3 n-side of the Detector

We now switch to the n-side of the detector. Many of the measurements at a bias voltage of 40V are very similar to those of the p-side. This may be an indication that the parasitic elements in the probe dominate over anything we are trying to see in the detector itself. The n-side is different in several ways from the p-side. There are two wafers whose strips run along the length of the detector, compared to the p-side where the strips run along the width of the detector. The AC metal strips above are connected together and while each strip still has its own bias resistor. This means that when we measure the impedance between the AC metal and the bias line, we measure twice the $C_{AC}$ since there are two strips connected together. In addition, we also see two $R_b$ in parallel and so the resistance appears to be half that of the bias resistance. We show some data for an n-side strip 252 at 1V input amplitude and full bias of 40V along with the data from the p-side for comparison in Fig. 15.

Figure 15: Impedance and phase measurements for both sides

From the plots, it can be seen that the most difference is found at low frequencies where the value of the bias resistance is effectively halved compared to the p-side. The value for $C_{AC}$ is about 96 pF for n-side, however, the probe at low frequencies effectively see twice that amount so that the effective value is 192pF which is about the same as $C_{AC}$ on the p-side.

We can do the same type of modeling for the n-side with a network of resistors and capacitors. If we assume that the parasitics are the same, meaning that we keep the values that we used for the p-side,

$$\begin{array}{ccc} C_1 &=& 80 \textrm{pF}\\ C_2 &=& 30 \t... ...\textrm{pF}\\ R_{big} &=& 150 \textrm{M}\Omega\\ \end{array}$$ (14)

we get somewhat good agreement as show in Fig. 16.

Figure 16: Simulation of a complicated network model vs. data

By altering some of the values shown above, we can get better agreement as shown in Fig. 17 using the values

$$\begin{array}{ccc} C_1 &=& 80 \textrm{pF}\\ C_2 &=& 50 \t... ...\textrm{pF}\\ R_{big} &=& 300 \textrm{M}\Omega\\ \end{array}$$ (15)

Figure 17: Simulation of a complicated network model vs. data

Another distinction between the n-side and the p-side is that the n-side has p-doped p-stops between the strips to prevent them from shorting together. The p-stops only work when the detector is fully depleted, which happens at around 30V. So as we vary the bias voltage on the n-side, the strips start shorting together and so we should see effectively more and more resistors in parallel. So the effective value of $R_b$ should drop and we can see this clearly from the plots shown in Fig. 18.

Figure 18: Impedance and phase at various bias voltages