N-type solar cell. Image: IMEC.
Why are the two c-Si solar cell concepts with the highest efficiency, IBC from SunPower and HIT from Panasonic, based on n-type technology and out there for a very long time? Why is almost 90% of PV production still based on p-type c-Si technology? Will that change in the future? The latter has been one of the most-discussed questions in c-Si PV for a few years now. From 23 March the SiliconPV conference and more specifically the 5th nPV workshop and HERCULES workshop will address this topic, showing improvements in p- and n-type Si technologies.
The reason why p-type technology is dominant today has definitely mostly a historical background. The history and past status was described in the 21st edition of Photovoltaics International in 2013. So we are very confident, also in agreement with the ITRPV consortium, that n-type technology will gain more and more importance in the share of technologies, as many companies are upgrading their p-type or n-type cell lines and even investing in new capacities, as summarised at the end of this blog post. As depicted in Figure 1 the PV world in future will be divided into high-performance (HP) p-type mc-Si cells with >22% efficiencies and n-type cells with >25% efficiencies.
Fig. 1: Relative market shares of casted and mono c-Si. Source: http://www.itrpv.net.
The fact that the two cell technologies featuring the highest efficiencies in industrial production are based on n-type Cz-Si wafer is a striking demonstration of why n-type wafers are the most suitable material for high-efficiency solar cells. Going more into details, there are some physical reasons for the superiority of n-type versus p-type, the most important are:
- due to absence of boron, there is no light induced degradation (LID) occurring in p-type Si wafers, due to boron-oxygen complexes
- as n-type Si is less sensitive to prominent metallic impurities, in general the minority carrier diffusion lengths in n-type Cz-Si are significantly higher compared to p-type Cz-Si
- n-type Si is less prone to degradation during high temperature processes such as B-diffusion.
Accordingly it can be assumed that, in order to guarantee an electrical quality in the wafers that is sufficiently high for the fabrication of solar cells with a cell efficiency of over 20% (in particular regarding the minority charge carrier diffusion length), ‘average’ to ‘high’ quality wafers are required for p-type, while for n-type ‘low’ to ‘average’ should be sufficient. Taking for granted that the wafer prices indicated in Figure 2 are related to the wafer quality, currently, the wafers for such solar cells should be priced in the range of US$1.08-1.22 in the case of p-type, while suitable n-type wafers have a price ranging from US$1.25-1.30.
Fig. 2: Updated market prices (USD/wafer) for p- and n-type Cz-Si wafers (according to http://www.pvinsights.com: "solar PV wafer weekly spot prices, last update: 2015-03-11).
Accordingly, depending on the individual supply situation, n-type wafers can be up to 20% more expensive than p-type wafers. However, experts in industrial Cz-Si crystal growth agree that – apart from a wider resistivity distribution over one crystal – there is no technological difference between the growths of p- and n-type crystals that would explain an increased manufacturing cost for n-type wafers. Therefore it all comes down to economy of scale: currently, more than 80% of the worldwide Cz-Si crystal production capacity for PV is dedicated to p-type. According to the expectations of the latest ITRPV (see Figure 1) parity between p- and n-type productions will be nearly achieved by 2018. Then, at the latest, all cell manufacturers should have access to n-type wafers at the same price as p-type wafers. Vertical integration into crystal growth and wafering would be a way for cell manufacturers to achieve the breakeven earlier.
The growing market share of n-type and the availability of n-type modules at standard price levels will also result in a higher awareness among end-users regarding the LID issue of p-type modules, highlighting another benefit of n-type in terms of levelised cost of energy (LCOE).
One of the possible solutions for avoiding LID in p-type, apart from using more costly Cz-Si with low oxygen content, is the permanent deactivation of the B-O-complexes by a combined heat-illumination treatment. Equipment that is capable of performing this treatment in a reliable way with industrial throughput is currently under development at some equipment suppliers, such as Centrotherm.
In conclusion, considering that cell concepts enabling efficiencies significantly exceeding 21% will require boron diffusion, the lower degradation of n-type Si during high temperature processes will make it the predominant wafer material for industrial solar cell manufacturing in the mid-term future.
The advantages and drawbacks of different diffusion technologies were already discussed in our last article for Photovoltaics International, cited above. Nothing much has changed since then – the B-diffusion has to be performed fast, cost-effectively and homogeneously which is not that easy. Centrotherm is focusing on low-pressure BBr3 diffusions and Sandvik has designed a BBr3 boron deposition process which has a unique gas flow configuration. Tempress and Semco also have solutions for tube furnace BBr3 and BCl3 diffusions respectively, while Schmid has also been very successful with its APCVD B-diffusion equipment where the diffusions are done from one-sided deposition of doped oxides. AMAT, Kingstone, Intevac and others are offering solutions for ion-implantation, however not all of them for boron.
The challenge after the process of a good B-diffused surface is to clean it properly, passivate it effectively and metallise without high losses in Voc. There are many solutions for special applications which are provided by RENA (cleaning, metallisation), Schmid (cleaning, metallisation), SINGULUS (cleaning, passivation, metallisation) centrotherm (passivation), R&R (passivation), Levitech (passivation), Solaytech (passivation), Dupont (metallisation), Hareaus (metallisation) and many others. In order to reach efficiencies in production exceeding 21-22%, the Al spiking has to be eliminated in future which leads to limited Voc of about 655mV. Also the edge isolation is not trivial on n-type solar cell processing and has to be implemented thoroughly into the cell process.
Increasing significantly the cell efficiencies compared to standard Al-BSF cell technology requires the introduction of additional process steps as discussed before. Consequently, advanced cell concepts such as PERC, MWT and PERT come with a higher cost of ownership (COO) in US$/cell. HIT and today's industrially implemented IBC-cell concepts feature in addition more complex (i.e. more expensive) process technologies.
As mentioned above, n-type concepts (n-PERT, HIT, IBC) have been – up to now – also disadvantaged by a higher wafer price. However, depending on the achievable module power and on the module manufacturing cost, the COO in US$/Wp at module level can still be economically interesting. This applies even more, if one considers the parameter that determines the return on investment of each PV system, irrespective of its exact type and size: the LCOE in US$/kWh. When building a PV system with high efficiency modules, less cabling, mounting structures, land, labour and other elements are required – in short: the related balance of system (BOS) cost, and consequently the total cost of the installed system, is reduced. A lower cost of the PV system (US$/Wp) results in a lower LCOE (US$/kWh).
In this respect, many advanced c-Si solar cell technologies with boron back surface field or emitter are inherently bifacial or can be easily made bifacial. Bifaciality reduces the LCOE even further, as it acts just like an "efficiency booster": as illustrated in Figure 3, assuming a moderate 15% bifacial gain (increase in kWh/kWp(front)), a bifacial BiSoN module with 280Wp (front Pmpp) features the same energy yield as a 320Wp monofacial module while featuring a COO that is in the range of standard mc-Si modules with a Pmpp of 250 Wp.
Fig. 3: Market price vs. Pmpp of 60-cells modules for various industrially implemented cell technologies compared to the calculated Coo for BiSoN (n-PERT) and ZEBRA (n-IBC) technology. Coo calculation based on 1.0 USD/wafer for p-type Cz-Si and 1.3 USD/wafer for n-type Cz-Si.
Newcomer companies and new R&D Highlights
Many companies have been involved in n-type cell and module production for many years such as SunPower, Panasonic, Yingli, PVGS, Neosolarpower and LG. Newcomers such as First Solar, Silevo, Mission Solar (Nexolon), SSNED, Motech and MegaCell are following quickly. Many of them will present their progress at the nPV workshop, showing >20% efficient n-type cell concepts that can be also used in glass-glass or glass-transparent foil modules, benefiting from the bifacial character of the cells.
Many institutes are developing similar cell concepts on large six-inch wafers (some still having some processes which are hard to transfer to production) and reaching efficiencies >22% for a simple n-type PERT structure (IMEC) or >22% for an IBC structure (ISFH). Results from IMEC´s, FhG’s, ISE´s and ISFH´s n-type developments are summarised in the 27th edition of Photovoltaics International. All these results will be also shown in presentations at the nPV workshop. In addition, ISC Konstanz together with MegaCell will show BiSoN (BIfacial Solar cell On N-type) cells in production and the newest improvements to the ZEBRA (diffused n-type IBC) technology with >21.5% efficiency (both concepts depicted in Figure 4). ECN will present their upgrade of nPASHA – the n-type MWT cell with >21% efficiency. Last but not least INES and EPFL/CSEM will summarise their excellent results on heterojunction cell and module concepts.
Fig. 4: PERT (BiSoN) solar cell and IBC (ZEBRA) solar cell of ISC Konstanz.
In summary we can say that n-type is on its way to rapidly taking off. The cell concepts have been out there for a long time, the wafers are constantly getting cheaper, the paste manufacturers are improving their n-type products quickly and the awareness of kWh thinking instead of Wp mentality is growing. So the ramp for take-off is prepared - and the n-type rocket is now starting its engines.