CORPORATE HEADQUARTERS
11578 SORRENTO VALLEY ROAD
SAN DIEGO, CA 92121-1311
USA
QuantumDesign
TEL: 619.481.4400
800.289.6996
FAX: 619.481.7410
E-MAIL: service@qdusa.com
http://www.qdusa.com/
Distorted low-level signal readback of AC signals in the PPMS in the temperature
range 25-35 K due to Inconel mitigation of inductive cross talk
Keywords: AC Transport, ACMS, Inconel, Harmonics, Cross Talk
Symptoms:
Signal (V)
Several PPMS users have noticed anomalous peaks in their data when measuring resistance
versus temperature of a low resistance sample with the AC Transport option. The peak occurs
between 25 K and 35 K, is several Kelvins wide, increases with increasing measurement
frequency, and is suppressed by application of magnetic fields as low as a few hundred Oe. The
precise temperature at which the peak occurs and the magnitude of the peak vary from system to
system. Upon closer investigation of the data, one also finds that the 2nd and 3rd harmonics of the
detected voltage across the sample are uncharacteristically high in this temperature range as well.
If the measured waveform were observed with an oscilloscope or recorded to a diagnostic data
file, one would see a systematic distortion of the readback signal as the temperature passes
through this temperature regime, such that the voltage across the sample in the time domain does
not resemble a sine wave (fig 1.) The effect can be seen when measuring a 4-point short or a very
small piece of
copper, for
Raw Signal Distortion by
example. Figure
the Inconel Feedthrough
2 shows a
-5
1.0x10
35.5 K
spurious peak in
the resistivity
36.5 K
data for a small
-6
40.0 K
5.0x10
piece of copper
due to the AC
Transport
0.0
software’s
attempt to reduce
-6
the distorted
-5.0x10
waveform to a
resistivity value.
-1.0x10
-1.5x10
-5
-5
time
Figure 1. The signal read back across the voltage leads during a 4-wire
resistivity measurement can be distorted in a certain temperature range.
1/5
Effects of Distorted Signal on Resistivity Measurements
of Copper with the AC Transport
Resistivity Calculated by
AC Transport Software (Ohm-cm)
0.000020
0.000018
0.000016
false peak due to
distorted signal readback
0.000014
0.000012
0.000010
0.000008
0.000006
0.000004
15
20
25
30
35
40
45
Temperature (K)
Figure 3. The distorted signal can lead to misleading resistivity values due to
inappropriate data reduction.
The same phenomenon also manifests itself in AC resistivity measurements performed in the
PPMS with third party instrumentation, as well as in AC susceptibility measurements performed
with the ACMS option. The phenomenon in ACMS measurements appears to be a small (< 0.5
degree) systematic change in otherwise steady phase data (see figure 3.) The effect can be seen
in the data from the ACMS calibration coil measurement steps, meaning that the ACMS can at
least partially correct for the effect. While noticeable, this effect tends to be below the ACMS
system’s true sensitivity, appearing only as small phase shifts, and therefore should not pose a
reasonable obstacle to collecting good AC susceptibility data in the ACMS.
Effects of Distorted Signal on AC Moment Measurements
of Dysprosium Oxide in the ACMS
0.00026
Amplitude
0.05
Phase
0.00022
0.00
0.00020
-0.05
0.00018
Phase (deg)
Moment (emu)
0.00024
0.00016
-0.10
0.00014
20
25
30
35
40
45
Temperature (K)
Figure 2. Signal distortion can cause small shifts in the phase data for AC
susceptibility measurements in the ACMS.
2/5
Cause
AC Moment, m' (emu)
The fact that the effect increases with measurement frequency indicates that it is inductive in
nature. But the temperature and field dependence suggest that some material’s physical
properties also play a role. The source of this behavior is in fact the inconel electrical
feedthrough at the bottom of the PPMS sample chamber that the puck interface plugs into. The
inconel in the electronic vacuum
feedthrough has a peak in its
Real Component of the AC Moment
of Inconel-X in Small Fields at Various Temperatures
instantaneous susceptibility curve,
0.005
Excitation = 1.5 Oe, 100 Hz
25 K
χac(T), around 25 K-30 K. The
Performed in MPMS
30 K
inconel mitigates inductive cross
0.004
talk between the excitation circuit
35 K
and the sensitive voltage detection
20 K
40 K
0.003
circuit where the lead twisting
stops to pass through the
45 K
50 K
feedthrough, distorting the
0.002
15 K
inducted waveform. The resulting
effect is similar to an inductor core
0.001
with changing permeability. The
distorted waveform is not handled
well by the AC Transport
0.000
-100
-50
0
50
100
software’s data reduction routines.
Field (Oe)
Imaginary Component of the AC Moment
0.00025
of Inconel-X in Small Fields at Various Temperatures
Excitation = 1.5 Oe, 100 Hz
Performed in MPMS
AC Moment, m'' (emu)
0.00020
20 K
0.00015
0.00010
15 K
25 K
0.00005
0.00000
30 K
-0.00005
-100
50 K 45 K
-50
35 K
40 K
0
50
100
Field (Oe)
Figure 4. The real (top) and imaginary (bottom) components
of the AC moment of a sample of the inconel feedthrough in
the PPMS sample chamber. The data indicate that the
instantaneous susceptibility of the sample has a peak around
between 25 and 30 K in zero magnetic field. The small offset
between zero field and the peaks is an instrumental offset in
field reporting.
3/5
Inconel is a standard material used
for vacuum-tight feedthroughs in
cryogenic applications. It is a
nickel super-alloy in the same
family as stainless steel. The
variability of the noted effect from
PPMS sample chamber to sample
chamber is believed to be due to
variations in the treatment of the
individual inconel feedthroughs.
The AC moment of a small piece
of this material when measured
with a constant alternating field of
1.5 Oe is shown in figure 4. When
changing temperature the changing
susceptibility of the material
clearly reverses directions between
25 K and 30 K. Also, fields of 100
Oe and less seem to saturate the
sample, so that its susceptibility
does not change much with further
changes in field or temperature.
The result of these material properties is that the feedthrough can amplify the flux through the
detection circuit enough to induce measurable amounts of cross-talk when the resistive voltage
across the sample itself is quite small and the current through the excitation circuit is quite large.
Figure 5 illustrates the geometry of the cross-talking circuits. As a magnetic field is applied, or
the temperature of the sample chamber moves outside the regime where the inconel has a
significant AC susceptibility, the inconel no longer mitigates the cross talk between the
excitation circuit and the detection circuit, so phase-sensitive detection will effectively filter out
the cross talk. It is only in zero magnetic field and at the temperatures noted above that the cross
talk cannot be filtered out due to the intervening physical properties of the inconel.
Figure 5. The approximate geometry of the electrical feedthrough on the bottom of the
PPMS sample chamber illustrates how alternating current through two contacts can create a
small oscillating magnetic dipole that is detected by voltage taps at another pair of contacts.
Changes in the inconel susceptibility with temperature amplify or distort the effect.
Solutions
Unfortunately, there is no practical way to replace the Inconel feedthrough on the PPMS sample
chamber. To physically avoid passing any electrical currents through the Inconel material, one
might wish to avoid the PPMS sample puck interface altogether and feed new wires into the
sample chamber from the top for electrical measurements. This could be accomplished using the
PPMS multifunction probe, for example.
Short of creating a new electrical feedthrough, other suggestions to minimize the effect are as
follows:
a) Keep the drive frequency as low as practically possible. Resistivity measurements will vary
with drive amplitude, but should not vary with frequency. Frequency dependence is a
reactive phenomenon. Because the phenomenon of concern here is reactive, keeping the
drive frequency low (or measuring with DC) will keep the cross talk to a minimum.
b) Increase the signal from the sample without increasing the drive amplitude (which causes the
cross talk). Try greater voltage lead separation, for example, or a sample with smaller crosssection.
4/5
Be sure to watch the fit quality for reduced data and values of 2nd and 3rd harmonic contribution
to the signal of interest. These are strong clues indicating the quality of the measured signal.
Reading Quantum Design Application Note 1584-202, “Inductive Cross-talk During Alternating
Current Measurements: Guidelines for Sensitive Measurements,” is also recommended.
5/5