ANALYTICAL EDITION
296
Vol. 2, No. 3
Revised Procedure for the Determination of Uncombined Lime in Portland Cement'" William Lerch a n d R. H. Bogue PORTLAND
I
K A former paper ( 4 )
CEMEST ASSOCIATIOS FELLOWSHIP, 'c. s. B U R ~ AOU B STANDARDS,
\vAS€IINGTOX,
A revised procedure for t h e d e t e r m i n a t i o n of u n -
D C
ammonium acetate in 1 liter
of absolute ethyl alcohol or c o m b i n e d l i m e in P o r t l a n d c e m e n t s b y t h e a m m o n i u m the authors presented anhydrous alcohol denatured a c e t a t e t i t r a t i o n m e t h o d is presented. Various f a c t o r s experimental evidence by Formula 3-a or Formula w h i c h bear on t h e precision of t h e t e s t a r e discussed. which indicated t h a t a 2-b. Standardize this solution A curve is given w h i c h shows t h a t t h e percentage by titrating against pure calmodification of the animocium o x i d e . T h e c a l c i u m average deviation f r o m t h e m e a n varies w i t h t h e m e a n nium acetate titration method oxide should be freshly prevalue for u n c o m b i n e d lime. of Einley 13) was suitable pared by calcining pure calcite The probable e r r o r s have been d e t e r m i n e d a n d a for the d e t e r m i n a t i o n of or calcium oxalate in a platicurve is given w h i c h shows t h e probable e r r o r of a n y num crucible a t 900" t o uncombined liine in Portland 1000" C. to constant weight. g r o u p of d e t e r m i n a t i o n s f r o m t h e m e a n , a t a n y m e a n cement. Ordinary c. P. CaO should not u n c o m b i n e d l i m e c o n t e n t up to 7 per c e n t . Diepschlag and Matting be used. Pulverize a few zrams (z), TvhOin\-estigat'ed a large of the calcite or calcium oialate nuniber of methods for the determination of uncombined to an impalpable powder and heat as before. Weigh about 0.1 gram and transfer to a 200-ml. Erlenmeyer flask containing 60 lime in slags and cement,s, conclude t,hat the aninioniuni ml. of the neutralized glycerol-alcohol solution. Disperse the acetate method is the only method available a t present which calcium oxide in the solution by shaking. Insert a water or gives satisfactory quantitative results. Rathke ( 7 ) agrees air reflux condenser in the neck of the flask, and boil the mixthat the aininoiiium acetate met'liod gives satisfactory results, ture for 20 minutes. Then remove the condenser and immebut suggests that an alcoholic solution of tartaric acid as the diately titrate the solution, while near boiling, x-ith the ammoacetate solution. Replace the condenser and boil the solutitrating solution eliminates the necessity of boiling to yola- nium tion again for 20 minutes. Repeat the titrations a t 20-minute tilize the annnonia liberat'ed during the titration. As boiling intervals. The titration is complete when no further color is required in either case t o bring the linie into solution, appears in the solution during continuous boiling for 1 hour. Then calculate the value of the solution in terms of grams of however! this method was not studied. lime equivalent to each milliliter of ammonium acetate solution. Since the publication of our earlier paper the aninioniuiii Grind the cement to be tested to an impalpable powder. acetat'e inet,hod has been employed extensively in this Iabora- n'eigh a 1-gram sample and transfer it t o a 200-ml. Erlentory and in cement and commercial laboratories fur the meyer flask containing GO ml. of the glycerol-alcohol solution, determination of uncombined lime. The method, n-itli Shake the flask to disperse the cement in the solution. Attach the reflux condenser and boil the mixture on the hot plate. some modifications, has been applied also to the deteriiiina- Titrate, as in the standardization, a t 20-minute intervals, until tion of calcium hydroxide in hydrat'ed cement ( 1 ) . As a no further color appears after boiling for a n hour. In the final result of the experience obtained certain modifications in the titrations, if there is doubt of the exact end point, add a drop of technic of the original procedure which advance t'he precision indicator to the quiescent liquid in the flask, carefully observing point a t which the indicator strikes the solution. If no and reproducibility of the results have been developed. The the color appears the end point has been reached. Then calculate revised procedure is presented a t this time, together with an the uncombined lime content of the cement from the known analysis of data obtained with the method. The procedure lime value of the ammonium acetate solution used. given is for uncombined lime. It does not consider the niodiC o m m e n t s on Procedure fications necessary for the determination of calcium hydroxide THE XoLrrIoss-It is essential that anhydrous alcolioI in hydrated cement. be used in tlie preparation of the glycerol-alcohol solution: Revised Procedure the aiiiiiioiiiuni acet,ate solution, and the plienolpht~haleiii The deterinination is based on the solution of the un- solution. Ordinary "pure" alcohol contaiiis 4 or 5 per cent combined lime in a hot soIution of glycerol and alcohol and of water. Absolute ethyl alcohol or anhydrous alcohoIs the subsequent titration of the dissolved lime with an alco- denatured by Formula 3-a or Formula 2-b of the Bureau of holic solution of aiiimonium acetate in accordance with the Internal Rerenue are satisfactory. The anhydrous alcohols are obtainable through industrial alcohol dealers on tlie reaction same permit requirements as obtain for the less concentrated CaO + 2CHsCOONH4 = (CH3C00)2Ca f H?O 2"s alcohols. If distillation permits are obtained, absolute Prepare a solution consisting of 1 part by volume of U. S. P. alcohol niay be prepared from 96 per cent ethyl alcohol Ly glycerol and 5 parts by volume of absolute ethyl alcohol or anhydrous alcohol denatured in accordance with Formula 3-a the following procedure (6):
+
or Formula 2-b of the Bureau of Internal Revenue. To each liter of this solution add 2 ml. of an indicator prepared by dissolving 1 gram of phenolphthalein in 100 ml. of absolute alcohol. It is essential that the glycerol-alcohol solution be neutral t o the indicator. If the solution is colorless, add a dilute alcoholic solution of sodium or potassium hydroxide until the pink color appears, and just remove this by a drop of an alcoholic solution of ammonium acetate. If the initial color is pink, just remove it by the alcoholic solution of ammonium acetate. A standard alcoholic solution of ammonium acetate, approximately 0.2 N , is required. Dissolve 16 grams of crystalline Received M a y 7, 1930. Publication approved by the Director of the Bureau of Standards, U. S . Department of Commerce. Paper No. 23 of the Portland Cement Association Fellonwhip a t the Bureau of Standards. 1
2
To each liter of 96 per cent ethyl alcohol add about 300 grams of freshly ignited lime. 4110w to stand with occasional shaking for 3 or more days. Decant the clear alcohol and reject the residue. Add 100 grams of freshly ignited lime and again allow to stand with occasional shaking for several days, protecting the alcohol from moist air, Siphon the clear liquid off with careful exclusion of moist air, Boil in a flask with a reflux condenser for 8 hours after adding 25 grams of freshly ignited finely powdered lime, Then distil the alcohol, protecting the distillate from moist air with tubes of calcium chloride. Reject the first and last fractions of about I0 ml. per liter. Keep well stoppered to prevent ingress of moisture.
U. S.P. glycerol is satisfactory but, as this is hygroscopic, care must be taken to insure that it has been well protect'ed
July 15, 1930
ISDUSTRIAL A N D ENGINEERING CHEMISTRY
from the moisture of the air. -1solution containing 1 part of glycerol to 5 partr of alcohol has been found most satisfactory. Although solution of the lime is more rapid when a higher concentration of glycerol is used, experiments have shown that some of the compounds of the cement are dissociated when the proportion of glycerol to alcohol is greater thaii 1 to 3. During the boiling operation some concentration of glycerol may occur owing to evaporation of the alcohol. As a precaution against dissociation by that process, i t appears best to use a solution of 1 part of glycerol to 5 parts of alcohol. GRIXDISGAXD ~~EIGHISG-Forgreater precision in establishing the standard values of the ammonium acetate solution an amount of calcite or calcium oxalate equivalent to about 0.1 gram of lime may be weighed into a platinum crucible, provided with a well-fitting cover. After calcination, ths crucible and contents are weighed, reheated, and rexeiglied rapidly with weights in place to avoid the possibility of any appreciable hydration or carbonation prior to obtaining the weight of lime. Kunierous tests with mixtures of known lime value, however. ha>.-eindicated that the error involved in the operations as given in the procedure are essentially negligible. Likewise, the possibility suggests itself that some hydration and carbonation of the free lime in the clinker may take place during the process of grinding. Mixtures of clinker with lime, however, n-ere found to have almost the theoretical free lime values, whether the lime was ground with the clinker or added directly to the reaction flasks. I t is well to emphasize, nevertheless, that the clinkers always should be freshly ground and placed in the glycerol-alcohol solution directly, with no standing in the laboratory before starting the analysis. ROILISG-It has been found necessary to prevent ingress of air to the flasks during boiling because the lime brought into solution reacts readily n-ith the carbon dioxide to form calcium carbonate. This carbonate does not react TTith the anirnonium acetate during titration, so that the lime results obtained are reduced in proportion to the amount of lime that reacts with the carboii dioxide. I t has been found that the amount of lime carbonated is inappreciable if the solution is kept boiling vigorously. This causes a current of vapor passing upward aiid prevents the ingress of air. Boiling in this manner in an open flask, however, brings about a loss in alcohol, and, by the consequent concentration of the glycerol solution (unless the alcohol is replenished a t frequent intervals), disintegration of the calcium coinpounds of the cement will follow. The use of a reflux condenser and positive, but not violent, ebullition combine to prevent ingress of carbon dioxide and a t the same time avoid loss of alcohol. But even with the use of the condenser, a very slon- boiling will perinit ingress of the carbon dioxide and give low results, while a violent boiling and loss of alcohol will tend to make the glycerol more concentrated and give high results. The condenser may be water cooled or air cooled. A glass tube about 6 inm. in diameter and 45 em. in length, inserted in a stopper and attached to the flask, is satisfactory. If new rubber stoppers are employed, they should be washed to remove the white powder on them before use. Cork stoppers are more satisfactory. If it is necessary to leale the test uncompleted, the condenser should be removed, the solution titrated to the end point, and the flask tightly stoppered. Titrations should not be repeated, on renewing the test, until the solution has been boiled in the usual manner. TITRATIOS-111order to prevent ingress of carbon dioxide, titrations are made iininediately upon reiiioving the condenser and while the solution is still near the boiling point. Samples titrated to completion while hot may develop a pink
297
color on cooling owing to a difference in the end points in hot and cold solutions. Hence, both in standardization and in the determinations, the solutions should always be titrated while hot. The end points are better recognized in bright white light, but not in direct sunlight. The time required for the determinations will vary froiii 3 to 8 or inore hours, depending on the amount of lime present. the fineness of the sample, and the nature of the cement. -1high lime content, a coarsely ground sample, or a high burned product each makes for slower reaction and a longer required boiling period. The time int,erval between tit'rat'ions is best inaintained a t 20 to 30 minutes. \Then boiling is continued for longer periods without titration crystals. prebably calciuin glyceride, are formed. These will go s l o ~ d yinto solution again as t'he lime of the solut,ion is reduced by the aninioniuni acetate, but the total time required for the determination may be iiicreased.
Figure 1-Relation
between Per Cent Average Deviation and Uncombined Lime
The introduction of ammonium acetate beyond the end point of each titration is to be avoided because the ammonium acetate reacts slowly with some compound or compounds other than lime in the cement. An excess of ainmonium acetate, therefore, causes lime, other than that uncombined in the cement, to react, so that high results will be obtained. For this reason. as pointed out by Steinour and Woods (8, it is not perniissible to add a n excess of the anirnonium acetate solution a t the start, and, after a g k e n time, titrate back with alcoholic potassium hydroxide. Results Obtained with Revised Procedure
Supplies of eight cements were obtained. Some were conimercial products and others were so prepared in the laboratory that a \vide range of uncombined lime values might be available. Samples of each cement mere subinitted in sealed bottles to seventeen laboratories, together with a copy of the revised procedure for the determination of uncombined lime. The mean uncombined lime values obtained by these laboratories on the eight cements are given in Talrle I, together with the maxiinuni and minimum values obtained, the average deviation from the mean, the per cent average deviation from the mean, and the probable error. The average deviations from the means are obtained by averaging the actual deviations of each laboratory without regard to sign. They represent the average spread of individual values from the mean for each sample of cement. The percentage average deviation from the mean is obtained b y dividing the average deviation by the inean value for each cement.
ANALYTICAL EDI TIOS
298
Table I-Results w i t h Revised Procedure f o r Uncombined L i m e Obtained on Eight Laboratory a n d Commercial C e m e n t s by S e v e n t e e n Laboratories MEAN MAX. MIN. Av. DEVIATION PROBABLE CEMENTVALUE VALUE VALUE FROM MEAN ERROR
70 7.4 6.6 4.3 2.4 1.6 0.9 0.3 0.1
8.5 7.8 5.3 .3.1 2.2 1.9 0.6 0.5
6 5 3 1 0 0 0
3 7 2 7 9 6 0
0 0
0.5 0 4 0 4 0 3 0.3 0 2 0 1 0 1
6.8 6.1 9.3 12.5 18.7 22.2 33.3
100.0
0 41 0 38 0 0 0 0 0
36 29 28 14 12 0 06
Vol. 2 , s o . 3
mean value for uncombined lime as obtained by severaI laboratories is 2.0 per cent, the chances are equal that a given laboratory will obtain a figure which differs from 2.0 per cent b y not more than 0 . 3 4 . e., 2.0 * 0.3 per cent. Of course, the chances also are equal that any laboratory will obtain a figure which differs from the mean by more than the probable error. Literature Cited Assarsson and Sundius, Swedish Geological Survey Arsboh, 23, KO 2 (1929) ; Personal communication from Work and Lasseter, Columbia University, New York, N. Y . Diepschlag and Matting, Cenlr. Hutten Walewerke, 31, 363 (1927). Emley, Trans. A m . Ceram. Soc., 17, 720 (1915). Lerch and Bogue, IND,EXG.CHEX.,18, 739 (1926). Merriman, “.Method of Least Squares,” p. 66, New York. Osborne. McKelvy, and Bearce, Bur. Standards, Sci. Paper 197 (1913). Rathke, Tonind. Ztg., 32, 1318 (1928). Steinour and Woods, Rock Products, 32 (Xo. 4 ) , 74 (1929).
Use of Tantalum as Cathode for the Electrodeposition of Copper’ B. Mears and P. R. Pine THOMPSON CHEMICAL LABORATORY, WILLIAMS COLLEGE, WILLIAMSTOWI, MASS.
Figure 2-Probable Error of Individual D e t e r m i n a t i o n s of Unc o m b i n e d L i m e f r o m t h e M e a n Value of Several Laboratories
The average deviations froin the means vary from 0.1 to 0.5, and the per cent average deviations vary from 6.1 to 100. By plotting the per cent average deviations against the mean values for uncombined lime, the curve shown in Figure 1 is obtained. The per cent average deviations vary in a regular manner with the mean value for the uncombined lime present. The curve makes it possible to read the most probable value for the per cent average deviation a t any content of uncombined lime up to about 7 per cent. The probable error, defined as that deviation from the mean within which one-half of a large number of deterininations may be expected to fall, also was calculated for each cement. Probable error is determined ( 5 ) by an application of the principIe of Ieast squares in accordance with the equation
where T is the probable error of a given group of deterniinations from the most probable value as represented by the mean of a large number of group determinations, Z v 2 is the sum of the squares of the residuals or deviations from the means, and n is the number of group determinations under consideration. The values obtained are shown in Table I. The probable errors as here determined represent the probable deviations from the best results that can be obtained b y the method and bear no necessary relation to the true values. No method is known by which the absolute accuracy of the method can be gaged. Close agreement of results on mixtures of known composition, and with the qualitative observations on White’s test and microscopic and x-ray examinations, however, lead to the assurance that the method is essentially reliable. I n Figure 2 the probable errors are plotted on the abscissa against the mean values for uncombined lime on the ordinates. From this figure the probable error of any group of determinations may be read a t any mean value for uncombined lime u p to about i per cent. For example, if the
TROCK and Lukens (2) have recommended tantalum as B satisfactory substitute for platinum in the electrodeposition of copper if the surface of the cathode is kept free of oxide. They have also recommended the use of abrasives and the mechanical removal of the oxide film as more satisfactory than chemical methods. This involves time, considerable care, and the loss of considerable tantalum. Their table shows a loss of from 0.002 to 0.0035 gram for each stripping and cleaning. Doughty and Freeman ( I ) have shown that an aqueous solution of trichloroacetic acid and ammonia readily dissolvcs copper, cadmium, zinc, and, to some extent, nickel. They have applied the reaction to the stripping of copper from the surface of silver cathodes used for the deposition of copper. Silver cathodes, however, have never been satisfactory in t h i i laboratory. The present writers have applied Doughty’s method to the stripping of copper from the surface of tantalum cathodes with entirely satisfactory results. A 1:1 aqueous solution of ammonia (sp. gr. 0.90) made 10 per cent with respect to trichloroacetic acid was employed. Ten consecutive depositions were made, stripping each time with the above solution. Each deposition was practically perfect, being smooth, highly adherent, and of excellent color. The electrode showed no deteriorating effect on continued use, as was found by Strock and Lukens, but actually improved with use. Aftcr stripping the tenth time the cathode exhibited a high metallic luster, entirely free of oxide, and showed no tendency toward brittleness. The loss of weight for the ten strippings was 0.001 gram, a n average loss for each stripping of 0.0001 gram, practically zero. Weight before 10 strippings Weight after 10 strippings Loss of weight
Grams 14 9953 14 9943 0 0010
The time required for each stripping was less than 5 minutes. Literature Cited (1) Doughty and Freeman, J . A m C h e m Soc , 43, 700 (1921). (2) Strock and Lukens, Trans. A m . Electrochem. Soc , 66, 409 (1929). 1
Received April 29, 1930.
