Insights into the structural complexity of colloidal cdse nanocrystal surfaces: Correlating the efficiency of nonradiative excited-state processes to specific defects

Article abstract of DOI:10.1021/jacs.7b10649

II-VI colloidal semiconductor nanocrystals (NCs), such as CdSe NCs, are often plagued by efficient nonradiative recombination processes that severely limit their use in energy-conversion schemes. While these processes are now well-known to occur at the surface, a full understanding of the exact nature of surface defects and of their role in deactivating the excited states of NCs has yet to be established, which is partly due to challenges associated with the direct probing of the complex and dynamic surface of colloidal NCs. Here, we report a detailed study of the surface of cadmium-rich zinc-blende CdSe NCs. The surfaces of these cadmium-rich species are characterized by the presence of cadmium carboxylate complexes (CdX₂) that act as Lewis acid (Z-type) ligands that passivate undercoordinated selenide surface species. The systematic displacement of CdX₂ from the surface by N,N,N′,N′-tetramethylethylene-1,2-diamine (TMEDA) has been studied using a combination of ¹H NMR and photoluminescence spectroscopies. We demonstrate the existence of two independent surface sites that differ strikingly in the binding affinity for CdX₂ and that are under dynamic equilibrium with each other. A model involving coupled dual equilibria allows a full characterization of the thermodynamics of surface binding (free energy, as well as enthalpic and entropic terms), showing that entropic contributions are responsible for the difference between the two surface sites. Importantly, we demonstrate that cadmium vacancies only lead to important photoluminescence quenching when created on one of the two sites, allowing a complete picture of the surface composition to be drawn where each site is assigned to specific NC facet locale, with CdX₂ binding affinity and nonradiative recombination efficiencies that differ by up to two orders of magnitude.

Full text of DOI:10.1021/jacs.7b10649

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Insights into the Structural Complexity of Colloidal CdSe
Nanocrystal Surfaces: Correlating the Efficiency of Non-
Radiative Excited-State Processes to Specific Defects
Mersedeh Saniepay, Chenjia Mi, Zhihui Liu, Emily Paige Abel, and Rémi Beaulac
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b10649 • Publication Date (Web): 02 Jan 2018
Downloaded from http://pubs.acs.org on January 2, 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 free 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 accessible to all
readers and 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.
Journal of the American Chemical Society 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 20
Journal of the American Chemical Society
1
2
3
4
5
6
7
Insights into the Structural Complexity of Colloidal CdSe Nanocrys-
tal Surfaces: Correlating the Efficiency of Non-Radiative Excited–
State Processes to Specific Defects
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
Mersedeh Saniepay, Chenjia Mi, Zhihui Liu, E. Paige Abel and Rémi Beaulac*
Department of Chemistry, Michigan State University, East Lansing, Michigan, 48824-1322, United States
ABSTRACT: II-VI colloidal semiconductor nanocrystals (NCs), such as CdSe NCs, are often plagued by efficient non-radiative
recombination processes that severely limit their use in energy-conversion schemes. While these processes are now well-known to
occur at the surface, a full understanding of the exact nature of surface defects and of their role in deactivating the excited states of
NCs has yet to be established, which is partly due to the challenges associated with the direct probing of the complex and dynamic
surface of colloidal NCs. Here, we report a detailed study of the surface of cadmium-rich zinc-blende CdSe NCs. The surfaces of
these cadmium-rich species is characterized by the presence of cadmium carboxylate complexes (CdX2) that act as Lewis acid (Z-
type) ligands to passivate undercoordinated selenide surface species. The systematic displacement of CdX2 from the surface by
1
N,N,N,N-tetramethylethylene-1,2-diamine (TMEDA) has been studied using a combination of H NMR and photoluminescence
spectroscopies. We demonstrate the existence of two independent surface sites that differ strikingly in the binding affinity for CdX2
and which are under dynamic equilibrium with each other. A model involving coupled dual equilibria allows a full characterization
of the thermodynamics of surface binding (free energy, as well as enthalpic and entropic terms), showing that entropic contributions
are responsible for the difference between the two surface sites. Importantly, we demonstrate that cadmium vacancies only lead to
important photoluminescence quenching when created on one of the two sites, which allows a complete picture of the surface com-
position to be drawn where each site is assigned a specific NC facet locale, with differing CdX2 binding affinity and non-radiative
recombination efficiencies.
atoms, most notoriously by providing low-lying electronic
states that lead to efficient non-radiative excitonic
recombination pathways.1,25-32 One approach to circumvent
these processes consists in growing inorganic shells properly
chosen so as to electronically insulate the NC core from its
surface;33-37 this „potential barrier‟ approach, which has proven
extremely powerful in the utilization of NCs for
photoemission applications, cannot be easily extended to
instances where charge-carriers need to be exchanged across
the NC surface. In principle, the detrimental contribution of
surfaces can be mitigated by providing adequate termination
of the surfaces by ligating species that push the surface
electronic states far from the lowest delocalized excited-state
energy,1,24,25,30 and indeed the literature is replete with
empirical approaches that lead to various levels of
improvement of the photoluminescence quantum yield
(PLQY),30,38-43 the main metric by which the efficiency of non-
radiative recombination processes is generally quantified. An
important issue that severely limits the development of
rational approaches to control the behavior of NC surfaces is
the complexity of their structure, which varies not only from
NC to NC, but also from facet to facet or even from site to
site.44-48 Furthermore, surface-bound species are often labile
and/or mobile, and thus exhibit complicated dynamical
variations across complex free-energy landscapes with
multiple nearly-equivalent minima.27,49 Detailed and
quantitative structural surveys of NC surfaces are
consequently notoriously challenging tasks to accomplish even
for the simplest cases and using state-of-the-art analytical
probes,29,42,48,50-61 highlighting the challenge of experimentally
1. INTRODUCTION
Colloidal semiconductor nanocrystals (NCs) share the
fascinating electronic, optical, and chemical properties that
characterize crystalline semiconductor materials, but with an
explicit, and often strong, dependence on the dimensionality of
the crystallite.1,2 These chemical species, which can now
readily be prepared efficiently through a variety of solution-
phase approaches, are particularly well-suited for a host of
applications
–
photosensing,3-5 solid-state lighting,4,6-9
photovoltaics,3-5,10,11 bio-imaging,12-14 or photo-catalytic
transformations,15-21 to name but a few – that rely on the
extraction of work, of one form or another, from
electronically-excited NCs. Such energy-conversion schemes
necessarily require control over the energetic configuration of
NC electronic structure as well as the ability to limit undesired
excited-state relaxation pathways that lead to direct energetic
losses by rapid excitonic recombination. Unlike bulk
semiconductors, where non-radiative processes generally arise
at localized lattice defects (vacancies, impurity doping,
dislocations…)22 that are not generally found in NCs due to
unfavorable energies of formation,23 non-radiative relaxation
in NCs generally originate at the surface of the crystal.24,25 At
least from a structural point of view, the molecular objects that
we refer to as “nanocrystals” are as much characterized by
those atoms that are regularly organized in space (the so-called
core atoms), than by those that are found on their surfaces and
which often amount to a significant part of the total mass of a
single NC. Importantly, surfaces can significantly impact the
desirable electronic properties associated with the NC core
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 20
1
2
3
4
5
6
7
morphology and the ensemble photodynamical behavior of
CdSe NCs. Our results directly support the proposal that Cd
vacancies on (100) facets are efficient trapping centers,29,32 but
also provide a wider context by allowing a direct quantitative
estimate of the overall (ensemble) PLQY of CdSe NCs from
the knowledge obtained in this study about the thermodynamic
stability and prevalence of specific surface defects.
8
9
2. EXPERIMENTAL SECTION
2.1.Chemicals. Cadmium nitrate tetrahydrate (Cd(NO3)∙4H2O,
98%), oleic acid (90%), octadecene (ODE, 90%), N,N,N,N,-
tetramethylethylene-1,2-diamine (TMEDA, 99%), myristic acid
(>99%), ferrocene (Fc, 98%) were purchased from Sigma Aldrich.
Ethyl acetate (HPLC grade), toluene (ACS grade), sodium hydroxide
(>98%), anhydrous methanol (HPLC grade) were purchased from
Macron Fine Chemicals. Pentane was purchased from Fisher Chemi-
cal. d-toluene was purchased from Cambridge Isotope Laboratories
(CIL). Ethyl acetate was degased by bubbling nitrogen for 2 h and
dried over molecular sieves for at least two days prior to use. Oleic
acid was degased through 3 cycles of vacuum/purge with nitrogen and
kept under nitrogen atmosphere just prior to use. Ferrocene was re-
crystallized from methanol solutions. Toluene and pentane were dried
using an alumina distill columns. TMEDA was stored in the glovebox
and taken out for sample preparation. Other chemicals were used as
received.
2.2. Synthesis of Cadmium Myristate. Cadmium myristate was
prepared by adapting a literature protocol.80 4.63 g (15 mmol) cad-
mium nitrate tetrahydrate is weighed and transferred to a 500 mL
beaker with 150 mL anhydrous methanol. The mixture was sonicated
until all cadmium nitrate was dissolved, and the solution was trans-
ferred to a 250 mL addition funnel. Separately, 1.2 g (30 mmol)
crushed NaOH, 6.85 g (30 mmol) myristic acid and 1.5 L anhydrous
methanol were added to a 2 L flask; the mixture was heated in water
bath at 40 oC for 2 min, then sonicated until a single phase solution of
sodium myristate is formed. The solution was then transferred to a 2 L
round bottom flask with a stir bar, and positioned over a magnetic
stirrer. The cadmium nitrate solution was then added into the sodium
myristate solution over the course of 3 h, under stirring. The resulting
cadmium myristate was then filtered and washed five times with an-
hydrous methanol (200 mL) and air-dried for 24 h, then crushed and
transferred to a 250 mL round bottom flask, and finally sealed and
dried under vacuum for 24 h at 50 oC. The yield was about 82%.
Product was characterized using melting point, 1H NMR and IR spec-
troscopies.
2.3. Synthesis and Purification of CdSe NCs. Cadmium selenide
was prepared following literature protocols.81 5.67 g (10 mmol) cad-
mium myristate, 1.11 g (10 mmol) selenium dioxide powder and
630 mL octadecene were placed in a 1 L three-neck round bottom
flask equipped with a water condenser. The flask was degassed in
three cycles of 20 min each, then heated to 240 oC over 7 min. Heat
was maintained until the desired size was reached, which was checked
by taking the absorption of a diluted aliquot of the reaction suspension
(vide infra). Heat was consequently removed and 10 mL (32 mmol) of
degassed oleic acid were injected over 20 min while the flask was
allowed to cool down back to room temperature, followed by conti-
nuous stirring for 15 h under nitrogen flow. Octadecene was then
removed by vacuum (50-70 mTorr) distillation at 130 oC, after which
20 mL dry toluene was added to the dark red concentrated and impure
NCs suspension, which was finally transferred into four test tubes and
centrifuged for 20 min at 6000 rpm.
The supernatant was transferred in 1 mL portions into test tubes,
and 10 mL ethyl acetate was added to each tube to precipitate the
NCs. After centrifugation at 6000 rpm, the supernatant was removed
and the NCs resuspended in 1 mL of pentane; the entire cleaning
cycle was repeated four times. Finally, 10 mL ethyl acetate was added
to each test tube (with only the particles), followed by sonication for
15 s. The test tubes were then centrifuged (6000 rpm) and the super-
natant removed. The NCs were dried over nitrogen for 24 h, and then
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
Scheme 1. Surface exchange processes involving two Lew-
is-basic surface sites, B1 and B2, with different binding af-
finities to a generic CdX2 Lewis-acidic moiety.
relating specific surface defects to non-radiative loss
mechanisms, which is one of the main objectives of the
present contribution.
We present here a detailed investigation of the surface of CdSe
NCs, one of the most studied materials of its class.29,32,60-76 We
specifically focused our attention to zinc-blende (cubic) CdSe
NCs obtained from carboxylate precursors; these materials
have a non-stoichiometric cadmium-rich composition, which
arises from the presence of neutral cadmium carboxylate
(CdX2) complexes on the NC surfaces. A valuable model to
describe the surface of these NCs has been formulated by
Owen and collaborators,42,52,58,77 whereby these CdX2
complexes are considered as Lewis acid acceptors (Z-type
ligands in the Covalent Bond Classification, CBC)78,79 that
bind to unpassivated surface selenide sites (Se2-), yielding
stable surfaces and likely playing an important role in favoring
the zinc-blende morphology during the growth process. These
surface-bound CdX2 complexes can be displaced by simple
(Lewis) acid/base exchange reactions taking the following
general form:
(1)
where B represents a generic Lewis-basic site at the surface of
the CdSe NCs (sometimes referred to as a “dangling bond”)
and L is a soluble Lewis base (L-type ligand in the CBC).78,79
The displacement of CdX2 from the surface induces a strong
reduction of the PLQY, betokening enhanced non-radiative
relaxation to surface states that have been associated to hole-
trapping processes.29,32
Using the general ligand exchange process given in eq. (1), we
demonstrate the existence of at least two types of sites at the
surface of zinc-blende CdSe NCs (Scheme 1). We show that
the two types of site not only differ strikingly in their binding
affinity to CdX2 complexes, but also lead to very different
non-radiative mechanisms. The thermodynamical analysis of
the surface binding and the photoluminescence spectroscopy
data allows a detailed model of the surface composition to be
proposed, along with a direct correlation between surface
2
ACS Paragon Plus Environment

Page 3 of 20
Journal of the American Chemical Society
1
suspended in d-toluene to make a 85.8 µM stock solution of NCs, as
2
3
assessed by UV-Vis spectroscopy, following the empirical calibration
curve established by Mulvaney et al.82
2.4. Optical Spectroscopy. Absorption spectra are collected on an
OLIS17 UV/vis/NIR spectrometer in 1 nm increments, using 1 cm
pathlength quartz cuvettes. Spectra were measured in toluene after the
toluene background was subtracted. Continuous wave (CW) photolu-
minescence (PL) measurements were performed using a 0.55 m focal
length spectrometer (iHR550, f/6.4, 150 grooves/mm grating blazed at
500 nm) equipped with a CCD detector (Horiba Symphony II nitro-
gen cooled deep depleted CCD), and exciting at 450 nm using a
tungsten lamp coupled to a 0.32 m focal length monochromator (Ho-
riba iHR320, f/4.1, 600 grooves/mm grating blazed at 500 nm).
PLQY were measured with an absolute PL quantum yield spectrome-
ter (Hamamatsu Quantasaurus, C11347), using toluene suspensions
prepared as described below and exciting at 400 nm.
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
2.5. Sample Preparation for 1H Nuclear Magnetic Resonance
(NMR) and Photoluminescence Spectroscopies. Toluene solutions
of TMEDA were prepared to cover a range of concentrations from
0.01 M to 6.7 M (neat TMEDA). Mixtures of CdSe NCs and TMEDA
in d-toluene were prepared, each containing 600 µL of 85.8 µM CdSe
NCs ([CdSe NC] = 78 µM), 30 μL of a given TMEDA solution
([TMEDA] varying from 0.1 mM to 273 mM, or about 1:1 to 1:3500
CdSe NC:TMEDA ratios) and 30 μL of 4.2 mM ferrocene solution
([Fc] = 0.19 mM); each sample was allowed to equilibrate after mix-
ing by waiting 30 min prior to measurements.42 1H NMR spectra were
collected on an Agilent DDR2 500 MHz NMR spectrometer equipped
with 7600AS 96 sample auto-sampler running VnmrJ 3.2A, using a
45o pulse angle / 10 s relaxation time sequence and 32 scans (see SI
for more details). Spectra were analyzed using MestreNova (Mestre-
lab Research S.L.) and Igor Pro 7.05 (Wavemetrics). Samples for
photoluminescence (PL) spectra were prepared by diluting 30 μL
aliquots from each NMR sample to 2 mL with toluene. Equilibrated
samples of CdSe NCs with various amounts of TMEDA exhibit sur-
prisingly sustained stability for colloidal materials: the same samples
measured over 8 months apart (stored in the glovebox) exhibited the
Figure 1. Absorption (solid) and photoluminescence (dashed)
of 3.8 nm CdSe NCs (1  0.1 µM) before (black) and after
(red) addition of TMEDA (3.6 mM), illustrating the impact of
displacing CdX2 complexes from the surface of CdSe NCs.
that of the carboxylate ligands. The ligand coverage was then calcu-
lated from the number of cadmium carboxylate ligands per NC (con-
centration of ligands divided by concentration of NC) and the surface
area of the NC, assuming spherical shape (as confirmed by TEM).
Although the surface coverage involves a mixture of both oleate and
myristate species, both types of ligands behave identically with regard
to exchange reactions,83 which we further confirmed by NMR by
monitoring terminal methyl protons (which account for both types of
carboxylates) as well as vinylic protons (which belong only to oleate
ligands). Specific ligand exchange reactions were monitored subse-
quently by monitoring exclusively the vinylic peaks. Residual
protonated carboxylates (oleic or myristic acid) were observed by
displacement of ethane with diethyl zinc added to isolated and
cleaned NC samples (assuming there are no other sources of protons);
the integration of the ethane 1H NMR peak suggested that these
species form at most 10% of the total carboxylate population, which
is consistent with values reported by Chen et al.61 This puts at ~5%
the overestimation of the cadmium coverage values reported herein.
Further details are given in Supporting Information and below.
1
same surface coverage (as measured by H NMR), PLQY, and ab-
sorption spectra, showing no change in the average size, surface com-
position, or concentration of suspended NCs.
2.6. Ligand Concentration and Surface Coverage Calculations.
1
The integration of terminal methyl (-CH3) peak at 1 ppm in H NMR
each spectra was recorded and referenced to the integration of the
ferrocene protons at 4 ppm, allowing for the calculation concentration
of total carboxylate ligands. From charge compensation considera-
tions, the number of cadmium carboxylate complexes (CdX2) is half
3. RESULTS AND DISCUSSION
3.1. CdX2 Displacement from CdSe NCs. The absorption
and photoluminescence spectra of a clean suspension (that is,
[CdX2]tot
B
(a)
(b)
(c)
F
100
[TMEDA]
0
80
0.35
1.05
[CdX2]B
2.1
4.2
60
40
25
50
[CdX2]F
20
137
274
0
8
7
6
5
4
3
2
1
6.0 5.8 5.6 5.4
0
100
200
300
[TMEDA] (mM)
Chemical Shift (ppm)
Chemical Shift (ppm)
Figure 2. (a) 1H NMR Spectra of 3.8 nm CdSe NCs (78 μM in d-toluene) in presence of different concentrations of TMEDA (10
points are selected, values in mM listed on the right-hand side). (b) Zoomed-in spectra (from panel (a)) between 5 and 6 ppm
showing the bound (B) and free (F) cadmium carboxylates (CdX2). (c) Total number of CdX2 units per CdSe NC (gray squares),
number of CdX2 units bound to the surface of CdSe NCs (red circles), and number of CdX2 units displaced by TMEDA from the
surface of CdSe NCs (blue circles).
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 20
1
2
3
4
5
6
7
8
without free ligands; vide infra) of 3.8 nm CdSe NCs are
at 5.47 ppm, which is readily assigned to carboxylate species
shown in Figure 1 (data for 3.0 and 4.1 nm are also given in
SI). Upon addition of 3,500 equivalents of TMEDA per NC,
the PLQY drops by more than two orders of magnitude, from
10% to <0.1%, which is associated with the displacement of
CdX2 complexes in solution (eq. (1)), which can be quantified
that are unbound from the surface of CdSe NCs (hereby
labeled “free”).64,84-86 Importantly, as shown in Fig. 2(c), the
decrease in intensity from the broad band is exactly matched
by the increase in the integrated intensity of the sharp band,
consistent with the assignment of a single species (the
carboxylate ligands) being exchanged dynamically under the
influence of the TMEDA species. The nature of this
equilibrium was discussed in great details previously42 as a L-
type ligand-induced displacement of surface-bound cadmium
carboxylates (CdX2) complexes into solution, as shown in a
general form by eq. (1), where the soluble Lewis base (L) was
chosen here to be N,N,N‟,N‟-tetramethylethylenediamine,
TMEDA, shown in Scheme 1 (we note that formally TMEDA,
a bidentate donor, would be referred to as a L2 ligand and the
resulting free complex as CdL2X2 in the CBC;78,79 we use the
shortcut notation throughout here). The CdX2 complex
behaves as a Lewis acid (Z-type ligand) and eq. (1) thus
describes the Lewis acid/base exchange of a Z-type ligand (the
cadmium carboxylate complex) between two Lewis base
groups: (i) the surface of the NC (B) and (ii) TMEDA (L).
Interestingly, the trends in Fig. 2(c) suggest that the simple
relation given by eq. (1) cannot account fully for this
dynamical exchange: whereas almost a third of the surface-
bound CdX2 can be displaced upon addition of 10 mM of
TMEDA (about 125 equivalents per NC here), the further
addition of over 250 mM of TMEDA has a much reduced
impact overall, bringing the total CdX2 displacement to 40%
of the initial value. This general feature is observed for all
samples we investigated, irrespective of the NC size (SI). We
show below that this behavior can be readily explained by
invoking the existence of two separate CdX2 binding sites at
the NC surface, each with their own free energy of binding
and role in deactivating the excited state of CdSe NCs.
1
1
by H NMR.42 The overall H NMR spectra of 3.8 nm CdSe
NCs, with and without TMEDA, are presented in Figure 2(a).
Given the nature of these colloidal NCs (zinc blende CdSe
9
NCs capped with oleate/myristate surfactants
–
see
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
Experimental Section and SI for more details), these spectra
give a detailed account of the surfactants bound to the surface
of the NCs, as well as of any organic species freely diffusing
in solution. Of particular interest are the integrated intensities
of bands associated with specific ligands, as they allow a
direct evaluation of the NC surface densities of these species
to be evaluated. Direct comparison of the integrated intensity
of terminal methyl (-CH3) groups, located at 1 ppm in
Fig. 2(a), with an internal standard (ferrocene), yields a value
of 190 ± 4 carboxylate (X-) ligands per NC, corresponding to
95 ± 2 CdX2 complexes per NC for the 3.8 nm CdSe NCs
sample shown in Fig. 2. This value corresponds to a surface
coverage of 4.2 ± 0.6 carboxylate∙nm-2, or, equivalently,
2.1 ± 0.3 CdX2∙nm-2, a value close to the largest coverage of
CdSe NC by carboxylates reported in the literature.42,55,84
Furthermore, the 1H NMR spectrum of as-prepared CdSe NCs
indicates that, prior to the addition of TMEDA, all carboxylate
species are quantitatively bound to the surface, without any
measurable trace of unbound carboxylate species. This is
indicated by the single broad feature around 5.7 ppm, better
seen in the zoomed-in perspective of Fig. 2(b), which is
assigned to the vinylic protons of surface-bound oleate
ligands. Upon addition of TMEDA to the CdSe NCs
suspension, the integrated intensity of this broad feature
decreases, concomitantly with the increase of a sharp feature
3.2. Analysis of the Ligand-Exchange Equilibrium. We
[TMEDA] = [L] + [L-CdX2] (mM)
[ TMEDA] / [ CdX2]
10 20 30
log10([TMEDA] / [CdX2])
-1 4
0
100
200
300
0
40
0
1
2
3
20
18
16
14
12
10
50
60
50
40
30
20
10
0
100
80
60
40
20
0
(a)
(b)
(c)
B1 + B2
B1 + B2
B1
B2
70
80
B1
B2
90
100
0
20 40 60 80 100
0
100
200
300
-1
0
1
2
3
4
5
φ = [L] / [L-CdX2]
log10([ TMEDA] / 1 mM)
[TMEDA] (mM)
Figure 3. (a) TMEDA titration of 3.8 nm CdSe NCs (78 µM in d-toluene), modeled with eq. (11) (black dashed line). The top
axis is a transformation of the ratio φ reported on the bottom axis (eq. (13), and represents the total TMEDA concentration in
solution (note the non-linear scale). The right axis is a reciprocal transformation of the left axis, showing the number of bound
CdX2 units per NC. (b) Calculated number of vacant sites at the surface of the 3.8 nm CdSe NC sample, for each type of site, B1:
blue, B2: red, all sites: black. The data for the total number of open sites, equivalent to the number of CdX2 ligands that are
removed from each surface site as a function of added TMEDA, is shown by the orange circles. The top axis is linearly scaled to
represent the TMEDA concentration in total equivalents of CdX2 complexes. (c) Same data/calculated curves as panel (b), but for
a wider range of TMEDA concentrations and on a logarithmic scale. The largest magnitudes shown here correspond to
unrealistically large concentrations, and are shown here for discussion purposes. The two arrows indicate the TMEDA
concentrations where temperature-dependent studies (Figure 4) where conducted (blue arrow: 2 mM, red arrow: 274 mM).
4
ACS Paragon Plus Environment

Page 5 of 20
1
Journal of the American Chemical Society
begin by describing quantitatively the equilibrium involving a
needed to describe the data in Figure 3(a). We note that the
2
3
4
5
single type of surface sites (B), which can be either open
(vacant) or occupied by a CdX2 complex, itself being involved
in a dynamic exchange between the free Lewis base (L) and
the Lewis-acidic surface site (B), as given by eq. (1) above.
The equilibrium constant for this equilibrium is given by:
two equilibria are not well separated, which implies that both
equilibrium regimes are coupled to each other (at least in the
intermediate region) and need to be considered
simultaneously.
6
As above, we define these exchange reactions as being
associative in terms of the CdX2 complex, and use the labels
B1 / B2 and B1-CdX2 / B2-CdX2 to distinguish the two types of
open (vacant) and occupied surface sites, respectively:
7
é
ùé ù
é
ù
B- CdX2
L
B- CdX2
ë
ûë û
ë
û
8
Keq
=
=
f
é ùé
ù
é ù
B
B
L- CdX2
9
ë ûë
û
ë û
(2)
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
where as an approximation molarities are invoked instead of
activities. We note that eq. (1) describes the binding of the
CdX2 complex onto the NC surface, which is the reversal of
the displacement process monitored in Fig. 2; as shown below,
this form yields simpler equations to describe quantitatively
the equilibrium process; this convention will be kept
throughout the study, where larger equilibrium constants are
associated with stronger NC-CdX2 bonds (i.e. less easily
displaced by TMEDA). The parameter ϕ in eq. (2) designates
the ratio of unbound (“free”) vs. bound TMEDA:
(6a)
(6b)
The equilibrium constants for each exchange reaction are
given by:
é
ë
ùé ù
é
ë
ù
û
B1 - CdX2
L
B1 - CdX2
ûë û
ù
K1 =
=
f
é
ùé
é
ù
B1
B1 L- CdX2
ë
ûë
û
ë
û
(7a)
é
ë
ùé ù
é
ë
ù
û
B2 - CdX2
L
B2 - CdX2
é ù
é ù
ûë û
L
L
ë û0
K2 =
=
f
ë û
f =
=
- 1
é
ùé
ù
é
ù
B2 L- CdX2
B2
é
ù
é
ù
ë
ûë
û
ë
û
L- CdX2
L- CdX2
(7b)
ë
û
ë
û
(3)
The analysis of 1H NMR does not allow us to distinguish
between the two types of sites (B1 and B2), as the
deconvolution of free and bound cadmium carboxylates peaks
yields only the total concentration of TMEDA-bound CdX2
species, [L-CdX2], as well as the total concentration of
surface-bound CdX2 species, [B-CdX2]:
and is a quantity directly obtainable from NMR data such as
that presented in Fig. 2 ([L]0 is the total concentration of
TMEDA). The total number of binding sites per NC, N, is
given by the sum of occupied and vacant (open) sites:
é
ù é ù
B- CdX2 + B
ë
û ë û
N =
é
ù
é
ù é ù
é
ù
B- CdX2 = B1 - CdX2 + B2 - CdX2
NC
ë
û
ë
û ë
The total number of surface sites, Ntot, is given by:
ù é
û
ë
û
(8)
(4)
Combining equations 2 and 4 yields an equation to describe
the equilibrium process in terms of directly measurable
quantities:
é
ù
B- CdX2 + L- CdX2
û ë
ë
û
N
tot = N1 +N2 =
é
ù
NC
é
ù
ë
û
NC
(9)
1
1
ë
û
=
+
f
where we assume full coverage of the surface in the absence
of TMEDA (vide infra). The coupled exchange reactions
given by eqs. (6a-b) are fully described by the following
equation, which invokes the same two observables as eq. (5)
above ([NC]/[B-CdX2] vs. φ):
é
ù
N
N×Keq
B- CdX2
ë
û
(5)
Eq. 5 shows that a plot of [NC]/[B-CdX2] (the inverse of the
number of bound CdX2 complexes per NC, directly obtained
from the NMR data in Fig. 2(b)) vs. ϕ should yield a line of
slope (N∙Keq)-1 and intercept N-1. Figure 3(a) reports the same
data shown in Fig. 2(c) for the 3.8 nm CdSe NCs using the
formalism of eq. (5). Consistent with the conclusion drawn
above in Section 3.1, the data shows two different regimes,
each converging to linear trends in the limits of small and
large TMEDA concentrations. Obviously, the assumption of a
single binding site is too restrictive, and a more sophisticated
equilibrium model involving two types of surface sites is
f 2 +(K1 +K2)×f +K1 ×K2
é
ù
NC
ë
û
=
é
ù
(N1 ×K1 +N2 ×K2)×f +Ntot ×K1 ×K2
B- CdX2
ë
û
(10)
Interestingly, we note that the two limiting linear regimes
observed at small and large [TMEDA] in Fig. 3 can be directly
extracted from eq. (10) above:
Table 1. CdSe NC surface-related data, obtained from analysis of TMEDA titration experiments, T = 293.15 K.
Surface
coverage
(CdX2/nm2)
2.1 ± 0.3
[NC]b
(µM)
[CdX2]c
(mM)
NC Diametera
(nm)
b
N1
N2
K1
K2
Ntot
3.0 ± 0.2
3.8 ± 0.2
4.1 ± 0.3
23 ± 1
34 ± 2
43 ± 4
34 ± 3
61 ± 4
78 ± 6
1.0 ± 0.3
0.9 ± 0.3
0.6 ± 0.3
(9 ± 5) × 102
(7 ± 4) × 102
(3 ± 1) × 102
57 ± 2
95 ± 2
121 ± 2
80 ± 2
78 ± 2
78 ± 2
4.6 ± 0.3
7.4 ± 0.7
9.4 ± 0.2
2.1 ± 0.3
2.2 ± 0.2
a From lowest excitonic transition energy in the absorption spectrum, using the calibration curve of ref. 81.
b From total excitonic absorption cross-section, using the calibration curve of ref. 81.
c From 1H NMR data of the as-prepared CdSe NCs, assuming no intrinsic vacancies; q.v. Experimental Section for further details.
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 20
1
2
Table 2. Thermodynamic parameters for surface-bound CdX2 exchange with TMEDA of CdSe NCs, T = 293.15 K.
3
4
5
6
7
8
a
a
b
b
c
c
ΔΗ1
ΔΗ2
ΔS1
ΔS2
ΔG1
ΔG2
NC Diameter
(nm)
(kJ∙mol-1)
(kJ∙mol-1)
(J∙mol-1∙K-1)
(J∙mol-1∙K-1)
106 ± 21
99 ± 10
(kJ∙mol-1)
0.0 ± 0.7
0.3 ± 0.8
1.2 ± 1.4
(kJ∙mol-1)
3.0 ± 0.2
3.8 ± 0.2
4.1 ± 0.3
11 ± 2
14 ± 4
38 ± 10
-17 ± 2
8 ± 2
13 ± 1
26 ± 10
-16 ± 2
10 ± 5
13 ± 4
30 ± 22
92 ± 17
-14 ± 1
9
a from the van ‟t Hoff analysis presented in Fig. 4; b from DS = DH - DG ×T- 1 , using ∆H and ∆G from this Table.
(
)
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
c from
, using Kn from Table 1.
DGn = - R×T ×lnKn
each surface site to the ratio φ, or, equivalently, to the total
amount of TMEDA added, cf. eq. (13):
æ
ö
K1 +K2
[NC]
1
lim
=
+
×f
f ®0 ç
÷
[B- CdX ] N1 +N2 (N1 +N2)K1 ×K2
è
ø
2
(11a)
(11b)
é
ù
Bi
Ni ×f
ë
û
=
i =1,2
æ
ö
÷
K1 +K2
[NC]
1
é
ù
f +Ki
NC
lim
=
+
×f
ë
û
f ®¥ ç
(12)
[B- CdX ] N1 ×K1 +N2 ×K2 N1 ×K1 +N2 ×K2
è
ø
2
Ntot f 3 +f 2 + N K +N K f 2 +f
(
)
(
)
(
)
1
2
2
1
The data in Fig. 3 has been fitted to eq. (10) using a non-linear
regression analysis, yielding values for N1, N2, K1, and K2 that
are listed in Table 1.
é ù
é
ù
NC
(13)
L
=
ë û0
ë
û
f 2 + K +K f +K K
(
)
1
2
1
2
The displacement of the CdX2 complexes from the CdSe NC
surfaces is plotted separately for each individual site (B1 and
B2) in Figure 3(b), showing clearly that full displacement of
CdX2 from the B1 surface sites is effectively achieved at
around 50 mM of added TMEDA, corresponding to about 650
equivalents per NC. Comparatively, less than 10% of the B2
sites are vacant under identical conditions; as shown in
Fig. 3(c), there is not any physically realistic scenario under
which TMEDA could displace all of the CdX2 complexes from
the B2 sites, as concentrations higher than the density of pure
liquid TMEDA being required.
The good agreement between the model and the observed
experimental data validates the assumption that each site on
the surface interacts independently from each other, as
strongly correlated surface binding would not be well modeled
by eq. (10) above. Importantly, different CdSe NC sizes are
equally well described by the model, and the relevant
parameters for each NC size are listed in Table 1. On the other
hand, the good agreement between the sum N1 + N2 obtained
from the nonlinear regression analysis and the value Ntot
extracted from the NMR quantitation analysis should not be
regarded as indicative of the validity of the model, as the
model is itself designed to yield such a good agreement: the
sum N1 + N2 is related to the y-axis intercept, which is
experimentally related to the value Ntot obtained from the
NMR data, see eq. (11) above for φ = 0. Rather, a subtle point
related to this issue is the question of the validity of the
assumption that there are no vacant sites prior to displacement
of CdX2 by TMEDA: as vacancies in our study are never
probed directly per se, it is not strictly possible to verify this
assumption, although we can evaluate the possible impact that
their presence would have on the exchange dynamics. Given
the large difference between K1 and K2, it is clear that under
equilibrium, any intrinsic vacant site would have to exist as a
B1 site. As a consequence, neglecting the presence of intrinsic
vacant B1 sites in our model artificially boosts the value of the
K1 equilibrium constants, that is, the mass action of each
equivalent of TMEDA in eq. (6a) is effectively reduced.
Consequently, the K1 values reported in Table 1 can be
regarded as upper-bound values for the actual equilibrium
constants, to the extent that intrinsic B1 vacancies, present in
the as-prepared NC samples, would need to be taken into
account. A more detailed analysis is presented in Section 3.5
below, allowing an estimate of the fraction of vacancies in the
as-prepared samples and of the actual equilibrium constant K1.
It is sufficient for now to regard the values N1 and K1 listed in
Table 1 as effective parameters that describe accurately the
equilibria of eqs. (6a,b).
Close inspection of the data in Table 1 reveals a few
interesting facts. First, the apportionment of each type of site
appears fairly regular across all sizes, with a rough 2:3 ratio
between N1 and N2, although we note that this likely depends
Temperature (K)
340
320
300
280
260
6
4
K2
2
0
K1
-2
2.8
3.0
3.2
3.4
3.6
3.8
Reciprocal Temperature (x10-3 K-1)
Figure 4. Temperature dependence of the equilibrium
constants K1 and K2 for 3.8 nm CdSe NCs, expressed in the
van ‟t Hoff formalism. Blue circles: [TMEDA] = 2 mM; red
circles: [TMEDA] = 273 mM; the error bars are at most as
large as the circles shown here. The black dashed lines are
linear least squares fitted curves (R2 = 0.885 for K1 and R2 =
0.830 for K2).
With the values of the equilibrium constants and site densities
in hand, a useful analysis of each separate equilibria can be
obtained, for instance by relating directly the occupancy of
6
ACS Paragon Plus Environment

Page 7 of 20
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
regimes are highlighted with arrows for the 3.8 nm CdSe NC
Table 3. Comparison of the thermodynamic binding stability
of CdX2 on the B2 vs. B1 surface sites of CdSe NCs.a
sample in Fig. 3(c). In each regime, and for all NC sizes
investigated here (see SI), NMR studies indicate that the
TMEDA-induced displacement of surface-bound CdX2
becomes more efficient the lower the temperature, indicating
that this process is exothermic (∆H < 0) or, in the framework
established above, that eqs. (6a,b) represent endothermic
processes (∆H > 0). A quantitative estimate of the enthalpy of
reaction can be obtained from a van ‟t Hoff analysis of the
temperature-dependent equilibria:
ΔΗ2-1
(kJ∙mol-1)
3 ± 4
ΔS2-1
(J∙mol-1∙K-1)
68 ± 23
ΔG2-1
(kJ∙mol-1)
-17 ± 2
NC Diameter
(nm)
3.0 ± 0.2
3.8 ± 0.2
4.1 ± 0.3
5 ± 2
73 ± 14
-16 ± 2
9
10
11
3 ± 5
62 ± 28
-15 ± 2
a Data referenced to eq. (15), ∆X2-1 = ∆X2 - ∆X1 ; T = 298.15 K.
dlnKeq
DH
= -
sensitively on sample preparation conditions, and NCs
prepared by different approaches, or those with much smaller
superficial coverage than those reported here, might not
exhibit the same trend. Second, whereas equilibrium constants
for a given type of sites generally appear to decrease with
increasing sizes, the accuracy of the analysis does not allow
for firm conclusions to be raised concerning the precise size
dependence of these parameters. Furthermore, we note that
from a free energy perspective (Section 3.3), there is not a
significant difference across the range of values reported for a
given site. A much more critical observation stemming from
this analysis, and a general feature common to all samples
(beyond the existence of two distinct surface sites), is the fact
that the equilibrium constants for the binding of CdX2 on each
of the two surface sites differ by about 2-3 orders of
magnitude, i.e., pK1 ~ 0 and pK2 ~ -2. This points to a
significant difference in the free energy of binding CdX2
between B1 and B2, which we address specifically in the next
Section.
d T - 1
R
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
(
)
(14)
Figure
4
reports the temperature-dependent exchange
equilibria data obtained from NMR in the formalism of eq.
(14). The behavior of lnK vs. T-1 over the whole temperature
range investigated (roughly 40 K above and beyond room
temperature) is linear, indicating that the enthalpy change is
nearly temperature-independent over this range, as expected
for the simple ligand exchange processes under consideration.
The enthalpy change for each exchange reaction is thus
directly related to the gradient of each fitted line in Fig. 4. The
Gibbs free energies for each process are directly obtained from
the equilibrium constants in Table 1, which then directly
yields the corresponding entropic changes; the thermodynamic
quantities for eqs. (6a-b) extracted from such van ‟t Hoff
analyses are reported in Table 2 for three different CdSe NCs
sizes.
Although the thermodynamic data in Table 2 is referenced to
TMEDA (more precisely, to differences between state
functions of the TMEDA-CdX2 complex and uncoordinated
solvated TMEDA), a direct comparison between the binding
stability of CdX2 to B1 vs. B2 sites can be obtained by
subtracting the two reactions in eqs. 6a and 6b, yielding the
following overall equilibrium:
3.3 Thermodynamics of Surface Site Binding. To probe
further into the specific nature of the two different types of
surface sites, temperature-dependent 1H NMR studies were
conducted as a way to modulate the equilibrium of the CdX2
exchange reactions and extract the enthalpic contribution to
the overall free energy of each exchange process. For each NC
sizes, two TMEDA concentrations were selected, each
corresponding to a regime where one of the two exchange
reactions dominates the overall equilibrium. These two
(15)
which is independent of the nature of the Lewis base species
used to displace the CdX2 complexes; the thermodynamic
1.0
60
50
40
30
20
10
0
60
0 mM
(a)
(b)
(c)
0.8
0.6
0.4
0.2
0.0
50
40
30
20
10
0
B2 vacancies
TMEDA
B1 vacancies
4 mM
625
525
575
Wavelength (nm)
0
1
2
3
4
0
1
2
3
4
0
10
20
30
CdX2 groups removed
[TMEDA] (mM)
[TMEDA] (mM)
Figure 5. (a) PL intensity of 3.8 nm CdSe NCs (~1.0 ± 0.1 µM), with varying amounts of TMEDA (0 -3.6 mM). Inset: PL spectra
of the same samples. Intensities have been normalized to the integrated PL intensity of the CdSe NCs-only sample (I0), after
baseline subtraction and correction for dilution effects. (b) Quantification of the TMEDA-induced CdSe NC PL quenching,
reported as the Stern-Volmer ratio, I0/I. (c) CdSe NC PL quenching efficiencies of each type of vacancies, as induced by
displacement of CdX2 by TMEDA. The dashed lines are linear least squares fitted curves, constrained to the initial quenching
region. The PL quenching correlates well with the creation of vacancies on the B2 sites.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 20
1
2
Table 4. CdSe NCs Photoluminescence and Stern-Volmer Analysis Data.
3
4
5
KSBV2
KSBV2 M- 1
a
b
c
NC Diameter
(nm)
PLQYbright
(%)
PLQYdark
(%)
Pdark
(%)
Observed
PLQY (%)
Predicted
KSBV1 M- 1
(
)
(
)
PLQYd (%)
B1
K
SV
6
7
8
9
3.0 ± 0.2
3.8 ± 0.2
4.1 ± 0.3
0.22 ± 0.04 170 ± 20 800 ± 200
17 ± 2
14 ± 2
22 ± 5
0.1 ± 0.1
0.3 ± 0.3
2 ± 1
0.4
1.0
2.0
16 ± 3
10 ± 2
12 ± 2
16 ± 2
13 ± 2
22 ± 5
0.18 ± 0.02
0.11 ± 0.03
53 ± 4
12 ± 1
290 ± 40
150 ± 60
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
a Calculated PLQY of a NC without a B2 vacancy, eq. (19); b Calculated PLQY of a NC with a single B2 vacancy, eq. (20); c Probability that a NC has a
single B2 vacancy; d Overall calculated PLQY
.
dark
= 1- P ×PLQYbright + Pdark ×PLQY
(
)
dark
parameters relevant to eq. (15) are listed in Table 3. The data
in Table 3 shows directly that the difference in the enthalpic
terms associated with the equilibrium of the B1 and B2 sites is
not only quite small, but also opposite to the thermodynamic
stability associated with the B2-CdX2 complex. Indeed, as
precise composition of the NC surface can be made, as
presented in Fig. 5(c). This figure clearly indicates a strong
correlation between the overall PL quenching and the creation
of B2 sites vacancies: comparatively, the removal of CdX2
species from B1 sites does not appreciably quench the NC PL
(the rise of the Stern-Volmer ratio at large B1 displacement
values is actually due to the concomitant creation of B2
vacancies). From the values of the Stern-Volmer constants
extracted from the onset of each regime, the efficiency of a B2
vacancy is estimated to be nearly 300 times larger than that of
a B1 vacancy: KSV = 0.18/CdX2 for B1 vs. 53/CdX2 for B2
(Table 4). This stark contrast between the two types of
vacancies underlines that not all surface defects are necessarily
equivalent, even for such defects that are in appearance quite
similar (here, CdX2 vacancies). Furthermore, although the
overall stability of the B2-CdX2 complex overwhelmingly
surpasses that of the B1-CdX2 (making B2 vacancies much
rarer species under equilibrium conditions), the creation of
even a single B2 vacancy leads to a greater impact on the PL
quenching (by over one order of magnitude) than would occur
from the removal of every CdX2 group bound to B1 sites.
Ultimately, the brightness of CdSe NCs is exceedingly
sensitive to B2 vacancies, and without the strong
thermodynamic stability of CdX2 complexes bound to these
sites, the use of Cd-based Z-type ligands to passivate
undercoordinated Se sites would unlikely be a successful
strategy to mitigate non-radiative recombination processes.
written, eq. 15 is
a
slightly endothermic reaction,
demonstrating, somewhat surprisingly, that the large
thermodynamic stability of the B2-CdX2 complex relative to
the B1 site analog is intrinsically an entropic effect; further
insight into the origin of this effect will be provided in Section
3.5, along with a discussion of the nature of the B1 and B2 sites
in terms of the specific surfaces of zinc-blende CdSe NCs.
3.4. Impact of Specific Surface Vacancies on PL Efficiency.
With the knowledge of the energetics of CdX2 surface binding
in hands, we proceed to analyze the impact that the
displacement of this group from the NC surfaces has on the
excitonic photoluminescence (PL). As shown in Figure 5(a),
the addition of TMEDA to CdSe NC suspensions dramatically
impacts the PL intensity, an indication that efficient non-
radiative recombination processes are becoming operative in
the presence of TMEDA.29,42 Quantitatively, the PL quenching
efficiency is better characterized through the Stern-Volmer
formalism, whereby the reciprocal of the PL intensities,
normalized to the PL intensity in the absence of the quencher
(I0), as reported in Figure 5(b):
I0
= 1+ KSV
Q
[
]
I
(16)
Interestingly, a systematic trend between the quenching
efficiency of single vacancies and the average NC diameter is
observed, whereby smaller NCs are most efficiently quenched
by each surface defects (Table 4); the same trend is also
observed in the ratios between the Stern-Volmer constants for
the two sites. The magnitude of the effect, which does not
scale directly with the surface-to-volume ratio, might possibly
indicate enhanced coupling to the surface states in more
strongly quantum confined systems; a quantitative evaluation
where [Q] is the quencher species concentration and KSV is a
constant characterizing the efficiency of the PL quenching
process; from the onset of the Stern-Volmer data for TMEDA
presented in Fig. 5(b), KSV = (57,000 ± 1,000) M-1. Of course
TMEDA is not itself the direct PL quenching species, but
rather acts indirectly by creating CdX2 vacancies on the NC
surfaces that are active trap centers, presumably as efficient
hole trapping species.32,42 As the impact of TMEDA on the
displacement of CdX2 has been fully characterized above, a
direct connection between the PL efficiency (which is
indirectly reporting on the presence of trap centers) and the
Table 5. CdSe NCs Surface Areas and Ligand Coverage.
eff
NC
Diameter
(nm)
d
a
b
c
l 111
Stot
(nm2)
S100
(nm2)
S111
(nm2)
N1,tot
(nm-2)
3.0 ± 0.2
3.8 ± 0.2
4.1 ± 0.3
28 ± 3 13 ± 1 15 ± 4 1.5 ± 0.4
45 ± 3 23 ± 1 22 ± 4 1.5 ± 0.3
47 ± 2
69 ± 4
75 ± 7
53 ± 5 29 ± 2 24 ± 7 1.8 ± 0.5
a
b
Surface area of the equivalent sphere; Calculated using eq. (17);
Scheme 2
d
eff
c S111 = Stot – S100
;
.
l 111 = N1 / S111
8
ACS Paragon Plus Environment

Page 9 of 20
1
Journal of the American Chemical Society
of this complicated effect goes beyond the scope of this study
PLQY of the as-prepared samples studied here (avg. PLQY ~
2
3
4
5
and will be reported upon in a future communication.
13%). The total (100) surface area, S100, is thus simply
calculated from:
We note finally that the PLQY of isolated NC samples
(following
repeated
ethyl
acetate/pentane
N2
S100
=
crashing/resuspension cycles) are always markedly higher than
for the same samples in the TMEDA/free CdX2 mixtures (see
SI). Interestingly, an analysis of the surface coverage after
isolation shows a concomitant increase in the surface coverage
which follows precisely the trend observed for the
NC/TMEDA/free CdX2 mixtures (see SI), suggesting that NCs
before and after isolation are behaving identically, provided
proper correction for the change in the surface coverage. This
also suggests that TMEDA does not impact the PLQY of CdSe
NCs, as there are not significant differences observed between
samples with very high concentrations of TMEDA and those
after isolation. The increase in the surface coverage for
isolated NCs is likely attributable to a reduction of the
solubility of the TMEDA-CdX2 complex in the non-
polar/polar solvent mix, which drives the rebinding of the
cadmium complex onto the surface.
l 100 / 2
(17)
6
7
8
9
where λ100 is the atom density of the (100) facet (the factor of
2 arising from the fact that only half of the (100) facets are Se-
terminated and able to coordinate CdX2). The calculated total
area for the (100) and (111) facets are listed in Table 5;
interestingly, the ratio of the (100) vs. (111) areas obtained
from this assumption is close to unity, which is consistent with
the observed spherical morphology (among the isomorphic
Archimedean solids, this specific shape falls in between that of
the ideal truncated octahedron and the cuboctahedron).
Following on our initial assumption that the B2 sites
correspond to vacancies on the (100) facet, we assign the B1
sites (initially bound to N1 CdX2 complexes) to vacancies on
the (111) facet. Since the total surface of each NC sample can
be directly estimated from the average diameter, a direct
quantification of the surface coverage for the (111) facets can
be obtained, as given in Table 5. The surface coverage of the
(111) facets (measured in CdX2 units) is lower than the
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
3.5 Microscopic Nature of the Surface Sites of CdSe NCs.
Some insights into the potential microscopic nature of each of
the two surface sites, B1 and B2 can be obtained from the
results presented above. As has been previously
recognized,25,27,29,32,70,87 the different crystal facets of
semiconductor NCs often exhibit strikingly different behaviors
regarding the stability of surface-bound species (ligands or
adatoms) and the generation of specific surface defects for
excited-state energy-trapping processes. The surface of zinc-
blende CdSe NCs is generally described in terms of the two
most stable facets, namely the (100) and (111) facets.28,32,88,89
The (100) facet exposes two-coordinated species, either Se or
Cd; the idealized “selenium-rich” (100) facet thus presents
rows of Se2- ions, each with two lone pairs located in sp3
hybrids. The (111) facets by contrast expose species that are 3-
coordinated to the underlying lattice atoms; the idealized
“selenium-rich” (111) facet thus presents arrays of Se2- ions
each with a single lone pair pointed perpendicular to the facet
plane (Scheme 2). Although the (100) facet is atomically
sparser than the (111) facet:
l
111 /2 @ 3.1 nm- 2
, which is expected
theoretical maximum,
given the low overall ligand coverage listed in Table 1. In
spite of the relatively large number of (111) sites that remain
unpassivated, it is likely that these sites are actually sterically
hindered, suggesting that the CdX2 coverages reported here are
perhaps actually close to the effective maximum that can be
achieved for the specific carboxylates and NC shapes under
consideration in this study. Importantly, as discussed above,
these unpassivated sites – assigned here as B1 vacancies – do
not significantly impact the brightness of the NCs, minimizing
the detrimental impact of the low (111) facet coverage. With
an estimate of the surface spanned by the (111) facets in hand,
the number of B1 sites that are not initially ligated by CdX2,
which was neglected in the analysis before, can now be
evaluated:
l
N1,tot
=
111 ×S111
2
(18)
2
a02
l 100
=
@ 5.4 atoms×nm- 2
where, as above, we acknowledge that only half of the surface
apportioned to (111) facets is Se-terminated. These values are
listed in Table 5, showing that there are many B1 vacancies on
these as-prepared NC samples; an estimate of the role that
these vacancies might play in the photoluminescence of these
“pristine” CdSe NCs is presented in the next Section. As
mentioned in Section 3.2 above, neglecting the presence of
these B1 vacancies effectively magnifies the values of K1
obtained therein; using the values of N1,tot listed in Table 5 to
analyze the data in Figure 2 leads to values of K1 that are
practically identical for all three sizes: K1,corr = 0.23 ± 0.03.
This actual value, which is 3-5 times smaller than the effective
values listed in Table 1, likely provides a better estimate of the
actual affinity of the B1 sites – that is, undercoordinated (111)
selenide sites – toward CdX2 ligation. We note here that
although we treat surfaces as uniform distributions, it is
possible, and perhaps likely, that specific locales over each
surfaces (say edges or corners) play a dominant role in
defining the ligand exchange kinetics and/or the efficiency of
the recombination processes. Given the likely high mobility of
4 3
3a02
l 111
=
@ 6.2 atoms×nm- 2
where a0 is the lattice constant of zinc blende CdSe
(0.608 nm),90 the (111) facet has a lower density of “dangling
bonds”, and recent studies have suggested that CdX2 vacancies
created on (100) facets of cadmium-rich CdSe NCs lead to the
introduction of electronic states in the bandgap that can
efficiently trap photo-generated valence-band holes on surface
Se2- ions.29, 32 Following this hypothesis, we assign the B2 sites
(the highly efficient PL quenching sites) to vacancies on (100)
facets. Given the large driving force to displace any CdX2
complexes from B1 sites onto B2 vacancies, it is reasonable to
assume that, at equilibrium, the (100) facets must be fully
covered with CdX2 complexes, a proposal which is also
implied by the Stern-Volmer analysis presented above, where
even a single vacancy on the B2 sites would yield a PL
quantum yield (PLQY) < 1/50, much lower than the observed
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 20
1
2
3
4
5
6
7
8
surface ligands over the whole surface of a NC and our
given the free energy cost, any NC with even a single of these
inability to distinguish specific surface sites beyond the two
types (B1 and B2) identified here, we limit our analysis to the
attribution of the chemical and photophysical behavior of
surface defects as mean values over the two main
crystallographic facets.
vacancies would have a significantly lower PLQY than NCs
with only B1/(111) vacancies, and would properly be described
as a “dark” NC compared to the ensemble. The probability
that a B2/(100) vacancy is found, Pdark can be calculated from
the equilibrium of eq. (15), using the number of B1/(111)
vacancies and the revised equilibrium constant K1 (~ 0.2)
obtained in Section 3.5; the calculated values for Pdark are
listed in Table 4, and are all found to be on the order of 1% for
the different NC sizes sampled here. The photoluminescence
of these “dark” NCs can be obtained from the Stern-Volmer
analysis data of Section 3.4:
The conclusion that the (100) facets are fully covered is
consistent with the highly organized nature of these facets,
which lend themselves well to full packing of the CdX2
ligands. On the other hand, as each carboxylate is required to
bridge two Cd atoms, the (100) faceting can only
accommodate half of the carboxylate ligands that are
associated with a full Cd coverage; this “enrichment” of Cd2+
on the selenide-rich (100) surface balances the net negative
charge of these facets, which likely provides a large amount of
the driving force favoring the (100) faceting of these zinc-
blende nanocrystals (everything else being equal, and prior to
reconstruction, (100) facets are always less stable than (111)
facets in diamond-like structures).91 Overall charge neutrality
can be maintained by the transfer of an equal amount of
carboxylates onto the opposite Cd-rich surfaces, which not
only provides a mechanism to passivate the 3-coordinated
cadmium atoms in the absence of L-type ligands, but also
reduces the net positive charge of these facets. Overall, the
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
1
PLQY
=
dark
1+ KSBV1
N
- N +1 + KB
2
SV
(
)
1,tot
1
(20)
where for simplicity only the case of single B2/(100) vacancy
per NC is treated (the probability that a NC is found with more
than one vacancy is lower than 0.1% in all cases). The values
listed in Table 4 show that photoexcited “dark” CdSe NCs are
at least ten times less likely to radiatively recombine than their
“bright” counterparts, with smaller NCs presenting the most
striking dark/bright contrasts. It is important to point out here
that the distinction made here between “dark” and “bright”
NCs is likely not linked to the ubiquitous “blinking”, or PL
intermittency, phenomenon exhibited by most NC
materials,26,92 as the “dark” PLQY values calculated here are
still too large to sufficiently explain that phenomenon. Recent
studies have suggested that PL blinking is probably linked to
electron trapping processes that lead to delayed PL on long
timescale;93-96 such electron trapping processes, the
microscopic origin of which is still under debate, are not taken
into account here, but perhaps explain partly the remaining
discrepancy between the PLQY values we simulate and the
observed ones. At any rate, we emphasize emphatically here
that the proposed model of the overall PLQY discussed in this
Section should not be taken too literally given the complexity
of the actual problem. Rather, these results should be taken as
representing part of the puzzle: inasmuch as the specific
surface sites we uncovered here do exist on CdSe NCs, there
will be unavoidable non-radiative losses associated with these;
using an objective assessment of the actual prevalence of each
site vacancy leads to qualitative agreement with the observed
PLQY, suggesting that our estimate of the non-radiative
recombination efficiencies of each type of site and of the
surface composition are reasonably accurate. In this case,
whereas full coverage of the Se-rich (100) facets appears
absolutely critical to limit efficient non-radiative
recombination in zinc blende CdSe NCs, for carboxylate-
capped species the (100) facets rarely exhibit vacancies, and it
is the low coverage of the (111) facets that effectively dictates
the relatively low PLQY of the as-prepared materials.
results presented herein allow to develop
a clearer
representation of the schematic dual exchange reaction scheme
used in the Introduction (Scheme 1), where the B1 and B2 sites
are localized on specific facets.
3.6 Correlating Specific Facet Vacancies with
Photoluminescence Quantum Yields: Dark vs Bright NCs.
In Section 3.4, we correlated how vacancies generated on each
of the two specific sites, B1 and B2, affected the
photoluminescence efficiency. Having now in hand a better
microscopic description of the nature of each of these sites, we
complete this study by correlating the proposed surface
composition of zinc-blende CdSe NCs with their observed
photoluminescence behavior. Although we found in Section
3.4 that B1/(111) vacancies are not as effective as B2/(100)
vacancies in quenching the PL of CdSe NCs, we were also led
in Section 3.5 to postulate the existence of a rather large
number of B1/(111) vacancies that are present in the as-
prepared samples (that is, before the deliberate displacement
of CdX2 complexes by TMEDA). As a consequence, assuming
that each of these vacancies provides the same types of non-
radiative relaxation pathways than those we found for B1/(111)
vacancies generated by the action of TMEDA, we can readily
estimate the starting PLQY of such as-prepared NCs (which
we label here as “bright” NCs, to distinguish with “dark” NCs
we will soon discuss):
1
PLQY
=
bright
1+ KSBV1
N
- N
(
)
1,tot 1
(19)
5. CONCLUSIONS
where the difference between N1,tot (Table 5) and N1 (Table 1)
corresponds to the number of B1/(111) vacancies per NC, and
we assume that a NC without any vacancies would possess
unity PLQY. The estimated PLQYbright obtained from eq. (19)
are listed in Table 4, and qualitatively compare well with the
observed PLQY of each samples, although being
systematically larger. One additional contribution to the PLQY
of CdSe NCs might arise from the presence of B2/(100)
vacancies. Although such vacancies cannot be predominant
In summary, we have demonstrated a direct correlation
between specific surface defects and non-radiative
recombination processes occurring at the surface of cadmium-
rich zinc-blende CdSe NCs. The model developed here
provides for the first-time direct evidence for the specificity of
different crystalline facets with regards to the free-energy
changes associated with the binding of cadmium carboxylate
complexes, allowing for a detailed account of the distribution
10
ACS Paragon Plus Environment

Page 11 of 20
1
Journal of the American Chemical Society
(17) Razgoniaeva, N.; Moroz, P.; Lambright, S.; Zamkov, M. J.
Phys. Chem. Lett. 2015, 6, 4352-4359.
of these species over the different facets. These results provide
2
3
4
5
insights into the complexity of the surface of CdSe NCs,
which are expected to impact the development of targeted
designs of efficient NC systems for energy-conversion
applications.
(18) Brown, K. A.; Harris, D. F.; Wilker, M. B.; Rasmussen, A.;
Khadka, N.; Hamby, H.; Keable, S.; Dukovic, G.; Peters, J. W.;
Seefeldt, L. C.; King, P. W. Science 2016, 352, 448-450.
(19) Jensen, S. C.; Homan, S. B.; Weiss, E. A. J. Am. Chem. Soc.
2016, 138, 1591-1600.
(20) Caputo, J. A.; Frenette, L. C.; Zhao, N.; Sowers, K. L.;
Krauss, T. D.; Weix, D. J. J. Am. Chem. Soc. 2017, 139, 4250-
4253.
6
7
8
ASSOCIATED CONTENT
Supporting Information
9
The Supporting Information is available free of charge on the
ACS Publications website at DOI:
(21) Pal, A.; Ghosh, I.; Sapra, S.; König, B. Chem. Mater. 2017,
29, 5225-5231.
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
(22) Queisser, H. J.; Haller, E. E. Science 1998, 281, 945-950.
(23) This statement is genelly valid for II-VI semiconductors,
where the ionicity of the lattice lowers the barrier to diffusion
enough to "self-anneal" such defects. There is nevertheless
evidence that the more covalent III-V semiconductors, such as
GaAs, do exhibit such localized lattice defects. See e.g.
Srivastava, V.; Liu, W.; Janke, E. M.; Kamysbayev, V.; Filatov,
A. S.; Sun, C.-J.; Lee, B.; Rajh, T.; Schaller, R. D.; Talapin, D. V.
Nano Lett. 2017, 17, 2094-2101.
(24) Smith, A. M.; Nie, S. Acc. Chem. Res. 2010, 43, 190-200.
(25) Boles, M. A.; Ling, D.; Hyeon, T.; Talapin, D. V. Nat.
Mater. 2016, 15, 141-153.
(26) Gomez, D. E.; van Embden, J.; Jasieniak, J.; Smith, T. A.;
Mulvaney, P. Small 2006, 2, 204-208.
(27) Voznyy, O. J. Phys. Chem. C 2011, 115, 15927-15932.
(28) Voznyy, O.; Thon, S. M.; Ip, A. H.; Sargent, E. H. J. Phys.
Chem. Lett. 2013, 4, 987-992.
(29) Busby, E.; Anderson, N. C.; Owen, J. S.; Sfeir, M. Y. J.
Phys. Chem. C 2015, 119, 27797-27803.
(30) Krause, M. M.; Kambhampati, P. Phys. Chem. Chem. Phys.
2015, 17, 18882-18894.
(31) Kilina, S. V.; Tamukong, P. K.; Kilin, D. S. Acc. Chem. Res.
2016, 49, 2127-2135.
(32) Houtepen, A. J.; Hens, Z.; Owen, J. S.; Infante, I. Chem.
Mater. 2017, 29, 752-761.
(33) Hines, M. A.; Guyot-Sionnest, P. J. Phys. Chem. 1996, 100,
468-471.
(34) Chen, Y.; Vela, J.; Htoon, H.; Casson, J. L.; Werder, D. J.;
Bussian, D. A.; Klimov, V. I.; Hollingsworth, J. A. J. Am. Chem.
Soc. 2008, 130, 5026-5027.
(35) Mahler, B.; Spinicelli, P.; Buil, S.; Quelin, X.; Hermier, J.-P.;
Dubertret, B. Nat. Mater. 2008, 7, 659-664.
(36) Reiss, P.; Protière, M.; Li, L. Small 2009, 5, 154-168.
(37) Chen, O.; Zhao, J.; Chauhan, V. P.; Cui, J.; Wong, C.; Harris,
D. K.; Wei, H.; Han, H.-S.; Fukumura, D.; Jain, R. K.; Bawendi,
M. G. Nat. Mater. 2013, 12, 445-451.
(38) Dannhauser, T.; O'Neil, M.; Johansson, K.; Whitten, D.;
McLendon, G. J. Phys. Chem. 1986, 90, 6074-6076.
(39) Donakowski, M. D.; Godbe, J. M.; Sknepnek, R.; Knowles,
K. E.; Olvera de la Cruz, M.; Weiss, E. A. J. Phys. Chem. C 2010,
114, 22526-22534.
(40) Morris-Cohen, A. J.; Donakowski, M. D.; Knowles, K. E.;
Weiss, E. A. J. Phys. Chem. C 2010, 114, 897-906.
(41) Wei, H. H.-Y.; Evans, C. M.; Swartz, B. D.; Neukirch, A. J.;
Young, J.; Prezhdo, O. V.; Krauss, T. D. Nano Lett. 2012, 12,
4465-4471.
Additional NMR data, PL quenching, data analysis for all
sizes of nanoparticles, nanoparticle characterization, XRD,
TEM, and FTIR (PDF)
AUTHOR INFORMATION
Corresponding Author
* beaulac@msu.edu
ACKNOWLEDGMENTS
Support from Michigan State University and its Center of Re-
search Excellence on Complex Materials (CORE-CM) is grateful-
ly acknowledged. The authors thank Dr. Daniel Holmes for help
and support with the NMR instrumentation and analyses, and
Profs. Benjamin G. Levine, Aaron L. Odom, and Jonathan S.
Owen for insightful discussions. E.P.A. acknowledges support
from the NSF “REU Site: Cross-Disciplinary Training in Sustain-
able Chemistry and Chemical Processes” program, under award
number CHE-1358842.
REFERENCES
(1) Bawendi, M. G.; Steigerwald, M. L.; Brus, L. E. Ann. Rev.
Phys. Chem. 1990, 41, 477-496.
(2) Klimov, V. I. Nanocrystal Quantum Dots; CRC Press: Boca
Raton, FL, 2010.
(3) Sargent, E. H. Adv. Mater. 2008, 20, 3958-3964.
(4) Talapin, D. V.; Lee, J.-S.; Kovalenko, M. V.; Shevchenko, E.
V. Chem. Rev. 2010, 110, 389-458.
(5) Colloidal Quantum Dot Optoelectronics and Photovoltaics;
Konstantatos, G.; Sargent, E. H., Eds.; Cambridge University
Press, 2013.
(6) Wood, V.; Bulovic, V. Nano Rev. 2010, 1, 5202-5208.
(7) Supran, G. J.; Shirasaki, Y.; Song, K. W.; Caruge, J.-M.;
Kazlas, P. T.; Coe-Sullivan, S.; Andrew, T. L.; Bawendi, M. G.;
Bulović, V. MRS Bull. 2013, 38, 703-711.
(8) Talapin, D. V.; Steckel, J. MRS Bull. 2013, 38, 685-691.
(9) Kambhampati, P.; Mack, T.; Jethi, L. ACS Photonics 2017, 4,
412-423.
(10) Kamat, P. V. J. Phys. Chem. Lett. 2013, 4, 908-918.
(11) Kramer, I. J.; Sargent, E. H. Chem. Rev. 2014, 114, 863-882.
(12) Alivisatos, P. Nat. Biotechnol. 2004, 22, 47-52.
(13) Michalet, X.; Pinaud, F. F.; Bentolila, L. A.; Tsay, J. M.;
Doose, S.; Li, J. J.; Sundaresan, G.; Wu, A. M.; Gambhir, S. S.;
Weiss, S. Science 2005, 307, 538-544.
(14) Chang, J. C.; Rosenthal, S. J. J. Phys. Chem. Lett. 2013, 4,
2858-2866.
(15) Han, Z.; Qiu, F.; Eisenberg, R.; Holland, P. L.; Krauss, T. D.
Science 2012, 338, 1321-1324.
(16) Wilker, M. B.; Shinopoulos, K. E.; Brown, K. A.; Mulder, D.
W.; King, P. W.; Dukovic, G. J. Am. Chem. Soc. 2014, 136, 4316-
4324.
(42) Anderson, N. C.; Hendricks, M. P.; Choi, J. J.; Owen, J. S. J.
Am. Chem. Soc. 2013, 135, 18536-18548.
(43) Vempati, S.; Ertas, Y.; Uyar, T. J. Phys. Chem. C 2013, 117,
21609-21618.
(44) Fang, C.; van Huis, M. A.; Vanmaekelbergh, D.;
Zandbergen, H. W. ACS Nano 2010, 4, 211-218.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 20
1
2
3
4
5
6
7
8
(45) Choi, J. J.; Bealing, C. R.; Bian, K.; Hughes, K. J.; Zhang,
W.; Smilgies, D.-M.; Hennig, R. G.; Engstrom, J. R.; Hanrath, T.
J. Am. Chem. Soc. 2011, 133, 3131-3138.
(46) Bealing, C. R.; Baumgardner, W. J.; Choi, J. J.; Hanrath, T.;
Hennig, R. G. ACS Nano 2012, 6, 2118-2127.
(47) Zhou, J.; Pu, C.; Jiao, T.; Hou, X.; Peng, X. J. Am. Chem.
Soc. 2016, 138, 6475-6483.
(48) Giansante, C.; Infante, I. J. Phys. Chem. Lett. 2017, 8, 5209-
5215.
(49) Chen, O.; Yang, Y.; Wang, T.; Wu, H.; Niu, C.; Yang, J.;
Cao, Y. C. J. Am. Chem. Soc. 2011, 133, 17504-17512.
(50) Berrettini, M. G.; Braun, G.; Hu, J. G.; Strouse, G. F. J. Am.
Chem. Soc. 2004, 126, 7063-7070.
(73) Dzhagan, V. M.; Lokteva, I.; Himcinschi, C.; Kolny-Olesiak,
J.; Valakh, M. Y.; Schulze, S.; Zahn, D. R. T. J. Appl. Phys. 2011,
109, 084334-084341.
(74) von Holt, B.; Kudera, S.; Weiss, A.; Schrader, T. E.; Manna,
L.; Parak, W. J.; Braun, M. J. Mater. Chem. 2008, 18, 2728-2732.
(75) Karel Čapek, R.; Moreels, I.; Lambert, K.; De Muynck, D.;
Zhao, Q.; Van Tomme, A.; Vanhaecke, F.; Hens, Z. J. Phys.
Chem. C 2010, 114, 6371-6376.
(76) Chen, O.; Yang, Y.; Wang, T.; Wu, H.; Niu, C.; Yang, J.;
Cao, Y. C. J. Am. Chem. Soc. 2011, 133, 17504-17512.
(77) Owen, J. S.; Park, J.; Trudeau, P.-E.; Alivisatos, A. P. J. Am.
Chem. Soc. 2008, 130, 12279-12281.
(78) Green, M. L. H. J. Organomet. Chem. 1995, 500, 127-148.
(79) Parkin, G. Classification of Organotransition Metal
Compounds. In Comprehensive Organometallic Chemistry III;
Crabtree, R. H., Mingos, D. M. P., Eds.; Elsevier: Oxford, 2007,
p. 1-57.
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
(51) Moreels, I.; Fritzinger, B.; Martins, J. C.; Hens, Z. J. Am.
Chem. Soc. 2008, 130, 15081-15086.
(52) Anderson, N. C.; Owen, J. S. Chem. Mater. 2013, 25, 69-76.
(53) Hens, Z.; Martins, J. C. Chem. Mater. 2013, 25, 1211-1221.
(54) Knittel, F.; Gravel, E.; Cassette, E.; Pons, T.; Pillon, F.;
Dubertret, B.; Doris, E. Nano Lett. 2013, 13, 5075-5078.
(55) Morris-Cohen, A. J.; Malicki, M.; Peterson, M. D.; Slavin, J.
W. J.; Weiss, E. A. Chem. Mater. 2013, 25, 1155-1165.
(56) Li, X.; Nichols, V. M.; Zhou, D.; Lim, C.; Pau, G. S. H.;
Bardeen, C. J.; Tang, M. L. Nano Lett. 2014, 14, 3382-3387.
(57) De Roo, J.; Justo, Y.; De Keukeleere, K.; Van den Broeck,
F.; Martins, J. C.; Van Driessche, I.; Hens, Z. Angew. Chem., Int.
Ed. 2015, 54, 6488-6491.
(58) Owen, J. Science 2015, 347, 615-616.
(59) Shen, Y.; Tan, R.; Gee, M. Y.; Greytak, A. B. ACS Nano
2015, 9, 3345-3359.
(60) Knauf, R. R.; Lennox, J. C.; Dempsey, J. L. Chem. Mater.
2016, 28, 4762-4770.
(61) Chen, P. E.; Anderson, N. C.; Norman, Z. M.; Owen, J. S. J.
Am. Chem. Soc. 2017, 139, 3227-3236.
(62) De Roo, J.; De Keukeleere, K.; Hens, Z.; Van Driessche, I.
Dalton Trans. 2016, 45, 13277-13283.
(63) Morris-Cohen, A. J.; Frederick, M. T.; Lilly, G. D.;
McArthur, E. A.; Weiss, E. A. J. Phys. Chem. Lett. 2010, 1, 1078-
1081.
(64) Hassinen, A.; Moreels, I.; de Mello Donegá, C.; Martins, J.
C.; Hens, Z. J. Phys. Chem. Lett. 2010, 1, 2577-2581.
(65) Gomes, R.; Hassinen, A.; Szczygiel, A.; Zhao, Q.;
Vantomme, A.; Martins, J. C.; Hens, Z. J. Phys. Chem. Lett. 2011,
2, 145-152.
(66) Gao, Y.; Peng, X. J. Am. Chem. Soc. 2014, 136, 6724-6732.
(67) Tavasoli, E.; Guo, Y.; Kunal, P.; Grajeda, J.; Gerber, A.;
Vela, J. Chem. Mater. 2012, 24, 4231-4241.
(68) Kim, W.; Lim, S. J.; Jung, S.; Shin, S. K. J. Phys. Chem. C
2010, 114, 1539-1546.
(69) Lim, S. J.; Kim, W.; Shin, S. K. J. Am. Chem. Soc. 2012,
134, 7576-7579.
(70) Manna, L.; Wang; Cingolani, R.; Alivisatos, A. P. J. Phys.
Chem. B 2005, 109, 6183-6192.
(71) Wuister, S. F.; de Mello Donegá, C.; Meijerink, A. J. Phys.
Chem. B 2004, 108, 17393-17397.
(80) Yang, Y. A.; Wu, H.; Williams, K. R.; Cao, Y. C. Angew.
Chem., Int. Ed. 2005, 44, 6712-6715.
(81) Chen, O.; Chen, X.; Yang, Y.; Lynch, J.; Wu, H.; Zhuang, J.;
Cao, Y. C. Angew. Chem., Int. Ed. 2008, 47, 8638-8641.
(82) Jasieniak, J.; Smith, L.; van Embden, J.; Mulvaney, P.;
Califano, M. J. Phys. Chem. C 2009, 113, 19468-19474.
(83) De Nolf, K.; Cosseddu, S. M.; Jasieniak, J. J.; Drijvers, E.;
Martins, J. C.; Infante, I.; Hens, Z. J. Am. Chem. Soc. 2017, 139,
3456-3464.
(84) Fritzinger, B.; Capek, R. K.; Lambert, K.; Martins, J. C.;
Hens, Z. J. Am. Chem. Soc. 2010, 132, 10195-10201.
(85) Hassinen, A.; Moreels, I.; De Nolf, K.; Smet, P. F.; Martins,
J. C.; Hens, Z. J. Am. Chem. Soc. 2012, 134, 20705-20712.
(86) Cass, L. C.; Malicki, M.; Weiss, E. A. Anal. Chem. 2013, 85,
6974-6979.
(87) Voznyy, O.; Sargent, E. H. Phys. Rev. Lett. 2014, 112,
157401-157404.
(88) Liu, L.; Zhuang, Z.; Xie, T.; Wang, Y.-G.; Li, J.; Peng, Q.;
Li, Y. J. Am. Chem. Soc. 2009, 131, 16423-16429.
(89) Peng, Z. A.; Peng, X. J. Am. Chem. Soc. 2001, 123, 1389-
1395.
(90) Madelung, O. Semiconductors: Data Handbook; 3rd ed.;
Springer-Verlag: New York, NY, 2004, 229.
(91) Harrison, W. A. Electronic Structure and the Properties of
Solids; Dover: New York, NY, 1989, 230-233.
(92) Frantsuzov, P.; Kuno, M.; Janko, B.; Marcus, R. A. Nat.
Phys. 2008, 4, 519-522.
(93) Rabouw, F. T.; Kamp, M.; van Dijk-Moes, R. J. A.; Gamelin,
D. R.; Koenderink, A. F.; Meijerink, A.; Vanmaekelbergh, D.
Nano Lett. 2015, 15, 7718-7725.
(94) Rabouw, F. T.; van der Bok, J. C.; Spinicelli, P.; Mahler, B.;
Nasilowski, M.; Pedetti, S.; Dubertret, B.; Vanmaekelbergh, D.
Nano Lett. 2016, 16, 2047-2053.
(95) Marchioro, A.; Whitham, P. J.; Knowles, K. E.; Kilburn, T.
B.; Reid, P. J.; Gamelin, D. R. J. Phys. Chem. C 2016, 120,
27040-27049.
(96) Tsui, E. Y.; Carroll, G. M.; Miller, B.; Marchioro, A.;
Gamelin, D. R. Chem. Mater. 2017, 29, 3754-3762.
(72) Li, R.; Lee, J.; Yang, B.; Horspool, D. N.; Aindow, M.;
Papadimitrakopoulos, F. J. Am. Chem. Soc. 2005, 127, 2524-
2532.
12
ACS Paragon Plus Environment

Page 13 of 20
Journal of the American Chemical Society
1
2
Table of Content Figure
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
13
ACS Paragon Plus Environment

Wavelength (nm)
Journal of the American ChemPaicgael S1o4coieft2y0
400 450 500 600 700
0.6
1
0.4
2
3
4
0.2
5
6
CdSe NC
CdSe NC + TMEDA
0.0
ACS Paragon Plus Environment
7
8
9
25000
20000
15000
Wavenumber (cm-1)

[CdX2]tot
[CdX2]B
B
Page 15 of 20
Journal of the American Chemical Society
(a)
(b)
(c)
F
100
80
60
40
20
0
[TMEDA]
0
0.35
1
12.05
2.1
3
4.2
4
525
[CdX2]F
6
50
7
137
8
274
9
ACS Paragon Plus Environment
10
11
12
8
7
6
5
4
3
2
1
6.0 5.8 5.6 5.4
0
100
200
300
[TMEDA] (mM)
Chemical Shift (ppm)
Chemical Shift (ppm)

[TMEDA] = [L] + [L-CdX2] (mM)
[TMEDA] / [CdX2]
log10([TMEDA] / [CdX2])
Journal of the American Chemical Society
Page 16 of 20
0
100
200
300
0
10
20
30
40
-1
0
1
2
3
4
20
18
16
14
12
10
50
50
40
30
20
10
100
80
60
40
20
0
(a)
(b)
(c)
B1 + B2
B1 + B2
1
2
3
4
5
6
7
8
9
60
70
B1
B2
B1
80
90
B2
100 ACS Pa0ragon Plus Environment
0
20 40 60 80 100
0
100
200
300
-1
0
1
2
3
4
5
10
11
12
φ = [L] / [L-CdX2]
log10([TMEDA] / 1 mM)
[TMEDA] (mM)

Temperature (K)
Page J1o7uornf 2a0l of the American Chemical Society
340
320
300
280
260
6
4
1
2
3
4
5
6
7
8
9
K2
2
0
K1
-2
ACS Paragon Plus Environment
3.0 3.2 3.4 3.6
Reciprocal Temperature (x10-3 K-1)
10
11
12
2.8
3.8

Journal of the American Chemical Society
Page 18 of 20
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
1.0
0.8
0.6
0.4
0.2
0.0
60
(b)
60
50
40
30
20
10
0
0 mM
(a)
(c)
50
40
30
20
10
0
B2 vacancies
TMEDA
B1 vacancies
4 mM
625
525
575
Wavelength (nm)
0
1
2
3
4
0
1
2
3
4
0
10
20
30
CdX2 groups removed
[TMEDA] (mM)
[TMEDA] (mM)
ACS Paragon Plus Environment

B1
ourPnaaglBe1o1fC9tdhXoefA2m0 erican Chemical Societ
2
CdSe
NC
CdSe
NC
B2
B2
CdX2
1
2
3
4
L5=
6
7
8
LCdX2
LCdX2
LCdX2
L
LCdX2
L
B1
CdSe
NC
X- =
O
ACS Paragon Plus Environment
B2
N
O
7
N
CdX2
7

Journal of the American ChemPicaagleS2o0cieotfy20
Se
Se
Cd
Cd
Se
ACS Paragon Plus Environment
Cd
Se
Se
Cd
Cd
Cd
Cd
Cd
Cd
Cd
(111) facet : B1 sites
(100) facet: B2 sites
1