The distribution of sodium in the oxide films of MOS capacitors was determined by etch-off procedures, using sodium-22 as a tracer. The effects on sodium distribution of ionizing radiation, of in-diffused aluminum, and of bias-temperature stressing were then determined and compared with C-V measurements of the total oxide charge in the same samples. Aluminum films sintered for 30 minutes at 500°C provided high aluminum concentrations (∼10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">20</sup> /cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> ) in the oxide near the surface, while sintering for 60 minutes caused aluminum to diffuse about three times deeper. In both cases low concentrations (∼10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">18</sup> /cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> ) were found next to the silicon. Aluminum-rich portions of the oxide were found to trap sodium, and thus the oxides with the 30-minute aluminum sinter had very little sodium concentrated near the silicon substrate (0.2 to 0.6 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> /cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ); the interfacial sodium accounted for only 2% of the oxide charge determined by C-V measurements. Oxides without aluminum or with 60-minute aluminum sinter had nearly 10-fold higher interfacial sodium concentrations (1 to 6 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> /cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) and this interfacial sodium accounted for 10% of the oxide charge from C-V measurements. Ionizing radiation (10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> rads in one hour from cobalt-60 at 25°C with a positive bias of 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> volts/cm) increased the total oxide charge by 14 to 17 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> electronic charges/cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , but caused very small increases, in interfacial sodium (0.2 to 0.9 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> ions/cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ). However, the combination of ionizing radiation followed by positive bias stress (10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> volts/cm) at 300°C for 10 minutes caused an appreciable movement of sodium to the interface in some of the oxides, especially those with no aluminum or with aluminum sintered for 60 minutes. Positive bias at 300°C for 10 minutes in aluminum-doped oxides caused large increases in negative voltage shifts (corresponding to an increase of 10 to 20 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> positive charges per cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) but in the same tests no more than 3 &time; 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> sodium ions per cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> moved into the oxide next to the silicon substrate. These results suggest that the drift of positively charged aluminum interstitials may be the cause of much of the voltage shift in 300°C tests.
Read full abstract