V O L U M E 2 4 , NO. 3, M A R C H 1 9 5 2 end points which are close together, When conditions are such that the wings of the curve are actually straight lines, but the end point is very poor, the ordinary method is sometimes superior to the differential method. For slow reactions, the differential method may be inapplicable. A buret is described which can be used either in the normal manner or to deliver solution a t a uniform and low rate of addition. ACKNOWLEDGMENT
This work was supported in part by the Wisconsin Alumni Re-
459 search Foundation and in part by grants-in-aid from E. I. du Pont de Nemours & Co. and the Atomic Energy Commission. LITERATURE CITED
(1) Blaedel, W. J., and Malmstadt, H. V., ANAL. CHEM.,22, 734 (1950). (2) rbid., P. 1410. (3) rbid., p. 1413. (4) Ibid., 23, 471 (1951). (5) Ibid., 24, 450 (1952). (6) Lingane, J. J., Ibid., 20, 285 (1948). RECEIVED for review June 16, 1951.
Accepted September 25, 1951.
Precipitation from Homogeneous Solution LOUIS GORDON, Syracuse University, Syracuse, iV. Y. The conventional or heterogeneous process of precipitation, in which a solution of the precipitant is added to the reaction medium, often results in excebsive coprecipitation due to localized concentration effects. In the process of precipitation from homogeneous solution, the precipitant is generated uniformly throughout the entire reaction region. This procedure thus avoids the concentration gradients which characterize the ordinary mode of precipitation. A comparison of the precipitates formed in the two processes indicates the superiority of the homogeneous method because there is much
A
less coprecipitation of interfering elements. The precipitate is also more readily filtered and washed, owing to its dense and compact character. The technique of precipitation from homogeneous solution can be an efficient method of separating a substance from other constituents, and therefore it should find further application in gravimetric analysis. The number of fractionation steps in certain precipitation processes can be reduced. Distribution studies of trace materials on carriers formed under slow precipitation conditions closely approximating equilibrium conditions are also possible.
ess, are reduced through the use of dilute solutions there will be some improvement. A much higher degree of improvement will be obtained when the precipitant is generated uniformly and slowly, so that the liquid phase remains homogeneous. The precipitate resulting from such a process will show much less coprecipitation of interfering ions. It can also be readily filtered and washed because of its more compact character. A visual comparison of the precipitates obtained by the two methods is shown in Figures 1 and 2. The apparent volumes of such precipitates and some typical data are compared in Tables I and 11. Because of the unsatisfactory results obtained whenever gelatTable I. Apparent-Volume Ratios of Precipitates inous precipitates similar t o hydrous ferric oxide were encounTime of Settling tered, it is obvious why so much early attention was focused on Months’ Ratio4 this class of substances. In a recent paper Willard (28) reviewed 2 20 Fer] IC oxide the early methods used t o effect a homogeneous change in the pH Thorium oxide 2 9 Stannic oxide 2 20 of a solution. The basic acetate and similar methods (18) and hlasnesium oxalate 0 . 5 hour8 51 2 . 5 hours 34 procedures which employ sodium thiosulfate (11) or hexamethyl4 5 hours 29 17 hours 20 enetetramine (20)are examples. With these methods a gelatinous 2 months 11 precipitate is still obtained, although there is usually a reduction Ratio of volume of ppt. produced by heterogeneous method to that produced by homogeneous method. by about one half in the a p - parent volume as compared to the ammonia precipitate. Table 11. Comparative Separations It was first demonstrated by Element Other Substance Coprecipitated Present, TT’illard and Tang (37) that Substance Pptd. Method of Precipitation Mg. Gram Error, mg. more than just a homogeneous Mn 1 .O Ammonia (24) A1 0 . 1 Aluminum oxide change in pH is required to preMn 1.0 Urea-succinate (36) A1 0 . 1 Ammonia, two pptn. (84) Fe 0 . 1 Co 0 . 0 5 Ferric oxide cipitate hydrous oxides or basic Urea-formate (36) Fe 0.112 c o 1.0 Urea-formate, 2-stage pptn. (36) Fe 0.112 co 1.0 salts so that there is minimum Oxalate-ammonia ( 9 8 ) Ca 0.0503 Mg 0 1 Calcium oxalate Urea (31) Ca 0.0503 .Mg 0.1 coprecipitation of interfering Methyl oxalate (16) Ca 0.0503 M g 0.1 ions. I n an experiment in which Li 0.1 Ammonium oxalate (6) hlagnesium oxalate Mg 0.0100 Ethyl oxalate (14) Li 0 . 1 Mg 0.0112 a solution containing alum and a Other substance determined by analysis of precipitate. urea was heated to boiling, they b Data obtained by Gordon and Wroczynski (18). obtained adense aluminum pre-
MOSG the desirable properties of a substance to be sepa-
rated from a solution by a precipitation process are minimum coprecipitation and maximum filterability. When a precipitate is produced by the conventional method of adding a reagent directly to a solution, the degree t o which these properties are attained often leaves considerable room for improvement. Such is the case when ammonium hydroxide is used to precipitate hydrous ferric oxide in a gelatinous form a t pH 2 in the presence of manganous ion which precipitates a t p H 8.5 (29). When the concentration effects, characteristic of this heterogeneous proc-
Q
ANALYTICAL CHEMISTRY
460 cipitate which could not be reproduced when aluminum chloride was substituted for the alum. Subsequent tests indicated that the presence of a particular anion was required in order t o obtain a dense precipitate. In this case it is the sulfate ion. When the required ion is furnished by an organic acid such as formic, there is the added advantage of buffering action which aids in controlling change of pH, and it is easier t o remove the organic anion during the ignition process.
pH by simply cooling the solution and the precipitate is removed by filtration. This constitutes the fint stage. More urea is then added and the hydrolysis is continued in order t o separate the 1t o 5 mg. of substance which remains in solution a t the pH a t which the hulk of the precipitate was removed. The small amount of precipitate obtained in this second stage does not have a high capacity for an impurity, even though there is at the higher pH a larger amount of coprecipitation per unit quantity of precipitate. In one experiment in which 1.3 mg. of copper were coprecipitated with basic iron formate, this was reduced t o 0.32 mg. when the two-stage process was employed. It is also possible t o substitute an ammonia precipitation for the aeeond stage. I n either case the second precipitate may be filtered through the paper containing the bulk of the precipitate. The use of urea has been successfully applied t o the separation of basic salte of aluminum (36, 57), gallium (80),thorium (33), and iron ($6). Zirconium ( l a ) and tin(1V) (fd) can also he separated as dense basic salts. A method for zirconium was not developed further, as many elements normally associated with it would also precipitate. Tin(1V) separates &B an exA A' B B' C D' tremely compact hasic sulfate, but satisfactory ipitate Figure 1. Effect of JMethod on Vol seoarations from other elements are not obtained. It is quautitatively precipitated at pH 1.3, but even a t this low pH considerable amounts of manganese(I1) are coprecipitated. Figure 4 indicates the residual manganese remaining with 273 mg. of tin(1V) precipitated by the urea method and then washed 15 times with hot 1% ammonium nitrate solution. I n view of the fact that manganous ion (89) precipitates a t p H 8.5, a surprising amount of manganese coprecipitates. The presence of these anions changes the pH a t which a parThe precipitation of basic salts with ureais characterized by the ticular substance begins t o precipitate. This can be advantsgeous formation of thin transparent films of precipitate which adhere when B basic =It is precipitated a t s lower pH than is the hydrous strongly t o glass surfaces. Nonnslly, the few milligram involved are dissolved with hydrochloric acid, reprecipitated with oxide, since i t affords opportunity for a better separation from ammonium hydroxide, and then filtered through the same filter interfering ions. The effects of difierent anions have been studied by Willard and coworkers. &me of the effect8are surprising. For example, when aluminum (36) in the Dreaence of formate a . . is precioitated . . dense product is olotainrd, a.hrrr.2~in the presrnrr of acetiie the .. . " . -. precipitate 18 flocculent. l l l e Same 1'6 true oi tliorium (I?, 33). Directions which do not utilize .a ''suitable anion" when the urea method is employed for the p ntcipitation of basic salts or hydrous oxideioverlook a very essenli d aspect. Urea is an almost ideal reagent for use in hydrolytic processes. It possesses negligible basic properties (Ks = 1.5 X 10-9,it is soluble in wster, and ita hydrolysis 13t e can be easily controlled. Urea resdilv -hvdrolvaes . " at 90"to 100'' C. (Z7);hydrolysis can he quickly terminated a t a desired pH b,y cooling the reaction mixture. The rate of change of pH can he modified through the use of varying quantities of urea and buffer. Figure 3 shows the variation of pH with t i e for two precipitittion processes. Becauseit is possible t o follow and iihus maintain control of the pH of the solutions in which urea is being hydrolyEed, one can utilize the technique of two-stage prmecipitittion (35). In effect, this technique obtains results compai?able t o a double precipits, tion without having t o redissolve the f i j:st precipitate. I n studying A the separation of hasic iron formate Ifrom homogeneous solution with urea, Willard and Sheldon (35)determinedthe relation heFigure 2. EReet tween pH and the amount of coprecipitation. They ahserved a small degree of coprecipitation a t low o H values at which the separation was 95 t o 99% complete a increased rapidly with increase in 2 method is used, the precipitation proc ~
V O L U M E 2 4 , NO. 3, M A R C H 1 9 5 2
461
paper containing the initial precipitate. It is a curious case in point that the adherent film formed when basic stannic sulfate is precipitated can only be removed either with hydrofluoric acid or by filling the beaker with an ammonium sulfate-ammonium hy-
620 6.00
'
-
.
. 2.00
5.80
.
560
.
7'"YN
2.20
:
I .80 ,,60
PH (Th) '"O
.1.20
520
- 1.00
5 00
-
4 ao
0.80
- 0.60 - 0.40
4.60 4.40
0
10
20
30 4 0 50
60 70
80 90 100 110 120
MINUTES
Figure 3.
Change of pH with Boiling Time of Solutions Containing Urea
droxide solution a t pH 8.5 and then warming for about 30 minutes until large transparent films break loose from the beaker walls. To those who have worked with the urea method, the existence of these films will always be a reminder of Willard's fond hope that
Table 111.
No.
WaCl Present, Grqms
1
1.75
2
1 .75
3
1 75
4 5
1.7
Coprecipitation of Sodium with Calcium Oxalate Conditions of Precipitation (0.1 Gram of C a + - Present) Pptd. at looo C. by adding 1 gram of ammonium oxalate to 250 ml. of solution containing 5 ml. of concd. hydrochloric acid: then neutralized with 4 N ammonia added dropwise 1 gram of ammonium oxalate in 50 ml. of water added to 250 ml. of calcium 8 0 1 ~ tion at 100° C., containing 1.0 ml. of eoncd. ammonium hydroxide. Time, 1 minute 1 gram of ammonium oxalate in 25 ml. of cold water added at once to calcium solution having total volume of 250 ml. Pptd. at room temperature, then digested at 100' C. for 2 days Pptd. from homogeneous solution with 10 grams of methyl oxalate in 250 ml. containing 100 ml. of solution 2.5 N in ammonium acetate and 2.5 N in acetio acid
NatCaOd %%/Gram rOcHrO 13.0"
late from sodium is to carry out the precipitation a t room temperature and then digest the precipitate for 2 days a t a higher temperature. This allows the fine crystals first formed to undergo an entire structural change due to transformation of the higher hydrates. Kolthoff's data are shown in Table I11 with data obtained when calcium is precipitated from homogeneous solution with methyl oxalate. While his mode of precipitation gives improved results in this caae, it nrould probably not be applicable with substances that postprecipitate on standing, as is the case with magnesium oxalate. The hydrolysis of esters can be used to produce a desired anion for a precipitation process. Killard and Freund (31)used triethyl phosphate to separate zirconium and hafnium fractionally. This method, in which trimethyl phosphate can be used equally well, affords a more efficient separation than does the method of Larsen, Fernelius, and Quill (22) which approximates the homogeneous process. This latter method utilized an atomizer system which simultaneously sprayed a dilute solution containing zirconium and hafnium and one containing phosphoric acid into a reaction vessel containing sulfuric acid. Although the resulting precipitate thus produced is described as sandy and easy to filter and wash (f), several more fractionation steps are required to ohtain a precipitate of desired hafnium content than when the ester hydrolysis method is used. Trimethyl phosphate has also been used to precipitate zirconium from homogeneous solution but does not yield separations comparable t o those obtained when zirconyl phosphate slowly and homogeneously precipitates from a solution containing the soluble metaphosphate complex (34). The hydrolysis of organic esters of oxalic acid has been utilized to obtain improved separations. Methyl oxalate precipitates coarse crystalline oxalates of thorium and the rare earths (33) and thus effects a separation from the large amount of phosphate present in monazite sand. The fractional separation of lanthanum and cerium(II1) and of lanthanum and praseodymium has also been accomplished with methyl oxalate (IS).
' 9
8.2"
2.w
4.26 4.6)
a Data by Kolthoff (.91): sodium determined by precipitation as zinc uranyl acetate. b Sodium determined by difference; 0.1006 gram of C a t + used: data from ( 1 6 ) .
I
0
50
100
150
200
250
300
I
Mn ADDED (mg)
he will some day find a way of making all the precipitate adhere to the beaker, so that it will only be necessary to dry and weigh the beaker after discarding the solution. The use of urea is not limited to the precipitation of basic salts. Calcium oxalate (3) can be precipitated in the form of large crystals, which according to TTTIlard and Furman (52) accomplish in one precipitation a satisfactory separation from magnesium even in dolomite. This method has been adapted to the determination of calcium in natural waters (19). Kolthoff (21) states that the separation of calcium oxalate from the alkali metals is a case in which the heterogenous method is capable of a better separation than is possible with the homogeneous method. Kolthoff's method for separating calcium oxa-
Figure 4. .Snalytical Coprecipitation of Manganese with Basic Stannic Sulfate 273 mg. of tin present; precipitate washed 15 times
Ethyl oxalate precipitates magnesium (14)in 85% acetic: acid solution. The resulting oxalate is strikingly different from that obtained by the precipitation procedure of Elving and Caley (6). There is a similar method for the precipitation of zinc oxalate ( 2 ) . For the generation of sulfate ion one can use sulfamic acid ( I O ) , potassium methyl sulfate (f2), or ethyl or methyl sulfate (8, 28). Elving and Van Atta (8) recently published directions for the separation and determination of barium, strontium, and calcium using dimethyl sulfate. In the presence of normally occurring
ANALYTICAL CHEMISTRY
.462
Table IV.
Element Precipitateda Aluminum (,Ti‘), gallium ( 3 0 ) . thorium (39), iron ( 3 6 ) , tin ( 1 8 ) , zirconium (18) Titanium (88) Thorium (20) Zirconium and hafnium ( 5 1 ) b Zirconium ( 3 4 ) Zirconium (3.4)
many more methods may soon become available. One may well quote Willard (88),who says, “when one considers how frequently precipitation processes are utilized, it is obvious that any improvement in the process is of much
Ethyl oxalate Urea and an oxalate
Thorium and the rare ea;ths (33), lanthanum and cerium ( 1 3 ) b , ,lanthanum and praseodymium ( f W calcium , (16) . Magneazum (141, zinc ( 8 ) , c a k i u m and magnesium ( 7 ) Calcium (31)
LITERATURE CITED
Dimethyl sulfate Sulfamic acid Potassium methyl sulfate Trichloroacetate Areenite Tetrachlorophthalic acid Iodine Urea and a dichromate
Barium (8),calcium (81, strontium (8) Barium (10) Barium (12) Lanthanum and praseodymium (26)b Zirconium ( 1 7 ) Thorium (16) Thorium ( I $ ) , Zirconium ( 4 ) Barium (9,E%)
Summary of Methods of Precipitation from Homogeneous Solution
Precipitant Hydroxide
Reagent Urea
Phosphate
Acetamide Hexamethylenetetramine Triethyl phosphate Trimethyl hosphate Metaphospgoric acid
,Oxalate
Methyl oxalate
Sulfate Carbonate Arsenate Tetrachlorophthalate Iodate Chromate a b
Methods for elementa in italics may be found in periodical literature. Fractionation procedures.
Audrieth, L. F., “Inorganic Syntheses,” Vol. 111, p. 71, New York, McGraw-Hill Book Co., 1950.
Caley, E. R., Gordon, L., and Simmons, G. A., Jr., ANAL. CHEM.,22, 1060 (1950).
Chan, F. L., dissertation, University of Michigan, elements, the precipitation of barium sulfate from homogeneous solution is a superior method of separation. However, the precipitation of barium sulfate in the presence of a trace quantity of radium is another instance where the heterogeneous method of precipitation can accomplish a better separation. Doerner and Hoskins (5) found that more radium sulfate is coprecipitated when barium sulfate is precipitated under conditions similar to the homogeneous process from a solution containing on the order of 10” gram of radium. While the total amount of radium is so small that it would not interfere with the gravimetric determination of barium, were it all to precipitate, this experiment indicates that the technique of precipitation from homogeneous solution might be useful in studying the coprecipitation of trace quantities on carriers formed under almost ideal conditions allowing distribution equilibrium between solid and solution. Carbonates can be precipitated from homogeneous solution by hydrolyzing the trichloroacetate ion to form carbon dioxide and chloroform. Quill and Salutsky (65) fractionated lanthanum and praseodymium by hydrolyzing their trichloroacetates; the praseodymium carbonate concentrated in the precipitate. This procedure is a very iapid method of fractionation. A desired precipitant may be synthesized by an oxidgionreduction reaction. Gump and Sherwood (17) oxidized arsenite to arsenate to precipitate zirconium in a dense form. The oxidation of iodine to iodate with chlorate has been utilized t o precipitate thorium iodate ( I d ) in a dense granular form. This method, in which a solid phase is present in the solution, is perhaps not a true homogeneous precipitation but it achieves the same results. Zirconium iodate (4) can be similarly precipitated by the oxidation of iodine. The precipitation of thorium from homogeneous solution with tetrachlorophthalic acid (16) produces a dense crystalline precipitate which effects a separation from the rare earths. The procedure is simple. Cold solutions of thorium and of tetrachlorophthalic acid are mixed and when the resulting clear solution is slowly heated to 70” C. the precipitate forms. The mechanism of the reaction is not known. The method itself is an adaptation of a procedure used by Lenher (23),who similarly fractionated the yttrium earths with succinic acid by allowing the rare earth succinates to slowly separate from solution. Table I V summarizes the methods of precipitation from homogeneous solution which have been studied since 1937, when t h e urea method for aluminum (36, 37) was published. These methods, mostly by Willard, are few when compared to the vast number of precipitation procedures described in the literature. Other investigators have recently begun to apply the general principles of this mode of precipitation to different substances and
1932. (4) Cooperstein, R., master’s thesis, Syracuse University, 1949. (5) Doerner, H. A., and Hoskins, W. M., J . Am. Chem. Soc., 47, 662 (1925). (6) Elving, P. J., and Caley, E. R., IND.ENG.CHEM.,ANAL. ED., 9, 558 (1937). (7) Elving, P. J., and Chao, P. C., ANAL.CHEM., 21, 507 (1949). (8) Elving, P. J., and Van Atta, R. E., Ibid., 22, 1375 (1950). (9) Firsching, H., master’s thesis, Syracuse University, 1951. (IO) Freund, H., dissertation, University of Michigan, 1945. (11) Furman, N. H., ed., “Scott’s Standard Methods of Chemical Analysis,” 5th ed., Vol. I, pp. 950-2, New York, D. Van Nostrand Co., 1939. (12) Gordon, L., dissertation, University of Michigan, 1947. (13) Gordon, L., Brandt, R. A., Quill, L. L., and Salutsky, M. L., ANAL.CHEM.,23, 1811 (1951). (14) Gordon, L., and Caley, E. R., Ibid., 20, 560 (1948). (15) Gordon, L., Vanselow, C. H., and Willard, H. H., Ibid., 21, 1323 (1949). (16) Gordon, L., and R’roczynski, A. F., ANAL.CHEM.,in press. (17) Gump, J. R., and Sherwood, G. R., Ibid., 22, 496 (1950). (18) Hillebrand, W. F., and Lundell, G. E. F., “Applied Inorganio Analysis,” p. 71, Kew York, John Wiley & Sons, 1929. (19) Ingols, R. S..and Murray, P. E., ANAL.CHEM.,21, 525 (1949). (20) Ismail, A. M., and Harwood, H. F., Analyst, 62, 185 (1939). (21) Kolthoff, I. M., and Sandell, E. B., J . Phys. Chem., 37, 443 (1937). (22) Larsen, E. M., Fernelius, W. C., and Quill, L. L., IND.ENG. CHEM.,ANAL.ED.,15, 512 (1943). (23) Lenher, V., J . A m . Chem. Soc., 3 0 , 5 7 2 (1908). (24) Lundell, G. E. F., and Knowles, H. B., Ibid., 45, 676 (1923). (25) Quill, L. L., and Salutsky, M. L., Symposium on Chemistry of
(26) (27) (28) (29) (30)
the Less Familiar Elements, Division of Physical and Inorganic Chemistry, 117th Meeting, AM. CHEM.SOC., Detroit, hIich., 1950. Tang, N. K., dissertation, University of Michigan, 1930. Warner, R. C., J . BioZ. Chem., 1 4 2 , 7 0 5 (1942). Willard, H. H., ANAL.CHEM.,22, 1372 (1950). Willard, H. H., and Diehl, H., “Advanced Quantitative Analysis,” p. 45, New York, D. Van Nostrand Co., 1943. Willard, H. H., and Fogg, H. C., J . Am. Chem. SOC.,59, 1197
(1937). (31) Killard, H. H., and Freund, H., IND.END.CHEM.,ANAL. ED., 18, 195 (1946). (32) Willard, H. H., and Furman, N. H., “Elementary Quantitative Analysis,” 3rd ed., pp. 342-3, 398, Kew York, D. Van Noatrand Co., 1940. (33) Willard. H. H., and Gordon, L., ANAL. CHEM.,20, 165 (1948). (34) Wllard, H. H., and Hahn, R., Ibid., 21, 293 (1949). (35) Willard, H. H., and Sheldon, J. L., Ibid., 22, 1162 (1950). (36) Willard, H. H., and Tang, N. K., IND.ENO.CHEM.,ANAL.ED., 9 , 357 (1937). (37) Willard, H. H., and Tang, N,K., J . Am. Chem. Soc., 59, 1190 (1937). RECEIVED for review May 29, 1951. Accepted October 2, 1951. Presented before the Division of Analstical Chemistry at the 119th Meeting of the AMERICAX CHEMICAL SOCIETY, Boston, Mass.