SUGAR. —Analysis—Glucose, Water, and Asit.
891
the optically active substances present in the solution.Assume that these consisted of cane sugar, havingdextro-rotatory power, which by heating with anacid becomes “invert sugar” (a mixture of equalweights of dextrose and hevulose), and dextrose, orlsevulose, or any mixture of them, or any other bodynot liable to have its rotatory power affected byheating with acids. Then the amount of change ofrotatory power will be a measure of the cane sugarpresent in the original solution. If, for example, wchad a mixture of 90 parts of cane sugar and 10 partsof dextrose, we should find on examining it aspreviously described, say with a Soleil-Duboscqpolarimeter, that the direct reading gave 97’3 percent, of sugar instead of 90, because the dextroselikewise rotated the ray to the right 7°-8 of the scale.On inversion the reading would be — 25°'5, supposingthe observation to be made at 15° C. The changeproduced by heating with acid was here from 97°3on the right to 2o°-5 on the left, or a total of122°-8. Now, 100 parts of pure cane sugar sotreated give a change of 136° p 5 at this temperature,
and = 90°, the actual percentage of
136-5
sugar present. As the left-handed rotatory powerof “invert sugar” decreases rapidly with the tem-perature (about of Soleil’s scale, for each 1° C.),it is necessary to take exact account of the tempera-ture at which the reading is made, and then torefer to Clerget’s table, which will be found inmost French and German works on the manufactureof beet sugar.
Estimation of Glucose. —The only method forestimating glucose in presence of cane sugar whichseems to be sufficiently accurate for use is that com-monly known as Fehi.ing’s, the best way of workingwhich is found by the writer to be this :—Use thesolution of sugar employed for polarizing, after re-moving all lead from the solution , and diluting to suchan extent that the liquid does not contain morethan 0 - 5 per cent, of the glucose. To remove thelead, measure off such a quantity of the liquid usedfor polarizing as experience shows will contain notmore than 0-5 gram, or less than 0-2 gram, glucose ;put it in a 100 c.c. flask, add solution of sulphurousacid till all lead is precipitated (if lead remains insolution, the results will come out much too low),add a little washed alumina, fill up to the markwith water, and filter. Fill the burette with thissolution. Measure off with a pipette 5 or 10 c.c. ofthe “copper solution” into a test tube, lj to 1^inch wide, and slightly blown out at the bottom, sothat it shall be less liable to crack; dilute withabout 3 volumes of water; put in a few fragmentsof clay pipe-stem to prevent bumping, and heat toboiling, ltun in the sugar solution from the burette,2 c.c. at a time, boiling up between each addition,till the blue colour has nearly disappeared; thenmore slowly, till on allowing the red cuprous oxide tosettle, the liquid above seems colourless, or very faintlyinclined to yellow, when the tube is held oppositea white surface (a distant white-washed wall is best).Worked thus the method gives very constant and
accurate results, providing the copper solution issound and the whole operation is quickly finished.As cane sugar likewise reduces the copper solutionwhen long boiled with it, this last condition isobviously important.
As 0-5 gram, glucose (either dextrose or tevulose)will reduce the copper contained in 100 c.c. of theFeeling’s solution, and as the above experimentgives the weight of the material containing 0-05gram, glucose, the method of calculating the per-centage of glucose is plain.*
Estimation of Water. —Weigh off 5 grams, of thesample (if it be a granular sugar) into a flat-bottomed platinum pan of good size (say with 3square inches of bottom), spread it about evenly,and set it in an air-bath, heated to about 60° C.(140° Falir.) in the case of cane sugars, or to 100° C.(212° Fabr.) in the case of “beets,” and weigh atintervals of two hours till the weight is constant.The loss calculate as water. Twelve hours will berequired for low “ canes.”
The estimation of water in low syrups—treacle,beet and cane juice, &c.—presents practical diffi-culties, owing to a skin forming over the surface,which prevents further drying at any moderatetemperature. It is best therefore to proceed in oneof the two following ways : —Weigh off 5 grams,of the sample, dissolve in water, and make up to100 c.c. Of this solution take 10 c.c., and pour itover about 10 to 15 grams, of clean silver sand,which has been ignited and tared in a flat platinumdish ; dry off at a temperature not exceeding 60° C.(140° Fahr.) till the whole ceases to lose weight onfurther heating. A small quantity being spreadover a large surface dries readily.
The drying takes place with very great readiness,and at a lower temperature, if the pan be heated ina pretty good air-pump vacuum. For this purposea convenient and cheap arrangement consists of awide glass tube, run through an outer metal casewhich will hold water, and communicating at one endwith a chloride of calcium tube and an air-pump,and which can be closed by a good rubber plug atthe other. The pans here require to be long andnarrow. In this apparatus the drying of half agram, of syrup may be finished in a few hours.
Estimation of Ash. —Take the same or a similarportion of sugar; moisten it all over with the leastpossible quantity of water, and then with pure con-centrated oil of vitriol; heat gently till the frothinghas ceased, and the mass has become converted intoa black cinder; and then introduce it into a muffleheated to very low redness (four ordinary Bunsengas burners will heat a good-sized muffle to fullredness, if the latter is so set in bricks that there isonly about half an inch space between the two).When the carbon is about half burnt away, takeout the pan and moisten its contents a second time
* The original statement of Fehi.ing that the moleculeglucose will reduce ten equivalents of cupric oxide, has beenfollowed in the text; but the writer’s own experience is thatwhen the process of analysis is worked in the way described,the quantity of glucose required to fully reduce the copperoxide is two per cent, greater than there assumed.