results suggested that the DMePDAPbI 4 -2
structure formed, as evidenced by the char-
acteristic low-angle diffraction peak, at ~8.5°
for DMePDAPbI 4 -2, rather than ~8.7° for
DMePDAPbI 4 -1 (Fig. 3A).
We also checked the 2D structures on top of
three other common perovskite compositions
of Cs0.05FA0.95PbI 3 , (FAPbI 3 )0.95(MAPbBr 3 )0.05,
and FAPbI 3 (fig. S12). For these compositions,
the characteristics peaks at (002), (004), and
(006) matched well to DMePDAPbI 4 -2, which
were absent in the DMePDAPbI 4 -1 spectrum.
Last, the low-angle diffraction peak associated
with the 2D structure from the thin-film XRD
results were further confirmed with grazing-
incidence wide-angle x-ray scattering (GIWAXS)
measurements (fig. S13). In terms of 2D surface-
layer topology and coverage, the scanning
electron microscopy (SEM) measurements
indicated that the treatment induced for-
mation of a thin surface layer with small
apparent grain sizes (figs. S14 and S15). The
conductive-atomic force microscopy (C-AFM)
measurements show that the current of the
treated film is much more uniform and lower
than that of the control film, which is con-
sistent with the formation of a capping layer
over the 3D perovskite layer (fig. S16).
To gain more insight into how the DMeP-
DAI 2 modification affects the optoelectronic
properties in perovskite films, we conducted
steady-state photoluminescence (PL), time-
resolved photoluminescence (TRPL), and TRMC
studies on these samples. The DMePDAI 2
treatment led to enhanced PL intensity (fig.
S17), longer TRPL lifetime (fig. S18 and table
S5), and improved TRMC mobility and life-
time (fig. S19) that were consistent with the
improved surface properties ( 8 , 30 ). In addition,
the ultraviolet photoelectron spectroscopy (UPS)
measurements showed that the 2D surface
treatment improved the energetics for hole
transport from the 3D perovskite to the 2D
surface layer (fig. S20).
The impact of the DMePDAI 2 treatment on
the perovskite surface chemistry was inves-
tigated with x-ray photoelectron spectroscopy
(XPS). Normalized core levels from key elements
identified on the sample surface are included
in figs. S21 and S22. The spectral shapes of
most core levels showed minimal change be-
tween the two samples, indicating similar
bonding environments, but surface treatment
caused change in the C 1s and N 1s core levels.
We fit the core levels (Fig. 3, B to E) using
constrained fitting procedures (summarized
in tables S6 and S7). The control sample had
a N 1s region whose relative peak areas were
dominated by a C=NH 2 +(FA) peak (~401 eV)
with a small shoulder to higher binding energy
(~403 eV) that corresponded to C–NH 3 (MA).
The DMePDAI 2 treatment increased the area
of the C–NH 3 peak and also led to two ad-
ditional peaks at a lower binding energy
consistent with that of C–NH 2 (~400 eV) and
the tertiary amine in DMePDAI 2 (~398 eV).
Concomitant with these changes, redistribu-
tion occurred in the features in the C 1s spectra
comprising four main peaks that are con-
sistent with primarily C–CorC–H (~285 eV),N–CH 3 (~287 eV), HC(NH 2 ) 2 (~289 eV), and
C–O or C=O bonds (~290 eV). The surface
treatment decreased the concentration of
HC(NH 2 ) 2 bonds from FA on the surface while
simultaneously increasing the amount of N–
CH 3 and C–CorC–H bonds. In addition, XPS
revealed that surface treatment increased
the amount of halide on the surface, from
about 2.6 halide-to-lead ratio for the control
to 3.1 for the DMePDAI 2 -treated film. Col-
lectively, these results suggest that both or-
ganic and halide components of the additive
incorporated into the top surface of the treated
films. Undercoordinated lead can cause donor
defects on the surface, resulting in downward
band bending and increased recombination
centers ( 31 ), so the increase in the halide-
to-lead ratio associated with the formation
of 2D interfacial component upon surface
treatment was consistent with a less defec-
tive surface.
We investigated the impact of DMePDAI 2
surface treatment on the PV performance by
fabricating PSCs using the standard n-i-p device
architecture, glass/FTO/electron transport layer
(ETL)/perovskite/hole transport layer (HTL)/
Au, where ETL is TiO 2 or SnO 2 and HTL is
spiro-OMeTAD, with more details in the sup-
plementary materials ( 23 ). Typical cross-section
SEM images of devices are shown in fig. S23.
In Fig. 4A, we compare theJ–Vcurves of the
PSCs on the basis of triple-cation–mixed-
halide FA0.85MA0.1Cs0.05PbI2.9Br0.1without and
with DMePDAI 2 treatment under simulated
100-mW/cm^2 air mass coefficient (AM) 1.5 G74 7 JANUARY 2022•VOL 375 ISSUE 6576 science.orgSCIENCE
5 10 15 20 25 30
2 Theta (degree)DMePDAPbI 4 -1
DMePDAPbI 4 -2Intenstiy (a.u.)Control PVK
PVK/DMePDAI 27 7.5 8 8.5 9 9.5 10Intensity (a.u.)2 Theta (degree)A10987
Intensity (counts x103 )406 404 402 400 398 396
Binding energy (eV)(^12) N 1s PVK/DMePDAI 2
11
10
9
8
Intensity (counts x10 7
3 )
292 290 288 286 284 282
Binding energy (eV)
C 1s PVK/DMePDAI^2
9
8
7
6
Intensity (counts x10^5
3 )
406 404 402 400 398 396
Binding energy (eV)
6.4 N 1s Control PVK
6.0
5.6
5.2
Intensity (counts x10
3 )
292 290 288 286 284 282
Binding energy (eV)
C 1s Control PVK
BC
DE
- Fig. 3. Surface layer treatment.(A) Comparison of grazing incident XRD
(GIXRD) patterns of thin films of DMePDAPbI 4 and perovskites without
(control PVK) and with DMePDAI 2 surface treatment (PVK/DMePDAI 2 ).
(Inset) Zoom-in view of the GIXRD pattern from 7° to 10°. The peak
labeled with an asterisk is from the FTO substrate. X-ray source,
Cu Karadiation. (BtoE) Comparison of the XPS spectra of N1s and
C1s for [(B) and (C)] the control and [(D) and (E)] the DMePDAI 2 -modified
perovskite thin film.
RESEARCH | REPORTS