28
T H E J O U R N A L O F I i V D U S T R I A L AiVD E N G I N E E R I N G C H E M I S T R Y TABLE TI-EFFECT O F BARIUMS U L F A T G
.4ND S I L I C A
Results in Percentages of Sulfur on the Dry Basis Sample KO.5--Arminius ore No. 6-Rio Tinto hlETIioD Determinations Av. Determinations Allen and Bishop 39.51 N E W FUSION METHOD: 39.56 39.41 39.20 '47.12 46.95 4 7 . 3 4 ' Unmodified 1 3 9 . 8 2 39.72 3 9 . 2 7 1 39.55 1 4 7 . 3 2 47.08 47.151. ( 4 7 . 8 6 47.17 4 7 . 1 9 ) /39.7039.85 . . . ! Added 0.2 g. BaSOc.. . . , 39.76 3 9 . 5 5 39.65 Added 0.2 g. SiO?.. . . , . . 39.64 39.23 39.43 Added 0.2 g. BaSOa and 0.2 g . SiOz., . , . . . . 4 7 . 3 4 47.40
ore Av. 47.45 47.36
... 4i:37
From Table I1 it would appear t h a t a n y barium sulfate present in t h e ore and decomposed b y t h e sodium peroxide would be precipitated b y the addition of hydrochloric acid. It is generally claimed t h a t the precipitation of barium sulfate under such conditions is incomplete b u t t h e results in Table I 1 shorn t h a t the addition of as much as 2 . 7 j per cent sulfur, as barium sulfate, and silica amounting t o about 13 per cent, has no appreciable influence on the results. T h e results in Table I 1 further show t h a t this method would serve a s a rapid approximate method for determining sulfur in pyrites and with some care and s t u d y would doubtless furnish quite accurate results. T h e effect of barium sulfate and silica was studied on samples of pyrites instead of cinder since these two samples had been very carefully standardized in other work. Doubtless such a method as this would meet admirably t h e needs of those who h a r e t o make b u t a n occasional determination of sulfur in cinder. It has been in use in our laboratory for about a year and has been found t o be entirely satisfactory. Prior to the adoption of the method as described above, we used a similar one for about two years. By t h e old method, solution of the sample was effected in t h e same manner b u t the iron was precipitated with ammonia and t h e precipitation of barium sulfate made in a hot solution. AEMOWR FERTILIZER \170RKS, ATLAKTA, GEORGIA
THE DETERMINATEON OF SUCROSE IN CONDENSED MILK' By G. W. KNIGHTA N D G. F O R M . 4 6 B K Received July 1, 1915
Agricultural and dairy chemists2 usually estimate sucrose in condensed milk b y deducting t h e milk solids from the total solids. This is unquestionably t h e simplest, a n d probably t h e most satisfactory way of estimating the quantity of cane sugar present. When sucrose is t h e only constituent t h a t one needs t o determine, however, it is necessary t o have resource t o some other method. W O R K O F P R E V I O U S IXVESTIGATORS
As early as 1893 Bigelow a n d RlcElroy3 published a biological method in which t h e sucrose is inverted b y fermentation with compressed yeast at j5' C. for five hours a n d t h e percentage calculated by Clerget's formula from the polariscopic readings before a n d after inversion. The milk proteins are precipitated b y an acetic acid solution of mercuric potassium iodide. 1 Contributed with the permission oE the Secretary of the Treasury a n d the U. S. Appraiser, Port of New Pork. 2 Bureau of Chemistry, BuU. 107 (Rev-.), 123. a J . A m Chem. Sac., 15, 668.
Vol. 8 . N o .
I
If invert sugar is present in appreciable amount t h e procedure is modified by conducting the fermentation for ten days a t a temperature of zj' t o 30' C. in t h e presence of potassium fluoride. Wiley claims t h a t it is fair t o assume t h a t no invert sugar is present in condensed milk. I n t h e samples we ha1.e analyzed t o determine this point we have been unable t o detect enough invert sugar t o materially affect the result. Baker and CHulton,' working with dilute sugar solutions of from 0.3 t o 3.0 per cent strength, have tested this method of fermentation with lactose alone and with lactose mixed with other fermentable sugars, allowing the fermentation to proceed for 60 to 70 hours a t a temperature of 2 7 ' C. The unfermented sugar, lactose, was determined volumetrically with Fehling's solution. They obtained low results in most cases, the difference between the amount of lactose known t o be present and t h a t found varying u p t o I O per cent of the lactose present. Other investigators have noticed similar losses of the lactose in attempting t o determine sugars by this fermentation method. Inasmuch as there is always danger of fermentation of the lactose due t o the possible presence of bacteria and symbiosis between t h e yeast a n d bacteria, one is never certain of obtaining reliable and accurate results b y a n y fermentation method. NowakZ has recently worked u p a n interesting method proposed b y Jolles3 which depends on the f a c t t h a t all mono- and disaccharides, with t h e exception of sucrose, are rendered optically inactive when heated under pressure in a steam b a t h for 4 5 minutes with 16 per cent sodium hydroxide solution. After destroying the lactose in this way, the amount of sucrose present is determined b y the polariscope. T h e solution polarized is so dilute t h a t the error of reading the instrument makes t h e final result uncertain t o * I or z per cent, and as Jolles particularly states t h a t t h e amount of sugars other t h a n sucrose present in t h e solution treated with t h e sodium hydroxide must not exceed z per cent, it would appear t h a t this method could not be made t o give the desired degree of accuracy. The A. 0. A. C.4 did some cooperative work a few years ago on two gravimetric methods in which t h e milk was prepared and clarified and t h e reducing sugars before inversion determined in the same manner as described in their provisional method5 for lactose in milk. I n one method t h e sucrose was inverted b y heating the solution on a water b a t h for 40 minutes with 4 g. of citric acid in t h e manner proposed b y Stokes and Bodmer;G i n ' t h e other method the sucrose was inverted in the cold with concentrated HC1 as described in t h e official method' of t h e A. 0. A. C. for t h e determination of sucrose in foods. After inversion the invert sugar was determined gravimetrically with :AnaEysl, 36, 512.
2. anal. Chem., 5 1 (1912), 610. Z . A'alw. Genussm., 20 (1910), 631. 4 Proc. of A . 0 . A . C., 1909, Bur. of Chem., Bull. 131, 171 6 Bur. of Chem., Bull. 107 (Rev.), 123. 6 Analyst, 10, 10. 7 Bur. of Chem., BnU. 107 (Rev.), 41. 2 3
Jan., 1916
T H E J O G R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G CHE,WISTRY
Fehling’s solution. The average percentage obtained from nine chemists b y t h e first method was 42.27 per cent, b y t h e second method 42.95 per cent. T h e percentage of sucrose obtained b y subtracting t h e average per cent milk solids from t h e average per cent total solids was 42.98. T h e agreement between t h e last two results is remarkable. T h e highest result obtained b y t h e first method was 44.36 a n d t h e lowest 39.90 per cent. T h e highest obtained by t h e second method was 44.30 a n d t h e lowest 41.23 per cent. Although this is a rather wide variation, t h e other determinations show a corresponding proportional variation. These discrepancies are undoubtedly due t o lack of ‘homogeneity in t h e condensed milk analyzeda matter which will be discussed later. From t h e results obtained in this cooperative work, i t would appear t h a t t h e second method is a reliable one; b u t a polariscopic method is preferable where a large number of samples have t o be analyzed, owing t o t h e greater rapidity a n d ease of manipulation of polariscopic methods. Of t h e polariscopic methods published, Cochran’ has described one in which he utilizes Wiley’s acid mercuric nitrate solution both a s clarifying and inverting agent. This method is based on his discovery t h a t this acid mercuric nitrate solution completely inverts sucrose when heated with i t on a water b a t h for eight minutes, while a t temperatures below 15’ C. it inverts very slowly. Quite recently Revis a n d Payne2 have published a polariscopic method for t h e determination of lactose a n d sucrose in condensed milk, in which they use acid mercuric nitrate solution similarly for clarifying a n d inverting t h e sucrose. Their formulas for calculating t h e percentages of t h e sugars include correction factors for t h e change in concentration caused by t h e volume of fat a n d protein present which involve a determination of t h e percentages of these substances. T o show t h e inadequacy of acid mercuric nitrate solution as a clarifying agent t o remove t h e proteins occurring in condensed milk, Richmond3 has published some interesting d a t a . He found t h a t on adding phosphotungstic acid t o a solution of milk, clarified b y acid mercuric nitrate, a voluminous precipitate was formed. F u r t h e r , on analyzing whey powders containing about 7 0 per cent of lactose, a n error amounting t o 4 per cent occurred, owing t o t h e presence of protein unremoved b y t h e acid mercuric nitrate solution. He recommends using phosphotungstic acid after t h e milk has been partially clarified with t h e acid mercuric nitrate solution. We have used phosphotungstic acid in conjunction with lead acetate for a number of years, t o remove troublesome proteins in making polariscopic determinations of sugars in meat extracts, milk preparations a n d galenicals, with very satisfactory results; t h e polariscopic determinations check very well with gravimetric determinations. Chemists familiar with commercial condensed milk know t h a t there is a tendency for lactose crystals t o 1 2
J . Am. Chem. Soc.. 29 (1907), 5 5 5 Analyst. 39, 476 I b d , 36, 516
29
collect a n d pack on t h e bottom of t h e can, a n d it is difficult t o obtain a representative sample of t h e milk for analysis. T h e official organ’ of the A. 0. A. C. recommends t h a t t h e sample be mixed thoroughly b y transferring t h e contents of t h e can t o a large evaporating dish a n d stirring it with a pestle until homogeneous. Bache2 has lately published analyses showing t h e inaccuracies t h a t may be caused b y failure t o obtain a sample representative of t h e contents of t h e can, through lack of homogeneity of t h e different kinds of milk. The greatest difficulty was encountered in t h e cases of milk having a thick consistency and t h i n milk t h a t had stood for a long time. He showed t h a t t h e only way t o obtain a n absolutely reliable sample was t o dilute t h e milk with water before analyzing. Bache states t h a t possibly a separation of t h e constituents of t h e condensed milk also takes place in t h e process of manufacture, and t h a t may explain t h e fact t h a t tins from t h e same manufacturer do not always analyze t h e same. The writers have sometimes obtained a higher percentage of sucrose t h a n t h a t p u t in b y t h e manufacturer; b u t generally t h e y obtained slightly less. These results could be explained only b y Bache’s theory. I n t h e manufacture of condensed milk, as in t h e manufacture of certain kinds of confectionery, there undoubtedly occurs a slight destruction of sucrose due t o contact with t h e steamheated surfaces used in evaporation, a n d this would account for slightly lower results. D I S C U S SI 0 N O F G E K E R A L C 0 pi S I D ER.4 TI0 NS
The lactose crystallized out in t h e cans of condensed milk evidently has assumed t h e /3 form. T h e official method of t h e A. 0. A. ( 2 . 3 specifically mentions t h e fact t h a t i t is not necessary t o heat t h e solution before polarizing .in t h e determination of lactose in milk. The writers have never noticed a n y evidence of mutarotation in a n y of t h e samples t h e y have analyzed t o determine this point. Invertase is unquestionably t h e most ideal inverting agent t o use in general work as it possesses a more or less selective inverting action, a n d inverts very few substances other t h a n sucrose. Hudson4 has worked out a method for determining cane sugar, using invertase. For practical analytical work, however, t h e use of invertase has many drawbacks. Among these are t h e time required a n d t h e inconvenience of preparing a n d preserving t h e solution; a n d t h e uncertainty as t o t h e strength of t h e solution, which necessitates t h e occasional running of blanks of known sugar content t o determine t h e inverting power of t h e solution, unless a new supply is made every few months. T h e use of organic acids also has serious drawbacks. Tolman6 has shown t h a t in the presence of soluble acetates, as when lead acetate is used t o clarify, t h e inverting action of citric acid is greatly retarded. The results-obtained in t h e cooperative work of t h e A. 0. A. C. mentioned above tend t o show incomplete in1 2
4 9
Bur. of Chem., Bull. 107 (Rev.), 122. Analyst, 36, 138. Bur. of Chem., Bull. 107 (Rev.), 119
THISJOURNAL, 2, 143. Bur. of Chem.. Bull. 73, 69.
T H E J O U R N A L O F I N D U S T R I A L A N D ENGIhrEERING C H E M I S T R Y
30
version of the sucrose b y t h e use of oxalic acid in the analysis of condensed milk, although pure sucrose treated similarly is completely inverted. Many reasons have been given for this lowering of inverting action. Perhaps t h e best explanation is t h a t offered b y t h e dissociation theory. Concentrated hydrochloric acid, acting in the cold a s described in t h e official methods of the A. 0. A. C., inverts sucrose completely except where a n unusual amount of lead acetate has been used t o clarify t h e solution. T h e writers have made a number of analyses of lactose under similar conditions a n d have been unable t o discover enough inversion of t h e lactose t o affect t h e polariscopic reading for a n y concentration. T h e invert reading should be made in neutral solution t o eliminate inaccuracies caused b y the effect of free acid on t h e polarization of levulose for different concentrations. The Clerget factor for neutral solutions for a concentration of 1 3 grams of sucrose in 100cc. is 141.7, the same a s t h a t determined by Hudson for invertase inversion. For lower concentrations the factor is slightly different, owing t o change in t h e specific rotation of invert sugar for different concentrations as shown b y Gubbe,' b u t t h e polarization is less so t h a t t h e difference obtained by using the more correct factor is less t h a n the error of reading t h e instrument. Unless t h e per cent of fat and protein in the sample of condensed milk has been determined, the double dilution method of Scheibler2 should be used t o correct for the error due t o t h e volume occupied b y these substances a n d a n y precipitate caused b y clarification, since the percentages of these substances vary over wide enough ranges t o seriously affect t h e accuracy of t h e determination of sucrose. EXPERIMENTAL WORK
A solution of condensed milk t h a t has been completely clarified b y t h e use of phosphotungstic acid and lead acetate does not show a n y trace of protein with picric acid or any of t h e other more delicate protein precipitants. Test analyses made with pure sucrose using t h e same quantities of phosphotungstic acid, lead acetate a n d potassium oxalate as t h a t found necessary t o thoroughly clarify t h e condensed milk showed t h a t t h e free acetic acid, acetates, and oxalates left in t h e solution do not reduce the concentration of the hydrogen ions sufficiently t o prevent thorough inversion of t h e sucrose with the amount of hydrochloric acid used. I t was found t h a t 0 . 5 gram of potassium oxalate was generally sufficient completely t o de-lead t h e solutions when the specified quantity of lead acetate solution was used, and t h a t a large excess should not be used, since it seems t o redissolve the lead precipitate and cause a milky filtrate, in addition t o causing errors b y diluting t h e solution. I n order t o determine a n y error due t o change in volume caused b y the addition of the potassium oxalate used in de-leading, 0.j gram of potassium oxalate was dissolved in IOO cc. of water and t h e increase in volume measured. The solution of 0 . 5 gram of potassium 1 2
B e 7 , 18, 2207. 2. Ver. Deut. Zuck., 26, 1054.
Vol. 8,
KO.I
oxalate caused a n increase in volume of 0.2 cc. A filtrate from condensed milk t h a t had been clarified with the required amounts of phosphotungstic acid and lead acetate was de-leaded with 0 . 5 gram of potassium oxalate, t h e precipitate filtered off, washed with water, alcohol, and ether, dried and then added t o I O O cc. of water. The addition of the dried precipitate caused a n increase in volume of 0 . 2 cc. Since t h e increase in volume caused b y the solution of t h e potassium oxalate needed t o de-lead the filtrate is compensated for b y the decrease in volume caused by the space occupied by t h e precipitate, no correction is needed for either source of error. To study the combined effect of the various constituents occurring naturally in the milk a n d those added incidental t o the clarification, de-leading, inverting and neutralizing on t h e specific rotation of t h e sugars and consequently on t h e Clerget factor, some chemically pure sucrose was added t o commercial samples of evaporated unsweetened whole milk, in t h e proportions t h a t it occurs in commercial condensed milk, and t h e Clerget factor calculated after t h e solutions had been treated exactly a s in the method given below. The factor obtained in each case was 141.7. The chemically pure sugar used was carefully prepared, contained no moisture, no ash, and had a n average direct polarization of 99.87, and analyzed 99.85 per cent sucrose b y Clerget, reading the invert solution neutralized. The readings were made a t 2 0 ' C. in standardized tubes of the Bates type, on a Bates t y p e polariscope, u s h g a bichromate cell a n d setting the instrument for maximum sensitiveness. All the apparatus used was carefully standardized. Since Bates and Jackson' have recently shown t h a t t h e 100' S. point is approximately 0.1' S. too high on t h e present types of saccharimeter, this sucrose may be considered pure enough for our purpose. The method given below was tested in the following manner. Four cans of unsweetened evaporated whole milk, dried in a desiccator and weighed full and e m p t y t o obtain the weight of milk in each can, were transferred t o volumetric flasks. The four cans were f r o m t h e same manufacturer and were of t h e same grade milk. The flasks were filled with water and shaken t o dissolve a n y crystallized lactose and make t h e solution homogeneous. An aliquot was evaporated t o dryness t o obtain t h e percentage of total solids. Then, using this percentage of solids a s a basis for calculation, such quantities of the C. P. sucrose mentioned above were added t o three other aliquot portions t h a t t h e respective solutions contained 33.40, 40.00 and 43.60 per cent sucrose of t h e evaporated milk and sucrose contained therein. After t h e sucrose had been thoroughly dissolved, t h e resulting solutions were analyzed for SUcrose b y t h e polariscopic method given below and also gravimetrically b y illunson and Walker's Method,2 using the corrected tables for l a ~ t o s e . ~The solutions clarified in t h e same way as in the polarimetric determination were used for the latter determination. of Standards, BULL. 10, 538. Bur of Chem , Bull. 107 (Rev 1, 241 Bur. of Chem , Czrc 82.
1 Bur 2
Jan., 1916
T H E JOCRNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY
T h e following results were obtained: PER CENTSUCROSE PRESENT ACTUALLY 33.40 40.00 43.60
PER CENT SUCROSE FOUND Polariscopic Gravimetric method method 33.25 33.20 40.24 39.99 43.51 43.56
M E T H OD
After removing t h e label, t h e can of condensed milk is heated for a short time in a 100' oven, cooled in a desiccator, weighed, t h e contents, weighing from 1 2 t o 1 5 ounces, transferred b y ' m e a n s of hot water a n d a suitable funnel t o a 5 0 0 cc. volumetric flask. This may be accomplished most readily b y punching two holes in opposite positions near t h e circumference of t h e t o p of t h e can, inverting t h e can over a funnel in a fixed position and allowing t h e can t o drain till practically e m p t y , t h e n removing t h e t o p a n d washing out t h e remainder. T h e e m p t y can is t h e n dried a n d weighed in t h e same manner as before a n d t h e weight of milk in t h e can ( W ) t h u s determined. T h e cont e n t s of t h e flask are shaken until all crystals of lactose a n d cane sugar are dissolved a n d t h e milk solution homogeneous; then t h e flask is filled t o t h e mark with water, cooled t o room temperature a n d shaken. Two aliquot portions of jo a n d IOO cc. measured in volumetric flasks, are transferred with rinsings t o zoo cc. volumetric flasks a n d clarified b y t h e addition first of I . j cc. of 5 per cent phosphotungstic acid solution for each I O grams of condensed milk contained in t h e aliquot, a n d t h e n after shaking, by t h e addition of 2 . 1 cc. of a 2 j per cent neutral lead acetate solution for each I O grams of condensed milk contained in t h e aliquot portion. T h e flasks are well shaken a n d t h e n made u p t o t h e mark, well shaken again, a n d filtered. T o t h e filtrates measuring about 100 cc. are now added potassium oxalate crystals in portions of 0.1gram a t a time, with constant shaking, until a curdy precipitate forms which quickly settles leaving a clear liquor. Usually 0 . j gram of t h e potassium oxalate is sufficient, a n d a large excess should be avoided. T h e solutions are again filtered, using a hardened filter containing 3 t o j grams Fuller's earth placed in t h e apex a n d testing t h e first I O cc. or so with more potassium oxalate crystals for complete removal of lead, a n d a portion of t h e filtrates polarized at 20' C., preferably on a Bates instrument set for maximum sensitiveness a n d using a bichromate cell. Multiplying t h e reading .of t h e dilute solution b y 4 a n d subtracting t h e reading of t h e stronger solution from t h e product gives t h e direct polarization (P) of t h e solution corrected for t h e volume of precipitate. T w o aliquot portions of 5 0 cc. of t h e filtrates are measured into I O O cc. flasks b y means of pipettes; 5 cc. of concentrated HC1 (38.8 per cent) added t o each, a n d t h e resulting solutions allowed t o s t a n d over night at room temperature. T h e room temperature should not be below 20' C., b u t preferably around 2 j ' C. I n t h e morning a few drops of phenolphthalein are added t o t h e solutions a n d t h e y are t h e n neutralized with strong N a O H solution. A few drops of N / I O HC1 are added t o dispel t h e pink color, t h e solutions are made u p t o t h e mark, cooled t o room temperature
if necessary, a n d t h e n polarized in t h e same manner as before inversion, preferably using 400 mm. polariscope tubes on other instruments t h a n those of t h e Bates type. T h e corrected invert polarization is obtained b y subtracting t h e polarization of t h e strong solution from 4 times t h e polarization of t h e weaker solution, in t h e same manner as before inversion. Multiplying this corrected invert polarization b y 2 , except where 400 mm. tubes have been used, gives (P') t h e invert polarization corresponding t o ( P )t h e polarization before inversion. Substituting these values i n t h e following equation gives t h e per cent of sucrose in t h e condensed milk. 2 6 0 0 0 ( P - P') Per cent sucrose = W (141.7 - T / 2 ) Where W = weight of condensed milk contained in can and T = temperature in degrees centigrade at which invert reading is made. All flasks a n d pipettes used should be carefully standardized for t r u e cubic centimeters at 20' C. U. S. CUSTOXS SERVICE PORT OF NEW Y O R K
THE DETERMINATION O F MOISTURE IN SYRUPS BY THE CALClUM CARBIDE METHOD By R. M. WEST Received June 4, 1915
INTRODUCTION
The determination of moisture in organic, as well as in inorganic, substances is a source of much difficulty. This is due largely t o t h e fact t h a t i n m a n y instances chemical changes, either increasing or decreasing t h e weight of t h e material, t a k e place at t h e temperature necessary for t h e evaporation of t h e moisture a n d , in others, compounds are present which volatilize with t h e water during t h e determination. Attempts have been made t o correct for t h e errors inherent in t h e ordinary drying process b y drying ( I ) i.n Z J U C U O , ( 2 ) in a n atmosphere of neutral gas, ( 3 ) at low temperatures over dehydrating agents, a n d (4) b y distilling t h e moisture from large samples, together with oil, a n d measuring t h e water obtained. None of these modifications, however, are entirely satisfactory. Syrups, fruit juices, substances with a high fat content, a n d those containing other volatile compounds are particularly troublesome. The syrups, after concentration t o a point a t which only five or t e n percent of t h e water remains, become so viscid as practically t o prevent further drying. The addition of sand or pumice, as prescribed by t h e A. 0. A. C. official method,' results more or less satisfactorily, depending upon t h e proportion used a n d t h e original moisture content of t h e syrup. Tables I a n d I1 show quite clearly t h e variation t o which results b y this method are subject. Furthermore, t h e long continued heating t o which it is necessary t o subject a syrup in order t o arrive at nearly constant weight, results in some inversion of t h e sucrose, t h u s decreasing t h e apparent moisture content. This is especially marked when considerable amounts of acid 1 U. S. Department of Agriculture, Bureau of Chemistry, Bull. 107, (revised) 64-65.
