fiber.raman_efficiency_at_shift

raman_efficiency_at_shift(
    shift: float | numpy.ndarray,
    *,
    peak_efficiency: Annotated[float, minimum=0, units='1/(W·km)'] = 0.4,
) -> numpy.ndarray | float

Raman gain efficiency against Stokes shift.

Fibre datasheets and amplifier papers quote the efficiency $g_R/A_{\rm eff}$ in 1/(W km) rather than $g_R$ itself, which sidesteps both the effective area and the factor of two for polarization scrambling. This takes the same convention.

The shape is a three-Gaussian approximation to the silica spectrum: the main peak at 13.2 THz, the shoulder above it, and the broad low-shift tail. The components are tuned so the sum peaks at 13.2 THz, which is not the same as centring the largest Gaussian there: an overlapping shoulder drags the maximum up, and an untuned fit peaks near 14 THz. Good enough for a WDM tilt or a pump offset, not a substitute for a measured spectrum.

Args

• shift: Stokes shift, in Hz. Positive means the signal is below the pump in frequency, which is the direction that sees gain.
• peak_efficiency: Efficiency at the 13.2 THz peak, in 1/(W km). About 0.4 for standard fibre with a 1450 nm pump, higher for smaller cores.

Returns

Efficiency in 1/(W km), same shape as shift, zero for shifts of the wrong sign.