Aug., 1954
SOLUBLE CRYSTALLINE POLYPHOSPHATES
GO3
SOLUBLE CRYSTALLINE POLYPHOSPHATES-THEIR PURIFICATION, ANALYSIS AND PROPERTIES BY OSCART. QUIMBY Miami Valley LaboTatories, The Procter & Gamble Company, Cincinnati SI,Ohio Received April 19, 1964
Application of tracer techniques has shown that sodium pyrophosphate is readily obtained in 99.9% purity, but that sodium triphosphate is not easily obtained in higher than 99% purity, the chief impurity being pyrophosphate formed during the triphosphate recrystallizations. While some orthophosphate is probably also formed during triphosphate purification, it is either more readily removed than pyrophosphate or formed in lesser uantities; a t any rate the orthophos hate content of a quadruply recrystallized triphosphate seldom exceeds 0.1% as N a b P 0 4 . The purity attainable for t i e newly-discovered, crystalline hexaguanidinium tetraphosphate has not been so extensively investigated, but is probably at least 95%, and may be as high as 99%. The hexaguanidine salt, derived from the easily available tetrametaphosphate? provides the tetraphosphate in stable and convenient form, ready for instant use in any situation where a tetraphosphate ion is desired. Methods of analysis have been developed recently which are specific for individual polyphosphate species. The most versatile of these is based on paper chromatography studies in Canada and in France. Chromatography is capable of determining ortho-, pyro-, tri-, tetra-, trimeta- and tetrametaphosphates in the presence of one another and is also capable of extension to other condensed phosphates. I n our own laboratories certain specific analytical method8 have been develo ed. For commercial sodium tripolyphosphate (or triphosphated detergents) isotope dilution methods specific for pyro- a n a for triphosphate have been developed. They are based on the isolation of the "pure'' sodium salts as already indicated. Attem ts to obtain a pure pyrophosphate by precipitation with zinc ions in the presence of triphosphate, or to obtain a pure tri gosphate by recipitation with tris-(ethylenediamine)-cobalt(II1) ion in the presence of pyrophosphate, failed because eitier polyphos f a t e tends to cocrystallize with the other. I n spite of such difficulties it was possible to develop a colorimetric method for triphosphate in commercial triphosphate or triphosphated detergents, using the tris-(ethylenediamine) cobalt(II1) ion. When samples are available in crystalline form, e.g., commercial tripolyphosphate, analysis by X-ray diffraction gives the polymorphic distribution in addition to an estimate of triphosphate content. The stability of polyphos phates toward hydrolysis in solution decreases in the order pyro-, tri-, tetraphosphate. I n the hydrolysis or reversion of triphosphate there is an important difference between solution hydrolysis and that which accompanies the dehydration of the hexahydrate NabP3OI0.6H20. I n the former, all evidence is consistent with cleavage of triphosphate to give equimolar amounts of pyro- and orthophos hate. In the latter, considerably more than a mole of pyrophosphate is often found for each mole of orthophosphate. &is fact necessitates a revision of the hexahydrate hydrolysis reaction offered by Raistrick and supported by Thilo and Seeman, namely, NasP3010~6Hz0 + Na4P207 f NaH2P04 5H20. Some additional reaction + Na4P207 2Na3HP20, 11HzO. A commust be invoked which produces more pyrophosphate, e.g., 2NasP3OIo.6Hz0 bination of the two can explain the observations below the temperature of recondensation (120'). Above 120' both simple dehydration directly to N.%aPrOlo(11) and formation of NaSP3010 (11) by recondensation (reversal of above reactions) occur, the one which dominates depending on the speed of escape of the water. The interaction of calcium and triphosphate ions in solution is briefly discussed.
++
+
Introduction phates8-13 in aqueous solutions. Ingerson and Condensed sodium phosphates have received Morey,2 however, state that the normal pyrophosconsiderable attention in recent years, with the re- phate Na4P207is thermodynamically stable in cersult that most of the crystallizable species have now tain parts of the Na20-PzOs-H20 system. I n spite of its lack of thermodynamic stability been fairly well characterized. The present survey deals with recent progress in characterizing the the tetraphosphate can be prepared in aqueous solusoluble crystalline polyphosphates : pyro-, tri- and tions by mild alkaline hydrolysis of cyclic sodium tetrametaphosphate, another metastable substance. tetraphosphate. Prior to the classic work of Partridge, Hicks and This has been accomplished by three groups of in~ ~ -none ~ 6 of them succeeded in Smith, little phase information was available on v e ~ t i g a t o r s , ~ ~but the occurrence of sodium triphosphate Na6P3OI0and isolating a crystalline tetraphosphate. Thilo and Ratz14 obtained insoluble calcium, silver and zinc still less on that of sodium tetraphosphate NaeP&. This study and that of Ingerson and more^^,^ tetraphosphate precipitates, all of which were amor~ Ebel16 demonstrated for the NazO-P205 system that Nab- phous to X-rays. Westman, et u Z . , ~ , ~ and P3010is stable below a temperature of about 620" obtained tetrametaphosphate hydrolysates, the and that Na6P& does not occur, a mixture of bulk of whose phosphate ions differed chromatoNasPsOlaplus Na3P309always being found in well- graphically from the common phosphate species, crystallized samples of the tetraphosphate composi- namely, ortho, pyro, tri, trimeta and tetrameta, in tion. This was confirmed by other less extensive just the way one would expect for tetraphosphate studies. -7 ions. In the NazO-Pz05-H20 system not only is the While the data summarized below emphasize (1) tetraphosphate unstable, but-so also are the"triphos- purification and analysis of sodium triphosphate, phate and pyrophosphate, as shown by numerous and (2) isolation and purification of crystalline tetstudies of the hydrolysis of the latter two phos(8) 8. J. Kiehl and W. C. Hansen, J . A m . Chem. Soc., 48,2802 (1926). (1) E.P.Partridge, V. Hicks and G. W. Smith, J . A m . Chem. Soc., 63, 454 (1941). (2) E.Ingerson and G. W. Morey, A m . Mineral., 28, 448 (1943). (3) G. W. Morey and E. Ingerson, A m . J . Sei., 242, 1 (1944). (4) K.R. Andress and K. WUst, 2. anorg. allgem. Chem., 237, 113 (1938). (5) 0.T.Quimby, unpublished studies (1938-1940). (6) P. Bonneman-BBmia, Ann. chim., 16, 395 (1941). (7) A. E. R. Westman, A. E. Scott and J. T. Pedley, Chemistry i n Canada. 35 (1952).
(9) J. Muus, 2. physik. Chem., 169A, 268 (1932). (10) 8. J. Kiehl and E. Claussen, Jr., J . A m . Chem. Soc., 67, 2284 (1935). (11) R. Watzel, Die C h e d e , 6 6 , 356 (1942). (12) R. N.Bell, Ind. Eng. Chem., 39, 136 (1947). (13) J. R. Van Wazer, E. 6. Griffith and J. F. McCullough, J . A m . Chem. Soc., 7 4 , 4977 (1952). (14) E. Thilo and R. Rlitz, 2.anorg. Chem., 260, 255 (1949). (15) A. E. R. Westman and A. E. Scott, Nature. 168, 740 (1951). (16) J. P. Ebel, BUZZ. 800. chim., 991, 1085 (1953).
OSCART. QUIMBY
604
raphosphates of organic bases, properties of both polyphosphates and analyses for pyrophosphate ions also receive attention. Triphosphate (and Pyrophosphate) Purification The compound NasPaOlo was made as early as 1895, as demonstrated by the optical properties reported by Schwarz.'7 However, lack of adequate methods of analysis has until recently prevented a critical examination of methods of purification. By using tracer methods to follow the purification of triphosphate,18it has been found that commercial triphosphate of 85-94% purity expressed as Na6P3010can be brought to a purity of 99% (anhydrous basis) by three to five crystallizations from water-ethanol mixtures a t room temperature, followed by air drying. Occasional samples of 99.5% purity have been obtained.
W
c a c_
&
0 W K
Vol. 58
anol slowly with stirring until ths ratio of ethanol to water is about */d by volume. Aft,er a total of 30 minutes of stirring the hexahydrate crystals are filtered off and washed twice with 1:I mixtures of ethanol and water, using an aspirator to remove most of the adhering liquid. The damp crystals are then redissolved in the minimum amount of water and the process repeated. If the original purity was below 90% the sample is given four or five crystallizations; if the original purity was above 90% one less crystallization is needed. Finally the crystals of hexahydrate are airdried a t room temperature, preferably a t relative humidities of 4040%. Heat should not be used in any part of the process; drying of the crystals is to be avoided in intermediate stages, as is vacuum-drying on the final stage. The yield of purified Na6P3010is usually 4045% of the weight of commercial triphosphate taken.
Removal of Ortho- or Trimetaphosphate.-When
10% ortho- or trimetaphosphate, tagged with P82,
is added to pure inactive triphosphate and the mixture purified by recrystallizations similar to the method described above, the added impurity is readily reduced to a level of 0.1% or less, as illustrated by Fig. 1. For the orthophosphate this has been confirmed by other methods" on inactive triphosphate preparations. Since these impurities are usually present in much smaller quantities than lo%, it is apparent that impurity levels below O.ly0 can be achieved easily by the recrystallization . . method. Removal of Pyrophosphate.-While tagged pyrophosphate is less readily removed (Fig. Z), never-
a 100
0. l-
v1
40 'OI
5 0
r
3
4
I a
W
-
t a a a
z
0 W
0
0'02
a a
s t
z W
0.01
. .
'\
-
0.007
a
0.004
a a
0.002
4 a
t
0.00l~
RIM ETA PHOSPHATE ( No3PZ209)
b
1
2
NUMBER
3
4
5
6
7
8
OF CRYSTALLIZATIONS.
Fig, 1,-Removal of added impurity (NarHP3204 or Na3P32aOo) from sodium triphosphate (STP) by fractional crystallization; phosphate/water ratio 1 g./S ml ethanol/ water volume ratio 1/4, room temperature 25-3jb. Method.-The method of purification is simple. One makes an aqueoufi solution containin 12-15% of commercial triphosphate (usually S5-94% P$a6P3010), filters to remove any insoluble matter, and preci itates most of the triphosphate as hexahydrate Na5PsO1o&ItO by adding eth(17) F. Sohwarz, Z. anorg. Chem., 9, 249 (1895). (18) 0. T. Quimby, A. J. Mabis and H. W. Lanipe, Anal. Chem., 88. 661 (1954).
I---0001
NUMBER OF CRYSTILLIZATIONS.
Fig. 2.-Removal of added pyrophosphate impurity (Na4P3220,)from STP by fractional cr stallization; phorphate/water ratio 1 g./S ml., ethanol/Ywater volume ratio 1/4, ezcept for curve B where it was 1/5, room temperature 25-35
.
theless one sees that four crystallizations bring the added pyrophosphate to a level of O.5y0or less provided the original sample contained no more than 20 g. of NadPz0, to 80 g. of Na6P3010. That, the total pyrophosphate is not necessarily reduced to this level is shown by a reverse experiment in which 10-30% tagged triphosphate, purified by a procesn
SOLUBLE CRYSTALLINE POLYPHOSPHATES
Aug., 1954
similar to the above, is added to inactive pyrophosphate, and the pyrophosphate purified by the same process except that more water is required to dissolve the phosphate. l8 As expected the activity t'alls rapidly during the first few crystallizations . (Fig. 3) because tagged triphosphate is being lost.
loDl 70
605
At least this would not be surprising in view of the labile nature of the hexahydrate.6*1g-21 Critical Ratio of NasPaOloto Na4P207.-For isolation of pure sodium. triphosphate from a mixture containing both pyro- and triphosphate, the initial weight ratio NasP3010/Na4P207 must exceed 7/3, preferably be 4/1 or greater.lS This is readily apparent from the fractional crystallization data of Fig. 4. In like manner this ratio must be less than 7 / 3 , preferably less than 3/2, in order to yield pure pyrophosphate.
w' t
a
t
E
0 W
a a n
a
VI
I c)
r
r
4
H
w
a >
r, a
3
n
5.
a
0
W
a a a
'10 S T P IN PHOSPHATE M I X T U R E ,
Fig. 4.-Purification diagram for sodium triphosphatesodium pyrophosphate mixtures by fractional crystallization. phosphate/water ratio 1 g./14 ml. for less than 70% STP in mixture, 1 g./8 ml. otherwise] ethanol/water volume ratio 1/4, room temperature 25-35'.
8 I-
z
w
K
a
004
-
n n
a
Analysis Until recently the available solution methods for determining triphosphate22p2awere indirect and -* F hence subject to interference from other polyphos0.01 1 0 1 2 3 4 5 6 7 phate ions. Accordingly several attempts have for the tribeen made to develop methods specific NUMBER OF CRYSTALLIZATIONS. . . Fig. 3.-Removal of added triphosphate impurity phosphate ion. ( Na:,Paz3Ol0)from sodium pyrophosphate (SPP) by fracX-Ray Diffraction.-Methods based on measurtional crystallization; phosphate/water ratio 1 g./14 ml., ing intensities of diffraction peaks have been in ethanol/water volume ratio 1/4, room temperature 25-35', use for some time.19s24 They are especially useful However, all of these curves eventually level out when information is desired on the polymorphic because the tagged triphosphate contains tagged nature of a solid sample, for they provide a convepyrophosphate (usually about lye) which is, of nient way of estimating relative amounts of tricourse, retained by the pyrophosphate. For exam- phosphate present as Phase I, Phase I1 and hexaple, a mixture supposedly containing 7Oy0 Na4Pz07 hydrate. Some examples are given on commercial 30% NasP323010 acts as if it contained 70% sodium triphosphate in Table I where the compoNa4P207 0.3% Na4Pa2z07 29.7% Na6P323010.nents listed have been adjusted to a total of 100%. Thus, additional pyrophosphate is formed during Such samples may also contain small amounts of (19) B. Raistrick, Roy. Coll. Sci. J . , 19, 9 (1949). the purification of triphosphate. Calculations (20) E. Thilo and H . Seeman, Z . anorg. allgem. Chem., 287, 85 based on the solution hydrolysis rate given for (1951). room temperature in the "Hydrolysis" section show (21) 0. T. Quimby, Chern. Reus., 40, 141 (1947). (See also the secthat hydrolysis would not account for so much new tion below on decomposition accompanying dehydration of NasPaO1o.Gpyrophosphate. Possibly the explanation is that Hz0.) (22) R. N. Bell, Ind. Eng. Chem., Anal. Ed., 19, 97 (1947). the pyrophosphate impurity formed during tri(23) B. Raistrick, F. J. Harris and E. J. Lowe, Analyst. 76, 230 phosphate purification arises from surface decom- (1951). position of hexahydrate crystals during drying. (24) A. J. Mabis and 0. T. Quimby, Anal. Chem., 16, 1814 (1953). 0.02
+
.
+
+
606
OSCART. QUIMBY
VOl. 58
four times continues to reduce, but does not eliminate, such contamination. (4) Precipitation of zinc pyrophosphate fails to occur in the presence of massive amounts of triphosphate. For the case of the two precipitations called for by the method, the weight yield of zinc precipitate is usually equal to or slightly higher than that expected for quantitative precipitation of the pyrophosphate alone, pro= 3 or vided the weight ratio of Na6PaOlo/Na4Pz07 less. Thus, the net effect of the two precipitations X-RAY ANALYSES OF SOME COMMERCIAL TRIPHOSPHATES is that the pyrophosphate remaining in solutions is % Triphosphate present as fully or a little more than compensated weight-wise Sample Na6PaO1a (I) NitsPaOia (11) NasPi0106HzO NarPn07, % by the triphosphate contamination of the zinc py29 None 5.5 A 65.5 rophosphate precipitate. Table I1 gives examples 83 None 7.5 B 9.5 of the recovery of P32put in as pyrophosphate. None 12 C 27.5 58.5 Note that a t weight ratios of 9 or higher pyrophosWhile the X-ray method is specific for triphos- phate often does not precipitate a t all, an observaphate, it is not a general-purpose tool for quantita- tion also recorded by 0 t h e r s . ~ ~ ~ ~ 8 tive analysis, because it requires that the sample be TABLEI1 solid and fully crystallized. The importance of this is shown in the analyses of partially dehydrated PYROPHOSPHATE RECOVERY IN BELLMETHOD hexahydrate discussed in the later section on stabil% of total pyro found Wt. Zn Countity of NasP3010.6HzO. Phosphate composition precipitate ing Pal On wholly crystallized samples, the X-ray dif- 10% NasP3010 90% Na4P23207 101 95 fractometer method involving an internal stand- 25% Na~P3010 75% N a 4 P ~ 3 ~ 0 ~ 95 90 ardz4gives a total triphosphate determination with 50% Na6Pa010 50% Na4PP07 108 97 an average deviation of =k3% absolute if the tri- 75% NasP3010 25% Na4PP0, 99 84 phosphate is present in one form and of =k5% if 90% Na6Pdh 10% NadPPO, No pptn. all three forms are present. This method is of course not limited to deterIn view of the uncertainties in the pyrophosphate mining the triphosphate species, for pyrophosphate precipitation and the fact that pyrophosphate as crystalline Na4Pz07can also be determined with contributes two moles of Hfto approximately one slightly greater precision.z4 It will be seen from H + for the triphosphate, it is remarkable that the Table I that pyrophosphate is the chief impurity method has given such reasonable estimates of in commercial sodium triphosphate. Other crys- triphosphate content, based on the acidity not talline species, such as trimetaphosphate, might accounted for by the pyrophosphate as calculated also be determined by the X-ray method if present from the weight of zinc precipitate. However, when pyrophosphate fails to precipitate the titrain sufficient concentration. Precipitation by Zinc.-The Bell method of tion is all calculated to triphosphate and gives imanalysis has never been claimed to give high pre- possibly high values. cision in determining either tri- or pyrophosphate ; It is possible that a method for determining pyrofor even the author used the word ‘(estimation” phosphate might still be evolved by modifying the rather than “determination” in the title of his pa- Bell procedure. For instance, by using a limited per.zz Nevert#heless, it is true that this method excess of zinc a t pH 3.8 one could estimate pyrophosphate by centrifuging off the precipitate and often gave a fair estimate of triphosphate contentno mean achievement in the light of the mutual determining the excess of zinc in the supernatant. interferences revealed by Paztracer studiesz6to be By suitable control of NasP3010/Na4Pz07ratio and by making calibration curves in such a way as to discussed below. The Bell method was developed by modifying correct for triphosphate interference, a reliable dethe older method of Britzke and Dragunovz6for de- termination of pyro- in the presence of triphosphate, termining pyrophosphate by titrating the hydrogen may yet result, after the manner of the cobalt ions liberated by adding excess zinc ion a t p H 3.8. method developed for triphosphate by Weiserz9as But since both pyro- and triphosphate liberate indicated in the next section. Precipitation by Tris-(ethylenediamine)-cobalthydrogen ions, it was necessary to obtain an indesolutions containing only tripendent measure for one of them and this was done (111) Ion.-From by weighing the pyrophosphate as the zinc salt phosphate ions a t p H 3-4 the cobalt reagent precipitates triphosphate (practically quantitatively) after two precipitations. This ~ ~is. shown ~ H ~ in O .Fig. 5 , While the precipitation of pyrophosphate by zinc as C O ( ~ ~ ) ~ H ~ P ~ O a t p H 3.8 is quantitative in absence of triphos- which also shows that pyrophosphate by itself does phate, tracer studies showed that, in the presence of not precipitate under these conditions. Neverthetriphosphate, (1) the precipitation is incomplete (27) A. B. Gerber and F. T. Miles, Ind. Eng. Chem., Anal. Ed., 13, in both the first and the second precipitation; (2) 406 (1941). (28) H. Schmid and W. Dewald, Fette u. Seifen, 16, 19 (1953). both precipitates are contaminated with triphos(29) H. J. Weiser, unpublished reaults, Methods Development phate; (3) repeating the precipitation as many as Group, Chemical Division, The Procter & Gamble Company. (30) H. W. McCune and G. J. Arquit, presented before the Division (25) 0. T. Quimby and H. W. McCune, unpublished results. orthophosphate and/or trimetaphosphate, but their sum seldom exceeds 2%. Since commercial triphosphate contains no hexahydrate as made, Xrays may be used to determine the progress of hydration from exposure to humid atmosphere. Such a method shows not only the amount of hexahydrate produced a t any stage, but indicates that Phase I hydrates faster than Phase 11. TABLE I
+ + + + +
(26) E. V. Britake and S. S. Dragunov, J . Chem. Ind. (Moscoiu), 4, 29 (1927).
of Physical and Inorganic Chemistry of the Am. Chem. SOC. a t Chicago. September, 1953.
I
SOLUBLE CRYSTALLINE POLYPHOSPHATES
Aug., 1954
.
607
Iess, tracer studies30 (cf. also Table 111) have shown that a similar precipitate from a mixture of pyroand triphosphate (1) does not contain all of the triphosphate ; (2) is contaminated by pyrophosphate; (3) contains less pyrophosphate with each succeeding precipitation, but requires many precipitations (6-10) for production of a triphosphate of 99% purity, and (4) is not precipitated a t all if the ratio of pyro- to triphosphate is too high. TABLE I11 k PRECIPITATION OF C o ( e n ) 3 H ~ P ~IN O ~PRESENCE ~ OF PYRO- E V PH
3.5 3.5 3.5 2.5 3.0 4.0
4.5
Molar ratio tri/pyro
Ppt. yield, % of theory
6.5/1 1,08/1 1/3,23 4/1 4/1 4/1 1 /2
100 101 73 104 107 106 34
40'
II
W
PHOSPHATE
Contamination
a8
% Co(en)aHP%Oi
2.7 10.6 15.5 6.9
n LL
0
n J
w>
\
I I
K
20-
I I I I
n
.
..
8.1 24
Thus, the behavior with the cobalt reagent is a replica of that with zinc ions, except that pyro- and triphosphate have exchanged roles. These two condensed phosphates are so similar in structure that each is prone to enter a crystal lattice of the other. It is therefore difficult to obtain a pure pyrophosphate or a pure triphosphate precipitate, if the other species is present even in small quantities. I n spite of the pyrophosphate interference with the triphosphate precipitation by C ~ ( e n ) ~ Wei+~, ser29was able to develop a colorimetric method for determining triphosphate in materials such as commercial triphosphate and triphosphated detergents, both of which contain pyrophosphate. He accomplished this by preparing a calibration curve using mixtures of purified pyro- and triphosphate. To either such a known mixture or the sample an excess of the cobalt reagent is added, the precipitate is removed by filtration and the excess of reagent in the filtrate is measured with a colorimeter. Translation of the colorimeter reading into triphosphate content by means of the special calibration curve proved to be an effective way of correcting for the pyrophosphate contamination of the precipitate. Replicate analyses showed an average deviation, expressed as parts per hundred parts of triphosphate, of 0.5 in commercial triphosphate and of 0.6 in triphosphated detergents. The uncertainty in the true value is somewhat larger since the triphosphate used for calibration probably had a purity near 99% rather than the assumed 100%. The presence of 10% tetraphosphate or of as little as 2% sodium polyphosphate glass prevents completely the precipitation of tris-(ethylenediamine)-cobalt dihydrogen triphosphate in the colorimetric method. 29 Smaller amounts interfere, causing the triphosphate analysis to come out low. This may sometimes be overcome by adding some pure triphosphate to reduce the contaminant to triphosphate ratio. Isotope Dilution.-The fastest approach to a pure tri- or pyrophosphate phase known a t present occurs upon recrystallization of the sodium salts from aqueous media. As indicated in the section on
purification of triphosphate, this was accomplished a t room temperature by adding ethanol to an aqueous tri- or pyrophosphate solution to induce crystallization as NasP3010.6Hz0 or Na4P207.10Hz0. If the initial weight ratio of XasP3010 to Na4P207 is less than 3/2 four crystallizations under the conditions recommended1*reduces the triphosphate impurity to less than 0.5% as NasP3010. Similarly, when this ratio is 4/1 or greater, four crystallizations reduce the pyrophosphate originally present as such to an insignificant level. Unfortunately, however, some pyrophosphate is made in the recrystallization process, l8 so that it is difficult to reduce the pyrophosphate content of the anhydrous salt below 1% as Na4P207. Since a reproducible state of purity is attainable for both species, it is therefore possible to utilize the recrystallization process as the basis for determination of both pyro- and triphosphate by isotope dilution.18 The results have a precision of 1-1.5% absolute on such samples as commercial tripolyphosphate and synthetic detergents containing triphosphate. The method has the advantage of being specific for the species being determined. However, when either the pyro- or triphosphate content of the phosphate is less than 20% the relative error becomes rather large because of the necessity of adding pure inactive pyro- or triphosphate for ratio adjustment. Chromatography.-Two l a b o r a t ~ r i e s ~have , ' ~ recently developed paper chomatography to the point where quantitative analyses can be made for each of t.he following phosphate species: ortho, pyro, tri, tetra, trimeta and tetrameta. It involves an unambiguous separation of the species, followed by determining the total phosphorus in each separated phosphate. The accuracy is of the order of 3-57&. An advantage of the chromatographic method, as applied to determination of pyro- and triphosphates for example, is that tetraand higher polyphosphates do not interfere provided the conditions of experiment (pH, temperature, etc.) are chosen so as to make negligible the
OSCART. QUIMBY
608
VOl. 58
production of the lower members by hydrolysis of the higher polyphosphates. Solubility Table I V contains phase data on the metastable binary system NasP3Ol0-Hz0. These were obtained a t the Armour Research Foundationla' using
steady state between two periods of increasing solubility. Because of this limitation the data given a t 55" and higher are not trustworthy; significant hydrolysis occurred before solubility equilibrium was attained. Approximate values at 55 and 70" were 14.8 and 16.5%, respectively. At temperatures below 40" the data of Table IV agree reasonably well with the "approximate soluTABLE IV bility" data reported graphically by Van Wazer,a2 HETEROGENEOUS EQUILIBRIA I N THE SYSTEMNabPaOlr but differ from them in showing a shallow miniHz0 A T 0 TO 50" mum a t about 20". (Data of Jones, Cook and McCrone3I) Temp., NasPaOia, Time Crystal Stability OC. allowed, hr. phase wt. % Since triphosphate is metastable in contact with -0.45 3.03 Ice F.p. detn. water, it is of interest to examine its rate of hydroly-0.74 5.54 F.p. detn. Ice sis under various conditions. -0.93 7.36 F.p. detn. Ice Hydrolysis in Aqueous Solutions.-It has been -1.08 8.91 F.p. detn. Ice shown that triphosphate hydrolyzes first to a mix-1.11 9.44 Ice F.p. detn. ture of ortho- and pyrophosphate, and ultimately -1.22 10.45 Ice F.p. detn. to orthophosphate. At room temperature, how-1.33 12.13 F.p. detn. Ice ever, the hydrolysis is rather slow as shown by the -1.41 13.92 Ice very near F.p. detn. data for three concentrations and three methods euctectic of analysis in Table V. The concentration de0.00 Na5P3Oi0~6HZ0creases linearly with time, ie., appears to follow a 13.98 50 9.04 13.19 Nad'301~.6Hz0 zero-order law. I n 150 days the extent of de5.5 14.70 Na5P3Oio~6Hz0composition was so slight that the pH remained 116 13.00 17.82 NajP3O1o.6Hz0 practically unchanged at a value near that for pure 13.02 53.5 20.10 Na6Pa0106H20 NaaPaOlo, namely, 10.0. Thus, a t pH values of 143.2 12.92 24.78 24.87 25.00 29.84 30.00 30.01 30.51 30.81 30.87 30.93 31.11 31.31 31.81 32.39 33.95 34.83 35.03 39.83 40.06 44.97 44.98 50.07
NajP3OIo.6H20 TABLE V NahP3010.6HzO Na5Pa010~6H20HYDROLYSIS OF SODIUM TRIPHOSPHATE (STP) IN AQUEOUS Na~P~O1o~6Hz0 SOLUTIONS AT 25-28' Initial concn. Initial concn. Initial concn. Na5P3010.6H~0 0.1 g./100 ml. 0.7 g . / l O O ml. 10 g./lOO ml. Anal. for pyro A d . for tri Anal. for tri NasPaOlo.6Hq0 by iso. diln." by Co method b by G-M titrn.C Na6PaO1o.6Hz0 Calcd. % of Na5P3010.6Hz0 total STP Total NaJ'301~.6H20 reST9 Total STP NasPa010.6HzO Time, NarPnO7, main- Time, remaining, Time, remaining, days % ing days % days % NaJ'301~.6H~0 0 0.37 99.5 0 98.9,98.6 0 99 (assumed) NasPa010.6HzO 14 98 .49 99.3 7 99.1 Na5P30io~6Hz0 9 20 .40 99.4 14 98.6 41 100 NasP301n*6HzO 43 -77 98.9 35 98.0 70 100 Na5Ps01n.6H~O 76 .92 98.7 70 97.0 78 98 Na6Pr010.6H~0 103 1 . 2 98.5 106 97.55 121 92 Na6P3010.6H~0 146 3 . 3 95.5 146 97.25 148 94.5 Na5P3010.6H20 k = 0.010% k = 0.013% k = 0.02% Na6PaOio~6Hz0 NabP3010/day Na5P3010/day Na5PaOdday NasPa010.6Hz0 Analysis made for pyrophosphate by inverse isotope Nad'301~.6H20 dilution18 and calculating the triphosphate assuming one Na5Pa010.6H~0 mole of triphosphate disappeared for each mole of pyrophos hate formed. Initial orthophosphate content = sodium triphosphate from commercial tripolyphos- 0.154 as NazHPOl. Thus, 100 - 0.37 - 0.15 = 99.48% phate which had been purified by four crystalliza- NaSPSO10 present initially. b Analysis made by the Weiser2B tions similar to those described under "Purifica- colorimetric cobalt method. Analysis made by a modified tion," except that a larger ethanol :water volume Gerber-Miles titration.27 12.97 12.96 12.96 13.02 13.06 13.23 13.26 13.26 13.30 13.26 13.44 13.28 13.48 13.32 13.30 13.38 13.50 13.57 13.68 14.02 13.85 14.34
F.p. detn. 51 219 72 74.8 75 67 3 3 3.5 21.1 67 26 3.5 77.2 66.5 3.5 65.8 3.0 3.0 67 1.0
ratio (1:3) was used. While the purified triphosphate was not analyzed by the more reliable methods, the purity expected from the method of purification would be 98-99%. The solubility increases slowly with time and reaches the saturation value in a day or two a t 0", and in a few hours a t 45-50'. Since hydrolysis, detected by analysis for orthophosphate, also causes an increase in solubility with time, the equilibrium solubility was recognized as a (31) 8.P. Jones, J. W. Cook and W. C. McCrone, reaearch project a t Armour Research Foundation, Chicago, Illinois, sponsored by the Procter & Gamble Co.
(I
9-10 and at concentrations of 0.1 to 10.0 g./ml. the initial specific reaction rate k corresponds to the disappearance of 0.01 to 0.02% of the total triphosphate per day. The specific reaction rate appears to increase somewhat with increasing concentration, a trend contrary to that previously reported by Greena3 on very dilute solutions (6.5 to 65 p.p.m.) a t pH 9 and 190°F. (88°C.)) based on rate (32) J. R . Van Wazer, "Encyclopedia of Chem. Technology," Vol. X , edited by Kirk and Othmer, Interacience Publishers. Inc., 1953. p. 413. (33) J. Green, Znd. En& Cham., 48, 1542 (1950).
%
Aug., 1954
SOLUBLE CRYSTALLINE POLYPHOSPHATES
of appearance of orthophosphate. Probably this simply means that dependence of rate on concentration cannot be determined from the data of Table V, primarily because of the use of analytical methods of widely differing accuracy. The least accurate rate, i e . , that dependent on the titration method, may be in error by a factor of two. Suffice it to say that the widely different methods have given about the same hydrolysis rate. It is well known that rate of hydrolysis of any condensed phosphate increases rapidly with increasing temperature. Following the disappearance of triphosphate a t 180°F. (82°C.) by means of the cobalt method29 in a solution initially containing 10 g. of NasP3010per 100 ml. gave the data of Table VI. Again the initial decrease of triphosphate concentration with time is approximately linear and the specific reaction rate derived from this part of the curve corresponds to disappearance of about 4oy0 of the triphosphate during the first day (1.70j0/ hr.). This calculation ignores any effect of the change in pH which had fallen to about 7.5 by the end of the first day. The 2,000- to 4,000-fold increase in rate resulting from raising the temperature from 27 to 82" corresponds to a 4- to 4.5-fold increase for each 10" rise in temperature. For comparison, it may be noted that the above value appears consistent with the higher rate (5.8'%/hr.) reported by Green33for triphosphate at a concentration of 65 p.p.m., controlled pH of 7, and temperature of 190°F. (88°C.). Both the lower average pH and the higher temperature used by Green would increase the rate. Van Wazer and co-worke r reported ~ ~ ~an even faster rate (14y0/hr.) a t a still higher temperature (90") for a 1.29% NasP3010 solution held a t pH 7. Allowing for the small effect of the temperature difference, the rates determined by Green and by Van Wazer still differ markedly for some unknown reason.
609
and pyrophosphates occurs. At 95-105" there is much hydrolysis regardless of how fast the water is removed. In one room temperature experiment nearly half of the mater was removed from hexahydrate in a vacuum over PzO5; the pH of the product (1% solution in COz-free water) fell from 10.0 to 8.4 and about 40% of the triphosphate was converted to a mixture of ortho- and pyrophosphates. Because of this tendency to hydrolyze it is impossible to obtain satisfactory equilibrium measurements of the partial pressure of water vapor over hexahydrate crystalsjaleven a t relatively Surface hydrolysis low temperatures of 40-50 always modifies the system so that one always has more than the three phases demanded by the equation Na5P3OIo.6Hz0J_ NajP30lo(II) 6Hz0 (a) Such surface hydrolysis becomes appreciable a t 70-80" and quite rapid a t 90" as demonstrated by the weight-loss data of Bonneman-BBmia.6 From studies of the dehydration of hexahydrate in open containers a t temperatures near loo", RaistrickIg and Thilo and SeemanZ0conclude that what happens is
".
+
+ NaH2P04+ 5 H ~ 0 (b)
Ka5Pa010.GHz0 -+ Na4P207
According to Thilo and Seeman, heating the resultant equimolar mixture of pyro- and orthophosphate for a long time a t the same temperature, or better for a shorter time a t a slightly higher temperature, e.9., 105-120") simply results in recondensation thus
+
+
Na4P2O7 KaH2P04+ Na5PaOlo HzO ( c )
Looking first a t the recondensation reaction (c), experiments were made with equimolar mixtures of pure Na4P207and pure NaH2P04 to see whether any triphosphate forms at 105-120". In one case the mixture was put into solution and spray-dried to ensure an intimate mixture in the solid state; the only phase revealed by X-ray diffraction was TABLEVI Na4Pz07, but diffuse halos indicated the presence HYDROLYSIS O F SODIUM TRIPHOSPHATE (STP) I N AQUEOUS of amorphous material. Upon analysis for triSOLUTION AT 180°F. (82°C.) phosphate it was found to contain essentially no Initial concn. IO g./lOO ml. triphosphate (Table VII). Furthermore, heating Time, hr. Total S T P remaining, % this spray-dried mixture for 6 hours at 120" caused 0 98.6,08.9 the appearance of Na2HP04 in the X-ray pattern, 1 91.4 7.5 16 24 40 48 64 72
84.9 7'2.0,69.2 51.1,52.9 29.7 21.1 6.3 4.0 Initial k = 1.7%/hr.
Hydrolysis during Dehydration of Hexahydrate Crystals.-It has become increasingly evident6.19-21 that hexahydrate does not lose water a t any temperature below some point in the interval 130-140' without undergoing more or less hydrolysis. A t 120" the amount of hydrolysis is relatively small if a shallow layer of hexahydrate crystals is quickly brought to temperature and the mater vapor pumped off or swept away; however, in an ordinary oven the escape of water vapor is so delayed that almost complete decomposition into ortho-
TABLEVI1 ATTEMPTEDCONDENSATION OF A N APPROXIMATELYEQUIMOLAR MIXTURE OF Na4P20,A N D iYaH2P04 Triphos. as NasPaO~a,
%"
Loss in oven htg.,
% wt.
A.
Tot. pH of HgO prod. (ign. (1% loss), % soln.)
Prod. htg. hesahydrate, 12 hr., 95' 34,35,30 (12 8) 11 50 8 11 B. Prod. htg. hexa5 18.1 ... hydrate, 72.5 hr., 95' C. Mech. mixt. 0.98 mole X ~ ~ H P P1.00 O~: ... .. . . . 7.66 iYa4P207 D. Prod. C spray8.87' 7 42 dried from solution 290 E. Prod. D heated 6 2,2, 1 ( 2 . 8 ) 6.22 7.23 hr. at 120" a By colorimetric Co(en)3+smethod of W e i ~ e r . Results ~~ of 2% or less are probably not significantly different from zero.
..
610
OSCART. QUIMBY ROUGHANALYSES OF Sample
Oveno temp., C.
PRODUCTS O F
Heating, hr.
Vol. 58
TABLE VI11 TO '/a INCH LAYERS O F HEXAHYDRATE I N OPEN DISHES
HEATINGl/g
Ha0 loss, %
Producb
l%HA. Molybd. total
PaOs, wt. 5% Color Bell ortho pyro
~ ~ tri
1 Moles 1 pyre Moles ortho
20-77 R.T. None None 10.0 45.6 .. N.F. 41.2 95 6 10.5 8.5 .. 4 17 25 2 28-3 A .. 14 34 3 1.2 28-33 95 20 19.0 7.75 28-35 95 69 19.9 7.79 .. 17 34 10 1.0 28-3N 95 162 19.6 7.82 .. 16 39 3 1.2 20-80A 105 0.5 1.8 9.73 44.1 1 6 39 ... 20-80B 105 1 2.9 9.51 46.3 3 11 32 2 20-8OC 105 2 3.6 9.22 46.4 3 11 32 2 45-24A 105 2.5 13.0 ... 51.9" 8 24b 17 1.5 45-24B 105 24 19.4 ... 57.6" 16 38' 0 1.2 20-80F 105 70 20.1 8.23 56.3 13 42 6 1.6 20-80G 120 0.5 9.3 8.78 .. 7 13 28 0.9 20-80H 120 1 17.2 8.23 54.4 12 36 6 1.3 20-801 120 2 19.6 8.00 54.4 12 36 6 1.3 20-8OL 120 70 20.5 8.20 57.0 10 41 7 2 28-5 120 6 18.5 7.9 .. 11 3gd 7 1.8 120 G 19.9 7.95 55.6 8 40" 3 2.5 45-1 1 a Two end-point determination of total PtO,, c Andrews.36 Quantitat,ive X-ray analysis18 (Be0 internal standard) revealed 23% Na4P107, 19% NabP3010.6H20, no &a&Olo or orthophosphate. X-Ray analysis (EkO): 41% Na4PzOt, nothing else crystalline. After 2.5 years aging in bottle, X-ray analysis (BeO) revealed 28% Na4Pp07 a larger amount of Na3HPZO7.H20, but no NasPsOlo or orthophosphate. X-Ray analysis (BeO): 43% Na4Pz07,nothing else crystalline.
+
TABLB IX OF PRODUCTS OF HEATING VARIOUS DEPTHSOF HEXAHYDRATE AT 95' ROUGHANALYSES Sample
Depth of cryst. layer, in.
Heating, hr.
28-38 '/8 28-3B /8 28-3C 2 28-3 D 8 28-33 /8 = /8 28-3F 28-30 2 8 28-3H 28-31 '/a '/8 28-35 28-3K 2 28-3L 8 28-3 M 1/8 28-3N 6/8 28-3P 2 28-3R 8 Ortho content probably
Ha0 loss, 7%
Product p H in 1% s o h
6 10.5 8.50 6 3.4 9.00 6 0.5 9.47 6 1.3 8.91 20 19.0 7.75 20 17.5 7.75 20 17.5 8.01 20 11.0 7.98 69 19.9 7.79 69 19.1 2.65 69 17.8 8.03 69 16.7 7.91 162 19.6 7.82 162 19.5 7.71 162 17.5 8.06 162 16.7 7.90 too high because sum of ortho
Molybd. total
..
but no formation of triphosphate. Thus, recondensation to triphosphate is unimportant a t 120" or below. I n fact, there mas probably some hydrolysis of the pyrophosphate, because both the spraydrying and the further heating at 120' caused the 0.98:l.OO molar mixture of NaH2P04 and Na4P207 t o drop in pH (Table VII). Turning back to the initial decomposition, which is said to occur according to reaction (b), one can say that it is qualitatively adequate, but is really an oversimplification. It calls for a product with a pH of 7.6-7.7, as compared with hexahydrate decomposition products having a pH of 7.7-8.2 (Tables VI11 and IX). Equation (b) calls for a water loss of 18.9%, as compared with observed losses up to 20.5%, depending on depth of hexahydrate crystals, temperature and time of heating. It also calls for a product containing an equimolar
+
4 0 0 0.5 14 27"
IN
OPENDISHES
--.
Bell tri.
17 25 2 41 .. 0 44 46.7 3 41 .. 34 3 .. 32 3 54.6 4 41 9 50.7 14" 41 5 .. 17 34 10 .. 21 36 1 .. 6 43 3 .. 10 44 1 .. 16 39 3 .. 18 38 2 .. 5 46 2 .. 10 43 3 pyro tri-Pz06 exceeds the total P206. 47.0
+
Pa?,, wt. % Color in Bell ortho PYro
Moles pyro Moles ortho
2
...
...
... 1.2 0.6 5 1.5 1.0 0.9 4 2 1.2 1.1 5 2
mixture of pyro- and orthophosphate, whereas hexahydrate decomposition products made at 95-120' usually contain relatively more pyrophosphate even in the early stages as shown by the approximate analyses in Tables VI11 and IX. One remarkable fact about such products of dehydrating hexahydrate a t 90-120' to a steady state is that the X-ray diffraction pattern shows sharp lines for but one species, namely, Na4P207.This observation has been reported by 0thers.'9.~~Attempts t o make the orthophosphate evident by slurrying such products with a little water, allowing them to dry slo~vlya t room temperature, and reX-raying the air-dried product fails to reveal orthophosphate in the vast majority of cases. Product 28-5 of Table VIII, like so many others with an oven water loss of 17-21y0 and a pH of 7.7-8.2, initially gave sharp diffraction linm for Na4P207only, plus
SOLUBLE CRYSTALLINE POLYPHOSPHATES
Aug., 1954
the usual broad halo showing the presence of considerable material amorphous to X-rays. After aging 2.5 years in a screwcapped bottle, quantitative analysis by X-ray diff ractionZ4showed 28% Na4P207plus what appeared to be an appreciably larger amount of Na3HP207.H20; no orthophosphate was detected; in fact, all lines were accounted for by Na4P207and Na3HP207.H20. Upon slurrying the aged 28-5 with water and air-drying, the sample now showed only diffraction lines for Na4P207.10H20; again no orthophosphate was evident. On the other hand, synthetic mixtures made from pure Na4P207and NaH2P04or from pure Na3HP207 and Na2HP04 in 1:1 molar ratio, slurried with water and air-dried a t room temperature or oven-dried always showed Na3HPz07.H20 Na2HP04and/or NazHP04.2H20. Thus, qualitative X-ray evidence says that orthophosphate is usually present in concentrations below that called for by equation (b). The rough analyses for pyro- and orthophosphate in Tables VI11 and IX tell a similar story. The pyrophosphate analyses are the more nearly correct the closer the triphosphate composition approaches zero, but are probably high by no more than 5% relative error in any case. The orthophosphate data, obtained by a colorimetric molybdate determination without separation from the condensed phosphates, are probably a little high due to hydrolysis of condensed phosphates during the development of the blue color; in the case of products 28-3F and 28-3H in Table IX, the ortho P205appears to be significantly high because too much P205 is accounted for. It is probable that the pyro/ortho ratios given in Table VI11 are approximately correct in most cases. This ratio usually exceeds 1:1.
+
TABLE X TRACER ANALYSES OF A PRODUCT MADEBY HEATINGA '/*-INCH LAYEROF HEXAHYDRATE CRYSTALSFOR Two HOURSAT 105' Phosphate species
Reported as
Tri NajP3010 Pyro NaiPzOr Ortho NsHPOj Moles pyro/moles ortho
Wt. %
Method of anal.
24 58 19 I 6
Isotope diln.18 Inverse iso. diln.18 Extractive c0lorim.3~
61 1
much more decomposed than sample 20-8OC in Table VI11 or sample 10-105B in Table X I is not known, but all three samples clearly show a molar pyro/ortho ratio much greater than 1/1. Another series of partially dehydrated samples was prepared a t 95-140" by choosing a time of heating that would yield a product about '/3 to decomposed. The products were analyzed for pyroand triphosphate by a modified Gerber and Miles titration method,27for orthophosphate by the Martin and Doty extractive procedure,a4and for total Pz06by the two-end-point method.36 The results (Table X I ) show that a t all temperatures above 95' the molar ratio of pyro- to orthophosphate is greater than 1.00. This is true even a t 150" where appreciable recondensation eventually takes place as indicated by rise in pH and by increase in intensity of Na5P3010(II) diffraction lines if the heating is prolonged. It is therefore evident by a variety of methods that products of dehydrating hexahydrate a t 90120' are extensively decomposed into pyro- and orthophosphates, but, unlike the solution hydrolysis of triphosphate, they often contain more moles of pyrophosphate than orthophosphate. It therefore seems highly probable that a t least two mechanisms are involved in this solid state hydrolysis of triphosphate. One probably is the simple hydrolytic cleavage of the triphosphate ion equivalent to equation (b) and the initial step can be represented thus 0-
0OPOOPO___-__________ 0 0 0 H+ OPOOPO 0 0H
OPO 0
+
H
0-
+ OPO H 0-
(d)
This may be followed by a second reaction such as
+ Na2HP04+Na4P207+ NaHzP04 (e)
Na3HP207
But some other reaction capable of giving more pyrophosphate must be involved. Conceivably it could be some such reaction as 2Na5P801o
+ H20 +Na4P2O7+ 2Na3HP207 ( f )
Ignoring the question of mechanism of such a reac-
TITRATION ANALYSES O F PRODUCTS
OF
TABLE XI HEATING'/2-INCH ..
Sample
Ove? temp., C.
Time of htg.
Prod. Ha0 content
10-29A 10-105A 10-105B IO-24B 10-24C 10-105C 10-24D
95 95 105 105 120 150 150
7 hr. 12 hr. 2 hr. 2 . 5 hr. 50 min. 20 min. 25 min.
19.0 11.5 14.0 13.5 12.8 5.3 4.7
Prod. p H 1% s o h .
8.6 8.1 8.4 8.1 8.3 8.3 8.0
As a further check a sample made by heating hexahydrate for two hours a t 105" was analyzed by recently developed methods known to be both reliable and specific for the species being determined.1*r34s35 Why this sample (Table XS is SO (34) J. B. Martin and D. M. DOBY,Anal. Chem., 21, 9G5 (1949). (35) H. W. Lampe, unpublished work, Research Department, Chemical Division, The Procler & Gamble Co.
-
L.4YERS O F
Total 2 EP
47.4 51.3 49.4 49.1 51.3 54.7 55.4
HEXAHYDR4TE I N OPEN DISHES
PzOr, wt. % Colorim. Titrn. ortho pyre
4.2 10.8 6.1 9.2 7.1 7.2 11.3
6.9 20.8 17.8 23,8 17.8 20.5 29.2
Titrn. tri
36.3 19.7 25.5 16.4 26.4 26.8 14.9
'
Moles PYro Moles ortho
0.82 0.95 1.47 1,29 1.25 1.42 1.29
tion, one finds that it is reasonably consistent with the observed facts as follows (1) The Na4P207phase is always detected by Xrays and orthophosphate is not. Recall that oven dwiiie: a 1/1 molar mixture of Na4P20, and NaH2PG4 &- of' Na3HP20, and Na2HPOa fails to give (36) J. T. R . Andrews, J . A n . Oil Cdrem. Roc., 91, 192 (1954).
(312
OSCART. QUIMBY
Vol. 58
+
Na4P207 amorphous but shows Na3HP207.H20 some hexahydrate in addition to considerable Naband Na2HP04by X-ray diffraction. P3olO(II),but a t 150-230' r\ra6P3Ol0(1I)was the (2) The material amorphous to X-rays may well only crystalline species evident. None of these be largely Na3HPz07 which is known not to crystal- products had a pH below 9.4. A somewhat similar lize readily. Also recall that an aged decomposi- product, 36 made by heating hexahydrate for two tion product (28-5 in Table IX) contained much hours in an oven a t 200", gave the following analyNa3HP20,.H~0. Some amorphous NazHP04 and/ sis or KaH2P04is probably present also. 94% NasPaOlo (isotope dilution'*) (3) Equation (f) together with (b) or (d) ac5 . 2 % Na4P207 (inverse isotope dilutionlg) 1 .8% NazHPOa (Martin and Dotys4) counts for water losses between 5 and 5.5 moles quite as well as the recondensation hypothesis An alternative mechanism of triphosphate de(equation (e)). composition might be a disproportionation in (4) The range of pH values 7.7-8.3 is better ac- which both more and less condensed phosphates counted for by a combination of equations (f) are formed. The only slight indication of such a with (d) or (b) than by (b) alone (Table XII). reaction is the failure of hexahydrate decomposiRecall that heating a 1/1 molar mixture of Na4P20, tion products to give a precipitate with tris-(ethyland NaH2P04 actually pushed the pH downward enediamine)-cobalt(II1) ion under the conditions from 7.6 (Table VII). specified by Wei~er.~gThe inhibitory effect is . large enough to prevent precipitation when 1-2 g. TABLE XI1 of pure triphosphate is added to each gram of deSOMEPERTINENT pH VALUES composition product and a suitable aliquot of the Molar pH of Salts mixed ratio 1% s o h . resulting solution taken for the test. While this Na4P207 NaH2P04 1:l 7.63 does not prove the presence of higher polyphosNaaHP~07 NazHPO4 1:l 7.63 phates, it is true that they can prevent precipitation NarP20, Na3HP207 1:2 8.11 of C O ( ~ ~ ) ~ H ~and P ~ therefore O , ~ , such compounds may be present. However, the amount of more (5) Most of the molecular weight data of Thilo condensed phosphate need not be large. Hence, and Seeman20 are accounted for. Thus, hexahy- decomposition of hexahydrate at 90-120 " should drate samples which have been heated at 95-98' so regarded as due mainly to reactions such as (b) as to lose 5 moles of water give molecular weights be and (f). close to the 193 demanded by equation (b) or (d). Those which have been heated a t 100-120Oso as to Reaction with Calcium Ions in Dilute Solutions lose 5 to 5.5 moles (18.9-20.8%) of water have moFor dilute systems containing CaC12 and Nabpalecular weights between 193 and 251, the latter 0 1 0 the boundary between homogeneous and heterovalue corresponding to complete conversion via geneous regions at GO" is shown in Fig. G ; it mas deequation (f). Only when hexahydrate is heated so termined turbidimetricallya7 after attainment of that more than 5.5 moles of water is lost does one steady state, which often required only 10-30 minhave to invoke the recondensation to triphosphate utes, but sometimes required 1-3 hours. The hetervia equation (e) or (g). ogeneous region is rather wide and comes close to Na4PzO7 2Na3HPz07 2NaeP3Olo HzO (g) the calcium axis, so that the area of clear solutions Recondensation has been detected a t temperatures is very narrow on this side (Fig. GA). Of more inabove 120" by X-rays. Thus, products made by terest is the behavior on the other side DE where, in heating hexahydrate for l/z hour a t 150" show little general, more than a mole of triphosphate per mole or no Na6P3olo(II),besides the Na4P20,plus amor- calcium is required to prevent precipitation (Fig. phous material. But similar products heated for a GB). The right branch DE of this curve shifts to much longer time, e.g., 70 hours, at 150" reveal an the right as sodium salts are added to the solutions increase in Na5P3Ol0(II)content by X-rays. Such (Table XIII), that is, more sodium triphosphate is samples have lost at least 5.5 moles of water and required to clarify the sodium rich solutions. This their pH has risen from a minimum near 8.0 (1/2 rightward shift increases with increasing concentrahour) to 8.4-8.6 (20 hr. or more). Thus, in dehyTABLEXI11 drating triphosphate hexahydrate the temperature EFFECT OF SODIUM SALTS ON SOLUBILITY OF CERTAIN CAI~ 120 O causes only hydrolysis, but higher temperaCIUM TRIPHOSPHATE PRECIPITATES tures may cause appreciable recondensation after the initial extensive hydrolysis. PsOia -6 concn. Concn. of mmoles/l. Lest it be supposed that reaction (a) is never a Kind of Na salt, Ca concn.. required t o Na salt e./l. mmoles/l. suppress pptn. factor another set of experiments must be mentioned. At 135-230' NaJ'aOlo(II) is formed if the None .. 1.2 1.2 water is quickly flashed out. As an example of Na2S04 5 1.2 1.5 this, hexahydrate in l/Z-inch depth in a 10-mm. testNaps04 10 1.2 2.0 tube was heated suddenly to temperature (oil-bath), None .. 3.6 4.0 held there for 10 minutes to allom the water vapor NazS04 2.5 3.6 5.5 Na2S04 to escape into the laboratory atmosphere, and the 5.0 3.6 8.0 samples chilled by means of an ice-bath. By X-rayNaCl 10 3.6 14.0 ing a t room temperature it was shown that such (37) J. A. Gray and I
