World Journal of Engineering
Monolayer molecular doping for cost-effective photovoltaics
Fei Xiu1, Miao Yu2, Johnny C. Ho1,*, Zhiyong Fan2,*
Department of Physics and Materials Science, City University of Hong Kong, 83 Tat Chee Ave., H.K. SAR,
China.
2
Department of Electronic and Computer Engineering, Hong Kong University of Science and Technology, Clear
Water Bay, Kowloon, H.K. SAR, China.
*
E-mail: johnnyho@cityu.edu.hk; eezfan@ust.hk
1
Abstract
Tremendous effort has been focused on the cost-effective
processing approach for photovoltaics using silicon (Si), the
dominant material for semiconductor industrials and current
PV technologies. In this regard, here, we utilize our recently
reported low-cost, scalable, monolayer surface doping of
semiconductor materials to fabricate Si photovoltaic devices
without using conventional costly doping schemes such as
the ion-implantation. This novel technique is based on the
formation of self-assembled monolayers of dopantcontaining molecules on the surface of Si followed by the
subsequent thermal diffusion of dopant atoms via rapid
thermal annealing. As an example, we demonstrate a high
quality surface (sub-100 nm) n+/p junction Si solar cells
with 10nm thick Ti/Au top electrode. This cell has a
respectable energy conversion efficiency under the
illumination of AM1.5G and confirmed by the device
simulation, the conversion efficiency can be enhanced to >20
% with further optimization in junction profile, top-contact
transparent electrode and anti-reflective coating while
ongoing investigation will be presented to improve the cell
performance.
physics modeling has shown that the formation of ultrashallow junction is preferred to achieve high PV efficiency
since the majority part of electron-hole pair optical
generation occurs just below the surface of the devices. Thus
the shallow junction can effectively separate charge carriers.
Recently, we have developed an innovative approach to
achieve the low-cost, large-scale, high throughput and
controllable monolayer doping for shallow junctions in Si,
which is based on the formation of self-assembled
monolayers of dopant-containing molecules on the Si surface
followed by the subsequent thermal diffusion of dopant
atoms via rapid thermal annealing.[6,7] Notably, in contrast
to the competing technologies, monolayer doping does not
require expensive vacuum equipment and long vacuum
process time which is attractive to be used as a massproduction platform.
2. Results and Discussion
1. Introduction
Solar energy is one of the most promising alternative energy
sources. It is clean, renewable and the most abundant energy
source on earth.[1] However, it is also the least harvested
form of renewable energy.[2] The major obstacles come
from the substantial fabrication and materials costs of
existing high performance silicon (Si) solar cells which are
not economical for terrestrial daily applications, due to the
fact that solar cell modules already take up approximately 60
% of the total cost for the solar power installation.[3] Since
Si is the second most abundant element, making up 25.7%
by mass, in Earth's crust [4] and has been studied thoroughly
in electronic industries for decades, Si will remain as the
major work-horse in photovoltaic infrastructure in future. In
this regard, tremendous effort has been made to develop
highly efficient and low-cost Si photovoltaic devices.
Typical solar cells rely on p/n junctions to collect the photogenerated electron-hole pairs and then convert them into
electricity.[5] During fabrication processes, impurities such
as dopants are thermally introduced to the Si crystal to
produce those junctions; however, this is not possible to
control the junction profile near the surface region to attain
so-called shallow junctions to push the ultimate energy
conversion to the thermodynamic efficiency limit.[5] Device
The
monolayer doping (MLD) process is based on the formation
of self-assembled dopant-containing monolayer on the
crystalline silicon surfaces, followed by the subsequent
diffusion of dopants from the surface into the lattice by a
Figure 1. Process schematic of the wafer-scale monolayer doping
approach.
thermal annealing step (Figure 1). In detail, for the
phosphorus- MLD (P-MLD) process, 4 inch p-type Si wafers
were first treated with dilute hydrofluoric acid (∼ 1%) to
remove the native SiO2. The Si surface was then reacted with
diethyl 1-propylphosphonate (DPP, Alfa Aesar) and
mesitylene as a solvent (25:1, v/v) for 2.5 h at 120 °C to
assemble a P-containing monolayer. The details of this
reaction and the monolayer formation kinetics have been
reported elsewhere.[6,7] Then, a layer of ∼50 nm thick SiO2
is electron-beam evaporated as a cap, and the substrate is
spike annealed between 900-1050 °C in Ar ambient to drive
1235
World Journal of Engineering
in the P atoms and achieve n+/p USJs. The spike annealing
is performed in a rapid thermal processing tool (AG
Associate, model 410) with a fast ramping rate of 100 °C/s
to the target temperature.
Secondary ion mass spectrometry (SIMS) measurements
were performed to characterize the dopant profiles. Figure 2
illustrates the phosphorus SIMS profiling for P-MLD
withspike anneal temperatures of 950-1050 °C. The obvious
trend observed from the SIMS profiling is a monotonic
increase in the junction depth and areal dopant dose with the
diffusion temperature, which indicates that the junction
characteristics can be easily controlled by the annealing
parameters.
3
[P] (atoms/cm )
10
10
10
10
21
20
Figure 3. (A) The device structure of shallow junction Si solar
cell. (B) Corresponding junction profile. (C) Photovoltaic
performance assessment of the shallow junction Si solar cells
under AM1.5G.
19
18
3. Conclusions
10
17
5
10
15
20
Depth (nm)
In summary, shallow junction Si solar cells has been enabled
by a nanoscale monolayer doping technique. The resultant
cell has a respectable energy conversion efficiency and the
device simulation has confirmed that the efficiency greater
than 23% is achievable with this cell configuration.
25
Figure 2. Secondary ion mass spectrometry (SIMS) profile of
phosphorus atoms for different spike anneal temperatures
Using this doping technique, as shown in Figure 3, we have
successfully fabricated a high quality (sub-100 nm) n+/p
junction Si solar cells. In specific, we first monolayer doped
a p-type crystalline Si wafer with phosphorus to form a sub100 nm n+/p USJ and then deposit 10nm thick Ti/Au and
100 nm thick Ni layer for the top and bottom electrical
contacts, respectively. The resultant cell has an energy
conversion efficiency of ~4 % under the illumination of
AM1.5G. Although its cell efficiency and structure are far
from the ideal mainly because of the non-optimized junction
profile and ~50 % optical transparency loss of the top metal
contact, it illustrates the feasibility and potency of the
proposed fabrication process for the low-cost, high
throughput and efficient Si solar cell applications. Further
studies are currently ongoing to explore the appropriate
junction profile and transparent top electrode. Importantly,
the device modeling has confirmed that the energy
conversion efficiency greater than 23% is achievable with
this monolayer doping enabled shallow junction silicon solar
cells.
4. Acknowledgements
This work is financially supported by Innovation and
Technology Fund (ITS/049/10).
5. References
[1] G.W. Crabtree, N.S. Lewis, Phys. Today, 60, 37-42 (2007).
[2] R.E.H. Sims, MRS Bull., 33, 389-395 (2008).
[3] A. R. Jha, Solar Cell Technology and Applications, Florida:
Taylor & Francis Group (2010).
[4] S.E. Manahan, Environmental Chemistry, Florida: CRC Press.
(2005).
[5] A.L. Fahrenbruch, R.H. Bube, Fundamentals of Solar Cells:
Photovoltaic Solar Energy Conversion, New York: Academic
Press, Inc. (1983).
[6] J.C. Ho*, R. Yerushalmi*, Z.A. Jacobson, et al., Nat. Mater., 7,
62-67 (2008).
[7]. J.C. Ho, R. Yerushalmi, G. Smith, P. Majhi, et al., Nano Lett.,
9, 725-730 (2009).
1236