Optical waveguides – Optical fiber waveguide with cladding
Reexamination Certificate
1999-08-20
2001-11-27
Lee, John D. (Department: 2874)
Optical waveguides
Optical fiber waveguide with cladding
C385S033000, C385S043000, C372S006000, C359S341100
Reexamination Certificate
active
06324326
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to optical amplifiers and lasers, and in particular to fiber lasers.
2. Technical Background
Optical fiber is increasingly becoming the favored transmission medium for telecommunications due to its high capacity and immunity to electrical noise. Silica optical fiber is relatively inexpensive, and when fabricated as a single-mode fiber can transmit signals in the 1550 nm band for many kilometers without amplification or regeneration. However, a need still exists for optical amplification in many fiber networks, either because of the great transmission distances involved, or the optical signal being split into many paths.
Erbium-doped fiber amplifiers (EDFAs) have been found quite effective in providing the required optical gain. As illustrated schematically in
FIG. 1
, a conventional EDFA is interposed between an input transmission fiber and an output transmission fiber
14
. Both transmission fibers
12
,
14
need to be single-mode, because higher-order modes exhibit much greater dispersion (typically the limiting factor for the fiber transmission distance at high data rates). The EDFA includes a length (on the order of tens of meters) of an erbium-doped silica fiber
16
, as is well known in the art. The doped fiber
16
should also be single-mode in order to maintain the transmission signal integrity. The doped fiber
16
is optically active due to the presence of Er
3+
ions, which can be excited to higher electronic energy levels when the doped fiber
16
is pumped by a strong collinearly-propagating optical pump signal. Typically, an optical pump source
18
inputs the pump signal into the doped fiber
16
through a pump source fiber
20
coupled to either the undoped upstream fiber or the doped fiber
16
through a wavelength-selective directional coupler
22
, but downstream coupling is also known. Again, for integrity of the transmission signal, the pump source fiber
20
should be single-mode. An operative EDFA may contain some additional elements (such as an isolator) which are well known to the art but not relevant to the understanding of the background of the present invention.
Conventionally, one typical pump source
18
has been an edge-emitting semiconductor laser that includes a waveguide structure (in what is called a “stripe” structure) that can be aligned with the single-mode pump source fiber
20
to provide effective power coupling. However, this approach has failed to keep up with modern fiber transmission systems incorporating wavelength-division multiplexing (WDM). In one approach to WDM, a number of independent lasers inject separately-modulated optical carrier signals of slightly different wavelengths into the transmission fiber
12
. The EDFA has sufficient bandwidth to amplify carrier signals within about a 40 nm bandwidth. A large number of multiplexed signals to be amplified require in aggregate a proportionately large amount of pump power. Over the past decade, the number of WDM channels preferably utilized in a standard network has increased from about four to current levels of forty or more, but at best the output power from a single-stripe laser source has only doubled. Derivative designs such as a master oscillator power amplifier (a single-mode stripe followed by a broad stripe amplifier) or flared-semiconductor devices are capable of producing more than one watt of optical output power, but many of these designs have been subject to reliability problems (such as back-facet damage caused by feedback) that have hindered their practical deployment as fiber amplifier pumps.
Another approach uses WDM technology to combine pump signals. Multiple single-stripe lasers are designed to emit light at narrowly-spaced wavelengths, usually within the wavelength bands of 970-990 nm or 1460-1500 nm. Wavelength-dependent directional couplers combine these multiple optical waves into a single (somewhat broadband) pump signal. While this approach increases the power available for optical amplifiers, it greatly adds to the complexity of the pump source, and requires additional components such as thermoelectric coolers, fiber gratings, and directional couplers. As a result, this approach increases cost.
An alternative approach for high-power pump lasers has involved fiber lasers that are pumped through their cladding. That is, a large outer cladding supports the pump signal from a primary pump source, and an inner cladding supports a single-mode output signal that is used as the secondary pump source for the EDFA. The core is typically doped to provide lasing capability. Typically, a neodymium- or ytterbium-doped double-clad fiber is pumped with a high-power diode optical source (at 800 nm or 915 nm) to produce a single transverse mode (at 1060 nm or 1120 nm, respectively). One of these modes then pumps a cascaded Raman laser to convert the wavelength to around 1480 nm, which can then pump erbium. To date, such a design by itself (that is, without an additional Raman oscillator) does not produce an output in any of the appropriate absorption bands for EDFAs.
Double-clad fiber lasers offer superior performance for four-level lasing (that is, where the lasing occurs in a transition between two excited states). In such a case, the doped core is still transparent at the laser signal wavelength when not being pumped. As a result, the power threshold for lasing depends essentially on the dimensions of the doped core, and the background loss in the fiber over the pump absorption length. However, ytterbium and neodymium ions (Yb
+3
and Nd
+3
) provide three-level lasing systems at around 980 nm and 940 nm, respectively. In a three-level system, the lasing occurs from an excited level to either the ground state or a state separated from it by no more than a few kT (that is, thermally mixed at operating temperature). As a result, an unpumped doped core strongly absorbs at the laser wavelength, and the lasing power threshold can become a problem.
Ytterbium has offered much promise as a pump for high-powered EDFAs. It is well known that Yb
3+
ions exhibit gain in a narrow 6 nm-wide three-level transition at 976 nm, and in a broad quasi-three-level transition peaked at 1030 nm (but extending as far as 1140 nm). The latter transition requires a population inversion of only a few percent for transparency, while the former requires at least a fifty-percent inversion.
Thus, a source based on the 976 nm Yb
+3
transition has long been suggested as a pump for EDFAs. However, a single-stripe diode laser remains the most efficient pump structure. The problem potentially lies in the relationship between the gains in the two transitions and the pump absorption. As a representative example, the gains at the two wavelengths in Yb-doped germano-alumino-silicate glass (assuming homogeneous broadening) are related by the equation:
G
1030
=
0.25
G
976
+
0.74
α
P
Γ
S
Γ
P
(
1
)
where G
1030
and G
976
are the gains at 1030 nm and 976 nm, respectively, &agr;p is the partially-bleached absorption in decibels (dB), and &Ggr;
S
and &Ggr;
P
are the respective overlap factors of the signal mode and pump mode with the dopant profile.
Double-clad fibers allow coupling from diode bars and other similar active structures. However, this is accomplished by a greatly-reduced pump overlap with the doping profile relative to the signal overlap, since the doping needs to be confined in or close to the signal core in order to obtain sufficient optical gain for the core mode at the signal wavelength. Typically, the core is uniformly doped, and the area ratio between the pump waveguide and the signal core is on the order of 100:1. As a result, &Ggr;
S
≈1 and &Ggr;
P
<0.01. Using these values in Equation (1), each 1 dB of pump absorption produces about 74 dB of gain at 1030 nm. Even with weak pumping, amplified spontaneous emission (ASE) at 1030 nm will saturate the amplifier and prevent a buildup of the populat
Dejneka Matthew J.
Minelly John D.
Zenteno Luis A.
Agon Juliana
Corning Incorporated
Lee John D.
LandOfFree
Tapered fiber laser does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Tapered fiber laser, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Tapered fiber laser will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2600589