Formation and stability of lanthanide complexes and their

---

J . Phys. Chem. 1992, 96, 8626-8631

8626

(58) De Schryver, F. C.; Demeyer, K.; van der Auweratr, M.; Quanten,

E. Ann. N . Y. Acad. Sei. 1981, 366, 93. (59) Seki,K.; Ichimura, Y.; Imamura, Y. Macromolecules 1981,14, 1831. (60)Seki, K.; Imamura, Y. Bull. Chem. Soc. Jpn. 1982, 55, 37 1 1 . (61) Valeur, B.; Monnerie, L. J . Polym. Sci., Polym. Phys. Ed. 1976, 14.

11. (62) Viovy, J. L.; Monnerie, L. Macromolecules 1983, 16, 1845. (63) Sasaki, T.; Yamamoto, M.; Nishijima, Y. Makromol. Chem., Rapid Commun. 1986, 7, 345. (64) Hyde, P. D.; Waldow, D. A.; Ediger, M. D.; Kitano, T.; Ito, K. Macromolecules 1986, 19, 2533. (65) Waldow, D. A.; Johnson, B. S.; Hyde, P. D.; Ediger, M. D.; Kitano, T.; Ito, K. Macromolecules 1989, 22, 1345. (66) Heatley, F.; Wood, 8.Polymer 1978, 19, 1405. (67) Allerhand, A.; Hailstone, R. K. J . Chem. Phys. 1972, 56, 3718. (68) Laupretre, F.; Noel, C.; Monnerie, L. J. Polym. Sci., Polym. Phys. Ed. 1977, 15, 2127.

(69) Heatley, F.; Begum, A. Polymer 1976, 17, 399. (70) Matsuo, K.; Kuhlmann, K. F.; Yang, H. W. H.; Geny, F.; Stockmayer, H.; Jones, A. A. J . Polym. Sci., Polym. Phys. Ed. 1977, 15, 1347. (71) Glowinkowski,S.;Gisser, D. J.; Ediger, M. D. Macromolecules 1990, 23, 3520. (72) Wmsner, D. E. J . Chem. Phys. 1962, 36, 1 . (73) Valeur, B.; Jarry, J. P.; Geny, F.; Monnerie, L. J. Polym. Sci., Polym. Phys. Ed. 1975, 13, 667. (74) Schaefer, J. Macromolecules 1973,6, 882. (75) Gronski, W. Makromol. Chem. 1979, 180, 1119. (76) Denault, J.; Prud'homme, J. Macromolecules 1989, 22, 1307. (77) Tekely, P. Macromolecules 1986, 19, 2544. (78) Lewis, R. J.; Pecora, R.; Eden, D. Macromolecules 1986, 19, 134. (79) Mahr, H.; Hirsch, M. D. Opr. Commun. 1975, 13.96. (80) Gochanour, C. R.; Fayer, M. D. J . Phys. Chem. 1981, 85, 1989.

(81) Todd, D. C.; Jean, J. M.; Rosenthal,S.J.; Ruggiero, A. J.; Yang, D.; Fleming, G. R. J . Chem. Phys. 1990, 93, 8658.

Formation and Stability of Lanthanide Complexes and Their Encapsulation into Polymeric Microspheres Russell J. Mumpert**.* and Michael Center of Membrane Sciences and Division of Medicinal Chemistry and Pharmaceutics, College of Pharmacy, University of Kentucky, Lexington. Kentucky 40536 (Received: February 1 1 , 1992; In Final Form: June 16, 1992)

The complexation of lanthanides (Ln) with dicarbonyl compounds (acetylacetone, acac; ethyl acetoacetate; 3-ethyl-2,4and diethyl malonate) was investigated using a potentiometric pentanedione; 2,4-hexanedione; 3-methyl-2,4-pentanedione; titration technique. The ability of a dicarbonyl compound to complex with the lanthanide elements was greatly dependent on its pK, and on the pH of the titrated solution. Selected lanthanide complexes (Ln complexes) were incorporated into spherical poly(~-lacticacid) (PLA) matrices and irradiated in a nuclear reactor with neutrons to produce short-lived high-energy ~-particle-emittingradioisotopes. The lanthanidesinvestigated (Ho,Dy, Sm, and La) were chosen on the basis of their physical and nuclear properties. A transition element (Re) was also studied. The small decrease in the ionic radii of the lanthanides with increasing atomic number led to (a) greater ability to extract and complex from an aqueous solution with complexing agents, (b) larger formation and stability constants for the Ln complexes, (c) increased solubility of the Ln complexes in chloroform, and (d) increase in the maximum percent incorporation of the stable lanthanides in PLA spheres. Ho(acac), was found to be the most promising candidate of the complexes studied on the basis of the above observations and due to the favorable physical properties of 1 6 5 Hand ~ nuclear properties of laHo.

Introduction The complexation and extraction of the lanthanides (Ln) utilizing a variety of complexing agents has been reported.'-3 The dicarbonyl chelates of the tripositive lanthanides, such as those of acetylacetone, are among the most stable of the known complex species! Complexes of this sort are of use in ion-exchange chromatography,s separation of by-products of fissioned uranium? and, more recently, as catalyst systems.' The &diketone chelates, and other dicarbonyl compounds, are of particular interest in our work since their complexes with lanthanide metals have favorable physical stability and relatively high thermal stability, and consist of a high weight percentage of the lanthanide metal. In addition, these complexes have suitable solubility in organic solvents. We have incorporated these Ln complexes into biodegradable polymeric matrices for use as radiotherapeutic agents.*-' Although the lanthanide metals have a similar ground-state electronic configuration, which leads to similar chemical reactivities, small differences in their ionic radii dramatically affect their ability to be complex4 with these chelating agents. The stable elements in Table I can absorb neutrons to become short-lived high-energy 8-particle-emitting radioisotopes (Le., half-lives 0.8 MeV). Each of the radioisotopes has a small photon yield (4-29%) associated 'Center of Membrane Sciences. *Division of Medicinal Chemistry and Pharmaceutics. i Current address: Center for Bioengineering, FL-20, University of Washington, Seattle, WA 98195.

TABLE I: Physical and Nuclear Properties of Candidates for hcomoratioo into Biodmndable PLA Microspberes stable

natural

nuclide

abundance

89Y

'39La lslpr %m IMDy

165H~ lsJRe I8'Re

100 99.9 100 26.7 28.1 100 37.0 63.0

ai

(b) 1.3

radionuclide

9.0

%a I4Pr

11.5 206 2700 64 112 73

"Y 15'Sm 16JDy

166H~ IE6Re lssRe

half-life

E@max

(h)

(MeV)

64.1 40.1 19.2 46.0 2.3 26.9 90.9 16.7

2.28 2.18 2.16 0.81 1.29 1.84 1.08 2.12

with its decay enabling one to quantify the radioactivity by yscintillation or externally image with a y-camera. This paper investigates the differences in lanthanide ionic radii and how these differences influence the formation of large quantities of stable Ln complexes for encapsulation into poly(Llactic acid) microspheres for later neutron irradiation.

Experimental Seetion Matori.bn Holmium chloride hexahydrate (99.9%),dysprosium chloride hexahydrate (99.9%), samarium chloride hexahydrate (99.9%),lanthanum chloride heptahydrate (99.9%), and anhydrous rhenium chloride (99.9%) were obtained from AESAR/Johnson Matthey. The following complexing agents, acetylacetone (acac), 2,4-hexanedione, ethyl acetoacetate, 3-ethyl-2,4-pentanedione, 3-methyl-2,4-pentanedione, and diethyl malonate, were purchased

0022-365419212096-8626$03.00/0 0 1992 American Chemical Society

The Journal of Physical Chemistry, Vol. 96, No. 21, 1992 8627

Formation and Stability of Lanthanide Complexes

R3

0

0

II

- IIC -

- C-

CH

R1

I R2

(A)

acetylacetone

(333

H

a 3

(B)

ethylacetoacetate

WISH3

H

a 3

(C)

3-ethyl 2,4-pentanedione

CH3

(3.k"

a

(D)

2,4-hexancdione

a 3

H

m 3 m Z

(E)

3-methyl 2,4-pentanedione CH3

CH3

M 3

3

(F) diethyl malonate 3" H a33cHzo Figure 1. Structures of possible complexing agents for the tripositive lanthanides.

from Aldrich and used without further purification. The structures of these complexing agents are shown in Figure 1. Poly@-lactic acid) (PLA, MW (as supplied) = 57 000) was obtained from the Henley Company. Poly(viny1 alcohol) (PVA; 88% hydrolyzed; MW = 78 OOO), chloroform (technical grade), inositol, and standard solutions of sodium hydroxide (1.038 N) and ammonium hydroxide (4.96 N) were purchased from Aldrich. Preparation of La Complexes. Ln complexes were formed by a modified method of Brown et al.' The general procedure was as follows: 46.5 mL of a solution at pH 2.5 containing approximately 5 g of the lanthanide chloride, approximately 6 mL of the complexing agent, and 500 pL of hydrochloric acid was titrated with a standard solution of sodium hydroxide (1.038 N) or ammonium hydroxide (4.96 N) to raise the pH to 7.34. The lanthanide ion concentration was 0.28 M, and the complexing agent concentration was 1.25 M. All titrations were performed with continuous stirring at 25 OC. Sufficient time was allowed after each addition of base so that the solution could reach equilibrium, and the pH of the titrated solution was recorded. The precipitated complexes were collected on preweighed filter paper, washed with 300 mL of distilled and deionized water to remove excess complexing agent, and stored over calcium sulfate for 3 days. The efficiency of a complexing agent to complex the lanthanides from the aqueous solutions was calculated by dividing the mass of the complexed lanthanide by the original lanthanide available in the titrated aqueous solution. The structures of the Ln complexes were confirmed by infrared spectroscopy, mass spectroscopy, and both 'H-NMR and I3C-NMR spectroscopy.I0 Neutron Irradiationof La Complexes by Californium-252 for Analysis. The 1.25-mg 252Cfneutron source at the University of Kentucky Department of Chemistry was used to neutron irradiate the complexed lanthanides to produce their corresponding neutron-rich short-lived radionuclides. Spontaneous fission of 252Cf produces 3.8 neutrons per fission, which is capable of activating lanthanides. The small neutron flux density of lo6 n/(cm2 s) produced sub-microcurie amounts of activity that was used for determining the percent incorporation of the lanthanide complexes in microspheres and for producing a radiolabel that was used to determine the solubilities of Ln complexes. The solubilities of the neutron-irradiated Ln complexes in chloroform were determined by the phase solubility technique as described by Mader.12 For determining incorporation percentages, microsphere samples were neutron irradiated along with lanthanide standards to saturation and m t e d in a y-scintillation counter (NaI(Tl) detector). Preparation of PLA Microspheres with Ln Complexes. Microspheres containing Ln complexes were prepared by the solvent evaporation technique as described previously.8-'I Briefly, a dispersed phase consisting of 50-500 mg of Ln complex and 5 0 0 mg of PLA in 10 mL of CHC13 was slowly added to a stirring

continuous phase consisting of 80 mL of 2% w/v PVA in deionized H20. Stirring was maintained for 15 min at 92&1140 rpm. The oil-in-water emulsion was then transferred to a 1000-mL roundbottom flask and diluted with 100 mL of deionized HzO, and the chloroform was removed by rotary evaporation. The precipitated spheres were filtered and collected on 20-pm nylon filter paper. To remove unincorporated lanthanide complex, the fdtcred spheres were washed with 800 mL of 0.1 N HCl and refiltered. The spheres were then rinsed with deionized HzO and stored in a desiccator. High Neutroa Flux M t y Irradiation of S e W d Sphereswith Ln Complexes. To produce therapeutic amounts of activity (25-mCi levels and greater), selected sphere samples were irradiated at (1) the University of Missouri Research Reactor (MURR) in a thermal neutron flux density of 8.0 X 1013n/(cm2 s) and an epithermal neutron flux density of 2.0 X 10l2n/(cm2 s) or at (2) the TRIGA Reactor at the University of Illinois in a thermal neutron flux density of 8.9 X 10l2n/(cmz s) and an epithermal neutron flux density of 7.1 X 10" n/(cm2 s). For all irradiations, 50 mg of PLA spheres and 150 mg of inositol, used as a diluent, were placed in high density polyethylene vials. The amount of activity produced (Aloai) when spheres containing lanthanides were irradiated in a reactor was calculated by A,,,

= N4ut(1 - c X T i )

(1)

where A is 0.693/tl,z. AIoaIis directly related to the number of target atoms (N),the thermal neutron flux density (I$, in n/(cm2 s), and the neutron capture cross section of the atoms (ut, in barns where 1 barn = cm2). Therefore, to reduce the time of irradiation (Ti,in hours) needed to produce therapeutic amounts of activity, it is essential to maximize these factors.

Results and Discussion Physical and Nuclear Properties of Stable and Irradiated Lanthanides. The use of radionuclides for internal radiation therapy has become increasingly popular in the past three decades, primarily in the fields of oncology, rheumatology, and endocrinology. With the advancement of particulate carriers such as nanospheres, microspheres, and liposomes, and the emergence of molecular carriers such as monoclonal antibodies, selective targeting has been made possible. We have been investigating biodegradable microspheres as carriers for radiotherapeutic agents to be used as improved treatments for hepatic metastases and rheumatoid arthritis.*v9 Treatment of these conditions using internal radiation therapy with a variety of radionuclides has been reported. I3,l4 In determining the most appropriate lanthanide to be encap sulated into biodegradable microspheres for later neutron flux density irradiation, it is apparent that many factors must be considered. One must consider (a) its ability to be complexed and incorporated into spheres, (b) its potential toxicity, (c) the physical characteristics of the stable isotope which enable it to be neutron activated to its corresponding radionuclide in a short irradiation time, and (d) the nuclear characteristicsof the activated radionuclide which make effective radionuclide therapy possible for its intended application. To reduce the time of irradiation necessary to produce therapeutic amounts of activity, the appropriate stable lanthanide must be incorporated in the spheres in sufficient masses and have associated with it a high neutron capture cross section. It is not necessary that the stable lanthanide have a large natural abundance since enriched stable target nuclides can be obtained, although it is desirable. The higher neutron capture cross sections for 1 3 9 b,l4lR,152sm, IwDy, 165H0,la5Re,and IS7Re(as compared make these stable lanthanides especially attractive for their to 89Y) incorporation and subsequent neutron irradiation in a biodegradable PLA matrix due to the fact that excessive irradiation may have unfavorable chemical and physical effects on PLA. However, I4IPr was not considered for further study since the corresponding radioisotope, 142Pr,has a high y-energy of 1576 keV (3.7% photon yield) which would cause unnecessary radiation exposure to healthy cells.

Mumper and Jay

8628 The Journal of Physical Chemistry, Vol. 96, No. 21, 1992 3.0

2.5 2.0 1.5 1.o

0.5 L

I

I

I

I

1.04 1.06 1.08 1.10 1 . 1 2

Ionic Radius

1.14

7

(A) of Ln

On the basis of preliminary investigations of the toxicological, physical, and nuclear properties of the above-mentioned lanthanides, we preceded with the evaluation of Ln complexes of lanthanum, samarium, dysprosium, and holmium and of those complexes with the transition element rhenium. Acetylacetonate Complexes. The acetylacetonate complexes of holmium, dysprosium, samarium, and lanthanum were freeflowing powders of high purity. Rhenium did not form a complex with acetylacetone when the pH was raised. In all preparations with acetylacetone, there was no evidence of hydroxide formation. It was concluded that hydrolysis of the lanthanides occurred only to an insignificant extent. This conclusion can be demonstrated by an estimate of the relative amounts of [Ln(OH)J2+and [Ln(acac)12+in a titrated solution at pH 7.5. The most strongly hydrolyzed lanthanide is Lu3+with a hydrolysis constant of The formation constant, kl,for [Lu(acac)l2+is lo6, and by derivation, it can be shown that [Lu(0H)l2+

1.00

1 . 1 6 1.18

Figure 2. Effect of lanthanide ionic radius on their ability to be complexed and extracted from an aqueous phase with acetylacetone. The aqueous solutions of the lanthanides were raised to pH 7.34 with (D) NaOH and ( 0 )NH40H.

[Lu(acac)] 2+

RA2 = 0.991

0.0

= 0.02

Thus, at pH 7.5,even the most strongly hydrolyzed lanthanide, Lu3+,is mostly in the form of the complex, owing to the large formation constants of the L n ( a ~ a c complexes. )~ The ability of acetylacetone to complex the lanthanides was found to be dependent on several factors including (a) the base used (NaOH of NH40H) to titrate the solution, (b) the pH of the titrated solution, (c) the ionic radius of the metal ion, and (d) the ionic strength of the titrated solution. Maximum Complexation Percentages of Lanthanides with Acetylacetoae. The results of the acetylacetonecomplexation of lanthanide ions in aqueous solutions using NaOH or NH40H as the titrating base are shown in Figure 2. Previous investigators have reported the use of NaOH2JSand NH40H'v4.'6as the titrating base; however, no prior work has utilized both bases and compared the differences in the maximum lanthanide complexed when they are used under identical conditions. Stites, McCarty, and Quillls used N H 4 0 H as the titrating base in their work published in 1948. The group utilized ammonium to first solubilize acetylacetone (S = 1 part/l parts of H 2 0 ) as ammonium acetylacetonate in solution at pH 5.5 and then followed with the addition of NH40H. Aqueous ammonia or NH4+forms a complex with acetylacetone and presents a competing complexation reaction with lanthanide ions in solution. Therefore, one would expect decreased complexation percentages of the lanthanides when NH40H is used. On the other hand, Rydberg16has reported that no complexes are formed between sodium and acetylacetone. Only acetylacetone in the form of acac- can complex Ln3+,and since the concentration of acac- present in a solution is a function of pH, the complexation of acac- and Ln3+ should also be a function of pH. We have reported, in fact, that the amount of Ho3+ complexed with acetylacetone is a function of pH with maximum complexation occurring at pH 7.34.1° This pH was the most practical point for end titration of all the lanthanides since complexation with acetylacetone was optimized (Le., the

1.04

1.08

1.12

1.16

1.20

Ionic Radius (A) of Ln Figure 3. Effect of lanthanide ionic radius on the Ln(acac)3 solubility (S) in chloroform.

agent was more ionized) and the hydroxide of the lanthanides had not yet formed. Another factor that has been found to influence the ability of the lanthanides to be complexed with acetylacetone is the ionic radius (A) of the metal ion. An overall decrease in the ionic radii occurs from lanthanum to holmium. The effect as shown in Figure 2 is a decrease in the ability of the lanthanides to be complexed with acetylacetone as the ionic radius of the lanthanidesincreases. The two intersecting negatively sloped lines experienced is similar to the results obtained by Brown et al.' Brown suggested that variation in the maximum percent complexed could be attributed to the extent of hydrolysis of the metal ion or changes in solubility of the complexes. It was expected that since holmium (of the lanthanides studied here) has the smallest ionic radius, it would experience the largest extent of hydrolysis. This was also predicted theoretidy whereby the hydroxide of holmium should precipitate at a lower pH than the corresponding lanthanide hydroxides. In fact, Brown et al. concluded that the reverse is true, that holmium exhibits the smallest amount of hydrolysis when extracted and complexed with acetylacetone. Due to the lanthanidecontraction, ions of the heavier lanthanides such as holmium and dysprosium are smaller than the methyl groups of an octahedral acetylacetonate complex. Thus, the methyl groups would be closer and perhaps hinder the attachment of water, resulting in anhydrous complexes. The trend of smaller lanthanides to form less hydrated complexes has been discussed by Brittaine3 Finally, these anhydrous complexes formed with the heavier lanthanides in the aqueous solution, having less water solubility, would tend to precipitate leading to the higher overall complexation percentage observed for these lanthanides. The reduced complexation percentages for monohydrated lanthanide acetylacetonates, as reported by Brown et al., provide evidence for the solubility hypothesis as stated above. Additional evidence for this hypothesis is provided by solubility data as shown in Figure 3. There is an excellent correlation between the ionic radius of the lanthanides and their corresponding L n ( a c a ~ complex )~ solubility in CHC13. F a r " lad SbMlity Chwmts of Ia(acacbcomplexes The formation constants (kl, k2,and k3) and the stability constants (B k,k2k3)were determined by the average number method" described by Bjerrum. Titration curves were obtained by adding a strong base to a solution of acetylacetone, and to another containing acetylacetone and lanthanide ions, and then plotting the pH against the volume of base added. The results of the titration curves may be treated quantitatively in order to detennine the stoichiometricratio of acetylacetone to lanthanide and to obtain a quantitative expression of the stability constant for complex formation of the Ln(acac)3 complexes. The average number of acetylacetone molecules bound per metal ion present is denoted as fi and is defined by [acac b o ~ n d l ~ ~ ~ ~ l ii= (3) [Ln3+1tota1 The formation constants can be substituted for the individual complex concentrations, and if may be quantified as kl[acac-] + 2klk2[acac-12+ 3kIk2k3[acac-l3

=

if=

I

+ k,[acac-] + klk2[acac-12+ klk2k3[acac-13 (4)

Formation and Stability of Lanthanide Complexes

The Journal of Physical Chemistry, Vol. 96, No. 21, 1992 8629 7.0 3

41

: 3 .5 3.0

1.5

2.5

3.5

4.5

5.5

8.0

6.5

8.5

9.0 9.5 10.0 1 0 . 5 11.0

p[acac-]

Figure 4. Formation curve for the Ho(acac), complex. The average number ( 8 ) of acetylacetone molecules bound per metal ion as a function of p[acac-1.

TABLE Ik log k, Values for the Complexation of Tripositive Lanthanides with Acetyhcetone complexation conditions" metal T = 25 OC T = 25 OCb T = 30 OC' T = 30 O C d ion log k, p = 1.70 M p = 1.70 M p = 0.1 M p = 0.0 M HO IOR k, 5.81 f 0.02 6.07 f 0.01 6.05 f 0.01 4.68 f 0.01 log k; 4.64 f 0.04 ND 3.40 f 0.10 log k3 3.54 f 0.01 ND Dy log kl 5.36 f 0.03 5.77 f 0.00 6.03 f 0.01 4.67 f 0.02 log k2 4.22 f 0.1 1 ND 3.34 f 0.04 log k3 3.45 f 0.30 N D 5.9 Sm log k, 5.20 f 0.03 5.71 f 0.04 5.59 f 0.02 4.46 f 0.01 4.5 log kz 3.99 f 0.08 N D 3.2 2.90 f 0.01 log k3 3.04 f 0.08 ND La log kl 3.90 f 0.05 4.50 f 0.04 4.96 f 0.03 5.1 3.45 f 0.05 3.8 log kz 2.80 f 0.02 ND 2.50 f 0.10 3.0 log k3 1.63 f 0.58 ND "The standard deviations in the log k values are obtained as the maximum deviations from the different mean values obtained by using different sets of 8 and placac-] values in the simultaneous equations. All data shown were the result of titrations using NaOH unless stated otherwise. * p H raised with NH40H. 'Results of Grenthe et a1.2 dReSults of Jzatt et al.I9

The value ii at any pH can be determined from the titration curves since the horizontal distance between the two curves at any given pH gives the amount of OH- consumed in the reaction, which is exactly equal to the concentration of acetylacetone bound at any pH. The concentration of free acetylacetone, as acac-, at any pH is obtained with knowledge of the acid dissociation constant for acetylacetone. The concentration [acac-1, at any pH, is equal to the difference between the initial concentration [acac-Ii of acetylacetone and the concentration of [OH-] added. It follows that [acac-] =

Ka([acacli - [OH-]) [H+l

and taking the logarithm of both sides leads to -log [acac-] = p[acac-] = pK, - pH - log ([acacIi - [OH-]) (6) Therefore, since acetylacetone is a weak acid (pKa 8.9),values of if and p[acac-] at any pH can be obtained with knowledge of [H+] and the original concentrations of the substances in the solution. Values of fi and p[acac-] can then be plotted as in Figure 4. The formation curve reaches a limit for ii of 3, which signifies that three molecules of acetylacetone can combine with one lanthanide ion. Similar plots and ii values were obtained for all complex titrations of the lanthanides. The determinant method described by Block and McIntyre'* was used to find kl, kz, and k3 from pairs of ii and p[acac-] values. The results of our analysis of the potentiometric titrations of the lanthanides and acetylacetone are contained in Table I1 and are compared with those reported by Grenthe et aL2and Izatt et aI.l9 The data indicate that there is a relationship between the formation constants obtained and the ionic radius of the lanthanide. Holmium, having the smallest ionic radius, has the highest values for

Ionic Potential (Z2/r)

Figure 5. log K,as a function of ionic potential, P/r,for the Ln(acac), complexes. The aqueous solutions of the lanthanides were raised to pH 7.34 with (m) NaOH and (0)NH40H. Results are compared to data (0)of Grenthe et al.*

the formation constants. Conversely, lanthanum has the largest ionic radius and has the lowest values for the formation constants. Furthermore, the logarithm of the stability constant, log 8, obtained for the acetylacetonates of Ho, Dy, Sm, and La of 13.99, 13.03, 12.23, and 8.33,respectively, provide evidence for the strong complexation between acetylacetoneand the heavier lanthanides of holmium and dysprosium. Another parameter that may be studied in order to obtain some knowledge of the stability of the Ln(acac)3 is the bond strength of acetylacetone to the lanthanide. If the lanthanidmcetylacetone bond is truly ionic, that is (acac- 'Ln), the strength of the bond should increase linearly with increasing ionic potential P / r , where 2 is the charge of the lanthanide ion and r is the radius. Typically, a measure of the bond strength for the reaction (Ln3++ acac= [Ln(acac)12') is AH, or the difference in bond energies or attractive energies between product and reactants. The free energy change (AFj is sometimes used in place of AH due to the experimental difficulties in obtaining AH. Therefore, AF = -RT In K1,where K1 is equal to the first formation constant kl, since K, = klk2k k,. This equation for AFis a good approximation since the entropy changes AS for these reactions of acetylacetone and lanthanide ions are usually very smalll2and thus, AF = AH - T U becomes A F AH. In Figure 5 , the ionic potential of the lanthanides is plotted against the log KI values from Table 11. Although there does exist a positive relationship between ionic potential of the lanthanides and log KI, the relationship is only linear initially and appears to become relatively constant. Grenthe et al.z observed this same phenomenon for the complexation of acetylacetone with the lanthanides, and only Ln complexes of ethylenediaminetetraacetic acid (EDTA) and cyclohexanediaminetetraaceticacid (DCTA) show a linear relationship between ionic potential and log K1. The plateau observed may be due to changes in the number of attachment points of the acetylacetone ligand around the smaller lanthanides. These configurational changes in the ligands will result in changes in both AH and AS, and AF a AH would no longer be valid. Infrared studies by Moeller et al.zoand thermodynamic work by Betts and Dahlinges" proved that configurational changes can occur in some ligands. However, it has been suggested that steric effects, inducing potential configurationalchanges, for the acetylacetone ligand was improbable and that the less than expected stability for the higher lanthanides cannot be explained by electrostatic bond theory.2 Nevertheless, Figure 5 d m emphasize the increased stability and bond strength of the acetylacetone to the heavier lanthanides such as holmium and dysprosium. Finally, another factor that influences the ability of acetylacetone to extract and complex the lanthanides from an aqueous solution is the ionic strength of the solution. The large ionic strengths ( p = 1.70 M) used in this study were the direct result of the very large initial concentrations of lanthanides ions and acetylacetone. The data in Table I1 indicate that the formation constants decrease with increasing ionic strength of the solution. This result appears to agree with classic DebyeHuckel predictions; however, the analysis is more complicated due to the fact that the data obtained in this study were done so at both decreased

,...

-

8630 The Journal of Physical Chemistry, Vol. 96, No. 21, 1992 TABLE 111: Influence of Ln(ncnc)3Solubilities (S) on the Incorporation in Spheres (% I w/w) and the Time of Irradiation (Ti) Needed To Produce Therapeutic Amounts of Activity (A ,J)" Ti when S(CHC13) max % I stable (w/w) A. = 25 mCi (h) nuclide (mg/mL) 1.3 400b 11.3 139La 6.4 21.8 13.8 152Sm 31.8 0.02 103.4 164Dy 1.5 235.3 36.3 1 6 5 ~ ~

" Irradiation of 50 mg of spheres in reactor power of 600 kW corresponding to a thermal neutron flux density of 8.88 X 10l2 n/(cm2 s) with an additional epithermal neutron flux density of 7.10 X IO" n/ (cm2 s). bAo= 2.2 mCi. temperature and increased ionic strength. In addition, the Debye-Huckel theory for the influence of ionic strength on the reaction rate stipulates that the reaction occur in dilute aqueous solutions of