Synthesis 2020; 52(15): 2147-2161
DOI: 10.1055/s-0039-1690898
short review
© Georg Thieme Verlag Stuttgart · New York

Manifestation of the β-Silicon Effect in the Reactions of Unsaturated Systems Involving a 1,2-Silyl Shift

Rūdolfs Beļaunieks
,
Mikus Puriņš
,
Institute of Technology of Organic Chemistry, Faculty of Materials Science and Applied Chemistry, Riga Technical University, P. Valdena str. 3, Riga, 1048, Latvia   Email: Maris.Turks@rtu.lv
› Author Affiliations

This work was supported by a grant from Latvian Council of Science (grant No. LZP-2018/1-0315) and doctoral student grant from Riga Technical University (grant No. DOK.MLKF/19).
Further Information

Publication History

Received: 31 January 2020

Accepted after revision: 30 March 2020

Publication Date:
20 April 2020 (online)

 


Abstract

Many chemical transformations of organosilicon compounds proceed due to the capability of silyl substituents to stabilize a positive charge in its β-position. This short review provides an overview of the present understanding of the β-silicon effect and focusses on the synthetic applications of 1,2-silyl shifts resulting from non-vertical stabilization of alkylcarbenium ions and vinyl cations. The reactions of silicon containing unsaturated starting materials, alkenes, allenes, and alkynes, involving β-silyl group stabilized cationic intermediates, transition metal carbenes, or vinylidene complexes will be discussed.

1 Introduction

2 Origins of the β-Silicon Effect

3 Reactions of Allenylsilanes

4 Reactions of Alkynes

4.1 Propargylsilanes

4.2 Alkynylsilanes

5 Reactions of Alkenes

5.1 Allylsilanes

5.2 Vinylsilanes

6 Conclusions


# 1

Introduction

Zoom Image
Rūdolfs Beļaunieks(left) obtained his B.Sc. and M.Sc. from Riga Technical University (RTU) in 2017 and 2019, respectively. Currently he is pursuing his doctoral degree under the supervision of Prof. M. Turks. His research is focused on novel synthetic methodologies involving 1,2-silyl shift.
Mikus Puriņš (middle) obtained his B.Sc. and M.Sc. from RTU under the supervision of Prof. M. Turks in 2017 and 2019, respectively. Currently he is pursuing his doctoral degree under the supervision of Professor Jérôme Waser at Swiss Federal Institute of Technology, Lausanne (EPFL), Switzerland. His research interests include tethered strategies for selective functionalization of small molecules.
Professor Māris Turks (right) obtained his B.Sc. in chemistry from the University of Latvia in 1998 and his M.Sc. from Riga Technical University in 2000. In 2005 he earned his Dr. ès. sc. degree in synthetic organic chemistry from EPFL, Switzerland under the direction of Professor Pierre Vogel. Then he spent one year as SNSF Postdoctoral Fellow at Stanford University in the group of Professor Barry M. Trost. In 2007, he accepted an academic position at the Faculty of Materials Science and Applied Chemistry, RTU, where he is currently Dean of Faculty and Director of the Institute of Technology of Organic Chemistry. Prof. Turks’ current research interests are in organosilicon chemistry, application of sulfur dioxide in organic synthesis, and purine and triterpenoid chemistry.

In the family of group IV elements, carbon plays a central role as the main building block of organic chemistry. Silicon, albeit related, offers some curious differences. Although, silicon and carbon share the archetypal tetravalency, silicon’s position in the 3rd period of Mendeleev’s periodic table unlocks many pathways impossible for its neighbor, carbon. The clear differences are its increased size,[2] higher energy p-orbitals,[3] accessible vacant d-orbitals,[4] and diminished electronegativity.[5] This amounts to remarkable differences in the bonding properties of silicon. The C–Si bond (1.85 Å) is considerably longer than the C–C bond (1.54 Å),[6] although of comparable strength. Silicon also forms high strength bonds with electronegative elements, which is a driving force for many organosilicon-based reactions.[7] The unique affinity of silicon towards fluorine offers orthogonal deprotection conditions, which cements the silyl group’s position as one of the most useful protecting groups for oxygen-containing compounds.[8] Silicon even holds a near monopoly for the protection of sp hybridized carbons.[8] Moreover, the activation via ‘ate’ complexes has realized the possibility of harnessing the intrinsic polarity of the C–Si bond, exemplified by the Tamao–Fleming oxidation or the Hiyama–Denmark cross-coupling reactions.[9] [10]

The most intriguing property of silicon is that of the stabilization of electron-deficient carbon atoms in β-silicon carbenium ions, widely recognized as the β-silicon effect. In fact, this effect is responsible for most of the observed reactivity of unsaturated organosilicon compounds.[11] However, historically the explanation of the β-silicon effect has not been a trivial task. Various experimental observations of products where silicon had changed its position on the carbon backbone led to different interpretations of the structure of the β-silicon carbenium ion (vide infra). This question was parallel to the famous debate of Winstein and Brown on the non-classical carbocations, albeit not as vocal and publicized.[12] Just as for the question on the structure of the 2-norbornyl cation, the resulting experimental and theoretical studies have increased the understanding of β-silicon carbenium ions.

Synthetic methods involving 1,2-silyl migration have been developed since the 1980s and have been previously reviewed by Aye and co-workers for the reactions of allenyl, propargyl, and vinylsilanes;[13] Knölker and co-workers[14] [15] and Landais and co-workers[16] have reviewed allylsilane annulation reactions. This review aims to introduce the origins of the β-silicon effect and its manifestation in some synthetic methodologies of the last decade that involve a 1,2-silyl shift, yet discussing them in the context of previous research.


# 2

Origins of the β-Silicon Effect

The first observation of enhanced reactivity or a change in chemoselectivity caused by a β-silyl substituent was reported by Sommer and Whitmore.[17] [18] They observed that when treating β-silicon alkyl chloride 1a with sodium hydroxide only the elimination product, propylene (2), was obtained (Scheme [1]). Under the same reaction conditions, the regioisomeric silylalkyl chlorides 3a and 4a afforded only the hydrolysis products of the Si–Cl bond. Further investigations concluded that similar elimination reactions proceed via the E1 reaction mechanism, where the rate determining step is the dissociation of C–X bond.[19]

Zoom Image
Scheme 1 Hydrolysis of silylalkyl chlorides

The twelve-fold increase of the solvolysis rate can be explained by the formation of a carbenium ion that is stabilized by the β-silyl group. This intermediate is described by two possible structural proposals: (1) a vertically stabilized carbenium ion (i.e., hyperconjugation by the neighboring C–Si bond) where no notable changes in lengths of bonds are observed; (2) a non-vertically stabilized silonium ion, where the lengths of the bonds alter significantly (Scheme [2]).[20] The question between vertical and non-vertical stabilization is that of whether the non-vertically stabilized form can be regarded as a feasible transition state for two different structures in rapid equilibrium, or as a true intermediate (energy minimum).

Zoom Image
Scheme 2 Vertical and non-vertical stabilization of carbenium ions by β-silicon substituent. Here and further in text substituent ‘ Si ’ is used as generic description of the R3Si group (R = alkyl and/or aryl).

The geometry required for the manifestation of the β-silicon effect was convincingly demonstrated with cyclic and conformationally restricted model substrates possessing a distinct dihedral angle between the leaving group and the β-silyl substituent (Figure [1]).[21] [22] [23] [24] Compared to non-silylated­ analogues a greater increase of the solvolysis rate was observed in anti-periplanar conformation 12, up to 1012 times. The syn-periplanar conformation 8 showed a significant increase as well (105 times), though not as notable as the anti-periplanar conformation. The elimination from compound 8 indicates the existence of vertical stabilization in the intermediate β-silyl carbenium ion. On the contrary, the non-vertical stabilization of the intermediate during the reaction 813 is disturbed by the syn-peri­planar placement of the silyl group and the mesylate leaving group. However, the observations do not exclude non-vertical stabilization effects in anti conformation 12. The increase of the rates in gauche (9) and anticlinal (11) conformations by a factor of 104 clearly states that the partial overlap can increase the reaction rates as well. As predicted, in orthogonal conformation 10, where neither hyperconjugation, nor formation of silonium ion can take place, no increase in the reaction rate was observed. In conclusion, the experimental evidence indicates that formation of carbenium ion in the above-mentioned and other conformationally unrestricted β-silyl systems is mainly due to both vertical and non-vertical stabilization (Scheme [2]), but not from the induction effects of the electron-donating silyl group.

Zoom Image
Figure 1 Comparison of solvolysis rate constants in CF3CH2OH at 25 °C (a, d, and e are extrapolated from experiments at other temperatures)
Zoom Image
Figure 2 Calculation of stabilization energy for different β-silylcarbenium systems

These suggestions are in the same vein as computer-aided modeling (Figure [2]). MP3/6-31G* level calculations of simple systems showed a 38.0 kcal/mol increase in the stabilization energy for the vertical (open) transition state 19b in comparison to the unsubstituted ethyl cation (17). The primary carbenium ion 19a, in which the C–Si bond is orthogonal to the empty orbital, is stabilized by 8.9 kcal/mol, suggesting the existence of induction and polarization effects arising from the C–Si bond. However, the cyclic silonium ion 19c was found to be even more stable by 2.4 kcal/mol than the open structure 19b. In spite of the latter, further calculations showed that for tertiary carbenium ions (e.g., intermediate 20) the added donating substituents compete for their role in the overall stabilization of the system. Therefore, the necessity for silonium ion formation is diminished. Finally, the non-vertically and vertically stabilized secondary carbenium ions 21a and 21b differ by 3.9 kcal/mol with the latter being more stable.[25] [26]

To determine transition state structures, the solvolysis kinetics of β-silyl systems 22 were measured for variously substituted β-aryldimethylsilyl systems (Figure [3]).[27] [28] [29] [30] The obtained data were interpreted with the Yukawa–Tsuno equation, a modification of the Hammett equation that accounts for the resonance effects on a reactive center (Equation 1). The parameter ρ describes the sensitivity of the transition state towards the electronic effects of its substituents. It shows the relative amount of charge on the silicon center: a strongly negative ρ-value indicates an enhanced positive charge on the silicon (i.e., greater importance of the non-vertical stabilization).

Zoom Image
Figure 3 Kinetic studies of transition states of β-silylcarbenium systems
Zoom Image
Equation 1 Yukawa–Tsuno equation: k x = reaction rate constant for X-substituted arylsilane; k h = reaction rate constant for unsubstituted phenylsilane, ρ = reaction constant, σ° = Hammett substituent constant; r + = enhanced resonance parameter, Δσ+ R = Hammett resonance substituent constant.

It was concluded that for α-unsubstituted systems 23 ρ = –1.75, alkyl-substituted systems 24 ρ = –1.50, and aryl-substituted systems 25 ρ = –1.10 to –0.80. Combined quantum chemical calculations and kinetic data showed that the less stabilized, unsubstituted systems prefer non-vertical stabilization, but for the benzyl carbenium ion vertical stabilization is the dominant one. The alkyl-substituted system behaved as a non-discrete structure between vertical and non-vertical stabilization. These findings correlate with previously described solvolysis and DFT calculation results.

The stabilization of vinyl cations by the β-silyl group initially was studied by DFT calculations. Compared to the unsubstituted vinyl cation 26 its counterpart, the vinyl­silane 29, is by 28.6 kcal/mol more stable (Figure [4]).[25]

Zoom Image
Figure 4 Stabilization of vinyl cation by β-silicon effect
Zoom Image
Scheme 3 Formation of silylvinyl cations for studying effects of hyperconjugation

To assess the stabilization effects of the β-silyl carbocation, Stone and co-workers analyzed thermodynamic data of electrophile [TMS+ (TMS = trimethylsilyl) or H+] addition to differently substituted alkynes and alkenes (Scheme [3]). From the measured ΔH f of cations 31 and 32 and alkenes 33 and 34 the corresponding stabilization energies: ΔH stab were calculated. The β-silyl effect stabilizes vinyl carbo­cation 32 by 8.8–11.6 kcal/mol in comparison to its non-silylated­ counterpart 31. In a similar way ΔΔH were determined between sp2-carbenium ions 36 and 37. In this case, the stabilization resulting from the presence of the silyl group was doubled and the intermediate 37 was by 26.2 kcal/mol more stable than 36. Finally, if a substituent at α-position is more stabilizing, e.g. aryl groups, the participation of the silicon in the stabilization is lower when compared to α-alkyl substituents (Scheme [3]).[31]

It was discovered that β,β-silyl disubstituted vinyl cations 41 are surprisingly stable in solution and as anhydrous salts with tetrakis(pentafluorophenyl)borate as counterion. Their NMR,[32] X-ray diffraction,[33] and IR spectroscopy[34] studies revealed further correlation between cation stabilization from an α-aryl substituent and hyperconjugation from the β-silyl group (Scheme [4]). In 29Si NMR the presence of a single signal indicates the formation of symmetrical species. A significant low-field shift difference of Δδ 29Si ~ 29–42 ppm can also be observed between compounds 40 and 41. The such significant deshielding was attributed to the localization of the positive charge on the silicon due to hyperconjugation effects. In addition, the correlation between deshielding and donating/withdrawing effects can be observed as the most deshielded silicon was with R = fluorophenyl and the least deshielded, with R = ferrocenyl. The same goes with the 1 J C-Si coupling constants showing lower values for more electron-deficient systems indicating reduced degree of bonding between Si–Cβ.[32] Similar conclusions can be drawn when studying these substances with IR spectroscopy. By decreasing the electron-donating effects, the C=C+ bond order increases indicating a growing importance of hyperconjugation by silicon.[34]

Zoom Image
Scheme 4 Formation of β,β-silyl disubstituted vinyl cations

In summary, the β-silicon effect is an established phenomenon that has resulted in useful synthetic transformations of vinyl, allyl, propargyl, allenyl, and other silanes via a β-silyl carbenium ion intermediates. On several occasions, the lower energy of the intermediate β-silyl carbenium ion is due to non-vertical stabilization. This, in turn, can lead to 1,2-silyl shift that produces a more stable ionic species than the initial one.


# 3

Reactions of Allenylsilanes

The first reports of 1,2-silyl migration come from the Danheiser group back in 1981.[35] Treating the α,β-unsaturated ketone 43 with allenylsilanes 42 forms the cationic intermediate 44 which undergoes silyl migration to form the novel vinyl carbocation ion 45 that ultimately affords the favorable cyclopentane 46 (Scheme [5]). This formal use of allenylsilanes 42 as a 1,3-dipole in reactions with 1,2-unsaturated systems 43 was established further. Following the reaction pattern of [3+2] addition, Danheiser developed methods for the formation of other 5-membered rings 46 like azulene,[36] pyrrole,[37] furan,[37] [38] and isoxazole derivatives.[39]

Zoom Image
Scheme 5 Cyclopentene annulation reaction; the Danheiser annulation

In 2018, a novel method using the Danheiser annulation was developed to synthesize cis-hydrindan-2,4-diones that could further be used for the synthesis of lycopodium alkaloids.[40]

Zoom Image
Scheme 6 Reaction and plausible mechanism of [3+3] and [3+2] addition of cyclopropanes and allenylsilanes

The use of allenylsilanes as 1,3-dipoles was extended to other cycloadditions. Under Lewis acidic conditions cyclopropanes 47 can serve as 1,3-dipolar synthons and participate in cycloaddition reactions with allenylsilanes 48 (Scheme [6]).[41]

The cyclopropane 47a underwent ring-opening and annulation at –78 °C with (trimethylsilyl)allene 48a using TiCl4 as a catalyst; these catalytic conditions induced protodesilylation. To overcome this problem, a mixture of TiCl4 and Et2AlCl was used and product 52 was obtained. This selectivity of [3+2] addition was switched by increasing the reaction temperature to 25 °C and using Et2AlCl as the only Lewis acid. This change promoted intermediate 50 to undergo complete migration of the silyl group to the intermediate 51 that, after cyclization, resulted in [3+3] addition product 53.

The reaction was also studied with unsubstituted allenylsilane 48c. This reaction afforded only the [3+2] addition products 52, as the transition state 51 for this compound is hypothesized to be much higher in energy than transition state 50. The reaction scope could be extended to spiro­cyclic cyclopropanes to construct the corresponding spiro[4.4], -[4.5], and -[5.5] compounds from the corresponding cyclopropanones.

1,4-Dipoles can be obtained in a similar fashion from cyclobutanones as previously described with cyclopropanes. The Matsuo group worked on various [4+3] annulation reactions in order to uniquely obtain 8-oxabicyclo-[3.2.1]octan-3-ones via Lewis acid mediated reactions involving 1,2-silyl migration (Scheme [7]).[42]

Zoom Image
Scheme 7 Reaction and mechanism of cyclobutanone and allenylsilanes

The reaction mechanism proceeds with an initial Lewis acid promoted ring-opening of the butanone cycle. This results in the addition of the allene 55 to the oxocarbenium ion 58. Further 1,2-silyl migration promotes an intramolecular transetherification along with EtCl formation. The resulting cyclization and hydrolysis afford the desired product 56 (Scheme [7]). The most successful results (product 56 yields up to 67% and product 57a as low as 3%) were achieved by using TiCl4. A series of other Lewis acids yielded no product with the only exception being EtAlCl2 which in combination with TiCl4 gave product 56 in 52% yield.

The highest yields of product 56 were obtained with the tert-butyldiphenylsilyl (TBDPS) group. The relatively smaller triisopropylsilyl (TIPS) and tert-butyldimethylsilyl (TBS) groups showed a decrease in overall reactivity and greater formation of product 57a, even after prolonged heating. Larger substituents at the α-position of the silicon decreased yields of the bicyclic product 56, whereas the unsubstituted allenylsilane 55 ( Si = TBDPS, R1 = R2 = Me, R3 = H) gave only the [4+2] addition product 57b without 1,2-silyl­ migration.

Allenylsilanes can also be used in cycloisomerization reactions. Gevorgyan and co-workers used DFT calculations to develop a method that employs a 1,2-silyl migration (Scheme [8]).[43]

Zoom Image
Scheme 8 Gold-catalyzed furan synthesis with 1,2-silyl migration

Calculations in gas phase and solution both revealed that a 1,2-silyl migration is also possible for the late transition metal stabilized carbenes and they are strongly favored over 1,2-shifts of Me, Ph, and even H (Table [1]). The results showed that α-alkyl, α-aryl, and even unsubstituted allenylsilanes 64 proceeded smoothly to afford the corresponding 3-silylfurans 66 in good yields (Scheme [8]).

Table 1 Predicted Relative Activation Energies of Silicon vs Substituent R3 for 1,2-Migrations in Intermediate 65 (R1 = R2 = H, R3 = H, Me, Ph)

1,2-Migrating group

ΔΔG (kcal/mol)

Gold catalyst

Gas phase

Solution

Me

22.7

 –

[Au] = AuCl3

Ph

18.1

 –

H

11.1

12.8

TMS

 0.0

 0.0

H

 –

12.2

[Au] = H3PAu+

TMS

 –

 0.0


# 4

Reactions of Alkynes

4.1

Propargylsilanes

Rearrangement reactions with 1,2-silyl migration in propargylsilanes 68 were first observed in 1985 by Miginiac and co-workers as a side reaction between propargylsilanes and acetals.[44] Further investigations on reactions of propargylsilanes were performed by Danheiser and co-workers, continuing their success in annulation chemistry. This annulation reaction, with a similar mechanism to that of allenylsilanes, affords a variety of 5-membered rings 72, cyclopentenes, 1,2,5,7a-tetrahydro-3H-pyrrolizin-3-ones, isoxazoles, and azulenes (Scheme [9]).[45]

Zoom Image
Scheme 9 Synthesis of 5-membered rings from propargylsilanes

Evans and Aye developed a catalytic, enantio- and diastereoselective method for synthesis of highly functionalized vinylepoxides from propargylsilanes and N-phenylglyoxamide (Scheme [10]).[46] The propargylsilane 73 and the glyoxamide 74 gave the Hosomi–Sakurai product 77 under Sc(OTf)3 catalysis at ambient temperature. However, when lowering the temperature to –55 °C the reaction afforded only the vinylepoxide 78. The reaction starts with the π-nucleophile addition to the activated carbonyl group. The intermediate 75 undergoes 1,2-silyl shift to the more stable allyl cation 76, whereas, at –55 °C 76 is trapped kinetically. When the reaction is carried out at ambient temperature deprotonation of 76 takes place and thermodynamic product 77 is obtained.

Zoom Image
Scheme 10 Epoxide synthesis from propargylsilanes

To make this reaction enantioselective, it was concluded that highest stereoselectivity was achieved using the TBDPS moiety. Although Sc(III) showed the highest reaction chemoselectivity and yield, Al(III)–sal-BINAM (Figure [5]) in combination with AgOTf was used for the enantioselective catalytic system, as it resulted in much higher enantiomeric excess.

Zoom Image
Figure 5 Structure of Al(III)–sal-BINAM

Ferreira and co-workers developed a synthesis of highly functionalized alkenes from α-hydroxypropargylsilanes in electrophilic activation reactions (Scheme [11]).[47] Initially, the isomerization of α-hydroxypropargylsilanes 79 to the corresponding α,β-unsaturated ketones 82 was promoted by transition metals. The highest yields of products 82 were obtained with PtCl2. The reactions proceeded smoothly with high stereoselectivity [up to 19:1 favoring (Z)-isomer] even with highly functionalized alkynes and tolerated a wide variety of functional groups.

Zoom Image
Scheme 11 Synthesis of highly functionalized alkenes from propargylsilanes (DMPS = SiMe2Ph)

In addition to Pt(II) salts, the alkyne moiety was also activated with other electrophiles. Halogenating electrophiles such as N-bromosuccinimide (at ambient temperature), N-iodosuccinimide (at –10 to 0 °C) were successful for this transformation affording the products 83 with high stereoselectivity [>19:1 favoring (E)-isomer]. The reactions proceeded smoothly with a wide variety of substituents at alkynyl and propargylic positions.[48] The obtained alkenes could easily undergo cross-coupling reactions to achieve high complexity of either tri- or tetrasubstituted alkenes.

Ferreira and co-workers also developed a cyclomerization reaction for generating α,β-unsaturated platinum carbenes 85 that would afford the corresponding furans 86. Depending on the solvent, it was possible to migrate either a hydrogen atom or the TBS group (Scheme [12]).[49] It was found that the highest selectivity for silyl group migration could be obtained using PtCl2 with oct-1-ene in toluene. On the other hand, the highest H-migration was observed using same system in THF.

Zoom Image
Scheme 13 Synthesis of dienes and indenes from propargylsilanes
Zoom Image
Scheme 12 Synthesis of furans from propargylsilanes

A method for generating an allyl cation as a useful intermediate from propargylsilanes using a Brønsted acid catalyst was developed by Turks and co-workers (Scheme [13]).[50] Protonation of propargylsilanes 87 with a strong Brønsted acid formed vinyl cation 88 which underwent 1,2-silyl migration to afford more stable allyl cation 89. Deprotonation of the latter provided diene 90, when terminal alkynes (R1 = H) were used.

The reactivity of aryl-substituted systems 87 (R1 = Ar) was altered either by changing the proton source or the electronic character of the aryl substituent. The use of HNTf2 on the nitrophenyl-substituted propargylsilane 87 in CHCl3 gave the corresponding dienes 90 in up to 92% yield with (E,E)/(E,Z) 10:1. The use of any other less electron-withdrawing substituent on the phenyl ring resulted in exclusive­ formation of the indene 92. When changing the TBS group to the more sterically hindered TIPS group in combination with HCTf3 as the proton source gave silylindenes 92 in up to 90% yield with a wide variety of electron-withdrawing substituents on the aryl group. A change in the protonation regioselectivity was observed with electron-donating substituents, as the 4-methoxyphenyl derivative leads to the vinyl cation at the benzylic position followed by hydrolysis.

Synthesis of indanones 98 from in situ generated propargylsilanes was developed by Ballesteros and co-workers. In the presence of a gold catalyst, the cyclization reaction between alkynylsilanes 93 and acylsilanes 94 takes place. It is operational with a wide variety of either electron-withdrawing or -donating aryl substituents on the alkynylsilane and with different alkyl and halogen substituents on the acylsilane. The reaction proceeds with good to excellent yields (45–93%). The transformation can also be performed with fused acylsilanes and with a silyl(2-thienyl)methanone (Scheme [14]).[51]

Zoom Image
Scheme 14 Synthesis of indenes from in situ generated propargylsilanes

The catalytic cycle is initiated by a silicon–gold transmetalation of the alkynylsilane to form the gold acetylide 95. The latter attacks the activated carbonyl group of the acylsilane 94 and forms the intermediate 96 that is further complexed by the gold catalyst to initiate the silyl migration. The following C–H functionalization affords the final indene 98 (Scheme [14]).

The proof of concept was established by verifying the separate reaction steps. The reaction proceeds with 5 mol% isolated gold acetylide 95, as a catalyst, in the presence of TMSNTf2. Without TMSNTf2 the reaction did not proceed. The migration of the silyl group was confirmed by using the corresponding TBS acylsilane 94. The results showed complete migration of TBS group in the final product. Additionally, (D3C)3Si alkynylsilane 93 was used with TMS acylsilane 94 and showed no scrambling of the silyl­ groups.


# 4.2

Alkynylsilanes

The combination of alkynes with transition-metal catalysis can lead to highly reactive intermediates, such as, metal–vinylidene complexes. In the context of this review their formation is determined by the phenomenon of metal complex back-donation in combination with the 1,2-migration of silyl substituents.

Matsubara and co-workers reported isoindole synthesis from alkynylsilanes 99 and phthalimides 100 via decarbonylative alkylation.[52] The developed catalytic system consists of Ni(cod)2/PMe3 in combination with methylaluminum bis(2,6-di-tert-butyl-4-methylphenoxide) (MAD) as a sterically bulky and monomeric homogeneous Lewis acid. Products 105 were obtained with high E/Z selectivity in 44–90% yield (Scheme [15]).

Zoom Image
Scheme 15 Synthesis of isoindoles from alkynylsilanes

The proposed reaction mechanism starts with the oxidative addition of nickel to the C(O)–N bond followed by decarbonylation. Then, the complexation of the alkyne affords the nickel complex 101. The complexation with MAD as a Lewis acid endorses the formation of the acyclic cationic nickel intermediate 102 and the 1,2-silyl shift promotes the formation of vinylidene complex 103. Further insertion of the vinylidene in the neighboring C–Ni bond affords the six-membered nickelacycle 104 that undergoes reductive elimination to form the desired product 105.

The double bond geometry is established when the vinylidene inserts into the C–Ni bond. It is understood to be controlled by the steric interactions between the silyl group and the phosphine ligand of the nickel intermediate.

The reaction tolerates a wide variety of substituents on the alkynyl substrate, including a multitude of functional groups.

In a similar fashion, Xie and co-workers developed a method where TMS alkynylsilanes can be used in [2+2+1] cycloaddition reactions with unactivated alkenes and carboryne.[53] Alongside the desired [2+2+1] addition product 109, they observed also formation of the [2+2+2] product 110. They suggest that a Zr-catalyzed cycloisomerization provides intermediate 112. The latter undergoes transmetalation with the nickel source affording the cyclic nickel species 113. The alkyne coordinates to intermediate 113 and undergoes 1,2-silyl migration to form the corresponding vinylidene 114. The latter inserts into the Ni–C bond to form the cyclic intermediate 115, which after reductive elimination provides the [2+2+1] addition product 109 (Scheme [16]).

Zoom Image
Scheme 16 Annulation reaction of carboryne, unactivated alkenes, and alkynylsilanes

In the presence of PPh3, the nickelacycle 113 can also enter [2+2+2] cyclotrimerization and provide product 110. This can be altered by choosing right phosphine ligand. In this case, the use of PMe3 leads to the formation of product 109.

Alkynylsilanes 116 similarly to allenylsilanes can be used to obtain furans 119 in a gold-catalyzed transformation (Scheme [17]). Based on DFT calculations, Gevorgyan and co-workers developed counterion and solvent dependent reactions towards silylated furans 119 and 121. Thus, the use of SbF6 directs the mechanism via a possible formation of allene 117 that cyclizes afterwards. On the other hand, TfO in a non-polar solvent favors a direct cyclization to Au–furan 120 which, after protodeauration, leads to the formal 1,2-H migration product 121. However, the use of polar solvents leads to Au–carbene 118, from which the triflate ligand dissociates, thus facilitating 1,2-silyl migration.[43]

Zoom Image
Scheme 17 Synthesis of furans from alkynylsilanes

Indeed, the experimental results agreed with the calculations. The use of Ph3PAuSbF6 smoothly afforded the desired furans 119 with a wide variety of substituents in good to high yields (65–91%).

Zoom Image
Scheme 18 Double migration in the synthesis of furans and pyrroles

Further expansion of this concept to homopropargylic ketones or imines led to reactions involving double rearrangements.[54] When studying gold-catalyzed [(C6F5)3PAuSbF6] reactions with substrates of type 122 possessing a cyclic substituent at propargylic position, the formation of the fused ring 129 was observed (Scheme [18]).[54] It was suggested that this reaction proceeds via an initial activation of the triple bond followed by cyclization to obtain intermediate 124. The latter undergoes alkyl group migration and proton elimination affords intermediate 126. Further protonation of the silylated carbon initiates 1,2-silyl migration and deauration yields product 129.

Similarly, dihydrofurans 132 can be obtained by combining alkynylsilanes 130 and the corresponding aldehydes 131 (Scheme [19]).[55] The reaction tolerates a variety of Ar1 substituents, phenyl and heterocyclic motifs with both electron-withdrawing and -donating effects, affording the desired product in good to excellent yields. However, larger Ar1 groups can affect the reactivity with bulkier silyl groups, like a TBS group. Ar2 groups are required to induce significant electrophilicity on the carbonyl group as benz­aldehyde and aliphatic aldehydes did not participate in this reaction.

Zoom Image
Scheme 19 Synthesis of dihydrofurans from alkynylsilanes

It was proposed that the reaction mechanism is initiated by the formation of gold complex 133. Intramolecular deprotonation yields Au–allene 134 that further reacts with the aldehyde 131 to afford alkyne 135. The latter undergoes a gold-catalyzed dihydrofuran formation resulting from the 1,2-silyl shift in a similar fashion to Scheme [18].

Tanaka and co-workers developed a synthetic procedure similar to the previously described methods for benzofuran and indole synthesis using a rhodium catalyst (Scheme [20a]).[56] Various transition metal complexes could be obtained using a similar reaction, as reported by Wong and co-workers. The first ever isolable Ru(II)–indole zwitterion 139a complex was obtained from aniline-tethered alkynes (Scheme [20b]).[57]

Zoom Image
Scheme 20 (a) Synthesis of benzofurans and indoles from alkynylsilanes. (b) Synthesis of ruthenium complex from alkynylsilanes. (c) Reaction mechanism for rhodium- and ruthenium-catalyzed reactions for synthesis of 137 and 139.

Both reactions follow a similar pattern (Scheme [20c]). First, the transition metal coordinates to the alkyne moiety inducing 1,2-silyl migration and vinylidene 141 formation. Cyclization of the latter affords the desired benzofuran or indole scaffold 142. Depending on the transition metal, the heterocycle undergoes protodemetalation or protodesilylation.

The rhodium-catalyzed reaction 136137 is highly applicable, it tolerates a wide selection of aryl substituents with a broad variety of silyl groups. For the ruthenium-catalyzed reaction, DFT calculations revealed that activation energy for the TMS migration was 7.0 kcal/mol lower than of that of hydrogen for a terminal alkyne. This was also proved synthetically, as N,N-disubstituted indole zwitterion 139a was obtained in 90% yield compared to 35–70% for its counterparts arising from terminal alkynes. Unfortunately, all Ru-catalyzed reactions underwent protodesilylation.

During the total synthesis of acylphloroglucinol, Barriault, Korobkov, and co-workers observed and isolated a vinyl–gold complex that was obtained from 1,2-silyl migration.[58] Treating the starting material 144 with a gold source effectively resulted in the coordination of the triple bond to the gold species. However, further migration to afford intermediate 147 in such a sterically demanding environment was only possible with small to moderate silyl groups. Thus, the TBS group showed a decrease in the reaction yields of the cyclization products 145. Furthermore, the TIPS-containing substrate gave only 35% yield (Scheme [21]).

Zoom Image
Scheme 21 Synthesis of stable gold complex from propargylsilanes

To confirm the intramolecular silyl group migration, an experiment was performed with a mixture of starting materials 144 with two different silyl groups. The crossover products were not observed thus confirming an intra­molecular process.

In addition, boron-based Lewis acids can activate alkynylsilanes similarly to transition metals in a Wrackmeyer reaction modification. Erker and co-workers developed a method where, from 1,2-bis(alkynyl)benzenes 148, the corresponding naphthalenes 149 are synthesized via consecutive 1,1-carboboration and silyl migration reactions. The entire process proceeds easily with high yields (Scheme [22a]).[59] The reaction scope can be further enhanced using heterocyclic bis-silylacetylenes resulting in the corresponding carbazoles, benzothiophenes, and quinolones.[60]

Curran and co-workers developed a method for bis-silyl­acetylene hydroboration reactions with N-heterocyclic carbene-borane that involves a 1,2-silyl migration (Scheme [22b]).[61]

Similarly to the borenium-catalyzed hydroboration of allylsilanes, it was expected that the alkynylsilanes would undergo hydroboration in the typical 1,2-fashion, but only the 1,1-hydroboration product was observed. Using the diprotected TMS,TIPS-acetylene, only a single hydroboration occurred and formed solely the (E)-stereoisomer, as confirmed by 1H-NOESY NMR.


#
# 5

Reactions of Alkenes

5.1

Allylsilanes

The initial reports for rearrangements using allylsilanes 157 came from Knölker and co-workers. The main byproduct of the Hosomi–Sakurai reaction on α,β-unsaturated ketones resulting in the [3+2] annulation product was observed.[62] Further major contributions to this field were done by Knölker,[63] [64] [65] [66] [67] Danheiser,[68] Monti,[69] and Meyers[70] (Scheme [23]).

Zoom Image
Scheme 22 (a) Naphthalene synthesis from alkynylsilanes. (b) Bis-silyl­acetylene hydroboration reactions.
Zoom Image
Scheme 23 [3+2] Annulation reactions of allylsilanes

The catalytic enantioselective annulation between oxindoles 162 and allylsilanes 163 was developed by Franz and co-workers.[71] Scandium(III)/pybox complexes with NaBArF allowed significant reactivity and selectivity for the formation of [3+2] annulation products 164 (Scheme [24]). These products were obtained in high yields (up to 97%), high dia­stereoselectivity (up to 99:1) and enantioselectivity (up to 99.5:0.5). The reaction tolerates a wide variety of oxindole substrates bearing ester and nitrile groups.[71]

Zoom Image
Scheme 24 Catalytic and enantioselective annulation reaction of allylsilanes

Another catalytic stereoselective reaction of allylsilanes was developed by Ohe and co-workers.[72] Five-membered [3+2] annulation products 169 were obtained from unsaturated β-silyl ketones 167 and allylsilanes 168 using Sc(OTf)3 (Scheme [25]).

The reaction proceeded with full conversion yielding the desired annulation product as the main product. The Hosomi–Sakurai pathway resulted in minor product formation. The transformation occurred with a wide variety of β-silyl ketones and allylsilanes bearing different silyl groups and different substituents on the ketone. The reaction with allylsilane 168 bearing the super silyl group [Si(SiMe3)3] afforded the annulation product in 99% yield.[72]

The reaction mechanism starts with the enone activation by the Lewis acid followed by nucleophilic 1,4-addition of allylsilane. Then, in a typical manner, 1,2-silyl migration takes place to allow further cyclization to the corresponding cycle 169.


# 5.2

Vinylsilanes

Lee and co-workers developed a method for the synthesis of allenes 173 from silylcyclopropenes 172. This method employs PtCl2 to perform the ring-opening with a variety of substituted cyclopropenes bearing aliphatic substitution and tolerating unsaturation and remote aryl substitutions. The only limitation to this reaction is substrates bearing unhindered alkene functions or a phenyl ring at close proximity to the Pt complexation site on the cyclopropane ring, possibly explained by interference with the formation of the productive Pt complex (Scheme [26]).[73]

Zoom Image
Scheme 25 Enantioselective cyclopropenyl ketone synthesis from allylsilanes­
Zoom Image
Scheme 26 Allenylsilane synthesis from cyclopropenylsilanes

Two possible pathways were originally proposed as confirmed by 13C labeling, this was reduced to one by DFT calculations. The reaction is initiated by platinum coordination to the double bond affording intermediate 174. Then, the reaction goes directly through 1,2-silyl shift to give intermediate 175 that, after rearrangement and Pt elimination, yields the desired product 173.[74]

An interesting reaction was developed by Ito and co-workers. Their strategy involves an intramolecular disilylation of propargyl silyl ethers 177 followed by ether cleavage under Brønsted or Lewis acidic conditions to ultimately initiate a 1,2-silyl migration yielding propargylsilanes 179 (Scheme [27]).[75]

Zoom Image
Scheme 27 Propargylsilane synthesis from vinylsilanes

The rearrangement is initiated by the activation of the oxygen with TMSOTf. This is followed by the cleavage of the C–O bond which is facilitated by the β-silicon effect and results in 1,2-silyl migration. The reaction has high stereoretention and, hence, it is hypothesized that the rearrangement takes place in a concerted manner as the syn-periplanar stabilization should be responsible for the syn-migration. The 1,2-silyl migration is also enhanced by the two sterically repulsive geminal TMS groups and by stabilizing the resulting cation by the β-effect. Finally, nucleo­phile-induced mono-desilylation provides the product 179.


#
# 6

Conclusions

This short review highlights the importance of the β-silicon effect in reactions involving 1,2-silyl migration. The interactions of the high energy β-C–Si σ-bond with the empty p orbital of a electron deficient carbon have been extensively studied. In conclusion for symmetrical systems the non-vertical stabilization (closed – silonium ion) is dominant. For non-symmetrical systems, both non-vertical and vertical (open – hyperconjugation) stabilization are possible. The more electron deficient the α-carbon and less electron deficient the β-carbon, the more dominant the non-vertical structure is and vice versa. In more complex systems, hyperconjugation can still play a crucial role in the stabilization of carbenium ions. In the reviewed examples the combination of multiple effects can explain why many reactions tend to undergo 1,2-silyl shift via a pseudo-non-vertically stabilized transition state.

Strategies involving 1,2-silyl shift can result in the formation of complex cyclic or unsaturated structures conveniently equipped with a handle for further functionalization. Currently, research of high interest revolves around the synthesis of metal–vinylidene complexes, where the 1,2-silyl migration plays a crucial role. Several experimental and theoretical studies show the importance of silicon migration for improved reactivities due to the reduction of their activation energies. For these reasons, further development in this area of organosilicon chemistry is expected in the future.


#
#

Acknowledgement

The authors thank Ms. Stephanie G. E. Amos for helpful discussions.



Zoom Image
Rūdolfs Beļaunieks(left) obtained his B.Sc. and M.Sc. from Riga Technical University (RTU) in 2017 and 2019, respectively. Currently he is pursuing his doctoral degree under the supervision of Prof. M. Turks. His research is focused on novel synthetic methodologies involving 1,2-silyl shift.
Mikus Puriņš (middle) obtained his B.Sc. and M.Sc. from RTU under the supervision of Prof. M. Turks in 2017 and 2019, respectively. Currently he is pursuing his doctoral degree under the supervision of Professor Jérôme Waser at Swiss Federal Institute of Technology, Lausanne (EPFL), Switzerland. His research interests include tethered strategies for selective functionalization of small molecules.
Professor Māris Turks (right) obtained his B.Sc. in chemistry from the University of Latvia in 1998 and his M.Sc. from Riga Technical University in 2000. In 2005 he earned his Dr. ès. sc. degree in synthetic organic chemistry from EPFL, Switzerland under the direction of Professor Pierre Vogel. Then he spent one year as SNSF Postdoctoral Fellow at Stanford University in the group of Professor Barry M. Trost. In 2007, he accepted an academic position at the Faculty of Materials Science and Applied Chemistry, RTU, where he is currently Dean of Faculty and Director of the Institute of Technology of Organic Chemistry. Prof. Turks’ current research interests are in organosilicon chemistry, application of sulfur dioxide in organic synthesis, and purine and triterpenoid chemistry.
Zoom Image
Scheme 1 Hydrolysis of silylalkyl chlorides
Zoom Image
Scheme 2 Vertical and non-vertical stabilization of carbenium ions by β-silicon substituent. Here and further in text substituent ‘ Si ’ is used as generic description of the R3Si group (R = alkyl and/or aryl).
Zoom Image
Figure 1 Comparison of solvolysis rate constants in CF3CH2OH at 25 °C (a, d, and e are extrapolated from experiments at other temperatures)
Zoom Image
Figure 2 Calculation of stabilization energy for different β-silylcarbenium systems
Zoom Image
Figure 3 Kinetic studies of transition states of β-silylcarbenium systems
Zoom Image
Equation 1 Yukawa–Tsuno equation: k x = reaction rate constant for X-substituted arylsilane; k h = reaction rate constant for unsubstituted phenylsilane, ρ = reaction constant, σ° = Hammett substituent constant; r + = enhanced resonance parameter, Δσ+ R = Hammett resonance substituent constant.
Zoom Image
Figure 4 Stabilization of vinyl cation by β-silicon effect
Zoom Image
Scheme 3 Formation of silylvinyl cations for studying effects of hyperconjugation
Zoom Image
Scheme 4 Formation of β,β-silyl disubstituted vinyl cations
Zoom Image
Scheme 5 Cyclopentene annulation reaction; the Danheiser annulation
Zoom Image
Scheme 6 Reaction and plausible mechanism of [3+3] and [3+2] addition of cyclopropanes and allenylsilanes
Zoom Image
Scheme 7 Reaction and mechanism of cyclobutanone and allenylsilanes
Zoom Image
Scheme 8 Gold-catalyzed furan synthesis with 1,2-silyl migration
Zoom Image
Scheme 9 Synthesis of 5-membered rings from propargylsilanes
Zoom Image
Scheme 10 Epoxide synthesis from propargylsilanes
Zoom Image
Figure 5 Structure of Al(III)–sal-BINAM
Zoom Image
Scheme 11 Synthesis of highly functionalized alkenes from propargylsilanes (DMPS = SiMe2Ph)
Zoom Image
Scheme 13 Synthesis of dienes and indenes from propargylsilanes
Zoom Image
Scheme 12 Synthesis of furans from propargylsilanes
Zoom Image
Scheme 14 Synthesis of indenes from in situ generated propargylsilanes
Zoom Image
Scheme 15 Synthesis of isoindoles from alkynylsilanes
Zoom Image
Scheme 16 Annulation reaction of carboryne, unactivated alkenes, and alkynylsilanes
Zoom Image
Scheme 17 Synthesis of furans from alkynylsilanes
Zoom Image
Scheme 18 Double migration in the synthesis of furans and pyrroles
Zoom Image
Scheme 19 Synthesis of dihydrofurans from alkynylsilanes
Zoom Image
Scheme 20 (a) Synthesis of benzofurans and indoles from alkynylsilanes. (b) Synthesis of ruthenium complex from alkynylsilanes. (c) Reaction mechanism for rhodium- and ruthenium-catalyzed reactions for synthesis of 137 and 139.
Zoom Image
Scheme 21 Synthesis of stable gold complex from propargylsilanes
Zoom Image
Scheme 22 (a) Naphthalene synthesis from alkynylsilanes. (b) Bis-silyl­acetylene hydroboration reactions.
Zoom Image
Scheme 23 [3+2] Annulation reactions of allylsilanes
Zoom Image
Scheme 24 Catalytic and enantioselective annulation reaction of allylsilanes
Zoom Image
Scheme 25 Enantioselective cyclopropenyl ketone synthesis from allylsilanes­
Zoom Image
Scheme 26 Allenylsilane synthesis from cyclopropenylsilanes
Zoom Image
Scheme 27 Propargylsilane synthesis from vinylsilanes