It is often difficult to know when a photoresist problem is affecting the productivity of a 300 mm PVD system. Even small amounts of residual resist, when introduced over a prolonged period, can reduce yields, cause particle contamination or slowly contaminate process chambers, thereby requiring premature kit changes. Major photoresist hits can be very costly, upwards of hundreds of thousands of dollars for lost wafers, especially if the tantalum/copper target becomes contaminated.

Resist can be inadvertently introduced to PVD systems in a number of ways. Rushing Hot Lots through opens the door to skipping steps like the resist strip or even the complete etch and strip cycle. This could result in contamination throughout the process tool if the wafer is not stopped at the degas chamber. Such process flow errors can impact PVD tool performance, immediately, from a single wafer. More common, however, is the incomplete removal of resist. This can happen due to a poorly controlled end point in the asher such as from firstwafer, cold-chamber effects, or from faults on integrated etch/strip tools.

Each of these mechanisms can leave resist contamination on wafers which eventually is seen as a loss of die or high particle densities in sputter etch process chambers. Resist contamination can be the cause of a series of factors, all of which work to reduce tool availability, increase COO, and decrease the overall yield of a process flow.

The 300mm Resist-Torr considers many different parameters of the degas step and combines that data to provide information in the form of a meaningful number, without the need for user-interpretation. This measurement, termed the PR Index, is a normalized measure of wafer borne photoresist contamination that can be presented in various forms to compare data over time.

The 300mm Resist-Torr can interrupt the processing of any wafer automatically, without operator intervention, if a wafer enters the vacuum system with detrimental levels of residual photoresist, as determined by the PR Index.

PR Index trace for wafer sequence run at a foundry with interspersed normal and known PR contaminated wafers; red (alarm) and yellow (warning) lines show the default control limits PR Index trace for a series of test wafers run at a 300mm PVD tool OEM. The PR hit event was much longer than the preceding normal wafers because the tool was stopped on that wafer.