Spectroscopic Study and Local Coordination of Polyphosphate

Jan 19, 2005 - Adriana P. Duarte , Marie Gressier , Marie-Joëlle Menu , Jeannette ... A. J. Barbosa , L. J. Q. Maia , B. Montanari , R. R. Gonçalves...
8 downloads 0 Views 149KB Size
Download PDF
1776

Langmuir 2005, 21, 1776-1783

Spectroscopic Study and Local Coordination of Polyphosphate Colloidal Systems Francisco A. Dias Filho,† Luis D. Carlos,‡ Younes Messadeq,† and Sidney J. L. Ribeiro*,† Institute of Chemistry, UNESP, P.O. Box 355, 14801-970, Araraquara-SP, Brazil, and Department of Physics, CICECO, University of Aveiro, 3810-193 Aveiro, Portugal Received September 16, 2004. In Final Form: November 17, 2004 The interaction between metaphosphate chains and the metal ions Ca2+ and Eu3+ has been studied in water by Eu3+ luminescence, infrared absorption, and 31P NMR spectroscopy. Two main families of sites could be identified for the metal ions in the aqueous polyphosphate colloidal systems: (1) cagelike sites provided by the polyphosphate chain and (2) a family which arises following saturation of cagelike sites. Occupation of this second family leads to supramolecular interactions between polyphosphate chains and the consequent destabilization of the colloidal system. In the polyphosphate-Ca2+ system, this destabilization appears as a coacervation process. Equilibrium existing between colloidal species as a function of the compositions could be reasoned based on the spectroscopic measurements. The determination of coordination numbers and the correlation of the results with the observation of coacervates show that Eu3+ luminescence properties can be used to probe in a unique way the coacervation process.

1. Introduction Commercial sodium polyphosphate, known as Graham’s salt, is obtained as a glassy solid by quenching molten NaPO3 (which in turn is obtained by heating NaH2PO4‚ H2O at 625 °C).1 The characteristic glass temperatures are easily identified in differential scanning calorimetry (DSC) scans at 498 and 595 K for the glass transition (Tg) and the crystallization temperatures (Tx), respectively. Sodium polyphosphate is the only water soluble inorganic polyphosphate, and 31P nuclear magnetic resonance (NMR) and chromatography results clearly show the presence of long (PO3)n phosphate chains. In this sense, they are classified as oligophosphates.1 In fact, the average chain length ranges from n ) 3 to n ) 30 for the commercial products. In the laboratory, samples can be prepared with variable average chain sizes ranging from 3 to ca. 300 phosphorus atoms. The interesting aspects of the solution chemistry of these polyphosphate chains in water have led to many different applications ranging from water softening to stabilization of mineral suspensions. At concentrations as low as in the range of parts per million (ppm), polyphosphates may be used to inhibit precipitation of calcite1 or in the stabilization of colloidal semiconductor or metal particles.2-4 In fact, the flexibility offered by the long chain, allowing metal ions to be fitted into the structure, and the complexing power of the middle PO3 units make polyphosphates strong complexing agents.1,5 Interestingly, enhanced interaction is observed between polyphosphate groups and alkaline earth ions when compared for example with alkali metals. * To whom correspondence should be addressed. E-mail: [email protected]. † Institute of Chemistry, UNESP. ‡ Department of Physics, CICECO. (1) Corbridge, D. E. C. Phosphorus: An Outline of Its Chemistry, Biochemistry and Technology; Studies in Inorganic Chemistry, Vol. 20; Elsevier: Amsterdam, 1995. (2) Dijken, A. V.; Vanmaekelbergh, D.; Meijerink, A. Chem. Phys. Lett. 1997, 269, 494. (3) Hengleinv, A.; Linnert, T.; Mulvaney, P. Ber. Bunsen-Ges. Phys. Chem. Chem. Phys. 1990, 94 (12), 1449. (4) Henglein, A. Chem. Phys. Lett. 1989, 154, 473. (5) Rashchi, F.; Finch, J. A. Miner. Eng. 2000, 13 (10-11), 1019.

Van Wazer and Callis6 interpreted this behavior in the 1950s, suggesting that alkaline earth ions could be held at specific sites of the chain. In materials preparation, polyphosphates have been used to produce metal polyphosphate gels from which different interesting products may be obtained. Gel formation is being studied as the result of a liquid-liquid phase separation.7-9 If high polyphosphate concentrations (>2 mol/L) are used, an interesting phenomenon may be observed connected also with liquid-liquid phase separation: the addition of electrolytes generally leads to the separation of two liquid phases with different densities (the same effect may be observed by the addition of a low dielectric constant solvent to mixtures of aqueous solutions of sodium polyphosphate and alkaline earth chlorides). The less dense liquid is the “equilibrium liquid” supposed to contain short polyphosphate chains. The high density phase, supposed to be colloid-rich, i.e., composed mainly of long polyphosphate chains, is currently called “coacervate” as proposed in 1929 by Bungenberg and coworkers.10 Early work by Umegaki and Kanazawa11,12 on magnesium and calcium polyphosphate coacervates concluded for a structure where polyphosphate chains might be coiled spherically and somewhat tightly. Alkaline earth ions would provide cross-linkages of the chains.11,12 The high density and also high viscosity (up to 105 cP) polyphosphate coacervates find numerous applications in materials science. They may be used as precursors for glass preparation,13,14 in the protection of metals against (6) Van Wazer, J. R.; Callis, C. F. Chem. Rev. 1958, 58, 1011. (7) Lima, E. C. O.; Galembeck, F. Colloids Surf., A 1993, 74, 65. (8) Masson, N. C.; de Souza, E. F.; Galembeck, F. Colloids Surf., A 1997, 121, 247. (9) Azevedo, M. M. M.; Bueno, M. I. M. S.; Davanzo, C. U.; Galembeck, F. J. Colloid Interface Sci. 2002, 248, 185. (10) Bungenberg de Jong, H. G.; Kruyt, H. R. Proc. K. Ned. Akad. Wet. 1929, 32, 849. (11) Umegaki, T.; Kanazawa, T. Bull. Chem. Soc. Jpn. 1975, 48 (5), 1452. (12) Umegaki, T.; Kanazawa, T. Bull. Chem. Soc. Jpn. 1979, 52 (7), 2124. (13) Vast, P.; Barbieux, F.; Gomez, F. Verre 1996, 2 (3), 3. (14) Willot, G.; Gomez, F.; Vast, P.; Andries, V.; Martines, M.; Messaddeq, Y.; Poulain, M. C. R. Chim. 2002, 5, 899.

10.1021/la0476837 CCC: $30.25 © 2005 American Chemical Society Published on Web 01/19/2005

Polyphosphate Colloidal Systems

corrosion,13 or in the immobilization and destruction of asbestos.15 The possibility of incorporation of different fillers opens a way to many different applications.13,16 In general, the formation of these coacervates may be understood as a compromise involving different effects such as electrostatic repulsive and van der Waals attractive and repulsive solvation forces. A simple DLVO (Derjaguin-Landau-Verwey-Overbeek) electrostatic model has been used to explain the phenomenon.17 Concerning the preparation of polyphosphate coacervates, the study of the interaction between the polyphosphate chain and bivalent metal ions such as Ca2+ is a very important step toward a better comprehension of the process mechanism. In fact, the existence of stable complexes between polyphosphate and alkaline earth ions was proposed a long time ago5,11,12 and is the base for the industrial application of polyphosphates in sequestering of metal ions. However no study has been able to clearly describe the nature of the sites occupied by alkaline earth ions in solution and their role in coacervation. The possibility of substituting spectroscopic active trivalent lanthanide ions for bivalent alkaline earth ions could be helpful to understand these metal ion-polyphosphate interactions. In fact, the rich site-dependent spectroscopic features displayed by Eu3+ and Tb3+ have led to the utilization of these ions as luminescent probes for Ca2+ sites in biological systems.18 In parallel, stable polyphosphate-lanthanide colloids can be employed in the preparation of lanthanide-based luminescent materials. The coordinating polyphosphate chain could provide an encapsulating environment where lanthanide compounds could be prepared with restricted grain sizes and protected from the well-known deleterious emission quenching effect of the water molecules. High energy vibrational modes of these molecules can connect excited and ground states, leading to quenching of emission by nonradiative paths. The evolution of either the Eu3+ 5 D0 or the Tb3+ 5D4 excited state decay time values can be easily connected to the number of water molecules in the first coordination shell of metal ions.19-21 In the so-called Horrocks method, the number of water molecules may be obtained from decay time measurements performed in H2O and D2O. The difference between these two values is connected with the number of water molecules around Eu3+, as will be further discussed in the results section. This feature is well explored when using Eu3+ as a probe for Ca2+ sites in biological systems since the determination of the number of water molecules in the metal ion’s first coordination shell is of paramount importance in understanding the Ca2+ chemistry in the biological medium. On the other side, hindering of the Eu3+ excited state quenching mechanism is highly desired when the preparation of Eu3+ luminescent materials must be performed in water. We present in this paper the results on the preparation and characterization of stable colloids and calcium coacervates from sodium polyphosphate solutions. The systems were prepared containing Eu3+, and luminescent properties (emission and excitation spectra together with emis(15) Vast, P. European Patent WO 98/21155, 22 May, 1998. (16) Vast, P. Phosphorus Lett. 1996, 26, 10. (17) Gomez, F.; Vast, P.; Barbieux, F. Phosphorus Res. Bull. 1995, 5, 143. (18) Richardson, F. S. Chem. Rev. 1982, 82, 541. (19) Horrocks, W. D., Jr.; Schmidt, G. F.; Sudnick, D. R.; Kittrell, C.; Bernheim, R. A. J. Am. Chem. Soc. 1977, 99, 2378. (20) Horrocks, W. D., Jr.; Sudnick, D. R. J. Am. Chem. Soc. 1979, 101, 334. (21) Supkowski, R. M.; Horrocks, W. D., Jr. Inorg. Chim. Acta 2002, 340, 44.

Langmuir, Vol. 21, No. 5, 2005 1777

Figure 1. Room-temperature emission spectra (λexc ) 394 nm): (a) Eu3+ (EuCl3 10-2 mol/L, pH ) 5); (b) PE ) 526; (c) PE ) 50; (d) PE ) 10; (e) PE ) 5; (f) PE ) 2. Numbers denote J levels for the 5D0 f 7FJ levels. Inset: Detail of the 5D0 f 7F2 transition.

sion decay time measurements) have been followed as a function of compositions. 31P NMR and Fourier transform infrared (FTIR) spectroscopy were the other experimental techniques employed in the characterization. 2. Experimental Section Stable europium sodium polyphosphate colloids were prepared by the addition of europium chloride aqueous solution (0.1 mol L-1) to a sodium polyphosphate water solution (2 mol L-1). The molar ratio P/Eu3+ (hereafter called PE) ranged from 2 to 526. Solid precipitates have been observed for PE < 4. Coacervates were prepared by mixing at room temperature (25 °C) calcium chloride 1 mol L-1 solution and Eu3+-containing (0.2 mol %) sodium polyphosphate (4 mol L-1) solution. The P/Ca (hereafter called PCa) molar ratio ranged from 2 to 20. Coacervates separate as dense viscous phases from mother solutions after stirring typically for 3 h solutions with PCa < 6.5. Excitation and emission spectra as well as Eu3+ 5D0 decay times were obtained with a SPEX 212I fluorimeter equipped with both continuous (450 W) and pulsed (5J per pulse, 3 µs bandwidth) Xe lamps. Emission decay times have been processed with a SPEX 1930 phosphorimeter. Infrared absorption spectra were obtained from the solutions with an attenuated total reflection (ATR) accessory in a FTIR Spectrum 2000 Perkin-Elmer spectrometer. 31P NMR spectra were obtained with a Bruker AC 200 spectrometer operating at 81.02 MHz. The samples were studied at 25 °C in freshly prepared aqueous solution using 5- or 13-mm spinning sample tubes, and the results are presented relative to aqueous (14 mol L-1) H3PO4.

3. Results and Discussion Polyphosphate-Eu3+ System. Figure 1 shows roomtemperature emission spectra (corrected for spectral response of the detection setup) obtained for the Eu3+ aqueous solution (Figure 1a) and also the sodium polyphosphate solutions with different PEs under 394 nm excitation (Figure 1b-f). This excitation line corresponds to the Eu3+ 7F0 f 5L6 absorption transition, and from the 5 L6 excited level nonradiative decay paths populate the 5 D0 level, from which emission is observed to the lower 7 F0-4 levels (Eu3+ energy levels are shown in Figure 2). Assignments are made in the figure. 5D0 experimental decay time values are shown in Table 1.

1778

Langmuir, Vol. 21, No. 5, 2005

Dias Filho et al.

Figure 2. Eu3+ energy levels. Data are from ref 23. Table 1. System Eu3+-Polyphosphate: Eu3+ Experimental and Calculated (Radiative) 5D0 Decay Times, 5D0 Quantum Efficiencies (η), and Number of Water Molecules (q) in the Eu3+ First Coordination Shell molar ratio P/Eu (PE) 526 200 100 50 10 5 4 3 2 1

5D

0 decay time ((0.01 ms)

1.15 1.13 1.11 1.06 0.50 0.70 0.39 0.39 0.31 0.31

5D radiative 0 decay time (ms)

η

q

5.30 5.31 5.34 5.38 6.47 5.86 5.42 5.12 5.55 5.67

0.22 0.21 0.21 0.20 0.08 0.12 0.07 0.08 0.06 0.05

0.4 0.4 0.4 0.5 1.7 1.1 2.3 2.3 3.1 3.1

As is well-known, in diluted aqueous solution (pH ≈ 5) and in the absence of coordinating anions the Eu3+ first coordination shell is composed by water molecules. Considering the lanthanide series, Eu3+ lies in a region where the coordination number changes from 9 for the lighter members to 8 for the heavier elements. Equilibrium must be in fact operative between eight- and ninecoordinated species, and a mean value of 8.3 water molecules is obtained from X-ray diffraction studies.22 The 5 D0 decay time in H2O is 0.12 ( 0.01 ms, and in D2O a value of 3.20 ( 0.01 ms is obtained. As mentioned in the Introduction, decay time measurements in H2O and D2O are usually used to obtain the number of water molecules in the Eu3+ first coordination sphere. The Horrocks approach19,20 considers the utilization of the two values with the empirical formula

[

q ) 1.05

]

1 1 τ(H2O) τ(D2O)

(1)

where q is the number of water molecules, τ(H2O) is the decay time (ms) in water, and τ(D2O) is the decay time in D2O. Recently a correction of the above formula was proposed by the same group21 which is then written

q ) 1.11

[

Considering this last formula and the decay time values, 8.5 ( 1.0 water molecules are obtained for Eu3+ in water, in excellent agreement with the experimental value. However, a currently found drawback for those working with Eu3+ as a structural probe lies in obtaining decay time values in D2O. Most laboratories do not have D2O easily available; it is expensive and hygroscopic in the sense that fast exchange D2O-H2O occurs with atmospheric water. This problem can be overcome if one considers that the decay time observed in D2O can be considered as a contribution of a purely radiative process. In this way, the Eu3+ emission spectrum fortunately displays a purely magnetic dipolar transition at around 590 nm (5D0 f 7F1). This transition does not depend on the ligand field, and therefore its intensity may be considered constant.23 Einstein’s spontaneous emission coefficient for that transition (A01) is readily determined from the coefficient calculated in vacuum (A01′)23 and the refractive index n (A01 ) A01′n3). A01 is 34.4 s-1 in water. In this way, considering the 5D0 f 7FJ (J ) 0-4) transitions in the emission spectrum the total radiative coefficient (AT or τ-1 RAD) is obtained from the experimental intensities taking the 5D0 f 7F1 transition as a reference. The 5D0 f 7 F5,6 contributions were too small and therefore not -1 considered. By doing that, τ-1 RAD is 0.12 m s . Substituting -1 -1 τRAD for τ (D2O) in the Horrocks formula (eq 2), one obtains 8.8 ( 1.0 water molecules for Eu3+ in water, also in good agreement with the experimental value. In the -1 following, therefore, we will use τ-1 RAD instead of τ (D2O) in the Horrocks formula as proposed before in the literature.24 The emission quantum efficiency of the 5D0 level (η) may be evaluated as the ratio of the experimental and the -1 pure (calculated) radiative decay values (η ) τ-1 EXP/τRAD). A value of 0.02 is then obtained for Eu3+ in water (if τ-1(D2O) is used instead of τ-1 RAD, a value of η ) 0.04 is obtained, which in this case is a negligible difference). Table 1 shows the values so obtained. In polyphosphate solution, the Eu3+ emission spectrum is completely changed due to the complexation of the polyphosphate chains as shown in Figure 1b. The spectrum is dominated by the Laporte forbidden electric dipolar 5D0 f 7F2.23 The enhanced intensity observed for this wellknown hypersensitive transition 5D0 f 7F2 (∆J ) 2) is usually related either to a less symmetric ligand field or the increasing degree of covalence in the interaction between the lanthanide ion and the ligands,25-27 and in this sense the strong Eu3+-polyphosphate interaction is put forward. As Table 1 shows, the number of water molecules obtained with eq 2 for PE ) 526 is surprisingly low (0.4) and the quantum efficiency is 0.22. This is a very impressive result in water and can be explained by the encapsulation of Eu3+ by polyphosphate chains in cagelike environments. As mentioned in the Introduction, the sequestering effect of polyphosphate chains toward alkaline earth ions in water is well-known and in fact gives

]

1 1 - 0.31 τ(H2O) τ(D2O)

(2)

(22) Habenshuss, A.; Spedding, F. H. J. Chem. Phys. 1980, 73 (1), 442. (23) Carnall, W. T.; Crosswhite, H.; Crosswhite, H. M. Energy structure and transition probabilities of the trivalent lanthanides in LaF3; Argonne National Laboratory Report, Unnumbered; Argonne National Laboratory: Argonne, IL, 1977. (24) Hazenkamp, M. F.; Blasse, G. Chem. Mater. 1990, 2, 105. (25) Judd, B. R. J. Chem. Phys. 1979, 70, 4830. (26) Reisfeld, R.; Jo¨rgensen, C. K. In Handbook on the Physics and Chemistry of Rare Earths; Gschneidner, K. A., Eyring, L., Eds.; NorthHolland: Amsterdam, 1987; Vol. 9, Chapter 58, and references therein. (27) Malta, O. L.; Carlos, L. D. Quim. Nova 2003, 26 (6), 889.

Polyphosphate Colloidal Systems

Langmuir, Vol. 21, No. 5, 2005 1779

rise to some of the different applications found for Graham’s salt. In water, polyphosphate chains (PP) are hydrated and electric neutrality is provided by electrostatic interaction with Na+ ions. With the addition of Eu3+, some of the water molecules and Na+ ions must be displaced from the surroundings of the polyphosphate chain and the simplified equilibrium shown below must be established, illustrating the role played by the polyphosphate chain as an ionic exchanger.

PP(H2O)Na + xEu3+(aq) / PP(H2O)Eu + 3xNa+(aq) (3) [PP(H2O)Na] and [PP(H2O)Eu], rather than chemical formulas, are just schematic representations of the colloidal species present in the system. Therefore the pseudo-equilibrium represented by eq 3 is governed by the well-known equilibrium laws of thermodynamics and also by electrostatic interactions in the aqueous medium (DLVO theory).17 Just by changing the solvent, for example, the situation could be dramatically changed. Further information concerning the nature of metal ion sites in water solution can be obtained taking as a starting point the description of the environment of metal ions in amorphous polyphosphate systems (glasses).28,29 Planar polyphosphate chains are proposed to wind around metal ions, and that encapsulation could then be taken as a base to define the local site for Eu3+ in solutions with high PE (PE ) 526). Increasing the Eu3+ content from PE ) 526 to PE ) 50, similar spectra and similar decay time values are observed, whereas changes in relative intensities, bandwidths, and decay time appear for lower PE values (Figure 1). The inset of Figure 1 shows the continuous broadening observed for the 5D0 f 7F2 transition on going from PE ) 526 to PE ) 2. Moreover, destabilization of the colloidal system is observed with precipitation of an amorphous white solid for P/Eu < 5. These results suggest that apart from the occupation of a family of sites (called site I hereafter) characterized by spectra b-d in Figure 1 for high PE values, a second family of sites (called site II hereafter) appears with increasing intensity in spectra e and g in Figure 1. Considering the equilibrium described by eq 3 completely shifted to the right (saturation of sites I), a second situation must be established with further addition of Eu3+ and is simply described by

PP(H2O)Eu(I) + Eu3+ / PP(H2O)Eu(I)Eu(II)

(4)

where Eu(I) and Eu(II) refer to Eu3+ occupying the family of sites I and II, respectively. As observed for eq 3, thermodynamics and electrostatics rules control the stability and relative content of the colloidal species. The same suggestion can be made from excitation spectra shown in Figure 3. The observed bands could be easily assigned to intra-4f6 transitions arising from the ground state 7F0 (and also the first excited states 7F1,2 that could be thermally populated at room temperature) to several excited states.23 Assignments are done in the figure. The observation is the same as for the emission spectra; that is, starting with low Eu3+ content (PE ) 526) up to PE ) 50 general features (relative intensities and splitting of levels) are the same. Increasing further the Eu3+ content, continuous changes occur, suggesting the increasing contribution of a second general family of (28) Ropp, R. C. Inorganic Polymeric Glasses; Studies in Inorganic Chemistry, Vol. 15; Elsevier Science: Amsterdam, 1992. (29) Brow, R. K. J. Non-Cryst. Solids 2000, 263-264, 1.

Figure 3. Room-temperature excitation spectra (λem ) 615 nm): (a) PE ) 526; (b) PE ) 50; (c) PE ) 10; (d) PE ) 5; (e) PE ) 2. Inset: Stokes and anti-Stokes vibronic bands associated with the 7F0 f 5D2 central line (see the text for further explanation). Numbers denote excited levels for transitions arising from the lower 7F0-2 levels [1, (5I, 3H)6; 2, 3P0, 5FJ; 3, 5H5; 4, 5D4; 5, 5GJ, 5I7; 6, 5L6; 7, 5D3; 8, 5D2].

sites. The inset of Figure 3 shows a selected region of the excitation spectrum obtained for the sample with PE ) 10. Sidebands are observed for the 7F0 f 5D2 electronic transition. The ordinate scale has been changed to wavenumbers in order to show that both bands occurring at 20 580 and 22 650 cm-1 can be assigned to anti-Stokes and Stokes vibronic transitions associated with the electronic line centered at 21 525 cm-1. It would correspond to a 1035 cm-1 vibrational mode. The probe nature of Eu3+ ions is clearly put forward by the observation of the vibronic spectrum in Figure 3, and FTIR and NMR spectra presented hereafter will show that interactions between the polyphosphate chain and metal ions occur preferentially with middle PO3 groups. Figure 4a shows the evolution of the Eu3+ 5D0 lifetime as a function of the PE ratio. Values are also displayed in Table 1. It is observed that for high values of PE the decay time values go asymptotically to 1.15 ms. As the PE values decrease, the lifetime remains constant until around a value of 100, whereas for lower values a sudden change is observed, with the decay time values decreasing to 0.31 ms (Table 1). In fact, decay curves obtained for samples with low PE values could not be fitted by a singleexponential function. At least two rate constants could be extracted from the experimental data, suggesting the contribution of at least two families of sites in agreement with emission and excitation spectra. Decay time values presented in Figure 4a and also in Table 1 for samples with PE < 5 refer to the short value (the longer one is ≈1.1 ms). Figure 4b shows typical fits for one sample (PE ) 526) displaying just one decay constant and one sample displaying the biexponential behavior (PE ) 2). Following the suggestion made above (eqs 3 and 4), the appearance of the contribution of a short decay time in addition to the longer one in samples with high concentrations of Eu3+ could be due to saturation of polyphosphate

1780

Langmuir, Vol. 21, No. 5, 2005

Dias Filho et al.

Figure 5. Relative occupations of site I (pEu 1 ) as a function of the PE ratio. The curve shows the fit of experimental data with eq 7. Figure 4. Eu3+ 5D0 decay time results for systems with different PE values. (a) Evolution of the 5D0 decay time with the PE value. (b) Typical decay curves obtained for samples with PE ) 2 and 526. Adjusted decay times are shown in the figure.

“cagelike” coordination sites (family I). Additional Eu3+ ions would locate in a second family of sites where they are no more completely protected from water molecules, leading to emission quenching by nonradiative paths through OH oscillators. As shown in Table 1, the number of water molecules calculated with the Horrocks formula increases, reaching the value of 3.1 for PE ratio 1. -1 The contribution of the two rate constants, τ-1 I and τII , -1 to the experimental value, τEXP, can be written -1 -1 p1τ-1 I + (1 - p1)τII ) τEXP

(5)

where p1 defines the relative occupation of site I, and the two limit decay time values are τI ) 1.15 ms and τII ) 0.30 ms. Expressing p1 in terms of the rate constants, we have

p1 )

-1 τ-1 EXP - τII

τ-1 I

-

τ-1 II

(6)

Figure 5 shows a plot of p1(Eu) as a function of PE. A sigmoidal-like function is suggested, and in this way experimental data could be fitted by the following expression:

p1 )

1 1+e

-(PE-x0)/dx

cagelike local environments. In the metaphosphate chain, each phosphate group must coordinate to a Eu3+ by one oxygen atom and therefore the value of 8.9 obtained for x0 could well be understood as defining the contribution of phosphate species to the coordination number for the Eu3+ ions in those sites. Therefore 8.9 oxygen donor atoms coming from phosphate species and 0.4 oxygen donor atoms coming from residual water molecules, as obtained with the Horrocks expression, give a total coordination number of 9.3. Considering usual coordination numbers, this is a reasonable value for Eu3+. The saturation of sites I and the consequently progressive occupation of sites II with the increasing of Eu3+ content is given by dx(Eu). This quantity should be governed by the unknown relative equilibrium constants associated with the different Eu3+-first neighbor’s interactions. From eq 3, a first equilibrium constant, K1, can be written as

K1 )

[PP(H2O)Na][Eu3+(aq)]x

(8)

After saturation of sites I, eq 4 defines the relative content of sites I and II and a second equilibrium constant, K2, can be written as follows:

(7)

In this function, x0 defines an inflection point and dx controls the slope of the curve. By fitting experimental data with eq 7, the values x0(Eu) ) 8.9 ( 1.0 and dx(Eu) ) 3.9 ( 0.9 are obtained. Considering the two main families of sites being occupied as a function of the relative Eu3+ concentration, we suggest that the value so obtained for x0 defines the saturation point for the occupation of the family of sites I by Eu3+ ions. That is, at that point all available phosphate groups are involved in the coordination to the Eu3+ ions in the

[PP(H2O)Eu(I)][Na+(aq)]3x

K2 )

[PP(H2O)Eu(I)Eu(II)] [PP(H2O)Eu(I)][Eu3+]

(9)

From eq 5,

K2 )

1 - p1 p1[Eu3+]

(10)

Equation 10 may be rewritten in order to obtain an expression for p1,

Polyphosphate Colloidal Systems

p1 )

Langmuir, Vol. 21, No. 5, 2005 1781

1 1 + K2[Eu3+]

(11)

Remembering that PE was defined before as the molar ratio (P5+/Eu3+) and that the initial polyphosphate content is constant, eq 11 is then written

p1 )

1 K2 1+ PE

(12)

Comparing eqs 7 and 12 shows that

e-(PE-x0)/dx )

K2 PE

or

-

( )

(PE - x0) K2 ) ln dx PE

(13)

When PE ) x0, K2 ) PE. Therefore the value of PE at the inflection point in Figure 5 gives the relative content of the two colloidal species. Polyphosphate-Ca2+ System. The model describing the Eu3+ local coordination could be used to understand the behavior observed for the polyphosphate solution with the addition of Ca2+ which, as mentioned before, leads to the formation of the so-called coacervates. Figures 6 and 7 show emission spectra and the evolution of the 5D0 lifetime, respectively, obtained for Eu3+containing solutions (0.2 mol % Eu3+) with different PCa values. For high PCa values, the emission spectrum closely resembles the spectrum obtained for high values of PE in Figure 1. Also the 5D0 lifetime is similar (1.07 ms). Increasing the Ca2+ amount, the emission spectra are observed to change continuously. The 5D0 decay time values decrease drastically for PCa < 10. Table 2 summarizes the results obtained. The same approach used for polyphosphate-Eu3+ which led to eq 6 can be used here in the polyphosphate-Ca2+ system. In this way, Figure 8 shows a plot of p1(Ca) as a function of PCa where p1(Ca) defines the occupation of polyphosphate cages available for Ca2+ and probed by Eu3+ doping ions (site I). By fitting experimental data with the sigmoidal-like function described by eq 7 (substituting PCa for PE), the values x0(Ca) ) 5.4 ( 0.2 and dx(Ca) ) 1.5 ( 0.2 are obtained. Considering those cagelike local sites for Ca2+ in solution with high values of PCa, an interesting interpretation from the results could be proposed taking into account the arguments presented in the analysis of the polyphosphate-Eu3+ system. Starting with solutions with the highest value of PCa, the added Ca2+ ions progressively occupy the cagelike sites until saturation. At that point, Eu3+ probe ions are displaced from the cages giving place to Ca2+. That equilibrium can be schematically shown by

PP(H2O)Eu(I) + Ca2+ / PP(H2O)Ca(I)Eu(II) (14) Interestingly, the x0(Ca) value obtained from the fit, in the same way as for the Eu3+-polyphosphate, can be related to the coordination number of Ca2+ in the polyphosphate system. Therefore 5.4 oxygen donor atoms from phosphate groups and 0.5 oxygen donor atoms from water molecules (Table 2) give a total number of 5.9 for the Ca2+ coordination number. To our knowledge, there is no experimental value for the coordination number of Ca2+

Figure 6. Room-temperature emission spectra (λexc ) 394 nm): (a) PCa ) 20; (b) PCa ) 8; (c) PCa ) 6; (d) PCa ) 2. Inset: Detail of the 5D0 f 7F2 emission transition.

Figure 7. Eu3+ 5D0 decay times as a function of the PCa ratio.

in polyphosphate solutions, but in phosphate glasses a value of 6 is obtained by X-ray diffraction28 that could be assumed for our polyphosphate systems. This result highlights the probe nature of the Eu3+. Still concerning the coordination sites for metal ions, FTIR spectra obtained for selected samples are displayed in Figure 9a. A stretching mode ν(P-O) related to nonbridging PO groups in the middle of the chain is observed in the 1070 cm-1 region. A blue shift is observed with increasing Ca2+ concentration, which supports the

1782

Langmuir, Vol. 21, No. 5, 2005

Dias Filho et al.

Figure 8. Relative occupation of site I (pCa 1 ) as a function of the PCa ratio. The curve shows the fit of experimental data with eq 7. Table 2. System Ca2+-Polyphosphate: Eu3+ Experimental and Calculated (Radiative) 5D0 Decay Times, 5D0 Quantum Efficiencies (η), and Number of Water Molecules (q) in the Eu3+ First Coordination Shell molar ratio P/Ca (PCa) 50 20 10 8 6 5 4.5 4 3.5 3 2

5D

0 decay time ((0.01 ms)

1.15 1.10 1.07 0.92 0.80 0.78 0.69 0.65 0.64 0.58 0.45

5D radiative 0 decay time (ms)

η

q

5.30 5.35 5.34 5.75 5.71 6.00 5.31 5.41 5.81 6.00 8.00

0.22 0.20 0.21 0.16 0.14 0.13 0.13 0.12 0.11 0.10 0.07

0.4 0.5 0.4 0.7 0.8 0.9 1.0 1.3 1.2 1.4 2.0

main interaction of the metal ions with that middle chain phosphate group. The vibronic spectrum shown in Figure 3 displays a vibrational mode of 1035 cm-1. The energy mismatch between modes observed in infrared absorption (or Raman scattering) and vibronic spectra is usually due to the large effective mass of the local vibration system.30 Figure 9b shows 31P NMR spectra. The middle phosphate group is identified as a Q2 resonance in the 21-22 ppm region,29 and the figure shows that with the increase of Ca2+ content a shift of the Q2 peak from -21.4 to -21.9 is clearly observed in agreement with the FTIR spectrum. It is interesting to note that for PCa ) 6 the phenomenon of coacervation is observed. From 31P NMR results, we observed that the initial polyphosphate solution presents polyphosphate chains with a mean value of 23 phosphate units. After coacervation, 31P signal is still observed from the equilibrium liquid, but related to shorter chains with 5-10 phosphate units. Therefore coacervation occurs mainly for the long chain structures which concentrate in coacervate. Attending to the features of the coordination model proposed, and based on the spectroscopy results here (30) Blasse, G. Int. Rev. Phys. Chem. 1992, 11 (1), 71.

Figure 9. (a) ATR FTIR spectra for (I) NaPO3, (II) PCa ) 20, and (III) PCa ) 5. (b) 31P NMR spectra for (I) PCa ) 20, (II) PCa ) 10, and (III) PCa ) 5.

presented, coacervation can be understood and predicted. Therefore for high PCa values, Ca2+ ions are located in the cagelike sites provided by the polyphosphate chains leading to a stable colloidal system. Figure 8 and the fit using eq 7 suggest that the available cagelike sites are completely occupied for PCa ) 6. Ca2+ ions added after the saturation of these sites would then locate outside the polyphosphate cages (site II). Supramolecular interactions, that is, cross-linking between adjacent chains, would be promoted by Ca2+ ions, leading therefore to destabilization of the colloidal system and coacervation as suggested before.11,12 We have shown here that photoluminescence features of Eu3+ could be used to monitor the formation of coacervates in polyphosphate-Ca2+ systems. 4. Conclusion The interaction between metaphosphate chains and the metal ions Ca2+ and Eu3+ has been studied in water by Eu3+ luminescence, infrared absorption, and 31P NMR spectroscopy. Concerning polyphosphate-europium systems, Eu3+ locates at cagelike sites provided by the polyphosphate chain for high relative polyphosphate content (PE molar ratio > 10). Those cagelike sites hinder the interaction of the active ion with water molecules, leading to an increase in the emission quantum efficiency. For high Eu3+ content, changes are observed in luminescence spectra and excited state decay time values, leading to the proposition that Eu3+ occupies new sites where the metal ions are no more encapsulated and interacts with water molecules strongly. These additional Eu3+ also lead to destabilization of the colloidal system and consequently precipitation of a white amorphous solid. Polyphosphatecalcium systems were also studied by Eu3+ luminescence with the probing atoms being added to the initial system in small proportion. In that case, spectroscopy results lead

Polyphosphate Colloidal Systems

to the suggestion that for high PCa molar ratio (>6) Ca2+ would locate in similar cagelike sites with coordination number of 6 provided by phosphate units of the metaphosphate chain. With further addition of Ca2+, Eu3+ luminescence shows changes attributed to occupation of a second family of sites. Ca2+ ions at those sites would promote supramolecular interactions, which destabilizes the colloidal system leading to coacervation. The equilibrium existing between colloidal species as a function of the compositions could be discussed based on the spectroscopic measurements. Therefore Eu3+ luminescence properties can be used to probe in a unique way the

Langmuir, Vol. 21, No. 5, 2005 1783

coacervation process. That information can contribute to elucidate the intricate and not yet fully understood process of coacervation. Acknowledgment. The authors thank Professor Dr. Pierre Vast and Vitor Amaral for help on discussions about coacervates and the Brazilian agencies CNPq, CAPES, and FAPESP for financial support. Additional support from the Brazil-Portugal cooperation program (CAPESGRICES Project 061/00) is also acknowledged. LA0476837