HETEROJUNCTION

TECHNOLOGY
THE SOLAR CELL
OF THE FUTURE
G. Roters1, J. Krause1, S. Leu2, A. Richter2, B. Strahm3

1
Roth & Rau AG, An der Baumschule 6-8, 09337 Hohenstein-Ernstthal, Germany
Meyer Burger Technology Ltd, Schorenstrasse 39, CH-3645 Gwatt (Thun), Switzerland
HETEROJUNCTION TECHNOLOGY
2

3
Roth & Rau Research AG, Rouges-Terres 61, CH-2068 Hauterive, Switzerland

Standard process Sel. Emitter process MWT process MB-PERC HJT process
ABSTRACT
Texture Texture Laser Drilling Texture Texture
Wafer-based silicon photovoltaic (PV) production has only changed slightly in the last forty years. The
standard concept comprises p-type silicon wafers, fired contacts and encapsulation. Cost reduction
Doping / Diffusion Doping / Diffusion Texture Doping / Diffusion a-Si Front / Rear Side
is necessary if PV is to survive without feed-in tariffs and be competitive with grid electricity costs.
Therefore levelised cost of electricity (LCOE) is one of the primary metrics for the cost of electricity Edge Isolation
Additional
Doping / Diffusion Edge Isolation
TCO /
4 Process Steps Metal Rear Contact
produced by both utility scale and distributed power systems. The fastest path to lower LCOE is to
introduce high efficiency solar cell concepts like the heterojunction technology (HJT). Photovoltaic PSG Etch Edge Isolation PSG Etch PSG Etch Print Front Side
systems using heterojunction technology (HJT) modules outperform any other PV systems and this
SiNx Capping- & AlOx
AR Coating PSG Etch AR Coating Curing
paper will explain why. passivation-layer

Print Front Grid

Print Rear Side AR Coating AR Coating Test & Sort
and p-Contact

Laser contact
Print Front Side Print Front Side Firing
Introduction opening

Print Front- & Rear-

Meyer Burger develops solar technology – from wafers to com- Recently Panasonic demonstrated efficiencies Firing Print Rear Side Edge Isolation (Front)
side Significant
plete PV systems – with the aim of promoting the widespread use of 24.7% on laboratory scale cells [1]. With the potential for
cost savings!
of photovoltaics and making solar power a first-choice source of discontinuation of the basic technology patent, Test & Sort Firing Edge Isolation (Rear) Firing

renewable energy. To this end, the company focuses strongly on heterojunction technology was opened to the
Test & Sort Test & Sort Test & Sort
developing solar cell technologies which allows highest efficiencies public in 2010. Meyer Burger is now offering this
and highest energy output at lowest production costs. appealing technology as a high performance key
technology in the photovoltaic value chain.
Heterojunction cell technology combines the advantages of mono
crystalline silicon (c-Si) solar cells with the good absorption and The simple structure of a HJT cell is shown in
CZ: 20 – 21% CZ n-type:
the superior passivation characteristics of amorphous silicon (a-Si) Figure 1. The thin intrinsic a-Si:H layers deposited CZ: 20 – 21% MC: 19 – 20% 21 ~ 25%
known from a-Si thin film technology using readily available ma- between c-Si wafer and doped layers are key to MC: 18 – 19%
CZ: 19 – 20%
terials. The HJT design is not new. Sanyo (now Panasonic) first achieving maximum performance from the cell MC: 18 – 19% Figure 2:
pushed this technology into mass production achieving around structure. They result in reduced interface state CZ: 19% Process sequence of differing PV technologies
20% cell efficiency. density, decreased surface recombination losses MC: 16.8 – 17.5%

and lower the emitter saturation currents.

Thin wafers
The relatively straightforward HJT production process takes place Today, wafers with a thickness of ~180 µm are in
at low temperatures and requires fewer production steps compared standard use in cell manufacturing. A wafer thick-
to other high efficiency designs as shown in Figure 2, which is eco- ness even below 100 µm is enough for absorb-
ITO Texture nomically attractive as it results in significant energy cost savings. ing light in silicon solar cells. Using thinner wafers
Ag (screen print) leads to significantly lower material costs because
a-Si Front / Rear Side An important technological advantage of HJT cells is the excellent a higher number of wafers can be cut from one
surface passivation of a-Si:H which results in high open-circuit volt- brick. With decreasing wafer thickness, the influ-
TCO /
(p) a-Si:H Metal Rear Contact ages and high cell efficiencies. The superior temperature coefficient ence of the bulk material quality also decreases
of TC = -0.20%/K ensures higher energy yield during module op- resulting in less tight requirements for e.g. minor-
(i) a-Si:H Print Front Side erating conditions. As previously mentioned, low temperature pro- ity charge carrier lifetimes. Diamond wire sawing
n-type c-Si
(i) a-Si:H cessing (<250°C) saves energy during manufacturing, prevents bulk is the best approach for thin wafers. The use of
Curing degradation and enables the use of thin wafers. Integrated develop- diamond wire leads to fewer microcracks and the
ITO
Ag (PVD) (n) a-Si:H ment throughout the PV value chain (diamond wire wafering, het- depth of microcracks is shallower. With decreas-
Test & Sort erojunction cell technology and SmartWire Connection Technology ing wafer thickness and the corresponding higher
(SWCT) guarantees the maximum performance of HJT modules. surface to bulk ratio the surface recombination
loss becomes dominant compared to recombi-
Figure 1: Meyer Burger technology for high performance solutions ensures nation loss in the bulk. Consequently superior
Schematic drawing of the HJT cell and the process steps required the perfect integration and harmonisation of all processes with the surface passivation techniques are mandatory.
goal of lowest LCOE.

2 3
 HELiAPECVD: a-Si:H coating for front and rear side HELiAPVD: TCO/ Metal coating for rear side contact

In general there is an optimum wafer thickness for 25.0 For heterojunction cell concepts, intrinsic a-Si:H layers are depos- doped p a-Si:H) and a back surface field (BSF)
every cell technology. However as shown in Figure 3, ited on the front and rear sides of the wafer respectively, providing on the rear side (phosphorous doped n a-Si:H).
24.0
the final cell efficiency for heterojunction cells is excellent surface passivation. The emitter enables the separation of the charge
independent of the wafer thickness at least in the 23.0 carriers while the BSF drives the minority charge

cell efficiency [%]

range of wafer thicknesses achievable with sawing Doped a-Si:H layers are deposited on the intrinsic a-Si:H layers carriers away from the rear side to reduce rear
processes. 22.0 in order to create an emitter on the front side of the cell (boron side recombination losses.

Figure 3: Impact of wafer thickness on efficiency of 21.0

heterojunction cells including the potential with im-

20.0
proved light management 80 100 120 140 160 180 200 220 240 260 Fz1: 12378 µs
Fz2: 12841 µs
mean cell thickness [µm]
ST27-2205: 4472 µs
10-2 ST27-2209: 5639 µs
ST27-2222: 6769 µs
ST27-2318: 4837 µs
Process flow / Texturing ST27-2324: 4752 µs
ST27-2325: 4365 µs
An optimal texturing and cleaning process forms the 16

Measured lifetime [s]

basis for a successful production process for highly 14
efficient HJT cells. Meyer Burger process know- 12
how sets the foundation for optimum passivation.
SRV Seff [cm/s]

10
The initial steps in the HJT process sequence are 8
wet chemistry processes for saw damage removal 6
(SDR), texturing, cleaning and hydrogen termination.
4
The subsurface damages and microcracks have to
2
be almost completely removed to achieve high ef-
0
ficiency HJT cells. Measuring the surface recombina- 4 6 8 10 12 14
tion velocity (SRV) indicates the necessary saw dam- SDR [µm/side]
age removal independent of the quality of the bulk. 10-3
Figure 4 shows the impact of saw damage removal Figure 4: Impact of saw damage removal on SRV 1014 1015 1016 1017
on the surface recombination velocity. Minority Carrier Density [cm-3]
Figure 5:
The saw damage removal can be optimised with the Si:H contributes to passivation due to its field effect properties. Sinton lifetime data of an i-passivated polished FZ and CZ wafers (Numbers for internal reference of experiment)
texturing process. A final short dip in hydrofluoric
acid terminates the silicon surface until final passiva- Roth & Rau, a member of the Meyer Burger Group, has devel- Stripe Stripe Stripe Stripe
4 3 2 1
tion in the subsequent PECVD process. The surface oped the modular high throughput HELiAPECVD system specifi-
after this wet chemical treatment is crucial for the cally for wafer based silicon HJT cell concepts [2]. The heart of
quality of c-Si / a-Si:H interfaces and therefore for the HELiAPECVD system is the patented S-Cube™, a sophisticat-
PM PM PM PM
the surface passivation. ed parallel plate plasma reactor with a box-in-box arrangement n i p i In the single stripe:
providing ultra-pure and uniform amorphous silicon layers. A 1. Automatic carrier movement
2. Process sequence automatic
13.56 MHz RF source is used to minimise plasma damage
PECVD: during layer deposition. Thus the required quality of the amor-
a-Si:H coating for front and rear side phous silicon layers in terms of minority charge carrier lifetime
In order to minimise energy loss within the solar cell, and band gap are ensured. The key parameter to determine
the surface must be highly passivated. Low tempera- the quality of the passivation with the a-Si:H layer is the minor- HTO WF HTI

ture passivation with a-Si:H deposited in a tempera- ity charge carrier lifetime. The HELiAPECVD system has achieved
ture range between 150-250°C results in outstanding values of more than 10 ms effective lifetime on FZ-wafers and Automatic wafer transfer

surface recombination velocities. This a-Si:H is able more than 4 ms on CZ production wafers as shown in Figure Automatic wafer transfer
Automatic wafer transfer
to passivate all levels of n-type and p-type silicon. 5. These sets of data both for the polished FZ-wafers and for between Stripes
Doped a-Si:H is used to form both the emitter and the textured CZ-wafers demonstrate obviously the passivation
the back surface field (BSF). Additionally, doped a- capability of the HELiAPECVD system. Figure 6: Wafer flow within the HELiAPECVD system

4 5
 Lift WL Double LM BMi TCO BMo TM BMi Metal BMo UM WU Lift

D - Unload

Figure 7: HELiAPVD system configured for HJT cell processing C - Process

B - Process
To achieve a high quality surface passivation of a HJT cell, it is es- while providing a good electrical contact with the
sential to avoid any cross doping. The ­HELiAPECVD system therefore doped a-Si:H layers. In a second chamber of the
performs these process steps in dedicated process strips to use same HELiAPVD system, a metal layer such as sil-
an optimised process setting for each layer and prevent from any ver, or a stack of different metals such as a com- A - Load CALiPSO: flexible heat treatment furnace
surface contamination.” Figure 6 shows a schematic of a HELiAPECVD bination of silver and nickel vanadium is applied
system including the wafer flow. on the TCO layer as rear side metallisation. The
composition of the metal stack is dependent on
The HELiAPECVD system in Figure 6 is designed for a throughput up the downstream module process. The rear side
Curing
to 2,400 wafers per hour. As the a-Si:H layers are in the range of metallisation provides an excellent electrical con- After printing the front side, contact firing is not required with this ers on the cells as well as the solderability of the
only a few nanometers, the deposition process is very short and tact to the TCO layer as well as a good internal HJT cell concept. The curing of printed HJT cells is a simple ther- cells. The Roth & Rau curing system is tailored to
the gas consumption is low. reflectance of long wavelength light. This is ac- mal process at temperatures <250°C in order to outgas the sol- the HJT process in terms of process and produc-
complished without having to break the vacuum vents of the low temperature paste. The temperature profile affects tivity requirements [2].
or to turn the wafer between these coating pro- the conductivity of the screen printed lines and of the TCO lay-
PVD: TCO / Metal coating for rear side contact cesses. Figure 7 shows a HELiAPVD system con-
In the modular HELiAPVD system, a sputter process is used in one figured for HJT cell processing. This configuration
of the chambers to apply a transparent conductive oxide (TCO) enables to achieve the edge isolation simultane-
layer to the front and rear side of the wafer [2]. In addition to col- ously avoiding extra laser or chemical steps.
Testing
lecting the photo-generated current and forming an ohmic contact Rotating cylindrical sputter targets for TCO and HJT cells are high capacitance cells which require a measurement measure the power of high capacitive modules,
on the cell, the TCO layer on the front side acts as an anti-reflection metals on magnetrons enable a high target uti- time of 400-600 ms. This is significantly longer compared to stand- taking into account production tact-time, low
layer. Especially for the front side of the HJT cell, the optical and lisation of over 85% to be achieved, ensuring ard low capacitance cells. Pasan SA [3], a member of the Meyer TCO and high measurement accuracy. This in-
electronic properties of a-Si:H and TCO layers need to be adjusted a cost effective coating process. The HELiAPVD Burger Group, in cooperation with the Institute of Micro Technology novative solution has been successfully validated
with respect to each other. Indium tin oxide (ITO) is a favorable TCO and HELiAPECVD systems are perfectly matched in (IMT) at the University of Neuchâtel in Switzerland [4], has devel- by PI Berlin.
material for HJT cells because it is very transparent and conductive terms of capacity and layer properties. oped a new I-V curve cell tester series known as SpotLIGHT, which
is available in two formats: SpotLIGHT 1 sec and SpotLIGHT HighCap.
With its high quality A+A+A+ 5 ms length pulsed Xe light source,
SpotLIGHT 1 sec is dedicated to the high-speed measurements that
Print front side are required for in-line applications, such as end-of-line quality con-
102.0
Conventional crystalline silicon solar cell technologies use front trol in solar cell production lines or beginning-of-line quality control
101.5
side collector lines (“fingers”) and busbars. Reducing the con- in solar module production lines.
sumption of the silver paste used in the fingers and busbars is 101.0 High quality Xenon
key to reducing costs. 100.5 SpotLIGHT HighCap is dedicated to testing solar cells with high ca- flash for calibration
Power [%rel]

100.0 pacitances such as heterojunction cells. The SpotLIGHT HighCap of LEDs

Meyer Burger offers the busbarless SmartWire Connection combines light-emitting diodes (LEDs) to increase the pulse length
Technology (SWCT) to minimise cell-to-module (CTM) loss and 99.5
from 5ms to 600ms (Figure 9). This hybrid light source measure-
to optimise the module efficiency. Only the finger grid pattern 99.0 ment process had been validated by Fraunhofer ISE. The result is a
on to the front surfaces of HJT cells will be formed by using a 98.5 system which provides dependable solar cell measurements while
conventional screen printer and an epoxy-based low tempera- maintaining the total COO of the system at the same level achieved
98.0
ture paste. The busbarless design facilitates a fine-lined screen Screen A Screen B
by the SpotLIGHT 1 sec. This is a unique characteristic among long-
Finger width: 90-100 um Finger width: 50-60 um
printed grid pattern. By replacing the busbars with a lined grid 110 mg of Ag paste 40 mg of Ag paste pulse solar simulators.
pattern with thinner and smaller fingers, silver consumption is
significantly reduced. It also increases the cell conversion effi- Figure 8: Power gain combined with reduced Measurement of high efficiency modules has different constraints:
ciency by decreasing shadow loss on the cells. Figure 8 shows silver consumption of Heterojunction SmartWire illuminating a large area during several hundreds of millisecond LED flash
the performance of ­HJT SWCT modules. The module power is Connection Technology modules would not be cost effective, neither accurate. On the other hand, up to 600ms
increased while silver consumption is reduced. the capacity of a module is much lower than that of a cell. There-
be combined with SWCT without causing signifi- fore, Pasan has developed the DragonBack® solution which is
With the application of SWCT, the requirements for finger conduc- cant power losses. The shorter the distance be- a dynamic sweep methodology that can be used with standard Figure 9:
tivity become less stringent compared to the requirements for con- tween two finger contacts, the lower the impact HighLIGHT module tester with a 10 ms pulse length of A+A+A+ qual- Measuring methodology of the new
ventional busbar designs. Finger resistance of up to 100 Ω/cm can of power losses in fingers. ity. Such industry-oriented and cost-effective solution enables to Pasan spotLIGHT HighCap [3]

6 7
 VOC : 44.38 V
ISC : 9.01 A HETEROJUNCTION TECHNOLOGY
FF : 76.8%
PMPP : 307.1 W

HJT – A breakthrough in levelized cost of electricity

LCOE is one of the primary metrics used to measure Figure 11 shows a comparison of both regions. For both
the cost of electricity produced by both utility-scale and regions, the LCOE achieved using the HJT PV system
Milestone: 307 watt module distributed power systems. A simulation for 100 kW PV clearly outperforms the other PV systems. The HJT PV
At 307 watts, Meyer Burger set a new photovoltaic record power systems using silicon wafer and thin film PV mod- system benefits from its excellent temperature coefficient,
using a standard 60 cell solar module shown in Figure 10 ules was done for two different climatic conditions based higher conversion efficiency, higher energy yield and lower
[5]. A high level of process integration between the wafer, on Germany which represents a cold climate and on India balance of system costs.
cell and module technologies made this developmental which represents a humid, subtropical climate. The calcu-
leap possible. lation used PV SOL 6.0 Expert and the following param- As a result of this simulation LCOE of 9.1 $cent/kWh for
eters: Inverter Fronius Agilo 100.0-3, Modules Poly: 245 Germany and 5.2 $cent/kWh for India are calculated.
The Meyer Burger HJT modules deliver an increased en- W, HJT 290 W, CdTe 90 W, CIGS 95 W and the following
ergy yield by combining HJT cells with a conversion ef- financial assumptions: 80% bank loan, 20% self invest, These attractive numbers underline the superiority of the
ficiency of 21% and a very low temperature coefficient of capital interest: 2%, loan interest 4%, 0,5% degradation/ Meyer Burger heterojunction technology compared to ex-
just TC = -0.20%/K, together with the revolutionary Smart- year, 0,7% running cost/yr., base system cost 2015 with isting PV technologies currently available on the market.
Wire Connection Technology. Compared to standard cells 20% margin.
which have a value of TC = -0.43%/K, a solar module us-
ing HJT cells from Meyer Burger can achieve >10% more
energy yield on average. This results is a significant com-
petitive advantage for cell and module manufacturers, as
well as for end customers. SGS Fresenius Institute has Site: Germany Site: India (Delhi)
already IEC certified HJT modules with SmartWire Con-
nection Technology. 0.140 0.140
0.120 0.120
0.100

LCOE [$/kWh]
0.100

LCOE [$/kWh]
0.080 0.080
0.060 0.060
0.040 0.040
Record Module! 0.020 0.020
60 cells, 156 mm x 156 mm 0.000 0.000
307 watt Standard Poly HJT Standard Poly HJT
(Meyer Burger) (Meyer Burger)
CdTe CdTe
CIGS CIGS

Figure 10: 307 watt record HJT module from Meyer Burger Figure 11: Levelized cost of electricity for a 100 kW PV power system

8 9
 ABOUT US AUTHORS

Meyer Burger
Meyer Burger Technology Ltd. is the technology leader at every Benjamin Strahm graduated in Material Sciences from Swiss Federal Institute of Technology (EPFL) in 2004 and
stage of the photovoltaic (PV) value chain – from wafers to building- achieved his PhD in industrial plasma physics in 2007 in the field of silicon deposition in large area reactors. In 2008,
integrated PV systems. Processes and systems from the Meyer he joined Roth & Rau Research (formerly Roth & Rau Switzerland) working in the field of plasma reactor design and
Burger group play a vital role in increasing overall performance and crystalline silicon surface passivation by amorphous silicon for heterojunction solar cells. Since 2012, he is heading R&D
efficiency throughout the value chain. activities in Roth & Rau Research.

All core expertise under one roof Sylvère Leu was born and educated in Switzerland, and graduated from ETH Polytechnic with a degree in Electronic
Engineering and in Business administration at the University of St.Gallen (HSG). Sylvère Leu began working in Photovol-
Comprehensive, integrated technology portfolio
taics 25 years ago. At the end of 2005 he was charged with building up an integrated 250 MWp facility. In 2008 Sylvère
Meyer Burger covers all important technology stages in the photo- Leu joined 3S Industries AG as COO. The company merged in 2010 into Meyer Burger Technology, where he is working
voltaic value chain. The interdisciplinary research and development as CIO/CTO of the group.
work from wafers to the installed modules results in new standards
in production procedures and processes and guarantees efficiency
and performance throughout the entire system. We rapidly incor- Jens Krause studied semiconductor physics & technology at the University of Technology in Chemnitz, Germany.
porate research results into commercial, integrated solutions. After spending three year at the Center for Microtechnologies of TU Chemnitz, he worked for more than ten years in the
semiconductor industry.  In 2009 he joined Roth & Rau, a member of the Meyer Burger Group, working now as head of
strategic product management.

USA Europe Asia Georg Roters studied physics at the Westfälische Wilhelms – Universität Münster, Germany. He received his PhD
in physics from the Ruhr – Universität Bochum, Germany and an Executive Master from the Boston Business School,
Zürich. Georg Roters worked as Process Engineer, Project Manager and Product Marketing Manager in the semicon-
ductor industry for 12 years. He joint Roth & Rau AG, a member of the Meyer Burger Group, in 2009 were he was
Moscow (RU) building up and heading the Product Management department until 2012. Currently he is heading the Sales department
Dresden (DE) within the Roth & Rau B.V.
Hohenstein-Ernstthal (DE)
Hillsboro (OR) Zülpich (DE)
Colorado Springs (CO) Eindhoven (NL)
Columbia (NJ) Reichelsheim (DE) Tokyo (JP)
Umkirch (DE) Seoul (KR) André Richter has a degree in electronic engineering (communication engineering, process measuring and control
Thun (CH) Zhubei City (TW)
Shanghai (CN) technology and environmental measurement) and he managed his own company specialising in electronic education
Neuchâtel (CH)
Pune (IN) systems for eleven years. Since 2001 his focus has been on photovoltaics. At Conergy AG he was involved in develop-
Singapore (SG)
ment and third level support of solar plants where he held the position of Technical Director in the Conergy solar plant
located in Frankfurt (Oder). He then assumed the role of CEO for Conergy Electronics GmbH. In 2008 Andre Richter
joined the Geneva based company, SES, as a consultant for the planning and construction of module lines in the United
States. Since 2010 he has held the position of Senior Technical Developer at Meyer Burger Technology Ltd where he
focusses on the establishment and realisation of strategic photovoltaic projects.

References
[1] Panasonic (http://panasonic.co.jp/corp/news)
[2] Roth & Rau AG (http://www.roth-rau.com)
[3] Pasan SA (http://www.pasan.ch)
[4] IMT, Photovoltaics and Thin Film Electronics Laboratory
(PV-Lab), EPFL, University of Neuchâtel, Switzerland
With around 1800 employees worldwide, Meyer (http://pvlab.epfl.ch)
Burger maintains close proximity to its customers.
[5] Meyer Burger AG (http://www.meyerburger.com)

10 11
PASSIONATE ABOUT PV

Meyer Burger Technology Ltd

Schorenstrasse 39 / 3645 Gwatt (Thun) / Switzerland
HJT-EN-01

Phone +41 33 221 28 00 / Fax +41 33 221 28 08

[email protected] / www.meyerburger.com