R Helen Zha1, Bas F M de Waal1, Martin Lutz2, Abraham J P Teunissen1, E W Meijer1. 1. Institute for Complex Molecular Systems and Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of Technology , 5600 MB Eindhoven, The Netherlands. 2. Crystal and Structural Chemistry, Bijvoet Center for Biomolecular Research, Utrecht University , 3584 CH Utrecht, The Netherlands.
Abstract
Monodisperse oligodimethylsiloxanes end-functionalized with the hydrogen-bonding ureidopyrimidinone (UPy) motif undergo phase separation between their aromatic end groups and dimethylsiloxane midblocks to form ordered nanostructures with domain spacings of <5 nm. The self-assembly behavior of these well-defined oligomers resembles that of high degree of polymerization (N)-high block interaction parameter (χ) linear diblock copolymers despite their small size. Specifically, the phase morphology varies from lamellar to hexagonal to body-centered cubic with increasing asymmetry in molecular volume fraction. Mixing molecules with different molecular weights to give dispersity >1.13 results in disorder, showing importance of molecular monodispersity for ultrasmall ordered phase separation. In contrast, oligodimethylsiloxanes end-functionalized with an O-benzylated UPy derivative self-assemble into lamellar nanostructures regardless of volume fraction because of the strong preference of the end groups to aggregate in a planar geometry. Thus, these molecules display more classically liquid-crystalline self-assembly behavior where the lamellar bilayer thickness is determined by the siloxane midblock. Here the lamellar nanostructure is tolerant to molecular polydispersity. We show the importance of end groups in high χ-low N block molecules, where block-copolymer-like self-assembly in our UPy-functionalized oligodimethylsiloxanes relies upon the dominance of phase separation effects over directional end group aggregation.
Monodisperseoligodimethylsiloxanes end-functionalized with the hydrogen-bonding ureidopyrimidinone (UPy) motif undergo phase separation between their aromatic end groups and dimethylsiloxane midblocks to form ordered nanostructures with domain spacings of <5 nm. The self-assembly behavior of these well-defined oligomers resembles that of high degree of polymerization (N)-high block interaction parameter (χ) linear diblock copolymers despite their small size. Specifically, the phase morphology varies from lamellar to hexagonal to body-centered cubic with increasing asymmetry in molecular volume fraction. Mixing molecules with different molecular weights to give dispersity >1.13 results in disorder, showing importance of molecular monodispersity for ultrasmall ordered phase separation. In contrast, oligodimethylsiloxanes end-functionalized with an O-benzylated UPy derivative self-assemble into lamellar nanostructures regardless of volume fraction because of the strong preference of the end groups to aggregate in a planar geometry. Thus, these molecules display more classically liquid-crystalline self-assembly behavior where the lamellar bilayer thickness is determined by the siloxane midblock. Here the lamellar nanostructure is tolerant to molecular polydispersity. We show the importance of end groups in high χ-low N block molecules, where block-copolymer-like self-assembly in our UPy-functionalized oligodimethylsiloxanes relies upon the dominance of phase separation effects over directional end group aggregation.
In an effort to overcome
the size limitations facing top-down microelectronics fabrication,
the bottom-up phenomenon of block copolymer nanophase separation is
an attractive strategy for creating nanolithography templates.[1−4] Numerous experimental and theoretical studies have shown that the
morphology and domain spacing of nanophase-separated block copolymers
can be predictably controlled by varying the degree of polymerization
(N), the volume fraction (f), and
the interaction parameter between blocks (χ).[5−10] Block copolymers containing polydimethylsiloxane (PDMS) are particularly
desirable, as the generally high incompatibility between siloxanes
and many organic components facilitates ordered phase separation at
low N, yielding correspondingly small features.[11,12] For example, our group has utilized the high χ and etch contrast
between blocks in PDMS-b-poly(lactic acid) to create
ordered arrays of sub-20 nm features.[13] However, while studies have shown that high χ–low N block copolymers can exhibit domain spacings as small
as 9 nm,[14] the structure size achievable
by traditional block copolymers is still limited by the order–disorder
transition at low molecular weight. Recent work with discrete dimethylsiloxane–lactic
acid oligomers suggests that perfect monodispersity can help push
ordered nanophase separation toward domain spacings as small as 6.8
nm.[15] Sinturel, Bates, and Hillmyer similarly
expressed that features smaller than 5 nm may be better obtained by
using small, well-defined molecules that self-assemble instead of
classical block copolymers.[16] This class
of low-molar-mass molecules is dominated by liquid crystals, which
can exhibit ordered phases with domain spacings of a few nanometers
in bulk.[17−19] Here, self-assembly is strongly guided by directional
interactions resulting from shape anisotropy and enthalpic forces
(e.g., hydrogen bonding and π stacking). Thus, liquid crystal
morphology generally depends on the molecular geometry rather than
on the interplay between entropic and enthalpic forces to minimize
the interfacial energy at a given volume fraction. For example, discotic
mesogens predominantly exhibit columnar morphologies while calamitic
mesogens are usually needed for nematic or smectic morphologies.[18] Interestingly, molecules that bridge the gap
between liquid crystals and block copolymers have been rarely studied.
Hillmyer and co-workers showed that small (2.5–8.9 kg/mol)
poly(ethylene oxide)–polyethylethylene block copolymers exhibit
phase transition behavior resembling those of both high molecular
weight block copolymers and low molecular weight nonionic lytropic
liquid crystals.[20] However, the ability
of small, discrete molecules to exhibit block-copolymer-like self-assembly
behavior has yet to be demonstrated.Here we describe the synthesis
and characterization of novel siloxane-based “block molecules”
consisting of monodisperse dimethylsiloxane oligomers (ODMS) end-capped
with ureidopyrimidinone (UPy), an aromatic motif capable of dimerizing
via quadruple hydrogen bonds.[21−23] Notably, we utilize a stepwise
procedure to synthesize ODMS of perfectly defined length and thereby
obtain oligomeric molecules with up to 24 dimethylsiloxane repeat
units without any polydispersity. We show that these well-defined
molecules exhibit phase separation between the dimethylsiloxane chain
and the aromatic end groups into ordered morphologies with sub-5 nm
domain spacings. By varying the molecule length and volume fraction,
we further show data suggesting that the self-assembly of these UPy-functionalized
ODMS resembles the behavior of block copolymers rather than liquid
crystals. We investigate the role of end group interactions in facilitating
this block-copolymer-like self-assembly by additionally studying O-benzylated
UPy end groups.
Results and Discussion
To synthesize
well-defined UPy-functionalized siloxanes with varying molecular length
and volume fraction, perfectly monodisperse ODMS with N = 4, 8, 16, or 24 repeat units was utilized (Scheme ). While monodisperse ODMS with four dimethylsiloxane
repeat units can be commercially purchased, longer monodisperse ODMS
are not readily available. Thus, we developed a stepwise procedure
to synthesize ODMS in discrete lengths using chlorosilane and silanol
building blocks. Synthetic details are presented in the Supporting Information as well as a recent paper,[15] and the general strategy for the facile synthesis
of monodisperse ODMS with various reactive or stable end groups will
be reported in greater detail in further publications. UPy bearing
a terminal olefin was synthesized for attachment to hydride-terminated
ODMS chains via hydrosilylation (Scheme ). However, in order to avoid attachment
of ODMS to the oxygen atom of the enol tautomer of UPy, a benzyl protecting
group was added prior to hydrosilylation. After purification, this
benzyl group was removed by palladium-catalyzed reduction in hydrogen
gas to give the final bis-UPy oligodimethylsiloxanes (USiN).
Scheme 1
Synthesis of UZSiN and USiN
Previous studies have shown
that PDMS functionalized with UPy exhibits phase separation between
the polymer backbone and the UPy motifs.[24,25] High-molecular-weight supramolecular polymers were obtained, and
the formation of UPy microdomains dramatically improved the elastic
properties of the material. Thus, our well-defined USiN materials were expected to be polymeric in nature and contain nanoscale
phase-separated domains. Indeed, we found that USiN molecules behave as high molecular weight copolymers instead of
oligomers. When cast from chloroform, USi4 and USi8 are clear, brittle
films while USi16 and USi24 are tacky, viscoelastic solids (Figure ). Thermal analysis
by differential scanning calorimetry (DSC) showed polymeric glass
transition temperatures ranging from −14 to 64 °C depending
on the number of dimethylsiloxane repeat units (Figure S1). An IR band at 1700 cm–1 characteristic
of the pyrimidinone carbonyl stretch vibration indicates that UPy
end groups exist in the strongly dimerizing keto form,[21,24] and an IR band at 3216 cm–1 further confirms hydrogen
bonding of the end groups to form supramolecular polymers (Figure S2).
Figure 1
(A–D) Images of bulk (A) USi4,
(B) USi8, (C) USi16, and (D) USi24. (E–G) Polarized optical
microscopy images showing birefringence in (E) USi4, (F) USi8, and
(G) USi16 materials. (H) USi24 lacks birefringence in polarized optical
microscopy.
(A–D) Images of bulk (A) USi4,
(B) USi8, (C) USi16, and (D) USi24. (E–G) Polarized optical
microscopy images showing birefringence in (E) USi4, (F) USi8, and
(G) USi16 materials. (H) USi24 lacks birefringence in polarized optical
microscopy.In stark contrast to
the liquid nature of unfunctionalized ODMS and even high molecular
weight PDMS, the solidity of USiN suggests the existence
not only of dimerized UPy end groups but also of phase-separated UPy
“hard” domains. Interestingly, polarized optical microscopy
(POM) showed USi4, USi8, and USi16 to be birefringent (Figure ), presumably resulting from the presence of optically anisotropic
material structures with periodic domains having one refractive index
within a matrix having a second refractive index. This result further
suggests that the UPy domains are well-ordered in these materials.
Small-angle X-ray scattering (SAXS) confirmed that UPy and dimethylsiloxane
moieties can phase separate into ordered structures with domain spacings
(L0) as small as 2.2 nm (Table and Figure ). Surprisingly, the phase morphology varies
depending on the volume fraction. Specifically, USi4 exhibits a lamellar
morphology, USi8 a hexagonally packed columnar morphology, and USi16
a body-centered cubic morphology. Highly asymmetric USi24 is disordered,
as it exhibits a broad, low-intensity Lorentzian peak typical of correlation
hole scattering in disordered block copolymer melts. This result agrees
with the lack of birefringence of USi24 in POM. Stacking of UPy dimers
with an interdisc distance of 3.5 Å can be seen in the SAXS spectra
of USi4 and USi8, although there is no evidence for crystallization
of the UPy domains. Such lateral aggregation is logical given the
lamellar and columnar morphologies of USi4 and USi8, respectively.
UPy stacking is less apparent in the spectrum of USi16 and is nearly
absent in the spectrum of USi24. These results likely reflect both
the constraints against significant lateral aggregation in body-centered
cubic and disordered morphologies as well the low UPy volume fraction
contribution to the scattering intensity.
Table 1
USiN Volume Fractions and Phase Characterizationsa
sample
fUPy
L0 (nm)
phase
USi4
0.52
2.2
LAM
USi8
0.35
3.2
HEX
USi16
0.21
4.8
BCC
USi24
0.15
–
DIS
The domain spacing, L0, is the bilayer
thickness for the lamellar (LAM) morphology and the center-to-center
spacing for the cylindrical (HEX) and spherical (BCC) morphologies.
The calculation of fUPy is described in Supporting Information.
Figure 2
Small-angle X-ray scattering of USiN materials.
Intensity is given in arbitrary units, and curves have been shifted
for visual clarity. The siloxane backbone and the UPy interdisc scattering
are marked with arrows.
The domain spacing, L0, is the bilayer
thickness for the lamellar (LAM) morphology and the center-to-center
spacing for the cylindrical (HEX) and spherical (BCC) morphologies.
The calculation of fUPy is described in Supporting Information.Small-angle X-ray scattering of USiN materials.
Intensity is given in arbitrary units, and curves have been shifted
for visual clarity. The siloxane backbone and the UPy interdisc scattering
are marked with arrows.Our observations suggest that unlike conventional liquid
crystal self-assembly, USiN self-assembly is not
dominated by directional interactions. Indeed, previous studies of
UPy-functionalized poly(ethylene butylene) and polycaprolactone indicate
that UPy dimers have relatively low propensity for ordered aggregation
without additional directionality provided by adjacent hydrogen-bonding
urea or urethane motifs.[26,27] Instead, the self-assembly
of USiN into ordered nanoscale morphologies appears
to be driven predominantly by phase separation and can be considered
analogous to the self-assembly of a linear diblock copolymer in which
one block is significantly more rigid than the other. In our USiN molecules, the UPy end groups represent short, rigid A
blocks and the ODMS segments represent flexible B blocks. Studies
of linear block copolymers suggest that the ABA topology exhibits
identical phase behavior as an AB topology in the strongly segregated
regime.[28,29] However, conformational asymmetry of the
blocks shifts the phase boundaries because of the tendency of the
interface to curve toward the more flexible block.[30] Thus, the appearance of lamellar, hexagonal close packed,
body-centered cubic, and disordered phases with varying USiN volume fraction is generally consistent with block copolymer
theory. It is probable that the high chemical incompatibility of dimethylsiloxane
and UPy is necessary for such ordered nanophase separation. The high
conformational freedom of ODMS chains likely further facilitates the
polymeric behavior of USiN on such molecular length
scales, as the DSC data suggest even short ODMS segments to be amorphous
on the basis of a lack of sharp dimethylsiloxane melting transitions
(Figure S1). In other words, the unusual
flexibility of dimethylsiloxane backbones[31] may allow the USiN molecules to behave more like
Gaussian polymer chains, which is an underpinning requirement in classical
block copolymer mean-field theory.[5,32] Interestingly,
narrow dispersity of the ODMS segment is vital for ordered morphologies.
Mixtures of USiN molecules of various lengths to
create artificial dispersity suggested that structures become disordered
with increased dispersity, with all order lost by Mw/Mn = 1.14 (Figure S3). Therefore, unlike conventional block copolymers,
which may in some cases form ordered morphologies even with Mw/Mn > 1.3,[33−35] the nanodomain ordering of USiN is sensitive to
polydispersity. The thermal phase behavior of USiN molecules and extraction of an effective χ value are interesting
topics for further investigation.To understand the self-assembly
of USiN molecules, the intermolecular interactions
of the end groups must be carefully considered. We thus further investigated
the self-assembly of O-benzylated USiN molecules
(UZSiN), which were synthetic precursors originally
designed to avoid improper attachment of ODMS chains to UPy during
hydrosilylation (Scheme ). SAXS showed that UZSiN molecules form lamellar
layers regardless of volume fraction (Figure A and Table ), which is additionally evident from the highly birefringent
circular focal conic and fan textures seen in POM (Figure ). Diffraction peaks of O-benzylated
UPy end groups can be observed in the SAXS spectra of samples with
high fUPy. Accordingly, UZSiN molecules have liquid-crystalline rather than polymeric attributes.
UZSi4 and UZSi8 are white powders, while UZSi16 and UZSi24 are waxy
solids (Figure ) that
may be deformed by shear at room temperature in order to align crystallographic
orientations (Figure S4). DSC showed sharp
order–disorder transitions rather than glass transitions (Figure S1). In context, previous research suggested
that liquid crystal molecules incorporating just two or three dimethylsiloxane
units prefer smectic phases because of phase separation of the dimethylsiloxanes.[36−39] However, ODMS segments longer than a few repeat units have not been
previously studied, and long aliphatic chains tend to give a nematic
morphology in conventional liquid crystals.[40,41] Thus, the propensity for lamellar morphology in UZSiN can likely be attributed to the packing preference of the O-benzylated
UPy end groups. The X-ray crystal structure of O-benzylated UPy alone
(compound 3; Figure S5) shows
that this moiety dimerizes through two hydrogen bonds rather than
the four hydrogen bonds seen in UPy dimerization (Figure S6). Dimers of O-benzylated UPy can further interact
through slipped stacking of phenyl and pyrimidine rings as well as
CH···π interactions (Figure S7). This arrangement results in a planar crystal structure
in which ODMS chains can extend away from the plane formed by the
end group motifs (Figure ). The overall nanostructure of UZSiN thus
consists of alternating layers of amorphous ODMS and crystalline O-benzylated
UPy. The incorporation of a benzyl group provides an additional source
of directional intermolecular interaction that leads to crystallization
of the UZSiN end groups in a planar geometry, thereby
driving a lamellar morphology regardless of volume fraction.
Figure 3
(A) Small-angle X-ray scattering of UZSiN materials.
Intensity is given in arbitrary units, and curves have been shifted
for visual clarity. The siloxane backbone scattering is marked with
an arrow. (B) Fit of the lamellar spacing (L0) vs the average number of dimethylsiloxane units in pure
and mixed UZSiN.
Table 2
UZSiN Volume Fraction and Phase Characterizations
sample
fUPy
L0 (nm)
phase
UZSi4
0.62
1.7
LAM
UZSi8
0.44
2.2
LAM
UZSi16
0.28
3.5
LAM
UZSi24
0.20
4.8
LAM
Figure 4
(A–D) Images of bulk (A) UZSi4, (B) UZSi8, (C) UZSi16, and
(D) UZSi24. (E–H) Polarized optical microscopy images showing
strongly birefringent smectic textures in (E) UZSi4, (F) UZSi8, (G)
UZSi16, and (H) UZSi24.
Figure 5
Planar geometry of crystallized compound 3 as seen by
viewing unit cells lying in the XZ plane. Dotted
lines are hydrogen bonds. Purple shading highlights a pair of phenyl
rings in slipped stacking, green shading highlights a pyrimidine dimer,
and blue shading highlights olefin terminals, where ODMS chains are
attached by hydrosilylation. The planar geometry can also be viewed
in the crystallography data included in the Supporting Information.
(A) Small-angle X-ray scattering of UZSiN materials.
Intensity is given in arbitrary units, and curves have been shifted
for visual clarity. The siloxane backbone scattering is marked with
an arrow. (B) Fit of the lamellar spacing (L0) vs the average number of dimethylsiloxane units in pure
and mixed UZSiN.(A–D) Images of bulk (A) UZSi4, (B) UZSi8, (C) UZSi16, and
(D) UZSi24. (E–H) Polarized optical microscopy images showing
strongly birefringent smectic textures in (E) UZSi4, (F) UZSi8, (G)
UZSi16, and (H) UZSi24.Planar geometry of crystallized compound 3 as seen by
viewing unit cells lying in the XZ plane. Dotted
lines are hydrogen bonds. Purple shading highlights a pair of phenyl
rings in slipped stacking, green shading highlights a pyrimidine dimer,
and blue shading highlights olefin terminals, where ODMS chains are
attached by hydrosilylation. The planar geometry can also be viewed
in the crystallography data included in the Supporting Information.Even though UZSiN molecules generally exhibit properties characteristic of thermotropic
liquid crystals, some comparisons to block copolymer self-assembly
can still be made. Mixing UZSiN molecules to create
artificial polydispersity yields lamellae with L0 that scales as Nav2/3 (Figures S3 and 3B), where Nav is the average number of
dimethylsiloxane repeat units. This mirrors the scaling law between L0 and the degree of polymerization in a strongly
segregated block copolymer,[6,42−44] suggesting that benzylated UPy and siloxane domains are compositionally
pure with narrow interface widths. However, mixtures containing >25%
UZSi4 deviate from this trend, suggesting that UZSi4 molecules are
too short to exhibit the conformational freedom needed to give freely
jointed polymeric behavior.The balance between secondary interactions
and phase separation in our functionalized siloxanes plays a crucial
role in directing the morphology. As shown by UZSiN molecules, when the strong planar packing preference of end group
motifs dominates structure formation, lamellae are formed regardless
of the molecular volume fraction and dispersity. However, as shown
by USiN molecules, when end groups lack a strong
propensity for directional aggregation, block-copolymer-like phase
separation behavior becomes more apparent. In this case, the molecules
exhibit volume-fraction-dependent phase behavior reminiscent of linear
diblock copolymers, where the volume fraction is readily related to
the molecular weights of the individual blocks. In contrast to amphiphilic
surfactants, our USiN “block molecules”
are entirely hydrophobic and do not require solvent as a driving force
for aggregation and interface minimization. However, in some aspects
they may be regarded as similar to surfactants,[17,45] which are frequently described by a packing parameter.[46] While the essence of a packing parameter does
reflect volume fraction, it is designed to use aspects of molecular
geometry such as headgroup size or tail branching to define the interfacial
curvature. Calculations of packing parameter require estimates of
molecular volume, hydrocarbon chain length, and effective headgroup
area. Because these parameters must be determined empirically[47,48] and are not readily applicable to our functionalized siloxanes,
the ability to describe our molecules using accessible and well-established
block copolymer concepts opens attractive avenues for understanding
and predicting their self-assembly behavior.
Conclusion
We
have presented a class of well-defined monodisperse end-functionalized
siloxane oligomers that can exhibit ordered nanodomains with sub-5
nm spacings. Depending on the balance between directional interactions
of the aromatic end groups and phase separation of the siloxane midblock
from the end groups, the self-assembly behavior of these molecules
can be similar to that of liquid crystals or linear diblock copolymers.
Thus, these functionalized siloxane oligomers bridge the gap between
classical liquid crystals and block copolymers.
Authors: Joan M Widin; Adam K Schmitt; Andrew L Schmitt; Kyuhyun Im; Mahesh K Mahanthappa Journal: J Am Chem Soc Date: 2012-02-16 Impact factor: 15.419
Authors: Bas van Genabeek; Bas F M de Waal; Mark M J Gosens; Louis M Pitet; Anja R A Palmans; E W Meijer Journal: J Am Chem Soc Date: 2016-03-21 Impact factor: 15.419
Authors: Jimmy Lawrence; Sang-Ho Lee; Allison Abdilla; Mitchell D Nothling; Jing M Ren; Abigail S Knight; Carolin Fleischmann; Youli Li; Austin S Abrams; Bernhard V K J Schmidt; Michael C Hawker; Luke A Connal; Alaina J McGrath; Paul G Clark; Will R Gutekunst; Craig J Hawker Journal: J Am Chem Soc Date: 2016-05-06 Impact factor: 15.419
Authors: Abraham J P Teunissen; José Augusto Berrocal; Christiaan H W A Corbet; E W Meijer Journal: J Polym Sci A Polym Chem Date: 2017-07-03 Impact factor: 2.702
Authors: José Augusto Berrocal; R Helen Zha; Bas F M de Waal; Jody A M Lugger; Martin Lutz; E W Meijer Journal: ACS Nano Date: 2017-04-07 Impact factor: 15.881
Authors: Brigitte A G Lamers; Robert Graf; Bas F M de Waal; Ghislaine Vantomme; Anja R A Palmans; E W Meijer Journal: J Am Chem Soc Date: 2019-09-12 Impact factor: 15.419
Authors: Zhengyuan Shen; Ke Luo; So Jung Park; Daoyuan Li; Mahesh K Mahanthappa; Frank S Bates; Kevin D Dorfman; Timothy P Lodge; J Ilja Siepmann Journal: JACS Au Date: 2022-05-31
Authors: José Augusto Berrocal; Joan Teyssandier; Olga J G M Goor; Steven De Feyter; E W Meijer Journal: Chem Mater Date: 2018-05-01 Impact factor: 9.811