Genesis of electron deficient Pt1(0) in PDMS-PEG aggregates

While numerous single atoms stabilized by support surfaces have been reported, the synthesis of in-situ reduced discrete metal atoms weakly coordinated and stabilized in liquid media is a more challenging goal. We report the genesis of mononuclear electron deficient Pt1(0) by reducing H2PtCl6 in liquid polydimethylsiloxane-polyethylene glycol (PDMS-PEG) (Pt1@PDMS-PEG). UV–Vis, far-IR, and X-ray photoelectron spectroscopies evidence the reduction of H2PtCl6. CO infrared, and 195Pt and 13C NMR spectroscopies provide strong evidence of Pt1(0), existing as a pseudo-octahedral structure of (R1OR2)2Pt(0)Cl2H2 (R1 and R2 are H, C, or Si groups accordingly). The weakly coordinated (R1OR2)2Pt(0)Cl2H2 structure and electron deficient Pt1(0) have been validated by comparing experimental and DFT calculated 195Pt NMR spectra. The H+ in protic state and the Cl− together resemble HCl as the weak coordination. Neutralization by a base causes the formation of Pt nanoparticles. The Pt1@PDMS-PEG shows ultrahigh activity in olefin hydrosilylation with excellent terminal adducts selectivity.

This paper describes the easy method to prepare the single atom catalyst of Pt. In this method, the use of polydimethylsiloxane-polyethylene glycol (PDMS-PEG) copolymer, which weakly supports the Pt atom, as an amphiphilic solvating agent successfully prevent the aggregation of Pt metals. Formation of the Pt1 species is supported by various spectroscopic data. For example, the full conversion of the starting platinum chlorides were confirmed by UV-vis and 195Pt NMR. Dispersion of Pt atoms were also supported by both UV and NMR. The reduced oxidation state of Pt is supported by XPS. DFT calculation was also performed to support these experimental data. However, although I am not an expert of this field, the identification seems to be not sufficient considering previous reports often measured EXAFS, EXANES, STEM etc, which directly support the formation of Pt1 species. It is of interest the developed Pt1 catalysts exhibit higher activity in the olefin hydrosilylation reaction. In this point, the easy preparation method is a great advantage. To enhance the importance of this work, it is recommended to show additional data on the stability and the reusability of the catalyst.
Reviewer #2 (Remarks to the Author): The manuscript by Bai, Zhang and co-workers reports on the synthesis of "electron deficient "Pt1" in PDMS-PEG" aggregates. It is claimed that in this work the authors generate "discrete platinum atoms Pt1 by reducing H2PtCl6 in PDMS-PEG aggregates. Later the authors write that the Pt1 species exists as a pseudo octahedral structure of the type (ROR)2PtCl2H2. The Pt1@PDMS-PEG showed ultra-high activity in olefin hydrosilylation. In the following introduction, the authors point the interest of making single reduced metal atoms rather than metal clusters. They then state that "reduced metal atoms in their conventional organometallic complexes are strongly bonded by electron rich ligands". What is the meaning of "conventional organometallic complex"? Moreover, unlike what is written by the authors, in fact reduced metal centers are classically stabilized by electron accepting ligands, such as CO they are using later on. They claim that being able to make "discrete reduced metal atoms, distinguishable from organometallic complexes, in liquid media may lead to precise synthesis of new materials at atomic scale". They are apparently confused with the definition of a metal complex, which is a molecule made of a metal center (in any formal oxidation state: positive, 0 or even negative, 0 being the "reduced metal atom" they refer to) and ligands (any type: neutral L type, anionic X type). There is no such thing as a discrete reduced metal atom in a liquid medium, free of any coordination by a ligand. These would agglomerate and make bulk material. In fact, coordination of solvent molecules acting as ligands to the metal center is very common in organometallic chemistry. In agreement, they propose that the Pt1 reduced species is (ROR)2PtCl2H2, which is nothing else than a complex. Therefore, the whole introduction should be modified in depth.
The authors have carefully studied the reduction of the Pt(IV) starting complex by EtOH in the presence of the PDMS-PEG polymer in EtOH/H2O mixture. They state on page 3 line 80, that "complete reduction of the Pt ions" is seen after 3h at 105°C. I do not understand what the authors mean by "loss of coulombic Pt-Cl bonds". There is still a signal in far IR for the Pt-Cl bond, and they show by ion chromatography that Cl/Pt ratio is 2. It is shown that Pt clusters are not formed. The question at this point is the nature of the Pt species. The authors claim that the absence of signals near 0 and -1617 ppm "indicates full reduction of platinum". It is not correct. It shows the efficient reduction of Pt(IV), but not the full reduction to Pt(0). Indeed, the Karstedt complex, a Pt(0) complex, features a signal at much higher field (-6130 ppm). They observe a signal at -2680 ppm. 195Pt NMR chemical shifts are indeed really sensitive to the electronic environment of the Pt center and therefore to the ligands. Notably, NMR signals for Pt(II) complexes have been observed at much higher field than -2680 ppm (see for ex JOMC 2001, 622, 180). The authors state that in Karstedt complex, "the platinum atoms are highly electron rich due to strongly electron donating by 1,3-divinyl-1,1,3,3-tetramethyldisiloxane". This is not correct. The Pt center is electron rich because it is a formal Pt(0), but the dvtms ligands are both sigma donor and pi acceptors via the two double bonds. In fact, it is because the ligand is (good pi) acceptor that the Pt(0) complex is stable and isolable. In order to further characterize the Pt species Pt1, the authors made a very interesting experiment using labelled and non labelled CO. They observed a significant upfield shift (-3231ppm), proving coordination of CO to the Pt center. Interestingly, the Pt NMR proves the coordination of only one CO ligand to the Pt center. Note that in the JOMC paper cited above, PtCl2(CO)2 is said to be colorless, just like the solution obtained upon reduction of H2PtCl6 here. Moreover, for example cis-PtBr2(CO)2, a defined Pt(II) complex, has a chemical shift of -4243 ppm which is a much lower field than the -3231 ppm for the Pt1 species. The authors propose a structure (ROR)2PtCl2H2 for Pt1, based on DFT calculations of Pt chemical shifts. However the structures proposed are Pt(IV) complexes with formally 2 Cl-and 2 H-ligands. This is not consistent with the NMR and XPS data, which both point to lower oxidation state at Pt. There is no experimental evidence for the 2 H ligands at Pt. The authors should record the proton coupled 195Pt NMR spectrum. If the two H are indeed bound to the Pt center, a triplet should be observed. Alternatively, Pt-H are also very well characterized by 1H NMR spectroscopy at high field. The corresponding signal should present the expected Pt satellites. It is difficult to trust the NMR chemical shifts of 195 Pt predicted by DFT for proposed structures, when the prediction of the chemical shift of CO is -11.2 ppm, compared to real chemical shift at ca 185 ppm. To me the NMR and XPS data is more consistent with either monometallic square planar Pt(Cl)2(L)2, (Pt(II) center) or bimetallic Pt complexes with bridging Cl ligands, (L)(Cl)Pt(mu2-Cl)2Pt(Cl)(L) with L being the PDMS "shell". Several bridging Cl Pt dimers are known in the literature. In these types of complexes, the poorly donating "PDMS" ligand (via O) would be readily displaced by CO. The Pt1 species being a Pt(II) complex would also be consistent with Cl to Br exchange using HBr. Finally, it is very doubtful that Pt1 is a discrete Pt(0) complex because the reactivity with octane was done in the presence of CHCl3, a very oxidizing solvent for low oxidation state Pt center. On the other hand Pt(II) complexes are stable toward CHCl3. To me a very important question is not addressed: what is the size and shape of the PDMS-PEG "vesicle" in the EtOH/H2O mixture. Does it accept more than one Pt complex? DOSY NMR experiments should be carried out to answer this question.
In conclusion for the synthesis/characterization of the Pt1 species, the data is not consistent with the proposed structure.
In conclusion, I believe that the generation of a highly active Pt complex in PDMS-PEG for hydrosilylation of alkenes is of high interest. However, I am not convinced at all by the structure proposed by the authors, and several additional experiments should be carried out. Comparison with data in the literature for Pt(II) complexes should also be done. Moreover, the introduction needs to be modified in depth. For these reasons, I recommend rejection of the article, but I believe it would make a nice contribution to Nat. Comm once the points raised are addressed.

Reviewer #1 (Remarks to the Author):
This paper describes the easy method to prepare the single atom catalyst of Pt. In this method, the use of polydimethylsiloxane-polyethylene glycol (PDMS-PEG) copolymer, which weakly supports the Pt atom, as an amphiphilic solvating agent successfully prevent the aggregation of Pt metals. Formation of the Pt1 species is supported by various spectroscopic data. For example, the full conversion of the starting platinum chlorides were confirmed by  NMR. Dispersion of Pt atoms were also supported by both UV and NMR. The reduced oxidation state of Pt is supported by XPS. DFT calculation was also performed to support these experimental data. However, although I am not an expert of this field, the identification seems to be not sufficient considering previous reports often measured EXAFS, EXANES, STEM etc, which directly support the formation of Pt1 species.

Response:
The TEM and the EDX analysis in STEM have been used to characterize the morphology  Table 5). It should be noted that the Pt atoms tend to agglomerate under intense electron beam or X-ray radiation, making TEM, STEM and synchrotron X-ray measurements infeasible in providing unambiguous evidence of discrete Pt atoms in our system 35 ."

(a) (b)
It is of interest the developed Pt1 catalysts exhibit higher activity in the olefin hydrosilylation reaction. In this point, the easy preparation method is a great advantage. To enhance the importance of this work, it is recommended to show additional data on the stability and the reusability of the catalyst.

Response:
We tested used catalyst for three times. The results are given in Supplementary Figure   10. A discussion is given in the revised manuscript: "Moreover, through three recycled uses,  Table 5). Further magnification of the image was not available due to formation of Pt nanoparticles from the aggregation of Pt 1 single atoms, induced by the high energy electron beams (For STEM, no phosphotungstic acid was used).
The procedure is given in Methods: "After the hydrosilylation reaction between 1-octene and (Me 3 SiO) 2 MeSiH (1-octene: 0.4488 g, (Me 3 SiO) 2 MeSiH: 0.89 g) finished, the solution was centrifuged. 0.6694 g reaction mixture was taken out from the upper solution, and analyzed by 1 H NMR. The remaining solution (0.6694 g) was used for the subsequent hydrosilylation reaction. For the reuse test, 0.4488 g 1-octene and 0.89 g (Me 3 SiO) 2 MeSiH was added. After reaction, solution was centrifuged. 1.3388 g (equal to the amount of added reactants) reaction mixture was taken out from the upper solution, and analyzed by 1 H NMR, while the remaining solution (0.6694 g) was used for the subsequent hydrosilylation reaction.
For the following reuses, the added amounts of 1-octene and (Me 3 SiO) 2 MeSiH were always 0.4488 g and 0.89 g, respectively. After each reaction, the amount of reaction mixture taken out for 1 H NMR from the upper solution after centrifugation was always 1.3388 g (equal to the amount of added reactants), and the remaining solution was kept at 0.6694 g for the subsequent hydrosilylation reaction."

Reviewer #2 (Remarks to the Author):
The manuscript by Bai, Zhang and co-workers reports on the synthesis of "electron deficient "Pt1" in PDMS-PEG" aggregates. It is claimed that in this work the authors generate "discrete platinum atoms Pt1 by reducing H 2 PtCl 6 in PDMS-PEG aggregates. Later the authors write that the Pt1 species exists as a pseudo octahedral structure of the type (ROR) 2 PtCl 2 H 2 . The Pt1@PDMS-PEG showed ultra-high activity in olefin hydrosilylation.
In the following introduction, the authors point the interest of making single reduced metal atoms rather than metal clusters. They then state that "reduced metal atoms in their conventional organometallic complexes are strongly bonded by electron rich ligands". What is the meaning of "conventional organometallic complex"? Moreover, unlike what is written by the authors, in fact reduced metal centers are classically stabilized by electron accepting ligands, such as CO they are using later on. They claim that being able to make "discrete reduced metal atoms, distinguishable from organometallic complexes, in liquid media may lead to precise synthesis of new materials at atomic scale". They are apparently confused with the definition of a metal complex, which is a molecule made of a metal center (in any formal oxidation state: positive, 0 or even negative, 0 being the "reduced metal atom" they refer to) and ligands (any type: neutral L type, anionic X type). There is no such thing as a discrete reduced metal atom in a liquid medium, free of any coordination by a ligand. These would agglomerate and make bulk material. In fact, coordination of solvent molecules acting as ligands to the metal center is very common in organometallic chemistry. In agreement, they propose that the Pt1 reduced species is (ROR) 2 PtCl 2 H 2 , which is nothing else than a complex. Therefore, the whole introduction should be modified in depth.

Response:
We thank the reviewer's comments, and the term of "conventional organometallic complex" has been removed from the revised manuscript. Instead, we emphasize that (ROR) 2 PtCl 2 H 2 has a weakly coordinated mononuclear electron deficient Pt 1 .
Detailed characteristics of the (ROR) 2 PtCl 2 H 2 structure have been fully discussed in the revised manuscript. What is important for the Pt 1 structure of this work is that the H is protic, instead of the hydride as in conventional metal-hydride bond. More evidence and further discussion are given in the revised manuscript to clarify the differences. More details are also given below in responses to the specific questions. Computationally, several possible structures of PtCl 2 (R 1 OR 2 ) 2 and Pt 2 (μ-Cl) 2 Cl 2 (R 1 OR 2 ) 2, in which the Pt is in Pt(II) state, are optimized and the results are compared with the (R 1 OR 2 ) 2 PtCl 2 H 2 in which the Pt is in reduced state as proposed in this work. The new computational results for Pt(II)Cl 2 (R 1 OR 2 ) 2 and Pt 2 (μ-Cl) 2 Cl 2 (R 1 OR 2 ) 2 were added to     The authors state that in Karstedt complex, "the platinum atoms are highly electron rich due to strongly electron donating by 1,3-divinyl-1,1,3,3-tetramethyldisiloxane". This is not correct.

The Pt center is electron rich because it is a formal Pt(0), but the dvtms ligands are both sigma donor and pi acceptors via the two double bonds. In fact, it is because the ligand is (good pi) acceptor that the Pt(0) complex is stable and isolable.
Response: We thank the reviewer's comments and have removed the related statements about electron rich in the Pt center of Karstedt complex. We added the discussion in the revised manuscript: "The Bader charge (~0.6 eV) on the Pt 1 center in (R 1 OR 2 ) 2 PtCl 2 H 2 (Entries 2-5, Supplementary center when Pt 1 @PDMS-PEG was exposed to CO. In our work, the 195 Pt NMR, 13 C NMR and CO infrared spectra all consistently showed the preference of mono-carbonyl coordination to the Pt 1 center in Pt 1 @PDMS-PEG. To rationalize the domination of mono-carbonyl coordination to the Pt 1 center, we added a statement to the revised manuscript, based on our new calculation, as: "… replacing the first R 1 OR 2 group with CO is energetically favored (ΔE = -0.18 eV). However, further replacement of the remaining R 1 OR 2 with a CO resulted in negligible energy difference (ΔE = -0.04eV), consistent with fact that only one CO binds to Pt 1 atom. " The calculated 195 Pt chemical shifts for the hypothetical structures of trans-Pt(II)Cl 2 (CO) 2 , cis-Pt(II)Cl 2 (CO) 2 and cis-Pt(II)Br 2 (CO) 2 , where two CO's are coordinated, are given in the We added the discussion in the Supplementary Materials (below Supplementary Table   4) "Clearly, the 195 Pt chemical shift of the CO di-coordinated Pt(II) structures, trans-PtCl 2 (CO) 2 and cis-PtCl 2 (CO) 2 , could not produce predicted chemical shifts in agreement with the experimentally determined value (-3231 ppm, Figure 2b in the manuscript).

Supplementary
If the CO treated Pt 1 @PDMS-PEG were in the form of cis-Pt(II)Cl 2 (CO) 2 , we would anticipate the appearance of the carbonyl bands at 2178 and 2138 cm -1 in DRIFT spectrum 33 .
However, only one peak at 2084 cm -1 appeared for the CO treated Pt 1 @PDMS-PEG ( Figure   3b in the manuscript).
It should be pointed out that, for Pt(II)Cl 2 L 2 (L = R 1 OR 2 ), the coordination of two CO  Table 4). In addition, unlike (R 1 OR 2 )Pt(0)Cl 2 H 2 , the formation of dicarbonyl complex, Pt(II)Cl 2 (CO) 2 , from Pt(II)Cl 2 (R 1 OR 2 ) is energetically favorable by the DFT calculations (Supplementary Table 4)." Summary discussion about the DRIFT results was added to the revised manuscript: "the characteristic peaks at 2178 and 2138 cm -1 corresponding to the cis-Pt(II)Cl 2 (CO) 2 were absent 33 " The authors propose a structure (

Response:
The reviewer raised an excellent point about 1 H-195 Pt coupling, which led to an in-depth investigation of the issue. Our explanation based on further computational has been added in the revised version: "The Bader charge of hydrogen is positive for all optimized (R 1 OR 2 ) 2 PtCl 2 H 2 structures (Entries 2-5, Supplementary Table 3), suggesting the H in (R 1 OR 2 ) 2 PtCl 2 H 2 is similar to labile proton rather than to hydride in nature.
In the presence of abundant protons such as in residual H 2 O and OH groups of the solvent, the protic H in the (R 1 OR 2 ) 2 PtCl 2 H 2 could be readily hydrated, e.g. H 3 O + . The calculated 195 Pt chemical shift (-2705 ppm) of (R 1 OR 2 ) 2 Pt(0)Cl 2 (H 3 O + ) 2 (Entry 6, Supplementary Table 3) is in an excellent agreement with the experimental shift (-2680 ppm), suggesting that (R 1 OR 2 ) 2 Pt(0)Cl 2 (H 3 O + ) 2 may exist as a hydrated form of (R 1 OR 2 ) 2 PtCl 2 H 2 in the system. This calculated result is a further validation of protic H, rather than a hydride. The absence of J( 195 Pt-1 H) spin-spin couplings in the 195 Pt spectra may be explained by the rapid chemical exchanges between the protic H in (R 1 OR 2 ) 2 PtCl 2 H 2 and abundant labile protons of water and hydroxyl protons in the system. Furthermore, the protic H in the (R 1 OR 2 ) 2 PtCl 2 H 2 is consistent with a recent report that showed HCl acted as a weak ligand in the stabilization of Ir nanoparticles with the H in protic state 30 ." It is difficult to trust the NMR chemical shifts of 195 Pt predicted by DFT for proposed structures, when the prediction of the chemical shift of CO is -11.2 ppm, compared to real chemical shift at ca 185 ppm.

Response:
We simply used the absolute 13 C NMR shielding constant (-11.2 ppm) in the original version, instead of the chemical shift, for CO in the text. To address the reviewer's concern, we calculated the 13 C chemical shieldings for Si(Me) 4 (as a reference), the CO in gas phase, and the (R 1 OR 2 )PtCl 2 H 2 CO. The 13 C NMR chemical shifts were given in Entries 9-12,

Response:
We have stated that no solvent was used in the hydrosilylation reaction in the section of "General Procedure for the hydrosilylation of 1-octene catalyzed by Pt 1 @PDMS-PEG" in Methods.
Only an adequate amount of CDCl 3 was used during the 195 Pt NMR characterization of the 1-octene treated Pt 1 @PDMS-PEG to enhance the solubility. If Pt 1 @PDMS-PEG is oxidized by small amount of CHCl 3 , one would expect a change in the 195 Pt chemical shift to down field after 1-octene treated Pt 1 @PDMS-PEG. In fact, a 195 Pt NMR shift of -2706 ppm was observed n, very close in value to that of the Pt 1 @PDMS-PEG (-2680 ppm), indicating Pt (0) exists throughout the processes.  The blank test was carried out and reported in the revised manuscript. The blank test procedure was inserted into the "General Procedure for the hydrosilylation of 1-octene catalyzed by Pt 1 @PDMS-PEG" in Methods as follows "Blank test was run as follows:

To me a very important question is not addressed: what is the size and shape of the PDMS-PEG "vesicle" in the EtOH
hydrosilylation reaction was carried out with PDMS-PEG/ethanol-water as the catalyst (without Pt) following the same procedure as for the hydrosilylation of 1-octene catalyzed by Pt 1 @PDMS-PEG, No conversion of 1-octene was observed (Supplementary Table 6)." We also added the results and discussion in the revised manuscript "No 1-octene conversion was observed in a blank test using PDMS-PEG alone (Supplementary To further confirm no agglomeration of catalysts, the 195 Pt NMR spectrum of the Pt 1 @PDMS-PEG after the hydrosilylation reaction [n (1-octene): n(silane) = 5 : 3] was acquired. A possible explanation for why 1-octene still bound to Pt 1 center is added in the caption of Supplementary Figure 9 "Because of the excess amount of olefin (1-octene), the Pt 1 maintained the (olefin)(R 1 OR 2 )PtCl 2 H 2 state ( 195 Pt chemical shift), not (R 1 OR 2 ) 2 PtCl 2 H 2 state." In the revised Manuscirpt, we add the ratio of 1-octene to silane "n 1-octene : n silane = 5 :

3"
Lewis et al. 44 have reported the aggregation of Karstedt catalyst to form Pt nanoparticles with excessive silane at the end of the hydrosilylation reaction (after full conversion of olefin). In our work, 1-octene was added and mixed with the Pt 1 @PDMS-PEG first; see "General Procedure for the hydrosilylation of 1-octene catalyzed by Pt 1 @PDMS-PEG" in Methods. When silane was added first to a concentrated Pt 1 @PDMS-PEG solution, as inquired by the reviewer, the solution turned into a mud-like slurry with light brown color and was not suited for further characterization. The procedure is given in Methods: "After the hydrosilylation reaction between For the following reuses, the added amounts of 1-octene and (Me 3 SiO) 2 MeSiH were always 0.4488 g and 0.89 g, respectively. After each reaction, the amount of reaction mixture taken out for 1 H NMR from the upper solution after centrifugation was always 1.3388 g (equal to the amount of added reactants), and the remaining solution was kept at 0.6694 g for the subsequent hydrosilylation reaction." In conclusion for the synthesis/characterization of the Pt1 species, the data is not consistent with the proposed structure.
confusion. We hope the reviewer would find these revisions satisfactory. (The rewritten Introduction in the revised manuscript follows: "Supported single atoms have been prepared using various methods 1,2 . The reduced metal atoms are stabilized on solid surfaces through weak interaction 3,4 or by strong ligand coordination 5,6 . Weakly coordinated noble metal atoms are often capable of activating adsorbed reactants in catalysis 7,8 . Reduced single atoms tend to aggregate into nanoparticles in liquid media in the absence of strong ligand coordination 9 . Full and strong coordination is not desired for metal catalysts 10 . Therefore, synthesis of weakly coordinated and yet stable metal atoms in liquid media may pave a new path for facile preparation of highly dispersed metal catalysts by overcoming the long-standing challenge of metal aggregation. The readily removable ligands from the discrete metal atoms may also potentially enable atomically controllable synthesis of metallic materials by design.
Numerous surfactants with diverse functionalities have been used in the synthesis of metallic nanoparticles 11,12 or metal ion-surfactant complexes 13,14 in micro-emulsions system. added results as outlined in the revised manuscript allowed us to acquire a deeper understanding about Pt 1 @PDMS-PEG. We hope the reviewers and the editor would find these revisions satisfactory.