 |
| Image Source: https://www.123rf.com/photo_54736993_integrated-circuit-on-binary-background-ic-chip.html |
Antenna violation is a critical concern, especially as we advance to lower technology nodes where gate oxide is made thin. This article explains antenna violation and how to fix it.
Antenna violation occurs due to the build-up of charges on the gate terminal during metal etching. During the etching process, unwanted metal is removed to create the required mask pattern. It is done using plasma gas and is thus known as plasma etching. Plasma is an ionized gas and contains free, charged particles. These charges get deposited on the gate metal during the plasma etching process. This causes the gate capacitance to get charged up. As a result, the voltage across the gate capacitance increases and damages the gate oxide. So the built-up charges on the gate terminal are getting discharged through the gate oxide causing its damage. This is known as the antenna effect.
However, this charge build-up results in an antenna violation only if the threshold antenna ratio specified in the process development kit(PDK) is exceeded. The antenna ratio is the ratio of metal area to gate area.
Antenna Ratio = Metal Area / Gate Area
A larger metal area means it is more prone to antenna violation. Here a larger metal area does not necessarily mean the wire connected to the gate has to be long. Even a wide but short wire can cause antenna violation provided the area is large enough to cause the antenna ratio to exceed the specified threshold.
To fix antenna violation, we need to limit the area of gate terminal metal. There are 2 methods of reducing antenna violation.
The first one is by using a metal jumper of a higher metal layer. In this method, a small cut is made in the wire causing antenna violation, and replaced with a metal wire that is one layer higher. For example, if a violation occurs in the metal 2 layer, then the cut made is replaced with a jumper wire of the metal 3 layer. The cut is made near the gate. This method is used to fix antenna violations for layers in the middle of the metal stack. The logic behind this solution is that higher metal layers are fabricated after at least a 1-2 days gap after manufacturing the present layer. By this time the charges would have dissipated. Theoretically, we can use a jumper of a lower metal layer to fix the violation as well but practically this is not possible as the lower layer would have been manufactured already and a discharge path through the gate oxide will be available.
Fig.1. Fixing antenna violation using jumper wire method.
If antenna violation is occurring in the topmost metal layer of the stack, we cannot use the jumper method as there is no higher metal layer available to create a jumper wire. So here we go for the diode insertion method. A reverse bias diode acts as a resistor when subject to high temperatures during the manufacturing process. Thereby it provides an alternate path for charge dissipation and protects the gate oxide. The diode is inserted between the metal layer and the substrate close to the gate. If the violating metal layer is metal 3, the diode is inserted between metal 2 and the substrate. However, inserting the diode at metal 2 fixes the antenna violation for the layers above it as well to the end of the stack, so might as well just insert the diode at metal 1 so the violation is fixed for all layers.
Fig.2. Fixing antenna violation using diode insertion method.
The jumper method is most preferred for fixing antenna violations as it is just a small change in routing
whereas diode insertion is a change to the circuit’s logic and can cause timing violations which again have
to be fixed. Hence we should use the latter only when it’s really necessary.
Copyright © Silicon Cat 2023
Comments
Post a Comment