T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
596
pletely hydrate t h e magnesium oxide and t h e basic silicates, aluminates, and ferrites derived from t h e combinations of the basic and acidic oxides. The results reported in t h e first paper were summarized as follows: I-That magnesite is not completely dissociated at 500' C. in I hr. under the conditions used in the experiments. 2-That dissociation of the magnesium carbonate is complete a t 600' C., while that of calcium carbonate is not. 3-That the hydration of magnesium oxide burned at 600°, 7o0°, o r 800" C. is practically complete in 3 days. 4-That between 800' and goo' C. the calcium carbonate is dissociated, and that combination takes place between basic and acidic oxides, resulting in the formation of silicates or aluminates. The silicates or aluminates so formed combined with more water than would be required for the complete hydration of the basic oxides alone. 5-That a change in the constitution of the magnesium oxide C., resulting in a marked sets in between 1000' and 1 1 0 0 ~ decrease in the rate of hydration, and that this change becomes more marked with rise of temperature of burning, until at 1450' C., or nearly the temperature required for burning Portland cement, the magnesium oxide after 18 mos. immersion in water has combined with only 61.4 per cent of the water required for complete hydration.
/ZOO"
c L70
f\
I
I
I
I
I
I
Vol.
IO,
No. 8
a slight, b u t steady, increase in weight of t h e samples, due t o absorption of carbon dioxide. This increase in weight amounted in t h e course of 4 yrs. t o about 3 per cent i n the cases of all those samples which had been completely hydrated. Further increase in weight due t o absorption of carbon dioxide after 4 yrs. was prevented by replacing t h e distilled water in t h e desiccator with a dilute solution of potassium hydroxide. The total percentage gain of weight of the samples burned a t g o o o C. or above a t t h e end of each year between I and 6 yrs. is given i n Table I . TABLEI-PERCENTAGE GAIN O F WEIGHT Time of Hydration , Temperature of Burning-Years 900° 1000a llOOo 1200O 1300' 1400° 1 44.88 44.92 45.33 41.34 28.99 28.12 2 45.25 45.18 46.20 42.75 32.75 32.05 45.13 34.43 3 46.27 46.27 47.25 34.08 4 47.88 47.45 47.90 47.33 36.90 35.26 5 47.50 47.49 47.77 47.61 37.36 35.94 47.54 47.77 6 47.33 47.67 37.71 36.58
1
1450' 23.31 26.31 27.97 30.25 30.99 31.82
Since all samples burned a t or bdow 1100' C. were completely hydrated within 3 mos., t h e percentage of complete hydration after long time periods has been. computed only for those samples burned at or above 1200' C. I n computing t h e percentages of complete hydration of these latter samples, correction has been made for t h e increase in weight due t o t h e slight absorption of carbon dioxide. T h e percentages of complete hydration of these latter samples computed in this way are given in Table 11. TABLB11-PERCENTAGEOF COMPLETE HYDRATION Temperature ----Time of Hydration in Years of Burning Degrees 1 2 3 4 5 6 1200 91,8 94.9 100 0 1300 64.0 72.4 76:2 si:7 8i:7 si:, 1400 62.0 70.8 75.3 78.0 79.5 81.0 1450 51.2 58.0 61.7 66.8 66.9 70.3
The results given in Table I1 are shown graphically in Fig. I, in which the ordinates give the percentage of total hydration and t h e abscissae t h e length of time t h e samples were kept in water. These results show clearly why materials containing fret magnesium oxide, if burned a t temperatures approaching t h a t used for t h e production of Portland cement, will not become completely hydrated, even when continuously immersed in water, until the lapse of probably 2 0 yrs. or more. FIG.I-CURVES SHOWING THE INFLUENCE OF BURNING AT TEMPERATURESABOVE 1000° C. ON THE RATEOF HYDRATION OF MgO
The object of this second paper is t o record t h e results obtained after continuing t h e hydration of t h e samples described in t h e first paper up t o a period of '6 yrs. A study of t h e d a t a reported in t h e first paper shows t h a t all samples burned a t temperatures not exceeding 1100' C. were completely hydrated within 3 mos., very slow hydration taking place only in t h e case of samples burned a t 1200' C. or above. During t h e first 4 yrs. t h e desiccator, in which were placed the crucibles with t h e samples just covered with water, was partially filled with distilled water, b u t it was noted a t the end of t h e 4 yrs. t h a t there had been
CHEMICAL LABORATORY UNIVERSITY OF MICHIGAN ANN ARBOR,MICH.
THE DETERMINATION OF PHTHALIC ANHYDRIDE IN CRUDE PHTHALIC ACID By CHARLESR. DOWNSAND CHAR~ES G. STuap Received February 28, 1918
I n connection with t h e control of a plant producing phthalic anhydride it became necessary t o develop a method whereby t h e crude phthalic acid, containing mineral impurities, sulfur compounds, and other organic acids, might be assayed for t h e amount of phthalic anhydride present. A search of t h e literature was made, b u t t h e methods described, with t h e exception of t h a t given by BOSwell (noted below), were not applicable t o t h e crude acid resulting from t h e manufacture of phthalic acid.
Aug., 1918
*
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Several attempts were made t o sublime a mixt u r e of pure phthalic acid and anhydride, in such ordinary laboratory subliming apparatus as is used in investigating other materials. The apparatus was made up of heated dishes, inverted funnels, etc., with t h e object of weighing t h e sublimate, t h u s ob' taining quantitative results. I n most of these trials a current of air was passed over the subliming surface, heated t o a temperature of zoo0 t o 2 2 0 ' C. Various types of condensing chambers were used. I n all cases i t was found t h a t either t h e sublimed . crystals, built up on t h e condensing surface, fell back from their own weight onto a hot portion of the apparatus and remelted, or else they were carried through into the collecting chamber in the form of a fine suspended dust, which was hard t o collect quantitatively. A detachable tube appaB ratus was then made up as shown in Fig. I. A weighed amount of t h e phthalic acid t o be tested was placed in t h e weighed glass capsule D a n d the ' t u b e C was inserted, as per sketch, after i t came t o cons t a n t weight in a steam oven. Air was drawn through t h e tube C by connecting t o a suction p u m p a t the point B, t h e air entering the apparatus through t h e annular space E. The apparatus was then immersed t o t h e point A in a low-meltidg metallic bath kept a t a tempera t u r e from 2 0 0 " to 2 2 0 ' C. I Any vapors formed during sublimation were drawn FIG.I through the cotton and held there instead of escaping through t h e annular space E. Using this apparatus i t was expected t h a t t h e weight of t h e residue could be obtained, t h a t this residue would be in a convenient form for testing for t h e presence of unsublimed phthalic left behind, and t h a t after drying t h e cotton t u b e t o constant weight, i . e . , when t h e water of decomposition had been evaporated, t h e weight of t h e sublimed material or phthalic anhydride would be given.
W
RESULTSOBTAINED ON A SAMPLEOB CRUDEPHTHALIC TemTime Weieht Per perature Immerof cent of Per sion Sample Subli- cent of of Bath C. Hours Gram mate Residue No. 1.. * . 200-220 0.2574 72.9 10.4 2 200-220 0.2453 72.0 10.2
. ......... .... .. .... ......
;$*
The difference of 0 .2 per cent in t h e weight of t h e residue represents a n actual w.eight in t h e case of only 0.0005 g., which shows t h a t practically all of t h e sublimable material was off after '/z hr. immersion. The speed of the air drawn through was ap.proximately 3 bubbles per sec. from a 1/4 in. glass tube through a 3-in. layer of water in a Woulff bot-
597
tle. This rate of air was found t o be satisfactory. A higher rate tendkd t o carry some phthalic through the cotton plug. I n connection with weighing t h e sublimate in t h e cotton tube, t h e following tests were made on pure "commercial" anhydride. Weighed portions of the anhydride were placed on watch glasses, 'moistened with water, and placed i n an oven. Temperature of Oven
c.
88 88
100
Time of Heating Hours 1.5 18 3
Weight Taken Gram
Loss Per cent
0.2500 0.2500 0.2500
97 96
24
From these results i t was evident t h a t a direct determination by weighing t h e sublimate, after reaching constant weight in a steam oven, could not be used as a method of analysis. Colorimetric tests were attempted on t h e residue, using phenol or resorcinol with a dehydrating agent, but dark green solutions were obtained t h a t did not a t all resemble the colors of phenolphthalein or fluorescein. I n fact they were so dark t h a t small amounts of phthalic could not be detected. Attempts were also made t o determine phthalic anhydride present in t h e crude phthalic acid b y colorimetric tests, b u t this method was found t o be inapplicable. Attention was then turned t o the titration method described by C. Boswelll and this method was used with t h e substitution of t h e detachable tube described above. The details of t h e method finally adopted b y us are given as follows: o 250 g. of the sample to be analyzed shall be weighed into the glass capsule, and I . 5 g. of the prepared cotton boiled in I O per cent NaOH and then washed and dried, shall be packed not too tightly in the inner tube. The length of this tube shall be about 5l/2 in. and the column of cotton shall extend to within 3/4 in. of the bottom. The tube itself shall be inserted into the capsule to within l / 2 in. of the bottom of the latter. Suction is then applied to the top of the inner tube and air drawn through as described before, so that it bubbles through a 1/4 in. glass tube in a suction bottle a t the rate of three bubbles per second. This apparatus shall be then transEerred to the heating bath of melted Rose metal and adjusted so that the capsule is immersed to a depth of about l / 4 in. It is important that the apparatus should be assembled and the air be passing through it, and that the bath temperature be regulated to zooo to 220' C. before the immersion occurs. Only in this way can accuracy of results be assured. The heating shall then be continued 45 min., during which time the phthalic acid is completely decomposed and the sublimed anhydride collected in the cotton tower. The tube is then removed from the bath, the outside of the capsule cleaned of adhering metal and the weight of residue determined. The cotton plug containing the sublimate is pushed out into a beaker containing 45 cc. of standard N / I ONaOH. About 50 cc. of water are added and the solution boiled 1 / 2 hr. In case some anhydride adheres to the glass tube, the latter may be left in the caustic solution during the boiling. After I/$ hr. the anhydride is completely dissolved in the alkali, and a small amount of solid phenolphthalein (alcoholic solution of phenolphthalein cannot be used) shall then be added for an indicator and a known excess of standard acid added. This excess must be a t least 5 cc. Boiling shall then be continued 15 min. longer and the titration completed by adding alkali in the hot until the end-point is reached. The 1
J . Am. Chem. SOC., 29 (1907), 235.
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
598
phthalic anhydride titer of an exactly N / I O NaOH solution
is 0.0074.
It will be found t h a t t h e cotton plugs can be used again and again provided they be washed carefully with neutralized water after each run. Occasional titrations must be made on t h e cotto4 blanks t o .be sure of their uniformity and their acid equivalent. New cotton must be thoroughly boiled with N / I O caustic soda and washed well with neutral water. Where there are a number of determinations t o be run daily it will be found t h a t this method using the detachable tube considerably simplifies the method as described by Boswell and in addition the weight of t h e unsublimable residue is obtained. This is a n important feature in connection with plant control. I t is important t o note t h a t when a crude phthalic is obtained containing sulfuric and sulfurous acids it must be washed free of these compounds before testing. The presence of sulfuric acid in t h e crude is generally indicated by a charring of t h e cott o n and, if this happens, t h e test must be repeated. This method has been thoroughly checked b y analyses of pure phthalic anhydride. A further proof t h a t this method is correct, when applied t o crude phthalic acid, is t h a t t h e actual sublimation in t h e plant, where there are small known losses, has given, a t t h e lowest, 95 per cent of t h e analytical figure for phthalic anhydride as determined b y the above method. RBSEARCH. DEPARTMSNT
THEBARRETT COMPANY, 17 BATTERY PLACE NEWYORKCITY
AN IMPROVED DISTILLATION METHOD FOR THE DETERMINATION OF WATER IN SOAP By RALPHHART Received M a y 20, 1918
Water in soap is usually calculated after determining all the other ingredients. I n many instances, however, it is found directly by drying the soap t o constant weight in an oven at 105' C. This procedure, of course, takes considerable time and t h e result indicates not only water b u t also any other volatile constituents t h a t may be present. A quicker method%ist o heat the sample over a free flame until a watch glass held over i t for a moment shows no condensation of vapor, or until t h e odor of acrolein is just noted. This method, evidently, depends a good deal upon the personal equation. A method occasionally employed and originally suggested by Marcusson' is t o distill with xylene. The distillate containing t h e water is received in a special flask having a graduated leg in which t h e water settles, and the reading of t h e lower layer is taken as t h e water content of t h e sample. I n t h e case of soaps containing alcohol OT ammonia, these are found partly in t h e water layer; corrections are made by taking t h e specific gravity in t h e case of alcohol, or b y titrating with standard acid in t h e case of ammonia. The distillation method has been employed quite 1
Mitt. k. Material$uUfungsaml. 23 (1905), 58.
Vol.
IO,
Yo. 8
successfully for the determination of moisture in such materials as foods, oils, tars,l creosoted wood,2 etc. A comparison of this method with t h a t of drying t o constant weight is given in t h e following table taken from S. S. Sadtler:3 Distillation Tests Per cent Analyses I I1 Egg Albumen 15.90 15.35 Cheese.. 29.75 29.90 Butter.. . . . . . . . . . . . . .1 1 , 4 8 11.88 Linseed Meal A , . . . . . . 5.90 5.90 Linseed Meal B . . . . . . . 12.00 11.80 Sawdust . . . . . . . . . . . . . . 1 7 . 2 0 .,.
......... .............
~~
Oven Tests at loo0 C. Per cent I TT 15.70 29.25 29.51 12.25 13.22 5.83 5.72 9.85 16.50 ...
__ ... ...
I n t h e case of soaps, however, the distillation method gives considerable trouble on account of excessive foaming during t h e heating. The operation becomes tedious and slow, and considerable care is necessary t o prevent t h e foam from entering the condenser tube. Another objection t o this method is the very viscous condition of t h e xylene-soap solution towards t h e end of t h e distillation, thereby hindering the free escape of the vapor. The solution a t the end of the distillation usually gelatinizes on cooling. These obstacles, t h e writer found, are very satisfactorily overcome b y the addition of red oil or oleic acid before distilliqg. T h e foaming is entirely eliminated and the solution remains very fluid even a t low temperatures. The addition of red oil presents another advantage in t h a t t h e xylene-red oil mixture is a much better solvent for soap than xylene alone; under like conditions it takes considerably less time t o dissolve a sample of soap in the mixture t h a n in pure xylene. The quantity of- red oil required is about t h e same as t h e weight of t h e soap taken for analysis. I n t h a t case, the soap is quickly and completely dissolved b y t h e xylene, and t h e foaming is entirely eliminated, allowing t h e distillation t o be carried out a t any desired rate. The results compare favorably with t h e oven method for soap as recommended by t h e U. S. Bureau of standard^.^ The two methods are compared in the following table: Distillation Test with Red Oil Per cent Soft Soap 42.5 Degumming S o a p . . 69.5 Fulling Soap ...................... 16.2 3 cc. 0 5 N NaOH 96.4 (e) Theoretical.
........................
............... .................
Oven Test at 105' C , Per cent 42.7 70.6 17.1 98.O(a)
Miche15 finds i t necessary in the determination of water in foods t o apply a coqrection of 0 . 1 2 5 cc. t o t h e water reading t o allow for t h e shape of t h e meniscus. This correction is not necessary in soap analysis as the meniscus is very nearly flat, probably due t o traces of soap mechanically carried over; for it was noticed t h a t t h e meniscus between fresh water and xylene became flat on adding a few drops of a liquid soap. However, a correction of 0.05 cc. may be allowed for moisture left in t h e condenser, since the results in t h e table are uniformly lower t h a n by t h e oven test. J . A m . Chem. SOC.,25 (1903), 814. U. S. Dept. Agr., Forest Service, Circ. 134 (1908). 8 THIS JOURNAL, 2 (1910). 66, 4 CircuZar 62 (1916), 21. s J . SOC.Chem. I n d . , 32 (19131, 445. 1
2
,
