CA1048134A - Solar cell devices - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE:

Tubular solar cells are provided which can be coupled together in series and parallel arrays to form an integrated structure.
Solar energy concentrators are combined with the solar cells to maximize their power output. The solar cells may be cooled by cir-culating a heat exchange fluid through the interior of the solar cells and the heat captured by such fluid may be utilized, for example, to provide hot water for a heating system. The coolant circulating system of the solar cells also may be integrated with a solar thermal device so as to form a two-stage heating system, whereby the coolant is preheated as it cools the solar cells and then is heated further by the solar thermal device.

Description

This inven-tion rela-tes to apparatus for converting solar energy into electrical energy and more particularly to improved solar cells and solar cell arrays.
PRIOR ART
It is well known that radiation of an appropria-te wavelength falling on a p-njunction of a semiconductor body serves as a source of external energy to generate hole-electron pairs in that body. Because of the potential difference which exists at a p-n junction, holes and electrons move across the junction in opposite directions and thereby give rise to flow of an electric current that is capable of delivering power to an external circuit. Accordingly, it is presently common practice to provide an array of solar cells to generate electri-cal energy from solar radiation. Most solar cells are made of silicon but cells made of other materials, e.g., cadmium sulfide and gallium arsenide, have also been developed and ''- tested. Silicon is a favored material since it has a band gap of approximately 1.1 electron volts and -thus responds quite favorably to electromagnetic energy having a wavelength in the visible and ultraviolet regions of the spectrum.
At the state of the art prior to this inverrtion, solar cells are most commonly fabricated as separate physical entities with light gathering surface areas in the order of 4-6 cm . For this reason it is standard practice for power generating applications to mount the cells in a flat array on a supporting substrate or panel so that their li~h-t ga-ther-ing surfaces provide an approximation of a single large ligh-t gathering surface. Also since each cell itself generates only a small amount of power (a silicon solar cell has an open circuit voltage of abou-t 0.52 volt), the required voltage and/
or current is realized by interconnecting the cells of the array in a series and/or parallel ma-trix.

cb/

Ano-ther me-thod is to fabricate in-tegra-ted solar cell panels wherein one region of semiconductivi-ty of each cell is formed by a portion of a con-tinuous body of semiconductor material. Such integrated panels may be used singly but more - commonly are connected in series to obtain the desired curren-t.
A number of problems have been encountered in the manufacture of solar cell panels using individual solar cells.
Among the more significan-t problems and limitations are rela-tively low packing density due to consumption of space by cell interconnections, poor current collecting efficiency, heating ..
up of the cells due to absorption of radiation of wavelengths . greater than about 1.1 micron, energy loss due to reflection of incident solar radiation from the light gathering surfaces ':'.
. of the solar cells, eclipsing of portions of the cells by the cell interconnecting means, physicaldamage to cells and cell ~ interconnections due to thermal cycling or physical stress, .` and high cost of manufacture. Some of the approaches pro-~ posed to overcome such problems are set forth, for example, .~ in the following U. S. Patents and the references cited ,.:,.
;?', 20 thereln:
~ DATE OF
.`` PATENT NUMBER INVENTOR'S NAME ISSUE

~ 3,359,137 Stephen Kaye et al 12/19/67 : 3,575,721 Alfred E. Mann 4/20/71 :, 3,150,999 Rudenberg et al 9/64 3,778,312 Siegfried Karius 12/11/73 3,502,507 Alfred E. Mann 3/24/70 3~489,615 Alfred E. Mann et al 1/13/70 3,378,407 Richard V. Keys 4/16/68 3,819,417 Joseph G. Haynos 6/25/74 3,546,542 Robert K. Riel et al 12/8/70 3,811,954 Joseph Lindmayer 5/21/74 ;, 3,457,427 Rrishan S. Tarneja et al 7/22/69 ,' ~ cb/ - 2 -.
`

3~
.
DAT~ OF
PATENT N~MBER INVENTOR'S N~ME ISSUE
.
3,459,597 Wilfred R. Baron 8/5/69 3,411,050 Arthur E. Middleton et al 11~12/68 3,175,929 David A. Kleinman 3/30/65 3,361,594 Peter Albert Iles et al 1/2/68 3,615,853 John V. Goldsmith et al 10/26/71 3,682,708 Allan I. Bennett 8/8/72 3,574,925 Jens R. W. Schneider et al 4/13/71 3,129,061 Steve N. Dermatis et al 4/14/64 3,162,507 Steve N. Dermatis et al 12/22/64 ` 3,394,994 John W. Faust, Jr. et al 7/30/68 3,116,171 Robert J. Nielsen 12/31/63 - 3,089,070 Eugene L. Ralph 5/07/63 3,015,590 Calvin S. Fuller 1/02/62

2,984,775 Sheldon L. Matlow et al 5/16/61

3,046,324 Sheldon L. Matlow 7/24/62 3,005,862 Charles A. Escoffery 10/24/6]

Certain of the foregoing problems are quite more important than others. For example, it is known that the power of a solar cell increases with increasing intensity of the impinging radiation as long as the temperature remains constant, but decreases again with increasing temperature.
Also common solders used for interconnections are commonly of the soft variety which cannot withstand eleva-ted tempera-tures and also tend to absorb infrared radiation which, as is well known, has a heating effect. Hence, it is also essential or at least desirable to provide some means for cooling -the solar cells and also to protect the panel from infrared radia-tion. It also is important to maximize the amount of ultra-violet radiation absorbed by the solar cells so as to maxi-mize their electrical power output. It also is impor-tant from the standpoint of providing a reliable power source to , cb/ - 3 -3~
have an array which can withstand -thermal expansion and con-traction and mechanical stress of components. For space application, it also is essential to provide solar panels with a high power-to-weight ratio. The la-tter is also important if solar cells are to compete with o-ther means of generating electricity for terrestrial use. Although silicon is an abun-, .
dant material, the cost of silicon in the purity required for `` cell manufacture is quite high ancl its production consumes large amounts of electricity. Hence, it is desirable from ; 10 the standpoint of cost and to reduce resistive losses (which , have the effect of decreasing conversion efficiency) to produce - solar panels wherein the bulk of the semiconductor material in each cell is minimized.
Unfortunately, certain of the foregoing problems and limitations can be overcome only with difficulty or at relatively great expense.
SUMMARY OF THE INVENTION
Accordingly, a primary object of this invention is to provide a solar cell of new and unique configuration which substantially avoids or overcomes a number of the problems encountered in the manufacture and use of solar cells made according to prior art techniques. A further object is to provide solar cell arrays which comprise a plurality of such cells electrically interconnected in a series and/or parallel matrix. More specific objects of the inven-tion are to provide solar cells and arrays thereof which have a modular form, can be easily coolecl, have structural integrity, can be made by existing techniques, and are capable of witnstanding changes in dimensions due to thermal cycling. A further object is - 30 to provide a solar cell unit wherein current leakage is mini~

mized by the use of a geometry which minimizes the ratio of exposed active surface area to exposed junc-tion region area.

~ 4 -:' 3~
Still another objec-t is to provide a solar cell module which can be integrated with a solar -thermal system. ~e-t another objec-t is to provide solar cell modules which can be easily and efficiently interconnected physicall.y and electrically.
: Another important object is to provide solar cells and solar cell arrays of -the type described in combination with radiant energy concentrators for maximizing the in-tensity of radiation received by such cells and also for distributing the concen-tration of such radiation.
This invention provides a solar cell which comprises a tubular structure with a P-N junction formed close to its outer light gathering surface. To collect current from the solar cell, a first electrode comprising a grid of conduc-tors is provided on and forms an ohmic contact with the outer sur-face of the tubular structure, and a second electrode in the form of a layer of conductive material is provided on and forms an ohmic contact with its inner surface. Each tubular struc-ture may comprise one or more photovoltaic cells and a number of such structures may be physically attached end-to-end with appropriate means provided for series and/or parallel electrical connection of cells. Because of the tubular stru.cture, a fluid ,- coolant may be circulated through the interior of each tube so as to provide cooling by direct conduction of heat.
Still other objects of the inven-tion are set forth or rendered obvious by the following detailed description of the invention whichshould be considered together with the .. accompanying drawings, wherein like numbers refer to like : parts and:
Figure 1 is a perspective view with a portion broken away of a preferred form of a cylindrical solar cell constructed in accordance with this invention;
Figure 2 is a longitudinal sectional view of an .~

cb/ - 5 -'?

3~
array of photo cells of the type shown in Figure l;
FiguLes 2A, 2B and 2C are enlarged sectional views of certain components of the array of Figure 2;
Figure 3 is a longitudinal sectional view of a second form of solar cell array;
Figure 4 is a view similar to Figure 3 of solar cell array with a central support;
Figure 4A is a fragmentary sec-tional view showing an alternative form of solar cell array with a cylindrical `~ 10 mandrel support;
Figures 5-8 are fragmentary sectional views on an enlarged scale showing different methods of interconnecting tubular solar cells in an array;
Figure 9 is a view like Figure 3 of a parallel-connected array;
Figure 10 is a perspective view showing several solar cell arrays combined with solar energy concentrators;
and Figure 11 is a cross-section of a Schottky-barrier solar cell.
i DESCRIPTION OF THE INVENTION
The present invention is predicated on the use of semiconductor-grade silicon (or other suitable semiconductor material as hereinafter described) in tubular form. As is already known to persons skilled in the ar-t, silicon and other semiconductor materials may be grown as hollow, i.e., tubular, substantially monocrystalline bodies with cylindrical, rect-:~ angular or other cross-sectional shapes by the processes des-- cribed and claimed in U. S. Patent Nos. 3,471,266 and 3,591,348 issued to Harold E. LaBelle, Jr. on 10/7/69 and 7/6/71 res-pectively (see also U. S. Patent 3,826,625 issued 7/30/74 to J. S. Balley). By controlling the growth environment and ~ .

cb/ - 6 -3~
using a high purity melt, it is possible to grow tubular bodies wi-th a purity suitable for semiconductor purposes.
Also by introducing sui-table conductivity-type-determining impuritiesl i.e., dopants, to the melt it is possible to pro-duce tubular bodies by the aforesaid processes which have a P-or N-type conductivity and a predetermined resistivity. The addition of a dopant to a melt from which a crys-tal is grown is conventional, for example, with Czochralski-type processes and also is exemplified by U. S. Patents ~os. 3,129,061, 3,]62, 507 and 3,334,994.
In the preferred mode of practiclng this invention, - a tubular body of one type conductivity is provided initially, and such body is then treated to provide one or more zones of opposite type conductivi-ty so that a P-N junction is created between such zone or zones and the adjacent portion or portions of the hollow body. The zone of opposite type conductivity may be formed in various ways known to persons skilled in the art, e.g., by diffusion or ion implantation of dopants or by epitaxial deposition of opposite type conductivity ma-terial.
Preferably, the opposite type conductivity zone is formed at the outer surface of the hollow body, preferably by diffusing a suitable dopant into such surface. Thus if the hollow body is a P-type semiconductor, a suitable N-type dopant is diffused - into it to create an N-type conductivity zone. Similarly, if the hollow body is an N-type semiconductor, a suitable P-type dopant is diffused into it to create a P-type conductivity zone. The choice of dopant used depends on the material of which the hollow body is composed and also its conductivi-ty -type.
Thus, for example, boron may be diffused into N-type silicon to produce a zone of P-type conductivity while phosphorus may be diffused into P-type silicon to produce a zone of N-type conductivity. The several -types of dopants used for modifying cb/

........... the conductivi-ty of silicon and how such conduc-tivity-modifying impurities may be diffused into a sillcon body are well known (see for example, U. S. Patents Nos. 3,162,507; 3,811,95~;
;.; 3,089,070; 3,015,590; and 3,546,5~2). The types of dopants required to modify the conductivi-ty type of other ma-terials, e.g., gallium arsenide, cadmium telluride, etc., also are well known to persons skilled in the art. In accordance with the prior art knowledge, the concentration of dopants in the P
and N regions of the tubular structures is controlled to obtain 10 the desired resistivity of the P- and N-type regions. Pre-ferably, the resistivity of such regions is held -to less than about 100 ohm-cm and for best conversion efficiency is between about 0.001 to about 10 ohm-cm. In order to improve -the effi-. ciency of collecting the photoelectrically produced carriers, . the depth of the P-N junction from the outer surface is made < small, preferably in the order of 1/2 micron. After the P N
. junction is formed, the hollow body is provided with ohmic contacts or electrodes for its P- and ~I-type zones whereby the resulting solar cell unit may be connected to an exterior 20 circuit. Additionally, the hollow body may be coated with some sort of anti-reflection or interference film to reduce reflec-~'~ tion losses or to block absorption of infrared radiation.
. In the case of a silicon solar cell for -terrestrial use, it is preferred that the hollow body by N-type silicon and the P-type zone be doped with boron and created at its outer sur-.. face, since the reflectivity of boron-doped silicon is only about 5% as contrasted with the normal reflectivity of 35%
for undoped silicon. In contrast, if an N-type zone is pro-duced by diffusing phosphorus in-to the outer surface of a P-type silicon body, the reflectivity of tha-t surface will be reduced only a minor amount. However it appears that P on N
cells are less resis-tant to radiation deterioration -than N

cb/ - 8 -3~
on P cells. Hence, for space applications, it may be preferred to employ ~1 on P cells wi-th an an-ti-reflection coating or else a P on N cell with an interference film or filter that narrows the wave-length of the incident radiation according to the spectrum of the solar radiations in space and -the spectral res-ponse of the solar cell.
Referring now to Figure 1, the illus-trated solar cell comprises a cylindrical silicon tube 2 of N-type conduc-tivi-ty which has been subjected to diffusion of boron in-to i-ts outer surface to form an outer P-conductivity type region 4 and a P-N
junction 6. The inner surface of the cylindrical tube is pro-vided with a first electrode in -the form of an adherent me-tal conductive film 8 which forms an ohmic contact with the tube.
The film 8 covers the entire inner surface of the tube and consists of a selected metal or metal alloy having relatively high conductivity, e.g., gold, nickel, aluminum, copper or the like, as disclosed in U. S. Patents 2,984,775, 3,046,324 and 3,005,862. The outer surface is provided with a second electrode in the form of a grid consisting of a plurality of circumferentially extending conductors 10 which are connected together by one or more longitudinally-extending conductors 12. The opposite ends of the outer surface of -the hollow tube are provided with two circumferentially-extending terminal conductors 14 and 16 which intercept the longit-udinally-extending conductors 12. The spacing of the circumferentially-extending conductors 10 and the longitudinally extending conductors 12 is such as to leave relatively large areas 18 of the outer surface of the tube exposed to solar radiation.
Preferably, but not necessarily, the conductors 12, 14 and 16 are made wider than the circumferentially-extending conductors 10 since they carry a greater curren-t than any of the latter.
These conductors are made oE an adherent me-tal film like the cb/ - 9 -3~
inner elec-trode 8 and form ohmic contacts with -the outer surface of the tube. The several conduc-tors 10, 12, 14 and 16 and the film 8 may be applied by any of several suitable techniques well known in the art, e.g., by evaporation deposi tion.
As is obvious to a person skilled in the art, the unit of Figure 1 consti-tutes a discrete photovoltaic solar - cell. When the unit is connected by its inner and outer elec-trodesinto an exterior circuit and the exposed portions 18 of the outer surface of the unit are exposed to solar radiation, electron-hole pairs are generated in -the tube with -the result that current will flow through the exterior eircuit via the inner and outer electrodes. The open circuit potential of the unit is approximately 0.52 volt. The same results will be produced if the solar cell unit is made by providing a tube made of P-type eonduetivity and treating its outer sur-faee to provide regions of N-type conductivity with an inter-vening P-N junetion.

~,. .
~, A plurality of P on N or N on P units as shown in ....
Figure 1 may be eombined to form a solar eell array, Witil the individual solar eells being intereonneeted elee-trically _.
either in series or in parallel aecording to the output voltage and the output eurrent desired. Preferably, but no-t necessar-ily, the several units are meehanieally eonnec-ted end-to-end to form an integra-ted strueture.
Figure 2 illustrates a solar eell array eomprising three of the units shown in Figure 1 interconneeted eleetrically in series with one another. As seen in Figure 2, the three units are disposed end-to-end with the eentral unit 20B mecnan-ieally eoupled to the two end units 20A and 20B by means oftwo like eoupling members 22 whieh preferably are made of an eleetrieally insulating material sueh as plastie, but also eb/ - 10 -3~
may be made of an elec-trically conducting ma-terial WiliCh iS
provided with an insulating coating. As seen bes-t in Figure 2A, the coupling members 22 are ma.de of electrically insulating ma-terial and comprise an annular portion 24 formed with cylindri-cal flanges 26 and 28 at its inner and outer edges respectively.
The inner flange extends within and engages -the inner surface of one solar cell unit while the outer flange surrounds and engages -the outer surface oE the adjacent solar cell uni-tO
Interposed be-tween the annular portion of the coupling member . 10 and the solar unit which is surrounded by the outer flange 28 is a ring 30 which also is made ofinsulating material and has a width in its radial direction which is sufficient to span the entire end surface of the solar unit which it engages, thereby preventing short circuiting across the P-lN junction of the solar unit. Each of -the coupling members is provided with a plating 32 of conductive material on the inner surface of i-ts outer flange, the adjacent surface of its annular portion, and the inner, end and outer surfaces of its inner flange. The inner and outer fIanges of each coupling member make a tight 20 fit with the two solar cell units between which i-t is disposed, with the result that a direct ohmic contact is made between the outer conductor 16 of unit 20A and the inner electrode 8 of unit 20B, and similarly between the outer conductor 16 of unit 20B and the inner electrode of unit 20C. A first end member 36 is attached to the free end of uni-t 20A and another : end member 38 is attached to the free end of the third unit 20C.
- As seen in Figure 2~, end member 36 comprises a circular ring portion which engages the end surface of unit 2A and a cylindri-cal ring portion which engages the inner surface of the same 30 unit. End member 36 is provided with a conductive metal coating 40 on the exposed annular end surface of its ring por-tion and also on the inner, end and outer surfaces of i-ts flange portion.

cb/ - 11 -The opposite end member 38 is formed as a cylindrical sleeve with a circular groove 42 in one end to receive the free end of the third solar cell unit 20C. The outer surface of this end member is provided with a concluctive metal coa-ting 44 which extends around the outer part of i-ts inner end surface and along the outer side of groove 42 as shown in Figure 2C.
.. End members 36 and 38 make a tight fit with solar cell units 20A and 20C so that their conductive metal layers 40 and 44 make direct onmic contacts with the inner electrode of unit 10 2A and the outer conductor 16 of uni-t 20C.
., .`. The above-described array is coupled to an exterior circuit (not shown) by means of terminal leads 46 and 48 which are conductively secured to the conductive coatings on end members 36 and 38. As will be obvious to a person skilled .
in the art, the terminal lead 46 is connected to the N side a; of unit 20A while the other conductive lead 48 is connected to the P side of unit 20C. Furthermore, the P side of unit 20A
is electrically connected to the N side of unit 20B, while the P side of unit 20B is connected to -the N side of unit 20C.
As a consequence, the three units are connected elec-trically in series with the result that the open circuit voltage of the array is equal to the sum of the voltages generated by the three solar cell uni.ts, i.e. abou-t 1.56 volts.
The three units 20A-20C of Figure 2 may be held fixed in end-to--end relation in several ways. One way is to connect the coupling members 22, 36 and 38 to the solar cell units by means of a conductive cement located where electrical :.coupling i.s desired. Another approach is to encircle the coupling members and the ends of tubes 2 with mechanical clamp-.30 ing rings, e.g., split rin~s with screw means for drawing -the ends of the rings together so as to radially compress the tubes and coupling members together. Still ano-ther method ci/ - 12 -:

3~
is to provide means for axially compressing the tubes -toge-ther.
A fourth approach is -to force fit the coupling members to the hollow tubes. Still other techniques obvious to persons skilled in the art may be used to hold the assembled solar cell uni-ts together so as to form an in-tegrated struc-ture. Preferably the mode of holding a number of tubular solar cell units assem-bled end-to-end so as to form a sturdy structure is such as to permit a coolant to be circulated through the interior of the units. Three such modes are illustrated in Figures 3, 4 and 4A.
Figure 3 also illustrates how P on iN and N on P cells may be combined in one array.
Turning now to Figure 3, -there is shown an array of : tubular solar cell units 50A, B, C and D which are like the solar cel units ofFigure 2 except that units 50B and 50D are N on P cells whereas units 50A and 50C are P on N cells. Thus, units 50A and 50C are like the solar cell unit of Figure 1 while unit 50B comprises a tube of P-type silicon with the outer surface treated to provide a cylindrical N-type region separated from the interior portion of the tube by a P-N junc-tion which is the reverse of the junction 6. The end cells 50A
and 50D are provided with end members 36 and 3g as described above while four coupling members 52A-D are located between the mutually confronting ends of successive tubes. Coupling members 52 are made of electrically insulating material and are in the form of cylindrical sleeves with a groove in each end face to accommodate a tube end. Coupling members 52A and C
are provided with a conductive metal coating (represented by the heavy line 54) in Figure 3 which covers its outer surface ` and extends around the outer portion of each of its end edges - 30 and along the outer side of each of its end grooves. Thus the end conductors 16 and 14 of uni-ts 50A and 50B and -the corresponding conductors of units 50C and 50D engage and make cb/ ~ 13 -' a direct ohmic contac-t with the conductive coa-ting 54 on coupling members 52A and C. Coupling members 52B and 52D
are like members 52A and 52C excep-t that each is provided with a conductive metal coating (represented by heavy line 56) which covers its inner surface and ex-tends around -the inner portion of each of its end edges and along the inner side of each of its end grooves. Thus, the inner electrodes 8 of units 50B and 50C engage and make an ohmic contact with the conductive coating on coupling member 52B and a similar . 10 contact is made by the conductive coating on coupling member 52D with the inner electrodes of uni-ts 50C and 50D. AS a consequence, the several P-N junctions are connected in series so that the open circuit potential of the array is the sum of the open circuit potentials of the individual cells.
The several solar cell units 50A-50 may be secured together in the same manner as the units of Figure 2 and may be cooled by passing a suitable fluid through the several units via the openings provided by members 36, 38 and 52A-D.
Figure 3 also illustrates how a radiation filter 20 may be combined with a solar cell or solar cell array construct-ed in accordance with this invention. In this case, the radia-. tion filter is formed as a cylindrical tube 58 which is slipped over the several units and is secured, e.g., by mechanical means or by bonding with a suitable cement, to a-t least the two end members 36 and 38 so as to hold the array toge-ther.
For this modification the end member 36 is modified as shown in dotted lines so as to provide a surface for engaging tube 58. AS an optional feature, the filter tube 58 may also be secured to the coupling members 52. The tube 58 is made of 30 a suitable material, eOg., a selected glass, which is trans-parent to radiation with a wavelength which will produce electron-hole pairs and thereby produce the desired photovoltaic cb/ - 1~ -effect but will pass lit-tle or no infrared radiation. Thus, inthe case of silicon, -the filter is made preferably of a material which will block radia-tion of wavelengths grea-ter than about 1.2 microns.
Figure 4 shows a solar cell array like tha-t of Figure 2 wherein the several units are mounted on a central suppor-t.
In this case, three like units 20A-C are separated by coupling members 60A and 60B which are similar to coupling members 22 except that their inner diameters are sized so that they make 10 a snug sliding fit with a center support rod or mandrel 62.
Additionally, each of the coupling members 60 is provided with one or more apertures 64 so as -to permit a coolant to pass from the interior of one solar cell unit to -the next solar cell unit. The coupling members 60 may be plated like the coupling members 22, in which case insula-ting spacer rings like those shown at 30 in Figure 2 may be introduced between each coupling member and the adjacent solar cell unit which is em-braced by the outer flange of the coupling member. Alter-natively, the coupling members may be plated wi-th a conductive 20 metal film which covers the inner, end and outer surfaces of the outer flange and extends to and covers the ou-ter surface of the inner flange, as represented by the heavy line 66 in Figure 4. In such case, a circular spacer 30A made of electri-cal insulating material is interposed be-tween each coupling member and the solar cell uni-t which fi-ts over the inner flange of the coupling member so as to prevent shor-t circuiting of - -the P/N junction by the metal film 66. In this way, each i coupling member provides an ohmic connection between -the end conductor 16 of one unit and the inner electrode 8 of the adja-30 cent unit. The opposite ends of the array are fitted with - end members 68 and 70. The end member 68 is essen-tially a ~ cylindrical plug with a reduced diameter axial ex-tension 72 :' cb/ - 15 -. .~
J

~ " .

at one end and a peripheral flange 74 a-t the other end. rrhe circumferential surface of -the flange 74 is coa-ted with a conductive metal film represented by heavy line 7~ which extends to and covers the circumferen-tial surface of that portion of . ".
the plug which fits wi-thin the uni-t 20A, Whereby an ohmic con-tact is made to the inner electrocle 8 of that unit. The end member 68 is provided with an axial bore 76 and one end of the center support 62 is provided with a reduced diameter section which fits within the inrler end of bore 76. Addition-ally, the member 68 is provided with one or more radially-- extending passageways 78 which intersect the axial bore 76.
A non-conductive spacer 80 is interposed between -the flange 74 of end member 68 and the adjacent end surface of unit 20A, so as to prevent short circuiting of the P/N junction. l'he member 68 is secured in place by bonding it and the spacer 80 to the unit 20A and/or by bonding its extension 72 to the central support 62. The opposite end member 70 is also formed with an axial ex-tension 82 and a central bore 84. One or more radially-extending ports 86 are provided which intersect bore 84, and the adjacent end of center support 62 has a reduced diameter section which fits wi-thin the axial bore 84.
` The end member 70 is formed with a cylindrical flange 87 which is sized to fit over and engage the adjacent end of unit 20C. The inner surface of the flange of end member 70 is coated with a conductive metal film represen-ted by heavy line 89 which extends around the edge surface of that flange and covers -the cylindrical ou-ter surface of -the same flange, whereby an ohmic contact is made to the end conduc-tor 16 of unit 20C.
End member 70 is bonded to the unit 20C and/or to the center support 62. As a result, the several units and the center support 62 form an integrated structure. r~'erminal leads 46 and 48 may be coupled to the conduc-tive metal films on end cb/ - 16 -3~
members 68 and 70 as shown, whexeby -the il.lustra-ted array ' may be connected onto an exterior circuit (not shown).
The above-described array offers -the advantage that the end members 68 and 70 not only are used to form a s-turdy mechanical assembly but also function as means for circulating a coolant fluid through the interior of the array. A coolant may be introduced, for example, through the axial bore 76 and radial ports 78 and removed via radial ports 86 and axial bore 84, with the coolant passing from one unit to the other via the - 10 passageways 64 of coupling members 60A and B.
Figure 4A shows a further modifica-tion of the inven-tion. In this case, the center support 62 extends through an end member 88 which is similar to end member 68 except that it lacks the reduced diameter extension 72. An O-type seal 90 is located in a groove surrounding the axial bore in end member 88 and tightly engages the central support 62. The outer end of the center support 62 is threaded as shown at 92 to receive a nut 94 which cooperates with the central : support to urge the end member 88 against it in a direction to compress the spacer 80 between it and the end of -the solar cell unit 20A. The central support 62 is provided with a blind axial bore 96 and one or more radial ports 98 which intersect bore 96. At the opposite end of -the array, an end member lO0 is employed which is similar to end member 38 . and has a conductive coating 101 like coating 44. A second nut 94 at the adjacent end of support 62 urges end member lO0 agains-t the end of solar cell unit 20C. Hence, the several . solar cell units are held together by the axial compression . exerted on end members 88 and 100 by coaction of nuts 94 . 30 and center support 62. A coolant may be introduced in-to one end of the array via axial bore 96 and ports 98 and is - withdrawn from the other end of the array by the corresponding 3~
por-ts and axial bore in the opposite end of center suppor-t 62.
The use of a center support 62 wi-th end members as shown in Figures 4 and 4A is advantageous regardless of whether the solar cell array comprises P on N or N on P cells or a combina-tion of P on ~ and N on P cells.
Figure 5 shows one alternative method of electrically and mechanically coupling together two solar cell units of -the -type shown in Figure 1. In this case a non-conductive coupling member 104 is employed which is in the form of a cylindrical annulus having an inner diameter sized to mahe a close sliding fit with the center support 62. The coupling member 10~ is provided with passageways 64 as shown for permitting a coolant to flow from one solar cell unit to the other. The ou-ter surface of the coupling member 10~ is provided with a rib 106 which fits between and forms two oppositely disposed shoulders for engaging the two solar cell units. In this case each of the solar cell units 20A and 20B is modified so that at one end its inner electrode 8 terminates a short distance from its end edge, while at the other end the metal film which forms the electrode is extended around the end edge and up oyer the outer surface of the hollow tube so as to form a tab as shown at 107. ~owever, between the extended portion of the inner electrode 8 and the end and outer s~rfaces of the tube 2, a thin layer of insulating material 108 is provided so as to prevent shor-t circuiting of the P/N junction. By way of example, if the tubes of solar cell units 20A and 20s are made of silicon, the insulating material 108 may be a film or layer of silicon dioxide (SiO2). In this case also, -the end conductor 16 is spaced from the end edge of the tube 16 so that a gap exists between it and the extended portion 107 of the inner electrode. The adjacent ends of the two -tubes are fit-ted over the coupling member 104 so tha-t they abut the cb' - 18 -shoulders forrned by its rib 106, and a sui-table non-conduc-tive cement or adhesive may be applied between the rib and the adja-cent end surfaces of the two tubes as shown at 110 so as to bond the two tubes to the coupling member 104. Thereafter, a direct electrical connection is ma,de between the inner electrode 8 of the tube 20A and the outer conductor 14 of unit 2OB by means of one or more conduc-tive straps 112 which are secured to tab 107 of unit 20A and conductor 14 of unit 20B by solder-ing or by a conductive cement or by other suitable means known . 10 to persons skilled in the art. To allow for expansion and contraction due to temperature variations, the cement 110 may be omitted and the conductive strap 112 may be formed with a . bowed portion as shown in phantom at 114, whereby endwise movement of one tube relative to coupling member 104 and the other tube may be compensated for by flexing of the bowed por-tion 114.
Figure 6 shows still another way of providing electri-' cal connections between two adjacent tubular units. In this ;.- case the coupling member 116 is similar to coupling member 104 ,, 20 except that its outer rib 117 is bevelled as shown. The outer - .
.. , surface of coupling member 116 is provided with a coating of ,' a conductive metal as shown at 118 which is soldered to and ~'" makes an ohmic contact with the inner electrode 8 of -the unit .' 20B. The other unit 20A has its inner surface bonded to coupl-'.' ing member 116 by a non-conductive cemen-t as shown at 119.
. ,.
'~ The end conductor 16 of unit 20A is coupled to the me-tal film 118 on coupling member 116 by one or more flexible conductive wire straps 120. If desired, the straps 120 may be replaced by a flexible conductive cylinder with one end large enough to surround and engage the end conductor 16 of unit 20A and the other end small enough to surround coupling member 116 and , be conductively bonded to the metal film 118.
'' cb/ 19 -Figure 7 shows an arrangemen-t wherein the central support 62 extends through spacer elemen-ts 122 whic}l are similar to coupling members 104 and 116 except -that they do not extend between -two solar cell units. Preferably, but no-t necessarily, the spacer elements 122 are bonded to units 20A and 20s and preferably are slzed to make a close but sliding fit with center support 62. Spacers 122 are provided with passageways 64 to permit flow of coolant as previously described. Inter-posed between and connecting the two solar cell units is an accordian-type bellows 124. One end of the bellows has a cylindrical extensian 126 which fits over and is bonded to the end conductor 16 of unit 20A. The other end of the bellows has a cylindrical extension 128 which fits within and is bonded to the inner electrode 8 of unit 20B. Preferably, bellows 124 is made entirely of a conductive metal or metal alloy; alter-natively, it may be made of a non-conductive material but plated with a conductive metal so that a direct electricalconnection : is made between the conductor 16 of unit 20A and the inner electrode 8 of unit 20B. The cylindrical portions 126 and 128 are preferably soldered but may be bonded by a conductive cement to units 20A and ~OB so that a good ohmic contact is assured.
This modification offers the advantage that -the bellows 124 allows one or both of the coupled units 20A and -j 20B to shift lenythwise to compensate for shock or temperature-; induced expansion or contraction without rup-turing the connect-ions between the coupled unit.
Figure 8 shows a modification of the invention which is like that of Figure 7 except that the bellows 124 is replaced with a bowed flexible sleeve 130 which has a cylindrical end section 132 which is bonded to end conductor 16 of unit 20A
; and a smaller cylindrical end section 13~ which is bonded to cb,/ - 20 the inner elec-trode 8 of unit 20B. The sleeve 130 may be made of a conductive material or of an insulating ma-terial with conductive surface coatings so as to provide a direct electrical path between end conduc-tor 16 of uni-t 20A and inner elec-trode 8 of unit 20B. If desired, insulating spacers 136 may be bonded to the confronting end surfaces of unL-ts 20A
and 20B as shown in Figures 7 and 8 so as to prevent portions of the bellows 124 and sleeve 130 from making elec-trical con-tact with those end surfaces; in this way short circuiting of the P/l~ junctions by the elements 124 and 130 is avoided in -the event the units 20A and 20B are moved toward one another.
The embodiments of Figures 7 and 8 offer the advantage that the units 20A and 20B are free -to move lengthwise of the central support 62 to a limited extent, thereby preventing rupture of the electrical connections between them when the units are subjected to shock or vibra-tions or when they con-tract or expand due to changes in temperature. The different ways of coupling together adjacent solar cell. units shown :.
` in Figures 5-8 may be employed in arrays where the opposite . 20 ends of the center support 62 are connected to end members as shown in Figures 4 and 4A, or otherwise.
A further advantage of the use of center support 62 iS that it may be used as a common conductor or bus for the inner electrodes 8 of several solar cell units where it is desired to electrically connect the several units in para-~`. llel. Thus, as shown in Figure 9, three tubular solar cell units 20A-20C are connected end-to-end by means of coupling ~. members 140 which are shaped generally like the coupling ele-.. ments 52A and C of Figure 3 and have corresponding conduc-tive coatings 54. However, the inner diameters of coupling member 140 are sized so that their inner surfaces tightly grip center support 62, and passageways 64 are provided -to allow a coolant cb/ - 21 -to be passed through the several units. Additionally, the inner surface of each coupling member is coated with a conduc-t-ive metal film, which as represent:ed by the heavy line 142/
extends around the inner portions of its opposi-te end surfaces and along the inner sides of its -two grooves. ~'he me-tal films 54 are bonded -to -the end conductor 16 of one uni-t and -the oppo-site end conductor 14 of the adjacent unit, while the metal films 142 are bonded to the inner electrodes 8 of the correspond-ing units and tightly grip the center support 62. The la-tter is made of an electrically conductive material or else has an electrically conductive coating so that it will serve to electrically connect the inner electrodes 8 of the three units to~nd member 68A. The latter is like end member 68 except that it is made of a conductive material. End member 70 is made the same as the correspondingly numbered element in Figure

4. Terminal leads 46 and 48 are bonded to the end member 68 and the conductive metal film 89 of member 70 respectively.
As a consequence, the three cells are connected in parallel with one another so that when the array is connected to an ; 20 exterior circuit, the total current output will be the sum of the currents generated by the individual solar cell units.
Figure 10 illustrates how tubular solar cel]s as provided by the present invention may be combined with solar energy concéntrators, the solar cells acting as energy receivers.
The embodiment of Figure 10 comprises four solar cell arrays or bat-teries 150 like the one shown in Figure 4 and, for con-venience of illustration, only tllree of the arrays are com-bined with concentrators. Each concentrator 152 is affixed to a support plate 151 and comprises flat opposite end walls 154 and 156, opposi-te side walls 158 and 160 which are para-bolically curved in cross-section, and a bottom wall 162 which is circularly curved in cross-section. The open upper end of cb,' - 22 -each concentrator forms an entrance pupil with a width dl.
Each concentrator is made so that the inner surfaces of its end, side and bottom walls are capable of func-tioning as reflectors of solar radiation. Thus, for example, each concentrator may be made of sheet metal wi-th a mirror surface, e.g. aluminum, or may be made of a plastic with a reflec-tive metal film deposited on its inner surfaces. The junction of the side walls 158 and 160 with bottom wall 162 forms an exi-t pupil with a width d2. The curved bottom wall 162 forms a chamber to receive the associated solar cell array 150 which is centered in the chamber. The radius of curvature of the bottom wall is great enough to provide a space between it and the associated array which is large enough to permit its inner surface to receive and reflect a substantial portion of whatever radiation passes into the exit pupil. Preferably but not necessarily, the outer diameter of the tubes which form each solar cell array is about one-half of the width d2.
Preferably but not necessarily the width of the entrance and exit pupils are set so as to provide a ratio of dl/d2 equal to l/sin emax, where emax is the angle formed between -the center axis of the concentrator and a line extending from one edge of the entrance pupil to the opposite edge of the exit pupil. The concen-trator accepts radiation (diffused or collimated) over an angle of 2emax and concentrates it all in the exit pupil. This type of concentrator is described in a preprint of an article by Roland Winston, "Solar Concentrators of a Novel Design", scheduled for publication in the Octo 197~ issue of Solar Energy Journal.
The opposite ends of each array extend through insulating sleeves 164 mounted in -the opposite end walls of the associated concentrator and conduits 166 and 167 are attached to the end members 68 and 70. The condui-ts 166 and cb/ - 23 -3~
167 are connected -to header pipes 168 and 169 respe~-tively.
The latter are connected to condui-ts 171 and 173 whereby coolant is fed into one end of each array and fed out of the opposite end of each array. The coolant circulating system is preferably of the closed loop type comprising an exterior heat exchanger shown schema-tically as box 170 and a pump 172 for circulating the coolant through the solar cell arrays and the heat exchanger. With such a circulating system, the coolant absorbs heat from the solar cell arrays and is relieved of heat in the heat exchanger. For terrestrial in-stallations, the heat exchanger may be replaced by a refriger-ation plant or a large reservoir of coolant which is adapted to give up the heat recovered from the solar cells by radiative cooling or by heat exchange with a solid or fluid medium, e.g.
stones, water, air, etc.
As an alternative measure, the coolant circulating system may be arranged so that coolan-t circulates -through the several arrays in series instead of in parallel. However, a parallel cooling system as shown in Figure 10 is preferred since it enables all of the arrays to be maintained at sub-stantially the same temperature.
Still referring -to Figure 10, the three cells in each array are connected in series in the manner shown in Figure 4, but the four arrays are connected in parallel, whereby a series-parallel ma-trix is formed. The parallel connections are provided by ~a) connecting together -the coupl-ing members 68 with conduc-tive straps 17~ which are bonded to and make ohmic contacts with the metal films 75 of the coupling members, and (b) connecting together -the coupling members 70 with conductive s-traps 176 which are similarly secured to the metal films 89. Terminal leads 46 and ~8, similarly connected to one of the coupling members 68 and 70, are provided -to cL,/ - 24 -connect the solar cell matrix to an exterior circuit. The la-tter may comprise a power consuming load such as, for example, a d.c. motor, an electric heater or electric lights, or a power storage means such as a rechargeable s-torage battery.
In the embodiment of Figure 10, some of the solar radiation entering the entrance pupil of a concentra-tor may pass direc-tly through the exit pupil and be received by the associated solar cells either directly or after reflection from the bottom wall 162. The remainder of the radiation entering the entrance pupil is reflected by the end or side walls of the concentrator into the exit pupil where it strikes the solar cells either directly or after reflection from bottom wall 162. The lat-ter wall functions -to direct radiation onto the bottom half of the solar cell array so that each solar cell is irradiated substantially uniformly over its entire circumference. This ~ has the dual effect of maximizing the current output and avoid-; ing local hot spots. Simultaneously, the circulating coolant removes any heat generated in the solar cells by absorption of infrared radiation or by resistive losses, whereby the solar cell arrays are maintained at an even temperature. The coolant employed and the rate at which it is circulated are selec-ted so as to maintain the solar cells at a tempera-ture which will enable the cells to operate with a satisfac-tory conversion efficiency Also the coolant must be a non-conductor of electricity since otherwise i-t might cause shor-t-circuiting of the cells. By way of example but no-t limitation~ the coolant may be de-ionized water, a fluorina-ted hydrocarbon, a silicone oil, Freon ~, air or nitrogen.
It is to be understood that the -tubular solar cells and arrays may be combined with other forms of solar energy concentrators. Thus the concentrator may take -the form of a simple trough-like reflector which has a parabolic cross-section, cb/ - 25 -with a tubular solar cell or reflector extending lengthwise of the trough subs-tantially coaxially with the focus of -the parabola. Furthermore, a -transparen-t cover may be mounted over the concentrator(s~ to provide protec-tion from rain, ; dus-t, etc.
While the invention as herein described preferably takes the forms of silicon P-N junction solar cells, it is not limited to devices made of silicon or to devices with homo-junctions, Instead the tubular cells may be made of 10 other semiconductor materials and comprise a hetero-junction or a surface barrier junction (e.g., a Schottky-barrier) in place of a homo-junction. Furthermore, the semiconductor material need not be substantially mono-crystalline since photovoltaic devices are known which comprise polycrys-talline semiconductor materials, e.g., cadmium telluride. Thus, for example, tubular solar cells may be made which essentially comprise gallium arsenide P-N junctions, gallium arsenide phosphide P-N junctions, cadmium telluride P-N junctions, cadmium/sulfide/copper sulfide and gallium arsenide/gallium 20 phosphide hetero-junctions. Similarly, for example, the tubular solar cells may be surface barrier devices which com-pxise metal or metal oxide/semiconductor junctions, e.g., solar cells using gold on N type silicon, aluminum on P~type silicon, tin oxide on N-type silicon, chromium on P-type ; silicon, and indium oxide on cadmium telluride. The afore-; said P-N and hetero-junctions may be made by providing a tubular body of one junction material and forming a layer of the other junction material at or on the inner or outer sur-face of such body by methods well known in the art, e.g., by 30 diffusing an opposite~conductiyity-type dopant into the outer surface in the case of a homo-junction or epitaxially growing a thin layer of the other junction material on the outer surface cb/ - 26 -3~
of the tubular body in the case o:E a he-tero-junction. Sim-ilarly, solar cells with surface barrier junctions may be made by depositing a metal or metal ox:ide barrier material on -the outer surface of a -tubular semiconduc-tor body by vacuum deposi-tion, sputtering, electroless plating or other suitab]e tech-nique. ~n essential requirement of the barrier ma-teri.al is that it have a suitable optical t:ransmission capability so -that - the device will exhibit a pho-tovoltaic behavior.
; E`igure 11 il]ustra-tes a cross-section of a -tubular '; 10 Schottky-barrier solar cell which comprises a tubular body ~ .
180 of P-type silicon, an aluminum ohmic contact layer 182 . on its inner surface, and a layered Schottky barrier on its outer surface which is made according to the teachings of W. A. Anderson et al, An 8~ Efficient Layered Schottky-Barrier Solar Cell, Journal of Applied Physicsl Vol. 45, No. 9, pp.
3913-3915, Sep~. 1974. The layered barrier consists of a chromium barrier layer 184, a copper conductive layer 186, a chromium oxidation layer 188 over the copper layer, an alumi-num ohmic contact or current collector 190, and a silicon :~. 20 monoxide anti-reflection coating 192. The ohmic contact is:, .- represented as several discrete sections since i-t is fabri-. cated as a grid, preferably a grid with sec-tions corresponding to conductors 10, 12 and 14 and 16 of the cell unit shown in .:.
~- Figure 1, whereby a plurality of relatively large areas of -the - chromium barrier layer 184 (corresponding to areas 18 of Figure :~ 1) are exposed for stimulation by solar radiation. The silicon monoxide layer is applied over the copper layer in the spaces . between the discrete sec-tions of the grid-like outer contact 190. Preferably the ohmic con-tact 182 covers most if not all 30 of the inner surface of the silicon body and has a thickness of about 1 micron or less. The thickness of chromium layer 184, copper layer 186, chromium oxidation layer 188, ohmic ",(, cb/ - 2~ -3~
contact 190 and the SiO coating 192 are preferably rnade with thicknesses of 44A~, 58A, 23A, 1000A, and 690A respec-tively.
Obviousiy, a plurality of -tubular hetero-junction and surface barrier junction solar cells may be arranged to form arrays and be combined with concentrators in the various ways illus-trated in Figures 2-lG.
It is to be appreciated also that the tubular solar cells need not be cylindrical, but instead, for example, they may have an el~iptical, square, rectangular, or o-ther cross-sec~ional configura-tion. The essential requirement of the invention is that the solar cell unit comprise a tubular semi-conductor body adapted to exhibit a photovoltaic behavior.
In this connection, it is to be noted further that the term "photovoltaic se~ieonductor barrier device" is intended to encompass devices which have a homo-junction, a hetero-junction, o~ a surface barrier junction, and also that the term "surface barrier junction" ineludes metal/semiconductor barrier devices and metal oxide/semieonduetor barrier deviees, and notably Schottky barrier devices.
The following specific example illustrates a pre-ferred mode of practicing the invention.
E X A M P L E
A cylindrical substantially monocrys-talline P-type silicon tube is grown according to the me-thod described in U. SO Patent 3,591,348. The tube is made with a length of .:
about 6 inches, an outside diameter of 0.50 inch and a wall - thickness of about 0.01 ineh. The interior surface is plated with a 0.001 ineh layer of nickel and phosphorus is diffused ; into the outer surface of the tube to a depth of about 0.5 micron to form an N-type outer region with a distinct P-N
junction. Then aluminum is vacuum deposi-ted on-to the outer surfaee of the tube in -the form of a grid consis-ting of a ~/ - 28 -plurality of longitudinally- and circumferen-tially-extending conductorsO The aluminum grid is formed with a thickness of about 4.0 microns. The inner and outer conductors are connected to a measuring circui-t and -the device irradiated by sunlight.
The device exhibits an open circui.t voltage of about 0.5 volts and a conversion efficiency of about 10%.
The advantages of the invention are numerous. The - tubular structure renders the cell units self-supporting even with tubes of relatively small wall thicknesses (e.g., 1/2 inch silicon tubes with a wall thickness of 100-200 microns), thereby obviating the need for a supporting tubular substrate.
The absence of a supporting substrate reduces weight, cost and also facilitates mechanical and electrical interconnection of two or more cells. Furthermore, the tubular cells may be ' connected electrically in parallel or in series, and by means - of inexpensive reflectors such as a parabolic reflector it is possible to achieve energy concentration ratios of 10 or more, By connecting a plurality of tubular cells in series, it is possible to obtain a high electrical power output at a moderate current level and at a vol-tage level sui-table for charging conventional batteries, thereby obvia-ting the need for heavy conductors on the cells. It is to be appre-~, ,; ciated also that to achieve a reasonable voltage outpu-t from - a photovoltaic array, individual cells must be connected in series. In the case of planar cells, a rec-tangular hea-t ex changer is required Eor cooling purposes if solar concentration is used, but the heat exchanger typically must be electrically insulated from the solar cells. The intervening insula-ting layer reduces the rate at which heat can be conducted away from the solar cells and also tends to complicate -the heat exchanger design. The instan-t invention facili-tates cooling since the coolant is in direct contact wi-th the inner surface c~/ - 29 -of the tubular solar cell. Ilence, no auxiliary heat exchanger structure need be mounted immediately adjacent to the solar cell unit. A further advantage is that whatever portion of the received solar energy is absorbed by the solar cells as heat may be recovered by the coolant. ~lence, the coolant for the solar cell arrays may advanta~eously be coupled to solar thermal devices of the type which are designed to hea-t a fluid by solar energy and to use the heated fluid as a heat supply or, if it is steam, to drive a turbine and thereby an electri-cal generator. More specifically~ the exit coolant from tnesolar cells may be used as the entry heat absorber for a solar thermal device. As a result of the preheating of the fluid by its transit through the solar cells, less heating is required to be accomplished in the solar thermal device to produce a selected fluid temperature, e.g., 200F, and therefore, -the solar thermal device can be correspondingly reduced in size from what it would have to be if the fluid was not preheated in the solar cells.

cl/ - 30 -

Claims (24)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A solar cell unit of the type comprising a tubular semiconductor body having an outer radiation-receiving region of a first conductivity type and an inner region of a second opposite conductivity type separated by a P-N or N-P junction, and first and second electrodes respectively carried by said body on opposite sides of said junction for coupling said outer and inner regions to an external circuit, said first electrode comprising a plurality of contacts electrically-connected to one another and to said outer radiation-receiving region.

2. A solar cell unit according to claim 1 wherein said body is cylindrical.

3. A solar cell unit according to claim 1 wherein said body is made of silicon.

4. A solar cell unit according to claim 1 further including a radiation filter surrounding said outer radiation-receiving region.

5. A solar cell unit according to claim 4, wherein said radiation filter is a self-supporting tubular element, and further including means holding said tubular element fixed with respect to said tubular semiconductor body.

6. A solar cell unit according to claim 1 wherein said body has outer and inner surfaces and said contacts are attached to said outer surface.

7. A solar cell unit according to claim 6 wherein said first electrode comprises a grid of electrical conductors bonded to and covering portions of said outer surface.

8. A solar cell unit according to claim 7 wherein said grid includes a first conductor at one end of said body and a second conductor at the other end of said body.

9. A solar cell unit according to claim 1 further including radiation-receiving means positioned adjacent to said body for directing received radiation onto said radiation-receiving region.

10. A solar cell unit according to claim 1 wherein said outer and inner regions are substantially concentric with one another.

11. A solar cell unit according to claim 10 wherein said electrodes form ohmic contacts with said regions.

12. A solar cell unit according to claim 10 wherein the outer radiation-receiving region and said inner region have P-type and N-type conductivities, respectively.

13. A solar cell unit according to claim 10 wherein the outer radiation-receiving region and said inner region have N-type and p-type conductivities, respectively.

14. A solar cell unit according to claim 1, wherein said junction is a surface barrier junction.

15. A solar cell unit according to claim 14 wherein said junction comprises a conductive metal or metal oxide layer.

16. A solar cell unit according to claims 1, 2 or 3 wherein said junction is a hetero-junction.

17. A solar cell array comprising a plurality of solar cell units according to claim 1 wherein at least some of the electrodes of said units are interconnected so that at least some of said solar cell units are connected electrically in parallel and/or in series.

18. A solar cell array according to claim 17 wherein at least some of said units are disposed so that their tubular bodies are disposed end-to-end, and means for mechanically interconnect-ing said end-to-end bodies.

19. A solar cell array according to claim 17 further in-cluding a coupling member disposed between and connecting the end of one body with the adjacent end of another body.

20. A solar cell unit according to claim 19 wherein said coupling member provides an electrical connection between an electrode on said one body and an electrode on said another body.

21. A solar cell array according to claim 18 wherein said interconnecting means comprises an elongate support member dis-posed within and extending lengthwise of the bodies of at least some of said units, and means extending radially of said support member for preventing movement of the bodies of said at least some units radially of said support member.

22. A solar cell array according to claim 21 wherein said radially extending means has openings to permit a fluid to flow lengthwise within and between the bodies of said at least some units.

23. A solar cell array according to claim 22 wherein said radially extending means are mounted on said support member.

24. A solar cell array according to claims 17, 18 or 19, further including means for circulating a fluid coolant through said hollow bodies.