Stereochemical outcomes of C-F activation reactions of benzyl fluoride

Article abstract of DOI:10.3762/bjoc.14.6

In recent years, the highly polar C-F bond has been utilised in activation chemistry despite its low reactivity to traditional nucleophiles, when compared to other C-X halogen bonds. Paquin's group has reported extensive studies on the C-F activation of benzylic fluorides for nucleophilic substitutions and Friedel-Crafts reactions, using a range of hydrogen bond donors such as water, triols or hexafluoroisopropanol (HFIP) as the activators. This study examines the stereointegrity of the C-F activation reaction through the use of an enantiopure isotopomer of benzyl fluoride to identify whether the reaction conditions favour a dissociative (SN¹) or associative (SN²) pathway. [²H]-Isotopomer ratios in the reactions were assayed using the Courtieu ²H NMR method in a chiral liquid crystal (poly-γ-benzyl-L-glutamate) matrix and demonstrated that both associative and dissociative pathways operate to varying degrees, according to the nature of the nucleophile and the hydrogen bond donor.

Full text of DOI:10.3762/bjoc.14.6

Stereochemical outcomes of C–F activation reactions of
benzyl fluoride
Neil S. Keddie‡1, Pier Alexandre Champagne‡2, Justine Desroches2,
Jean-François Paquin*2 and David O'Hagan*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2018, 14, 106–113.
1School of Chemistry, Biomedical Sciences Research Complex,
University of St Andrews, North Haugh, St Andrews, Fife KY16 9ST,
United Kingdom and 2PROTEO, CCVC, Département de chimie,
1045 Avenue de la Médecine, Université Laval, Québec, QC G1V
0A6, Canada
doi:10.3762/bjoc.14.6
Received: 20 October 2017
Accepted: 12 December 2017
Published: 09 January 2018
Email:
This article is part of the Thematic Series "Organo-fluorine chemistry IV".
Associate Editor: J. A. Murphy
Jean-François Paquin* - jean-francois.paquin@chm.ulaval.ca;
David O'Hagan* - do1@st-andrews.ac.uk
* Corresponding author ‡ Equal contributors
© 2018 Keddie et al.; licensee Beilstein-Institut.
License and terms: see end of document.
Keywords:
benzylic fluorides; C–F activation; chiral liquid crystal; 2H NMR;
PBLG; stereochemistry
Abstract
In recent years, the highly polar C–F bond has been utilised in activation chemistry despite its low reactivity to traditional nucleo-
philes, when compared to other C–X halogen bonds. Paquin’s group has reported extensive studies on the C–F activation of
benzylic fluorides for nucleophilic substitutions and Friedel–Crafts reactions, using a range of hydrogen bond donors such as water,
triols or hexafluoroisopropanol (HFIP) as the activators. This study examines the stereointegrity of the C–F activation reaction
through the use of an enantiopure isotopomer of benzyl fluoride to identify whether the reaction conditions favour a dissociative
(SN1) or associative (SN2) pathway. [2H]-Isotopomer ratios in the reactions were assayed using the Courtieu 2H NMR method in a
chiral liquid crystal (poly-γ-benzyl-L-glutamate) matrix and demonstrated that both associative and dissociative pathways operate
to varying degrees, according to the nature of the nucleophile and the hydrogen bond donor.
Introduction
The C–F bond is the strongest carbon–halogen bond known [1]. in C–F bond activation [2], with a view to using organic bound
Its low reactivity, in comparison to other C–X bonds, means fluoride as a leaving group in substitution reactions that typical-
that it is inert to all but the most harsh reaction conditions, and ly require more activated leaving groups. Such an approach
fluorine can generally be carried through multistep syntheses could circumvent the requirement for protecting groups in
without concern over side reactions (the exception being SNAr multistep synthesis by capitalizing on the low reactivity of the
reactions). In recent years, there has been an increasing interest C–F bond. Paquin et al. have published extensively on non-
106

Beilstein J. Org. Chem. 2018, 14, 106–113.
metal based methods for benzylic C–F bond activation [3-7]. center and perfect enantiospecificity, while an SN1 mechanism
The reactivity relies on protic activation driven by the capacity would yield a fully racemized product. Any mixed pathway
of organic fluoride to form hydrogen bonds [8,9]. Protocols would generate products with partially racemized stereocenters.
using water/isopropanol [3], optimally coordinated triols [4,5], In this context, we decided to explore the stereointegrity of the
and hexafluoroisopropanol (HFIP) [6,7] as the corresponding aforementioned reactions using enantiopure 7-[2H1]-(R)-benzyl
hydrogen bond donors have shown considerable success. This fluoride ((R)-1, Figure 2) as a primary, yet chiral electrophile
mode of activation has been demonstrated for amination [3-5] [10].
and Friedel–Crafts reactions [6,7] on benzylic fluoride sub-
strates (Figure 1), producing the corresponding substituted
products in moderate to good yields. The water/isopropanol
system was also shown to be amenable to phenolate and thio-
late nucleophiles [3].
Figure 2: 7-[2H1]-(R)-Benzyl fluoride ((R)-1).
Previously, Paquin et al. undertook density functional theory
(DFT) studies on the mechanism of C–F amination reactions
employing water/isopropanol [3] and triols [4,5] as hydrogen- Substitution reactions of benzyl fluoride (1) will generate
bond donor activators. Through these studies, the authors sug- substituted products that retain the deuterium atom, and the
gested that multiple donors (even when using a triol) surround degree of stereointegrity can be determined by examining the
the fluorine atom of the benzyl fluoride, thus stabilising the enantiopurity of the isotopically labelled product. Quadrupolar
transition state through substantial F···HOR hydrogen bond 2H-nuclei can serve as a particularly useful NMR probe for
interactions, rather than through electrostatic stabilisation only assaying enantiopurity. If the 2H NMR is recorded in a
[3]. This stabilisation was suggested to lead to a purely associa- lyotropic liquid crystalline solvent, where tumbling of the solute
tive bimolecular (SN2) mechanism. The authors also studied the is restricted, then the 2H NMR signal splits into a doublet due to
C–F activated Friedel–Crafts reactions [6,7] using very strong differential interactions of the quadrupolar nuclei with the elec-
hydrogen bond donors, namely HFIP, in the presence or tric field gradient associated with the oriented media [10].
absence of trifluoroacetic acid (TFA). For both of these activa- When placed in an enantiomerically enriched liquid crystalline
tors, Paquin et al. proposed a dissociative unimolecular (SN1) environment, the enantiomeric isotopomers interact unequally
mechanism, whereby the strong hydrogen bond donor associ- with the electric field gradients associated with the orientated
ates with the benzyl fluoride, leading to ionisation of the mole- media, creating anisotropy and resolving into two sets of
cule, generating a benzylic carbocation and a formal equivalent doublets. If there is sufficient resolution between these
of HF (which behaves in an autocatalytic manner as a stronger quadrupolar couplings, then the enantiomeric ratio can be re-
hydrogen bond donor than HFIP or TFA).
corded. We have used poly-γ-benzyl-L-glutamate (PBLG) pre-
viously as the liquid-crystalline matrix for the determination of
Overall, there are three possible mechanistic pathways that ee of samples of deuterated benzyl alcohols, benzyl fluorides,
these C–F activation reactions could follow: SN1, SN2, and a and esters of fluoroacetic acid [11] by 2H NMR and found it to
mixed SN1/SN2 pathway. Typically, benzylic substitutions be effective for the resolution of enantiomers. In this study, we
would be expected to display a significant level of SN1 char- explore various nucleophilic substitutions and a Friedel–Crafts
acter. However, given the particularly poor properties of fluo- reaction on enantiomerically labelled [2H1]benzyl fluoride.
ride as a leaving group, developing a better understanding of the
dissociative nature of these transformations remains of consid- Results and Discussion
erable interest. A direct bimolecular SN2 substitution would Highly enantiomerically enriched 7-[2H1]-(R)-benzyl fluoride
result in a complete inversion of configuration of the stereo- ((R)-1) was synthesised in two steps from benzaldehyde (2), as
Figure 1: C–F activation of benzylic fluorides to generate benzylamine or diarylmethane products.
107

Beilstein J. Org. Chem. 2018, 14, 106–113.
described previously [11]; the procedure is summarised in hydrogen bond donor. In addition, three reactions of racemic
Scheme 1.
substrates (Table 1, entries 1–3), were performed in order to
ensure that sufficient resolution could be obtained in the
Aldehyde 2 was reduced under Noyori’s conditions [12] using 2H{1H}-PBLG-NMR, therefore allowing the ee of the products
(S,S)-Ru(DPEN)2 as catalyst and [2H2]-formic acid as the to be determined. A representative example of the 2H NMR
deuterium source. This afforded the corresponding 7-[2H1]-(S)- spectra (107.5 MHz) is displayed in Figure 3, using N-methyl-
benzyl alcohol ((S)-3) in moderate yield (81%) and high ee aniline as a nucleophile, showing the spectra of both a racemic
(95%), as evidenced by 2H-PBLG-NMR. Benzyl alcohol 3 was sample (Figure 3A) and an enantioenriched sample (Figure 3B)
converted to the corresponding benzyl fluoride (1) using a mod- of 6. However, as evidenced by entry 3 (Table 1), and entries 3
ification of Bio’s method [13,14] to promote the SN2 reaction and 7 (Table 2), the analysis revealed that some nucleophiles
exclusively, using TMS-morpholine and DAST, in moderate (such as N-methylbenzylamine and morpholine [not shown])
yield (51%) and high ee (94%).
were unsuitable for this study, as the resulting products 7 could
not be resolved in the 2H NMR with PBLG assay.
The isotopically enriched [2H1]-benzyl fluoride ((R)-1, 95% ee)
was then subjected to a range of C–F activation reactions using Two different activator systems were investigated for the
a mixture of nucleophiles (for direct substitutions) and aryls (for nucleophilic substitution of 1: a mixture of water and
Friedel–Crafts reactions) to give products 5–9. The nucleo- isopropanol (Table 2, entries 1–5) and tris(hydroxymethyl)pro-
philic substitution reactions of 1 are shown in Table 1 and pane (Table 2, entries 6–8). Using water/isopropanol as the acti-
Table 2, and were all conducted using either a mixture of water/ vator afforded the benzylated products 5–9 in moderate yields
isopropanol, or tris(hydroxymethyl)propane as the activating after 18 h. The ee values of all of the resulting products was
Scheme 1: Synthesis of enantioenriched 7-[2H1]-(R)-benzyl fluoride ((R)-1) from benzaldehyde (2).
Table 1: Nucleophilic substitution reactions of racemic 7-[2H1]-benzyl bromide (4).
Entry
Reaction
ee (%)
1
racemic
2
3
racemic
nda
aee could not be determined as a result of poor 2H{1H} NMR resolution.
108

Beilstein J. Org. Chem. 2018, 14, 106–113.
Table 2: Nucleophilic substitution reactions of 7-[2H1]-(R)-benzyl fluoride ((R)-1).
Entry
Reaction
ee (%)
1
94
2
3
4
5
90
nda
93
91
6
7
8
87
nda
89
aee could not be determined as a result of poor 2H{1H} NMR resolution.
very close to that of the original benzyl fluoride ((R)-1, 95%), on a coordinate anti to the C–F bond resulting in an inversion of
indicating that a highly associative SN2-like pathway was oper- the configuration. These results are in good agreement with the
ating, where the incoming nucleophile must have approached transition state proposed by Paquin [3-5]. Unfortunately,
109

Beilstein J. Org. Chem. 2018, 14, 106–113.
Figure 3: Partial 2H{1H} NMR (107.5 MHz) with PBLG in CHCl3 (13% w/w). (A) racemic sample of 6 (from Table 1, entry 2) and (B) enantioenriched
sample of 6 (from Table 2, entry 2). The magnitude of the quadrupolar splittings for the (R)- and (S)-enantiomers are labelled ΔνQ(ent). The ee of each
sample was determined by deconvolution of the line shapes and subsequent integration.
N-methylbenzylamine (Table 2, entry 3) afforded a product 7 At room temperature (Table 3, entry 1), benzyl fluoride (R)-1
that did not resolve by 2H NMR, and thus the ee could not be was activated by HFIP, affording biarylmethane 10 in a good
determined.
yield (88%) after 18 h. The ee of the product was low (24%),
but not racemic. The proposed stereochemistry of the product
Changing the activator from water/isopropanol to tris(hydroxy- was verified by independent synthesis of the (S)-isomer from
methyl)propane was anticipated to increase the stability of the the unsymmetric diphenyl ketone 11 (Scheme 2).
triol–benzyl fluoride complex, and hence a tendancy towards an
associative mechanism was expected. However, on performing Corey–Bakshi–Shibata reduction of diaryl ketone 11, afforded
the reactions with nitrogen nucleophiles (Table 2, entries 6 and the (R)-alcohol 12 in moderate to good yield and moderate ee
8) and the triol as the hydrogen bond donor, slightly lower ee’s [15,16]. The absolute stereochemistry of 12 was confirmed by
were obtained relative to those obtained using the same nucleo- X-ray crystallography of the 4-bromophenyl ester derivative 13.
philes with the water/isopropanol system (Table 2, entries 2 and Alcohol 12 was activated as the tosyl ester at −20 °C, and then
5). These minor differences in ee may be due to the higher tem- immediately displaced by LiAlD4 [17], inverting the stereo-
perature leading to a minor, but noticeable dissociative path- center to afford the (S)-diarylmethane 10 isotopomer in 18% ee.
way. Once again, using N-methylbenzylamine as the nucleo- 2H{1H} NMR in a PBLG matrix indicated that the dominant
phile (Table 2, entry 7) afforded 7, which could not be resolved isomer was the same as was produced in entry 1, Table 3.
by 2H NMR. Overall, the nucleophilic substitution of 1, using Therefore, this analysis showed that the dominant enantiomer of
either of the described hydrogen bond activating systems, 10 arose from an inversion, rather than retention, of configura-
afforded enantioenriched benzylated products with little erosion tion of the original stereocenter of 1.
in stereointegrity.
There may be four different reaction mechanisms operating in
In contrast to the above nucleophilic substitutions, which all these Friedel–Crafts reactions as shown in Figure 4. (A) Coordi-
proceeded with good stereointegrity, the Friedel–Crafts reac- nation of the fluorine atom with the hydrogen bond donor, fol-
tions of 1 with p-xylene gave very different results, as shown in lowed by backside attack of the nucleophile leads to SN2 reac-
Table 3.
tion and inversion of configuration. (B) Hydrogen bond donor
110

Beilstein J. Org. Chem. 2018, 14, 106–113.
Table 3: Friedel–Crafts reactions of 7-[2H1]-(R)-benzyl fluoride ((R)-1).
Entry
Reaction
ee (%)
1
2
24
19
28
3
Scheme 2: Synthesis of enantioenriched (S)-diarylmethane 10 from diaryl ketone 11 and confirmation of configuration of (R)-13 by single crystal
X-ray structure.
coordination to fluorine leads to ionisation of 1, producing an poor and k4 > k3, a solvent-separated ion pair will be formed,
intimate ion pair, which only permits backside attack of the where the HBD-coordinated fluorine atom is loosely associated
nucleophile on the benzylic cation. (C) If the nucleophile is with the solvated cation, allowing a nucelophilic attack to occur
111

Beilstein J. Org. Chem. 2018, 14, 106–113.
Figure 4: Possible reactive intermediates for C–F activation of benzyl fluoride 1 with strong hydrogen bond donors.
from more trajectories, leading to a mixture of inversion suggesting that at lower temperatures the separation of the ions
(predominant) and retention products. (D) Fully solvated is less favoured in solution, presumably for entropic reasons.
cation, where attack of the nucleophile can freely occur from
either face, leading to racemization of the product in an SN1 The nature of the poorer nucleophile, coupled with the stronger
reaction.
hydrogen bond donor in the Friedel–Crafts reaction allows the
solvent-separated ion-pair mechanism to predominate, signifi-
We propose that the actual attack of the nucleophile does not cantly eroding the stereointegrity of the biarylmethane products
occur on the coordinated benzyl fluoride (A), or the fully 10. However, the products were not racemic, showing that the
solvated carbocation (D), as these scenarios would incur 100% nucleophilic attack also occurs via an associated ion pair, rather
or 0% enantiospecificity, respectively. Rather, the data suggest than the fully solvated carbocation.
that attack occurs on a mixture of intimate (B) and solvent-sepa-
rated (C) ion pairs. The partial racemization observed in Table 3 Conclusion
suggests that the solvent-separated ion-pair intermediate (C) is In summary, we have analyzed the stereochemical outcomes of
most likely the reactive species, as it would naturally lead to a substitution and Friedel–Crafts reactions of 7-[2H1]-(R)-benzyl
partial racemization of the substrate stereocenter.
fluoride ((R)-1), mediated by C–F activation using hydrogen-
bond donors. When strong nucleophiles are used in conjunction
When the activator was changed from HFIP to a mixed system with hydroxyl-based donors, an associative SN2-like reaction
of HFIP and TFA (3 mol %, Table 3, entries 2 and 3), the reac- mechanism predominates, with almost complete inversion of
tions were complete in a significantly shorter time, i.e., the the configuration at the stereogenic center. Poorer aryl nucleo-
initial induction period observed when only HFIP was used [7] philes can be used for Friedel–Crafts reactions if strong hydro-
disappeared in each case. The ee of entry 2 was lower (19%), gen bond donors (such as HFIP or TFA) are used to activate the
showing that the increased hydrogen bonding strength of the C–F bond. In these cases, a dissociative mechanism operates,
TFA, and thus the more rapid generation of HF (vide infra), probably via a solvent-separated ion pair, rather than a fully
promotes a dissociative pathway via the solvent-separated ion solvated benzylic carbocation. The products arising from this
pair. The greater ionic strength of the solution may also play a mechanism are only partially enantioenriched, suggesting
part in stabilising the partially dissociated carbocation. Pleas- that there is still a steric influence for backside attack of the
ingly, decreasing the temperature (Table 3, entry 3) did not slow nucleophile in the solvent-separated ion pair, arising from the
the reaction down, and it completed after 3 h. However, the de- large, congested hydrogen bond networks around the fluorine
creased temperature lead to a slightly higher ee (28%) for 10, atom.
112

Beilstein J. Org. Chem. 2018, 14, 106–113.
9. Champagne, P. A.; Desroches, J.; Paquin, J.-F. Synthesis 2015, 47,
306–322. doi:10.1055/s-0034-1379537
Supporting Information
10.Canet, I.; Courtieu, J.; Loewenstein, A.; Meddour, A.; Pechine, J. M.
J. Am. Chem. Soc. 1995, 117, 6520–6526. doi:10.1021/ja00129a015
11.Wadoux, R. D. P.; Lin, X.; Keddie, N. S.; O’Hagan, D.
Tetrahedron: Asymmetry 2013, 24, 719–723.
The Supporting Information features experimental
protocols and 1H, 19F (where appropriate) and 2H{1H}
NMR spectra of benzyl fluoride 1 and adducts 5–10. The
methods for measurement of the ee by 2H{1H} NMR are
also described.
doi:10.1016/j.tetasy.2013.05.001
12.Yamada, I.; Noyori, R. Org. Lett. 2000, 2, 3425–3427.
doi:10.1021/ol0002119
13.Bio, M. M.; Waters, M.; Javadi, G.; Song, Z. J.; Zhang, F.; Thomas, D.
Synthesis 2008, 891–896. doi:10.1055/s-2008-1032181
14.Bresciani, S.; O’Hagan, D. Tetrahedron Lett. 2010, 51, 5795–5797.
doi:10.1016/j.tetlet.2010.08.104
Supporting Information File 1
Experimental protocols.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-14-6-S1.pdf]
15.Corey, E. J.; Helal, C. J. Tetrahedron Lett. 1996, 37, 5675–5678.
doi:10.1016/0040-4039(96)01198-7
16.The enantiomeric excess varied slightly when the reaction was run on
different scales, due to variations in temperature.
Supporting Information File 2
2H NMR analysis of enantiopurity.
17.Bolshan, Y.; Chen, C.-y.; Chilenski, J. R.; Gosselin, F.; Mathre, D. J.;
O'Shea, P. D.; Roy, A.; Tillyer, R. D. Org. Lett. 2003, 6, 111–114.
doi:10.1021/ol0361655
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-14-6-S2.pdf]
License and Terms
Acknowledgements
NK and DOH acknowledge support from the University of St
Andrews, Engineering and Physical Sciences Research Council
(EPSRC, Grant No.: EP/L017911/1), and the EPSRC UK
National Mass Spectrometry Facility at Swansea University.
This work was also supported by the Natural Sciences and
Engineering Research Council of Canada (NSERC), the
FRQNT Centre in Green Chemistry and Catalysis (CGCC), and
the Université Laval.
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
ORCID® iDs
Neil S. Keddie - https://orcid.org/0000-0002-9502-5862
The definitive version of this article is the electronic one
which can be found at:
Jean-François Paquin - https://orcid.org/0000-0003-2412-3083
doi:10.3762/bjoc.14.6
References
1. Blanksby, S. J.; Ellison, G. B. Acc. Chem. Res. 2003, 36, 255–263.
doi:10.1021/ar020230d
2. Amii, H.; Uneyama, K. Chem. Rev. 2009, 109, 2119–2183.
doi:10.1021/cr800388c
3. Champagne, P. A.; Pomarole, J.; Thérien, M.-E.; Benhassine, Y.;
Beaulieu, S.; Legault, C. Y.; Paquin, J.-F. Org. Lett. 2013, 15,
2210–2213. doi:10.1021/ol400765a
4. Champagne, P. A.; Saint-Martin, A.; Drouin, M.; Paquin, J.-F.
Beilstein J. Org. Chem. 2013, 9, 2451–2456. doi:10.3762/bjoc.9.283
5. Champagne, P. A.; Drouin, M.; Legault, C. Y.; Audubert, C.;
Paquin, J.-F. J. Fluorine Chem. 2015, 171, 113–119.
doi:10.1016/j.jfluchem.2014.08.018
6. Champagne, P. A.; Benhassine, Y.; Desroches, J.; Paquin, J.-F.
Angew. Chem., Int. Ed. 2014, 53, 13835–13839.
doi:10.1002/anie.201406088
7. Hemelaere, R.; Champagne, P. A.; Desroches, J.; Paquin, J.-F.
J. Fluorine Chem. 2016, 190, 1–6. doi:10.1016/j.jfluchem.2016.08.003
8. Schneider, H.-J. Chem. Sci. 2012, 3, 1381–1394.
doi:10.1039/c2sc00764a
113