Synthetic Variation and Structural Trends in Layered Two-Dimensional Alkylammonium Lead Halide Perovskites

Article abstract of DOI:10.1021/acs.chemmater.9b01318

We report the cooling-induced crystallization of layered two-dimensional (2D) lead halide perovskites with controllable inorganic quantum-well thicknesses (n = 1, 2, 3, and 4), organic-spacer chain lengths (butyl-, pentyl-, and hexylammonium), A-site cations (methylammonium and formamidinium), and halide anions (iodide and bromide). Using single-crystal X-ray diffraction, we refined crystal structures for the iodide family as a function of these compositional parameters and across their temperature-dependent phase transitions. In general, lower-symmetry crystal structures, increasing extents of organic-spacer interdigitation, and increasing organic-spacer corrugation tilts are observed at low temperature. Moreover, greater structural distortions are observed in lead halide octahedra closest to the organic-spacer layer, and higher-n structures exhibit periodic variation in Pb-I bond lengths. These structural trends are used to explain corresponding temperature-dependent changes in the photoluminescence spectra. We also provide detailed guidance regarding the combination of synthetic parameters needed to achieve phase-pure crystals of each composition and discuss difficulties encountered when trying to synthesize particular members of the 2D perovskite family containing formamidinium or cesium as the A-site cation. These results provide a foundation for understanding structural trends in 2D lead halide perovskites and the effects these trends have on their thermal, electronic, and optical properties.

Full text of DOI:10.1021/acs.chemmater.9b01318

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Synthetic Variation and Structural Trends in Layered Two-
Dimensional Alkylammonium Lead Halide Perovskites
Watcharaphol Paritmongkol, Nabeel S Dahod, Alexia Stollmann, Nannan
Mao, Charles Settens, Shao-Liang Zheng, and William A. Tisdale
Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.9b01318 • Publication Date (Web): 10 Jul 2019
Downloaded from pubs.acs.org on July 10, 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 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
Synthetic Variation and Structural Trends in Layered Two-
Dimensional Alkylammonium Lead Halide Perovskites
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
,2
2
2
2,3
Watcharaphol Paritmongkol , Nabeel S. Dahod , Alexia Stollmann , Nannan Mao , Charles
4
5
2
Settens , Shao-Liang Zheng , William A. Tisdale *
1Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts
0
2139, United States
2Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139, United States
3Department of Electrical Engineering and Computer Science, Massachusetts Institute of
Technology, Cambridge, Massachusetts 02139, United States
4Center for Materials Science and Engineering, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, United States
5Department of Chemistry and Chemical Biology, Harvard University, Cambridge,
Massachusetts 02138, United States
*
Correspondence to: tisdale@mit.edu
Abstract
We report the cooling-induced crystallization of layered two-dimensional lead halide perovskites
with controllable inorganic quantum-well thickness (n = 1, 2, 3, 4), organic spacer chain length
(butyl-, pentyl-, hexylammonium), A-site cation (methylammonium, formamidinium), and halide
anion (iodide, bromide). Using single-crystal X-ray diffraction, we refined crystal structures for
the iodide family as a function of these compositional parameters and across their temperature
dependent phase transitions. In general, lower symmetry crystal structures, increasing extents of
organic-spacer interdigitation, and increasing organic-spacer corrugation tilts are observed at low
temperature. Moreover, greater structural distortions are seen in lead halide octahedra closest to
1
ACS Paragon Plus Environment

Chemistry of Materials
Page 2 of 61
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 organic spacer layer, and higher-n structures exhibit periodic variation in Pb-I bond lengths.
These structural trends are used to explain corresponding temperature-dependent changes in the
photoluminescence spectra. We also provide detailed guidance regarding the combination of
synthetic parameters needed to achieve phase-pure crystals of each composition, and discuss
difficulties encountered when trying to synthesize particular members of the 2D perovskite family
containing formamidinium or cesium as the A-site cation. These results provide a foundation for
understanding structural trends in 2D lead halide perovskites and the effect these trends have on
their thermal, electronic, and optical properties.
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 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
Introduction
Layered two-dimensional (2D) lead halide perovskites (LHPs) have emerged as an alternative to
their bulk or three-dimensional (3D) counterparts due to improved stability,1,2 rich structural
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
,4
3,5
flexibility, and greater tunability of optical properties. Their structures consist of 2D inorganic
slabs of ABX perovskite unit cells separated by large monovalent cations. Their general formula
3
is L [ABX ] BX or L A B X3n+1, where L is a large monovalent cation, A is a small cation
2
3 n-1
4
2
n-1
n
(i.e., methylammonium {MA}, formamidinium {FA} or cesium {Cs}), B is a divalent cation of
lead (Pb), X is a halide anion (Cl, Br, or I), and n determines the inorganic quantum-well thickness
defined by the number of corner-sharing BX octahedral layers. Since L is usually a large organic
6
molecule with an amine functional group in its protonated form, L acts as an electronically
insulating organic spacer, and 2D LHPs form natural quantum-well structures. Due to the spatial
confinement and the low dielectric screening of organic spacers, 2D LHPs have high exciton
6,7
binding energies (>400 meV for single-layered 2D LHPs). Moreover, both optical and electronic
properties of 2D LHPs can be tuned by varying the identities of each constituent ion (L, A, B, and
X) as well as n.3,5 Hence, 2D LHPs have been applied in various optoelectronic devices, such as
solar cells,1,2,8–12 light-emitting diodes,13–16 and photodetectors.17–20
Electronic and optical properties of 2D LHPs are determined by their structure.3,21–24 The structural
evolution of 3D LHPs and single-layered (n = 1) 2D LHPs with the absence of small A-site cations
has been extensively investigated and shows interesting temperature-dependent
transformations.25–30 For example, 3D MAPbI perovskites adopt cubic, tetragonal, and
3
orthorhombic phases at high, intermediate, and low temperatures, respectively.25–27 On the other
hand, the number of crystal phases and phase transition temperatures of single-layered
(
C H
NH ) PbI (m = 4, 5, 6, 7, 8, 9, 10, 12, 14, 16, and 18) are dependent on the chain length
2m+1 3 2 4
m
3
ACS Paragon Plus Environment

Chemistry of Materials
Page 4 of 61
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
of the organic spacer.28–30 Higher members (n ≥ 2) of 2D LHPs contain both organic spacers and
A-site cations, and lie in between these two limits. Interesting questions include how the structures
of n ≥ 2 2D LHPs evolve as a function of temperature and how the structure is influenced by
changing the organic spacer, A-site cation, or quantum-well thickness.31–33 This understanding has
been hampered, in part, by the difficulty of synthesizing phase-pure variants of n ≥ 2 2D LHP
single crystals.24
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
D LHPs can be prepared by a variety of methods.3,21,24,34–38 Polycrystalline thin films of 2D LHPs
can be made by spin casting a mixture of PbX , AX, and LX dissolved in dimethylformamide
2
and/or dimethylsulfoxide onto a spinning substrate. The same mixture can also be added drop-wise
to a non-polar solvent, such as toluene, to produce 2D nanoplatelets via ligand-assisted
reprecipitation.5,7,39 Colloidal solutions of 2D nanoplatelets can be prepared by a modified hot-
injection method,40,41 similar to the approach used to synthesize halide perovskite quantum dots.42
4
3
In addition, 2D sheets of LHPs can be grown directly on a substrate by solvent evaporation as
well as chemical vapor deposition.44,45 Inverse temperature crystallization, which employs the
retrograde solubility of perovskite reagents, was successfully used to make high-quality single
crystals of 3D LHPs. However, the reagents of 2D LHPs do not show retrograde solubility, and
46
47
the method is not feasible to grow 2D LHP single crystals due to the difference in surface-solvent
interaction. Anti-solvent vapor assisted growth has also been a successful approach for growing
4
8
49
3
D LHPs as well as n = 1 2D LHPs, but successful application of this method to the preparation
of n ≥ 2 2D LHPs has not been reported (our own attempts were unsuccessful). Recently,
synthesis of 2D LHPs with quantum-well thicknesses of n ≤ 3 was reported using controlled ratios
5
0
of solvent and anti-solvent. However, the method that has demonstrated the greatest versatility
so far is cooling-induced crystallization, which has been shown to produce pure single crystals of
4
ACS Paragon Plus Environment

Page 5 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
D methylammonium lead iodide perovskites with precise thickness control of up to n = 5.51–53
2
3
2
Recently, this limit has been extended further to n = 7, and has been applied to a variety of
organic spacers.10,18,33,51,54 In this method, perovskite ions are dissolved in hot aqueous hydrohalic
acid, and the solution is then allowed to cool slowly to induce crystallization.
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
Herein, we report successful cooling-induced crystallization and structural evolution of
(
C H NH ) (MA)Pb I ,
(C H NH ) (MA) Pb I ,
(C H NH ) (MA) Pb I ,
4 9 3 2 3 4 13
4
9
3 2
2 7
4
9
3 2
2
3 10
(
C H NH ) (MA)Pb I , (C H NH ) (MA)Pb I , and (C H NH ) (FA)Pb I . Henceforth, these
5
11
3 2
2 7
6
13
3 2
2 7
4
9
3 2
2 7
chemical formulae will be abbreviated as n = 2 BA-MAPbI, n = 3 BA-MAPbI, n = 4 BA-MAPbI,
n = 2 PA-MAPbI, n = 2 HA-MAPbI, and n = 2 BA-FAPbI, respectively. Optical absorption,
photoluminescence, and X-ray diffraction further confirm their chemical compositions and
structures. Moreover, we discuss factors that influence the phase purity of 2D LHP crystals and
provide guidance for achieving 2D LHP crystals of desired composition.
Furthermore, structural dependence of 2D LHPs on quantum-well thicknesses, organic spacer
lengths, and A-site cations are discussed. We found that (i) there is a general trend of lower crystal
symmetry as the temperature is reduced; (ii) PbI octahedra on a surface (i.e., proximal to the
6
organic spacer layer) have higher distortion and tilting angles than their middle-layer counterparts;
(
iii) there is an alternation of axial Pb-I bond lengths, with terminal bonds being the shortest; (iv)
there is a cohesive trend in the change of Pb-(µ-I)-Pb bond angles as the crystal symmetry is
lowered; and (v) there are increases in the extents of interdigitation and corrugation tilts of organic
spacer molecules as 2D LHPs adopt their lower-temperature structures. Increasing the length of
the organic spacer molecule affects the number of phase transitions of n = 2 samples and the tilting
angles of PbI octahedra, while replacing the A-site cation from MA to FA has little effect on the
6
5
ACS Paragon Plus Environment

Chemistry of Materials
Page 6 of 61
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
structures – except in the tilting angles of the PbI octahedra. In addition, structural distortion is
6
used to explain photoluminescence peak shifts across phase transitions.
Finally, we conclude by further modification of the cooling-induced crystallization to prepare
bromine-based 2D LHPs. Successful syntheses of n = 1 BA-PbBr, n = 2 BA-MAPbBr, and n = 3
BA-MAPbBr are reported, along with their absorption and photoluminescence spectra and powder
X-ray diffraction (PXRD) patterns.
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
Experimental Section
Synthesis of iodide 2D LHPs. Crystals of iodide 2D LHPs were synthesized using a modified
procedure based on a cooling-induced crystallization method, as reported before.18,51,52 For the
synthesis of iodide 2D LHPs, a large volume of lead (II) iodide (PbI ) was prepared by dissolving
2
lead (II) oxide (PbO, 99.9+%, (trace metal basis) <10 microns, powder, ACROS Organic) in 55%
hydrogen iodide solution (HI, contains no stabilizer, ACS reagent, 55%, MilliporeSigma) under
reflux for 15 mins. A small volume of this solution was then transferred into a small vial, and
mixed with hypophosphorus acid (H PO ). After the mixed solution turned bright yellow, a small
3
2
volume of long chain organic spacer (L = butylamine {BA}, pentylamine {PA}, or hexylamine
HA}) was added in an ice bath, and an orange precipitate of n = 1 L-PbI formed. To synthesize a
{
higher member of iodide 2D LHPs, a solution of the A-site cation iodide salt (AI = MAI or FAI)
was prepared in a separate vial by dissolving the salt in HI. This AI solution was subsequently
added into the solution containing the orange precipitate to form a combined solution. Additional
volume of HI may be added before the final solution was heated on a hotplate set at 130 °C for 4
mins or until clear. Crystals of iodide 2D LHPs were then grown by storing the clear solution in a
thermos filled with hot sand at 110 °C. After a day, the crystals were isolated by suction filtration
6
ACS Paragon Plus Environment

Page 7 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
and dried under reduced pressure for at least 12 hours. The quantities of reagents used for making
iodide 2D LHPs can be found in Table S1.
Synthesis of bromide 2D LHPs. To synthesize bromide 2D LHPs, lead (II) bromide (PbBr ) was
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
prepared by dissolving PbO (99.9+%, (trace metal basis) <10 microns, powder, ACROS Organic)
in hydrogen bromide solution (HBr, ACS reagent, 48%, MilliporeSigma)) under reflux for 15
mins. A small volume of this solution was then transferred into a small vial. Without any addition
of H PO , a small volume of BA was added in an ice bath and white precipitate of n = 1 BA-PbBr
3
2
formed. In a separate vial, a solution of the A-site cation bromide salt (ABr = MABr) was prepared
by dissolving the salt in HBr. This ABr solution was subsequently added into the solution
containing the white precipitate to form a combined solution. Additional volume of HBr was added
before the final solution was heated on a hotplate set at 130 °C for 4 mins or until clear. Crystals
of bromide 2D LHPs were then grown by storing the clear solution in a thermos filled with hot
sand at 110 °C. After a day, the crystals were isolated by suction filtration and dried under reduced
pressure for at least 12 hours. The quantities of reagents used for making bromide 2D LHPs can
be found in Table S2.
Mechanical exfoliation. Exfoliated flakes of 2D LHPs were prepared by tapes. A crystal of 2D
LHPs was mounted on a piece of single-sided adhesive tape (Ultron Systems, 1007R-6.0). The
two ends of the tape were then folded together and separated to exfoliate the crystal. This step was
repeated 5-10 times before the tape was gently pressed on a substrate. After the tape was removed
from the substrate, thin flakes of 2D LHPs were transferred onto the substrate.
Absorption measurement. Transparent flakes of 2D LHPs were mechanically exfoliated onto fused
quartz substrate for absorption measurements. The absorption spectra were measured using a
7
ACS Paragon Plus Environment

Chemistry of Materials
Page 8 of 61
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
transmission mode on a home-built micro-absorption setup. The incident white light beam
(
Thorlabs, SLS201L) was focused onto the sample through an inverted lens (10X, NA=0.25). The
transmitted light was collected by a 100x lens (NA= 0.90) and analyzed by a Raman spectrometer
T64000, Horiba) with 150 g/mm grating. The absorbance (A) of a 2D LHP sample was calculated
by A = -log (I/I ), where I and I are the transmitted light intensities on the 2D LHP flake and
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
0
0
0
nearby quartz substrate.
Photoluminescence. Steady-state photoluminescence measurements were performed on an
inverted microscope (Nikon, Ti-U Eclipse) in air or vacuum. The samples were excited by focusing
the output of a 405 nm pulsed laser diode (Picoquant, LDHDC-405M, 2.5 MHz repetition rate) by
an objective lens (Nikon, CFI S Plan Fluor ELWD, 40x, 0.6 NA) to <1 µm spot. The
photoluminescence was collected in the epi configuration, and passed through a dichroic mirror
and a long-pass filter. Then, it was directed into a spectrograph (SP-2500, Princeton Instruments)
mounted with a cooled charge-coupled detector (Princeton Instruments, Pixis). Laser power was
kept below ~125 nW throughout the measurements. For the measurement on n = 1 BA-PbBr, the
excitation source was replaced by a 375 nm LED (Thorlabs, M375F2). Photoluminescence
measurements under vacuum were achieved by mounting 2D LHP crystals in a microscopy
cryostat (Janis Research, ST-500), and the sample was cooled by flowing liquid helium through a
cold finger attached to the base of the sample. Photoluminescence spectra measured under air and
vacuum were identical.
Powder X-ray diffraction (PXRD). Powder X-ray diffraction data was collected using a
PANalytical X’Pert Pro MPD X-ray diffractometer (Cu Kα radiation, λ = 1.54184 Å) with High-
Speed Bragg-Brentano Optics. A 0.04 rad soller slit, a 1° anti-scatter slit, a 10 mm mask and a
programmable divergence slit with an illuminated length of 6 mm were used in the incident beam
8
ACS Paragon Plus Environment

Page 9 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
path. The diffracted beam optics included a 0.04 rad soller slit, a Ni Filter and an automatic
receiving slit. The detector was an ultrafast X’Celerator RTMS detector. The angular step in 2θ
was 0.04°. For temperature dependent measurements, an Oxford Cryosystems PheniX cooler and
an Anton Paar HTK-1200N heating stage were used. All temperature-dependent measurements
were taken during heating, and the reversibility was checked by comparing room-temperature data
at the beginning and the end of the experiment. At each temperature, the temperature was stabilized
for ~15 minutes during the data collection.
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
Single-crystal X-ray crystallography. A crystal was mounted on a diffractometer at 300 K. The
temperature was raised/lowered down to measurement temperatures at a rate of 5 K/h before the
data sets were collected. The intensities of the reflections were collected by means of a Bruker
APEX II CCD diffractometer (MoK radiation,  = 0.71073 Å) equipped with an Oxford
Cryosystems nitrogen flow apparatus. The collection method involved 0.5 scans in  at 28 in
2. Data integration down to 0.84 Å resolution was carried out using SAINT V8.37A (Bruker
diffractometer, 2016) with reflection spot size optimization. Absorption corrections were made
with the program SADABS (Bruker diffractometer, 2016). The structure was solved by the
2
55
Intrinsic Phasing methods and refined by least-squares methods F using SHELXT-2014 and
5
6
57
SHELXL-2014 with OLEX 2 interface. Non-hydrogen atoms were refined anisotropically, and
hydrogen atoms were allowed to ride on the respective atoms. Crystal data as well as details of
data collection and refinement can be found in the supporting information. Further details on
refinement are also available in the CIF files provided. Visualization of crystallographic data was
performed by using VESTA.58
Results and Discussion
9
ACS Paragon Plus Environment

Chemistry of Materials
Page 10 of 61
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
1
. Syntheses of iodide 2D LHPs
Iodide 2D LHPs were synthesized using a modified procedure based on the cooling-induced
crystallization method18,51,52 (Figure 1a). Detailed tables specifying reagent quantities and
conditions used to synthesize each 2D LHP composition can be found in the Supporting
Information. Additional procedural details are included in the experimental section. Briefly, a large
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
volume of lead (II) iodide (PbI ) solution was prepared by the reaction of lead (II) oxide (PbO)
2
and 55% hydrogen iodide (HI) solution under reflux. A small volume of PbI solution was then
2
transferred into a vial, and hypophosphorus acid (H PO ) was added to reduce any residual iodine
3
2
(I2) back to HI. After a few seconds, the solution turned bright yellow. Next, the long chain organic
cation (L) was added, forming an orange precipitate of n = 1 L-PbI. To make thicker members of
the iodide 2D LHP family (n > 1), a solution of the A-site cation iodide (AI) salt was prepared in
a separate vial by dissolving the salt into HI. This AI solution was subsequently added into the
solution containing the orange precipitate. This combined solution was then heated until clear, and
slowly cooled to produce crystals of iodide 2D LHPs. The crystals were then collected by suction
filtration and dried under reduced pressure. Photographs of crystals obtained by this method are
shown in Figure 1b.
1
0
ACS Paragon Plus Environment

Page 11 of 61
Chemistry of Materials
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
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
Figure 1. (a) Schematic describing cooling-induced crystallization of 2D lead halide perovskites
(LHPs). (b) Photographs of 2D LHP crystals. The spacing between minor ticks on the ruler is 1
mm. (c) Schematic illustrations showing the structures of layered 2D LHPs and a 2D LHP single
layer. MA = methylammonium; FA = formamidinium; BA = butylammonium; PA =
pentylammonium; HA = hexylammonium.
We found that controlling relative quantities of reagents was the key to obtaining pure crystals.
Deviating from the optimal concentrations in the starting solutions resulted in crystals containing
impurities of thicker (e.g., n + 1) and/or thinner (e.g., n - 1) iodide 2D LHPs. Because of the
differences in solubilities among different reagents, each solidifies at a different rate and at
different temperature during cooling. Consequently, the optimal starting concentrations are not
simply the stoichiometric amounts found from the chemical formulae of iodide 2D LHPs; to obtain
pure crystals of the desired iodide 2D LHPs, the precipitation rates of precursors must be balanced,
which can be achieved by adjusting concentrations.
1
1
ACS Paragon Plus Environment

Chemistry of Materials
Page 12 of 61
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 overall effect of changing long chain organic spacer concentrations can be predicted from the
chemical formulae of iodide 2D LHPs. As we go from n = 1 iodide 2D LHP to bulk LHP, the L:Pb
ratio decreases from 2:1 to 0:1. This decrease suggests that the higher concentration of long chain
organic cation favors the formation of thinner (lower n) iodide 2D LHPs, which was indeed
observed experimentally (Figure S1). The long chain organic spacer is the least soluble component
of the reaction mixture and is the first to precipitate as the solution is cooled. When a stoichiometric
amount of long chain organic spacer was used, a complete crystallization of orange n = 1 iodide
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
D LHP or thinner (i.e. n – 1) iodide 2D LHPs was observed. Consequently, the amount of long
chain organic spacer used in our synthesis was always kept below that predicted from
stoichiometry (Table S1).
Because the solubility of the long chain organic spacer in HI decreases with increasing
alkylammonium chain length, decreasing concentrations are needed as the chain length increases.
The percentages of long chain organic spacers used with respect to their stoichiometric amounts
for making n = 2 BA-MAPbI, n = 2 PA-MAPbI, and n = 2 HA-MAPbI decreased from 65% to
6
0% to 50%, respectively (Table S1). The trend in solubility also affects the crystal shape. As HA
is the least soluble among the three organic spacers, it precipitated out fastest during cooling –
resulting in thinner, smaller crystals. On the other hand, large crystals with lateral sizes of up to 5
mm of n = 2 BA-MAPbI and n = 2 PA-MAPbI (Figure 1b) could be obtained easily with this
method. These crystals were large enough that lamellar structure is evident in macro-scale
morphology. This observation agrees with the reflections along (h00) planes in X-ray diffraction
data as reported previously18,47 and shown again in this work.
Thicker (higher-n) 2D LHPs can be obtained by decreasing the concentration of the long chain
organic spacer in the reaction mixture. The optimal percentage of long chain organic spacer with
1
2
ACS Paragon Plus Environment

Page 13 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
respect to the stoichiometrically required amount shows a downward trend (Table S1) due to
higher solubility of the A-site cation and lower solubility of long chain organic spacer. We found
that fixing the percentage of long chain organic spacer while changing n results in thinner (e.g. n
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) impurities.
The effect of changing A-site cation concentration can be similarly understood. We observed that
increasing the concentration of the A-site cation favors the formation of thicker (higher-n) iodide
2
D LHPs (Figure S2). In most of our syntheses, stoichiometric amounts of A-site cations were
used. However, due to the low solubility of HA, a more saturated solution of MA – more MA in
the starting mixture – was needed to match the precipitation rate of HA. As a result, 175% of MA
stoichiometric amount was needed to form phase-pure n = 2 HA-MAPbI. When a stoichiometric
amount of MA was used, orange crystals of n = 1 iodide 2D LHP were observed to be the main
products, regardless of the initial concentration of HA (Figure S3).
Adjusting the total volume of HI also affected the quality of the final crystallization product. We
found that a small volume of HI often resulted in mixed-phase iodide 2D LHPs and crowded
solutions with small crystals, whereas increasing HI volume promoted the formation of larger
crystals. However, when too much HI was added, the A-site cation did not precipitate out of
solution, resulting in the formation of n = 1 iodide 2D LHP crystals (Figure S4). We also found
that greater HI volume (per gram of PbO) was needed for the weakly-soluble longer-chain organic
spacers. When tuning n, HI volume was kept the same to minimize the number of changing
parameters. Detailed tables specifying reaction conditions and reagent quantities can be found in
the Supporting Information.
1
3
ACS Paragon Plus Environment

Chemistry of Materials
Page 14 of 61
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
When the A-site cation was changed from MA to FA or Cs, additional difficulties emerged. In the
case of FA cation, the precipitation rate was found to be faster, and the resulting crystal sizes
tended to be smaller. To grow larger crystals of high-purity n = 2 BA-FAPbI, the volume of HI
was increased relative to the amounts of other reagents. Attempts to synthesize n > 2 BA-FAPbI
were unsuccessful; formation of n = 2 BA-FAPbI was always observed, regardless of how the
synthesis conditions were varied (Figure S5). This observation suggests that n = 2 BA-FAPbI
might be the most thermodynamically stable product among the iodide 2D LHPs with FA cation.3
Similar observations have been reported in the case of n = 5 BA-MAPbI, which undergoes
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
2
disproportionation to form n = 3 BA-MAPbI and bulk MAPbI . Finally, we note that attempts to
3
synthesize iodide 2D LHPs with Cs cation were unsuccessful due to the poor solubility of CsI in
HI. These attempts resulted in a mixture of white CsI solid and orange n = 1 BA-PbI 2D LHP
(Figure S5).
In general, we observed some seasonal variation in the synthesis product. During high-humidity
seasons, needle-liked crystals with yellow color emerged as a by-product (Figure S6). Due to the
5
9,60
shape and color of this by product, it may be a water complex such as MAPbI ∙ H O
or
3
2
6
0–63
(MA) PbI ∙ H O.
Comparison of its powder X-ray diffraction (PXRD) pattern to known
4
6
2
structures of candidate compounds in the Cambridge Crystallographic Data Center are shown in
Figure S7. The formation of these unknown crystals could be avoided by slightly increasing the
AX fraction by approximately 5%.
2. Stability of iodide 2D LHP crystals and exfoliated flakes:
All of the iodide 2D LHP crystals synthesized by this method were stable to storage inside a N2
glovebox for more than one year (Figure S8). Additionally, the large crystals shown in Figure 1b
1
4
ACS Paragon Plus Environment

Page 15 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
could be mechanically exfoliated to form thin 2D LHP flakes (Figure S13; see Experimental
Section). However, when the crystals were exfoliated and kept under ambient conditions, newly
generated surfaces underwent a transition after ~3 days to form a yellow solid (Figure S9). We
hypothesize that, because the interaction between the long chain organic spacer and the lead halide
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
6
4
octahedra has a hydrogen-bonding character, the long chain organic spacer can be easily stripped
off during exfoliation. The newly exposed surface is unprotected, and susceptible to attack from
H O and O in the air to form PbI . This hypothesis was tested by storing exfoliated crystals either
2
2
2
in inert atmosphere or in oil. No apparent color change was observed under either condition (In
the case of oil, the experiment was performed for one week). These observations suggest that the
stability of exfoliated 2D LHP flakes could be further improved through encapsulation by a
polymer layer65,66 or another 2D material, such as hexagonal boron nitride (hBN).65,67–70
3
. Optical characterization of iodide 2D LHP crystals:
Room temperature absorption and photoluminescence spectra of exfoliated iodide 2D LHPs are
shown in Figure 2. All variants exhibit a single sharp excitonic absorption resonance 100-500 meV
below the continuum band and Stokes-shifted emission close in energy to the 1s excitonic
absorption feature. 2D alkylammonium LHPs exhibit large exciton binding energies (470 meV for
n = 1 butylammonium lead iodide) due to spatial confinement of electrons and holes within the 2D
6
inorganic lattice, and low dielectric screening from the long chain organic spacer layer. Increasing
n red-shifts the excitonic resonance in both absorption (Abs) and photoluminescence (PL) spectra
(Figure 2a). These shifts are the result of two partially offsetting factors: quantum confinement,
which widens the electronic band gap as n decreases; and dielectric confinement, which leads to
an increase in exciton binding energy as n decreases – thereby reducing the magnitude of the
overall blue-shift. The n = 2-4 samples show similar photoluminescence (PL) full width at half
1
5
ACS Paragon Plus Environment

Chemistry of Materials
Page 16 of 61
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
maximum (FWHM) of 25-27 nm, as well as similar Stokes shifts of ~10 nm (Table S3). In contrast,
the n = 1 sample shows narrower PL FWHM of 17±2 nm.
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
6
ACS Paragon Plus Environment

Page 17 of 61
Chemistry of Materials
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
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
Figure 2. Absorption (dotted) and photoluminescence (solid) spectra of exfoliated 2D LHP flakes
at room temperature. Comparisons are shown for (a) n = 1 BA-PbI, n = 2 BA-MAPbI, n = 3 BA-
MAPbI, and n = 4 BA-MAPbI; (b) n = 2 BA-MAPbI, n = 2 PA-MAPbI, and n = 2 HA-MAPbI;
(c) n = 2 BA-MAPbI and n = 2 BA-FAPbI.
For the n = 2 iodide samples, Abs and PL spectra were largely unaffected when the organic spacer
length increased from four to six carbons (Figure 2b). All samples showed Abs peaks centered
1
7
ACS Paragon Plus Environment

Chemistry of Materials
Page 18 of 61
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
near ~570 nm, PL peaks centered near ~580 nm, PL FWHM of ~25 nm and Stokes shift of ~10
nm (Table S3). However, when the A-site cation was changed from MA to FA (Figure 2c), the PL
FWHM decreased from 25±2 nm to 21±2 nm.
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 observation of single Abs and PL peaks in each sample also indicates the high purity of these
samples. Occasionally, when checking PL from multiple locations under a microscope, regions
exhibiting PL from thicker (e.g. n + 1) or thinner (e.g. n - 1) iodide 2D LHPs could be observed.
These observations were rare in high-purity samples (Figure S10), and phase impurity was not
detected by powder X-ray diffraction (section 4). However, red-shifted emission attributed to edge
states66,71 (Figure S11) could be observed under air exposure, in agreement with the report from
Dou et al.66
4. Structural insights from powder X-ray diffraction
Initial structural characterization was performed using powder X-ray diffraction (PXRD) on finely
ground crystals (Figure 3a-c). All samples showed (111) peaks near 14° 2θ angle and characteristic
stacking peaks of 2D layered materials, which manifest as periodic (h00) diffraction peaks in the
51
PXRD pattern. Below the (111) peak, we observe either two, three, four, or five of the evenly
spaced (h00) reflections for samples with BA organic spacer and n = 1, 2, 3, and 4 respectively
5
1
(
Figure 3a). The increasing number of reflections in this region indicates a larger unit cell
parameter along the stacking direction, and this number can be used to confirm the value of n and
the purity of the samples. When the organic spacer is lengthened from four to five and six carbons,
the evenly spaced reflections shift to lower 2θ angles (Figure 3b), reflecting the larger spacing
distances in samples with longer alkylammonium chains. On the other hand, the substitution of A-
1
8
ACS Paragon Plus Environment

Page 19 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
site cation from MA to FA does not induce a significant change in the PXRD patterns as shown in
Figure 3c.
Using the positions of (h00) reflections, the stacking period of each iodide 2D LHP can be
calculated. The stacking periods of n = 1 BA-PbI, n = 2 BA-MAPbI, n = 3 BA-MAPbI, and n = 4
BA-MAPbI were deduced to be 1.39 nm, 1.96 nm, 2.60 nm, and 3.22 nm, respectively. This
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
incremental increase of ~0.6 nm corresponds to the thickness of an additional PbI octahedral layer
6
as n increases.18,51 Using the same approach, stacking periods of n = 2 PA-MAPbI and n = 2 HA-
MAPbI were found to be 2.07 nm and 2.25 nm, larger than that of n = 2 BA-MAPbI by 0.11 nm
and 0.29 nm, respectively. This non-stepwise increase as the alkylammonium chain length
increases is due to the different extents of interdigitation in the organic spacer layers as the organic
spacer length is changed, as determined by single crystal X-ray diffraction (SCXRD – see section
5
.4). When changing the A-site cation from MA to FA, the stacking period remained constant at
1.96 nm. This is not surprising, since MA and FA have similar ionic radii of 2.70 Å and 2.79 Å,
respectively.72
1
9
ACS Paragon Plus Environment

Chemistry of Materials
Page 20 of 61
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
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
Figure 3. Powder X-ray diffractograms of (a) iodide n-series, (b) iodide L-series, and (c) iodide
A-series. All peaks are labelled with their corresponding diffraction planes. Evenly spaced (h00)
peaks correspond to the periodicity in the stacking direction.
2
0
ACS Paragon Plus Environment

Page 21 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
In addition to the shifts in the stacking peaks, changing sample identities can lead to a similar
effect on the position of the (111) peak. As the period of the (111) planes is influenced by the unit
cell parameter in the stacking direction, increasing n results in the (111) peaks moving to lower 2θ
angles (Figure 3a). When the organic spacer molecule is changed from BA to PA and HA, 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
(
111) peak starts to split to two peaks of (111) and (11 ) reflections (Figure 3b). This split arises
1
from the fact that n = 2 BA-MAPbI, n = 2 PA-MAPbI, and n = 2 HA-MAPbI have different crystal
symmetries: n = 2 BA-MAPbI adopts orthorhombic Cmcm, but n = 2 PA-MAPbI and n = 2 HA-
MAPbI adopt monoclinic C2/c at room temperature. Similar to the effect on the stacking peaks,
replacing MA by FA has a marginal effect on the (111) peak position (Figure 3c).
Temperature-dependent PXRD below room temperature was then performed to find the phase
transitions of the samples. The data (Figure 4 and Figures S14-S20) show that (i) n = 2 BA-MAPbI,
n = 3 BA-MAPbI, and n = 4 BA-MAPbI have two phase transitions below room temperature; (ii)
n = 2 BA-FAPbI has one phase transition below room temperature; and (iii) n = 2 PA-MAPbI and
n = 2 HA-MAPbI have no phase transition below room temperature. The higher-temperature phase
transitions of n = 2 BA-MAPbI, n = 3 BA-MAPbI, and n = 4 BA-MAPbI fall between 280 K and
290 K, and the phase transition of n = 2 BA-FAPbI happens at a slightly lower temperature range
31
between 266 K and 280 K. Recent work from our group described these phase transitions as
partial melting transitions of the organic spacer layers, similar to what has been observed in n = 1
7
3
2
D LHPs. Using differential scanning calorimetry (DSC), we assigned the phase transition
temperatures to be 283 K for n = 2 BA-MAPbI, 282 K for n = 3 BA-MAPbI, 283 K for n = 4 BA-
MAPbI, and 268 K for n = 2 BA-FAPbI. We also showed that n = 2 PA-MAPbI and n = 2 HA-
MAPbI exhibit analogous phase transitions at 351 K and 369 K, respectively. These phase
transition temperatures agree with recently published data from Kanatzidis & co-workers.32,33 On
2
1
ACS Paragon Plus Environment

Chemistry of Materials
Page 22 of 61
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 other hand, the temperatures of the lower-temperature phase transitions of n = 2 BA-MAPbI,
n = 3 BA-MAPbI, and n = 4 BA-MAPbI have a decreasing trend with ~186-196 K for n = 2 BA-
MAPbI, ~156-176 K for n = 3 BA-MAPbI and ~136-146 K for n = 4 BA-MAPbI. Recent work by
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
33
Kanatzidis & co-workers has suggested that this lower-temperature phase transition corresponds
to a structural change to a triclinic structure and its behavior is similar to the transition from β-
phase to γ-phase in bulk MAPbI3.
Figure 4. Selected temperature-dependent powder X-ray diffractograms of n = 2 BA-MAPbI
showing two phase transitions at 280-290 K and 186-196 K. The full-range and selected data sets
of n = 2 BA-MAPbI, n = 3 BA-MAPbI, n = 4 BA-MAPbI, n = 2 PA-MAPbI, n = 2 HA-MAPbI,
and n = 2 BA-FAPbI are shown in Figure S14-20.
5
. Analysis of iodide 2D LHP crystal structures
Crystal structures of all iodide 2D LHP samples were acquired by performing SCXRD on single
crystals obtained directly from their mother liquors. Using the phase transitions temperatures
identified by powder XRD and DSC discussed in section 4, we collected the crystal structures by
mounting crystals at 300 K. Then, the temperature was raised, if needed, to a temperature above
2
2
ACS Paragon Plus Environment

Page 23 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
the highest phase transition temperature before the temperature was lowered to below a phase
transition. We were able to observe the structural transformation at and above room temperature
in all samples. However, we did not observe the phase transitions at ~136-196 K of n = 2 BA-
MAPbI, n = 3 BA-MAPbI, and n = 4 BA-MAPbI by SCXRD even when the temperature was
lowered to 100 K. Useful crystal models could not be obtained due to poor crystal qualities at
lower temperature caused by the phase changes and possible X-ray damages. The space groups
and phase transitions of all iodide 2D LHP samples are summarized in Figure 5.
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
Figure 5. Summary of space groups and phase transitions identified by temperature-dependent
powder X-ray diffraction (PXRD), temperature-dependent photoluminescence (PL), differential
scanning calorimetry (DSC), and single-crystal X-ray diffraction (SCXRD).
When refining crystal structures, we note that there is a space group choice dilemma between a
centrosymmetric and non-centrosymmetric space group. For example, Cmcm and Ama2 space
groups have the same systematic absences of reflections. However, we decided to present the
2
3
ACS Paragon Plus Environment

Chemistry of Materials
Page 24 of 61
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
centrosymmetric space group over the non-centrosymmetric space group in this work because the
E-value statistics and the combined figure of merit (CFOM) values suggest the centrosymmetric
space group. Moreover, we found that there is not significant improvement in the final refinement
model for the non-centrosymmetric space group with a ~50% inversion twinning. In addition,
when it is unclear between a centrosymmetric and non-centrosymmetric space group, the
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
1
centrosymmetric space group is usually recommended. Choosing the centrosymmetric space
group may lead to a refinement with a disorder model, which suggests that the crystal consists of
3
1
two or multiple undistinguished structures. Nevertheless, the non-centrosymmetric space group
is still possible if the presence of a polar space group is confirmed by other techniques.
All of the refined crystal structures show that every iodide 2D LHP sample consists of stacks of
inorganic slabs separated by bilayers of organic molecular spacers. In the inorganic sublattice, PbI6
octahedra share their corners in both lateral and stacking directions, and the number of PbI6
octahedra layers increase with n. The inorganic sublattice shows distortion with respect to the
perfect cubic unit cell of the iodide 3D LHP, and adjacent PbI octahedra have out-of-phase out-
6
of-plane tilting with respect to each other. In all samples, adjacent inorganic slabs are staggered,
and all samples are categorized as members of Ruddlesden-Popper perovskites.3,51

5.1 Structural comparison of iodide 2D LHP crystals with different n
At 300 K, n = 2, 3, and 4 BA-MAPbI crystallizes in orthorhombic space groups Cmcm, Cmca, and
Cmcm, respectively (Figure 6). For n = 2 BA-MAPbI, all PbI octahedra are equivalent, and can
6
be related by symmetry operations. When the quantum-well thickness increases to n = 3 and 4,
two distinct types of PbI octahedra emerge: surface PbI octahedra, which are adjacent to the
6
6
organic spacer layer; and middle-layer PbI octahedra, which are surrounded on all sides by other
6
2
4
ACS Paragon Plus Environment

Page 25 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
PbI octahedra. Using this terminology, all PbI octahedra in n = 2 iodide 2D LHPs can be treated
6
6
as surface PbI octahedra. The surface and the middle-layer PbI octahedra differ by their out-of-
6
6
plane tilting angles. The out-of-plane tilting angles of the Pb-I bonds on the surfaces are between
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
9
4
.5° and 10°; and, for the non-surface Pb-I bonds, these numbers decreases to ~6-7° in n = 2 and
samples, and 0° in n = 3 sample (Figure S22). Although the middle-layer of n = 3 BA-MAPbI
does not have any out-of-plane tilting, it still shows distortion in the form of in-plane, out-of-phase
rotation between adjacent PbI octahedra. The difference in out-of-plane tilting angles in the
6
surface PbI octahedra and the middle-layer PbI octahedra reflects the difference in the interaction
6
6
between the PbI octahedral framework with the smaller MA cation and the larger organic spacer
6
cations.
Table 1. Crystal data and structure refinement for n = 2 BA-MAPbI, n = 3 BA-MAPbI, and n =
4
BA-MAPbI
n = 2 BA-MAPbI
Pb
1483.04
n = 3 BA-MAPbI
n = 4 BA-MAPbI
Chemical formula
C
9
H
30
I
7
N
3
2
C
10
H
36
I
10
N
4
Pb
3
C
11
H
42
I
13
N
5
Pb
4
M
r
2103.00
2722.95
Temperature (K)
Crystal system
Space group
a (Å)
300
250
300
250
300
250
Orthorhombic
Cmcm
39.277 (4),
8.9317 (8)
8.8533 (8)
90
Triclinic
P-1
Orthorhombic
Cmca
Triclinic
P-1
Orthorhombic
Cmcm
64.305 (5)
8.9327 (6)
8.8835 (6)
90
Triclinic
P-1
8.7432 (10)
9.0448 (10)
19.669 (2)
102.905 (3)
95.093 (3)
91.393 (3)
1508.6 (3)
2
51.963 (8)
8.9349 (13)
8.8648 (13)
90
8.7799 (13)
9.0236 (14)
25.436 (4)
90.030 (3)
95.685 (3)
91.316 (3)
2004.8 (5)
2
8.8047 (11)
8.9991 (12)
31.917 (4)
97.912 (2)
93.216 (2)
91.222 (2)
2499.9 (6)
2
b (Å)
c (Å)
α (°)
β (°)
90
90
90
γ (°)
90
90
90
V (Å3)
3105.9 (5)
4
4115.8 (10)
4
5102.9 (6)
4
Z
-
1
µ (mm )
17.79
18.31
19.75
20.27
21.04
21.47
Crystal size (mm)
0.10 × 0.06 × 0.10 × 0.06 0.18 × 0.09 × 0.18 × 0.09 0.12 × 0.08 × 0.12 × 0.08
0.02 × 0.02 0.06 × 0.06 0.02 × 0.02
2
5
ACS Paragon Plus Environment

Chemistry of Materials
Page 26 of 61
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
No. of measured
reflections
24804
1506
1100
25836
27562
1859
1428
28054
7091
5876
34882
2459
1677
34039
8853
6378
No. of independent
reflections
5344
4101
No. of observed [I >
2
(I)] reflections
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
R
int
0.053
0.596
0.052
0.597
0.045
0.597
0.042
0.597
0.049
0.597
0.044
0.596
(
sin θ/λ)max (Å-1)
2
2
R[F > 2(F )],
0.039, 0.101,
1.07
0.057,
0.138, 1.08
0.042, 0.115,
1.09
0.056,
0.136, 1.09
0.041, 0.109,
1.05
0.062,
0.166, 1.07
2
wR(F ), S
No. of reflections
No. of parameters
No. of restraints
1506
92
5344
222
1859
109
7091
276
2459
133
8853
355
71
126
68
124
84
113
-3
Δmax, Δ min (e Å )
1.32, -0.93
2.18, -1.56
0.62, -2.00
2.21, -4.06
1.11, -1.98
2.57, -3.98
In addition to the out-of-plane tilting, the degree of distortion can be quantified by bond length
distortion74–77 and bond angle variance.74,78 The bond length distortion (Δ푑) of an octahedron is
defined as
2
1
푑푛 ― 푑
푑
Δ푑 = (6)
∑
[
]
(Equation 1)
where 푑 is the individual Pb-I bond distance, and 푑 is the mean Pb-I bond distance. Additionally,
푛
2
the bond angle variance (휎 ) of an octahedron can be calculated by
(
휃푖 ― 90)2
휎2 = ∑12
(Equation 2)
푖 = 1
11
where 휃 is the individual I-Pb-I angle. Nonzero values of these two parameters indicate distortion
푖
from a perfect octahedron, which is characterized by six equal bond lengths and 90° bond angles.
2
Across n = 2, 3, and 4 BA-MAPbI, both Δ푑 and 휎 are larger for surface octahedra than in middle-
layer octahedra (Figure 6d left panel), in agreement with the extent of out-of-plane tilting discussed
above.
2
6
ACS Paragon Plus Environment

Page 27 of 61
Chemistry of Materials
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
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
Figure 6. Analysis of crystal structures for n = 2, 3, and 4 BA-MAPbI. (a) Structural phase
transition of n = 2 BA-MAPbI from orthorhombic Cmcm at 300 K to triclinic P-1 at 250 K. (b)
Structural phase transition of n = 3 BA-MAPbI from orthorhombic Cmca at 300 K to triclinic P-1
at 250 K. (c) Structural phase transition of n = 4 BA-MAPbI from orthorhombic Cmcm at 300 K
to triclinic P-1 at 250 K. Structurally similar PbI octahedra are labelled by blue, orange, red, and
6
green colors. (Disorder is omitted in all illustrations for clarity.) (d) Bond length distortion (훥푑,
2
top) and bond angle variance (휎 , bottom) of each type of PbI octahedra of n = 2, 3, and 4 BA-
6
MAPbI at 300 K (left) and 250 K (right). (e) Schematic diagram showing the orbital overlap of a
terminal Pb-I bond and a linking Pb-I bond. (f) Plot of axial Pb-I bond length starting from the top
to the bottom of the inorganic quantum wells of n = 2 BA-MAPbI (blue), n = 3 BA-MAPbI (green),
and n = 4 BA-MAPbI (red). The number on the plot indicates the order of the bond starting from
the top of the inorganic quantum wells. (g) Evolution of Pb-(µ-I)-Pb angle in axial (left) and
equatorial (right) directions of n = 2 BA-MAPbI (blue), n = 3 BA-MAPbI (green), and n = 4 BA-
MAPbI (red).
2
7
ACS Paragon Plus Environment

Chemistry of Materials
Page 28 of 61
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 closer inspection of the bond lengths across samples with different n also sheds light on how the
bonding orbitals of Pb and I interact. In iodide 2D LHPs, Pb-I bonds can be classified as terminal-
axial (T-Ax) Pb-I bonds, linking-axial (L-Ax) Pb-I bonds, and linking-equatorial (L-Eq) Pb-I
bonds (Figure S23), where axial and equatorial directions are defined as out-of-plane and in-plane
directions, respectively. A general trend observed across all iodide 2D LHP samples is that the T-
Ax Pb-I bond is always the shortest bond among the three types. This can be explained by noting
that the orbitals of a terminal I atom are only interacting with the orbitals from one Pb atom,
whereas those of a linking I atom are shared with the orbitals from two Pb atoms (Figure 6e).
Hence, orbital overlap is greatest in the T-Ax Pb-I bond, and this bond is the shortest and strongest.
Going from the top to the bottom of the inorganic slab, the axial Pb-I bond lengths exhibit a zigzag
pattern with a reflection plane at the middle of the inorganic slab (Figure 6f). This behavior 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
explained by noting that the two Ax I atoms in each PbI octahedron are interacting with the same
6
orbitals of the Pb atom (Figure 6e). Thus, stronger orbital overlap in one Ax Pb-I bond will make
the orbital overlap of the other Ax Pb-I bond weaker, resulting in two Ax Pb-I bonds with different
bond lengths. This effect is called the trans effect and is commonly found in transition metal
7
9
complexes. This alternation of bond length is not unique to 2D LHPs and has been previously
8
0
observed in both cubic and tetragonal phases of MAPbI . The Pb-I bond lengths in cubic and
3
tetragonal MAPbI alternate between 3.21 Å and 3.10 Å for the former and between 3.20 Å and
3
3
.13 Å for the latter. These Pb-I bond lengths are also similar to the Pb-I bond lengths of 3.19 Å
and 3.12 Å found in the middle-layer PbI octahedra. However, an exception of bond-length
6
alternation was observed in the middle-layer PbI octahedra of n = 3 BA-MAPbI, which has equal
6
Ax Pb-I bond lengths due to symmetry. On the other hand, L-Eq Pb-I bond lengths have a small
2
8
ACS Paragon Plus Environment

Page 29 of 61
Chemistry of Materials
1
2
3
4
5
6
7
8
9
variation and are 3.16 0.02 Å long, which is similar to those found in cubic and tetragonal
±
MAPbI3.80
When n = 2, 3, and 4 BA-MAPbI are cooled down to 250 K, the 2D LHPs retain their staggered
geometry but undergo phase transitions to a lower symmetry space group of triclinic P-1 (Figure
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
6
a-c). As a result, both surface PbI octahedra and middle-layer PbI octahedra bifurcate into two
6 6
families, each with different distortions yet similar magnitudes (Figure 6d right panel). The phase
transformation is accompanied by a significant change in the Pb-(µ-I)-Pb bond angles (Figure 6g):
(i) the Pb-(µ-I)-Pb bond angles in axial direction reduce to ~160°, leading to significant increases
in out-of-plane tilting across the three samples; and (ii) the Pb-(µ-I)-Pb bond angles in equatorial
direction with initial values of ~164-172° separate into two groups with lower bond angles of
~153-160° and higher bond angles of ~175-180°. This alteration in bond angles suggests a higher
degree of distortion with respect to the ideal perovskite structure with all Pb-(µ-I)-Pb bond angles
being 180°. In addition, the middle-layer PbI octahedra of n = 3 BA-MAPbI in this phase shows
6
an out-of-plane tilting, unlike in the higher-temperature phase.
2
9
ACS Paragon Plus Environment