• Emitter sheet resistance to touch new highs – from the current level of 90 to 100 ohm/sq to 135 ohm/sq in 10 years
  • Uniform emitters realized with traditional tube furnaces has been and will dominate the market
  • All selective doping put together will attain 15% share in 10 years, while about double this level was predicted in previous roadmap
  • ITRPV dramatically revised its forecast for ion implantation – max of 2% by 2023
  • In boron diffusion, thermal diffusion employing BBR3 to dominate; current roadmap sees brighter chances for solid precursor + drive-in approach over ion implantation

Moving to high sheet resistance emitters is perhaps the most significant change solar cell processing has undergone and the march seems to continue, at least according to the International Technology Roadmap for Photovoltaics. The latest 7th edition released in late March emphasized that increasing emitter sheet resistance is an important technology trend within the diffusion segment, and it is expected to go on for the next 10 years.

As for the process employed for phosphorus diffusion, the well establishing tube diffusion method employed for forming homogeneous emitters has been and will lead the market, actually with a more prominent presence than expected in last year’s study (see article Many Ways Of Diffusing Impurity). For the remaining portion of the market, though small, ITRPV still foresees good prospects for selective emitters, especially using laser processes. At first, it appears that the 7th roadmap allocates higher market shares for selective laser doping and selective etching over ion implantation technology, but in reality, the estimates for selective doping methods have not changed considerably, while the current roadmap has significantly lowered its prospects for ion implantation methods. Under the section inducing dopant impurity, the current roadmap, like the previous one, also discussed methods of diffusing boron, mainly employed for creating the emitter of n-type cells. However, emitter sheet resistance has been a perennial subject.

Emitter sheet resistance to touch new heights
Solar cell emitters have to meet two opposite requirements. The area under the contact needs a high doping density in order to facilitate good contact formation, while the open area is lightly doped to enhance the blue response. This variable doping is nothing but the formation of a selective emitter. However, an even easier method is to develop a paste that can also establish reliable contacts with lightly doped emitters, which eliminates the hassle of forming selective emitters with similar performance. The second approach has been the driving force for recent developments in cell processing lines, and the feasibility actually killed the prospects for selective emitters in an early stage. Emitter sheet resistances between 90 and 100 ohm/sq are already in production, says the 7th version of ITRPV. The further rise is expected – to 110 and 120 ohm/sq in 2018 and 2020. The estimates until this point are basically at last year’s predictions. However, the previous two roadmaps expected an increase in emitter sheet resistance, which would reach a saturation point at 120 ohm/sq, while the latest study sees a scope to increase it further. While sheet resistance is expected to be stagnant at 120 ohm/sq between 2020 and 2023, it would take an upward path in the subsequent 3 years touching 135 ohm/sq in 2026. A comparative development for metallization pastes is necessary to realize emitters with such high sheet resistance.

Diminishing prospects for being “selective”
Selective emitters can also facilitate this march to very high sheet resistance emitters. It actually enables configuring sheet resistance of the cell’s open area independent from the portion under the metallization lines. The previous 6th roadmap expected selective emitters would have a share of about 6% in 2015, but the current document says it was about 3.5% and above 5% in 2006. The 7th roadmap estimates that the share of selective emitters would cross 10% in 2020 and account to 15% in the coming 10 years, whereas the 6th edition anticipated this level to be attained already in 2017 and to nearly double in 2025.

Old still holds good for uniform phosphorus doping
As for the methods to induce phosphorus dopant, the traditional gas-phase diffusion realized with tube furnaces is and will retain its dominance. The current roadmap expects the share of this time-tested technology to form a homogeneous emitter will lose only 10% points in 10 years, from 95% in 2015 to 85% in 2026. The 6th version also estimated the same starting point, but for 2014 and expected it to fall significantly to just above 60% in 2024.

Standard will be standard: Thermal diffusion realized in tube furnaces, which is the current state of the art, is also expected to play a leading role even after 10 years, estimates the 2016 ITRPV study.

Standard will be standard: Thermal diffusion realized in tube furnaces, which is the current state of the art, is also expected to play a leading role even after 10 years, estimates the 2016 ITRPV study.

Theoretically, inline furnaces using liquid precursor can also be employed for phosphorus diffusion, but the technology has nearly lost its ground for tube furnaces, as the latter is considered to result in higher efficiency. However, ITRPV has not considered inline diffusion in its study. Then, ion implantation is yet another alternative for forming uniform emitters. The 6th edition of ITRPV was optimistic about the prospects of ion implantation – from almost nothing to grab a 10% market share in 10 years. This year’s study surprisingly predicts the share of ion implantation process would stay just above 1% till 2020 and touch 2% in 2023, then fall back to be above 1% in 20206. While ITRPV has not provided the reason for its revised forecast, it is obvious that the withdrawal of one of the major proponents of this process and at the same time a leading supplier of PV ion implantation tools weighed into the forecast. Last summer, Applied Materials said it would stop its activity in this segment.

Lasers to lead in processing selective emitter
ITRPV’s scenario is nearly same with the application of ion implantation for the formation of selective emitters. Introduction of the technology is expected in 2020, in contrast to 2017 as estimated in the previous roadmap, and its market presence is projected to be between 1% and 2%, while 6th edition allocated 9% by 2024. As for the traditional methods of realizing selective emitters, the 7th roadmap has eliminated selective printed dopants, whereas laser doping and chemical etch back methods are considered the major technologies. Among the duo, laser based processes is leading, increasing its presence from close to 2% in 2015 to about 8% in 2026. While the previous roadmap has also estimated similar long term prospects, but anticipated the technology would attain 7% in 2017, the current study expects it to be at 4% in 2018. However, the current and former roadmaps have similar opinion about application of etch back methods to realize selective emitter. With a small share of just above 1% in 2015, the process would progressively gain a 5% share by 2026, in line with the progress expected in the 6th roadmap.

BBr3 thermal diffusion dominates n-type cell processing
While all the above updates are related to employing phosphorus as a dopant, mainly related to p-type wafers, ITRPV has also provided an update on boron diffusion, which is a key step in n-type cells processing. Thermal diffusion using boron tribromide (BBr3) as the precursor is currently the most widely used process with a share exceeding 80%. The previous roadmap estimated the same outcome. The difference is that the latest version predicts that thermal diffusion retains its dominance with a 60% share even after 10 years, while the previous version expected the thermal process to lose considerably to ion implantation, both reaching close to 40% in 2025. In fact, the thermal process is not easy – the byproduct of the process using BBR3 acts as a glue to quartz, thus creating a lot of maintenance issues. Then, boron diffusion typically has higher process times, lowering the throughput of the diffusion furnace by about half compared to POCl process. Ion implantation, on other hand, addresses majority of these shortcomings. The technology offers an effortless switch between the precursor elements and it also simplifies overall n-type cell processing, especially by eliminating masking and cleaning steps required with thermal diffusion. However, according to ITRPV, implants are not expected to make great impact on PV manufacturing. The 7th roadmap estimates that ion implantation technology would attain a peak market share of about 17% for boron diffusion in 2018, increasing from 11% in 2015, but falls again to about 11% in 2026. The high capital costs associated with ion implantation as well as the complexity of equipment that makes operation and maintenance more difficult are blockages the technology is facing.

A lot has changed: Compared to the previous study, the 2016 ITRPV roadmap considerably changed its assumptions for technology shares of boron diffusion – thermal diffusion will dominate even after 10 years and the ‘solid precursor + drive-in’ approach is expected to become more popular than ion implantation.

A lot has changed: Compared to the previous study, the 2016 ITRPV roadmap considerably changed its assumptions for technology shares of boron diffusion – thermal diffusion will dominate even after 10 years and the ‘solid precursor + drive-in’ approach is expected to become more popular than ion implantation.

Intermediate approach for boron diffusion
There is an intermediate method for boron diffusion that avoids the cleaning and masking steps to an extent and can also take advantage of low cost thermal process. It is a hybrid approach – it decouples the process of applying dopant element from the drive-in step. The latter can be accomplished in traditional diffusion furnaces. Actually, ion implants also require furnaces for dopant activation. The difference is that it is based on a solid precursor, Boron Silicate Glass (BSG), which is applied onto the wafer surface by means of Atmospheric Pressure Chemical Vapor Deposition (APCVD). The advantage of this method, especially for n-type cell processing is that both boron emitters as well as phosphorus Back Surface Fields (BSF) required for this type of cell architectures can be realized in a single drive in step called co-diffusion. The current ITRPV sees a slight upward potential for this approach. The technology with its share of close to 5% in 2015 would capture more than 1/4th of the market by 2026, while the previous report also estimated the same starting point, but was expecting to attain a 20% presence by 2025.

Giving n-type hard times
All these market shares appear significant in relative terms, but are much smaller in absolute terms. Because, the n-type market itself is small. That means the real term progress of these approaches mainly depend on a switch from p-type to n-type. On the other hand, p-type technology is progressing in many ways that is diminishing the incentives for switching to n-type. One such advancement is moving to very high sheet resistances. The current roadmap expects sheet resistances of 120 ohm/sq is attained already 2020. Maintaining this level for the following 3 years, the emitter sheet resistance is expected to touch new heights of 135 ohm/sq. When this happens, emitter sheet resistance would again remain as the most significant contribution to the performance increase of standard solar cells and become a major bonding factor for sticking standard cell structures. At the same time, it also creates a high level entry barrier for any new technologies.