was located at 495 nm, corresponding to the formation of InP
nanocrystals of approximately 2 nm in diameter.[15] With an
increasing ratio of myristic acid to indium the first absorption
feature shifted to longer wavelengths of up to 650 nm,
corresponding to InP nanocrystals of 4.3 nm in diameter.
The wide-angle X-ray scattering (WAXS) confirmed a zinc
blende InP structure, and Scherrer analysis of the peak shapes
corresponded to the appropriate InP nanocrystal crystalline
coherence lengths. In addition, the InP nanocrystal size and
shape was probed by transmission electron microscopy
(TEM), and further corroborated our assignment of size
and structure to the InP nanocrystals.
(TMS)3P flows), providing a 15-second residence time per
injection, and 4 minutes (for 5 mLminÀ1 additional flows),
providing a 40-second residence time per injection. The
temperature at the additional injection points was 808C, and
for the aging process was 3208C. This continuous sequential
injection process resulted in a growth of the first absorption
peak from 495 nm to 595 nm corresponding to a size increase
from 2 to 3.2 nm while maintaining a homogeneous size
distribution (Figure 4). The growth of InP nanocrystals,
The addition of excess myristic acid was found to be the
dominant experimental parameter in the control of InP
nanocrystal size. The excess myristic acid may promote the
dissolution of active nonmolecular InP species, such as
monomers or small clusters, from the InP nanocrystal surface.
The active InP species can subsequently act as a source of
precursors for the growth of InP nanocrystals in a classical
ripening process.[16] However, other nonmolecular processes
such as the coalescence of particles may also contribute to the
growth process. The drastic effect of free myristic acid on
particle size is consistent with an interparticle ripening model
for InP nanocrystal growth.
Another route to the synthesis of larger InP nanocrystals
is the subsequent injection of additional molecular precursors.
As the molecular phosphorus precursors are immediately
depleted following mixing, additional injections can be a
source of In or P precursors.[7] By using a method analogous to
the SILAR method of overcoating nanocrystals, we alter-
natively supply additional monomers of (TMS)3P and
In(MA)3 through six injection ports. In these reaction
schemes, we utilized a continuous three-stage microfluidic
system which utilizes the third reactor stage for sequential
injections (Figure 1c) following the mixing and aging stages.
InP nanocrystals of approximately 2 nm in diameter were
produced in the first two reaction stages, and then this
solution was directly injected into the third sequential
injection microreactor for further growth. In(MA)3 and
(TMS)3P were injected through six alternating subinjections.
The flow resistance of each of the side injections was made to
be one order higher than the resistance of the main channel by
narrowing the channel widths to 80 mm and elongating the
channel lengths to obtain uniformly distributed injections and
to prevent any backflows. The flow resistances were calcu-
lated with a series solution of Navier–Stokes equation for
rectangular channel dimensions.
Initially, InP nanocrystals were synthesized from 50 mm
In(MA)3 and 25 mm (TMS)3P with a 30 mLminÀ1 flow rate at
3208C aging temperature. As a result of the enhanced mixing
in supercritical octane, only brief residence times were
necessary between alternating injections of 80 mm of
In(MA)3 and 50 mm (TMS)3P (in total six injections). The
amount of the additional In and P precursors that were added
was controlled by tuning the injection flow rates from 5–
30 mLminÀ1, corresponding to a ratio of additional (TMS)3P
to initial (TMS)3P ranging from 0.3–2.0. The total residence
time at the sequential injection reactor varied from
1.5 minutes (for 30 mLminÀ1 additional In(MA)3 and
Figure 4. Absorption spectra of InP nanocrystals for various injection
flow rates of In(MA)3 and (TMS)3P in the sequential injection stage of
the microreactor. The InP nanocrystals were synthesized using a
temperature gradient in the mixing stage followed by aging at 3208C
in the aging stage. Spectra are offset for clarity; absorbance is valid for
the lower spectrum.
through the method of sequential injection, allows for precise
control over the growth of larger InP nanocrystals with size
distributions as narrow, or narrower, than the InP nano-
crystals grown by the ripening process.
In summary, we have developed a continuous three-stage
microfluidic system that separates the mixing, aging, and
subsequent injection stages of InP nanocrystal synthesis. The
microfluidic system operates at high temperature and high
pressure enabling the use of solvents such as octane operating
in the supercritical regime for high diffusivity resulting in the
production of high-quality InP nanocrystals in as little as
2 minutes. We have found that the synthesis of InP nano-
crystals is largely independent of many experimental param-
eters that are significant in II–VI CdSe nanocrystal syntheses,
such as mixing temperature and reagent concentrations. The
dominant experimental parameter in the synthesis of InP
nanocrystals is the concentration of free myristic acid in
solution, which significantly contributes to the degree of
interparticle ripening processes. We speculate InP nanocrystal
growth is dominated by nonmolecular processes such as the
coalescence of particles and interparticle ripening. This work
will help in the design of future III–V nanocrystal syntheses,
as reagents that promote interparticle ripening process may
Angew. Chem. Int. Ed. 2011, 50, 627 –630
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
629