Light-triggered thermal conductivity switching in azobenzene polymers
- aDepartment of Materials Science and Engineering, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
- bMaterials Research Laboratory, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
- cBeckman Institute for Advanced Science and Technology, University of Illinois at Urbana–Champaign, Urbana, IL 61801;
- dX-ray Science Division, Argonne National Laboratory, Argonne, IL 60439;
- eMaterials and Manufacturing Directorate, Air Force Research Laboratory, Wright Patterson Air Force Base, OH 45433
See allHide authors and affiliations
Edited by Timothy M. Swager, Massachusetts Institute of Technology, Cambridge, MA, and approved February 12, 2019 (received for review October 19, 2018)

Significance
Heat is carried as diffusion of vibrational modes in insulating polymers, a process that is highly dependent on the macromolecular ordering of a polymer. As a result, changes in macromolecular ordering have potential to significantly change the thermal transport property of a polymer. Here, we design and synthesize a thermally switchable azobenzene polymer that exhibits a reversible crystal-to-liquid transition in response to UV and visible light. By driving a transition between the planar (trans) and nonplanar (cis) conformational states of azobenzene moieties attached to the polymer, we modulate interchain π-π bonding, resulting in fast and reversible thermal and structural transitions. This work unravels the pathway of crystal-to-liquid transitions of the azobenzene polymer and the resulting thermal and physical property changes.
Abstract
Materials that can be switched between low and high thermal conductivity states would advance the control and conversion of thermal energy. Employing in situ time-domain thermoreflectance (TDTR) and in situ synchrotron X-ray scattering, we report a reversible, light-responsive azobenzene polymer that switches between high (0.35 W m−1 K−1) and low thermal conductivity (0.10 W m−1 K−1) states. This threefold change in the thermal conductivity is achieved by modulation of chain alignment resulted from the conformational transition between planar (trans) and nonplanar (cis) azobenzene groups under UV and green light illumination. This conformational transition leads to changes in the π-π stacking geometry and drives the crystal-to-liquid transition, which is fully reversible and occurs on a time scale of tens of seconds at room temperature. This result demonstrates an effective control of the thermophysical properties of polymers by modulating interchain π-π networks by light.
At the most fundamental level, the chemical structure of a polymer dictates its properties (1). Thus, stimuli-modulated reversible chemical transitions can drive reversible property changes (2). Light (3⇓–5), electric (6) and magnetic (7) fields, temperature (8), redox reactions (9), mechanical force (10), and changes in pH (11) have all been demonstrated as triggers for reversible physical and chemical property transitions for applications of polymers in sensing, drug delivery, actuation, and self-healing (12⇓–14). However, no polymers have been shown to undergo extreme changes in macromolecular ordering, e.g., crystal-to-liquid, in response to nonthermal stimuli. Here, we describe a photoresponsive azobenzene-based polymer (azopolymer) exhibiting an unprecedented reversible crystal-to-liquid transition driven by UV (375 nm) and green (530 nm) light-triggered modulation of interchain π-π interactions. This is an observation of a reversible phototriggered crystal-to-liquid transition in a polymeric material. In conjunction with the structural transition, the polymer exhibits a notable thermal conductivity contrast, r, between the crystal, Λhigh and liquid, Λlow states, with r = Λhigh/Λlow ∼ 3.5. In comparison, r of a giant-magneto-thermoresistive material is ∼1.8 (15); r of a liquid crystal network switched with a magnetic field is ∼1.5 (16); and r of electrochemically lithiated/delithiated LixCoO2 is ∼1.5 (17, 18). Our results demonstrate powerful control of the thermophysical properties of polymers by light. The fast (seconds), reversible crystal-to-liquid transition may also provide a class of polymers engineered to switch physical, optical, and thermal properties on demand, which may enable controlled molecular release, mechanical bonding, and reconfigurable thermal routing.
Results and Discussion
Light-Triggered Phase Transition.
We hypothesized that photoisomerization of azobenzene groups attached to a polymer backbone could be used to modulate interchain bonding strength, driving nonthermal switching of polymers between crystalline and amorphous/fluidic states in a fast and reversible manner. We based our hypothesis on prior works showing that small molecule azobenzenes (19, 20) could undergo optically triggered crystal-to-crystal and crystal-to-liquid transitions, and azobenzene-containing polymers underwent optically triggered glass-to-liquid transitions (21). This phase-transition mechanism of azobenzene molecules has been studied to optically trigger the phase transition of organic phase-change materials in thermal energy storage systems (22, 23).
Here, we synthesize and investigate both the structural and thermal properties of a photoresponsive azobenzene-based polymer (azopolymer). Fig. 1 shows the chemical structure of the azopolymer we synthesized, as well as optical microscopy (OM) and cross-polarized OM (POM) images of a spin-coated azopolymer film in the cis (after UV exposure) and trans (after green light exposure or time) states (sample preparation details are provided in the Method and SI Appendix).
Light-triggered phase transition of azopolymer. (A and B) Trans-azopolymer (A) and cis-azopolymer (B) structures. (C and D) Corresponding appearances of trans-azopolymer (C) and cis-azopolymer (D) films under cross-polarized OM (Upper) and OM under continuous UV and green light illuminations (Lower) (Movie S1). (Scale bars, 200 μm.)
Starting from the crystalline ground state (Fig. 1A), UV light triggers a trans-to-cis photoisomerization leading to melting of the polymer (Fig. 1B). Upon green light triggered reconversion of the cis-azobenzene groups to trans-azobenzene, the polymer returns to the crystalline state (at room temperature in the dark, the cis-to-trans isomerization will also take place over the course of several hours). The light-triggered melting and crystallization is a direct result of differences in the stacking of the planar trans- and nonplanar cis-states of the azobenzene group. The planar trans-azobenzene readily undergoes π-π stacking, while the bent cis-azobenzene, in which the two phenyl rings align in different planes (24), does not (19). As shown in the OM and POM images (Fig. 1 C and D), spherulites in the crystalline trans-azopolymer film disappear upon exposure to UV light and reappear upon exposure to green light. Under our UV and green light illumination conditions, the melting and crystallization are complete within 10 s (Movie S1 and SI Appendix, Figs. S1 and S2). Hereafter, we refer to the crystalline and liquid states of azopolymer as trans- and cis-azopolymer, respectively.
Thermal Conductivity Switching.
Across the crystal-to-liquid transition, the most significant property change observed, other than the crystal-liquid transition, is a threefold change in the thermal conductivity in a matter of seconds at room temperature. We measured the dependence of the out-of-plane thermal conductivity of the azopolymer film on exposure to UV and green light using in situ time-domain thermoreflectance (TDTR) (16, 25). Fig. 2A shows a schematic illustration for the experimental configuration of in situ TDTR measurements for azopolymer/Al/polyimide/sapphire samples under green and UV light illumination. Fig. 2B shows measured and fitted TDTR curves for trans- and cis-azopolymer films after green and UV light illumination. The thermal conductivity reversibly switches between 0.35 ± 0.05 W m−1 K−1 in trans-azopolymer to 0.10 ± 0.02 W m−1 K−1 in cis-azopolymer. These values are within the values expected for polymers: for example, the thermal conductivity of polyethylene varies from 0.1 W m−1 K−1, when randomly oriented to 90 W m−1 K−1 along the draw direction of a highly oriented crystalline fiber (26, 27).
Thermal conductivity switching of azopolymer. (A) Schematic illustration for in situ TDTR measurements for an azopolymer film under green and UV light illumination. (B) Measured and fitted TDTR data for trans- and cis-azopolymer films after green and UV light illumination, respectively. (C) Temperature-dependent thermal conductivity of trans- and cis-azopolymer films. (D) Thermal conductivity of the azopolymer film under alternating green and UV light illumination. The reversible trans-to-cis transitions between the crystalline and liquid states occur within 10 s under UV and green light illumination (320 W cm−2). Error bars represent temporal signal fluctuations and experimental uncertainty. (E) Thermal conductivity switching rate τ0−1 of azopolymer film with increasing UV intensity.
Fig. 2C shows temperature-dependent thermal conductivity of trans- and cis-azopolymers. Fig. 2D shows the thermal conductivity of the azopolymer under alternating UV and green light illumination; 90% of the change in thermal conductivity occurs within a few tens of seconds when the light intensity is 320 mW cm−2. The rate of thermal conductivity switching, τ0−1, which is the rate to achieve 90% of the total thermal conductivity change, increases from 0.016 to 0.126 s−1 as the UV light intensity increases from 38 to 630 mW cm−2 (Fig. 2E). The thermal conductivity of the liquid state azopolymer measured by TDTR agrees with the value obtained from a frequency-domain probe beam deflection (FD-PBD) method (SI Appendix, Fig. S3) (28).
UV-visible spectroscopy and differential scanning calorimetry (DSC) measurements confirm the trans-cis photoisomerization (SI Appendix). We observe an appearance of a glass transition temperature (Tg,cis) at −48 °C after exposure to UV light, which disappears upon exposure to green light. This transition temperature is markedly lower than the temperature of a small endothermic inflection in the DSC data at 31 °C, which we interpret as Tg of residual amorphous trans-azopolymer. (The trans-azopolymer melting point is 80 °C.) Previous studies of polymers with similar azobenzene side-chain structures in trans states have reported glass-transition temperatures in the range of 35 < Tg < 80 °C (29⇓–31). We note that similar large shifts in Tg as a function of isomer state have been previously observed in polybutadiene (ΔTg ∼ 90 °C) (32) and azobenzene functionalized polymers (ΔTg ∼ 60 °C) (21). All measured physical parameters, including the molecular weight, the polydispersity indices (PDIs), heat capacity per unit volume (C), Tg, Tm, and longitudinal and transverse speed of sounds (Vl and Vt) and elastic moduli (C11 and C44) for trans- and cis-azopolymers, are described in SI Appendix, Figs. S4–S7 and summarized in Table 1.
Physical properties of trans- and cis-azopolymer
While C can have a significant impact on the thermal conductivity, we observe that C is only ∼10% greater in the amorphous state than the crystalline state, and yet the crystalline trans-azopolymer exhibits an 80% higher thermal conductivity than the amorphous trans-azopolymer (Λamorphous = 0.19 W m−1K−1) (SI Appendix, Fig. S8), suggesting that crystallinity plays the dominant role in the change in thermal conductivity. We speculate that the higher thermal conductivity of crystalline trans-azopolymer is the result of stronger dispersion and longer lifetimes of vibrational modes created by the out-of-plane alignment of side chains with planar azobenzene groups. The thermal conductivity contrast of azopolymer, r ∼ 3.5, is similar to what is observed at the melting transition of n-alkanes (n = 9 ∼ 19) with r ∼ 2 ∼ 3 (33, 34). This comparison supports our assertion that we can attribute the enhancement in the out-of-plane thermal conductivity to the side-chain alignment where the fully stretched side-chain length (21 Å) is comparable to n-alkanes with n = 16–17.
Due to the limitations in the experiments associated with a very low sensitivity of the TDTR signals to in-plane thermal conductivity, our experimental data are limited to the out-of-plane thermal conductivity of azopolymer films. Nevertheless, we expect that the in-plane thermal conductivity of azopolymer film is lower than the out-of-plane thermal conductivity since the direction of thermal conduction is normal to the aligned side chains where most of heat would be carried by interchain interaction across the side-chain networks.
Macromolecular-Ordering Transitions.
Polymer films with azobenzene side-chain groups often form smectic or lamellar structures with an out-of-plane side-chain arrangement (35, 36), exhibiting characteristic side-chain interdigitation due to the π-π stacking of azobenzene groups (37). Upon UV light illumination, these interchain π-π interactions decrease dramatically with the torsional rotation of the azobenzene groups (19). Fig. 3A illustrates a possible mechanism for crystal-to-liquid transition, associated with disruption of the π-π–stacked azobenzene groups in azopolymer during the trans-to-cis isomerization. Upon UV excitation, the trans-azobenzene groups transform to nonplanar cis-azobenzene isomers with various (C-N = N-C) dihedral angles between the two phenyl rings. This torsional rotation of the azobenzene groups causes a steric hindrance for azobenzene π-π stacking (24). As the population of the cis-isomer grows, the azobenzene stacking decreases, and long-range crystalline order is lost (19). Upon exposure to the green light (cis-to-trans transition), crystalline order is recovered by stacking of the trans-azobenzene groups.
Evolution of π-π stacking of azobenzene groups during trans-to-cis photoisomerization. (A) Schematic illustration of the conformation changes of the azopolymer driven by photoisomerization. (B) 2D GIWAXS diffraction pattern of the trans-azopolymer film. (C) Azimuth integrated in situ transmission WAXS intensity of the azopolymer during the trans-to-cis transition driven by UV illumination (630 mW cm−2). (D) dπ-π and FWHM of the π-π stacking peak as a function of trans-to-cis photoisomerization time. Error bars represent the uncertainty of q. This time-dependent plot represents the dπ-π diffraction shift in C.
We performed in situ synchrotron grazing-incidence wide-angle X-ray scattering (GIWAXS), wide-angle X-ray scattering (WAXS), grazing-incidence small-angle X-ray scattering (GISAXS), and small-angle X-ray scattering (SAXS) measurements to test this hypothesis and study the temporal evolution of short- and long-range macromolecular structure created by UV and green light illumination. Fig. 3B shows a 2D GIWAXS diffraction pattern of a trans-azopolymer film. Two sets of diffraction rings can be observed. One corresponds to in-plane diffraction associated with the lateral side-chain interdigitation of azobenzene π-π stacking (dπ-π) at q = 1.68–1.78 Å−1. The second arises from out-of-plane diffraction at integral multiples of qz = 0.233 Å−1, consistent with a 2.7-nm (001) lamellar spacing of the azopolymers. The diffraction changes associated with dπ-π and lamellar spacing represent macromolecular ordering of the azopolymer at different length scales. Fig. 3C shows the time-dependent, azimuth-integrated WAXS signal intensity. Under UV light illumination, the dπ-π peak gradually broadens and shifts to lower q, while the unit cell lattice (010) reflection remains at q = 1.22 Å−1. After ∼20 s of UV illumination, the dπ-π peak fades, followed by damping of the (010) peak, leaving only a broad diffuse scattering ring. Fig. 3D shows dπ-π spacing and the dπ-π peak full width at half-maximum (FWHM) during the trans-to-cis photoisomerization. Subsequent green light illumination triggers a cis-to-trans transition in which all diffraction patterns reappear, indicating recovery of the interdigitated structure. The time scale of the phase transition corresponds to the time scale of the thermal conductivity switching observed in TDTR measurements (SI Appendix, Figs. S9–S11).
In addition to the short-range order-disorder transition of side-chain azobenzene groups revealed by the WAXS measurements, we performed in situ SAXS to study the long-range order-disorder transition of the azopolymer backbone under UV and green light illumination. Fig. 4A shows a 2D GISAXS diffraction pattern of a trans-azopolymer film. We observe integral multiples of qz = 0.233 Å−1 and qz = 0.224 Å−1, consistent with a double periodic array at short (001)s and long (001)l spacings for trans-azopolymer lamellae.
Phototriggered reversible lamellar crystallization pathway. (A) 2D GISAXS diffraction pattern of the trans-azopolymer film. (B) Azimuth integrated in situ transmission SAXS intensity of azopolymers during trans-to-cis photoisomerization (UV light). (C) Azimuth integrated transmission SAXS intensity of azopolymers during the cis-to-trans photoisomerization (green light). (D) Schematic showing the suggested crystallization pathway under green light illumination (right to left). (D, i–iii) Randomly distributed cis-azopolymer chains (i) initially form an intermediate phase consisting of noninterdigitated polymer clusters (ii), followed by the formation of the interdigitated structure with long (001)l and short (001)s spacings (iii).
Fig. 4 B and C shows the changes of azimuth-integrated SAXS intensity of azopolymers during trans-to-cis and cis-to-trans transitions. The interdigitated (001) and (002) peaks decay under UV illumination in ∼10 s. Crystallization into the lamellar structure occurs under subsequent green light illumination on a comparable time scale. Before formation of the final lamellar structure, we observe a transient diffraction peak at q = 0.189 Å−1, which can be attributed to an intermediate phase consisting of a weakly ordered structure (d = 3.3 nm) that does not give rise to higher-order diffraction peaks. The diffraction peak of this intermediate phase decays rapidly as the fully interdigitated (001)s and partially interdigitated (001)l peaks grow in intensity (SI Appendix, Figs. S12–S14). The transient nature of the intermediate phase before the formation of the lamellar structure supports a crystallization pathway that passes through a weakly ordered phase. We speculate that in this intermediate state, side chains form a noninterdigitated structure during the cis-to-trans isomerization before forming the interdigitated structure. After the azopolymer chains are interdigitated at the nucleation site, they exhibit outward growth of the spherulite structure (Fig. 1). Fig. 4D is a schematic illustration of the assembly and crystallization process outlined here. The intensity of the (001)s peak is an order of magnitude higher than the (001)l peak, suggesting that (001)s is the most dominant structure. This split peak can be attributed to the two distinct populations of lamellar spacings originating from slightly different azobenzene interdigitation configurations (37, 38) (SI Appendix, Figs. S15 and S16).
Conclusion.
In summary, we observe reversible crystal–liquid transitions in an azopolymer, which occur on the order of 10 s with a UV and green light under illumination intensities on the order of 100 mW cm−2. This transition is associated with a threefold change in thermal conductivity. UV light illumination induces a transition in the azobenzene side chains from the planar trans state to the nonplanar cis state; green light illumination reverses this transition. The torsional rotation of the phenyl rings in cis-azobenzene disrupts π-π interactions between the azobenzene groups, resulting in formation of an isotropic liquid with a thermal conductivity of 0.10 W m−1 K−1. Subsequent green light illumination induces crystallization of trans-azopolymer, which is found to have a thermal conductivity of 0.35 W m−1 K−1 and consists of crystals containing interdigitated aligned planar-azobenzene side chains. Excitingly, we find that the conformational state of specific functional groups on polymer chains regulates not only short- and long-range ordering but also the thermal transport properties of the polymer.
Methods
Sample Preparation.
All materials were purchased from Sigma-Aldrich. 4-Phenylazophenol (98%), methacrylol chloride (97%), 6-bromo-1-hexanol (97%), and cyanoisopropyl dithiobenzoate were used as received. 2,2′-Azobisisobutyronitrile (98%) was recrystallized in chloroform before use. A complete synthetic procedure for azopolymer is available in SI Appendix.
We prepared ∼280-nm thick, trans-azopolymer thin films by spin-coating 5 wt% azopolymer dissolved in a mixture of cyclopentanone and cyclohexane (9:1 weight ratio) on Al/polyimide/sapphire substrates. The substrates were first prepared by spin-coating a 250-nm polyimide film on a sapphire wafer, followed by a 60-nm Al film deposition using magnetron sputtering. The purpose of the polyimide film is to reduce heat flow between the Al transducer and the sapphire substrate and thereby to increase the sensitivity of the TDTR measurements to the thermal conductivity of the azopolymer layer. The same specimens were used throughout the in situ OM, in situ TDTR, and in situ synchrotron X-ray scattering measurements.
OM.
The OM images and movies were recorded using a VHX-5000 series Keyence Digital Microscope at 200× magnification. The phase transition of azopolymers was characterized under green (530 nm) and UV (375 nm) light illumination. We used collimated LEDs with wavelengths of 530 and 375 nm for green and UV light sources, respectively (M530L3 and M375L3; Thorlabs). We focused the light on a 0.16-cm2 area. The illuminated light intensity was controlled between 38 and 630 mW cm−2 using a variable power controller LEDD1B (Thorlabs). Movies S1 and S2 show the phase transition of the spin-coated trans-azopolymer film and powder under UV and green light illumination (630 mW cm−2).
TDTR.
We performed in situ TDTR measurements with our two-tint pump-probe Ti-sapphire laser system to heat and sense the temperature excursions of the azopolymer/Al (60 nm)/polyimide (250 nm)/sapphire samples (25, 39). We used a bidirectional temperature model in which the pump and probe beams arrive from and reflect through the transparent sapphire substrate side of the sample (16). We collected in-phase voltage (Vin) and out-of-phase voltage (Vout) from a radio frequency lock-in amplifier synchronized to the modulation of the intensity of the pump beam at f = 1.12 and 11 MHz. The temperature-dependent thermal conductivities of trans- and cis-azopolymers were measured using a temperature-controlled Instec Hot/Cold stage having an optical window with the heating and cooling rates of 10 °C min−1 under continuous green and UV illuminations (630 mW cm−2). We used the measured ratio of Vin and Vout with a varying time delay between pump and probe pulses from 80 ps and 3.6 ns to calculate the thermal properties of trans- and cis-azopolymers. The 1/e2 intensity radius was 10 μm, and the intensities of the pump and probe beam were 5 and 3 mW, respectively. The steady-state temperature rise, ΔTSS, of the probed region was ∼10 K. The thermal penetration depths are
DSC.
DSC was performed to study the phase-transition temperature of the azopolymer with trans and cis conformational states of the attached azobenzene groups. The DSC measurements were performed using a TA Instrument Q20 Differential Scanning Calorimeter equipped with a Liquid Nitrogen Cooling System. Tzero aluminum pan and lids were used as sample containers. Al2O3 was used as a reference material for calculating the specific heat of azopolymers. Dry nitrogen was used as a sample purge gas. The heating and cooling rates were 10 °C min−1. We determined the glass transition temperature (Tg), melting temperature (Tm), and crystallization temperature (Tc) of the trans- and cis-azopolymers from the inflection points in the DSC data.
Speed of Sound Measurements.
Transverse (Vt) and longitudinal (Vl) speeds of sound were measured using picosecond acoustic and surface acoustic wave measurements for crystalline azopolymer and picosecond interferometry for liquid azopolymer using a TDTR setup (41, 42). Vl was calculated from the acoustic echoes from Al/azopolymer film interphase using the relationship Vl = 2 h/Δt, where h is the thickness of azopolymer and Δt is the time between echoes. Vt was calculated by generating and measuring surface acoustic waves using a periodic elastomeric PDMS mask on trans-azopolymer film. The longitudinal (C11) and transverse (C44) elastic moduli were calculated as C11 = ρVl2 and C44 = ρVt2. For liquid-state azopolymer, Vl of liquid is given by Vl = νp/(2λnl), where νp is the in-phase TDTR probe beam oscillation frequency, λ is the wavelength of light, and nl is the refractive index of liquid.
Synchrotron X-Ray Scattering.
To study macromolecular order and the effect of light on chain alignment, we brought the azopolymer/Al/polyimide/sapphire samples to the 12-ID-B and 12-ID-C beamlines at the Advanced Photon Source at the Argonne National Laboratory for in situ synchrotron X-ray scattering measurements. We carried out GIWAXS, transmission WAXS, GISAXS, and transmission SAXS to study the short- and long-range macromolecular structural evolution of azopolymers under illumination by UV and green light. Samples were probed using 13.3 keV (12-ID-B) and 18 keV (12-ID-C) X-rays. The X-ray beam width was 200 μm. The samples were placed on a temperature-controlled stage at a temperature of 25–30 °C. The UV and green LEDs were mounted 15 cm above the sample stage. The LEDs were remotely controlled during X-ray scattering measurements. The data collection time of the detector was set at 0.5 s for the GISAXS, SAXS, and GIWAXS experiments and 1 s for the WAXS experiments. The time intervals between each data collection were 1 and 3 s for the SAXS and WAXS measurements, respectively.
Data Availability.
We deposited raw synchrotron X-ray scattering datasets in the Globus Materials Data Facility (MDF) Open, dx.doi.org/doi:10.18126/M2VH2X (43). All data are available in the main text, the SI Appendix, and MDF Open.
Acknowledgments
Sample preparation and characterization was performed at the Materials Research Laboratory and the Beckman Institute for Science and Technology at the University of Illinois at Urbana–Champaign. This work was supported by the National Science Foundation Engineering Research Center for Power Optimization of Electro-Thermal Systems, with Cooperative Agreement EEC-1449548, and Air Force Office of Scientific Research Grant FA9550-16-1-0017. This research used resources of the Advanced Photon Source, a US Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract DE-AC02-06CH11357. Authors declare no competing interests.
Footnotes
↵1J. Shin and J. Sung contributed equally to this work.
- ↵2To whom correspondence may be addressed. Email: pbraun{at}illinois.edu or d-cahill{at}illinois.edu.
Author contributions: J. Shin, J. Sung, K.M.L., and T.J.W. designed research; J. Shin, J. Sung, M.K., and B.L. performed research; J. Shin, X.X., and D.G.C. contributed new reagents/analytic tools; J. Shin, J. Sung, M.K., B.L., C.L., N.R.S., P.V.B., and D.G.C. analyzed data; and J. Shin, J. Sung, P.V.B., and D.G.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Data deposition: The raw synchrotron X-ray scattering datasets of the 2D diffraction images reported in this paper have been deposited in Globus MDF Open, dx.doi.org/doi:10.18126/M2VH2X.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1817082116/-/DCSupplemental.
Published under the PNAS license.
References
- ↵
- Koros WJ,
- Burgess SK,
- Chen Z
- ↵
- ↵
- ↵
- ↵
- ↵
- Tanaka T,
- Nishio I,
- Sun ST,
- Ueno-Nishio S
- ↵
- Thévenot J,
- Oliveira H,
- Sandre O,
- Lecommandoux S
- ↵
- ↵
- ↵
- ↵
- Dai S,
- Ravi P,
- Tam KC
- ↵
- ↵
- Zhai L
- ↵
- Wei ML,
- Gao YF,
- Li X,
- Serpe MJ
- ↵
- Kimling J, et al.
- ↵
- Shin J, et al.
- ↵
- Cho J, et al.
- ↵
- Wehmeyer G,
- Yabuki T,
- Monachon C,
- Wu JQ,
- Dames C
- ↵
- Hoshino M, et al.
- ↵
- Bushuyev OS,
- Corkery TC,
- Barrett CJ,
- Friscic T
- ↵
- Zhou H, et al.
- ↵
- Han GGD,
- Li H,
- Grossman JC
- ↵
- Han GGD,
- Deru JH,
- Cho EN,
- Grossman JC
- ↵
- ↵
- ↵
- Wang XJ,
- Ho V,
- Segalman RA,
- Cahill DG
- ↵
- Shrestha R, et al.
- ↵
- Xie X,
- Dennison JM,
- Shin J,
- Diao Z,
- Cahill DG
- ↵
- Miao TF, et al.
- ↵
- ↵
- Wu S, et al.
- ↵
- Makhiyanov N,
- Temnikova EV
- ↵
- Konstantinov VA
- ↵
- Velez C,
- Khayet M,
- de Zarate JMO
- ↵
- Guo SF,
- Sugawara-Narutaki A,
- Okubo T,
- Shimojima A
- ↵
- Seki T
- ↵
- Hassan S,
- Anandakathir R,
- Sobkowicz MJ,
- Budhlall BM
- ↵
- Bobrovsky A, et al.
- ↵
- ↵
- Xie X, et al.
- ↵
- Li DY,
- Zhao P,
- Zhao JC,
- Cahill DG
- ↵
- ↵
- Shin J, et al.
Citation Manager Formats
Article Classifications
- Physical Sciences
- Engineering