In our recent publication [1], we gave a physical interpretation to the century-old Curie–von Schwei-
dler law of dielectrics which interestingly may also be seen as an electrical analogue of Nutting’s law
from rheology [2, 3]. The derivation of the Curie–von Schweidler law required a revision of the classical
charge–voltage relation of capacitors, Q(t) = C0V(t), where Qis the charge, C0is the constant capac-
itance of the capacitor, Vis the voltage, and tis the time. It is worthwhile to note that the classical
relation is more than two-centuries old [4] and it was originally proposed for a constant capacitance
capacitor. In contrast most capacitors of practical applications exhibit a time-varying capacitance
but in lack of a better relation the same-old classical relation has been inadvertently used. In order
to overcome this lacking we proposed a new charge–voltage relation which is equally applicable for a
constant capacitance capacitor as well as for a capacitor with a time-varying capacitance. The relation
is expressed by Eq. (9) in our paper as [1]:
Q(t) = C(t)∗˙
V(t),(1)
where ∗represents the convolution operation and C(t) is the time-varying capacitance of the capacitor.
In their Comment on our paper, Allagui et al. [5] have raised certain concerns with the use of the
convolution integral of linear time-varying capacitance with the time-derivative of voltage. In this
Response, for the sake of clarity, completeness, and correctness, we prove that each of their observations
and the inferences that they have drawn from them are quite questionable and therefore not valid.
1. According to Allagui et al. capacitances being added in parallel is a frequency domain assertion,
and that they cannot be added in the time domain [5]. Moreover, they have the opinion that
circuit synthesis is not done in the time domain.
Response: We would like to emphasize that as long as the linearity of the system holds, circuit
synthesis could be done in the time domain as well as in the frequency domain. We transform
a problem from one domain to another domain depending upon the domain in which finding a
solution to the problem is relatively easier. For example, modeling a digital filter is often easier
in the frequency domain as it facilitates a convenient framework for the mathematical analysis
of its properties. In contrast, a sampling process is meaningful in the time domain. The two
domains should be seen as the two different viewpoints from which to understand the problem.
It is also worthwhile to respect the fact that the time domain is the domain of physical reality
and it is probably more intuitive than the frequency domain.
It has already been shown that our expression for current matches exactly with the predictions
2