Journal of the American Chemical Society
Communication
adatom addition. Other unknown factors (e.g., surface-8a or
ligand-mediated processes,8b see Supporting Information) likely
play key roles. Yet, in all cases, product crystal structures
replicated those of the cores (Figure S2). For alloys, this
suggests that cation substitution proceeds without lattice
rearrangement. (Zinc blende products are ∼tetradedral,8a and
wurtzite are more symmetric; see Figure 2a,c insets.)
To evaluate the stability of CdSe/ZnSe core/shell QDs to
alloying postsynthesis, we annealed the products at high
temperature (310 °C) and exposed them to high pressure (10.7
GPa). Annealing does not cause appreciable alloying (minimal
PL blue-shifting, Figure 2f); whereas, under high pressure, the
core and shell behave independently. Namely, consistent with
literature,11 the wurtzite CdSe core transforms to the rock salt
crystal structure above ∼5.6 GPa and then to ∼zinc blende
upon return to 0 GPa (Figures 2g and S3). As expected for
ZnSe (phase-transition pressure ≈ 13 GPa12), the shell is
unaffected by our experiment (∼10 GPa). The postsynthesis
integrity of core/shell CdSe/ZnSe further supports the
hypothesis that surface-related processes during shelling are
key to effecting shell vs alloy formation (see extended
discussion, Supporting Information).
formation of ∼0.4 nm CdSe shell by cation exchange) or an
alloy of composition Zn0.4Cd0.6Se. Thus, cation exchange causes
composition redshifting that outpaces electronic-structure
effects.
We modified SILAR growth to avoid cation exchange: lower
temperatures or use of a single-source (SS) precursor
(cadmium diethyldithiocarbamate). Lower temperatures limit
cation exchange as indicated by spectral-shift and elemental
analysis data (Figure 3c). However, PL quantum yields are
reduced (<10% vs >40%, Supporting Information). A SS
precursor eliminates Zn−Cd exchange, but optical properties
suffer and the particle shape is asymmetric (rod-like, Figure
S4).
Finally, CdSe/ZnSe and ZnSe/CdS gQDs were assessed for
single-dot optical properties: blinking and relative biexciton
emission efficiency as determined by second-order photon-
correlation (g(2)) experiments (Supporting Information).16
Individual CdSe/ZnSe gQDs exhibit PL intensity fluctuation,
but rarely blink all the way off (Figure 4a, on-time fractions =
In contrast with CdSe/ZnSe QDs, the key challenge to
ZnSe/CdS core/(thick)shell synthesis is to avoid Cd
exchanging for Zn. Almost complete cation exchange has
been shown for specific experimental conditions, i.e., high
temperature (220 °C) and excess Cd,13 and is supported by the
more favorable Cd−Se compared to Zn−Se bond energy (310
vs 136 kJ/mol).14 SILAR growth does not entail using excess
Cd, and the inclusion of a S precursor means that Cd−S bond
formation would compete with Zn−Se bond breaking/Cd−Se
bond making. Thus, we did not a priori expect cation exchange
to interfere with shell growth. After all, ZnSe/CdS core/
(thin)shell QD synthesis had been reported previously.15
However, optical and elemental analysis data suggest that
cation exchange is active in the ZnSe/CdS system. The large
redshift in absorption and PL after addition of a single CdS ML
(Figure 3a,b, >130 nm, 240 °C SILAR) is inconsistent with
Figure 4. (a) Representative PL time trace for a CdSe/ZnSe gQD and
(b) corresponding g(2) trace (obtained for average per-dot carrier
population < 0.2; see Supporting Information). (c and d) Same for a
ZnSe/CdS gQD. Dashed orange line in a and c is detection limit
(below which gQD is off state).
0.4−0.8 and Figure S5). The large fluctuations between bright
and less-intense gray states are the subject of further
investigation, but may result from severe QD charging. In
contrast to conventional QDs, charged gQDs are emissive
(rather than dark) as a result of suppressed AR, where the
efficiency of the charged-state emission depends on the extent
of AR suppression17a and/or the degree of charging.17b Also
indicative of reduced AR efficiency, this system is characterized
by high g(2) values (here ∼0.4, Figure 4b), as anticipated for the
combination of a quasi-type II electronic structure and thick-
shell.
ZnSe/CdS gQDs grown at high temperature exhibit strongly
suppressed shell-ML-dependent blinking, with the fraction of
nonblinking gQDs approaching 60% for long interrogation
times (Figures 4c and S6), though g(2) values are unremarkable
(Figure 4d). In contrast, other ZnSe/CdS QDs blink and
rapidly photobleach. Thus, high-temperature growth paired
with partial cation exchange substantially improves optical
performance in this system.
Figure 3. (a) Absorption (inset: TEM) and (b) PL spectra for ZnSe
core (black trace) and after CdS shelling (red trace: 1 ML, then even
MLs 2−12). (c) Summary of PL progression with CdS-shell ML for
different shelling temperatures. Zn/Se ratio (2.3 nm radius core) = 0.6
(240 °C), 0.6 (215 °C), 0.7 (2 ML at 150 °C then 215 °C to 12 ML),
1.0 (215 °C, SS precursor), and 0.8 (160 °C).
calculations considering type II electronic structure alone. The
final PL peak position (627 nm) is also red-shifted compared to
theory, which requires a bandgap of 2.1 0.05 eV (577−605
nm) for an ideal core/(thick) shell product. Elemental analysis
reveals a Zn/Se ratio of 0.4, rather than the 1:1 of an intact
ZnSe core. We surmise that the core transforms to either a
ZnSe/CdSe core/shell (3.1 nm ZnSe reduced to 2.3 nm with
We have shown two new examples of blinking/AR-
suppressed QDs based on the gQD core/shell motif and type
II electronic structure. Through careful control of shelling
conditions, competing processes of alloying or cation exchange
can be minimized or tuned for optimal performance.
3757
J. Am. Chem. Soc. 2015, 137, 3755−3758