The Challenge of Variable Dielectrics
Imagine looking at the capacitor from the side. We have two square plates of side a, separated by a distance d. The lower triangular region is filled with a dielectric of constant K. Because the thickness of the dielectric varies continuously as we move along the plate, we cannot use the standard formula C=dε0A directly. The electric field lines experience different amounts of dielectric material depending on their horizontal position.
Slicing the Capacitor
The Differential Element
To find the total capacitance, we must slice the capacitor into infinitesimally thin vertical strips. Each strip acts as a tiny, independent capacitor. Since these strips are placed side-by-side between the same two equipotential plates, they are connected in parallel. The total capacitance will simply be the integral (sum) of the capacitance of these tiny strips.
Let's look at one such strip of width dx at a distance x from the left edge. Inside this strip, the electric field passes through the dielectric and then through the air. This means we have two capacitors in series: the lower part filled with dielectric of height y, and the upper part filled with air of height d−y.
The Series Combination Inside the Strip
First, we need to express the height y in terms of x. By observing the similar triangles formed by the dielectric boundary, the ratio of y to x is equal to the ratio of the total height d to the total length a.
Now, let's write the capacitance for the two parts of our strip. The area of the strip is a⋅dx (since the plates are square, the strip extends a distance a into the page).
The air part has capacitance C1=d−yε0(a⋅dx).
The dielectric part has capacitance C2=yKε0(a⋅dx).
Since they are in series, their equivalent capacitance dC is given by C1+C2C1C2. Substituting C1 and C2 into the series formula and simplifying the algebra, we get:
dC=y+K(d−y)Kε0adx=Kd+(1−K)yKε0adx
The Master Integral
To find the total capacitance C, we integrate dC from x=0 to x=a. We substitute y with adx in the denominator to express everything in terms of x.
C=∫0aKd+a(1−K)dxKε0adx
This is an integral of the form ∫A+Bx1dx. Integrating it gives a natural logarithm. Let's use substitution. Let u=Kd+a(1−K)dx. This implies dx=(1−K)dadu.
When x=0, u=Kd. When x=a, u=Kd+(1−K)d=d.
Evaluating the Logarithmic Integral
Substituting these into our integral, we get:
C=∫KdduKε0a(1−K)dadu=(1−K)dKε0a2∫Kddu1du
Evaluating the integral yields lnu from Kd to d:
C=(1−K)dKε0a2(lnd−lnKd)
Using the property of logarithms, lnd−lnKd=ln(Kdd)=ln(K1)=−lnK.
Finally, absorbing the negative sign into the denominator gives us our elegant final expression: