On the effect of backbone modifications in 3,3-dimethyl-1-(trifluoromethyl)-3H-1λ3,2-benziodaoxole

Article abstract of DOI:10.1002/ejoc.201403634

We report on the effect of small side-chain modifications to the structure of 3,3-dimethyl-1-(trifluoromethyl)-3H-1λ³,2-benziodaoxole (1b) on its reactivity, as expressed by the initial rate v₀ in a model reaction, and show how the latter can be successfully correlated to an easily determined physical parameter p, a ¹³C NMR chemical shift. The relationship v₀～ p is already present in the simplest starting material devoid of the hypervalent bond and the iodine core and, therefore, presents an interesting approach towards the future scaffold-optimization of this class of reagents. The reactivity of hypervalent-iodine-based trifluoromethylating agents, as expressed by the initial rate v₀ in a model reaction, correlates to an easily determined physical parameter p, a ¹³C NMR chemical shift.

Full text of DOI:10.1002/ejoc.201403634

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
FULL PAPER
DOI: 10.1002/ejoc.201403634
On the Effect of Backbone Modifications in 3,3-Dimethyl-1-(trifluoromethyl)-
3H-1λ3,2-benziodaoxole
Nico Santschi,*[a][‡] Coraline Matthey,[a] Rino Schwenk,[a][‡‡] Elisabeth Otth,[a][‡‡] and
Antonio Togni*[a]
Keywords: Hypervalent compounds
/ Kinetics / Reactivity predictions / Electrophilic addition / Iodine /
Trifluoromethylation
We report on the effect of small side-chain modifications to
the structure of 3,3-dimethyl-1-(trifluoromethyl)-3H-1λ3,2-
benziodaoxole (1b) on its reactivity, as expressed by the ini-
tial rate v0 in a model reaction, and show how the latter can
be successfully correlated to an easily determined physical
parameter p, a 13C NMR chemical shift. The relationship v0
~ p is already present in the simplest starting material devoid
of the hypervalent bond and the iodine core and, therefore,
presents an interesting approach towards the future scaffold-
optimization of this class of reagents.
Introduction
Hypervalent iodine compounds have become increas-
ingly popular in the synthetic community, as they provide
a stable scaffold for several electrophilic-group transfer re-
actions. For example, recent years have seen the rise of elec-
trophilic trifluoromethylating,[1] alkynylating,[2] and azide-
transfer agents[3] and their widespread applications in syn-
thesis. To further establish this chemistry, fundamental in-
formation relating the reactivity of the reagent to the struc-
ture of the scaffold is desirable. In 2010, the first attempts
were undertaken in this direction by comparison of the data
derived from solid-state structures (bond lengths) and
the initial rates (v0) determined in the model reaction of
para-toluenesulfonic acid (pTsOH) with a variety of struc-
turally diverse electrophilic trifluoromethylating reagents
(Scheme 1).[4]
Scheme 1. Hypervalent iodine trifluoromethylating agents 1b–6b
and model reaction for the assessment of their reactivity.
Although vague trends could be elucidated, no simple
descriptor correlating with the reactivity v0 was found that
would allow for a directed optimization of the scaffold.
For example, the I–Cl bond lengths in the precursor chloro-
iodinanes 1a–6a are loosely associated with the I–CF3 bond
lengths in the final reagents 1b–6b. More recently, a reexam-
ination of the original data suggested that the difference in
the I–X (X = Cl or CF3) and I–O bond lengths, ΔI–X/I–O, is
a suitable measure of the reactivity as it reflects the bonding
situation in the hypervalent motif.[1b] In particular, when
comparing two reagents, a larger ΔI–X/I–O value corresponds
to lower p character of the I–X bond and, thereby, leads to
a more-labile reagent. Unfortunately, however, the pre-
viously reported reagents 1b–6b are structurally too diverse
to be described by a tabulated classical physicochemical pa-
rameter.
Therefore, an additional small library was required with
modifications to the backbone minimal such that the Taft
inductive parameters σ* could be used to adequately de-
scribe these new structures and uncover a relationship of
the form σ* ~ p.[5] Therein, p is an easily determined, struc-
ture-related parameter, for example, a specific chemical
shift, that accurately describes the electronic features of the
reagent. Ideally, p should be solution-phase-based to cir-
cumvent the problems associated with crystal-structure-
[a] Department of Chemistry and Applied Biosciences,
Swiss Federal Institute of Technology, ETH Zürich,
Vladimir-Prelog-Weg 2, 8093 Zürich, Switzerland
http://www.tognigroup.ethz.ch
[‡] Current address: Organisch-Chemisches Institut, Westfälische
Wilhelms-Universität Münster,
Corrensstrasse 40, 48149 Münster, Germany
E-mail: santschi@uni-muenster.de
[‡‡] Crystallographic work.
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201403634.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
N. Santschi, C. Matthey, R. Schwenk, E. Otth, A. Togni
FULL PAPER
based reactivity analysis such as errors caused by packing
The oxidations of 7b–11b to the corresponding 1-chloro-
effects (vide infra). Consequently, by determining p for all 1λ3,2-benziodaoxoles (7c–11c) with tBuOCl in dichloro-
available reagents, including 1b–6b, and comparing p versus methane (DCM) at 0 °C again proceeded smoothly in 73–
the initial rates v0 observed in the model trifluoromethyl- 86% yield (Table 1).[1a] Finally, the introduction of the CF3
ation of pTsOH, a structure–reactivity relationship should moiety was achieved through in situ preparation of the cor-
result and unveil the impact of backbone modifications.
responding 1-fluoro-1λ3,2-benziodaoxoles by treatment of
7c–10c with a large excess of anhydrous KF in dry MeCN
overnight, followed by the addition of Ruppert–Prakash
reagent [trifluoromethyltrimethylsilane (TMSCF3)] at
0 °C.[7] Then, the crude reagents 7d–10d were purified by
high-vacuum sublimation followed by recrystallization from
pentane at low temperature to afford the pure products in
35–61% yield (Table 1). 1-Chloro-1λ3,2-benziodaoxole 11c
was not further converted into the corresponding reagent,
as only small amounts were available owing to the poor
yield of the ortho-iodination (11%). Thus, four new reagents
(7d–10d) were accessed with overall yields of 8–14% over
four steps. The ortho-iodination was found to be a limiting
factor but still outperformed strategies relying either on
Grignard additions to 2Ј-iodoacetophenone or metalation
of 1,2-diiodobenzene followed by quenching with appropri-
ate ketones. Overall, the reagents 7d–10d were accessed in
gram amounts, sufficient for additional kinetic studies.
Results and Discussion
Synthesis of Reagents 7d–10d
The key iodinated alcohols 7b–11b were accessed by a
slightly modified Seebach ortho-lithiation protocol starting
from the tertiary alcohols 7a–11a (Scheme 2).[6] Hence,
bromobenzene was lithiated with nBuLi in tetrahydrofuran
(THF) at –78 °C and quenched with the appropriate ketone
to afford the desired alcohols in good-to-excellent yields
(Table 1). Subsequently, 7a–11a were treated with nBuLi
and tetramethylethylenediamine (TMEDA) overnight. Af-
ter the further addition of nBuLi and stirring at room tem-
perature, the reactions were quenched with solid iodine.
Whereas the polarities of alcohols 7a–11a and the corre-
sponding iodine-functionalized compounds 7b–11b are al-
most identical and, therefore, render chromatographic puri-
fication tedious, the differences in the boiling points by ca.
50 °C render the products amenable to purification by dis-
tillation. Consequently, column chromatography was only
applied for smaller-scale experiments (Ͻ2 g of starting
material) and fractional high-vacuum distillation was em-
ployed for larger-scale experiments (ca. 10 g of starting ma-
terial). Thus, iodoalcohols 7b–11b were obtained in 11–38%
yield (Table 1).
X-ray Analysis of Chlorides 7c–11c
Compounds 7c–11c were subjected to solid-state analysis
by X-ray crystallography, as the bond length difference
ΔI–Cl/I–O was previously proposed as a rough measure of
reactivity.[1b] Gratifyingly, suitable crystals were readily
grown by the vapor diffusion of pentane into saturated
solutions of the chlorides in diethyl ether. The important
bond lengths I–Cl and I–O, ΔI–Cl/I–O, and the O–I–Cl bond
angle are given in Table 2, and ORTEP representations are
shown in Figure 1.
Table 2. Bond lengths, bond length differences, and angles of 7c–
11c.
R1
I–O [Å]
I–Cl [Å]
ΔI–Cl/I–O [Å] O–I–Cl [°]
Et (7c)[a]
2.043(5)
2.043(4)
2.046(5)
2.5343(17)
2.544(16)
2.5350(19)
2.533(2)
0.491
0.501
0.489
0.490
0.512
0.484
0.495
0.456
170.41(15)
171.80(14)
171.51(16)
171.39(18)
172.12(18)
171.87(5)
171.29(7)
172.01(8)
nPr (8c)
iBu (9c)[a] 2.043(6)
Scheme 2. Synthesis of reagents 7d–10d.
2.031(6)
2.0433(16) 2.527(17)
2.543(2)
iPr (10c)
Cy (11c)[a] 2.047(2)
2.5422(10)
2.5131(10)
Table 1. Yields for the syntheses of 7d–10d and 11c.
2.057(2)
R1
a [%][a]
b [%]
c [%]
d [%]
Overall [%]
[a] Asymmetric unit contains two independent molecules.
Et (7)
87
78
85
90
90
28
36
29
38
11
75
86
80
70
73
61
35
55
58
–
11
8
11
14
7[b]
nPr (8)
iBu (9)
iPr (10)
Cy (11)
A comparison of the bond lengths and ΔI–Cl/I–O values
presented in Table 2 with those of the standard reagent 1a
[I–O 2.042(2) Å, I–Cl 2.5491(8) Å, ΔI–Cl/I–O = 0.507 Å] re-
veals no special deviations. Indeed, whereas the reported
ΔI–Cl/I–O descriptors for structures 3a–6a range from 0.328
[a] The descriptors a, b, c, and d refer to the explicit structures of
7–11, for example, 7a, 7b, and so on. [b] Overall yield over three
steps to 11c.
2
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
Electrophilic Trifluoromethylating Agents
Figure 2. ORTEP representation of the asymmetric unit of 11c.
Hydrogen atoms are omitted for clarity, and thermal ellipsoids are
set to 50% probability. Short van der Waals contacts of 2.92 and
2.87 Å are found between I1–O2 and I2–O1, respectively. The paral-
lel orientation of the cyclohexyl rings suggests that there is further
van der Waals stabilization.
Figure 1. ORTEP representations of 7c–11c. Hydrogen atoms are
omitted for clarity, and thermal ellipsoids are set to 50% prob-
ability.
Parameter p
In search for a solution-phase parameter, p, to correlate
to the reactivity of the reagents 1b, 3b–6b, and 7d–10d, we
turned to 13C NMR spectroscopic data. Although struc-
tural diversity is present, two carbon atoms in proximity to
the reactive center can be identified as a common element
in all structures: the aryl carbon atom (Cipso) attached to
the iodine core and the quaternary carbon atom (Cmeso) of
the alkoxide ligand (Figure 3). Although “Cmeso” in 6b is a
homobenzylic position, it will be referred to as the meso
position for simplicity and consistency.
to 0.548 Å, the corresponding values for 7c–11c display a
narrower distribution around (0.490Ϯ0.016) Å. Moreover,
neither the electronic nature of the substituent R1 (Et/nPr/
iBu vs. iPr/Cy) nor its steric bulk (Et/nPrϽiPr/CyϽiBu)
seem to affect these values or the O–I–Cl angles signifi-
cantly. On the basis of this qualitative comparison, the rea-
gents 7d–10d are expected to be as reactive as the standard
reagent 1b in the electrophilic trifluoromethylation of para-
toluenesulfonic acid (Scheme 1).
Interestingly, 7c, 9c, and 11c have two independent mol-
ecules in their asymmetric units. Short van der Waals con-
tacts between I1–O2 and I2–O1 of 2.88/2.84, 2.79/2.76, and
2.92/2.87 Å, respectively, suggest that the iodine cores are
stabilized through the coordination of an additional oxygen
ligand to yield four-coordinate hypervalent iodine(III) spe-
cies (Figure 2). In addition, the two molecules in the asym-
Figure 3. The common carbon atoms found in all reagents are
labeled Cipso and Cmeso
.
Thus, we first investigated whether the 13C NMR shifts
metric unit of 11c form a homodimer (S/S or R/R) with for Cipso and Cmeso in 1b and 7d–10d can be described as a
the cyclohexyl rings oriented in a parallel fashion. This may function of the reported Taft inductive parameter (σ*) and
grant further stabilization by van der Waals contacts (Fig- the Taft steric parameter (ES) of R1.[5a] If applicable, this
ure 2). However, as these packing effects cannot be directly would allow the description of all available structures by an
quantified, the errors that affect reactivity predictions de- readily determined quantity. However, the initial consider-
rived from solid-state structural parameters are unknown ation of the Cipso data dismissed this carbon atom as a valid
and, consequently, small differences in reactivity may be ob- descriptor as, irrespective of R1, roughly the same chemical
scured.
shift was found [δ = 110.6–111.2 ppm, maximal Δδipso(CF3)
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
N. Santschi, C. Matthey, R. Schwenk, E. Otth, A. Togni
FULL PAPER
= 0.6 ppm]. Thus, we concentrated on Cmeso as a more Table 3. Compilation of 13Cmeso NMR spectroscopic data for com-
pound series 1 and 7–11.
promising candidate with a maximal Δδmeso(CF3) = 4.8 ppm
between R1 = iPr and R1 = Me. In Table 3, a summary of
1b
7d
8d
9d
10d
11d
all relevant 13C NMR spectroscopic data is presented, and
δmeso(X) refers to the chemical shift of the meso-carbon
atom in the (trifluoromethyl)iodinanes (X = CF3, e.g., 7d),
chloroiodinanes (X = Cl, e.g., 7c), iodoalcohols (X = I, e.g.,
7b), or parent noniodinated benzyl alcohols (X = H, e.g.,
7a).
Multidimensional regression analysis according to
δmeso(CF3) = δ1·σ* + δ2·ES + δ3 for the data set consisting
of 1b and 7d–10d delivered the row vector (δ1, δ2, δ2) =
(–26.2, –0.1, 76.5) with R2 = 0.98; this analysis implies that
δ2 (sterics) has a less-pronounced effect on the prediction
than δ1 (electronics; Figure 4, A). In addition, a simpler lin-
ear model including only the electronic term, δmeso(CF3) =
δipso(CF3) [ppm] 110.6 111.2 111.0 110.8 111.0
–
–
δmeso(CF3) [ppm] 76.5
79.2
7c
79.1
8c
79.6
9c
81.3
10c
1a
11c
δmeso(Cl) [ppm]
δmeso(I) [ppm]
δmeso(H) [ppm]
85.2
12[a]
73.6
13[b]
87.8
7b
87.7
8b
88.2
9b
88.8
10b
78.0
10a
76.9
89.8
11b
77.9
11a
76.8
76.2
76.0
8a
76.4
9a
7a
72.4[8] 75.1
74.8
75.4
[a] Compound 12 is 2-(2-iodophenyl)propan-2-ol. [b] Compound
13 is 2-phenylpropan-2-ol.
δ1·σ* + δ2, produced (δ1, δ2) = (–24.7, 76.5) with R2 = 0.94 ing values for 1b did not reveal any impact of the electronic
for the same data set. Therefore, both models suggest a di- or steric nature of R1 (vide supra). However, 13C NMR
rect relationship between the Taft parameter and the 13C spectroscopic analysis clearly distinguishes between the in-
NMR chemical shift of Cmeso in (trifluoromethyl)iodinanes; dividual side chains and is, therefore, considered a more
hence, δmeso(CF3) may be utilized as a measure p for the accurate methodology. Moreover, the data presented in
electronic and steric environments in all reagents examined Table 3 also suggests that the relationship δmeso(X) ~ σ* +
(1b, 3b–6b, 7d–10d). Importantly, a comparison of the bond ES is already present in the simple alcohol precursors 7a–
length differences ΔI–Cl/I–O in 7c–10c with the correspond- 11a (X = H).
Figure 4. Graphic representation of the data listed in Table 3. The data were correlated to the Taft electronic (σ*) and steric (ES) param-
eters according to δmeso(X) = δ1·σ* + δ2·ES + δ3, and the results presented in row vector form as (δ1, δ2, δ2). The dashed lines are
projections to the σ*–δmeso(X) plane (δ2 = 0).
4
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
Electrophilic Trifluoromethylating Agents
A graphical depiction of these findings is given in Fig- data set consisting of 1b, 3b–4b, and 7d–10d suggests that
ure 4, in which projections onto the δmeso(X) versus σ* there is a nonlinear relationship between the reactivity and
plane are presented by choosing δ2 = 0. The full row vectors the chemical shift of the meso-carbon atom (Figure 5, B).
(δ1, δ2, δ2) for the two-dimensional fits and the R2 values The curvature is understood on the basis of the relationship
are presented in the lower left corners, and the general com- δmeso(X) ~ σ* with the small steric contribution neglected
pound structures are given in the upper right corners of (vide supra). Therefore, the Taft equation log(k/k0) = ρ*·σ*
Figure 4 (A–D). Irrespective of the specific structure of the with δ = 0 can be rewritten to log(v0/ms–1) ~ ρ*·δmeso(X)/
compounds analyzed, good correlations between the 13C ppm and an exponential fit to the data proposed (dashed
NMR spectroscopic data and the Taft parameters were line). Indeed, a plot of log(v0/ms–1) versus δmeso(H)/ppm and
found. Most intriguingly, on the basis of these results, the a fit according to a linear model delivered log(v0/ms–1) =
hypothesis that the δmeso(H) values of starting alcohols 7a– –(8.79Ϯ1.04)ϫ10–2·δmeso(H)/ppm
–
(3.78Ϯ77.9)ϫ10–2
11a can be used to describe the reactivity v0 of the final with R2 = 0.94. Furthermore, analysis of the differences
reagents can be formulated.
between the predicted and actual values (residuals) further
corroborated the goodness of fit, and there was no clear
relationship between δmeso(H) and the residuals. The devia-
tions of the predicted initial rates vЈ0 from the measured v0
values, as measured by the ratio vЈ0/v0, are in the range 0.84
to 1.23 for 1b, 3b–6b, and 7d–10d (Table 4). As expected,
however, the reactivities of reagents 5b and 6b are underesti-
mated by factors of 0.43 and 0.03. A likely explanation for
the reactivity of 6b is found in its solid-state structure, in
which the alkoxide ligand at the iodine atom is part of a
six-membered ring that adopts an energetically unfavorable
half-chair conformation.[4a]
The Correlation δmeso(X) ~ v0
As suggested by the analysis presented above, the 13C
NMR chemical shift δmeso(X) may be a suitable solution-
phase parameter to describe the reactivity of reagents 1b,
3b–6b, and 7d–10d. Thus, to further test this hypothesis, the
initial rates v0 for 7d–10d for the electrophilic trifluoro-
methylation of para-toluenesulfonic acid were obtained
(Scheme 1, further details concerning the experimental
setup and data processing can be found in the Supporting
Information).[4] Each set of experiments was repeated at
least three times to ensure reproducibility and allow for er-
ror estimation. Hence, the v0 values for 1b and 7d–10d were
obtained with relative uncertainties of 5.5 to 10.5%. A sum-
mary of the initial rates for 1b, 3b–6b, and 7d–10d, standard
deviations (σ) if available, and the 13C NMR chemical shift
δmeso(H) for the noniodinated parent alcohols, such as 7a–
11a, is provided in Table 4.
Table 4. Compilation of 13C NMR shifts δmeso(H) of noniodinated
parent alcohols, initial rates v0, standard deviations, and predicted
initial rates vЈ0.
Reagent δmeso(H) [ppm] v0 [ms–1
]
σ [ms–1
]
vЈ0 [ms–1][a]
1b
3b
4b
5b
6b
7d
8d
9d
10d
72.4[8]
78.8
4.69ϫ10–7
1.2ϫ10–7[4a]
3.4ϫ10–7[4a]
5.3ϫ10–6[4a]
9.4ϫ10–7[4a]
2.17ϫ10–7
1.99ϫ10–7
2.21ϫ10–7
1.65ϫ10–7
4.92ϫ10–8 3.96ϫ10–7 (0.84)
–
–
–
–
1.09ϫ10–7 (0.91)
3.44ϫ10–7 (1.01)
1.76ϫ10–7 (0.03)
4.05ϫ10–7 (0.43)
73.1[9]
76.4[10]
72.3
75.1
74.8
75.4
76.9
1.87ϫ10–8 2.30ϫ10–7 (1.06)
1.97ϫ10–8 2.44ϫ10–7 (1.23)
1.47ϫ10–8 2.16ϫ10–7 (0.98)
9.05ϫ10–9 1.59ϫ10–7 (0.96)
[a] The vЈ0/v0 ratio is given in parentheses.
Figure 5. The chemical shifts δmeso(H) and the reactivities v0 corre-
late exponentially (B) if the initial data set (A) is reduced by remov-
ing outliers 5b and 6b.
An inspection of v0 for 7d–10d reveals that, contrary to
the initial prediction based on ΔI–Cl/I–O for 7c–10c (Table 2),
there is a clear order of reactivity: 1b (R1 = Me)Ͼ7d (R1
Therefore, the initial hypothesis that the δmeso(H) values
= Et)/8d (R1 = nPr)/9d (R1 = iBu)Ͼ10d (R1 = iPr). Further of the parent alcohols devoid of iodine may serve as a solu-
analysis of the subset 7d–9d also suggests that the steric tion-phase descriptor p for the reactivity (v0) of hyperva-
demand of R1 has a negligible effect on the reactivity (σ*: lent-iodine-based trifluoromethylating agents was verified,
Et ≈ nPr ≈ iBu, ES: Et ≈ nPrϾϾiBu). Finally, a plot of v0 although some restrictions apply. An additional phenyl
13
versus δmeso(H) delivers Figure 5 (A). As
C
meso
in 5b is a group should not be attached to the meso-carbon atom, and
double-benzylic position and 13Cmeso in 6b is a tertiary alkyl the I–O bond of the hypervalent interaction should be part
position, both data points constitute structural outliers and of a five-membered ring, as in the original reagent 1b, to
will be neglected in the following analyses. The reduced allow for an accurate prediction.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
N. Santschi, C. Matthey, R. Schwenk, E. Otth, A. Togni
FULL PAPER
Considering the electronic nature of the individual sub- clearly. Furthermore, the predicted reactivities vЈ0 for rea-
stituents, one identifies the general trend that reagents fea- gents 1b, 3b–4b, and 7d–10d were in excellent agreement
turing a more-electron-donating substituent are less reactive with the experimentally determined ones. For two reagents,
and feature lower initial rates v0 in the model reaction (Fig- 5b and 6b, vЈ0 was underestimated by factors of 0.43 and
ure 5, B). For example, the standard reagent 1b (R1 = Me, 0.03, respectively; this underestimation is attributable to sig-
σ* = 0) reacts 2.8 times faster than the slightly electron- nificant variations in the molecular structures of these rea-
richer reagent 10d (R1 = iPr, σ* = –0.19). Moreover, the gents compared with that of reference compound 1b. Most
introduction of the strongly electron-withdrawing NO2 intriguingly, however, the correlation δmeso(X) ~ v0 already
group (σ* = 4.66) para to the iodine atom in reagent 2b exists at the earliest stage of the synthesis (X = H) and,
leads to increased labilization of the hypervalent bond un- thereby, allows for future optimization studies of the scaf-
der acidic conditions.[11] This further substantiates that elec- fold before the actual reagents are accessed. On the basis
tron-releasing substituents are generally expected to lead to of these findings, endeavors towards the synthesis of more-
diminished reactivity.
reactive structures are ongoing and will be reported in due
On first sight, however, on the understanding that the course.
trifluoromethyl group is strongly electron-withdrawing (σ*
= 2.00–3.00), the bis-trifluoromethyl derivative 3b does not
Experimental Section
satisfy this criterion as it reacts 3.9 times slower than 1b.
However, in 1967, Pople and Gordon noted that fluorine
does not display a standard halogen-inductive effect, as
shown in Figure 6 (A), that is, it does not lead to a decreas-
ing positive polarization (black) along an alkyl chain.[12] In-
stead, alternating charges along the chain are found such
that the β and δ carbons feature a partial negative charge
(grey), contrarily to the intuitively expected positive one
(Figure 6, B). Holmes et al. also observed this effect experi-
mentally for the CF3 moiety in (trifluoromethyl)benzene
model systems (Figure 6, C).[13] Indeed, a consideration of
the structure of 3b reveals that Cmeso is at the β position
relative to the fluorine substituents. Hence, on the basis of
these considerations, Cmeso would feature a (small) negative
partial charge owing to the six electron-donors and, there-
fore, further corroborate the overall observed trend.
General: Reactions with moisture-sensitive chemicals were con-
ducted under an atmosphere of argon by using Schlenk techniques.
High-vacuum (HV) refers to a pressure of less than 0.01 mbar. Dry
solvents refer to solvents distilled before use either from CaH2
(MeCN) or from Na/benzophenone (THF). All commercially
available chemicals were used without further purification. Chro-
matographic purification was performed with silica gel 60 (230–
400 mesh) from Fluka or MP Alumina B, Act. I from MP Biomedi-
cals Germany GmbH. TLC plates were obtained from Merck (sil-
ica gel 60 F254). The NMR spectra were recorded with Bruker DPX
300, DPX 400, Avance III HD Nanobay-300, and Avance III HD
Nanobay-400 spectrometers. The 1H and 13C NMR chemical shifts
are referenced to residual solvent signals. The 19F NMR chemical
shifts are referenced to external CFCl3. Mass spectra (EI, ESI,
MALDI) were recorded by the MS service of the Laboratory of
Organic Chemistry (ETH Zurich). Elemental analyses were per-
formed by the Laboratory of Microelemental Analysis (ETH
Zürich). Melting points were determined with a Büchi Melting
Point B-540 apparatus.
CCDC-1031953 (for 7c), -1031954 (for 8c), -1031955 (for 9c),
-1031956 (for 10c), and -1031957 (for 11c) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Figure 6. Fluorine displays a nonintuitive inductive effect such that
an alkyl chain is polarized alternatively (B) rather than linearly (A).
Therefore, reagent 3b features a comparably electron-rich Cmeso
atom, which accounts for its low reactivity.
2-Phenylbutan-2-ol (7a): A heat-gun-dried 500 mL three-necked
round-bottomed flask equipped with a magnetic stirring bar, an
argon inlet, a pressure-equalized dropping funnel, and a rubber
septum was charged with bromobenzene (16 mL, 0.15 mol,
1 equiv.) and THF (300 mL, pro Analysis). The solution was cooled
to –78 °C (dry ice/acetone), and nBuLi (100 mL, 1.6 m, 0.15 mol,
1.05 equiv.) was slowly added by cannula transfer to yield a pale
yellow solution, which was stirred at –78 °C for an additional 2 h.
Conclusions
Four new electrophilic trifluoromethylating agents (7d– Then, butan-2-one (16.4 mL, 0.18 mol, 1.2 equiv.) in THF (20 mL,
pro Analysis) was added dropwise over 15 min, and the mixture
was warmed to ambient temperature overnight (16 h). The reaction
mixture was diluted with diethyl ether (200 mL) and washed with
distilled water (500 mL). The aqueous layer was back-extracted
twice with diethyl ether (2ϫ 500 mL), and the combined organic
layers were concentrated under reduced pressure. The crude prod-
uct was then purified by high-vacuum distillation (75 °C bath tem-
perature, 40–50 °C vapor temperature, 0.05 mbar) to afford the
product (19.9 g, 87%) as a viscous colorless oil. The spectroscopic
data agree with those previously reported.[10] 1H NMR (400 MHz,
10d) were accessed through the formal replacement of a
methyl group in 3,3-dimethyl-1-(trifluoromethyl)-3H-1λ3,2-
benziodaoxole (1b) by different alkyl chains. Subsequently,
by relating the 13C NMR chemical shifts to the Taft induc-
tive parameter σ*, δmeso(X) was identified as a suitable solu-
tion-phase parameter p to adequately describe the reactivity
of a reagent in a model trifluoromethylation reaction, as
measured by the initial rate v0. In stark contrast to predic-
tions derived from solid-state structures, this method al-
lowed the reactivities of 1b and 7d–10d to be distinguished CDCl3): δ = 7.48–7.45 (m, 2 H), 7.39–7.35 (m, 2 H), 7.29–7.25 (m,
6
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
Electrophilic Trifluoromethylating Agents
1 H), 1.93–1.82 (m, 2 H), 1.81 (s, 1 H), 1.58 (s, 3 H), 0.83 (t, J =
7.4 Hz, 3 H) ppm. 13C NMR (101 MHz, CDCl3): δ = 147.9, 128.2,
126.6, 125.0, 75.1, 36.8, 29.7, 8.4 ppm.
mixture was vacuum-filtered through a densely packed Celite pad.
The latter was washed with additional MeCN, and the combined
organic phases were concentrated under reduced pressure. The
brown residue was dissolved in pentane (pro Analysis) and vac-
uum-filtered through neutral alumina, and the resulting solution
was concentrated to dryness. The solid obtained was then subjected
to high-vacuum sublimation (30 °C, 0.02 mbar) and finally recrys-
tallized from pentane at –17 °C to afford the product (2.7 g, 61%)
2-(2-Iodophenyl)butan-2-ol (7b): A heat-gun-dried 500 mL three-
necked round-bottomed flask equipped with a magnetic stirring
bar, an argon inlet, a pressure-equalized dropping funnel, and a
rubber septum was charged with nBuLi (200 mL, 1.6 m, 0.32 mol,
2.5 equiv.) and TMEDA (4.8 mL, 32 mmol, 0.25 equiv.). The re-
sulting solution was stirred for 45 min at ambient temperature to
afford an off-white suspension that was subsequently cooled to
0 °C (ice bath). Then, 7a (19.2 g, 0.13 mol, 1 equiv.) in hexane
(60 mL, pro Analysis) was added over the course of 60 min, and
the mixture was slowly warmed to ambient temperature (16 h). Ad-
ditional nBuLi (100 mL, 1.6 m, 0.16 mol, 1.25 equiv.) was added,
and stirring was continued for 24 h. Finally, the orange solution
was cooled to 0 °C (ice bath), and solid iodine (100 g, 0.39 mol,
3.1 equiv.) was added under vigorous stirring. Then, the mixture
was warmed to ambient temperature over the course of 60 h. The
resulting brown mixture was washed with aqueous sodium thio-
sulfate (50 wt.-%, 500 mL), and the aqueous layer was back-
extracted with diethyl ether (2ϫ 500 mL). The combined organic
layers were concentrated under reduced pressure. Two subsequent
fractional high-vacuum distillations (170 °C bath temperature, 80–
95 °C vapor temperature, 0.02 mbar) afforded the product (9.8 g,
28%) as a viscous yellow oil. 1H NMR (400 MHz, CDCl3): δ =
7.97 (d, J = 7.47 Hz, 1 H), 7.64 (dd, J = 8.0, 1.2 Hz, 1 H), 7.34–
7.30 (m, 1 H), 6.90–6.86 (m, 1 H), 2.33 (dq, J = 14.5, 7.3 Hz, 1 H),
2.04 (s, 1 H), 2.00 (dq, J = 14.5, 7.3 Hz, 1 H), 1.72 (s, 3 H), 0.80
(t, J = 7.4 Hz, 3 H) ppm. 13C NMR (101 MHz, CDCl3): δ = 147.9,
142.9, 128.6, 128.1, 128.0, 93.1, 76.2, 33.7, 27.95, 8.5 ppm. HRMS
(EI): calcd. for C10H13IO [M]+ 276.0006; found 276.0009.
1
as colorless crystals; m.p. 77–78 °C. H NMR (300 MHz, CDCl3):
δ = 7.55–7.50 (m, 2 H), 7.45–7.39 (m, 1 H), 7.33–7.30 (m, 1 H),
1.85 (dq, J = 14.8, 7.4 Hz, 1 H), 1.74 (dq, J = 14.0, 7.4 Hz, 1 H),
1.43 (s, 3 H), 0.80 (t, J = 7.4 Hz, 3 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 148.1, 130.6, 130.0, 128.0, 127.9 (q, JC,F = 2.8 Hz),
111.2 (q, JC,F = 3.1 Hz), 110.5 (q, JC,F = 395.9 Hz), 79.2, 36.6,
29.8, 8.4 ppm. 19F NMR (282 MHz, CDCl3): δ = –40.1 ppm.
HRMS (EI): calcd. for C9H7F3IO [M – C2H5]+ 314.9489; found
314.9485. C11H12F3IO (344.11): calcd. C 38.39, H 3.51; found C
38.47, H 3.52.
Supporting Information (see footnote on the first page of this arti-
cle): full experimental data including copies of NMR spectra of all
new compounds.
Acknowledgments
ETH Zürich is gratefully acknowledged for financial support. Dr.
Katrin Niedermann is acknowledged for crystallographic measure-
ments and helpful discussions during manuscript preparation.
[1] a) P. Eisenberger, S. Gischig, A. Togni, Chem. Eur. J. 2006, 12,
2579–2586; b) N. Santschi, A. Togni, Chimia 2014, 68, 419–
424; c) J. Charpentier, N. Früh, A. Togni, Chem. Rev. 2015,
115, 650–682, and references cited therein.
[2] J. P. Brand, J. Waser, Chem. Soc. Rev. 2012, 41, 4165–4179.
[3] V. V. Zhdankin, C. J. Kuehl, A. P. Krasutsky, M. S. Formaneck,
J. T. Bolz, Tetrahedron Lett. 1994, 35, 9677–9680.
[4] a) K. Niedermann, J. M. Welch, R. Koller, J. Cvengrosˇ, N.
Santschi, P. Battaglia, A. Togni, Tetrahedron 2010, 66, 5753–
5761; b) R. Koller, Q. Huchet, P. Battaglia, J. M. Welch, A.
Togni, Chem. Commun. 2009, 5993–5995.
[5] a) C. Hansch, A. Leo, D. Hoekman, Exploring QSAR, vol. 2:
Hydrophobic, Electronic, and Steric Constants, American
Chemical Society, Washington DC, 1995; b) E. V. Anslyn,
D. A. Dougherty, Modern Physical Organic Chemistry, Univer-
sity Science Books, Sausalito, 2006.
[6] N. Meyer, D. Seebach, Angew. Chem. Int. Ed. Engl. 1978, 17,
521–522; Angew. Chem. 1978, 90, 553.
[7] V. Matousˇek, E. Pietrasiak, R. Schwenk, A. Togni, J. Org.
Chem. 2013, 78, 6763–6768.
[8] Y. Zhang, J. Feng, C.-J. Li, J. Am. Chem. Soc. 2008, 130, 2900–
2901.
[9] T. Maekawa, H. Sekizawa, K. Itami, Angew. Chem. Int. Ed.
2011, 50, 7022–7026; Angew. Chem. 2011, 123, 7160.
[10] H. Zong, H. Huang, J. Liu, G. Bian, L. Song, J. Org. Chem.
2012, 77, 4645–4652.
[11] N. Santschi, R. C. Sarott, E. Otth, R. Kissner, A. Togni, Beil-
stein J. Org. Chem. 2014, 10, 1–6.
1-Chloro-3-ethyl-3-methyl-3H-1λ3,2-benziodaoxole (7c): Compound
7b (9.8 g, 35 mmol, 1 equiv.) was dissolved in DCM (80 mL, pro
Analysis) in a round-bottomed flask wrapped in aluminum foil.
The resulting solution was cooled to 0 °C (ice bath), tBuOCl
(4.1 mL, 36.4 mmol, 1.04 equiv.) was added in one portion, and
the mixture was slowly warmed to ambient temperature overnight
(16 h). The neon-yellow solution was concentrated under reduced
pressure, and the resulting residue was washed with pentane to af-
ford the product (8.3 g, 75%) as a yellow solid. X-ray quality crys-
tals were obtained by the vapor diffusion of pentane into a satu-
rated ethyl acetate solution at 4 °C; m.p. 100–102 °C. 1H NMR
(400 MHz, CDCl3): δ = 8.03 (dd, J = 8.0, 0.7 Hz, 1 H), 7.57 (td, J
= 7.6, 1.5 Hz, 1 H), 7.52 (td, J = 7.3, 0.8 Hz, 1 H), 7.10 (dd, J =
7.3, 1.5 Hz, 1 H), 1.93 (dq, J = 14.4, 7.3 Hz, 1 H), 1.80 (dq, J =
14.4, 7.2 Hz, 1 H), 1.49 (s, 3 H), 0.88 (t, J = 7.4 Hz, 3 H) ppm. 13
C
NMR (101 MHz, CDCl3): δ = 148.8, 130.9, 130.6, 128.6, 126.8,
115.1, 87.8, 35.2, 27.8, 8.3 ppm. HRMS (EI): calcd. for C9H9ClIO
[M – CH3]+ 294.9382; found 294.9383. C10H12ClIO (310.56): calcd.
C 38.68, H 3.89; found C 38.42, H 3.93.
3-Ethyl-3-methyl-1-(trifluoromethyl)-3H-1λ3,2-benziodaoxole (7d):
Potassium fluoride (2.2 g, 37.9 mmol, 2.9 equiv.) was added to a
Schlenk flask and flame-dried under high-vacuum. Then, 7c (4 g,
12.9 mmol, 1 equiv.) was added under an argon counterflow, and
the solids were suspended in dry MeCN (72 mL) and stirred vigor-
ously for 62 h at ambient temperature. The resulting white suspen-
sion was cooled to 0 °C (ice bath), and Ruppert–Prakash reagent
(TMSCF3, 4.2 mL, 28.4 mmol, 2.2 equiv.) was added in one por-
tion. Stirring was continued for 1 h at 0 °C, and then the brown
[12] J. A. Pople, M. Gordon, J. Am. Chem. Soc. 1967, 89, 4253–
4261.
[13] S. A. Holmes, T. D. Thomas, J. Am. Chem. Soc. 1975, 97, 2337–
2341.
Received: December 16, 2014
Published Online: ᭿
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7

Job/Unit: O43634
/KAP1
Date: 12-02-15 13:22:46
Pages: 8
N. Santschi, C. Matthey, R. Schwenk, E. Otth, A. Togni
FULL PAPER
Reactivity Patterns
The reactivity of hypervalent-iodine-based
trifluoromethylating agents, as expressed
by the initial rate v0 in a model reaction,
correlates to an easily determined physical
parameter p, a 13C NMR chemical shift.
N. Santschi,* C. Matthey, R. Schwenk,
E. Otth, A. Togni* ............................ 1–8
On the Effect of Backbone Modifications
in
3,3-Dimethyl-1-(trifluoromethyl)-3H-
1λ3,2-benziodaoxole
Keywords: Hypervalent compounds / Kin-
etics / Reactivity predictions / Electrophilic
addition
/
Iodine
/
Trifluoromethylation
8
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0