University of Groningen Chemo and enantioselective addition of grignard reagents to ketones and enolizable ketimines Ortiz, Pablo

Chemo and enantioselective addition of grignard reagents to ketones and enolizable

ketimines

Ortiz, Pablo

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2017

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Ortiz, P. (2017). Chemo and enantioselective addition of grignard reagents to ketones and enolizable ketimines. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

If we knew what it was we were doing, it would not be called research, would it? Attributed to Albert Einstein

The most common solvents for the asymmetric catalysis using Grignard reagents

are Et2O, tBuOMe and CH2Cl2. While the solution structure of Grignard reagents in

the former is known, that is not the case for tBuOMe and CH2Cl2. This chapter

describes the use of 1D and 2D NMR spectroscopy to determine the structure of alkyl Grignard reagents in them. It has been found that at conditions commonly used in asymmetric catalysis (0.15-0.5 M, -78 °C) Grignard reagents in tBuOMe and

CH2Cl2 are monomeric.

Chapter 8:

The Solution Structure of Alkyl Grignard Reagents in

tBuOMe and CH

Cl

8.1. Introduction

Grignard reagents have played a central role in the development of organic chemistry. Their importance was recognized by scientific community by awarding

the Nobel Prize in Chemistry to Victor Grignard shortly after his discovery.[1] _Since

then, many other non-stabilized nucleophiles have been prepared, but organomagnesium compounds continue to be the reference organometallics,

cherished by both academia and industry.[2] Importantly, despite that more than a

century has passed since they were first prepared, innovation is still taking place:

among the most remarkable ones, their use in asymmetric reactions[3-6] and the

preparation of functionalized Grignard reagents and turbo-Grignards (complex

with LiCl).[7]

Organomagnesium halides, although generally represented as RMgX, have more complex structure in solution. Wilhelm Schlenk and Wilhelm Schlenk Jr. found that

in diethyl ether RMgX is in equilibrium with R2Mg and MgBr2,[8] in what is now

termed the “Schlenk equilibrium”. In the 1970s extensive work was carried out by Ashby and others using ebulloscopic, kinetic and spectroscopic measurements to determine the position of the Schlenk equilibrium and the aggregation state of

organomagnesium compounds in tetrahydrofuran (THF) and diethyl ether (Et2O).[9]

This led to the proposal of the “extended Schlenk equilibrium” to accommodate the new findings (Scheme 1). It was also recognized that the composition of the Grignard reagent depends on the nature of the organic group, the halide, temperature, solvent and concentration, being these last two the most important

ones. In THF, monomeric species (RMgX, and R2Mg+MgX2) were found to

predominate over a wide range of concentrations. Monomeric species were detected

as well in dilute Et2O solutions, but in this case they consisted mainly of RMgX.[9]

On the contrary, at concentrations above 0.3-0.5 M extensive association was observed. Consequently, it was postulated that with a stronger Lewis basic solvent the coordination with the magnesium atom is stronger and the halogen bridge is

less stable, leading to monomeric species (Scheme 1). And indeed, when Et3N was

used as solvent, only RMgX was detected. Since its introduction, NMR spectroscopy has been the preferred tool to study Grignard reagents, as it allows observing them

in solution. 1_{H NMR,}[10]25_{Mg NMR}[11] _{and DOSY experiments}[12,13a-c] _{have been used}

to determine the structure of Grignard reagents and their turbo variants in THF or

Et2O.

Scheme 1. Extended Schlenk equilibrium. (Solvent omitted) R = Alkyl or aryl. X = Halogen.

As mentioned above, one of the most important recent advances in the use of Grignard reagents has been their implementation in asymmetric synthesis, more

specifically their use for asymmetric conjugate addition[3] _{and allylic substitution}

(Scheme 2).[3a-b,4] _{The pioneering work of Lippard et al.}[3d-e] _{and van Koten et al.}[3f-g]

in the 90´s and the development of methodology by Feringa et al.[3a] _{and Alexakis et}

al.[3b] _{(with important contributions from others) in the beginning of this century has} made these reactions a common practice in both Organic Chemistry laboratories

and industry.[2,3c] _{More recently, Minnaard and Harutyunyan have developed the}

1,2-asymmetric addition of Grignard reagents to carbonyl compounds[3h,5a] and the

latter also to imines[5b-c] (Scheme 2, see also chapters 2 and 7).

Scheme 2. Overview of the most prominent uses of Grignard reagents in asymmetric catalysis: asymmetric allylic alkylation, conjugate addition and 1,2-addition. LG = Leaving group. PG = Protecting group.

Altogether, the use of Grignard reagents has allowed to access a variety of chiral, valuable molecules using readily available reagents. Interestingly, in almost all the reported examples of catalytic asymmetric additions using Grignard reagents, THF as a solvent gives poor results in terms of chemo and enantioselectivity. Instead,

Et2O, used in early reports, allowed better results, and later dichloromethane

8.1. Introduction

Grignard reagents have played a central role in the development of organic chemistry. Their importance was recognized by scientific community by awarding

the Nobel Prize in Chemistry to Victor Grignard shortly after his discovery.[1] _Since

then, many other non-stabilized nucleophiles have been prepared, but organomagnesium compounds continue to be the reference organometallics,

cherished by both academia and industry.[2] Importantly, despite that more than a

century has passed since they were first prepared, innovation is still taking place:

among the most remarkable ones, their use in asymmetric reactions[3-6] and the

preparation of functionalized Grignard reagents and turbo-Grignards (complex

with LiCl).[7]

Organomagnesium halides, although generally represented as RMgX, have more complex structure in solution. Wilhelm Schlenk and Wilhelm Schlenk Jr. found that

in diethyl ether RMgX is in equilibrium with R2Mg and MgBr2,[8] in what is now

termed the “Schlenk equilibrium”. In the 1970s extensive work was carried out by Ashby and others using ebulloscopic, kinetic and spectroscopic measurements to determine the position of the Schlenk equilibrium and the aggregation state of

organomagnesium compounds in tetrahydrofuran (THF) and diethyl ether (Et2O).[9]

This led to the proposal of the “extended Schlenk equilibrium” to accommodate the new findings (Scheme 1). It was also recognized that the composition of the Grignard reagent depends on the nature of the organic group, the halide, temperature, solvent and concentration, being these last two the most important

ones. In THF, monomeric species (RMgX, and R2Mg+MgX2) were found to

predominate over a wide range of concentrations. Monomeric species were detected

as well in dilute Et2O solutions, but in this case they consisted mainly of RMgX.[9]

On the contrary, at concentrations above 0.3-0.5 M extensive association was observed. Consequently, it was postulated that with a stronger Lewis basic solvent the coordination with the magnesium atom is stronger and the halogen bridge is

less stable, leading to monomeric species (Scheme 1). And indeed, when Et3N was

used as solvent, only RMgX was detected. Since its introduction, NMR spectroscopy has been the preferred tool to study Grignard reagents, as it allows observing them

in solution. 1_{H NMR,}[10]25_{Mg NMR}[11] _{and DOSY experiments}[12,13a-c] _{have been used}

to determine the structure of Grignard reagents and their turbo variants in THF or

Et2O.

Scheme 1. Extended Schlenk equilibrium. (Solvent omitted) R = Alkyl or aryl. X = Halogen.

As mentioned above, one of the most important recent advances in the use of Grignard reagents has been their implementation in asymmetric synthesis, more

specifically their use for asymmetric conjugate addition[3] _{and allylic substitution}

(Scheme 2).[3a-b,4] _{The pioneering work of Lippard et al.}[3d-e] _{and van Koten et al.}[3f-g]

in the 90´s and the development of methodology by Feringa et al.[3a] _{and Alexakis et}

al.[3b] _{(with important contributions from others) in the beginning of this century has} made these reactions a common practice in both Organic Chemistry laboratories

and industry.[2,3c] _{More recently, Minnaard and Harutyunyan have developed the}

1,2-asymmetric addition of Grignard reagents to carbonyl compounds[3h,5a] and the

latter also to imines[5b-c] (Scheme 2, see also chapters 2 and 7).

Scheme 2. Overview of the most prominent uses of Grignard reagents in asymmetric catalysis: asymmetric allylic alkylation, conjugate addition and 1,2-addition. LG = Leaving group. PG = Protecting group.

Altogether, the use of Grignard reagents has allowed to access a variety of chiral, valuable molecules using readily available reagents. Interestingly, in almost all the reported examples of catalytic asymmetric additions using Grignard reagents, THF as a solvent gives poor results in terms of chemo and enantioselectivity. Instead,

Et2O, used in early reports, allowed better results, and later dichloromethane

standard ones. In contrast with THF and Et2O, little or nothing is known about the

structural composition of Grignard reagents in CH2Cl2 and tBuOMe.

This chapter describes the combination of 1D and 2D NMR spectroscopy techniques to determine the structural composition of Grignard reagents in the reaction conditions commonly used in asymmetric catalysis, which typically are -78 ±C and

0.15 M concentration of the Grignard reagent.[3a-c,3h-i,4,5] _{(The range is between 0.07-0.8}

M, but most examples are around 0.15 M). Regarding the organomagnesium halide, we have selected MeMgBr and EtMgBr because alkyl magnesium bromides are the most employed Grignard reagents and among those the short length of the alkyl chain reduces the error when using diffusion to determine the molecular weight (MW).

8.2. Results and discussion 8.2.1 Schlenk equilibrium

Grignard reagents are generally formed in ethereal solvents that facilitates their formation due their Lewis basicity. However, once formed, Grignard reagents can

be diluted in other solvents, e.g. CH2Cl2. Throughout this chapter, when referred to

Grignard reagents in CH2Cl2, keep in mind that Grignard reagents are prepared or

purchased in Et2O and then diluted in CH2Cl2. Alkyl Grignard reagents do not react

with the relatively acidic protons of the solvent, and can be stored at room temperature for days without decomposition (tested in this thesis). To start with the

determination of the structure of Grignarnd reagents in CH2Cl2 we first studied the

Schlenk equilibrium. The 1H NMR of 0.15 M solution of MeMgBr and EtMgBr, both

in Et2O and CD2Cl2, at 25 and -78 °C, only showed the presence of

organomagnesium halide RMgBr. By separately preparing and recording the spectra of dialkylmagnesium species, which had different chemical shift to that of organomagnesium halides (Figure 1), we have verified their absence.

Figure 1: Difference in the chemical shift between 0.15 M RMgBr and R2Mg (left R = Me, right

R = Et) at -78 °C.

1: 0.15 M solution of 3 M RMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570 μL).

2: 0.15 M solution of 1 M R2Mg in Et2O (90 μL) dissolved in CD2Cl2 (510 μL).

3: 0.15 M solution of 3 M RMgBr in Et2O (30 μL) dissolved in Et2O(570 μL).

4: 0.15 M solution of 1 M R2Mg in Et2O (90 μL) dissolved in Et2O(510 μL).

8.2.2. Intereaction between Et2O and Mg atom

Ruling out the presence of dialkylmagnesium compounds reduced the number of possible structures of Grignard reagents that we might encounter in solution (Scheme 1). Nevertheless, the options were still broad, and before proceeding to diffusion measurements we decided to determine the extent of association of the reagent with the solvent. X-ray crystallography data starting from the 1960s have shown that Grignard reagents crystallize out with ethereal solvents in the crystal

unit.[14] _{However, the extent of the interaction in solution is not known, so NOESY}

experiments were conducted. They showed positive cross peaks between the Grignard and diethyl ether at 25 °C and negative cross peaks at -78 °C (red peaks,

standard ones. In contrast with THF and Et2O, little or nothing is known about the

structural composition of Grignard reagents in CH2Cl2 and tBuOMe.

This chapter describes the combination of 1D and 2D NMR spectroscopy techniques to determine the structural composition of Grignard reagents in the reaction conditions commonly used in asymmetric catalysis, which typically are -78 ±C and

0.15 M concentration of the Grignard reagent.[3a-c,3h-i,4,5] _{(The range is between 0.07-0.8}

M, but most examples are around 0.15 M). Regarding the organomagnesium halide, we have selected MeMgBr and EtMgBr because alkyl magnesium bromides are the most employed Grignard reagents and among those the short length of the alkyl chain reduces the error when using diffusion to determine the molecular weight (MW).

8.2. Results and discussion 8.2.1 Schlenk equilibrium

Grignard reagents are generally formed in ethereal solvents that facilitates their formation due their Lewis basicity. However, once formed, Grignard reagents can

be diluted in other solvents, e.g. CH2Cl2. Throughout this chapter, when referred to

Grignard reagents in CH2Cl2, keep in mind that Grignard reagents are prepared or

purchased in Et2O and then diluted in CH2Cl2. Alkyl Grignard reagents do not react

with the relatively acidic protons of the solvent, and can be stored at room temperature for days without decomposition (tested in this thesis). To start with the

determination of the structure of Grignarnd reagents in CH2Cl2 we first studied the

Schlenk equilibrium. The 1H NMR of 0.15 M solution of MeMgBr and EtMgBr, both

in Et2O and CD2Cl2, at 25 and -78 °C, only showed the presence of

organomagnesium halide RMgBr. By separately preparing and recording the spectra of dialkylmagnesium species, which had different chemical shift to that of organomagnesium halides (Figure 1), we have verified their absence.

Figure 1: Difference in the chemical shift between 0.15 M RMgBr and R2Mg (left R = Me, right

R = Et) at -78 °C.

1: 0.15 M solution of 3 M RMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570 μL).

2: 0.15 M solution of 1 M R2Mg in Et2O (90 μL) dissolved in CD2Cl2 (510 μL).

3: 0.15 M solution of 3 M RMgBr in Et2O (30 μL) dissolved in Et2O(570 μL).

4: 0.15 M solution of 1 M R2Mg in Et2O (90 μL) dissolved in Et2O(510 μL).

8.2.2. Intereaction between Et2O and Mg atom

Ruling out the presence of dialkylmagnesium compounds reduced the number of possible structures of Grignard reagents that we might encounter in solution (Scheme 1). Nevertheless, the options were still broad, and before proceeding to diffusion measurements we decided to determine the extent of association of the reagent with the solvent. X-ray crystallography data starting from the 1960s have shown that Grignard reagents crystallize out with ethereal solvents in the crystal

unit.[14] _{However, the extent of the interaction in solution is not known, so NOESY}

experiments were conducted. They showed positive cross peaks between the Grignard and diethyl ether at 25 °C and negative cross peaks at -78 °C (red peaks,

takes places, resulting in EXSY (Exchange spectroscopy) cross peaks, while at low temperature the cross peaks are from the NOE (blue peaks, Figure 2, right).

Figure 2. NOESY of 0.15 M solution of 3 M MeMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570

μL) at 25 °C (left) and at -78 °C (right). The result for EtMgBr was analogous.

To gather further insight we measured the 1D-NOESY at different temperatures (from 25 °C to -78 °C, Figure 3) and saw the change from positive to negative, with the inversion taking place around -25 °C (Figure 3). Further proof of the close

association between Et2O and the Mg atom could be extracted from the variation of

the chemical shift of Et2O signals. Et2O bound to Grignard reagent appeared at

higher chemical shift than free Et2O because it donates electrons to the Mg atom

(Figure 4). These results are in agreement with the tetracoordinate Mg atom found

both in crystals[14] _{and in computational studies.}[15] _{The results also indicate that the}

nature of the coordination is temperature-dependent. Moreover, a significant effect

in T2 relaxation is observed for Et2O when a Grignard reagent is present. It drops an

order of magnitude compared to free Et2O (no Grignard reagent), while T1 is

lowered by a factor of 2-3. This is observed at 25 °C as well as at -78 °C. Similar

trend was found for both 0.15 M and 0.5 M concentrations as well as when Et2O was

partially removed. All together, T1/T2 experiments point in the direction of very fast

exchange processes between Et2O and the Grignard reagent.

-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 f1 (ppm) E t₂O T M S CH3C H2M gB r C H_{3CH2M gBr} 25 °C 0°C -25° C -5 0°C -7 8°C

Figure 3. Changes in the 1D-NOE experiment with temperature. Top: 1_{H NMR of 0.15 M}

solution of 3 M EtMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570 μL). Stacked:

1DNOESY (irradiation at the methylene carbon of the Grignard reagent) at 25, 0, 25, -50, and -78 °C.

Figure 4. 1_{H NMR sprectra showing changes in chemical shift of Et2O peaks due to association}

with MeMgBr (left) and EtMgBr (right) at -78 °C. Top: Et2O (30 μL) dissolved in of

CD2Cl2 (570 μL). Middle: 0.15 M solution of commercial 3 M MeMgBr in Et2O (30

μL) dissolved in CD2Cl2 (570 μL). Bottom: 0.15 M solution of commercial 3 M

MeMgBr in Et2O (30 μL), Et2O removed and then dissolved in CD2Cl2 (570 μL).

(Note: more Et2O could be evaporated but if the Grignard was left as a solid it did

takes places, resulting in EXSY (Exchange spectroscopy) cross peaks, while at low temperature the cross peaks are from the NOE (blue peaks, Figure 2, right).

Figure 2. NOESY of 0.15 M solution of 3 M MeMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570

μL) at 25 °C (left) and at -78 °C (right). The result for EtMgBr was analogous.

To gather further insight we measured the 1D-NOESY at different temperatures (from 25 °C to -78 °C, Figure 3) and saw the change from positive to negative, with the inversion taking place around -25 °C (Figure 3). Further proof of the close

association between Et2O and the Mg atom could be extracted from the variation of

the chemical shift of Et2O signals. Et2O bound to Grignard reagent appeared at

higher chemical shift than free Et2O because it donates electrons to the Mg atom

(Figure 4). These results are in agreement with the tetracoordinate Mg atom found

both in crystals[14] _{and in computational studies.}[15] _{The results also indicate that the}

nature of the coordination is temperature-dependent. Moreover, a significant effect

in T2 relaxation is observed for Et2O when a Grignard reagent is present. It drops an

order of magnitude compared to free Et2O (no Grignard reagent), while T1 is

lowered by a factor of 2-3. This is observed at 25 °C as well as at -78 °C. Similar

trend was found for both 0.15 M and 0.5 M concentrations as well as when Et2O was

partially removed. All together, T1/T2 experiments point in the direction of very fast

exchange processes between Et2O and the Grignard reagent.

-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 f1 (ppm) E t₂O T M S CH3C H2M gB r C H_{3CH2M gBr} 25 °C 0°C -25° C -5 0°C -7 8°C

Figure 3. Changes in the 1D-NOE experiment with temperature. Top: 1_{H NMR of 0.15 M}

solution of 3 M EtMgBr in Et2O (30 μL) dissolved in CD2Cl2 (570 μL). Stacked:

1DNOESY (irradiation at the methylene carbon of the Grignard reagent) at 25, 0, 25, -50, and -78 °C.

Figure 4. 1_{H NMR sprectra showing changes in chemical shift of Et2O peaks due to association}

with MeMgBr (left) and EtMgBr (right) at -78 °C. Top: Et2O (30 μL) dissolved in of

CD2Cl2 (570 μL). Middle: 0.15 M solution of commercial 3 M MeMgBr in Et2O (30

μL) dissolved in CD2Cl2 (570 μL). Bottom: 0.15 M solution of commercial 3 M

MeMgBr in Et2O (30 μL), Et2O removed and then dissolved in CD2Cl2 (570 μL).

(Note: more Et2O could be evaporated but if the Grignard was left as a solid it did

8.2.3. The solution structure of Grignard reagents

In order to determine the association degree of Grignard reagents in solution we conducted DOSY experiments and determined the MW from their diffusion

coefficients.[16] _{We used a set of external calibration curves}[13d,e] _{to get normalized}

diffusion coefficients by using two internal references, tetramethylsilane (TMS) and 1,2,3,4-tetraphenylnaphthalene (TPhN) (See Experimental section 8.4.3.)

We first validated the method by determining the aggregation state of MeMgBr and

EtMgBr in Et2O at 25 °C, which are known to be monomeric at low concentrations

and of greater aggregation at higher concentrations.[9] To our delight, this is

precisely what we observed. The determined MW of a 3 M commercial solution of

methyl- and ethylmagnesium bromide in Et2O was close to that calculated for the

dimeric structure (a fractional number of association is possible, as previous ebulloscopic methods have shown) while the determined MW of a 0.15 M solution fitted very well with the calculated MW of the monomer (Graph 1, blue dots). Once confident with the reliability of the method we moved to conditions used in asymmetric catalysis by cooling down the samples to -78 °C. We did not observe any change in the association, so, most likely, monomers are the competent species

in catalysis (0.15 M, -78 °C) when Et2O is the solvent.

8.2.3.1 The solution structure of of MeMgBr and EtMgBr in CD2Cl2

Then we proceed to analyze the composition of the Grignard reagents in CD2Cl2

(Graph 1, red squares). 0.15 M solutions (diluted from Grignard reagent in Et2O)

both at 25 and -78 °C contained Grignard reagents as monomers, and removing

Et2O did not change it. 0.5 M solutions (upper range of concentration in catalysis)

were measured too and the determined MW fitted with the monomeric structure as well. Mg Br Br Et2O Mg R R OEt2 Monomeric Dimeric Br Et₂O Mg R OEt2 MW = 267 (R = Me) = 281 (R = Et) MW = 387 (R = Me) = 415 (R = Et)

Graph 1: Determined MW of MeMgBr (left) and EtMgBr (right) in Et2O and CD2Cl2 at

different concentrations and temperatures.

From the data obtained one can conclude that Grignard reagents diluted in CH2Cl2,

as typically used in asymmetric catalysis (<0.5 M, -78 °C) are monomeric, as they are

in Et2O. This means that the agreagation state alone is not responsible for variations

in the reaction outcomes in terms of chemo and enantioselectivities when the

catalytic asymmetric reactions are carried out in CH2Cl2 and Et2O. Intriguingly, we

observed that diluting the commercial Grignard reagents (3 M in Et2O) in CH2Cl2

resulted in a clear solution whereas in Et2O, even if seems counterintuitive, it was

more likely to form a suspension, especially at low temperatures. In fact, all the

attempts to prepare 0.5 M dilution of MeMgBr and EtMgBr in Et2O at -78 °C failed

because precipitation took place when cooling down.[17] The Schlenk equilibrium is

influenced by the amount of coordinating solvent (Scheme 3, S = Et2O).

Consequently, the dilution of Grignard reagents in a non-coordinating solvent (such

as CH2Cl2) is likely to have a very different impact compared with dilution in a

8.2.3. The solution structure of Grignard reagents

In order to determine the association degree of Grignard reagents in solution we conducted DOSY experiments and determined the MW from their diffusion

coefficients.[16] _{We used a set of external calibration curves}[13d,e] _{to get normalized}

diffusion coefficients by using two internal references, tetramethylsilane (TMS) and 1,2,3,4-tetraphenylnaphthalene (TPhN) (See Experimental section 8.4.3.)

We first validated the method by determining the aggregation state of MeMgBr and

EtMgBr in Et2O at 25 °C, which are known to be monomeric at low concentrations

and of greater aggregation at higher concentrations.[9] To our delight, this is

precisely what we observed. The determined MW of a 3 M commercial solution of

methyl- and ethylmagnesium bromide in Et2O was close to that calculated for the

dimeric structure (a fractional number of association is possible, as previous ebulloscopic methods have shown) while the determined MW of a 0.15 M solution fitted very well with the calculated MW of the monomer (Graph 1, blue dots). Once confident with the reliability of the method we moved to conditions used in asymmetric catalysis by cooling down the samples to -78 °C. We did not observe any change in the association, so, most likely, monomers are the competent species

in catalysis (0.15 M, -78 °C) when Et2O is the solvent.

8.2.3.1 The solution structure of of MeMgBr and EtMgBr in CD2Cl2

Then we proceed to analyze the composition of the Grignard reagents in CD2Cl2

(Graph 1, red squares). 0.15 M solutions (diluted from Grignard reagent in Et2O)

both at 25 and -78 °C contained Grignard reagents as monomers, and removing

Et2O did not change it. 0.5 M solutions (upper range of concentration in catalysis)

were measured too and the determined MW fitted with the monomeric structure as well. Mg Br Br Et2O Mg R R OEt2 Monomeric Dimeric Br Et₂O Mg R OEt2 MW = 267 (R = Me) = 281 (R = Et) MW = 387 (R = Me) = 415 (R = Et)

Graph 1: Determined MW of MeMgBr (left) and EtMgBr (right) in Et2O and CD2Cl2 at

different concentrations and temperatures.

From the data obtained one can conclude that Grignard reagents diluted in CH2Cl2,

as typically used in asymmetric catalysis (<0.5 M, -78 °C) are monomeric, as they are

in Et2O. This means that the agreagation state alone is not responsible for variations

in the reaction outcomes in terms of chemo and enantioselectivities when the

catalytic asymmetric reactions are carried out in CH2Cl2 and Et2O. Intriguingly, we

observed that diluting the commercial Grignard reagents (3 M in Et2O) in CH2Cl2

resulted in a clear solution whereas in Et2O, even if seems counterintuitive, it was

more likely to form a suspension, especially at low temperatures. In fact, all the

attempts to prepare 0.5 M dilution of MeMgBr and EtMgBr in Et2O at -78 °C failed

because precipitation took place when cooling down.[17] The Schlenk equilibrium is

influenced by the amount of coordinating solvent (Scheme 3, S = Et2O).

Consequently, the dilution of Grignard reagents in a non-coordinating solvent (such

as CH2Cl2) is likely to have a very different impact compared with dilution in a

dialkylmagnesium compounds, as these were perfectly soluble in Et2O and 0.5 M solutions could be prepared without precipitation.

Scheme 3. Solvent take-in and expel between species in the Schlenk equilibrium.

As already mentioned, in diluted Et2O solutions the Schlenk equilibrium was

measured to lie heavily on the side of RMgBr, with small amount of R2Mg and

MgBr2.[9a] However, it was shown that dialkylmagnesium compounds react at least

10 times faster than the corresponding organomagnesium halides.[9c] The presence

of Mg2+ as well as the halogen atoms has been proven to be essential for successful

asymmetric conjugate addition of Grignard reagents.[18] The worse results obtained

when using THF in these reactions have been explained by the Schlenk equilibrium being shifted to the dialkylmagnesium compounds, which were shown to have a

deleterious effect on the catalysis.[18] In order to test it, we titrated the Grignard

reagents with THF, and we found a strong effect of the solvent on the outcome. In

CH2Cl2 gradual shift from EtMgBr to Et2Mg was observed as the amount of THF

increased (Figure 5, left). It is worth mentioning that even after addition of 7 equivalents of THF the solution remained clear. On the contrary, when using a

solution of Grignard reagents in Et2O, upon addition of 1 equivalent of THF the

equilibrium was shifted completely to the dialkylmagnesium compound and extensive precipitation was observed (Figure 5, right). Very recently it has been shown using computational methods that the solvent has a crucial role in assisting the transition from organomagenesium halides to dialkylmagnesium species in the

Schlenk equilibrium.[19] _{Our observations and experiments point to the fact that the}

transition to the dialkylmagnesium compounds is more difficult in CH2Cl2 as

compared with diethyl ether. Consequently, more RMgBr could get involved in the

catalytic pathway, minimizing the non-catalyzed reaction with R2Mg. This might be

of special importance when dealing with slow reactions, while fast catalytic systems would be less sensitive to this effect.

Figure 5. Titration of EtMgBr using increasing amounts of THF at -78 °C. Grignard dissolved

in CD2Cl2 on the left and in Et2O on the right. Top: EtMgBr. Stacked: increasing

amount of THF added to the solution. Bottom: Et2Mg (separately prepared) + THF.

Experiments with MeMgBr gave analogous results.

8.2.3.2 The solution structure of of MeMgBr and EtMgBr in tBuOMe

Next we studied the solution structure of Grignard reagents in tBuOMe, which

compared to Et2O, is a poorer Lewis base, and structurally more bulky. As a

consequence, the coordination with the organomagnesium compounds is expected to be weaker, which could rationalize why it is more difficult to form Grignard reagents in tBuOMe, or sometimes, impossible. Nevertheless, especially in industry,

it is of high interest to be able to replace Et2O by tBuOMe due to its higher boiling

point, and therefore, safer profile. In academia, the choice of tBuOMe is dictated by

the better results obtained in some cases with it. Contrary to CH2Cl2, Grignard

reagents can be prepared in tBuOMe, being this then the only solvent of the reaction. We started by determining the MW of the commercial 1 M solution of EtMgBr in tBuOMe. It is worth noting that it is not a coincidence that 1 M concentration is the commercial molarity. As already mentioned, making Grignard reagents in tBuOMe is more difficult than in ether, and getting concentrations such

dialkylmagnesium compounds, as these were perfectly soluble in Et2O and 0.5 M solutions could be prepared without precipitation.

Scheme 3. Solvent take-in and expel between species in the Schlenk equilibrium.

As already mentioned, in diluted Et2O solutions the Schlenk equilibrium was

measured to lie heavily on the side of RMgBr, with small amount of R2Mg and

MgBr2.[9a] However, it was shown that dialkylmagnesium compounds react at least

10 times faster than the corresponding organomagnesium halides.[9c] The presence

of Mg2+ as well as the halogen atoms has been proven to be essential for successful

asymmetric conjugate addition of Grignard reagents.[18] The worse results obtained

when using THF in these reactions have been explained by the Schlenk equilibrium being shifted to the dialkylmagnesium compounds, which were shown to have a

deleterious effect on the catalysis.[18] In order to test it, we titrated the Grignard

reagents with THF, and we found a strong effect of the solvent on the outcome. In

CH2Cl2 gradual shift from EtMgBr to Et2Mg was observed as the amount of THF

increased (Figure 5, left). It is worth mentioning that even after addition of 7 equivalents of THF the solution remained clear. On the contrary, when using a

solution of Grignard reagents in Et2O, upon addition of 1 equivalent of THF the

equilibrium was shifted completely to the dialkylmagnesium compound and extensive precipitation was observed (Figure 5, right). Very recently it has been shown using computational methods that the solvent has a crucial role in assisting the transition from organomagenesium halides to dialkylmagnesium species in the

Schlenk equilibrium.[19] _{Our observations and experiments point to the fact that the}

transition to the dialkylmagnesium compounds is more difficult in CH2Cl2 as

compared with diethyl ether. Consequently, more RMgBr could get involved in the

catalytic pathway, minimizing the non-catalyzed reaction with R2Mg. This might be

of special importance when dealing with slow reactions, while fast catalytic systems would be less sensitive to this effect.

Figure 5. Titration of EtMgBr using increasing amounts of THF at -78 °C. Grignard dissolved

in CD2Cl2 on the left and in Et2O on the right. Top: EtMgBr. Stacked: increasing

amount of THF added to the solution. Bottom: Et2Mg (separately prepared) + THF.

Experiments with MeMgBr gave analogous results.

8.2.3.2 The solution structure of of MeMgBr and EtMgBr in tBuOMe

Next we studied the solution structure of Grignard reagents in tBuOMe, which

compared to Et2O, is a poorer Lewis base, and structurally more bulky. As a

consequence, the coordination with the organomagnesium compounds is expected to be weaker, which could rationalize why it is more difficult to form Grignard reagents in tBuOMe, or sometimes, impossible. Nevertheless, especially in industry,

it is of high interest to be able to replace Et2O by tBuOMe due to its higher boiling

point, and therefore, safer profile. In academia, the choice of tBuOMe is dictated by

the better results obtained in some cases with it. Contrary to CH2Cl2, Grignard

reagents can be prepared in tBuOMe, being this then the only solvent of the reaction. We started by determining the MW of the commercial 1 M solution of EtMgBr in tBuOMe. It is worth noting that it is not a coincidence that 1 M concentration is the commercial molarity. As already mentioned, making Grignard reagents in tBuOMe is more difficult than in ether, and getting concentrations such

as 3 M, common for Et2O or THF, is if not impossible, uncommon in tBuOMe. 1 M EtMgBr in tBuOMe at 25 °C was found to be dimeric (Graph 2, blue dot, left). Interestingly, dilution by half and to 0.15 M both resulted in detection of monomeric species at 25 °C as well as at -78 °C (Graph 2, blue dots). Again, conditions typical of asymmetric catalysis (0.15-0.5 M, -78 °C) showed monomeric Grignard reagent as the prevalent species when they are dissolved in tBuOMe.

Graph 2: Determined MW of EtMgBr in tBuOMe at different concentrations and temperature.

Although some Grignard reagents can be prepared in tBuOMe, only EtMgBr is commercially available (from Sigma-Aldrich). On the contrary, dozens are available

in Et2O. Hence, if the reaction tolerates a small amount Et2O, it is common practice

to dilute the commercially available Grignard reagent (prepared in Et2O) in tBuOMe

when the latter is a better solvent. Besides, sometimes it might be easier to prepare a

Grignard reagent in Et2O and dilute in tBuOMe rather than preparing it in tBuOMe.

To emulate this situation we diluted commercial Grignard reagents (3 M in Et2O) in

tBuOMe to a final concentration of 0.15 M. Not surprisingly considering the

previous measurements, the determined structure was monomeric both at 25 and at -78 °C.

However, contrary to the previous cases there were two competing solvents in this

setup. What is then the fate of the original Et2O? Does it stay bound to the Grignard

reagent due to its higher Lewis basicity or is it displaced by the more abundant

tBuOMe? Although the determined MW fits pretty well with the calculated

structure solvated by tBuOMe (Graph 2, red dots) the low difference in weight

between tBuOMe and Et2O (14 g/mol), together with the possibility of having both

coordinated, makes it risky to state which one is bound to the magnesium atom. In

order to shed light on it we conducted 1D-NOESY experiments.[20] _{We prepared a}

0.15 M solution of EtMgBr in tBuOMe from 3 M in Et2O (Figure 6, top), cooled down

to -78 °C and upon irradiation of the methylene carbon of EtMgBr a positive NOE effect was seen with the tBuOMe. Under the signal corresponding to the tert-butyl group of tBuOMe there is also the methyl peak of the Grignard, which would give negative NOE, and thus the shape of the signal (Figure 6, middle). On the other

hand, if the methylene peak of the Et2O was irradiated no NOE effect was seen

toward the Grignard reagents, but a small positive NOE to the tBuOMe was observed (Figure 6, bottom). This has to come from exchange at Mg atom of

tBuOMe and Et2O. Together, these experiments suggests that exchange of tBuOMe

and Et2O takes place at the Mg atom, but because there number of tBuOMe

molecules is higher than those of Et2O, the former is preferentially bound. This

could explain why sometimes tBuOMe is better solvent than diethyl ether: by replacing it at the magnesium atom the resulting species is less nucleophilic than the former (due to the lower Lewis basicity of tBuOMe), and therefore, it could slow down the uncatalyzed (racemic) reaction. On the other hand, the bulkier tBuOMe was shown to improve the chemoselectivity towards the addition to ketones

compared with Et2O, which yielded more reduction product.[21] These two effects,

electronic and steric, apply to the organomagensium halide species that is in solution. However, changing the solvent would most likely have an influence too on the whole (extended) Schlenk equilibrium (Scheme 3) as suggested above when

as 3 M, common for Et2O or THF, is if not impossible, uncommon in tBuOMe. 1 M EtMgBr in tBuOMe at 25 °C was found to be dimeric (Graph 2, blue dot, left). Interestingly, dilution by half and to 0.15 M both resulted in detection of monomeric species at 25 °C as well as at -78 °C (Graph 2, blue dots). Again, conditions typical of asymmetric catalysis (0.15-0.5 M, -78 °C) showed monomeric Grignard reagent as the prevalent species when they are dissolved in tBuOMe.

Graph 2: Determined MW of EtMgBr in tBuOMe at different concentrations and temperature.

Although some Grignard reagents can be prepared in tBuOMe, only EtMgBr is commercially available (from Sigma-Aldrich). On the contrary, dozens are available

in Et2O. Hence, if the reaction tolerates a small amount Et2O, it is common practice

to dilute the commercially available Grignard reagent (prepared in Et2O) in tBuOMe

when the latter is a better solvent. Besides, sometimes it might be easier to prepare a

Grignard reagent in Et2O and dilute in tBuOMe rather than preparing it in tBuOMe.

To emulate this situation we diluted commercial Grignard reagents (3 M in Et2O) in

tBuOMe to a final concentration of 0.15 M. Not surprisingly considering the

previous measurements, the determined structure was monomeric both at 25 and at -78 °C.

However, contrary to the previous cases there were two competing solvents in this

setup. What is then the fate of the original Et2O? Does it stay bound to the Grignard

reagent due to its higher Lewis basicity or is it displaced by the more abundant

tBuOMe? Although the determined MW fits pretty well with the calculated

structure solvated by tBuOMe (Graph 2, red dots) the low difference in weight

between tBuOMe and Et2O (14 g/mol), together with the possibility of having both

coordinated, makes it risky to state which one is bound to the magnesium atom. In

order to shed light on it we conducted 1D-NOESY experiments.[20] _{We prepared a}

0.15 M solution of EtMgBr in tBuOMe from 3 M in Et2O (Figure 6, top), cooled down

to -78 °C and upon irradiation of the methylene carbon of EtMgBr a positive NOE effect was seen with the tBuOMe. Under the signal corresponding to the tert-butyl group of tBuOMe there is also the methyl peak of the Grignard, which would give negative NOE, and thus the shape of the signal (Figure 6, middle). On the other

hand, if the methylene peak of the Et2O was irradiated no NOE effect was seen

toward the Grignard reagents, but a small positive NOE to the tBuOMe was observed (Figure 6, bottom). This has to come from exchange at Mg atom of

tBuOMe and Et2O. Together, these experiments suggests that exchange of tBuOMe

and Et2O takes place at the Mg atom, but because there number of tBuOMe

molecules is higher than those of Et2O, the former is preferentially bound. This

could explain why sometimes tBuOMe is better solvent than diethyl ether: by replacing it at the magnesium atom the resulting species is less nucleophilic than the former (due to the lower Lewis basicity of tBuOMe), and therefore, it could slow down the uncatalyzed (racemic) reaction. On the other hand, the bulkier tBuOMe was shown to improve the chemoselectivity towards the addition to ketones

compared with Et2O, which yielded more reduction product.[21] These two effects,

electronic and steric, apply to the organomagensium halide species that is in solution. However, changing the solvent would most likely have an influence too on the whole (extended) Schlenk equilibrium (Scheme 3) as suggested above when

Figure 6. Solvent exchange when tBuOMe and Et2O are present in the solution of a Grignard

reagent at -78 °C. Top: 0.15 M solution of 3 M EtMgBr in Et2O (30 μL) dissolved in

tBuOMe(570 μL). Middle: 1D-NOE effect upon irradiation of the methylene peak of

EtMgBr. Bottom: 1D-NOE effect upon irradiation of the methylene peak of Et2O.

Experiments with MeMgBr gave analogous results.

8.3. Conclusion

In conclusion, we have determined the structure of alkyl Grignard reagents diluted

in CH2Cl2 and tBuOMe. At room temperature, solutions of MeMgBr and EtMgBr are

dimeric both in Et2O and in tBuOMe at high concentrations (3 M and 1 M

respectively). Dilution leads to decrease in the association of alkyl Grignard reagents to monomeric state. At conditions typically used in asymmetric catalysis (0.15-0.5 M, -78 °C) Grignard reagents are monomeric, in the form of the magnesium halide (with two molecules of ethereal solvent coordinated). This was known for

Et2O, and with this research we have shown that this is also the case for the solvents

CH2Cl2 and tBuOMe, largely employed in asymmetric catalysis. Catalytic

asymmetric reactions are complex systems and the solvent can have influence on several parameters apart from the Grignard reagents, such as the catalyst (metal and ligand), the reaction intermediates and the transition state. Thus, the different results observed in catalysis when moving from one solvent to another cannot be attributed simply to their effect on the Grignard reagents. Having said that, on the ground of the experiments performed some correlations between the solvent and

the reactivity can be made: we have noticed that CH2Cl2 is a better solubilizing

medium for RMgBr·(OEt2)2 than Et2O and proven that in the latter the shift of

Schlenk equilibrium towards the dialkylmagnesium species is more favored when

compared with CH2Cl2. As R2Mg is more reactive than RMgBr the blank reaction

derived from addition of R2Mg is more likely in Et2O than in CH2Cl2. On the other

hand, Grignard reagents in tBuOMe (prepared in it or by dilution from a

concentrated solution in Et2O) are less nucleophilic because tBuOMe is poorer Lewis

base donor than Et2O.

8.4. Experimental section 8.4.1. General information

MeMgBr (3 M in Et2O), EtMgBr (3 M in Et2O), EtMgBr (1 M in tBuOMe), dry

tBuOMe, dry Et2O, CD2Cl2, 1,2,3,4-tetraphenylnaphthalene, tetramethylsilane, 2-phenylpiridine, diphenylacetylene and dioxane were purchased from

Sigma-Aldrich. Dry THF was distilled from sodium metal/benzophenone. Me2Mg and

Et2Mg in Et2O were prepared according to a literature procedure.[22] When

non-deuterated solvents were used (tBuOMe and Et2O) a stick containing CD2Cl2 was

placed in the NMR tube to lock the sample. Anhydrous conditions were used when preparing the samples for NMR measurements: a vacuum dried 5mm NMR tube

(with CD2Cl2 stick if needed) was closed with a septum and three cycles of

N2/vacuum applied. Then, 0.0225 mmol (9.7 mg) of tetraphenylnaphthalene were

weighted in a GC-MS vial, closed, and N2 atmosphere achieved by repeated cycles

of N2/vacuum. The main solvent of the reaction (amount dependent on the final

concentration of the solution) was added to the vial, and the 1,2,3,4-tetraphenylnaphthalene solution of it transferred to the NMR tube. Then 0.0225 mmol (3 μL) of tetramethylsilane were added and after the organomagnesium compound was added. The total volume of the sample was always 600 μL. The sample was stirred using a Vortex and if necessary, it was cooled down to -78 °C. All NMR experiments were recorded on an Agilent VNMR 500 MHz spectrometer equipped with a HF-X indirect probe that can run low temp experiments. It had

z-gradients that can perform a maximum gradient strength of 65 G cm-1. 5mm NMR

tubes were used and no spinning was done in any of the experiments.

All NOESY experiments used 0.6 s in mixing time and 1 s in delay time before first scans in experiments. Both the NOESY and the 1D-NOESY used a zero-quantum

Figure 6. Solvent exchange when tBuOMe and Et2O are present in the solution of a Grignard

reagent at -78 °C. Top: 0.15 M solution of 3 M EtMgBr in Et2O (30 μL) dissolved in

tBuOMe(570 μL). Middle: 1D-NOE effect upon irradiation of the methylene peak of

EtMgBr. Bottom: 1D-NOE effect upon irradiation of the methylene peak of Et2O.

Experiments with MeMgBr gave analogous results.

8.3. Conclusion

In conclusion, we have determined the structure of alkyl Grignard reagents diluted

in CH2Cl2 and tBuOMe. At room temperature, solutions of MeMgBr and EtMgBr are

dimeric both in Et2O and in tBuOMe at high concentrations (3 M and 1 M

respectively). Dilution leads to decrease in the association of alkyl Grignard reagents to monomeric state. At conditions typically used in asymmetric catalysis (0.15-0.5 M, -78 °C) Grignard reagents are monomeric, in the form of the magnesium halide (with two molecules of ethereal solvent coordinated). This was known for

Et2O, and with this research we have shown that this is also the case for the solvents

CH2Cl2 and tBuOMe, largely employed in asymmetric catalysis. Catalytic

asymmetric reactions are complex systems and the solvent can have influence on several parameters apart from the Grignard reagents, such as the catalyst (metal and ligand), the reaction intermediates and the transition state. Thus, the different results observed in catalysis when moving from one solvent to another cannot be attributed simply to their effect on the Grignard reagents. Having said that, on the ground of the experiments performed some correlations between the solvent and

the reactivity can be made: we have noticed that CH2Cl2 is a better solubilizing

medium for RMgBr·(OEt2)2 than Et2O and proven that in the latter the shift of

Schlenk equilibrium towards the dialkylmagnesium species is more favored when

compared with CH2Cl2. As R2Mg is more reactive than RMgBr the blank reaction

derived from addition of R2Mg is more likely in Et2O than in CH2Cl2. On the other

hand, Grignard reagents in tBuOMe (prepared in it or by dilution from a

concentrated solution in Et2O) are less nucleophilic because tBuOMe is poorer Lewis

base donor than Et2O.

8.4. Experimental section 8.4.1. General information

MeMgBr (3 M in Et2O), EtMgBr (3 M in Et2O), EtMgBr (1 M in tBuOMe), dry

tBuOMe, dry Et2O, CD2Cl2, 1,2,3,4-tetraphenylnaphthalene, tetramethylsilane, 2-phenylpiridine, diphenylacetylene and dioxane were purchased from

Sigma-Aldrich. Dry THF was distilled from sodium metal/benzophenone. Me2Mg and

Et2Mg in Et2O were prepared according to a literature procedure.[22] When

non-deuterated solvents were used (tBuOMe and Et2O) a stick containing CD2Cl2 was

placed in the NMR tube to lock the sample. Anhydrous conditions were used when preparing the samples for NMR measurements: a vacuum dried 5mm NMR tube

(with CD2Cl2 stick if needed) was closed with a septum and three cycles of

N2/vacuum applied. Then, 0.0225 mmol (9.7 mg) of tetraphenylnaphthalene were

weighted in a GC-MS vial, closed, and N2 atmosphere achieved by repeated cycles

of N2/vacuum. The main solvent of the reaction (amount dependent on the final

concentration of the solution) was added to the vial, and the 1,2,3,4-tetraphenylnaphthalene solution of it transferred to the NMR tube. Then 0.0225 mmol (3 μL) of tetramethylsilane were added and after the organomagnesium compound was added. The total volume of the sample was always 600 μL. The sample was stirred using a Vortex and if necessary, it was cooled down to -78 °C. All NMR experiments were recorded on an Agilent VNMR 500 MHz spectrometer equipped with a HF-X indirect probe that can run low temp experiments. It had

z-gradients that can perform a maximum gradient strength of 65 G cm-1. 5mm NMR

tubes were used and no spinning was done in any of the experiments.

All NOESY experiments used 0.6 s in mixing time and 1 s in delay time before first scans in experiments. Both the NOESY and the 1D-NOESY used a zero-quantum

filter steady-state scans to prevent artifacts. The 1D-NOESY used 64 scans and the 2D-NOESY experiments had 4 scans and 2x256 scans in the indirect dimension. The diffusion (DOSY) experiments used a bipolar stimulated echo with convection compensating and gradient compensating. The combined acquisition time and delay time between experiments was 6 s and the diffusion time, Δ, varied from 50 ms, suitable for diffusion at room temperature, to 100 - 150 ms at -78 °C. The length of gradients, δ/2, was kept constant (1 ms) during the experiments and the strength of the gradients was arrayed in 16 steps to obtain good diffusion resolution for single exponential decays. 16 scans were used in each step and the full phase cycling of the Dbppste_cc pulse sequence was used. All diffusion experiments were processed and evaluated with the DOSY package in VNMRJ and showed to fit well to single components. Here also a calibration for non-uniform gradient was also used in the DOSY processing.

T1 Relaxation was done with an inversion-recovery experiment and had 11 steps of

delay time between the 180° and 90° pulse. T2 Relaxation was using CPMG

experiment with 10 steps of a series of 180 ° pulses to get a increased time in the x,y-plane. Both relaxation experiments used 4 scans and had a delay time of 20s before

the first pulse. Relaxation data was fitted accordingly to T1 and T2 function

equations in the VNMRJ software. All values for relaxation and diffusion were processed and extracted from the VNMRJ software. Nevertheless, all fits were also analyzed individually to see that the fits gave reliable results.

8.4.2. T1/T2 relaxation experiments 0.15 M EtMgBr (3 M in Et2O) in CD2Cl2 at 25 °C T1 T2 CH3CH2MgBr 1.4 0.15 CH3CH2MgBr 3.7 0.12 CH3CH2OCH2CH3 5.2 0.02 CH3CH2OCH2CH3 4.6 0.03 TPhN 3.6 2.8 0.15 M EtMgBr (3 M in Et2O, solvent removed) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.84 0.02 CH3CH2MgBr 1 0.12 CH3CH2OCH2CH3 0.48 0.04 CH3CH2OCH2CH3 0.57 0.03 TPhN 2.2 0.6 0.5 M EtMgBr (3 M in Et2O) in CD2Cl2 at 25 °C T1 T2 CH3CH2MgBr 0.42 0.15 CH3CH2MgBr 1.9 0.12 CH3CH2OCH2CH3 3.9 0.02 CH3CH2OCH2CH3 3.5 0.03 TPhN 2.9 2.8

8.4.3. Determination of Molecular Weights (MW)

1  = 

2 = 10

, 

 

3 log, = log,− log+ log 

0.15 M EtMgBr (3 M in Et2O) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.9 0.3 CH3CH2MgBr 1.2 0.05 CH3CH2OCH2CH3 0.8 0.04 CH3CH2OCH2CH3 0.8 0.02 TPhN 2.1 0.64 Et2O (30 µL) in CD2Cl2 at -78 °C T1 T2 CH3CH2OCH2CH3 1.5 2.9 CH3CH2OCH2CH3 1.6 0.7 TPhN 1.7 0.6 0.5 M EtMgBr (3 M in Et2O) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.9 0.5 CH3CH2MgBr 1.1 0.1 CH3CH2OCH2CH3 0.5 0.04 CH3CH2OCH2CH3 0.7 0.02 TPhN 2.2 0.5

filter steady-state scans to prevent artifacts. The 1D-NOESY used 64 scans and the 2D-NOESY experiments had 4 scans and 2x256 scans in the indirect dimension. The diffusion (DOSY) experiments used a bipolar stimulated echo with convection compensating and gradient compensating. The combined acquisition time and delay time between experiments was 6 s and the diffusion time, Δ, varied from 50 ms, suitable for diffusion at room temperature, to 100 - 150 ms at -78 °C. The length of gradients, δ/2, was kept constant (1 ms) during the experiments and the strength of the gradients was arrayed in 16 steps to obtain good diffusion resolution for single exponential decays. 16 scans were used in each step and the full phase cycling of the Dbppste_cc pulse sequence was used. All diffusion experiments were processed and evaluated with the DOSY package in VNMRJ and showed to fit well to single components. Here also a calibration for non-uniform gradient was also used in the DOSY processing.

T1 Relaxation was done with an inversion-recovery experiment and had 11 steps of

delay time between the 180° and 90° pulse. T2 Relaxation was using CPMG

experiment with 10 steps of a series of 180 ° pulses to get a increased time in the x,y-plane. Both relaxation experiments used 4 scans and had a delay time of 20s before

the first pulse. Relaxation data was fitted accordingly to T1 and T2 function

equations in the VNMRJ software. All values for relaxation and diffusion were processed and extracted from the VNMRJ software. Nevertheless, all fits were also analyzed individually to see that the fits gave reliable results.

8.4.2. T1/T2 relaxation experiments 0.15 M EtMgBr (3 M in Et2O) in CD2Cl2 at 25 °C T1 T2 CH3CH2MgBr 1.4 0.15 CH3CH2MgBr 3.7 0.12 CH3CH2OCH2CH3 5.2 0.02 CH3CH2OCH2CH3 4.6 0.03 TPhN 3.6 2.8 0.15 M EtMgBr (3 M in Et2O, solvent removed) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.84 0.02 CH3CH2MgBr 1 0.12 CH3CH2OCH2CH3 0.48 0.04 CH3CH2OCH2CH3 0.57 0.03 TPhN 2.2 0.6 0.5 M EtMgBr (3 M in Et2O) in CD2Cl2 at 25 °C T1 T2 CH3CH2MgBr 0.42 0.15 CH3CH2MgBr 1.9 0.12 CH3CH2OCH2CH3 3.9 0.02 CH3CH2OCH2CH3 3.5 0.03 TPhN 2.9 2.8

8.4.3. Determination of Molecular Weights (MW)

1  = 

2 = 10

, 

 

3 log, = log,− log+ log 

0.15 M EtMgBr (3 M in Et2O) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.9 0.3 CH3CH2MgBr 1.2 0.05 CH3CH2OCH2CH3 0.8 0.04 CH3CH2OCH2CH3 0.8 0.02 TPhN 2.1 0.64 Et2O (30 µL) in CD2Cl2 at -78 °C T1 T2 CH3CH2OCH2CH3 1.5 2.9 CH3CH2OCH2CH3 1.6 0.7 TPhN 1.7 0.6 0.5 M EtMgBr (3 M in Et2O) in CD2Cl2 at -78 °C T1 T2 CH3CH2MgBr 0.9 0.5 CH3CH2MgBr 1.1 0.1 CH3CH2OCH2CH3 0.5 0.04 CH3CH2OCH2CH3 0.7 0.02 TPhN 2.2 0.5

4 = 1 −_

 × 100

We have determined the MW of the different molecules from their diffusion

coefficients (equation 1).[16] _{The MW of a molecule can be determined using equation}

2.[13d,e] _{For that logDx,norm has to be known, and this can be obtained from equation 3.}

Here, Dx is the measured diffusion coefficient of the molecule, while Dx,norm refers to

the normalized diffusion coefficient (normalized to the internal reference diffusion

coefficient Dref). In our case we have two Dref,, namely DTMS (for TMS) and DTPhN (for

tetraphenylnaphthalene, TPhN). (See table below) Therefore, we also have two

normalized diffusion coefficients (Dx,norm) one for TMS (Dx,norm for TMS) and another for

TPhN(Dx,norm for TPhN). Dref,fix is the fixed diffusion coefficient of the internal standard,

measured independently to the sample containing the analyte. Log K and α are

calculated from the plot of the logDx of external references against their logMW

(external calibration curve, see below). Because we have two Dx,norm we also have

two determined MWdet, one determined using TMS as internal standard and the

other using TPhN. Consequently, there are two errors of measurements (calculated

using equation 4). MWcalc refers to the theoretical MW of the molecule. We found

that the errors are smaller when TPhN is used as internal reference, and those are bold in the table and the ones used. However, we have also included in the table the

MWdet if TMS is chosen as internal reference and the errors obtained with it. Only in

the cases in which we have determined the MW of Grignard reagents directly taken from the bottle (higher concentration and density) we have used TMS as internal standard to avoid problems of solubilizing TPhN.

Constructed external calibration curves (ECC):

ECC for measurements in CD2Cl2 at 25 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 2,36E-009 -8,63 88 1,94

TPhN 9,85E-010 -9,01 433 2,64

Log K and α were calculated to be 7.56 and

-0.55 respectively, which match very well with the values previously reported for

this extended calibration curve.[13e]_{(-7.55 and -0.54 respectively)}

-9,1 -9 -8,9 -8,8 -8,7 -8,6 1,5 2,5 Log Dx LogMW

For CD2Cl2 at -78 °C and the other solvents ECC were constructed with four

references of different MW and shape.

ECC for measurements in CD2Cl2 at -78 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 3,42E-010 -9,47 88 1,94

2-Phenylpyridine 2,19E-010 -9,66 155 2,19

Diphenylacetylene 2,28E-010 -9,64 178 2,25

TPhN 1,42E-010 -9,85 433 2,64

Log K: -8,444; α: -0,536; r= 0,97

ECC for measurements in Et2O at 25 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 4,38E-009 -8,36 88 1,94

2-Phenylpyridine 3,00E-009 -8,52 155 2,19

Diphenylacetylene 2,90E-009 -8,54 178 2,25

TPhN 1,70E-009 -8,77 433 2,64

Log K: -7,22; α: -0, 588; r= 0,99

ECC for measurements in Et2O at -78 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 7,14E-010 -9,15 88 1,94 2-Phenylpyridine 3,92E-010 -9,41 155 2,19 Diphenylacetylene 4,16E-010 -9,38 178 2,25 TPhN 2,29E-010 -9,64 433 2,64 Log K: -7,84; α: -0,691; r= 0,96 -9,9 -9,8 -9,7 -9,6 -9,5 -9,4 1,5 2,5 Log Dx LogMW -8,8 -8,7 -8,6 -8,5 -8,4 -8,3 1,5 Log Dx LogMW -9,70 -9,60 -9,50 -9,40 -9,30 -9,20 -9,10 1,5 2,5 Log Dx LogMW

4 = 1 −_

 × 100

We have determined the MW of the different molecules from their diffusion

coefficients (equation 1).[16] _{The MW of a molecule can be determined using equation}

2.[13d,e] _{For that logDx,norm has to be known, and this can be obtained from equation 3.}

Here, Dx is the measured diffusion coefficient of the molecule, while Dx,norm refers to

the normalized diffusion coefficient (normalized to the internal reference diffusion

coefficient Dref). In our case we have two Dref,, namely DTMS (for TMS) and DTPhN (for

tetraphenylnaphthalene, TPhN). (See table below) Therefore, we also have two

normalized diffusion coefficients (Dx,norm) one for TMS (Dx,norm for TMS) and another for

TPhN(Dx,norm for TPhN). Dref,fix is the fixed diffusion coefficient of the internal standard,

measured independently to the sample containing the analyte. Log K and α are

calculated from the plot of the logDx of external references against their logMW

(external calibration curve, see below). Because we have two Dx,norm we also have

two determined MWdet, one determined using TMS as internal standard and the

other using TPhN. Consequently, there are two errors of measurements (calculated

using equation 4). MWcalc refers to the theoretical MW of the molecule. We found

that the errors are smaller when TPhN is used as internal reference, and those are bold in the table and the ones used. However, we have also included in the table the

MWdet if TMS is chosen as internal reference and the errors obtained with it. Only in

the cases in which we have determined the MW of Grignard reagents directly taken from the bottle (higher concentration and density) we have used TMS as internal standard to avoid problems of solubilizing TPhN.

Constructed external calibration curves (ECC):

ECC for measurements in CD2Cl2 at 25 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 2,36E-009 -8,63 88 1,94

TPhN 9,85E-010 -9,01 433 2,64

Log K and α were calculated to be 7.56 and

-0.55 respectively, which match very well with the values previously reported for

this extended calibration curve.[13e]_{(-7.55 and -0.54 respectively)}

-9,1 -9 -8,9 -8,8 -8,7 -8,6 1,5 2,5 Log Dx LogMW

For CD2Cl2 at -78 °C and the other solvents ECC were constructed with four

references of different MW and shape.

ECC for measurements in CD2Cl2 at -78 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 3,42E-010 -9,47 88 1,94

2-Phenylpyridine 2,19E-010 -9,66 155 2,19

Diphenylacetylene 2,28E-010 -9,64 178 2,25

TPhN 1,42E-010 -9,85 433 2,64

Log K: -8,444; α: -0,536; r= 0,97

ECC for measurements in Et2O at 25 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 4,38E-009 -8,36 88 1,94

2-Phenylpyridine 3,00E-009 -8,52 155 2,19

Diphenylacetylene 2,90E-009 -8,54 178 2,25

TPhN 1,70E-009 -8,77 433 2,64

Log K: -7,22; α: -0, 588; r= 0,99

ECC for measurements in Et2O at -78 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 7,14E-010 -9,15 88 1,94 2-Phenylpyridine 3,92E-010 -9,41 155 2,19 Diphenylacetylene 4,16E-010 -9,38 178 2,25 TPhN 2,29E-010 -9,64 433 2,64 Log K: -7,84; α: -0,691; r= 0,96 -9,9 -9,8 -9,7 -9,6 -9,5 -9,4 1,5 2,5 Log Dx LogMW -8,8 -8,7 -8,6 -8,5 -8,4 -8,3 1,5 Log Dx LogMW -9,70 -9,60 -9,50 -9,40 -9,30 -9,20 -9,10 1,5 2,5 Log Dx LogMW

ECC for measurements in tBuOMe at 25 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 2,80E-009 -8,55 88 1,94

2-Phenylpyridine 2,19E-009 -8,66 155 2,19

Diphenylacetylene 2,02E-009 -8,70 178 2,25

TPhN 1,11E-009 -8,95 433 2,64

Log K: -7,39; α: - 0,589; r= 0,99

ECC for measurements in tBuOMe at -78 °C

Reference Dx (m2/s) LogDx MW Log MW

TMS 2,90E-010 -9,54 88 1,94 2-Phenylpyridine 2,16E-010 -9,67 155 2,19 Diphenylacetylene 1,92E-010 -9,72 178 2,25 TPhN 1,14E-010 -9,94 433 2,64 Log K: -8,38; α: - 0,591; r= 0,99 -9,00 -8,90 -8,80 -8,70 -8,60 -8,50 1,5 2,5 Log Dx LogMW -10,00 -9,90 -9,80 -9,70 -9,60 -9,50 1,5 Log Dx LogMW T(° C ) Sa m ple D x Log D x D TMS D TP hN Log D x, nor m (fo rT Ph N ) Log D x, nor m (f or T M S) M w ca lc M w de t (f ro m T M S) M w erro r (f ro m T M S) M w de t (fro m T Ph N ) M w erro r (f ro m T Ph N ) 25 0. 15 M M eM gB r ( 3 M in E t 2 O ) i n M eM gB r 1, 42E -00 9 -8 ,848 -8 ,624 -8 ,976 -8 ,879 -8 ,851 267 224 16 253 5 TM S 2, 38E -00 9 -8 ,624 -8 ,624 -8 ,976 -8 ,655 -8 ,627 88 87 1 98 -12 Et 2 O 2, 08E -00 9 -8 ,682 -8 ,624 -8 ,976 -8 ,713 -8 ,684 74 112 -51 125 -70 TP hN 1, 06E -00 9 -8 ,976 -8 ,624 -8 ,976 -9 ,007 -8 ,978 433 383 12 430 1 -7 8 0. 15 M M eM gB r ( 3 M in E t 2 O ) i n C D 2 C l 2 M eM gB r 2, 08E -01 0 -9 ,682 -9 ,428 -9 ,828 -9 ,702 -9 ,720 267 240 10 222 17 TM S 3, 73E -01 0 -9 ,428 -9 ,428 -9 ,828 -9 ,448 -9 ,466 88 81 8 75 15 Et 2 O 2, 78E -01 0 -9 ,556 -9 ,428 -9 ,828 -9 ,576 -9 ,593 74 139 -88 129 -74 TP hN 1, 49E -01 0 -9 ,828 -9 ,428 -9 ,828 -9 ,848 -9 ,865 433 448 -3 415 4 -7 8 0. 15 M M eM gB r ( so lve nt re m ove d) in C D 2 C l 2 M eM gB r 2, 09E -01 0 -9 ,680 -9 ,442 -9 ,796 -9 ,732 -9 ,703 267 223 16 252 6 TM S 3, 61E -01 0 -9 ,442 -9 ,442 -9 ,796 -9 ,494 -9 ,466 88 81 8 91 -3 Et 2 O 2, 29E -01 0 -9 ,640 -9 ,442 -9 ,796 -9 ,692 -9 ,664 74 188 -154 213 -187 TP hN 1, 60E -01 0 -9 ,796 -9 ,442 -9 ,796 -9 ,848 -9 ,819 433 367 15 415 4 -78 30μ o f E t 2 O in 570μ o f C D 2 C l 2 TM S 3, 31E -01 0 -9 ,480 -9 ,480 -9 ,779 -9 ,549 -9 ,466 88 81 8 115 -31 Et 2 O 4, 20E -01 0 -9 ,376 -9 ,480 -9 ,779 -9 ,445 -9 ,362 74 52 30 74 1 TP hN 1, 66E -01 0 -9 ,779 -9 ,480 -9 ,779 -9 ,848 -9 ,765 433 291 33 415 4 DO SY m ea sur em ents in CD 2 Cl 2