Organic Letters
Letter
phosphine oxide (TEPO) as a Lewis basic probe.14−16 First
reported by Gutmann and Beckett to measure Lewis acidic
solvents,14 we and others have since adapted this method to
quantify the H-bonding ability and Lewis acidity of various
catalysts.15,16 Previous work in our group has successfully
demonstrated the proportional relationship of H-bonding
ability and catalytic activity for a variety of H-bonding donors
using 31P NMR spectroscopy (Figure 1D).16 Compared to the
extensive application for H-bonding, the investigation of X-
bonding ability for organocatalysts is limited.5j,17 Here, we
report using the 31P NMR spectroscopy method to systemati-
cally quantify X-bonding ability and correlate the catalytic
ability for XBD organocatalysts (Figure 1E). This work
examines representative XBD compounds including neutral
(1−4), cationic benzimidazolium- (5),5e,f imidazolium- (6−
7),5e,f,i and bis(imidazolium)-based5e,g (8−9) structures
(Figure 2).
Figure 3. (A) Example 31P NMR spectra for the downfield 31P NMR
shifts upon TEPO binding to XBD (in CD2Cl2). (B) Titration
experiments comparing XBDs. Calculation of binding constant: 8c·
possess enhanced binding ability21 and saturation was observed
for each at ∼5 equiv.
To assess that TEPO binding can probe X-bonding ability,
the Δδ(31P) values for XBDs with different electrophilic
halogen atoms were measured (Figure 2, 5a−5b·OTf, 7e−7g·
OTf, 8c−8d·OTf). The more electrophilic halogen atom (I >
Br > Cl) should correlate with larger Δδ(31P) values. Indeed,
switching from iodo to bromo and chloro (in 7e−7g·OTf)
significantly decreased Δδ(31P) values (Table 1), matching a
decrease in polarizability of the halogen substituents and hence
the X-bonding ability.
To validate 31P NMR spectroscopy and the ability of TEPO
binding to quantify electronic effects for X-bonding, a
Hammett plot was created for a series of 2-iodoimidazolium
triflate salts (7a−7e·OTf) with electronically varied substitu-
ents on the conjugated benzene ring (Table 1, Figure 4). A
linear relationship (R2 = 0.988) was observed, indicating that
Δδ(31P) values can accurately quantify the electronic changes
affecting X-bonding ability.
Results and Discussion. To build a scale to quantify the X-
bonding ability of organocatalysts, Δδ(31P) values were
measured for a variety of XBD compounds (Table 1). TEPO
binding with most neutral XBDs (e.g., 1−2, 4) results in very
low Δδ(31P) values, indicating low X-bonding ability; however,
the X-bonding ability was notably increased with a high
electronegative atom attached to the halogen atom (e.g., 3a).
Without the cationic charge on the XBD core, the Δδ(31P)
values are <1.0 ppm and interactions are too weak to observe
trends, highlighting the importance of the cationic charge for
increasing the X-bonding ability (e.g., 4a−4d). The larger
Δδ(31P) value of cationic benzimidazolium-based XBD (5a·
OTf, Δδ(31P) = 4.94 ppm) compared to cationic imidazolium-
based XBD (6·OTf, Δδ(31P) = 4.03 ppm) suggested higher X-
bonding ability. This is attributed to the extended aromaticity
of the benzimidazolium core.22 By increasing the electron
deficiency of the benzimidazolium core (5d·OTf), the largest
Figure 2. Halogen-bond donors studied.
Method Validation. To validate 31P NMR spectroscopy as a
method to quantify X-bonding ability, factors such as the
XBD−TEPO equilibrium, solvent interferences, and the
competition of other noncovalent interactions from impurities
were carefully examined.18 Downfield 31P NMR shifts (Δδ) are
observed upon TEPO binding to XBDs (Figure 3A).19
Stronger X-bonding ability is expected to correlate to larger
Δδ(31P) values. To assess the XBD−TEPO binding equili-
brium,20 titration experiments of XBD (relative to TEPO)
were investigated (Figure 3B). For imidazolium- (6·OTf) and
bis(imidazolium)-based triflate (8c·OTf) XBDs, saturation
occurs at approximately 10 and 15 equiv, respectively. With the
weakly coordinating counteranion, 7e·BArF and 8c·BArF XBDs
B
Org. Lett. XXXX, XXX, XXX−XXX