Observation of a promethium complex in solution

Lanthanide rare-earth metals are ubiquitous in modern technologies1–5, but we know little about chemistry of the 61st element, promethium (Pm)6, a lanthanide that is highly radioactive and inaccessible. Despite its importance7,8, Pm has been conspicuously absent from the experimental studies of lanthanides, impeding our full comprehension of the so-called lanthanide contraction phenomenon: a fundamental aspect of the periodic table that is quoted in general chemistry textbooks. Here we demonstrate a stable chelation of the 147Pm radionuclide (half-life of 2.62 years) in aqueous solution by the newly synthesized organic diglycolamide ligand. The resulting homoleptic PmIII complex is studied using synchrotron X-ray absorption spectroscopy and quantum chemical calculations to establish the coordination structure and a bond distance of promethium. These fundamental insights allow a complete structural investigation of a full set of isostructural lanthanide complexes, ultimately capturing the lanthanide contraction in solution solely on the basis of experimental observations. Our results show accelerated shortening of bonds at the beginning of the lanthanide series, which can be correlated to the separation trends shown by diglycolamides9–11. The characterization of the radioactive PmIII complex in an aqueous environment deepens our understanding of intra-lanthanide behaviour12–15 and the chemistry and separation of the f-block elements16.

which can be exploited in its separation and its applications, which are burgeoning.Also, its frames promethium in terms of the lanthanide contraction, a feature in the periodic table that has significant consequences, not only in the chemistry of the rare-earth elements but also in that for neighbouring metals such as zirconium and hafnium.
The chemistry presented is very difficult and is only achievable in appropriately equipped laboratories with researchers who are experts.This is the main aspect that makes this work stand out and the authors should be applauded for not only isolating the Pm starting materials and diglycolamide product but also ensuring the validity of the characterisation.As the authors admit, these are not the first Pm complexes to be prepared but they are the first that allows a close inspection of the variation in bond distances for the whole lanthanide series (including La).
There are some weaknesses in the manuscript.1.The authors state that there is limited comprehension of Pm chemistry.While I agree that there is limited Pm chemistry existing, the general chemistry of this element could be straightforwardly inferred from the related chemistry of its closest congenors, Nd and Sm.
2. I have some misgivings about the EXAFS data analysis as these show only one M-O bond distance that is an average of the 2 x M-O(amide) and 1 x M-O(ether) bonding -the lack of differentiation is ascribed to fluxionality.No evidence is provided for a dynamic process occurring in solution, and while ligand exchange on lanthanides is diffusion controlled, this should not be relevant to complexes of chelating ligands.If exchange is happening (e.g. with water) then this should be taken into account as aquo ligands may be present.
3. The authors describe the subtle changes in Ln-O bond distance across the series (Fig 4a) and state that there are deviations from the standard quadratic model, with a particular deviation prior to Pm.It is interesting that a similar step in the curve is also seen between Tb and Ho, but this is not commented on.Also, if La is removed from the sequence (it is a group 3 element with 0 f electrons) then a more defined line for the series is seen.Can the authors comment on these facets?Some more general points.1.I do not like the title -I do not think it reflects the work presented as this is a characterisation of a complex, not just its observation.2. There appears to be some confusion of what is a rare-earth element and what is a lanthanide.maybe use the term lanthanoid or talk about the lanthanides + La 3. Fig. 1 cation.Pm(NO3)3 is a little too simplistic, more likely to be [Pm(H2O)9][NO3]3.4. Fig. 2 caption.the complex is not organometallic, but a coordination complex. 5. Fig. 2 caption.It is not clear why the diglycolamide should provide aqueous solubility, especially as the amido oxygen atoms are coordinated.
Referee #3 (Remarks to the Author): Popovs and co-authors have presented an interesting study of the Promethium complex synthesized with a novel PyDGA complexing agent.The resulting [Pm (PyDGA3)]3+ complex was analyzed using experimental X-ray absorption spectroscopy methods, covering XANES and EXAFS regions.The authors have also made great efforts to conduct DFT simulations, shedding light on the nature of observed electronic transitions in the XANES region.Additionally, the EXAFS data was fit with the help of ab initio molecular dynamics simulations.The agreement between theoretical and experimental EXAFS data is excellent.The authors went deeper into their analysis by examining the nature of the Pm-O bond using natural bond orbital calculations.I found the results, particularly the insight that Pm-O bonds originate from an electron density donation from O lone pairs to the Pm center, very exciting.
The manuscript is well-written with flawless English and clear explanations.I believe it will be of interest to the broader Nature community.All the necessary information for understanding the synthesis and characterization of this novel material is provided.Therefore, I am pleased to recommend this paper for publication.However, I have a few remarks and several questions that might be addressed: However, I don't observe this in the Ln(PyDGA3)]3+ particular case.Nevertheless, the EXAFS data does show such trends (Fig. 4).I'm curious as to why the position of post-edge features in XANES does not appear sensitive to these differences.It might be helpful if the authors could add dashed vertical lines to indicate the trends in post-edge features.Do the authors have an explanation for this?
2) I found Extended Figure 4 a bit misleading.The authors mention that the figure should show region "I" corresponding to the 2p-5d transitions and region "III" attributed to transitions involving Pm 4f orbitals, while the origin of broad feature "IV" is quite complex with leading components from 2p to 5d/ligand and Pm 4f dz3 orbitals (Extended Data Fig. 4).However, only the plot of molecular orbitals is provided, which might not be understandable to non-expert readers.I suggest plotting the corresponding density of states related to the 5d and 4f contributions below the spectrum.
3) I'm curious if the pre-edge region of the Pm compound only shows 5d states reflections (please see my comment above; it would be better to examine the DOS with respect to the Fermi energy).It is somehow known fact that region "I" in Ln3+ should contain pre-edge structure due to the quadrupole 2p-4f excitations (which might only be visible in high-energy resolution XAS mode like HERFD, as authors may not be able to experimentally resolve it).However, DFT calculations should show the 4f contribution to Region I in the spectrum.1817-1830 (2022).and citations there) Out of curiosity, I took the Pm2O3 structure from 1972 and ran FEFF calculations to get an idea of the 4f and 5d states distributions.I believe your Pm L3 spectrum might contain 4f states in Region I. 4) I recommend including information about the ground state configuration.How many 5d and 4f electrons does Pm have?I ask because I am a bit confused.Figure 4c indicates that the Pm3+ compound contains 4f1 electrons.However, the known ground state configuration of Pm is [Xe] 4f5 6s2 with the term symbol 6H5/2.Therefore, Pm 3+ should contain 4 electrons at the 4f level.Do authors see it differently?
Minor comments: • What was the Pm activity per sample?(perhaps in Bq) • What was the final Pm concentration in the sample measured by XAS?
• I understand that the sample was double confined, but with which material (I see polyamide mentioned later, but what was the thickness)?How did the sample holder look?Perhaps authors can add a photo of it to the Extended Data file?(I only saw photo on how the synthesis was done) How was it transported to the beamline from the lab? • Samples were measured at room temperature.Did the authors observe any radiation damage, or has this issue been checked?• How long was the EXAFS data collected (few min or 30 minutes per scan)?Was it taken from one spot on the sample or from several spots?How homogeneous was the sample?• What was the beam size?
• Line 259 in the Data Collection section: It is written that "the data were energy-calibrated to the main edge from the spectra of Ln oxide standards," but earlier in the text, authors mention that energy calibration was done using the Fe foil.Please add more clarity.• Beamline 6-BM of NSLS was used, but no citation to the BL is given.Please add it.
• Extended Figure 8 is identical to Figure 4c.I believe the Extended Data can be removed.

Author Rebuftals to Inifial Comments:
Referee #1 (Remarks to the Author): This manuscript describes the synthesis and characterisafion of a novel promethium compound, with a view to stabilising it for long enough to probe the chemical coordinafion by X-ray Absorpfion Spectroscopy.The data detail, for the very first fime, the experimentally observed Pm-O bond distance, which is close to previous theorefical predicfions.It seems remarkable these days that not all element-O bonding environments are known -that these authors have been able to elucidate this is testament to their approach (and probably a great deal of pafience in the radiochemistry laboratory) and the ufility of the 6BM beamline at NSLS-II.These results, together with a wider assessment of the enfire lanthanide series, are verified by atomisfic modelling, which give new, and confirmatory insights to this important series of elements.We appreciate the referee's crifical review of our manuscript.
The methodology and data produced are sound and well-jusfified.The authors are to be applauded for their separafion approach, which gave rise to such "clean" solufions for analysis.I also understand the rafionale behind examining aqueous solufions rather than solid samples; however, do the authors expect there to be any effects from radiolysis of the solufions?
We thank the referee for the valid comments.No radiafion damage was observed during our XAS measurements, as we ufilized a low-flux-density, unfocused beam at 6-BM.This was verified by comparing individual XAS scans for all Ln samples, which did not show any abnormal changes.Addifionally, a comparison of individual Pm-PyDGA XANES scans from the L3-and L1-edge spectra is also provided in Extended Data Fig. 13.Also, I wonder whether the erroneous datapoints in the EXAFS region, at 8 A-1 in k space, are actually mulfi-electron excitafion peaks rather than the absorpfion edges of Sm or Nd?The lafter is fully jusfified (although the presence of the edges was not evidenced in the data as far as I could see), but can the authors comment on whether they observed any evidence of the former?
We have aftached the Pm XAFS plot with the magnified region (inset), poinfing to the small presence of Sm L3 (6716 eV)/Nd L2 (6722 eV) edges.
In Fig. 3b (main text), the XAS data are presented in k -space, and the χ(k) is amplified by k 3 , making the adjacent lanthanides' features more apparent (quick calculafions show that ~8.2 Å -1 is in line with the energy range for the Sm L3/Nd L2 edges).We also note in the Methods secfion that "a small quanfity of 146 Nd was present in the sample due to challenging separafions of the adjacent lanthanides using the aforemenfioned techniques.Traces of Sm present in the sample on the moment of XAS measurements originated from the radioacfive decay of promethium to the daughter samarium according to the following process: 147 Pm (β -→) 147 Sm.Approximately 77 days had passed between the Pm purificafion and XAS data collecfion.Based on the τ ½= 2.62 years of the radioacfive decay, up to 5.632% of the starfing 147 Pm has decayed into 147 Sm at the fime of the sample measurements at NSLS-II." Regarding the mulfi-electron excitafion peaks menfioned by the referee, we indeed observe them at approximately 5.7 Å -1 to 6.0 Å -1 for the light lanthanides (Extended Data Fig. 7).This is consistent with the previous study by Solera et al. (Mulfielectron Excitafions at the L Edges in Rare Earth Ionic Aqueous Solufions.Phys. Rev. B 1995, 51, 2678-2686), where the mulfi-electron excitafion features were seen at about 5-7 Å -1 .
The conclusions are robust and well-jusfified.I find the references to be appropriate and the abstract to be both enficing and well-wriften.This is an excellent paper and I fully recommend its publicafion in Nature, following an answer to the two very minor quesfions above.We appreciate the referee's posifive evaluafion of our work.

Referee #2 (Remarks to the Author):
This manuscript reports the synthesis and characterisafion in solufion diffracfion measurements of a set of rare-earth complexes of diglycolamide ligands.Importantly, and for the first fime, this set of elements includes promethium, the element missing from all previous examples of rare-earth chemistry.This work is therefore highly original and significant, as it provides new knowledge in promethium chemistry which can be exploited in its separafion and its applicafions, which are burgeoning.Also, its frames promethium in terms of the lanthanide contracfion, a feature in the periodic table that has significant consequences, not only in the chemistry of the rare-earth elements but also in that for neighbouring metals such as zirconium and hafnium.
We appreciate the referee's comments.
The chemistry presented is very difficult and is only achievable in appropriately equipped laboratories with researchers who are experts.This is the main aspect that makes this work stand out and the authors should be applauded for not only isolafing the Pm starfing materials and diglycolamide product but also ensuring the validity of the characterisafion.As the authors admit, these are not the first Pm complexes to be prepared but they are the first that allows a close inspecfion of the variafion in bond distances for the whole lanthanide series (including La).We appreciate the referee's posifive evaluafion of our work.
There are some weaknesses in the manuscript.1.The authors state that there is limited comprehension of Pm chemistry.While I agree that there is limited Pm chemistry exisfing, the general chemistry of this element could be straighfforwardly inferred from the related chemistry of its closest congenors, Nd and Sm.
We parfially agree with the comment from the reviewer.We note that Pm is situated between Nd, which has accessible oxidafion states of +3 and +4, and Sm, which exhibits oxidafion states of +2 and +3.By stafing that Pm chemistry had been "virtually unexplored", we tried to draw the community's aftenfion to the fact that this element is indeed squarely located in the lanthanide series, however very liftle is known about its chemical behavior.In fact, it is plausible that Pm could exhibit chemical properfies, such as an accessible oxidafion state, that are different from one or, perhaps, even both of its neighbors and further studies are warranted to explore this possibility.For instance, it is not known whether complexes with formal Pm(II) or Pm(IV) can be prepared.
2. I have some misgivings about the EXAFS data analysis as these show only one M-O bond distance that is an average of the 2 x M-O(amide) and 1 x M-O(ether) bonding -the lack of differenfiafion is ascribed to fluxionality.No evidence is provided for a dynamic process occurring in solufion, and while ligand exchange on lanthanides is diffusion controlled, this should not be relevant to complexes of chelafing ligands.If exchange is happening (e.g. with water) then this should be taken into account as aquo ligands may be present.We appreciate the reviewer's comment and would like to provide clarificafion regarding our use of the term 'dynamic process' concerning the behavior of ligands around the metal ion in solufion.The interacfions governing ion coordinafion in an aqueous solufion differ significantly from those in a crystalline solid.In solid-state environments, structural arrangements are predominantly stafic, allowing for disfinct resolufion of Ln-O (amide) and Ln-O (ether) from corresponding single-crystal XRD data, often collected for solid-state complexes at ~100 K.However, in a solufion, molecules and ions lack organizafion on a laftice, causing the bonds formed between the metal ion and the donor groups of the ligand to exhibit dynamic behavior.
Our EXAFS data for the Pm complex were obtained in the solufion phase at room temperature.Under these condifions, the primarily ionic bonding interacfions between Pm(III) and the donor atoms of PyDGA are expected to be more dynamic compared to the solid-state.This dynamic behavior contributes to the challenge of resolving individual Pm-O (amide) and Pm-O (ether) distances in the complex.Our fifting model, describing the Pm EXAFS with an average Pm-O distance, proved safisfactory, and no stafisfically significant value could be added to the EXAFS analysis by considering separate Pm-O (amide) and Pm-O (ether) distances.
Moreover, the dynamic nature of the Pm(III)-ligand interacfions is visually demonstrated in the aftached movie (mp4 file), illustrafing the Pm coordinafion complex and water molecules within 5 Å of the metal ion.This visualizafion is based on our ab-inifio molecular dynamics simulafions, which also reproduce the experimental EXAFS spectra (Fig. 3b,c).Addifionally, we have included a plot showing the evolufion and overlap of the Pm-O (amide) and Pm-O (ether) distances over fime.As depicted, it is stafisfically challenging to disfinguish individual Pm-O (amide) and Pm-O (ether) bonds in solufion.Therefore, we aftribute the lack of differenfiafion to the fluxionality of these interacfions in an aqueous environment at room temperature.
3. The authors describe the subtle changes in Ln-O bond distance across the series (Fig 4a) and state that there are deviafions from the standard quadrafic model, with a parficular deviafion prior to Pm.It is interesfing that a similar step in the curve is also seen between Tb and Ho, but this is not commented on.Also, if La is removed from the sequence (it is a group 3 element with 0 f electrons) then a more defined line for the series is seen.Can the authors comment on these facets?
We appreciate the insighfful comment from the reviewer.The EXAFS data in Fig. 4c are presented with error bars, specifically 1 error bars associated with each data point.These are based on EXAFS fifting uncertainty and were computed from the covariance matrix of the non-linear minimizafion of the EXAFS fit.Therefore, drawing definifive conclusions from a small deviafion from the standard quadrafic model becomes challenging, especially considering potenfial influences such as the restricted k-window for the Pm and Dy cases (see Extended Data Table 4).
In the manuscript, we just state that there is a "somewhat accelerated shortening of bonds at the beginning of lanthanide series" and "filling the 4f orbitals apparently influences shielding of the nuclear charge and according to our data this effect was most pronounced early in the series from La to Pm, accounfing for as much as ≈ 36% of the overall Ln contracfion."These findings align with Shannon's effecfive ionic radii decrease (at a coordinafion number of nine), which is more substanfial at the beginning of the series than at the end.
It is important to note that our use of a quadrafic fit was aimed at obtaining parameters necessary for the derivafion of a shielding constant for f electrons (s = 0.74).While we agree that removing La from the sequence could result in a more defined line, including La was crucial for comparing the f-electron shielding constants obtained by different methods.To the best of our knowledge, a generally accepted value of s = 0.69 was obtained from the Ln ionizafion energies, including La in the dataset.Some more general points.1.I do not like the fitle -I do not think it reflects the work presented as this is a characterisafion of a complex, not just its observafion.
We thank the reviewer for this suggesfion.Mirriam-Webster dicfionary provides the following definifion of the noun "observafion"-an act of recognizing and nofing a fact or occurrence often involving measurement with instruments.We believe the fitle corresponds to the spirit of our work.
2. There appears to be some confusion of what is a rare-earth element and what is a lanthanide.maybe use the term lanthanoid or talk about the lanthanides + La.
We appreciate the reviewer's aftenfion to details and the opportunity to clarify the terminology used in our manuscript.In the context of our work, we have chosen to use the term "lanthanide" to refer to the series of chemical elements with atomic numbers 57 to 71, inclusive.We have included the following text in the introducfion paragraph to elucidate our use of the term "lanthanide": "One reason promethium (Pm) was so elusive for many years, despite a relafively low atomic number (Fig. 1), is that it is the only element in the lanthanide series (elements with atomic numbers 57-71) with no stable isotopes."3. Fig. 1 cafion.Pm(NO3)3 is a liftle too simplisfic, more likely to be [Pm(H2O)9][NO3]3.
We thank the reviewer for this correcfion.The promethium nitrate aqueous solufion was slowly evaporated in a negafive-pressure radiological glovebox upon heafing to obtain an anhydrous salt.The photograph was taken right after the sample of an anhydrous salt cooled to reach ambient temperature.There were no addifional precaufions taken to ensure the complete removal of water from the glovebox.Addifionally, knowing that anhydrous Pm(NO3)3, similar to other lanthanides, is expected to be hygroscopic, we cannot rule out that a small amount of water was reabsorbed by the salt to form a hydrate.Therefore, per reviewer's suggesfion we have changed figure capfion to reflect this fact: "Fig. 1 | Summary of lanthanide elements, year of their discovery, and electron configurafion after [Xe] core along with a photograph of purified Pm III compound prepared in this study.The depicted pink-colored Pm(NO3)3•nH2O (n < 9) solid residue was obtained after mulfiple purificafion steps and used in a Pm III complexafion with the diglycolamide ligand."4. Fig. 2 capfion.the complex is not organometallic, but a coordinafion complex.
We have changed Fig. 2  We thank the reviewer for the comment.PyDGA ligand was designed and synthesized to serve several purposes.First, the ligand's subsfitufion paftern was specifically ensuring good aqueous solubility while maintaining strong binding and similar funcfionality as in a very promising reagent used In lanthanide separafion, TODGA (N,N,N',N'-tetraoctyl diglycolamide) that is extremely lipophilic.The present PyDGA ligand has experimentally verified aqueous solubility in excess of 200 g/L.Second, we used a neutral chelafing ligand PYDGA to ensure high aqueous solubility of the complex formed with the lanthanides to prevent the possible precipitafion of the metal complex.
We thank the referee for thoughfful review and posifive feedback on our work.The insighfful comments have significantly contributed to the improvement of the manuscript.
However, I have a few remarks and several quesfions that might be addressed: 1.I am somewhat surprised by the Extended data in Fig. 2 on Sm, Nd, and Pm compounds.The L3 XANES on different Lns appear very similar.Our experience in invesfigafing a series of Ln ions in the same structure typically shows differences between them, and this is usually reflected in the XANES data, specifically in the posifions of the post-edge features.(c.f.Zasimov et al, Inorg.Chem.2022).However, I don't observe this in the Ln(PyDGA3)]3+ parficular case.Nevertheless, the EXAFS data does show such trends (Fig. 4).I'm curious as to why the posifion of post-edge features in XANES does not appear sensifive to these differences.It might be helpful if the authors could add dashed verfical lines to indicate the trends in post-edge features.Do the authors have an explanafion for this?
We indeed observe changes in the posifions of the post-edge features, which was not much evident from Extended Data Fig. 3, because of the plofting scale.We have introduced data for the La and Lu complexes and added verfical lines to make the shift in the post-edge posifions more apparent.The trends in post-edge XANES features along the lanthanide series agreed well with the previous study by Zasimov et al. (cited in the revised version of the manuscript).The following sentence was added: "Furthermore, the shrinkage of the Ln-O bonds is corroborated by the trend in the relafive energy posifions of the Ln L3-edge XANES spectral features (Extended Data Fig. 3), consistent with the results of a recent study 41 on some isostructural Ln compounds using high-energy-resolufion fluorescencedetected XANES (HERFD-XANES) 42,43 measurements."Extended Data Fig. 3 was also updated accordingly.4 a bit misleading.The authors menfion that the figure should show region "I" corresponding to the 2p-5d transifions and region "III" aftributed to transifions involving Pm 4f orbitals, while the origin of broad feature "IV" is quite complex with leading components from 2p to 5d/ligand and Pm 4f dz3 orbitals (Extended Data Fig. 4).However, only the plot of molecular orbitals is provided, which might not be understandable to non-expert readers.I suggest plofting the corresponding density of states related to the 5d and 4f contribufions below the spectrum.

I found Extended Figure
Since we focus on an isolated molecular system (the Pm complex), the concept of density of states (DOS), primarily used in the solid-state physics to represent the number of states in unit energy interval with energy levels being configuous, is not very relevant here because in a molecular system the energy levels are discrete.Therefore, we followed the original DFT/ROCIS approach and XANES interpretafion by Dr. F. Neese (the DFT/ROCIS code developer), where the assignment of XANES features is interpreted on the basis of natural difference orbitals (NDOs).
Addifionally, following the reviewer's recommendafion, we performed mulfiple scaftering theory calculafions (FEFF9 code) to reproduce XANES and explore the origin of the peaks using DOS.The corresponding FEFF9-simulated XANES spectrum and the local DOS are plofted in Extended Data Fig. 4, consistent with our DFT/ROCIS results.We anficipate that presenfing results in this way (both molecular orbitals and DOS are presented) will be helpful to a general reader.
3. I'm curious if the pre-edge region of the Pm compound only shows 5d states reflecfions (please see my comment above; it would be befter to examine the DOS with respect to the Fermi energy).It is somehow known fact that region "I" in Ln3+ should contain pre-edge structure due to the quadrupole 2p-4f excitafions (which might only be visible in high-energy resolufion XAS mode like HERFD, as authors may not be able to experimentally resolve it).However, DFT calculafions should show the 4f contribufion to Region I in the spectrum.Can the authors comment on why they don't observe it?(please see more work done at the pre-edges : Hämäläinen, K., Siddons, D. P., Hasfings, J. B. & Berman, L. E. Eliminafion of the inner-shell lifefime broadening in x-ray-absorpfion spectroscopy.Phys.Rev. Left. 67, 2850Left. 67, -2853Left. 67, (1991)).Kvashnina, K. O., Butorin, S. M. & Glatzel, P. Direct study of the f-electron configurafion in lanthanide systems.J. Anal. At. Spectrom. 26, 1265(2011), Zasimov, P. et Inorg. Chem. 61, 1817-1830(2022).and citafions there) Out of curiosity, I took the Pm2O3 structure from 1972 and ran FEFF calculafions to get an idea of the 4f and 5d states distribufions (please see aftached file).I believe your Pm L3 spectrum might contain 4f states in Region I.
Since the convenfional L3-edge XANES of Pm is likely broadened by the large 2p core-hole lifefime, we acknowledge that high-energy-resolufion fluorescence-detected XANES (HERFD-XANES) measurements could potenfially offer befter resolufion of XANES features for Pm and other lanthanides.However, achieving this would necessitate access to specific beamline/instrumentafion.Addifionally, the challenges associated with handling the Pm radionuclide would warrant a separate publicafion on this topic.We thank the reviewer for providing relevant references on HERFD-XANES works for the Ln, which we cite in the revised version of our manuscript.
We also appreciate the reviewer for supplying addifional FEFF calculafions for the Pm2O3 structure.These calculafions demonstrate that the pre-edge feature originates from both f and d states, as seen in the projected DOS.Our FEFF9 calculafions for the Pm(PyDGA)3 complex (Extended Data Fig. 4b) led to similar conclusions, indicafing that the pre-edge feature (I) in the Pm complex arises from quadrupole 2p-4f and dipole 2p-5d electronic transifions.
A detailed analysis of our DFT/ROCIS results using the natural difference orbitals (NDOs) revealed that transifions from 2p to 5d orbitals mainly contribute to the pre-edge region (I), with a smaller fracfion of 2p to 4f transifions.This is generally consistent with the FEFF9-DOS calculafions.While quadrupole transifions (2p-4f) are typically two orders of magnitude weaker than dipole transifions (2p-5d), the DFT/ROCIS theory might introduce variafions in relafive values of 4f and 5d transifion energies, depending on the chosen density funcfional and/or basis set.Thus, although esfimafing the exact contribufion of 4f and 5d transifions in region I is challenging, the consistent results from the molecular orbital approach (DFT/ROCIS) and DOS (mulfiple scaftering theory) indicate the involvement of both f and d states in region I.
We have modified the following sentence in the main text: "Based on our density funcfional theory restricted open shell configurafion interacfion singles (DFT/ROCIS) and mulfiple scaftering theory calculafions (Extended Data Fig. 4), region "I" corresponds to transifions from Pm 2p to 4f/5d orbitals and the most intense peak "II" is dominated by 2p core electron excitafions to 5d but with some PyDGA orbital contribufions.Less visible peak "III" can be aftributed to transifions involving Pm 4f/5d/ligand orbitals, while the origin of broad feature "IV" is quite complex with leading components from 2p to 5d/ligand and Pm 4f dz 3 orbitals."Extended Data Fig. 4 was also updated accordingly.
4. I recommend including informafion about the ground state configurafion.How many 5d and 4f electrons does Pm have?I ask because I am a bit confused.Figure 4c indicates that the Pm3+ compound contains 4f1 electrons.However, the known ground state configurafion of Pm is [Xe] 4f5 6s2 with the term symbol 6H5/2.Therefore, Pm 3+ should contain 4 electrons at the 4f level.Do authors see it differently?
We confirm that the Pm(III) complex contains 4f4 electrons (not 4f1).Figure 4c in the original manuscript was also correct.X-axis in Figure 4c corresponds to the number of Ln f-electrons in the studied complexes; i.e. 4f1 corresponds to Ce(III) and 4f4 to Pm(III).Computafional details Methods secfion also indicates the ground state configurafion of the Pm(III) complex: "… the complex was treated as a triply charged quintet with four unpaired f-electrons."

Minor comments:
• What was the Pm acfivity per sample?(perhaps in Bq) For 8.5 mM 147 Pm, the acfivity was 4.0 GBq (108 mCi).
• What was the final Pm concentrafion in the sample measured by XAS?
This informafion is provided in the Methods secfion: "Approximately 77 days had passed between the Pm purificafion and XAS data collecfion.Based on the τ ½= 2.62 years of the radioacfive decay, up to 5.632% of the starfing 147 Pm has decayed into 147 Sm at the fime of the sample measurements at NSLS-II." Inifially, 8.5 mM Pm was used for the sample preparafion right after the purificafion procedure, and thus the final Pm concentrafion measured by XAS was ~8 mM.
• I understand that the sample was double confined, but with which material (I see polyamide menfioned later, but what was the thickness)?How did the sample holder look?Perhaps authors can add a photo of it to the Extended Data file?(I only saw photo on how the synthesis was done) How was it transported to the beamline from the lab?
This informafion has been provided in the Methods secfion.We also added the capillary thickness and manufacturer (0.05 mm, Cole-Parmer): "To ensure the full complexafion of promethium, a solufion (~90 µL) of 8.5 mM 147 Pm(NO3)3 was added to 180 mM PyDGA.The obtained solufion was then loaded into a polyimide capillary (1.8 mm inner diameter by 5 cm long, 0.05 mm thickness, Cole-Parmer) using a Hamilton syringe and then sealed twice with Devcon 2 Ton epoxy (Extended Data Fig. 1).Once the epoxy had dried completely, the sample was transferred from a glovebox to a radiological fume hood for further decontaminafion.The sample was then surveyed and doubly contained for shipment to the XAS beamline." We have provided addifional photos related to the sample transportafion and our sample holder in Extended Data Fig. 1.
• Samples were measured at room temperature.Did the authors observe any radiafion damage, or has this issue been checked?
We did not observe any radiafion damage during our XAS measurements, as the low-flux-density, unfocused beam was used at 6-BM.This was checked by comparing individual XAS scans for all Ln samples, which did not show any abnormal changes.Addifionally, a comparison of individual Pm-PyDGA XANES scans from the L3-and L1-edge spectra is also provided in Extended Data Fig. 13.
• How long was the EXAFS data collected (few min or 30 minutes per scan)?Was it taken from one spot on the sample or from several spots?How homogeneous was the sample?
We spent approximately 10 minutes per scan, focusing the beam on one spot.As we only had liquid samples, homogeneity was not an issue.
• What was the beam size?6 mm x 0.3 mm • Line 259 in the Data Collecfion secfion: It is wriften that "the data were energy-calibrated to the main edge from the spectra of Ln oxide standards," but earlier in the text, authors menfion that energy calibrafion was done using the Fe foil.Please add more clarity.This statement is correct -calibrafion for the Pm sample was done using an Fe foil.For other lanthanides, we used the corresponding Ln oxide standards.This is now clarified in the text: "The data for Ln, except Pm, were energy-calibrated to the main edge from the spectra of Ln oxide standards." • Beamline 6-BM of NSLS was used, but no citafion to the BL is given.Please add it.
There isn't one yet.The beamline was fully acknowledged in the acknowledgment secfion.
• Extended Figure 8 is idenfical to Figure 4c.I believe the Extended Data can be removed.
We used Extended Data Fig. 8 to illustrate the quadrafic fit and the derived parameters from this fit, which were then employed to calculate the shielding constant for f electrons.The figure capfion has been slightly modified to accurately reflect this.
1) I am somewhat surprised by the Extended data in Fig.2 on Sm, Nd, and Pm compounds.The L3 XANES on different Lns appear very similar.Our experience in investigating a series of Ln ions in the same structure typically shows differences between them, and this is usually reflected in the XANES data, specifically in the positions of the post-edge features.(c.f.Zasimov et al, Inorg.Chem.2022) capfion to coordinafion complex: "Fig. 2 | The synthesized mulfidentate ligand bipyrrolidine diglycolamide (PyDGA) chelates Pm III to form the homolepfic coordinafion complex in an aqueous solufion."5. Fig. 2 capfion.It is not clear why the diglycolamide should provide aqueous solubility, especially as the amido oxygen atoms are coordinated.
Glatzel, P. Direct study of the f-electron configuration in lanthanide systems.J. Anal.At.Spectrom.26, 1265 (2011), Zasimov, P. et al.HERFD-XANES and RIXS Study on the Electronic Structure of Trivalent Lanthanides across a Series of Isostructural Compounds.Inorg.Chem.61, & al. HERFD-XANES and RIXS Study on the Electronic Structure of Trivalent Lanthanides across a Series of Isostructural Compounds.