C60 building blocks with tuneable structures for tailored functionalities

Darius Kayley1 and Bo Peng2, ∗

1
Fitzwilliam College, University of Cambridge, University of Cambridge,
Storey’s Way, Cambridge CB3 0DG, United Kingdom
2
Theory of Condensed Matter Group, Cavendish Laboratory, University of Cambridge,
J. J. Thomson Avenue, Cambridge CB3 0HE, United Kingdom
(Dated: January 6, 2025)
We show that C60 fullerene molecules can serve as promising building blocks in the construction
of versatile crystal structures with unique symmetries using first-principles calculations. These
phases include quasi-2D layered structures and 3D van der Waals crystals where the molecules
arXiv:2501.01494v1 [cond-mat.mes-hall] 2 Jan 2025

adopt varied orientations. The interplay of molecular arrangement and lattice symmetry results in
a variety of tuneable crystal structures with distinct properties. Specifically, the electronic structures
of these phases vary significantly, offering potential for fine-tuning the band gap for electronics and
optoelectronics. Additionally, the optical properties of these materials are strongly influenced by
their crystalline symmetry and molecular alignment, providing avenues for tailoring optical responses
for photonics. Our findings highlight the potential of fullerene-based building blocks in the rational
design of functional materials.

I. INTRODUCTION cal properties of four phases with distinct crystal sym-

metries from orthorhombic to trigonal and to cubic. The
Atoms are the elemental building blocks that govern C60 molecules in these structures form various lattices
the chemical bonding and thus the structure of matter. including quasi-2D layered structures and van der Waals
The intrinsic properties of the atoms, including the elec- plastic crystals with different orientations, leading to dis-
tronic configuration, electronegativity, and bonding ten- tinct lattice dynamics, varied electronic band gaps, and
dencies, dictate the preferred packing arrangement and tuneable optical properties. Our work provides promising
crystal symmetry. As an example, an element like cop- molecular building blocks for designing tuneable struc-
per naturally adopts a face-centred cubic (fcc) structure tures with tailored functionalities.
and cannot be reconfigured to form a hexagonal close-
packed (hcp) lattice [1, 2]. This structural constraint sig-
nificantly limits the design of new functional materials II. METHODS
at the atomic level where diverse crystal structures are
desired for greater structural and functional tunability. The Vienna ab initio Simulation Package (VASP) [20,
Instead, molecular or colloidal building blocks allow di- 21] was used for first-principles calculations. The
rect control and precise tuning over crystal structures and projector-augmented wave (PAW) basis set [22, 23] was
functionalities [3, 4]. However, the complex experimental used with the Perdew-Burke-Ernzerhof (PBE) func-
synthesis and assembly of these building blocks struggle tional [24] under the generalized gradient approximation
to provide clear answers to the rational design of stable (GGA). The zero damping DFT-D3 method of Grimme
structures with tailored functionalities. was included to describe the van der Waals interac-
Among all the molecular building blocks, the cage-like, tions [25]. A plane-wave cutoff energy of 800 eV was used
hollow C60 fullerene with nearly spherical symmetry [5] with the Brillouin zone sampled using a well-converged
provides an atom-like, stable unit, which can act as a ver- Monkhorst-pack k-point grid of 3 × 3 × 3. Both the
satile platform for constructing functional materials [6? lattice parameters and the atomic positions were fully
? –12]. With inherent symmetry, high stability, and re- relaxed until the difference between the electronic and
markable electronic properties, these fullerene building ionic steps reached 10−6 eV and 10−2 eV/Å respectively.
blocks can form various lattice structures with diverse The vibrational properties were calculated under the har-
symmetries, enabling precise control over the physical monic approximation using PHONOPY [26, 27] with in-
and chemical properties [13–18]. This adaptability opens teratomic force constants computed from a finite dif-
new avenues for material design beyond atoms, offering ference method [28]. A supercell size of 2 × 2 × 2 was
unprecedented opportunities to explore the interplay be- used for all three phases with one C60 molecule in the
tween symmetry and functionality. primitive unit cell, while a 1 × 1 × 1 cell of (14.181 Å)3
In this work, we use C60 fullerene as an “ultimate was used for the P a3̄ phase with four C60 molecules
building block” for designing functional materials. We in the primitive unit cell. Unscreened hybrid func-
explore the structural, vibrational, electronic, and opti- tional PBE0 [29] calculations were conducted for elec-
tronic structures. This approach has predicted compa-
rable band gaps for monolayer fullerene networks [14, 17]
to those from computationally-heavy many-body pertur-
∗ bp432@cam.ac.uk bation theory [30]. Excitonic effects were included us-
 2

FIG. 1. 3D and top views of the crystal structures of (a) monolayer and (b) bilayer fullerene networks visualised by VESTA [19].

TABLE I. Symmetry information, lattice constants with conventional unit cell, volume per C60 , and cohesive energy Ec for
different phases of solid C60 .
Crystal system Space group No. a (Å) b (Å) c (Å) V (Å3 ) Ec (eV/atom)
Orthorhombic Immm 71 9.038 9.137 15.289 631.309 -7.66107
Trigonal R3̄m 166 9.190 9.190 24.796 604.489 -7.65026
Cubic F m3̄ 202 14.232 14.232 14.232 720.657 -7.66258
Cubic P a3̄ 205 14.181 14.181 14.181 712.950 -7.66431

ing the time-dependent Hartree-Fock (TDHF) method Within each layer, the C60 cages along a and b directions
based on the Casida equation [31] from a basis of 16 are connected by [2 + 2] cycloaddition bonds, forming a
highest valence bands and 16 lowest conduction bands 2D quasi-square lattice. Along the c direction, the lay-
with a k-mesh of 8 × 8 × 8 (4 × 4 × 4) for the phases ers closely stack in a space-efficient manner. The trigo-
with one (four) C60 molecule(s) in the primitive unit nal phase in Fig. 1(b) also has a van der Waals layered
cell. The PBE0 + TDHF approach replaces the screened structure with space group R3̄m (No. 166). In the 2D
exchange W in the Bethe-Salpeter equation (BSE) with plane, the C60 cages form a quasi-triangular lattice linked
the exchange-correlation kernel to describe the screen- through [2 + 2] cycloaddition bonds with slightly larger
ing of the Coulomb potential. This has yielded exciton lattice constants a and b (Table I). The R3̄m phase is
binding energies for C60 systems [14, 17] quantitatively in a closely-packed rhombohedral lattice with the small-
agreeable with GW + BSE results [30] but, crucially, is est volume among all four phases, similar to an hcp lat-
computationally more feasible. tice. However, it has the highest cohesive energy. These
two layered structures can be obtained experimentally
in the pressure range of 2 − 8 GPa at 500 − 900 K as
III. RESULTS AND DISCUSSION mixtures, with higher pressures yielding more R3̄m crys-
tals [8, 12, 32].
A. Crystal structures Beyond the layered structures, there are two fcc phases
in Fig. 1(c) and (d) with no intermolecular bonds. The
Figure 1 shows the crystal structures of solid C60 in dif- molecules in these structures, which have much higher
ferent phases. There are two layered structures. The or- volume, are held together purely by van der Waals inter-
thorhombic phase in Fig. 1(a) crystallises in space group actions, with lower cohesive energies than the two lay-
Immm (No. 71) with a van der Waals layered structure. ered phases, as shown in Table I. Considering an isolated
 3

FIG. 2. Low-frequency phonon spectra for solid C60 with space groups (a) Immm, (b) R3̄m, (c) F m3̄, and (d) P a3̄.

fullerene molecule, the icosahedral symmetry leads to a as a spherical shell with the same F m3̄m symmetry.
point group of Ih (m3̄5̄), which, after including transla- This crystal symmetry can be found in superconducting
tional symmetry that is incompatible with five-fold ro- K3 C60 [6, 37]. For the P a3̄ phases in Fig. 2(d), there are
tational symmetry, has a maximal subgroup of m3̄ crys- twenty four low-frequency modes as there are four C60
talline symmetry. The F m3̄ (No. 202) phase in Fig. 1(c) molecules in the primitive unit cell, leading to even more
represents the simplest crystalline structure for which complex rotational modes with imaginary frequencies.
all four icosahedra have the same orientation. For the Experimentally, a phase transition to the P a3̄ phase at
P a3̄ (No. 205) phase in Fig. 1(d), the four sites are in- 255 K has been reported [7]. Therefore, it is expected that
equivalent with different orientations, further reducing the imaginary phonons in this orientationally-ordered
the symmetry. As a result of the extra rotational degree phase will disappear after including anharmonic effects
of freedom in the C60 molecules, the F m3̄ phase is more at increased temperatures.
symmetric than the P a3̄ phase, with one C60 molecule in
the primitive unit cell, leading to a typical fcc Brillouin
zone (BZ), as shown in Fig. 1(g). On the other hand,
the P a3̄ phase has four inequivalent C60 molecules in the
primitive unit cell with a cubic BZ, as shown in Fig. 1(i). C. Electronic structures
The F m3̄ phase has a slightly larger lattice constant and
a higher cohesive energy than the P a3̄ phase.
We then compare the electronic structures of the four
phases. As shown in Fig. 3(a), the Immm fullerite has
B. Lattice dynamics an indirect band gap of 2.075 eV. The valence band max-
imum (VBM) is located at the T high-symmetry point,
We next examine dynamic stability of the four phases and the conduction band minimum (CBM) is located at
by comparing the phonon spectra. For the two layered the R high-symmetry point [the exact positions of these
structures in Fig. 2(a) and (b), the phonons exhibit no high-symmetry points are shown in Fig. 1(e)]. The R3̄m
imaginary modes, indicating that both structures are in a fullerite in Fig. 3(b) has an indirect band gap of 2.027 eV
local minimum of the potential energy surface [33–36]. In with the VBM at T and the CBM at L. The PBE0 band
contrast, both the fcc van der Waals phases are dynam- gaps of the two layered structures are comparable to each
ically unstable with imaginary modes across the entire other.
BZ. In the two fcc van der Waals structures, the band gaps
For the F m3̄ phase in Fig. 2(c), there are six acoustic- are much larger than those of the layered structures be-
like modes, and the three imaginary modes correspond to cause of the weaker screening effects with increasing in-
the rotational motion of molecules. The rotational dis- termolecular distance. The F m3̄ fullerite in Fig. 3(c) has
order corresponds to two possible orientations, leading a direct band gap of 2.293 eV at X and an energy gap
to an “average” molecule with 120 half-occupied atom of 2.588 eV at L, agreeing well with previous GW cal-
sites, thus raising the symmetry to F m3̄m [7]. The ori- culations and photoemission spectra [39–41]. The P a3̄
entational disorder can be further increased by assuming fullerite in Fig. 3(d) has a direct band gap of 2.461 eV at
completely random orientations, which can be viewed R, agreeing well with previous calculations [42].
 4

FIG. 3. Electronic structures for solid C60 with space group (a) Immm, (b) R3̄m, (c) F m3̄, and (d) P a3̄.

FIG. 4. Optical absorption of solid C60 with space group (a) Immm, (b) R3̄m, (c) F m3̄, and (d) P a3̄ without (denoted
as PBE0) and with (denoted as TDHF) the excitonic effects. The extraterrestrial solar irradiance [38] is also shown as the
background colour map.

photoemission studies [43]. The peaks for polarisation

TABLE II. PBE and PBE0 band gaps (eV) for different
along the in-plane directions are mainly above the band
phases of solid C60 .
gap, and have stronger absorption. For the R3̄m phase,
Space group PBE Eg PBE0 Eg the in-plane polarised light has a strong absorption peak
Immm 0.845 (indirect) 2.075 (indirect) slightly below 1.9 eV, as shown in Fig. 4(b). For the van
R3̄m 0.742 (indirect) 2.027 (indirect) der Waals fcc F m3̄ and P a3̄ phases, the strongest ab-
F m3̄ 1.116 (direct) 2.293 (direct) sorption peaks are above the band gap, almost reaching
1.5 × 105 cm−1 .
P a3̄ 1.273 (direct) 2.461 (direct)
We then compare the low-energy excitons by listing the
first bright and first dark states. As shown in Table III,
all phases possess a high exciton binding energy above
D. Optical Absorption 0.206 eV, which is expected in molecular systems with
weaker screening effects and thus stronger Coulomb inter-
Figure 4 summarises the optical absorption of all four actions between electron-hole pairs. The R3̄m phase has
phases in the independent-parcticle approximation and in the lowest bright exciton eigenvalue of 1.871 eV, which
the presence of excitonic effects. For the Immm phase is even lower than that of the first dark excitons. This
in Fig. 4(a), the inclusion of excitons leads to a strong leads to strong exciton absorption below the band gap.
optical absorption peak around 1.9 eV, below the band In contrast to the R3̄m fullerite, the eigenvalue of the
gap for polarisation along z, agreeing well with previous first dark exciton in the other three phases is lower than
 5

IV. CONCLUSIONS
TABLE III. Eigenvalue Ex (eV) and binding energy Eb (eV)
of the first bright and first dark excitons for different phases
of solid C60 . In conclusion, we demonstrate the versatility of
fullerene building blocks for engineering materials with
Space group Exbright Ebbright Exdark Ebdark
a diverse range of crystal symmetries from quasi-2D lay-
Immm 1.912 0.293 1.808 0.374 ered arrangements to 3D van der Waals plastic crystals.
R3̄m 1.871 0.275 1.900 0.347 The chemical bonding, intermolecular distance, and ori-
F m3̄ 2.158 0.206 1.987 0.304 entations of C60 molecules play a crucial role in deter-
P a3̄ 2.223 0.295 2.050 0.413 mining the lattice dynamics, electronic structures, and
optical properties of these materials. By leveraging the
structural tuneability of these phases, we open pathways
that of the first bright exciton. for designing materials tailored to specific applications in
electronics, photonics, and energy systems. Our results
show the potential of fullerene-based building blocks as
a highly tuneable platform for exploring new physical
E. Discussion and chemical phenomena, and developing advanced func-
tional materials.
The diverse fullerite structural phases with distinct
symmetries unlock various applications. One notable ap-
plication lies in tuning the electronic and optical band CONFLICTS OF INTEREST
gaps, which is critical for optimising performance in semi-
conductors and photovoltaic devices. By manipulating There are no conflicts to declare.
the crystalline symmetry and the interfullerene distance,
electronic properties such as effective masses at band
edges can be tuned to enhance carrier mobility and en- ACKNOWLEDGEMENTS
ergy conversion efficiency. Additionally, tailored optical
properties can be obtained by varying the space groups B.P. acknowledges support from Magdalene College
that govern the optical selection rules. As an example, Cambridge for a Nevile Research Fellowship. The calcu-
for the two layered systems with anisotropic optical prop- lations were performed using resources provided by the
erties, the orthorhombic Immm phase has stronger out- Cambridge Service for Data Driven Discovery (CSD3)
of-plane exciton absorption, whereas the rhombohedral operated by the University of Cambridge Research Com-
R3̄m structure exhibits enhanced in-plane absorption in puting Service (www.csd3.cam.ac.uk), provided by Dell
the entire visible spectrum. This enables their applica- EMC and Intel using Tier-2 funding from the Engi-
tions in the design of polarisation filters and in the di- neering and Physical Sciences Research Council (capi-
rectional control of light emission and photodetection. tal grant EP/T022159/1), and DiRAC funding from the
These molecular crystals also offer a promising platform Science and Technology Facilities Council (http://www.
for energy storage such as batteries and supercapacitors dirac.ac.uk ), as well as with computational support
because of their large surface area, high chemical stabil- from the UK Materials and Molecular Modelling Hub,
ity, and environmental friendliness. The ability to design which is partially funded by EPSRC (EP/T022213/1,
versatile C60 lattices opens a new route towards the de- EP/W032260/1 and EP/P020194/1), for which access
velopment of functional materials for the future, under- was obtained via the UKCP consortium and funded by
scoring their transformative potential. EPSRC grant ref EP/P022561/1.

[1] F. Jona, X. Z. Ji, and P. M. Marcus, Phys. Rev. B 68, Nature 351, 632 (1991).
172101 (2003). [7] P. A. Heiney, Journal of Physics and Chemistry of Solids
[2] G. Grimvall, B. Magyari-Köpe, V. Ozoliņš, and K. A. 53, 1333 (1992).
Persson, Rev. Mod. Phys. 84, 945 (2012). [8] M. Núñez-Regueiro, L. Marques, J. L. Hodeau,
[3] E. A. Doud, A. Voevodin, T. J. Hochuli, A. M. Champ- O. Béthoux, and M. Perroux, Phys. Rev. Lett. 74, 278
saur, C. Nuckolls, and X. Roy, Nature Reviews Materials (1995).
5, 371 (2020). [9] C. H. Xu and G. E. Scuseria, Phys. Rev. Lett. 74, 274
[4] T. Hueckel, G. M. Hocky, and S. Sacanna, Nature Re- (1995).
views Materials 6, 1053 (2021). [10] M. Springborg, Phys. Rev. B 52, 2935 (1995).
[5] H. W. Kroto, J. R. Heath, S. C. O’Brien, R. F. Curl, [11] L. Marques, J. L. Hodeau, M. Núñez Regueiro, and
and R. E. Smalley, Nature 318, 162 (1985). M. Perroux, Phys. Rev. B 54, R12633 (1996).
[6] P. W. Stephens, L. Mihaly, P. L. Lee, R. L. Whetten, [12] M. Álvarez Murga and J. Hodeau, Carbon 82, 381 (2015).
S.-M. Huang, R. Kaner, F. Deiderich, and K. Holczer,
 6

[13] L. Hou, X. Cui, B. Guan, S. Wang, R. Li, Y. Liu, D. Zhu, [30] A. Champagne, M. Camarasa-Gómez, F. Ricci, L. Kro-
and J. Zheng, Nature 606, 507 (2022). nik, and J. B. Neaton, Nano Lett. 24, 7033 (2024).
[14] B. Peng, J. Am. Chem. Soc. 144, 19921 (2022). [31] T. Sander and G. Kresse, J. Chem. Phys. 146, 064110
[15] B. Peng, Nano Lett. 23, 652 (2023). (2017).
[16] E. Meirzadeh, A. M. Evans, M. Rezaee, M. Milich, C. J. [32] Y. Iwasa, T. Arima, R. M. Fleming, T. Siegrist, O. Zhou,
Dionne, T. P. Darlington, S. T. Bao, A. K. Bartholomew, R. C. Haddon, L. J. Rothberg, K. B. Lyons, H. L. Carter,
T. Handa, D. J. Rizzo, R. A. Wiscons, M. Reza, A. Zan- A. F. Hebard, R. Tycko, G. Dabbagh, J. J. Krajewski,
giabadi, N. Fardian-Melamed, A. C. Crowther, P. J. G. A. Thomas, and T. Yagi, Science 264, 1570 (1994).
Schuck, D. N. Basov, X. Zhu, A. Giri, P. E. Hopkins, [33] B. Peng, H. Zhang, H. Shao, Z. Ning, Y. Xu, G. Ni,
P. Kim, M. L. Steigerwald, J. Yang, C. Nuckolls, and H. Lu, D. W. Zhang, and H. Zhu, Materials Research
X. Roy, Nature 613, 71 (2023). Letters 5, 399 (2017).
[17] C. Jones and B. Peng, J. Phys. Chem. Lett. 14, 11768 [34] O. I. Malyi, K. V. Sopiha, and C. Persson, ACS Appl.
(2023). Mater. Interfaces 11, 24876 (2019).
[18] J. Wu and B. Peng, J. Am. Chem. Soc. , (2024). [35] D. Luo, K. Yin, and R. Dronskowski, J. Am. Chem. Soc.
[19] K. Momma and F. Izumi, Journal of Applied Crystallog- 144, 5155 (2022).
raphy 44, 1272 (2011). [36] I. Pallikara, P. Kayastha, J. M. Skelton, and L. D. Whal-
[20] G. Kresse and J. Furthmüller, Phys. Rev. B 54, 11169 ley, Electron. Struct. 4, 033002 (2022).
(1996). [37] M. Mitrano, A. Cantaluppi, D. Nicoletti, S. Kaiser,
[21] G. Kresse and J. Furthmüller, Computational Materials A. Perucchi, S. Lupi, P. Di Pietro, D. Pontiroli, M. Riccò,
Science 6, 15 (1996). S. R. Clark, D. Jaksch, and A. Cavalleri, Nature 530,
[22] P. E. Blöchl, Phys. Rev. B 50, 17953 (1994). 461 (2016).
[23] G. Kresse and D. Joubert, Phys. Rev. B 59, 1758 (1999). [38] C. Gueymard, D. Myers, and K. Emery, Solar Energy
[24] J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. 73, 443 (2002).
Lett. 77, 3865 (1996). [39] E. L. Shirley and S. G. Louie, Phys. Rev. Lett. 71, 133
[25] S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, J. Chem. (1993).
Phys. 132, 154104 (2010). [40] R. Schwedhelm, L. Kipp, A. Dallmeyer, and M. Ski-
[26] A. Togo, F. Oba, and I. Tanaka, Phys. Rev. B 78, 134106 bowski, Phys. Rev. B 58, 13176 (1998).
(2008). [41] T. L. Makarova, Semiconductors 35, 243 (2001).
[27] A. Togo and I. Tanaka, Scripta Materialia 108, 1 (2015). [42] E. L. Shirley, L. X. Benedict, and S. G. Louie, Phys.
[28] Y. Le Page and P. Saxe, Phys. Rev. B 65, 104104 (2002). Rev. B 54, 10970 (1996).
[29] C. Adamo and V. Barone, J. Chem. Phys. 110, 6158 [43] J. H. Weaver, J. L. Martins, T. Komeda, Y. Chen, T. R.
(1999). Ohno, G. H. Kroll, N. Troullier, R. E. Haufler, and R. E.
Smalley, Phys. Rev. Lett. 66, 1741 (1991).