Structural study of trivalent lanthanide and actinide complexes formed upon solvent extraction

Abstract

The coordination of the trivalent 4f ions, Ln = Nd³⁺, Eu³⁺ and Yb³⁺, as well as the trivalent 5f ion, Am³⁺, with diamide and dialkylphosphoric acid extractants, individually and in combination, was studied by use of X-ray absorption spectroscopy. These studies provide metrical information about the interatomic interactions between the f-ions (M³⁺) and the ligands, dihexylphosphoric acid (HDHP) and N,N′-dimethyl-N,N′-dioctylhexylethoxymalonamide (DMDOHEMA), that is of practical relevance to the control of metal–ligand binding in liquid–liquid extraction systems for the separation of trivalent actinide ions, An³⁺, from trivalent lanthanide ions, Ln³⁺. Through systematic variations of extraction conditions and extractant combinations, we have found that the HDHP complexes with M³⁺ involve MO₆ coordination and distant M⋯P interactions, whereas the DMDOHEMA complexes with M³⁺ involve MO₈ coordination. The combination of the EXAFS results with ancillary extraction data and IR results facilitates descriptions of the stoichiometries and structures of the molecular species formed in solution upon liquid–liquid extraction and leads to a new understanding of the binary extraction systems in terms of the strength and selectivity of An³⁺- vs. Ln³⁺-ligand interactions. This fundamental structure information affords insight into solvent extraction processes that are of contemporary and practical importance in heavy element chemistry and to environmentally related issues arising from the separation and disposal of radioactive materials, particularly actinides and selected fission products, in the field of nuclear waste reprocessing research.

---

Introduction

Current research in the field of nuclear waste reprocessing aims at separating long lived radionuclides, specifically the trivalent actinide ions, An³⁺, from the trivalent lanthanide ions, Ln³⁺, by solvent extraction. One of the newest processes developed to address this longstanding challenge in radioanalytical sciences involves the contact of an aqueous phase, containing both An³⁺ and Ln³⁺, with an organic phase, containing a mixture of two extractants, particularly a diamide and a dialkylphosphoric acid, in an aliphatic diluent.^1–3 This binary mixture is more efficient and more selective at separating An³⁺ from Ln³⁺ than other solvents based singularly on either diamides, for which there is no An³⁺/Ln³⁺ separation, or on dialkylphosphoric acids, for which there is no extraction from very acidic aqueous solutions. Nonetheless, because the extraction system is rather complex, liquid–liquid extraction results are not easily interpreted⁴ largely because of a lack of structure information about the coordination chemistry of An³⁺ and Ln³⁺. Such details are pivotal to a predictive understanding of the chemical interactions responsible for the selectivity and the extraction properties of different liquid–liquid extraction systems^5,6 and, ultimately, to advances in the contemporary understanding of synergistic effects.

To provide metrical knowledge that is otherwise not available, we have characterized the coordination environments of the complexes formed by two light lanthanides, Nd³⁺ and Eu³⁺, one heavy lanthanide, Yb³⁺ and one trans-plutonium actinide, Am³⁺, with two different extractants—both alone and in combination. Through use of EXAFS, we studied the Ln³⁺ and Am³⁺ complexes with DMDOHEMA (N,N′-dimethyl-N,N′-dioctylhexylethoxymalonamide) and HDHP (dihexylphosphoric acid), which are illustrated below.

Although no structural studies of 4f/5f-ion (abbreviated collectively hereafter as M) complexes with either of these two extractants are available, reports for the HDEHP extractant, di-2-ethylhexylphosphoric acid, of similar molecular composition to HDHP show that different organic complexes are formed in n-dodecane with An³⁺ and Ln³⁺. Specifically, an EXAFS investigation has revealed that the Nd³⁺ coordination environment in complexes with HDEHP changes as a function of the Nd : HDEHP ratio.⁷ Similarly, EXAFS studies of M³⁺ complexes of dialkyldithiophosphinic acid, dialkylmonothiophosphinic acid and dialkylphosphinic acid in n-dodecane demonstrate that coordination numbers and bond lengths are important molecular-level indicators of solvent-extraction performance.^8,9 With regard to the other extractant of interest here, i.e., DMDOHEMA in n-dodecane, much less ancillary structure information is available. The solvent-phase complexes of Ln³⁺ and Am³⁺ with TEMA (N,N,N′,N′-tetraethylmalonamide) are of similar functionality and have been successfully investigated by several techniques, including EXAFS.¹⁰

In addition to studies of the two single-extractants—DMDOHEMA and HDHP—on their own, we have investigated the mixed extractant system, which combines DMDOHEMA and HDHP, in various ratios. No metrical information about the organic phase M³⁺ speciation with this extractant combination is available. We wish to determine whether the Ln³⁺ and Am³⁺ environments in the binary-extractant systems consist of hybrid complexes with structures and stoichiometries that are different from those in the corresponding single-extractant systems. Alternatively, the M complexes in the binary systems may be simple admixtures of the M complexes that prevail in the two single systems. Because extractants having similar structures can exhibit different coordination modes⁹ when employed in simultaneous combination, in contrast to independent use, the direct comparison of EXAFS results for both single- and binary-extractant systems provides a deliberate means by which to search for a structure basis with regard to the effects of synergism and antagonism in liquid–liquid extraction processes.

Results and discussion

Single-extractant systems

M³⁺–HDHP. The k³χ(k) EXAFS data for the HDHP complexes (of series no. 5, Table 1) with Yb³⁺, Eu³⁺, Am³⁺ and Nd³⁺ (in order of increasing ionic radii) and the corresponding FTs (Fourier transforms) are provided in Fig. 1(a) and 2(a). The FT data show two peaks of physical significance—an intense one at 1.82–1.92 Å attributed to the nearest O neighbors, and a medium one at 3.32–3.41 Å attributed to the distant P atoms.¹¹ Based upon the appearance of M–O and M–P interactions in the FT data, all of the M k³χ(k) EXAFS were fit using a two-shell (O and P) backscattering model. The fits are shown in the plots provided as ESI.† The metrical results are collected in Table 2.

Table 1 Sample identifications, compositions, and solution concentrations in mol L⁻¹; n-dodecane was used as organic diluent. The water content for the single extractant systems of sample series 5 and 8 is ca. 0.05 and 0.2 mol L⁻¹, respectively, and ca. 0.5 mol L⁻¹ for the binary extractant systems of series 18, 19 and 20

		Concentrations/M
		Organic phase		Aqueous phase
Series no.	M^3+	HDHP	DMDOHEMA	M^3+	HNO3
5	Am	0.3		0.01	0.1
8	Am		0.7	0.01	3
18	Am	0.3	0.7	0.01	0.1
19	Am	0.3	0.7	0.01	1
20	Am	0.3	0.7	0.01	3
5	Nd	0.3		0.01	0.1
8	Nd		0.7	0.025	3
18	Nd	0.3	0.7	0.05	0.1
20	Nd	0.3	0.7	0.01	3
5	Yb	0.3		0.01	0.1
8	Yb		0.7	0.01	3
19	Yb	0.3	0.7	0.01	1
20	Yb	0.3	0.7	0.01	3
5	Eu	0.3		0.01	0.1
8	Eu		0.7	0.01	3
20	Eu	0.3	0.7	0.01	3

---

Table 2 Two-shell (M–O/P) fits to the k³χ(k) EXAFS, Fig. 1(a), for the 5-series M³⁺ complexes with HDHP. Values shown in parentheses are the estimated standard deviations (esds) at the 95% (3σ) confidence level

M^3+	Atom	r ^a/Å	CN^b	σ ^c/Å^2	ΔE^d/eV
a Interatomic distances. b Coordination numbers. c Debye–Waller factors. d Energy threshold values.
Nd	O	2.39(1)	6.7(9)	0.007(1)	−0.3(9)
	P	3.94(2)	5.1(20)	0.008(3)
Am	O	2.37(1)	5.6(6)	0.006(1)	−0.1(9)
	P	3.93(2)	2.9(18)	0.007(3)
Eu	O	2.30(1)	6.3(6)	0.006(1)	−0.6(9)
	P	3.86(1)	3.1(12)	0.004(3)
Yb	O	2.20(1)	6.0(6)	0.004(1)	0.6(9)
	P	3.77(2)	5.7(20)	0.008(3)

---

Fig. 1 M L₃-edge k³χ(k) EXAFS for single extractant complexes of trivalent Am, Nd, Eu and Yb with (a) HDHP and (b) DMDOHEMA. The Nd and Eu EXAFS data sets are shorter than those for Yb and Am (k_max = 14.3 Å⁻¹) due to the onset of their L₂-absorption edges. The data for Nd, Eu and Yb are offset for clarity of display.

Fig. 2 Fourier transforms of the k³χ(k) EXAFS data of Fig. 1 for single extractant complexes of trivalent Am, Nd, Eu and Yb with (a) HDHP and (b) DMDOHEMA. The O and P peaks in the FT data for the Nd and Eu complexes are broader than those for Yb and Am complexes because of their shorter EXAFS range.

The results indicate that each M³⁺ ion is six-coordinate with O atoms. The average EXAFS-determined M³⁺–O bond lengths exhibit a direct and essentially linear correlation with the M³⁺ IR (ionic radii, CN = VI)¹² as shown in Fig. 3(a). The linear least-squares fit (R² = 0.965) is shown as the solid line in this Figure, which also shows (open circles connected with dashed line) the predicted M–O distances based upon the sum of the M³⁺ and O²⁻ IR (1.35 Å, CN = II).¹² Comparison of the predicted and experimental distances reveals an overall correspondence that is typical for electrostatic metal–ligand interactions.^13,14 The predicted values for the Yb³⁺–O₆ and Eu³⁺–O₆ distances are within the error of the measurements, whereas the values for Am³⁺–O₆ and Nd³⁺–O₆ distances are somewhat shorter than the experimental ones. Nevertheless, the absolute magnitudes of the M–O distances provide support for the EXAFS-determined O coordination numbers of 6 ± 1. For example, the aquo ions of Yb³⁺, Eu³⁺, Am³⁺ and Nd³⁺, which are known to have higher (8–9) O coordination numbers,^15,16 have longer distances (+0.10–0.13 Å) consistent with their increased CNs. More to the point, there is precedent for 6 O coordination of Ln³⁺ and An³⁺ ions in complexes with dialkylphosphoric acids. A 6 O coordination of Eu³⁺ in the complex with HDEHP is known,¹⁷ and 6 O coordination of Nd³⁺, Sm³⁺ and Cm³⁺ with Cyanex-272 has been demonstrated.⁹ In addition, crystal structures of Ln³⁺ tris(diethylphosphate) and tris(dimethylphosphate) complexes show that the Ln ions are in six-coordinate, octahedral environments with O in Nd[O₂P(OEt)₂]₃¹⁸ and Ln[O₂P(OMe)₂]₃ for Ln = La,¹⁹ Sm,²⁰ and Eu.²¹ Although crystals of M³⁺ complexes with long-chain dialkylphosphates, such as di(2-ethylhexyl)phosphate, have proven to be unsuitable for conventional crystallography, texture analyses of powder X-ray diffraction data,²² in combination with ancillary spectroscopic data,²³ is consistent with Ln–O₆ coordination. As such, 6 O coordination of M³⁺ with HDHP is consistent with precedent. The question that arises with regard to the low coordination number is the source of the O. That is, does it arise from the HDHP ligand in either of its monodeprotonated (DHP⁻) or monodeprotonated dimer (HDHP·DHP⁻) forms? Alternatively, does either water and/or nitrate ion, or some combination of all of these possibilities, contribute to the inner coordination spheres about M? For insights into this issue we turn to the metrical analyses of the distant M–P interactions.

Fig. 3 (a) EXAFS-determined M³⁺–O bond lengths for HDHP complexes of M³⁺vs. M³⁺ IR (CN = VI). The linear regression (solid line) provides a slope of 1.6(2) and intercept of 0.8(2) Å. The dashed line is the sum of the M³⁺ IR and the IR (1.35 Å) for O²⁻ (CN = II). (b) EXAFS-determined M³⁺–O bond lengths for DMDOHEMA complexes of M³⁺vs. M³⁺ IR (CN = VIII). The linear regression (solid line) provides a slope of 1.41(3) and intercept of 0.91(4) Å. The dashed line is the sum of the M³⁺ IR and the IR (1.35 Å) for O²⁻ (CN = II).

The results of Table 2 show that the M–P distances span the range from 3.77 to 3.94 Å. As shown in Fig. 4, the plot of the average EXAFS-determined M³⁺–P bond lengths vs. the M³⁺ IR (CN = VI)¹² exhibits a direct and linear (R² = 0.967) correlation. This trend suggests that the mode of HDHP coordination to M via O is the same for all. Even in view of the large estimated standard deviations (±1.2–2.0 atoms at the 3σ, 95%, confidence level) for the P CNs, they distribute into two groups: 3 ± 1 for the Am³⁺ and Eu³⁺ samples, and 6 ± 2 for the Yb³⁺ and Nd³⁺ samples. The latter value of 6 P is consistent with the coordination of three monodeprotonated dimers to M in the form of the neutral species M(HDHP·DHP)₃ that has been reported beforehand.^9,17 In this mononuclear structure, illustrated in Fig. 5(a), three of the O atoms are associated with HDHP and three with DHP⁻. Moreover all six of these O are in monodentate, corner-sharing coordination with M. Such corner-sharing of O atoms in [O₂P(OR)₂]⁻ ligands with different M ions is a common motif for a number of network forming Ln³⁺ dialkylphosphate solids of general composition Ln[O₂P(OR)₂]₃.^18–22,24 Alternatively, a multinuclear structure could develop wherein the –OPO– moiety serves as a bridge between two Nd (and Yb) ions in the form illustrated in Fig. 5(b), wherein the bridging H atoms of Fig. 5(a) are replaced by M. The general stoichiometry of this extended structure, {Ln(DHP)₃}_n, is equivalent to that proposed for Eu tris(di-2-ethylhexyl)phosphate,²² wherein the O and P CNs are both six. The average degree of aggregation, n, for such Ln complexes in aliphatic naphtha can reach values of the order of hundreds.²⁵ Although we see no evidence in the FT data of Fig. 2(a) for either Nd⋯Nd or Yb⋯Yb interactions (expected at somewhat less than twice the M–P distances—ca. 6 Å) that would indicate a multinuclear species, such distant peaks are difficult if not impossible to observe in room-temperature EXAFS data of solution specimens.¹⁵ Nevertheless, our analysis of the small-angle X-ray scattering (SAXS) data obtained for the solutions of these Nd³⁺ and Yb³⁺ extractant complexes show evidence of polydispersity²⁶ that is consistent with the presence of both forms shown in Fig. 5.

Fig. 4 EXAFS-determined M³⁺–P bond lengths vs. M³⁺ IR (CN = VI) for HDHP series of complexes of M³⁺. The linear regression (solid line with R² = 0.968) provides a slope of 1.5(2) and intercept of 2.5(2) Å.

Fig. 5 Structures proposed for Nd³⁺ and Yb³⁺ complexes formed upon extraction with HDHP, including (a) neutral, mononuclear M(HDHP·DHP)₃ species obtained from M³⁺ coordination with three monodeprotonated dimers, and (b) neutral, multinuclear {M(DHP)₃}_n species obtained when M–O bonds replace the hydrogen bonds in (a) forming an extended structure, …O–M–O–P–O–M–O…,^{22, 23} with phosphate bridging groups of the three deprotonated ligands, DHP⁻. In both cases, the O CN is six and the P CN is six for M³⁺, consistent with the EXAFS results for Nd³⁺ and Yb³⁺ of Table 2.

The P CNs of three obtained for the Am³⁺ and Eu³⁺ samples is precisely and ostensibly explained by the formation of neutral species, Am(DHP)₃ and Eu(DHP)₃, assembled with bidentate, edge-sharing O coordination of three DHP⁻ ligands. In the structure illustrated in Fig. 6, each of the six O atoms arises from the anions. Although this mode of coordination is not uncommon for orthophosphates of Ln,²⁷ the problem here is that the M–P and M–O interatomic distances are inconsistent with this structure model. Specifically, the M–P distances are too long and the M–O interatomic distances are too short. Consider, for example, the average distances obtained from three Eu^27–29 and three uranyl(VI)^30–32 structures that exhibit both monodentate, corner sharing of phosphate O and bidentate, edge-sharing of phosphate O as illustrated in Fig. 7. There is a 0.5 Å difference between the M–P distances and a 0.2 Å difference between the M–O distances that is associated with the change of coordination. For our Eu³⁺ sample with a 2.30 Å Eu–O distance, which is the same as that for the M–corner-sharing-O interaction illustrated in Fig. 7 (bottom), the associated 3.86 Å Eu–P distance is consistent with the long distance expected for monodentate O coordination. Moreover, because there are no discontinuities in the variation of M–P distances with M IR (Fig. 4) that would be indicative of a change in the mode of coordination, the same mode is relevant throughout the series with Yb³⁺, Eu³⁺, Am³⁺ and Nd³⁺. With the P CN (three) accounted for in neutral clusters of stoichiometries Am(DHP)₃ and Eu(DHP)₃, and with the structure of Fig. 6 disregarded, the problem reduces to the realization of 6 O coordination. With three O attributable to the three DHP⁻ monoanions, we suggest that the remaining three O are accounted for by the presence of bound water, which is in 5× excess of the [M³⁺], Table 1. This is illustrated in the M[(DHP)₃(H₂O)₃] structure of Fig. 8, wherein H bonding is visualized (dashed lines) to provide stabilization. The combination of dialkylphosphate and water coordination about Ln has precedent in the structure of La[(O₂P(OMe)₂)₃(H₂O)]²⁴ wherein the difference between the Ln–OH₂ and Ln–OPO(OR)₂ interatomic distances is smaller than the expected resolution of our EXAFS measurement, 0.13–0.14 Å.

Fig. 6 Possible structure model of neutral, mononuclear M(DHP)₃ species obtained from bidentate coordination of Am³⁺ and Eu³⁺ by three deprotonated ligands. For M³⁺, the O CN is six and the P CN is three, consistent with the EXAFS results for Am³⁺ and Eu³⁺ of Table 2.

Fig. 7 (Top) Bidentate, edge-sharing of two phosphate O with M. (Bottom) Monodentate, corner-sharing of one phosphate O with M. The M–O and M–P distances are diagnostic of the phosphate denticity. The EXAFS results are consistent with the monodentate interaction shown at the bottom.

Fig. 8 Structure proposed for Am³⁺ and Eu³⁺ complexes formed upon extraction with HDHP, consisting of a neutral, mononuclear M[(DHP)₃(H₂O)₃] species in which the O CN is six and the P CN is three. Hydrogen bonds with the terminal and linking O atoms are shown as dashed lines.

M³⁺–DMDOHEMA. The k³χ(k) EXAFS data for the DMDOHEMA complexes (of series no. 8, Table 1) with Yb³⁺, Eu³⁺, Am³⁺ and Nd³⁺ and the corresponding FTs are provided in Fig. 1(b) and 2(b). The FTs show one intense peak at 1.87–1.97 Å of physical significance attributed to the nearest O neighbors.²² Unlike the FTs of the M³⁺ EXAFS data for the HDHP complexes (Fig. 2(a)), the FT data for the M³⁺ DMDOHEMA complexes shown in Fig. 2(b) present no strong evidence for systematic distant correlations. There are, however, a number of peaks in the range ca. 2.6–3.7 Å of possible structural significance. Their weak and varying intensities and positions are likely the result of spectral congestion, wherein many different scattering paths are smeared out and dampened by myriad variations of absorber–backscatter phases. Because of this complicated interference it is not practical to fit discrete interatomic interactions to the high r′ features in the FT data. Instead, we turn our attention to the details obtained from the one-shell nonlinear least squares curve fitting analyses of the large and intense O backscattering peaks. The fits are shown in the plots provided as ESI†. The results of the metrical analyses are collected in Table 3.

Table 3 One-shell (M–O) fits to the k³χ(k) EXAFS, Fig. 1(b), for the 8-series M³⁺ complexes with DMDOHEMA. Values shown in parentheses are the esds at the 95% (3σ) confidence level. The metrical parameters have the same definitions as those in Table 2

M^3+	Atom	r/Å	CN	σ/Å^2	ΔE/eV
Nd	O	2.47(1)	7.9(9)	0.009(1)	−0.8(9)
Am	O	2.44(1)	7.5(9)	0.010(2)	−0.9(9)
Eu	O	2.40(1)	8.3(12)	0.008(2)	−0.6(9)
Yb	O	2.29(1)	7.8(8)	0.007(1)	−0.3(9)

---

Each of the four M³⁺ ions is eight coordinate with O atoms. The average EXAFS-determined M³⁺–O bond lengths exhibit a direct and linear correlation (R² = 0.999) with the M³⁺ IR (CN = VIII) as shown as solid squares in Fig. 3(b). In this Figure, the least squares fit is shown as the solid line. The predicted M–O distances based upon the sum of the M³⁺ and O²⁻ IR are also shown as open squares connected with a dashed line. The predicted and the experimental distances are in close agreement for the three largest ions Eu, Am and Nd, and less so for Yb³⁺. Nevertheless, the correspondence is typical for electrostatic M–O interactions throughout the series of 4f and 5f ion complexes.^13,14 As shown in Fig. 3(b), without exception, all of the EXAFS-determined M³⁺–O distances for the DMDOHEMA complexes are 0.07–0.10 Å longer than the corresponding M³⁺–O distances for the HDHP complexes, shown in Fig. 3(a). This is consistent with the familiar, direct correlation between bond length and coordination number,¹² and manifest herein with O coordination numbers of six for the M³⁺–HDHP complexes and eight for the M³⁺–DMDOHEMA complexes.

It is apparent from the data of Table 3 that the O CNs for the Nd, Eu and Yb complexes have a best integer value of 8, whereas the CN of 7.5 for Am is consistent with either 7 or 8, reflecting the typical uncertainty with CN number determination by EXAFS.¹⁵ The general quandary here is how to account for the distribution of eight O interactions about M³⁺. That is, how many of these eight are from the DMDOHEMA ligand itself, and what about the presence of nitrate ions as well as water molecules in the inner coordination sphere of M? The elementary molecular motif suggested by the data at hand is that the M³⁺ ions are coordinated by four DMDOHEMA ligands serving in bidentate fashion as illustrated in Fig. 9. The charge on the tripositive molecular cluster would be balanced by nitrate ions distributed in the outer coordination sphere, which is not resolved in the FT data, Fig. 2(b). Precedence for this structure type and stoichiometry comes from single-crystal X-ray diffraction studies of the Sm and Er complexes Ln(TMMA)₄(PF₆)₃, where TMMA (N,N,N′,N′-tetramethylmalonamide).³³ The reported Sm–O₈ bond distance of 2.41 Å is in exact agreement with our Eu–O₈ distance of 2.40 Å, in view of the 0.013 Å smaller IR for Eu³⁺.¹² So, too, the mean Er–O₈ distance (2.31 Å) is in agreement with our Yb–O₈ distance (2.29 Å), in view of the 0.019 smaller IR for Yb³⁺.

Fig. 9 Possible structure model for M³⁺ complexes formed upon extraction with DMDOHEMA, consisting of a cationic, mononuclear [M(DMDOHEMA)₄]³⁺ species obtained from bidentate coordination of M³⁺ by four DMDOHEMA ligands. For each M³⁺, the O CN is eight.

Although no structures have been reported for M³⁺ complexes with DMDOHEMA, the structure and stoichiometry of Fig. 9 are different from a number of other single-crystal structures of La³⁺, Pr³⁺, Nd³⁺ and Yb³⁺ complexes with related malonamides, including TEMA,^10,34–36 TEEEMA,³⁷ TMMA,³⁸N,N′-dicyclohexyl-N,N′-dimethylmalonamide and N,N′-dimethyl-N,N′-diphenylmalonamide,³⁹ as well as butyl-N,N′-dimethyl-N,N′-diphenylmalonamide and ethoxy ethyl-N,N′-dimethyl-N,N′-diphenylmalonamide.⁴⁰ Comparisons of these crystallographic structures with our EXAFS results show differences in the nearest-neighbor O coordination numbers. In the five structures with Nd,^10,39,40 the CNs are ten (we report eight for DMDOHEMA), and in the three with Yb,^10,35,39,40 they are nine (we report eight for DMDOHEMA). The Nd complexes are neutral molecular species with bidentate O coordination of either one or two malonamide ligands at 2.43–2.49 Å, distances that are consistent with the EXAFS-determined value of 2.47(1) Å for Nd-DMDOHEMA. In the published structures, the cluster neutrality is realized through bidentate O coordination of three nitrate ions at 2.54–2.64 Å. In the lone case with water in the coordination sphere,³⁹ the average M–OH₂ distance, 2.56 Å, is indistinguishable from the bidentate, M–O₂NO, interactions. The Yb coordination environment in the three crystal structures^10,35,39 has bidentate O coordination of either one or two malonamide ligands at 2.23–2.30 Å, distances that are consistent with the EXAFS-determined value of 2.29(1) Å for Yb-DMDOHEMA. Except for the monocationic complex of [YbTEMA(NO₃)₂(H₂O)₃]⁺,³⁵ the other published structures are neutral complexes with three nitrate ions in either monodentate and/or bidentate O coordination with mean Yb–ONO₂ and Yb–O₂NO distances of 2.32 and 2.46 Å, respectively, and water molecules. The mean Yb–OH₂ distance, 2.35 Å, is essentially indistinguishable from the monodentate nitrate coordination.

With reference to ancillary information from complementary physical and spectroscopic measurements of the DMDOHEMA series of extractant species with M³⁺,^1,3 it is possible to exclude the hypothetical cationic molecular species proposed in Fig. 9, which was visualized from the one-dimensional EXAFS results alone. With just one exception,³⁸ precedent shows that diamide extractants function in bidentate mode and that, also with one exception,³⁵ the M–diamide complexes are neutral. Moreover, the extraction process³ data are consistent with the following stoichiometric reaction:

M³⁺ + 3 NO₃⁻ + 2 DMDOHEMA ⇌ [M(NO₃)₃(DMDOHEMA)₂]

This information about the 1:3 M:NO₃⁻ ratio and the 1:2 M:DMDOHEMA ratio, in combination with the total O CN = 8, facilitates the visualization of two neutral molecular entities with bidentate O coordination of DMDOHEMA. These include M(NO₃)₃(DMDOHEMA)₂ and M(NO₃)₃(DMDOHEMA)₂H₂O, which are illustrated in Fig. 10(a) and (b), respectively. In the structure of Fig. 10(a), two of the three nitrate ions have monodentate O coordination to M and the third one has bidentate O coordination to M. In the structure of Fig. 10(b), all three of the nitrate ions have monodentate O coordination to M. To determine which one of the two plausible structures is more likely to prevail, we turn to the results of Table 4 obtained from IR spectroscopy. In particular, the separation (Δν) between the asymmetric and symmetric frequencies of the O–N–O stretching vibrations can be used as a diagnostic tool for determining the mode of nitrate coordination to a metal ion.^41–43 A Δν value greater than 186 cm⁻¹ indicates bidentate coordination, whereas a value that is less than or equal to 115 cm⁻¹ indicates monodentate coordination. For example, in a study of third phase formation in the extraction of UO₂²⁺ and Th⁴⁺ by tri-n-butylphosphate, Δν was found to be 254 ± 4 cm⁻¹, indicating bidentate coordination of nitrate to both cations.⁴⁴

Table 4 Most important vibration bands (cm⁻¹) of 0.7 mol L⁻¹ DMDOHEMA in n-dodecane before and after contact with an aqueous solution containing 0.01 mol L⁻¹ Eu(NO₃)₃ and 2 mol L⁻¹ LiNO₃

Before contact	After contact	Assignment
m = medium intensity; w = weak intensity.
1654 (m)	1654 (m)	Free C=O, stretch
	1619 (m)	Bound C=O, stretch
	1485 (w)	O–N–O asym, stretch
1400 (w)	1404 (w)	C–N
	1300 (m)	O–N–O sym, stretch
1111 (m)	1111 (m)	C–N and/or C–O–C and/or CH3–N, stretch

---

Fig. 10 Structures proposed for M³⁺ complexes formed upon extraction with DMDOHEMA, consisting of neutral, mononuclear M³⁺ species with (a) two bidentate DMDOHEMA, one bidentate NO₃⁻ and two monodentate NO₃⁻; (b) two bidentate DMDOHEMA, three monodentate NO₃⁻ and one H₂O. In each of these structures, the O CN of M³⁺ is eight. Based upon independent results from extraction experiments¹ and IR spectroscopy, Table 4, the species illustrated in (a) is most likely to prevail, although both may coexist in solution.

As shown in Table 4, the value of Δν = 185 cm⁻¹ does not conclusively differentiate between the two possibilities because it is not large enough for pure bidentate nitrate coordination, on the one hand, and it is larger than that expected for complete monodentate coordination, on the other. Therefore, the structure of Fig. 10(a), with mixed, bidentate and monodentate nitrate ion coordination is more compatible with the IR information than is the structure shown in Fig. 10(b), with only monodentate nitrate ion coordination, although both may coexist in solution.

Binary-extractant systems

In view of the difficulties at hand in modeling the EXAFS data for the four M³⁺ single-extractant systems with DMDOHEMA alone, it is impractical to attempt any conventional EXAFS analyses of the data obtained for the binary-extractant systems of Table 1, specifically the 18, 19 and 20 sample series of DMDOHEMA and HDHP with Am³⁺. Instead, we put these three Am L₃-edge k³χ(k) EXAFS spectra together with those for the two single-extractants—namely, the 5Am (HDHP) and 8Am (DMDOHEMA) sample series of Table 1—and subjected the entire collection to principal component (factor) analysis (PCA) as described elsewhere.^45–47 The full range of data (to k_max = 14.3 Å⁻¹) for all five EXAFS spectra were first analyzed and then, to minimize the possibility of adverse effects of experimental noise at high k on the PCA, the EXAFS data truncated at k = 11 Å⁻¹ were re-analyzed for comparison. The results are the same in both treatments—the first two principal components account for all of the data—5Am, 8Am, 18Am, 19Am and 20Am. Moreover, the PCA indicates that the spectra for 5Am and 8Am contain the most pure versions of those components. The higher-order components are experimental noise and artifact, see the ESI†, with no structural significance. This result, from a purely mathematical treatment, is consistent with independent preparative, physical and spectroscopic evidence showing that the single extractant systems with HDHP (5Am) and DMDOHEMA (8Am) alone are the so-called end-member species.¹ Most important, the implication of the PCA result is that the EXAFS data for the binary-extractant systems—e.g., 18Am, 19Am, 20Am—are adequately described as admixtures of the EXAFS data for the two aforementioned single-extractant systems.

Linear regression analyses of M³⁺ EXAFS

With this pivotal information in hand, we have performed linear combination regression analyses on the primary, normalized spectra, I_f/I₀vs. energy (keV), of the binary extractant systems using the spectra for samples with HDHP (5M) and DMDOHEMA (8M) alone as the two limiting end members. The results are collected in Table 5, and the best fits are shown as ESI†. In Table 5, the samples are listed for increasing HNO₃ concentration in the aqueous phase and for increasing atomic number of the lanthanide ions.

Table 5 Fractional, relative concentrations of HDHP and DMDOHEMA complexes of M³⁺ as 5M/HDHP and 8M/DMDOHEMA, respectively, in the binary extractant systems of 18M, 19M, 20M (Table 1, for M = Nd, Am, Eu and Yb) as obtained using linear combination fits^a

		Relative concentrations
Series no./M	Aqueous [HNO3]/M	5M/HDHP	8M/DMDOHEMA
a Estimated standard deviations are shown in parentheses. M L3-edge data ranges used in fits: Am, 18.400–18.750 keV, 235 points; Nd, 6.221–6.401 keV, 134 points; Eu, 6.991–7.20 keV, 155 points; Yb, 8.958–9.180 keV, 176 points.
18Am	0.1	0.84(2)	0.16(2)
19Am	1.0	0.36(2)	0.64(2)
20Am	3.0	0.18(2)	0.82(2)
18Nd	0.1	0.65(3)	0.35(3)
20Nd	3.0	0.13(2)	0.87(2)
20Eu	3.0	0.12(2)	0.88(2)
19Yb	1.0	1.00(1)	0.00(1)
20Yb	3.0	0.33(1)	0.67(1)

---

The DMDOHEMA extractant is known to exhibit a higher affinity for Am³⁺ than for Ln³⁺.^48–50 Moreover, its extracting power increases with the third power of the aqueous HNO₃ concentration.¹ Conversely, the HDHP extractant exhibits a higher affinity for Ln³⁺ than Am³⁺ and for heavier than lighter lanthanides, and its extraction power declines with the third power of the aqueous acidity.¹ The results of the regression analysis reported in Table 5 fully agree with the behavior expected from the distribution studies.

The results for the three Am samples are consistent with an increasing relative concentration of the Am–DMDOHEMA complex (8Am) and decreasing contribution of the Am–HDHP complex (5Am) upon increasing the aqueous HNO₃ concentration from 0.1 to 3 M. The same trend, as expected, is shown by the results from the neodymium and ytterbium samples. The ytterbium case is particularly striking: the affinity of HDHP for Yb³⁺ is so high that only the HDHP complex (5Yb) is needed to fit the EXAFS data for the extractant mixture 19Yb from 1 M HNO₃. When the aqueous acidity is increased to 3 M, as for all the samples of the 20M series, however, the Yb–DMDOHEMA complex (8Yb) becomes predominant (67%), although the contribution of the Yb–HDHP complex remains significant (33%), and considerably higher than for the Am, Nd and Eu samples under the same conditions of high, 3 M, aqueous acidity.

Conclusions

Even for the single and purportedly simple dialkylphosphoric acid extractant, HDHP, there are at least two different M³⁺ coordination environments: one for Am and Eu, and a different one for Nd and Yb. The two stoichiometries and structure models of neutral molecular species, illustrated in Fig. 5 and 8, respectively, have the same O CN (six) and differ only in the number of distant P atoms (three and six). Although we do not fully understand why these ions are partitioned in this fashion, we note that the electronic configurations of Am³⁺ and Eu³⁺ are the same, f⁶. The metrical results from the analyses of the EXAFS data for the M³⁺ complexes with the diamide, DMDOHEMA, alone are surprisingly less complicated than those for the corresponding HDHP systems, due to the absence of distant structure features in the FT data. One stoichiometry and structure type prevails throughout the M³⁺ series. This is depicted in Fig. 10(a), which precisely accounts for the EXAFS response of eight O about M, and the IR results, which suggest a combination of monodentate and bidentate O coordination of the nitrate ions to M. The M³⁺ EXAFS data for the three binary extractant systems were shown to be linear combinations of the M³⁺ EXAFS for the single extractant complexes with HDHP and DMDOHEMA. The apparent synergy that is observed in liquid–liquid extraction is rationalized by the conformational flexibility afforded by the simultaneous presence of both chelants, which accommodate M³⁺ ions with O coordination environments of either six or eight.

Experimental

Materials and sample preparation

The DMDOHEMA was provided by the PANCHIM Chemical Company, France. All experiments were performed using a single lot having a >99% purity. The HDHP was synthesized at Marcoule at a purity level higher than 99%.⁵¹ Among the several published syntheses of alkylphosphoric acids, the method based on POCl₃ was followed.⁵² All other chemicals were of analytical grade and used without further purification. Equal volumes of the organic phase containing HDHP or DMDOHEMA or their mixtures were equilibrated in screw-cap tubes at 25 °C by vortexing for several minutes, a time sufficient for equilibrium attainment. After centrifugation and phase separation, the organic phases were subjected to the measurements described in this paper. The extractions were made under conditions where almost quantitative extraction of the metal took place. The isotope ²⁴³Am was obtained from Argonne National Laboratory stocks and purified before use as described elsewhere.⁵³ The sample compositions, including metal, extractant, and acid concentrations, are given in Table 1.

IR spectroscopy

Spectra of selected DMDOHEMA samples were obtained by using a Bruker Equinox 55 FT-IR spectrometer with a golden gate single reflection ATR (Attenuated Total Reflection) sampling system. The ATR crystal is a 2 × 2 mm² diamond with a 0.6 × 0.6 mm² active area. The head is fitted with a flat sapphire anvil for compressing the sample. Samples were prepared by extraction and analyzed in the ATR cell, 16 scans were summed in order to obtain good signal-to-noise ratio.

X-Ray absorption spectroscopy

The Ln³⁺ L₃-edge data were obtained for solutions contained in micro X-ray cells (SPEX 3577) with Kapton® film windows, 0.0003″ (SPEX 3511). The Am³⁺ L₃-edge data were obtained for solutions contained in purpose-built cells-constructed of polychlorotrifluoroethylene (PCTFE), aluminum window frames, Teflon® o-ring seals and PCTFE film windows, 0.010″ (available from Boyd Technologies, CA)⁵⁴—which were subsequently doubly-enclosed in a multi-sample containment tower. All measurements were made at the Advanced Photon Source beam line 12-BM–B⁵⁵ (Argonne National Laboratory) using a 13-element fluorescence detector (Canberra). The EXAFS was analyzed in consistent fashion in the usual manner with EXAFSPAK⁵⁶ and, also, by use of principal component analysis^46,47 and through linear regression analysis with WinXAS.⁵⁷ The conventional metrical analysis of the M k³χ(k) EXAFS was performed with a fixed scale factor (S₀² = 0.9) in a series of step-wise fits of increasing complexity and numbers of O and P coordination shells, using theoretical phase and amplitude functions calculated with FEFF8.0.⁵⁸ Ultimately, all of the EXAFS spectra were fit with either four parameters: M–O interatomic distance, r; O coordination number, CN; M–O Debye–Waller factor, σ²; energy threshold parameter, ΔE₀—for the one-shell (M–O) model, or seven parameters—the previously mentioned four plus the M–P interatomic distance r; P coordination number, CN; M–P Debye–Waller factor, σ²—for the two shell (M–O and M–P) model. Even for the shortest, Nd L₃-edge, EXAFS (k_max = 11 Å⁻¹) the number of refined parameters was less than the number of relevant independent data points, N_I = 8, available in the primary spectra.⁵⁹ For the longest, Am and Yb L₃-edge, EXAFS (k_max = 14.3 Å⁻¹) N_I = 10, and for the intermediate, Eu L₃-edge, EXAFS (k_max = 12.5 Å⁻¹) N_I = 9.

Acknowledgements

We thank the staff of the Actinide Facility and Dr M. P. Jensen (Argonne) for assistance and Dr S. R. Wasserman (Structural GenomiX, Inc.) for the PCA. This work is supported by the U.S. DOE, BES-Chemical Sciences and Materials Sciences under contract number W-31–109-ENG-38, for the part performed at Argonne National Laboratory; and by the CEA, DRCP/SCPS, for the part performed at Marcoule. This collaboration was realized in the framework of the CEA-DOE agreement (C5096).

References

1. B. Gannaz, R. Chiarizia, M. R. Antonio, C. Hill and G. Cote, Solvent Extr. Ion Exch., 2006, submitted.
2. X. Heres, C. Nicol, I. Bisel, P. Baron and L. Ramain, Global'99: \“Nuclear Technology-Bridging the Millennia\”, Proceedings of the International Conference on Future Nuclear Systems, Jackson Hole, WY, United States, Aug. 29–Sept. 3, 1999, 1999, pp. 585–591.
3. C. Madic, JAERI-Conf., 2002, 2002–004, 27–37.
4. G. Cote, Radiochim. Acta, 2003, 91, 639–643.
5. G. Ionova, R. Guillaumont, S. Ionov, C. Madic and M. J. Hudson, J. Alloys Compd., 1998, 275–277, 785–791.
6. H. Narita, T. Yaita and S. Tachimori, Solvent Extraction for the 21st Century, Proceedings of ISEC'99, Barcelona, Spain, July 11–16, 1999, 2001, vol. 1, pp. 693–696.
7. M. P. Jensen, R. Chiarizia and V. Urban, Solvent Extr. Ion Exch., 2001, 19, 865–884.
8. T. Guoxin, Z. Yongjun, X. Jingming, Z. Ping, H. Tiandou, X. Yaning and Z. Jing, Inorg. Chem., 2003, 42, 735–741.
9. M. P. Jensen and A. H. Bond, J. Am. Chem. Soc., 2002, 124, 9870–9877.
10. C. Den Auwer, M. C. Charbonnel, M. G. B. Drew, M. Grigoriev, M. J. Hudson, P. B. Iveson, C. Madic, M. Nierlich, M. T. Presson, R. Revel, M. L. Russell and P. Thuéry, Inorg. Chem., 2000, 39, 1487–1495.
11. The quoted distances are not corrected for phase shift, and are used here only for a descriptive discussion of the FT features in Fig. 2.
12. R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751–767.
13. M.-H. Chiang, M. R. Antonio, C. W. Williams and L. Soderholm, Dalton Trans., 2004, 801–806.
14. M.-H. Chiang, C. W. Williams, L. Soderholm and M. R. Antonio, Eur. J. Inorg. Chem., 2003, 2663–2669.
15. M. R. Antonio and L. Soderholm, in The Chemistry of the Actinide and Transactinide Elements, ed. L. R. Morss, N. M. Edelstein and J. Fuger, Springer, Dordrecht, 3rd edn, 2006, vol. 5, pp. 3086–3198.
16. S. Ishiguro, Y. Umebayashi and M. Komiya, Coord. Chem. Rev., 2002, 226, 103–111.
17. V. A. Cocalia, M. P. Jensen, J. D. Holbrey, S. K. Spear, D. C. Stepinski and R. D. Rogers, Dalton Trans., 2005, 1966–1971.
18. V. G. Lebedev, K. K. Palkina, S. I. Maksimova, E. N. Lebedeva and O. V. Galaktionova, Russ. J. Inorg. Chem. (Engl. Transl.), 1982, 27, 2980–2982 (Refcode BOVREJ in Cambridge Structural Database).
19. G. Zeng, X. Guo, C. Wang, Y. Lin and H. Li, Jiegou Huaxue, 1994, 13, 24–27 (Refcode HEBDEX in Cambridge Structural Database).
20. L. Li, Y. Lin, G. Zeng and A. Ma, Yingyong Huaxue, 1989, 6, 53–57 (Refcode JEVVOV in Cambridge Structural Database).
21. L. Li, Y. Lin, G. Zeng and A. Ma, Jiegou Huaxue, 1991, 10, 155–158 (Refcode KIXGON in Cambridge Structural Database).
22. Y. I. Trifonov, A. G. Tutov, E. K. Legin and D. N. Suglobov, Sov. Radiochem., 1991, 33, 260–266.
23. V. I. Tsaryuk, V. F. Zolin, E. N. Lebedeva, O. V. Galaktionova and V. G. Lebedev, Sov. J. Coord. Chem. (Engl. Transl.), 1989, 15, 236–242.
24. C. Liu, G. Wu, Z. Tang, Y. Han, Z. Pan, N. Shi and L. Liao, Huaxue Xuebao, 1990, 48, 116–120 (Refcode JIGVEA in Cambridge Structural Database).
25. T. Harada, M. Smutz and R. G. Bautista, J. Chem. Eng. Data, 1972, 17, 203–204.
26. B. Gannaz, R. Chiarizia, M. R. Antonio, C. Hill and G. Cote, Solvent Extr. Ion Exch., 2006, to be submitted.
27. Y. X. Ni, J. M. Hughes and A. N. Mariano, Am. Mineral., 1995, 80, 21–26.
28. D. F. Mullica, D. A. Grossie and L. A. Boatner, J. Solid State Chem., 1985, 58, 71–77.
29. D. F. Mullica, D. A. Grossie and L. A. Boatner, Inorg. Chim. Acta, 1985, 109, 105–110 (f-Block Elem. Art. Lett).
30. J. A. Danis, H. T. Hawkins, B. L. Scott, W. H. Runde, B. E. Scheetz and B. W. Eichhorn, Polyhedron, 2000, 19, 1551–1557.
31. M. B. Doran, C. L. Stuart, A. J. Norquist and D. O'Hare, Chem. Mater., 2004, 16, 565–566.
32. A. J. Locock and P. C. Burns, J. Solid State Chem., 2004, 177, 2675–2684.
33. E. E. Castellano and R. W. Becker, Acta Crystallogr., Sect. B, 1981, 37, 61–67.
34. P. Byers, M. G. B. Drew, M. J. Hudson, N. S. Isaacs and C. Madic, Polyhedron, 1994, 13, 349–352.
35. M. G. B. Drew, M. J. Hudson, P. B. Iveson and C. Madic, J. Chem. Crystallogr., 2000, 30, 455–458.
36. P. Thuéry, M. Nierlich, M. C. Charbonnel, C. Den Auwer and J. P. Dognon, Polyhedron, 1999, 18, 3599–3603.
37. P. Thuéry, M. Nierlich, M. C. Charbonnel and J. P. Dognon, Acta Crystallogr., Sect. C, 1999, 55, 1434–1436.
38. E. E. Castellano and R. W. Becker, Acta Crystallogr., Sect. B, 1981, 37, 1998–2001.
39. G. Y. S. Chan, M. G. B. Drew, M. J. Hudson, P. B. Iveson, J.-O. Liljenzin, M. Skaalberg, L. Spjuth and C. Madic, J. Chem. Soc., Dalton Trans., 1997, 649–660.
40. P. B. Iveson, M. G. B. Drew, M. J. Hudson and C. Madic, J. Chem. Soc., Dalton Trans., 1999, 3605–3610.
41. J. Chatt, L. A. Duncanson, B. M. Gatehouse, J. Lewis, R. S. Nyholm, M. L. Tobe, P. F. Todd and L. M. Venanzi, J. Chem. Soc., 1959, 4073–4080.
42. N. F. Curtis and Y. M. Curtis, Inorg. Chem., 1965, 4, 804–809.
43. B. M. Gatehouse, S. E. Livingstone and R. S. Nyholm, J. Inorg. Nucl. Chem., 1958, 8, 75–78.
44. M. Borkowski, J. R. Ferraro, R. Chiarizia and D. R. McAlister, Solvent Extr. Ion Exch., 2002, 20, 313–330.
45. L. Soderholm, M. R. Antonio, C. Williams and S. R. Wasserman, Anal. Chem., 1999, 71, 4622–4628.
46. S. R. Wasserman, J. Phys. IV, 1997, 7, 203–205.
47. S. R. Wasserman, P. G. Allen, D. K. Shuh, J. J. Bucher and N. M. Edelstein, J. Synchrotron Radiat., 1999, 6, 284–286.
48. M. C. Charbonnel, J. L. Flandin, S. Giroux, M. T. Presson, C. Madic and J. P. Morel, International Solvent Extraction Conference, Cape Town, South Africa, Mar. 17–21, 2002, 2002, pp. 1154–1160.
49. B. F. Myasoedov, T. A. Maryutina, M. N. Litvina, D. A. Malikov, Y. M. Kulyako, B. Y. Spivakov, C. Hill, J. M. Adnet, M. Lecomte and C. Madic, Radiochim. Acta, 2005, 93, 9–15.
50. K. Van Hecke and G. Modolo, J. Radioanal. Nucl. Chem., 2004, 261, 269–275.
51. The HDHP was synthesized by Sebastien Garcia-Argote (CEA-Valrhô, DEN/VRH/DRCP/SCPS/LCAM).
52. G. W. Mason, A. F. Bollmeier and D. F. Peppard, J. Inorg. Nucl. Chem., 1970, 32, 1011–1022.
53. M. Miguirditchian, D. Guillaneux, D. Guillaumont, P. Moisy, C. Madic, M. P. Jensen and K. L. Nash, Inorg. Chem., 2005, 44, 1404–1412.
54. See http://www.boydtechnologies.com/.
55. M. A. Beno, M. Engbretson, G. Jennings, G. S. Knapp, J. Linton, C. Kurtz, U. Rutt and P. A. Montano, Nucl. Instrum. Methods Phys. Res. A, 2001, 467–468, 699–702.
56. For documentation and availability, see http://www-ssrl.slac.stanford.edu/exafspak.html.
57. T. Ressler, J. Synchrotron Radiat., 1998, 5, 118–122.
58. J. J. Rehr and R. C. Albers, Rev. Mod. Phys., 2000, 72, 621–654.
59. E. A. Stern, Phys. Rev. B, 1993, 48, 9825–9827.

---

Footnote

† Electronic supplementary information (ESI) available: Results from the M³⁺ L₃-edge k³χ(k) EXAFS curve fitting, PCA of Am k³χ(k) EXAFS and linear regression analyses of M³⁺ X-ray absorption data for M = Nd, Am, Eu, Yb. See DOI: 10.1039/b609492a

---