|
Lasers
that produce pulses with femtosecond durations (1 fs = 10-15s)
have found a large
variety of scientific and commercial
applications. Length, distance and optical frequencies are
now
routinely measured with fs-lasers. Optical frequency combs produced by
fs lasers are key technology to probe and manipulate
the quantum state
of atoms or molecules, and in vivo 3D imaging with micrometer
resolution is achieved by optical coherence tomography using fs-lasers
as light sources. Fs pulses are also indispensable for the
investigation of electron dynamics in solid-state materials, such as in
novel nano-structured metamaterials.
An increasing number of applications require local oscillators with
ultra-low timing jitter. An example of these applications is ultra-high
speed analog to digital conversion (ADC), which is extensively used in software
defined RADAR
and radios. The performance of such high-speed ADCs is ultimately
limited by the timing jitter and power of their local oscillators. For
example, a 12 bit ADC at a 10GHz input frequency requires a timing
jitter of less than 4 fs. This low jitter cannot be obtained by
state-of-the art room-temperature microwave oscillators but it is
possible to achive sub-fs jitters with actively stabilized optical
frequency combs.
Our chip-scale
mode-locked lasers project targets the development of monolithic
femtosecond sources based on integrated optics,
ultimately leading to fs-lasers on a chip. Such sources enable not only
a dramatic reduction in size but also a vastly improved immunity
against environmental influences. This leads to mode-locked lasers that
operate close to the quantum limit. Such sources could find
their way in our households similar to the cw-laser diodes some thirty
years ago.
Mode-locked semiconductor lasers certainly
would qualify for applications outside the protected environment of the
optics laboratory. However, their performance is severely restricted by
the short lifetime of the gain inversion (typically 10 ns or less).
Compact and rugged fiber lasers on the other hand have recently proven
suitable for precision optical measurements and for generating fs-combs
at the highest level of precision, but none of these systems qualify
for integration in an instrument nor could they be operated by
laypersons. A variety of applications we are foreseeing for these
sources are:
- Optical telecommunication, possibly
employing DWDM, optical-phase-shift keying.
- Optical A/D conversion on a chip at
hundreds of GHz; Optical signal processing.
- Remote sensing and portable distance
measurements with ultrahigh resolution.
- Portable gas analyzers based on
optical frequency combs.
- THz generation, THz imaging and
tomography with built-in chemical analysis.
- Pump-probe systems with no moving
parts using two synchronized chip-scale lasers.
- Optical arbitrary waveform
generation.
- Turn-key optical frequency combs.
Challenges:
Making
state-of-the-art solid-state lasers smaller brings a variety of
challenges. Two of the most severe challanges are material damage
(especially if III-V semiconductor absorbers are used), and dynamic
instabilities, such as Q-switching instabilities. Both of these issues
are related to the high repetition rate, and the limited intracavity
pulse energy in compact lasers. Latter can be overcome by passive means
(namely intra-cavity dispersion management, and inverse saturable absorption; see Fig. 3), or by active feedback methods (see Fig. 4).
To overcome both, the material damage and Q-switching issues, we have
developped graphene-based saturable absorbers, that enable active
modulation of the intracavity loss, as well as passive mode-locking
owing to the intrinsic saturable loss of graphene.
|
|
|
|
Fig. 1: Artist’s
vision of a fully integrated and packaged fs-laser. Control-electronics
in form of a CMOS chip could also be integrated in the same package.
|