Carbon Nanotubes as Optical Probes:
Preparation:
In order to use carbon nanotubes (CNTs) as near-field optical probes we must first attach them to AFM probes. CNTs are
grown on a silicon wafer using chemical vapor deposition. The wafer is scanned
by an AFM tip in tapping mode until a nanotube is picked up. From the
images of the phase and height trace it is fairly simple to notice when
a nanotube is picked up. (Fig. 1).
Fig 1: The height and phase trace of the AFM. When a nanotube is picked up
there is an abrupt change in the phase trace and the resolution of the
features of the height trace also change.
In
order for CNT probes to yield good images, they
typically have to be shortened to protrude from the tip less than
~100 nm. This is done by applying voltage pulses to the
tip, which allow us to controllably ablate short sections of
the CNT.
Fig 2: SEM image
of an AFM tip with a long nanotube attached.
Although
the tip's length can be monitored by an SEM(Scanning Electron Microscope), it is not time or cost
efficient. Typically a force curve is preformed to detect the tip's
length. The deflection of the tip is monitored while the tip is lowered
into the silicon substrate. When the nanotube initially touches the
substrate the deflection changes at once. As the tip keeps being
"pushed" into the substrate the nanotube buckles. Eventually the
pyramid of the tip touches the substrate and the deflection changes at
once again. By measuring the distance between the two "kinks" (Fig. 3), one can measure the CNT's protrusion length.
Fig. 3: Red curve is a force curve of a long CNT (~440 nm) while the blue curve is after shortening (CNT length is ~40 nm).
Imaging:
The
carbon nanotube tip then can be used as an AFM and optical probe at the same
time. The AFM resolution for 5nm QDs(Quantum Dots) was ~ 40 nm (Fig.
4). All optical images of QDs obtained with CNT probes show
strong
reduction in the fluorescence signal due to the nanoscale interaction
between the QD and the tip-attached CNT. The optical resolution
was about ~
20 nm.
Fig 4: (a) AFM topography of a 5nm quantum dot and (c) the optical
image. the cross sections at the yellow line are shown in (b) and (d).
A
reduction of signal is also seen in our setup when using metal tips.
However, for the CNT tips a much stronger reduction in signal is
observed. From the cross section of the optical images one can see that
the value of the near-field pixel is roughly the same as the
background noise level. Typically the tip oscillates ~ 120 times per
pixel above the
QD at an amplitude of ~24 nm. A force curve above a QD shows that the
reduction of signal begins at about 25 nm away from the QD (Fig. 5).
This would indicate that the total signal of a near-field pixel
should be
higher
than the background noise level since the CNT spends a considerable
time away from
the QD. However, it seems as if the QD switches "off" the whole time
when
scanned by a CNT
tip.
We
believe that this phenomena might be an indication of an energy
transfer between the CNT and the QD in particular charge transfer. We are currently investigating the precise mechanism of this observed
recovery time effect of the QD when brought in close proximity with the
CNT.
Fig. 5: The red line is the photon count rate and the blue is the
deflection. The signal starts drooping at 25nm away from the QD. When
the CNT touches the QD the photon signal is roughly the same as the
background.
Further information about this project can be found at:
"Nanoscale
Fluorescence Microscopy Using Carbon Nanotubes,"
C. Mu, B. D. Mangum, C. Xie, and J.M. Gerton, IEEE J Sel Top Quant 206,
Volume 14 (2008). [PDF]
