By Ian Latchford, J. Bruce True, Intevac, Inc.
The value for in-line PL inspection is large for the solar industry. Its uses are expected to grow as quality measures make a difference in manufacturing approaches that reduce overall cost/watt.
Today, silicon cell factories have moved toward the gigawatt (GW) scale. These factories churn out hundreds of thousands of cells each day and competitiveness is based around fractions of cents per watt in manufacturing cost, and tenths of points on the efficiency scale. Every little bit counts. Quality control is now becoming a critical factor to the future survival of solar companies' fortunes.
Efficiency: the ultimate quality measure
The ultimate value and quality measure of a solar cell manufacturing line is the cell efficiency. In a solar manufacturing facility, cell efficiency is typically only measured once – at the end of the cell processing line after a series of processing steps that add about 80 cents per wafer of manufacturing cost. Ironically, one of the largest impacts on cell efficiency is the quality of the silicon wafer itself at the very beginning of the line. Today, solar cells are indiscriminately processed the same way, regardless of silicon wafer quality, so lines cannot be optimized around silicon quality because it isn’t being measured. Because the silicon has a wide distribution in quality, the resulting efficiency distribution is usually quite a broad-skewed distribution with a long tail down to lower efficiencies (Fig. 1).
FIGURE 1. The typical solar cell broad-skewed distribution of cell efficiencies including a long, low, efficiency tail indicating lower quality silicon).
Quality control creates money
Two opportunities exist in a broad-skewed distribution from a quality control point of view. One is to cut out the long tail (low quality) and recycle bad silicon, thus saving additional processing costs for cells that might not be used. The second quality control technique is quality streaming, where the wider distribution is cut into narrower streams, which allow the subsequent processes to be tuned or optimized for higher and lower grade silicon (Fig. 2). Optimized narrow streams of silicon will likely result in higher cell efficiency in the line due to further line optimization. A calculation of the value for inline PL inspection in a 100 MW/Yr cell processing line results in the following potential financial gains: 1) a 0.2 percent efficiency improvement will increase the value of the manufacturing line by $1.2 million additional revenue annually, and an additional $0.3 million will be saved annually for each one percent of low quality silicon wafers not processed in the line. Simple quality control measures at the beginning of the line can have large effects on the bottom line.
FIGURE 2. Streaming wafers in to high and medium quality lines produces the highest efficiency at the lowest cost.
Photoluminescence
Photoluminescence (PL) is an inspection technique that can be used to predict cell efficiency at the beginning of the line. The technique is both accurate and very fast and can be run at line speeds (3600 wph) without stopping and without contacting the wafer (on the fly).
PL is the re-emission of light after absorbing light of a higher energy (shorter wavelength). Most of the photo-generated electrons give up their energy as heat, but a small fraction of the electrons recombine with holes, emitting a photon (radiative recombination). More defects will result in more energy lost as heat, and fewer emitted photons. Fewer defects will result in more radiative recombination and more emitted photons.
Silicon solar wafers and cells create luminescence (light) when excited with laser light. The luminescence from the silicon is emitted from 950 – 1250 nm with the peak occurring at approximately 1150 nm. The emission is caused by a band-to-band transition, and is a relative measure of the minority carrier lifetime and efficiency. Areas of silicon with a higher minority carrier lifetime have brighter luminescence and higher conversion efficiency (Fig. 3). Therefore, the luminescence image is a rapid measure of the overall quality of the wafer.
FIGURE 3. PL image of a multi-crystalline silicon solar wafer.
The need for speed and sensitivity
For fast and accurate PL to cell efficiency prediction a combination of sensor range and sensor speed is required. The range of the sensor is dictated by the sensor material, the wider the range capability the more likely that the information needed for correlation can be picked up at high speed. Intevac has developed a photoluminescence inspection system capable of acquiring high resolution photoluminescence images at very high speed (Fig. 4). This new PL inspection system enables faster solar cell inspection much more quickly with less illumination, and with higher accuracy than current camera technology. It has significantly higher quantum efficiency than silicon CCD cameras, and much higher sensitivity and lower dark current than InGaAs focal plane arrays.
Not all defects are bad defects
To a first approximation, simply adding up all the bright and dark pixels in the wafer image should lead to an estimate of the performance parameters of the finished cell. However, some types of defects have a large impact on the cell performance, while other defects may have only minimal impact, and still others may be mitigated during processing of the wafer. To achieve an accurate prediction of the cell performance, the impact of the defects must be quantified by analyzing their brightness and shapes. In addition, different process lines may be impacted more or less by different defect types. A robust image analysis algorithm must account for the differing effect of various defects, and provide a means to adapt to a given process line.
FIGURE 4. An InGaAsP sensor has high sensitivity to all silicon PL wavelengths.
Conclusion
The value for in-line PL inspection is large for the solar industry. Its uses are expected to grow as quality measures make a difference in manufacturing approaches that reduce overall cost/watt.
Ian Latchford holds a BSc in chemical engineering from London South Bank U., England and now runs the Solar Marketing team at Intevac, 3560 Bassett St., Santa Clara CA 95054 USA; ph.: 408-588-2138; email ilatchford@intevac.com.
J. Bruce True received his PhD in chemistry from the U. of Arizona and his BA degree in chemistry from the U. of Texas at Austin.
http://www.renewableenergyworld.com/rea/news/article/2011/08/an-eye-on-quality
The value for in-line PL inspection is large for the solar industry. Its uses are expected to grow as quality measures make a difference in manufacturing approaches that reduce overall cost/watt.
Today, silicon cell factories have moved toward the gigawatt (GW) scale. These factories churn out hundreds of thousands of cells each day and competitiveness is based around fractions of cents per watt in manufacturing cost, and tenths of points on the efficiency scale. Every little bit counts. Quality control is now becoming a critical factor to the future survival of solar companies' fortunes.
Efficiency: the ultimate quality measure
The ultimate value and quality measure of a solar cell manufacturing line is the cell efficiency. In a solar manufacturing facility, cell efficiency is typically only measured once – at the end of the cell processing line after a series of processing steps that add about 80 cents per wafer of manufacturing cost. Ironically, one of the largest impacts on cell efficiency is the quality of the silicon wafer itself at the very beginning of the line. Today, solar cells are indiscriminately processed the same way, regardless of silicon wafer quality, so lines cannot be optimized around silicon quality because it isn’t being measured. Because the silicon has a wide distribution in quality, the resulting efficiency distribution is usually quite a broad-skewed distribution with a long tail down to lower efficiencies (Fig. 1).
FIGURE 1. The typical solar cell broad-skewed distribution of cell efficiencies including a long, low, efficiency tail indicating lower quality silicon).
Quality control creates money
Two opportunities exist in a broad-skewed distribution from a quality control point of view. One is to cut out the long tail (low quality) and recycle bad silicon, thus saving additional processing costs for cells that might not be used. The second quality control technique is quality streaming, where the wider distribution is cut into narrower streams, which allow the subsequent processes to be tuned or optimized for higher and lower grade silicon (Fig. 2). Optimized narrow streams of silicon will likely result in higher cell efficiency in the line due to further line optimization. A calculation of the value for inline PL inspection in a 100 MW/Yr cell processing line results in the following potential financial gains: 1) a 0.2 percent efficiency improvement will increase the value of the manufacturing line by $1.2 million additional revenue annually, and an additional $0.3 million will be saved annually for each one percent of low quality silicon wafers not processed in the line. Simple quality control measures at the beginning of the line can have large effects on the bottom line.
FIGURE 2. Streaming wafers in to high and medium quality lines produces the highest efficiency at the lowest cost.
Photoluminescence
Photoluminescence (PL) is an inspection technique that can be used to predict cell efficiency at the beginning of the line. The technique is both accurate and very fast and can be run at line speeds (3600 wph) without stopping and without contacting the wafer (on the fly).
PL is the re-emission of light after absorbing light of a higher energy (shorter wavelength). Most of the photo-generated electrons give up their energy as heat, but a small fraction of the electrons recombine with holes, emitting a photon (radiative recombination). More defects will result in more energy lost as heat, and fewer emitted photons. Fewer defects will result in more radiative recombination and more emitted photons.
Silicon solar wafers and cells create luminescence (light) when excited with laser light. The luminescence from the silicon is emitted from 950 – 1250 nm with the peak occurring at approximately 1150 nm. The emission is caused by a band-to-band transition, and is a relative measure of the minority carrier lifetime and efficiency. Areas of silicon with a higher minority carrier lifetime have brighter luminescence and higher conversion efficiency (Fig. 3). Therefore, the luminescence image is a rapid measure of the overall quality of the wafer.
FIGURE 3. PL image of a multi-crystalline silicon solar wafer.
The need for speed and sensitivity
For fast and accurate PL to cell efficiency prediction a combination of sensor range and sensor speed is required. The range of the sensor is dictated by the sensor material, the wider the range capability the more likely that the information needed for correlation can be picked up at high speed. Intevac has developed a photoluminescence inspection system capable of acquiring high resolution photoluminescence images at very high speed (Fig. 4). This new PL inspection system enables faster solar cell inspection much more quickly with less illumination, and with higher accuracy than current camera technology. It has significantly higher quantum efficiency than silicon CCD cameras, and much higher sensitivity and lower dark current than InGaAs focal plane arrays.
Not all defects are bad defects
To a first approximation, simply adding up all the bright and dark pixels in the wafer image should lead to an estimate of the performance parameters of the finished cell. However, some types of defects have a large impact on the cell performance, while other defects may have only minimal impact, and still others may be mitigated during processing of the wafer. To achieve an accurate prediction of the cell performance, the impact of the defects must be quantified by analyzing their brightness and shapes. In addition, different process lines may be impacted more or less by different defect types. A robust image analysis algorithm must account for the differing effect of various defects, and provide a means to adapt to a given process line.
FIGURE 4. An InGaAsP sensor has high sensitivity to all silicon PL wavelengths.
Conclusion
The value for in-line PL inspection is large for the solar industry. Its uses are expected to grow as quality measures make a difference in manufacturing approaches that reduce overall cost/watt.
Ian Latchford holds a BSc in chemical engineering from London South Bank U., England and now runs the Solar Marketing team at Intevac, 3560 Bassett St., Santa Clara CA 95054 USA; ph.: 408-588-2138; email ilatchford@intevac.com.
J. Bruce True received his PhD in chemistry from the U. of Arizona and his BA degree in chemistry from the U. of Texas at Austin.
http://www.renewableenergyworld.com/rea/news/article/2011/08/an-eye-on-quality
No comments:
Post a Comment