ARTICLE

https://doi.org/10.1038/s41467-021-26078-1 OPEN

Band-selective Holstein polaron in Luttinger liquid

material A0.3MoO3 (A = K, Rb)
L. Kang1,8, X. Du 1,8, J. S. Zhou1, X. Gu1, Y. J. Chen1, R. Z. Xu1, Q. Q. Zhang1, S. C. Sun1, Z. X. Yin1, Y. W. Li2,3,
D. Pei4, J. Zhang2, R. K. Gu 4, Z. G. Wang 5, Z. K. Liu 2,3, R. Xiong6, J. Shi6, Y. Zhang 5,
Y. L. Chen1,2,3,4 ✉ & L. X. Yang1,7 ✉
1234567890():,;

(Quasi-)one-dimensional systems exhibit various fascinating properties such as Luttinger

liquid behavior, Peierls transition, novel topological phases, and the accommodation of unique
quasiparticles (e.g., spinon, holon, and soliton, etc.). Here we study molybdenum blue bronze
A0.3MoO3 (A = K, Rb), a canonical quasi-one-dimensional charge-density-wave material,
using laser-based angle-resolved photoemission spectroscopy. Our experiment suggests that
the normal phase of A0.3MoO3 is a prototypical Luttinger liquid, from which the charge-
density-wave emerges with decreasing temperature. Prominently, we observe strong
renormalizations of band dispersions, which are recognized as the spectral function of Hol-
stein polaron derived from band-selective electron-phonon coupling in the system. We argue
that the strong electron-phonon coupling plays an important role in electronic properties and
the charge-density-wave transition in blue bronzes. Our results not only reconcile the long-
standing heavy debates on the electronic properties of blue bronzes but also provide a rare
platform to study interesting excitations in Luttinger liquid materials.

1 State Key Laboratory of Low Dimensional Quantum Physics, Department of Physics, Tsinghua University, Beijing 100084, China. 2 School of Physical

Science and Technology, ShanghaiTech University and CAS-Shanghai Science Research Center, Shanghai 201210, China. 3 ShanghaiTech Laboratory for
Topological Physics, Shanghai 200031, China. 4 Department of Physics, Clarendon Laboratory, University of Oxford, Oxford OX1 3PU, UK. 5 International
Center for Quantum Materials, School of Physics, Peking University, Beijing 100871, China. 6 Department of Physics, Wuhan University, Wuhan 430072,
China. 7 Frontier Science Center for Quantum Information, Beijing 100084, China. 8These authors contributed equally: L. Kang, X. Du.
✉email: [email protected]; [email protected]

NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1

D
imensional confinement can strongly influence the elec- various Q1D systems, molybdenum blue bronze A0.3MoO3
tronic properties of many-body systems. In drastic con- (A = K, Rb, Tl) provides a suitable platform to investigate the LL-
trast to two- or three-dimensional systems, low-energy to-CDW transition. It exhibits a CDW transition at
particle-hole pair in the one-dimensional system can be excited TCDW ≈ 183 K21,22 (Supplementary Figure 1), accompanied by a
only if its wave vector |q | ≈ 0 or |q | ≈ 2kF with kF being the Fermi Kohn anomaly23 and a sharp peak in the Lindhard function24.
momentum of electrons (Fig. 1a). The reduction of scattering Angle-resolved photoemission spectroscopy (ARPES) measure-
phase-space and electronic screening gives rise to rich fascinating ments suggest an important role of FS nesting in the CDW
properties, such as power-law correlation function and spin- transition25,26. Nevertheless, the absence of the Fermi edge in the
charge separated collective bosonic excitations, as properly density of states27,28 has challenged this scenario and sparked
described by Luttinger liquid (LL) theory1,2. On the other hand, heavy debates about non-Fermi liquid behavior29–31 and
quasi-one-dimensional (Q1D) metals with well nested Fermi (pseudo)gap above TCDW in the system25,27,32,33. Moreover, the
surface (FS) are highly susceptible toward a charge-density-wave underlying microscopic interaction that is crucial for the CDW
(CDW) state with periodic lattice distortion and an energy gap transition is also under heavy debate25,26,33,34. These con-
near EF (Fig. 1b)3, in which electron–phonon coupling (EPC) troversial results call for a comprehensive understanding of the
usually has a crucial role. Despite the intensive theoretical single-particle spectral properties of the system.
efforts4–7, the LL-to-CDW transition is rarely studied In this work, we report a high-resolution laser-based ARPES
experimentally8–10, mainly owing to the lack of a suitable mate- study of the blue bronze A0.3MoO3 (A = K, Rb). By tracking the
rials platform. Moreover, with stronger EPC, electrons can be band dispersion over a wide temperature range, we provide
dressed by local lattice distortion, forming novel composite spectroscopic evidence of the CDW transition near TCDW ≈ 183
quasiparticles—the alleged Holstein polarons that can strongly K. Above TCDW, we observe a linear band dispersion with power-
renormalize the electronic structure of the system and induce law scaling behavior of ARPES spectral weight, which can be
energy gaps and “flat bands” at −nΩ0 (n = 1, 2…) with Ω0 being precisely described by the LL model35. Below TCDW, the LL
the energy of the strongly coupled phonon involved in the for- property of the spectra revives after the CDW gap is suppressed
mation of the polaron (Fig. 1c). Up to date, although polarons by surface doping of Rubidium (Rb) atoms. Interestingly, we
have been widely observed and well understood in two- or three- observe strong renormalizations of the band dispersions, which
dimensional materials11–16, they are yet to be discovered in Q1D are identified as the spectral function of Holstein polaron derived
materials. from band-selective EPC in the system. We conclude that the
On account of the assumptions of LL theory, it is essential for a strong EPC22,34,36,37, in addition to the FS nesting25,26, has an
real material to be a highly anisotropic metal with linear band important role in electronic properties and the LL-to-CDW
dispersion. Several crystalline materials meeting these criteria transition in blue bronzes. Our results not only help understand
have been demonstrated to be LL candidates, such as 1D the long-standing mysteries in blue bronzes, including the non-
cuprate17, Lithium purple bronze18, Chromium-based Fermi liquid behavior, the pseudogap above TCDW, and the
superconductor19, and organic metallic systems8,20. Among the mechanism of the CDW transition but also provide a rare

Fig. 1 Physics in one dimension and basic band structure of K0.3MoO3. a Luttinger liquid (LL) behavior in one dimension, where only low-energy particle-
hole pairs with wave vector |q | ≈ 0 or |q | ≈ 2kF can be excited. b Charge-density-wave (CDW) and periodic lattice distortion, leading to a gap opening near
the Fermi energy (EF). c An electron dressed by the local lattice distortion, forming the Holstein polaron and leading to band gaps and flat dispersions near
multiples of the phonon energy (Ω0). d Crystal structure of A0.3MoO3 (A = K, Rb, Tl) showing one-dimensional chains of MoO6 octahedrons. Black lines
indicate the unit cell. e Experimental Fermi surface (FS) by integrating ARPES intensity in an energy window of 20 meV near the Fermi energy (EF), with
surface Brillouin zone appended. f, g Band dispersions along and X 
Z showing the bonding (B) and anti-bonding (AB) bands measured along with cuts #1
and #2 as indicated by the orange lines in e. The dashed lines indicate the momentum position of high symmetry points. Data in e–g were collected at 80 K
with a He lamp (hν = 21.2 eV).

2 NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1 ARTICLE

Fig. 2 CDW transition in K0.3MoO3. a Temperature evolution of the band dispersion along the Γ
Y  direction. The black arrows indicate the CDW gap
opening below TCDW ≈ 183 K. The dashed lines indicate EF. b, c Energy distribution curves (EDCs) near Fermi momenta of B (b) and AB (c) bands at
different temperatures showing the CDW gap opening below TCDW. The dashed lines indicate EF. d CDW gap as a function of temperature. The CDW gap
is determined by the shift of the EDC leading edge with respect to the spectrum at high temperatures. The error bars are determined by the combination of
energy resolution and deviations between samples. The red dashed line is the fit to the BCS-type gap equation. Data were collected with a linear-
horizontally polarized laser at 7 eV.

platform to study the intriguing EPC in LL materials, which will which can be directly recognized by ARPES9,18,19,31,41. In blue
also shed light on the understanding of rich physics in other Q1D bronzes, the absence of Fermi edge was previously attributed
materials. either to non-Fermi liquid behavior30 or the pseudogap induced
by the CDW fluctuations25,27,32,33. However, compelling evidence
Results for both scenarios is yet to be derived. In Fig. 3b, we approximate
CDW transition in blue bronze. As shown in Fig. 1d, A0.3MoO3 the EDCs integrated along  Γ
Y  near EF with a LL spectral
crystallizes into the monoclinic structure with space group C2/m. function at finite temperatures35:
It consists of 1D chains of MoO6 octahedrons extending along the 

α þ 1 þ iϵπ
[010] direction and MoO3 layers intercalated by A atoms stacking ρðϵ; T Þ / T α Re ð2iÞαþ1 β ;
α ð1Þ
along the ½201 direction38. Fig. 1e shows the Q1D FSs running 2
perpendicular to b* (along kx) measured at 80 K, consistent with where ϵ ¼ ðE
EF Þ=kB T is the scaled energy, β is the beta
the Q1D crystal structure and in good agreement with our cal- function, and α is the LL anomalous exponent. The EDCs fit
culation (Supplementary Figure 3). Fig. 1f and g show band perfectly to the LL spectral function convoluted with the energy
dispersions along the Γ
Y and X 
Z directions. The bonding
resolution in our experiment with α = 0.6 ± 0.1 (Supplementary
(B) band shows a linear dispersion in a large energy range, Figure 5) as shown in Fig. 3b. Moreover, Eq. (1) suggests a scaling
whereas the anti-bonding (AB) band shows a relatively narrow relation of the EDCs, that is, ρðϵ; T Þ=T α is independent of the
bandwidth with a band bottom near 370 meV below EF, in good temperature, which is well established in Fig. 3c with the same α,
agreement with our calculation (Supplementary Figure 4). proving the LL nature of blue bronzes in the normal state.
To investigate the CDW transition in A0.3MoO3, we track the At temperatures slightly below TCDW, the spectra still fit nicely
temperature evolution of the band dispersion using laser-based to the LL model after introducing an extra fitting parameter Δ
ARPES with superb resolutions, as shown in Fig. 2. At 304 K, we accounting for the CDW gap in the spectral function (Supple-
observe a strong suppression of the spectral weight near EF and mentary Figure 6), suggesting the intimate relationship between
the absence of Fermi edge. With decreasing temperature, the the CDW state and the LL phase. Fig. 3d shows the temperature
spectral weight is further suppressed and an energy gap (ΔCDW) evolution of α. It increases slightly with decreasing temperature
gradually opens below TCDW, as also shown by the energy (above TCDW), in contrast to the two-band model proposed in
distribution curves (EDCs) near the Fermi momenta kF of the B Li0.9Mo6O1718,42. Far below TCDW, the EDCs cannot be fitted to
and AB bands (Fig. 2b, c). The CDW gap is about 42 meV at 83 K the LL model owing to the strong modification of the LL phase by
(Fig. 2d) and follows the BCS-type gap equation, in accordance the CDW ordering (Supplementary Figure 6). Interestingly, the
with previous reports27,33,39,40 and the estimation based on our LL property revives after the CDW gap is suppressed by doping
resistivity measurements (Supplementary Figure 1). Rb atoms on the sample surface as shown by the shift of the
leading edge of the EDCs in Fig. 3e and f, which saturates to a
LL nature of the normal state. As schematically shown in Fig. 3a, value of about 20 meV (Fig. 3f and Supplementary Figure 7).
the spectral density in LLs shows a power-law behavior near EF, After that, the spectrum again fits well to the LL model as shown

NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1

Fig. 3 Luttinger liquid behavior of A0.3MoO3. a The spectral density of LLs (orange curve) showing a power-law dependence on the energy. b EDCs of
Rb0.3MoO3 integrated along Γ
Y with fits to finite temperature LL model (black curves). The curves are vertically offset for clarity. c Scaling plot of the
EDCs at different temperatures. The scaling factor is Tα with α = 0.6. d The anomalous exponent α as a function of temperature measured with He lamp
(circles) and laser (diamonds and triangles). The error bars are determined by the combination of the confidence intervals of the data fitting and deviations
between samples. e, f Evolution of the EDCs near kF of the B band with surface Rb doping. The inset in f shows the fit of the integrated EDC after 6th (67 s)
Rb doping to the LL model. Data in panels b and c were collected using 7 eV laser, whereas the data in panels e and f were collected using He lamp
(Supplementary Figure 7).

in the inset of Fig. 3f, further proving our conclusion that the confirming the formation of Holstein polaron in the system. It is
normal state of the blue bronze is a LL. noteworthy that the band renormalization near −85 meV is
concealed by the suppression of the spectral weight near EF and is
instead manifested by a kink-like structure in our experiment and
Spectral function of Holstein polaron. To understand the simulation.
microscopic interactions underlying the LL-to-CDW transition,
we perform further spectroscopic analyses in Fig. 4. Fig. 4a shows
the ARPES spectrum of the AB band (left) and its second deri- Discussion
vative with respect to the energy (right). Interestingly, we observe Although the electronic properties are radically different in Fermi
a flat band near −150 meV and a discontinuity in the dispersion liquids and LLs, we observe the formation of Holstein polaron in
near −170 meV, which are clearly resolved in the EDCs (Fig. 4b), a LL material, similar to that in Fermi liquids11. To understand
but absent in the ab initio calculation (Supplementary Note 3) the mechanism of Holstein polaron in blue bronzes, it is crucial to
and previous ARPES measurements. For a more-detailed reveal the nature of the phonon mode involved. The phonon
inspection of the spectral function, we approximate the mode at ≈85 meV is identified as the bridge Mo-O-Mo stretching
momentum distribution curves (MDCs) of the AB band with mode as observed in Raman scattering and infrared spectroscopic
Lorentzians. Despite the strong suppression of the spectral experiments44,45. Owing to the Q1D crystal structure of blue
weight by the LL nature of the system, we can still extract the bronzes, this phonon mode is spatially localized perpendicular to
band dispersion and the spectral broadening ΔkðEÞ near EF. As the atomic chains, thus is short-range in nature. Along the atomic
shown in Fig. 4c and d, we observe a kink-like structure chains, however, it can be relatively extended. The size or the
near −85 ± 10 meV. coherence length of the Holstein polaron can be estimated by
The doubled energy positions of the band renormalizations at l = τvp, where τ and vp are the lifetime and velocity of the qua-
≈−85 meV and ≈−170 meV mimic the characteristic spectral siparticles. τ and vp can be obtained from the spectral broadening
function of Holstein polarons that are usually manifested by a and Fermi velocity of the AB band, respectively. With vp ≈ 3 eV Å
series of band renormalization at −nΩ0 (n = 1, 2,…), where Ω0 is and MDC width Δk = 0.12 Å−1 near EF, we estimate τ ≈ 11 fs and
the energy of the phonon mode involved in the formation of the thus l ≈ 8.3 Å, slightly larger than the lattice constant along the
polaron (Fig. 1c)11,13,43. To verify the formation of Holstein atomic chains.
polaron in the blue bronze, we simulate its spectral function by a Moreover, the strong band renormalizations and the formation
simple momentum-average model (see Methods for details)11,43. of Holstein polaron are absent in the B band and the AB band is
Considering the energy position of the kink-like structure and the much more broadened than the B band, suggesting an interesting
dispersion discontinuity, we adopt a phonon mode with the band-selective EPC. It is very likely that the phonon mode at
energy of ≈85 meV in the simulation44,45. Fig. 4e–g elucidate the ≈85 meV involves mainly the Mo II atoms whose orbitals con-
evolution of the band renormalizations with increasing EPC tribute significantly to the AB band (Supplementary Note 3),
parameter λ. At λ ≈ 0.6 (Fig. 4g), both the flat band and the kink- giving rise to the band-selective formation of Holstein polarons.
like structure are excellently reproduced by our simple approx- Even so, the B band is also influenced by the strong EPC in the
imation (also see the EDCs of the simulated spectrum in Fig. 4h), system. By quantitative analyses, we show that the electron self-

4 NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1 ARTICLE

Fig. 4 Holstein polarons in K0.3MoO3. a ARPES intensity map of the AB band along X 
Z  (left) and its second derivative (right) with respect to the energy
axis measured with linear-vertically polarized 7 eV laser at 75 K. The orange and green dashed curves are the guides to the eyes for the band discontinuity
(emphasized with blue and red arrows). b Stacking plot of EDCs showing observed dispersions. The green bar and orange triangles are the guides to the
eyes for dispersions. c, d Extracted AB band dispersion (c) and the spectral broadening (d) showing a kink-like feature near 85 meV below EF (blue arrow).
The red line in panel c is the guide to eyes for the bare band dispersion. The error bars are determined by the combination of the confidence intervals of the
data fitting and deviations between samples. e–g Simulation of the spectral function (left) and its second derivative (right) of Holstein polarons at different
EPC parameters. h Simulated EDCs showing good agreement with the raw data in a. The green bar and orange triangles are the guides to the eyes for
dispersions.

energy of the B band near EF strongly depends on the tempera- blue bronzes. Our work suggests that the physics in blue bronze
ture (Supplementary Note 8), further confirming the strong EPC may be properly described by the 1D Holstein model, in which
in blue bronzes22,34,36,37. Likewise, the estimated coherence EPC is a vital parameter in the phase diagram7.
length of the Holstein polaron is in good comparison with the
CDW period of ~9 Å, and the phonon mode involved in the Methods
Holstein polaron is intimately related to the CDW Sample preparation. The A0.3MoO3 single crystals were grown by an electrolytic
transition33,44,45, suggesting that the strong EPC and its band- reduction method49. A2CO3 (99.5% purity) and MoO3 (99% purity) were used as
selectivity may play an important role in the electronic properties reagent materials in standard electrolysis run with a molar ratio of 1:4.3. During a
typical run, the furnace temperature is increased to 600°C for 40 min and main-
and the LL-to-CDW transition in blue bronzes. tained for 2 h. Then the temperature was brought tardily (1°C/min) to 5–10 °C
It is worth noting that we do not observe clear evidence of above the solidification temperature of the liquid of A2CO3 and MoO3 mixture. A
spin-charge separation, although it has been suggested in some current of 25 mA was applied through the melt during the electrolytic process and
previous ARPES measurements26. Although the spin-charge maintained for 2–2.5 h. This process of crystal growth can be repeated many times
during several days, after which platelet crystals were obtained.
separation is evidenced in a few LL materials17,20,46, it is diffi-
cult to observe in the blue bronze: On the one hand, the LL phase
is only well established in blue bronzes at high temperatures ARPES. High-resolution ARPES measurements were performed at beamline 13U
of National Synchrotron Radiation Laboratory (NSRL), China, beamline 9 A of
where the strong thermal fluctuation challenges the observation Hiroshima Synchrotron Radiation Center (HSRC), Japan, and Peking University,
of spin-charge separation. On the other hand, the dispersions of China. Data were collected with Scienta R4000 (DA30) electron analyzers at NSRL
the separated spinon and chargon are expected to be well resol- and HSRC (Peking University). The overall energy and angle resolutions were set
vable only in LL materials with α much smaller than 0.519,31. For to 15 meV and 0.2°, respectively.
example, it is likewise absent in the prototypical LL material
Li0.9Mo6O17 with a similar value of α = 0.6518. Finally but also Laser-based ARPES. Laser-based ARPES measurements were performed using
importantly, the strong EPC may smear the spin-charge separa- DA30L analyzers and vacuum ultraviolet 7 eV lasers in Tsinghua University and
ShanghaiTech University, China. The overall energy and angle resolutions were set
tion in the system by strongly broadening the dispersion of the to 6 meV and 0.2°, respectively. The samples were cleaved in situ and measured
chargon and reducing its spectral weight47,48. under ultra-high vacuum below 1.0 × 10−10 mbar. Surface Rb doping was per-
In summary, using high-resolution laser-based ARPES, we formed in situ at 80 K using a SAES alkali-metal source after well outgassing. The
systematically investigate the electronic structure of blue bronze. current was set to 5.6 A.
Our results indicate a LL phase of the system in the normal state.
We observe Holstein polarons that strongly renormalize the band Single-crystal X-ray diffraction. Single-crystal XRD was performed using Mo
dispersion. We, therefore, conclude that the strong EPC has an target by the Rigaku Oxford Diffraction at the Department of Physics, University of
important role in the spectral properties and CDW transition of Oxford. The beam spot size is 10–200 μm in diameter. The data were collected and
analyzed by the CrysAlisPro software.

NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1

Transport measurement. Transport measurements were taken in the Physical 11. Kang, M. et al. Holstein polaron in a valley-degenerate two-dimensional
Properties Measurement System of Quantum Design. semiconductor. Nat. Mater. 17, 676–680 (2018).
12. Chen, C., Avila, J., Frantzeskakis, E., Levy, A. & Asensio, M. C. Observation of
First-principles calculation. Electronic structure calculations were performed a two-dimensional liquid of Fröhlich polarons at the bare SrTiO3 surface. Nat.
using density functional theory50 with the projected augmented wave method as Commun. 6, 8585 (2015).
implemented in the QUANTUM ESPRESSO package51,52. The exchange- 13. Franchini, C., Reticcioli, M., Setvin, M. & Diebold, U. Polarons in materials.
correlation functional was approximated within the Perdew–Burke–Ernzerhof Nat. Rev. Mater. 6, 560–586 (2021).
scheme53. Experimental structural parameters were relaxed with a force threshold 14. Moser, S. et al. Tunable polaronic conduction in anatase TiO2. Phys. Rev. Lett.
of 0.01 eV/Å. The cutoff energy for the plane-wave basis was set to 600 eV and the 110, 196403 (2013).
Monkhorst-Pack k-point mesh of 9 × 9 × 9 was used to get a self-consistent charge 15. Riley, J. M. et al. Crossover from lattice to plasmonic polarons of a spin-
density. The FS was calculated with a denser mesh of 16 × 16 × 16 and checked by polarised electron gas in ferromagnetic EuO. Nat. Commun. 9, 2305 (2018).
the tight-binding-based calculation supplied by the Wannier90 code54. 16. Cancellieri, C. et al. Polaronic metal state at the LaAlO3/SrTiO3 interface. Nat.
Commun. 7, 10386 (2016).
17. Kim, B. J. et al. Distinct spinon and holon dispersions in photoemission
Simulation of the spectral function of Holstein polaron. The spectral function
spectral functions from one-dimensional SrCuO2. Nat. Phys. 2, 397–401
calculation of Holstein polaron is performed within momentum-average
(2006).
approximation43, which is demonstrated to feature high efficiency and
accuracy11,43. The model has a simple form of an electron coupled with a dis- 18. Wang, F. et al. New Luttinger-liquid physics from photoemission on
persionless phonon mode, i.e., Holstein polaron. The Green’s function of the Li0.9M06O17. Phys. Rev. Lett. 96, 196403 (2006).
Holstein model is represented as: 19. Watson M. D. et al. Multiband one-dimensional electronic structure and
spectroscopic signature of tomonaga-luttinger liquid behavior in K2Cr3As3.
1 Phys. Rev. Lett. 118, 097002 (2017).
Gðk; ωÞ ¼ ð2Þ
ω
ϵk
ΣMA ðωÞ þ iη 20. Claessen, R. et al. Spectroscopic signatures of spin-charge separation in the
where ω is the energy, ∈k is the bare dispersion, η is the spectral broadening quasi-one-dimensional organic conductor TTF-TCNQ. Phys. Rev. Lett. 88,
parameter, and the momentum-averaged self-energy ΣMA ðωÞ has simple analytic 096402 (2002).
form for 1D systems: 21. Pouget, J. P., Kagoshima, S., Schlenker, C. & Marcus, J. Evidence for a Peierls
transition in the blue bronzes K0.30MoO3 and Rb0.30MoO3. J. Phys. Lett. 44,
g 2 Ω0 g 0 ðω
Ω0 Þ 113–120 (1983).
ΣMA ðωÞ ¼ 2g 2 g 0 ðω
Ω0 Þg 0 ðω
2Ω0 Þ ð3Þ
1
3g 2 g 0 ðω
2Ω0 Þg 0 ðω
3Ω0 Þ
22. Mankowsky, R. et al. Dynamical stability limit for the charge density wave in
1
1
¼ K0.3MoO3. Phys. Rev. Lett. 118, 116402 (2017).
23. Pouget, J. P., Hennion, B., Escribe-Filippini, C. & Sato, M. Neutron-scattering
sgnðωÞ
g 0 ðωÞ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð4Þ investigations of the Kohn anomaly and of the phase and amplitude charge-
ðω þ iηÞ2
4t 2 density-wave excitations of the blue bronze K0.3MoO3. Phys. Rev. B 43,
8421–8430 (1991).
where Ω0 is the energy of the phonon involved in the formation of Holstein 24. Guster, B., Pruneda, M., Ordejón, P., Canadell, E. & Pouget, J.-P. Evidence for
polarons, g 2 is related to EPC constant λ by g 2 ¼ 2tΩ0 λ and t is the hopping the weak coupling scenario of the Peierls transition in the blue bronze. Phys.
amplitude. Rev. Mat. 3, 055001 (2019).
In the spectral function simulation, the quadratic energy dependence of η is 25. Ando, H. et al. Angle-resolved photoemission study of K0.3MoO3: direct
adopted and a LL type spectral weight is included to mimic the experimental observation of temperature-dependent Fermi surface across the Peierls
spectral function of AB band. transition. J. Phys. Condens. Matter 17, 4935–4940 (2005).
26. Mou, D., Konik, R. M., Tsvelik, A. M., Zaliznyak, I. & Zhou, X. Charge-density
Data availability wave and one-dimensional electronic spectra in blue bronze: Incoherent
The data sets that support the findings of this study are available from the corresponding solitons and spin-charge separation. Phys. Rev. B 89, 201116 (2014).
author upon reasonable request. 27. Dardel, B. et al. Temperature dependence of the spectral function through the
peierls transition in quasi-one-dimensional compounds. Europhys. Lett. 19,
525–530 (1992).
Received: 4 May 2021; Accepted: 14 September 2021; 28. Grioni, M., Perfetti, L. & Berger, H. Strong electron–electron and
electron–phonon interactions in one- and two-dimensional solids. J. Electron
Spectros Relat. Phenom. 137–140, 417–423 (2004).
29. Allen, J. W. Quasi-particles and their absence in photoemission spectroscopy.
Solid State Commun. 123, 469–487 (2002).
30. Gweon, G. H. et al. ARPES line shapes in FL and non-FL quasi-low-
References dimensional inorganic metals. J. Electron Spectros. Relat. Phenom. 117-118,
1. Giamarchi T. Quantum physics in one dimension. Clarendon Press (2003). 481–502 (2001).
2. Haldane, F. D. M. ‘Luttinger liquid theory’ of one-dimensional quantum 31. Gweon, G. H., Allen, J. W. & Denlinger, J. D. Generalized spectral signatures
fluids. I. Properties of the Luttinger model and their extension to the general of electron fractionalization in quasi-one- and two-dimensional molybdenum
1D interacting spinless Fermi gas. J. Phys. C Solid State Phys. 14, 2585–2609 bronzes and superconducting cuprates. Phys. Rev. B 68, 195117 (2003).
(1981). 32. Tanaka, S., Ueda, E. & Sato, M. STM·STS studies of molybdenum oxides
3. Gruner G. Density Waves in Solids. (Perseus Publishing, 1994). conductors. Solid State Commun. 87, 877–881 (1993).
4. Hohenadler, M., Wellein, G., Bishop, A. R., Alvermann, A. & Fehske, H. 33. Perfetti, L. et al. Mobile small polarons and the Peierls transition in the quasi-
Spectral signatures of the Luttinger liquid to the charge-density-wave one-dimensional conductor K0.3MoO3. Phys. Rev. B 66, 075107 (2002).
transition. Phys. Rev. B 73, 245120 (2006). 34. Yang, L. X. et al. Bypassing the structural bottleneck in the ultrafast melting of
5. Ejima, S., Hager, G. & Fehske, H. Quantum phase transition in a 1D transport electronic order. Phys. Rev. Lett. 125, 266402 (2020).
model with Boson-affected hopping: luttinger liquid versus charge-density- 35. Orgad, D. Spectral functions for the tomonaga-luttinger and luther-emery
wave behavior. Phys. Rev. Lett. 102, 106404 (2009). liquids. Philos. Mag. B 81, 377–398 (2001).
6. Hohenadler, M., Assaad, F. F. & Fehske, H. Effect of electron-phonon 36. Degiorgi, L. et al. Fluctuation effects in quasi-one-dimensional conductors:
interaction range for a half-filled band in one dimension. Phys. Rev. Lett. 109, optical probing of thermal lattice fluctuations. Phys. Rev. B 52, 5603–5610
116407 (2012). (1995).
7. Bursill, R. J., McKenzie, R. H. & Hamer, C. J. Phase diagram of the one- 37. Beyer, R., Barišić, N. & Dressel, M. Charge-density fluctuations probed by
dimensional holstein model of spinless fermions. Phys. Rev. Lett. 80, vibronic modes of K0.3MoO3. Phys. B Condens. Matter 407, 1823–1826 (2012).
5607–5610 (1998). 38. Graham, J. & Wadsley, A. D. The crystal structure of the blue potassium
8. Ito, T. et al. Temperature-dependent luttinger surfaces. Phys. Rev. Lett. 95, molybdenum bronze, K0.28MoO3. Acta Crystallogr. 20, 93–100 (1966).
246402 (2005). 39. Brutting, W. & Nguyen, P. H. Rie beta W, Paasch G. DC-conduction
9. Nicholson, C. W. et al. Dimensional crossover in a charge density wave mechanism and Peierls gap in organic and inorganic charge-density-wave
material probed by angle-resolved photoemission spectroscopy. Phys. Rev. conductors. Phys. Rev. B 51, 9533–9543 (1995).
Lett. 118, 206401 (2017). 40. Schwartz, A. et al. Fluctuation effects on the electrodynamics of quasi-one-
10. Mou, D. et al. Discovery of an unconventional charge density wave at the dimensional conductors above the charge-density-wave transition. Phys. Rev.
surface of K0.9Mo6O17. Phys. Rev. Lett. 116, 196401 (2016). B 52, 5643–5652 (1995).

6 NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26078-1 ARTICLE

41. Wang, F. et al. Case for bulk nature of spectroscopic Luttinger liquid Author contributions
signatures observed in angle-resolved photoemission spectra of Li0.9Mo6O17. L.X.Y. and Y.L.C. conceived the experiments. L.K. and X.D. carried out ARPES mea-
Phys. Rev. B 74, 113107 (2006). surements with the assistance of J.S.Z., X.G., Y.J.C., R.Z.X., Q.Q.Z., S.C.S., Z.X.Y., Y.W.L.,
42. Chudzinski, P., Jarlborg, T. & Giamarchi, T. Luttinger-liquid theory of D.P., J.Z., R.K.G., Z.G.W., Y.Z. and Z.K.L.; L.K. and X.D. performed the data analysis on
purple bronze Li0.9Mo6O17 in the charge regime. Phys. Rev. B 86, 075147 the ARPES results; X.D. performed ab initio calculations; R.X. and J.S. synthesized the
(2012). single crystals. Y.W.L. characterized the single crystals. L.K. and X.D. wrote the first draft
43. Goodvin, G. L., Berciu, M. & Sawatzky, G. A. Green’s function of the Holstein of the paper. L.X.Y., Y.L.C. and J.S. contributed to the revision of the manuscript. All
polaron. Phys. Rev. B 74, 245104 (2006). authors contributed to the scientific planning and discussion.
44. Yin, D. et al. Polarized Raman scattering study of blue bronze Tl0.3MoO3.
Phys. B: Condens. Matter 405, 3782–3786 (2010).
45. Richard Mortimer, James G. Powell, Martha Greenblatt, McCarroll WH,
Competing interests
The authors declare no competing interests.
Ramanujachary AKV. Variable-temperature infrared spectroscopic study of
some molybdenum bronzes: evidence for electron-phonon coupling. J. Chem.
Soc. Faraday Trans. 89, 3603–3609 (1993). Additional information
46. Lorenz, T. et al. Evidence for spin–charge separation in quasi-one- Supplementary information The online version contains supplementary material
dimensional organic conductors. Nature 418, 614–617 (2002). available at https://doi.org/10.1038/s41467-021-26078-1.
47. Assaad, F. F. Spin, charge, and single-particle spectral functions of the one-
dimensional quarter filled Holstein model. Phys. Rev. B 78, 155124 (2008). Correspondence and requests for materials should be addressed to Y. L. Chen or
48. Ning, W.-Q., Zhao, H., Wu, C.-Q. & Lin, H.-Q. Phonon effects on spin-charge L. X. Yang.
separation in one dimension. Phys. Rev. Lett. 96, 156402 (2006).
49. Wang, J. et al. Growth of high quality and large-sized Rb0.3MoO3 single Peer review information Nature Communications thanks Jose Avila, and the other,
crystals by molten salt electrolysis method. J. Solid State Chem. 178, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer
1440–1444 (2005). reviewer reports are available.
50. Kohn, W. & Sham, L. J. Self-consistent equations including exchange and
correlation effects. Phys. Rev. 140, A1133–A1138 (1965). Reprints and permission information is available at http://www.nature.com/reprints
51. Giannozzi, P. et al. Advanced capabilities for materials modelling with
quantum ESPRESSO. J. Phys: Condens. Matter 29, 465901 (2017). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
52. Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source published maps and institutional affiliations.
software project for quantum simulations of materials. J. Phys: Condens.
Matter 21, 395502 (2009).
53. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation
Open Access This article is licensed under a Creative Commons Attri-
made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
bution 4.0 International License, which permits use, sharing, adaptation,
54. Pizzi, G. et al. Wannier90 as a community code: new features and applications.
distribution and reproduction in any medium or format, as long as you give appropriate
J. Phys: Condens. Matter 32, 165902 (2020).
credit to the original author(s) and the source, provide a link to the Creative Commons
license, and indicate if changes were made. The images or other third party material in
this article are included in the article’s Creative Commons license, unless indicated
otherwise in a credit line to the material. If material is not included in the article’s Creative
Acknowledgements
The authors thank the helpful discussion with P.Z. Tang. This work was supported by the Commons license and your intended use is not permitted by statutory regulation or
National Natural Science Foundation of China (Grants no. 11774190, no. 11427903, and no. exceeds the permitted use, you will need to obtain permission directly from the copyright
11634009), the National Key R&D program of China (Grants no. 2017YFA0304600, no. holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
2017YFA0305400, and no. 2017YFA0402900), and EPSRC Platform Grant (Grant no. EP/
M020517/1).
© The Author(s) 2021

NATURE COMMUNICATIONS | (2021)12:6183 | https://doi.org/10.1038/s41467-021-26078-1 | www.nature.com/naturecommunications 7