Content uploaded by Anatoli Kurkin
Author content
All content in this area was uploaded by Anatoli Kurkin on Aug 14, 2023
Content may be subject to copyright.
Unravelling the Dynamic Crosslinking Mechanism in Polyborosiloxane
Anatoli Kurkin,a Yulia Lekina,b David G. Bradley,c Geok Leng Seah,a Kwan Wee Tan,a Vitali Lipik,a John V.
Hanna,c Xin Zhang,*d and Alfred Iing Yoong Tok **a
a School of Materials Science and Engineering, Nanyang Technological University, Singapore 639798, Singapore
b School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 637371, Singapore
c Department of Physics, University of Warwick, Coventry, CV4 7AL, United Kingdom
d Department of Mechanics and Aerospace Engineering, Southern University of Science and Technology,
Shenzhen 518055, Guangdong, China
*Co-corresponding author. E-mail: zhangx8@sustech.edu.cn, 8F South Tower, college of Engineering, 1088
Hueyuan Avenue, Shenzhen 518055, Guangdong, China (X. Zhang)
**Corresponding author. E-mail: miytok@ntu.edu.sg, School of Materials Science and Engineering, Nanyang
Technological University, Singaporeб 50 Nanyang Avenue Block N4.1-02-23 Singapore 639798 (A. Tok)
Abstract: This study offers a comprehensive investigation into the mechanism behind the dynamic crosslinking of
polyborosiloxane (PBS). Despite this material being well known for over 70 years, the origin of PBS's unique
viscoelastic properties has been a topic of debate in the literature. Through combined FTIR and solid-state 29Si and
11B MAS NMR analyses, this study provides evidence that the formation of Si—O—B dynamic covalent bonds, along
with their associative exchange with neighboring hydroxyl-bearing species (free silanol, water, alcohol, etc.), drives
the gelation and viscoelastic behavior of PBS. Results show no indication of hydrogen or dative bonding, instead
the low energy barrier for formation and breakage of Si—O—B bonds allows for easy exchange at room
temperature. Moreover, the study finds that the viscoelastic properties of PBS can be adjusted by the choice of
boron B—O functionality, leading to n-functional dynamic crosslinking through Si—O—B bonds. This research
provides a clear understanding of the mechanism of dynamic crosslinking in PBS, resolving long-standing
controversies in the field.
Graphical abstract:
Key words: polyborosiloxane, dynamic covalent bonding, supramolecular polymers, Fourier Transform Infrared
spectroscopy, Nuclear Magnetic Resonance
2
Introduction
Polyborosiloxane (PBS) is a material that has been studied for over 70 years, and was first produced as
the main ingredient of the Silly Putty
TM
toy in 1947 by a researcher from Corning Inc. [1] Many years later,
researchers started paying attention to the intriguing viscoelastic properties of PBS and how they can be
utilized for various applications, such as impact protection [2], flexible electronics [3–6], sensors [7–11] etc.
PBS flows like a viscous liquid at low deformation rates, however once transition rate is exceeded it stiffens.
The transition is attributed to the behavior of dynamic bonds. At a lower rate of deformation, these bonds
undergo relaxation and subsequent reformation, causing the material to exhibit liquid-like flow behavior.
However, when the deformation rate surpasses a critical value, the bonds become locked and exhibit
characteristics akin to those of permanent crosslinks, resulting in the material behaving like an elastic
rubber.
Despite extensive research into its properties, the origins of its dynamic crosslinking mechanism remain
a topic of debate in the literature. Over the years, various mechanisms (Scheme 1) have been proposed,
including hydrogen bonding between borono (R—O—B(OH)
2
, where R is PDMS) end-groups
[12–14], dative
bonding between boron and oxygen
[15–17], boroxine formation, and dynamic covalent bonding
[18].
Scheme 1. Five possible types of dynamic crosslinking mechanisms in polyborosiloxane. (a) Hydrogen bonding mechanism
involves the formation of dimeric hydrogen bonds between two borono end groups (R—O—B(OH)2), resulting in a temporary
chain extension. (b) Dative bonding of Type 1 involves the donation of a lone pair of electrons from an oxygen atom from the
main siloxane chains to a boron atom with an empty p orbital. (c) In a dative bonding of Type 2, another oxygen donates a lone
pair of electrons — from borono end group. (d) Formation of boroxines through dehydration or transesterification and
subsequent exchange with neighboring boroxine without the presence of any diol molecules (metathesis). (e) Dynamic
covalent bonding results in the formation of trifunctional Si—B—O crosslinks (in the case of boric acid) that participate in the
exchange through neighboring silanol groups.
Hydrogen bonding was one of the first proposed supramolecular interactions to be responsible for the
gelation of PBS
[12]. It was believed that after the synthesis, PDMS chains become terminated with boron
9
functionality through a single Si—O—B covalent bond, whereas the two remaining hydroxyl groups
participate in a dimeric hydrogen bonding with other borono modified chains ends (Scheme 1, a).
Dative bonding, alongside hydrogen bonding, has been suggested in the literature as an explanation for
the unique viscoelasticity of PBS
[15]. It is known that boronic acid can change its hybridization from sp
2
to
sp
3
upon interaction with a strong nucleophile, like oxygen, due to a vacant p-orbital of the boron atom
[19–21]. Thus, boron could become anionic 4-coordinate species which allows it to form a fourth additional
bond of dative nature with a nucleophile. This suggests that the borono end group in PBS can form
reversible interactions with a Lewis's base atom (such as oxygen) resulting in a sticker-like effect. Two types
of dative bonding have been proposed as the main supramolecular interactions: Type 1 involves the boron
atom bonding with the oxygen of the siloxane (Scheme 1, b), while Type 2 involves bonding with the oxygen
of the B—OH group from another borono group (Scheme 1, c), which was proposed more recently
[22].
The transformation from neutral to negatively charged structures during the dative bonding of boron
atoms can be explained through the change in coordination geometry from 3-coordinate (trigonal) to 4-
coordinate (tetrahedral) ion species. In the neutral state, boron typically forms three covalent bonds,
resulting in a trigonal planar geometry. However, during dative bonding, boron accepts an electron pair
from a donor atom, such as oxygen (which, in fact, becomes positively charged), to form a coordinate
covalent bond.
In parallel, a whole new branch of boron chemistry research has been developed, which brought new
insight into the possible mechanism of reversible boron crosslinking in polymers [21,23,24]
Some of the
most elaborated mechanisms are the reversible formation of boronic esters and boroxines. Boronic esters
could be formed through dehydration of boronic acid in the presence of available diol groups; this reaction
is reversible and can even be dynamic if the system is in equilibrium[25]. Furthermore, these esters could
participate in a transesterification exchange with externally added diol molecules or even directly exchange
between themselves (metathesis) [19–21,26,27]. Boroxines could also reversibly be formed through
dehydration and participate in metathesis [21,24]. In both cases, the addition of a Lewis base could
facilitate and tune rates of boronic transesterification [28] and boroxine metathesis [21,29]. Boroxines
could also be formed in PBS under specific conditions (i.e., severe dehydration), some other researchers
suggested (Scheme 1, d) [13,14].
Recently, another mechanism of dynamic crosslinking in PBS was proposed by Bloomfield[18], which
resembles transesterification exchange in boronic esters. He suggested that all PDMS chains are covalently
bonded through trifunctional boron crosslinks like in silicone rubber, but in the case of PBS, these covalent
bonds are dynamic. They detach and frequently re-attach through oxygens ligand exchange with free
silanol groups (Scheme 1, e).
The vastness of boron chemistry complicates the determination of the dominant dynamic crosslinking
mechanism in PBS. This study aims to clarify the conflicting proposed mechanisms for dynamic crosslinking
in PBS and classify them from unlikely to dominant. The study employs various synthesis conditions,
hydrolysis, and solution studies, paired with FTIR and solid-state
29
Si and
11
B MAS NMR to probe the
possible mechanisms and their contributions. In addition, the effect of B–O functionality on the viscoelastic
properties of PBS is analyzed in order to provide additional evidence pertaining to the mechanistic
pathways.
Materials and experimental methods
Chemicals and samples synthesis
4
Hydroxy terminated polydimethylsiloxane (HO—PDMS—OH) with three different kinematic viscosities
(25, 65, and 750 cSt), boric acid (BA), phenylboronic acid (PBA), toluene, hexane, isopropyl alcohol,
dimethylformamide, tetrahydrofuran, chloroform, acetone, DI water were purchased from Sigma-Aldrich
and used as received without further purification.
The synthesis of PBS samples was performed via a straightforward condensation between hydroxyl
groups of HO—PDMS—OH and boron compounds. The stoichiometric concentration of BA and PDMS was
calculated based on gas-permeation chromatography (GPC) molecular weight data. Two methods were
used: (1) Room temperature (RT) synthesis: HO—PDMS—OH and BA were mixed in a beaker with a
magnetic stirrer until the bar could no longer rotate. (2) Synthesis with heating: A Dean-Stark apparatus
was used to shift the chemical equilibria by removing the water produced during the reaction. Toluene was
used as the solvent media to form an azeotrope with water. The solution was heated to 140 ⁰C and stirred
for 24h. Six samples were synthesized: B25, B65, B750, P25, P65, and P750, labeled according to the
kinematic viscosity of PDMS and the first letter of the boronic compound. The RT label was added to
samples synthesized at room temperature without the Dean-Stark apparatus, such as B25RT.
Upon completion of the reaction, the solution was cooled down first and dried in a fume hood. Despite
the clear state of the solution, small amounts of powder residues were seen after the drying step due to
partial hydrolysis of boronic acids under ambient humidity. The samples were refined by dissolving them
in hexane and filtering through 0.22 µm PVDF filters several times until no residues were seen.
Characterization
The solvent study was conducted by dissolving B25 in six different solvents (IPA, hexane, chloroform,
acetone, DMF, and DI water). The time for complete dissolution was measured by placing 2 g pieces of B25
gel in small bottles containing 10 ml of solvent and a magnetic bar, then starting the timer as magnetic
stirring was initiated.
Gel permeation chromatography (GPC) was used to determine the number-average molecular weight
(M
N
) of the synthesized samples and PDMS precursors. It was done similarly to An Agilent 1260 Infinity
with a TOSOH TSKgel GMHHR-M column was used, with the samples dissolved in THF at 1 mg/ml and
analyzed at a 1 ml/min flow rate. Monodisperse polystyrene (PS) calibrants were used to create a
calibration curve. The usage of THF as an eluent and PS standards as calibrants is in accordance with the
other reported GPC data on PBS [30,31]. Calculated M
N
values can be found in Table S1 and calibrated
molecular weight distribution curves in Fig. S1.
Functional group analysis was performed using Perkin Elmer Frontier Fourier transforms infrared
microscope (FTIR) in Attenuated Total Reflection (ATR) mode. Spectra were scanned 32 times for each
sample, with the force gauge varied from 20 to 60 for liquids, gels, and powders respectively.
Rheological characterization was conducted using an Anton Paar MCR 501 rotational rheometer.
Dynamic analysis was performed using 1 % amplitude frequency sweep tests on gel samples and shear rate
ramp tests on liquid samples, with 25 mm parallel steel plates and a gap setting of 0.5 mm and 0.3 mm,
respectively. Tests were conducted at laboratory humidity and 25 ⁰C.
All
11
B MAS NMR measurements were performed at 16.4 T using a Bruker 700 MHz Advance III HD
spectrometer operating at a Larmor frequency of 224.7 MHz, and a Bruker 3.2 mm HXY probe which
enabled MAS frequencies of ~10 kHz.
Data from pulse experiments were acquired using selective π/12
pulses along with recycle delays of 200 s in order to obtain quantitative measurements of the B speciation.
The measured 11B MAS NMR data were referenced against the IUPAC recommended primary reference of
15 % boron trifluoride etherate (BF
3
.Et
2
O) in deuterated chloroform (CDCl
3
) (
𝛿
iso
= 0.0 ppm) via a secondary
reference of sodium borohydride (NaBH
4
) (
𝛿
iso
= -42.06 ppm). Corresponding
29
Si MAS NMR measurements
9
were performed at 7.05 T using a Bruker 300 MHz Advance III HD spectrometer operating at a
29
Si Larmor
frequency of 59.6 MHz, and a Bruker 7 mm HX probe which facilitated MAS frequencies of ~5 kHz to be
obtained (measurements on the PDMS precursor were performed under static conditions). Quantitative
29
Si data were acquired using single pulse experiments which used 4 μs π/4 single pulses and a recycle delay
of 5 s. All data were referenced against the IUPAC recommended primary reference of TMS (
𝛿
iso
= 0.0 ppm)
via a secondary solid reference of kaolinite (
𝛿
iso
= -92.0 ppm). All NMR measurements were undertaken at
room temperature.
Results and discussion
Mechanisms unlikely to be present in PBS
Analysis was initiated by excluding dynamic crosslinking mechanisms that are not present in PBS, based
on experimental data. To demonstrate the absence of hydrogen bonding and dative bonding in our
samples, FTIR measurements coupled with
11
B and
29
Si solid-state MAS NMR were used to probe site-
specific bonding arrangements. In this study, we primarily discuss the findings of the analysis of B25, which
is composed of the shortest siloxane chains reacted with boric acid. B25 possesses the highest crosslinking
density, resulting in higher intensities for the terminal units, thereby simplifying the analysis. B25 is a PBS
polymer product of the reaction between PDMS25 and BA.
FTIR is a highly sensitive technique that can distinguish between different boron species based on the
B—O band in IR spectra. The FTIR spectra of the resulting gel B25 along with the siloxane precursor PDMS25
were then collected and analyzed (Fig. 1a). The disappearance of the bounded -OH band (3281 cm
-1
) upon
condensation, which is a distinctive characteristic of PDMS hydrogen bonding in the system. Appearance
of new bands in the B—O trigonal asymmetric stretch region (1400-1300 cm
-1
) [32] were detected in the
B25 spectra. A strong peak around 1334 cm
-1
indicates the formation of Si—O—B covalent bonds, as has
been previously reported [13,14,31]. Additionally, hydrogen bonding cannot explain gelation alone since it
would only provide dimeric chain extension.
The tetrahedral boric acid (BA) species were shown to have different B—O vibrational bands
(asymmetric stretching in the range 1050-850 cm
-1
and symmetric in the range 700-850 cm
-1
) from trigonal
ones (asymmetric stretching in the range 1450-1300 cm
-1
and symmetric in the range 1050-950 cm
-1
) [32].
The results showed that while the trigonal B—O bands was well pronounced, the tetrahedral B—O
counterparts were not present in B25, thus eliminating the possibility of any dative bonding arrangements.
The free state of boric acid has a C
3h
symmetry, which results in one strong peak in the B—O asymmetric
stretching zone on an FTIR spectrum. However, if a BA proton is substituted by a heavy atom, the symmetry
lowers to C
2v
, which generates two peaks in the B—O asymmetric stretching band (Fig. 1, b). If two protons
are substituted, the symmetry remains the same but results in two vibrations of slightly different energies
[32]. Substituting all three protons with one type of atom restores the original symmetry but with a
frequency shift. In the case of PBS, the general symmetry is broken due to the long flexible siloxane chains,
although it is close to C
3h
in the nearest proximity to the boron centers. Broadening and some splitting of
the FTIR asymmetric stretching are expected owing to the variety of siloxane conformations in PBS.
6
Fig. 1 (a) FTIR spectra of B25 synthesized with heating in toluene in a Dean-Stark apparatus against precursors (boric
acid and PDMS25). The zoom-in region shows an expanded plot of the tetrahedral B—O asymmetric stretching region
and indicated the absence of the tetrahedral B—O asymmetric stretching bands. Inset illustrates the hypothesized
outcome of the synthesis reaction of PBS B25 with chemical formula of the original precursors. (b) Splitting of the
asymmetric B—O stretching band based on the symmetry of boron compounds (free BA, one, two, and three proton-
substituted BA molecules). (c) Gaussian peak fitting of the B25 FTIR spectrum in the trigonal B—O asymmetric
stretching zone (R
2
= 0.9999).
The FTIR spectra of PBS were deconvoluted using Gauss function fitting (Fig. 1, c). The two low-intensity
peaks at 1404 and 1414 cm
-1
originate from the siloxane vibrations and are present in the PDMS precursor
spectra. The broad peak at 1320-1380 cm
-1
only appears in the PBS spectra and is assigned to the trigonal
B—O asymmetric stretching mode, consistent with literature [32]. The deconvolution revealed three
components: a broad one centered at 1363 cm
-1
and two sharp peaks at 1331 and 1348 cm
-1
. The latter
peaks are associated with the Si—O—B bond stretching
[14,30,33] and confirms the bond formation
between PDMS and BA. The origin of the broad peak at 1363 cm
-1
is more uncertain. The peak corresponds
to one of B—O asymmetric stretching vibrations [13,14,33]. In some studies [13,14], this band was assigned
9
to B—O—B bond stretching in boroxines. However, the formation of these anhydrides requires severe
dehydration and probably cannot be formed at room temperature. It has been assigned to B-O-C stretching
in boronic esters before [34]. Additionally, the peak is located within the Si—O—B bond range as well, and
the increased FWHM can be related to the siloxane conformations. However, the exact assignment of this
peak remains an open question and it does not help clarify the mechanism of the dynamic crosslinking.
The absence of the B—O—B bond cannot be confirmed based on FTIR analysis alone, therefore it was
additionally tested through stoichiometric analysis. The formation of boroxines requires a three-times
excess concentration of boric acid. However, in the B25 sample, all hydroxyl groups (Si—OH and B—OH)
disappeared due to condensation at stoichiometric concentrations, indicating that boroxines or any other
boron anhydrides are unlikely to be present. Furthermore, it has been shown that excessive concentration
of boronic acid beyond the saturation level does not affect mechanical properties [18].
Solid state MAS NMR is a highly sensitive technique to probe local environments of specific nuclei, even
in disordered materials such as polymers and gels.
11
B is a useful tool for determining the presence of BO
3
and BO
4
units, based on their characteristic isotropic chemical shift ranges (δ
iso
(BO
3
) ≈ 7 – 20 ppm and
δ
iso
(BO
4
) ≈ 0 – 5 ppm) [35]. Fig. 2 shows that both the BA precursor and the B25 sample contain only BO
3
units. In comparison to the boric acid system which comprises of well-
defined ordered ring and non-ring species with characteristic second order quadrupolar broadened
lineshapes, the B25 sample exhibits a broad gaussian centered at a δ
iso
value typical of non-ring BO
3
units.
Fig. 2. (Left) Single pulse
11
B MAS NMR of the BA precursor and B25, and (right) single pulse
29
Si MAS NMR of the
PDMS precursor and B25. Red and blue lines in the
11
B spectra represent simulations of the experimental data and
individual components, respectively. Asterisks in the
29
Si spectra denote the remaining molecular structure. Simulated
NMR parameters are presented in Table S2.
8
The local disorder present in the B25 network structure results in a large chemical shift dispersion that
defines the lineshape rather than any quadrupolar broadening. Unfortunately, the disorder is such that the
presence or absence of B—O—B or B—O—Si species cannot be determined from this data, only that the
boron exists and a non-ring BO
3
species which rules out the possibility of any dative bonding (Scheme 1,
b).
29
Si solid state NMR offers an additional probe for analyzing bonding networks, and is sensitive to
structural changes further that the first coordination sphere.
29
Si MAS NMR spectra for the PDMS precursor
and the B25 sample are also shown in Fig. 2, and the internal Si(CH
3
)
2
(OR)
2
and terminal Si(CH
3
)
2
(OR)(OH)
units are shown clearly for the PDMS precursor. When differentiating between Si-species in NMR spectra
it is common to use M
n
, D
n
, T
n
and Q
n
notation to represent SiC
3
O, SiC
2
O
2
, SiCO
3
and SiO
4
environments,
respectively, where n represents the number of bridging oxygens. All such species have well defined
chemical shift ranges, and clearly in the systems studied here only D
2
and D
1
species are present due to the
internal and terminal Si-species. Upon condensation of the PDMS precursor with the BA precursor, no
change in chemical shift is observed for the internal Si-species, however a considerable upfield shift is
observed for the terminal species. This confirms that the condensation and formation of B—O—Si bonds
occur at the terminal groups. Furthermore, data were acquired quantitatively, and the integral ratios of
70:30 for both samples are consistent with the molecular weight (MW) measurements (Table S1) and
supplier data (550 Da). The fact that the whole of the terminal species have shifted upfield upon
condensation with the same integral intensity indicates that every terminal borono species is used in B—
O—Si formation.
Thus, the possibility of hydrogen bonding (Scheme 1, a), dative bonding (Scheme 1b, c) and boroxine
formation (Scheme 1, d) can be excluded. The elimination of these mechanisms has narrowed down the
possible candidates for the dynamic crosslinking mechanism in PBS. Dynamic covalent bonding is the most
likely mechanism as it has been extensively studied and shown to be present in many comparable systems
(boronic esters and boroxines)
[21]. The following part of this study aims to provide evidence confirming
this mechanism.
Dynamic covalent bonding
To determine whether dynamic covalent bonding is present, the properties that the material is expected
to exhibit if this is the dominant mechanism are tested. The main hypothesis for the mechanism involves
the endless substitution of oxygens atoms in the B—O bond with oxygens in neighboring hydroxyl bearing
moieties, which do not involve breakage of covalent bonds but rather a ligand substitution (Scheme 1, e)
The high energy of the B—O bond (537.6 kJ mol
-1,
which is even higher than Si—O bond energy 422 kJ mol
-
1
) [35–37] together with the fact that methanol is released during condensation synthesis from trimethyl
borate [18] supports this hypothesis. This mechanism is aligned with ligand substitution observed in
boronic esters, which were shown to participate in associative exchange either through transesterification
(in the presence of free diols) or metathesis and certainly occurs without any bond cleavage [21]. When
oxygen from a free silanol group aligns with the trifunctional boron crosslinking sights, oxygen lone electron
pairs facilitate ligand swap resulting in a new attachment of Si to the boron atom through oxygen (Scheme
1, e). Though, it is important to highlight that in this mechanism, some amount of free -OH groups are
necessary to be present to keep the dynamic equilibrium of the exchange.
Considering the proposed mechanism of dynamic crosslinking, we anticipated that only a boronic
compound with three B—O functionalities will lead to gelation of PBS, while a two-functional crosslinker
will rather provide chain extension. However, it would be different in case if dative bonding plays a role in
the mechanism since in this scenario boron would become 4-cordinate species and could still bond to 3
separate PDMS molecules which in fact is sufficient to give crosslinked material and gelation. Hydrogen
bonding mechanism, in contrast, would only yield chain extension in both cases of three- and two-
functional boronic compounds.
9
Therefore, the synthesis involving a bifunctional crosslinker is essential to support the hypothesis of a
dynamic covalent bonding mechanism in PBS. If our hypothesis of dynamic covalent bonding is correct, we
expect to observe increased viscosity, but not gelation as seen in the previous experiments. Three hydroxy-
terminated PDMS were reacted with stoichiometric concentrations of phenylboronic acid, which is
bifunctional. The results showed an increase in viscosity, but no gelation was observed, confirming the
hypothesis. The apparent viscosities of the filtered and dried samples were measured using a shear mode
with parallel plate geometry (Fig. 3, a) and all showed Newtonian behavior in the measured range of shear
stresses, allowing for easy comparison of apparent viscosities. Despite the increase of apparent viscosity
up to 8 times with respect to silicone oil precursors, the number average molecular weight
𝑀
𝑁
of PBSs was
found to be similar to the precursor values (Table S3). The viscosity of P25 increased by a factor of 4,
corresponding to an increase in molecular weight by a factor of 8, as calculated using the Barry method
[38]. Similar results were reported by Gridina et al. [39] where the intrinsic viscosity of PBS from PBA
increased 5 times, corresponding to a ten-fold increase in molecular weight. They suggested that
bifunctional boron compounds like PBA provide only linear chain extensions, while trifunctional
crosslinking leads to a rubbery state. Therefore, in our case the bifunctional crosslinking resulted in chain
extension only, which explains the absence of gelation (Fig. 3, b).
Fig. 3 (a) Apparent viscosities as a function of shear rate. In the selected range of rates, all PBA-based PBS
demonstrated Newtonian flow. Numbers indicate an n-times increase in apparent viscosities with respect to the
silicone oil precursor. The apparent viscosities values against MN obtained using GPC analysis could be find in Table
S3. (b) Illustration of the condensation reaction outcome between phenylboronic acid (PBA) and hydroxy-terminated
PDMS of various molecular weights based on dynamic covalent bonding mechanism. Difunctional dynamic bonding
results in a linear extension only. The number of monomer units in PDMSs was calculated based on the GPC
measurements.
FTIR analysis of the synthesized samples confirms the completion of the condensation between
hydroxyls of phenylboronic acid and silanol groups. The strong O-H stretching peak around 3300 cm
-1
disappeared and new peaks in the range of trigonal asymmetric B—O stretching (1400-1300 cm
-1
) appeared
as was observed for B25 (Fig. 4, a). Gaussian fitting of these peaks revealed two additional bands at 1316
and 1297 cm
-1
(Fig. 4, b). The 1316 cm
-1
band was assigned to organic vibrations of the phenyl ring, based
on literature [40] and its absence in the PBA precursor could be due to its crystalline form. The 1297 cm
-1
band is due to in-plane C-H deformations of the phenyl ring, also observed in the PBA spectrum. The same
three B—O asymmetric stretching bands from B25 were observed in P25, indicating bifunctional chain
extension through Si—O—B bonds. The intensities of these bands changed with the concentration of Si—
O—B bonds, in the same manner as shown in our previously published work
[41]. The presence of the
broad
𝜈
3
band at 1368 cm
-1
in P25 indicates the absence of boroxines or other anhydrides, as their
10
formation is impossible with only two available hydroxyl groups in PBA. These peaks were not present in
another study [17] where chain ends of siloxane were modified with phenylboronic acid through an
ethylene bridge, indicating that they are inherent to direct attachment of boron to silicon through oxygen.
The rest of the new peaks appeared with low intensities and at the same positions as in PBA spectra,
suggesting the presence of unreacted PBA in bulk. Two small in-plane B—OH bending bands were also seen
in P25, from attached (1140 cm
-1
) and free PBA molecules (1184 cm
-1
).
Fig. 4 (a) FTIR spectra of PBS samples synthesized with bifunctional phenylboronic acid (PBA) from different MW PDMS. The
inset is the zoomed-in plot in the 1200-1500 cm-1 range. (b) Gaussian fits of the P25 spectra in a trigonal B—O asymmetric
stretching zone (R2 = 0.9999). (c) 29Si (left) and 11B (right) single pulse MAS NMR data of the P25 sample in comparison to the
BA and PDMS precursors as shown in Fig. 2.
No tetragonal B—O asymmetric stretching bands (1050-850 cm
-1
) were observed in the P25 FTIR
spectra. The absence of 4-coordinate boron species was additionally confirmed through solid-state
11
B
NMR. A single broad peak of 3-coordinate boron species was seen (Fig. 4, c), with a pronounced downfield
shift compared to B25. This difference is caused by large deshielding effects caused by the directly bonded
aromatic ring, and the chemical shift is comparable to that of boronic acids [42]. In the corresponding
29
Si
9
NMR spectra, the peak representing terminal Si units shifted further upfield than in B25 due to the higher
proximity of the phenyl ring in its second coordination sphere.
To verify that dynamic covalent bonding is the primary mechanism contributing to the observed
viscoelastic behavior, the dynamic moduli of B25 and B65 (Fig. 5, a) were fitted with Maxwell viscoelastic
model (Fig. 5, b). The results indicated that the model is suitable for describing the behavior of systems
that rely solely on dynamic exchange mechanisms. When additional mechanisms such as permanent
crosslinking are present, the dynamic behavior becomes more complex, and the Maxwell model is no
longer applicable. Thus, the absence of significant contribution from other mechanisms confirms t hat
dynamic covalent bonding is the primary contributor to the observed behavior.
Fig. 5 (a) Maxwell viscoelastic model fit of storage moduli (G’) and loss moduli (G”) of B25 and B65 samples obtained
during small amplitude oscillatory frequency sweep tests. Parameters of the fit can be found in Table S4. (b)
Illustration of condensation reaction outcome between boric acid (BA) and hydroxy-terminated PDMS of various
molecular weights based on dynamic covalent bonding mechanism. This reaction leads to the gelation of PBS product
due to the trifunctional dynamic crosslinking of PDMS molecules. The number of monomer units in PDMSs was
calculated based on the GPC measurements.
Rheology analysis and research have indicated that all dynamic crosslinks in PBS are of the same nature
and are temporary, with the same lifetime. They all participate in the exchange and are of a covalent
nature, with no hydrogen bonds, dative bonds, or anhydride complexes found in PBS.
The dynamic covalent bonding suggests that Si—O—B bonds are easy to form and break. Reports
suggest that the formation of these bonds in the presence of hydroxyl end groups occurs easily at room
temperature (RT) due to the low energy barrier [16,18,39]. However, in the absence of available
condensation sites, these bonds cannot be formed at RT without cleavage of Si—O—Si bonds. Despite this,
most studies have not considered RT as a reaction condition.
To test the hypothesis that the bond formation has a low energy barrier, we reacted PDMS 25 with boric
acid using only mechanical stirring at room temperature for two hours. Gel formation was observed after
two hours of stirring, and the stirring was stopped when the magnetic bar could no longer rotate. We
compared the FTIR spectra of this sample (B25RT) with conventionally synthesized B25. Their spectra
showed a remarkable match (Fig. S2), indicating that the condensation reaction occurs similarly at room
temperature and during heating. Moreover, their complex viscosities are in a good match as well (Fig. S3).
Thus, the Si—O—B bond formation and breaking indeed is highly achievable at room temperature, which
is in good agreement with the dynamic covalent bonding mechanism.
12
Six solvents (IPA, hexane, chloroform, acetone, DMF, and DI water) were tested for their ability to break
dynamic crosslinks in PBS (Table S5). IPA was the most efficient solvent, dissolving PBS quickly with no
visible BA residues. Chloroform and hexane were also effective, with some BA residue seen in chloroform
solution. Acetone dissolved PBS slower but with no visible residues. Water initially had no effect on PBS,
but after a few days it completely hydrolyzed it. DMF showed no effect on PBS. IPA's efficiency is due to its
moderate polarity, allowing it to diffuse into PBS, partially hydrolyze Si—O—B bonds and partially disrupt
exchange by competing with silanol groups. Chloroform, being more polar, enters the bulk at a slower rate
and hydrolyzes boron crosslinks to a greater extent. Hexane's non-polar nature allows BA to remain
attached to PDMS molecules without any hydrolysis. Acetone's polarity means it takes longer to penetrate
the hydrophobic PDMS surface, with potential dissolution of released BA after hydrolysis. DMF's high
polarity index probably prevented it from having any effect. Water, despite its high polarity, was able to
penetrate and hydrolyze all Si—O—B bonds due to small molecule size, hence higher mobility. Alcohols
have been found to be better solvents for PBS than ketones, with no effect from amides. Alkanes are good
solvents which preserve molecular structure without hydrolysis of B—O bonds. The effectiveness of the
solvent is an interplay between its polarity and the type of oxygen-bearing group, with -OH being the best
due to its higher reactivity towards the B—O oxygen exchange. Therefore, these results provide additional
evidence towards a dynamic covalent bonding mechanism.
To summarize, the results of this study suggest that hydrogen bonding is unlikely to be the mechanism
of dynamic crosslinking in PBS. Hydrogen bonding between hydroxyl groups of BA was one of the first
proposed mechanisms [12], but later it was found that dimeric hydrogen bonding between two boronic
acids is preferred over monomeric bonding due to sterical advantages [19]. However, the presence and
contribution of H-bonding to dynamic properties were debated. The disappearance of both Si—OH and
B—OH vibration bands on FTIR suggest that gelation in PBS is due to a condensation reaction between
silanol groups and boric acid, rather than hydrogen bonding. Furthermore, a comparison of viscoelastic
properties between PBS and permanently crosslinked silicone rubber with the same crosslinking density
suggests that all dynamic crosslinks in PBS are covalent in nature [18].
The ability of boron forming dative bonds with oxygen was proposed as another explanation of dynamic
crosslinking in PBS alongside H-bonding. It was believed that the dative bonding takes place between a
boron atom and oxygen of siloxane chains only [16,39,43,44], until it was found that boron can form dative
bonds with the oxygen of another BA molecule [17]. In addition, dative bonds are identical to covalent in
terms of length and strength, which makes it difficult to distinguish between them [45]. However, the
absence of 4-coordinate boron species in the
11
B NMR spectra and tetragonal B—O bands on FTIR indicates
that dative bonds are not present in the PBS samples, and other mechanisms should be considered.
Boroxines and boron anhydrides were also considered as possible mechanisms of dynamic crosslinking
in PBS.
According to Liu et al. [13], boroxines can form in PBS when the sample is significantly dehydrated,
but once the sample is exposed to air moisture, these boroxines hydrolyze into BA.
The presence of
boroxines in PBS was correlated with the appearance of an additional time scale on the loss modulus and
the absence of a clear rubbery plateau [13]. The efficient gelation of PBS without dehydration together
with the absence of free hydroxyl moieties at a stoichiometric concentration of BA and the absence of an
additional time scale on the dynamic moduli curves (Fig. 5a) suggests that boroxines are unlikely to be
present in the PBS samples.
Boronic acids are widely used as a building block in supramolecular assemblies and polymers due to
their dynamic covalent bonding ability [19,21,23,25]. Boronic esters and boroxines are the most studied
B—O-containing compounds, which participate in dynamic exchange through both dissociative [21] and
associative mechanisms [21,29,46]. The mechanism proposed by Bloomfield [18], where newly formed Si—
O—B bonds participate in an associative exchange with each other through ligand exchange, resembles
transesterification exchange in boronic esters. This mechanism is called dynamic covalent bonding and
9
explains the experimental data and observations. We demonstrated the absence of other types of dynamic
bonds (hydrogen and dative) and boroxines. The efficiency of alcohol and water molecules in disrupting
dynamic bonds in PBS suggests that hydroxyl-bearing moieties interfere with the dynamic exchange by
competing with silanol groups. Linear chain extension without signs of gelation was demonstrated in the
synthesis of PBS from bifunctional PBA molecules. The initial formation of Si—O—B bonds occurs through
a dissociative mechanism, while the associative bonds exchange occurs through oxygen ligand exchange.
However, the kinetics of Si—O—B bonds associative exchange needs further investigation.
Conclusions
The analysis of spectroscopic and rheological data from this study has shed light on the mechanism of
dynamic crosslinking in PBS. All dynamic crosslinks in PBS were shown to be temporary, to participate in
the exchange, and to be covalent in nature, with three-functional Si—O—B bonds being the key bonding
species responsible for gelation and viscoelastic behavior. This study proved the absence of hydrogen
bonds, dative bonds, and anhydride complexes in PBS, and demonstrated that the only mechanism present
is dynamic covalent bonding as observed in boronic esters. This mechanism includes oxygen ligand
exchange reactions between -OH bearing molecules and covalently attached boron trifunctional crosslinks
through B—O functionalities. The study also found that the viscoelastic properties of PBS can be adjusted
by the choice of boron B—O functionality, leading to n-functional dynamic crosslinking through Si—O—B
bonds. In summary, the results provide a clear understanding of the mechanism of dynamic crosslinking in
PBS, resolving long-standing controversies in the field. The research highlights the importance of boron
chemistry for the development of new materials with tunable mechanical properties for a wide range of
applications, including biocompatible materials, coatings, and adhesives.
Author Contributions
Anatoli Kurkin
: Conceptualization, Methodology, Investigation, Writing - Original Draft.
Yulia Lekina
:
Writing - Review & Editing, Data Curation.
David G. Bradley
: Investigation, Writing - Review & Editing.
Geok
Leng Seah
:
Investigation.
Kwan Wee Tan
: Resources.
Vitali Lipik
: Supervision, Project administration.
John
V. Hanna
: Data Curation, Writing - Review & Editing,
Xin Zhang
: Writing - Review & Editing.
Alfred Tok
:
Supervision, Project administration.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
JVH acknowledges financial support for the solid state NMR instrumentation at Warwick used in this
research which was funded by EPSRC (grants EP/M028186/1 and EP/K024418/1), the University of
Warwick, and the Birmingham Science City AM1 and AM2 projects, which, in turn, were supported by
Advantage West Midlands (AWM) and the European Regional Development Fund (ERDF). YL acknowledges
Ministry of Education (MOE) for funding the research through the following grants: AcRF Tier 1 (Reference
No: RG57/21); AcRF Tier 2 (Reference No: MOE-T2EP50220-0020, and MOE-T2EP50122-0005).
14
Declaration of Generative AI and AI-assisted technologies in the writing process.
Statement: During the preparation of this work the authors used Chat GTP in order to improve
readability and language. After using this tool, the authors reviewed and edited the content as needed and
take full responsibility for the content of the publication.
References
[1] R.R. McGregor, E.L. Warrick, Treating dimethyl silicone polymer with boric acid, 2431878, 1947.
[2] C. Zhao, X. Gong, S. Wang, W. Jiang, S. Xuan, Shear Stiffening Gels for Intelligent Anti-impact Applications,
Cell Rep Phys Sci. 1 (2020) 100266. https://doi.org/10.1016/j.xcrp.2020.100266.
[3] S. Wang, L. Ding, X. Fan, W. Jiang, X. Gong, A liquid metal-based triboelectric nanogenerator as stretchable
electronics for safeguarding and self-powered mechanosensing, Nano Energy. 53 (2018) 863–870.
https://doi.org/10.1016/j.nanoen.2018.09.035.
[4] J. Chen, G. Zhu, W. Yang, Q. Jing, P. Bai, Y. Yang, T.-C. Hou, Z.L. Wang, Harmonic-Resonator-Based
Triboelectric Nanogenerator as a Sustainable Power Source and a Self-Powered Active Vibration Sensor,
Advanced Materials. 25 (2013) 6094–6099. https://doi.org/10.1002/adma.201302397.
[5] G. Zhu, B. Peng, J. Chen, Q. Jing, Z. Lin Wang, Triboelectric nanogenerators as a new energy technology:
From fundamentals, devices, to applications, Nano Energy. 14 (2015) 126–138.
https://doi.org/10.1016/j.nanoen.2014.11.050.
[6] S. Wang, L. Ding, Y. Wang, X. Gong, Multifunctional triboelectric nanogenerator towards impact energy
harvesting and safeguards, Nano Energy. 59 (2019) 434–442.
https://doi.org/10.1016/j.nanoen.2019.02.060.
[7] S. Wang, S. Xuan, W. Jiang, W. Jiang, L. Yan, Y. Mao, M. Liu, X. Gong, Rate-dependent and self-healing
conductive shear stiffening nanocomposite: a novel safe-guarding material with force sensitivity, J Mater
Chem A Mater. 3 (2015) 19790–19799. https://doi.org/10.1039/C5TA06169E.
[8] S. Wang, L. Gong, Z. Shang, L. Ding, G. Yin, W. Jiang, X. Gong, S. Xuan, Novel Safeguarding Tactile e-Skins
for Monitoring Human Motion Based on SST/PDMS-AgNW-PET Hybrid Structures, Adv Funct Mater. 28
(2018) 1707538. https://doi.org/10.1002/adfm.201707538.
[9] S. Wang, S. Xuan, M. Liu, L. Bai, S. Zhang, M. Sang, W. Jiang, X. Gong, Smart wearable Kevlar-based
safeguarding electronic textile with excellent sensing performance, Soft Matter. 13 (2017) 2483–2491.
https://doi.org/10.1039/C7SM00095B.
[10] C.S. Boland, U. Khan, G. Ryan, S. Barwich, R. Charifou, A. Harvey, C. Backes, Z. Li, M.S. Ferreira, M.E.
Möbius, R.J. Young, J.N. Coleman, Sensitive electromechanical sensors using viscoelastic graphene-
polymer nanocomposites, Science (1979). 354 (2016) 1257–1260.
https://doi.org/10.1126/science.aag2879.
[11] F. Yuan, S. Wang, S. Zhang, Y. Wang, S. Xuan, X. Gong, A flexible viscoelastic coupling cable with self-
adapted electrical properties and anti-impact performance toward shapeable electronic devices, J Mater
Chem C Mater. 7 (2019) 8412–8422. https://doi.org/10.1039/C9TC01980D.
9
[12] R.L. Vale, The synthesis and irradiation of polyborosiloxanes, Journal of the Chemical Society . (1960)
2252–2257. https://doi.org/10.1039/jr9600002252.
[13] Z. Liu, S.J. Picken, N.A.M. Besseling, Polyborosiloxanes (PBSs), synthetic kinetics, and characterization,
Macromolecules. 47 (2014). https://doi.org/10.1021/ma500632f.
[14] G.A. Zinchenko, V.P. Mileshkevich, N.V. Kozlova, Investigation of the synthesis and hydrolytic degradation
of polyborodimethylsiloxanes, Polymer Science U.S.S.R. 23 (1981) 1421–1429.
https://doi.org/10.1016/0032-3950(81)90109-X.
[15] M. Wick, Bor-Siloxan-Elastomere, Kunststoffe. 50 (1960) 433–436.
[16] L.A. Mitrofanov, Ye.A. Sidorovich, A.V. Karlin, A.I. Marei, The question of molecular interaction in
polyborondimethylsiloxanes (PBDMS), Polymer Science U.S.S.R. 11 (1969) 882–888.
https://doi.org/10.1016/0032-3950(69)90312-8.
[17] G.Y. Foran, K.J. Harris, M.A. Brook, B. Macphail, G.R. Goward, Solid State NMR Study of Boron Coordination
Environments in Silicone Boronate (SiBA) Polymers, Macromolecules. 52 (2019) 1055–1064.
https://doi.org/10.1021/acs.macromol.8b02442.
[18] L.A. Bloomfield, Borosilicones and viscoelastic silicone rubbers: network liquids and network solids,
ArXiv.Org [Cond-Mat.Soft]. (2018).
[19] S.D. Bull, M.G. Davidson, J.M.H. van den Elsen, J.S. Fossey, A.T.A. Jenkins, Y.-B. Jiang, Y. Kubo, F. Marken, K.
Sakurai, J. Zhao, T.D. James, Exploiting the Reversible Covalent Bonding of Boronic Acids: Recognition,
Sensing, and Assembly, Acc Chem Res. 46 (2013) 312–326. https://doi.org/10.1021/ar300130w.
[20] W.L.A. Brooks, B.S. Sumerlin, Synthesis and Applications of Boronic Acid-Containing Polymers: From
Materials to Medicine, Chem Rev. 116 (2016) 1375–1397. https://doi.org/10.1021/acs.chemrev.5b00300.
[21] A.P. Bapat, B.S. Sumerlin, A. Sutti, Bulk network polymers with dynamic B–O bonds: healable and
reprocessable materials, Mater Horiz. 7 (2020) 694–714. https://doi.org/10.1039/C9MH01223K.
[22] L. Zepeda-Velazquez, B. Macphail, M.A. Brook, Spread and set silicone–boronic acid elastomers, Polym
Chem. 7 (2016) 4458–4466. https://doi.org/10.1039/C6PY00492J.
[23] M. Gosecki, M. Gosecka, Boronic Acid Esters and Anhydrates as Dynamic Cross-Links in Vitrimers, Polymers
(Basel). 14 (2022) 842. https://doi.org/10.3390/polym14040842.
[24] A.L. Korich, P.M. Iovine, Boroxine chemistry and applications: A perspective, Dalton Trans. 39 (2010) 1423–
1431. https://doi.org/10.1039/B917043J.
[25] X. Zhang, Y. Zhao, S. Wang, X. Jing, Cross-linked polymers based on B–O bonds: synthesis, structure and
properties, Mater Chem Front. 5 (2021) 5534–5548. https://doi.org/10.1039/D1QM00514F.
[26] R. Nishiyabu, Y. Kubo, T.D. James, J.S. Fossey, Boronic acid building blocks: tools for sensing and
separation, Chemical Communications. 47 (2011) 1106. https://doi.org/10.1039/c0cc02920c.
[27] R. Nishiyabu, Y. Kubo, T.D. James, J.S. Fossey, Boronic acid building blocks: tools for self assembly, Chem.
Commun. 47 (2011) 1124–1150. https://doi.org/10.1039/C0CC02921A.
16
[28] O.R. Cromwell, J. Chung, Z. Guan, Malleable and Self-Healing Covalent Polymer Networks through Tunable
Dynamic Boronic Ester Bonds, J Am Chem Soc. 137 (2015) 6492–6495.
https://doi.org/10.1021/jacs.5b03551.
[29] W.A. Ogden, Z. Guan, Recyclable, Strong, and Highly Malleable Thermosets Based on Boroxine Networks, J
Am Chem Soc. 140 (2018) 6217–6220. https://doi.org/10.1021/jacs.8b03257.
[30] M. Tang, W. Wang, D. Xu, Z. Wang, Synthesis of structure-controlled polyborosiloxanes and investigation
on their viscoelastic response to molecular mass of polydimethylsiloxane triggered by both chemical and
physical interactions, Ind Eng Chem Res. 55 (2016) 12582–12589.
https://doi.org/10.1021/acs.iecr.6b03823.
[31] P. Qu, C. Lv, Y. Qi, L. Bai, J. Zheng, A highly stretchable, self-healing elastomer with rate sensing capability
based on a dynamic dual network, ACS Appl Mater Interfaces. 13 (2021) 9043–9052.
https://doi.org/10.1021/acsami.1c00282.
[32] D. Peak, G.W. Luther, D.L. Sparks, ATR-FTIR spectroscopic studies of boric acid adsorption on hydrous ferric
oxide, Geochim Cosmochim Acta. 67 (2003) 2551–2560. https://doi.org/10.1016/S0016-7037(03)00096-6.
[33] X. Li, D. Zhang, K. Xiang, G. Huang, Synthesis of polyborosiloxane and its reversible physical crosslinks, RSC
Adv. 4 (2014) 32894–32901. https://doi.org/10.1039/C4RA01877J.
[34] M.K. Smith, B.H. Northrop, Vibrational properties of boroxine anhydride and boronate ester materials:
Model systems for the diagnostic characterization of covalent organic frameworks, Chemistry of Materials.
26 (2014) 3781–3795. https://doi.org/10.1021/cm5013679.
[35] L. Li, J. Zhao, H. Li, T. Zhao, Synthesis of hyperbranched polymethylvinylborosiloxanes and modification of
addition-curable silicone with improved thermal stability, Appl Organomet Chem. 27 (2013) 723–728.
https://doi.org/10.1002/aoc.3070.
[36] X. Zhao, C. Zang, Y. Sun, K. Liu, Y. Wen, Q. Jiao, Borosiloxane oligomers for improving adhesion of addition-
curable liquid silicone rubber with epoxy resin by surface treatment, J Mater Sci. 53 (2018) 1167–1177.
https://doi.org/10.1007/s10853-017-1589-1.
[37] G.D. Soraru, F. Babonneau, C. Gervais, N. Dallabona, Hybrid RSiO1.5/B2O3 Gels from Modified Silicon
Alkoxides and Boric Acid, J Solgel Sci Technol. 18 (2000) 11–19.
https://doi.org/10.1023/A:1008733511697.
[38] A.J. Barry, Viscometric Investigation of Dimethylsiloxane Polymers, J Appl Phys. 17 (1946) 1020–1024.
https://doi.org/10.1063/1.1707670.
[39] V.F. Gridina, A.L. Klebanskii, L.P. Dorofeyenko, L.Ye. Krupnova, The synthesis and properties of
borosiloxane polymers with low boron content, based on boron compounds of different structure,
Polymer Science U.S.S.R. 9 (1968) 2196–2202. https://doi.org/10.1016/S0032-3950(68)80019-X.
[40] J.A. Faniran, H.F. Shurvell, Infrared spectra of phenylboronic acid (normal and deuterated) and diphenyl
phenylboronate, Can J Chem. 46 (1968) 2089–2095. https://doi.org/10.1139/v68-341.
9
[41] A. Kurkin, V. Lipik, K.B.L. Tan, G.L. Seah, X. Zhang, A.I.Y. Tok, Correlations Between Precursor Molecular
Weight and Dynamic Mechanical Properties of Polyborosiloxane (PBS), Macromol Mater Eng. 306 (2021)
2100360. https://doi.org/10.1002/mame.202100360.
[42] J.W.E. Weiss, D.L. Bryce, A solid-state 11B NMR and computational study of boron electric field gradient
and chemical shift tensors in boronic acids and boronic esters, Journal of Physical Chemistry A. 114 (2010)
5119–5131. https://doi.org/10.1021/jp101416k.
[43] P. Qu, C. Lv, Y. Qi, L. Bai, J. Zheng, A Highly Stretchable, Self-Healing Elastomer with Rate Sensing Capability
Based on a Dynamic Dual Network, ACS Appl Mater Interfaces. 13 (2021) 9043–9052.
https://doi.org/10.1021/acsami.1c00282.
[44] Q. Wu, H. Xiong, Y. Peng, Y. Yang, J. Kang, G. Huang, X. Ren, J. Wu, Highly Stretchable and Self-Healing
“Solid–Liquid” Elastomer with Strain-Rate Sensing Capability, ACS Appl Mater Interfaces. 11 (2019) 19534–
19540. https://doi.org/10.1021/acsami.9b05230.
[45] A. Haaland, Covalent versus Dative Bonds to Main Group Metals, a Useful Distinction, Angewandte Chemie
International Edition in English. 28 (1989) 992–1007. https://doi.org/10.1002/anie.198909921.
[46] P.M. Iovine, C.R. Gyselbrecht, E.K. Perttu, C. Klick, A. Neuwelt, J. Loera, A.G. DiPasquale, A.L. Rheingold, J.
Kua, Hetero-arylboroxines: the first rational synthesis, X-ray crystallographic and computational analysis,
Dalton Transactions. (2008) 3791. https://doi.org/10.1039/b804705g.
