AN-AMP-003 Rev B1
Amptek SuperSDD and SiPIN at Low Energies
2
2
Amptek’s silicon X-ray 25 mm SuperSDD and 6 mm SiPIN can be used to measure light elements.
Their energy resolution is very good, with a noise threshold of 140 eV and about 50 eV FWHM of electronic
noise for the SuperSDD. Figure 1 and Figure 2 show spectra of several different light elements, obtained
with a SuperSDD. The biggest challenge in light element detection is sensitivity, arising mostly from
attenuation in the Be window and in air.
Normalized Counts
Ca Ka
Ca Kb
Chalk
Ar Ka
Ar Kb
Air
Cl Ka
Table salt
Cl Kb
S Ka
Epsom salt
Al Ka
Ar Ka
Al alloy
Mg Ka
Mg alloy
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Energy (keV)
2
Figure 1. Plot showing spectra measured by Amptek’s 25 mm silicon drift diode (SDD) from light element
samples.
6.0E+04
Aluminum
Magnesium
5.0E+04
Al Ka
Counts
4.0E+04
3.0E+04
Mg Ka
Ar Ka
2.0E+04
1.0E+04
Si Ka
0.0E+00
0.0
0.5
1.0
1.5
2.0
Energy (keV)
2.5
3.0
3.5
Figure 2. Plot showing spectra measured from Mg and Al samples with a SuperSDD. The clear separation
between the Mg and Al Ka peaks at this low energy is critical for analysis.
Amptek Inc.
1
AN-AMP-003 Rev B1
Sensitivity
Figure 3 shows the sensitivity of the Si detectors as a function of energy for several different
configurations. Amptek’s standard products, with a ½ mil Be window, have a sensitivity of only 20% at Na
(1.04 keV). With 1 cm of air between the sample and Be window, the sensitivity falls to 1%. For Al, at 1.49
keV, the sensitivity is 60% in vacuum but 16% with 1 cm of air. It is certainly possible to detect these low
energy X-rays, even in air, but the sensitivity is very low. Since the K X-ray yield drops for light elements,
and the X-rays from these elements have short attenuation lengths in the samples, the count rate is low.
1.0
1.0E+00
Windowless
in vacuum
Windowless
Vacuum
1/3 mil window
1/2 mil window
1 mil window
0.6
Vacuum
1/3 mil window
1/2 mil window
1 mil window
1.0E-01
1 cm air
1/3 mil window
1/2 mil window
1 mil window
Efficiency
Efficiency
0.8
0.4
1 cm air
1/3 mil window
1/2 mil window
1 mil window
1.0E-02
0.2
Be
B
C
N
O
F Ne Na Mg Al Si P S Cl Ar Ca Ti Cr Fe Cu Ge
Be
0.0
B
C
N
O
F Ne Na Mg Al Si P S Cl Ar Ca Ti Cr Fe Cu Ge
1.0E-03
0.1
1.0
10.0
0.1
1.0
Energy (keV)
10.0
Energy (keV)
Figure 3. Plot showing computed sensitivity of Amptek’s SiPIN and SDD as a function of energy for various
configurations.
Figure 4 shows the sensitivity for no air path, for 1 cm, and for 1.5 cm. For Al, the sensitivity falls from
16% for 1 cm of air to 9% for 1.5 cm of air. Only 5 mm of air attenuates the signal by a factor of two. This
makes it very difficult to obtain quantitative results: if the geometry of the sample under test and of the
calibration reference changes by a fraction of a millimeter, significant errors result. This is not unique to
these particular detectors, of course, but is an intrinsic limitation of measurements carried out in air. This is
why low energy XRF is usually carried out under a He purge or in vacuum. Figure 4 also shows that, with 1
cm of air, the 1/3 and 1/2 mil windows yield negligibly different sensitivities. The 1/3 mil window is
advantageous only if the measurement is in vacuum or a He purge.
1.0E+00
1 cm air
1/3 mil window
1/2 mil window
Vacuum
1/3 mil window
1/2 mil window
Efficiency
1.0E-01
1.5 cm air
1/3 mil window
1/2 mil window
1.0E-02
Be
B
C
N
O
F
Ne Na Mg Al Si P S Cl Ar Ca Ti
Cr Fe Cu Ge
1.0E-03
0.1
1.0
10.0
Energy (keV)
Figure 4. Plot showing the efficiency of Amptek’s SiPIN and SDD behind 1/2 and 1/3 mil Be windows in
vacuum and with 1 and 1.5 cm air paths.
Amptek Inc.
2
AN-AMP-003 Rev B1
Figure 5 compares a Mg spectrum, measured in air, with the result of exciting only an air column. Argon
is an important component of air, yielding the strong peak at 3.0 keV, present in essentially all
measurements. Note the proximity of the Ar escape peak (1.21 keV) is to the Mg K a peak (1.25 keV). High
energy resolution is critical for processing such spectra.
3.0E+03
Argon (in air)
Magnesium
2.5E+03
Ar Ka
Counts
2.0E+03
1.5E+03
Mg Ka
Ar escape
peak
1.0E+03
5.0E+02
0.0E+00
0.0
0.5
1.0
1.5
2.0
Energy (keV)
2.5
3.0
3.5
Figure 5. Spectrum of an Mg target compared to an air spectrum, showing the proximity of the Ar escape
peak to the Mg Ka peak.
Super Low Energy
Spectra of Sodium (Na) in vacuum, with 0.5 mil Be and C1 window.
Amptek Inc.
3
AN-AMP-003 Rev B1
Resolution
2
Figure 6 (left) shows the energy resolution of the 25 mm SDD for the light elements, at several different
2
peaking timings. The 6 mm SiPIN at its optimum, 32 s, is comparable to the SuperSDD at 0.8 s. The
energy resolution arises from Fano broadening and electronic noise, which add in quadrature. Fano
broadening is the dominant term for elements above Na, for Tpeak=11.2 sec. For heavier elements, the
noise is less important, so the resolution does not depend as much on the signal processing settings. To run
at high count rates, one should use a short peaking time. This will degrade the resolution most for the light
elements. Figure 6 (right) plots, on the vertical axis, the ratio of the energy resolution (eV FWHM) and the
spacing between the Ka lines of adjacent elements. This is an indicator of the ability of the system to resolve
the lines of closely spaced peaks.
180
5
160
Fano Limit
Tpeak=0.8 s
140
4
Tpeak=2.4 s
Resolution (eV FWHM)
Resolution (eV FWHM)
Tpeak=4.8 s
Tpeak=11.2 s
120
100
Fano Limit
80
60
40
3
Tpeak=0.8 s
Tpeak=2.4 s
Tpeak=4.8 s
2
Tpeak=11.2 s
1
20
Na
Al
P
Ar
Ti
Ca
Fe
Cr
Ni
Na
Zn
0
Al
P
Ar
2
3
Ti
Ca
Fe
Cr
Ni
Zn
0
0
1
2
3
4
5
Energy (keV)
6
7
8
9
10
0
1
4
5
6
7
8
9
10
Energy (keV)
Figure 6. Plots showing the resolution of the SDD as a function of energy and peaking time. The plot on the
left shows the raw resolution values. The plot on the right shows the ratio of the energy resolution (eV
FWHM) to the spacing between nearest Ka lines. This is a measure of the resolvability of the peaks.
Amptek Inc.
4