Thermally-activated recombination in one component of (CH3NH3)PbI3/TiO2 observed by photocurrent spectroscopy

Article abstract of DOI:10.1039/c5cc00751h

Photocurrent measurements on devices containing perovskite (CH₃NH₃)PbI₃ show two distinct spectral responses when deposited in a mesoporous oxide matrix, compared with one response for planar perovskite alone. With a TiO₂ matrix, the shorter wavelength response has an inverted temperature response with increasing performance on cooling.

Full text of DOI:10.1039/c5cc00751h

ChemComm
View Article Online
View Journal
COMMUNICATION
Thermally-activated recombination in one
component of (CH NH )PbI /TiO observed
3
3
3
2
Cite this:DOI: 10.1039/c5cc00751h
by photocurrent spectroscopy†
Received 26th January 2015,
Accepted 22nd March 2015
a
a
b
b
Patrick Cottingham, David C. Wallace, Ke Hu, Gerald Meyer and
acd
Tyrel M. McQueen*
DOI: 10.1039/c5cc00751h
www.rsc.org/chemcomm
Photocurrent measurements on devices containing perovskite the oxide is well above that of perovskite, achieve similar or
8
(
CH
3 3 3
2
NH )PbI show two distinct spectral responses when deposited better performance than devices on TiO . Given the performance
in a mesoporous oxide matrix, compared with one response for planar of Al O -containing devices, the role of the oxide matrix in charge
2
3
perovskite alone. With a TiO matrix, the shorter wavelength response separation remains unclear.
2
has an inverted temperature response with increasing performance on
cooling.
A recent study of the structure of (CH
3 3 3 2
NH )PbI in a TiO
matrix shows that the matrix has a profound effect on the
14
structure of perovskite. PDF (pair distribution function) analysis
Organometallic halide perovskites have been the subject of of X-ray total scattering data suggests that when (CH NH )PbI is
3
3
3
intense research in recent years. Interest in these materials deposited into mesoporous TiO it forms two components with
2
stems in part from their ability to function as efficient light the same crystal structure: a relatively disordered component
absorbers and electron- and hole-transport materials in photovoltaic consisting of B1.4 nm diameter nanocrystals which are confined
(PV) devices. These materials are especially attractive for PV by pores of the matrix and one with much longer-range coherence.
applications because of their earth-abundant components and More recent work suggests that the catalytic action of water may be
1
–4
15
the wide variety of low-energy synthetic techniques by which essential for the formation of the perovskite phase.
they can be produced. Early studies of PV devices containing In order to explore the mechanisms of charge separation in
organometallic halide perovskites used perovskite (CH NH )PbI planar devices versus devices based on TiO or highly-insulating
Later studies oxides, we report temperature-dependent photocurrent mea-
demonstrated that efficient devices could be constructed using surements on a range of devices. The planar perovskite (Fig. 1a)
CH NH )PbI and larger-bandgap oxides such as Al O and ZrO and perovskite on mesoporous TiO or ZrO (Fig. 1b) devices
3
3
3
2
5
–7
deposited into a mesoporous TiO2 matrix.
(
3
3
3
2
3
2
2
2
6
in which the energy of the conduction band minimum is too were prepared following published procedures. X-ray diffraction was
high for electron injection from (CH
8
–11
3
NH
3
)PbI
3
to occur.
used to check film purity (see Fig. S1, ESI†). Briefly, 40.3 weight%
Perovskite PV cells not containing any oxide matrix (‘planar’ perovskite precursor solution was prepared with an equimolar ratio
1
,2,9,12
devices) have also produced high efficiencies.
It has been suggested that in addition to providing a driving FTO (fluorine-doped tin oxide) substrate or onto mesoporous TiO
force for charge separation at the (CH NH )PbI /TiO interface, TiO ZrO on an FTO substrate. (see ESI† for details). The substrates were
of CH
3
NH
3
I and PbI
2
in g-butyrolactone and spin-coated onto an
2
or
3
3
3
2
2
2
enhances the performance of devices by acting as an electron- mechanically compressed against a nickel foil electrode using a
13
transport material. However, devices with (CH NH )PbI deposited spring-loaded clamp.
3
3
3
2 3
in highly-insulating Al O , where the conduction band minimum of
a
b
c
Department of Chemistry, Johns Hopkins University, 3400 N. Charles St.,
Baltimore 21218, USA. E-mail: mcqueen@jhu.edu
Department of Chemistry, University of North Carolina at Chapel Hill,
1
23 South Rd., Chapel Hill, NC 27599, USA
Department of Physics and Astronomy, Johns Hopkins University,
400 N. Charles St., Baltimore 21218, USA
3
Fig. 1 Schematic illustration of two devices containing (CH
devices contain an absorber layer compressed between one FTO electrode
and one nickel electrode. The device in (a) contains a planar (CH NH )PbI
absorber layer. The device in (b) contains (CH NH )PbI embedded in a
mesoporous MO matrix, where M is Ti or Zr.
3 3 3
NH )PbI . All
d
Department of Materials Science and Engineering, Johns Hopkins University,
400 N. Charles St., Baltimore 21218, USA
Electronic supplementary information (ESI) available. See DOI: 10.1039/
3
3
3
3
†
3
3
3
c5cc00751h
2
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
1
4
Choi et al.; this is also consistent with our observation that
planar (CH NH )PbI devices with a compact, but not meso-
3
3
3
porous, TiO layer also only show a single response (Fig. S2,
2
ESI†). The peak-like shape of the longer-wavelength response is
reminiscent of a quantum dot, suggesting that it derives from the
short-range ordered component. However, quantum confinement
effects blue-shift the absorption of nanocrystals, suggesting that the
photocurrent at shorter wavelengths derives from the short-range
ordered component. If the shorter-wavelength response does corre-
19
spond to the short-range ordered component, the Brus equation
for the energy of transitions in a semiconductor cluster (eqn (1))
ꢀ
ꢁ
h2
1
1
DEðrÞ ¼ Egap þ
þ
(1)
r2 me
ꢀ
mh
ꢀ
8
can be used to calculate the effective mass of a free carrier in
this component by assuming that the electron and hole effec-
tive masses, m * and m *, are equal. DE(r) is the energy of the
e h
transition in the cluster, Egap is the bandgap of the bulk
semiconductor, r is the radius of a cluster, and h is the Planck
constant. The effective mass calculated using this equation is
m* = 15m , which differs substantially from the values of m * =
0
e
0
.23m
0
and m
h
* = 0.29m
0
determined from density functional
Fig. 2 (a), (b), and (c) Respectively show the photocurrent of devices containing
20
theory (DFT) calculations. This discrepancy suggests that the
difference in the photocurrent edge between the two compo-
nents derives from differences in structure, composition, or
disorder or from space-charge effects rather than quantum
confinement.
3 3 3 3 3 3 2 3 3 3 2
planar (CH NH )PbI , (CH NH )PbI in TiO , and (CH NH )PbI in ZrO as a
function of temperature. (d), (e), and (f) Show the photocurrent of those devices
at l = 500 nm, and l = 760 nm as a function of temperature. All lines are guides
to the eye. Noise at T = 280 K is due to failure of the light bulb.
As it is not possible to definitively assign the two features in
Photocurrent measurements were conducted by mounting the photocurrent spectrum to either the short-range ordered or
the samples on a bespoke probe and illuminating the samples the medium-range ordered component of (CH NH )PbI , they
using a monochromated lamp. All data is normalized per Watt will be referred to below as component 1 (giving rise to the
of incident light (see ESI† for details). feature at shorter wavelengths) and component 2 (giving rise to
Fig. 2a shows the photocurrent of the device containing the feature centered at l = 760 nm).
planar (CH NH )PbI . The photocurrent is approximately constant Plotting the photocurrent at l = 500 nm and l = 760 nm as a
3
3
3
3
3
3
with wavelength from l = 400 nm until the photocurrent onset at function of temperature shows profound differences between
l = 800 nm. This spectral shape is commonly associated with the different devices. For planar (CH NH )PbI (Fig. 2d), the
3
3
3
photoconductivity of a direct-bandgap semiconductor. In devices of responses at both wavelengths vary almost linearly with temperature
this kind, containing neither a strong interfacial electric field nor a and have approximately the same positive slope. For (CH NH )PbI
dedicated electron- or hole-transport layer, charge separation may in TiO , however, the response at l = 500 nm decreases with
occur by diffusion. As the temperature of the sample is decreased, increasing temperature while the response at l = 760 nm increases
the photocurrent decreases at all wavelengths above the photo- with increasing temperature (Fig. 2e). For (CH NH )PbI in ZrO , the
current edge by an approximately constant factor. two response at both wavelengths both increase with increasing
Fig. 2b shows the photocurrent of a device with mesoporous temperature (Fig. 2f).
TiO . The photocurrent spectrum for this device contains two The short-wavelength photocurrent efficiency decreases with
distinct responses. The first response is centered at shorter increasing temperature for the TiO -containing devices and
3
3
3
2
16
3
3
3
2
2
2
wavelengths and shows a gradual decrease in intensity on scanning increases for all other samples; this implies that TiO
2
is respon-
to longer wavelengths. The second response has a peak-like sible for this behaviour.
shape and is centered at around l = 760 nm. This kind of
distinct upturn near the band-edge has been previously observed containing devices can be described well by an expression for
in devices containing (CH NH )PbI and oxide when the data is thermally activated recombination containing a constant term
reported in terms of external quantum efficiency.
photocurrent of (CH NH )PbI in mesoporous ZrO2 (Fig. 2c) and an Arrhenius-type term which accounts for recombination
2
The changes in photocurrent at l = 500 nm for the TiO -
3
3
3
2
,17,18
The which represents the photocurrent in the absence of recombination
3
3
3
shows two similarly shaped responses.
A plausible explanation for the appearance of two responses in incident light intensity, E
the photocurrent when (CH NH )PbI is deposited on a mesoporous C is a constant which represents the photocurrent in the absence of
matrix is the formation of two components as observed by recombination, and A is a free scaling parameter. The sign of A is
(eqn (2)). Here Di is the measured photocurrent normalized for the
a
is the energy barrier for recombination,
3
3
3
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
involving TiO
2
, perhaps involving percolation through a net-
work of nano-scale domains. It is plausible for such networks
1
4
to exist given Choi et al.’s observation
that B70% of
(
CH NH )PbI on TiO2 consists of the short-range ordered
3
3
3
component.
The thermally-activated recombination of carriers from
component 1 of (CH NH )PbI in TiO likely consists of a
combination of back-electron transfer and/or recombination
related to transport through TiO . Experimentally determined
values for the offset between the conduction band minima of
3
3
3
2
2
7
21
(
CH NH )PbI and TiO range from 0.07 eV to 0.4 eV. DFT
3 3 3 2
including spin–orbit coupling has been used to calculate an
2
2
offset of 0.2 eV. With an energy offset in this range electron
transfer from TiO to (CH NH )PbI is possible. This is in
contrast to ZrO where the conduction band minimum is more
than 1 eV above that of TiO
is unlikely. Studies of transport in TiO
solar cells (DSSCs) have also found activation energies for
2
3
3
3
2
2
2
3
and electron transfer to the oxide
-based dye-sensitized
2
2
4
electron mobility of E = 0.15 eV. The thermal activation of
a
recombination may involve transport to recombination sites in
TiO or to an interface where recombination or back-electron
2
transfer occurs.
It is unclear from the data above why charge injection to
TiO
2
does not occur from component 2. One explanation is that
, likely by
this component is physically separated from TiO
2
component 1. Alternately, it is possible that the conduction
band minimum of component 2 is below that of TiO2 and
carriers diffuse away from the (CH NH )PbI /TiO interface
3
3
3
2
before hot-carrier or thermally-activated injection can occur.
One aspect of the photocurrent data which is not fully
explained by the model in Fig. 3b is the substantially larger
photocurrent at all wavelengths of the device containing
Fig. 3 (a) Shows a fit to the change in photocurrent as a function of
temperature for the device containing (CH
00 nm. (b) Shows a model of charge recombination in component 1 of
CH NH )PbI in TiO , compared with planar (CH NH )PbI , (CH NH )PbI
in ZrO , and component 2 of (CH NH )PbI in TiO
3 3 3 2
NH )PbI in TiO at l =
5
(
3
3
3
2
3
3
3
3
3
3
(
CH
ences in sample thickness. Scanning electron microscopy
SEM) imaging shown in Fig. 4 indicates that the TiO2-
opposite to that of Di and C because the Arrhenius term represents containing samples were B2.5 times thicker than ZrO
recombination which reduces the overall photocurrent.
containing samples. This difference is approximately the same
3 3 3 2
NH )PbI on ZrO . The difference may result from differ-
2
3
3
3
2
.
(
-
2
as the difference in photocurrent between the samples at longer
ꢁ
Ea
kBT
Di ¼ C þ Ae
(2)
wavelengths where injection into TiO is not a factor. Similar
2
decreases in incident photon conversion efficiency (IPCE) have
If the value of C = 0.06(1) A/W determined from fitting is been observed with increasing (CH
3 3 3 2
NH )PbI /TiO layer thick-
subtracted from Di, the Arrhenius term can be isolated and ness for complete PV cells and has been attributed to increased
2
5
shown in a typical ln(k) vs. 1/T -type plot (Fig. 3a). The energy recombination.
barrier determined from the fit is E = 0.17(5) eV. The results of The observation of two distinct responses in the photocur-
the fit can be used to construct a model of charge recombina- rent spectra of devices containing (CH NH )PbI on an oxide
a
3
3
3
tion for the three devices (Fig. 3b). For all of the devices, free matrix, compared to one response for planar (CH
charge carriers are generated when the absorption of a photon supports previous crystallographic observations of two distinct
generates an electron–hole pair. For planar (CH NH )PbI components in (CH NH )PbI on TiO . The inverted tempera-
component 2 of (CH NH )PbI in TiO , and both components of ture dependence of the photocurrent from one component of
CH NH )PbI in ZrO , recombination proceeds through a pathway (CH NH )PbI in TiO indicates that thermally activated recom-
3 3 3
NH )PbI ,
3
3
3
,
3
3
3
2
3
3
3
2
(
3
3
3
2
3
3
3
2
in which thermal activation is not the rate-limiting step. In contrast, bination occurs when charge injection and transport occur
for component 1 of (CH NH )PbI in TiO , recombination occurs through TiO . The other component of (CH NH )PbI in TiO
2
3
3
3
2
2
3
3
3
with an effective energy barrier of E = 0.17(5) eV.
appears to separate and transport charge through a mechanism
The absence of thermally-activated recombination in com- not involving TiO . Fitting the activation energy of recombina-
ponent 2 of (CH NH )PbI in TiO suggests that charge collec- tion with an Arrhenius term gives an energy barrier for recom-
a
2
3
3
3
2
a
tion from this component proceeds through a pathway not bination of E = 0.17(5) eV.
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
Notes and references
1
2
3
4
M. H. Kumar, N. Yantara, S. Dharani, M. Gr ¨a tzel, S. Mhaisalkar,
P. P. Boix and N. Mathews, Chem. Commun., 2013, 49, 11089.
J. You, Z. Hong, Y. Yang, Q. Chen, M. Cai, T.-B. Song, C.-C. Chen,
S. Lu, Y. Liu, H. Zhu and Y. Yang, ACS Nano, 2014, 8, 1674.
J. Burschka, N. Pellet, S.-J. Moon, R. Humphrey-Baker, P. Gao,
M. K. Nazeeruddin and M. Gr ¨a tzel, Nature, 2013, 499, 316.
P. Docampo, F. Hanusch, S. D. Stranks, M. D o¨ blinger, J. M. Feckl,
M. Ehrensperger, N. K. Minar, M. B. Johnston, H. J. Snaith and
T. Bein, Adv. Energy Mater., 2014, 4, 1400355.
5
6
7
A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc.,
2
009, 131, 6050.
J. H. Im, C. R. Lee, J. W. Lee, S. W. Park and N. G. Park, Nanoscale,
011, 3, 4088.
2
H.-S. Kim, C.-R. Lee, J.-K. Im, K.-B. Lee, T. Moehl, A. Marchioro,
S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Gr ¨a tzel
and N.-G. Park, Sci. Rep., 2012, 2, 591.
8
9
M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and
H. J. Snaith, Science, 2012, 338, 643.
J. M. Ball, M. M. Lee, A. Hey and H. J. Snaith, Energy Environ. Sci.,
2013, 6, 1739.
1
0 D. Bi, S.-J. Moon, L. Haggman, G. Boschloo, L. Yang, E. M. J. Johansson,
M. K. Nazeeruddin, M. Gr ¨a tzel and A. Hagfeldt, RSC Adv., 2013, 3, 18762.
1 H.-S. Kim, I. Mora-Sero, V. Gonzalez-Pedro, F. Fabregat-Santiago,
E. J. Juarez-Perez, N.-G. Park and J. Bisquert, Nat. Commun., 2013, 4, 2242.
2 M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395.
3 E. Edri, S. Kirmayer, A. Henning, S. Mukhopadhyay, K. Gartsman,
Y. Rosenwaks, G. Hodes and D. Cahen, Nano Lett., 2014, 14, 1000.
4 J. J. Choi, X. Yang, Z. M. Norman, S. J. I. Bilinge and J. S. Owen, Nano
Lett., 2014, 14, 127.
1
1
1
1
1
1
1
1
5 K. K. Bass, R. E. McAnally, S. Zhou, P. Djurovich, M. E. Thompson
and B. Melot, Chem. Commun., 2014, 50, 15819.
6 T. Ripolles-Sanchis, A. Guerrero, J. Bisquert and G. Garcia-
Belmonte, J. Phys. Chem. C, 2012, 116, 16933.
7 A. Abrusci, S. D. Stranks, P. Docampo, H.-L. Yip, A. K.-Y. Jen and
H. J. Snaith, Nano Lett., 2013, 13, 3124.
8 H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, Z. Hong,
J. You, Y. Liu, T. Liu and Y. Yang, Science, 2014, 345, 542.
9 L. Brus, J. Phys. Chem., 1986, 90, 2555.
1
2
0 G. Giorgi, J.-I. Fujisawa, H. Segawa and K. Yamashita, J. Phys. Chem.
Lett., 2013, 4, 4213.
2
1 R. Lindblad, D. Bi, B.-W. Park, J. Oscarsson, M. Gorgoi, H. Siegbahn,
M. Odelius, E. M. J. Johansson and H. Rensmo, J. Phys. Chem. Lett.,
2014, 5, 648.
Fig. 4 Scanning electron microscope (SEM) images of (a) planar
(
CH
3
NH
3
)PbI
3
, (b) (CH
3
NH
3
)PbI
3
in TiO
2
, and (c) (CH
3
NH
3
)PbI
3
in ZrO
2
.
22 J. Evan, L. Pedesseau, J.-M. Jancu and C. Katan, J. Phys. Chem. Lett.,
013, 4, 2999.
2
2
2
2
3 L. Chai, R. T. White, M. T. Greiner and Z. H. Lu, Phys. Rev. B:
Condens. Matter Mater. Phys., 2014, 89, 035202.
4 H. G. Agrell, G. Boschloo and A. Hagfeldt, J. Phys. Chem. B, 2004,
108, 12388.
5 J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal, C.-S. Lim, J. A. Chang,
Y. H. Lee, H.-J. Kim, A. Sarkar, Md. K. Nazeeruddin, M. Gr ¨a tzel and
S. I. Seok, Nat. Photonics, 2014, 7, 486.
Work supported by the American Chemical Society Division
of Inorganic Chemistry Exxon-Mobil Solid State Faculty Fellow-
ship and the National Science Foundation (NSF) Division of
Materials Research (DMR) CAREER grant #1253562.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015