Conference PaperPDF Available

High Efficiency Phosphorus Emitters for Industrial Solar Cells: Comparing Advanced Homogeneous Emitter Cells and Selective Emitters Using Silicon Ink Technology

Authors:
  • Jiangxi Jiayin

Abstract and Figures

Recent advancements in silver paste technology, which can contact low surface concentration phosphorus emitters, have enabled higher efficiency homogeneous emitter solar cells. This paper presents a comparison between two industrial approaches for high efficiency front contact solar cells: (i) lightly doped emitter cells using advanced silver paste and (ii) selective emitter cells using silicon ink technology. Optimal emiter profiles were determined for the lightly doped emitter cells in order to compare both technologies. For lightly doped emitter cells, it was found that phosphorus surface concentrations can be lowered to ~2.5e20 atoms/cm3 providing a 5mV gain over a conventional cell while maintaining good metal contact. It was also found that a surface concentration above 2e20 atoms/cm3 is required to minimize recombination at the metal contact. Through process optimization average efficiencies of 18.9% are reported for lightly doped emitter cells compared to 18.6% for optimized conventional cells and 19.2% for selective emitter cells using silicon ink. The gains achieved with lightly doped emitter cells shrink the efficiency gap between the homogeneous emitter and selective emitter cell from 0.6% down to 0.3%. Selective emitters, with their dual doped regions, continue to provide superior Voc, maintain excellent contact and minimize recombination at the metal contact region. However, with the continuous advancements in Ag paste technology the gap between selective emitter and homogenous emitter solar cells may continue to shrink.
Content may be subject to copyright.
High Efficiency Phosphorus Emitters for Industrial Solar Cells: Comparing Advanced Homogeneous
Emitter Cells and Selective Emitters using Silicon Ink Technology
Giuseppe Scardera1, Andreas Meisel1, Kurt R. Mikeska2, Alan F. Carroll3, Michael Z. Burrows1, Dmitry Poplavskyy1,
Malcolm Abbott1, Francesco Lemmi1, and Homer Antoniadis1
1Innovalight, Inc. 965 E Arques Ave, Sunnyvale, CA 94085, USA
2DuPont Central Research and Development, Wilmington, DE 19880-0304
3DuPont Electronic Technologies, Research Triangle Park, NC 27709
ABSTRACT: Recent advancements in silver paste technology, which can contact low surface concentration
phosphorus emitters, have enabled higher efficiency homogeneous emitter solar cells. This paper presents a
comparison between two industrial approaches for high efficiency front contact solar cells: (i) lightly doped emitter
cells using advanced silver paste and (ii) selective emitter cells using silicon ink technology. Optimal emiter profiles
were determined for the lightly doped emitter cells in order to compare both technologies. For lightly doped emitter
cells, it was found that phosphorus surface concentrations can be lowered to ~2.5e20 atoms/cm3 providing a 5mV
gain over a conventional cell while maintaining good metal contact. It was also found that a surface concentration
above 2e20 atoms/cm3 is required to minimize recombination at the metal contact. Through process optimization
average efficiencies of 18.9% are reported for lightly doped emitter cells compared to 18.6% for optimized
conventional cells and 19.2% for selective emitter cells using silicon ink. The gains achieved with lightly doped
emitter cells shrink the efficiency gap between the homogeneous emitter and selective emitter cell from 0.6% down to
0.3%. Selective emitters, with their dual doped regions, continue to provide superior Voc, maintain excellent contact and
minimize recombination at the metal contact region. However, with the continuous advancements in Ag paste
technology the gap between selective emitter and homogenous emitter solar cells may continue to shrink.
Keywords: diffusion, phosphorus emitter, emitter saturation current, crystalline silicon, selective emitter, silicon ink,
lightly doped emitter.
1 INTRODUCTION
Two industrial approaches for high efficiency front
contact solar cells are discussed in this paper: (i) Lightly
Doped Emitter (LDE) cells using advanced silver paste
(DuPontSolamet® PV17x) and (ii) Selective Emitter
(SE) cells using Silicon Ink Technology. Selective
Emitter solar cells made with Silicon Ink Technology
have already been shown to be an industrially viable
approach to making high efficiency solar cells [1,2].
Recent advancements in Ag paste technology enable
contacting homogeneous emitters with phosphorus
surface concentrations lower than those used in
conventional solar cells [3].
SelectiveEmitter
(SE)
LightlyDoped
Emitter(LDE)
Contactlow
surfaceconc.
emitters
Heavilydoped
contactregions
&
Verylowsurface
con.emitter
Optimize
diffusion
AdvancedAg
paste
SiInkProcess
Optimize
diffusion
Features
Process
Requirements
HighEfficiency
FrontContact
Solarcell
Figure 1: Approaches to high efficiency front contact
solar cells.
As highlighted in Figure 1, both approaches allow for
phosphorus emitters with lower surface concentration
without losing contacting capabilities. Implementing
these approaches requires optimization of the diffusion
process in order to have high performance phosphorus
emitters.
Without the constraints of heavy doping for
maintaining good ohmic contact, emitter profiles can be
optimized to achieve lower emitter saturation current
densities (Joe). As a result, such emitters should yield
higher Voc values in completed solar cells. It is well
known that lower Joe values can be achieved for
phosphorus emitters by lowering the surface
concentration [4]. On the other hand, a heavy diffusion
(high surface concentration and deep emitter) is required
to minimize the emitter saturation current underneath the
metal contact (Joe-metal). [5, 6] These two opposing
requirements impose design constraints on the emitter
profile used in a homogenous emitter.
In this paper, three classes of phosphorus emitters
relevant to screen printed solar cells are studied:
Standard, with high surface concentration, Lightly Doped
Emitter (LDE), with low surface concentration, and High
Efficiency Emitter, with the lowest surface concentration
(used for SE cells). These three classes are categorized by
their surface concentration ranges which also determine
their Joe ranges. This work also discusses the shrinking
efficiency deltas between homogenous and selective
emitter solar cells given the recent entry of the LDE cell
into manufacturing.
2 EXPERIMENTAL
Phosphorus emitters were created using a
conventional POCl3 diffusion quartz tube furnace.
Emitter surface concentration and depth were modulated
27th European Photovoltaic Solar Energy Conference and Exhibition
923
by varying the POCl3 diffusion processing parameters. Joe
values were extracted from qss-PC measurements at high
injection levels measured on textured symmetric
structures with double-sided diffusion and silicon
nitrideand subjected to a firing condition used in normal
cell processing. Phosphorus profiles were measured using
Secondary ion mass spectroscopy (SIMS) on polished
wafers processed alongside the textured wafers. The
sheet resistance of the emitters were also monitored using
four point probe measurements.
Standard and LDE solar cells were fabricated with
various emitter profiles in order to monitor the impact of
the modulated Joe on cell performance. A set of SE cells
made with Si Ink were also fabricated as a reference.
The cell fabrication process was completed on
industrial scale equipment using commercially available
silicon wafers. P-type Cz-Si pseudo-square
(156mmx156mm,180μm thick) wafers with a bulk
resistivity of about 2 ·cm were used as substrates. The
as-cut wafers were cleaned and textured in an alkaline
bath to create a random pyramid textured surface.
Phosphorus emitters were subsequently performed in an
MRL Industries quartz diffusion tube, followed by
phosphosilicate glass (PSG) removal performed in a
RENA inline wet cleaning InOx tool. The front surface
was then passivated with an antireflective silicon nitride
coating using a Roth and Rau inline PECVD system.
Subsequently aluminum back-surface field (Al-BSF) and
silver front metal contacts were screen printed using a
conventional Baccini screen printer. Finally contact co-
firing was done in a Despatch belt furnace, followed by
laser edge isolation. The SE cells were fabricated with
the same process flow with the addition of a Si Ink
printing and baking step prior to diffusion. A three busbar
front contact design was used for all cells. All cells were
fabricated using the new advanced Ag paste for the front
grid.
3 RESULTS AND DISCUSSION
3.1 Emitter profiles and Joe
Representative SIMS profiles for each emitter class
are shown Figure 2. All of the emitter profiles shown
have sheet resistance values of approximately 70
Ohm/sq. The POCl3 diffusion process parameters were
varied in order to achieve a range a surface concentration
values of each emitter class.
1.E+16
1.E+17
1.E+18
1.E+19
1.E+20
1.E+21
0 0.1 0.2 0.3 0.4 0.5 0.6
[P](atoms/cm3)
depth(microns)
Standard
LDE
HEE(SE)
Figure 2: SIMS phosphorus profiles for a Standard
emitter, a Lightly Doped Emitter (LDE) and a High
Efficiency Emitter (HEE), which is used for SE cells. All
of these emitters have a sheet resistance value of
approximately 70 Ohm/sq as measured by four point
probe.
Figure 3 shows the SIMS profiles for the emitter
region and the contact region under the Si Ink used for
SE cells. The phosphorus surface concentration for the
emitter region is ~1.6 e20 atoms/cm3 while that for the Si
Ink defined contact region is ~3.9e20 atoms/cm3. These
two profiles essentially meet the requirements to achieve
low Joe, good contact and minimize recombination at the
metal contact.
1.E+16
1.E+17
1.E+18
1.E+19
1.E+20
1.E+21
0 0.2 0.4 0.6 0.8
[P](atoms/cm3)
depth(microns)
emitter
Ink
Figure 3: SIMS profiles for the emitter region and the
contact region under the Si Ink used for the SE cell.
Joe values for the three emitter classes as a function of
sheet resistance are shown in Figure 4. There is a clear
offset in Joe values between the three classes of emitters.
The Standard emitters with the highest surface
concentration range span the highest Joe values, while the
HEE, with the lowest surface concentration values, span
the lowest Joe values. There is a significant offset between
the Standard and LDE Joe values suggesting significant
Voc improvement if the metal contact can be maintained.
Standard
HEE (SE)
LDE
Figure 4: Emitter saturation current densities, Joe, for
three classes of phosphorus emitters as a function of
emitter sheet resistance.
3.2 Cell performance
Figure 5 shows Voc results for Stanard and LDE cells
with varying emitter sheet resistance as well as reference
SE cells. Significant improvements are made to Voc for
both Standard and LDE cells by dropping the surface
27th European Photovoltaic Solar Energy Conference and Exhibition
924
concentration shown as an increasing sheet resistance in
Figure 5. Voc values for LDE approach those of the SE
cell. The offsets observed in Joe values between Standard,
LDE and HEE is reflected in the corresponding Voc
results. The lightest LDE condition suffers from reduced
Voc. This significant drop is likely due to increased Joe-
metal indicating that a surface concentration above 2e20
atoms/cm3 is required to minimize minority carrier
recombination at the metal contact.
Voc (V)
Figure 5: Voc results for Standard and LDE cells with
varying emitter sheet resistance and of SE cells.
Figure 6 shows the corresponding series resistance
values for the cells represented in Figure 5. For the
Standard emitter cells the Rseries only slightly increases
with lighter diffusion owing to its relatively high surface
concentrations. For the LDE cells the series resistance
increases dramatically for the 68 Ohm/sq condition
indicating that surface concentrations above 2.4e20
atoms/cm3 are required to avoid contacting problems.
Se
r
ies
r
esistance (Ohm-cm2)
Figure 6: Series resistance results for Standard and LDE
cells with varying emitter sheet resistance and of SE
cells.
Based on the above results further diffusion and front
grid optimizations were performed in order to capitalize
on the Voc gains demonstrated. IV results (average
values) for optimized cells are shown in Table 1. It should
be noted that laser edge isolation was used for all cells,
which we have found to lower cell efficiencies by up to
0.2% absolute compared to edge isolation using inline wet
chemistry. There is a clear offset in Voc between the
Standard, LDE and SE cells. FF for the Standard cell is
lower because the Voc and Jsc were maximized via the
emitter to give the highest efficiency.
Cell Voc
(mV) Jsc
(mA/cm2) FF
(%) *Eff.
(%)
Standard 630 37.4 79.0 18.6
LDE 635 37.3 79.9 18.9
SE 639 37.6 79.9 19.2
Table 1: IV results (average values) for optimized cells.
*Laser edge isolation used.
4 CONCLUSIONS
Two approaches for driving screen printed front
contact solar cells to higher efficiencies have been
discussed in this work: (i) the lightly doped emitter
(LDE) with advanced Ag paste and (ii) the selective
emitter (SE) using silicon ink technology. Both
approaches use higher performance phosphorus emitters
and require diffusion optimization targeting lower surface
concentration phosphorus emitters.
The superior contacting capability of new advanced
Ag pastes enables lower phosphorus surface
concentration emitters thus significantly increasing the
Voc and efficiency of homogenous emitter solar cells.
The gains achieved with LDE cells shrink the efficiency
gap between the homogeneous emitter and SE cell from
0.6% down to 0.3%.
For LDE it was found that phosphorus surface
concentrations need to be maintained above 2.4e20
atoms/cm3 to maintain good metal contact. It was also
found that a surface concentration above 2e20 atoms/cm3
is required to avoid minority carrier recombination at the
metal contact. The offset between these two limits may
leave room for further Ag paste improvement.
Selective emitters, with their dual doped regions,
continue to provide superior Voc, maintain excellent
contact and avoid recombination at the metal contact
region. However, with the continuous advancements in
Ag paste technology the gap between selective emitter
and homogenous emitter solar cells may continue to
shrink. This efficiency gap will also need to be re-
evaluated as advances are made in the area of fine line
printing.
5 REFERENCES
[1] H. Antoniadis, F. Jiang, W. Shan and Y. Liu, “All
Screen Printed Mass Produced Selective Emitter
Solar Cells”, Proc of 35th IEEE Photovoltaic
Specialist Conference, Honolulu, USA, 2010.
[2] G. Hahn, Proc. of 25th PVSEC Conference, Valencia,
Spain, 2010.
[3] K. R. Mikeska, Z. R. Li, P.D. VerNooy, et al. Proc. of
26th PVSEC Conference, Hamburg, Germany, 2011.
[4] R. R. King, R. A. Sinton, and R. M. Swanson, IEEE
Trans. Electron Devices, 37(2) , 365-371, 1990.
[5] A, Cuevas, P. A. Basore, G. Giroult-Matlokowski, C.
Dubois, J. Appl. Phys. 80(6) 3370-3375 (1996).
[6] U. Jäger, S. Mack, A. Kimmerle, A. Wolf and R.
Preu, Proc of 35th IEEE Photovoltaic Specialist
Conference, Honolulu, USA, 2010.
27th European Photovoltaic Solar Energy Conference and Exhibition
925
... As a result more aggressive emitter profiles with lower surface concentrations can be contacted. Excellent metal contacts and high efficiency cells have been reported for emitters with surface concentrations as low as 2.4 Á 10 20 cm À 3 and 1.4 Á 10 20 cm À 3 when implementing a deeper profile [104,105]. ...
... Optimization of implant dose and annealing conditions has led to low surface concentration P emitters [137][138][139] more closely resembling those of advanced POCl 3 processing [104,105]. Oxygen is often introduced into the damage anneal ambient in order to simultaneously grow a thin oxide passivation layer with reported thicknesses in the 10-30 nm range [129,134,137,139]. ...
... Manipulation of the emitter profile during diffusion can mitigate this [101,128]. As mentioned in the emitter section, the trend for monocrystalline silicon cells has been towards lower surface concentration and deeper profiles which reduce emitter recombination losses and maintain good metal contact [100][101][102][103][104][105][106]255]. ...
Article
This article is the second article in a three-part series dedicated to reviewing each process step in crystalline silicon (c-Si) photovoltaic (PV) module manufacturing process: feedstock and wafering, cell fabrication, and module manufacturing. The goal of these papers is to identify relevant metrology techniques that can be utilized to improve the quality and durability of the final product. The focus of this article is on the cell fabrication process. In this review, the fabrication of c-Si PV cells is divided into four steps: (1) wet chemical processes; (2) emitter formation; (3) anti-reflection coating (ARC) and passivation deposition; and (4) metallization. Each of these processing steps can impact the final reliability and durability of PV modules deployed in the field, and here the failure modes and degradation mechanisms induced during cell manufacturing are explored. Additionally, a literature review of relevant measurement techniques aimed at reducing or eliminating the probability of such failures occurring is presented along with an assessment of potential gaps wherein the PV community could benefit from new research and demonstration efforts.
... Innovalight developed a SE technology based on Si ink containing Si nanoparticles [12,13]. In this process the Si ink is selectively applied on the front, either by ink jet or screen printing. ...
... During the following P diffusion, a lightly doped emitter is formed in the non-inked areas, whereas the inked areas become heavily doped. The Innovalight SE technology has achieved an efficiency of 19.2% on screen-printed c-Si solar cells [13]. Ion implantation based SE technology has been reported by Varian [14], using in-situ masking to achieve patterned doping during ion implantation. ...
... To date, the above-mentioned SE approaches have not achieved wide-scale acceptance in the PV industry [5]. The reasons for the industry's reluctance to adopt these technologies are the relatively complex laser processing [2,4,6,7] and the use of relatively expensive materials [12,13] and equipment [9,14,15]. The most industrially implemented SE technology was originally developed at University of Konstanz (UKN) [16]. ...
... In the present, commercially available researches involve dopant ink through an additional screen printing step and laser-based SE process, which are available for forming different regions including heavily-doped emitter region with about below 50 Ω /Sq and lightly-doped emitter region with about 90-110 Ω /Sq. Si ink based SE technology has achieved an efficiency of up to 19.2% on screen-printed c-Si solar cells [7,8]. However, both of the SE processes have relied on more masking steps, laser ablate openings and scribe grooves. ...
Article
Full-text available
In the present, a novel cost-effective process scheme for single step selective emitter diffusion was implemented. It is based on the fabrication of acid-resist pattern using a stamping technique with collaboration of a spin on dopant (SOD) and chemical etched-back emitter methods. The SOD diffusion process provided heavily doping n-emitter. Acid-resist pattern without exploitation of a complex method as a photolithography, was stamped as a metal contact pattern for prevention of a localized heavy-dope region from etching back. Phosphorus doping profiles were controlled by etching back time to provide the formation of n-type selective emitter. Sheet resistance is tunable from 10 to 180 Ohm/Sq on localized n-layer. After removal of the patterned acid-resist, the selective n-emitter solar cell structure was obtained under one-step diffusion to achieve a better blue-light response and low contact resistance.
Article
A selective emitter attempts to optimize emitter losses by using two doping levels: high doping under the metal finger and low doping between the fingers. While a selective emitter is an improvement over a homogeneous doped emitter, it is not a full optimization. A graded emitter is introduced here, in which the doping is continually varied between the finger and midpoint between the fingers. The two main emitter power loss mechanisms, recombination and lateral resistance (I2R), vary in severity from the metal finger to the midpoint between two fingers. Thus, in a graded emitter, the doping level is also varied to minimize the total power loss in a continuous manner. The benefits of the graded emitter include a higher efficiency of up to 0.5% abs_{\text{abs}} compared with a homogeneous emitter. In comparison with a typical POCl 3_{3} diffused selective emitter, the ion-implanted graded emitter can achieve a higher efficiency by 0.2% abs_{\text{abs}} at a wider finger pitch, representing a reduction of the front silver paste consumption by 8%. Finally, an industrially feasible method of manufacturing graded emitters using a commercially available ion implantation tool for solar cells is described.
Chapter
Silicon wafer-based technologies comprise more than 90 percent of annual photovoltaic (PV) production. Most silicon PV modules are fabricated using p-type multi-crystalline silicon wafers with screen-printed metal electrodes. More advanced cell structures which use new screen-printing approaches, ion-implantation, selective emitters and rear-surface passivation will continue to increase in future years as manufacturers seek to differentiate themselves in a competitive and diversifying market. The market for higher-efficiency crystalline silicon PV modules is being driven by space-constrained solar PV deployment, typified by the growing Japanese consumer market. The PV market remains dominated by the production of screen-printed p-type crystalline silicon modules with most manufacturing currently focused on production of the cheaper multi-crystalline modules. Although challenges remain for these technologies in terms of cost-effective manufacturing and yield, they represent an opportunity as the return on investment for research and development is potentially higher.
Conference Paper
Full-text available
The fundamental mechanisms of iron impurities in silicon have been thoroughly studied and are well explained in the literature. Of interest to solar cell manufacturers is to understand how these mechanisms manifest in a production environment and, more importantly, how to quickly diagnose and mitigate iron contamination as it occurs. This paper presents examples of iron contamination using p-type CZ wafers processed in production-style environments. The impact of iron on the IV performance of industrial screen printed solar cells is presented, including the time dependence of these effects and how they manifest in the various characterisation techniques that are typically used to diagnose solar cell performance. Examples are given of potential sources of iron contamination and the impact of subsequent processing on the redistribution of those contaminants. The paper demonstrates that iron contamination can occur in a variety of ways, can spread quickly and is severely detrimental to solar cell efficiency. Additionally, it is shown that the fundamental properties of iron in silicon can be used to quickly identify the root cause of contamination in a production environment.
Conference Paper
Silicon solar cell technology has advanced by improvements to all aspects of cell design including metallization. Improvements to metal contacts have been synergistic with other cell improvements such as passivation, lightly doped emitters, and boron doped emitters. This paper briefly recounts the history of c-Si PV metallization technology and its synergy with other cell technologies. It concludes with a perspective on various metal contact mechanisms.
Conference Paper
This paper quantifies the recombination losses associated with industrial POCl3 emitters. We examine three standard (STD) recipes and four of DuPont's lightly doped emitter (LDE) recipes. We find that an STD emitter has a higher effective surface recombination velocity than an LDE recipe of the same sheet resistance because it has a higher surface concentration. More significantly, we find that STD emitters have greater SRH recombination within the doped region, probably because they contain a greater concentration of inactive phosphorus atoms which are known to form silicon phosphide precipitates. These conclusions are drawn from simulations and experiments on the lifetime of test wafers and the quantum efficiency of solar cells. It is shown how this data can be used to distinguish the SRH recombination that occurs at the surface from the SRH that occurs inside an emitter.
Data
Full-text available
This paper describes a galvanic deposition method for the formation of uniformly thin nickel seed layers for silicon solar cells. Unlike the previously-reported electroless and light-induced plating methods, where the thickness of nickel seed layers can be vary significantly due to non-uniform nucleation of metal which can be exacerbated by variations in surface dopant concentrations and roughness, this method saturates resulting in Ni seed layers of uniform thicknesses in the range of 250-400 nm depending on the plating time and patterning process. The method was successfully used in the fabrication of nickel/copper plated homogeneous 100 Ω/□ emitter Si solar cells and laser-doped selective emitter (LDSE) cells, both with screen-printed, Al-alloyed back surface fields. Average cell efficiencies of 18.4% and 18.9% were achieved for the homogeneous emitter cells and LDSE cells, respectively, with the best LDSE cell having an efficiency of 19.2%.
Conference Paper
The Innovalight Cougar™ Platform is a portfolio of simple to implement technologies that, when combined with Innovalight Silicon Ink, enables the manufacture of a selective emitter solar cell with a non-masking single-step diffusion. Innovalight Silicon Ink is a highly engineered silicon nanoparticle colloidal dispersion, implemented for both high volume ink-jet and screen-print deposition, and further optimized to be produced and delivered in commercial volumes. With the addition of an industrial high throughput screen printing system, used to deposit and dry the Silicon Ink, the Cougar Platform requires the same tools and materials used in a standard screen printed solar cell manufacturing line. Consequently, the Cougar Platform improves the efficiency of standard screen printed solar cells by about 1% absolute at a manufacturing cost lower (in terms of $/W) than that of conventional homogeneous emitter front-contact solar cells. Cell efficiencies, on 125mm × 125mm Cz-Si wafers, of up to 19% have been recorded for Cougar cells processed initially at the Innovalight 10MW pilot line. Mass production data coming from the JA Solar manufacturing site, demonstrate median cells efficiencies of 18.6%.
  • A Cuevas
  • P A Basore
  • G Giroult-Matlokowski
  • C Dubois
A, Cuevas, P. A. Basore, G. Giroult-Matlokowski, C. Dubois, J. Appl. Phys. 80(6) 3370-3375 (1996).
  • R R King
  • R A Sinton
  • R M Swanson
R. R. King, R. A. Sinton, and R. M. Swanson, IEEE Trans. Electron Devices, 37(2), 365-371, 1990.