Researchers from the National Chin-Yi University of Technology in Taiwan proposed a novel heater design for Czochralski silicon crystal growth equipment to control and decrease oxygen concentration.
The method was simulated and validated finding that the optimal design enabled oxygen reduction of 6 parts per million (Ppm) atoms.
For both multicrystalline and monocrystalline silicon photovoltaic applications, oxygen is a key impurity issue. For example, it can cause silicon oxide formation, which increases the hardness of crystals, which can complicate downstream processing.
“Our findings indicated that certain oxygen defects reduce the bulk lifetime and enhance recombination activity at dislocations,” Amir Reza Ansari Dezfoli, the research's first author said.
There are several ways to tackle this type of problem. “In our study, we focused on controlling, mainly reducing, the oxygen impurity by modifying the heater design in the Czochralski (CZ) puller,” said Dezfoli, noting that an oxygen reduction of 6 Ppm atoms could be achieved by “simply altering” the heater design configuration.
“It was particularly interesting to discover that a low-cost modification to the heater could significantly impact oxygen impurity control, showing how a simple adjustment can have a major effect,” he added.
The study first investigated heater design and distribution of heat sources through simulations, and then performed testing on an experimental setup to validate. The simulation was based on a CZ crystal growth of a silicon ingot with a diameter of 200 mm and a length of 700 mm, with four different heater designs. The heaters were made of graphite and varied top section lengths, ranging from 500 mm to 200 mm.
To validate, the experimental setup used a heater height (H1) of 500 mm. The Fourier transform infrared spectroscopy (FTIR) data analysis was used to measure oxygen concentrations in the crystals along the axial direction at the ingot centerline. Measurements of heat transfer and impurity transport, the heater power, maximum crystal front deflection, and oxygen concentration were conducted. Temperature profiles within the heater was analyzed, as well as the impact on silicon melt and crystal.
The validation comparison confirmed a “strong agreement between simulation and experimental data.”
Concluding that the heater design “significantly influences temperature distributions and melt patterns, affecting oxygen distribution and its transport mechanisms,” the paper listed the effects of varying the H1 length, noting it was clear that controlling the temperature profiles and melt flow patterns “significantly” impacts oxygen distribution.
The group is now preparing to introduce the first discrete model to simulate the formation of crystal originate particles (COP) and void defect formation during the CZ process. “It will be the first model capable of functioning similarly to laser particle counting in the silicon wafer production industry, opening up new possibilities for simulation-based quality control,” said Dezfoli.
The proposed technology was presented in “Engineering insights into heater design for oxygen reduction in CZ silicon growth,” published in Case Studies in Thermal Engineering,
