Solar Energy 77 (2004) 931–938

www.elsevier.com/locate/solener

Solar cell research and development using the hot wire

CVD process
A.H. Mahan *

National Renewable Energy Lab, 1617 Cole Blvd., Golden, CO 80401, USA

Received 6 February 2004; accepted 3 June 2004

Available online 3 August 2004

Communicated by: Associate Editor T.M. Razykov

Abstract

This article reviews the research and development of solar cells fabricated using the hot wire CVD (HWCVD) proc-
ess. A short history of the technique is given, and the advantages of this deposition technique as they relate to solar cell
fabrication are quantified. A chronological history of solar cell fabrication involving three major research groups, with
a long history in HWCVD solar cell fabrication, is then given, as such histories illustrate the evolution of solar cell
research involving HWCVD. Illustrative examples of activities of new research groups entering this field are then dis-
cussed, and the outlook for this field is given.

2004 Elsevier Ltd. All rights reserved.

Keywords: Hot wire CVD; Amorphous silicon; Microcrystalline silicon; Deposition site; Staebler-Wronski Effect; Large area
deposition

1. Introduction material was improved considerably (Matsumura, 1986),

but this time the film properties were very similar to
The hot wire CVD (HWCVD) technique has recently those deposited by PECVD, so again not much interest
emerged as an attractive alternative to the standard was generated by these results. However, in 1991 the
PECVD technique for the production of silicon based deposition of a-Si:H containing a very low film H con-
materials for solar cell fabrication. It is a considerably tent (CH), (Mahan et al., 1991) and which exhibited a
newer technique than the PECVD technique. Although smaller Staebler-Wronski Effect (SWE) (Mahan and
US patents on the HWCVD process date back to 1979 Vanecek, 1991), generated significant worldwide interest
(Weismann, 1979), with the first deposition of amor- in HWCVD, and shortly afterwards solar cell fabrica-
phous silicon (a-Si:H) using silane (SiH4) and a hot fila- tion commenced. In 1993 (Papadopoulos et al., 1993)
ment, this work initially did not attract much attention and 1996 (Rath et al., 1997), the first cells were fabri-
because the quality of the latter a-Si:H films, for reasons cated using HWCVD a-Si:H and microcrystalline silicon
which are now understood, was very poor. In 1986 this (lc-Si) as i-layer materials, with the doped layers still
deposited by PECVD. By 1998 several multicham-
ber, load locked deposition systems dedicated to
*
Tel.: +1 303 384 6697; fax: +1 303 384 6531. HWCVD became operational, and all-HWCVD devices
E-mail address: harv_mahan@nrel.gov (Wang et al., 2000) as well as tandem device structures

0038-092X/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2004.06.008
932 A.H. Mahan / Solar Energy 77 (2004) 931–938

(Schroeder et al., 2001) began to be reported. At this low pressure nature of the technique, dust formation is
time, the unique properties of HWCVD as a high depo- not an issue (Gallagher, 2001), thus avoiding the peri-
sition rate (Rd) technique for a-Si:H became apparent. odic chamber cleaning problems associated with
In addition, high efficiency heterojunction structures uti- PECVD solar cell processing.
lizing HWCVD emitters (Schroeder et al., 2001) were For completeness, the issue of filament lifetime needs
also reported. Since 2000, a significant increase in lc- to be kept in mind. However, considerable progress has
Si research and device fabrication has occurred, and been made in this area for both a-Si:H and lc-Si. For a-
the versatility of the technique as regards to solar cell Si:H, this lifetime is already long, due to the use of high
fabrication has begun to be utilized (Holt et al., 2003). filament temperatures, and can be further extended with
Finally, deposition uniformity over large areas has rap- the additional use of a H filament ÔtreatmentÕ, which also
idly progressed (Ledermann et al., 2001; Matsumura et results in the added benefit of enhancing device repro-
al., 2003), and large area machines are being readied ducibility (Iwaniczko, 2003). In addition, progress in de-
for potential commercial production. For a Ôstatus re- vice quality lc-Si deposition using elevated filament
portÕ of solar cell activity in HWCVD prior to 2003, temperatures (translating into a regime where filament
the reader is referred to the excellent review article of alloying has been significantly reduced) has been re-
B. Schroeder in the Proceedings of the Ô2nd Interna- cently reported (Iwaniczko et al., 2003a), and new fila-
tional Conference on Cat-CVD (Hot-Wire CVD) Proc- ment holder designs (Matsumura, 2001), where the
essÕ, held in Denver, CO (USA) in September 2002 filament ends are shielded from exposure to SiH4, have
(Schroeder, 2003). In that work are tables of solar cell eliminated alloying at the extreme filament ends, thus
efficiencies (gs) of every group working in HWCVD extending the filament lifetime indefinitely for both
prior to 2003, along with their cell structures and layers a-Si:H and lc-Si.
deposited by HWCVD. In the present work, therefore,
the results will be presented in a slightly different fash-
ion. In particular, the advantages of HWCVD as they 3. Solar cell results
relate to solar cell fabrication will first be briefly dis-
cussed. After this, the research activities of three groups It is instructive to examine in a chronological fashion
with a long history of solar cell fabrication will be high- the research activities of three groups with a long history
lighted chronologically, as this approach will also serve of solar cell fabrication by HWCVD; in this way, not
to illustrate how the field is evolving in HWCVD. Final- only can particular research results by highlighted, but
ly, new research efforts will be discussed, and the out- also trends in this field can be discerned. Accordingly,
look for this field will be given. Tables 1–3 show the research activities (and solar cell re-
sults) for the University of Kaiserslautern, NREL, and
the University of Utrecht, starting in 1993 and progress-
2. Advantages of the HWCVD process as they relate ing for 10 years. Given in these tables are the HWCVD
to solar cell fabrication material, the cell structure and g, the i-layer Rd, and any
pertinent remarks related to the particular investigation.
Several aspects of the HWCVD process set it apart As can be seen, all three groups share several things in
from thin film deposition using other (PECVD) tech- common. First, all investigated HWCVD a-Si:H early
niques. Three of these have been discussed previously on as the active layer, with the doped layers deposited
(Mahan and Schropp, 2001), and deal with the high by PECVD. Early cells needed air breaks to enable de-
Rd nature of the process, the lack of energetic particle vice fabrication, due to the unavailability of load-locked
bombardment (film damage), and the high gas phase uti- systems dedicated to HWCVD. Nevertheless, reasonable
lization of SiH4 (Matsumura, 2001). In particular, SiH4 gs were achieved, particularly by Kaiserslautern in their
utilizations as high as 80% have been reported for device p–i–n structure (g > 10%) (Bauer et al., 1998). In addi-
quality a-Si:H device fabrication (Mahan et al., 2002). tion, a 6.8% cell was deposited on flat SS by NREL uti-
Three other aspects of this technique are also important lizing their low CH material (Mahan et al., 1998), and
from a commercial point of view. First, deposition sca- such cells exhibited an improved SWE (10%), which
lability has been achieved by the use of multiple fila- was considered significant considering the i-layer thick-
ments and/or filament grids and, unlike the problems ness (4000 Å). Many investigations during this time were
associated with VHF-CVD, there are no fundamental characterized by high a-Si:H Rds. By 1998, load-locked
limits to scale-up. Second, due to the absence of a plas- systems became available to all groups and a-Si:H gs be-
ma needed for gas dissociation, the substrates in came state of the art, as evidenced by the >7% n–i–p
HWCVD are totally decoupled from the deposition cells deposited at high Rd (up to 20 Å/s) on flat SS by
process, suggesting that substrate movement through both NREL (Wang et al., 2000) and Utrecht (van Veen
the deposition chamber can be easily accomplished with- and Schropp, 2001); due to the absence of light trapping,
out affecting the deposition process. Finally, due to the such gs represent the limit on cell g for this structure.
 A.H. Mahan / Solar Energy 77 (2004) 931–938 933

Table 1
University of Kaiserslautern (B. Schroeder)
Year i-Layer Cell structure Doped layers g (%) Rd (Å/s) Remarks References
+
1993 a-Si:H p–i–n on Asahi ÔUÕ PECVD p 4.3 8 First HWCVD cell Papadopoulos et al. (1993)
1997 a-Si:H p–i–n on Asahi ÔUÕ PECVD p+, n+ 8.0 8 Two air breaks Bauer et al. (1998)
1998 a-Si:H p–i–n on Asahi ÔUÕ PECVD p+, n+ 10.2 2.5 Two air breaks Bauer et al. (1998)
1st 10% cell
1999 a-SiGe:H p–i–n on Asahi ÔUÕ PECVD p+, n+ 6.1 3–5 Two air breaks Lill and Schroeder (1999)
2000 a-Si:H p–i–n on Asahi ÔUÕ All HWCVD 7.7 4–5 lc-Si p+ Schroeder et al. (2001)
8.8 3 a-SiC:H p+ Schroeder et al. (2001)
2000 a-Si:H p–i–n/p–i–n tandem All HWCVD 7.0 – – Schroeder et al. (2001)
on Asahi ÔUÕ
2001 a-Si:H Heterogeneous Cell p+ c-Si 15.2 – Highest g for Schroeder et al. (2001)
HWCVD cell structure
2002 lc-Si n–i–p on flat SS All HWCVD 5.4 1.5 1.4 lm i-layer Kupich et al. (2004)

Also in 1998, doped layers began to be deposited by ration current density (JO) and the diode quality factor
HWCVD, with the one exception being Utrecht, which (n), for i-layer film thicknesses in the range 1–2 lm,
still uses PECVD doped layers in all their structures. are significantly less than their PECVD counterparts;
During this time, the 9.8% NREL a-Si:H n–i–p cell both suggest a smaller number of recombination centers
was fabricated on a Ag/ZnO coated SS back reflector, (bulk defects) in the HWCVD i-layers. In addition, light
with the HWCVD i-layer Rd of 16.5 Å/s; at the time this soaking measurements for this cell have indicated a deg-
represented the highest g to date for this (high) i-layer Rd radation after 1000 h of 8%. Interestingly, the amount
(Mahan et al., 1998). By 1999, different materials (a-Si- of this degradation is seen to depend upon the device
Ge:H) as well as different device structures (heterojunc- Voc; for cells exhibiting a lower Voc (0.518 V versus
tion cells, tandem cells) were reported. In this realm, the 0.568 V for the record cell), indicating a (slightly) higher
15.2% heterojunction (active area) cell is quite notewor- crystalline volume fraction, the light induced degrada-
thy (Schroeder et al., 2001), as is the 11.7% tandem cell tion is negligible (Klein et al., 2003).
g, with the HWCVD a-SiGe:H bottom layer deposited Returning again to a-Si:H, recent work by Utrecht
at 10Å/s; this cell stabilized after 1000 h of light soaking (van Veen, 2003) also needs to be highlighted. An n–i–
to 9.6% (Wang et al., 2001). Finally, by 2001–2002 all p device structure on flat SS was fabricated, using a high
three groups started investigating lc-Si, due primarily Rd (10 Å/s) i-layer deposited at a low Tf (1800 C) and a
to the much smaller SWE exhibited by such i-layers low substrate temperature (Ts). This i-layer contained a
when incorporated into devices. Results for the first significant fraction of the H bonded in the IR ÔshiftedÕ
HWCVD triple tandem solar cell utilizing lc-Si were re- (2100 cm 1) mode, as evidenced by the high value of
ported at the Third World Conference on PV Energy the microstructure factor (R = 0.2 to 0.3); this high R
Conversion (Schropp et al., 2004). suggests that this film contained a significant microvoid
While the HWCVD lc-Si single junction cell gs of volume fraction. Nevertheless, as seen in Fig. 2, for an i-
these three groups are still in the 5–7% range (note that layer L of 0.3–0.4 lm, an initial (best) device FF of 0.72
the Utrecht cell is still on flat SS), it is vital to highlight was obtained, and this FF degraded only to 0.66 after
at this stage the results of Juelich, which started 1000 h of light soaking. In addition, the device Jsc in-
HWCVD activities much later than the three groups. creased slightly with light soaking, resulting in a de-
They initiated HWCVD deposition in 1999 and, in con- crease of only 6–8% in g for this device. Experiments
trast to the other groups, emphasized film deposition at are in progress to transfer this film structure to a tex-
low substrate temperatures by running their filaments at tured substrate and to understand the improved stability
low temperatures (Tf < 1700 C) and heating the sub- of this unusual i-layer.
strates exclusively by filament radiation. One year later At this stage it is important to note other recent and
they had achieved a 9.2% a-Si:H p–i–n efficiency, and projected work in HWCVD cell fabrication. Ecole Poly-
then turned their efforts to lc-Si. Their recent 9.4% sin- technique has reported a 6.1% single junction lc-Si g for
gle junction lc-Si g (Klein et al., 2002) stands as a record an n–i–p device (with the light entering through the n-
in HWCVD for this type of cell, and device analysis on layer). While this g is not high, the cell was fabricated
these cells suggest that their HWCVD i-layer material is with two air breaks using PECVD doped layers, and is
superior in quality to their PECVD material of equiva- remarkable for generating >27 mA/cm2 Jsc for a 3.5
lent film thickness and l-volume fraction. These results lm i-layer thickness (Bouree, 2003). While this work
are seen in Fig. 1(a and b). In particular, both the satu- has been discontinued, primarily due to the lack of
 934
Table 2
NREL (A.H. Mahan, Q. Wang)
Year i-Layer Cell structure Doped layers g (%) Rd (Å/s) Remarks References
1995 a-Si:H p–i–n on SnO PECVD 6.2 5 High film CH Nelson et al. (1994)
1997 a-Si:H n–i–p on flat SS PECVD p+ (UniSolar) 6.8 8 Two air breaks, low film CH, lower SWE Mahan et al. (1996)
1998 a-Si:H n–i–p on Ag/ZnO SS PECVD p+ (UniSolar) 9.8 16.5 Record g at high Rd Mahan et al. (1998)
1998 a-Si:H n–i–p on flat SS All HWCVD 7.2 18 State-of-art g for this cell structure Wang et al. (2000)
1999 a-Si:H n–i–p on flat SS All HWCVD 5.9 80 Working devices at 120 Å/s Mahan et al. (2002)
2000 a-SiGe:H n–i–p tandem on Ag/ZnO SS PECVD p+, top cell (UniSolar) 11.7 10 Stabilized at 9.5% Wang et al. (2001)
2002 lc-Si n–i–p on Ag/ZnO SS All HWCVD 6.6 5 0.7 lm i-layer Iwaniczko et al. (2003b)

A.H. Mahan / Solar Energy 77 (2004) 931–938

2003 lc-Si n–i–p on Ag/ZnO SS PECVD p+ (UniSolar) 5.5 8 0.7 lm i-layer, no filament alloying Iwaniczko et al. (2003a)
2003 lc-Si n–i–p on Ag/ZnO SS PECVD p+, top cell (UniSolar) 9.6 8 0.7 lm i-layer, no filament alloying Iwaniczko et al. (2003a)

Table 3
University of Utrecht (R.E.I. Schropp, J.K. Rath)
Year i-Layer Cell structure Doped layers g (%) Rd (Å/s) Remarks References
1997 lc-Si n–i–p on n+ c-Si PECVD p+ 3.2 5 First HWCVD lc-Si cell Rath et al. (1997)
1998 a-Si:H n–i–p on flat SS PECVD n+, p+ 5.6 20 – Feenstra et al. (1998)
1999 a-Si:H n–i–p on flat SS PECVD n+, p+ 7.2 10 State-of-art g for this cell structure van Veen and Schropp
(2001)
2001 a-Si:H a/a tandem on flat SS PECVD n+, p+ 8.5 10 Record tandem g (%) for HWCVD cell van Veen and Schropp
(2001)
2002 lc-Si n–i–p on flat SS PECVD n+, p+ 4.8 1.5 – Schropp (2003)
2002 a-Si:H, lc-Si a/lc tandem on flat SS PECVD n+, p+ 8.1 10 (a-Si) 5 (lc-Si) – Schropp (2002)
2003 a-Si:H, lc-Si a/lc/lc tandem on flat SS PECVD n+, p+ 8.5 10 (a-Si) 2 (lc-Si) First HWCVD triple tandem cell Schropp (2002)
 A.H. Mahan / Solar Energy 77 (2004) 931–938 935

at high Rd. i-layer deposition using different filament

materials is currently being explored. MVSystems will
also explore this device structure, using their graphite fil-
aments and an ÔoscillatingÕ substrate to enable uniform
deposition over large areas (>20 cm · 20 cm) (Morrison
and Madan, 2000). In Japan, the New Energy and
Industrial Technology Organization (NEDO) has
funded a collaborative program, this time to explore
high g HWCVD a-Si:H solar cell development (Mat-
sumura et al., 2003). The participants in this program in-
clude JAIST and the Universities of Osaka and Gifu.
Currently, load-locked deposition systems dedicated to
HWCVD have been designed and installed, properties
of doped and i-layers are being collected, and proof of
concept, all HWCVD cells (with gs > 7.5 %) have been
fabricated using a single chamber system (Matsumura,
2003).
Two other variations in solar cell fabrication illus-
trate the versatility of the HWCVD technique. The first
is the potential of combining HWCVD with other well
established technologies to i.e. increase the c-Si film
Rd. As an example, Kyocera Corporation has used the
Ôcatalyzer assistedÕ PECVD technique to enhance the
conversion g and Rd of both a-Si:H and lc-Si solar cells.
Fig. 1. Saturation current density (JO) versus i-layer thickness Their method involves placing a catalyzer (hot filament)
(d) for HWCVD lc-Si devices (solid symbols) and PECVD lc- inside the gas shower head, which acts directly upon the
Si devices (open symbols). The data is from Klein et al. (2002). H2 diluent gas. In this design, the SiH4 is introduced into
(b) Diode quality factor (n) versus i-layer thickness (d) for the plasma separately so that it will not be Ôpredissoci-
HWCVD lc-Si devices (solid symbols) PECVD lc-Si devices atedÕ by the catalyzer. Using this novel design, an a-
(open symbols). The data is from Klein et al. (2002). Si:H single junction initial g of 10.1 % (in the p–i–n con-
figuration) has been achieved at an Rd of 28 Å/s, with i-
multichamber, load-lock capabilities, other groups have layer material improvements shown by both IR and Ra-
initiated HWCVD work. In particular, ENEA has al- man measurements (Komoda et al., 2004). In this cell,
ready purchased and installed a fully computer control- the device Jsc is 16.4 mA/cm2 for a 0.3 lm i-layer thick-
led, multichamber system dedicated to HWCVD, and its ness. A similar advancement was made in their lc-Si
research is aimed at single junction lc-Si cells fabricated cells, where the g achieved using their standard VHF
technology (7.8%, 2 Å/s) was extended to 8.1% at 12
Å/s with the addition of a catalyzer in the configuration
mentioned above (Hakuma et al., 2004). In the latter
cell, no Ôrear sideÕ texturing was used, as natural textured
light confinement was claimed due to the (220) preferen-
tial crystallite orientation. Another combination tech-
nique with solar cell implications has been examined at
Iowa State University (Ring et al., 2004). Here,
HWCVD has been combined with soft (<10 eV) He
ion bombardment generated by a remote plasma; with
this geometry, most of the deposition (8–10 Å/s) came
from the hot wire, and only minimal deposition (1
Å/s) from the ECR plasma. Using this deposition com-
bination, a-Si:H and a-SiGe:H deposited at low Ts
(<275 C) exhibited narrower Urbach edges and im-
proved microstructure factors (R) compared to those
deposited by HWCVD alone, with the R for Ge improv-
Fig. 2. Fill factor versus light soaking time (van Veen, 2003) for ing the most. The thesis was that high Rd films need high
single junction n–i–p cells with the i-layer deposited from energy at the growing film surface to remove the excess
undiluted silane (s) and from H2 diluted silane (j). surface H. This is typically done by raising the film Ts,
936 A.H. Mahan / Solar Energy 77 (2004) 931–938

but these results show that this can also be accomplished

by moderate ion bombardment.
The second variation in HWCVD solar cell fabrica-
tion utilizes the benign nature of the process (lack of
energetic ion bombardment). In particular, Cal. Tech.
is using HWCVD a-SiN as both a passivation layer
and an anti-reflection coating for c-Si solar cells (Holt
et al., 2003). In this variation, an a-SiN HWCVD film
is deposited onto a diffused c-Si n+/cSi string ribbon
p+/Al back contact device structure. This composite
structure is then annealed which generates the top con-
tact (deposited on top of the a-SiN) and allows H from
the SiN layer to diffuse into the c-Si bulk to enable defect
passivation while continuing to serve as an antireflection
coating. A 12.4% g has been achieved using this ap- Fig. 4. Schematic drawing of a HWCVD deposition Ôshower-
proach with HWCVD a-SiN, and a higher g (larger headÕ (Matsumura et al., 2003). The small vertical lines
Jsc) is projected by improving the uniformity of the represent short filaments, and the dark squares gas inlets for
a-SiN deposition. the (primarily) SiH4 gas. Uniform deposition over a 90 cm ·
Finally, recent progress in large area deposition scale 40 cm area has been demonstrated.
up of HWCVD needs to be highlighted. Not mentioned
in the chronology of the three HWCVD cell fabrication cently, deposition over much larger areas has been dem-
groups was the evolution of deposition area. Initially, all onstrated by Anelva, utilizing a modified version of their
groups investigated HWCVD deposition using very ÔshowerheadÕ. A schematic picture of this ÔshowerheadÕ is
small substrate areas (<20–30 cm2). Kaiserslautern pre- shown in Fig. 4. With this apparatus, uniform deposi-
sented the first large area study (experimental and theo- tion over a 90 cm · 40 cm area has been achieved, and
retical) using a filament array, and found that with the in addition a high gas utilization of SiH4 has been dem-
proper filament arrangements and source gas introduc- onstrated (Matsumura et al., 2003).
tion, excellent solar cell uniformity over a 20 cm · 20
cm area was achieved (Ledermann et al., 2001). The very
narrow distribution of gs for a-Si:H devices over this 4. Conclusions
area is shown in Fig. 3. The use of 2 air breaks in device
fabrication is one factor limiting the cell gs. Later, Solar cell fabrication using the HWCVD technique
MVSystems showed that, with a substrate that Ôoscil- has been summarized. A large part of this presentation
latedÕ under a single filament, a similar film uniformity has summarized chronologically the activities of three
was achieved (Morrison and Madan, 2000). More re- groups with a long history of HWCVD solar cell fabri-
cation. All groups followed similar research and devel-
opment paths. Early work was on a-Si:H cells, with
the doped layers deposited by PECVD. Many high Rd
devices were fabricated in this time period, with the n–
i–p devices on flat SS achieving state of the art perform-
ance (g > 7%). Within five years, all groups had multi-
chamber, load-locked systems dedicated to HWCVD
device fabrication, and during this period different types
of devices were investigated. At the present time lc-Si n–
i–p devices have been the main topic of research, and a
fourth group has fabricated a state of the art single junc-
tion cell with a 9.4% g. Several groups have successfully
demonstrated very large area, uniform deposition capa-
bilities, and machines are being readied for potential
commercial production.

Fig. 3. P–i–n a-Si:H solar cell uniformity (Ledermann et al.,

2001) over a 20 cm · 20 cm deposition area. The low solar cell Acknowledgments
gs are due primarily to the air breaks which were introduced at
the n–i and i–p interfaces; these were necessary to enable solar The author thanks many of his colleagues for sharing
cell fabrication with this large area HWCVD apparatus. their unpublished and/or recent results with him, as well
 A.H. Mahan / Solar Energy 77 (2004) 931–938 937

as discussing candidly their new research ideas and Lill, M., Schroeder, B., 1999. Preparation of amorphous
directions. This work is funded by the US DOE under hydrogenated silicon-germanium material and solar cells
subcontract number DE-AC36-99-GO10337. using the thermocatalytic chemical vapor deposition. Appl.
Phys. Lett. 74, 1284.
Mahan, A.H., Schropp, R.E.I., 2001. Preface. Thin Solid Films
395, ix.
References Mahan, A.H., Vanecek, M., 1991. A reduction in the Staebler–
Wronski effect observed in low H content a-Si:H films
Bauer, S., Schroeder, B., Herbst, W., Lill, M., 1998. A deposited by the hot wire technique. AIP Symp. Proc., vol.
significant step towards fabrication of high efficient, more 234, p. 195.
stable a-Si:H solar cells by thermo-catalytic CVD. Proc. Mahan, A.H., Carapella, J., Nelson, B.P., Crandall, R.S.,
Second World Conf. on PV Solar Energy Conversion, Balberg, I., 1991. Deposition of device quality, low H
Vienna, p. 363. content amorphous silicon. J. Appl. Phys. 69, 6728.
Bouree, J.E., 2003, private communication. Mahan, A.H., Iwaniczko, E., Nelson, B.P., Reedy Jr., R.C.,
Feenstra, K.F., Rath, J.K., van der Werf, C.H.M., Hartman, Crandall, R.S., Yang, J., Guha, S., 1996. Hot wire deposited
Z., Schropp, R.E.I., 1998. Thermally stable amorphous hydrogenated amorphous silicon solar cells. Proc. 25th
silicon n-layers for application in n–i–p solar cells with hot- IEEE PVSC, Washington, DC, p. 1065.
wire photoabsorbers. Proc. Second World Conf. on PV Mahan, A.H., Reedy Jr., R.C., Iwaniczko, E., Wang, Q.,
Solar Energy Conversion, Vienna, p. 956. Nelson, B.P., Xu, Y., Gallagher, A.C., Branz, H.M.,
Gallagher, A.C., 2001. Some physics and chemistry of hot-wire Crandall, R.S., Yang, J., Guha, S., 1998. H out-diffusion
deposition. Thin Solid Films 395, 25. and device performance in n–i–p solar cells utilizing high
Hakuma, H., Niira, K., Senta, H., Nishimura, T., Komoda, M., temperature hot wire a-Si:H i-layers. MRS Symp. Proc., vol.
Okui, H., Aramaki, K., Okada, Y., Tomita, K., Higuchi, 507, p. 119.
H., Arimune, H., 2004. Microcrystalline-Si solar cells by Mahan, A.H., Xu, Y., Iwaniczko, E., Williamson, D.L.,
newly developed novel PECVD method at high deposition Nelson, B.P., Wang, Q., 2002. J. Non-Cryst. Sol. 299–302,
rate. Proc. Third WCPEC Conf., Osaka, p. 1796. 2.
Holt, J., Goodwin, D.G., Gabor, A.M., Stavola, M., Atwater, Matsumura, H., 1986. Catalytic chemical vapor deposition
H.A., 2003. Hot-wire chemical vapor deposition of high (CTL-CVD) method producing high quality hydrogenated
hydrogen content silicon nitride for solar cell passivation amorphous silicon. Jpn. J. Appl. Phys. 25, L949.
and anti-reflection coating applications. Thin Solid Films Matsumura, H., 2001. Summary of research in NEDO Cat-
430, 37. CVD project in Japan. Thin Solid Films 395, 1.
Iwaniczko, E., 2003, private communication. Matsumura, H., 2003, private communication.
Iwaniczko, E., Mahan, A.H., Yan, B., Gedvilas, L.N., Will- Matsumura, H., Umemoto, H., Izumi, A., Masuda, A., 2003.
iamson, D.L., Nelson, B.P., 2003a. Deposition of device Recent progress of Cat-CVD research in Japan––bridging
quality lc-Si films and solar cells at high rates by HWCVD between the first and second Cat-CVD conferences. Thin
in a W filament regime where W/Si formation is minimal. Solid Films 430, 7.
MRS Symp. Proc. 762, p. 643. Morrison, S., Madan, A., 2000, Deposition of device-quality
Iwaniczko, E., Xu, Y., Schropp, R.E.I., Mahan, A.H., 2003b. amorphous and microcrystalline silicon films with a new
Microcrystalline silicon for solar cells deposited at high Ôhot wireÕ CVD technique. Proc. 28th IEEE PV Spec. Conf.,
rates by hot-wire CVD. Thin Solid Films 430, 212. Anchorage, p. 837.
Klein, S., Finger, F., Carius, R., Rech, B., Houben, L., Nelson, B.P., Iwaniczko, E., Schropp, R.E.I., Mahan, A.H.,
Luysberg, M., Stutzmann, M., 2002. High efficiency thin Molenbroek, E.C., Salamon, S., Crandall, R.S., 1994.
film solar cells with intrinsic microcrystalline silicon pre- Amorphous silicon solar cells incorporating hot-wire depos-
pared by hot wire CVD. MRS Symp. Proc. 715, A26.2. ited intrinsic layers. Proc. 12th European PVSEC, Amster-
Klein, S., Finger, F., Carius, R., Dylla, T., Rech, B., Grimm, dam, p. 679.
L., Houben, L., Stutzmann, M., 2003. Intrinsic microcrys- Papadopoulos, P., Scholz, A., Bauer, S., Schroeder, B.,
talline silicon prepared by hot-wire chemical vapour depo- Oechsner, H., 1993. Deposition of device quality a-Si:H
sition for thin film solar cells. Thin Solid Films 430, 202. films with the hot-wire technique. J. Non-Cryst. Sol. 164–
Komoda, M., Niira, K., Senta, H., Nishimura, T., Hakuma, H., 166, 87.
Okui, H., Aramaki, K., Okada, Y., Tomita, K., Higuchi, Rath, J.K., van Zutphen, A.J.M.M., Meiling, H., Schropp,
H., Arimune, H., 2004. Amorphous silicon solar cells at R.E.I., 1997. Application of hot wire deposited intrinsic
high deposition rates using newly developed PECVD. Proc. poly-silicon films in n–i–p cells and TSTs. MRS Symp.
Third WCPEC Conf., Osaka, p. 1733. Proc., vol. 467, p. 445.
Kupich, M., Grunsky, D., Kumar, P., Schroeder, B., 2004. Ring, M., Dalal, V.L., Muthukrishnan, K.K., 2004. Properties
Preparation of microcrystalline single junction and amor- of a-(Si,Ge):H films grown using combined hot wire-ecr
phous–microcrystalline tandem silicon solar cells entirely by plasma processes. J. Non-Cryst. Sol. 338–340, 61.
hot-wire CVD. Solar Energy Mater. Solar Cells 81, 141. Schroeder, B., 2003. Status report: solar cell related research
Ledermann, A., Weber, U., Mukherjee, C., Schroeder, B., 2001. and development using amorphous and microcrystalline
Influence of gas supply and filament geometry on the large- silicon deposited by HW(Cat)CVD. Thin Solid Films 430, 1.
area deposition of amorphous silicon by hot-wire CVD. Schroeder, B., Weber, U., Seitz, H., Ledermann, A., Mukher-
Thin Solid Films 395, 61. jee, C., 2001. Current status of the thermo-catalytic
938 A.H. Mahan / Solar Energy 77 (2004) 931–938

(hot-wire) CVD of thin silicon films for photovoltaic van Veen, M.K., Schropp, R.E.I., 2001. Higher efficiency of n–
applications. Thin Solid Films 395, 298. i–p solar cells by hot-wire CVD at moderate temperatures.
Schropp, R.E.I., 2002. Advances in solar cells made with hot wire MRS Symp. Proc., vol. 664, p. A11.2.
chemical vapor deposition (HWCVD): superior films and Wang, Q., Iwaniczko, E., Xu, Y., Gao, W., Nelson, B.P.,
devices at low equipment cost. Thin Solid Films 403–404, 17. Mahan, A.H., Crandall, R.S., Branz, H.M., 2000. Efficient
Schropp, R.E.I., 2003, private communication. 18 Å/s solar cells with all silicon layers deposited by hot-wire
Schropp, R.E.I., Hartman, G.Z., Franken, R.H., Jimenez chemical vapor deposition. MRS Symp. Proc., vol. 609,
Zambrano, R.J., Li, H., Loeffler, J., Rath, J.K., Rusche, p. A4.3.
M., Stolk, R.L., van Veen, M.K., van der Werf, K.H., 2004. Wang, Q., Iwaniczko, E., Yang, J., Lord, K., Guha, S., 2001.
First hot-wire deposited triple junction thin film solar cell. High quality amorphous silicon germanium alloy solar cells
Proc. Third WCPEC Conf., Osaka, p. 1847. made by hot-wire CVD at 10 Å/s. MRS Symp. Proc., vol.
van Veen, M., 2003. Tandem solar cells deposited using hot- 664, p. A7.5.
wire chemical vapor deposition. Ph.D. Thesis, Utrecht Weismann, H., 1979. Method of producing hydrogenated
University, Faculteit Natuur-en Sterrenkunde, p. 69. amorphous silicon film. US Patent 4,237,150.