Selective Extraction of Hafnium over Zirconium with Dialkylphosphinic Acids from H2SO4 Media

Designing highly efficient dialkylphosphinic acid extractant for zirconium/hafnium separation relies on systematic structure-property studies. In this paper, dialkylphosphinic acids with different substituents at β-C, γ-C and δ-C for zirconium and hafnium extraction and separation from H2SO4 media were investigated. The results show that substituents at β-C, γ-C and δ-C reduce the extraction ability of dialkylphosphinic acids for both zirconium and hafnium. The substituent effect at β-C is greater than that at γ-C and δ-C. The larger steric hindrance of the substituents (ethyl > methyl > H), the weaker extraction ability of the dialkylphosphinic acids. The zirconium and hafnium separation behavior is related to the extraction ability of dialkylphosphinic acids. Dialkylphosphinic acids with stronger extraction ability show better zirconium/hafnium separation performance at higher acidity of 2.0 mol L-1 H2SO4, while those with weaker extraction ability show better zirconium and hafnium separation performance at lower acidity of 0.25 mol L-1 H2SO4. The highest hafnium/zirconium separation factor (βHf/Zr) in the current study occurs with (2-ethylhexyl)(2,4,4-trimethylpentyl)phosphinic acid (USTB-1), which reaches 19.2.

Organophosphorous acids, of which typical commercial products are D2EHPA, EHEHPA, and Cyanex 272, selectively extract the minor hafnium over zirconium from H 2 SO 4 media.They have much stronger extraction ability for zirconium and hafnium than MIBK.They also have higher boiling points and flashing points, and lower solubility in water (see Table 1).Besides, organophosphorous acid-H 2 SO 4 systems have no problems of HSCN oxidation.It is the most promising kind of extractants to explore new zirconium/ hafnium separation systems with commercial prospects.
The current studies on zirconium/hafnium separation by organophosphorous acids are limited to D2EHPA, 11,12 EHEHPA, 13,14 Cyanex 272, 15,16 Cyanex 572 (a mixture of EHEHPA and Cyanex 272), 17 diisooctylphosphinic acid (Dio-PA) 32 and diphenyl phosphate (DPPA). 32D2EHPA has strong extraction ability for zirconium and hafnium, which leads to hard stripping of the extracted hafnium.Most importantly, its hafnium/zirconium separation factor (β Hf/Zr ) is only 2-3 in industrial production scale, much lower than that of MIBK-HSCN system (β Hf/Zr = 4-5). 7PPA has even much stronger extraction ability than D2EHPA, while Dio-PA has nearly equivalent extraction ability to D2EHPA. 32Their hafnium/zirconium separation factors have not been reported.Many studies [33][34][35] show that di-(2,4,4-trimethylpentyl) phosphinic acid (Cyanex 272) has weaker extraction ability and better separation performance for similar metals (such as Co/Ni, rare earths) than D2EHPA and EHEHPA.We can expect more efficient extraction for zirconium/hafnium separation from dialkylphosphinic acids.In addition, dialkylphosphinic acids have no O-P bond and will not hydrolyze in acidic media.They are more stable in acidic media than D2EHPA and EHEHPA.Designing highly efficient dialkylphosphinic acid extractant for zirconium/hafnium separation relies on systematic structure-property studies.As far as we know, there has not been such research yet.
For dialkylphosphinic acids, their performances mainly depend on two factors: the number of carbon atoms and the branches.The former mainly determines the extractant solubility in water, while the latter determines the extractant extraction ability, selectivity, stripping, etc.The carbon atoms of commercial organophosphorous acid extractants D2EHPA, EHEHPA, Cyanex 272, Cyanex 301 and Cyanex 302 are all 16, which means 16 is the proper number of carbon atoms for dialkylphosphinic acid extractants.As to the branches, based on our previous studies, the substituents at α-C lead to too weak extraction ability. 36Therefore, the effect of substituents at β-C, γ-C and δ-C on zirconium/hafnium extraction and separation performance are more important.
In this work, seven dialkylphosphinic acids were synthesized in our lab or kindly provided by others.They all have 16 carbon atoms with different substituents at β-C, γ-C or δ-C.Their name, structure, molecular formula and the reported pK a are summarized in Table 2. Their zirconium/ hafnium extraction and separation behaviors from H 2 SO 4 media were investigated.The relationship between the  38 5.58 37 structure and their extraction ability and selectivity for zirconium and hafnium were revealed.
The zirconium and hafnium feed solutions were prepared by dissolving ZrCl 4 and HfCl 4 in H 2 SO 4 solution with different concentrations and diluting to volume in 1-L volumetric flasks.The H 2 SO 4 solutions were prepared by precisely diluting the concentrated H 2 SO 4 with deionized water.The H 2 SO 4 concentrations were confirmed by titration with standard NaOH solution, which were 0.25, 0.51, 0.75, 1.02, 1.49 and 2.01 mol L -1 , respectively.The feed solutions were let stand more than 24 h after preparation and then the extraction were carried out.The concentrations of zirconium and hafnium in the feed solutions were 64 ± 2 and 68± 4 mg L -1 , respectively.
The organic phases were 10 mmol L -1 of each corresponding dialkylphosphinic acid.They were prepared as follows: (i) preparation of 0.1 mol L -1 dialkylphosphinic acid: 2.90 ± 0.01 g of the corresponding dialkylphosphinic acid was dissolved in n-octane and diluted to volume in a 100-mL volumetric flask; (ii) stepwise dilution: 10 mL of the obtained 0.1 mol L -1 of dialkylphosphinic acid was measured and diluted to 100 mL in another 100 mL volumetric flask.

Instrumentation
A Multi-tube Vortexer UMV-2 (Usun Technologies Co., Ltd., Beijing, China) was used for mixing aqueousorganic phases (2500 r min -1 for 20 min).A Guanghe TD4C Low Speed Tabletop Centrifuge (Jintan Liangyou Instrument Co., Ltd., Changzhou, Jiangsu Province, China) was used for accelerating phase separation (3000 rpm for 5 min).An iCAP 7400 inductively coupled plasma optical emission spectrometer (ICP-OES, Thermo Fisher Scientific, Waltham, MA, USA) was adopted to determine the zirconium and hafnium concentrations in aqueous solutions.The parameters during Zr and Hf determination were as follows: the plasma power was 1150 W; the wavelengths for Zr, Hf determination were 339.198 and 339.980 nm, respectively; the pump speed was 50 rpm; the auxiliary gas flow rate was 0.5 L min -1 ; the nebulizer or carrier gas flow rate was 0.6 L min -1 ; the plasma view was axial and the sample washing time was 30 s.

Extraction procedure
Certain volumes (4 mL/2 mL at A/O (phase ratio, namely the volume ratio of aqueous phase to organic phase) = 2, and 3 mL/3 mL at A/O = 1) of aqueous feed solution and extractant organic solution were sealed in a 10-mL centrifugal tube and oscillated on the Multi-tube Vortexer at 2500 rpm for 20 min, which was sufficient to reach extraction equilibrium. 40Then, the mixture was centrifugated at 3000 rpm for 3 min to accelerate phase separation.After that, the aqueous phase was separated for zirconium and hafnium concentration analysis with ICP-OES.All the samples were measured 3 times, and we took the average value as the final result.Zirconium and hafnium concentration in the organic phase was calculated on the basis of mass balance (equation 1).The calculation equations of zirconium/hafnium extraction percentage (E), distribution ratio (D), and hafnium/zirconium separation factor (β Hf/Zr ) are shown in equations 2-4. (1) where C e(o) stands for the zirconium/hafnium concentration in the organic phase after extraction; C i(a) and C e(a) represent the zirconium/hafnium concentrations in the aqueous phase before and after extraction, respectively; V a and V o denote the volumes of the aqueous and organic phases, respectively; A/O is the phase ratio, namely the volume ratio of aqueous phase to organic phase (V a /V o ); D Hf and D Zr stand for the distribution ratios of hafnium and zirconium, respectively.
All the extraction experiments were carried out at room temperature.

Results and Discussion
β-Substituent effect on the extraction ability for zirconium and hafnium To discuss conveniently, the dialkylphosphinic acids were classified into two groups.As P208, INET-1, P218, USTB-1 and P227 have different substituents at the β-C, they were discussed together to reveal the β-substituent effect on the extraction ability of dialkylphosphinic acids for zirconium and hafnium.P218, P2132 and Cyanex 272 all have methyl at the β-C, while different substituents at the γ-C or δ-C.Hence, they were compared to investigate the γ-, δ-substituent effect on the extraction ability of dialkylphosphinic acids for zirconium and hafnium.
The feed solutions were zirconium-hafnium mixed solutions with different H 2 SO 4 concentrations in the range of 0.25-2.0mol L -1 .The phase ratio A/O was first set at 2. The results are shown in Figure 1.P208 has the highest zirconium extraction percentages, followed by INET-1, and then P218.USTB-1 and P227 have almost equivalent zirconium extraction percentages, which are both lower than P218 (Figure 1a).Similar phenomena occur to hafnium extraction (Figure 1b).These indicate that β-substituents reduce the extraction ability of dialkylphosphinic acids for zirconium and hafnium.The larger steric hindrance of the substituents (ethyl > methyl > H), the weaker extraction ability of the dialkylphosphinic acids.USTB-1 has an identical alkyl to P227, while the other alkyl of USTB-1 is a 5-carbon main chain with one methyl at the β-C and two methyls at the δ-C.The substituents at the δ-C also weaken the extraction ability of dialkylphosphinic acids for zirconium and hafnium, which will be discussed in the next subsection.This leads to the almost equivalent extraction ability of USTB-1 and P227 for zirconium and hafnium.It has also been reported that USTB-1 and P227 show nearly the same extraction ability for lutetium (Lu). 38t the phase ratio A/O of 2, the extraction percentages of zirconium and hafnium by P218, USTB-1 and P227 are relatively low (Figure 1).Therefore, we further investigated the zirconium and hafnium extraction behaviors by these three extractants at A/O of 1.The results are shown in Figure 2. The extraction percentages of zirconium and hafnium by P218 are greatly improved, especially in low acidities (both ca.30% at 0.25 mol L -1 H 2 SO 4 ).In contrast, the extraction percentages of zirconium and hafnium by USTB-1 and P227 are both increased by only ca.10% at 0.25 mol L -1 H 2 SO 4 .This indicates that P227 and USTB-1 have much weaker extraction ability for zirconium and hafnium than P218, especially at the acidities of more than 1.0 mol L -1 H 2 SO 4 (see Figure 2).P227 and USTB-1 also show equivalent extraction ability for zirconium and hafnium at A/O of 1.

γ, δ-Substituent effect on the extraction ability for zirconium and hafnium
To investigate the γ, δ-substituent effect on the extraction ability of dialkylphosphinic acids for zirconium and hafnium, we also extracted zirconium and hafnium from their mixed solutions with different H 2 SO 4 concentrations in the range of 0.25-2.0mol L -1 .The concentrations of P218, P2132 and Cyanex 272 were also all 10 mmol L -1 .The phase ratio A/O was first set at 2. The results are shown in Figure 3.For both zirconium and hafnium, P218 shows the strongest extraction ability among these three extractants, followed by P2132, and then Cyanex 272.This indicates that the ethyl at the γ-C and the two methyls at the δ-C both apparently reduce the extraction ability of dialkylphosphinic acids for zirconium and hafnium.
Cyanex 272 has very weak extraction ability for zirconium and hafnium.Its extraction percentages for zirconium and hafnium are apparently lower than those of P227 and USTB-1 (compared Figures 1 and 3).Hence, we also repeated the extraction experiments at A/O of 1.The results are shown in Figure 4.All the extraction percentages of zirconium and hafnium apparently increased compared to corresponding those A/O of 2, especially at lower acidities (< 1.0 mol L -1 H 2 SO 4 ).Through comparing Figure 3 with Figure 1, and Figure 4 with Figure 2, it can be seen that P2132 has a slightly higher extraction ability for zirconium and hafnium than P227.P227 has an ethyl on its β-C, while P2132 has a methyl on its β-C and an ethyl on its γ-C.This indicates that the extraction ability of dialkylphosphinic acids   for zirconium and hafnium is more influenced by the substituent at the β-C than that at the γ-C.

Analysis of the zirconium and hafnium extraction by the dialkylphosphinic acids
Tetravalent zirconium (Zr 4+ ) and hafnium (Hf 4+ ) are hard Lewis acids, while sulfate ion (SO 4

2-
) is hard Lewis base.All of them are high valence ions.Therefore, SO 4 2-has a strong affinity to Zr 4+ and Hf 4+ to form various complexes.Ma et al. 41 predicted the existent forms of zirconium and hafnium at different acidities in H 2 SO 4 media using MEDUSA software. 42According to their results, the prevailing existent forms of zirconium and hafnium are both M(SO 4 ) 3 2-(M = Zr or Hf) at acidities of < 3 mol L -1 H 2 SO 4 .Organophosphorous acids extract zirconium and hafnium through cation exchange mechanism from low acidic H 2 SO 4 solutions, and the extraction reaction can be written as: 15,[43][44][45][46] where M represents Zr or Hf, n = 2 or 3, (HL) 2 denotes dimer form of dialkylphosphinic acid molecule, L represents one dialkylphosphinic acid molecule losing one proton (similarly hereinafter).Hf 4+ has a weaker interaction with SO 4 2-than Zr 4+ , which leads to selective extraction of hafnium over zirconium by organophosphorous acids from H 2 SO 4 media. 45This is consistent with our results.For all the investigated dialkylphosphinic acids (P208, INET-1, P218, P2132, USTB-1, P227 and Cyanex 272), hafnium has higher extraction percentages than zirconium (see Figures 1-4).
The extraction percentages of zirconium and hafnium all decrease as the H 2 SO 4 concentration is increased (see Figures 1-4).This phenomenon is in accordance with the extraction mechanism of cation exchange.According to equation 6, hydrogen ions are released during zirconium and hafnium extraction by organophosphorous acids.Therefore, increasing acidity is against the zirconium and hafnium extraction.
Alkyls are electron-donating groups.Alkyl branches increase the electron cloud density of the oxygen atoms, which leads to weaker acidity of dialkylphosphinic acids and more difficulty in dissociation during extraction.According to equation 6, this is unfavorable for zirconium and hafnium extraction.Ethyl has astronger electron donor effect, followed by methyl and then H. β-C is nearer to the oxygen atoms of dialkylphosphinic acids than γ-C.Therefore, electron of the substituents at β-C is easier to transmit to the oxygen atoms of -POOH group than that of the substituents at γ-C.Therefore, it is easy to understand the effect of β, γ, δ-substituent effect on zirconium and hafnium extraction from low acidic H 2 SO 4 media (< 2 mol L -1 ) and the regularities.
Zirconium/hafnium separation performance by dialkylphosphinic acids at A/O of 2 Separation factor (β) is always used to evaluate the separation performance of two metals.In addition, the corresponding extraction percentages of the two metals and their differences ΔE should also be considered, because high separation factors might occur at high extraction percentages of the two metals.For example, one is 99.21%, the other is 99.98%.Despite their separation factor is about 40, the ΔE is only 0.77%, which has nonsense in industrial applications.Similarly, high separation factor might also occur at both low extraction percentages of the two metals.Therefore, we not only summarized the hafnium/zirconium separation factors (β Hf/Zr ) in Table 3, but also drew the corresponding differences of their extraction percentages ΔE Hf-Zr (namely, E Hf -E Zr ) for all the dialkylphosphinic acids (see Figure 5).When the E Hf is > 99% or the E Zr is < 1%, the error of β Hf/Zr might be big.Hence, their corresponding β Hf/Zr are not listed in Table 3.
Zirconium/hafnium separation performance by dialkylphosphinic acids at A/O of 1 At the phase ratio A/O of 1, the ΔE Hf-Zr values and the corresponding β Hf/Zr values by P218, USTB-1, P227, P2132 and Cyanex 272 at the acidities in the range of 0.25-2.0mol L -1 H 2 SO 4 are summarized in Figure 6 and Table 4, respectively.For P218, its maximum ΔE Hf-Zr , which is 24.8%, also occurs at acidity of 2.0 mol L -1 H 2 SO 4 and the corresponding β Hf/Zr is 3.3.

Conclusions
To reveal the structure-activity of dialkylphosphinic acids, zirconium and hafnium extraction and separation behaviors by seven dialkylphosphinic acids with different substituents at β-C, γ-C or δ-C from H 2 SO 4 media were investigated.The following conclusions are drawn: (i) β-Substituents reduce the extraction ability of dialkylphosphinic acids for zirconium and hafnium.The larger steric hindrance of the substituents (ethyl > methyl > H), the weaker extraction ability of the dialkylphosphinic acids.The extraction ability of the investigated dialkylphosphinic acids with different β-substituents is in the sequence P208 > INET-1 > P218 > USTB-1 ca.P227.The substituents at the γ-C and δ-C also apparently reduce the extraction ability of dialkylphosphinic acids for zirconium and hafnium.The extraction ability of the investigated dialkylphosphinic acids with different γ, δ-substituents is in the order P218 > P2132 > Cyanex 272.The extraction ability of dialkylphosphinic acids for zirconium and hafnium is more influenced by the substituent at the β-C than that at the γ-C.
(ii) As the acidity is increased in H 2 SO 4 media, the variation trend of ΔE Hf-Zr depends on the extraction ability of dialkylphosphinic acids.Stronger extraction ability leads to an upward trend of ΔE Hf-Zr at acidities in the range of 0.25-2.0mol L -1 H 2 SO 4 , and the corresponding best zirconium/hafnium separation performance occurs at 2.0 mol L -1 H 2 SO 4 .On the contrary, weaker extraction ability causes a downward trend of ΔE Hf-Zr , and the corresponding best zirconium/hafnium separation performance occurs at 0.25 mol L -1 H 2 SO 4 .For the dialkylphosphinic acids with stronger extraction ability, their zirconium/hafnium separation performance is in the order P218 > INET-1 > P208.For the dialkylphosphinic acids with weaker extraction ability, their zirconium/hafnium separation performance is in the sequence USTB-1 > Cyanex 272 > P227 > P2132.
(iii) USTB-1 has stronger extraction ability and better zirconium/hafnium separation performance than the commercialized Cyanex 272.The highest β Hf/Zr in the current study occurs to USTB-1, which reaches 19.2.

Figure 6 .
Figure 6.Difference of zirconium and hafnium extraction percentages ΔE Hf-Zr at A/O of 1.

Table 2 .
Name, structure, molecular formula and pK a of dialkylphosphinic acids

Table 4 .
Hafnium/zirconium separation factors (β Hf/Zr ) in H 2 SO 4 system at A/O of 1