Schematic of (a) plan-view and (b) cross-sectional EBIC current collection geometries of a solar cell p-n junction.

Contexts in source publication

Context 1

... EBIC acquisition structures were used here: planview and cross-sectional. In Fig. 1a, the plane of the collecting p-n junction is perpendicular to the irradiated surface, which represents the plan-view; whereas in Fig. 1b the junction lies in the plane of the irradiated surface to obtain cross-sectional view. In a subset of specimens, FIB (focused ion beam) under the same instrument was used to mill the samples surface. ...

Context 2

... were then acquired at various accelerating voltages (4-16 kV) using the software package DISS5 supplied by Point Electronic. Two EBIC acquisition structures were used here: planview and cross-sectional. In Fig. 1a, the plane of the collecting p-n junction is perpendicular to the irradiated surface, which represents the plan-view; whereas in Fig. 1b the junction lies in the plane of the irradiated surface to obtain cross-sectional view. In a subset of specimens, FIB (focused ion beam) under the same instrument was used to mill the samples surface. The Zeiss NVision 40 is equipped with a gallium liquidmetal ion source used for milling at different current and accelerating voltages ...

Context 3

... 6 shows the relative collection efficiency between the two regions for accelerating voltages ranging from 4 kV to 16 kV. Histograms for all beam energies are given in the supplementary materials, Figure S1. It is clear that the relative efficiency depends on electron energy, with the ratio being as low as 0.49 at 4 kV, but 0.95 at 10 kV. ...

Context 4

... model 2, a laser-induced redistribution of phosphorus dopants was included besides setting the SRV to thermal velocity. The phosphorus profiles for both regions can be found in the supplementary materials, Figure S1. Fig. 8 shows that the J sc ratio for model 2 is lower and closer to the experimental data than that of model 1, which indicates that the increased Auger recombination at the surface leads to a significant decrease in collection efficiency. ...

Context 5

... front metal fingers were fabricated in the laboratory with a recently acquired screen-printer designed for mass production. The new screen printing enabled 25 µm wide silver fingers, similar to those reported in ( Beaucarne et al., 2019). The combined SE/ EBIC images of standard 35 µm fingers, and new 25 µm fingers in PERC cells are shown in Fig. 10a and b. The horizontal darker lines are artefacts induced by contact instability during the imaging experiment at Oxford. In both cells the laser-doped regions are evident from the loss in collection efficiency. In addition, there is a region with even lower EBIC signal vertically adjacent to the finger (<5 µm, labelled dark region in ...

Context 6

... in Fig. 10a and b. The horizontal darker lines are artefacts induced by contact instability during the imaging experiment at Oxford. In both cells the laser-doped regions are evident from the loss in collection efficiency. In addition, there is a region with even lower EBIC signal vertically adjacent to the finger (<5 µm, labelled dark region in Fig. 10). Such regions observed under an optical microscope does not appear to be residual metallisation paste. It therefore may be caused by the migration of contaminants, such as Ag ( Hilali et al., 2004) contained in the frits, during the process of firing or may be caused by infrared radiation during firing that is reflected off the finger ...

Context 7

... Ag ( Hilali et al., 2004) contained in the frits, during the process of firing or may be caused by infrared radiation during firing that is reflected off the finger ( Fields et al., 2016). This region with low collection efficiency is wider around the 25 µm finger than around the 35 µm one, which means a potential larger loss in cell performance. Fig. 10c shows two regions where the EBIC signal is considerably higher, within the usual laser-doped selective emitter. These regions most probably received insufficient laser doping because the laser head moved too fast between laser pulses, in the typical nanosecond switched lasers used in industry ( Lin et al., 2020). Such lower doping ...

Context 8

... ( Lin et al., 2020). Such lower doping concentration can lead to locally higher contact resistance with the metal contact, which does not affect cell efficiency because current flow across the contact interface is dominated by the regions with sufficient laser doping. However, such regions may cause recombination at the metal interface. In Fig. 10c, a ~100 µm long laser-doped region is surrounded by two ~15 µm long non-lasered gaps, leading to 13% of the area under the contact that would cause excess recombination. These findings are in agreement with previous works where pulsing of the laser was found to produce an intermittent pattern ( Bonse et al., 2013). However, experience ...

Context 9

... experience shows that laser pulses that overlap excessively are more detrimental than small gaps ( Brand et al., 2018). These micrometre scale features are only observed by EBIC imaging with submicron resolution, which allows linking them to specific laser processing. These findings help understanding and improving cell fabrication processes. Fig. 11 shows the structure of an n-type i-TOPCon cell fabricated by Trina Solar (Chen et al., , 2019. The p-n junction of the cell was fabricated by boron diffusion. It is well known that both boron diffusion and screen-printing are more difficult to attain than phosphorus diffusions. Mapping the p-n junction is therefore important to ensure ...

Context 10

... conducted using a cross sectional approach. 5 kV was chosen to overcome the noise due to electrical interference and avoid any potential shift in peak position. FIB milling using a Ga beam at 30 kV energy and 1 pA current, was carried out for 4 min to enhance the resolution of the junction imaging, and new EBIC images were acquired again at 5 kV. Fig. 12 shows combined SE/ EBIC local images of the junction, (a) before, and (b) after FIB milling. After FIB, the full width at half maximum of the EBIC signal is reduced from 1.05 µm to 0.53 µm, showing that FIB milling is a functional method of improving the EBIC resolution. In Fig. 12b, the average distance between the Ag-Si interface and ...

Context 11

... FIB, the full width at half maximum of the EBIC signal is reduced from 1.05 µm to 0.53 µm, showing that FIB milling is a functional method of improving the EBIC resolution. In Fig. 12b, the average distance between the Ag-Si interface and the maximum of the EBIC signal, i. e. the emitter junction depth, is 1.4 µm, with the junction deeper at the top of the pyramids than at the valleys, as is also found in the phosphorus emitter in Fig. 5. This also shows that screen-printing does not cause any spikes to the p-n junction, which was obtained only after a long series of paste and firing development. ...