Published online 8 June 2007
Published in Soil Sci Soc Am J 71:1137-1140 (2007)
DOI: 10.2136/sssaj2006.0109
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
SOIL CHEMISTRY
On the Solubility Constant of Strengite
Mauro Iuliano*,
Liberato Ciavatta and
Gaetano De Tommaso
Dipartimento di Chimica, dell'Università Federico II, via Cinthia 45, 80126 Napoli, Italy
* Corresponding author (miuliano{at}unina.it).
 |
ABSTRACT
|
|---|
Strengite, FePO4·2H2O, commonly occurs in soils and represents a source of P for plants. The solubility of strengite is thus considerably important in the field of agriculture and environmental geochemistry, but few solubility data are available. Using absorption spectrophotometric techniques, the solubility equilibrium was studied at 25°C by measuring the total Fe(III), mFe(III), dissolved while keeping strengite in contact with H3PO4 solutions. The acid concentration, mP, ranged from 0.001 to 0.1 mol kg–1. In solutions of mP 
0.01 mol kg–1 mFe(III), results were constant within the limits of experimental error. This was ascribed to the predominance of a soluble species FePO4(aq). Soluble complexes, mainly Fe(H2PO4)3(aq) and FeH3(PO4)2(aq), were responsible for the increased solubility at mP 
0.02 mol kg–1, as evidenced in previous investigations. Using the constants evaluated in these investigations, the solubility data can be explained with the ion product equal to 10–6.70 mol–1 kg at the infinite dilution reference state.
|
INTRODUCTION
|
|---|
Strengite, FePO4·2H2O, commonly occurs in soils (Sposito, 1989) and represents a source of P for plants (Tiessen, 1995). The solubility of strengite is thus considerably important in the field of agriculture and environmental geochemistry, but few solubility data are available. This is probably due to difficulties inherent in preparing pure strengite. Often, preparations contain other phases such as phosphosiderite or colloidal ferric(III) phosphate. In addition, the intrinsic ions tend to form a series of complex compounds. Finally, the presence of extrinsic ions in the test solution causes phase transformations.
The first attempt to determine the solubility product of strengite was by Chang and Jackson (1957), who concluded from pH and solubility measurements that at 23 to 25°C, the equilibrium constant for
 | [1] |
had a value between 10–33.5 and 10–35. For strengite and amorphous FePO4, Egan et al. (1961) measured heat capacities at temperatures between 8 and 310 K and the heat of formation at 298 K. From the calorimetric data, they calculated a constant of 10–34.56 for Eq. [1]. Nriagu (1972) measured the solubility of strengite in dilute H3PO4 solutions of molality ranging from 0.04 to 0.1 m. Assuming the protolysis constants of H3PO4 and the complexation reaction Fe3++H2PO4–
FeH2PO42+, Nriagu (1972) found a constant of 10–34.88 for Eq. [1]. This agreed well with results from Egan et al. (1961). The agreement, however, may have been fortuitous. The value by Egan et al. (1961) is the result of numerous steps in the combination of thermochemical data, so that its estimated uncertainty is quite great. Furthermore, recalculations of Nriagu's data by taking into account more detailed models on the Fe(III)–phosphate complex species in solution, as established by Ciavatta et al. (1992) and Ciavatta and Iuliano (1995), give for the equilibrium constant
 | [2] |
a value of 10–28.16. This value agrees well with the results of the present study.
The present investigation was undertaken to shed more light on the heterogeneous equilibria involving strengite. The crucial points are (i) the preparation of pure strengite and (ii) the detailed knowledge of the solution chemistry of Fe3+ ion in phosphate solutions. The solubility [i.e., the molality of total Fe(III), mFe(III)] was determined at 25°C as a function of the H3PO4 molality, mP, in the range 0.001 to 0.1. Thus, a complicated system was tackled without the use of the traditional constant ionic medium method. This was to avoid complications due to phase transformations in the presence of extrinsic ions like Na+, NH4+, or Cl–.
|
MATERIALS AND METHODS
|
|---|
Solution Preparation
Equilibrations of the solid with dilute H3PO4 solutions were done at 25.00 ± 0.05°C with the simple apparatus shown in Fig. 1. Stock H3PO4 solutions were prepared by diluting 85% H3PO4 (Mallinckrodt Baker Inc., Phillipsburg, NJ). The concentration was determined by gravimetric analysis as MgNH4PO4·6H2O according to Winkler (1931, p. 145). The analyses agreed to within 0.1%. All experiments were conducted in triplicate. A sample of 0.05 to 0.1 g of strengite was lodged in a bag of highly retentive, no. 42 filter paper (Whatman, Middlesex, UK). Bags were prepared by folding a square, 3- by 3-cm piece of filter paper in half along the diagonal, and then once again through the center to produce a bag in the shape of a right-angled triangle (Fig. 1). The solid was placed in the inside tip of the bag, and the open top was folded (3–4 mm) and closed by sewing. The bag was fastened with cotton threads and kept in the upper part of the solution to avoid grinding by the revolving magnet bar. The bag also made filtration unnecessary because it separated the solid from the solution when equilibrated. Preliminary experiments indicated a steady increase in strengite solubility during periods of months when the solid is kept in direct contact with the magnetic bar. Even though a steady state was finally attained, results were not reproducible. The erratic behavior can be ascribed to the transformation of the original solid into a phase of smaller, more soluble particles. The irreproducibility, on the other hand, was probably due to different terminal particle sizes in parallel experiments.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 1. The saturator: the solid was lodged in a bag of filter paper, prepared from a square piece of filter paper folded in half along the diagonal and then once again through the center; the open top was then folded and closed up by sewing.
|
|
Total Fe in solutions was determined by two spectrophotometric techniques. The first was based on the absorption at 470 nm of the Fe(III) oxinate complex extracted in chloroform. Following the procedure suggested in Vogel (1989), 0.1 to 0.2 µmol of Fe were analyzed with an accuracy of 1 to 2%. The second procedure was a spectrophotometric titration (Vogel, 1989) at 278 nm, at which Fe(III) phosphate complexes absorb appreciably. The titrant was prepared to contain a known concentration of Fe, while mP was identical to that of the test solution. Iron(III) perchlorate was used for preparation of Fe(III) solutions as described by Ciavatta et al. (1992). The end-point was assessed by extrapolation to zero ordinate in the linear plot of absorbance x total volume against added volume of reagent. With cells 4 to 10 cm thick, 0.1 µmol of Fe could be determined with an error of about 1%. Analyses by the two procedures agreed within 2%. The results of a series of measurements are given in Table 1. Three replicates were run to each point.
Absorbance measurements of the Fe(III) were made on a Hitachi Model U-2000 UV-visible double beam recording spectrophotometer (Hitachi High Technologies American, San Jose, CA). Matched quartz cells of 2-, 4-, or 10-cm thickness were used. Absorbance was recorded to 0.001 units. The reference was a H3PO4 solution of concentration identical to that of the test solution. The measuring cell was also the titration cell. The titrant was added in small portions with a weight burette.
Preparation of the Colloidal Phase
The colloidal phase was prepared by dissolving 4N Fe powder (Sigma-Aldrich, St. Louis, MO) in a slight excess of 57% HI (Mallinckrodt Baker Inc., Phillipsburg, NJ). After the metal was dissolved, H3PO4 was added in an amount exactly equimolar to that of the Fe. The clear solution was then poured into a large volume (5 L) of 0.7 mmol L–1 H3PO4. The Fe(II) and I ions were oxidized by 35% H2O2 in excess, and a fine colloidal phase developed. After removal of I and excess H2O2 by boiling, the colloid was washed several times by decantation with 0.7 mmol L–1 H3PO4. Portions of the suspension were put in glass-stoppered flasks and heated at 100°C for 30 d. The pale lavender resulting product was washed by decantation at frequent intervals for 4 d with 0.1 mol L–1 H3PO4 to remove soluble matter and very fine particles, then washed with water and air dried at room temperature.
Strengite Preparation
Crystalline strengite was obtained by digesting a suspension of colloidal Fe(III) phosphate in 0.7 mmol L–1 H3PO4 under mild hydrothermal conditions (100°C and 0.2 MPa). The acidity of this solution corresponds to pH 
3.2 and represents optimal conditions for crystallization (Cate et al., 1959). The critical point is the use of pure colloidal phases.
Strengite Characterization
The x-ray powder diffraction pattern for strengite (Fig. 2) shows that the amorphous phase was negligible. The terminal gravitational settling rate was 2.5 x 10–4 m s–1. Assuming the solid has a density of 2.74 g cm–3, this corresponds to 2 to 3 m2 g–1, a value at which no surface energy effects on the solubility are expected. Surface area measurements were made by application of the BET equation to conventional volumetric data for N (Brunauer et al., 1938). Solids kept in contact with H3PO4 solutions for periods necessary to reach a heterogeneous equilibrium gave x-rays patterns identical to the original phase.
|
RESULTS
|
|---|
Since [mFe(III)/mP] 0.004 in the saturated solutions, the principal ionic species are H+ and H\2PO4–, arising from the protolysis of H3PO4. The ionic strength I is then given by
 | [3] |
where mH and mA are the molality of H+ and H2PO4– ions, respectively. Values of mH and I can be obtained from equilibrium calculations based on the first protolysis constant of H3PO4, Ka1=mH2
HA/mHAHA, where mHA stands for the molality of the undissociated acid and H, A, and HA symbolize activity coefficients of H+ and H2PO4– ions and H3PO4, respectively. In the evaluations, mHA = mP – mH, Ka1 = 10–2.148 (Smith and Martell, 1989, p. 444), whereas the activity coefficient i of the ith ion of charge zi was calculated with Eq. [4], known as the Güntelberg approximation:
 | [4] |
A preliminary mH was deduced, setting activity coefficients equal to unity, and the process was reiterated to convergence. The results of the mH computations are given in Table 1. The quotient HA/HA in H3PO4 solutions was also calculated using the more accurate expressions by Pitzer and Silvester (1976). The differences at the dilutions in question, I < 0.03, were <1%. Thus the results from Eq. [4] have adequate accuracy.
The solubility is explained with the formation of soluble complexes, which due to the low Fe molality, most probably are mononuclear in Fe. Assuming the formation of a series of complexes with a general formula FeH–p(H3PO4)q3–p, the mass balance condition yields
 | [5] |
where ßp,q is the equilibrium constant for
 | [6] |
and 10p(p–5)D is the quotient of activity coefficients according to Eq. [4]. The solubility can be expressed as a function of mH only
 | [7] |
introducing *Ks, the constant for
 | [8] |
and Ka1 of H3PO4. The tendency of mFe(III) to reach the limiting value (3.94 ± 0.07) x 10–6 as mP approaches zero is consistent with the evanescence of the term *KsKa1108DmH and with the presence of complexes whose composition meets the condition *Ks 
ßp,qKa11–qmH(2q–p+1)10[8+p(p–5)–2q]D = constant. This is the case if the two conditions (i) 2q – p + 1 = 0 and (ii) [8 + p(p – 5) – 2q]D = 0 are met at the same time. Both conditions are met only for the couple (p,q) = (3,1), or in chemical terms by the presence of the soluble complex FePO4(aq). We may thus conclude that at mP 0.01 FePO4(aq) is the predominant soluble complex. The limiting mFe(III) corresponds to *Ksß3,1, hence
 | [9] |
The constraints Eq. [3] and [8] make it difficult to recognize the complexes responsible for the solubility increase at mP 0.02 mol kg–1. On the other hand, the models deduced from previous potentiometric and spectrophotometric studies of the system (Ciavatta et al., 1992; Ciavatta and Iuliano, 1995) help explain the solubility curve and evaluate the solubility constant *Ks.
The model best describing the Fe(III) orthophosphate system in dilute metal solutions, mFe(III) < 0.001 mol kg–1, is given in Table 2. In the mentioned investigations, the equilibria were studied at 25°C in a constant 3 mol L–1 ionic medium. The constants valid in the infinite dilution reference state are values extrapolated by the specific ion interaction theory (Scatchard, 1976, p. 143–148; Ciavatta, 1980, 1990) from the constants determined in 3 mol L–1 NaClO4. The specific interaction coefficients, b (kg mol–1), needed for the extrapolation are given in Table 2. The crucial point in these estimates is the exact knowledge of interaction coefficients of complex species. For mononuclear complexes with one or two ligands, like FeHPO4+ and Fe(H2PO4)2+ or FeH3(PO4)2 
FeHPO4H2PO4, the empirical rule suggested by Ciavatta (1990) yields interaction coefficients to within ±0.05 kg mol–1. Interaction coefficients equal to zero are often assumed for neutral species such as Fe(H2PO4)3(aq). There is, however, experimental evidence that they might be significantly different from zero (Grenthe et al., 1992, p. 281; Silva et al., 1995, p. 140). It therefore seems more realistic to resort to analogies with neutral complexes of the same ligand. For Fe(H2PO4)3(aq), a value of 0.08 kg mol–1 was adopted, which is between 0.06, the effect of NaClO4 on UO2(H2PO4)2(aq), and 0.1, the effect of NH4Cl on Am(H2PO4)3(aq). These coefficients were calculated from the variation with ionic strength of equilibrium constants selected by Grenthe et al. (1992, p. 281) and Silva et al. (1995, p. 140). The limits ascribed to the extrapolated constants reflect in large part the uncertainty introduced by interaction coefficients of complex species. The FeH7(PO4)3+ species is given an interaction coefficient similar to that of Fe(H2PO4)3(aq). This is somewhat arbitrary, but should have no sensible influence on calculation results due to a limited existence range of FeH7(PO4)3+ complex in our solutions.
View this table:
[in this window]
[in a new window]
|
Table 2. Equilibrium formation constants ßp,q, Eq. [6], of soluble complex species. Extrapolated values at I = 0 by the specific ion interaction model. Interaction coefficients b used in the extrapolation procedure: b(H+, ClO4–) = 0.14 kg mol–1, b(Fe3+, ClO4–) = 0.56 kg mol–1, b(Na+, H2PO4–) = –0.06 kg mol–1, b(Na+, HPO42–) = –0.16 kg mol–1, assumed from Ciavatta (1980); b(H3PO4, NaClO4) = 0.028, from the variation with ionic strength of the first protolysis constant of H3PO4, Ka1. For interaction coefficients of complex species bp,q=b[FeH-p(H3PO4)q3–p, NaClO4], see text.
|
|
The solubility constant, *Ks, was deduced for each point by inserting in Eq. [7] *Ksß3,1 from Eq. [9] and the constants from Table 2. Calculation results are in Table 1. As most probable, we propose the average
 | [10] |
where the error in large part reflects the uncertainties of the ßp,q values. By combining Eq. [10] with Eq. [9], we find
 | [11] |
|
DISCUSSION
|
|---|
The solubility curve of strengite in the H3PO4 molality range 0.001 to 0.1 mol kg–1 is explained by the formation of FePO4(aq) and the complexes in Table 2. The latter are known from independent sources, and the former exists in this study. Strictly, the present solubility data are explainable by any of the (FePO4)n species, with n = 1, 2, ... . The choice of n = 1 is suggested by the low metal concentration, but the formula FePO4(aq) should not be considered conclusive. We plan in the near future to investigate the system by other methods in a range of concentrations where such complexes appreciably form.
The high stability of FePO4(aq) is in part due to the favorable entropy change of the formation equilibrium. A contribution to stability might be envisaged also in some particular structure of the complex, in which the phosphate group acts as bi- or tridentate, or hydroxyl groups are present. As a matter of fact, the results from equilibrium analysis methods cannot tell anything about the structure of the complex so that FePO4(aq) is equivalent to Fe(OH)(HPO4)(aq) or Fe(OH)2(H2PO4)(aq).
From the value of *Ks in Eq. [9] and the protolysis constants of H3PO4 selected by Smith and Martell (1989, p. 444), we calculate
The solubility product obtained in this study differs by orders of magnitude from the literature values. Combining the equilibrium constant for Eq. [1] and the protolysis constants of H3PO4, Nriagu (1972) found logKs = –26.43. From the results by Egan et al. (1961), logKs can be evaluated as –26.1. Nriagu (1972) assumed FeH2PO4+ as the soluble complex. Using Eq. [9] and the model given in Table 2, logKs = –28.16 ± 0.33 is recalculated from Nriagu's data, which is a value that overlaps the constant proposed in this study within the limits of uncertainty. Egan et al. (1961) calculated the solubility product from the free energy of formation of strengite by combining calorimetric data in numerous and uncertain steps. A not improbable error of 10 kJ mol–1 for the free energy of formation of strengite would account for the difference.
The low solubility of strengite can explain the fixation in acid soils of soluble phosphates, added as fertilizer, by Fe(III) hydroxides. Assuming for 
-FeOOH (goethite) the solubility constant proposed by Baes and Mesmer (1976), logK[-FeOOH(cr) + 3H+Fe3+ + 2H2O] = 0.5, we evaluate logK[-FeOOH(cr) + H3PO4FePO4·2H2O(cr)] = 7.2. Thus, strengite can coexist with goethite in contact with solutions where mHA = 10–7. mol kg–1. This implies that at pH = 5, mP = 4.5 · 10–5 mol kg–1.
The availability of phosphate in soil depends on the pH and pE (the negative logarithm of the electron activity). The interactions of phosphate with Fe(II) and Fe(III) are very important. We can consider the solubility of phosphate as strengite and vivianite [Fe3(PO4)2·8H2O(cr)], in the presence of magnetite [Fe3O4(cr)] (this example is taken from Stumm and Morgan, 1996):
 | [12] |
 | [13] |
Considering a soil pH of 5, we can report log[H2PO4–] vs. pE, from Eq. [12] and [13]. As Fig. 3 shows, the solubility of phosphate is least at positive pE values, assuming logKs = –28.40 instead of –26.43.
|
NOTES
|
|---|
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.
Received for publication March 10, 2006.
|
REFERENCES
|
|---|
- Baes, C.F., and R.E. Mesmer. 1976. The hydrolysis of the cations. John Wiley & Sons, New York.
- Brunauer, S., P.H. Emmet, and E. Teller. 1938. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 60:309–319.[CrossRef][Web of Science]
- Cate, W.E., E.O. Huffman, and M.E. Deming. 1959. Preparation of crystalline ferric phosphates. Soil Sci. Soc. Am. J. 88:130–132.
- Chang, S.C., and M.L. Jackson. 1957. Solubility product of iron phosphate. Soil Sci. Soc. Am. Proc. 21:265–268.
- Ciavatta, L. 1980. The specific interaction theory in evaluating ionic equilibria. Ann. Chim. 70:551–567.[Web of Science]
- Ciavatta, L. 1990. The specific interaction theory in equilibrium analysis. Some empirical rules for estimating interaction coefficients of metal ion complexes. Ann. Chim. 80:255–263.[Web of Science]
- Ciavatta, L., and M. Iuliano. 1995. On the formation of mononuclear iron(III) orthophosphate complexes. Ann. Chim. 85:235–255.[Web of Science]
- Ciavatta, L., M. Iuliano, and R. Porto. 1992. The complexation of iron(III) with phosphate ions in 3 M sodium perchlorate medium at 25°C. Ann. Chim. 82:447–461.[Web of Science]
- Egan, E.P., Z.T. Wakefield, and B.B. Luff. 1961. Low temperature heat capacity and heat of formation of crystalline and colloidal ferric phosphate dihydrate. J. Phys. Chem. 65:1265–1270.[CrossRef][Web of Science]
- Grenthe, I., J. Fuger, R.J.M. Konings, R.J. Lemire, A.B. Muller, C. Nguyen-Trung, and H. Wanner. 1992. Chemical thermodynamics of uranium. Elsevier, Amsterdam.
- Nriagu, J.O. 1972. Solubility equilibrium constant of strengite. Am. J. Sci. 272:476–484.[Abstract]
- Pitzer, K.S., and L.F. Silvester. 1976. Thermodynamics of electrolytes. VI. Weak electrolytes including H3PO4. J. Solution Chem. 5:269–278.[CrossRef]
- Scatchard, G. 1976. Surface and colloid chemistry. Harvard Univ. Press, Cambridge, MA.
- Silva, R.J., G. Bidoglio, M.H. Rand, P.B. Robouch, H. Wanner, and I. Puigdomenech. 1995. Chemical thermodynamics of americium. Elsevier, Amsterdam.
- Smith, R.M., and A.E. Martell. 1989. Critical stability constants. Vol. 6. 2nd Suppl. Plenum Press, New York.
- Sposito, G. 1989. The chemistry of soils. Oxford Univ. Press, New York.
- Stumm, W., and J.J. Morgan. 1996. Aquatic chemistry. John Wiley & Sons, New York.
- Tiessen, H. 1995. Phosphorus in the global environment: Transfers, cycles and management. John Wiley & Sons, New York.
- Vogel, A.I. 1989. Vogel's textbook of quantitative chemical analysis. 5th ed. G.H. Jeffery et al. (ed.). Longmans Scientific, Harlow, Essex, UK.
- Winkler, L.W. 1931. Ausgewählte untersuchungsverfahren für das chemische laboratorium. F. Enke Verlag, Stuttgart.
This article has been cited by other articles:
|
|
|
|
 
M. Iuliano, L. Ciavatta, and G. De Tommaso
The Solubility Constant of Variscite
Soil Sci. Soc. Am. J.,
January 25, 2008;
72(2):
343 - 346.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
