Thermodynamic Control in the Catalytic Insertion Polymerization of

---

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Article

Thermodynamic Control in the Catalytic Insertion Polymerization of Norbornenes as Rationale for the Lack of Reactivity of endo-Substituted Norbornenes Jonathan Potier, Basile Commarieu, Armand Soldera, and Jerome P. Claverie ACS Catal., Just Accepted Manuscript • Publication Date (Web): 23 May 2018 Downloaded from http://pubs.acs.org on May 23, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Thermodynamic Control in the Catalytic Insertion Polymerization of Norbornenes as Rationale for the Lack of Reactivity of endo-Substituted Norbornenes Jonathan Potier, Basile Commarieu, Armand Soldera, Jerome P. Claverie* Université de Sherbrooke, Quebec Center for Functional Materials, Dept of Chemistry, Sherbrooke, Qc, J1K2R1 Canada

ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT The catalytic insertion polymerization of substituted norbornenes (NBEs) leads to the formation of a family of polymers which combine extreme thermomechanical properties as well as unique optical and electronic properties. However, this reaction is marred by the lack of reactivity of endo-substituted monomers. It has long been assumed that these monomers chelate the metallic catalyst, leading to species which are inactive in polymerization. Here we examine the polymerization of cis-5-norbornene-2,3-dicarboxylic anhydride (so-called carbic anhydride, CA) with a naked cationic Pd catalyst. Although exo-CA can be polymerized, the polymerization of endo-CA stops after a single insertion. Surprisingly, no chelate is formed between the catalyst and endo-CA. Using DFT calculation, it is shown that while the insertion of exo-NBEs is exergonic, the insertion of two endo-CA in a row is endergonic. In this latter case, the enthalpy gain corresponding to the insertion of a double bond is not sufficient to overcome the entropic penalty associated with ligand binding. Thus, the different reactivity between endo and exo NBEs is thermodynamic in nature and it is not controlled by kinetic factors. Interestingly, thermodynamics is also the main factor controlling the stereochemistry of the chain. For CA polymerization, and even for unsubstituted NBE polymerization, the formation of r and m dyads are respectively exergonic and endergonic resulting in a polymer which is essentially disyndiotactic. Thus, this study demonstrates that thermodynamics can control the chemo and stereoselectivity of a catalytic polymerization.

KEYWORDS insertion polynorbornene, catalytic insertion, DFT, disyndiotactic polymer, stereospecific polymerization

ACS Paragon Plus Environment

Page 2 of 29

Page 3 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

INTRODUCTION The catalytic polymerization of olefins has proved to be a formidable tool for the efficient, cost-effective and sustainable conception of organic materials with an unsurpassed degree of control on microstructure.1-16 Recently, significant progress has been made to expand the scope of polymerization catalysts to functional monomers.17-31 Among all polyolefins, cyclic olefin polymers such as insertion polynorbornenes (PNBEs) offer outstanding thermo-mechanical properties which resemble those of costly ultra high-performance polymers.32-35 These polymers are utilized as 157 nm photoresists,36–38 CO2 separation membranes,39 or ultra-high Tg thermosets.40 These desirable properties originate from the chemically-inert saturated backbone and from the rigidity of the chain imposed by the restricted rotation between two neighboring monomers.41 These polymers are prepared by the catalytic polymerization of norbornene monomers (NBEs) which are in turn obtained by a Diels-Alder reaction between cracked dicyclopentadiene and a variety of dienophiles leading to the formation of two isomers, endo and exo. The endo isomer is known to deactivate the catalyst. Consequently, the polymerization of a mixture of endo and exo NBEs is order of magnitudes slower than the polymerization of pure exo NBEs.34,42 The absence of an efficient and facile separation process for these two isomers has prevented the development of these remarkable functional polymers on a large scale. Several hypotheses have been put forward to explain the catalyst deactivation by the endo NBE. Hennis et al proposed that the inserted endo-NBE forms a stable chelate with the catalyst via its functionality.43 In support of this hypothesis, the structure of a Pt chelate was determined by X-ray crystallography. Interestingly, this Pt complex, although inactive for polymerization, is analogous to the Pd active site formed during polymerization. This explanation was however questioned by the same research group when it was also observed that endo alkyl NBEs (which are unable to form chelates) polymerize more slowly than exo alkyl NBEs.44 To further muddy

ACS Paragon Plus Environment

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the waters, Nozaki et al demonstrated that a tBu3P-ligated cationic palladium complex polymerizes at the same rate endo and exo NBEs bearing a CO2Me substituent.45 However, we recently demonstrated that so-called ‘naked’ cationic Pd complexes are able to promote the conversion of endo NBEs to exo NBEs via retro Diels-Alder reaction.46 Thus, starting from a pure endo NBE bearing CO2Me groups, alternating endo-exo copolymers were obtained, indicating that two endo monomers cannot be inserted in a row. Additionally, it was spectroscopically observed that a chelate was formed between the cationic Pd catalyst and the endo face of the monomer by coordination to its double bond and its ester functionality. Therefore, the lack of reactivity of the endo isomer remains controversial, with no satisfactory explanations for this phenomenon. In this study, we examine the polymerization mechanism of a substituted NBE monomer, cis5-norbornene-2,3-dicarboxylic anhydride (so-called carbic anhydride, CA, Scheme 1). We first show that endo-CA does not form any chelate with the classical family of naked catalysts developed by Risse et al.35, however its polymerization stops after a single insertion in contrast to exo-CA which readily polymerizes (Scheme 1). We demonstrate that the reactivity is limited by thermodynamic factors: the insertion of endo-CA is surprisingly endergonic. In fact, the polymerization of CA, albeit strongly exothermic, is controlled by entropic factors, which not only affects the chemoselectivity but also the stereoselectivity of the polymerization. In catalytic polymerizations, as well as in most chain-growth polymerizations, thermodynamic considerations are usually circumvented to the profit of kinetic factors: this study demonstrates that it should not always be the case.

ACS Paragon Plus Environment

Page 4 of 29

Page 5 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

O 1

O O exo-CA

Pd

O O

RT, 1h CH3NO2

L

+ Pd

n

L

O

O

exo-2

O O PNBE

no polymer O O O

+ Pd

O

1

RT, 1h O CH3NO2

L

O

endo-CA L 1=

Pd

L L = CH3NO2

L + Pd

O

endo-2 O

SbF6-

O

O

O

O chelate, not formed

Scheme 1. Reactivity of catalyst 1 with CA

RESULTS AND DISCUSSION Reaction of endo and exo-CA with one equivalent of the cationic naked Pd complex 1 leads to the quantitative formation of endo and exo-2 respectively (Scheme 1). These complexes were fully characterized spectroscopically (Figure S1-8), but could not be isolated in the solid state due to the loss of coordinated nitromethane upon drying. Such a weak coordination is expected for so-called naked Pd catalysts. The 3J coupling value between H2, in α position to Pd and H3, in β position (see Figure S1 and S5 for atom numbering), is respectively 6.8Hz and 6.9 Hz in exo-2 and endo-2, indicating that H2 are H3 are cis to each other. In both complexes, H2 and H3 are only weakly coupled to the bridgehead protons (3J < 3 Hz) which indicates that H2 and H3 are in endo position (3J coupling between bridgehead and exo protons are typically comprised between 6 and 8 Hz).47 Therefore, the newly formed Pd-C and C-C bonds are on the exo face, indicating that monomer insertion occurs in a cis exo fashion.

ACS Paragon Plus Environment

5

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

When 2 (either exo or endo) is reacted with one equivalent of endo-CA, no reaction occurs at room temperature, and in particular, no chelate is formed (Scheme 1). In CD3NO2, the chemical shift of endo-CA olefinic protons is 6.28 ppm (Figure 1). When one equivalent of endo-CA is added to a solution of endo-2 or exo-2, the chemical shift of these olefinic protons does not change, hinting that no reaction occurs between endo-CA and endo or exo-2, even after 72 hours at room temperature. By contrast, when endo-cis-bicyclo[2,2,1]hept-5-ene-2,3-dicarboxylic acid (NBE(CO2H)2, obtained upon reacting endo CA with water, is added to a mixture of endo-2 and endo-CA, immediate reaction between NBE(CO2H)2 and endo-2 occurs, as shown by a downfield displacement of the resonance of NBE(CO2H)2 olefin protons from 6.36 to 6.75 ppm. This downfield chemical shift is the hallmark of a chelated double-bond (see reference 46 for other examples of chelated double-bonds in analogous complexes). By contrast, endo-CA does not form a chelate (Scheme 1 and Figure 1). Thus, as NBE monomer, CA is unique in the sense that it is a polar endo-substituted norbornene which is unable to form a chelate, probably because the geometrically constrained C=O groups cannot form σ bonds with a Pd atom located in the endo face.

ACS Paragon Plus Environment

6

Page 7 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Figure 1. 1H NMR of pure endo-CA, a mixture of endo-2 and endo-CA (1:1), a mixture of exo-2 and endo-CA (1:1), a mixture of endo-2, endo-CA and endo-NBE(COOH)2 (1:1:1.2)

Figure 2. Kinetic plot for the polymerization of endo-CA (blue curve) and exo-CA (red curve) by 1 in CD3NO2 ([CA]/[1] = 10, [CA] = 2.2x10-1 mol.L-1, T = 70°C). Despite the absence of chelate, the polymerization of endo-CA stops after ca 10% conversion, corresponding to a single insertion as the monomer:catalyst ratio is set to 10:1 (Figure 2). By contrast, the polymerization of exo-CA proceeds to completion. The first insertion proceeds in ca 40 minutes for both endo-CA and exo-CA, indicating that reactivity differences become prominent only after the first insertion. After the first insertion of endo-CA, the yield marginally increases from 10% to 17% in 500 minutes. As we previously demonstrated that retro DielsAlder reaction is taking place during the polymerization,46 this increase is due to the slow conversion of endo-CA into exo-CA.48 As soon as exo-CA is formed, it is inserted.

ACS Paragon Plus Environment

7

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

In order to clarify the salient reasons for the absence of reactivity of endo-CA after a single insertion, theoretical calculations using density functional theory (DFT) were performed. As a model for calculations, the cationic Pd complex 3 was used (Scheme 2), whereby formaldehyde was used as ancillary ligand instead of nitromethane. Experimentally, the polymerization of polar norbornenes such as CA is either performed in the absence of solvent46,49 or using a noncoordinating solvent such as nitromethane. Under these conditions, we expect the cationic Pd complex to be coordinated to pendant carbonyl groups present in the monomer or in the polymer. These carbonyl groups are modeled here by formaldehyde molecules which are σ-coordinated to Pd. In order to select the most appropriate computing method, B3LYP, M05, M05-2X, M06 and M06-2X functionals50,51 were compared, using either 6-31g(d,p) or 6-311g(d,p) basis sets (see Table S1 and S2 in ESI). The Gibbs free energy for the transformation 3 to 6 (Scheme 2) was found to be 0 (±1) kcal/mol with the B3LYP functional, -2.5 (±1) kcal/mol with M05 functional, and -9 (±2) kcal/mol with the three other functionals. The insertion of NBE within Pd-allyl is known to be a quantitative, irreversible reaction22 (as shown also above for reaction of 1 with CA). Therefore, the B3LYP functional, which predicted an endergonic reaction, was not further considered. The M05 functional was also not selected, because Pd-O bond lengths in 6 calculated with this functional (trans = 2.36 Å, cis = 2.32 Å) are much longer than bond lengths computed with all other functionals (trans: 2.27-2.32 Å, cis: 2.21-2.28 Å). In analogous cationic Pd complexes containing σ-coordinated ester and characterized by single crystal X-ray diffractometry, Pd-O bond lengths were found to be in the 2.13-2.25 Å range, in good agreement with our calculated structures.46 For our systems, no major differences between M05-2X, M06 and M06-2X functionals were found. We decided to use the M06 functional as it has recently been proved to be appropriate for consistent prediction of late-transition metal catalytic ACS Paragon Plus Environment

8

Page 9 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

pathways.52–56 Thermal and entropy contributions have been computed for each complex using the usual approximations (ideal gas, rigid rotor and harmonic oscillator from frequency calculations). Solvation was taken into account during optimization through the use of the polarizable continuum model,13,57 considering a nitromethane environment (dielectric constant of 35.9). For transition states, the presence of a single imaginary frequency was checked. Its corresponding eigenvector was associated with the bond being formed.

L L

L

L

Pd

L

L

Pd

Pd

Pd

Pd L 3 L = CH2O

+ Pd

6

L

7r

TS4-5

L

+ Pd

TS7-8m

7m + Pd

4

L

+ Pd

L

TS7-8r

5

L

+ Pd

L

8m + Pd

6

+ Pd

9m L L

8r

+ Pd

PdL3

L

10m

L

9r

L3Pd

10r

Scheme 2. Modeling the insertion polymerization of NBE

Before analyzing the insertion of CA, the insertion of NBE was first examined (Scheme 2, Figure 3). The first step involves the coordination of NBE on 3, resulting in the formation of NBE-coordinated complex 4 which is 9 kcal/mol more stable than 3. This decrease in energy reflects the soft metal preference for the norbornene double bond rather than for the hard carbonyl. Insertion of NBE in the Pd-allyl occurs via a standard 4-center planar transition state (TS4-5) which is located 24 kcal/mol above the olefin complex 4. The insertion occurs on the

ACS Paragon Plus Environment

9

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

exo face of NBE, leading to coordinatively unsaturated complex 5 (as observed experimentally with CA). Coordination of formaldehyde on the vacant position of 5 leads to complex 6. Alternative routes from 3 to 6 were considered. Insertion on the endo face of NBE did not converge to stable structures. An associative mechanism whereby NBE coordinates to 3, forming a pentacoordinated intermediate which then undergoes insertion was also examined, but once again did not lead to stable structures. In complex 5, the formaldehyde ligand is trans to the allylic double-bond, an expected manifestation of the trans-influence. The geometrical isomer of 5 having a formaldehyde in cis position was found to be higher in energy. The insertion of the first NBE unit from 3 to 6 leads to ∆H = -23.4 kcal/mol and ∆G = -10 kcal/mol. The large negative ∆H value is consistent with the strong exothermicity observed experimentally for NBE polymerization. This exothermicity stems from the transformation of the π bond into σ bond during polymerization, and from the release of strain when the C=C bond (133 pm) of NBE is transformed into C-C bond (155 pm in 6).

ACS Paragon Plus Environment

10

Page 11 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Figure 3. Enthalpies (top, A) and Gibbs free energies (bottom, B) for the polymerization of NBE (2 first insertions). See scheme 2 for structure numbering. Values can be found in Table S5. When the second monomer coordinates, the bridge of the incoming monomer can be on the same side as the bridge of the inserted monomer (complex 7m), or on the opposite side (complex 7r). Interconversion between these two isomers is facile, with less than 1 kcal/mol between both, and inversion barriers less than 4 kcal/mol (Figure S9 in ESI). After insertion, complexes with m (isotactic) and r (syndiotactic) dyads are obtained (Scheme 2). The transition state (TS7-8) of the

ACS Paragon Plus Environment

11

ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

r isomer is slightly lower than the one of the m isomer (∆∆H‡ = 1.5 kcal/mol and ∆∆G‡ = 1.7 kcal/mol). Thus, there is a slight kinetic preference for the formation of the r dyad. As the catalyst is a naked square planar Pd catalyst, this preference is not due to enantiomorphic site stereocontrol, but it is a priori an expression of chain-end stereocontrol.58 In chain-end control, the steric bulk of the last inserted unit (in red in Scheme 2) influences the placement of the incoming monomer (in blue) as well as its insertion barrier. However, this conclusion will be reevaluated below when considering the energy of insertion products 8, 9 and 10 which are respectively bound to one, two and three formaldehyde molecules. In 10, Pd is no more coordinated to the allyl double bond. The six-membered chelate in 6 is most likely stable but the eight-membered chelate in 9 could be more labile, and could open to form 10. The transformation from 9 to 10 restores the rotation around the bond linking the two NBE units, thus allowing the placement of the two adjacent monomer in a more favorable zig-zag conformation (dihedral angle of 175 and 151 deg. in 10m and 10r respectively). However, 10 is higher in energy than 9, once again because of the preference of Pd for the soft double bond over the formaldehyde. As expected, the enthalpy of the coordinatively unsaturated 8 is higher than for tetracoordinated 9. When considering the Gibbs energy, 8 (1 formaldehyde) is more stable than 9 (2 formaldehydes) and 9 is more stable than 10 (3 formaldehydes), indicating that the entropic penalty to coordinate a formaldehyde molecule is high. As a consequence, the stable configuration of the cationic Pd(II) catalyst is not the usual 16 e- tetracoordinated square planar complex, but a 14 e- tricoordinated complex, which is a rare configuration for a Pd(II) complex.59 The tricoordinated complex 8 adopts a distorted T shape whereby the allyl group and the formaldehyde are trans to each other. Why are tricoordinated complexes 8 more stable than tetracoordinated 9? It should of course be mentioned that the entropic cost is mostly due to the ACS Paragon Plus Environment

12

Page 13 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

loss of translational and rotational entropy of formaldehyde (from 8 to 9, T∆Stranslation = -10.7 kcal/mol and T∆Srotation = -4.6 kcal/mol for both m and r complexes). However, the loss of translation entropy is not sufficient to explain the unusual stability of 8 because the terms T∆Stranslation and T∆Srotation are expected to be of the same magnitude for any ligand binding reactions.60 Yet, 8 is more stable under its dissociated form whereas the vast majority of Pd(II) complexes are more stable under their associated (tetracoordinated) form. The influence of steric bulk in complexes 8, 9 and 10 was assessed through the use of topographic steric maps developed by Cavallo et al and calculated via a freely accessible online interface, SambVca 2.61 In such maps, the volume occupied by a ligand is mapped using an isocontour representation projected in a plane perpendicular to the metal-ligand axis. Positive values indicate that the considered ligand moiety is penetrating into the opposite coordination face. Large negative values indicate that the moiety is within its coordination face but far from the coordination site. This analysis also provides another descriptor, the %Vburied which is the percentage of the ligand volume contained within a sphere of 3.5 Å radius centered on the metal atom. High values indicate that the ligand impinge on adjacent coordination sites. In our cases, we chose the –NBENBE-allyl fragment (growing polymer chain) as ligand. Steric maps and %Vburied can be found in Figure S10 and S11. It should be first noted that the volume of the –NBE-NBE-allyl fragment is the same (179.6 Å3) in all six complexes 8, 9 and 10 (m or r). However, the %Vburied decreases from 51.1% (resp. 52.3%) to 50.4% (resp. 52.1%) when comparing 8r (resp. 8m) to 9r (resp. 9m) indicating that in 9, the –NBE-NBE-allyl fragment is globally further away from the coordination plane. However, the lengths of the Pd-C bonds are virtually unchanged between 8 and 9 (difference