Si Nanowires for All-Si Tandem Solar Cells

Name: Rufi Kurstjens

• Prof. dr. ir. Robert Pierre Mertens (promotor)
• Prof. dr. ir. Jozef Poortmans (mede-promotor)

Single junction Si solar cells have a maximum theoretical efficiency of31% under 1 sun illumination. This is generally referred to as the Shockley-Queisser limit. A significant part of the losses of these devices are related to the thermalization of hot electrons generated by photons with an energy higher than the forbidden bandgap. A proven way of reducing these thermalization losses consists of stacking materials with different bandgaps on top of each other in a multi-junction device. III-V materials are a convenient choice of material for this stack since the bandgap can be tuned in a straightforward manner. Finding an all-Si based alternative would however allow significant cost advantages. Earlier attempts in literature to make all-Si based tandem solar cells focused on Si quantum dots. However, the level of control required to optimize both thequantum confinement in the quantum dots and the tunneling between quantum dots has so far proven to be beyond the current technological capabilities.

Towards the goal of an all-Si based tandem solar cell, this study investigates the potential use of Si nanowires as a higher bandgap crystalline material to form the top cell. The innovative device concept relies on quantum confinement in two directions to tune the bandgap while maintaining the third direction for current collection. This alleviates the problems related to tunneling between quantum dots. The scope of this study is centered on answering 2 main questions: is it possible to fabricate the necessary type of nanowires, and can the nanowires serve as active material for a top cell?

A proof of concept process flow is developed based on a combination of DUV lithography, dry etching and oxidation to obtain the right size of nanowires. This process results in nanowires that taper slightly from a diameter of 8-13 nm at the foot of the nanowires to 2.5-4.5 nm at the top of the nanowires. A dense array of nanowires is created with a pitch of 90 nm between the nanowires and a nanowire length of 500 nm. Electrical contacts to the nanowires are obtained by a combination of PECVD oxide filling of the gaps between the nanowires and chemical-mechanical polishing of the overfill and thermal oxide.

The quality of the Si/SiO2 interface is investigated byESR. The density of dangling bonds along the vertical interface of the nanowires themselves were measured as well as their passivation kinetics. The effective interface defect density [Pb0NW] = 3.3 x 1012 cm-2, witha passivation activation energy Ea of 1.53 eV and a standard deviation s>Ea of 0.23 +/- 0.02 eV. This high value of s>Ea is indicative of astrained interface. Considering the values for the passivation kinetics, the prognosis is that [Pb0] can only be decreased ~ 25 times by hydrogen or forming gas annealing. This means that the nanowires can't be passivated better than [Pb0] = 1.3 x 1011 cm-2.

Optical pump THz probe spectroscopy shows that this high density of interface traps cause almost immediate trapping (picosecond time-scale) and recombination of photogenerated carriers. An optical time-of-flight technique on the other hand, measured the time it took carriers to travel through the nanowires. From the transit time measured by this technique, the mobility is extracted. The mobility is 0.06 +/- 0.009 cm2/Vs. This much lower value is representative of the impact of all the dangling bonds on the mobility.

From these material properties it seems highly unlikely that the nanowires, as currently produced by a top-down approach and size-reduction by oxidation, can serve as active material for a top cell. In the devicemeasurements performed on the nanowires their role could not clearly bedistinguished. In the I-V curves they could only be distinguished as anadditional series resistance. They did not seem to form part of the active layer.

https://lirias.kuleuven.be/handle/123456789/364403

•   Prof. dr. ir. Robert Pierre Mertens (promotor)
•   Prof. dr. ir. Jozef Poortmans (mede-promotor)
•   Prof. dr. Carlo Vandecasteele (voorzitter/chairman)
•   Prof. dr. ir. Johan Driesen (secretaris/secretary)
•   Prof. dr. Andre Stesmans
•   Prof. dr. ir. Guido Groeseneken
•   Marc Burgelman , UGent
•   Prof. dr. Abdelilah Slaoui , University of Strasbourg - CNRS