MODULE FOUNDATIONS
From photovoltaic cell to power-generating module
A photovoltaic cell performs the energy conversion; a module turns many cells into a protected, electrically connected product that can operate outdoors.
Cells are normally connected in series to raise voltage and may be divided into parallel strings to manage current and reliability. Interconnect ribbons carry current between cells, while busbars and a junction box provide the module's external electrical connection. Bypass diodes commonly protect cell groups when part of a module is shaded.
The package must admit useful sunlight while limiting moisture ingress, mechanical damage, corrosion and electrical leakage. A conventional glass-backsheet module therefore combines front glass, encapsulant, the interconnected cell circuit, a second encapsulant layer and a rear backsheet. Glass-glass, flexible, bifacial and building-integrated products alter this stack to meet different optical, structural and environmental requirements.
This module explorer uses the same chronological, layer-focused structure as the PV-cell background. Future entries can compare how module materials, interconnection methods and formats changed as cell technologies and applications evolved.
Schematic basis: compared with the U.S. DOE module-assembly description, ARPA-E common module-construction diagram, and IEA PVPS module-component definitions.
1954–1958 · FIRST OPERATIONAL SPACE ARRAYS
From a practical silicon cell to a dependable power assembly
The 1950s turned the silicon solar cell from a laboratory device into an interconnected electrical system capable of operating beyond Earth.
Bell Laboratories demonstrated a practical silicon cell in 1954 at about 6% conversion efficiency. That cell milestone did not by itself make a module: engineers still had to connect many small cells, support them mechanically, route their combined output and qualify the assembly for its operating environment.
Vanguard I, launched on 17 March 1958, supplied the decisive system demonstration. NASA's historical solar-array review records six small panel units distributed around the spherical spacecraft. Each panel contained 18 p-on-n silicon cells measuring approximately 2 × 0.5 cm, giving 108 cells in total. The complete array produced less than one watt, yet powered a transmitter for more than six years.
This was not a modern glass–encapsulant–backsheet module. Mass and area were extremely constrained, so small cells were interconnected on insulated supports and mounted directly to the spacecraft body. The design problem was already recognisably modular: combine fragile cells into a mechanically supported, electrically routed and testable power-producing unit.
The success of Vanguard I established solar arrays as a credible long-life spacecraft power source. Later space modules added larger deployable substrates, coverglass protection, radiation-resistant cells, redundant interconnects and increasingly sophisticated deployment and pointing mechanisms.
Historical basis: NASA, History of Solar Cell Arrays, NASA Science, solar-power history, and NASA, Project Vanguard history.
1959 · FIRST SOLAR-CELL PADDLES
Move the generating area beyond the spacecraft body
Explorer 6 introduced solar-cell paddles: rigid photovoltaic panels that extended away from the spacecraft to provide more collecting area than its body could carry.
Vanguard I proved that body-mounted cells could power equipment for years, but the available cell area was limited by the spacecraft shell. NASA's historical array review identifies Explorer 6, launched on 7 August 1959, as the first spacecraft to use solar-cell paddles.
Four approximately 46 × 46 cm paddles were mounted around the satellite's equator and intended to recharge its storage batteries. The paddles introduced several new module-level functions: a rigid substrate for the cell field, deployment hinges, positive locking, flexible conductors across moving joints and a charging interface between intermittent sunlight and the spacecraft load.
The mission also demonstrated why those functions matter. Only three paddles fully extended and locked, leaving the initial power system at about 63% of its nominal level. Radiation-belt exposure then degraded the cells, and the falling charging current contributed to the end of transmissions on 6 October 1959.
Despite that short life, the architectural step was decisive. Photovoltaic area no longer had to follow the spacecraft body. Later systems developed larger rigid wings, multiple-fold panels, sun-pointing drives and eventually flexible blankets and roll-out arrays.
Historical basis: NASA, History of Solar Cell Arrays, NASA, Explorer VI mission report, and NASA spacecraft catalogue record.
1962 · RADIATION-PROTECTED REPEATABLE MODULES
Package cells for a harsher and longer mission
Telstar I treated groups of solar cells as preassembled, testable modules and protected them against radiation, micrometeorites, launch vibration and extreme temperature change.
Launched on 10 July 1962, Telstar I relayed the first live transatlantic television pictures. Its spherical exterior carried a solar plant of 3,600 cells, but the engineering advance was smaller and repeatable: 300 modules containing 12 cells each.
Each module connected twelve 1 × 2 cm n-on-p silicon cells in a shingled series string. The n-on-p structure and shallow junction preserved more useful blue-green response after radiation exposure, while a silicon-monoxide antireflection coating reduced optical loss.
The thin cell string could not survive launch by itself. Bell Laboratories mounted it on a metallised ceramic plate through compliant silver U-shaped links and small Nilvar or Invar interfaces that accommodated the different thermal expansion of silicon and ceramic.
A clear synthetic-sapphire plate completed the package. Brazed into a platinum frame and soldered around the ceramic assembly, it shielded the cells from micrometeorites and reduced damaging electron energy without blocking useful sunlight. This made protection, interconnection and environmental qualification part of the photovoltaic module rather than spacecraft-level afterthoughts.
Historical basis: NASA NTRS, The Solar Cells and Their Mounting, Smithsonian, Telstar photovoltaic module, and NASA, Telstar history.
1963 · STANDARDISED TERRESTRIAL MODULE PRODUCTION
Move photovoltaic modules from spacecraft projects to remote infrastructure
Sharp's 1963 production programme turned monocrystalline-silicon cell assemblies into repeatable terrestrial modules suited to unattended navigation equipment.
Sharp began photovoltaic research in 1959 and started mass production of standardised solar modules in 1963. In May of that year, the modules entered their first marine application on a lighted buoy in the Tsurumi Shipping Lane at Yokohama Port.
This changed the engineering target. A terrestrial marine module did not need Telstar's heavy charged-particle shielding, but it did need to withstand salt spray, humidity, rain, wind-driven contamination and long periods without routine maintenance. Transparent protection, sealed edges, corrosion-resistant conductors and electrically isolated support became central module functions.
The surviving Sharp photograph shows a compact clear enclosure containing rows of circular monocrystalline cells. It also illustrates an important manufacturing transition: the module is a recognisable, repeatable unit rather than a cell circuit designed around one spacecraft surface.
Standardisation let completed units be manufactured, inspected and replaced consistently, while multiple modules could be combined to match a buoy, lighthouse or remote-station power requirement. In 1966 Sharp supplied a 225 W lighthouse installation on Ogami Island, then the world's largest lighthouse photovoltaic installation according to the company's historical record.
Historical basis: Sharp, 1963 standardised solar modules, Sharp, surviving 1963 module photograph, and Sharp's IEEE Milestone history.
1973 · LOW-COST TERRESTRIAL SUBSTRATE MODULE
Design photovoltaic modules for practical remote power
Solar Power Corporation's first product made a deliberately economical terrestrial module from five silicon wafers, a circuit board and silicone-rubber encapsulation.
By the early 1970s, photovoltaic modules had proved themselves in space and specialist navigation systems, but their price still prevented broad terrestrial use. Elliot Berman's Exxon-backed Solar Power Corporation approached the module as a manufacturable remote-power product rather than a spacecraft component.
In 1973 the company manufactured and shipped its first product: five silicon wafers mounted on a circuit board and encased in silicone rubber. The cells used silicon rejected by the semiconductor industry, converting material that did not meet microelectronics specifications into a useful photovoltaic generator.
The substrate architecture was simple and comparatively inexpensive. A rigid board located the cell circuit, while transparent silicone admitted sunlight, insulated live parts and protected them from weather. Modules of this emerging commercial generation served oil platforms, remote telecommunications, boats and rural water pumping—loads for which avoiding fuel delivery and battery replacement could justify the initial cost.
Contemporary NASA and JPL histories also record the limitations of early silicone-potted substrate modules: soft exposed surfaces retained dirt, adhesion could fail, hail could damage cells, and mismatched thermal expansion could fracture them. Those lessons helped drive the later move towards factory-laminated glass-superstrate packages.
Historical basis: NPR/TPR interview on Solar Power Corporation's 1973 product, JPL/DOE history of terrestrial module construction, and NASA history of early low-cost terrestrial modules.
1976 · DURABLE DOUBLE-GLASS TERRESTRIAL MODULE
Protect the silicone-encapsulated circuit behind glass
The Philips BPX 47A placed 34 series-connected circular silicon cells in transparent resin between two glass plates, surrounded by a protective belt and anodised-aluminium frame.
This was a distinct response to the weaknesses of early substrate modules with exposed silicone surfaces. Glass supplied a hard, cleanable optical face and a protected rear boundary, while transparent resin still cushioned and electrically insulated the cell circuit. The edge belt and frame controlled the vulnerable glass perimeter and provided a mounting interface.
Philips' 1976 engineering data specify 34 cells of 57 mm diameter, all connected in series. The completed BPX 47A measured 468 × 365 × 15 mm, weighed 2.4 kg and used a one-metre, two-core output cable. At 1 kW/m² irradiance and 25 °C cell temperature it delivered a typical 11 W at 15.5 V and 700 mA.
The cell design was also documented: boron-doped p-type silicon received a shallow phosphorus-diffused n-type surface, a front grid covering about 7% of the illuminated face, a TiO2 anti-reflection coating and a fully metallised rear contact. This retained space-cell metallisation practice while packaging it for terrestrial service.
Reliability was becoming measurable rather than aspirational. Manufacturer tests included −40 °C cold exposure, cycling between −40 and +85 °C, +85 °C dry heat, cyclic humidity, salt mist, sand-laden wind, frost with water and pressure equivalent to 280 km/h airflow. The BPX 47A helped establish glass-covered modules as a credible route towards long outdoor service.
Historical and scientific basis: Philips 1978 BPX 47A manufacturer data, Philips cell construction and operating analysis, JPL module-encapsulation engineering framework, and peer-reviewed history of silicon module reliability.
1977 · QUALIFIED GLASS-SUPERSTRATE LAMINATE
Laminate the cell circuit to its load-bearing front glass
Spectrolab's Block II module placed a PVB-encapsulated silicon circuit behind a glass superstrate and closed it with a Mylar rear sheet—a recognisable ancestor of the modern glass/encapsulant/cell/backsheet module.
The date marks documented Block II deliveries in 1977 and early 1978. Unlike the earlier cast-silicone substrate modules, the sun-facing glass now carried the laminate, polyvinyl butyral (PVB) bonded and insulated the circuit, and a light polymer sheet completed the rear environmental package.
JPL records identify Spectrolab model 022962-G as a 1.168 × 0.388 m module with 0.453 m² area, 6.1 kg mass, an aluminium frame and external terminals. Its 120 round, 50 mm crystalline-silicon cells formed three parallel branches of 40 cells in series; the documented cell-packing factor was 0.69.
Development was tied to qualification rather than appearance alone. Block II testing combined thermal cycling, humidity, mechanical loading and electrical checks, and a parallel NASA environmental programme found the Spectrolab modules more resistant than the other contemporary module types tested.
The architecture also accumulated real field evidence: 19.5 kW of Spectrolab Block II modules entered the 100 kW Natural Bridges system, and later examinations reported modules still operating close to their original specification after twenty years. PVB, expanded-metal interconnects, silicon-oxide anti-reflection coatings and durable solder joints were identified as contributors—although later high-voltage moisture experience led the industry towards EVA and improved multilayer backsheets.
Historical and scientific basis: JPL Block II environmental testing report, JPL ten-year module-development record, NASA comparative environmental-test programme, JPL Natural Bridges system record, Sandia twenty-year diagnostic analysis, and JPL voltage-isolation and corrosion research.
1981–1984 · BLOCK V EVA RESIDENTIAL MODULE
Standardise the glass/EVA/cell/backsheet module
JPL's 1981 Block V requirements and Mobil Solar's documented 1984 Group II hardware capture the transition to a frameless, glass-supported EVA laminate with a multilayer weather barrier—the construction logic that came to dominate terrestrial crystalline-silicon modules.
The 1981 timeline date marks release of the Block V residential and intermediate-load module specifications. The particular module reconstructed here is the final Group II Mobil design reported in 1984, so the page distinguishes the programme milestone from the later documented hardware rather than back-dating the finished article.
A 3/16-inch (4.76 mm) sheet of tempered low-iron glass, 70.5 × 47.375 inches (1.791 × 1.203 m), served as both front cover and structural support. The frameless glass carried 432 rectangular EFG ribbon-silicon cells wired as twelve parallel strings of 36 cells in series. Two solder-plated copper straps per cell and copper-mesh cross-ties every twelve cells added current-path redundancy.
Modified, peroxide-crosslinked EVA bonded the circuit to the glass and enclosed its rear. Behind it, a Tedlar/aluminium-foil/polyester/adhesive composite separated functions: the outer Tedlar weather skin resisted exposure, aluminium foil supplied the principal vapour barrier, and polyester insulated the metallic foil from the live circuit.
Qualification drove the design. Earlier Group I laminates exposed rear-sheet adhesion, junction-box insulation and corrosion weaknesses; Group II changes produced modules that passed Block V qualification, although Mobil's proposed roof-mounting system still failed because thermal cycling caused movement and distortion. That distinction is important: the module package matured before every part of its building integration did.
Historical and scientific basis: Mobil Solar Block V Group II final report, JPL 1981 Block V specification status, JPL EVA encapsulation report, and JPL Block Program assessment.
1984 · LARGE-AREA AMORPHOUS-SILICON POWER MODULE
Package deposited thin-film submodules for outdoor power
The 1984 Hughes/JPL programme replaced hundreds of separately handled wafers with four one-square-foot glass submodules whose amorphous-silicon cell stripes were deposited and series-connected on the glass itself.
This is a specific research-and-qualification module, not a claim that Hughes invented amorphous-silicon photovoltaics. Its historical value is that it records the practical transition from small-area thin-film products to a 1 ft × 4 ft outdoor power module. A contemporary SERI review separately records that one-square-foot, 5 W amorphous-silicon power modules entered the US market in 1984.
Each of the four one-foot-square submodules contained 28 series-connected a-Si cells deposited on a glass superstrate. Hughes placed the four 28-cell series circuits in parallel along the module's long sides, joined their corner terminals with redundant copper ribbons and carried the combined output to an AMP junction box with provision for an integral bypass diode.
The outer package deliberately borrowed proven crystalline-module materials: tempered low-iron cover glass, EVA on both sides of the submodule assembly, a Tedlar rear cover, an aluminium frame, Kapton edge tape and a rubber isolation gasket. The important innovation was therefore inside the familiar weather package—a monolithically patterned thin-film generator rather than a matrix of soldered crystalline wafers.
The prototype also exposed what had not yet matured. Submodule edge defects caused glass breakage during lamination; conductive lubricant and gasket material contributed to high-voltage insulation failures; and the active circuit sat too close to the frame. The redesign increased the inactive glass border and improved edge isolation. Amorphous silicon also required technology-specific spectral measurement and management of light-induced Staebler–Wronski degradation.
Historical and scientific basis: Hughes/JPL module design, fabrication and test report, SERI history and 1984 commercialisation record, and NREL amorphous-silicon technology and stability review.
1992 · LARGE-AREA MONOLITHIC CdTe MODULE
Coat, scribe and laminate one full-size glass power circuit
Solar Cells, Inc. (SCI, a predecessor of First Solar) produced its first interconnected 60 × 120 cm CdS/CdTe module in April 1992 and reached 6% total-area efficiency that June—a documented step from thin-film tiles to one 0.72 m² monolithic circuit.
Development had begun in late 1990. A 5 mm soda-lime glass superstrate arrived already coated with a silica barrier and fluorine-doped tin oxide (SnOâ‚‚:F). SCI then deposited a thin n-type CdS window and a much thicker p-type CdTe absorber from hot vapours while the glass travelled on rollers through its elemental-vapour-deposition chamber. The contemporary report describes roughly ten seconds for CdS and forty seconds for CdTe, followed by a post-deposition treatment near 400 °C.
Three laser-patterning steps turned the continuous films into 116 cells, each about 1 cm × 58 cm, connected in series across the sheet. P1 isolated the transparent conductor; P2 opened a path through the semiconductors for the following metal to contact the neighbouring front electrode; and P3 separated the Ni/Al rear metal. The circuit typically operated near 65 V, demonstrating how monolithic integration could build useful module voltage without soldering 116 separate cells.
After a one-centimetre perimeter deletion provided electrical isolation, bus bars collected the two ends of the series circuit. Thin leads passed through holes in a second glass sheet. A 0.018-inch EVA interlayer bonded that rear glass to the coated superstrate, and the rear junction box contained the external termination and a bypass diode.
The achievement was still a pilot-line result, not yet mass production. The 1993 annual report records an independently confirmed 6.6% module and a 53 W in-house best result, while identifying shunts, film-thickness uniformity and optical absorption in the glass/TCO as important losses. SCI also established occupational, environmental and recycling controls for its cadmium-bearing process—an essential part of scaling CdTe responsibly.
Historical and scientific basis: SCI/NREL 1993 module-development report, NREL/First Solar CdTe manufacturing retrospective, US Department of Energy CdTe technology overview, and NREL photovoltaic technology history.
1997 · FLEXIBLE TRIPLE-JUNCTION MODULE PLATFORM
Move a spectrum-splitting solar stack through roll-to-roll production
United Solar changed its production cell from a same-bandgap double junction to an a-Si/a-SiGe/a-SiGe triple junction in 1997 and brought a 5 MW-per-year roll-to-roll line into operation—turning thin stainless-steel foil into flexible, polymer-laminated modules.
The innovation joined two ideas. First, three series-connected semiconductor junctions shared the solar spectrum: an approximately 1.8 eV amorphous-silicon top cell absorbed blue light, an approximately 1.6 eV amorphous-silicon-germanium middle cell absorbed green light, and an approximately 1.4 eV a-SiGe bottom cell extended response into the red and near-infrared. Together, the nine doped and intrinsic semiconductor films were only about half a micrometre thick.
Second, the complete cell stack was deposited continuously on a 5 mil (about 0.13 mm) thick, 14-inch-wide, 2,500-foot stainless-steel web. A textured metal/ZnO back reflector improved optical path length. Nine semiconductor layers followed in one radio-frequency plasma-enhanced chemical-vapour-deposition pass, then an indium-tin-oxide transparent contact was sputtered on top. The production report gives a web speed of roughly two feet per minute.
The coated roll was not itself a finished module. A press cut it into 35.2 × 35.6 cm slabs and punched interconnection features. The back-end line etched ITO borders, passivated local shorts under reverse bias, screen-printed collection grids, applied clear lacquer and interconnected selected slabs into a cell block. Vacuum-laminated EVA and Tefzel formed the transparent front cover; the laminate was then trimmed and fitted with a junction box and product-specific connectors or framing.
This exploded view is therefore a representative flexible module from the 1997 platform, not one catalog model. The same front-end material fed rigid, flexible, folding and building-integrated products. NREL's FY 1998 overview records commercial 3–64 W modules at about 7.5% stable efficiency, while later reports document the continuing challenge of light-induced degradation, roll-wide uniformity, yield and reliable polymer encapsulation.
Historical and scientific basis: United Solar/NREL history of the 1997 production transition, roll-to-roll cell and module manufacturing report, DOE/NREL FY 1998 commercial-module overview, and DOE PV:BONUS flexible roofing-module report.
1998 · COMMERCIAL BUILDING-INTEGRATED PV ROOF SHINGLE
Make the photovoltaic module part of the weather-shedding roof
United Solar's SHR-17 joined its flexible triple-junction laminate to a roofing-shingle format, so one product could generate electricity and replace part of the conventional roof covering.
The 1998 date marks a documented commercial and field-validation milestone, not the start of the idea. DOE's PV:BONUS programme entered its final development phase in August 1995, and a shingle demonstration was installed at Southface in Atlanta in June 1996. By fiscal year 1998 the products had entered United Solar's 5 MW production line, the solar roof shingle received an R&D 100 award, and NREL completed a monitored SHR-17 installation on 30 June 1998.
Each monitored shingle was 12 inches wide by 86.5 inches long and rated at 17.3 W, 2.0 A and 8.5 V under 1,000 W/m². The electricity-producing core retained United Solar's spectrum-splitting a-Si/a-SiGe/a-SiGe cells on thin flexible stainless steel, protected by polymer encapsulation. The long, shallow package was arranged to overlap neighbouring roof courses, shed water and route its leads through the roof deck.
The product also records an important design evolution. Early prototypes divided the visible face with physical cuts; development reports describe replacing those cuts with embossments and matching conventional shingle dimensions more closely. That preserved the familiar roofing rhythm without unnecessarily interrupting the continuous flexible laminate. An electroplated collection grid also reduced encapsulation material and cost.
Installation crossed two trades: roofers positioned, overlapped and attached the shingles, while qualified electrical workers completed the hidden wiring. Retrospectively, NREL reported no shingle-module reliability problems during six years of monitoring and roughly one percent annual power degradation. The result was an early commercial example of building-integrated photovoltaics (BIPV)—an electrical generator designed as part of the building envelope rather than mounted above it.
Historical and scientific basis: NREL's six-year SHR-17 performance and reliability study, DOE PV:BONUS Phase 3B final report, DOE PV:BONUS Phase II design report, and NREL's R&D 100 historical timeline.
1999 · COMMERCIAL LARGE-AREA MONOLITHIC CIS MODULE
Pattern a complete series circuit directly into thin films on glass
Siemens Solar Industries expanded its commercial CIS family in 1999 with the approximately 30 × 120 cm ST40, turning one coated glass plate into a roughly 40 W series-connected module without assembling separate wafer cells.
The date marks scale-up rather than the invention of CIS. Siemens had pursued CuInSe2-based photovoltaics since 1980 and introduced 5 W and 10 W products in 1998. In 1999 it added the ST20 and ST40; NREL verified more than 12% efficiency for a commercial large-area module, beyond the DOE year-2000 goal of 10%. The CIS module family also received an R&D 100 award shared by Siemens Solar, NREL and the California Energy Commission.
The active circuit began with ordinary soda-lime window glass. A thin SiO2 barrier controlled sodium diffusion and improved adhesion, followed by a sputtered molybdenum rear electrode. Copper-gallium and indium precursors were reacted in selenium- and sulfur-containing atmospheres to form a compositionally graded Cu(In,Ga)(Se,S)2 absorber. A very thin CdS buffer and conductive ZnO window completed the junction.
Three aligned patterning operations made the plate into a series circuit. A laser P1 scribe isolated the molybdenum; mechanical P2 opened a via through the absorber so ZnO could contact the neighbouring molybdenum strip; mechanical P3 then isolated adjacent top electrodes. The repeated dead-zone is narrow compared with each active stripe, allowing monolithic voltage build-up without soldered ribbons between discrete cells.
The 1999 production package was still recognisably a framed module: tempered cover glass and EVA protected the circuit plate, a second EVA layer bonded a TPAT backsheet, and aluminium extrusions formed the frame. The backsheet and the deliberate offset between live circuit and frame provided electrical isolation and a semi-hermetic seal. The report also treats water-vapour ingress, edge deletion, glass breakage, film peeling and lamination bubbles as real manufacturing and durability constraints rather than hiding them behind the efficiency result.
Historical and scientific basis: Siemens Solar/NREL commercialisation and process report, 1999 Thin-Film Partnership CIS final report, earlier large-area SSI module design record, and NREL's 1999 CIS Solar Module R&D 100 timeline.
2000 · SECOND-GENERATION FACTORY-INTEGRATED AC MODULE
Make each photovoltaic module deliver utility-compatible AC
The SunSine AC 325 paired one large ASE Americas crystalline-silicon laminate with a compact grid-tied inverter fixed directly to its rear. Instead of exposing a module's direct-current terminals, the factory-integrated product presented an alternating-current connection ready to operate with other AC modules and the utility grid.
This is a 2000 redesign milestone, not the invention date of the AC module. Ascension Technology began module-scale inverter development in 1991; its original SunSine 300 entered production in September 1997 and first shipped that December. Under a later NREL manufacturing programme, Applied Power Corporation delivered the first redesigned prototype for outdoor testing in fiscal year 2000 and began a 110-unit pilot run, including 28 units installed at the University of Texas Houston Science Center.
The photovoltaic side was based on a modified ASE-300-DG/50 module: a framed double-glass laminate approximately 1.892 × 1.283 m containing 216 edge-defined film-fed growth (EFG) multicrystalline-silicon cells. The documented module family used tempered low-iron front glass, tempered rear glass and six bypass diodes. The diagram keeps these physical layers separate while reducing the visible cell count for clarity.
On the rear, a redesigned die-cast aluminium enclosure spread heat and protected the power electronics. A high-frequency PWM buck stage shaped a half-sine current from the PV source; a current-fed MOSFET H-bridge and transformer converted it to alternating current. Grid-voltage and frequency sensing, isolated control, a solid-state disconnect, electromagnetic-interference filtering, surge protection and a fuse completed the utility interface.
The final compliant design reduced the enclosure contact footprint from 169 to 93 in² and the total component count from 442 to 310. It achieved 88.0% full-power inverter efficiency and a 235 WAC PTC module rating after FCC emissions work. An earlier soft-switching prototype reached 90.8% before those modifications; the final report candidly concluded that the available data did not establish zero-voltage switching as the optimum approach. That distinction keeps the displayed performance tied to the production-intent configuration.
Historical and scientific basis: NREL/Schott Applied Power SunSine final report, NREL's fiscal-year 2000 manufacturing review, original SunSine 300 commercial-production report, and the ASE-300-DGF/50 manufacturer datasheet.
2001 · HIGH-THROUGHPUT COMMERCIAL CdTe MODULE
Move thin-film CdTe from pilot plates into continuous commercial manufacture
First Solar's FS-50C represented the transition from Solar Cells, Inc.'s earlier full-size CdTe prototypes to a listed, production-built 60 × 120 cm module. Its semiconductor circuit was deposited and laser-integrated directly on one glass plate, then sealed behind a second glass sheet and fitted with rear C-channel mounting rails.
The date marks initial manufacturing, not the invention of CdTe photovoltaics. SCI began CdTe research in 1991, produced the 0.72 m² module shown earlier in this timeline in June 1992 and began limited pilot production in 1994. A high-rate vapour-transport deposition (VTD) process developed in 1997 enabled the next step. First Solar opened its new production facility in 2000, and NREL records initial polycrystalline CdTe module manufacturing in 2001.
Production began with fluorine-doped tin-oxide-coated soda-lime front glass. An optional resistive buffer and thin n-type CdS window preceded the approximately 3–3.5 µm p-type CdTe absorber. VTD grew CdTe at roughly 1 µm per second. After CdCl2 activation, an interfacial layer and sputtered metals formed the rear contact.
Laser scribes divided the coatings into 116 narrow cells and connected them in series on the glass itself. The 1.2 m-wide production coater ran at about 2.5 m/min and was designed for four 0.72 m² coated plates per minute; finishing equipment operated at one plate per minute. This was the essential manufacturing advance: a complete high-voltage circuit emerged from sequential coating and patterning rather than from handling and soldering 116 separate cells.
EVA bonded a 3 mm rear-glass cover over the circuit. Two leads passed through the rear glass into a sealed cord plate, replacing the earlier potted-polyurethane pigtail termination. In April 2001 the FS-50C and FS-50Z completed the mounting tests needed for the UL Listed mark; the letters identify rear aluminium C-channel and Z-bar mounting attachments. California programme listing followed in August.
During the documented production period, sustained output was about 7% total-area efficiency, consistent with the 50 W product class, while a champion module reached 9.3% total-area efficiency. The diagram therefore shows a historically representative FS-50C architecture, not a later Series module and not the champion plate's exact measured specimen.
Historical and scientific basis: First Solar/NREL high-throughput CdTe final report, NREL's manufacturing-history and device-stack report, PVMaT qualification record for the FS-50C and FS-50Z, and NREL's Cord Plate and finishing-line development record.
2002–03 · MONOLITHIC WRAP-AROUND-CONTACT STRING RIBBON PROTOTYPE
Print the cell interconnect onto the backskin
Evergreen Solar's September 2002 manufacturing programme explored a different route to a low-cost crystalline-silicon module: grow String Ribbon silicon without sawing wafers, bring both cell contacts to the rear and bond the cells directly onto conductive-adhesive bars printed on a frameless polymer backskin.
This date is the start of the documented monolithic-module programme, not the invention of String Ribbon or the first frameless Evergreen module. Evergreen had already reported frameless glass/polymer modules in the 1990s. The 2002–03 step was the integration of wrap-around-contact (WAC) cells and a printed rear interconnection pattern intended to remove conventional wires between adjacent cells and their associated handling.
In String Ribbon growth, two high-temperature strings stabilised the edges of a thin silicon sheet pulled vertically from molten silicon. The process avoided wafer sawing and its kerf loss. The related Gemini programme grew two ribbons back-to-back from one crucible; by the final project report it had progressed from research through pilot operation to a 15 MW/yr production phase. That growth equipment is manufacturing context—the diagram deliberately shows only one module.
The module concept placed printed conductive-adhesive bars on the proprietary backskin. Front fingers wrapped around a cell edge so both polarities could be reached from behind, then bonded to the printed bars. A transparent encapsulant and glass superstrate completed the documented exploded stack. Early WAC String Ribbon cells reached 13.57% efficiency on 60 cm² half cells.
The report also records the concept's central limitation. String-containing ribbon edges were less flat than saw-cut wafer edges, allowing encapsulant to intrude between the wrap contact and adhesive. Work therefore moved the wrap to a flatter laser-cut edge and used half cells. Two 36-cell demonstrators made with saw-cut cast-polycrystalline cells retained essentially all power after 400 thermal cycles; that result validated the conductive adhesive, not a completed full-size String Ribbon module.
The reconstruction below follows the report's Figure 7 layer order and separates the front grid, wrapped contact and printed rear circuit so their functions remain visible. The visible cell count, gap size, backskin fold and terminal geometry are explanatory rather than claims about a surviving prototype.
Historical and scientific basis: Evergreen/NREL first annual String Ribbon manufacturing report (September 2002–March 2003), Evergreen/NREL final String Ribbon manufacturing report, and Evergreen/NREL Advanced Polymer PV System report.
2004 · COMMERCIAL INTERDIGITATED-BACK-CONTACT SILICON MODULE
Move every cell contact behind an all-black 210-watt module
SunPower's SPR-210 translated the A-300 interdigitated-back-contact (IBC) cell into a commercial 72-cell module. Moving both polarities and all metal fingers to the cell rear removed front-grid shading, created a uniform dark face and let the module interconnect remain in one plane behind the cells.
This date marks commercial module introduction, not the invention of the back-contact cell. SunPower's research lineage extended through earlier point-contact and rear-contact devices. A Texas pilot line operated from 2003; the 25 MW/yr Philippines factory was completed in May 2004, and SunPower's SEC filing records the first commercial A-300 cell shipments late that year. The SPR-210 product line was announced in June for fourth-quarter delivery.
A preliminary April 2004 supply specification defines 72 A-300 cells as six strings of twelve, a 210 W rating, 4 mm solar glass, a black 1000 VDC-proof TPE backsheet, black-anodised aluminium frame and an IP65 junction box with three Schottky bypass diodes. The specified module measured 1558 × 798 × 30 mm. NREL's May 2004 market survey listed the SPR-210 at 16.9% module efficiency—the highest entry in that survey.
Inside each n-type monocrystalline A-300 cell, alternating p-type emitter and n-type contact regions occupied the rear. Silicon-dioxide passivation covered both surfaces, with small rear contact openings limiting recombination. A textured, antireflection-coated front trapped incident light while rear metallisation also acted as an optical mirror. Because minority carriers had to diffuse to a rear junction, high carrier lifetime and excellent front-surface passivation were essential.
A separate 2004 conference paper reported a 32-cell research module at 95 W and 18.2% efficiency after NREL measurement, plus completion of IEEE 1262 qualification sequences and a 5 kW field array. Those results establish the cell and module engineering basis, but they are not the commercial 72-cell SPR-210's rating.
The reconstruction shows one SPR-210-class module and its exploded functional layers. The exact encapsulant chemistry, ribbon metallurgy, internal junction-box layout and frame extrusion were not disclosed in the cited preliminary specification, so those details are labelled generically or illustrative rather than reverse-engineered.
Historical and scientific basis: SEC-filed April 2004 SunPower–Conergy module supply specification, SunPower's 2005 SEC registration statement, NREL's 2004 commercial-module survey, SunPower's 2004 A-300 manufacturing paper, and the March 2005 A-300 manufacturer datasheet.
2005 · COMMERCIAL TANDEM THIN-FILM SILICON MODULE
Stack two silicon junctions to use more of the solar spectrum
Sharp's NA-8501P brought a glass-superstrate amorphous-silicon / microcrystalline-silicon tandem into commercial module production. The upper junction harvested higher-energy visible light; longer-wavelength light transmitted through it reached the lower microcrystalline junction.
This date marks Sharp's product introduction and industrialisation, not the invention of tandem thin-film silicon. Sharp's company history records the NA-8501P launch in 2005, and its production history fixes the start of tandem mass production at the Katsuragi Plant to September 2005 with 15 MW annual capacity.
The manufacturer documented a five-part optical stack: glass substrate, transparent electrode, amorphous-silicon cell, microcrystalline-silicon cell and backside electrode. A Sharp technical paper received in September 2005 describes textured ZnO made by sputtering and wet etching, and reports a 10.0% one-square-centimetre microcrystalline research cell after optimisation. That laboratory result explains the lower junction's materials science; it is not presented as the NA-8501P module efficiency.
The two junctions are electrically in series, so the same current must pass through both. Their absorber thicknesses, optical confinement and internal junction therefore have to balance photocurrent while limiting recombination. Hydrogenated amorphous silicon has a wider bandgap and suits the top cell; hydrogenated microcrystalline silicon extends response towards the red and near-infrared. Sharp reported that the tandem module achieved about 1.5 times the conversion efficiency of its conventional amorphous-silicon modules.
Thin semiconductor films were deposited directly on glass rather than sawn from wafers. Sharp stated that this reduced silicon use to roughly one hundredth of crystalline-silicon cells and shortened the manufacturing sequence. The continuous films could also be divided into long, series-connected cell stripes, giving useful module voltage without soldering a matrix of separate wafers.
The exploded view reproduces the documented active-layer order and a scientifically representative monolithic thin-film circuit. Sharp's surviving sources do not disclose the NA-8501P's exact rear encapsulation, frame profile, terminal-box geometry or product dimensions; those package elements are therefore labelled as functional reconstructions rather than original specifications.
Historical and scientific basis: Sharp's 100-Year History: NA-8501P introduction and documented stack, Sharp's tandem-production record, Sharp Annual Report 2008 tandem-layer diagram, Sharp Technical Journal 93 microcrystalline-silicon research, Sharp 2004-priority tandem-device patent, and NREL thin-film silicon tandem review.
2007–08 · MASS-PRODUCIBLE TRIPLE-JUNCTION THIN-FILM SILICON
Add a second amorphous junction before scaling the coated glass
Sharp's 2007 platform inserted a second amorphous-silicon photovoltaic junction above the microcrystalline-silicon bottom junction. It was therefore a true three-junction, two-terminal all-silicon device: a-Si:H / a-Si:H / µc-Si:H.
This date marks the documented development of mass-production technology, not the first scientific triple-junction experiment. Sharp's history records the achievement in 2007; by November that year the company reported that it was making the triple-junction cells. A January 2007 Sharp announcement, preserved by contemporary technical news services, placed the first production introduction in May.
Sharp reported that moving from one to two amorphous junctions raised voltage, while the microcrystalline junction retained response to longer-wavelength light and helped reduce light-induced efficiency loss. Cell efficiency increased from 11% for the preceding tandem platform to 13%; module efficiency rose from 8.6% to about 10%. Those are manufacturer-reported development values, not independent certification of the illustrative package below.
The three photovoltaic junctions are connected internally in series. Their voltages add, but one common current passes through all three, so absorber thickness, light trapping and the two recombination junctions must be designed together. The second a-Si:H junction does not simply duplicate output: dividing absorption between thinner upper junctions can increase voltage while limiting the stability penalty of one thicker amorphous absorber.
Manufacturing continuity was central to the change. Sharp reported using the same plasma-CVD production equipment as for its tandem cells. In October 2008, the Katsuragi scale-up began production on 1,000 × 1,400 mm glass substrates and lifted annual thin-film capacity from 15 to 160 MW. The large dimensions shown here refer to that production substrate, not to an independently verified finished-module outline.
The exploded view follows Sharp's published active order: glass, transparent electrode, two amorphous-silicon cells, one microcrystalline-silicon cell and back electrode. The two required internal series junctions and a representative monolithic stripe circuit are shown functionally. Rear sealing and terminal geometry remain explicitly reconstructed because the surviving Sharp sources do not disclose one complete 2007 production-module drawing.
Historical and scientific basis: Sharp's 2007 triple-junction history, Sharp Annual Report 2007, Sharp's November 2007 production release, Sharp's October 2008 large-substrate production release, Sharp Annual Report 2008 active-stack and production diagrams, and NREL thin-film silicon multijunction review.
2008 · COMMERCIAL CYLINDRICAL CIGS MODULE ARRAY
Replace the flat laminate with forty light-collecting tubes
Solyndra's Series 100 panel arranged forty approximately metre-long cylindrical photovoltaic modules in a spaced structural frame. Each module was a concentric glass-tube package containing a circumferential, monolithically integrated copper-indium-gallium-diselenide (CIGS) thin-film cell.
This date marks commercial shipment of the cylindrical system, not the invention of CIGS or tubular solar cells. Solyndra's SEC filing records low-volume commercial production and shipments beginning in July 2008 after safety certification in the first half of that year. The company had produced Series 100 prototypes in 2007, so the chronology uses the point at which the architecture became a sold module product.
Terminology matters here. Solyndra called one sealed photovoltaic tube a module and the rectangular forty-tube assembly a panel. The September 2008 manufacturer datasheet specified a 1.82 × 1.08 × 0.05 m panel weighing 31 kg, with initial models rated from 150 to 191 W. The assembled schematic enlarges ten representative tubes so their spacing and cylindrical topology remain legible; the documented production panel contained forty.
The active radial sequence began at the inner soda-lime glass tube: molybdenum back contact, CIGS absorber, a thin CdS junction-partner or window layer, and a transparent conducting oxide. Laser or mechanical scribes divided the circumferential coatings into more than one hundred serially interconnected cells along each tube. A transparent middle sleeve, index-matched optical coupling liquid and outer glass casing then surrounded the active tube.
The cylindrical surface could receive direct, diffuse and roof-reflected light around its circumference. Spacing admitted light to the rear of the tubes and let wind pass through the panel, allowing horizontal, low-profile mounting on reflective commercial roofs. Solyndra reported roughly 50% more light reaching the inner photovoltaic surface through its concentric optical package; that figure and the term “self-tracking” are manufacturer claims, not independent measurements reproduced here.
CIGS is moisture-sensitive, so packaging was part of the device physics. Stainless-steel end caps were fused into glass-to-metal hermetic seals, the tube was evacuated and filled with optical coupling agent, and every assembly was leak-tested. The exploded cutaway follows Solyndra's patents, SEC disclosures and a U.S. Customs manufacturing record; colours, apparent thicknesses, the reduced tube count and wiring route are explanatory.
Historical and scientific basis: Solyndra's SEC registration statement: July 2008 shipments and complete panel description, September 2008 Series 100 manufacturer datasheet, Solyndra patent US20070215195A1: concentric cell and package stack, Solyndra patent US20080110487A1: panel frame and elongated-module array, U.S. Federal Register: production materials, sleeve, seals and forty-tube assembly, and NREL CIGS layer-stack and monolithic-interconnection science.
2009 · FLEXIBLE MONOLITHIC CIGS ON POLYIMIDE
Move a complete thin-film circuit from rigid glass onto a roll
Ascent Solar's FAB1 platform deposited a copper-indium-gallium-diselenide (CIGS) photovoltaic stack on electrically insulating, high-temperature polyimide, then patterned neighbouring cell stripes into a serially connected flexible module.
This entry marks a limited-production and scale-demonstration milestone, not a mass-market launch or the invention of flexible CIGS. Ascent's 2009 annual filing records limited production beginning in the first quarter, a five-metre encapsulated module in June, and independent NREL verification of 10.4% efficiency on 429 cm² FAB1 modules in July. Flexible CIGS research and portable products pre-dated this platform.
The June demonstrator was reported as approximately five metres long, two kilograms and 123 W under standard test conditions, with 9.1% aperture-area efficiency. Those size, mass and power figures were internal Ascent measurements; the separately stated 10.4% result applies to smaller modules measured by NREL. The schematic keeps these two evidence sets distinct.
The insulating polymer web made monolithic series integration possible without first cutting the coated web into discrete cells and wiring them with ribbons. Ascent's patent describes a rear contact, photovoltaic stack and transparent front contact divided by isolation and via scribes. Insulating ink filled isolation channels, while conductive ink in a via and surface grid connected one cell's front contact to the next cell's rear contact.
The active sequence shown here follows the documented CIGS substrate geometry: transparent top contact, a thin window/buffer junction partner, CIGS absorber and molybdenum rear electrode on polyimide. Public 2009 records do not fix the production window-layer chemistry or flexible barrier formulation, so those regions are labelled functionally rather than assigned an unsupported material recipe.
Removing glass reduced mass and allowed a long roll-format product, but it also removed glass's excellent moisture barrier. Ascent reported an internally qualified flexible package that completed 1,000 hours of 85 °C/85% relative-humidity damp heat in August 2009. The exploded front and rear barriers therefore represent a documented packaging function; their apparent thicknesses, colour, exact polymers, lead geometry and number of visible cell stripes are explanatory.
Historical and scientific basis: Ascent Solar 2009 Form 10-K: limited production, five-metre demonstrator, efficiency and damp-heat chronology, Ascent patent US20080314439A1: monolithic isolation, via and printed-grid architecture, Ascent patent family WO2008157807: flexible monolithic thin-film cell array, NREL FY2009 report: flexible CIGS module science and moisture-reliability challenge, NREL/ITN/Global Solar report: roll-to-roll CIGS deposition and layer sequence, and U.S. DOE 2008–2012 plan: flexible CIGS manufacturing and encapsulation context.
2011 · COMMERCIAL METAL-WRAP-THROUGH MODULE
Move both cell polarities to the rear—while retaining a light front grid
Canadian Solar launched its ELPS metal-wrap-through (MWT) module family in September 2011 after first showing the cells that June. The architecture retained p-type mono- or multicrystalline silicon, but routed the front emitter contact through laser-drilled, metallised holes so both electrical polarities could be interconnected from the rear.
This was a commercial module-architecture milestone, not the invention of MWT. Earlier research had already established wrap-through cells. ELPS matters here because Canadian Solar documented commercial assembly beginning in the third quarter of 2011 and completed IEC 61215, IEC 61730 and UL 1703 certification for MWT-cell modules that year.
Fine silver fingers still collected electrons across the illuminated face, so ELPS was not an interdigitated-back-contact cell. Selected fingers converged on conductive vias; screen-printed silver filled the holes and carried the emitter contact to negative pads on the rear. Aluminium rear metallisation and positive pads collected the base contact.
At module level, rear-side soldered ribbons joined the two rear polarities into the series circuit. Removing broad front busbars exposed more active area, shortened current paths and could reduce series resistance. The launch material reported cell efficiencies up to 19.5%, module power up to 265 W and module efficiency above 17%.
The manufacturing challenge moved into the wafer and rear interconnect: laser damage and edge isolation had to be controlled, vias had to remain conductive, rear pads had to align accurately, and soldering could not crack the thin cells. The drawing separates those functions without claiming production dimensions or a proprietary ribbon route.
Architecture boundary: this reconstruction follows the documented ELPS rear-ribbon module. It does not substitute the separate ECN conductive-foil and conductive-adhesive MWT module family, even though both place cell polarities on the rear.
Historical and scientific basis: Canadian Solar, 1 September 2011: official ELPS launch, MWT operation and performance; Canadian Solar 2011 Form 20-F: Q3 commercial assembly, rear-side soldering and 2011 certification; Canadian Solar investor presentation: p-type mono / multi ELPS cell, connecting hole and rear point contacts; Lamers et al. (2011): MWT cell process, SiNx:H, emitter, Ag via and rear metallisation; MWT module patent: rear interconnection principles; and NREL: MWT via, isolation and manufacturing development.
2012 · MICROCELL HIGH-CONCENTRATION MODULE SCALE-UP
Print microscopic III–V cells beneath a two-stage lens array
In January 2012 Siemens and Semprius announced a 33.9% module-efficiency result for Semprius’s high-concentration photovoltaic module. The result had been certified by IES-UPM and corroborated outdoors by ISFOC, while a North Carolina pilot-production facility was being prepared for operation later that year.
This is a 2012 record-announcement and manufacturing-scale-up milestone, not the invention date of concentrated photovoltaics or micro-transfer printing. Semprius had published the underlying two-stage optical and transfer-printed receiver architecture in 2010, and the 33.9% result was already being measured and discussed during 2011.
The module replaced a continuous field of large cells with an array of optical channels. A plano-convex silicone-on-glass primary lens focused direct sunlight towards a tiny glass-ball secondary lens. That secondary optic homogenised and redirected the concentrated beam onto a GaAs-based multijunction microcell mounted on a small receiver.
The 2012 record-module description reported triple-junction microcells approximately 600 µm by 600 µm and less than 10 µm thick, covering only about 0.1% of module area while receiving roughly 1,000-fold concentrated light. The precise alloy sequence of that record cell was not disclosed in the announcement, so the drawing labels its three bandgap junctions functionally rather than assigning later material compositions.
Microcells were grown epitaxially above a release layer, lifted from the GaAs source wafer and printed in parallel onto receiver substrates. Reusing the source wafer reduced compound-semiconductor use; the microcell scale shortened current and thermal paths, allowing heat to spread through the distributed receiver and backplane without a separate heat sink at every cell.
High concentration also imposed strict system conditions. Only direct-normal irradiance can be focused, each primary lens must remain aligned with its secondary lens and microcell, and the module must operate on a two-axis tracker within a narrow angular acceptance. Diffuse skylight largely passes through the optical train without reaching the microcells.
Historical and scientific basis: U.S. Department of Energy: 2012 Semprius 33.9% HCPV module award; Siemens Energy Sector, 31 January 2012: independent 33.9% module result and pilot-production plan; NREL PV Incubator final report: 3J cells, transfer yield, module scale-up, enclosure construction and 2012 plant; NREL Concentrator PV overview: microcell area, primary and secondary optics, transfer printing and 1,111× prototype; DOE final technical report: two-stage pupil-imaging optics, microcells, tracking systems and passive thermal strategy; and Furman et al. (IEEE PVSC 2010): original Semprius module and receiver architecture.
2014 · FIRST SOLID-STATE PEROVSKITE MODULE
Scale a new absorber into five monolithically connected cells
On 7 January 2014, Matteocci and colleagues published what they identified as the first solid-state modules based on an organometal-halide perovskite. The work had been submitted in August 2013, only about a year after efficient solid-state perovskite cells emerged, making module integration part of the technology’s earliest development.
The laboratory module divided one FTO-coated glass substrate into five series-connected subcells. Their combined active area was 16.8 cm2 inside a 25 cm2 aperture. Both the P3HT and spiro-OMeTAD variants reached 5.1% active-area efficiency; the later process analysis reports a 67% aperture ratio for this first design.
Light entered through glass and fluorine-doped tin oxide (FTO). Compact TiO2 blocked direct contact and collected electrons; screen-printed mesoporous TiO2 provided a scaffold infiltrated by CH3NH3PbI3−xClx. A perovskite-rich overlayer absorbed additional light, P3HT or spiro-OMeTAD transported holes, and evaporated gold formed the rear electrode.
The module’s important architectural step was its integrated series connection. A 1064 nm Nd:YVO4 laser made the P1 cut through FTO. Compact TiO2 was lifted off at the interconnection, while perovskite and hole-transport layers were chemically removed to expose FTO for the P2 contact. A patterned gold deposition completed the connection and kept neighbouring rear electrodes separated at P3.
This first layout also exposed the scaling problem. Wide manually cleaned interconnection lanes consumed aperture area, while larger coated areas increased the chance of pinholes, thickness variation and series resistance. Later modules replaced much of the manual patterning with tighter P1–P2–P3 laser processes.
The device was a research module rather than a weather-sealed commercial panel. Under continuous one-sun testing in air, the P3HT version lost about 80% of its initial efficiency after 170 hours, while the spiro-OMeTAD version retained more than 60% after 335 hours. The result therefore established feasibility while clearly revealing the need for improved materials, patterning and encapsulation.
Historical and scientific basis: Matteocci et al., PCCP 2014: first solid-state organometal-halide perovskite modules, 5.1% efficiency and stability tests; University of Rome Tor Vergata CHOSE: materials and large-area fabrication summary; Palma et al., IEEE Journal of Photovoltaics: retrospective reconstruction of the first module’s P1, P2 and P3 process, area and aperture ratio; and Green, Ho-Baillie and Snaith, Nature Photonics 2014: chronology of the rapid perovskite-PV emergence.
2015 · FIRST FLEXIBLE PEROVSKITE MODULE
Move the monolithic perovskite stack from glass onto plastic
Published on 7 January 2015, Di Giacomo and colleagues reported the first demonstrated flexible perovskite photovoltaic module. The change was architectural as well as mechanical: a complete monolithic module had to be fabricated on heat-sensitive PET/ITO rather than rigid conductive glass.
The 5.6 × 5.6 cm plastic substrate carried four series-connected cells. Each cell had 1.98 cm2 of active area, giving 7.92 cm2 in total. The unsealed research module reached 3.1% efficiency, a 71% fill factor and 3.4 V open-circuit voltage; the best large flexible cell reached 4.3%.
Its documented stack was PET/ITO/compact TiO2/mesoporous TiO2/CH3NH3PbI3−xClx/spiro-OMeTAD/Au. Plasma-assisted atomic-layer deposition formed an approximately 11 nm compact TiO2 contact at 150 °C. A printable TiO2 scaffold was then UV-treated and heated at only 140 °C, avoiding the high-temperature sintering used on glass.
Patterning was adapted to the plastic substrate. Protective masks and a scanning CO2 laser isolated ITO and the compact oxide. Mesoporous TiO2, perovskite and hole transporter were removed from connection lanes, and shadow-masked gold completed the four-cell series circuit.
The result joined two early perovskite strengths—thin solution-processed absorbers and monolithic voltage addition—with a lightweight, bendable support. Its modest efficiency came partly from non-uniform flexible-cell output and current mismatch: the lowest-current stripe constrained every cell in the series string.
This was a laboratory demonstration, not a roll-to-roll product or an outdoor-qualified laminate. A separate cell made with the same architecture retained its initial efficiency through 300 bends at radii of 30, 20 and 15 mm; tighter bending mainly damaged the brittle ITO. The paper did not report cyclic bending of the complete four-cell module, so the schematic keeps that evidence distinct.
Historical and scientific basis: Di Giacomo et al., Advanced Energy Materials 2015: original first flexible perovskite module report; Open full paper: exact stack, four-cell geometry, fabrication, electrical results and cell bend test; HOPV15 scientific record: plastic-compatible low-temperature process and 3.1% module; and Li et al., Energy & Environmental Science 2016: flexible-perovskite development and module context.
2016 · PRINTED ORGANIC MODULE WITH RECOMBINATION INTERCONNECTS
Use continuous transport layers to recover active module area
Published on 5 January 2016, Hong and colleagues demonstrated a new series-connection architecture for printed organic photovoltaic modules. Instead of patterning every functional layer into separate stripes, the design left both charge-transport layers continuous and used their direct junction to connect neighbouring cells.
The 60 × 44.5 mm research module contained three inverted organic subcells on patterned ITO glass. Its certified module efficiency was approximately 7.5% over 4.15 cm2, with a 90% geometric fill factor. The best laboratory measurement reached 8.1% active-area efficiency, 7.3% total-area efficiency and 2.36 V open-circuit voltage.
The active stack was glass/ITO/ZnO/PTB7-Th:PC70BM/MoO3/Ag. PTB7-Th donated electrons after light absorption, while the fullerene PC70BM accepted them inside an approximately 125 nm bulk-heterojunction blend. Sol-gel ZnO selected electrons, MoO3 selected holes, and patterned silver formed the rear anodes.
The architectural novelty sat in the series-connection region. Where the photoactive blend was absent, continuous ZnO and MoO3 touched between overlapping ITO and silver electrodes. Electrons and holes arriving from adjacent subcells recombined at this transport-layer junction, allowing the neighbouring cell voltages to add without a direct metal-to-metal connection.
Only the ITO, photoactive blend and silver required stripe definition. The 13.5 mm active stripes were slot-die coated with 0.5 mm offsets; ZnO and MoO3 could be deposited as unpatterned films using blade-compatible processes. This removed additional post-patterning steps and reduced the inactive interconnection width that commonly penalised printed modules.
This was a glass-based laboratory module, not a flexible encapsulated commercial panel. Its contribution was a manufacturable interconnection concept: it showed how bulk-heterojunction organic chemistry and continuous selective contacts could preserve 90% of the aperture while retaining approximately three times the voltage of one subcell.
Historical and scientific basis: Hong et al., Nature Communications 2016: original open-access architecture, materials, printing and certified module result; Original paper PDF: figures, TEM cross-sections, fabrication methods and electrical measurements; DOI record: publication chronology and article provenance; and Andersen et al., Energy & Environmental Science 2014: preceding scalable organic-module manufacturing context.
2017 · COMMERCIAL-SCALE SHINGLED-SILICON MODULE
Turn cut silicon strips into dense, ribbon-free supercells
Shingled photovoltaic cells were not invented in 2017: the principle appeared in a 1956 patent and later in specialist space and solar-racing modules. The 2017 milestone was industrialisation. SunPower was scaling its Performance Series production, publishing production-oriented patent architecture and presenting shingled panels as a commercial platform rather than a niche construction.
Processed crystalline-silicon cells were divided into narrow rectangular strips, approximately one inch across in the contemporary product description. Each strip overlapped the long edge of the next like roof tiles. Electrically conductive adhesive joined the rear contact of one strip to the front contact of its neighbour, creating a long series-connected supercell.
Overlap hid the front busbar beneath the next active strip and removed the conventional copper ribbons and gaps from the illuminated field. Narrower strips also carried less current than full cells, reducing resistive loss. The 2017 P-Series presentation listed framed multicrystalline panels at 350 W and 17.1% efficiency and a framed monocrystalline configuration at 400 W and 19.4%.
The published patent example arranged six long supercells as parallel rows. Each row contained many series-connected strips; transverse rear conductors tied the six rows together and connected intermediate circuit sections to bypass diodes in the junction box. This parallel-series topology redistributed current under partial shading while retaining module-scale voltage.
Packaging remained recognisable: aluminium perimeter frame, transparent front glass, front encapsulant, the shingled supercell circuit, rear encapsulant, polymer backsheet and rear junction box. The architectural change occurred inside the laminate, where high cell packing and compliant adhesive contacts replaced ribbon-tabbed full wafers.
Shingling also introduced a different reliability problem. Adjacent strips are tightly constrained, so thermal expansion creates shear in the thin adhesive joint rather than bending a free copper ribbon. Contemporary scientific analysis therefore favoured electrically conductive adhesives with sufficient compliance and shear strength over a stiff solder joint. Cell cutting, edge damage, adhesive cure, overlap control and moisture protection became central manufacturing controls.
Historical and scientific basis: SunPower patent publication US20170213929A1: shingled strips, ECA joints, six-row topology, laminate and rear interconnections; SunPower 2017 Performance Series technical presentation: module package, 350/400 W configurations, one-inch strips and reliability concept; SunPower 2017 Form 10-K: Performance Series launch and production ramp; Beaucarne, Energy Procedia 2016: peer-reviewed ECA joint mechanics and historical context; and Fraunhofer 2017: scientific review of shingled-cell packing, resistance and conductive-adhesive interconnection.
2018 · HALF-CUT BIFACIAL MONO-PERC MODULE
Split the cell current and admit useful light through the rear
Neither half-cells nor bifacial photovoltaics began in 2018. The milestone was their documented convergence in a mainstream commercial module platform. LONGi launched Hi-MO3 at SNEC on 28 May 2018 as a half-cut bifacial p-type monocrystalline PERC module, rated up to 320 W from the front with bifaciality above 75%.
A conventional square PERC cell was separated into two equal half-cells. Because each piece carried roughly half the current of the original cell, resistive heating in cell metallisation and interconnecting ribbons fell in proportion to I2R. The nominal 60-cell format therefore used 120 half-cell pieces without doubling the module voltage.
The circuit was divided across the module centre. Two mirrored 60-half-cell groups operated in parallel, while each group retained series-connected substrings. Three paired substring stages and three bypass diodes preserved the familiar output voltage but reduced conductor current and changed the module's response to partial shading.
The p-type mono-PERC cells used a front phosphorus emitter and passivated rear surface. Unlike a monofacial PERC cell with an opaque full-area aluminium rear, the bifacial version used patterned rear metallisation and exposed passivated silicon between the fingers. Rear irradiance reflected from the ground could therefore enter through a transparent rear sheet and produce additional current.
The package was correspondingly double-glass: aluminium frame, coated front glass, front encapsulant, the 120-piece half-cell circuit, rear encapsulant, transparent rear glass and a rear junction-box system. Front and rear optical transmission, cell spacing, mounting height, ground reflectance and rack shading all influenced the extra bifacial yield.
Independent 2018 modelling found that the two-parallel-string half-cell arrangement could reduce the maximum simulated hot-spot temperature by 28 °C compared with a 60-full-cell series module under the studied shading case. The architecture also created new controls: laser-cut edge damage, half-cell alignment, ribbon fatigue, centre-bus insulation, rear-side potential-induced degradation and non-uniform rear irradiance.
Historical and scientific basis: LONGi, 28 May 2018: Hi-MO3 half-cut bifacial p-mono PERC launch, front rating, bifaciality and power rationale; Qian et al., IEEE Journal of Photovoltaics 2018: two parallel 60-half-cell strings and hot-spot comparison; Fraunhofer CSP: measured and simulated half-cell shading, optical gain and connector-loss basis; Janssen et al., Solar Energy Materials and Solar Cells 2018: bifacial glass-glass optics, temperature and rear-irradiance behaviour; and LONGi installation documentation: framed bifacial double-glass module construction and LR6 series context.
2019 · MASS-PRODUCED HALF-CUT SILICON HETEROJUNCTION MODULE
Passivate crystalline silicon with carrier-selective amorphous layers
Silicon heterojunction cells pre-date 2019, and Sanyo had commercialised HIT modules decades earlier. The 2019 milestone was a new module combination at production scale: REC launched Alpha in May and began mass production that October, documenting it as the first mass-produced panel to combine silicon heterojunction cells with a half-cut format.
Each cell placed a phosphorus-doped n-type monocrystalline-silicon wafer between nanometre-scale intrinsic and doped hydrogenated amorphous-silicon layers. The intrinsic films passivated dangling bonds at the crystalline surface. Boron-doped p-type amorphous silicon on the illuminated side and phosphorus-doped n-type amorphous silicon on the rear acted as carrier-selective contacts, suppressing recombination while establishing the junction.
Transparent conductive oxide layers spread current laterally above the poorly conducting amorphous films and also contributed to optical management. Fine printed fingers transferred charge into a low-temperature foil-and-round-wire interconnection. Many wire contact points reduced the distance current travelled in the fingers and avoided the high thermal load of conventional ribbon soldering.
The commercial panel retained REC's centre-split Twin architecture: 120 rectangular half-cells formed six strings of twenty cells, divided into upper and lower sections connected in parallel. Three bypass-protected series stages produced a 60-cell-class module voltage while lowering current in each half-cell branch.
The documented package used a black anodised-aluminium frame, 3.2 mm antireflection-treated solar glass, encapsulant, the HJT half-cell circuit, rear encapsulant, a resistant polymer backsheet, rear support bars and a three-part junction-box system with three bypass diodes. The first Alpha range reached 380 W and 21.7% module efficiency in a 1.75 m² panel.
HJT processing occurs at comparatively low temperature and its n-type wafer avoids boron–oxygen light-induced degradation. The 2019 datasheet reported a power temperature coefficient of −0.26% per degree Celsius. Production nevertheless required tight control of amorphous-silicon passivation, TCO deposition damage and resistance, low-temperature metallisation, foil-wire contact pressure, encapsulant compatibility, moisture ingress and backsheet adhesion.
Historical and scientific basis: REC, September 2019: Alpha launch, 120 half-cut HJT cells, advanced connection and 380 W / 21.7% result; REC, October 2019: Alpha mass-production start on a 600 MW HJT cell-and-module line; Dated REC Alpha datasheet: six strings of twenty, glass-backsheet package, three-part junction box and certified dimensions; REC Alpha technical retrospective: HJT layer functions, low-temperature foil-wire connection, Twin circuit and support-frame design; Faes et al., Energy Procedia: peer-reviewed HJT metallisation for SmartWire module interconnection; Battaglia et al., Energy & Environmental Science: silicon heterojunction band structure and high-efficiency cell physics; and Arriaga Arruti et al., Progress in Photovoltaics: scientific review of HJT module reliability and interconnection risks.
2020 · VOLUME-PRODUCED M10 LARGE-FORMAT UTILITY MODULE
Scale the cell, interconnection and package as one system
PERC cells, bifacial modules and half-cell circuits all pre-date 2020. The milestone was their coordinated move to a much larger industrial format. On 24 June seven manufacturers proposed the 182 mm M10 wafer standard; LONGi released its Hi-MO 5 technical brief on 29 June and started volume production on 8 September.
The LR5-72HBD reference module used 144 rectangular half-cells in a 6 × 24 orientation. Each half came from a 182 mm gallium-doped p-type monocrystalline wafer and retained a bifacial PERC structure: a phosphorus-diffused front emitter, front antireflection/passivation, a passivated rear surface and local rear grid contacts that admitted reflected light.
Gallium replaced boron as the p-type dopant to avoid the familiar boron–oxygen light-induced degradation mechanism. This is not the same as complete immunity to ageing: contemporary peer-reviewed work found that fired gallium-doped PERC material could still undergo light- and elevated-temperature-induced degradation, making hydrogenation, firing and stabilisation conditions important production controls.
Nine front busbars and segmented solder ribbon formed the cell interconnection. A triangular ribbon section above each active cell redirected some incident light towards the silicon; the section crossing the approximately 0.6 mm inter-cell gap was flattened and softened. The geometry reduced inactive spacing and tensile stress without electrically overlapping adjacent cells like a shingle.
The bifacial utility package used a 35 mm anodised-aluminium frame, 2.0 + 2.0 mm heat-strengthened glass, transparent encapsulation, the 144-half-cell circuit and an IP68 junction-box system with three bypass diodes. The 2020 technical brief documented 2256 × 1133 mm, 32.3 kg, 540 W, more than 21% efficiency, 49.5 V open-circuit voltage and approximately 13 A operating current.
This architecture helped establish 182 mm as a utility-module format and shifted design attention beyond cell efficiency. Wafer fracture, cut-edge damage, ribbon fatigue, glass deflection, tracker compatibility, inverter current limits, manual handling and transport packing all had to be evaluated together as module dimensions and current increased.
Historical and scientific basis: LONGi chronology: seven-manufacturer M10 proposal on 24 June, product release on 29 June and production start on 8 September 2020; LONGi, 29 June 2020: Hi-MO 5 technical-brief release, M10 gallium-doped PERC, 540 W, 21%+ and certified dimensions; LONGi, 9 September 2020: first Hi-MO 5 production module and volume-production milestone; Dated 2020 Hi-MO 5 technical brief: 182 mm M10, half-cell 9BB format and electrical ratings; Contemporary LR5-72HBD datasheet: 144-cell orientation, 2+2 mm glass, frame, mass and dimensions; LONGi segmented-ribbon white paper: triangular optical section, flat micro-gap connection and mechanical testing; and Ciesla et al., Solar Energy Materials and Solar Cells: peer-reviewed degradation study of industrial gallium-doped PERC cells.