9:00am-9:15am

"UE2

Zero-Dispersion Wavelength Uniformity and

Four-Wave Mixing in Optical Fiber
J. B. Schlager, S. E. Mechels, andD. L. Franzen
National Institute of Standards and Technology, 325 Broadway, Boulder, CO 80303

The efficiency of four-wave mixing (FWM) in optical fiber is maximum with optimal phase
matching at the zerodispersion wavelength (ho),but this wavelength can vary along the fiber's length.' In
this paper, we describe measurements of partially degenerate four-wave mixing efficiency in a 10 km fiber
with varying ho. In partially degenerate four-wave mixing the pump provides two of the four optical waves
at the same optical frequency. The other two waves are the probe and the generated anti-Stokes light.
With the probe light fixed at a given wavelength, we determine the pump wavelength which gives maximum
anti-Stokes light intensity in the ho region of the fiber. We compare this pump wavelength with accurate
measurements of the path-average made with the frequencydomain phase-shift (FDPS) technique .2.3 In
general, these wavelengths do not match exactly, but simulations of FWM which include more than one
path-average ho from FDPS measurements more accurately predict the observed peak FWM efficiency
wavelength.

The experimental configuration for obtaining FWh4 efficiency is shown in Fig. 1. Two tunable
lasers provide the pump and probe lights, The pump laser is amplified with an erbium-doped fiber
amplifier (EDFA) to a power of +10 a m . The probe laser light at -3 dBm is mechanically chopped to
allow narrow-band detection of the FWM signal with lock-in amplifier techniques. Light from each source
passes through an optical bandpass filter to minimize the amplified spontaneous emission (ASE) light,
which can overwhelm the FWM signal, and polarization controllers are used to bring the polarization states
into coincidence. The pump and probe lights are combined in a fused-fiber coupler, traverse the test fiber,
and continue on to the photoreceiver. An additional optical bandpass filter with sufficient bandwidth
(2 nm, full width at half maximum) for the wavelength scan is used at the photoreceiver to further isolate
the FWM signal. The detected signal is passed to a lock-in amplifier. A computer automatically tunes the
pump laser and acquires the FWM efficiency data, Laser wavelengths are measured with an
interferometric wavemeter having an accuracy of 1 ppm.

The FDPS system has been described in more detail e l s e ~ h e r e . ~In. ~this system, the output of a
tunable laser is intensity modulated by an integrated-optic Mach-Zehnder modulator. The phase of the RF
modulation after a test fiber is monitored by a vector voltmeter as the laser is tuned through ho. The test
fiber is temperature controlled to a . 1 5 "C. Wavelengths are monitored by a wavemeter and matched to the
corresponding relative group delays with a temporal resolution of 0.15 ps. The system can determine LO
with a precision of better than 0.025 nm (three standard deviation^).^

Variation of ho with.length in a 10 km fiber as measured by the FDPS system is shown in Fig. 2.

The path-average ho for the 10 km fiber is plotted with the path-average L's of four 2.5 km and ten 250 m
segments. Zero-dispersion wavelength increases linearly by about 1.2 nm over the 10 km length and varies
by as much as 0.3 nm for the final 2.5 km. Figure 3 shows the corresponding four-wave mixing efficiency
curve obtained for the 10 km fiber. A theoretical curve assuming uniform ho and one based on the
measured ho values from the four 2.5 km segments are also plotted. These curves were obtained by solving
the wave equation which approximates FWh4 behavior in fiber with constant ho and applying the
appropriate boundary conditions at the segment ends.' Although discrepancy between experiment and

166
 theory still exists, simulations approximatingthe observed linear increase in ho more accurately predict the
wavelength of maximum FWM efficiency.

1. K. Inoue, “Four-wave mixing in an optical fiber in the zerodispersion wavelength region,” J

Lightwave Technol., vol. 10, no. 11, pp. 1553-1561, 1992.
2. S . E. Mechels and D.L. Franzen, “Accurate measurements of zero-dispersion wavelength in
single-mode fibers,” 3rd Optical Fiber Measurement Conference Technical Program, paper
IV.2, 1995.
3. S.E. Mechels, J. B. Schlager, and D. L. Franzen, “Accurate zero-dispersion wavelength
measurements in single-mode fibers: two frequency-domainmethods,” LEOS 1995 Meeting
Proceedings, vol. 2, pp. 75-76, 1995.
4. S.E. Mechels, J. B. Schlager, and D. L. Franzen, “Accurate measurements of zerodispersion
wavelength in optical fibers,” submitted to J. Res. Nat ’1. Inst. Stands. Technol.

Contribution of NIST, not subject to copyright.

Optical bandpau
filter
Tunable lawr
(Pump) Fad-fiber coupler

Tunable laser

Optical bandpnu

1 Qce
controller
Optical bandpau
filter
p
3J
-
Figure 1. Experimental codiguration for obtaining
FWM efficiency: PC, polarization controller,
CPU,personal computer.

1551.0 r
02 5 km segments
- 1550 5}
-.-.-..10 km path-average value
Section IV
0.a lambda zero @ 1549.44nm
-
E

2C 1550.0 =.
- 0.6
W .-0
._
2 1549.5 c
L

c
0 0.4 FDPS 10 km
._
e
$
6
1549.0 2 0.2
0
4 1548.5
0.0

1548.0
0 1 2 3 4 5 6 7 8 9 10 1547 I 548 1549 1550 1551
Length (km) Wavelength (nm)
Figure 2. FDPS measurementsof zero-dispersion Figure 3. Observed FWM efficiency in 10 km
wavelengths in sections of a 10 km fiber compared fiber and curves from FWM simulations.
with the path-average value for the entire length.

167