Either a prism or a diffraction grating is used
as the wavelength dispersive element in most
optical spectrometers.
This page contrasts and compare the
performance characteristics of each.
Prism Versus Diffraction Grating?
Prism Versus Diffraction Grating Performance
Detailed Diffraction Grating
Characteristics
Grating “Pros”
•
Off the shelf optional groove densities and
wavelength dispersion options
•
Wide selection of blaze wavelengths
Grating “Cons”
•
Produced either mechanically with burnished
grooves (incorporates “ruling errors”) or
holographically with limited groove density
potential
•
All have a peak efficiency at just one
wavelength (the “blaze” wavelength).
Diffraction efficiency drops off rapidly to shorter
and gradually to longer wavelengths. See
Figure 1
•
Grating efficiency curves present “anomalies”
or discontinuities in the efficiency profile.
•
Classical “mechanically ruled” gratings are
more efficient that those produced
holographically
•
The efficiency curves of all diffraction gratings
are affected by light diffraction into higher
“orders”
•
Each order is one octave, for example, 400 to
800-nm. Second order will be 200 to 400-nm
and overlay first order. See Figure 2
•
To remove second order overlap order sorting
filters are required
•
In actual practice peak efficiency of a diffraction
grating is almost always less than 70% and can
drop to near zero at the extremities of its
spectral range
•
Wavelength dispersion (nm/mm) is non-linear
varying as the diffraction angle and the
distance to focus at each wavelength. In
practice, wavelength dispersion is linear
enough to present near constant resolution
especially for low resolution instruments
Bottom line: Ruled gratings are more efficient
than many holographic gratings when considered
over a wide wavelength range. Nevertheless
diffraction gratings cannot be used over a
wavelength range greater than an octave without
order sorting filters.
Detailed Prism Spectrometer
Characteristics Pros and Cons
The major attraction of a prism is the near <90%
average transmission efficiency at all
wavelengths above ~365-nm. The efficiency
profile is flat with no drop-off after ~400-nm. In
terms of efficiency a prism will outperform all
diffraction gratings.
Prism “Pros”
Ironically non-linear wavelength dispersion! As
the QE of a camera decrease at longer
wavelengths bandpass falls to compensate.
Consequently, a prism delivers significantly
higher signal to noise ratio over an extended
wavelength range than a diffraction grating. See
Figure 3.
Transmission efficiency is a flat > 90% over the
bulk of a wavelength range above ~400-nm
outperforming all diffraction gratings
Refraction does not result in “overlapping
orders,” consequently a prism operates over
greater than one octave without requiring
filtering. Prisms work from 365 to 920-nm or
above.
Figure 4. Diffraction gratings split up incoming
light into diffraction orders reducing efficiency.
To see the PARISS imaging prism spectrograph
click here
Prism “Cons”
•
Compared to gratings, prisms are very
expensive. Only high end instruments
addressing challenging applications use a
prism.
•
Wavelength dispersion is non-linear,
consequently bandpass and resolution
change from high in the blue to lower in the
red. Linearizing dispersion is trivial in the
software, but does not compensate for impact
of changing bandpass.
•
Prism spectrometers share non-linear
dispersion with both AOTF and LCTF devices
(Acousto optic tunable filters and Liquid
Crystal Tunable filters)
Best Uses For Prism and Diffraction
Grating Spectrometers FAQ
Q:When is a prism spectrometer preferable to
diffraction grating spectrometer?”
A: It is best to use a prism in low-light applications
where high sensitivity and high signal-to-noise
ratios are of maximum importance.
High efficiency ensures the highest sensitivity. If
we consider the PARISS prism, the light
throughput efficiency between 365 to 1000-nm is
up to 90%.
A typical diffraction grating offers a peak efficiency
at the blaze wavelength of around 60% with a
drop-off on both the long and short wavelength
sides.
Q: When is a diffraction grating spectrometer
preferable to a prism spectrometer?
A: Best for use with diverse bright-light
applications demanding different spectral ranges
and spectral resolution.
Interchangeable gratings are available with
various groove densities and blaze wavelengths
to meet spectral range and resolution
requirements. (Spectral resolution varies linearly
with groove density)
Q: Do all diffraction gratings suffer from 2nd order
overlap with 1st order.”
A: All diffraction orders are an inherent form of
spectral pollution that degrades diffraction grating
performance.
Reducing the wavelength range to just one
octave (2x initial wavelength such as 400 to 800-
nm), eliminates higher-order overlap.
The inherent presence of higher orders
significantly reduces diffraction grating
efficiency, therefore reducing sensitivity.
Q: Hpw does the wavelength dispersion of a
prism compare to that of a diffraction grating?
A: The wavelength dispersion of a prism is non-
linear. Spectral resolution (bandpass) is highest in
the UV and blue and decreases at longer
wavelengths, which is actually an advantage!
The increasing bandpass at longer wavelengths
compensates for the falling quantum efficiency of
a CCD or CMOS over the same wavelength
range.
The net result is a flatter efficiency profile and
significantly higher signal-to-noise ratio over a
wider wavelength range than is possible with a
diffraction grating.
The wavelength dispersion of a diffraction grating
is close to linear, resulting in a gradual reduction
in signal-to-noise ratio with increasing
wavelength. (Actual bandpass varies as the
cosine of the diffraction angle)
Figure 1: Prism vs diffraction grating
efficiency curves. Diffraction grating
efficiency profiles vary, but never equal the
efficiency of a prism.
Figure 2: With some exceptions most diffraction
gratings are used in first order. However, gratings
also diffracts light into plus and minus orders that
overlap first order
Figure 3: All spectrometer components can present
wavelength efficiency issues. Signal to noise ratio S/N is
a product of bandpass and efficiency. The efficiency curve
of most cameras falls with increasing wavelength.
The spectral resolution of a prism goes a long way to
compensate and offer high S/N at long wavelengths that
are a problem for diffraction gratings.