Colloidal CdSe Nanoplatelets, A Model for Surface Chemistry/Optoelectronic Property Relations in Semiconductor Nanocrystals

Article abstract of DOI:10.1021/jacs.8b07566

While the surface termination of quasi-spherical metal chalcogenide nanocrystals or quantum dots has been widely investigated, it remains unclear whether the ensuing surface chemistry models apply to similar nanocrystals with anisotropic shapes. In this work, we report on the surface-chemistry of 2D CdSe nanoplatelets, where we make use of an improved synthesis strategy that yields stable and aggregation free nanoplatelet suspensions with a photoluminescence quantum yield as high as 55%. We confirm that such nanoplatelets are enriched in Cd and, by means of ¹H nuclear magnetic resonance spectroscopy, we show that the Cd-rich surface is terminated by X-type carboxylate ligands. Not unlike CdSe quantum dots (QDs), entire cadmium carboxylate entities can be displaced by the addition of amines, and the desorption isotherm points toward a considerable binding site heterogeneity. Moreover, we find that even the slightest displacement of cadmium carboxylate ligands quenches the nanoplatelet photoluminescence. These experimental findings are further confirmed by density functional theory (DFT) calculations on a 5 monolayer model CdSe nanoplatelet. These simulations show that the most labile ligands are located in the vicinity of facet edges, and that the displacement of ligands from such edge sites creates midgap states that can account for the observed photoluminescence quenching. Next to extending surface chemistry insights from colloidal QDs to nanoplatelets, this work indicates that CdSe nanoplatelets constitute a unique nanocrystal model system to establish a comprehensive description of midgap trap states, which includes their structural, chemical, and electronic properties.

Full text of DOI:10.1021/jacs.8b07566

Subscriber access provided by University of Sunderland
Article
Colloidal CdSe Nanoplatelets, a Model for Surface Chemistry/
Opto-Electronic Property Relations in Semiconductor Nanocrystals
Shalini Singh, Renu Tomar, Stephanie ten Brinck, Jonathan De
Roo, Pieter Geiregat, Jose C. Martins, Ivan Infante, and Zeger Hens
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b07566 • Publication Date (Web): 26 Sep 2018
Downloaded from http://pubs.acs.org on September 26, 2018
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 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Colloidal CdSe Nanoplatelets, a Model for Surface Chemistry/Opto-
Electronic Property Relations in Semiconductor Nanocrystals
¶
†
*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
‡
‡
‡
‡
∥
∥
∥
Shalini Singh, Renu Tomar, Stephanie ten Brinck, Jonathan De Roo, ∥ Pieter Geiregat, José C. Martins,
¶
‡
∥
Ivan Infante and Zeger Hens
‡ Physics and Chemistry of Nanostructures, Ghent University, 9000 Ghent, Belgium
∥ Center for Nano and Biophotonics, Ghent University, 9000 Ghent, Belgium
¶ Department of Theoretical Chemistry, Vrije Universiteit, 1081 HV Amsterdam, Netherlands
† NMR and Structural Analysis Unit, Ghent University, 9000 Gent, Belgium
Supporting Information Placeholder
ABSTRACT: While the surface termination of quasiꢀspherical metal chalcogenide nanocrystals or quantum dots has been widely
investigated, it remains unclear whether the ensuing surface chemistry models apply to similar nanocrystals with anisotropic
shapes. In this work, we report on the surfaceꢀchemistry of 2D CdSe nanoplatelets, where we make use of an improved synthesis
strategy that yields stable and aggregation free nanoplatelet suspensions with a photoluminescence quantum yield (PLQY) as high
1
as 55%. We confirm that such nanoplatelets are enriched in Cd and, by means of H nuclear magnetic resonance spectroscopy
(NMR), we show that the Cd rich surface is terminated by Xꢀtype carboxylate ligands. Not unlike CdSe quantum dots (QDs), entire
cadmium carboxylate entities can be displaced by the addition of amines, and the desorption isotherm points towards a considerable
binding site heterogeneity. Moreover, we find that even the slightest displacement of cadmium carboxylate ligands quenches the
nanoplatelet photoluminescence. These experimental findings are further confirmed by density functional theory (DFT) calculations
on a 5 monolayer model CdSe nanoplatelet. These simulations show that the most labile ligands are located in the vicinity of facet
edges, and that the displacement of ligands from such edge sites creates midgap states that can account for the observed photolumiꢀ
nescence quenching. Next to extending surface chemistry insights from colloidal QDs to nanoplatelets, this work indicates that
CdSe nanoplatelets constitute a unique nanocrystal model system to establish a comprehensive description of midꢀgap trap states,
which includes their structural, chemical and electronic properties.
ꢀ INTRODUCTION
reversed by the reꢀadsorption of these Zꢀtype ligands. More
recently, DFT studies indicated that the loss of PLQY upon
MX2 displacement is linked to the formation of underꢀ
coordinated surface Se.11 In particular, the formation of 2ꢀ
coordinated Se results in localized states that can behave as
traps for electrons or holes, depending on the Fermi level
position. This model seems generally applicable to IIꢀVI
QDs.12 In the case of CdTe QDs, charge carrier trapping was
linked to the presence of 2ꢀcoordinated Te,13 whereas similarly
2ꢀcoordinated Te were shown to create optically active localꢀ
ized states that make possible nearly thresholdless optical
amplification in HgTe QDs.14
The binding of ligands to the surface of colloidal semiconducꢀ
tor nanocrystals has attracted significant research interest
because ligands are essential to optimize, tune, and tailor the
physical and chemical properties of these functional nanoꢀ
materials.1ꢀ3 The multitude of possible surface chemistries was
recently classified based on the nature of the ligandꢀ
nanocrystal bond.4 Here, the starting point was the covalent
bond classification, which denotes ligands as L, X or Z deꢀ
pending on the number of electrons the neutral ligand provides
to the ligandꢀnanocrystal bond.5 Most importantly, this classiꢀ
fication led to a better understanding of nanocrystalꢀligand
binding, postꢀsynthesis ligand exchange reactions and the
Research that relates insight in surface chemistry to optoꢀ
electronic properties has been mainly focused on QDs. If
anything, it is tacitly assumed that the concepts used to deꢀ
scribe this relation for QDs apply equally to oneꢀdimensional
nanorods, twoꢀdimensional nanoplatelets and other nanoscale
objects.15 However, QDs are nanometerꢀsize crystallites that
have strongly curved surfaces composed of crystal facets that
often measure a few square nanometers only. Hence, a sizable
fraction of surface atoms will occupy sites at facet edges or
vertices. This differs markedly from colloidal 2ꢀD nanocrysꢀ
tals. CdSe nanoplatelets, for example, can be synthesized as
structures that are just a few monolayers thick, but with atomiꢀ
cally flat top and bottom surfaces that can measure several
hundreds of square nanometer.16ꢀ18 Several authors have reꢀ
ported on a pronounced interplay between the ligand capping
interplay between the composition of the core nanocrystal and
the ligand capping.4,
Moreover, the same approach was
6ꢀ9
applied later to describe the surface termination of metal oxide
nanocrystals.10
For the case of colloidal metal chalcogenide nanocrystals or
quantum dots (QDs), a comprehensive picture has emerged
over the past 10 years on the relation between surface pasꢀ
sivation and photoluminescence quantum yield (PLQY).4, 11 It
was shown that the surface of divalent metal chalcogenides
such as CdSe or PbS is metal rich and passivated by 2 equivaꢀ
lents of monovalent Xꢀtype ligands, such as carboxylates,
phosphonates or hydroxides.3 In the case of CdSe QDs, it was
demonstrated that displacing such MX2 salts from the CdSe
surface results in a marked drop of the PLQY, which could be
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 17
Synthesis of CdSe nanoplatelets. In a typical synthesis, cadmiꢀ
of nanoplatelets and their optical properties. For example, the
absorption spectra of CdSe nanoplatelets were shown to be
highly sensitive to changes in surface ligation.15, 19 In addition,
in spite of the large surface area, nanoplatelets can have a
PLQY of up to 40%, even without overgrowth with a large
bandꢀgap shell.20ꢀ24 On the other hand, research on CdSe nanoꢀ
platelets is often hampered by the limited colloidal stability of
nanoplatelet dispersions and the unwanted stacking of nanoꢀ
platelets in multiplatelet bundles,25, 26 two aspects that may be
linked to the specific way ligands bind and pack to the nanoꢀ
platelet surface. This combination of highly promising properꢀ
ties and cumbersome processing calls for an inꢀdepth study of
the nanoplatelet surface chemistry. Here, the question stands
out as to whether the concepts developed for highly curved,
multifaceted 0D QDs can be transferred to nanoplatelets,
which are terminated solely by atomically flat (100) interfaces.
um myristate (0.340 g, 0.6 mmol) and 25 mL of ODE were introꢀ
duced in a threeꢀneck flask and degassed under vacuum at room
temperature. The solution was then heated to 250 °C under an
inert atmosphere. At 250 °C, 0.024 g (0.3 mmol) of Se dispersed
in 1 mL of ODE were swiftly injected into the solution. A minute
later, 0.240 g of Cd(Ac)2 (0.9 mmol) were introduced. The reacꢀ
tion was allowed to proceed for 6 min at 250 °C, after which 1.5ꢀ2
mL of Cdꢀoleate, preꢀheated at 100 °C, was injected. To stop the
reaction, the heating mantle was removed immediately after this
second injection and the solution was allowed to cool down to
50 °C. The nanoplatelets were washed thoroughly by the purificaꢀ
tion techniques as described in detail in the Supporting Inforꢀ
mation Section S1. The nanoplatelets were then dispersed in
hexane for further use.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Electron Microscopy. Bright field transmission electron microsꢀ
copy (TEM) images and selected area electron diffraction (SAED)
measurements were made using a Csꢀcorrected JEOL 2200FS
TEM.
In this study, we present a combined experimental and theoretꢀ
ical analysis of the surface chemistry of 2D nanoplatelets. First,
we modified the classical synthesis protocol for 2ꢀD CdSe
nanoplatelets to achieve 5 monolayer (ML) nanoplatelets with
~55% quantum yield that form stable, aggregation free nanoꢀ
colloids. Using nuclear magnetic resonance (NMR) spectrosꢀ
copy, we demonstrate that these nanoplatelets are stabilized by
tightly bound cadmium carboxylates. We further analyzed the
distribution of binding sites on the surface by displacing these
Zꢀtype cadmium carboxylates through the addition of Lꢀtype
ligands. The displacement isotherm points towards binding
site heterogeneity on the nanoplatelet surface,27 where at least
two types of binding sites are required to model the displaceꢀ
ment equilibrium, not unlike CdSe QDs.27 This finding conꢀ
curs with DFT calculations on a 5 ML model nanoplatelet,
which enables us to localize the weaker binding sites at the
platelet edges and the stronger sites on the facets. In addition,
we show that the PL efficiency of nanoplatelets is highly senꢀ
sitive to the displacement of Zꢀtype ligands, where the removꢀ
al of a very small fraction of Zꢀtype ligands from edge sites
indeed leads to the appearance of localized states in the band
gap and quenching of PLQY. We thus conclude that in the
case of CdSe nanoplatelets, the combination of a wellꢀdefined
morphology, judicious ligand displacement experiments and
DFT simulations, yields in a complete picture of the surface
termination that combines structural, chemical and electronic
insights. As such, this work constitutes a next step towards an
atomistic understanding of colloidal semiconductor nanocrysꢀ
tals.
X-Ray Diffraction. Xꢀray diffraction (XRD) has been used to
analyze the crystal structure of the CdSe nanoplatelets. Diffractoꢀ
grams were recorded using a Thermo scientific ARL X’TRA
model.
Dynamic Light Scattering. For dynamic light scattering (DLS)
experiments, a Malvern Nano ZS was used in backscattering
mode (173°). The samples were suspended in nꢀhexane and measꢀ
urements were done at 25 °C.
Opto-Electronic Characterization. Absorption spectra were
taken with a PerkinElmer Lambda 950 spectrometer. The nanoꢀ
platelet concentration was obtained using the intrinsic absorption
coefficientꢀ ꢀꢁ,ꢂꢃꢃ = 5.24 × 105 cm−1 for 5 ML CdSe nanoplateꢀ
lets at 300 nm.28 Photoluminescence measurements were done on
an Edinburgh Instruments FLSP920 UV–Vis–NIR spectrofluoꢀ
rometer using a 450 W xenon lamp as the excitation source. An
excitation wavelength of 365 nm was used for all steadyꢀstate
emission spectra. The photoluminescence quantum yield was
measured using an integrating sphere.29
Rutherford Backscattering Spectrometry. The stoichiometry of
the CdSe nanoplatelets was determined by Rutherford backscatꢀ
tering spectrometry (RBS) using a 1.523 MeV 4He+ beam at a tilt
angle of 11°, with the detector positioned at a scattering angle of
170°. The atomic ratio between CdSe in the film was calculated
from the integrated backscattering intensity on Cd and Se via the
formula:
ꢋ
ꢄꢅꢆ ꢉꢅꢆ ꢊꢇꢈ
=
×
ꢀ EXPERIMENTAL SECTION
ꢋ
ꢄꢇꢈ ꢉꢇꢈ
ꢊꢅꢆ
Materials. Cadmium nitrate tetrahydrate Cd(NO3)2.4H2O, cadmiꢀ
um acetate tetrahydrate Cd(Ac)2·4H2O, CdO ( ≥ 99.99) were
purchased from SigmaꢀAldrich. Myristic acid, sodium hydroxide
NaOH, 1ꢀoctadecene (ODE, ≥90%) selenium powder (Se, −200
mesh, 99.999%) and oleic acid (OA, 90%) were purchased from
Alfa Aesar. Hexane, methanol and ethanol were purchased from
Fiers. Cyclohexaneꢀd12 (99.5% deuterated) was purchased from
Eurisoꢀtop.
Nuclear Magnetic Resonance Spectroscopy. Samples for NMR
spectroscopy were prepared by drying the purified QDs dispersion
with a nitrogen flow and reꢀdispersing the nanoplatelets in 500 ꢁL
deuterated cyclohexane. The details on purification of nanoplateꢀ
lets to make an NMR grade sample are given in Supporting Inꢀ
formation Section S1. The sample temperature was set to 298.15
K. NMR measurements were recorded on a Bruker Avance III
1
spectrometer operating at a H frequency of 500.13 MHz and
Synthesis of Cadmium Carboxylates. Cadmium myristate was
prepared as previously reported.16 For the synthesis of cadmium
oleate, CdO (0.642 g, 5 mmol) was mixed with 4.73 mL of OA
and 5 mL of ODE in a three neck round bottom flask. The mixture
was degassed at 110 °C for one hour, after which it was heated to
250 °C under nitrogen. When the mixture turned colorless, the
heating mantle was removed and the mixture was transferred to a
vial for further use.
equipped with a BBIꢀZ probe or on a Bruker Avance II spectromꢀ
1
eter operating at a H frequency of 500.13 MHz and equipped
1
with a TXOꢀZ probe. Quantitative H spectra were recorded with
a 20 s delay between scans to allow full relaxation of all NMR
signals. The quantification of the spectra was done by using the
Digital ERETIC method.30, 31 Diffusion measurements (2D diffuꢀ
sion ordered spectroscopy, DOSY) were performed using a douꢀ
bleꢀstimulated echo sequence for convection compensation and
ACS Paragon Plus Environment

Page 3 of 17
Journal of the American Chemical Society
with bipolar gradient pulses. Smoothed rectangle gradient pulse
indexed by means of the CdSe zincꢀblende crystal structure.
1
2
3
4
5
6
7
8
shapes were used throughout. The gradient strength was varied
linearly from 2 to 98% of the probe’s maximum value in 64 steps,
with the gradient pulse duration and diffusion delay optimized to
ensure a final attenuation of the signal in the final increment of
less than 15% relative to the first increment. The NMR spectra are
processed in IGOR Pro, version 6.32. To illustrate the diffusion
resonances from all the protons of bound ligand clearly in the 2ꢀD
This assignment is confirmed by the xꢀray diffractogram,
which shows diffraction peaks at the angles expected for zinc
blende CdSe (Supporting Information Section S2). Analyzing
200 nanoplatelets in such TEM images, we find that, for the
example shown, top surfaces measure on average 27.1×7.2
nm2, with standard deviations of 1.2 nm and 0.8 nm on width
and length, respectively (see Figure 1bꢀc). Other batches used
in this study contain nanoplatelets with comparable dimenꢀ
sions, see Supporting Information Section S2.
1
DOSY spectrum, 3 contour plots of different H shift range were
merged together as follows. The 1H shift range for these contour
plots are 1) ꢀ0.5 – 1.8 ppm, 2) 1.8ꢀ 4 ppm, 3) 4ꢀ7 ppm. The conꢀ
tours were intensified by two fold in 2) and 3). Twoꢀdimensional
1H–1H NOESY (nuclear Overhauser effect spectroscopy) spectra
were acquired using a standard pulse sequence from the Bruker
library, noesygpphpp. The NOESY mixing time was set to 300 ms
with 4096 data points set in the direct dimension and 512 data
points set in the indirect dimension. For 2D processing, before
Fourier transformation, the 2D spectra were multiplied with a
squared cosine bell function in both dimensions.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Density Functional Theory Calculations. For the DFT calculaꢀ
tions, we constructed a nanoplatelet model of a 5 monolayer CdSe
nanoplatelet by cutting a 3.07×3.07×1.25 nm3 nonstoichiometric
nanocrystal out of a bulk CdSe crystal. This model contains 295
Se atoms and 392 Cd atoms. To preserve charge neutrality, 194
chloride atoms were added to passivate the nanocrystal surface,
leading to a final chemical composition of [CdSe]295(CdCl2)97.
The nanoplatelet model exposes six (100) facets, terminated at the
vertices with either Cdꢀrich or Seꢀrich (111) facets. These strucꢀ
tures have been optimized in the gas phase using the PBE exꢀ
change correlation functional32 and a doubleꢀzeta basisꢀset augꢀ
mented with polarization functions.33 All calculations were carꢀ
ried out with the CP2k software package.34 Further details on the
models employed are given in the text.
ꢀ RESULTS AND DISCUSSION
We synthesized 5 ML CdSe platelets by the sequential injecꢀ
tion of elemental Se dispersed in 1ꢀoctadecene and cadmium
acetate in a hot reaction mixture containing cadmium
myristate, following the method originally proposed by Ithurꢀ
ria et al.17 However, we arrested the reaction by injecting
preheated cadmium oleate [Cd(OA)2] instead of oleic acid and
cooling down the entire reaction mixture. This adaptation was
inspired by the idea that excess oleic acid can only exchange
with ligands already bound to the CdSe surface – thus preservꢀ
ing the ligand surface concentration – whereas Cd(OA)2 is a
Zꢀtype ligand that can bind to CdSe surfaces and raise the
ligand surface concentration. We found that using cadmium
oleate instead of oleic acid as a reaction quencher resulted in
nanoplatelet dispersions with a markedly improved colloidal
stability. No platelet precipitation was observed over a time
period of six months, and absorbance spectra were free of any
background scattering. Both findings suggest that stopping the
nanoplatelet synthesis by injection of Cd(OA)2 is an effective
way to prevent the gradual formation of nanoplatelet stacks
and bundles, which give rise to light scattering and drive naꢀ
noplatelet precipitation over time. Possibly, this is due to the
formation of a more densely packed Cd(OA)2 ligand shell.
Figure 1. (a) Bright field TEM image of CdSe nanoplatelets with
SAED pattern in the inset; (b) Size histogram of the lateral dimenꢀ
sions of a batch of nanoplatelets as obtained from a set of TEM
overview images; (c) Absorbance and emission spectra of the
CdSe nanoplatelets; (d) Photographs of a vial containing CdSe
nanoplatelets in natural light and UV light.
The absorbance and emission spectra of CdSe nanoplatelet
dispersions as shown in Figure 1c are characteristic of 5
monolayer platelets. In line with literature reports, the absorpꢀ
tion spectra feature two wellꢀresolved transitions that correꢀ
spond to the heavyꢀhole exciton at 545 nm and the lightꢀhole
exciton at 516 nm.16, 35 The emission spectrum of these nanoꢀ
platelets consists of a single narrow band centered at 547 nm,
with a fullꢀwidth at halfꢀmaximum (FWHM) of 11.5 nm. Note
that the Stokes shift between the heavyꢀhole exciton absorbꢀ
ance and emission line amounts to 2 nm at the most. After the
first purification using ethanol/cyclohexane as the nonꢀ
solvent/solvent combination, a PLQY exceeding 50% is typiꢀ
cally obtained, which compares favorably with the best literaꢀ
ture reports on coreꢀonly CdSe nanoplatelets (<40%).20ꢀ22 By
any means, these efficiencies are surprisingly high as comꢀ
pared to core CdSe QDs. Whereas these tend to have similar
surface to volume ratios as few ML platelets, PLQYs typically
attain a mere 5ꢀ10% at best. This difference suggests that
CdSe nanoplatelets may have a nearly complete passivation of
the (100) outer surfaces, or are better at avoiding underꢀ
coordinated surface selenium, or other surface defects, from
forming when the surface termination is incomplete.
Figure 1a shows a low resolution bright field transmission
electron microscopy (TEM) image of CdSe nanoplatelets
synthesized according to the adapted protocol. Similar dark
field images are shown in Supporting Information Section S2.
Nanoplatelets lying mostly faceꢀdown on the TEM grid can be
readily identified. The selected area electron diffraction
(SAED) pattern (inset, Figure 1a) contains rings that can be
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 17
To study the surface passivation of asꢀsynthesized CdSe nanoꢀ if the top and side nanoplatelet (100) surfaces had a full excess
1
2
3
4
5
6
7
8
platelets, we first analyzed their composition using Rutherford
Backscattering Spectrometry (RBS). Figure 2a shows an RBS
spectrum of CdSe nanoplatelets that underwent up to 6 purifiꢀ
cation cycles using hexane and ethanol as the solvent and antiꢀ
solvent. We implemented such a thorough purification scheme
in order to obtain clean dispersions that are free of unbound
ligands. The RBS spectrum features two pronounced signals
that correspond to backscattering from Se and Cd on top of a
nearꢀ
Cd coverage (see Figure 2b). In addition, we found that the
PLQY dropped after each purification cycle, to reach about 10%
after six cycles, see Supporting Information Section S1. This
PLQY drop is probably linked to the lowerꢀthanꢀexpected Cd
excess. At least in the case of CdSe QDs, it is known that the
removal of Cdꢀcarboxylate ligands induces surface traps that
quench the photoluminescence.4, 12 On the other hand, it is
surprising to find the PLQY still at 10% ꢀ a value typical for
asꢀsynthesized CdSe QDs ꢀ after such extensive purification.
9
Figure 3a displays a representative 1H NMR spectrum of a
dispersion of 5 ML CdSe nanoplatelets in cyclohexaneꢀd12.
Apart from the single narrow resonance † at 1.38 ppm of cyꢀ
clohexane, the spectrum only features the broadened resoꢀ
nances characteristic of bound ligands. The sole occurrence of
broadened resonances suggests that the dispersion is well
purified and only contains bound ligands and solvent. All
resonances were assigned to protons from the oleyl chain as
indicated in the inset of Figure 3a; an assignment that agrees
with literature studies on oleate capped CdSe QDs. The charꢀ
acteristic resonance 5 of the alkene protons appears at 5.35
ppm and the methylene protons adjacent to the double bond at
2.055 ppm; the bulk of the methylene protons yields the inꢀ
tense resonance 3 at 1.320 ppm, and the resonance 6 of the
methyl protons is retrieved at 0.912 ppm. A striking feature of
the 1D 1H NMR spectrum of the CdSe nanoplatelets is the line
shape of the resonances, which consist of a relatively narrow
center part atop an excessively broadened pedestal. This broad
basis is notable in the region of the aliphatic protons, where
the upfield side of the resonance extends well beyond the CH3
resonance 6, and around the alkene proton resonance 5, whose
downfield side only reaches the baseline at ~7.5 ppm. Such
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. (a) Rutherford backscattering spectrum recorded on a
thin film of 5 ML CdSe nanoplatelets spun on a Si substrate. The
two signals correspond to backscattering from Se and Cd. The Z2
weighted ratio of the backscattering intensities yields a Cd:Se
ratio of 1.21. (b) Estimated Cd:Se ratio for 5 ML platelets as a
function of the platelet width and length. The markers indicate the
samples studied in this manuscript, where the filled marker repreꢀ
sents the RBS sample.
1
broad pedestals are absent in the case of 1D H NMR spectra
of oleateꢀcapped CdSe QDs.36 Possibly, this enhanced broadꢀ
ening reflects the reduced solvation of densely packed ligands
ꢀzero background signal. From the intensity of both backscatꢀ
tering signals, we obtain a Cd:Se ratio of 1.21±0.02, a number
that is somewhat smaller than the 1.255 ratio we would expect
on the flat (100) surfaces of the nanoplatelets
.
Figure 3. (a) 1D 1H NMR spectrum of 5 ML CdSe nanoplatelets in cyclohexaneꢀd12. Inset: Zoom on the resonance of the alkene protons at
∼5.35 ppm and the chemical structure of oleic acid, indicating the proton labeling used for assigning the resonances. (b) DOSY spectrum
of 5 ML CdSe nanoplatelets in cyclohexaneꢀd12. The dashed line indicates the diffusion coefficient as obtained from fitting the decay of the
intensity of the methyl protons x to a monoꢀexponential decay (see Supporting Information Section S4); (c) 2D NOESY spectrum of 5 ML
CdSe nanoplatelets, featuring the intense, negative nOe cross peaks that are characteristic of bound oleate.
Regardless of the interpretation, the particular lineꢀshape of
the oleyl proton resonances complicates the determination of
the ligand surface concentration from integrated resonance
intensities. Table 1 lists the surface concentration as estimated
for 5 different samples using the integrated intensity I6 and I5
of the methyl and the alkene resonance, respectively. To obꢀ
ACS Paragon Plus Environment

Page 5 of 17
Journal of the American Chemical Society
ands/nm2, 30% more than the corresponding selfꢀassembled
tain these numbers, we used a baseline correction as outlined
monolayer of thiols on flat gold surfaces.37 Experimental studꢀ
ies effectively found surface concentrations of 5.3 nmꢀ2 for the
16ꢀmercaptohexadecanoic acid/gold system,38 and comparable
numbers have been reported for other systems.39 Similarly,
molecular dynamics simulations showed that CdSe NCs pack
4.8 oleate ligands per nmꢀ2 on their (100) facets, even if the
oleate ligands are allowed to bind to sparsely occupied (111)
facets.40
1
2
3
4
5
6
7
8
in Supporting Information Section S3. As can be seen in Table
1, we found that the ratio I6/I5 showed a pronounced sample to
sample variation, ranging from 1.97 to 2.54, while always
exceeding the value of 1.5 that would be expected for oleyl
chains. As we will discuss later, addition of butylamine to a
nanoplatelet dispersion results in a partial displacement of
Cd(OA)2 from the nanoplatelet surface. As shown in Supportꢀ
ing Information Section S3, this leads to a progressive drop of
the ratio I6/I5 from a value of 2.36 down to 1.66 for the examꢀ
ple shown. This trend suggests that part of the signal intensity
contained in the broad pedestal of the bound alkene resonance
is lost in the background noise. In addition, any remaining
cadmium myristate on the nanoplatelet surface will also conꢀ
tribute to the methyl resonance 6, and not to the alkene resoꢀ
nance 5, and thus increase the intensity ratio I6/I5. In either
case, the implication is that the methyl resonance 6 will yield
the more reliable estimate of the ligand surface concentration.
We further analyzed to binding of carboxylate ligands to CdSe
nanoplatelets using 2D DOSY and NOESY. As can be seen in
Figure 3b, a DOSY spectrum recorded on the same dispersion
of CdSe nanoplatelets in cyclohexane indicates that the bound
ligand resonances come with a single diffusion coefficient that
is about 50 times smaller than the selfꢀdiffusion of cyclohexꢀ
ane. As shown in Supporting Information Section S4, the
bound ligand resonance intensities exhibit a monoꢀexponential
decay as a function of the square of the applied field gradient
strength, from which a diffusion coefficient of 34.8 µm2/s is
calculated. This value is about 12 times smaller than the diffuꢀ
sion coefficient of free oleic acid (see Supporting Information
Section S4) in the absence of nanoplatelets (D= 411 µm2/s).
Moreover, it compares well to the diffusion coefficient that
can be calculated for a randomly oriented diffusing sheet with
the given dimensions of the CdSe nanoplatelets (see Supportꢀ
ing Information Section S4).41
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Comparison of integrated peak intensities and ligand
surface concentration calculated from methyl proton and alkene
proton resonances along with the estimated Cd:Se ratio for the
given dimensions of the samples.
σmethyl (nm-2)
σalkene (nm-2)
Cd:Se*
Sample
I6/I5
1
2
3
1.96
2.28
2.18
4.74
5.54
5.47
3.75
4.16
3.64
1.255
1.245
1.250
In a separate set of experiments, we employed dynamic light
scattering (DLS) as a simpler alternative technique to measure
translational diffusion coefficient of the nanoplatelets (See
supporting information, Section S4). The diffusion coefficient
values extracted from single−exponential decay of the correlaꢀ
tion function of the scattered light intensity compares very
well again with the calculated diffusion coefficient of unꢀ
stacked randomly oriented nanoplatelet of the same dimension
passivated by a monolayer of oleate ligands (described in
section S4, supporting information). We thus conclude that
similar to CdSe QDs, CdSe nanoplatelets can be seen as nonꢀ
stoichiometric nanocrystals stabilized by tightly bound carꢀ
boxylate ligands (X) or, equivalently, as stoichiometric
nanostructures stabilized by tightly bound cadmium carboxꢀ
ylates (CdX2).6, 34 As can be seen in Figure 3c, all resonances
assigned to bound carboxylate protons feature strongly negaꢀ
tive NOE crossꢀpeaks. Hence, also in the case of nanoplatelets,
soꢀcalled negative nOe’s can be seen as a sensitive indicator
for ligandꢀnanocrystal interactions.
4
5
2.54
1.97
5.25
5.57
4.26
4.30
1.250
1.250
On average, integration of the methyl resonance 6 yields a
ligand surface concentration of 5.3 nmꢀ2 (see Supporting Inꢀ
formation Section S3 for details on surface concentration
calculations). A ligand surface concentration of 5.3 nmꢀ2 closeꢀ
ly matches the surface concentration of 5.4 nmꢀ2 of Cd2+ catiꢀ
ons at the CdSe (100) surface, indicating that the single excess
monolayer of Cd2+ is charge compensated by two layers of
carboxylates. On the other hand, such a surface concentration
is significantly larger than the 4 nmꢀ2 one expects at best based
on the molecular volume of a single oleic acid molecule in
liquid oleic acid. However, several studies have shown that
due to the strong surface curvature, nanocrystal surfaces can
accommodate more ligands than the corresponding macroꢀ
scopic flat surface. In the case of gold for example, nanocrysꢀ
tals are expected to accommodate up to 6 dodecanethiol ligꢀ
Figure 4. (a) Zoom on the alkene resonance 5 in the 1H NMR spectra of a CdSe nanoplatelet dispersion (nanoplatelet concentration: 11.7
µM, bound carboxylate concentration of 33.7 mM) to which BuNH2 has been added with concentrations as indicated. The resonances
assigned to bound and free cadmium oleate have been indicated. (b) The same, but zooming in on the resonances 1 (αꢀCH2) and 4 of the
oleyl chain. (c) Fractional surface coverage of oleate at the surface of CdSe nanoplatelets, calculated relative to the initial surface concenꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 17
tration, as a function of the BuNH2 concentration in solution. The full line represents a fit to a twoꢀsite adsorption model. (d) Photolumiꢀ
nescence quantum yield (PLQY) as a function of the BuNH2 concentration. Inset: photoluminescence spectrum at (bright red) the start and
(dark red) the end of the titration. The dashed line is a guide to the eye.
1
2
3
4
5
6
7
8
9
In the case of CdSe QDs, it is known that the photoluminesꢀ
cence efficiency is linked to the surface coverage by cadmium
carboxylates.4, 42ꢀ44 Moreover, such CdX2 ligands can be disꢀ
placed by Lꢀtype ligands such as alkylamines, which is an
equilibrium reaction that enables the ligand surface concentraꢀ
tion to be finely adjusted.27 Here, we used the NMR spectra
recorded before and after addition of butylamine (BuNH2) to a
dispersion of CdSe nanoplatelets in cyclohexane to analyze the
occurrence of amineꢀinduced CdX2 displacement, BuNH2
being an Lꢀtype ligand with a high displacement potency.42
Importantly, addition of BuNH2 did not affect the nanoplatelet
morphology, nor did it result in shifts of the excitonic features
in the absorption spectra (See supporting information, section
S5). BuNH2 addition did, however, lead to significant changes
in the NMR spectrum. Focusing on the resonances 4 and 5, it
can be seen that a second, narrow resonance develops at the
upfield side of the original, broad resonances (see Figure 4aꢀb)
during a BuNH2 titration. These additional resonances gain
intensity as the concentration of BuNH2 is increased, at the
expense of the intensity of the broad resonances. Similar
changes occurred with CdSe QDs, in which case the occurꢀ
rence of two resonances was assigned to the presence of bound
and displaced CdX2 exhibiting a slow displacement equilibriꢀ
um.27 Interestingly, while the αꢀCH2 resonance 1 of bound
CdX2 cannot be resolved due to excessive line broadening,
displaced CdX2 exhibits a well resolved αꢀCH2 resonance 1,
which appears as a triplet at around 2.15 ppm. Importantly, the
NOESY spectrum of the CdSe sample after the displacement
experiment features negative NOE crossꢀpeaks for oleate
chain protons and amine protons (See Supporting Information
Section S5). This indicates that displaced CdX2 still interact
with the platelet surface, most likely since displacement reꢀ
sults in a dynamic equilibrium between dissolved CdX2 coorꢀ
dinated by BuNH2 and surfaceꢀbound CdX2, similar to what
was demonstrated for CdSe QDs. We thus infer that the two
sets of oleate resonances reflect two pools of CdX2 – free and
bound – that exhibit a slow chemical exchange.
empty surface site. Moreover, L stands for BuNH2 and we
have explicitly introduced a 2ꢀequivalent displacement as
found in the case of CdSe QDs.27 Figure 4c represents a simuꢀ
lation of the displacement isotherm using this coupled, twoꢀ
site displacement equilibrium. Similar to the case of CdSe
QDs, this description captures the combination of a rapid
reduction of surface coverage at low BuNH2 with the persisꢀ
tence of bound CdX2 at high BuNH2 concentration. As comꢀ
pared to CdSe QDs, we find a similar equilibrium constant ꢑꢐ
for displacement from the weaker binding sites in the case of
CdSe nanoplatelets, yet the latter feature a notably smaller
fraction of weak binding sites (see Table 2). In combination
with the large surface concentration of CdX2, this yields a
surface concentration of strongly bound CdX2 of ≈ 3.4 nmꢒꢋ
in the case of nanoplatelets, a number that is markedly larger
than the ≈ 1.5 nmꢒꢋ we obtain for CdSe QDs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4d represents the evolution of the relative PLQY during
the different steps of a BuNH2 titration that was run in parallel
to the determination of the displacement isotherm. Importantꢀ
ly, we observe a sharp drop of the photoluminescence effiꢀ
ciency even during the initial stages of the titration were only
a minor fraction of surface ligands has been displaced. For
example, using the simulation of the experimental isotherm,
we estimate that the removal of a mere 1% of the surface
ligands reduces the relative PLQY by 25% in the first step of
the titration. Hence, opposite from CdSe QDs, there seems to
be no threshold surface coverage above which the PLQY can
be preserved.4 Moreover, this again indicates towards the large
fraction of strongly binding oleates on the surface which preꢀ
vents any change in the strain state of nanoplatelet crystal
structure.15, 44
Table 2. Initial surface concentration ꢓ of carboxylate ligands and
parameters used to simulate the BuNH2 displacement isotherm
using a 2 site model for the case of CdSe nanoplatelets and CdSe
QDs, including (ꢑꢐ, ꢑꢋ) the equilibrium constant for displacement
of weakly and strongly bound CdX2 and (ꢔꢏ the fraction of weakꢀ
ly bound CdX2. Data for CdSe QDs were taken from ref 27
.
We used the intensity of the αꢀCH2 resonance 1 of displaced
CdX2 to quantify the fraction of the originally surface bound
cadmium carboxylate at each step in a BuNH2 titration of a
CdSe nanoplatelet dispersion. This enabled us to determine a
displacement isotherm, which represents the CdX2 surface
coverage as a function of the BuNH2 concentration. As can be
seen in Figure 4c, this isotherm exhibits a pronounced drop in
surface coverage at low BuNH2 concentrations, in combinaꢀ
tion with a persistent surface coverage of 50ꢀ60% up to
BuNH2 concentrations of 0.8 mol/L. We analyze this isotherm
with the twoꢀsite model introduced by Drijvers et al to account
for binding site heterogeneity in the case of quasiꢀspherical
CdSe nanocrystals.27 This model assumes a coupled equilibriꢀ
um, where BuNH2 can displace CdX2 from two sets of binding
sites, a first with a low and a second with a high binding enerꢀ
gy:
ꢕ ꢎꢖꢗꢒꢘ
)
System
ꢙ
ꢚꢛ
ꢚꢘ
Platelets
QDs
5.4
3.2
0.375
0.525
3.0
3.0
0.002
0.01
ꢌ
ꢌ
ꢍꢎ
ꢏ
ꢌ
ꢍꢎ ꢏ
CdSe CdXꢋ ꢐ + 2L ⇌ CdSe ∘ ꢐ + LꢋCdXꢋ
(1)
ꢍꢎ ꢍꢎ ꢏ
CdSe CdXꢋ ꢋ + 2L ⇌ CdSe ∘ ꢋ + LꢋCdXꢋ
ꢏ
ꢌ
ꢌ
ꢍ
Here, CdSe represents the stoichiometric nanocrystal core,
ꢎ ꢏ
ꢎ
ꢏ
CdXꢋ a surface bound cadmium carboxylate, and ∘ an
ACS Paragon Plus Environment

Page 7 of 17
Journal of the American Chemical Society
um constant between QDs and nanoplatelets may originate
1
2
3
4
5
6
7
8
from the way the isotherm is measured. In both cases, we used
purified dispersions. This could mean that ligands with a bindꢀ
ing energy smaller than a certain threshold have been stripped,
which would create an identical starting point for both titraꢀ
tions. In this respect, the larger fraction of strongly binding
ligands in the case of CdSe nanoplatelets may explain why,
even after 6 purification cycles, the ligand surface density
exceeds 5 nmꢀ2.
Given the indications that BuNH2 strips Cd oleate ligands
from sites close to the platelet edges and concomitantly
quenches the photoluminescence efficiency, we analyzed in a
second step the electronic structure of the model nanoplatelet
at different ligand surface concentrations. Since the model
nanoplatelet offers large and flat (100) top and side planes,
many different patterns of successive ligand displacement are
possible, which can all result into different outcomes. To
minimize this effect, we displaced ligands only from the edge
sites as these are the most labile according to the above analyꢀ
sis. We then followed a twoꢀstep displacement, where we
reduced the surface concentration on the top (100) facet from
5.50 ligands/nmꢀ2 first to 4.90 nmꢀ2 and then to 4.20 nmꢀ2. At
each removal, we performed a structural relaxation and
checked for the presence of diꢀcoordinated Se, as these types
of surface atoms are the likely origin of midꢀgap states.11
9
10
11
12
13
14
15
16
Figure 5. The exposed top and side facet of with the computed
total displacement (desorption and coordination) energies in
kJ/mol calculated for different binding sites.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To better understand the passivation of the nanoplatelets by
ligands and the concomitant electronic structure of nanoplateꢀ
lets, we performed a density functional theory (DFT) analysis.
As shown in Figure 5, we built a 5 ML model platelet with a
3.07×3.07 nm (100) top facet and a surface fully passivated
with CdCl2. Using CdCl2 instead of Cd oleate as a Zꢀtype
ligand limits the computational effort and avoids the additional
degrees of freedom that come with the different possible bindꢀ
ing modes of carboxylates to Cd surface sites. We focused
first of all on the displacement of a single Zꢀtype ligand by
BuNH2 according to the equilibrium (1) and calculated the
displacement energy for different locations on the nanoplatelet
surface. Figure 5 represents the thus obtained displacement
energies. The small size of the nanoplatelet top surface makes
that most of the lattice positions remain relatively close to
edges or vertices. Nevertheless, a qualitative trend shows up,
where the most labile Zꢀtype ligands are found at the edges
and vertices of the platelet. More quantitatively, we found an
average displacement energy of 18.1 kJ/mol for the binding
sites within 2 atomic lines of a (100) edge, and 74.7 kJ/mol for
the binding sites further away from these edges. Clearly, this
result confirms that the picture of binding site heterogeneity
put forward for CdSe QDs applies to CdSe nanoplatelets as
well, and indicates that the weak bindings sites are mostly
related to the nanoplatelet edges.
In Figure 6, we show the electronic structures of the model
nanoplatelet with a pristine surface composition (left, model
1), followed by the nanoplatelet with 10 (center, model 2) and
22 (right, model 3) CdCl2 ligands removed. Note that in the
case of a 27×7 nm nanoplatelet, 10 and 22 ligands this would
amount to about 0.25 and 0.5% of the ligand surface coverage,
numbers that agree with the ligand displacement during the
initial stage of the BuNH2 titration. Each line in the figure
represents one molecular orbital (MO), and the colored secꢀ
tions measure the contribution of different atom types to this
MO. Each plot is centered around its Fermi level to allow for a
more straightforward comparison between the different modꢀ
els. One readily sees that model 1 features a band gap free of
midgap states. At ~0.8 eV, this band gap is quite small and
rather different from the experimental one. This is a known
problem of density functional theory within the generalized
gradient approximation, like the PBE employed here. Espeꢀ
cially for CdSe, such an approach tends to largely underestiꢀ
mate the bandgap energy. At the same time, however, this
method allows to tackle large systems, for which it provides
accurate geometrical structures and energetics, and correct
composition of the MOs in terms of the constituting atomic
orbitals.12
At this point, it is interesting to compare the simulation results
with the experimental findings on ligand displacement of
CdSe QDs and nanoplatelets. First, starting from the idea that
weak binding sites are related to surface Cd within the first or
the second atomic line from an edge, we estimate that a 7×27
nm 5 ML platelet will have 32.5% weak and 67.5% strong
ligands (see Supporting Information, Section S6). This number
comes close to the estimated fraction α=0.375 of weak ligands
we obtained from simulating the displacement isotherm (see
Table 2); a point that supports the link between weak binding
sites and nanocrystal edges. Second, the average displacement
energy of the weakest half of the binding sites amounted to ꢀ3
kJ/mol in the case of CdSe QDs, whereas we obtain 18.1
kJ/mol in the case of CdSe nanoplatelets. While this merely
reflects the smaller number of edge and vertex sites, it also
implies that the correspondence of the displacement equilibriꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 17
Zꢀtype ligands from the edges will readily generates midꢀgap
1
2
3
4
states that will reduce the efficiency of radiative recombinaꢀ
tion. Clearly, this conclusion agrees with our experimental
findings where the initial ligand stripping – attributed to CdX2
removal from edge or nearꢀedge sites – concurs with an imꢀ
mediate loss of photoluminescence efficiency.
5
6
7
8
9
Understanding the structural and chemical aspects of ligand
binding and elucidating the impact of ligand binding on elecꢀ
tronic structure are long standing goals in nanocrystal research.
In this respect, CdSe nanoplatelets appear as a unique model
system to study the properties of surface defects. Mostly due
to the straightforward nanoplatelet morphology, weak binding
of cadmium carboxylates can be ascribed almost exclusively
to nearꢀedge sites; a conclusion that is supported by the comꢀ
bined experimental/theoretical approach to nanocrystal surface
chemistry implemented here. This localization of labile ligꢀ
ands leads to realistic predictions on the nature of the trap
states created by ligand displacement. Importantly, these trap
states have properties such as trapping times or characteristic
defect emission, that can be tested in future experiments. We
believe such studies can provide the atomically precise deꢀ
scription of trap states needed to support the wide range of
optical, magnetic or charge transport properties of nanocrysꢀ
tals and nanocrystals assemblies assigned to trap states.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ CONCLUSION
By means of an improved protocol to quench the formation of
5 monolayer CdSe nanoplatelets we obtain nanoplatelet disꢀ
persions with longꢀterm colloidal stability and a high photoꢀ
luminescence quantum yield. Based on such dispersions, we
analyze the surface chemistry of CdSe nanoplatelets through a
combination of elemental analysis and solution NMR specꢀ
troscopy. Similar to CdSe QDs, we find that CdSe nanoplateꢀ
lets are Cdꢀrich crystallites, terminated by oleate ligands. Even
if solution NMR resonances exhibit a pronounced line broadꢀ
ening, ligand surface concentrations could be estimated on
average at 5.3 nmꢀ2. We show that addition of an Lꢀtype ligand
such as BuNH2 leads to cadmium carboxylate displacement,
where even a reduction of the surface coverage by 1% impacts
on the photoluminescence quantum yield. The displacement
isotherm points towards a heterogeneous set of binding sites,
which features a larger fraction of strong binding sites than
CdSe QDs. This finding is corroborated by DFT calculations
on a 5 monolayer model nanoplatelet, which predict signifiꢀ
cantly weaker binding at sites within 2 atomic lines from the
nanoplatelet edges. Moreover, we find that displacement of
ligands from these edge sites leads to the formation of mid gap
states that can be linked to the creation of diꢀcoordinated seleꢀ
nium. This localization of trap states at the platelet edges puts
forward CdSe nanoplatelets as a nanocrystal model system for
obtaining a comprehensive, atomistically precise description
of midꢀgap states.
Figure 6. Atomistic and electronic structure of the model nanoꢀ
platelet for which (left) no, (center) 10 and (right) 22 CdCl2 ligꢀ
ands were removed. The black arrows indicate the midgap states
that form upon removal of CdCl2 from the pristine nanoplatelet.
At the bottom, contour plots of the highest occupied molecular
orbitals of the three systems are shown.
Opposite from the pristine nanoplatelet model, 3 and 4 midgap
states appear in the band gap when CdCl2 ligands are removed
from facet edges in model 2 and 3, respectively. This concurs
with the respective formation of 3 and 4 diꢀcoordinate Se
atoms. To verify the relation between midgap states and diꢀ
coordinated surface Se, we analyzed the orbital composition of
these midgap states (black arrows lines in Figure 6) and found
that these states are localized on a single atom, or a small
group of atoms, and closely resemble uncoupled pꢀorbitals on
diꢀcoordinated Se (Figure 6, orbitals belonging to model 2 and
3). Note that the frontier orbitals in the pristine nanoplatelet
(Figure 6, orbitals belonging to model 1) shown in Figure 6
seem to present a facet specific localization on the (111) surꢀ
faces. Apparently, the larger dimension of the (100) surfaces
against the smaller (111) leads to a reduced orbital mixing
between the atoms lying on these facets. This was evident
when we projected the density of states on the (100) and (111)
facets (Supporting Information Section S7). Clearly, near the
valence band edge, MOs are dominated by atomic orbitals
lying on the (111) facets, while deeper in the valence band, the
(100) related states emerge. Overall, despite this unusual feaꢀ
ture, we do not regard any of the (111) states as traps or releꢀ
vant for our purposes. At more realistic sizes than the model
chosen for these calculations, the number of (111) MOs reꢀ
mains constant, as the size of this facet is unchanged, while
the number of (100) MOs grows significantly with the size of
the platelet. Overall, these calculations indicate that stripping
ASSOCIATED CONTENT
Supporting Information
Additional Experimental details, XRD spectrum, DLS measꢀ
urements, NMR analyses, diffusion coefficient calculations,
electronic structures of CdSe model nanoplatelets etc.
AUTHOR INFORMATION
Corresponding Author
ACS Paragon Plus Environment

Page 9 of 17
*Sh.S. Email: Shalini.Singh@ugent.be
Journal of the American Chemical Society
17. Ithurria, S.; Tessier, M. D.; Mahler, B.; Lobo, R. P. S.
1
2
3
4
5
6
7
8
M.; Dubertret, B.; Efros, A. L. Nat Mater 2011, 10, 936ꢀ941.
18. Gerdes, F.; Navío, C.; Juárez, B. H.; Klinke, C. Nano
Letters 2017, 17, 4165ꢀ4171.
19. Zhou, Y.; Wang, F.; Buhro, W. E. J. Am. Chem. Soc.
2015, 137, 15198ꢀ15208.
20. Bouet, C.; Tessier, M. D.; Ithurria, S.; Mahler, B.;
Nadal, B.; Dubertret, B. Chem. Mater. 2013, 25 (8), 1262ꢀ
1271.
21. Mahler, B.; Nadal, B.; Bouet, C.; Patriarche, G.;
Dubertret, B. J. Am. Chem. Soc. 2012, 134, 18591ꢀ18598.
22. Polovitsyn, A.; Dang, Z.; Movilla, J. L.; MartínꢀGarcía,
B.; Khan, A. H.; Bertrand, G. H. V.; Brescia, R.; Moreels, I.
Chem. Mater. 2017, 29, 5671ꢀ5680.
23. Tessier, M. D.; Javaux, C.; Maksimovic, I.; Loriette, V.;
Dubertret, B. ACS Nano 2012, 6 (8), 6751ꢀ6758.
24. She, C.; Fedin, I.; Dolzhnikov, D. S.; Dahlberg, P. D.;
Engel, G. S.; Schaller, R. D.; Talapin, D. V. ACS Nano 2015,
9, 9475ꢀ9485.
25. Antanovich, A.; Prudnikau, A.; Matsukovich, A.;
Achtstein, A.; Artemyev, M. J. Phys. Chem. C 2016, 120,
5764ꢀ5775.
26. Guzelturk, B.; Erdem, O.; Olutas, M.; Kelestemur, Y.;
Demir, H. V. ACS Nano 2014, 8, 12524ꢀ12533.
27. Drijvers, E.; De Roo, J.; Martins, J. C.; Infante, I.; Hens,
Z. Chem. Mater. 2018.
28. Achtstein, A. W.; Antanovich, A.; Prudnikau, A.; Scott,
R.; Woggon, U.; Artemyev, M. J. Phys. Chem. C 2015, 119,
20156ꢀ20161.
29. de Mello, J. C.; Wittmann, H. F.; Friend, R. H. Adv.
Mater. 1997, 9, 230ꢀ232.
30. Akoka, S.; Barantin, L.; Trierweiler, M. Anal. Chem.
1999, 71, 2554ꢀ2557.
31. Silvestre, V.; Goupry, S.; Trierweiler, M.; Robins, R.;
Akoka, S. Anal. Chem. 2001, 73, 1862ꢀ1868.
32. Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett.
1997, 78, 1396ꢀ1396.
33. VandeVondele, J.; Hutter, J. J. Chem. Phys. 2007, 127,
114105.
34. Hutter, J.; Iannuzzi, M.; Schiffmann, F.; VandeVondele,
J. Wiley Interdiscip. Rev: Comput Mol. Sci. 2014, 4, 15ꢀ25.
35. Gao, Y.; Weidman, M. C.; Tisdale, W. A. Nano Letters
2017, 17, 3837ꢀ3843.
36. Fritzinger, B.; Capek, R. K.; Lambert, K.; Martins, J. C.;
Hens, Z. J. Am. Chem. Soc. 2010, 132, 10195ꢀ10201.
37. Djebaili, T.; Richardi, J.; Abel, S.; Marchi, M. J. Phys.
Chem. C 2013, 117, 17791ꢀ17800.
38. Hinterwirth, H.; Kappel, S.; Waitz, T.; Prohaska, T.;
Lindner, W.; Lämmerhofer, M. ACS Nano 2013, 7, 1129ꢀ
1136.
39. Smith, A. M.; Johnston, K. A.; Crawford, S. E.;
Marbella, L. E.; Millstone, J. E. Analyst 2017, 142, 11ꢀ29.
40. De Nolf, K.; Cosseddu, S. M.; Jasieniak, J. J.; Drijvers,
E.; Martins, J. C.; Infante, I.; Hens, Z. J. Am. Chem. Soc.
2017, 139, 3456ꢀ3464.
Present Addresses
J.D.Rꢀ Department of Chemistry, Columbia University, New
York, New York 10027, United States
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The authors acknowledge support by the European Commisꢀ
sion via the MarieꢀSkłodowska Curie action Phonsi (H2020ꢀ
MSCAꢀITNꢀ642656), the Belgian Science Policy Office (IAP
7.35, photonics@be) and BOFꢀUGent (GOA no. 01G01513).
Sh.S. would like to acknowledge FWO postdoctoral funding
(FWO17/PDO/184). I.I. would like to thank The Netherlands
Organization of Scientific Research (NWO) for providing
financial support within the Innovational Research Incentive
(Vidi) Scheme. DFT calculations were carried out on the
Dutch national eꢀinfrastructure with the support of SURF
Cooperative.
REFERENCES
1. Talapin, D. V.; Lee, J.ꢀS.; Kovalenko, M. V.; Shevchenko,
E. V. Chem. Rev. 2010, 110, 389ꢀ458.
2. Kovalenko, M. V.; Manna, L.; Cabot, A.; Hens, Z.;
Talapin, D. V.; Kagan, C. R.; Klimov, V. I.; Rogach, A. L.;
Reiss, P.; Milliron, D. J.; GuyotꢀSionnnest, P.; Konstantatos,
G.; Parak, W. J.; Hyeon, T.; Korgel, B. A.; Murray, C. B.;
Heiss, W. ACS Nano 2015, 9, 1012ꢀ1057.
3. Hens, Z.; Martins, J. C. Chem. Mater. 2013, 25, 1211ꢀ1221.
4. Anderson, N. C.; Hendricks, M. P.; Choi, J. J.; Owen, J. S.
J. Am. Chem. Soc. 2013, 135, 18536ꢀ18548.
5. Green, M. L. H.; Parkin, GJ. Chem. Edu. 2014, 91, 807ꢀ
816.
6. Ji, X.; Copenhaver, D.; Sichmeller, C.; Peng, X. J. Am.
Chem. Soc. 2008, 130, 5726ꢀ5735.
7. Hughes, B. K.; Ruddy, D. A.; Blackburn, J. L.; Smith, D.
K.; Bergren, M. R.; Nozik, A. J.; Johnson, J. C.; Beard, M.
C. ACS Nano 2012, 6, 5498ꢀ5506.
8. Bealing, C. R.; Baumgardner, W. J.; Choi, J. J.; Hanrath,
T.; Hennig, R. G. ACS Nano 2012, 6, 2118ꢀ2127.
9. Saniepay, M.; Mi, C.; Liu, Z.; Abel, E. P.; Beaulac, R., J.
Am. Chem. Soc. 2018, 140, 1725ꢀ1736.
10. De Roo, J.; Van den Broeck, F.; De Keukeleere, K.;
Martins, J. C.; Van Driessche, I.; Hens, Z. J. Am. Chem. Soc.
2014, 136, 9650ꢀ9657.
11. Houtepen, A. J.; Hens, Z.; Owen, J. S.; Infante, I. Chem.
Mater. 2017, 29, 752ꢀ761.
12. Giansante, C.; Infante, I. J. Phys. Chem. Lett. 2017, 5209ꢀ
5215.
13. Boehme, S. C.; Azpiroz, J. M.; Aulin, Y. V.; Grozema,
F. C.; Vanmaekelbergh, D.; Siebbeles, L. D. A.; Infante, I.;
Houtepen, A. J. Nano Letters 2015, 15, 3056ꢀ3066.
14. Geiregat, P.; Houtepen, A. J.; Sagar, L. K.; Infante, I.;
Zapata, F.; Grigel, V.; Allan, G.; Delerue, C.; Van
Thourhout, D.; Hens, Z. Nat Mater 2017, 17 (1), 7.
15. Zhou, Y.; Buhro, W. E. J. Am. Chem. Soc. 2017, 139,
12887ꢀ12890.
41. Bagheri, G.; Bonadonna, C.. Powder Technol. 2016, 301,
526ꢀ544.
42. Chen, P. E.; Anderson, N. C.; Norman, Z. M.; Owen, J.
S. J. Am. Chem. Soc. 2017, 139, 3227ꢀ3236.
43. Campos, M. P.; Hendricks, M. P.; Beecher, A. N.;
Walravens, W.; Swain, R. A.; Cleveland, G. T.; Hens, Z.;
Sfeir, M. Y.; Owen, J. S. J. Am. Chem. Soc. 2017, 139, 2296ꢀ
2305.
16. Ithurria, S.; Bousquet, G.; Dubertret, B. J. Am. Chem. Soc.
2011, 133, 3070ꢀ3077.
ACS Paragon Plus Environment

Journal of the American Chemical Society
44. Yao, Y.; Zhou, Y.; Sanderson, W. M.; Loomis, R. A.; Buhro, W. E. Chem. Mater. 2018, 30, 2848ꢀ2857.
Page 10 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 11 of 17
Journal of the American Chemical Society
TOC
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 13 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 15 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 17 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment