Note: Descriptions are shown in the official language in which they were submitted.
IMPROVED SOLAR CELL
This invention relates to solar or photovoltaic cells
and more particularly to an improved solar c~ll uni~uely
designed for optimum efficiency.
Fossil fuels have in the past provided the bulk of
the world's energy needs. However, as the price of fossil
fuels has increased and their supply decreased, increasing
attention has been directed toward the development of
alternate energy sources. One such alternate energy source
Ls the solar cell, which directly converts the energy from
the sun into usable electrical power.
Solar cells have been long used in prior-art solar-
energy systems. In these systems, the typical solar cell
consists of a large area p-n junction formed in a wafer of
monocrystalline material such as silicon. The junction is
~15 formed parallel to the upper surface of the cell, and this
upper surface receives incident radiation from the sun, to
produce current flow across the p-n ~unction in a well-known
18 manner. These conventional solar cells suffer from many
~1
.
.. , . - ,, ~ .
disadvantages. Such disadvantages have been well documented
and include large series resistance, inefficient operation
at high concentrations of incident light, and the need for a
contact grid which is formed on the upper surface of the cell,
thereby reducing the cell area available to receive incident
radiation.
Vertical-junction solar cells have been developed in an
attempt to circumvent the various problems encountered with
conventional solar cells. One type of vertical-junction solar
cell is fabricated from stacks of silicon wafers which have
been appropriately doped to form p+ and n+ surface layers on
opposite sides of the wafers and the wafers are then stacked
and sintered together. The wafers are sliced into segments
to create a plurality of solar cells with p-n junctions
normal to the cell surface. With silicon, the resultant
vertical-junction solar cell has the potential for more effi-
cient high-intensity operation, does not require the contact
grid and also provides low series resistance. However, these
~ cells, as they have been commerically available, typically
2n ha~e a measured efficiency of only 8~ at one sun.
Our U. S. patent no. 4,042,417 issued August 16, 1977,
and entitled "Photovoltaic Structure Including a Lens
Structure", discloses a technique for achieving greater
efficiency in such vertical junction solar cells. This
technique includes the use of a cylindrical lens which focus-
es the solar radiation into a narrow beam incident on the
active, light-receiving area of the solar cell at an optimum
region, ad~acent to, but offset from the vertical junction
plane. This focusing technique achieves approximately a 2:1
improvement in solar cell efficiency. Improvement results
'
, , , ' ' .: '
.. . .
~o~
because incident radiation is not lost on the "dead space"
of the cell's p-n junction which consists of the p+, n~
regions and the electrode therebetween, and because incident
radiation is focused at an optimum region near the p-n
junction. Focusing incident radiation at this optimum region
rather than further away from the junction creates carriers
with a greater probability of being collected than carriers
created at a point further away from the junction. Notwith-
standing the dramatic improvement in cell efficiency achieved
by this invention, such imp~ovement required the utilization
of a lens array which increased cost and decreased reliability
due to the problem of lens degradation over a long period of
time.
A further improvement in the area of solar-cell fabrica-
tion is disclosed in U. S. Patent No. 4,110,122 issuedAugust 29, 1978, and entitled "High-Intensity Solid State
SoIar Cell". In this invention, a plurality of solar-cell
units are fabricated from a common substrate, and the body
material of each unit advantageously has the same positional
relation in the finished cell as existed in the original sub-
state. This technique provides solar-cell units having iden-
tical characteristics as to materiall orientation and physical
properties, as well as units having a predetermined fixed
positional relation to eàch other, thereby increasing the
accuracy of focusing arrangements. Again, however, the solar
ceIl of this invention requires a focusing lens for optimum
performance.
It is therafore, an object of an aspect of this invention
to achieve improved solar-cell performance without requiring
30~ special focusing by lenses or other means.
!
.,. ' ' ',' ' .
' ~ . . . ' ' ',, . '' '' ' " ~ ' , '
.,
. . ' , . . ',' ' .''. , , , ' " ~, . .
It is an object of an aspect of this invention to
provide solar-cell Wlits having identical physical character-
istics, and efficient operation when connected in series.
It is an object of an aspect of this invention to
provide an improved solar-cell which operates at high
incident light intensity.
It is an object of an aspect of this invention to
provide an improved solar-cell having greatly reduced "dead
space" when compared to prior art vertical-junction solar-
cells.
It is an object of an aspect of this invention to
provide an improved solar-cell in which the ra-tio of "dead `
space" in the cell to active area in the cell is greatly
reduced.
It is an object of an aspect of this invention to
co~cisely match the dimensions of the cell units and the
carrier diffusion lengths to thereby achieve high collection
efficiency.
It is an object of an aspect of this invention to
provide a solar-cell which will provide relatively large
amounts of electrical energy without requiring focusing
lenses of the character indicated.
An object of an aspect of this invention is to
provide an improved array of solar-cell units presenting an
overall exposure area upon which solar energy may be con-
oentrated in the order of at least 100 suns, to produce an
~; electrical output at relatively great efficiency.
In accordance with an aspect of the invention
individual solar-cell units are formed from a common sub-
strate, and the body material of each of the units formedrom the substrate advantageously has the same spaced relation
,, , , , . : : ~'
to the body material o~ each of the other units as existed
in the original substrate. Each unit has spaced elongate
sidewalls, and the "dead space" area between adjoining side-
walls of adjacent cells is made substantially smaller than
the active area between the opposite sidewalls of each in~
dividual unit. Moreover, the distance between the opposite
sidewalls of each unit (i.e., the width of the active area)
is concisely limited to a predetermined optimum distance
(related to minority--carrier diffusion length, including
surface recomblnation effects) such that radiation incident
at any point on the active area is incident at a point spaced
from at least one of the unit sidewalls by no more than the
predetermined optimum distance. Limiting the width of the
active area in this manner ensures that light is always in-
cident at a point near the p-n junction to thereby create
carriers with~a high probability of being collected. Reducing
the "dead space" area while correctly choosing the width of
the active area provides improved cell performance without
requiring focusing lenses.
In accordance with another aspect of this invention
there is provided a semiconductor solar-cell array comprising
a plurality of spaced, elongate, parallel units formed from
a common substrate, the body material of each of said units
being comprised of a first conductivity-type and having the
same spaced relation to the body material of other of said
units as in the original substrate from which they are formed, ~ -
each unit having upstanding sidewalls, and having there-
between an upper surface adapted for exposure to receive in-
cident radiation and a lower surface, adjacent sidewalls of
adjacent units having therebetween a first space portion
extending from the lower surface to a point close to but short
- 4a -
of the upper surface and a second substantially smaller space
portion, connected to the first space portion and extending
from said point to the upper surface, at least one sidewall
of each unit including a localized reg.ion of a second con-
ductivity-type, and ohmic connections extending between the
second conductivity-type region of one unit and a first con-
ductivity-type region of an adjacent unit.
In accordance with another aspect of thi.s invention
there is provided a semiconductor solar-cell array comprising
; 10 h plurality of spaced, elongate, parallel units formed from
a common substrate, the body material of each of said units
being comprised of a first conductivity-type and having the
` same spaced relation to the body material of other of said
units as in the original substrate from which they are formed,
: each unit having upstanding sidewalls, and having there-
between an upper surface adapted for exposure to receive
incident radiation and a lower surface, adjacent sidewalls
of adjacent units being more closely spaced at juncture with
. said upper surface than their spacing at juncture with said :~
lower surface, at least one.sidewall of each unit including
a region of a second conductivity~type, and ohmic connections
extending between the second conductivity-type region of one
unit and a first conductivity-type region of an adjacent
wall~
In accordance with another aspect of this invention
there is provided a semiconductor solar-cell array comprising
spaced, elongate, parallel units formed from a common sub-
strate, the body material of each of said units being com-
prised of a first conductivity-type and having the same
spaced relation to the body material of other of said units
as in the original substrate from which they are formed, each
- 4b -
~0~
unit having upstanding sidewalls, and having therebetween an
upper surface adapted for exposure to receive incident
radiation and a lower surface, adjacent sidewalls of adjacent
units being more closely spaced at juncture with said upper
surface than their spacing at junction with said lower
surface, corresponding first sidewalls of said units each
including a region of a second conductivity-type at least
at a point close to but short of the upper surface, corres-
ponding second sidewalls of said units each including a
region of said first conductivity-type at least at a point
in opposed adjacency to said second conductivity-type point,
and an ohmic connection between said first and second
conductivity-type points at each inter-unit space.
In accordance with another aspect of this invention
there is provided a semiconductor solar-cell comprising a
plurality of spaced, elongate, parallel units formed from a
common substrate, the body material of each of said units
being comprised of a first conductivity-type and having the same
spacedrelation to the body material of other of said units
: 20 as in the original substrate from which they are formed,
each unit having spaced, elongate sidewalls and adjacent
sidewalls of adjacent units haviny therebetween a dead :
space area comprised of an upper portion and a lower portion,
said upper portion of the dead space area being substantially
narrower than sa~d lower portion of the dead space area, each
: sidewall of each unit containing a region of a second con-
ductivity-type, and connection means including a conductive
element having ohmic contact with said second conductivity-
type region of one unit and ohmic contact with a first con-
30 ductivity-type region of an adjacent unit. ~ .
In accordance with another aspect of this invention
- 4c -
.
there is provided a semiconductor solar-cell, comprising a
plurality oE spaced, elongate, parallel units formed from a
common substrate, the body material of each of said units
being comprised of a first conductivity-type and having the
same spaced relation to the body material of other of said
units as in the original substrate from which they are formed,
each unit having upstandi~g sidewalls wherein upper edges of
the sidewalls of each unit are spaced a first fixed distance
from each other by an elongate upper surface extending between
the upper edges of the sidewalls and wherein lower edges of
the sidewalls are spaced a second fixed distance from each :~
other by an elongate lower surface extending between the
lower edges of the sidewalls, said first fixed distance
being greater than said second fixed distance, said upper
surface being adapted or exposure to receive incident
radiation, and said first fixed distance being limited to
twice a predetermined optimum distance, whereby radiation
incident at any point on said upper surface is incident at
a point spaced from at least one of the unit sidewalls by
~o no more than said predetermined optimum distance, each unit
further comprising a region of a second conductivity-type
in at least one of the sidewalls of each unit, and separate
conductive connections establishing ohmic contact between
the second conductivity-type region of one unit and the
first conductivity-type region of another unit.
In accordance with another aspect of this invention
there is provided a semiconductor solar-cell comprising a
plurality of spaced, elongate, parallel units formed from a
- common substrate, the body material of each of said units
being comprised of a first conductivi$y-type and having the
same spaced relation to the body material of other of said
- 4d -
.
, ~ ~
units as in the original substrate from which they are formed,
each unit having upstanding sidewalls and h~ving therebetween
an upper surface adapted for exposure to receive incident
radiation and a lower surface, said upstanding sidewalls
being arranged to provide between the adjacent upper surfaces
of adJacent units a space which is substantially smaller than
the corresponding space between the adjacent lower surfaces
of the adjacent units, each unit further comprising a region
of a second conductivity-type in at least one sidewall of
each unit, and separate conductive connections establishing
ohmic contact between the second conductivity-type region of
one unit and the first conductivity-type region of another
unit.
In accordance with another aspect of this invention
there is provided a semiconductor solar-cell, comprising a
plurality of spaced, elongate, parallel units formed from a
common substrate, the body material of each of said units
being comprised of a first conductivity-type and having the
same spaced relation to the body material of other of said ~ ~
20 units as in the original substrate from which they are form- ~;
ed, each unit having upsta~ding sidewalls and having there-
between an upper surface adapted for exposure to receive
incident radiation and a lower surface, each unit sidewall
comprised of a localized region of a second conductivity-type,
said second conductivity-type region extending along the
sidewall from said lower surface to a point close to but
shoxt of said upper surface, whereby said second conductivity-
type region is buried beneath the upper surface, and ohmic
connections ~xtending beneath said second conductivity-type
region of one unit and said first conductivity-type region
o~ an adjacent unit.
'
- 4e -
.. . ..
, ~ .
.
, ., . . .: . . , - . , , ...... , .. , .:, . . . . .
L~
To the accomplishment of the above and to further
objects that may hereinafter appear the present invention
relates to an improved solar-cell as defined in the appended
claims and as described in the following specification when
considered in conjunction with the accompanying drawings,
in which:
Fig. 1 is a schematic diagram as viewed in cross-
section of a multiple-unit fragment of a solar-cell array
previously designed by the instant inventors;
Fig. 2 is a similar sectional view of a substrate
showing a first wide groove cut into the substrate as a
first procedure in the construction of an improved solar-
cell of the present invention;
- , , : : .,, :, ~ : .
Figs. 3 and 4 are further sectional views of the sub~
strate of Fig. 2, showing (in Fig. 3) p~ and n~ regions
diffused into the substrate, and showing (in Fig. 4) a
final narrow groove cut into the substrate to form a
5 completed improved solar cell;
Fig. 5 is an enlarged sectional view of one area of
the improved solar cell;
Fig. 6 is a plot of the short-circuit current
response of one particular vertical-junction solar-cell
10 unit of the array of Fig. 4;
Fig. 7 is a view similar to Fig. 5 to show a modifi-
cation; and ~ ~-
Figs. 8 and 9 are views similar to Fig. 4 to show
further difications.
The embodiments of the invention described hereinbelow
involve a solar cell which has an n-type silicon substrate. -~
It is to be understood, however, that the solar cell of the
invention may also be implemented with a substrate of p-type
polarity, in which case the polarity of the other regions in
20 the cell would be reversed, with n replaced ~y p, n+ replaced
by p+ and so on~ It will also be understood that o~her types ~ -
of semiconductor material may also be employed and, that a
heterojunction, as well as the homojunction structure des-
cribed may be employed.
Before turning to a detailed description of the inven- -
tion, it is helpful in achieving a greater understanding of
theinvention's unique advantages to describe the solar cell
shown in Fig. 1. Fig. 1 il~lustrates a solar-cell arrange-
ment which is one particular embodiment of a solid-state
30 solar cell described in our U. S. patent No. 4,110,122.
-6~
: . . . . .
:, . . .
More particularly, the solar cel:L illustrated in Fiy. 1
consists of a series o~ individual unit solar cells 105,
each o~ which is formed from the same single wafer of
semiconductor material, the units being subsequently
connected in series, or in series-connected subunits
which are later connected in parallel.
Each of the unit solar cells 105, illustrated in
Fig. l, is comprised of an n-type substrate 107. The
unit solar cells 105 are separated from one another by
grooves lO1 which are formed by anisotropic etching to
create grooves having straight parallel walls extending
completely through the common substrate. P+ regions 103
are formed along the wall of eac~l groove and an n+ region
102 is formed along the lower surface of each unit cell~
An oxide layer 108 is formed on the lower surface of each
cell, and a passivation and anti-reflective coating 106
may be formed on the upper surface of each cell. The
grooves may be subsequently filled with an insulating
material lO9 which may be epoxyl glass or other suitable
insulating material to achieve electrical isolation
between the cell units. Alternatively, the grooves may be
left partially or completely open, with other means being
provided for structural support and maintenance of subunit
alignment. Cell construction in a series-connected group
of units in an array is completed by interconnecting the p+
reyions of each cell with a first connection schematically
indicated at 100, the n+ region of each cell being connected
to the p+ regions of the next-adjacent cell via a second
connection, schematically indicated at 104. Series-parallel
combination connections of unit groups with the array are also
": .' ,' ' ' : ,
,. ~
: :
~1~8~
possible; such constructions are the subject of another of
our applications, and are not necessary to a description of
the present invention.
Light is incident on th~ active region 110 of each
cell, and the active region is defined as the area of the
substrate surface available to receive incident radiation.
The incident light causes a ~low of carriers across the p-n
junction formed by the p-~ regions and substrate, to the n+
regions, and from the latter to the p+ regions in succeeding
solar cells via connection 104. The conduction of current
proceeds in series across the individual solar cells to a
current collector (not shown). The connections indicated
schematically in Fig. 1 are, of course, illustrations of
actual connections between adjacent cells wherein such
actual connections may be formed by known selective etching
and metalization techniques, or other microelectronic
techniques.
The solar cell shown in Fig. 1, although providing a
marked improvement over prior-art solar cells, requires a
special focusing arrangement to operate at maximum eficiency.
Such a focusing lens arrangement is described in our said
U. S. patent No. 4,042,417. This lens arrangement includes a
plurality of lens elements which focus the incoming radiation
into a plurality of narrow beams which are directed to be
incident on the upper surface of ~he cell at a location adja-
cent to, but offset from the plane of each p-n junction.
Such a focusing arrangement improves cell performance for two
reasons. Firstl incident radiation is not lost on the large
"dead space" area 111 indicated in Fig. 1 which contributes
essentially nothing to cell output. As indicated, this area
is comprised
~,
.
of the adjoining p~ regions of adjacent cells and the
material therebetween. Second, incident radiation is
not directed to just any point on the active area 110
but rather the incident radiation is directed to an
optimum region on the active area which is at least
generally adjacent to but offset from the plane of the
p-n junction. Direc-ting the incident radiation to this
optimum point improves cell performance because the
radiation focused at that point (or closer to the junction)
creates carriers with a much higher probability of being
collected than carriers created at a location which is
further from the junction.
Extensive investigation by the instant inventors
has determined where this optimum point should lie on the
active area in order to achieve maximum performance in a
silicon vertical junction solar cell. The result of these
investigations has been reported in an article entitled
"Improved Performance of Solar Cells for High Intensity
Application", published in "The Conference Record of the
Twelfth IEEE Photovoltaic Specialists Conference", held
November 15-18, 1976. As reported in this article, the
instant inventors were concerned with the nature of the
response of vertical-junction solar cells to focused incident
light. ~ccordingly an array of vertical-junction solar cells
2~ was obtained from Semicon, Inc. of ~urlington, Massachusetts.
Contacts were made to electrodes on either side of an individual
cell, and the cell array was mounted on a micrometer stage. A
light spot, provided by focusing the beam from a ~e-Ne laser,
and having a diameter of approximately 1 to 2 mils was directed
normal to the upper surface of the array, and the micrometer
.
staye was moved in 0.5-mil increments to e~fectively scan
the ~ocused laser beam across the s~lected individual cell.
The short-circuit cell current ISC was measured for each
location of the inci~ent spot using an external variable
load resistor for the measurements.
The solid curve A' shown in Fig. 6 is a plot of the
measured short-circuit current output, ISC in milliamperes,
versus traverse position of the light spot, in mils from
the electrode (horizontal axis I). This curve illustrates
that cell output ISC is much greater in the n region adjacent
to but offse* ~rom the p-n junction, whose approximate location
is shown by legend at dotted line A. More particularly, the
solid-line plot in Fig. 6 illustrates that cell performance
shows a marked increase when light is incident 1 to 2 mils
from the p-n junction.
The decrease in ISC shown by the solid-line plot in
~ig. 6j near the junGtion location at A, but within the n
type region to the left of A, is attributable to the use of
a 1 to 2 mil diameter light spot to scan the cell surface.
A light spot of such diameter will extend into the "dead space"
area between adjacent cells as it nears the junction location.
The result is a "vignetting" effect, in that only a portion of
the llght spot is incident on the active area of the cell when
; the light spot is near the junction location, thereby causing
a reduction in the v~lue of ISC- It will be understood, of
course, that if a light spot of infinitesimal diameter could
be used, the solid-line plot of Fig. 6 would not fall off as
rapidly as shown in Fig. 6, but rather ISC would remain at or
near its maximum value until the infinitesimal-diameter light
spot were much closer to the ~unction location. As the use of
.
--10--
::
8!3~L9~
an infinitesimal-diameter light spot is experimentally
impossible, the plot in Fig. 6 is offered as practical experi-
mental proof that IsC does show a rnarked increase approximate-
ly 1 to 2 mils from the junction location. Also, the use of
5 a lens array (per said U. S. patent No. 4,042,417) produces a
focused beam having an approximate width of 1 to 2 mils and,
therefore, the use of a light spot of 1 to 2 mils diameter
represents a realistically attainable situation.
It should also be understood that the optimum unit-cell
width is dependent on the minority-carrier diffusion length
of the base material. A base material having different
minority-carrier diffusion lengths would change this distance
I accordingly, as would processing the base material to improve
f the carrier-diffusion length. Generally speaking, it is
desirable that the maximum distance that a carrier must travel
f to reach and be collected at the p-n junction shall be less
than the minority-carrier diffusion length. For the sample
used in connection with Fig. 6, the diffusion length (including
surface-recombination effects) was in the order of 1 to 2 mils.
Based on the foregoing, it can be seen that improved solar
cell performance can be achieved without the use of focusing
lens (a) if the "dead space" area can be drastically reduced
and, moreover, (b) if substantially all of the light falling on
the active area can be incident at a point which is no more than
1 to 2 mils from the p-n junction. The 1 to 2 mils distance is
optimum for this particular base materlal and, of course, this
distance would vary depending on the properties of the base
material used. The following description focuses on an improved
solar cell which advantageously provides the two elements, (a)
1, '
,
:,
and (b), defined above which are necessary Eor improved
solar-cell performance.
Turning now to a description of the invention, there
is illustrated in Fig. 2 a common n-type silicon substrate
204 from which will be created a plurality of individual
unit solar cells. These unit cells are all created from
common substrate 204, and thus the body material of each
- cell unit will advantageously have the same physical prop-
erties and crystallographic orientation in the finished
cell structure as existed in the original substrate. This
technique provides solar cell units having identical
characteristics as to material, orientation and physical
properties, a feature which is particularly important for
series-connected cell units. A procedure for fabricating
cells from a common substrate has been described in detail
in our U. S. patent No. 4,110,122~ Therefore, the details
of fabrication will not be repeated here, but it is under-
stood that fabrication techniques described in this patent
may be utilized in fabrication of the improved solar cell
of the instant invention.
Referring to Fig. 2, it can be seen that into sub-
strate 204 there has been etched a plurality of grooves 203.
Grooves 203 are not etched completely through substrate
204 as was done in Fig. 1, but rather are etched from the
25 lower surface of the substrate to a point close to but ~
short of the upper surface of the substrate. Note that ~.
the top of each groove i5 wedge-shaped, rather than flat~
This wedge-shaped groove, along with the vertical side-
walls, is a property of anisotropic etching of
-12-
. .
- - .- . ,
silicon having a ~ 110 > surface orientation; of course,
oth~r shaped groove bottoms may result, depending upon the
properties of the particular type of etchant used. If the
etching for the gxooves were continued all the way through
the silicon substrate the resultant structure would appear
identical to the structure shown in Fig. 1 with the grooves
completely cutting through the substrate.
Grooves 203 partially define individual n-type substrate
units 202 which are to be formed into completed solar-cell
units. ~he grooves can, of course, be etched into the substrate
at any desired interval to form the individual units. The
particular interval chosen is indicated as Wl in Fig. 2, and
as will be hereinafter described, Wl approximately equals
between 2 and 4 mils, or other specific dimensions consistent
with the minority-carrier diffusion length. Formed on the
upper surface of the substrate in Fig. 2 is an anti-reflective
coating 200 and formed on the lower surface is an oxide layer
201.
Subsequent to groove formation, p+ regions are formed
along the walls of each groove by diffusion, as is indicated
in Fig. 3 at 300. These p+ diffusions follow the groove
contours, thereby forming a p-n junction completely around
each groove between the p~ region and the n-type substrate.
N~ regions are also formed along the lower surface of each
unit solar c~ll at 301, in the same manner as described for
the case of Fig. 1. Subsequent to the diffusion process, a
plurality of unit solar cells 302 exist in the substrate,
requiring only final separation and interconnection to form
29 completed individual cell units.
-13-
. :- . ' ' :
Fig. 4 shows the completed improved solar cell.
Separation of each unit solar cell has been achieved
with the formation of very narrow grooves 410. These
narrow grooves are advantageously etched ~rom the upper
surface of the substrate to intersect the much wider
grooves 407 which were etched ~rom the lower surface
o~ the subs~rate. Cell separation, therefore, is achieved
with a two-step etching process. A first wide groove is
etched from the bottom of the substrate to a point close
to, but short o~ the upper surface of the substrate. A
second, substantially narrower groove is then etched from
the top of the substrate to intersect and thus open the
much wider groove. Alternatively, of course, the order
of etching the narrow and wide grooves could be reversed
lS or both yxooves might be etched concurrently. It is, of
course, understood that etching a single narrow groove
through the substrate would be extremely di~ficult in this
situation, in that the width to depth ratio for a narrow
groove would be typically in the order of 60:1 for two-step
etching, or 30:1 for two-sided simultaneous etching, it
being understood that the ratio could be much higher the
thicker the substrate used.
Cell construction is completed by filling the wide
and narrow grooves with an insulating material to providP ~-
~5 electrical isolation of the cells; alternatively, the
grooves may be left partially or completely open, with
other means being provided for structural support and
maintenance of subunit alignment. The insulating material
is indicated by shading within the groove regions 407 and
410. Interconnection of the various cells is achieved by
-14-
.
: ' ~ ' ' . ', .......................... ~
.
interconnecting the p-~ regions 403 of each unit cell,
using a first connection schematically indicated at ~08,
and by connecting the n~ region ~04 of each cell to the
p+ regions of tne next-adjacent cell, using the connection
schematically indicated at ~09; and as noted above, it
should be recognized that series~parallel combination
connections can alternatively be made. However, regardless
of the interconnection of cell units, the flow of carriers
through each subunit in response to incident light, is as
ordinarily understood for p+, n, n+ silicon cells. To
complete the description of Fig. 4, the anti-reflection
coating and the oxide layer previously described above are
indicated at 406 and 405, respectively.
It will be understood that the connections shown
schematically in Fig. 4 can be made by selective etching
and metalization techniques well known in the art. The
remaining portions of the unit solar cells may be fabricated
by known diffusion, anisotropic etching, thermal oxidation,
vacuum deposition, chemical vapor deposition, photolithographic
and other techniques, such as those used in the field of micro-
electronics and which are accordingly not otherwise disclosed
herein.
The unit solar cells 400 of Fig. 4 will be seen to be
generally similar to the unit solar cells 105 described in
connection with Fig. 1. However, each unit solar cell now
has a very small "dead space" area 402 in comparison to the
active area 401. This ratio of "dead space" area to active
area is much smaller than the ratio of "dead space" area to
active area in Fig~ 1, due to the formation of each cell
with a large groove etched from the lower surface of the
-lS-
'' -
substrate and a very small groove etched from the upper
surface of the substrate. Moreover, the "dead space"
area on the upper surface of each cell is further reduced
because it is now comprised of only the narrow grooves
(which may be filled with insulating material) and does
not include the p~ diffused area of the junction itself.
Rather the p-n junction is "buried" beneath the light
incident surface, thereby increasing the active area of
the cell. What has been achieved therefore is a dramatic
reduction in the "dead space" area with a corresponding
decrease in the ratio of "dead space" area to active area.
This reduction of the "dead space" area results in a direct
and immediate improvement in solar cell performance by
reducing the losses due to incident radiation falling on
the "dead space" area, and such improvement is achieved
without the use of focusing lenses.
Fig. 5 provides an expanded view of the large and
narrow grooves which separate the individual solar cells
of Fig. 4. As is shown therein, the buried p-n junction
lies at a distance "d" beneath the light incident surface
of the solar cell. Since about 80 percent of th~ absorbable
portion of the incident sunlight is absorbed within 1 mil of
the surface of the cell (for the assumed case of silicon),
"d" is advantageously chosen to be in the range of l mil.
~5 Fig. 5 also illustrates the path of light ray incident
on the upper surface of the solar cell. As is shown therein,
the incident ray penetrates into the solar cell and when
striking the groove boundary is partially reflected back into
the base material and partially transmitted into the wider
groove areaO The light lost due to the transmission into the
-16-
9~
wider groove can be shown to be on the order o~ several
percent or less of the total light incident on the upper
surface o~ the cell. This is to be compared with the
amount of light that would be lost into the much larger
"dead space" area that would result if the wide grooves
had been etched all the way through the substrate as
indicated in Fig. 1. By inspection, it can be seen that
light loss from such a large "dead space" area would be
much larger than the light loss resulting from transmission
into the wide groove illustrated in Fig. 5. In addition,
even that small loss can be further reduced by metallizing
or otherwise coating the sloping walls of the grooves, the
latter b~ing suggested by legend in ~ig. 5.
From the preceding discussion, it will be recalled
that light incident at a point adjacent to but offset from
the p-n junction results in improved solar-cell performance,
due to the ~eneration of carriers with a greater p~obability
o~ being collected than carriers generated at a point more
distant from the p n junction. This point of optimum light
incidence beyond which cell response drastically decreases
has been experimentally determined to lie no more than 1 to
2 mils from the p-n junction for the particular case illus-
trated in Fig. 6. With this experimental determination in
mind, the distance W3 between opposite sidewalls of each
individual solar-cell unit (i.e., the width o~ the active
area) is advantageously chosen to be between 2 and 4 mils.
Placing W3 in this range results in the midpoint of each
active area being no more than 1 to 2 mils lthe maximum
acceptable distance from the junction for incident light~ -
from the adjacent sidewalls of the unit solar cell and thus
from the adjacent p-n junction. ~herefore, all light
incident at approximately the midpoint of the active
area is incident at a point experimentally determined
to be suitable for good solar~cell performance. More-
over, referring to Fig. 6, it can be seen that lightincident on the solar cell surface at points closer to
the p-n junction will result in improved solar cell
performance when compared with light incident at locations
more remote from the p-n junction. From the foregoing,
therefore, it can be seen that solar-cell performance is
readily improved without the use of special unit-focusing
lenses by properly choosing the width of the active area.
An active area having a width of between 2 and 4 mils (for
a material like that shown in Fig. 6) results in all light
incident on the active area being incident at a point spaced
from at least one of the unit sidewalls (and, thus, from the
p-n junction) by not more than the predetermined acceptable
distance of 1 to 2 mils.
The dramatic improvement in solar-cell performance
ZO resulting from correctly choosing the width of the active
area is shown in Fig. 6 by superimposing a dotted-line plot
B' of short circuit-current response over the solid-line
plot A' previously discussed. More particularly~ the vertical
line at B in Fig. 6 illustrates a second junction location
~5 removed approximately 4 mils from the first junction location
A. Extending between these two junctions at A and B is an
active area indicated by legend in Fig. 6. This configuration
of two p-n junctions separated by an active area of approximately
4 mils is, of course, the configuration of one embodiment of the
instant invention discussed above. The dotted-line plot B' of
.
ISC shown in Fig. 6 is merely the same plot illustraked
as a solid plot A' in Fig. 6 but the dotted line plot is
reversed to correspond with the existence of a p-n junction
at location B. The relationship of these two plots of ISC
can be readily understood by reference to horizontal axis
II illustrated in Fig. 6 which shows ISC as a function of
the position of the light spot in mils from a junction
location. The numerical legend for curve A' is labeled
curve A' on horizontal axis II while the numerical legend
for curve B' is labeled curve B' on horizontal axis II.
The combination of curves A' and B' in Fig. 6
dramatically illustrates the improvement in solar-cell
performance resulting from choosing the width of the active
area to be equal to 4 mils. An active area of this width
lS ensures that all light incident on the active area is incident
at a point adjacent to but offset from the junction location,
whereby the probability of carrier collection is markedly
improved. It is, of course, apparent that the width of the
active area could vary in accordance with the instant invention
and then the positional relationship between curves A' and B'
would change in response thereto. Also, curves A' and B' do
not illustrate how ISC increases near the respective p-n
junctions.
~ssuming Wl equal to 4.0 mils, it is of interest to
assign approximate numbers to the remaining dimensions in
Fig. 5 in order to calculate the dramatic increase in active
area achieved through use of the narrow-groove fabrication
technique. More particularly, let W2 equal 0.1 mil and W4
equal 0.5 mils, W3 is then, of course, equal to 3.9 mils.
With these assigned values~ it is readily calculated that
--19-- :
, .
.. . :: . - ~. ,
,
~L~3~
the percentage reduction in acti~e area due to the
narrow-groove fabrication technique is equal to
(0.1/3.9) X 100 or 2.55~ while the percentaye reduction
in active area resulting from etching the wide groove
completely through the substrate is equal to (0.5/3.9)
X 100 or 12.8%. The narrow-groove fabrication technique
results in a dramatic improvement in increasing the active
area of each unit solar cell.
In conclusion, solar cell performance is improved
by constructing solar cells which differ markedly from
prior-art solar cells in two ways. First, the "dead space"
region is drastically reduced by forming a first large
groove from the lower surface of the cell and a second
very narrow groove from the upper surface of the cell.
Second, the active area of each cell is advantageously
chosen to be between 2 and 4 mils such that all light
incident on the active area is incident at a point spaced
from at least one of the unit sidewalls by not more than
a predetermined distance. ~hese two techniques result in
improved cell performance without requiring the use of
special unit-focusing lenses.
Fig. 7 will be recognized as being structurally the
same as Fig. 5, except for the fact that the p-n junctions
are no longer buried. Thus, in Fig. 7, the two groove
regions 410 and 407 will have been fully etched and defined
prior to diffusion to form the p+ regions which now extend
all the way to the antireflection coating at the exposure
surface of the substrate. Preferably, the diffusion step
is of short duration, whereby the p-n junction is relatively
close to the surface of the groove wall. For example, for
' ~' ' ' "
"', ' ~
silicon substrate, a p-n junction plane suitable ~or high-
intensity exposure may be established about 0.1 rnil beneath
the treated groove surface; and if a metal coating (e.g.,
electroless nickel) is applied, the p+ region may be even
more shallow, approaching 0.01 mil.
Thus, in Fig. 7, at the exposure sur~ace, the p-n
junction will be exposed at each sidewall of the narrow
grooves ~10, thereby reducing the usefully effective
exposure area of each cell unit, as suggested by the
slightly reduced "active-area" designation W3' and by the
slightly enlarged "dead space" designation ~2' in the
legends of Fig. 7. Nevertheless, ~or the case of p-n
junctions formed at such shallow depth in the substrate
material, the performance loss compared to the buried-
junction configuration of Figs. 4 and 5 can be small and,indeed, relatively unimportant. Certainly, manufacture
is simplified by performing the etching steps prior to the
p+ di~fusion, and the operational improvement over the
configuration of Fig. 1 is still so substantial as to
enable direct exposure to high-intensity light wi~hout
requiring a special unit-lens configuration as in said
U. S. patent No. 4,042,417.
Figs. 8 and 9 depict further modification of the
in~ention, such modi~ication being illustrated in appli-
cation to "buried" p-n junctions in Fig. 8, and in
application to "non-buried" p-n junctions in Fig. 9.
Thé advantage of these embodiments lies in the vastly
simplified problem of series-connecting adjacent cell
units 800 (900), arising from the fact that p-n junctions
are produced by p+ diffusion to one sidewll only 811-812
-21-
.
.. . . ..
. : . . :. ,
~0~
t911-912) for each successive groove 807 (907);
similarly, n+ regions are formed only in the other
sidewall 813-814 (913-914), for each successive groove
807 (907). In Fig. 8, the indicated p+ and n+ regions
terminate short of the radiation-exposure surface of
the solar-cell array, p-n junctions being thus buried
beneath said surfacei and in Fig. 9, the p+ and n+
regions extend to the exposure surface, but the number
of p-n junctions at the exposure surface is one half the
number which characterizes the embodiment of Fig. 7.
Series-connection of adjacent cell units is schematically
indicated by conductors 815 (915), but it will be under-
stood that various specific techniques may be adopted for
such connection, depending upon specific application of
the resulting cell array. For example, entire grooves
807 (907) may be filled with electrically conductive
material such as metal; alternatively, only the confined
regions near the exposure surface may be filled, as in
groove 8~0 to the depth Dl in Fig. 8, or in groove 910
~0 to the depth D2 in Fig. 9, thus leaving the substantial
included volume of the wider groove regions 807 (907)
wi~hin which to accommodate flow of heat-exchanging
liquid, to serve the dual purpose of extracting useful
heat and cooling the array, while also extracting its
25~ electrical output. Still further, the adjacent cell
units 800 t900) may fixedly be retained by glass, quartz
or other transparent pLate means, bonded to front or back
sur~aces of these units~ with interconnect means carried
by one such plate, thereby leaving unfilled grooves for
particular purposes, such as cooling.
g~
To produce the p~-n~ characterized groove-wall config-
urations of Figs. 8 and 9, groove etching may proceed as
above described, but to slightly more narrow groove-width
proportions than is ultimately to apply in the finished
array. A p+ diffusion may then be applied to both walls
of all grooves, one of the walls being thereafter cut to
establish the ultimate groove widthr by localized etching
to expose bare silicon. One way to accomplish this result
is to oxidize all groove walls after making the p+ diffu- -
sion, and by then using a more narrowly slotted (and
slightly laterally-offset positioned mask for localized
anisotropic etching behind the oxide coating of one wall,
using a material (e.g., hydrazine) which will not attack
the oxide coating of the other wall; the p+ diffusion on
said other wall is still protected by its oxide coating,
so that an n+ diffusion will only be operative on the bare-
silicon wall, whereupon the oxide coating on the p+
diffusion wall can be removed.
While the invention has been described in detail for
preferred forms shown, it will be understood that modifi-
cations may be made without departing from the scope of
the invention. For example, the Figs. 8-9 succession of
p+-n+ characterized groove walls may be readily applied -
to alternate walls wha*ever the groove configuration,
whether straight walled as in Fig. 1, or otherwise as in
said U. S. patent No. 4,110,122. Also, when using the same
materials in Fig. 8 ~and Fig. 9) as those discussed in con-
nection with Figs. 4 and 5, it should be noted that unit-
cell width may be less than in connection with Figs. 4 and ;
5, in recognition of the fact (shown in Fig. 6) that inci-
d~nt light should strike the exposure surface of each cell
unit within 1 to 3 mils (i.e., approximately 2+1 mils) of
the nearby p-n junction.
-23-
