H.
Zheng
et
al.
/
Applied
Surface
Science
405
(2017)
412–419
413
lower
temperature;
however,
to
our
best
knowledge,
no
concerning
study
has
been
reported
thus
far.
Taken
into
account
the
reaction
mechanism
of
those
solid-state
reaction
methods,
a
layer
of
MAI
is
usually
stacked
on
a
compact
PbI
2
(c-PbI
2
)
layer,
in
which
the
reaction
is
realized
through
a
long-
range
interdiffusion
between
PbI
2
and
MAI
[17,19,20].
However,
the
diffusion
is
ineffective
due
to
the
lack
of
kinetically
favorable
van
der
Waals
gap
in
the
formed
CH
3
NH
3
PbI
3
interlayer
between
MAI
and
PbI
2
[24,25].
As
a
result,
to
enhance
the
diffusion,
high
temperature
annealing
is
always
applied
in
solid-state
reaction
for
preparing
perovskite
films
[25,26].
In
short,
the
necessity
of
high-
temperature
annealing
in
the
conventional
solid-state
reaction,
lies
in
the
long
distance
for
diffusion.
Therefore,
instead
of
elevating
the
reaction
temperature
to
accelerate
the
diffusion,
shortening
the
distance
for
diffusion
is
another
choice,
while
avoiding
the
neg-
ative
effects
of
high-temperature
annealing.
That
can
be
realized
by
the
application
of
nanoporous
PbI
2
(n-PbI
2
)
films
in
the
solid-
state
reaction:
the
diffusion
distance
in
the
case
of
n-PbI
2
film
will
be
reduced
to
the
size
of
PbI
2
particles
rather
than
the
thickness
of
entire
PbI
2
layer.
Moreover,
many
reports
have
shown
that
the
n-PbI
2
will
facilitate
the
reaction
between
PbI
2
and
MAI,
as
much
larger
reaction
interface
is
present
in
a
n-PbI
2
film
than
the
con-
ventional
compact
one
[11,12,15,23,27].
In
our
previous
work,
we
developed
a
facile
way
to
prepare
n-PbI
2
films
with
the
antisolvent-solvent
extraction
method
[15].
In
this
work,
however,
we
managed
to
fabricate
a
new
type
of
n-PbI
2
with
much
worse
crystallinity
by
omitting
the
post-
annealing
process
of
PbI
2
films
after
preparation,
which
is
desirable
for
energy-saving.
Moreover,
we
found
that
the
less-crystallized
nanoporous
PbI
2
(ln-PbI
2
)
film
was
able
to
convert
into
perovskite
film
more
rapidly,
compared
with
the
well-crystallized
nanoporous
PbI
2
(wn-PbI
2
,
with
post-annealing).
It
can
be
ascribed
to
the
fact
that
good
crystallinity
of
PbI
2
will
retard
the
conversion
of
PbI
2
into
perovskite
[22,28,29].
Here,
we
apply
ln-PbI
2
films
to
realize
effective
low-
temperature
solid-state
reaction
(LT-SSR).
The
replacement
of
c-PbI
2
by
ln-PbI
2
,
leads
to
better
crystallized
and
PbI
2
-free
per-
ovskite
films,
compared
with
the
PbI
2
-remained
films
based
on
c-PbI
2
.
Resultantly,
those
ln-PbI
2
based
PSCs
deliver
both
of
much
higher
PCE
and
better
stability
in
ambient
air,
as
the
remnant
PbI
2
plays
an
important
role
in
the
deterioration
of
c-PbI
2
based
samples.
Furthermore,
the
annealing
duration
was
optimized,
which
was
found
to
be
a
critical
factor
to
obtain
efficient
PSCs.
With
the
optimal
process,
a
champion
PCE
of
13.8%
was
obtained
and
the
average
PCE
of
70
samples
reached
10.1%
with
a
satisfactory
deviation.
Addition-
ally,
compared
with
the
sample
prepared
at
75
◦
C,
non-radiative
defects
increased
significantly
when
the
sample
was
prepared
at
140
◦
C.
As
a
result,
the
champion
PCE
of
140
◦
C
decreased
to
11.8%
from
13.8%
(75
◦
C).
2.
Experimental
section
2.1.
Materials
and
reagents
MAI
was
prepared
in-house
according
to
the
procedure
reported
in
literature
[25].
PbI
2
,
anhydrous
chloroben-
zene,
Li-bis(trifluoromethanesulfonyl)imide
(Li-TFSI)
and
4-tert-butylpridine
(t-BP)
were
obtained
from
Sigma-
Aldrich.
2,2
,7,7
-Tetrakis[N,N−di(4-methoxyphenyl)amino]-9,
9
,-spirobifluorene
(Sipro-MeOTAD)
and
Tris(2-(1H−pyrazol-1-yl)
-4-tert-butylpyridine)-cobalt(III)Tris(bis(trifluoromethylsulfonyl)
imide))
(FK209
Co(III)
TFSI
salt)
were
purchased
from
Lumines-
cence
Technology
Corp.,
Taiwan.
Both
of
dimethylformamide
(DMF)
and
2-propanol
(IPA)
were
dehydrated
by
molecular
sieves
(4A)
before
use.
2.2.
Device
fabrication
Fluorine-doped
tin
oxide
(FTO;
15
sq
−1
,
Nippon
Sheet
Glass)
glass
substrates
were
patterned
by
etching
with
Zn
powder
and
diluted
HCl
solution.
The
patterned
FTO
substrates
were
cleaned
with
ultrasonication
sequentially
in
detergent
(Hellmanex
II,
2%),
water,
ethanol,
acetone
and
ethanol,
respectively.
Then
those
cleaned
FTO
substrates
were
treated
with
ultraviolet
(UV)
for
15
min
before
being
coated
with
a
compact
TiO
2
layer
by
spin-
coating
a
yellowish
solution
of
tetrabutyl
titanate
[30].
After
drying
at
70
◦
C,
they
were
annealed
at
500
◦
C
for
30
min.
Then,
the
films
were
treated
with
TiCl
4
solution
(40
mM,
70
◦
C)
for
30
min,
and
rinsed
with
deionized
water
and
ethanol
before
being
annealed
at
500
◦
C
for
another
30
min.
When
cooled
down,
the
substrates
were
treated
with
UV
for
15
min
again,
before
the
deposition
of
perovskite
layer.
Specifically,
a
layer
of
PbI
2
was
prepared
by
spin-coating
1
M
PbI
2
solution
(in
DMF
kept
at
70
◦
C)
at
3000
r.p.m,
from
which
the
c-PbI
2
films
were
obtained
by
drying
at
70
◦
C;
while
wn-PbI
2
and
ln-PbI
2
films
were
fabricated
with
additional
antisolvent-solvent
extraction
process
developed
by
our
group,
and
post-annealing
was
conducted
on
wn-
PbI
2
but
not
on
ln-PbI
2
[15].
Finally,
perovskite
films
were
obtained
by
sequentially
spin-coating
of
8
mg/ml
and
30
mg/ml
MAI
solution
(in
IPA)
onto
those
PbI
2
films
and
annealing
at
75
◦
C
or
140
◦
C
for
varied
duration.
For
annealing
at
140
◦
C,
the
duration
was
short-
ened
to
20
min
to
avoid
the
decomposition
of
perovskite
films.
For
the
two-step
dipping
method,
perovskite
films
were
fabricated
by
dipping
ln-PbI
2
films
into
MAI
solution
(10
mg/ml
in
IPA)
for
sev-
eral
minutes.
Then,
those
perovskite
films
were
washed
by
IPA
and
annealed
at
75
◦
C,
unless
stated
otherwise.
The
hole-transporting
layer
(HTL)
was
fabricated
by
spin-
coating
of
a
solution
of
Spiro-MeOTAD
(72.3
mg/ml)
at
4000
r.p.m
for
30
s
with
a
recipe
as
reported
[3],
and
was
oxidized
in
a
des-
iccator
overnight.
At
last,
a
layer
of
Ag
(∼100
nm)
was
evaporated
onto
the
top
of
HTL,
with
an
effective
area
of
0.15
cm
2
for
device.
For
preparing
perovskite
layer
and
HTL,
all
the
procedures
were
conducted
inside
an
Ar-filled
glovebox.
2.3.
Measurement
and
characterization
X-ray
diffraction
(XRD)
spectra
were
obtained
from
Rigaku
D/max
2550
V
(using
Cu
K␣
radiation,
with
a
step
size
of
0.02
◦
).
A
field
emission
scanning
electron
microscopy
(FESEM;
SU8220,
Hitachi)
was
used
to
characterize
the
morphology
of
perovskite
films
and
conduct
Energy
Dispersive
X-Ray
Spectroscopy
(EDS)
test.
A
UV/Vis/NIR
spectrophotometer
(Lambda
950,
Perkin
Elmer)
was
used
to
measure
the
absorption
spectra
of
CH
3
NH
3
PbI
3
films.
Photoluminescence
(PL)
measurement
was
conducted
on
a
spec-
trometer
(FLS
980,
Edinburg
Instruments).
The
topography
of
perovskite
films
was
measured
in
the
tapping
mode
with
an
atomic
force
microscope
(AFM;
ScanAsyst,
Veeco
Dimension
Icon).
Under
the
simulated
AM
1.5G
illumination
(100
mW/cm
2
;
Oriel
Sol3A
Class
AAA
Solar
Simulator,
Newport),
the
J-V
curves
of
PSCs
were
recorded
by
a
Keithley
2400
source
meter
through
reverse
scan
(1
V
to
−0.1
V)
with
a
step
size
of
11
mV
and
delay
time
of
20
ms.
And
the
simulated
illumination
had
been
calibrated
by
an
optical
power
meter
(Newport,
1918-R)
before
each
test.
An
incident
photon-to-electron
conversion
efficiency
(IPCE)
test
sys-
tem
(Zahner,
CIMPS-pcS3/IPCE)
was
used
to
measure
IPCE
spectra.
Electrochemical
Impedance
spectroscopy
(EIS)
were
recorded
by
an
electrochemical
workstation
(CHI660D,
CH
Instruments)
under
simulated
AM
1.5G
illumination,
in
the
frequency
range
from
1
MHz
to
1
HZ.
The
measurement
of
photovoltaic
performance
(J-V
curves
and
EIS)
was
conducted
at
room
temperature
with
a
controlled
relative
humidity
(RH)
of
20%.