contact for silicon solar cells SPREE Seminar Talk Udo Rmer - - PowerPoint PPT Presentation

Polycrystalline silicon as carrier selective contact for silicon solar cells

SPREE Seminar Talk Udo Römer 21.07.2016

Outline

Process optimisation poly-Si contacts Theory / understanding poly-Si contacts Local overcompensation via ion implantation Combination of these technologies

1

Motivation

• Removing front side metal shadowing

 ~2.5 mA/cm² gain in Jsc

• Voltage can be enhanced by reducing

recombination:

• SP-PERC cell has J0 of ~300 fA/cm²
• Poly-Si / c-Si-junction:

15 fA/cm² for p-type poly-Si1 and 20 fA/cm² for n-type poly-Si1  ~70 mV gain in Voc

ln 1

gen OC

J kT V q J        

PERC solar cell Poly-Si rear contact solar cell

[1] J. Y. Gan and R. M. Swanson, IEEE Trans. Electron Devices, vol. ED-37, pp.245-250 (1990)

2

Passivation:

• Low Dit at wafer surface due to high

quality oxide

• Field effect passivation due to

highly doped poly-Si layer Transport:

• Tunnel transport through the oxide1
• r
• holes in the oxide2

Poly-Si contact: working principle

Band diagram of a n-typ poly-Si contact (sketch)

[1] Steinkemper, Feldmann, Bivour, and Hermle, IEEE Journal of Photovoltaics, vol. 5, no. 5, pp. 1348 (2015) [2] Peibst, Römer, Hofmann, Lim, Wietler, Krügener, Harder, and Brendel, IEEE Journal of Photovoltaics, vol. 4, no. 3, pp. 841 (2014)

electron transport few interface defects interface oxide many defects interface oxide: larger tunnel barrier for holes than for electrons Depth [nm] Energy [eV]

3

Poly-Si contact: transport

RCA-oxide after 10 min at 950 °C RCA-oxide after 10 min at 1100 °C

• HR-TEM investigations show braking up of the oxide after high

temperature processing

4

Wolstenholme, Jorgensen, Ashburn, and Booker, J. Appl. Phys. 61, 225 (1987)

Process flow poly-Si test structures

1-step-process / 2-step-process Thin oxidation Deposition of doped poly-Si layers Annealing / cracking of the oxide

5

Thin oxidation Deposition of undoped poly-Si layers Annealing / cracking of the oxide Doping

Process flow poly-Si test structures

1-step-process / 2-step-process

5

Measurement of the recombination characteristics

• J0-determination with photoconductance measurement and Kane &

Swanson method1

Flash lamp Coil J0 = 5 fA/cm² 2.4 nm thermal oxide 30 min at 1050 °C J0-test structure

6

[1] Kane and Swanson, Proc.of the 18th IEEE PVSC, pp. 578–583 (1985)

If oxide is perfectly insulating: R4PP ~ Rpoly If oxide is perfectly conducting: R4PP ~ Rabs

Determination of the contact resistance

If oxide is perfectly insulating: R4PP ~ Rpoly

 rel ~ 1

If oxide is perfectly conducting: R4PP ~ Rabs

 rel ~ 0

Rpoly Rbulk Rpoly Rpoly Rbulk Rpoly Rpoly Rbulk Rpoly Rabs

4PP sheet resistance measurement Inductive sheet resistance measurement

7

Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Simulation of 4PP-measurements

Input parameters:

• Sample geometry
• Resistivities of wafer, poly-Si layer and oxide

Result:

• “Measured" sheet resistance

SENTAURUS-DEVICE 3D-simulation

8

Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Simulation of 4PP-measurements

• Calculation of relative contact

resistance from simulated 4PP resistance

• Plot vs. “real” (specified) contact

resistance

• Example: poly-Si layer with sheet

resistance of 280 Ω/□: rel < 0.05 corresponds to contact resistance of < 0.5 Ωcm²

9

Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Investigation of different oxides

• All oxides reach J0-values between 5 and 20 fA/cm²
• High temperatures needed for low contact resistance
• Boron-doped poly-Si contacts comparable to Phosphorus-doped

10

Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Investigation of different oxides

• Passivation stable up to 60 min annealing at 1050 °C
• Contact resistance decreases with increasing annealing duration
• Good combination of low J0 and rel possible

11

Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Influence of the metallization

• ILM measurements before and after metallization show comparable lifetime

level (apart from edge effects)

• Planar solar cell demonstrators show Voc of 714 mV1
• Series resistance of 0.6 Ωcm² not limited by poly-Si contact1

Lifetime distribution measured via Infrared Lifetime Mapping (ILM) Metallized pieces Solar cell demonstrator

12

[1] Römer, Peibst, Ohrdes, Lim, Krügener, Bugiel, Wietler, and Brendel, Solar Energy Materials and Solar Cells, vol. 131, pp. 85-91, Dec. 2014.

Process optimisation poly-Si contacts Theory / understanding poly-Si contacts Local overcompensation via ion implantation

13

Local counterdoping with in-situ patterned ion implantation

• Counterdoping: Overcompensating one polarity
• f dopants with dopants of the other polarity
• With in-situ patterned counterdoping local doping

possible without structured dielectric layers

• Counterdoping with in-situ masked ion

implantation enables elegant process flow for back contacted solar cells

• Process results in formation of lateral pn-junction

with heavily doped p- and n-regions

• Risk for band to band tunneling
• Risk for trap-assisted tunneling

14

Full area counterdoping

• Processing:
• 1.5 x 1015 cm-2 B implantation
• 3 x 1015 cm-2 P implantation
• Annealing at 1050 °C
• SIMS measurement:
• Phosphorus profile covers boron

profile over whole depth  Counterdoping works fine!

Römer, Peibst, Ohrdes, Larionova, Harder, Brendel, Grohe, Stichtenoth, Wütherich, et al., Proc. 39th ,IEEE PVSC, pp. 1280 (2013)

15

Local counterdoping

• Simulations featuring lateral pn-junction
• Variation of the lifetime in the “implanted” area
• Measurements on test structures and comparison with simulations

Simulations of the diode characteristics Measurements on test structures

p-type silicon Full area boron implantation Local phosphorus implantation Aluminum contacts Passivation layer Position x [µm] Position z [µm]

Römer, Peibst, Ohrdes, Larionova, Harder, Brendel, Grohe, Stichtenoth, Wütherich, et al., Proc. 39th ,IEEE PVSC, pp. 1280 (2013)

16

Local counterdoping

• Simulations show no detrimental influence of highly doped pn-junction:

n = 1 as long as implant damage is well annealed

• Otherwise strong recombination in space charge region with n = 2
• Measurements on test structures show n = 1  Everything is fine!

Simulations of the diode characteristics Measurements on test structures

Römer, Peibst, Ohrdes, Larionova, Harder, Brendel, Grohe, Stichtenoth, Wütherich, et al., Proc. 39th ,IEEE PVSC, pp. 1280 (2013)

17

Local counterdoping

• Further investigations (incl. influence of the

lateral doping profile & characteristics in reverse direction) published1

Everything is fine!

• Large area (156 mm x 156 mm psq.) ion

implanted IBC solar cells featuring local counterdoping reach efficiencies of 22.1 %2

[1] Römer, Peibst, Ohrdes, Larionova, Harder, Brendel, Grohe, Stichtenoth, et al., Proc. 39th ,IEEE PVSC, pp. 1280 (2013) [2] Bosch Solar Energy, ISFH, press release, Aug. 15th, 2013

18

Process optimisation poly-Si contacts Theory / understanding poly-Si contacts Local overcompensation via ion implantation Combination of these technologies

19

Process flow for ion implanted poly-Si with counterdoping

Thermal

• xidation

LPCVD poly-Si deposition Boron implantation Annealing/

• xide break-up

Boron implanted Test structure Phosphorus implanted test structure

20

Process flow for ion implanted poly-Si with counterdoping

Annealing/

• xide break-up

Thermal

• xidation

LPCVD poly-Si deposition Boron implantation Masked phosphorus implantation Boron implanted Test structure Phosphorus implanted test structure Test structure full area counterdoping

20

Process flow for ion implanted poly-Si with counterdoping

Annealing/

• xide break-up

Thermal

• xidation

LPCVD poly-Si deposition Boron implantation Masked phosphorus implantation Boron implanted Test structure Phosphorus implanted test structure Test structure local counterdoping Test structure full area counterdoping

20

Ion implantation in poly-Si

• Decreasing J0 with increasing dose
• Very low values of 1.1 fA/cm² (phosphorus) and 4.4 fA/cm² (boron)
• Increase at too high doses, especially for boron doping

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

21

Ion implantation in poly-Si

• Doping concentration constant inside poly-Si layer
• Oxide acts as diffusion barrier, in particular for phosphorus
• For high doses strong diffusion of boron into wafer

22

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

Recombination characteristics of boron-implanted test structure

• Implantation dose: 1x1015cm-2 B
• Highest Voc,impl. value reported so far for

p+ doped poly-Si junctions Best value F-ISE1: 694 mV

• High pFFimpl.-value of 84.6 %

(ideal value for the given Voc is 85%)

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015) [1] Feldmann, Müller, Reichel, and Hermle, Phys. Stat. Sol. RRL, pp. 1 (2014)

23

Recombination characteristics of phosphorus-implanted test structure

• Implantation dose: 5 x 1015 cm-2 P
• Very high Voc,impl. of 742 mV
• Due to J0,poly of only 1.1 fA/cm² and very

high bulk lifetime, recombination characteristics at MPP dominated by Auger recombination

nAuger ≈ 2/3 results in very high pFFimpl.

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

24

Counterdoping in poly-Si

• J0-values comparable to values without counterdoping
• Contact resistance of some samples very high
• For others comparable to samples without counterdoping

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

25

Counterdoping in poly-Si

• For some samples diffusion of boron into the wafer

 No contact between n+ poly-Si and n-type wafer

• High phosphorus doses prevent in-diffusion of boron

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

26

Recombination characteristics of counterdoped test structure

• Implantation doses: 1x1015cm-2 B

5x1015cm-2 P

• Very high Voc,impl.-value

Best value F-ISE: 682 mV1

• Despite J0,poly-value of 0.9 fA/cm²

pFFimpl.-value of “only” 84.7%

27

[1] C. Reichel, F. Feldmann, R. Müller, A. Moldovan, M. Hermle, and S. W. Glunz, 29th EUPVSEC (2014)

Recombination characteristics of masked counterdoped test structure

• Implantation doses: 1x1015cm-2 B

5x1015cm-2 P

• Curve only fitable by adding a further

recombination path with n>2

• Standard SRH-Theory: 1<n<2 in SCR

Non-standard behavior e.g. coupled

defects1

e h h h h h e e e e h h e e h h h e e e h h e

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015) [1] Steingrube, Breitenstein, Ramspeck, Glunz, Schenk, and Altermatt, Journal of Applied Physics, vol. 110, no. 1 (2011)

28

Recombination characteristics of masked counterdoped test structure

JL Jcell Jpara Rpara Rs

e h h h h h e e e e h h e e h h h e e e h h e

Römer, Peibst, Ohrdes, Lim, Krügener, Wietler, and Brendel, IEEE Journal of Photovoltaics, vol. 5, no. 2, pp. 507 (2015)

29

Reduction of pn-junction recombination

[1] Peibst, Römer, Patent application [2] Rienäcker, Merkle, Römer, Kohlenberg, Krügener, Brendel, and Peibst, 6th SiliconPV Conference 2016

• Lowering of poly-Si thickness
• Possibly problems with metallisation
• Adaption of implantation parameters
• Lower dose at pn-junction
• Not very helpful (see thesis)
• Removal of lateral pn-junction
• Oxidisation1
• Wet chemical etching

 η = 23.9 %2 Voc = 722 mV2  FF = 78.7 %2

e h h h h h e e e e h h e e h h h e e e h h e

JL Jcell Jpara Rpara Rs

30

Summary

Fast and non-destructing method for the estimation of the contact resistance between different layers Poly-Si contacts with J0-values of 0.66 fA/cm² and 4.4 fA/cm² for phosphorus and boron doping developed Full poly-Si contacted solar cell with Voc of 714 mV and low contact resistance fabricated

31

Summary

Local overcompensation via masked ion implantation investigated Overcompensated poly-Si contacts with J0-values of 0.9 fA/cm² fabricated Anomalous recombination behaivour in locally overcompensated poly-Si contacts investigated

32

Many thanks to

• You for your attention
• Many colleagues at ISFH and LUH for their help:

Tobias Wietler, Robby Peibst, Susanne Mau, Heike Kohlenberg, Miriam Berger, Tobias Ohrdes, Michael Häberle, Jan Krügener, Agnes Merkle, Bianca Lim, Yevgeniya Larionova, Sarah Spätlich, David Sylla, … …and everyone else for the always nice atmosphere

• The Laboratory for Nano- and Quantum Engineering Hanover
• The BMWi for funding

Back-up slides

Influence of process sequence

Solar cell results

• Best process used for production of proof of principle solar cells
• High implied Voc measured; implied PFF nearly ideal (85.0 %)
• impl. Voc = 716 mV
• impl. PFF = 84.5 %

full area wafer

Solar cell results

• impl. Voc = 716 mV
• impl. PFF = 84.5 %

full area wafer

• impl. Voc = 714 mV
• impl. PFF = 72.7 %

Voc = 705 mV PFF = 73.1 %

• Voc reduced further due to the edge effects (measured through 2 x 2 cm² mask)

 Measurement without mask yields Voc of 714 mV

• Large area solar cells would not suffer from this effect

laser-cut into a 2.5 x 2.5 cm² piece

Solar cell results

• Flat surface of the solar cell and rather thick poly-Si layer (> 200 nm thick)

 Low Jsc

• Edge recombination

 Low PFF

• Excellent passivation quality of the poly-Si layer

 high Voc

• Good transport through the poly-Si / c-Si junction

 low Rs

Area [cm²] Voc [mV] Voc [mV] (full area illumination) PFF [%] FF [%] Rs,FF [Ωcm²] Jsc [mA/cm²] η [%] 4.25 705 714 73.1 71.2 0.6 28.8 14.5