Instrumentation & Measurement Magazine 25-4 - 32

of electrodes in-vivo. If the marker of the pathology is not
accessible in external samples, a radioactive tracer, whose concentration
in the body can be imaged from outside by means of
emission tomography techniques (PET and SPECT) thanks to
the penetration of gamma rays, can be molecularly attached to
the probe molecule and injected in the patient.
Circuits
As the majority of modern instruments are based on a digital
processing device (such as a microcontroller or FPGA [3]), the
acquisition chain of a charge measuring system is composed of
an analog front-end stage and an analog-to-digital converter
(ADC). I will follow the signal path along this chain and will
discuss highlights and relevant applicative examples for each
block (Fig. 3). Let us start from the input. Since the ADC typically
operates in voltage mode, the purpose of the input stage
is to convert the input charge into a voltage. This could be
easily performed by storing the charge on a capacitor CD
, often
associated with the structure of the device collecting the
charge. The smaller is the capacitance, the larger is the conversion
factor. Analogously, to convert the input current into a
voltage, a resistor RC
could be used. However, the main limitation
of this approach is due to the sensor parasitic impedances
in both cases. The optimal solution to these issues is the adoption
of negative feedback. A classical active scheme based on
an operational amplifier and a passive feedback branch, so
called transimpedance amplifier, provides a very low input
impedance at the input (virtual ground) that is suitable to read
the input current and a very linear conversion factor from current/charge
to output voltage. The feedback impedance can be
resistive or capacitive. When analyzing the noise performance
of the transimpedance stage (see [2] for details), it becomes apparent
that the dominant noise sources are the thermal noise
of the feedback resistance and the input equivalent noise of
the operational amplifier which is differentiated across the input
capacitance and thus dominates at high frequency. In order
to reduce noise, the value of the feedback resistor can be increased.
This results in a decrease of the amplifier bandwidth
that can be compensated, for instance, with zero-pole cancellation
stages. The feedback resistor can be avoided by adopting a
purely capacitive feedback. This is the typical configuration to
read charge packets that are integrated in the feedback capacitance,
thus producing steps at the voltage output. In order to
avoid the saturation of the stage due to input dc currents, such
as unavoidable leakage currents, a reset mechanism is needed.
Discrete-time (periodic reset) or continuous-time (with additional
feedback branches) can be implemented, depending
on the constraints of the application [2]. Continuous reset is
more suitable for current sensing, avoiding dead-time during
the periodic reset, which instead is suitable for charge sensing.
Once the impact of the feedback resistor is addressed, it
is crucial to minimize the value of the total capacitance connected
at the input node: it includes the amplifier input
capacitance, the feedback capacitor and the parasitic capacitance
associated with the sensing electrode and its connection
to the preamplifier. There are three strategies for the reduction
of the input capacitance. The first is the reduction of size
of the charge-collecting electrode, whose capacitance typically
scales with the area. Unfortunately, the collected current signal
also often scales with electrode area, especially in electrochemical
applications. Two actions can be implemented to mitigate
the decrease of the signal: (1) an active mechanism drives the
molecules to be detected towards a small sensing electrode
(analogous to the case of the electric drift used in Silicon Drift
Detector to drive, in a depleted semiconductor, the charge generated
by the absorption of a photon or by an electron towards
a small collecting anode [4], thus combining large detection
area with small sensor capacitance); and (2) splitting the sensor
area into small pixels readout by parallel channels, whose outputs
are then summed to reconstruct the signal [5]. The latter
approach provides a SNR improvement at the expense of increased
area and power dissipation, proportional to the square
root of the number of parallel circuits since the signals combine
linearly, while the noise of the individual channels, being
uncorrelated, sum in power. Splitting a large sensor into subpixels
can be also beneficial when the rate of incoming charge
packets is large (e.g., millions of photons per second in x-ray
and gamma-ray detection), again at the price of parallelizing
the whole acquisition chain up to the digital processing stages.
In this context, it can be highlighted how analog processing, especially
when leveraging microelectronic integrated circuits,
though less versatile, is typically more compact and thus, scalable
compared with the digital one, often offering comparable
performance [6].
The second method is the reactive canceling of the capacitance
by means of an inductor [7]. This is a simple and effective
solution, though limited by the properties of the resonance:
the noise improvement is limited to about one order of magnitude
due to the quality factor of the inductor and is suitable
only for narrow-bandwidth single-frequency (the resonance
frequency) measurements such as impedance time tracking at
a single frequency.
The third approach consists in the reduction of the distance
Fig. 3. Typical architecture of a current readout chain composed of an input
transimpedance amplifier, an optional shaping filter and an ADC allowing data
sampling and processing by a digital embedded device.
32
between the electrode and the input node of the amplifier.
This is mandatory when the stray capacitance of this interconnection
is the dominant source of noise. Miniaturization
of electronic components and boards enables more compact
realizations that can be placed very close to the sensing electrodes.
Such miniaturization trend is aligned with similar ones
IEEE Instrumentation & Measurement Magazine
June 2022
Instrumentation & Measurement Magazine 25-4

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 25-4
