A high-absorption cross-section not only implies strong stimulated emission, but also guarantees that absorption from the first excited state to higher excited states is correspondingly weaker.
the first excited state to higher excited states is correspondingly
weaker. Molecules with small fluorescence quantum yields
undergo nonradiative processes efficiently, and our hope was
that one or more of these processes would lead to the initiation
of polymerization.
We tested a broad range of dye molecules for their ability to
initiate polymerization, either upon linear absorption of blue
or ultraviolet light or upon absorption of multiple photons of
near-infrared light. The dye molecules were chosen to prevent
linear absorption from occurring at the wavelength of the
near-infrared light. Under these circumstances, multiphoton
absorption can only occur when the photons that are required
for an absorption event are in the same place at the same time,
leading to a nonlinear intensity dependence. Thus, in multiphoton absorption polymerization (MAP), the combination
of this optical nonlinearity with the chemical nonlinearity we
discussed earlier allows photopolymerization to be confined
to the focal volume of a tightly-focused laser beam. The high
peak intensity of ultrafast laser pulses makes it possible to drive
this process efficiently at low average powers.
We discovered that many common dye molecules that are
not typically thought of as photoinitiators (including ones
with high fluorescence quantum yields) could initiate polymerization via linear or nonlinear absorption under the right
conditions. To test whether any of the molecules could also
be deactivated, we used the experimental setup shown in the
figure to the right. We used a Ti:sapphire oscillator producing
150-fs pulses with a center wavelength of 800 nm as the excitation source for MAP.
A second oscillator, which was synchronized to the first one
with a controllable delay time, produced tunable deactivation
pulses that were stretched out to a duration of approximately
10 ps (longer pulses can stimulate emission more efficiently than
shorter ones). MAP was used to create lines on a substrate, and
the deactivation beam was chopped so that we could assess its
effect on fabrication. One of the dye molecules we tested, malachite green carbinol base (MGCB), was proven to be capable
of being deactivated before it could initiate polymerization.
There are two intriguing aspects of the deactivation process
for MGCB. First, deactivation can be accomplished at the
same wavelength used for excitation—which implies that
the same laser can be used for both processes (as long as the
deactivation pulses have been stretched to a longer duration).
Second, varying the delay between excitation and deactivation
over the roughly 12-ns repetition time window of our oscillators had no influence on our ability to turn off polymerization.
This observation implies that the lifetime of the state that is
deactivated is much longer than 12 ns.
A molecule with a large absorption cross-section should
have a fluorescence lifetime that is on the order of nanoseconds, and, if its fluorescence quantum yield is low, then its
dominant nonradiative processes must be faster still. Thus, the
long time window for deactivation indicates that MGCB is not
deactivated by stimulated emission but rather by another photophysical process. The deactivatable state has a lifetime that
is considerably longer than 12 ns, and it cannot be the same as
the state that is excited initially.
Given that the state that can be deactivated has a long
lifetime, we explored whether deactivation required the use of
pulsed light. We found that deactivation could be accomplished
with a continuous-wave (CW) laser, and that CW radiation
not only deactivates as efficiently as pulsed radiation, but that
it also enables higher deactivation intensities to be used.
Synchronization
electronics
Oscillator 1
Sample and
translation
stage
Pulse stretcher
Chopper
3 mm
In our original experimental setup for photo-induced deacti-
vation (top), Ti:sapphire oscillators 1 and 2 generated indepen-
dently tunable, synchronized 150-fs pulses. Oscillator 1 was
used for two-photon excitation. The pulses from oscillator 2
were stretched to roughly 10 ps to deactivate the photoinitia-
tor. To test for deactivation in a photoresist containing MGCB,
we scanned the stage at a constant velocity while the deac-
tivation beam was chopped. The deactivation beam turned
off polymerization completely even using a CW deactivation
beam (bottom).
