High-temperature superconductivity in monolayer Bi2Sr2CaCu2O8+δ

r
r
r
Z
E eV
g
eV g
eV
( , =
) ≡ ( , +
)/ ( , −
), which eliminates systematic errors 
related to the tunnelling setpoint associated with directly mapping the 
conductance r
g
eV
( ,
) (ref. 
26
). Figure 4a displays an example of the
conductance ratio map of monolayer Bi-2212 obtained at E = 20 meV. 
The Fourier transform of the conductance ratio map, q
Z
E eV
( , =
), shows 
clear maxima at q
i
that are fully consistent with the octet model, except 
that peaks at q
4
and q
5
are too weak to be detected (Fig. 4f). As the tun-
nelling bias V is varied, we observe that the measured q
i
disperse with 
energy E = eV, and the dispersions q
i
(E) are again consistent with those 
expected from the octet model (Extended Data Figs. 7 and 8). The dis-
persions q
i
(E) allow us to extract the energy-dependent locations of the 
octet ends of the ‘bananas’ in k space, and the obtained loci can be inter-
preted as the normal-state Fermi surface
24
. Our result, shown in Fig. 4i, 
is consistent with a cylindrical Fermi surface centred at (π,π) that is 
observed in bulk Bi-2212 and various other bulk copper oxide
superconductors
26
. Finally, we determine the superconducting gap 
dispersion 
Δ kk
( ) from q E
( )
i
. Figure 4j displays the measured supercon-
ducting gap energy of the monolayer as a function of 
θ
k
along the Fermi 
surface. The data agree with the d-wave superconducting gap disper-
sion of bulk Bi-2212 at similar doping level
23
. We therefore conclude 
that reducing the material’s dimensions from three to two does not 
fundamentally alter the superconducting gap structure.