Yifeng Chen, Shengzhao Yuan, Yanfeng Cui, Zigang Wang, YunSheng, Chengfa Liu, Daming Chen, Pietro P. Altermatt, Zhiqiang Feng and Pierre J. Verlinden

Opportunity and Challenge: >21% large-area n-type PERT bifacial solar cells in research and production

BifiPV2016 workshop, Miyazaki

Sep 29th, 2016

2

Low cost

Reducing the Levelized Cost of Electricity

(LCOE)

High efficiency

High reliability

Good field performance

Key elements for a PV module

Module

3

First development of n-type bifacial in Lab

n-type Si base

ARC Front contacts

p+ emitter

n+ BSF

20

20.5

21

21.5

Eff

icie

ncy

[%]

Time

Avg EffMax Eff

Industrial feasible tools and process for stable mass production:

Tube boron diffusion

Tube phosphorus diffusion

PECVD double side passivation

Front & Back screen printing

4

J0,e 8% J0,e (metal)

13%

J0, bulk 52%

J0, bsf 15%

J0, BSF (metal)

12%

Power loss analysis

Fabricate reference sample to extract J0 in: Emitter

Auger, SRH, front surface (passivated), front contact

Base: Auger, SRH

BSF: Auger, SRH, rear surface (passivated), rear contact

J0 distribution of a 20.4% cell No. 1

No. 2

No. 3

5

Shallow defects in n-type material

n-type: lifetime may be influenced by shallow defects

• Thermal double donors (TDD). Defect levels from 50 – 69 meV below the conduction band edge, Ec, (neutral states), from 114 – 156 meV for singly ionized (positively charged) states. [1,2]

• Sometimes, thermal donors can form associated deep defect levels by reacting with impurities [3,4]

[1] D. Wruck, P. Gaworzewski, Phys. Stat. Sol. (a) 56, 557 (1979). [2] A.R. Bean, R.C. Newman, J. Phys. Chem. Solids 33, 255 (1972). [3] E. Simoen et al, J. Elchem. Soc. 150 , G520-G526 (2003). [4] M. Tomassini et al, J. Appl. Phys. 119, 084508 (2016).

6

Shallow defects in n-type material

n-type: lifetime may decrease with bias (with inj. level)

Higher injection Δn widens plateau, shallow defects become more

recombination active

Difference in lifetime between Vmpp and Voc is maximal at

Ed < 100 meV and at low Ndop.

7

Mechanism of FF changes

n-type: lifetime decreases with bias → increase of FF

Higher voltage Higher Δn (injection density)

inj.-dep. lifetime

increases FF W. Duan et al, 23nd EU PV Conf., paper 2CO.2.1 (to appear)

8

Efficiency improvements roadmap in Lab Improvements to avg. eff. 21.3% by: Base:

→Improving bulk lifetime

Emitter: →Optimization of surface passivation →Improving metallization →Boron diffusion thermal budget

BSF: →Selective BSF

n-type Si base

n++ BSF

n+ BSF

20

20.5

21

21.5

Eff

icie

ncy

[%]

Time

Avg Eff Max EffSelective BSF

Improving bulk lifetime

9

Demonstration of >21% efficiency in pilot line

Solar Cell Efficiency

Quality of Starting Material

Cleanliness of Facilities

Solar Cell Design Process

Operating Procedures

Stability of production equipments

Solar Cell design is not the only thing to improve

10

Demonstration of >21% efficiency in pilot line

Trina Solar，2016/9

Eff. [%]

Jsc [mA/cm2]

Voc [mV]

FF [%]

Average 21.23 39.48 668.4 80.74

Median 21.44 39.60 671.4 80.89

Best 21.74 39.81 673.1 81.49

We started pilot line from June 2016 Latest batch run in a pilot line in “Golden Line”, rear side efficiency 20.4-20.8%

20.6

20.8

21

21.2

21.4

21.6

21.8

0 50 100 150 200 250 300 350 400 450

Cel

l eff

icie

ncy

[%]

Cell number

11

First batch of double glass modules & reliability

Front side 280W 285W 290W 295W

Distribution 2% 22% 62% 14% Calibrated with reference modules tested in Fraunhofer CalLab.

Fabrication of the first batch 300pcs of modules Bifaciality of 93%

Rear contribution 10% 20% 30% Total power 317W 344W 370W

Reliability: LID < 1% TC300 < 2% Mechanical loading <2% PID <2% Hot spot defined by UL and IEC

DH+ML and TC+HF ongoing

Reliability: Dual Glass PV Module Design (DuoMax)

• Replacing the backsheet of standard modules with another glass panel significantly reduces the reliability risk in harsh climates

2.5mm front glass/2.5 back glass

Both heat-strengthened

Frameless

60 or 72 cells

Mono or Multi

White reflective EVA or transparent EVA

Designed for 1500V max system voltage (IEC 1500V & UL1000V certified)

900MW modules sold No.1 market share

13

108kW fish culture &PV system in Changzhou

Location: Menghe, Changzhou (31°59'11.4"N, 119°49'59.9"E)

Climate: subtropical Surroundings: over the lake System capacity:

36 kW multi 36 kW bifacial 36 kW bifacial

System starting: Oct, 2016 LCOE calculation Publication sharing

Trina farmland Lake

Same inverter Same location & Climate Same operators Same installation Same measurement system “Apple-to-Apple” comparison

14

Achievements:

Significant progress of n-type bifacial project in Trina Solar since 2016: Power loss analysis indicates that recombination in the bulk dominates

Improving of technologies yields avg. 21.2% efficiency and >670 mV Voc in pilot line, with best efficiency of 21.7%

Module power 285W-295W (front) demonstrated with promising reliability

Issues to be solved:

Standard! Standard! Standard! Black sheet is NOT that black

Reproducible, measureable, useful for system builders

Mounting designs still need optimization

LCOE should be demostrated

To enter the market, the rear side gain “X”% should be clear System level test

Conclusions

15

Issues of bifacial technologies

0

20

40

60

80

100

300 500 700 900 1100

Refle

ctan

ce (%

)

Wavelength (nm)

Black backplane (4%)

White Backplane (70%)

4%

4% 4%

Cell/module

4%

Cell/module

Measured Eff drop 0.2% Overestimation of Jsc!

Too many “21%” n-PERT cells with <665mV Voc

16

Achievements:

Significant progress of n-type bifacial project in Trina Solar since 2016: Power loss analysis indicates that recombination in the bulk dominates

Improving of technologies yields avg. 21.2% efficiency and >670 mV Voc in pilot line, with best efficiency of 21.7%

Module power 285W-295W (front) demonstrated with promising reliability

Issues to be solved:

Standard! Standard! Standard! Black sheet is NOT that black

Reproducible, measureable, useful for system builders

Module and mounting designs still need optimization

LCOE should be demonstrated

To enter the market, the rear side gain “X”% should be clarified System level test

“Apple-to-Apple” comparison

Conclusions

www.trinasolar.com

CHINA

SINGAPORE

JAPAN

CANADA

CHILE

U.S.A.

U.K.

GERMANY

FRANCE

ITALY

SPAIN SWITZERLAND

AUSTRALIA

yifeng.chen01@trinasolar.com