Exploring the Subtle Effect of Aliphatic Ring Size on Minor Actinide‐Extraction Properties and Metal Ion Speciation in Bis‐1,2,4‐Triazine Ligands

Abstract The synthesis and evaluation of three novel bis‐1,2,4‐triazine ligands containing five‐membered aliphatic rings are reported. Compared to the more hydrophobic ligands 1–3 containing six‐membered aliphatic rings, the distribution ratios for relevant f‐block metal ions were approximately one order of magnitude lower in each case. Ligand 10 showed an efficient, selective and rapid separation of AmIII and CmIII from nitric acid. The speciation of the ligands with trivalent f‐block metal ions was probed using NMR titrations and competition experiments, time‐resolved laser fluorescence spectroscopy and X‐ray crystallography. While the tetradentate ligands 8 and 10 formed LnIII complexes of the same stoichiometry as their more hydrophobic analogues 2 and 3, significant differences in speciation were observed between the two classes of ligand, with a lower percentage of the extracted 1:2 complexes being formed for ligands 8 and 10. The structures of the solid state 1:1 and 1:2 complexes formed by 8 and 10 with YIII, LuIII and PrIII are very similar to those formed by 2 and 3 with LnIII. Ligand 10 forms CmIII and EuIII 1:2 complexes that are thermodynamically less stable than those formed by ligand 3, suggesting that less hydrophobic ligands form less stable AnIII complexes. Thus, it has been shown for the first time how tuning the cyclic aliphatic part of these ligands leads to subtle changes in their metal ion speciation, complex stability and metal extraction affinity.


Introduction
Nucleare nergy offers ac lean, low carbon source of electricity that is becoming ag rowing part of the energy mix in many countries worldwide. However,t he spent fuel that is produced in nuclearf ission reactors is long-lived and highly radiotoxic. [1] Following reprocessing to remove uranium and plutonium, the minor actinides americium,c urium and neptunium are responsible for much of the long-term heat load and radiotoxicity of the remaining spent fuel material. Removing these elements before disposal would contribute to sustainable nuclear energy by significantly reducing the size of the final waste re-pository, and the time neededf or the remaining materialt o decay to the radiotoxicityl evel of natural uranium (from ca. 10 4 years to af ew hundred years). [2] Beyondt he currently used PUREXp rocess that recovers and recycles most of the uranium and plutonium, [3] future reprocessings cenarios seek to close the nuclear fuelcycle by separating (partitioning) the minor actinidesf rom the chemically similara nd less-radiotoxic lanthanides,p rior to their burning (transmutation)i nh igh neutron flux advanced fast reactorsori na ccelerator-driven systems. [4] Numerous soft N-and S-donor ligands have been evaluated to accomplish the challenging separation of the minor actinides from the lanthanides in af uture solvent extraction process. [5,6] The greater orbital overlap between the more radially extended 5f orbitals of the actinides and ligand lone pairs is thought to be the basis for this separation. [7] AmongN -donor ligands, bis-1,2,4-triazinel igands 1-3 ( Figure 1) fulfil most of the challenging criteria to date for use in such ap rocess. In particular, bis-triazinyl-phenanthroline ligands such as 3 [8] and its derivatives have been extensively investigated. [9] Recent research has focusedm ostly on the effects that substituents attached to the aromatic rings of 2 [10] and 3 [11] have on their extraction properties. However,t here has been less emphasis on modifying the aliphatic rings appended to the triazine rings of ligands 1-3. [12] We wished to determinew hat effect changing the aliphatic ring size would have on the actinide extraction properties and metal speciationo ft hese ligands. In this paper, we report our studies on novel bis-1,2,4-triazine ligands con- taining af ive-membered aliphatic ring appended to the outer triazine rings insteado fasix-membered ring (as in [1][2][3], and we show that this small but subtle modification to the ligand structurec an have unexpected effects on the extraction properties and metal ion speciation of these ligands.
Preliminary solvente xtraction experiments were then carried out to determine the ability of the ligands 6, 8 and 10 to extract An III and separate them from Ln III .T he distribution ratios (D)f or Am III and Eu III ,a nd the separation factors (SF Am/Eu )f or the extraction of Am III and Eu III from nitric acid by solutionso f terdentate ligand 6 in 1-octanola re presented in the Supporting Information (section 4.1). The distribution ratios for Am III increasedw ith increasing nitric acid concentration to am aximum D Am value of 1.57 at 3.1 m HNO 3 ,w hich corresponds to 61 %A m III extraction. Although these D values are rather low, they would be sufficient for use in am ulti-step, counter-current An III extractionp rocess depending on the conditions (numbero fs tages, flow rates, etc). The average separation factor for Am III over Eu III was approximately 10 between 0.1 m and 1 m HNO 3 and reached am aximum value at 3.1 m HNO 3 . The distribution ratios for Cm III were very similar,and no significant selectivity for Am III over Cm III was observed for 6 (see Supporting Information section 4.1). The maximum D Am value observed for 6 is slightly less than that reported previously for ligand 1 in 1-octanol (D Am = 3.9, 0.5 m HNO 3 ,c ontact time = 60 minutes). [16] This is probably because ligand 6 is slightly less hydrophobic than ligand 1,a nd thus forms less hydrophobic complexes.
Results for the extraction of Am III and Eu III by tetradentate ligand 8 at differentn itric acid concentrationsa re presentedi n the Supporting Information (section 4.2). Extraction of Am III and Eu III by 8 showed as imilar trend to that of ligand 2,w ith the D values for both metals increasing as [HNO 3 ]i ncreases. With ligand 8,amore efficient and selective extraction of Am III was observed at high nitric acid concentrationst han with ligand 6.The selectivity of 8 for Am III over Eu III was significantly higher than that of ligand 6,a nd the average separation factor was approx.1 00 between 0.1 m and 3 m HNO 3 .O ncea gain,n o significant selectivity for Am III over Cm III was observedw ith 8 (see Supporting Information section4.2). Interestingly,t he D values for Am III and Eu III for 8 werea pproximately an ordero f magnitude lower than those previously reported fort he more hydrophobic ligand 2 under similarc onditions. [17] The results cannotb ed irectly compared however,a sa na dditional co-extractant; N,N'-dimethyl-N,N'-dioctyl-2-hexyloxyethyl malonamide 11,w as used in the case of 2.T oa llow ad irectc omparison with 2,wecarriedout extraction experiments for 8 in 1-octanol in the presenceo f0 .25 m 11 (see Supporting Information section 4.2). This led to as lighti ncrease in the D values for Am III at ! 1 m HNO 3 ,b ut am arked decrease in the selectivity for Am III over Eu III (SF Am/Eu 57 at 1-4 m HNO 3 )c ompared to the results in the absence of 11.T his is due to the competing non-selectiveco-extraction of Am III and Eu III by 11,w hich lowers the separation factor.H owever,t he distribution ratios for Am III and Eu III werestill significantly lower with 8 than with 2. [17] Results for the extraction of Am III and Eu III by tetradentate ligand 10 at different nitric acid concentrations are presented in Figure 2. Ah ighly efficient and selective extraction of Am III over Eu III was observed across ar ange of nitric acid concentrations. The D values for Am III reached am aximum value of 112 at 1 m HNO 3 .Amaximum selectivity for Am III over Eu III wasa lso observeda t1 m HNO 3 (SF Am/Eu = 237). The selectivity for Am III over Eu III shown by ligand 10 was similar to that shown by the analogous,m ore hydrophobic ligand 3.
[8] However,t he D values for both Am III and Eu III were approximately an order of magnitude lower with ligand 10 at high acidity than with ligand 3 (D Am % 1000, D Eu % 5f or 3 at ! 1 m HNO 3 ; D Am % 100, D Eu % 0.5 for 10 at ! 1 m HNO 3 ). This could allow for easier back-extraction (stripping) of the metals from the loaded organic phase after the extraction stagesh ave been carried out. Ligand 10 did not show any significant selectivity forA m III over Cm III (SF Am/Cm 2.2, see SupportingI nformation section 4.3), in contrasttol igand 3. [18] The extractiono fA m III and Eu III by ligand 10 as af unction of contactt ime is presented in Figure 3. As shown, Am III extraction equilibrium was reachedw ithin 10 minutes of phase mixing,w hile Eu III extraction equilibrium was reached after a mixing time of 20 minutes.T hus, the rates of metal extraction were slightly faster for the less hydrophobic ligand 10 than for its more hydrophobic analogue 3 under the same conditions (15 minutes for D Am and ! 60 minutes for D Eu to reach equilibrium for 3). [8] Spent nuclear fuels olutionsc ontain large amountso fY III and light trivalent lanthanides, which must be separated from the minor actinides. We therefore measured the distribution ratios for Y III anda ll lanthanides (except Pm III )a sw ell as the trivalent actinides. For all three ligands 6, 8 and 10,t he D values for the lanthanides showed an extraction profile across the lanthanides eries of first increasing, then decreasing D values, in agreement with previousr esults for ligands 2 and 3. [8,17] In the extractionsf rom 3.1 m HNO 3 ,H o III exhibited the highest D values for ligands 8 and 10 (D Ho = 0.37 for 8, D Ho = 3.56 for 10), while Dy III exhibited the highest D value for ligand 6 (D Dy = 0.08). Thus ap ractical separation of Am III and Cm III from all the lanthanides could be feasible with ligands 6 and 8 (D Ln < 1). Althought he later lanthanides Tb III -Yb III are somewhat extracted by ligand 10 at high nitric acid concentrations, ah ighly selective separation of Am III and Cm III from all the lanthanides is feasible since selectivel anthanide back-extraction can be carried out at lower nitric acid concentrations (D Am and D Cm > 1, D Ln < 1 at 0.1 m HNO 3 ). Furthermore, the later lanthanides are not present in spent fuel solutions, so their extraction is less relevant than that of the early lanthanides.

NMR titrations and X-ray crystallography
To gain furtheri nsighti nto the solution speciation of these ligands with metal ions and to rationalise the extraction results, we carriedo ut some 1 HNMR titrationso ft he ligandsw ith Y III and the diamagnetic lanthanides La III and Lu III .W eh ave previously employed this methodt oi nvestigate the solution speciation of the analogousligands 2 and 3,and relatedtetradentate ligandsw ith trivalentl anthanides. [10a, 19, 20] We used deuterated acetonitrile due to the high cost of deuterated 1-octanola nd to compare with previousresults for 2 and 3. [19] For tetradentatel igand 8,b oth 1:1a nd 1:2 M:L species were observed duringt he 1 HNMR titration with Y(NO 3 ) 3 in deuterated acetonitrile. As ingles pecies waso bservedi nitially during the titration, and the disappearance of the free ligand resonances at am etal:ligand ratio of 0.5 indicates this was the 1:2s pecies [Y(8) 2 3 ]w ere observed at higher metal:ligand ratios, reaching am aximum of 35, 30 and 18 % for La III ,L u III and Y III ,r espectively.A lthough these resultsa re broadly in agreement with those reported previously for tetradentate ligands 2 and 3, [19] an otable differencei so bserved in the relative ratios of the 1:1a nd 1:2c omplexes formed in each case. These differences are summarized in Ta ble 1. For the present ligands 8 and 10,t he percentage of the 1:1c omplex [M(L)(NO 3 ) 3 ]f or ag iven metal ion is significantly higher than that observed for ligands 2 and 3 (36 %f or 10 versus 27 %f or 3 with La III ,1 8% for 10 versus 5% for 3 with Y III ). Since it is known that the extracted species is the more hydrophobic 1:2 complex[ M(L) 2 (NO 3 )] 2 + (L = ligand), this could suggestt hat the lower distribution ratios observed above for ligands 8 and 10 in comparison to the analogousl igands 2 and 3 could be due to the lower percentage of 1:2c omplexes being formed by these ligandsunder extraction conditions.
During the 1 HNMR titration of terdentate ligand 6 with Y(NO 3 ) 3 in deuterated acetonitrile,asingle complex species was observed initially (see Supporting Information section 5.1). The complete disappearance of the free ligand resonances at a metal:ligand ratio of between 0.3 and 0.4 suggests that this is the expected1 :3 M:L complex [Y(6) 3 ] 3 + .T hese 1:3c omplexes are the major solution species formed by terdentate bis-1,2,4triazine ligands with trivalent lanthanides. [16,21] Further evidence for the formationo ft his chiral racemic 1:3c omplex, which exists as ap air of L and D enantiomers, wast he appearance of four 6-proton singletsi nt he aliphatic region corresponding to the four sets of diastereotopic methyl groups. Minor traces ( 10 %) of as econd species were also observed on continued addition of metal. This was tentatively assigned as the 1:2s pecies, formed by partial dissociation of the 1:3s pecies.
As eries of 1 HNMR competition experiments were then carried out to determine if phenanthroline-derived ligand 10 formed thermodynamically more stable complexes with the lanthanides than bipyridine-derivedl igand 8,a si mplied by the higherd istribution ratios observed in the extraction experimentsf or 10.T he aliphatic region of the 1 HNMR spectrum of a1 :1:1 mixture of 8, 10 and La(NO 3 ) 3 in deuterated acetonitrile is presented in Figure 5. The spectrum displays resonances for the 1:2b is-complexo f8,t he 1:2b is-complex of 10,a nd an additional set of resonances (four methyl resonances,t wo methylene resonances) which werea ssigned to the heteroleptic 1:2 bis-complex [La(8)(10)(NO 3 )] 2 + .T he heteroleptic complex showedo ne singlet, one triplet,t wo doubletsa nd am ultiplet in the aromatic region (see Supporting Information section 5.4). These resonancesw eren ot previously observed in the 1 HNMR titration of 10 with La(NO 3 ) 3 .T he ratio of bis-8 complex/bis-10 complex/heteroleptic bis-complex was 1:1:2, indicating that as tatistical mixture of the three 1:2b is-complexes had been formed, in agreement with previous work on ligands 2 and 3. [19] NMR experiments were then carried out to determine if each of the ligands 8 and 10 was ablet od isplace the other    3 (prepared by addition of 1equivalent of 8 to 0.5equivalent of La III )g ave rise to am ixture of the 1:2b iscomplexo f10,t he heteroleptic 1:2b is-complexa nd free uncomplexed 8 (see Supporting Information section 5.4). The major species present was the 1:2b is-complex of 10,a nd no traces of the 1:2b is-complexo f8 were observed. Thus ligand 10 displaces ligand 8 from its La III complex and forms the thermodynamically more stable complex with La III than 8. When as olution of ligand 8 (1 equivalent) wasa dded to a solution of the 1:2b is-complexo fl igand 10 with La(NO 3 ) 3 (prepared by adding 1equivalent of 10 to 0.5 equivalents of La III ), am ixture almost identical in compositiont ot hat observed above was obtained( see SupportingI nformation section 5.4). The 1:2b is-complex of 10 was again the major speciesformed, and no traces of either the 1:2b is-complex of 8,o rf ree uncomplexed 10 were observed. Thus ligand 8 is at best able to displace one of ligand 10 from its 1:2b is-complexes but is never able to displace both. We have previously observed the same phenomenon with ligands 2 and 3. [19] These results suggest that the order of thermodynamic stabilityo ft he three 1:2 bis-complexes is: Similarly,a1:1:1m ixture of ligands 8, 10 and Y(NO 3 ) 3 in deuterated acetonitrile led again to the expected statistical mixture of the three 1:2b is complexes (1:2 bis-complex of 8,1 :2 bis-complex of 10,h eteroleptic 1:2b is-complex) in ar atio of 1:1:2( see Supporting Information section 5.4). However,i n contrast to La III ,o nly partial ligand displacement reactions were observed when either 8 or 10 was added to as olution of the Y III bis-complex of the other ligand. Addition of 10 to the 1:2b is-complexo f8 with Y III led to am ixture containing mostly the bis-complexo f8 and free uncomplexedl igand 10, as well as traces of the heteroleptic 1:2b is-complex [Y(8)(10)(NO 3 )] 2 + .A ddition of 8 to the 1:2b is-complex of 10 with Y III led to am ixture of primarily the bis-complex of 10 and uncomplexed 8,a sw ell as traces of the heteroleptic 1:2b iscomplex.T he partial ligand displacementr eactions observed here for Y III are likely due to its higherk inetic inertness towards ligand substitution compared to La III ,i na greement with the lower ligand exchange rate constant observed for the Y III aqua complex. [22] To further characterise the various species produced upon complexation, single crystal X-ray crystallography experiments were performed. Perhapss urprisingly,t he crystalsg rown from solutionso f10 with Y(NO 3 ) 3  Both structuresc rystallised as acetonitriled isolvates and were essentiallyi somorphous, sharing similar unit cell parameters, space groups and packing. The two structures are distinguishable, however,b yt he coordination of the three nitrate ligands aboutt he lanthanide ion. In the structure of [Y(10)(NO 3 ) 3 ]( see Supporting Information section 3.2), all three nitrate ligandsa re bidentate with LnÀOd istances in the 2.47-2.56 range for those in axial positionsa nd around2 .37 for the nitrate ligand in the equatorial positionw hich lies in roughlyt he same plane as the ligand.T he structure is very similar to that of the 10-coordinate[ Y(3)(NO 3 ) 3 ]c omplex published previously. [19] In contrast, only two of the nitrate ligands in [Lu(10)(NO 3 ) 3 ]a re bidentate, one in an axial and the other in an equatorial position( Figure 6). These ligands exhibit significantly shorter LnÀOb ond lengths compared to those of [Y(10)(NO 3 ) 3 ]l ying in the ca. 2.41-2.44 range for the axial positionsa nd around 2.33 for the equatorial position. The third nitratel igand is monodentate with ab ond distance of 2.288(2) ,g iving an ine-coordinate complexo verall. This phenomenonh as been observed previously in complexes of Lu III with tetradentateb is-1,2,4-triazine ligands, where aw ater molecule displaced one of the nitratel igands to the outer coordination sphere. [23] This discrepancy in coordination geometry can be rationalised by considering the ionic radii of the lanthanides in question. With an ionic radius of 86 pm, Lu III is slightly smallert han Y III ,w hich has ar adius of 90 pm, but this subtle difference may be all that prevents the coordination sphere of Lu III from accommodating at hirdb identaten itrate ligand.F urther evidence of this constraint on the coordination geometry is observed in the twisting of the triazine rings of 10,w hich is more pronounced in the structureo f[ Lu (10) gives rise to an angle of 161.02(7)8.T his very slight deviation is enough to reduce the space availablet ot he monodentate nitrate and preventi tb inding in ab identate fashion whilea lso providing greatera ccess to the metal ion to the nitrate trans to it allowing it to bind more strongly and with shorter contacts than the axial nitrate ligandsin[ Y (10) Attempts to obtain the structures of any of the 1:2b is-complex specieso bserved in the course of the NMRt itrations provedu nsuccessful. However,g ood quality single crystalso fa  (Figure 7). The asymmetric unit of the structure comprises two crystallographically independent molecules (Z' = 2), one of each of the D and L opticali somers. In terms of their coordination, the structure is very similar to those of 2 with Eu III [24] and 3 with Pu III . [25] As this is the case it is probably safe to assume that the coordination of the ligands about the lanthanide is similar across the series and that any effect of the lanthanide contraction will be manifesti nt he coordination of the nitrate as was observed in the structures of the 1:1s pecies.
The X-ray crystal structure of free ligand 8 (see Supporting Information section 3.2) shows that the ligand adopts the nonchelating trans conformationi nt he solid state with respect to the CÀCt orsion between the central pyridine rings.T his was also observedi nt he structure of the analogous ligand 2, [23] and is due to the high torsional barrier to rotation about this CÀCb ond when the ligand adopts the chelating cis conformer. [19]

TRLFS measurementsand DFT calculations
To gain furtheri nsighto nt he speciation in solution and support the NMR and X-ray crystallography findings, the complexation of Cm III and Eu III with ligand 10 was studied by time-resolved laser fluorescences pectroscopy.T his technique allows the study of the coordination chemistry of fluorescent metal ions. [26,27] Cm III and Eu III represent trivalenta ctinides and lanthanides, respectively with excellent fluorescencep roperties.

Complexation kinetics
Te tradentate bis-1,2,4-triazine derivatives such as 3 show relatively slow complexation kinetics. [28] Therefore, the fluorescence emissiono fC m III at ag iven ligand concentration was measureda safunction of time after addition of 10.C m III fluorescence spectra resulting from the 6 D 7/2 ! 8 S 7/2 transition are showni nF igure 8. Without addition of 10 the Cm III solvent spectrum at 599.1 nm was observed with as houlder at 595.4 nm. Upon addition of 10 the emission band at 599.1 nm decreased and new emission bands at 606.4 nm and 618.7 nm occurred. With time the emission band at 618.7 nm became dominant. No further changes of the Cm III fluorescences pectrum were observed after 23 h, indicating that the system was at equilibrium.
In the case of Eu III ,t he fluorescencee mission bands of the 5 D 0 ! 7 F n (n = 0, 1,2) transitions weres tudied as af unction of time after addition of 10 (see Supporting Information section 6.1). Without 10 an emission band at 578.9 nm for the 5 D 0 ! 7 F 0 transition of the Eu III solvents pecies was observed. Upon addition of 10 two new emissionb ands at 579.5 nm and 581.1 nm occurred. The 5 D 0 ! 7 F 1 and 5 D 0 ! 7 F 2 transitions exhibitedachange of shape and splitting of the emission bands due to complexation of Eu III with 10.N of urtherc hanges of the Eu III fluorescence emission spectra were observed after 4.5 h, confirming chemical equilibrium.

Complexation of Cm III and Eu III with ligand 10
To determine thermodynamic data for the complexation of Cm III and Eu III with 10,t he evolution of the fluorescences pectra of Cm III and Eu III as af unctiono ft he concentration of 10 was studied in nitrate free media. Batch samples containing increasingc oncentrationso f10 were equilibrated for 24 h before being measured. Then ormalized Cm III fluorescence spectra are shown in Figure 9. The formation of two speciesa t 606.4 nm and 618.7 nm was observed. Single component spectra for the Cm III solvents peciesa nd both complex speciesa re shown in the Supporting Information (section 6.2). The fluorescencei ntensity factor (FI) describes the fluorescence intensity of as pecies relative to ar eference. Due to the  high FI of species 2(FI = 56 AE 6), the speciation was determined from the overall fluorescencei ntensity. The speciation is shown in Figure 10. The formation of species 2s tarts at 8 10 À9 m of free 10 and becomes dominanta t2 .8 10 À7 m.T he relative ratio of species 1 is irrelevant (< 3%)u nder the applied conditions and is therefore not shown in Figure 10. Slope analysis according to Equation (1) was performed to determine the stoichiometry of species2.
The slope of the linear correlation indicates the number of coordinated ligand molecules. As lope of 2.11 AE 0.17 was ob-tained, showing that species 2i st he 1:2c omplex [Cm(10) 2 ] 3 + (see Supporting Information section6.2). The conditional stability constantf or the formation of the 1:2c omplex according to Equation (2) is log b' 2 = 13.1 AE 0.2. Fluorescences pectra of the Eu III 5 D 0 ! 7 F 0 transition are shown in the Supporting Information (section 6.2). Since neither the excited state ( 5 D 0 )n or the ground state ( 7 F 0 )a re split (J = 0), the number of emission bands accounts for the number of speciespresent in the system.
The Eu III solvent spectrum was observed at 578.9 nm. Upon addition of 10 two new emission bands at 579.5 and 581.1 nm occurred, indicating the formation of two different species. Eu III speciation (see SupportingI nformation section 6.2) was determined from the overall fluorescencei ntensity due to the high FI factor of species 2( FI 2 = 1325 AE 130). Again, species1is only present at irrelevant concentrations and is not shown in the speciation. Slope analysis resulted in as lope of 2.05 AE 0.06, confirming the formation of the 1:2c omplex[ Eu(10) 2 ] 3 + .T he conditional stability constant for this complex is log b' 2 = 10.3 AE 0.4.
Comparing both tetradentate phenanthroline-derived ligands 10 and 3 under the same conditions, it is evident that 3 is as tronger ligand than 10.T he stability constantsf or both the Cm III and the Eu III 1:2c omplexes are approximately one order of magnitude lower in the case of 10 ( Table 2).

Comparison of mono-andb iphasic experiments
Te tradentate bis-1,2,4-triazine ligands extract trivalenta ctinide and lanthanide ions from nitric acid or nitrate solutions as 1:2 complexes. [10b, 17, 29] With 2 and 3,t he extracted complexes were previously shown to be [ML 2 (NO 3 )] 2 + complexes containing one inner-sphere nitrate anion (L = 2 or 3). [28] The possible presence of an inner sphere nitratei nt he 1:2c omplexes with 10 wass tudied in as imilar manner by extracting Cm III or Eu III from solutionsc ontaining 0.1 m nitric acid and 1.9 m NH 4 NO 3 into solutions of 10 mm 10 in 1-octanol. After phase separation, the organic phases were studied by time-resolved laser fluorescencespectroscopy. Figure 11 compares the Cm III (top) and Eu III (bottom) spectra of the 1:2c omplexes of 10 in methanol with those from the solvente xtraction experiments.T he emission spectrum of the extracted Cm III complex shows an emission band at 620.1 nm, which is bathochromically shiftedb y1 .4 nm with respectt o the emission band of the [Cm(10) 2 ] 3 + complex (618.7 nm). In   the case of Eu III ,t he emission band of the 5 D 0 ! 7 F 1 and 5 D 0 ! 7 F 2 transitions of the [Eu(10) 2 ] 3 + complex and the complex in the organic phaseo ft he extraction experiment differ in shape and position. The emission band ( 5 D 0 ! 7 F 2 transition) of the complexf ormed duringt he extraction experiment displays a peak maximum at 613.3 nm while the emission band of the [Eu(10) 2 ] 3 + complexe xhibits ap eak maximum at 615.7 nm. Similar shifts and changes in position and shapeo ft he emission bands were observed for 2 and 3 and were assigned to the additional complexation of an itrate ion in the inner coordination sphere of Cm III . [28] Adding ammonium nitrate to as olution containing the [Cm(10) 2 ] 3 + complex resulted in ac omparable bathochromic shift of the emission band from initially 618.7 nm (no nitrate added) to 619.2 nm (9.97 10 À2 m of nitrate added), as shown in the Supporting Information (section 6.3).
In the case of Eu III ,achange in shape of the emission band resultingf rom the 5 D 0 ! 7 F 2 transition was observed (see Supporting Information section 6.3). With increasing nitrate concentration,anew peak at 613.3 nm appeared, indicating the formationo ft he same specieso bserved in the extraction ex-periments. Thus ligand 10 extracts Cm III and Eu III from acidic nitrate solutionsa s[ M(10) 2 (NO 3 )] 2 + complexes, as was previously observed for 2 and 3. [28] In an attemptt og ain further insight into why An III complexes of ligand 10 are less stable than those of ligand 3,D FT calculations were carried out on free ligands 3 and 10 and their respective [AmL 2 (NO 3 )] 2 + complexes (L = 3 or 10)u sing a level of theory successfully used in previous studies of actinide complexes. [30] Comparison of the energies of the complexes relative to the respective free ligand conformationso fl owest energy enabled the relative complexation energies of the ligands to be determined. The resultsi ndicate that the binding energy of 3 when forming [Am(3) 2 (NO 3 )] 2 + was 1.76 kJ mol À1 less favorable than the binding of 10 when forming [Am(10) 2 (NO 3 )] 2 + in an acetonitrile solventf ield (see Supporting Information section 7). Similarv alues were obtained from calculations in the gas phase and in a1 -octanol solventf ield. This suggestst hat the 1:2c omplexo f10 is marginally more stable than that of 3,b ut that neither has as ignificantly greater binding energy than the other.T his may indicate that there is little difference between the inherent metal binding energies of the ligands 3 and 10,a nd that the differences in extraction properties observed above arise insteadf rom specific solvent interactions.

Conclusion
We report on three novel bis-1,2,4-triazine ligandsd erived from af ive-membered ring diketone, and we show for the first time how tuningt he aliphatic ring size of bis-1,2,4-triazine ligands leads to subtle changes in the speciation of the ligands with trivalent f-block metal ions, the thermodynamic stabilities of the formed metal complexes, andt he trivalent actinide extractiona ffinities of the ligands. We proposet hat this insight could enable am ore rational design of actinide-selective ligands with tailoreds olvent extraction properties suitable for future spent nuclear fuel reprocessing to close the nuclear fuel cycle.