Victor Prajapati

Oxidation enhanced diffusion for screen printed silicon solar cells

If the worlds answer to alternative energy production is to be silicon photovoltaics, the fabricated devices need to be produced with the lowest dollar per watt peak. This may be translated to a high efficiency at competitive costs. An easily implementable approach to increase solar cell efficiency without incurring too much cost is to include thermal oxidation [1,2,3,4]. Thermally grown silicon oxide is well known in the microelectronics industry as one of the best dielectrics to passivate silicon surfaces. This paper will focus on two simultaneous properties related to thermal oxidation. The first is the phosphorus emitter and how it can be significantly altered even at low temperatures (800C). The second is passivation due to the thermal oxidation of both front and rear surfaces. A 60 and an 80 Ω/□ emitter (compatible with silver screen printed contact formation) will be investigated with various oxidation conditions using secondary ions mass spectroscopy (SIMS), scanning spreading resistance (SRP), emitter saturation currents (Joe) as well as final solar cell devices. We report that increasing oxidation significantly decreases phosphorus surface concentration (Ns) while increasing junction depth. The final cell results show that increasing oxidation increases both open circuit voltage (Voc) as well as fill factor (FF), however the current density (Jsc) is reduced partly due to front reflection loss. The cells studied in this paper are fabricated on 149cm2, 155μm thin 1.5 ohm.cm p-type Cz-Si wafers that have screen printed Ag front contacts and a thin thermal oxide on both sides with a rear deposited oxide/nitride stack The highest confirmed efficiency of the cells studied is 19.9% with a Voc of 654mV, Jsc of 38.4 mA/cm2 and a fill factor of 79.3%.

Advanced approach for surface decoupling in crystalline silicon solar cells

Texturing and polishing crystalline silicon in the PV industry are both time consuming and inefficient, the need for a streamlined solution is prevalent. To obtain high conversion efficiency surface decoupling is needed, one side of the wafer is ideally textured while the other is completely polished. The front is textured to collect the maximum number of photons while the rear acts as a perfect mirror to reflect all outgoing light back into the silicon for absorption. In industry, crystalline silicon wafers initially undergo a saw damage etch (SDR) in which 10?m of silicon is removed from each side. Wafers are then placed in an alkaline texturing solution which textures both sides which removes another 5µm from each side. Finally the wafers are single side polished, removing a final 10µm from one side. This industrial process results in a total silicon loss of 40µm and a complete surface decoupling cycle time in the order of an hour from as cut until diffusion ready. For an industrial wafer thickness of 180µm more than 20% of the silicon is lost. In addition to silicon loss, the multiple wet processing steps needed for decoupling are time consuming and hinder high throughput manufacturing needed to achieve the ultimate goal of a low per watt peak price. To increase efficiency more process steps are needed, cell concepts such as the i-PERC, i-PERL and the i2-BC all need an increasing number of process steps leading to higher efficiencies. As the number of process steps and complexity increases, the variability in the overall process will increase as well; this will ultimately make these concepts unrealizable by industry. It is desirable to develop processes and techniques that are of multi-purpose. Such a technique is proposed in this paper, this technique decouples the front and rear surfaces and utilizes only one wet process step for both the texturing and polishing. This process simplification leads to savings by decreasing costly cycle time, saving consumables and reducing the amount of silicon loss. The complete process takes a few minutes from an as cut wafer to a fully decoupled wafer, ready for further processing. Thus far, a top efficiency of 18.4% has been achieved (37.3 mA/cm2, 640 mV, FF 77.0 %) on full 156cm2 wafers with screen printed contacts and local Al-BSF, an encouraging proof of concept.

Illumination level independence in relation to emitter profiles on industrial high efficiency local Al-BSF cells

If the worlds answer to alternative energy production is to be silicon photovoltaics, the fabricated devices need to be robust and highly efficient in varying operating conditions. Rear oxide passivated local Al-BSF cells have a prevalent issue that hinders their performance in particular operating conditions, this issue being bias light dependence (reduced response at low illumination levels). It has been demonstrated by many [1,2,3] that to obtain high quantum efficiency at longer wavelengths, the cells need to be illuminated with a sufficient high level of bias light. If quantum efficiency is shown to be bias light dependant, at low light conditions the cell will simply underperform. Although most cells suffer from this type of degradation, in practice some high efficiency cells reach maximum spectral responsivity already at 0.3 suns and are considered to be bias independent. Regardless if in practice, cells can be named bias independent, the mechanisms for bias dependency is a very relevant characteristic of high efficiency solar cells of the present and future. In this paper we present a phenomenon that has been observed and repeated in separate experiments. We compare differences in rear passivation, specifically between a fresh deposited silicon oxide versus one that has been used also as a diffusion mask. We observe a relationship between bias dependence and the process flows as well as a relationship with the densification recipe. As expected the 90 ohm/sq emitter outperforms the higher doped emitters in the UV wavelength range. What is not expected is that when measuring without a bias light, the higher sheet resistance emitters outperform the lower sheet resistance emitters in wavelengths above 700nm. Another observation is that the cells that have been passivated with fresh silicon oxide indicate a large bias dependence at ultra low bias levels below .01 sun, but saturate performance above .05 sun. The diffused silicon oxide cells increase their quantum efficiency as the bias light is increased. The cells studied in this paper are fabricated using 148.25 cm2, 160μm p-type Cz-Si wafers with screen printed Ag front contacts and are rear-side passivated with a deposited rear silicon oxide/silicon nitride stack. The highest efficiency of the cells studied is 19.2 % with a Voc of 637mV, Jsc of 38.2 mA/cm2 and a fill factor of 79.1%

A process technology toolbox for next generation large area crystalline silicon solar cells

For further reduction of the crystalline Silicon solar cell cost/Wp, a dual approach is required: Further reduction of the silicon material by using thinner wafer and further increasing the conversion efficiency. Considering wafer thicknesses of 150µm and below the standard process with Ag screen-printed contacts on 50–60Ω/sq emitter and full Al BSF cannot provide the necessary efficiency increase. The reason for that is the increasing influence of the rear surface recombination current, which becomes a limited current loss mechanism. Within the Photovoltaic department in IMEC a research program has been launched with the goal of providing industrial processes for the next generation thin crystalline silicon solar cells. In this paper we are reporting on the development of a process toolbox that allows overcoming the full Al-BSF and the Ag-screen printing front-side metallization limitations. The next step towards higher efficiency targets is the implementation of novel emitter schemes and consequently advanced front-side metallization like electro-plating of copper for further photocurrent and fillfactor increase. By implementing Cu-plating as a front-side metallization, large area cells with efficiencies up to 18.4% have been fabricated. These are the initial steps for a cell concept that potentially can reach 19% efficiency in an industrial process flow. The progress towards an industrial Passivated Emitter and Rear Locally doped cell concept (i-PERL) is presented.

Direct laser doping for high-efficiency solar cells

Renewable energies is a prominent, emerging industry whose momentum suggests surging growth. The promise of a contact-less, precise, high throughput laser tool that enhances solar cell production will be vital in current and future photovoltaic manufacturing. Establishing lasers as an essential part of the manufacturing chain is the major goal of the European funded SOLASYS1 project (“Next Generation Solar Cell and Module Laser Processing Systems”).The selective emitter concept is one of the key subjects in the race to grid parity solar cell production. Instead of using an intermediate doping concentration to meet both the demand for high carrier lifetimes and for low contact resistance the doping concentration is increased locally at the positions of the front contacts. This can be achieved by laser doping the contact region before metal deposition. The selection of the proper laser parameters, such as wavelength, pulse length and laser power, allows control of the depth of the doping profile and the dopant concentration.Renewable energies is a prominent, emerging industry whose momentum suggests surging growth. The promise of a contact-less, precise, high throughput laser tool that enhances solar cell production will be vital in current and future photovoltaic manufacturing. Establishing lasers as an essential part of the manufacturing chain is the major goal of the European funded SOLASYS1 project (“Next Generation Solar Cell and Module Laser Processing Systems”).The selective emitter concept is one of the key subjects in the race to grid parity solar cell production. Instead of using an intermediate doping concentration to meet both the demand for high carrier lifetimes and for low contact resistance the doping concentration is increased locally at the positions of the front contacts. This can be achieved by laser doping the contact region before metal deposition. The selection of the proper laser parameters, such as wavelength, pulse length and laser power, allows control of the depth of the doping profile and the dop...