the change of molecular weight of the cellulose, the disaggregation of the undoubtedly large molecule of the original 'normal' cellulose—which effects are immediately recognised in the viscose reactions of such products—are of such otherwise limited character that they do not affect the constitution of the unit groups. We should also conclude that the cellulose type of constitution covers a very wide range of minor variations of molecular weight or aggregation.
The resistance of the normal cellulose to the action of alkalis under these hydrolysing conditions should be mentioned in conjunction with the observations of Lange, and the results of the later investigations of Tollens, on its resistance to 'fusion' with alkaline hydrates at high temperatures (180°). The degree of resistance has been established only on the empirical basis of weighing the product recovered from such treatment. The product must be investigated by conversion into typical cellulose derivatives before we can pronounce upon the constitutional changes which certainly occur in the process. But for the purpose of this discussion it is sufficient to emphasise the extraordinary resistance of the normal cellulose to the action of alkalis, and to another of the more significant points of differentiation from starch.
Chemical constants of cellulose sulphocarbonate (solution).—In investigations of the solutions we make use of various analytical methods, which may be briefly described, noting any results bearing upon special points.
Total alkali.—This constant is determined by titration in the usual way. The cellulose ratio, C6H10O5: 2NaOH, is within the ordinary error of observation, 2: 1 by weight. A determination of alkali therefore determines the percentage of cellulose.
Cellulose may be regenerated in various ways—viz. by the action of heat, of acids, of various oxidising compounds. It is purified for weighing by boiling in neutral sulphite of soda (2 p.ct. solution) to remove sulphur, and in very dilute acids (0.33 p.ct. HCl) to decompose residues of 'organic' sulphur compounds. It may also be treated with dilute oxidants. After weighing it may be ignited to determine residual inorganic compounds.
Sulphur.—It has been proved by Lindemann and Motten [Bull. Acad. R. Belg. (3), 23, 827] that the sulphur of sulphocarbonates (as well as of sulphocyanides) is fully oxidised (to SO3) by the hypochlorites (solutions at ordinary temperatures). The method may be adapted as required for any form of the products or by-products of the viscose reaction to be analysed for total sulphur.
The sulphur present in the form of dithiocarbonates, including the typical cellulose xanthogenic acid, is approximately isolated and determined as CS2 by adding a zinc salt in excess, and distilling off the carbon disulphide from a water bath. From freshly prepared solutions a large proportion of the disulphide originally interacting with the alkali and cellulose is recovered, the result establishing the general conformity of the reaction to that typical of the alcohols. On keeping the solutions there is a progressive interaction of the bisulphide and alkali, with formation of trithiocarbonates and various sulphides. In decomposing these products by acid reagents hydrogen sulphide and free sulphur are formed, the estimation of which presents no special difficulties.
In the spontaneous decomposition of the solution a large proportion of the sulphur resumes the form of the volatile disulphide. This is approximately measured by the loss in total sulphur in the following series of determinations, in which a viscose of 8.5 p.ct. strength (cellulose) was dried down as a thin film upon glass plates, and afterwards analysed:
(a) Proportion of sulphur to cellulose (100 pts.) in original.
(b) After spontaneous drying at ordinary temperature.
(c) After drying at 40°C.
(d) As in (c), followed, by 2 hours' heating at 98°.
(e) As in (c), followed by 5 hours' heating at 98°.
| a | b | c | d | e | |
| Total sulphur | 40.0 | 25.0 | 31.0 | 23.7 | 10.4 |
The dried product in (b) and (c) was entirely resoluble in water; in (d) and (e), on the other hand, the cellulose was fully regenerated, and obtained as a transparent film.
Iodine reaction.—Fresh solutions of the sulphocarbonate show a fairly constant reaction with normal iodine solution. At the first point, where the excess of iodine visibly persists, there is complete precipitation of the cellulose as the bixanthic sulphide; and this occurs when the proportion of iodine added reaches 3I2: 4Na2O, calculated to the total alkali.
Other decompositions.—The most interesting is the interaction which occurs between the cellulose xanthogenate and salts of ammonia, which is taken advantage of by C. H. Stearn in his patent process of spinning artificial threads from viscose. The insoluble product which is formed in excess of the solution of ammonia salt is free from soda, and contains 9-10 p.ct. total sulphur. The product retains its solubility in water for a short period. The solution may be regarded as containing the ammonium cellulose xanthate. This rapidly decomposes with liberation of ammonia and carbon disulphide, and separation of cellulose (hydrate). As precipitated by ammonium-chloride solution the gelatinous thread contains 15 p.ct. of cellulose, with a sp.gr. 1.1. The process of 'fixing'—i.e. decomposing the xanthic residue—consists in a short exposure to the boiling saline solution. The further dehydration, with increase of gravity and cellulose content, is not considerable. The thread in its final air-dry state has a sp.gr. 1.48.
Cellulose Benzoates.—These derivatives have been further studied by the authors. The conditions for the formation of the monobenzoate [C6H9O4.O.CO.Ph] are very similar to those required for the sulphocarbonate reaction. The fibrous cellulose (cotton), treated with a 10 p.ct. solution NaOH, and subsequently with benzoyl chloride, gives about 50 p.ct. of the theoretical yield of monobenzoate. Converted by 20 p.ct. solution NaOH into alkali-cellulose, and with molecular proportions as below, the following yields were obtained:—
| Calc. for Monobenzoate | ||
| (a) C6H10O5 : 2.0-2.5 NaOH : C6H5.COCl— | 150.8} | 164.0 |
| (b) C6H10O5 : 2.0-2.5 NaOH : 1.5 mol. C6H5COCl | 159.0} |
An examination of (a) showed that some dibenzoate (about 7 p.ct.) had been formed. The product () was exhaustively treated with cuprammonium solution, to which it yielded about 20 p.ct. of its weight, which was therefore unattacked cellulose.
Under conditions as above, but with 2.5 mol. C6H5COCl, a careful comparison was made of the behaviour of the three varieties of cotton, which were taken in the unspun condition and previously fully bleached and purified.
| Sea Island | Egyptian | American | |
| Aggregate yield of benzoate | 153 | 148 | 152 |
| Moisture in air dry state | 5.28 | 5.35 | 5.15 |
| Proportion of dibenzoate p.ct. | 8.30 | 13.70 | 9.4 |
| Yield of cellulose by saponification | 58.0 | 54.0 | 58.3 |
It appears from these results that the benzoate reaction may proceed to a higher limit (dibenzoate) in the case of Egyptian cotton. This would necessarily imply a higher limit of 'mercerisation,' under equal conditions of treatment with the alkaline hydrate. It must be noted that in the conversion of the fibrous cellulose into these (still) fibrous monobenzoates, there are certain mechanical conditions imported by the structural features of the ultimate fibres. For the elimination of the influence of this factor a large number of quantitative comparisons will be necessary. The above results are therefore only cited as typical of a method of comparative investigation, more especially of the still open questions of the cause of the superior effects in mercerisation of certain cottons (see p. 23). It is quite probable that chemical as well as structural factors co-operate in further differentiating the cottons.
Further investigation of the influence upon the benzoate reaction, of increase of concentration of the soda lye, used in the preliminary alkali cellulose reaction, from 20 to 33 p.ct. NaOH, established (1) that there is no corresponding increase in the benzoylation, and (2) that this ester reaction and the sulphocarbonate reaction are closely parallel, in that the degree and limit of reaction are predetermined by the conditions of formation of the alkali cellulose.
Monobenzoate prepared as above described is resistant to all solvents of cellulose and of the cellulose esters, and is therefore freed from cellulose by treatment with the former, and from the higher benzoate by treatment with the latter. Several of these, notably pyridine, phenol and nitrobenzene, cause considerable swelling and gelatinisation of the fibres, but without solution.
Structureless celluloses of the 'normal' type, and insoluble therefore in alkaline lye, treated under similar conditions to those described above for the fibrous celluloses, yield a higher proportion of dibenzoate. The following determinations were made with the cellulose (hydrate) regenerated from the sulphocarbonate:—
| Mol. proportions of reagents | Yield | Dibenzoate p.ct. |
| C6H10O5 : 2NaOH : 2BzCl | 145 | 34.7 |
| [Caustic soda at 10 per cent. NaOH] | ||
| C6H10O5 : 4NaOH : 2BzCl | 162 | 62.7 |
| [Caustic soda at 20 per cent. NaOH] |
Limit of reaction.—The cellulose in this form having shown itself more reactive, it was taken as the basis for determining the maximum proportion of OH groups yielding to this later reaction. The systematic investigations of Skraup [Monatsh. 10, 389] have determined that as regards the interacting groups the molecular proportions 1 OH: 7 NaOH: 5 BzCl, ensure complete or maximum esterification. The maximum of OH groups in cellulose being 4, the reagents were taken in the proportion C6H10O5: 4 [7 NaOH: 5 BzCl]. The yield of crude product, after purifying as far as possible from the excess of benzoic acid, was 240 p.ct. [calculated for dibenzoate 227 p.ct.]. On further investigating the crude product by treatment with solvents, it was found to have still retained benzoic acid. There was also present a proportion of only partially attacked cellulose (monobenzoate). The soluble benzoate amounted to 90 p.ct. of the product. It may be generally concluded that the dibenzoate represents the normal maximum but that with the hydrated and partly hydrolysed cellulose molecule, as obtained by regeneration from the sulphocarbonate, other OH groups may react, but they are only a fractional proportion in relation to the unit group C6H10O5. In this respect again there is a close parallelism between the sulphocarbonate and benzoyl-ester reactions.
The dibenzoate, even when prepared from the fibrous celluloses, is devoid of structure, and its presence in admixture with the fibrous monobenzoate is at once recognised as it constitutes a structureless incrustation. Under the microscope its presence in however minute proportion is readily observed. As stated it is soluble in certain of the ordinary solvents of the cellulose esters, e.g. chloroform, acetic acid, nitrobenzene, pyridine, and phenol. It is not soluble in ether or alcohol.
Hygroscopic moisture of benzoates.—The crude monobenzoate retains 5.0-5.5 p.ct. moisture in the air-dry condition. After removal of the residual cellulose this is reduced to 3.3 p.ct. under ordinary atmospheric conditions. The purified dibenzoates retain 1.6 p.ct. under similar conditions.
Analysis of benzoates.—On saponification of these esters with alcoholic sodium hydrate, anomalous results are obtained. The acid numbers, determined by titration in the usual way, are 10-20 p.ct. in excess of the theoretical, the difference increasing with the time of boiling. Similarly the residual cellulose shows a deficiency of 5-9 p.ct.
It is by no means improbable that in the original ester reaction there is a constitutional change in the cellulose molecule causing it to break down in part under the hydrolysing treatment with formation of acid products. This point is under investigation. Normal results as regards acid numbers, on the other hand, are obtained by saponification with sodium ethylate in the cold, the product being digested with the half-saturated solution for 12 hours in a closed flask.
The following results with specimens of mono- and dibenzoate, purified, as far as possible, may be cited:
| Combustion results | Saponification results | ||||||
| Calc. | C6H5.COOH | Calc. | Cellulose | Calc. | |||
| Monobenzoate | C | 56.60 | 58.65} | 46.0 | 45.9 | 58.0 | 60.8 |
| H | 5.06 | 5.26} | |||||
| Dibenzoate | C | 63.10 | 64.86} | 65.5 | 66.6 | 34.3 | 40.3 |
| H | 3.40 | 4.86} | |||||
The divergence of the numbers, especially for the dibenzoate, in the case of the hydrogen, and yield of cellulose on hydrolysis are noteworthy. They confirm the probability of the occurrence of secondary changes in the ester reactions.
Action of nitrating acid upon the benzoates.—From the benzoates above described, mixed nitro-nitric esters are obtained by the action of the mixture of nitric and sulphuric acids. The residual OH groups of the cellulose are esterified and substitution by an NO2 group takes place in the aromatic residue, giving a mixed nitric nitrobenzoic ester. The analysis of the products points to the entrance of 1 NO2 group in the benzoyl residue in either case; in the cellulose residue 1 OH readily reacts. Higher degrees of nitration are attained by the process of solution in concentrated nitric acid and precipitation by pouring into sulphuric acid. In describing these mixed esters we shall find it necessary to adopt the C12 unit formula.
In analysing these products we have employed the Dumas method for total nitrogen. For the O.NO2 groups we have found the nitrometer and the Schloesing methods to give concordant results. For the NO2 groups it was thought that Limpricht's method, based upon reduction with stannous chloride in acid solution (HCl), would be available. The quantitative results, however, were only approximate, owing to the difficulty of confining the reduction to the NO2 groups of the nitrobenzoyl residue. By reduction with ammonium sulphide the O.NO2 groups were entirely removed as in the case of the cellulose nitrates; the NO2 was reduced to NH2 and there resulted a cellulose amidobenzoate, which was diazotised and combined with amines and phenols to form yellow and red colouring matters, the reacting residue remaining more or less firmly combined with the cellulose.
Cellulose dinitrate-dinitrobenzoate, and cellulose trinitrate-dinitrobenzoate.—On treating the fibrous benzoate—which is a dibenzoate on the C12 basis—with the acid mixture under the usual conditions, a yellowish product is obtained, with a yield of 140-142 p.ct. The nitrobenzoate is insoluble in ether alcohol, but is soluble in acetone, acetic acid, and nitrobenzene. In purifying the product the former solvent is used to remove any cellulose nitrates. To obtain the maximum combination with nitroxy-groups, the product was dissolved in concentrated nitric acid, and the solution poured into sulphuric acid.
The following analytical results were obtained (a) for the product obtained directly from the fibrous benzoate and purified as indicated, (b) for the product from the further treatment of (a) as described:
| Found | Calc. for | ||||
| (a) | (b) | Dinitrate dinitrobenzoate | Trinitrate dinitrobenzoate | ||
| Total | Nitrogen | 7.84 | 8.97 | 7.99 | 9.24 |
| O.NO2 | " | 5.00 | 5.45 | 4.00 | 5.54 |
| NO2 | " (Aromatic) | 2.84 | 3.52 | 3.99 | 3.70 |
With the two benzoyl groups converted into nitro-benzoyl in each product, the limit of the ester reaction with the cellulose residue is reached at the third OH group.
The nitrogen in the amidobenzoate resulting from the reduction with ammonium sulphide was 4.5 p.ct.—as against 5.0 p.ct. calculated. The moisture retained by the fibrous nitrate—nitrobenzoate—in the air-dry state was found to be 1.97 p.ct.
The product from the structureless dibenzoate or tetrabenzoate on the C12 formula, was prepared and analysed with the following results:
| Calc. for Mononitrate tetranitrobenzoate | |||
| Total | Nitrogen | 6.76 | 7.25 |
| O.NO2 | " | 1.30 | 1.45 |
| NO2 | " (Aromatic) | 5.46 | 5.80 |
The results were confirmed by the yield of product, viz. 131 p.ct. as against the calculated 136 p.ct. They afford further evidence of the generally low limit of esterification of the cellulose molecule. From the formation of a 'normal' tetracetate—i.e. octacetate of the C12 unit—we conclude that 4/5 of the oxygen atoms are hydroxyl oxygen. Of the 8 OH groups five only react in the mixed esters described above, and six only in the case of the simple nitric esters. The ester reactions are probably not simple, but accompanied by secondary reactions within the cellulose molecule.
(p. 34) Cellulose Acetates.—In the first edition (p. 35) we have committed ourselves to the statement that 'on boiling cotton with acetic anhydride and sodium acetate no reaction occurs.' This is erroneous. The error arises, however, from the somewhat vague statements of Schutzenberger's researches which are current in the text-books [e.g. Beilstein, 1 ed. p. 586] together with the statement that reaction only occurs at elevated temperatures (180°). As a matter of fact, reaction takes place at the boiling temperature of the anhydride. We have obtained the following results with bleached cotton:
| Yield | Calc. for Monoacetate | ||
| C6H7O4O.C2H3O | |||
| Ester reaction | 121 p.ct. | 125 p.ct. | |
| Saponification | {Cellulose | 79.9 | 79.9 |
| {Acetic acid | 29.9 | 29.4 |
This product is formed without apparent structural alteration of the fibre. It is entirely insoluble in all the ordinary solvents of the higher acetates. Moreover, it entirely resists the actions of the special solvents of cellulose—e.g. zinc chloride and cuprammonium. The compound is in other respects equally stable and inert. The hygroscopic moisture under ordinary atmospheric conditions is 3.2 p.ct.
Tetracetate.—This product is now made on the manufacturing scale: it has yet to establish its industrial value.