3.3 Evaluation of developed contact passivation stacks on textured surfaces

Given the excellent passivation quality from our developed electron-selective and hole-selective passivated contacts on planar Cz silicon wafers, it is then of research and commercial interest to evaluate the performance of these layers on textured surfaces as well, to determine its viability for deployment on a conventional silicon solar cell structure which adopts a front-side textured surface and either a rear-side planar or textured surface. To evaluate that, the lifetime test structures as shown in Figure 2 are utilized, featuring either symmetrical planar surfaces or symmetrical textured surfaces and symmetrically capped by either the electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup> )) or hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup> )) passivated contacts. The objective is to identify the suitability of our developed electron-selective and hole-selective passivated contacts for textured surfaces as well and to determine the optimum configuration for a silicon solar cell considering contact passivation for both the front and rear surfaces.

The highlight of this evaluation is plotted in Figure 8. Firstly, considering the influence of surface conditions on the passivation quality, it can be observed consistently from Figure 8 and summarized in Table 4 that both the electron-selective and hole-selective passivated contact stacks exhibited significantly better passivation quality on planar surfaces than on textured surfaces and which were consistent with the best results shown in Tables 2 and 3. Based on a batch average of 18 samples for each investigated lifetime test structure shown in Figure 8, the holeselective passivated contacts on symmetrical planar lifetime test structures demonstrated an effective minority carrier lifetime τeff of 1650 μs, a single-sided J0, rear of 27.5 fA cm<sup>2</sup> , and an implied-VOC of 689 mV, which is a significant improvement over the textured case (τeff of <sup>170</sup> <sup>μ</sup>s, single-sided <sup>J</sup>0, rear of 265 fA cm<sup>2</sup> , and implied-VOC of 628 mV). Effectively, upon deploying the hole-selective passivated contact on a textured surface, the τeff and implied-VOC reduce by 90 and 8.9%, respectively. Similarly, while the electron-selective passivated contacts continued to exhibit excellent passivation quality on planar surfaces (τeff of 6030 μs, singlesided J0, rear of 5.4 fA cm<sup>2</sup> , implied-VOC of 723 mV), the passivation quality

#### Figure 8.

selected results are highlighted in Table 3, which demonstrates the potential of our developed hole-selective contact passivation layers as well (i.e., thermal-SiOx/poly-

the as-deposited state and a further enhancement to 713 mV with single-sided J<sup>0</sup> values down to 4 fA /cm<sup>2</sup> after applying symmetrical SiNx capping layers. This can be attributed to the hydrogenation process which occurs spontaneously during the deposition of the SiNx capping layer, which helps to reduce the interface defect densities and directly improves the passivation quality [70]. Comparing our results to the excellent results from the Fraunhofer ISE team [69], which adopts PECVD of p-doped a-Si:H layers followed by sintering and SiNx capping (with a high implied-

optimization potential for our LPCVD of intrinsic silicon capping layer and the

Centrotherm

The thickness of the tunnel oxides/doped poly-Si layer/SiNx layer is 1.5/250/80 nm, respectively.

Comparison of the passivation quality of electron-selective passivated contacts on planar Cz n-Si symmetrical lifetime samples, both prior to and after the additional hydrogenation process step via the symmetrical addition

(mV)

) PECVD RF-MAiA 740 – [69]

) LPCVD Tempress 719 9 This work

)/SiNx LPCVD + MAiA 730 5 This work

)/SiNx LPCVD + MAiA 732 4 This work

(mV)

) LPCVD Tempress 698 37 This work

LPCVD +MAiA 713 8 This work

PECVD Centrotherm 732 1 [69]

LPCVD + MAiA 737 5 This work

) LPCVD Tempress 719 6 This work

) LPCVD Tempress 729 9 This work

VOC values up to 732 mV and single-sided J<sup>0</sup> values <1 fA cm<sup>2</sup>

associated boron diffusion optimization thereafter.

After a hydrogenation/anti-reflection coating step by SiNx

Tunnel layer/capping layer Method iVOC

After a hydrogenation/anti-reflection coating step by SiNx

)/

)/

) ALD Solaytec +

LPCVD

The thickness of the tunnel oxides/doped poly-Si layer/SiNx layer is 1.5/250/80 nm, respectively.

Comparison of the passivation quality of hole-selective passivated contacts on planar Cz n-Si symmetrical lifetime samples, both prior to and after the additional hydrogenation process step via the symmetrical addition

)/

Tunnel layer/capping layer Method iVOC

) PECVD

) stacks) with implied-VOC approaching 700 mV in

), we do identify

) References

Total J<sup>0</sup> (fA cm<sup>2</sup>

719 2 – FhG-ISE

Total J<sup>0</sup> (fA cm<sup>2</sup> )

697 26 This work

References

Si(p<sup>+</sup>

Silicon Materials

) or ALD-AlOx/poly-Si(p<sup>+</sup>

Wet-SiOx/poly-Si(n<sup>+</sup>

Wet-SiOx/poly-Si(n<sup>+</sup>

Wet-SiOx/poly-Si(n<sup>+</sup>

Wet-SiOx/poly-Si(n<sup>+</sup>

of the SiNx capping layers.

Thermal-SiOx/poly-Si(p<sup>+</sup>

Thermal-SiOx/poly-Si(p<sup>+</sup>

Thermal-SiOx/poly-Si(p<sup>+</sup>

of the SiNx capping layers.

SiNx

SiNx

Table 3.

102

ALD-AlOx/poly-Si(p<sup>+</sup>

SiNx

Table 2.

Ozone-SiOx/poly-Si(n<sup>+</sup>

Thermal-SiOx/poly-Si(n<sup>+</sup>

Ozone-SiOx/poly-Si(n<sup>+</sup>

Thermal-SiOx/poly-Si(n<sup>+</sup>

Comparison of the passivation quality (i.e. (a) effective carrier lifetime at 10<sup>15</sup> cm<sup>3</sup> injection level, (b) rear side J0 values and (c) implied-VOC values) by electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup> )) and hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup> )) passivated contacts on both symmetrical planar and symmetrical textured lifetime test structures. It can be observed that electron-selective passivated contacts are suitable for applications on both planar and textured surfaces (with implied-VOC > 720 mV and > 695 mV respectively), while the holeselective passivated contacts are only suitable for planar surfaces at the moment (with implied-VOC approaching 700 mV, compare Table 3).


Table 4.

Summary of the average measured passivation quality for both electron-selective and hole-selective passivated contacts deployed on both symmetrical planar and symmetrical textured silicon lifetime test structures.

reduces on textured surfaces as well (τeff of 1750 μs, single-sided J0, rear of 17 fA cm<sup>2</sup> , implied-VOC of 696 mV). Effectively, upon deploying the electronselective passivated contact on a textured surface, the τeff and implied-VOC reduce by 71 and 3.7%, respectively. Utilizing the same textured lifetime test structures, the electron-selective passivated contacts experience lower degradation of the passivation quality than the hole-selective passivated contacts by a factor of 2.4 times in terms of the implied-VOC values. The lower passivation quality measured on textured surfaces is not too surprising, given that similar observations were observed when evaluating silicon dioxide thin-film passivation on either planar or textured surfaces [71]. In particular, this reduced passivation quality can be attributed to (i) increased surface area (73% more surface area for textured [111] surfaces than planar [100] surfaces), (ii) increased density of dangling bonds at a [111] surface, and (iii) a higher concentration of interface defects, which could originate from the mechanical stress in the dielectric-silicon interfaces at creases, edges, or vertices [72]. Despite this inherent limitation, we demonstrate in this work that our electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup> )) passivated contacts have a great potential for being deployed on both planar and textured surfaces, while our hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup> )) passivated contacts are currently only suited on planar surfaces, based on our current developments.

Thus, if a double-sided contact passivation scheme is to be considered, the results in this work suggest that it is preferable to implement a solar cell structure with a textured front surface and a planar rear surface, and adopting the electronselective passivated contacts at the textured front surface and the hole-selective passivated contacts at the planar rear surface, as will be shown in the next section.

(symmetrically textured lifetime test structures with symmetrical electron-selective passivated contacts) was able to demonstrate an implied-VOC of 701 mV and a

Summary of the average measured passivation quality for both electron-selective and hole-selective passivated

B Sym.-planar (250 nm) Hole-selective 1883 39.1 696 C Sym.-textured (250 nm) Electron-selective 1943 28.5 701

Front electron-selective Rear hole-selective

Comparison of the passivation quality (i.e. (a) effective carrier lifetime at 10<sup>15</sup> cm<sup>3</sup> injection level, (b) rear side J0 values and (c) implied-VOC values) when both the electron-selective (thermal-SiOx /poly-Si(n<sup>+</sup>

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach…

passivated contacts and hole-selective passivated contacts (thermal-SiOx /poly-Si(p<sup>+</sup>

Structure Surface (poly-Si thickness) Pass. contact type τeff

(symmetrically planar), and lifetime test structure C (symmetrically textured).

solar cell structure A (front-side textured, rear-side planar silicon wafer), lifetime test structure B

structure A (the double-sided passivated contact solar cell precursors) to exhibit a measured passivation quality that lies between that exhibited by structure B and structure C. However, the actual measured results revealed that structure A

exhibited a poorer passivation quality than both structure B and structure C. Nonetheless, structure A was able to demonstrate quite high implied-VOC of 688 mV

layers, which likely cannot be attained by conventional diffusion of silicon solar

With a closer look at the key process steps, the key difference between the symmetrical lifetime test structures and the solar cell structures is that the former structures can be done in a one-step diffusion process, while the latter structures would require a series of dielectric masking to achieve single-sided diffused poly-Si layers with different polarities, starting from the higher-temperature requirement

sion process with a lower-temperature requirement (i.e., phosphorus diffusion

an increased near-surface recombination and poorer passivation quality, as evident

)). The goal is to reduce the drive-in/out-diffusion of boron

. At the first thoughts, we would expect

(μs)

))

iVOC (mV)

)) are deployed on both

Total J<sup>0</sup> (fA cm<sup>2</sup> )

1273 43.4 688

) in this work), followed by the diffu-

, prior to any anti-reflection/passivation

) layer into the silicon bulk which is expected to lead to

single-sided <sup>J</sup>0, rear value of 14 fA cm<sup>2</sup>

A Front-textured (250 nm)

DOI: http://dx.doi.org/10.5772/intechopen.85039

contacts deployed on different wafer surfaces.

Rear-planar (250 nm)

and a total <sup>J</sup><sup>0</sup> value of 43 fA cm<sup>2</sup>

first (i.e., boron diffusion toward poly-Si(p<sup>+</sup>

cell precursors.

Figure 9.

Table 5.

toward poly-Si(n<sup>+</sup>

105

dopants from the poly-Si(p<sup>+</sup>

## 3.4 Deployment of double-sided passivated contacts at the solar cell level

Based on the findings from the previous section, the deployment of double-sided passivated contacts at the solar cell level had been experimentally realized on n-type silicon wafers with a textured front surface and a planar rear surface and adopting an electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup> )) passivated contacts at the textured front surface and a hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup> )) passivated contacts at the planar rear surface. This is further compared to reference lifetime test structures with either symmetrical planar surfaces with symmetrical hole-selective passivated contacts or symmetrical textured surfaces with symmetrical electron-selective passivated contacts, as sketched in Figure 9. As shown in Figure 9 and summarized in Table 5, the lifetime test structures within this second batch of samples processed similarly to Figure 8 were able to consistently deliver excellent passivation qualities for the planar and textured lifetime test structures. In particular, Table 5 shows that structure B (symmetrically planar lifetime test structures with symmetrical holeselective passivated contacts) was able to again demonstrate an implied-VOC of 696 mV and a single-sided <sup>J</sup>0, rear value of 19.5 fA cm<sup>2</sup> , while structure C

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach… DOI: http://dx.doi.org/10.5772/intechopen.85039

#### Figure 9.

reduces on textured surfaces as well (τeff of 1750 μs, single-sided J0, rear of

Structure Surface Pass. contact type τeff (μs) J0, rear (fA cm<sup>2</sup>

A Planar Hole-selective 1649 27.5 689 B Textured Hole-selective 170 265 628 C Planar Electron-selective 6030 5.4 723 D Textured Electron-selective 1756 17.4 696

that our electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup>

suited on planar surfaces, based on our current developments.

hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup>

an electron-selective (thermal-SiOx/poly-Si(n<sup>+</sup>

front surface and a hole-selective (thermal-SiOx/poly-Si(p<sup>+</sup>

696 mV and a single-sided <sup>J</sup>0, rear value of 19.5 fA cm<sup>2</sup>

selective passivated contact on a textured surface, the τeff and implied-VOC reduce by 71 and 3.7%, respectively. Utilizing the same textured lifetime test structures, the electron-selective passivated contacts experience lower degradation of the passivation quality than the hole-selective passivated contacts by a factor of 2.4 times in terms of the implied-VOC values. The lower passivation quality measured on textured surfaces is not too surprising, given that similar observations were observed when evaluating silicon dioxide thin-film passivation on either planar or textured surfaces [71]. In particular, this reduced passivation quality can be attributed to (i) increased surface area (73% more surface area for textured [111] surfaces than planar [100] surfaces), (ii) increased density of dangling bonds at a [111] surface, and (iii) a higher concentration of interface defects, which could originate from the mechanical stress in the dielectric-silicon interfaces at creases, edges, or vertices [72]. Despite this inherent limitation, we demonstrate in this work

Summary of the average measured passivation quality for both electron-selective and hole-selective passivated contacts deployed on both symmetrical planar and symmetrical textured silicon lifetime test structures.

great potential for being deployed on both planar and textured surfaces, while our

Thus, if a double-sided contact passivation scheme is to be considered, the results in this work suggest that it is preferable to implement a solar cell structure with a textured front surface and a planar rear surface, and adopting the electronselective passivated contacts at the textured front surface and the hole-selective passivated contacts at the planar rear surface, as will be shown in the next section.

3.4 Deployment of double-sided passivated contacts at the solar cell level

Based on the findings from the previous section, the deployment of double-sided passivated contacts at the solar cell level had been experimentally realized on n-type silicon wafers with a textured front surface and a planar rear surface and adopting

the planar rear surface. This is further compared to reference lifetime test structures with either symmetrical planar surfaces with symmetrical hole-selective passivated contacts or symmetrical textured surfaces with symmetrical electron-selective passivated contacts, as sketched in Figure 9. As shown in Figure 9 and summarized in Table 5, the lifetime test structures within this second batch of samples processed similarly to Figure 8 were able to consistently deliver excellent passivation qualities for the planar and textured lifetime test structures. In particular, Table 5 shows that structure B (symmetrically planar lifetime test structures with symmetrical holeselective passivated contacts) was able to again demonstrate an implied-VOC of

, implied-VOC of 696 mV). Effectively, upon deploying the electron-

)) passivated contacts have a

) iVOC (mV)

)) passivated contacts at the textured

)) passivated contacts at

, while structure C

)) passivated contacts are currently only

17 fA cm<sup>2</sup>

Silicon Materials

Table 4.

104

Comparison of the passivation quality (i.e. (a) effective carrier lifetime at 10<sup>15</sup> cm<sup>3</sup> injection level, (b) rear side J0 values and (c) implied-VOC values) when both the electron-selective (thermal-SiOx /poly-Si(n<sup>+</sup> )) passivated contacts and hole-selective passivated contacts (thermal-SiOx /poly-Si(p<sup>+</sup> )) are deployed on both solar cell structure A (front-side textured, rear-side planar silicon wafer), lifetime test structure B (symmetrically planar), and lifetime test structure C (symmetrically textured).


#### Table 5.

Summary of the average measured passivation quality for both electron-selective and hole-selective passivated contacts deployed on different wafer surfaces.

(symmetrically textured lifetime test structures with symmetrical electron-selective passivated contacts) was able to demonstrate an implied-VOC of 701 mV and a single-sided <sup>J</sup>0, rear value of 14 fA cm<sup>2</sup> . At the first thoughts, we would expect structure A (the double-sided passivated contact solar cell precursors) to exhibit a measured passivation quality that lies between that exhibited by structure B and structure C. However, the actual measured results revealed that structure A exhibited a poorer passivation quality than both structure B and structure C. Nonetheless, structure A was able to demonstrate quite high implied-VOC of 688 mV and a total <sup>J</sup><sup>0</sup> value of 43 fA cm<sup>2</sup> , prior to any anti-reflection/passivation layers, which likely cannot be attained by conventional diffusion of silicon solar cell precursors.

With a closer look at the key process steps, the key difference between the symmetrical lifetime test structures and the solar cell structures is that the former structures can be done in a one-step diffusion process, while the latter structures would require a series of dielectric masking to achieve single-sided diffused poly-Si layers with different polarities, starting from the higher-temperature requirement first (i.e., boron diffusion toward poly-Si(p<sup>+</sup> ) in this work), followed by the diffusion process with a lower-temperature requirement (i.e., phosphorus diffusion toward poly-Si(n<sup>+</sup> )). The goal is to reduce the drive-in/out-diffusion of boron dopants from the poly-Si(p<sup>+</sup> ) layer into the silicon bulk which is expected to lead to an increased near-surface recombination and poorer passivation quality, as evident

from our measurements as well (see Figure 9). Figure 10 shows a comparison of the ECV profiles done on the same poly-Si(p<sup>+</sup> ) layer in the as-diffused state and after an additional diffusion masking and front-side phosphorus diffusion step. It can be clearly seen in the latter that the boron dopants have out-diffused from the poly-Si(p<sup>+</sup> ) capping layer into the silicon bulk, which is consistent with the reduced passivation quality measured on the solar cell precursors. Unfortunately, this issue is inevitable for our current investigated approach of obtaining the doped silicon capping layers, although the dopant out-diffusion could be better controlled via diffusion recipe optimization.

For a conventional silicon wafer solar cell, suitable dielectric thin films or stacks of thin films (such as SiOx, SiNx, AlOx) would be deposited on the silicon wafer surfaces to serve as anti-reflection/passivation prior to the metallization step. Similarly, in this work, the double-sided passivated contact solar cell precursors shown in Figure 9 were symmetrically capped with PECVD of 70-nm-thick SiNx films. The resulting passivation quality before and after additional SiNx capping is plotted in Figure 11 and listed in Table 6.

It can be seen from Figure 11 that upon the deposition of an additional symmetrical SiNx capping layer, there is a striking improvement in the pre-metallized solar cell precursors, in which the τeff/J0/iVOC values improve from 1.5 ms/48 fA cm<sup>2</sup> / 690 mV to 2.4 ms/16.5 fA cm<sup>2</sup> /713 mV. This improvement can be attributed to the hydrogenation effects from the overlying SiNx films, which is expected to further reduce the interface defect densities and improve its corresponding interface passivation quality, as evident from the measured lifetime results presented earlier. This observation was also consistently observed on the symmetrical lifetime test structures, in which the textured samples with electron-selective passivated contacts exhibited improvement in the τeff/J0/iVOC values from 1.9 ms/28.5 fA cm<sup>2</sup> /701 mV to 5.5 ms/13.3 fA cm<sup>2</sup> /731 mV, while the planar samples with hole-selective passivated contacts exhibited improvement in the τeff/J0/iVOC values from 1.9 ms/39 fA cm<sup>2</sup> /696 mV to 3 ms/30.8 fA cm<sup>2</sup> /710 mV.

3.5 Addressing the parasitic absorption issue for highly doped poly-Si layers

Summary of the measured passivation quality of a double-sided passivated contact solar cell precursor, before

Structure Surface (poly-Si thickness) Pass. contact type τeff

Measured passivation quality of the deployed double-sided passivated contacts on the solar cell structure sketched in Figure 10, both in the as-deposited state and after a symmetrical SiNx capping layer, was applied. With the symmetrical SiNx capping, which leads to the pre-metallized solar cell precursors, an excellent implied-VOC of 713 mV was obtained. The corresponding total J<sup>0</sup> values improved from 48 to 16.5 fA cm<sup>2</sup>

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach…

Front-textured (250 nm) Rear-planar (250 nm)

Front-textured (250 nm) Rear-planar (250 nm)

necessary.

107

Before SiNx

Figure 11.

After SiNx

and after additional SiNx capping.

approximately threefold improvement.

DOI: http://dx.doi.org/10.5772/intechopen.85039

Table 6.

Despite the excellent passivation qualities from the developed passivated contacts, one of the key challenges identified for device integration is the issue of parasitic absorption by these highly doped poly-Si capping layers. This issue is found to be more critical when the layers are deployed at the front surface than the rear surface, as simulation studies will show in the later sections. Hence, in order to address the parasitic absorption issue, a thinning of the doped poly-Si thickness is

Front electron-selective Rear hole-selective

Front electron-selective Rear hole-selective

(μs)

Total J<sup>o</sup> (fA cm<sup>2</sup> )

1500 48 690

2400 16.5 713

iVOC (mV) , an

Two different experimental approaches have been investigated: (1) applying a

optimized thick layers, and (2) performing a diffusion re-optimization for ultrathin LPCVD of intrinsic poly-Si layers. The goal is to determine the threshold (lowest

slow silicon etch-back technology, thereby thinning down our already well-

#### Figure 10.

Measured ECV profile for the poly-Si(p<sup>+</sup> ) region, comparing the as-diffused profile after the first rear-side boron diffusion (i.e., same compared to the lifetime test structure) and the final boron diffusion profile (i.e., after additional steps of masking, the second front-side phosphorus diffusion, and the chemical mask removal process). For the solar cell precursors, the additional high-temperature process step (second diffusion) causes out-diffusion of boron dopants from the poly-Si(p<sup>+</sup> ) layer into the silicon wafer bulk, as evident from ECV measurements.

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach… DOI: http://dx.doi.org/10.5772/intechopen.85039

#### Figure 11.

from our measurements as well (see Figure 9). Figure 10 shows a comparison of

after an additional diffusion masking and front-side phosphorus diffusion step. It can be clearly seen in the latter that the boron dopants have out-diffused from the

passivation quality measured on the solar cell precursors. Unfortunately, this issue is inevitable for our current investigated approach of obtaining the doped silicon capping layers, although the dopant out-diffusion could be better controlled via

For a conventional silicon wafer solar cell, suitable dielectric thin films or stacks of thin films (such as SiOx, SiNx, AlOx) would be deposited on the silicon wafer surfaces to serve as anti-reflection/passivation prior to the metallization step. Similarly, in this work, the double-sided passivated contact solar cell precursors shown in Figure 9 were symmetrically capped with PECVD of 70-nm-thick SiNx films. The resulting passivation quality before and after additional SiNx capping is plotted

It can be seen from Figure 11 that upon the deposition of an additional symmetrical SiNx capping layer, there is a striking improvement in the pre-metallized solar cell precursors, in which the τeff/J0/iVOC values improve from 1.5 ms/48 fA

to the hydrogenation effects from the overlying SiNx films, which is expected to further reduce the interface defect densities and improve its corresponding interface passivation quality, as evident from the measured lifetime results presented earlier. This observation was also consistently observed on the symmetrical lifetime test structures, in which the textured samples with electron-selective passivated contacts exhibited improvement in the τeff/J0/iVOC values from 1.9 ms/28.5 fA

hole-selective passivated contacts exhibited improvement in the τeff/J0/iVOC values

boron diffusion (i.e., same compared to the lifetime test structure) and the final boron diffusion profile (i.e., after additional steps of masking, the second front-side phosphorus diffusion, and the chemical mask removal process). For the solar cell precursors, the additional high-temperature process step (second diffusion) causes

/696 mV to 3 ms/30.8 fA cm<sup>2</sup>

) capping layer into the silicon bulk, which is consistent with the reduced

) layer in the as-diffused state and

/713 mV. This improvement can be attributed

/731 mV, while the planar samples with

) region, comparing the as-diffused profile after the first rear-side

) layer into the silicon wafer bulk, as evident from ECV

/710 mV.

the ECV profiles done on the same poly-Si(p<sup>+</sup>

diffusion recipe optimization.

in Figure 11 and listed in Table 6.

/ 690 mV to 2.4 ms/16.5 fA cm<sup>2</sup>

/701 mV to 5.5 ms/13.3 fA cm<sup>2</sup>

from 1.9 ms/39 fA cm<sup>2</sup>

poly-Si(p<sup>+</sup>

Silicon Materials

cm<sup>2</sup>

cm<sup>2</sup>

Figure 10.

measurements.

106

Measured ECV profile for the poly-Si(p<sup>+</sup>

out-diffusion of boron dopants from the poly-Si(p<sup>+</sup>

Measured passivation quality of the deployed double-sided passivated contacts on the solar cell structure sketched in Figure 10, both in the as-deposited state and after a symmetrical SiNx capping layer, was applied. With the symmetrical SiNx capping, which leads to the pre-metallized solar cell precursors, an excellent implied-VOC of 713 mV was obtained. The corresponding total J<sup>0</sup> values improved from 48 to 16.5 fA cm<sup>2</sup> , an approximately threefold improvement.


#### Table 6.

Summary of the measured passivation quality of a double-sided passivated contact solar cell precursor, before and after additional SiNx capping.

## 3.5 Addressing the parasitic absorption issue for highly doped poly-Si layers

Despite the excellent passivation qualities from the developed passivated contacts, one of the key challenges identified for device integration is the issue of parasitic absorption by these highly doped poly-Si capping layers. This issue is found to be more critical when the layers are deployed at the front surface than the rear surface, as simulation studies will show in the later sections. Hence, in order to address the parasitic absorption issue, a thinning of the doped poly-Si thickness is necessary.

Two different experimental approaches have been investigated: (1) applying a slow silicon etch-back technology, thereby thinning down our already welloptimized thick layers, and (2) performing a diffusion re-optimization for ultrathin LPCVD of intrinsic poly-Si layers. The goal is to determine the threshold (lowest

thickness) of the poly-Si films necessary to achieve the same excellent passivation quality as the thicker counterparts while reducing the parasitic absorption issue as much as possible.

intrinsic LPCVD of poly-Si films was deposited on both symmetrical lifetime test structures (textured and planar) and solar cell precursors (i.e., front-side textured, rear-side planar surfaces), followed by the phosphorus diffusion optimization process as mentioned above. The best results from the optimization process are

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach…

Comparing these results to the thick (250 nm) thermal-SiOx/poly-Si(n<sup>+</sup>

contacts on similar lifetime test structures (textured and planar) also exhibited excellent passivation qualities, attaining an implied-VOC of 703 and 727 mV for the textured and planar case, respectively, after a symmetrical SiNx capping step. Excellent film and diffusion uniformity was observed from the photoluminescence images; an example is shown in Figure 13(c) for the solar cell precursor structure (i.e., front-side textured, rear-side planar) with an electron-selective passivated contact being deposited on both sides (no SiNx capping). However, it was observed that the absolute implied-VOC values are slightly lower (few millivolts) than the

Excellent passivation quality demonstrated from our in-house developed electron-selective (thermal-SiOx/poly-

capped?

A Sym. planar (10 nm) No 4229 16 719 A Sym. planar (10 nm) Yes 7277 10 727 B Sym. textured (10 nm) No 961 67 686 B Sym. textured (10 nm) Yes 1928 31 703

Summary of the measured passivation quality parameters (τeff, total J0, implied-VOC) for an electron-selective passivated contact comprising an in situ thermal-SiOx tunnel layer coupled with a thin (10 nm) poly-Si(n<sup>+</sup>

capping layer, evaluated on lifetime test structures which are symmetrically planar (structure A), symmetrically textured (structure B), and front-side textured and rear-side planar solar cell precursors

τeff (μs)

textured lifetime test structures, with iVOC reaching 686 mV, and (b) symmetrical planar lifetime test structures, with iVOC reaching 720 mV, which further improves to 703 and 727 mV, respectively, after an additional standard SiNx capping layer. Good film and doping uniformity can be observed from the PL images for both the symmetrical lifetime test structures and solar cell precursors (i.e., front-side textured, rear-side

) capping layers applied on both (a) symmetrical

Total J<sup>0</sup> (fA cm<sup>2</sup> )

No 2982 22 713

Yes 6557 Inj. dep 741

iVOC (mV)

)

)) passivated contact with thin (10 nm) poly-Si(n<sup>+</sup>

Structure Surface (poly-Si thickness) SiNx

(10 nm)

(10 nm)

C Asym. front txt., rear planar

C Asym. front txt, rear planar

sivated contacts (Table 6), the thin (10 nm) thermal-SiOx/poly-Si(n<sup>+</sup>

) pas-

) passivated

highlighted in Figure 13 and Table 7.

DOI: http://dx.doi.org/10.5772/intechopen.85039

thicker counterparts.

Figure 13.

planar) as shown in (c).

Si(n+

Table 7.

(structure C).

109

Using our slow silicon etch (SSE) solution (DIW:KOH (3.5%):NaOCL (63.25%) at 80°C), an etch rate of 0.1 nm/s was determined, which was consistently observed for both poly-Si(n<sup>+</sup> ) and poly-Si(p<sup>+</sup> ) capping layers. Figure 12 highlights the influence of the resulting doped poly-Si capping layer thickness on the measured passivation quality.

Interestingly, for hole-selective passivated contacts, the passivation quality can be preserved for a poly-Si(p<sup>+</sup> ) thickness from a thick 250 nm down to ultrathin layers of approximately 3 nm, with measured τeff of 1.5 ms and implied-VOC of 690 mV, respectively. This suggests that a simple SSE etch could be an effective approach to reduce the poly-Si(p<sup>+</sup> ) capping layer thickness to an ultrathin (i.e., some nm ony) level. However, for electron-selective passivated contacts, the passivation quality was preserved only from 250 nm down to 70 nm, with measured τeff > 6 ms and implied-VOC > 720 mV, respectively. A further thickness reduction (<70 nm) leads to a severe degradation of passivation quality. As an example, upon reduction of the poly-Si(n<sup>+</sup> ) layer from 69 nm to 47 nm, the measured τeff and implied-VOC reduce by 86 and 6.5%, respectively. Hence, considering the preference to deploy electron-selective passivated contacts (thermal-SiOx/poly-Si(n<sup>+</sup> )) on the textured surface, we have to investigate alternative approaches (as outlined in the following) to obtain ultrathin poly-Si(n<sup>+</sup> ) capping layers suitable for device integration at the front textured surface of a double-sided passivated contact solar cell.

One of the alternative approaches to obtain ultrathin poly-Si(n<sup>+</sup> ) layers is to directly deposit an ultrathin intrinsic poly-Si capping layer, followed by a further optimization of the phosphorus diffusion conditions. The goal is to obtain a highly doped thin poly-Si(n<sup>+</sup> ) capping layer which can achieve excellent passivation quality similar to the thicker poly-Si(n<sup>+</sup> ) counterparts while minimizing the in-diffusion of phosphorus dopants into the silicon bulk. To achieve this, ultrathin (10 nm)

#### Figure 12.

Influence of the decreasing doped poly-Si capping layer thickness via the slow silicon etch process on the measured (a) minority carrier lifetime τeff and (b) implied-VOC values for symmetrically planar lifetime test structures. Promising results are observed on hole-selective passivated contacts, in which the passivation quality is preserved for a poly-Si(p<sup>+</sup> ) capping layer thickness reduction from a thick 250 nm down to a thin 3 nm. In contrast, the passivation quality of the electron-selective passivated contacts with poly-Si(n<sup>+</sup> ) capping layer was preserved down to a thickness of 70 nm, beyond which there is a drastic drop in passivation quality. The "star" symbol refers to the case where there is no doped poly-Si capping layer (i.e., only the tunnel oxide SiOx layer).

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach… DOI: http://dx.doi.org/10.5772/intechopen.85039

intrinsic LPCVD of poly-Si films was deposited on both symmetrical lifetime test structures (textured and planar) and solar cell precursors (i.e., front-side textured, rear-side planar surfaces), followed by the phosphorus diffusion optimization process as mentioned above. The best results from the optimization process are highlighted in Figure 13 and Table 7.

Comparing these results to the thick (250 nm) thermal-SiOx/poly-Si(n<sup>+</sup> ) passivated contacts (Table 6), the thin (10 nm) thermal-SiOx/poly-Si(n<sup>+</sup> ) passivated contacts on similar lifetime test structures (textured and planar) also exhibited excellent passivation qualities, attaining an implied-VOC of 703 and 727 mV for the textured and planar case, respectively, after a symmetrical SiNx capping step. Excellent film and diffusion uniformity was observed from the photoluminescence images; an example is shown in Figure 13(c) for the solar cell precursor structure (i.e., front-side textured, rear-side planar) with an electron-selective passivated contact being deposited on both sides (no SiNx capping). However, it was observed that the absolute implied-VOC values are slightly lower (few millivolts) than the thicker counterparts.

#### Figure 13.

thickness) of the poly-Si films necessary to achieve the same excellent passivation quality as the thicker counterparts while reducing the parasitic absorption issue as

at 80°C), an etch rate of 0.1 nm/s was determined, which was consistently

the influence of the resulting doped poly-Si capping layer thickness on the mea-

) and poly-Si(p<sup>+</sup>

Using our slow silicon etch (SSE) solution (DIW:KOH (3.5%):NaOCL (63.25%)

Interestingly, for hole-selective passivated contacts, the passivation quality can

layers of approximately 3 nm, with measured τeff of 1.5 ms and implied-VOC of 690 mV, respectively. This suggests that a simple SSE etch could be an effective

some nm ony) level. However, for electron-selective passivated contacts, the pas-

measured τeff > 6 ms and implied-VOC > 720 mV, respectively. A further thickness reduction (<70 nm) leads to a severe degradation of passivation quality. As an

alternative approaches (as outlined in the following) to obtain ultrathin poly-Si(n<sup>+</sup>

directly deposit an ultrathin intrinsic poly-Si capping layer, followed by a further optimization of the phosphorus diffusion conditions. The goal is to obtain a highly

of phosphorus dopants into the silicon bulk. To achieve this, ultrathin (10 nm)

Influence of the decreasing doped poly-Si capping layer thickness via the slow silicon etch process on the measured (a) minority carrier lifetime τeff and (b) implied-VOC values for symmetrically planar lifetime test structures. Promising results are observed on hole-selective passivated contacts, in which the passivation quality

was preserved down to a thickness of 70 nm, beyond which there is a drastic drop in passivation quality. The "star" symbol refers to the case where there is no doped poly-Si capping layer (i.e., only the tunnel oxide SiOx

In contrast, the passivation quality of the electron-selective passivated contacts with poly-Si(n<sup>+</sup>

capping layers suitable for device integration at the front textured surface of a

sivation quality was preserved only from 250 nm down to 70 nm, with

sured τeff and implied-VOC reduce by 86 and 6.5%, respectively. Hence, considering the preference to deploy electron-selective passivated contacts

One of the alternative approaches to obtain ultrathin poly-Si(n<sup>+</sup>

) capping layers. Figure 12 highlights

) thickness from a thick 250 nm down to ultrathin

) capping layer thickness to an ultrathin (i.e.,

)) on the textured surface, we have to investigate

) capping layer which can achieve excellent passivation qual-

) capping layer thickness reduction from a thick 250 nm down to a thin 3 nm.

) counterparts while minimizing the in-diffusion

) layer from 69 nm to 47 nm, the mea-

)

) layers is to

) capping layer

much as possible.

Silicon Materials

observed for both poly-Si(n<sup>+</sup>

be preserved for a poly-Si(p<sup>+</sup>

approach to reduce the poly-Si(p<sup>+</sup>

example, upon reduction of the poly-Si(n<sup>+</sup>

double-sided passivated contact solar cell.

ity similar to the thicker poly-Si(n<sup>+</sup>

sured passivation quality.

(thermal-SiOx/poly-Si(n<sup>+</sup>

doped thin poly-Si(n<sup>+</sup>

Figure 12.

layer).

108

is preserved for a poly-Si(p<sup>+</sup>

Excellent passivation quality demonstrated from our in-house developed electron-selective (thermal-SiOx/poly-Si(n+ )) passivated contact with thin (10 nm) poly-Si(n<sup>+</sup> ) capping layers applied on both (a) symmetrical textured lifetime test structures, with iVOC reaching 686 mV, and (b) symmetrical planar lifetime test structures, with iVOC reaching 720 mV, which further improves to 703 and 727 mV, respectively, after an additional standard SiNx capping layer. Good film and doping uniformity can be observed from the PL images for both the symmetrical lifetime test structures and solar cell precursors (i.e., front-side textured, rear-side planar) as shown in (c).


#### Table 7.

Summary of the measured passivation quality parameters (τeff, total J0, implied-VOC) for an electron-selective passivated contact comprising an in situ thermal-SiOx tunnel layer coupled with a thin (10 nm) poly-Si(n<sup>+</sup> ) capping layer, evaluated on lifetime test structures which are symmetrically planar (structure A), symmetrically textured (structure B), and front-side textured and rear-side planar solar cell precursors (structure C).

To provide more insights, ECV measurements were performed on the thin poly-Si(n<sup>+</sup> ) layers on both the textured and planar surfaces and compared to the thick reference as shown in Figure 14. The following observations can be made: (i) the thin poly-Si(n<sup>+</sup> ) layer exhibits a higher phosphorus dopant concentration (<sup>5</sup> <sup>10</sup><sup>20</sup> cm<sup>3</sup> ) than the thicker counterpart (<sup>2</sup> <sup>10</sup><sup>20</sup> cm<sup>3</sup> ); and (ii) the poly-Si(n<sup>+</sup> ) layer on the textured surface exhibits a higher dopants in-diffusion than the planar surface, which could partially explain the lower measured implied-VOC values for the former (i.e., 686 mV as compared to 719 mV).

concepts, which in this work had been entirely realized on commercially available industrial tools. However, as mentioned earlier, one of the key issues if contact passivation is to be applied front-side also is to minimize parasitic absorption within the highly doped front-side poly-Si capping layer. Highly doped poly-Si is similar to transparent conductive oxide (TCO) layers deployed for silicon

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach…

heterojunction solar cells, non-zero extinction coefficients, resulting in the inevitable parasitic absorption. This is even more pronounced, if applied front-side, thereby directly reducing the absorbable photogeneration current in the silicon wafer bulk (as the incident light is then first entering the parasitically absorbing

Hence, the objective of this section is to utilize an appropriate numerical calculation method to determine the parasitic absorption as a function of the (rear or

front side) poly-Si capping layer thickness and then subsequently predict the

absorption) for the investigated solar cell precursors in this work.

ical composition.

Si(n<sup>+</sup>

111

poly-Si(n<sup>+</sup>

41.1 mA cm<sup>2</sup>

.

corresponding solar cell efficiency potential of the correspondingly optimized passivated contact, being rear-side-only or front- and rear-side deployed in a solar cell. To address the above, the simulation program SunSolve™, available on PV Lighthouse [53], was utilized to study the impact of the doped poly-Si capping layer thickness on the maximum absorbable current density within the silicon wafer bulk Jabsorbed, cell. Besides calculating Jabsorbed, cell, the various optical losses can also be determined (i.e., front-reflected, front-escaped, rear-escaped, parasitic absorption in each layer, edge

To enhance the accuracy of the optical calculations, ellipsometry measurements were performed on all in-house fabricated samples, i.e., measuring our deployed dielectric films (SiNx, SiOx, AlOx) as well as our optimized doped poly-Si capping layers, followed by a fitting and extraction of the wavelength-dependent optical refractive indices (n, k). These wavelength-dependent refractive indices were then imported into the SunSolve™ simulation program for a more realistic prediction of the current loss analysis, based on our own developed contact passivation layers. As an example, Figure 15 shows the fitted wavelength-dependent refractive indices for the doped poly-Si layers in this work, which is further compared to the crystalline silicon reference [73]. As seen, the doped poly-Si layers do exhibit a higher extinction coefficient (k) compared to a c-Si reference within the visible to nearinfrared region (400–900 nm). This again indicates that parasitic absorption is inevitable and should be minimized by thickness reduction while not compromising on the passivation quality. Further optimization work should also try to reduce the extinction coefficient of the poly-Si capping layers itself, i.e., by changing its chem-

The current loss analysis results for a rear-side passivated contact solar cell (using SunSolve™) are shown in Figure 16. In order to account for internal back reflection, a local full-area metal contact scheme has been assumed (see Figure 16) (this can be realized by local laser ablation, forming contact openings in the SiNx

It can be seen that for a solar cell with a conventional front-side boron-diffused junction and a rear-side electron-selective passivated contact (thermal-SiOx/poly-

can be directly addressed by reducing the rear-side poly-Si film thickness (e.g., the parasitic absorption current loss reduces from 0.55 mA cm<sup>2</sup> for a 250-nm-thick

reduction in the parasitic absorption directly enhances the potentially absorbable current in the wafer bulk (Jabsorbed, cell), which in this case improves from 40.7 to

Interestingly, it was observed that for a poly-Si capping layer thickness lower

) capping layer

) layer). The

. On hindsight, we would

)), the parasitic absorption arising from the rear-side poly-Si(n<sup>+</sup>

) layer to 0.02 mA cm<sup>2</sup> for a 10-nm-thick poly-Si(n<sup>+</sup>

passivation layer and a subsequent full-area metallization).

than 25 nm, the <sup>J</sup>absorbed, cell saturates at 41.1 mA cm<sup>2</sup>

poly-Si capping layer before entering the silicon wafer).

DOI: http://dx.doi.org/10.5772/intechopen.85039

Similar to the thick poly-Si(n<sup>+</sup> ) capped samples, an additional symmetrical SiNx capping further enhances the overall passivation quality, such that the textured and planar lifetime structures now exhibit an improvement in the implied-VOC by 17 and 8 mV, which is a relative improvement of 2.5 and 1.1%, respectively.

To summarize, we have demonstrated on a textured silicon surface the ability to obtain an excellently passivating SiNx-capped electron-selective passivated contact (thermal-SiOx/poly-Si(n<sup>+</sup> )) with sufficiently thin poly-Si(n<sup>+</sup> ) thickness (10 nm) to reduce the parasitic absorption issue while maintaining excellent passivation qualities (implied-VOC values exceeding 700 mV). Re-optimizing the diffusion recipe for an ultrathin LPCVD of intrinsic poly-Si layer therefore solves the limitations encountered when using slow silicon etch technology, which limited the obtainable poly-Si(n<sup>+</sup> ) layer thickness to 70 nm (i.e., observing a drastic drop in passivation quality for thinner layers).
