36
R. W. Taylor and V. Sandoghdar
Indeed, one can rigorously show that a single atom can extinguish a laser beam in
entirety [125].
The most fundamental requirement for reaching full extinction is spatial mode
matching: the wavefronts of the incident beam should match the dipolar emission
pattern of the atom [125]. An alternative approach is to modify the radiation pattern
of the atom by coupling it to an appropriate antenna [126]. In fact, a glass–water
interface is known to act as a primitive planar antenna that modifies the radiation
pattern of an oscillating dipole into one that is enhanced about the critical angle [127].
Optimization of the iSCAT contrast in recent efforts, where part of the reflected beam
is blocked with a pinhole in the back focal plane, can indeed be understood as the
result of a better mode matching [128–130]. A particularly compact antenna consists
of a subwavelength waveguide (nanoguide), such as a thinned glass fiber [131],
whereby the emission of an atom is efficiently coupled to the nanoguide mode, thus
improving its interference with the field propagating within with high efficiency.
Given the central role of spatial modes in this picture, these considerations are also
of immediate relevance for iSCAT detection of nanoparticles, which radiate with a
dipolar pattern. To our knowledge, this feature has not yet been explored although
the nanoguide arrangement has been demonstrated to perform well in dark-field
microscopy [132].
2.3.2 Detection Sensitivity and Signal-to-Noise Ratio (SNR)
Thus far, iSCAT has been successfully employed to detect individual unlabeled proteins as small as 50 kDa. An important question that arises is whether there exists
a fundamental limit in detection sensitivity, and what might be the accompanying
technical challenges. In what follows below, we give an overview of some of the
pertinent issues.
The central restriction in any sensitive measurement is signal fluctuation. In their
absence, arbitrarily small signals may be identified even on very large backgrounds.
While slow fluctuations such as thermal or mechanical drifts can be directly or
indirectly accounted for, fast random variations, which we call “noise” pose a serious
challenge. Some of the main sources of noise in an iSCAT measurement are as
follows:
Laser intensity noise: Even the best lasers have instrumental power fluctuations
and beam instabilities. It is indeed not easy to have a freely running laser with power
stability better than about 10
−3 . To detect iSCAT contrasts beyond this, one has
to account for laser intensity fluctuations through referencing or normalization. In
confocal imaging, a balanced photodiode pair employing common-mode rejection
can be as good as 1×10
−7 [133]. In wide-field camera-based detection, one can
normalize the total power recorded within each frame to similarly reject intra-frame
fluctuations. As the contrast signal is the ratio of two fields sharing the same laser
noise, they naturally self-reference.
R. W. Taylor and V. Sandoghdar
Indeed, one can rigorously show that a single atom can extinguish a laser beam in
entirety [125].
The most fundamental requirement for reaching full extinction is spatial mode
matching: the wavefronts of the incident beam should match the dipolar emission
pattern of the atom [125]. An alternative approach is to modify the radiation pattern
of the atom by coupling it to an appropriate antenna [126]. In fact, a glass–water
interface is known to act as a primitive planar antenna that modifies the radiation
pattern of an oscillating dipole into one that is enhanced about the critical angle [127].
Optimization of the iSCAT contrast in recent efforts, where part of the reflected beam
is blocked with a pinhole in the back focal plane, can indeed be understood as the
result of a better mode matching [128–130]. A particularly compact antenna consists
of a subwavelength waveguide (nanoguide), such as a thinned glass fiber [131],
whereby the emission of an atom is efficiently coupled to the nanoguide mode, thus
improving its interference with the field propagating within with high efficiency.
Given the central role of spatial modes in this picture, these considerations are also
of immediate relevance for iSCAT detection of nanoparticles, which radiate with a
dipolar pattern. To our knowledge, this feature has not yet been explored although
the nanoguide arrangement has been demonstrated to perform well in dark-field
microscopy [132].
2.3.2 Detection Sensitivity and Signal-to-Noise Ratio (SNR)
Thus far, iSCAT has been successfully employed to detect individual unlabeled proteins as small as 50 kDa. An important question that arises is whether there exists
a fundamental limit in detection sensitivity, and what might be the accompanying
technical challenges. In what follows below, we give an overview of some of the
pertinent issues.
The central restriction in any sensitive measurement is signal fluctuation. In their
absence, arbitrarily small signals may be identified even on very large backgrounds.
While slow fluctuations such as thermal or mechanical drifts can be directly or
indirectly accounted for, fast random variations, which we call “noise” pose a serious
challenge. Some of the main sources of noise in an iSCAT measurement are as
follows:
Laser intensity noise: Even the best lasers have instrumental power fluctuations
and beam instabilities. It is indeed not easy to have a freely running laser with power
stability better than about 10
−3 . To detect iSCAT contrasts beyond this, one has
to account for laser intensity fluctuations through referencing or normalization. In
confocal imaging, a balanced photodiode pair employing common-mode rejection
can be as good as 1×10
−7 [133]. In wide-field camera-based detection, one can
normalize the total power recorded within each frame to similarly reject intra-frame
fluctuations. As the contrast signal is the ratio of two fields sharing the same laser
noise, they naturally self-reference.