2
J . L. Copeland and L. Gutierrez
ctions of Condensed Phosphates with Molten Alkali Nitrates. tsichiometric I nvestigation' James L. Gopeland" and Leslie Gutierrez Deparlment of Chemisfry, Kansas State University. Manhattan, Kansas 66502 (Received June 5. 7972)
Publication costs assisted by the National Science Foundation
~ t o i ( ~ h ~ o m e t rofi e Lux-Flood s acid-base reactions of (NaP03),, Na3(P03)3, Na4(P03)4, Na5PsO10, and 10 with molten NaN03, and of Na3(P03)3, Na4P207, and NasPsOlo with molten LiN03 were studied. Kinetics investigations of reactions of Na3(P03)3 and Na~PsOlowith excess molten N a N 0 3 were 02)and by 3lP nmr analyses of performed by measurement of rate of total gas evolution (4NO2 quenched reaction residues. With NaN03 a t 400" and above, all phosphates indicated except P4010 were degraded to Na4PzO7 while P4010 at 325" was depolymerized only to Nas(P03)3. With LiN03 a cation effect was observed in that the final phosphate product was PO&, probably due to polarization of -PO-P- bonds in P ~ 0 7 ~Kinetics -. studies of the reaction of Na3(P03)3 with excess N a N 0 3 as solvent revealed two distinct stages. Stage I is the formation of Na5P3010 via a set of complex unidentified intermediates, among which is an apparently cyclic phosphate anion formed in large amount very early in the resonance, and a steadyreaction with a 31P nmr chemical shift very close to that of the parent state trace quantity of (PO&-. Some tentative concepts are advanced to interpret some intermediates fif stage I. Stage I1 is the direct depolymerization: 2P3010s(02-1 (kz), with second-order behavior in P3OlO5- and no detectable intermediates. At 400, 411, 425, and 437", kz is 3.63 X 5.14 x 10-2, 9.58 X 10-2, and 13.06 X 10-2 (arbitrary units)-l min-1, respectively, giving an activation energy of 34.1 kcal. During stage 11, 4N02 0 2 are evolved with pseudo-first-order behavior in total gas volume a6 constant pressure. Reaction of Na3(P03)3 with NaN03 proceeds very slowly, if at all, under vacum9,but proceeds equally vigorously under an ambient of either Ar or 0 2 . Rate of reaction of NasP3010 with h a N 0 3 appears unaffected by presence or absence of an ambient atmosphere.
+
+
-
+
Intrcoducticsn High-temperature systems containing oxides and/or oxyanions have bern characterized as acid-base systems by Lux2 and by Flood and Forland.3 In this concept an oxide ion donor is a base while an oxide ion acceptor is an acid. It bas been known for some time that the N o s - ion in molten nitrates is a good Lux-Flood base, decomposing by some as yet inccmpletely understood mechanism into NO2 and 0 2 gases in a 4:1 mole ratio, upon donating the 02uke, et al 4 felt that the basic mechanism might be NO3-$N02+ NOz' t NO3 -
0+ 0 4 M
+ -
+
+ 02-
(14
2N02 + 0
(Ib)
+M
(IC)
0 2
where M is a third body, and the equilibrium of eq l a is shifted to the right upon addition of a sufficiently strong oxide ion acceptor (Lux-Flood acid). Topol, et al.,5 have disputed the existence of NO2+ in nitrate melts, and have advanced an alternate hypothesis
These latter inirestigalors5 have also demonstrated the ability of NO&) atself to act as an acid, as in the reactions
+ NOzKOz(g) 4- NOz- + N o s - + NO(g) 2 ~ 0 ~ ( .+ g )X X ) * ~ -4b?03- -I-NO:,- + ~ r 2 0 7 2 2NOa(g) 4-02-+ N o s -
The Journal of Physicai Chemistry, Vol. 77, No. 7, 7973
(34 (3b) (3~)
Zambonin and Jordan6 claim that the 0 2 - ion cannot exist a t any appreciable concentration at all in equilibrium with nitrate melts. However, the purpose of the present investigation was not the nature of this mechanism, but rather the mechanisms of condensed phosphate depolymerizations resulting from treatment with excess nitrate melts acting in their known capacity as bases. Thus, for our purposes a reaction of the type of eg 2 will suffice to illustrate the net role of NO3- as an oxide ion donor in reaction, even though 0 2 -may never exist per se as an independent intermediate; i. e., 02- has been effectively transferred to the acid by 2N03- ions in excess nitrate melt. The family of condensed phosphates represents a very interesting series of potential Lux-Flood acids. Schematic structural formulas for the phosphates studied in the present work, and also orthophosphate, are as giver] (IVII). (The letter symbols OP, PP, TMP, 'T'TMP, and TPP
(1) Presented in part at the 8th Midwest Regional Meeting of the American Chemical Society, Columbia, Mo., Nov 1972. (2) H . Lux, 2. Electrochem., 45, 303 (1939). (3) H. Flood and T. Forland, Acta Chem. Scand., 1. 781 (7947). ( 4 ) (a) F. R. Duke and M. L. Iverson, J. Amer. Chem. SOC., 8 0 , 5061 (1958): (b) F. R. Duke and S. Yamamoto, ibid., 81, 6378 (1959); (c) F. R. Duke and E. A. Shute, J, Phys. Chem., 66, 2114 (1962): (d) F. R. Duke and J. Schlegel, ibid., 67, 2487 (1963): (e) F. R . Duke, J. Chem. Educ., 39, 57 (1962): (f) F . R. Duke in "Fused Salts," E. R. Sundheim, Ed., McGraw-tiill, New York, N. Y., 1964, pp 409-417; (9) R. N. Kust and F. R. Duke, J. Amer. Chem. Soc.. 85, 3328 (1963): (h) F. R. Duke and M. L. Iverson, Anal. Chem.. 31, 1233 (1959). (5) L. E. Topol, R. A. Osteryoung, and J. H. Christie, J. Phys. Chem,, 70, 2857 (1966). (6) (a) P. G. Zambonin and J. Jordan, Ana/. Left., 1, 1 (1967): (b) J. Amer. Chem. SOC.,89,6365 (1967): (c) ibid.. 91. 2225 (1969).
Acid-Base Reactions of Condensed Phosphates
21
~-!-. .' 3-
I.
orthophosphate
11.
(oP;, PO,,-
pyrophosphate
(PPL ~ ~ 0 , ~ -
the same as (P03)a3-, and P30105- berng somewhere between (PO&3- and P20T4Shams El Din, et ~ t l . have , ~ performed the more noted experiments concerning the degradations of various condensed phosphates and P4010 in molten salts. These workers have reported studies of the degradations of P40 1 0 , Na4Pz07, and Nap03 [apparently Na3(P03)sr but reported only as Nap03 by these authors] in various molten salts. Of particular interest to the present work were their potentiometric acid-base titrations of NaP03. The degradation was found to depend on both the base used and the molten salt solvent. Thus, when Nap03 was dissolved in a LiCl-KC1 eutectic and titrated with DIazO2 at 400", the acid-base reaction was represented as
2Po3- -I 02-
~
~
0
~
4
-
(4)
with no further degradation to Po43-. In molten KNO3, however, titration of Nap03 with NazOz apparently resulted in eq 4 followed by 0
V.
0
i5-
0
P2074-
+
02- ; . 12P043-
Finally, when Nap03 was dissolved in MN03 at 350" and titrated with any one of the bases C & - , HCO3-, HCOO -, (COO)22-, or CH3COO -, the apparent reaction sequence was
phosphorus "pentoxide"
3Po3- f 0'-
P4'AC
(61
f P30105
2P30105- f 02-;f3P2074-
VI1
8
polyphosphate
(classical " m e t a p h o s p h a t e " ) ,
Pxo3x+l(x+~)--* P,03,'- = xP0,-
(5)
as x = n+2
--+
y
will be used occasionally for brevity.) With the exception of P4O10, all these species are stable almost indefinitely in aqueous solution inem pH 7 and room temperature. Inspection of the structures reveals the existence of four basic types of phosplhate groups:7 ( a ) the isolated Pod3group (structure I), (b) the end group (as in structure 11), (c) the middle group (as in structures TI1 and IV), and (d) the branching point group (structure V). Species VI and VI1 are seen to be constructed from both middle and end groups. In many chemical respects the relative stabilities of such groups, appear to be ordered as isolated Po43group > end group > middle group > branching point group, apparently as' a result of decreasing resonance stabilization as more and more isolated oxygen atoms become involved ixn bridging bonds between phosphorus atoms.8 One also obrierves that, stoichiometrically speaking, the only differences among these types of groups are the relative lark or excess of 0 2 - ions. Thus, P20T4-becomes 21)043- upon addition of 0 2 - ; becomes P30106- upon addition of 02-, or becomes P 2 0 ~ ~+Po43- upon addition of 2 0 2 - , etc. In each case a P-0-P bridge would 1~ ruptured in some way by an attacking 0 2 - ion in some form. Based upon the relative ordering of chemical stability of the groups just given, then, one would expect ii Lux-Flood acid strength arrangement to be ordered in the reverse sequence. Thus, for the phosphates discussed here, acid strength in the presence of a given base (NOS- in the present work) should be P4010 > (PO3)s3- > P 2 0 7 4 > P043- with (P03)s4-being about
(7)
followed by eq 5 . The extent of depolymerization of the condensed phosphates was also found to depend on the associated cation. Titration of P4010 in fused KN03 at 350" was complicated by rapid initial degradation of P4OI0 by NOS-, so that only the degradation product could be titrated by any of the previously mentioned bases. This initial reaction was suggested as being P4010
+ 4N03- -4P03-
C 2N205(g)
($1
Markowitz, et ~ t 1 . , ~have 0 also studied acid-base degradations of some alkali pyrophosphates and metaphosphates with C104- as the base. Lip03 and KPOs were of the long-chain form of "metaphosphate" (structure VU) while Na3(P03)3 was the trimetaphosphate (structure 111). Degradation depended on the associated cation. Lip03 was degraded completely to Li3P04, whereas N a d P 0 3 ) ~ and K P 0 3 were depolymerized only to the pyrophosphates. These studies apparently indicated the following reaction sequence
2MP03
+ 2MC104
+
M4P287 -I C l ~ ( g+ ) (7/2)Oa(g)
(9)
and further, in the case of M = Li M4P207
+ 2MC10.4
4
2M3P04 -I Clz(g)
* (7/2)02(g) (10)
Traces of Li5P3010 were found, but its existence was at(7) J. R. Van Wazer, "Phosphorus and Its Compounds," Vol. 1 . lnterscience, New York, N. Y.,1958, pp 390-393. (8) Reference 7, pp 437-441. (9) (a) A . M. Shams El Din, Electrochim. Acta 7, 285 (1962); (b) A. M. Shams El Din, A. A. El Hosary, and A . A . A. Gerges. J . Eiectroana/. Chem., 6 , 131 (1963); (c) ibid., 8 . 312 (1964); (d) A . M. Shams El Din and A. A. A. Gerges, Eiecfrochim. Acta, 9, 123 (1964); (e) A . M. Shams El Din and A. A . El Hosary, ibid.. 13, 135 (1968); ( f ) A . A. El Hosary and A. M . Shams El Din, ibid.. 16, 143 (1971); ( 9 ) A. M. Shams El Din, H . D. Taki El Din, and A. A . El Hosary, ibid., 13, 407 (1968). (10) M. M. Markowitz, H. Stewart, Jr., and D. A . Boryta. inorg. Chem.. 2, 768 (1963). The Journal of Physlcai Chemistry, Voi. 77, No. 7 , 7973
22
J . L. Copeland a n d L. Gutierrez
tributed to
ratus filled with an Ar atmosphere (or with 0 2 in one spe:!MPDs + M3Po4 M~P3010 (11) cia1 case), initially at 1-atm pressure. This apparatus is described in detail elsewhere.16 On the basis of their study, Markowitz, et a l , assigned Infrared spectra of crystalline samples in KBr pellets, the relative acadity of the alkali pyrophosphates and meand of evolved gas samples, were performed with a Pertaphosphates as decreasing according to Li+ > > > Na+, K + kin-Elmer Model 337 grating infrared spectrophotometer. = 0 for P20T4-,and Li7- >>> Na+ > K + for PO3-, and fiGas chromatography runs were performed using a localnally meta >>pyro. ly fabricated gas chromatograph with e carrier gas. Mass In view of the observed spontaneous reactions of many spectra of evolved gas samples were obtained from an condensed phosphates with molten NaN03, and the foreElectronics Associates Inc. Model 250 quadrupole mass going background, the objectives of the present work were spectrometer. formulated as follow. Overall stoichiometries of the reacAll nmr analyses were performed using a Varian Assotions of Nai3(l3O3)y, Na4(P03)4, NasP3Olo (form I1 of ciates Model XL-100 nmr spectrometer provided with a structure 'VI),11 (NaP03)x (insoluble and glass forms of 31P probe and deuteron lock, necessitating use of D2O as structure VII),ll and P.4010 with N a N 0 3 were determined, sample solvent (a proton lock had not yet been received as well as the reactions of Na3(P03)3, NasFsOlo, and Na4for our instrument). The nmr spectra were accumulated P207 with LiN03. These studies were performed using with a Varian Associates Model C-1024 time average comweight loss techniques with careful analyses of evolved puter. Samples were contained in Wilmad .Glass Co. gases and crystalline residues. Kinetics of the reactions of 12-mm 0.d. nmr sample tubes with plastic caps. The Na3(P03)3 and 'Ma&&lo with excess molten NaN03 solchemical shift reference signal, at 0.0 ppm, was provided vent were studied by measurements of the rate of total gas by 85% H3P04 in a small capillary tube placed in the un0 2 ) evolution, and by 31P nmr analyses of aque(NO2 known sample tube. This reference sample was preserved ous solutions uf quenched reaction residues. The results and was the same one used for every nmr determination. shed some interesting light on the mechanism of an overA constant high-temperature bath, for quenched reacall reaction of Ihe type of eq 6 and ?, although much still tion studies, was comtructed in the form of a thermostated, stirred bath of molten NaNQ3.16 remains to be done. All temperature measurements were performed using a ~ x ~ ~ r ~ mSection enta~ Leeds and Northrup Co. No. 8691 millivolt potentiometer with chromel-alum.el thermocouples. An ice-beth referMaterials. Rcagent grade P4010 and the insoluble form ence junction was employed. of (NaP03), were obtained from the Baker and Adamson A Vacuum Atmosphere Corporation Model Co. Purified grade Na5P3O1o (form 11) was obtained from Lab drybox, provided with a dry Ar atmosphere, was used the Fisher Scientific (30. and reagent grade NaN03 and for all experimentation with Pq 10, and for handling of LiN03 were from the Mallinckrodt Co. These chemicals LiNQ3. were used without further purification other than oven Reaction Stoichiometrjl Procedure. n the assumption drying at ca. 125" for a t least 24 hr and storage in a Drier(subsequently confirmed) that NO3 - composes in acidite charged dessicator just prior to use. base reactions according to an overall reaction scheme of Na3(PO& was prepared by thermal treatment of the type of eq 2, stoichiometries of the reactions of Na3NaHzP04.H2(3 and was purified by dissolving in dis(P03)3,Na4(P03)4, NasP3Olo (form II), ( tilied water, filtering off any insoluble "metaphosphatz," uble and glass forms), and P4010 with Ma and recrystallization from ethanol.12 Na4(P03)4 was prepared by the low-temperature hydrolysis of P ~ Q ~Purio . ~ ~mined by a method based on weight loss due to evolved NO2 and 0 2 , and analyses of the fino1 solid residues. In ties of these prepared cyclic metaphosphates were estabaddition, similar investigations were perform.ed for Na3lished as a t least 95% from crystalline infrared spectra (Po3)3,NasP3010, and Na4P207 with LiNQ3. and from the occurrence of a single 31P nmr peak for an In the cases of Na3(P03)3 and (NaPOa)x, intimate aqueous solution at ca. f 2 1 ppm from 85% H 3 P 0 4for Na3(PO&, and at ea. -1-23.2 ppm from 85% H3P04 for Na4- mixtures of a phosphate with NaN03, of varying weight ratios, were prepared and weighed in C , o r s No. 00 porce(PC)&.L4 The glass form of (NaPQ31, was prepared by lain crucibles. The crucibles were placed in a crucible furfusion of the insolubk form in a platinum test tube foinace of fixed temperature, and the samples were allowed lowed by quenching with cold water. to react until no further weight losses were observed when Oxygen and ar,qon gases, used as ambients in gas evoluthe crucibles were withdrawn, cooled in a dessicator, and tion studies, were Crom the National Cylinder Gas Co., reweighed. A similar procedure was followed for NasPjOlo and were specified ab being of better purity thsn 99.9%. and Na*(P03)4 except that 13 x 100 mix Pyrex test tubes These were used without further purification except for placed inside 23 X 230 mm Vyeor test. tubes served as being passed tliroiigh a Mg(C104)2 charged drying tower reaction vessels. These were placed in the constant higha t the time of' use. temperature bath rather than a crucible furnace. No obApparatus and Instrumentation Rate of total gas evoluservable attack occurred to either the crucibles or the tion studies for lu'aa(P03)3 and NasPjOlo with NaN03 Pyrex test tubes. were performed by measuring the total volume o€ gases evolved as a function of time at fixed room temperature, and under a constant 1-atm total pressure. The apparatus ( 11) Reference 7, pp 605-606. was constructed in our laboratory, and was very similar to (12) R. N. Bell, lnorg. Syn., 3, 103 (1950). ( : 3 ) R. N. Bell, L. F. Audrieth, and 0. F. Hill. l n d . Eng. Chem.. 44, 568 that employed by Freemanl5 in his investigations of ther(1952), rad decompositions of NaNO2 and NaN03. Our appara(14) M. M.'CrutchfieId, C. H. Dungan, J. H. Letchsr, V. Mark, 2nd J. R . Van Wazer, "31 P Nuclear Maanetic Resonance," Interscience. New tus was provided with accessory traps and fittings which Y o r k , N. Y . , 1967. permitted sampling of the evolved gases for analysis at (15) E. S. Freeman, J. Phys Chem , 6 0 , 1487 (1956). the end of a run. Each run was performed with the 'appa(16) L. Gutierrez, Ph.D. Thesis, Kansas State University, 1972 + +
+
The Journai of Physical Che,vistry, Val. 77, No. 1, 7973
Acid-Base Reactions oi' Condensed Phosphates For P4OI0 with N a N 0 3 in the drybox, mixtures of varying weight ratios were prepared in the test tube vessels, which were then stoppered. These assemblies were placed in the constant high-temperature bath, and the stoppers were removed about 15 sec later. After reaction, each tube was flushed with dry Ar while still in the bath, stoppered, removed to the drybox, and reweighed. Again, no attack on the test tubes was noted. Reactions of phosphates with EiNO3 were performed using the test tube assemblies, however, the LiN03 was handled in the drybox. No attack on the test tubes was noted. Solid residues remaining after each run were analyzed by infrared spectroscopy utilizing KBr pellets, and by 31P nmr analyses 'of aqueous solutions. Evolved gas analyses were performed with the aid of the constructed gas line in conjunction with the rate of total gas evolution studies, to be discussed later. For each reaction series of a phosphate with NaN03 or LiN03, a plot of weight loss per gram of nitrate us. initial weight ratio of phosphate to nitrate gave a straight line of positive slope from the origin for weight ratios less than the stoichiometric ~ q u ~ v a l e n cpoint, e and a horizontal straight line for weight ratios greater than this point, The intersection of these lines gave the stoichiometric equivalence point in terms of weight ratio, which was then converted to a molar ratio. The (NaP03)x, Na3(PQ3)3, and Na&Olo reaclions were run at 450", the Na4(P03)4 reaction was a t 4CO", and the P4010 react,ion was a t 325". 400" was chosen for Na4(P03)4 to preclude any possible converto (PO3333-, known to occur slowly at ca. sion of (P6)3)44-~ 450"."[ 325" wm used for P 4 0 1 0 since this substance sublimes a t 360".xs Gas Evolution kat^ Procedure. The gas evolution apparatus was employed for this study.16 In the case of the Naa(PQ3js -t :NaNOs reaction, 1.7000 g of NaN03 was placed in the outer portion of the reaction vessel, and 0.1109 g of N ~ i ~ ( was P 0 held ~ ~ ~ suspended in the inner portion of this vessel., just above the nitrate.16 The entire system was evacuated for ca. I hr, and was then filled with dry Ar (or dry 0 2 in one special case) at 1-atm pressure. An electrical cylindrical furnace was raised to encircle the reaction chamber, causing fusion of the NaN03, and it and the phosphate to attain a common, specified temperature. The 1\1;13(PQ3)3 sample was then made to fall into the excess molten NaN03 (phosphate:nitrate mol ratio of ea. l:55), and timing commenced. At various times total gas volume (Ar iNO2 QZ), at constant 1-atm total pressure :and room temperature, was observed on the gas buret. The total gas volumes so obtained were TP and. plotted us. time. At room temperature the 2NQ2 = N&4 equilibrium was decidedly complicating and was unablle t o be corrected for in these studies, as well as in the similar ones using Na5P3019. This definitely limits the value of the gas evolution results to semi-. quantitative pLirpnseci. Reaction runs were performed et 400, 411, 425, 4.37,arid 448". In each case the Na3(PQ& was totally soluble iri the excess fused NaN03, resulting in completely h ~ ~ ~ ~reactions. ~ e ~ ~ ~ o u s In ?he case of t h e NasPsOlo i- NaNQ3 reaction, the procedure was the same except that 0.4000 g of this phosphate was reacted wit,h 5.0000 g of molten N a N 0 3 (still a 1:55 mol ratio) in each run, and two sets of runs were performed: one using oa, 250-,Uparticle size of NasPsQlo, and the ether using ca. 125-fi particle size. The particle size of t,his p ~ i c l ~ ~ had ~ I ~an , ~effect ,e on the reaction rate
+
23
since solubility in NaNO3 was not complete, resulting in some degree of heterogeneous reaction. This fact obviously renders the results somewhat less valuable than those obtained for the trimetaphosphate. Runs were performed at 401, 411, 418, 424, 435, and 442" for the ea. 250-p particles, and at 400, 411, 425, and 437" for the ea. 125-p particles. In both phosphate cases, the amount of excess NaN03 was made purposely large so as to be treated as a solvent of approximately constant concentration. The use of 0.1109 g of Na3(P03)3 and of 0.4000 g of NasP3010 was intentional, sirice these quantities depolymerize to Na4Pz07 with evolution of equal amounts of gaseous products. The same apparatus was employed for completely reacting all the phosphates studied with NaN03 and collecting the total gases evolved for analyses. These gaseous products were analyzed by infrared spectrometry, mass spectrometry, gas chromatography, and quantitatively for N&(N204) by the technique of Whitnack, et a1.,19 using standard solutions prepared according to the methods of Skoog and West.20 Quenched Reaction Rate Procedure. Reaction vessels consisted basically of 23 X 230 mm Vycor test tubes. Each such tube contained 1.7000 g of NaNO3 for the reaction with Na3(PO3)3, or 1.6400 g of NaNQ3 for the reaction with NasPsOlo, These tubes were thermostated in the constant high-temperature bath at appropriate temperatures. A dropping device, similar to that employed in the gas evolution rate studies,l6 containing either 0.1.109 g of Na3(Po& or 0.1333 g of Na5P3010, was positioned in each tube, and dry Ar was swept through each such assembly for ea. 3 hr. The Ar flow was terminated, and reaction was commenced by dropping the phosphate charge suspended above the molten NaN03 into the latter in each tube, as in the gas evolution work. A swirling motion was imparted to each tube for ca. 1 min to ensure rapid dissolution and mixing of the reactants. At various times reaction vessels were withdrawn and immediately thrust i.nto cold water, thereby quenching the reactions. The solid residue in each tube was dissolved in 5 ml of DzO, and the 31P nmr spectrum was obtained using 85% H3FO4 in a capillary as a standard. Twenty scans of a 1000-Hz sweep width spectrum were accumulated on the time average computer (CAT). Sweep rate was at 4 Hz sec-1. To increase the area of each resonance signal in the accumulated spectrum, the read out of the CAT was at 4 Hz sec-l and the spectrum was recorded a t 10 Hz sec-1. This had the effect of multiplying each of the accumulated peak areas by a factor of 2.5. A reference spectrum from 0.1109 g of Na3(PO& and 1.7000 g of NaN03 in D20, or from 0.1333 g of NasP3Q10 and 1.6400 g of NaN03 in I 3 2 8 was also obtained for the zero reaction time point. No observable attack occurred to the Vycor reaction vessels. Since 3IP nmr resonance signals are nicely identifiable by chemical shift,l4 end groups occurring at ea. 4-1 ppm upfield from &Pod, and middle groups a t ca. 18-24 ppm upfield from H S P O ~it ~was a simple matter to obtain the phosphorus content due to each such general type of group by peak area integrations. These areas were normalized by dividing each by the reference peak area due t o (17) D. E. C.Corbridge, M. S. Pearson, and C. Waliing, Top. PhosphorusChem., 3, 265 (1966). (18) Reference 7, p 274. (19) G. C. Whitnack, C. J. Haiford, E. S. Gantz, and 6. B. L. Smith, Ana/. Chem., 23, 464 (195'1). ( 2 0 ) D. A. Skoog and D. M. West, "Anaiytical Chemistry, An Introduction." Holt, Rinehart and WinStQn,New York, N . Y., 1965.
The Journal ot Physical Chemistry, Vol, 77, No. 7, 1973
24
J. L. Cooeland and L. Gutierrez
the constant amount of H3P04. En each case all phosphorus initially present in a reaction as either (P03)a3- or PaOlO5- was (accounted for, within experimental error, as end and/or middle group form with no PO43- other than that present in the standard H3P04. Not all specific phosphate species could be unambiguously identified by chemical shift and spin-spin splitting patterns. However, it was generally possible in the N a a ( P 0 3 ) ~case to trace the decay of (PO&3-, growth and subsequent decay of Pa0106--,and ;growth of' final p & ) ~ ~ product throughout the course of a run. In addition, some ( P o & - was found. An unidentified species consisting of all equivalent middle groups, close to (P03)33-, and some complex splitting patterns due to one or more unidentifiable transients were found during early times in a run. In the case of the Na5Pa010 reaction, all phosphorus was accounted for as Pa0105- and/or p & ~ ~ during - the course of a run. The latter' reaction was not quite so clean, kinetically speaking, being complicated by the solubility problem of NasP3010 in NaN03 as mentioned before. A given run consisted of 15 to 20 quenched reactions a t various times, and runs were performed at 389, 400, 411, 425, and 437" for N a ~ ( P 0 3 ) 3and , a t 389, 400, 406, and 411" for NaePaO~o.0.1333 g of NasPsOlo was chosen €or reaction with 1.6400 g of N a N 0 3 since these amounts represent the quantities of these substances that would remain if 0.1109 g of IVa3(P03)3reacted with 1.7000 g of NaNO3 to produce only Na5PaOio. As in the gas evolution studies, the mole ratio of N a ~ ( P 0 3 ) sto NaN03 was 1:55; however, the mole ratio of Na5P3Olo to NaN03 was only about 3 :53 for the reasons just cited.
Results andl ~ ~ S C u S S ~ Q n A. Reaction Stoichiometries. To within experimental error (ea. &5%) weight loss experiments, with analyses of solid and gaseous products, indicate that the reactions of the various phosphates with NaN03 are (NaPO,),
+
---450'
xNaNO,
(x/%!ha,P,O, 4- xNO,(g) I- (s/4)0,(g) 2Na3(P03),
+
(12a)
450'
GNaNO,
3Na,P,O,
+
6NO,(g)
+
+
2NO,(g)
+ 0.502(g)
1.502(g)
(12b)
150'
2Na,P30,, -i- 2 N a N 0 , 3Na,F,07
Na,(PO,),
+
400'
INaNC),
+
2Na,P,O, -I- 4NO,(g)
+
(12c)
O,(g)
(12d)
P,O,, I- 4Nah10,
325"
-----
(4/3\Na,(PO,),
+
4NO,(g)
+
O,(g) We)
where (NaP03)z denotes both glass and insoluble forms of polyphosphates. The reasons for the use of 400 and 325" for the Na4(P&)a and P401o reactions, respectively, have aiready been explained. It is apparent that the Na3(PO& 10 must be reasonably stable to further depolymeri sation at 325", since no additional detectable reaction occurred within a reasonable time period. Treatment of this trimetaphosphate residue with additional NaN03 at higher temperatures yielded Na4P207 in accordance with eq 12b. Reactions of Na4P207, Na3(P03)ar and NasP30lo with LiN& occur as The iournal of P h . m - a i Ciiemistry,
vol. 77,NO.
7, 7973
420"
P2OY4-+ 2 N O , - 7 2POd3-
-
Ll
+ 6N0,-
420'
+
2N02(g)
+
0.502(g)
/3P0,3- + 6 N O d g ) -t-
L,
(13a)
1.502(g) (13b)
+
P,0,,5-
+
420'
4NO,--
3F043- + 4NO,(g)
L1+
+ O,(g)
(13~)
where the mixed Na+ and Lif cations are omitted for clarity. Complete details of the data and plots leading to these results are found elsewhere.16 A strong cation effect is apparent. A reasonable interpretation of the results using LiN03 is that the small Li+ ion, with its high ionic potential, is capable of weakening P-0-P bridge bonds, via induced polarization, in any P20 ~ that ~ may - form as an intermediate. Such weakened bonds would be more susceptible to basic attack by 0 2 in whatever form it may exist. Further details of the reactions in Li+-containing molten media were not studied at this time. B. Gas Evolution Rate Studies. Figure 1 contains plots 2 002, ~ ~ corof total volume of evolved gases, N ~ 2 ~ N + rected to STP, us. time for the reaction of Nas(PO& with molten NaN03 a t 400, 411, and 425". Curves for 437 and 448" are similar in shape, but cannot be shown because of crowding. An expanded time scale piot would show their like profiles. Data for these plots, and similar ones for the NaN03 reaction, are tabulated elsewhere.lG NasP3010 From the form of the curves it is apparent that the rate of total gas evolution is characterized by two stages. In stage I initially rapid production of total gas i s followed by a deceleration in this evolution. At the onset of stage II the evolution rate is accelerated once again, followed by relaxation a s the reaction approaches completion (conversion to Na4Pz07). Reaction times a t the inflections between stages I and 11, obtained from intersections made by extrapolating upper and lower portions of stages I and I1 for each curve, are 29.0, 19.0, 11.5, 5.5, and 2.5 min for the 400, 411, 425, 437, and 448" plots, respectively. These times will later be seen to correlate significantly with the times of occurrence of maximum P3Q105- intermediate concentration in the reacting systems, as determined by the nmr quenched reaction studies. Indeed, from the nmr studies it will be seen that the reactions occurring just after these inflection point times, and proceeding on to completion (stage 11), are essentially simple second-order - , no other phosbreakdowns of P301o5- to P ~ O T ~with phate species being detectable in the systems. Thus, the gas evolution curves during most of stage BI are in reality the curves for the reaction of Na5F$&o with excess NaN03. Based on this latter observation, if one sets an adjusted time scale with t = 0 just after the inflection points on curves of the type of Figure 1, one has total gas evolution us. time plots for the NasP30,o reaction with NaN03 without the complication of the dissolution problem of solid tripolyphosphate in the molten nitrate, which occurred when these reagents were mixed directly, It is of empirical interest that the corrected total volume of evolved gas, Vt, at time t, during this stage I1 reaction o f NasPsOlo with NaNO3 obeys the equation
+
vt
=
V,(1 - e - k t )
(14)
where V , is the final total volume at infinite time and k is a constant. Plots of In (1 - Vt/V,) us. t permit calculations of the constant k as 4.59 X 10-2, 7.74 X 10-2, 10.8 X 10-2, 15.2 X 10-2, and 20.2 X 10-2 inin-1 a t 400, 411, 425, 437, and 448", respectively. A plot of In lz us. 1/T (T
Acid-Base Reactions of Condensed Phosphates
25
________?
I-’ Q
>-a0 u)
a
15 -0
a>
z >
IO
w
0
>
5
c
0 t - 0
a
25
50
75
IO0
Reaction Time, t (min) Figure 1. Total voluime of gas evolved at constant pressure, corrected to STP, V t , vs reaction time, t , for the Na3(P03)3f NaN03 reaction at 400. 411, and 425”. All open points are for r u n s under an AI ambient, while the solid points at 425” are for an 02 ambierit in “K) i s fairly linear** with a slope of -6.33 X 103 deg, yielding a pseudo “activation energy” of 28.9 kea1 mol-*. Direct reactions of NasP3010 with NaN03 are roughly approximated by eq 14 also. At a given temperature k for ea. 125-p Na5P30lo particles is distinctly smaller than that for stage I1 of the Na3(PO3)3 reaction, and k for 250-p particles is sitill smaller. As temperature is increased, k for both particle sizes approaches k for the stage I1 process, reflecting the decreaising rate determining role of the dissolution process which complicates the direct reaction of NasPsOlo in NaNO:!, as mentioned before. Equation 14 and its derived properties are of empirical observational interest only for both stage 11 of the Na3(P03)3 reaction and the direct reaction of Na5PsOl0, since this equation is based on the rate of production of total gaseous products (further coniplicated by the 2N02 = N204 equilibrium) rather than on amounts of remaining reactants. In view of this, it is not deemed worthwhile to devote additional time or space to further discussion of this aspect of the study which cannot lead to any substantive conclusions or interpretations at this time. Complete data and plots for the foregoing are, therefore, available elsewhere.16 It was discovered that the reaction of Na3(P03)3 with NaN013 in u m r m proceeded to an imperceptible extent, if at all. When stoichiometric amounts of these reagents [ca. 0.2400 g of Na::(P03)3 and 0.2000 g of NaNOs] were thoroughly mixed and reacted a t ca. 412” for 90 min under vacuum, no noticeable gaseous products were formed, and the 31P nmr spectrum of the residue showed that little of the t r i m e t a p h ~ ~ ~ phad ~ a ~been e degraded. When Na3(PO313 was canrsed to react with excess NaN03 (1:55 mole ratio) under vac*uum a t 412” for 60 min, examination of the 31P nmi spectrum of the residue showed that very little 1’2074- had formed, and that most of the phosphorus was in the form of P30105-. When the same starting reaction mixture at 412” was under a n Ar atmosphere, the conversion to 1 ’ ~ C ) i 7 ~ - was nearly complete after 60 min. Since the prescmce 01” absence of an ambient atmosphere had such marked effects on the rate of the N a ~ ( P 0 3 ) 3f NaNO3 reaction, an attempt was made to ascertain if the nature of the ambient was critical. A gas evolution rate study was performed for the reaction at 425” under a pure 0 2 atmosphere. Some points for this run appear as solid points on the 425” isotherm in Figure 1, and it is seen that there is negligible difference between the effect of using
Ar or 0 2 as an ambient for the reaction. No attempt was made to study the effect of ambient pressure (both Ar and 0 2 were a t ca. 1 atm), nor was the effect of a possibly more reactive gas, such as NOn, investigated at this time. Thus, the presence of an ambient atmosphere is essentiat for the reaction to proceed at a reasonable rate, but the nature of this ambient would seem to be of little importance, assuming no direct chemical reaction of the gas of course. It is unlikely that any trace of moisture in the gas could have assisted the reaction, since in all cases gases were passed through a magnesium perchlorate drying tower prior to entry into the system, as discussed earlier. There appeared to be no effect at all on the rate of the NasP3010 + N d 0 3 reaction if it was run under vacuum or an ambient gas, Thus, one must conclude that there is a fundamental difference between the initiating mechanism for the degradation of (P03)a3 by N o s - and that for P30io5- by Nos-. This will be made more apparent in the quenched reaction studies. Gas evolution rate data and plots for the direct reaction of NasPsOlo with NaN03 are not recorded here, but are available elsewhere.l6 Because of the complication of incomplete solubility already discussed, these results are inconclusive. However, as mentioned, plots of corrected total evolved gas volume us. time do follow eq 14. Fortunately stage 11 of the Na3(P03)3 f NaN03 reaction corresponds, as discussed, to the simple reaction of NaN03 with NasPsOlo produced as an intermediate. Thus, uncomplicated runs of this reaction are already available as parts of the trimetaphosphate reactions This will be more apparent from the quenched reaction results. Quenched Reaction Studies. Figures 2-4 consist of a series of time-averaged 31P nmr spectra of the residues of quenched reactions of Na3(P03)3 with excess N a N 0 3 at 400” at various times. The actual run consisted of 23 such samples at times from 0 to 105.00 min. However, for brevity here only representative highlight spectra are shown for 0, 1.00, 5.00, 8.00, 18.00, 27.00, 32.00, 41.00, 52.00, and 105.00 min. The series of spectra for runs a t all other temperatures are similar to that for the 400“ experiment. On each spectrum the common 85% Pi~P04reference peak occurs at 0.0 ppm and is labeled QP. At time zero, the blank residue consists of the single peak at ea. 20.6 ppm due to the three equivalent phosphorus atoms in the cyclic trimetaphosphate anion (TMP). At 1.00 min, the unreacted T M P peak is, of course, still present, a trace of cyclic tetrametaphosphate (TTMP) has arisen at ea. 22.9 ppm, and a pronounced, unidentifiable middle group peak at ca. 21.2 ppm has appeared, which is labeled CP for “complex” phosphate. Since virtually no end groups have yet formed, this GP peak must be due to a new, apparently cyclic phosphate structure consisting of all equivalent middle group phosphorus atoms and very similar to TMP. Indeed, in other runs at shorter times there is not even a trace of end groups present, only TMP, CP, and a trace of T T M P accounting for the total phosphorus content. (The T M P and T T M P peaks were identified as being such by later adding pure TMP and T T M P to the residue solutions and observing that the peaks so labeled grew appropriately in areas.) As time progresses, the T M P peak diminishes while CP grows, TTMP not changing significantly. In addition, unidentifiable end groups begin to appear between 4 and 5 ppm, as seen, for example, at 5 min. By the time 8 min have elapsed, the doubly split end groups of tripolyphosphate (TPP) are definitely showing a t ea. 5.4 and 5.9 ppm, and the single The Journal of Physical Chemistry, Vol. 77.
No. 1, 1973
J. L. Copeland and L. Gutierrez
26 TMP
OP
0 r'
0 mln
1.00 m l n
OP
O? CP
1
is.00 m i n
. 1 L - i i -1 0 I 4 5 6
l
i
7
1-__ J - _ L -L 17 I8 19 20
Figure 3. Series of time-averaged 31P nmr spectra of residues of quenched reactions at 18.00, 27.00, and 32 00 min. See capI
0
-I
4
7
6
!3
20
19
22
21
tion to Figure 2.
23
OP
CP
I !
--I -I
Q
LL
-..---LA
I
4
!$
6
r
19
20
' 21
22
23
Figure 2. Series of time-averaged 31P nmr spectra of residues of quenched reactions of the Na3(P03)3 4- NaN03 reaction at 400" at 0, 1.00, 5.00, and 8.00 min. 85% H3P04 is reference (OP) peak; clhemicai shift is recorded in ppm.
peak from tbe equivalent end groups of pyrophosphate (PR) is appearing a t ca. 6.4 ppm. These latter peaks continue to grow while TMP, CP, and T T M P diminish. All the while unidentifiable end and middle group peaks are growing in imd/or diminishing, giving fairly complex spectra as, for example, at 18 min. By the time 27 min are attained, T P P has essentially reached a maximum value, much of the unidentifiable phosphates have vanished, and the triplet peak of the middle group of TPP becomes more The Journal of Physical ,Phe!mistry, Vol. 77, No, 1, 1973
visible. From this time on to the end t
