Jan., 19511 FRYD AND KIFF: THE DETERMINATION OF MOISTURE IN TOBACCO 25 The Determination of Moisture in Tobacco BY C. F. M. FRYD AND P. R. KIFF (Presented at the meeting of the Society op2 Wednesdgy, February lst, 1950) Moisture in tobacco is commonly determined by the statutory method as 'loss in weight on drying at a defined temperature. During recent years attention has been directed to the effect of variations in the time of drying and in the degree of ventilation in ovens; such variations, it has been shown, may seriously affect the apparent moisture content. This sensitivity of the apparent moisture content to drying conditions is demonstrated to arise at least in part in a component which does not exist as moisture in the undried tobacco, b u t is produced by reactions of the Maillard type or by other reactions involving reducing sugar, during the process of drying.Apparatus is described by the use of which experimental variations in the apparent moisture content of tobacco can be minimised. FOR revenue purposes it is necessary to determine the moisture content of tobacco; and the relevant Acts of Parliament prescribe that this shall be taken as the decrease in weight when the tobacco is dried at a temperature of 212" by Fahrenheit's thermometer. As temperature is the only parameter mentioned, it is not surprising that until recently, interest was concentrated almost entirely upon it. Tatel drew attention to difficulties in the equalisa- tion of temperature within ovens, and Fig. 1, reproduced from his paper, illustrates the extent of the variation that may take place in readings of the temperature at various points within an approximately cubical electrically heated oven.An internal fan rotating at sufficient speed was shown to obviate these variations, and a marked improvement in the repeatability of moisture determinations made in ovens so fitted was the result. Iles and Sharman2 have recently described the effect of the rate of ventilation within ovens on the apparent moisture content of Bright Virginia tobacco. This type of tobacco has long been known to yield moisture results that vary with the apparatus in which the determination is made. Variations of this kind suggest that similarly discordant results may be expected in the moisture figures for other tobaccos and indeed for biological material in general.As a sequel to the work of Iles and Sharman, which had been made available t o us before publication, an oven was constructed in the Government Laboratory in which up t o six 10-g samples of tobacco could be submitted for any desired time and at any pressure between fairly high vacuum and atmospheric to a metered ventilation. It was arranged that the apparatus could be totally and deeply immersed in a well-stirred glycerine - water bath, the temperature of which could be held to within 0.2" C at any point between 60" and 105" C (Fig. 2). Preliminary experiments with this apparatus made it clear that the repeatability of moisture determinations (or more exactly loss-in-weight determinations) on cut tobacco was limited by the homogeneity of the samples. For certain reasons interest in the Government Laboratory was directed especially to cigarette tobacco, and so the raw material for the experiments was obtained from a large number of cigarettes that had been submitted as samples. Cigarette tobacco consists physically of finely shredded lamina of the leaf proper together with portions of the veins and midribs (known as stalk) contaminated with a small proportion of "sand," the residue of those particles of blown dust and soil which have not been detached during the process of manufacture. It is not difficult to mix this tobacco thoroughly; but in samples of 10 g (an amount frequently used for moisture determinations) variations in the proportion of stalk or sand from sample to sample of the mixed material still produces detectable variations in the apparent moisture.After several experiments, a technique to minimise these effects was devised. In a blender (Fig. 3) consisting of a closed drum some 56 cm in diameter and provided with some sixty inward-pointing 4-cm spikes attached to the inner periphery, the tobacco, in lots of about 1 kg, is tossed by rotation of the drum26 FRYD AND KIFF: THE D-ETERMINATION OF [Vol. 76 for 24 hours at such a speed that the tobacco is just not carried round by centrifugal force but falls continuously from the highest point. (One end of the drum is made wholly of quarter-inch Perspex t o facilitate the choice of the appropriate speed.) During this process the sand particles are shaken free and the shreds of tobacco are reduced in length more than are the portions of stalk and vein.The tobacco at this stage is removed and sieved, only the portion which passes 12 meshes to the inch and fails to pass 40 meshes to the inch being retained and returned to the blending drum for a further 30 minutes. After immediate removal at the expiry of this time, the tobacco is packed tightly into glass jars provided with # d 3 P TEMPERATURE. OF Fig. 1. Variation of temperature in electrically-heated oven rubber sealing rings and holding some 200g per jar. Experiments indicate that in these containers no noticeable alteration in apparent moisture takes place during storage of several months, as long as the containers themselves are stored in conditions of equable temperature. For large-scale replicate work, the weighing out of samples requires the utmost care if loss or absorption of moisture by the sample is to be avoided.Two operators are needed, one of whom places in the previously tared and numbered drying-pan an approximate 10 g of the sample; the other weighs it immediately to the required (1 mg) accuracy on an aperiodic optical balance, and returns it to his companion, continuing with the next pan which has been prepared in the interim. The weighed pans a:re stored immediately in airtight containers, in order that the whole series of perhaps one hundred samples may enter the ovens in the same condition. Precautions must be taken to ensure that neither sunlight nor currents of air affect the moisture content of the bulk from which portions are being drawn; and allJan., 19511 MOISTURE IN TOBACCO 27 portions removed from an overfull pan, or remaining in almost empty storage bottles, must be immediately discarded.With these precautions repeatability can be attained which for organic natural materials is of high standard (Table I). m Inner oven Outer Container- I I Fig. 2. Experimental oven Table I1 shows the extent to which variations in the apparent moisture content can occur in homogenised material if the conditions of drying are artificially varied. Results of this nature called not only for the development of apparatus in which moisture determinations could be made under standard conditions, but also for some rational explanation of the cause of the discrepancies. Table I11 shows some results obtained in the Government Laboratory which illustrate the chemical effect of heating for various times and under varied Fig. 3.Blender conditions. It is immediately apparent that the major quantitative changes have taken place in the sugars present. Iles and Sharman2 suggest that reactions of the MaillardS type may be responsible and point to the blackening of the heated tobacco in support of this suggestion; but in addition it may be mentioned that the production of dark and humic substances by the direct action of organic acids on reducing sugars is well Table IV gives some results obtained in the Government Laboratory for certain of the acids and sugars28 FRYD AND KIFF: THE IIETERMINATION OF [Vol. 76 known to be present in tobacco. Some preliminary work, still mainly of a qualitative nature, has been carried out on the volatile end-products of these reactions associated with the heating of tobacco.It is certain that the loss in weight, or apparent moisture, is for the main part accounted for by the liberation of water itself. There is little, if any, carbon dioxide, but TABLE I APPARENT PERCENTAGE OF MOISTURE IN HOMOGENISED BRIGHT VIRGINIA TOBACCO A series of 10-g samples dried at 100" C for 17 hours % f \ 11-63 11.62 11.59 11.56 11.61 11.56 11-56 11.63 11-59 11-66 11-52 11.58 11.55 11.56 11.58 11.59 11.66 11-63 11-61 11.54 11.63 11-53 11.53 11.55 11.59 11-61 11.58 11.55 11.48 11-52 11.55 11.56 11.59 11-58 11.58 11.55 11.53 11.49 11.55 11.57 11.60 11.63 11-61 11-59 11.55 11-57 11-59 11-61 some organic matter of an aldehydic nature is evolved which is not formaldehyde. I t is trapped by concentrated sulphuric acid, with which it reacts to form an insoluble carbonaceous residue.TABLE I1 Further work is in progress on this aspect of the problem. APPARENT PERCENTAGE OF MOISTURE IN HOMOGENISED BRIGHT VIRGINIA TOBACCO 10-g samples dried under varied conditions Treatment Loss in weight % 1. 17 hours drying a t 60" C with 4 litres/min. air flow . . .. . . 8-62 3. 17 hours drying a t 100" C with 0-05 litres/min. air flow . . 13.37 4. 30 hours heating at 100" C followed by 17 hours drying at 100; C with 4 litres/min. air flow . . .. .. .. .. .. . . . . 16.62 2. 17 hours drying a t 100" C with 4 litres/min. air flow . . .. . . 11-91 Sufficient evidence is available therefore to show that the operation of heating tobacco for any appreciable period at a temperature of about 100" C results not only in the removal of all or part of the moisture originally present, but also in a chemical reaction in which moisture is produced.As may be expected, the component of the total apparent moisture accounted for by the results of chemical decomposition varies not only, as has been shown,2 TABLE I11 EFFECT OF OVEN CONDITIONS ON THE COMPOSITION* OF HOMOGENISED BRIGHT CIGARETTE TOBACCO Dried at 100" C for 17 hours -----7 After pretreatment of 17 hours at 100" C Dried at 60" C, full ventilation, without ventilation, With immediate :4 % % Total soluble matter .. .. .. 48.8 45.6 44.0 Total insoluble matter . . .. .. 42.7 43.1 40.8 Apparent moisture .. .. .. .. 8.4 11-3 15.1 Sum of above . . .. .. .. 99.9 100.0 99.9 Total sugars . . . . . . .. .. 16.3 11.9 4.6 Reducing sugars . . .. .. .. 13.2 9.5 4.3 Free acid (ml of 0-1 N NaOH per log tobacco) 72 64 53 Light petroleum ext. .. . . . . , 2.75 2.77 2.66 Total nitrogen . . .. . . . . 1.47 1.39 1-32 Nicotine . . .. . . . . . . 1.51 1.23 1.20 Free ammonia . . .. .. . . 0.12 0.09 0.02 Nitrogen, other than ammonia and Resins .. .. .. .. .. 1.88 1-80 1.76 nicotine nitrogen . . .. . . . . 1.11 1-10 1-10 * For the analytical methods used, see appendix, p. 31.Jan., 19511 MOISTURE IN TOBACCO 29 with the degree of ventilation, but also with the temperature attained (Fig. 4) and with the time during which the tobacco is held at that temperature (Fig. 5). Other considerations, which have effects fortunately of a minor order, are the rate of initial heating of the sample and the barometric pressure within the oven.Thus it is apparent that the use of ovens to determine the moisture in tobacco does not result in the determination of anything that can TABLE IV EFFECT OF HEATING MIXTURES OF 1 G OF VARIOUS ORGANIC ACIDS WITH 5 G OF EITHER LAEVULOSE OR GLUCOSE FOR 36 HOURS IN A DRYING OVEN AT ABOUT 96°C Sugar Acid Loss, Colour Effervescence Residue Glucose 33 33 > 3 Laevulose >? >) 33 Malic Citric Oxalic Protein Malic Citric Oxalic Protein hydro1 y sate h y drolysate % 3.7 Straw 4-5 Straw 16.4 Black 16.6 Black 8.5 Dark brown 9.4 Dark brown 27.4 Black 12.1 Black nil nil Much Much nil nil Little Much Soluble Soluble Soluble Soluble Soluble Soluble Insoluble Partly in soluble Sufficient water was added to ensure solution of the reagents, but the loss-in-weight percentages quoted below are based on the weight of the residue compared with the sum of the weights of the sugar and acid ingredients only.be called the “true moisture” of the sample. The most to be expected from oven determina- tions is a series of consistent and repeatable figures in which the effect of oven decomposition is minimised and standardised. But, as for revenue purposes the loss-in-weight on drying at 212” C remains the legal moisture content of tobacco, the Government Laboratory and presumably the trade in general will continue to use oven drying; hence it may be of interest to describe briefly the type of apparatus that is in use for moisture determination to-day and the requirements that have led to the present design.It is clear that the major con- siderations must be- (1) Ventilation sufficient to ensure that at no time is evaporation from the surface of the tobacco itself slowed down, or the unavoidable chemical reactions allowed to occur at a high rate of velocity in the liquid phase. (2) Constancy of final temperature within the oven. (3) Identity of heat history for all the samples during the process of drying. (4) Speed in dealing with a large number of samples. (5) Economy of the time of skilled operatives, which in effect means drying overnight so that the working period may be devoted to weighing. This requirement itself necessitates that the oven shall be fully automatic.If the heating of an individual sample pan placed in a hot oven is considered, it is clear that heat may reach it by one of three paths-contact with the ambient air, conduction through metal, or radiation from the surrounding surfaces. Some early experiments carried out in the Government Laboratory indicated that the heat flow through the air, though not negligible, contributes only slowly to the rise in temperature of the sample. Conduction through metal is important, but is likely to become uncertaiqas the pristine surface of the metal suffers from routine use. Radiation from surrounding surfaces is probably the largest and most consistent contributor to the considerable amount of heat required to raise the temperature of the sample and supply energy for the evaporation of moisture. To ensure identity of heat history for each sample, the Government Laboratory has found it necessary to construct ovens in which each pan is equidistant from the radiating surfaces and in which the whole of the internal surface is a t the same temperature.In such an oven there is no balancing out of hot and cold spots as is necessary with, for instance, electric ovens of the normal type, and assurance can be felt that not even for short periods is any sample exposed to radiation from surfaces at a higher temperature than that finally attained by the whole oven. Acknowledgment must be made of the great help of the Imperial Tobacco Company, Limited, in this connection. In 1936, Dr. Jollyman of that company showed us an apparatus embodying an oven consisting of a cylindrical shell heated by an outer steam jacket; and30 FRYD AND KIFF: THE DETERMINATION OF [Vol.76 although his successors and ourselves have introduced modifications, his apparatus is very little different in essence from that shown in Fig. 6, which is a diagram of one of the ovens now in use in the Government Laboratory. LOSS OF WEIGHT 'x Fig. 4. Variation of decomposition with temperature; drying time 17 hours In this device the oven proper is the central chamber, circular in plan, into which an internal movable rack (on which a number of sample pans are arranged symmetrically in tiers) is lowered. Up the central axis of the oven rises a perforated tube, distributing radially a flow of preheated air forced in by a centrifugal, pump (not shown).Round the bottom, sides, and top of the chamber circulates a flow of steam generated from a container of distilled water below, into which the condensate returns. With ovens of this type, replicate determina- Fig. 5. Variation of decomposition with time at 100" CJan., 19511 MOISTURE I N TOBACCO 31 tions having a degree of repeatability of the order of those shown in Table I may be confidently expected as a routine operation. Steam 7j=-i+ heated lid from lid Steam inlet to lid +* Air preheating coil 7- Air distribution tube An apparatus including some of the features of such an oven is the subject of a patent application made at the instance of the Imperial Tobacco Company, Limited, and quoted by Iles and Sharman.2 APPENDIX NOTES ON THE METHODS USED IN OBTAINING THE DATA IN TABLE 111 Nicotine-Five grams of tobacco is steam distilled with 2 g of magnesium oxide and 50g of sodium chloride, and 250ml collected.A further 5ml is collected and tested for absence of nicotine. A 100-ml aliquot is acidified and the nicotine is precipitated with silicotungstic acid. The precipitate is filtered through a Gooch crucible, ignited, and weighed. Light petroleum extract-Tobacco, 13-33 g, is in contact with 100 ml light petroleum for 18 hours. Final residue is heated at 100" C for 14 hours. Total nitrogen-One gram of tobacco, 10 g of potassium sulphate, 0-5 g of copper sulphate and 20 ml of concentrated sulphuric acid are boiled in a Kjeldahl flask for 4 hours or until clear. The solution is then diluted, neutralised with sodium hydroxide, distilled into 25 ml of 0.1 N sulphuric acid and titrated back.75 ml of supernatant liquor is removed and evaporated on steam.32 FRYD AND KIFF [Vol. 76 The mass is made up to 200 ml with silicotungstic acid, 12 per cent. solution; 100 ml of filtrate is then treated with 10ml of neutral, 30 per cent. lead acetate solution. After filtration through a Buchner funnel, 100 ml is collected, 5 In1 of saturated potassium oxalate is added, the solution made up to 110 ml and again filtered (Solution A). This solution is used to titrate Fehling's solution, using methylene blue ;is indicator. Total sztgars-As above, but 50 ml of Solution A is boiled for 30 seconds with 3 or 4 drops of concentrated hydrochloric acid and made up to 100ml after cooling.Total soluble-Five grams of tobacco is digested with 150 ml of water overnight. The mass is filtered through a Buchner funnel and washing continued until 1OOOml of extract is obtained. Total insoluble-Fibrous matter from the above is carefully removed and dried at 100" C to constant weight. Ammonia-Five grams of tobacco, 20 ml of 1 : 4 hydrochloric acid and 20 ml of silico- tungstic acid, 12 per cent. is made up to 100 ml with water, 50 ml of filtrate collected, distilled with sodium hydroxide into 15 ml of 0.1 N sulphuric acid and titrated back. Resins-Five grams of tobacco is digested with 100 ml of ethanol overnight, 50 ml of filtrate evaporated to dryness and digested with water. After re-filtration, the insoluble resins are dried at 100" C and weighed.Reducing sugars-Tobacco, 12.1 g, is digested with 150 ml of water overnight. 100 ml of extract is dried for 18 hours on a steam-bath. We wish to thank our colleagues in the Government Laboratory for help in the preparation of this paper, and the Government Chemist for permission to publish. REFERENCES 1. 2. 3. 4. CLEMENTS INN PASSAGE Tate, F. G. H., Analyst, 1934, 59, 168. Iles, W. G., and Sharman, C. F., J . Soc. Chem. Ind., 1949, 68, 174. Maillard, L. C., Compt. rend., 1912, 154, 66. Evdokimov, A. G., Sci. Mem. St. Univ. Leningrad, Chem. Ser., 1938, 3, 38. GOVERNMENT LABORATORY STRAND, LONDON, W.C.2 D I sc u SSION MR. L. G. BECKETT pointed out that there was a considerable amount of oxygen at a pressure of 1 cm of mercury, and this could account for the blackening of the tobacco leaf in spite of the authors' contention to the contrary. He Considered it essential that any remark concerning pressure should be qualified by the words "gauge" or "absolute" to avoid difficulties that may be experienced by other workers in interpreting results. He drew attention to the necessity for care when referring to degrees of vacuum. DR. J. H. HAMENCE asked if the new oven really reached a tsmperature of 100" C. MR. FRYD replied that it did. If the oven was empty it would reach 100" C in a very short time When the oven was fully loaded with about 100 samples, it would take from 2& to 4 hours- DR. HAMENCE said that the meeting had served a valuable purpose because it had brought the air The latest conclusions for the estimation of water suggested indeed. depending naturally upon the water content of the sample material-to reach 100" C. flow methods to the notice of the Society. that the air flow method was superior to other available methods.