exposing the QD film to pulsed 400-nm exci-
tation (supplementary text S5). To quantify the
level of charging, we monitor the PL dynamics
(Fig. 2A) and determine an average per-dot
number of photoinjected electrons (
the amplitude of the slow neutral-exciton–related
component ( 9 )(insetofFig.2C).Inthelasing
experiments, photodoping occurs in situ because
of the cg-QD film exposure to pump pulses used
to excite population inversion. All photodoping
and lasing experiments are conducted in the
presence of ambient oxygen.
In Fig. 2A, we show PL dynamics of the cg-QD
film immersed into 0.2 M LiEt 3 BH/THF recorded
at different times (t) after the onset of photo-
reduction. During the first 5 min, the PL peak
intensity increases and the PL lifetime becomes
longer (Fig. 2A, red circles), which is accompanied
byanincreaseofthePLQYfrom~15to~60%.
These changes can be attributed to saturation
of intragap surface traps by injected electrons as
observed in previous QD studies ( 9 , 16 ). Att>5
min, PL decay accelerates suggesting that at this
point, the intragap states are fully filled, and the
injected electrons begin toaccumulate in intrinsic
quantized states, which activates nonradiative
Auger recombination (Fig. 2A, blue squares). After
a quick initial growth, <ne>stabilizesat~1.2(Fig.
2C), which indicates establishment of equilibrium
between charging and ambient-oxygen–induced
discharging. The maximal charging level depends
on multiple factors, including, in particular, the
LiEt 3 BH concentration and the excitation power
(figs. S5 to S7 and supplementary text S6). Be-cause reductant is not replenished during the
measurements, eventually it becomes depleted, at
which point discharging starts to overwhelm charg-
ing, leading to slowing of the PL decay (Fig. 2A).
To investigate the effect of charging on lasing
performance, we deposit cross-linked cg-QD
films of ~200-nm thicknessHon top of SiO 2 DFB
resonators fabricated by laser interferometric li-
thography (Fig. 3A, fig. S8, and supplementary text
S7). The DFB grating parameters (Fig. 3B) are
selected to provide strong in-plane feedback via
second-order diffraction;this simultaneously leads
tosurfaceemissioninthenear-normaldirection
due to first-order diffraction (Fig. 3A and fig. S9)
( 17 ). In our lasing experiments, cg-QD/DFB resona-
tors are immersed into a cuvette with THF and
excited at 400 nm by ~130-fs laser pulses (Fig. 3A
and supplementary text S7). In the absence of the
photoreductant, the lasing threshold (jp,las)is
~9mJcm−^2 , as indicated by the transition from
a broad PL band modulated by cavity resonances
to a single narrow feature with the resolution-
limited linewidth of <0.2 nm (Fig. 3C). This is
accompanied by a sharp growth in the emission
intensity, which increases by two orders of mag-
nitude asjpchanges by only 20% (Fig. 4A, black
circles). To quantify lasing thresholds in terms of
<N>, we analyze the pump-intensity dependence
of long-time (Dt>>tXX) PL and transient absorp-
tion (TA) signals (fig. S10). On the basis of this
analysis,jp,lasof 9mJcm−^2 corresponds to <Nlas>
of 1.3 ± 0.3. As expected, this is slightly above the
neutral-exciton gain threshold of ~1.2 estimated
for our cg-QD samples (supplementary text S8).
To initiate cg-QD charging, we add 0.2 M
LiEt 3 BH to the cuvette with the cg-QD/DFB sam-
ple. Keeping the cg-QD layer under continuous
illumination with 400-nm femtosecond pulses,
we wait for ~30 min, at which point the lasing
threshold stabilizes at a sub–single-exciton level of
0.5 ± 0.1 (Fig. 4A, blue squares). The effect of charg-
ing on the lasing threshold is illustrated in Fig. 4B.
The neutral QD sample excited with <N>=1does
not exhibit any signatures of lasing. However,
after the sample is charged, a lower pump level
of <N> = 0.6 leads to intense lasing at 626.5 nm.
To test the reversibility of the effects of photo-
doping, we initiate sample discharging by turn-
ing off the excitation and replacing the LiEt 3 BH
solution with neat THF. After the sample is
kept in dark for an hour, the lasing threshold re-
turns to its near original value of 1.2 ± 0.3 (jp,las
~8mJcm−^2 ; Fig. 4C), suggesting that the cg-QDs
recover their precharging properties. We then
repeat the photodoping procedure using a greater
amount of LiEt 3 BH (0.4 M). After a 60-min ex-
posure to laser light, <Nlas> decreases again to
the sub–single-exciton value, which now reaches
0.31 ± 0.07 (jp,las=2.1mJcm−^2 ; Fig. 4C, main panel
and inset). This is less than a quarter of <Nlas>
for neutral cg-QDs and almost half the value for a
smaller amount (0.2 M) of LiEt 3 BH. This indi-
cates a more complete bleaching of ground-state
absorption due to the increased level of doping.
However, a further increase in the amount of
LiEt 3 BH does not lead to an additional reduction
of the lasing threshold. Indeed, the use of 1 MKozlovet al.,Science 365 , 672–675 (2019) 16 August 2019 3of4
Fig. 4. Sub–single-exciton lasing realized with charged cg-QDs.(A) Intensity of surface emission
of a cg-QD/DFB device (spectrally integrated signal around the lasing wavelength) as a function of
pump fluence. Black circles are for neutral dots and blue squares and red triangles are for dots
charged using 0.2 and 0.4 M LiEt 3 BH/THF, respectively (corresponding
As
the latter corresponds to
not lead to lasing (left panel). However, upon charging with
at a lower pump level
threshold. The lasing threshold decreases from
exposure to 0.2 M LiEt 3 BH (
dark in neat THF for 60 min,
using a greater amount of LiEt 3 BH (0.4 M) leads to even stronger reduction of the lasing threshold
to
corresponds to sub–single-exciton lasing thresholds. The sharp, more than fourfold reduction of
the lasing threshold upon charging is apparent from the inset, which shows a linear–linear plot of the
surface emission intensity versusjpfor neutral (black) and charged (red,
RESEARCH | REPORT