Publication Date


Document Type



As the microelectronic industry progresses toward smaller devices, a decrease in the thickness of gate oxides accompanies them. High quality cannot be sacrificed as a result of this shrinkage. It is believed that oxide quality can be related to oxide charge density. Total oxide charge is related to a shift in Vt for transistors and believed to be related to oxide breakdown strength.

A surface charge analyzer on loan from the SemiTest Corp. was used to quantify oxide charge densities after thermal oxidation of 300 A oxides under various processing conditions. Variations to the base process included the temperature at which wafers were pulled from the furnace, leaving the endcap of the furnace tube ajar, whether the wafers had been handled with tweezers or a vacuum wand, whether trichloroethane (TCA) had been included during the oxidation, and the type of wafer (P or N . )

The base process was as follows: 1) TCA cleaning of the furnace tube for 30 min. at 1100 C before processing 2) inserting the wafers at 900 C at 12 inches/min. with 5 LPM N2 flowing 3) ramping up to 100 C with 5 LPM dry O2 4) soaking for 18 min. with 5 LPM dry O2 5) ramping down to 750 C with 5 LPM N2 6) and pulling at 12 in/min with 5LPM N2.

The resulting oxide thicknesses were measured using a Nanospec.

Early results indicated a strong dependence on wafer type. Generally P type wafers showed oxide charge densities (Qox) an order of magnitude lower than N type wafer (E11 vs E12.) This proved to be a result of the measuring equipment and the problem was corrected.

Final results indicate no difference in Qox values for P vs. N wafers, no difference in Qox due to handling procedures, differences related to pull temperature, and an ability to identify wafers that could be considered unfit for device fabrication.

Included in

Engineering Commons