From 2D to 1D Electronic Dimensionality in Halide Perovskites with Stepped and Flat Layers Using Propylammonium as a Spacer

Article abstract of DOI:10.1021/jacs.9b02846

Two-dimensional (2D) hybrid halide perovskites are promising in optoelectronic applications, particularly solar cells and light-emitting devices (LEDs), and for their increased stability as compared to 3D perovskites. Here, we report a new series of structures using propylammonium (PA⁺), which results in a series of Ruddlesden-Popper (RP) structures with the formula (PA)₂(MA)_n-1Pb_nI_3n+1 (n = 3, 4) and a new homologous series of step-like (SL) structures where the PbI₆ octahedra connect in a corner- and face-sharing motif with the general formula (PA)_2m+4(MA)_m-2Pb_2m+1I_7m+4 (m = 2, 3, 4). The RP structures show a blue-shift in bandgap for decreasing n (1.90 eV for n = 4 and 2.03 eV for n = 3), while the SL structures have an even greater blue-shift (2.53 eV for m = 4, 2.74 eV for m = 3, and 2.93 eV for m = 2). DFT calculations show that, while the RP structures are electronically 2D quantum wells, the SL structures are electronically 1D quantum wires with chains of corner-sharing octahedra insulated by blocks of face-sharing octahedra. Dark measurements for RP crystals show high resistivity perpendicular to the layers (10¹¹ ω cm) but a lower resistivity parallel to them (10⁷ ω cm). The SL crystals have varying resistivity in all three directions, confirming both RP and SL crystals' utility as anisotropic electronic materials. The RP structures show strong photoresponse, whereas the SL materials exhibit resistivity trends that are dominated by ionic transport and no photoresponse. Solar cells were made with n = 3 giving an efficiency of 7.04% (average 6.28 ± 0.65%) with negligible hysteresis.

Full text of DOI:10.1021/jacs.9b02846

Subscriber access provided by Bethel University
Article
From 2D to 1D Electronic Dimensionality in Halide Perovskites
with Stepped and Flat Layers Using Propylammonium as a Spacer
Justin M. Hoffman, Xiaoyang Che, Siraj Sidhik, Xiaotong Li, Ido Hadar, Jean-
Christophe Blancon, Hisato Yamaguchi, Mikaël Kepenekian, Claudine Katan, Jacky
Even, Constantinos C. Stoumpos, Aditya D. Mohite, and Mercouri G. Kanatzidis
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b02846 • Publication Date (Web): 05 Jun 2019
Downloaded from http://pubs.acs.org on June 5, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
From 2D to 1D Electronic Dimensionality in Halide Perovskites with Stepped
and Flat Layers Using Propylammonium as a Spacer
1
3
2
1
1
Justin M. Hoffman, Xiaoyang Che, Siraj Sidhik, Xiaotong Li, Ido Hadar, Jean-Christophe
2
3
4
4
5
Blancon, Hisato Yamaguchi, Mikaël Kepenekian, Claudine Katan, Jacky Even, Constantinos
C. Stoumpos, Aditya D. Mohite, and Mercouri G. Kanatzidis *
1
2
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
E-mail: m-kanatzidis@northwestern.edu
1
Department of Chemistry, Northwestern University, Evanston, IL 60208, United States
Department of Chemical and Biomolecular Engineering, Rice University, Houston, Texas
2
77005, United States
3
Los Alamos National Laboratory, Los Alamos, NM 87545, United States
Univ Rennes, ENSCR, INSA Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes)
4
–
UMR 6226, Rennes F-35000, France
Univ. Rennes, INSA Rennes, CNRS, Institut FOTON – UMR 6082, Rennes F-35000, France
5
Abstract
Two-dimensional (2D) hybrid halide perovskites are promising in optoelectronic applications,
particularly solar cells and light emitting devices (LEDs), and for their increased stability
compared to 3D perovskites. Here, we report a new series of structures using propylammonium
+
(
PA ) which results in a series of Ruddlesden-Popper (RP) structures with the formula
(
PA) (MA) Pb I (n = 3, 4) and a new homologous series of “step-like” (SL) structures
n 3n+1
2
n-1
where the PbI octahedra connect in a corner- and face-sharing motif with the general formula
6
(
PA)2m+4(MA) Pb
I
(m = 2, 3, 4). The RP structures show a blue-shift in bandgap for
m-2
2m+1 7m+4
decreasing n (1.90 eV for n = 4 and 2.03 eV for n = 3) while the SL structures have an even
greater blue-shift (2.53 eV for m = 4, 2.74 eV for m = 3, and 2.93 eV for m = 2). DFT
calculations show that, while the RP structures are electronically 2D quantum wells, the SL
structures are electronically 1D quantum wires with chains of corner-sharing octahedra
“
insulated” by blocks of face-sharing octahedra. Dark measurements for RP crystals show high
1
1
7
resistivity perpendicular to the layers (10 Ω·cm) but a lower resistivity parallel to them (10
Ω·cm). The SL crystals have varying resistivity in all three directions, confirming both RP and
SL crystals’ utility as anisotropic electronic materials. The RP structures show strong
photoresponse, whereas the SL materials exhibit resistivity trends that are dominated by ionic
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
transport and no photoresponse. Solar cells were made with n = 3 giving an efficiency of 7.04%
(average 6.28 ± 0.65%) with negligible hysteresis.
Keywords: solar cell, photovoltaic, quantum confinement, light emitting diode, thin films,
perovskitoid, photoluminescence
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Introduction
Halide inorganic-organic perovskites have shown remarkable capabilities to be used in a
wide variety of optoelectronic devices.1-11 Two-dimensional (2D) halide perovskites are of
particular interest because of increased stability in air and light, compared to the heavily studied
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
D perovskites, while maintaining properties conducive to making high-efficiency optoelectronic
devices.12-20 These materials derive from the 3D AMX , halide perovskites, in which A must be
+
3
+
+
+
+
+
2+
2+
2+
2+
-
-
CH NH (MA ), HC(NH ) (FA ), or Cs , M is generally Pb , Sn , or Ge , and X is Cl ,
3
3
2 2
-
-
Br , or I . When larger hydrophobic cations are used in the synthesis, new phases of perovskites
form which show lower dimensionality and significantly improved moisture stability.21
Commonly seen are 2D sheets where the 3D perovskite structure is cut along the <100>
plane such as with the most commonly employed cations phenylethylammonium (PEA)12, 17, 22 or
butylammonium (BA)13, 15 among others, which act as spacers between the forming slabs.23-25
2
6
These two cases along with other less studied ones use monoamines to form the Ruddlesden-
Popper (RP) perovskites with the general formula (A’) A M X (in which A’ is the large
3n+1
2
n-1
n
2
7
-
spacing cation). Here, n refers to the thickness of the layers in units of [MX ] octahedra.
6
/2
Other types of <100>-oriented perovskites include Dion-Jacobson (DJ) perovskites using certain
2
8
diamines with the formula (A’)A M X , alternating cations in the interlayer space (ACI)
n-1
n
3n+1
with the formula (A’) A M X ,29-30 and alkyl diammonium based with the formula
2
n-1
n
3n+1
3
1
(
H N(CH ) NH )A M X . Cases where the cut is along the <110> plane have been
3
2 n
3
n-1
n
3n+1
documented as well,25, 32-37 such as for (NH C(I)=NH ) , where the organic I atom interacts
2
2 2
strongly with the I atoms at the edges of the inorganic layer along the <110> plane, thus
stabilizing this uncommon structural perovskite type.38-39 A few examples also exist of
perovskites being cleaved along the <111> plane, for example with the trivalent metals in metal
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
39-40
deficient A M X perovskites.
The formulas for each of the structure types are shown in a
3
2
9
hierarchal manner in Scheme 1 using the variables n, m, and q to indicate the different
orientation of the octahedra. All these 2D materials are semiconductors that form natural
quantum wells between the alternating organic and inorganic layers, forcing spatial and dielectric
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
5
confinement of electron-hole pairs within the 2D perovskite plane. These materials can be
chemically tuned to give specific layer thickness corresponding to n in the general chemical
4
1
formula and can attain different electronic states and bandgaps. This allows further
optimization for optoelectronic applications including solar cells and light emitting diodes.13, 15,
4
2-43
Certain intermediate size cations, that are too large to form a 3D perovskite and too small
to intercalate between 2D perovskite sheets, can lead to unique structure types. Some of these
cations lead to structures which maintain 3D connectivity. For example, in the case of the tin
iodide system, several 3D tin-based hexagonal perovskites with the formula (A’)SnI have been
3
reported in which the octahedra connect in both the typical corner-sharing fashion as well as a
4
4
+
+
+
+
+
face-sharing fashion. This is done by using EA (CH CH NH ) , GA (C(NH ) ), ACA
3
2
3
2 3
+
(
CH C(NH ) ), and IM (C N H ) as spacing cations. These are not proper perovskites because
3 2 2 3 2 5
they do not feature exclusively corner-sharing octahedra and in order to distinguish them from
4
4
perovskites, they are referred to as perovskitoids. Other examples of 3D structure types with
intermediate cations include the so-called “hollow” perovskites in which the cation replaces
metal octahedra while maintaining corner-sharing 3D connectivity. A disordered structure can be
stabilized by shorter diamines such as in (en)MAMI (M = Sn, Pb)45-46 and (PN/TN)FASnI (PN
3
3
4
7
=
propylenediammonium; TN = trimethylenediammonium) while an ordered structure is seen
4
8
with alkanolamines such as ethanolamine. The same tendencies toward the disruption of the
4
ACS Paragon Plus Environment

Page 5 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
inorganic lattice can be seen in the 2D materials.49-50 For example, materials such as (PPA) Pb I
3
2 7
+
+
+
+
and (PBA) Pb I (PPA = C H (CH ) NH and PBA = C H (CH ) NH ), whose alkyl chain
3
2 7
6
5
2 3
3
6
5
2 4
3
length in these phenylalkylammonium cations varies, can also induce the formation of structures
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
51
with mixed edge-, face-, and/or corner-sharing. These types of materials are stabilized using
intermediate cations but are less explored. Furthermore, 1D perovskitoids derived from the
4
4
BaNiO structure type have fully face-sharing connectivity while other 1D halide materials
3
5
2
have mixed-sharing connectivity as well such as (C H CH NH ) Pb I ·2H O which has
6
5
2
3 4
5 14
2
5
3
corner- and edge-sharing connectivity. The shortest spacer length in the RP family which has
been studied crystallographically is the butylammonium cation (four carbon chain length).
Although one expects longer molecules to also stabilize this 2D halide perovskite family by
54
expanding the inorganic layer spacing, it is interesting to consider how short the spacer can get
before we observe some sort of structural instability. Such destabilization could arise for
example from the inorganic layers being pulled too close together and steric repulsion becoming
dominant.
Here, we present a new series of materials using propylammonium (PA) chosen
specifically because it is the shortest spacer cation among the n-alkylammonium family (three
carbon chain length) that can stabilize a perovskite in order to test the limits of phase formation
in the RP 2D (PA) (MA) Pb I perovskite family. We find that the “n = 1” and “n = 2”
n 3n+1
2
n-1
members cannot be formed, but instead a new series of step-like (SL) structures forms. In this
series, the parent “n = 2” structure, cleaved already along the <100> plane, is further cleaved
along the <110> plane, giving the general formula (PA)2m+4(MA) Pb
I
. These structures
m-2
2m+1 7m+4
have corner-sharing as well as face-sharing PbI octahedra, but is distinct from the previously
6
5
5
reported (PA) Pb I structure which also has mixed-sharing connectivity. For n > 2 the RP
3
2 7
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
materials are stabilized. For high m (m = 4), the structure is most like that of the RP structures
while m = 2 deviates significantly because of its increased face-sharing connectivity. Thus, this
system allows us to carefully tune the structure and compare it to the RP n = 3 and n = 4
structures using the same spacer, giving us the ability to test the effects of connectivity in a
logical, controlled manner. The optical properties of the RP materials are blue-shifted relative to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MAPbI , but for the new SL structures, the band gap is further blue-shifted with increasing
3
bandgap from m = 4 to 2, allowing for even further tuning of the bandgap. Band structure
calculations show dispersive bands and suggest the SL materials are electronically 1D quantum
wires while the RP materials are electronically 2D quantum wells. The RP materials exhibit a
strong photoresponse. Because of the good photoresponse of the RP n = 3 material, it was
employed in photovoltaic devices. By utilizing our previously studied “hotcasting” technique56
5
7
along with so-called solvent engineering, a preliminary assessment gave an efficiency of 7.04%
(average of 6.28 ± 0.65%) was obtained with limited hysteresis and practically no current
leakage.
Experimental
Synthesis.
See Table S1for a summary of synthetic procedures.
(
PA) Pb I (m = 2). PbO (2.232 g, 10 mmol) was dissolved in a mixture of 10.0 mL aqueous HI
8 5 18
and 1.7 mL aqueous H PO by heating at 190 ºC under constant magnetic stirring for 15 min. In
3
2
a separate vial, 5 mL aqueous HI was cooled in an ice bath. Then, PA (1.316 mL, 16 mmol) was
added slowly and allowed 5 min to fully neutralize. To the hot bright yellow solution was added
the neutralized PA. The resulting yellow solution was stirred for 5 min. Then, it was allowed to
6
ACS Paragon Plus Environment

Page 7 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
cool to room temperature and recrystallize for 45 min, giving yellow needles. The needles were
dried through vacuum filtration for 30 min before being dried under vacuum overnight. Yield
4.6979 g (61.80% based on mol Pb).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
PA) (MA)Pb I (m = 3). PbO (2.232 g, 10 mmol) was dissolved in a mixture of 8.0 mL
10 7 25
aqueous HI and 1.7 mL aqueous H PO by heating at 200 ºC under constant magnetic stirring for
3
2
1
5 min. To this was added MACl (196.9 mg, 2.91 mmol), which formed a black precipitate
which was allowed to redissolve. In a separate vial, 2 mL aqueous HI was cooled in an ice bath.
Then, PA (1.028 mL, 12.51 mmol) was added slowly and allowed at least 5 min to fully
neutralize. To the hot bright yellow solution was added the neutralized PA. The resulting yellow
solution was stirred until it began to boil. Then, it was allowed to cool to room temperature and
recrystallize overnight, giving orange needles. The plates were dried through vacuum filtration
for 30 min before being dried under vacuum overnight. XRD shows a mixture of m = 2, m = 3,
and a minor phase, (PA) Pb I .
3
2 7
(
PA) (MA) Pb I (m = 4). PbO (2.232 g, 10 mmol) was dissolved in a mixture of 10.0 mL
12 2 9 32
aqueous HI and 1.7 mL aqueous H PO by heating at 200 ºC under constant magnetic stirring for
3
2
1
5 min. To this was added MACl (338 mg, 5 mmol), which formed a black precipitate which
was allowed to redissolve. In a separate vial, 2 mL aqueous HI was cooled in an ice bath. Then,
PA (822 µL, 10 mmol) was added slowly and allowed at least 5 min to fully neutralize. To the
hot bright yellow solution was added the neutralized PA. The resulting yellow solution was
stirred until it began to boil. Then, it was allowed to cool to room temperature and recrystallize
overnight, giving red elongated plates. The plates were dried through vacuum filtration for 30
min before being dried under vacuum overnight. Yield 1.8494 g (24.81% based on total Pb).
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
PA) (MA) Pb I (n = 3). PbO (2.232 g, 10 mmol) was dissolved in a mixture of 10.0 mL
2
2
3 10
aqueous HI and 1.7 mL aqueous H PO by heating at 190 ºC under constant magnetic stirring for
3
2
15 min. To this was added MACl (450 mg, 6.67 mmol), which formed a black precipitate which
was allowed to redissolve. In a separate vial, 5 mL aqueous HI was cooled in an ice bath. Then,
PA (548 µL, 6.67 mmol) was added slowly and allowed at least 5 min to fully neutralize. To the
hot bright yellow solution was added the neutralized PA. The resulting yellow solution was
stirred for 5 min. Then, it was allowed to cool to room temperature and recrystallize for 1.5 hr,
giving dark red plates. The plates were dried through vacuum filtration for 30 min before being
dried under vacuum overnight. Yield 2.4019 g (34.73% based on total Pb).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(PA) (MA) Pb I (n = 4). PbO (2.232 g, 10 mmol) was dissolved in a mixture of 10.0 mL
2
3
4 13
aqueous HI and 1.7 mL aqueous H PO by heating at 190 ºC under constant magnetic stirring for
3
2
1
5 min. To this was added MACl (507 mg, 7.5 mmol), which formed a black precipitate which
was allowed to redissolve. In a separate vial, 3 mL aqueous HI was cooled in an ice bath. Then,
PA (411 µL, 5.0 mmol) was added slowly and allowed at least 5 min to fully neutralize. To the
hot bright yellow solution was added the neutralized PA. The resulting yellow solution was
stirred for 5 min. Then, it was allowed to cool to room temperature and recrystallize for 1.5 hr,
giving black plates. The plates were dried through vacuum filtration for 30 min before being
dried under vacuum overnight. XRD shows some impurities such as n = 3 and MAPbI . Yield
3
1
.4588 g (21.70% based on total Pb).
Materials Characterization
High resolution synchrotron powder X-ray diffraction (SXRD) data were collected at
beamline 11-BM-B of the Advanced Photon Source (APS), Argonne National Laboratory using
8
ACS Paragon Plus Environment

Page 9 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
λ = 0.412736 Å. Discrete detectors covering an angular range from -4 to 2° 2θ were scanned over
a 26° 2θ range, with data points collected every 0.001° 2θ and a scan speed of 0.1°/s.
Single-crystal X-ray diffraction experiments were performed using a STOE IPDS II or
IPDS 2T diffractometer with Mo Kα radiation (λ = 0.71073 Å) and operating at 50 kV and 40
mA. Integration and numerical absorption corrections were performed using the X-AREA, X-
RED, and X-SHAPE programs. The structures were solved by charge flipping and refined by
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
8
59
full-matrix least-squares on F2 using the Jana2006 package. The PLATON software was used
to identify the twinning domains and validate the space groups of the compounds.
A Shimadzu UV-3600 PC double-beam, double-monochromator spectrophotometer was
used to acquire room-temperature optical diffuse reflectance spectra of the powdered samples in
the range of 200−2500 nm. BaSO was used as a non-absorbing reflectance reference, and
4
reflectance data were converted to absorbance data via Kubelka−Munk transformation. Both
bandgaps and exciton energies were extracted from the data. The exciton energies were
estimated based on the position of the exciton peak residing below the bandgap. The linear
portion of the curve above the exciton was used to extract the bandgap based on its intersection
with the x-axis. Low-energy impurity peaks, appearing as tails in the spectra, were ignored.
First-principles calculations are based on density functional theory (DFT) as implemented
in the SIESTA package.60-61 Calculations have been carried out on experimental structures with
6
2
the GGA functional in the revPBE form. Core electrons are described with Troullier-Martins
6
3
pseudopotentials. The valence wavefunctions are developed over double-ζ polarized basis set of
64
finite-range numerical pseudoatomic orbitals. In our calculations, spin-orbit coupling is taken
6
5
into account through the on-site approximation as proposed by Fernández-Seivane et al. In all
cases, an energy cutoff of 150 Ry for real-space mesh size has been used.
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Current-voltage (I-V) curves were measured at room temperature in air using a Keithley
517b picoammeter/voltage supply under -10 to 10 V bias via a two-probe method. Devices
6
were made using single crystals of the compounds. Electrical contacts were applied through 100
μm copper wires adhered to the crystal specimens through colloidal graphite paste.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Measurements were made on single crystals with a specific direction chosen based on the faces
of the crystals and correlated using BFDH morphology simulations by the Mercury software.66
The devices were put inside a guarded dark box for dark measurements. Additionally, they were
measured under a white light source with a power setting of 7 W (Taotronics TT-DL11).
퐴
Resistivity was calculated using the equation 휌 = 푅 ∗ where ρ is resistivity, R is resistance
푙
calculated by the slope of the I-V curve collected above, and A and l are the cross-sectional area
of the crystal and the length from one contact to the other. It was assumed that the contact
resistance was similar for each crystal.
Thin Film Characterization
Thin films were prepared on glass substrates. XRD measurements were carried out on a
Rigaku MiniFlex600 X-ray diffractometer (Cu Kα radiation, λ = 1.5406 Å) operating at 40 kV
and 15 mA. GIWAXS measurements were performed at Beamline 8-ID-E of the Advanced
Photon Source at Argonne National Laboratory. Samples prepared on glass substrates were
exposed to an X-ray beam (λ = 1.6868 Å) at an incident angle of 0.15° for 1 s, and the scattered
light was collected by a Pilatus 1 M pixel array detector at 204 mm from the sample. The
GIXSGUI program was used to plot images of the patterns.67
Solar Cell Device Fabrication
The ITO-coated glass substrates (TEC7, 2.2 mm, Hartford Glass Co. Inc.) were cleaned
by sequential sonication in aqueous detergent, deionized water, acetone, and isopropyl alcohol
10
ACS Paragon Plus Environment

Page 11 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
for 10 min each, followed by a 3 min oxygen plasma treatment. A thin layer of PEDOT:PSS was
deposited on the precleaned ITO substrates by spin-coating at 6000 rpm for 30 s and annealed at
150 °C for 30 min in air. The substrates were then transferred to an argon-filled glovebox to
complete the rest of the device fabrication. Solutions of (PA) (MA) Pb I were prepared by
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
3 10
dissolving 60 mg of the oven-dried crystals in 100 µL of DMF and 300 μL of DMSO solvents.
Different volumes of DMSO were incorporated into the total 400 µL when optimizing solvent
choice (0, 100, 200, 300, and 4000). The perovskite layer was deposited by hot-casting 75 μL of
the as-prepared room-temperature solution on the ITO/PEDOT:PSS substrates heated to 110ºC
for 20 s at 5000 rpm. Within 6-7 s of spin coating, the films turned from yellow to brown. The
PC BM (Sigma Aldrich) layer was deposited on the perovskite film from a 30 mg/mL PC BM
6
1
61
solution in CB at 1000 rpm for 30 s. Finally, 100 nm of Al was thermally evaporated through a
−
6
shadow mask at a pressure of ∼1 × 10 Torr. Half of the area was encapsulated by epoxy
(Epoxy Technology) and UV-cured for 30 min. The active area of the device was 0.03 cm2.
Solar Cell Device Characterization
The current density versus voltage (J−V) characteristics were collected in air using a
2
Keithley 2400 source meter under simulated AM 1.5G irradiation (100 mW/cm ) generated by a
standard solar simulator (Abet Technologies). The light intensity was calibrated by using an
NREL-certified monocrystalline Si reference cell to reduce the spectral mismatch.
Stability characterization was performed on a device using 75% DMSO for the precursor
solvent. The device was made with the same architecture used previously and encapsulated with
the same method. The device was then held under 1 solar sun and measured periodically.
Results and Discussion
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Materials Synthesis
The materials were synthesized by first dissolving lead oxide (PbO) and
methylammonium chloride (MACl) in boiling hydroiodic acid (HI) and hydrophosphorous acid
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
H PO ). Then, propylamine (PA), neutralized by HI, was added to the solution. The reagents
3 2
were originally added in ratios corresponding to the stoichiometry for RP perovskites for
different n values using the general formula (A’) A M X3n+1. Because of the high solubility of
2
n-1
n
the products, concentrations were kept as high as possible (≥1 M). Once dissolved, the solutions
were cooled to room temperature by removing the solutions from the hot plate, leading to the
precipitation of the materials within minutes. Note that the crystals must be completely dried
within hours of precipitation because they can undergo a high rate of disproportionation. As seen
in Figure S2, over days, other n values began to appear in the PXRD for (PA) (MA) Pb I (n =
2
2
3 10
3
), and even within hours, a red tint, indicative of lower n structures, was seen on the edges of
the crystals. For n = 3 and (PA) (MA) Pb I (n = 4), plates colored dark red and black,
2
3
4 13
respectively, precipitated within minutes (Figure 1). For n = 4, however, it was fairly common to
have impurities of other n members while n = 3 was more likely to be pure. For n < 3, needles
formed instead of plates. For the experimentally stoichiometric “(PA) (MA)Pb I ” samples, red
2
2 7
flat needles formed instead and for “(PA) PbI ” yellow needles formed. The “(PA) PbI ” sample
2
4
2
4
matched the calculated PXRD for a previously reported material by Billing and Lemmerer,
6
8
(
PA) Pb I , but the “(PA) (MA)Pb I ” material did not match any known phase. After
8 5 18 2 2 7
resolving the crystal structure for this material, it was revealed that the so-called “(PA) PbI ” and
2
4
“
(PA) (MA)Pb I ” materials were part of a homologous series of SL materials with the general
2 2 7
formula (PA)2m+4(MA) Pb
I
.
m-2
2m+1 7m+4
12
ACS Paragon Plus Environment

Page 13 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
To further probe this series for new phases, syntheses were performed using
stoichiometric ratios for values of m from 2 to 4. Unexpectedly, doing so led to the precipitation
of the yellow (PA) Pb I (m = 2) (Pb:I is 1:3.60) phase in all cases. Using the stoichiometry for
8
5 18
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
“
(PA) (MA)Pb I ” (Pb:I is 1:3.50) was the most reliable way to precipitate the red
2 2 7
(
PA) (MA) Pb I (m = 4, Pb:I is 1:3.56). When using an intermediate ratio of reagents between
12 2 9 32
the stoichiometry for “(PA ) (MA)Pb I ” and (PA) (MA) Pb I , a mixture of three phases
3 2
2 7
12
2
9 32
formed: yellow needles identified as m = 2, orange needles identified as (PA) (MA)Pb I (m =
1
1
7 25
3
, Pb:I is 1:3.57), and white needles identified as the MA-free (PA) Pb I . The third phase was
3 2 7
an impurity, though it was previously reported as the primary product under synthetic conditions
5
5
similar to those used here to synthesize m = 2. No instances of m > 4 were seen as the
stoichiometry further approached that of “(PA) (MA)Pb I ”. Furthermore, using limiting or
2
2 7
excess of a single reagent simply led to mixtures of the title phases and no new phases were
detected.
The SL (PA)2m+4(MA) Pb
I
materials are derived from the hypothetical RP “n =
m-2
2m+1 7m+4
2
” structure. In this case, when m = ∞, the formula collapses to (PA) (MA)Pb I , which is the
2 2 7
equivalent to the RP (PA) (MA) Pb I “n = 2” member. All the SL structures begin to
n 3n+1
2
n-1
+
+
deviate from the RP series because of an increasing ratio of PA :MA from 2:1 for “n = 2” to
infinity for m = 2 and a slightly decreasing ratio of Pb:I from 1:3.5 to 1:3.6.
Description of Crystal Structures.
Step-like Structures. Using propylammonium as a spacer results in a SL structures instead of the
expected “n = 1” and “n = 2” RP homologous members. The SL structures involve both corner-
sharing and face-sharing PbI octahedra, which have been colored blue and pink, respectively, in
6
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. These compounds form their own homologous series within the “n = 2” window, which
are described by the general formula (PA)2m+4(MA) Pb
I
(m ≥ 2). Having identified this
m-2
2m+1 7m+4
6
8
new homology here we can see that the previously reported m = 2 compound is a member
along with the m = 3 and 4 compounds reported here for the first time. Each of these structures,
like the RP perovskites, can be imagined as a derivative of a parent structure. For the RP
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
perovskites, the parent structure is MAPbI , which has been cleaved along the <100> plane by
3
the stoichiometric addition of the iodide salts of the alkylammonium cations. For the SL
structures, however, the parent structure is the hypothetical “n = 2” RP structure which, as
mentioned earlier, is essentially MAPbI cleaved along the <100> plane. The SL structure can
3
then be imagined to be cleaved a second time along the <110> plane of MAPbI , which is
3
described by the m variable in the formula (PA)2m+4(MA) Pb
I
. The combination of
m-2
2m+1 7m+4
these two perpendicular cleaving events produces a 1D chain of PbI octahedra with thickness n,
6
which always equals 2, and a width of m, which varies between 2-4. The use of m is based on the
nomenclature previously reported for 2D corner-sharing perovskite structures cleaved along
<
110>.69
Looking down the monoclinic b axis shows a cross section of the 1D chains for m = 2,
seen as the blue corner-sharing octahedra in Figure 2. In the case of m = 2, the chains are 2 x 2
octahedra. The 1D chain for this series is approximately two PbI octahedra thick (~12.5 Å) or
6
2
aʹ in length, (where aʹ is the ideal cubic perovskite lattice constant) corresponding to n = 2
thickness. The width for m = 2, 3, and 4 is 9.1351 Å (2aʹ/√2), 13.7718 Å (3aʹ/√2), and 18.1148 Å
4aʹ/√2), respectively. Looking down the c axis in Figure 3b gives a side view of the chains,
(
which extend infinitely along the b axis. As m increases from 2 to 4, the pure zigzag connectivity
subsides as the chain gains more octahedra and the 1D chains begin to more closely resemble a
14
ACS Paragon Plus Environment

Page 15 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
perovskite-like structure. The zigzag connectivity is like that of the <110>-oriented 2D
3
9
perovskites with proper corner-sharing connectivity as in m = 2 [NH C(I)=NH ] Sn I .
2
2 2
2 8
The 1D chains in the SL structures are connected by a single face-sharing octahedron, for
which the face-sharing connectivity is at the two opposite corners of these rectangles along the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
D chain. This forms the so-called “steps,” as the rectangles are the flat portion of the steps and
the face-sharing octahedra are the raise in height for each step. In (PA)2m+4(MA) Pb
I
2m+1 7m+4
m-2
the face-sharing octahedra face a different direction than the corner-sharing octahedra with an
interval of one octahedron. The few examples of mixed-connectivity structures seen in the halide
perovskite literature have also seen different patterns within the context of SL structures.
Kamminga et. al. describe a phenylbutylammonium-based material with n = 1 and m = 3 which
49
is quite similar to the structures reported here. Interestingly, the shorter chain
phenylpropylammonium forms a new type of structure with n = 1 and m = 4, but with the face-
sharing octahedra at corners on the same side of the rectangle cross-section, forming a cis-
orientation. Using a different synthesis, Billing et. al. solved the structure with a
phenylpropylammonium spacer with n = 1 and m = 2, but this structure has chains of three face-
7
0
sharing octahedra which also arrange in a cis pattern. Additionally, an all-inorganic structure
+
+ 71
was reported in for a SnI system using a mixture of Rb and Cs . This appears to be the only
6
other known multilayered structure with mixed-sharing connectivity.
+
In this homology the MA always is located within the cages of 2 x 2 x 2 corner-sharing
PbI octahedra. To form cages, the width of the chains must be at least m = 3. With each
6
+
additional layer, another MA can be incorporated. For <100>-oriented structures in general, the
spacing amines align themselves in the interlayer space with the amine group facing the
inorganic layers and positioned between the four terminal iodide atoms of four different
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
octahedra. For the general <110>-oriented structure, the spacing cations orient themselves in a
similar manner, but the four octahedra have their edges facing the interlayer space. In the case of
+
the SL structures, the PA orient themselves at both the <100> and <110> cleavage of the 1D
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
chains. There are 2m PA along the <100> cleavage and 2 along the <110> cleavage, giving 2m
+
+
+
2 PA oriented toward the inorganic chain and 2 still unaccounted. These 2 PA are found in
the large space near the face-sharing octahedra. This large space exists because the SL structure
of the inorganic layers is suboptimal to pack tightly together. The intermediate size of the PA+
cations allows them to properly fill up the interlayer spacing. If the cation size were smaller, they
+
may be able to incorporate into the 3D structure in a similar way to MA . Conversely, if the size
+
was a bit larger or just longer such as BA , the interlayer spacing would be too small for all the
cations to fit and the structure would be destabilized.
When attempting to synthesize the m = 3 and 4 structures, a minor phase (PA) Pb2I7
3
5
5
forms as well (Figure S3). The structure has also mixed face- and corner-sharing motifs with
D chains of corner-sharing octahedra cut along the <100> and <110> axes, and there is no fully
face-sharing octahedron, causing the chains to be at alternating orientations to one another. The
PA) Pb I however was only made as an impurity and its study is out of the scope of this paper,
1
(
3
2 7
so the characterization is not reported here beyond the crystallographic data given in Table 1.
Ruddlesden-Popper Structures. The n = 3 and n = 4 structures can be classified as Ruddlesden-
Popper (RP) materials. Like other RP systems, the n = 3 and n = 4 structures are fully corner-
sharing structures belonging to the formula (PA) (MA) Pb I but with a considerably
n 3n+1
2
n-1
contracted stacking repeat length given the short spacer (Figure 3). Crystallographic data is
reported in Table 1 and selected bond distances, angles, etc. in Tables S2-16. Like other RP
16
ACS Paragon Plus Environment

Page 17 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
structures, the structures can be imagined as MAPbI cleaved along the <100> plane by spacing
3
cations, which in this case A’ is propylammonium. The PbI octahedra, both terminal and within
6
the layers, experience distortion along all three axes, with the Pb-I-Pb angles ranging from
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13
1
61.77º to 171.28º, similar to other alkylamine-based perovskites. The n = 3 structure
crystallizes in the monoclinic space group P2 /a, as opposed to the n = 3 BA compounds which
1
crystallize in the orthorhombic space group C2cb and n = 3 PEA compound which crystallize in
the triclinic space group P1. The decrease in symmetry for PA n = 3 is caused by the unique
alignment of the layers in the structure. In general, RP perovskites have layers shifted by half an
octahedral unit length in the layer plane ((½, ½) displacement) so that when looking
perpendicular to the layers, the successive layers alternate in position (Figure 3). In the case of
the PA system, however, as the layers are brought together in closer proximity, they are not fully
1
1
shifted and the displacement is closer to ( ⁄ , ⁄ ), lying in between the stacking sequences of RP
4
4
and DJ perovskites. A change in the stacking sequence may also occur as a function of
72
temperature for layered perovskites such as (C H NH ) PbI . Different layer alignments are
5
11
3 2
4
common in 2D halide perovskites when small spacer cations are employed, as seen in the case of
GA,29-30 in which the layers are (½, 0) (the ACI perovskite), and for hybrid DJ perovskites, in
which the layers have no shift (Figure 3c,d). In addition, in the monoclinic setting, the unit cell
no longer doubles but instead incorporates only one layer. A similar primitive monoclinic unit
cell without cell doubling along the stacking axis was also identified for ([p-
28
FC H C H NH ] PbI ). The reduced point symmetry of the lattice by comparison to RP, DJ, and
6
5
2
4
3 2
4
ACI structures is compensated by the smaller number of atoms in the unit cell.73-74 Finally, the
intermediate-sized propylammonium spacer also results in a relatively short interlayer spacing of
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
.03 Å, which is shorter than the width of a PbI in the structure (~6.3 Å) and an I-I distance of
6
only 6.75 Å.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Optical Properties
Optical Absorption. Freshly dried crystals were used to acquire the optical responses of
7
5
these materials. Diffuse reflectance of the bulk polycrystalline powders obtained at room
temperature were used to create Kubelka-Munk plots for bandgap determination (Figure 4a).
Note that the spectra here contain two features: a bandgap absorption and an excitonic peak at a
lower energy. The bandgap was then estimated by the intercept of the linear portion of the
spectra with the x-axis, considering only the portion of the spectrum at higher energy than the
excitonic peak. Note that strong impurity peaks, likely n = 1, 2, or 3, which may have formed on
the surfaces of the crystals for the SL structures, are seen as tails toward the low energy end of
the spectra. Despite not being able to make the n = 1 and n = 2 RP structures in bulk, they may
be stabilized at the surfaces of the SL compounds, which would influence the optical spectrum.
The exciton energy, when possible, was also estimated based on position of the low-energy
exciton peak. It should also be noted that the absorption edge extends well beyond the bandgap,
as has been seen in previous reports which seemed to depend on the preparation method, in
particular when using HI during the synthesis.31, 76
4
1
The SL materials show higher bandgap compared to other RP n = 2 materials, which
reflects the decrease in structural dimensionality, as the perovskite-like connectivity only extends
in one dimension, even though the compound remains 2D structurally. Within the homologous
series of SL materials, the bandgap increases as m decreases, showing that the width of the
quantum wire (and thus its cross-sectional area) plays a key role in the quantum confinement. As
18
ACS Paragon Plus Environment

Page 19 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
m decreases from 4 to 3 to 2, the bandgap increases from 2.53 eV to 2.74 eV to 2.93 eV,
indicating a further increase in the electronic confinement with decreasing cross-sectional area.
Interestingly, the m = 2 structure has a bandgap closer to that of the <110>-oriented GA m = 2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
77
structure (2.49 eV) rather than to BA n = 2 (2.17 eV), revealing a strong structural correlation
between the <110>-oriented GA and a marked difference with the <100>-oriented BA, despite
both orientations being present in PA.
As expected, in comparison to the MAPbI bandgap (1.52 eV), the n = 4 and n = 3
3
materials were blue-shifted with bandgaps of 1.90 eV and 2.03 eV, respectively. This is because
of increasing quantum and dielectric confinement in the z direction with progressively thinner
perovskite layers. The same trend can be seen from the excitonic peaks, with energies of 1.86 eV
and 2.00 eV.
Photoluminescence (PL). The PL of crystals of the different materials were measured at room
temperature (Figure 4b). The SL structures are again blue-shifted compared to the RP structures
and continue to blue shift with decreasing m. Similar to the optical absorption, n = 3 and n = 4
were shown to blue shift with decreasing layer thickness. A comparison of the bandgaps and PL
is shown in Figure 4c. A notable difference between the RP and SL structure is the intensity of
the PL peaks, with RP perovskites showing a strong narrow emission and the SL perovskites
exhibiting a weak emission. PL lifetime measurements show that the average lifetime decreases
from 3.10 ns to 2.37 ns to 0.29 ns from m = 4 to 3 to 2 (Figure S4). The short lifetimes suggest
that efficient non-radiative decay routes are significant in these materials, dominating that of the
radiative processes. Based on DFT calculations in the next section, it appears that these materials
have indirect bandgaps, suggesting that the PL is a result of a trap-assisted process not directly
produced from the bandgap, leading to weak emissions. This hypothesis is also consistent with
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the broadening of the PL peaks down to 1.8 eV, as a broad number of states beyond the bandgap
can participate in these processes.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Electronic Properties.
DFT electronic structure calculations were performed on each of these materials (Figure
5
). The SL materials have different electronic structure with significantly less dispersive bands
than RP materials. Using the m = 2 member as an example, the bands show some dispersion
parallel to the 1D chains of corner-sharing PbI octahedra, though less so compared to the RP
6
phases. This is because the 1D corner-sharing chains are similar to the corner-sharing in the RP
perovskite derivatives, while the face-sharing connections in the structure along the layer seem to
weaken the hybridization of electronic atomic states close to the bandgap. This can be most
readily visualized by the partial charge densities computed at the VBM and CBM in Figure 5,
where both the intralayer face-sharing and the interlayer directions lead to a localization of the
electron densities. As seen by the VBM of m = 2, the face-sharing octahedra have negligible
electron density, as the main fraction of the electron density localized on the corner-sharing
octahedra. Based on this result, we can conclude that the face-sharing octahedra affect the
electronic dimensionality of the perovskite-like corner-sharing 1D chains, acting as electronic
barrier, a conclusion which agrees well with previous reports of structures with face-sharing
5
3
connectivity. Furthermore, the third direction, running parallel to the chains of corner-sharing
octahedra, has little dispersion compared to the calculations for n = 3, indicating the zigzag
connectivity also leads to poor dispersion. This is consistent with other <100>-oriented layered
7
8
halide perovskites, such as ammonium-propyl-imidazole-based structures. As m increases, the
dispersion along the chains of corner-sharing octahedra increases, as expected because of the
20
ACS Paragon Plus Environment

Page 21 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
increased perovskite-like character of the chains. Furthermore, the m = 4 structure begins to
show small amounts of coupling between the chains, overcoming the face-sharing octahedra
barrier. The local density of states close to the band edges also becomes similar to that of RP
structures.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
For the n = 3 and n = 4 RP materials, the band structure show an almost direct bandgap
4
1
similar to that of the other RP materials. The calculations indicate that the n = 3 member is a
direct bandgap semiconductor at the  point of the Brillouin zone, whereas the n = 4 member has
a slightly indirect bandgap at the  point (Figure 5a). The n = 4 member shows a slight deviation
from the typical direct bandgap behavior, observed in halide perovskites because of strong
Rashba splitting caused by the noncentrosymmetric nature of its structure. Within the layers, the
electronic bands are dispersive because of their perovskite-like nature resulting from the corner-
sharing PbI octahedra. Across the layers in the z direction, however, the bands show virtually no
6
dispersion because of the structural discontinuity caused by the spacing of the propylammonium
cations. As was reported previously, the conduction band maximum (CBM) is made mainly of
6
p orbitals of Pb located on the top and bottom surface of the perovskite slabs while the valence
band maximum (VBM) is made mainly of 5p orbitals of I located in the center, similar to the
BA-based RP structures.30
It is worth pointing out that m = 2 shows a clear indirect bandgap while m = 3 and m = 4
both show almost direct bandgaps. Even though the calculated bandgap is underestimated, which
41
is common for this type of calculations, the general experimental trend is captured by the DFT
calculations, with the slight exception of m = 3 (Figure S4). In this case, the calculated bandgaps
for the materials are 0.51 eV for n = 4, 1.03 eV for n = 3, 1.75 eV for m = 4, 1.60 eV for m = 3,
and 2.27 eV for m = 2.
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Electrical Properties
Photoconductivity measurements were taken for n = 3 RP perovskite, m = 2, and m = 4
SL perovskites at room temperature using a two-probe setup. Individual faceted crystals of each
compound were measured along different directions with the intention to capture the highly
anisotropic properties of the crystals.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
For the SL materials, three unique directions are possible. For m = 2, the crystals are
needles which can be imagined as rectangular prisms elongated along the b-axis. From tip-to-tip
for these needles, the crystallographic direction is the b-axis, which is the axis running parallel to
the 1D chains of corner-sharing octahedra. The measured resistivity is high along this direction
10
(
ρ ~ 10 Ω·cm), but still an order of magnitude smaller than the one corresponding to the cross-
plane direction in RP phases discussed below, while measurements along the a- and c-axes gives
7
8
resistivities on the order of ρ ~ 10 - 10 Ω·cm. While the resistivity values are seemingly quite
low here, it has been noted that ion migration is common in perovskites79-80 and increased in 2D
8
1
perovskites because of octahedral distortion. In the SL structures, the octahedral distortions are
significant, as in the RP structures. Also, the layers are no longer flat and thus have more
exposure to the interlayer space, which could act as a pathway for ion migration. If this is the
case, the measurement would be heavily influenced by ion migration instead of electronic
conduction and would lead to inflated values. A similar situation occurs for m = 4, with values of
6
8
6
resistivity ρ ~ 10 Ω·cm for the b-axis, ρ ~ 10 Ω·cm for the a-axis, and ρ ~ 10 Ω·cm for the c-
axis. While these values are only estimates, they demonstrate the anisotropic effects in these
crystals.
22
ACS Paragon Plus Environment

Page 23 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The n = 3 structure crystallizes as thin rectangular plates in the typical morphology for
8
2
the RP structures. Measuring along the tetragonal facets, corresponding to the cross-plane
1
1
direction, a resistivity of ρ = 10 Ω·cm was obtained, whereas measuring along the thin sides of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the crystals, corresponding to the layer plane direction, resulted in a much smaller resistivity of ρ
8
=
10 Ω·cm (Figure 6a). This vividly illustrates the anisotropic nature of the RP perovskites, and
it is also reasonable in the sense that the resistivity of the layer plane is comparable to that
7
83
observed for the 3D perovskite (ρ = 10 Ω·cm). This is understandable, as the carriers easily
transport through the perovskite-like planes, whereas much higher bias is required to force the
charge carriers to jump across adjacent layers. This observation is in agreement with previously
asserted measurements, even though the anisotropy is higher in the PA system compared to the
29
GA system, possibly because of the very short interlayer spacing in the ACI perovskites. The
photoconductive response, measured using white light is clear along both measured directions.
Parallel to the layers, the resistivity decreased a hundredfold and perpendicular to the layers, it
decreased tenfold. Somewhat surprising is the photoresponse perpendicular to the layers, which
implies that some small current can vertically cross the layers. This suggests that certain cations
are not sufficiently insulating to completely prevent charge transfer across the layers, even if the
transfer is minimal.
It is important to note that the SL materials displayed no difference in the current under
dark and light conditions, showing a lack of photoresponse in all directions for both structures.
This shows that the SL compounds are inappropriate for photovoltaics, yet they may find
applications as dielectric gates in halide perovskite optoelectronics because of their tunable
resistivity as a function of the 1D chain thickness. Furthermore, perovskites have seen utility as
ferroelectrics and magnets, as well as being utilized for ion exchange and reactivity
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
applications.24, 84 Since little is understood about how hybrid halide perovskitoids can be utilized
for these applications, this system may be used as a tunable model to see how a departure from
the perovskite structure affects these materials’ behavior in these other applications.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Thin Films and Solar Cell Devices
Thin Films Because of the negligible photoconductivity of the SL materials, only solar cell
devices using the RP phases were considered. Because of the lower purity of n = 4, n = 3 was
chosen for device fabrication instead. In a previous report, RP phases were successfully made
8
5
into high-efficiency devices by a process known as hotcasting. In this process, the substrate is
heated (in this case to 110 ºC) immediately before spin coating begins, which allows for
increased grain passivation and more preferable perpendicular orientation of the layers to the
substrate. The devices were shown to have high short-circuit current density (J ) because of
SC
increased charge transfer. Furthermore, solvent engineering has been shown to increase the
57
crystallinity of the 2D materials in the films. This can be done by simply using mixtures of
DMF and DMSO as opposed to only using DMF. The DMSO functions to slow the
crystallization of the films from 1-2 s to 6-7 s and thus leads to better crystallinity. To assess
better conditions for devices, XRD of films was measured using different concentrations of
DMSO in DMF (0%, 25%, 50%, 75%, and 100%). As seen by the PXRD in Figure S6a, the
FWHM of the (-110) peak decreases immediately upon addition of DMSO but does not further
improve with concentrations of DMSO above 25%. DMSO does, however, lead to a decrease in
preferred orientation and a disproportionation of phases, seen by the low angle peaks
corresponding to n = 4 which grow in with increasing concentrations of DMSO. GIWAXS
(Figure 7) further backs these trends, though it shows the n = 3 material as well based on low
24
ACS Paragon Plus Environment

Page 25 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
angle peaks. The spots in the GIWAXS spectrum become rounder as DMSO increases, which
shows an increase in crystallinity. This data can be further analyzed by taking line-cuts of the 2D
spectrum in different directions. By taking the FWHM along the q and q directions, the
z
y
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
crystallinity perpendicular to the substrate and parallel to the substrate, respectively, can be
distinguished. As seen in Figure S7, the crystallinity along q does not change in a substantially
z
nor consistent manner, indicating that the grains did not change in size perpendicular to the
substrate. In contrast, along q , the crystallinity overall increased and is greater compared to
y
along q . This suggests that the crystallites in the film are plate-like and lie parallel to the
z
substrate. Furthermore, increasing the concentration of DMSO increases the width of the
crystallites while not increasing the depth. Further studies will be needed to understand the
mechanism which leads to these trends and the effect of DMSO.
Solar Cells. Encapsulated devices were prepared using an inverted structure:
ITO/PEDOT:PSS/Perovskite/PCBM/Al, in which PEDOT:PSS functioned as the hole-
transporting material (HTM) and PCBM as the electron-transporting material (ETM). By tuning
the percent of DMSO in solution, a respectable average device efficiency of 6.28 ± 0.65% was
found by using 75% DMSO, much higher than the 3.70 ± 0.37% efficiency for pure DMF.
Typical device curves are seen in Figure 8b, and the average parameters are seen in Table S32.
Furthermore, the average VOC decreased slightly from 1.07 ± 0.06 V to 1.00 ± 0.03 V, the
2
2
average J increased significantly from 5.33 ± 0.66 mA/cm to 10.95 ± 0.57 mA/cm , and the
SC
FF decreased slightly from 68.6 ± 4.6% to 57.0 ± 3.5%. The increase in J is attributed to a
SC
slower growth rate of the film from DMSO, which leads to better crystallinity, as mentioned
above. The decrease in preferred orientation seems to be less essential in this case than the
crystallinity. Even so, DMSO has also shown to lead to very large grain size on the order of over
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a micron, which may lead to small gaps between the grains, negatively affecting the VOC and FF,
5
7
as seen here. The external quantum efficiency (EQE) was also measured for the devices and is
-2
shown in Figure 8c with an integrated J of 9.23 mA cm for 75% DMSO which is slightly
SC
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-2
lower than the J of 11.86 mA cm based on the J-V curve from the same device. Another point
SC
of interest is the broadening in the EQE spectrum with addition of DMSO. Even at 25% DMSO,
the curve has broadened significantly at lower wavelength. However, all films show a very broad
spectrum, especially at higher wavelength, as seen by the appearance of a shoulder at around 720
nm. This broadening could be attributed to disproportionation in the films, as evidenced by the
presence of n = 4 in the XRD. The broader EQE implies that this disproportionation is
potentially advantageous to absorb a broader range of wavelengths of light.
The wide EQE of the films, potentially due to the disproportionation in the films based on
XRD, is promising. In fact, the disproportionation of the n values in the film based on XRD
implies that low n values are also in existence in the film, as has been reported for other systems
86
spectroscopically. Here, however, the RP n = 1 and n = 2 structures are less likely to form than
the SL structures, which have no photoresponse and are detrimental to device performance. The
fact that the J was not as high as expected, along with lower FF and V suggests that the
SC
OC
issues arising here derive from the film morphology and presence of SL structures, not the RP n
3 material itself. Furthermore, the increase of J from DMSO additive implies the changes in
=
SC
morphology and makeup alleviate some of the potential issues, suggesting that further tuning of
the conditions for film fabrication could lead to higher J . In fact, for films using the GA
SC
spacing cation, fine tuning using antisolvent addition has given efficiencies of 14.7% once the
determinantal (GA) PbI impurity was suppressed. Also like the GA structures, the short
2
4
interlayer space allows for a lower organic volume in the structure and thus a higher volume of
26
ACS Paragon Plus Environment

Page 27 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
perovskite character in the structure, which should allow for a high absorption coefficient. Each
of these observations implies the potential for greater J for PA-based devices once optimization
SC
takes place. As a preliminary way to analyze these devices and to better understand this system,
further device studies were performed.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Series resistance (R ) can be a major limitation for J and is here estimated by the slope
S
SC
at the open-circuit voltage, where lower slopes resulting from resistive “tails” near the VOC lead
to increased R . For films with only DMF, a high resistivity is seen, likely due to poor interfaces
S
or defects in film morphology such as low crystallinity and poor mosaicity. This can be tamed by
the previously mentioned processing techniques of hotcasting and solvent engineering. In fact,
by using solvent engineering through mixing DMF and DMSO, the resistive “tail” seen near the
open-circuit voltage disappears entirely under high concentrations of DMSO (Figure 9).
Likewise, the series resistance (R ) is estimated by the slope at the J , where a greater slope
SH
SC
shows more shunt pathways. Once again, this is an effect of the film quality, not the material,
and is improved substantially by the addition of DMSO, as seen clearly by the dark curves in
Figure 9c,f,i.
The final issue tackled in this system was the hysteresis, which is seen both by varying
the scan speed and direction. While using pure DMF gave large hysteresis, DMSO addition
diminished the hysteresis substantially. Though full optimization of these devices is out of the
scope of this paper, they show the potential for further improvements because of the excellent
properties of the crystalline materials and the widened EQE curve, allowing for a broader range
of the solar spectrum to be absorbed. It should be noted that, despite having similar bandgaps to
the BA system and even shorter interlayer spacing, the performance is still far from that of BA.85
Based on the above analysis, it appears that the intrinsic properties of the PA materials are not to
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
blame, but the film morphology and makeup hinders the performance. Once the optimal
conditions of film formation for this material are better understood, higher films made from pure
crystals of these materials can be prepared, and the wide spread of n values in the films can be
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
utilized to get high J and thus improved efficiencies.
SC
Conclusions
+
Using the short PA spacing cation in PAI/MAI/PbI reactions results in two homologous
2
series of 2D organic-inorganic hybrid lead iodide materials: RP perovskites (PA) (MA) Pb I
n 3n+1
2
n-1
(
n = 3 and 4) and SL perovskitoids (PA)2m+4(MA) Pb
I
(m = 2, 3, and 4). What is quite
m-2
2m+7 7m+4
remarkable is that, by simply changing the PA:MA ratio, this system gives both the
electronically 1D SL materials and the electronically 2D RP materials, showing the huge range
of electronic and structural tunability within the same series. These trends are not found in the
longer n-alkyl ammonium 2D perovskites. Along with GA and MA in the ACI 2D perovskites,
PA is among the shortest organic spacers in 2D perovskites. The SL structures in these materials
have a new type of connectivity in the lead iodide layers in which chains of corner-sharing PbI6
octahedra are connected by face-sharing octahedra to form structurally 2D SL slabs. Optically,
the PA n = 3 and n = 4 structures are like other known RP structures with bandgaps for n = 3 and
4
of 2.03 eV and 1.90 eV while the SL structures have bandgaps over a wider range for m = 2, 3,
and 4 of 2.93 eV, 2.74 eV, and 2.53 eV. This shows that the bandgap can be further tuned by
adjusting not only the thickness of the slabs n, but also the width m, which could be useful in
designing materials with very specific bandgaps without relying on halide mixing which can
result in phase segregation. Unlike the RP structures which are electronically 2D, the SL
structures are electronically 1D with dispersion along the corner-sharing chains. Even though the
28
ACS Paragon Plus Environment

Page 29 of 56
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
chains incorporate every component of the MAPbI structure for m = 3 and 4, the materials are
3
confined to one dimension, showing the possibility for 2D perovskite-like materials which are
electronically 1D, further enhancing the ability to play with the optical and electronic properties
of these materials by simply adjusting the spacing cation.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The RP materials have good charge transport along the lead-iodide slabs along with a
good photoresponse, allowing them to be used in solar cells. By using solvent engineering with
varying DMF/DMSO mixtures, crystalline films can be made, leading to solar cells with
respectable efficiencies. While the efficiencies are not yet competitive with other 2D perovskite
materials, the PA system seems to have greater disproportion than other cations such as BA
8
5
based on the EQE spectrum. If this property can be utilized with improved film morphology,
these materials have the potential for higher J by absorbing a greater range of wavelengths of
SC
light. Also, it is fundamentally interesting that, when compared to the BA system, by simply
removing a carbon atom from the BA spacer, the film kinetics change so dramatically. We can
therefore use these materials to better understand film formation and define the techniques
suitable for shorter spacer cations. A systematic comparison between the film forming properties
of PA and BA films will further understanding of the disproportionation of n values in films,
important for LED applications.17, 87
ASSOCIATED CONTENT
Supporting Information
Materials and methods, synthetic details, additional supplementary figures and tables about
material characterization, SEM images, X-ray diffraction measurements, photoluminescence
measurements, thermogravimetric analysis, profilometry measurements and film stability
studies.
29
ACS Paragon Plus Environment