Article Post

Photovoltaic Technologies – Past, Present and Future - Part 2

This is the second part of a paper. The first part was posted yesterday. Link to part 1 is below:

1.Present into the Future

The current state-of-the-art efficiency for silicon-based commercial solar panels is between 22 and 23%, and yet the best efficiencies for research cells is 46% for concentrated solar, and 38.8% for non-concentrated (see chart below).[1] Some of these cells use exotic chemistries that are too expensive for commercial applications currently using silicon PV panels. Also the lenses or reflectors required for sunlight concentration are currently too expensive for most commercial applications. However some current research cell techniques could evolve to produce future cost-effective solar panels. Some of these will be explored in this section.

The newer technologies reviewed in this section are facing major competition in that silicon technologies have become very efficient, and continue to incrementally improve (see the next subsection). For that reason, initial developments in multi-junction or tandem cells will probably incorporate silicon, and these hybrids may set the stage for non-silicon designs when production techniques for other technologies mature.

1.1.Recent Silicon Solar Cell Advancements

Two manufacturers have demonstrated "module-level" efficiency significantly exceeding 20%:

  • Panasonic has demonstrated 25.6% efficiency in a module larger than 100 cm2. This is Panasonic's HIT (heterojunction) silicon cell. The figure below explains how heterojunction cells work.[2]
  • Kaneka Corporation (Japan) has demonstrated 26.33% efficiency in a 180 cm2 module. This also uses a heterojunction silicon cell.[3]

    Indications are that Tesla/SolarCity will produce the Panasonic HIT modules in their Gigafactory 2 in Buffalo, NY, and use them in their current deployments plus their future solar roof tiles. Panasonic is a partner in both Gigafactory 1 and Gigafactory 2.[4]

    Current commercial HIT panels from Panasonic are rated at just below 20% efficiency. Indications are from Panasonic (and Kaneka) that the demonstrated heterojunction technologies (bullets above) can be ramped up to volume production without substantially impacting production costs. This would result in panel-efficiencies significantly above 20% at only a small price-premium.

1.2.Multi-Junction Solar Cells

The "junction" in multi-junction refers to the P-N junction in a solar cell where electrons and holes are generated (see the figure in the prior subsection). The reason for having multiple junctions is that (1) each junction is limited to a theoretical light-conversion efficiency of about 34% by the Shockley–Queisser limit and (2) each junction can use materials that allows it to respond to different parts of the sunlight's spectrum. The Shockley–Queisser equation describes all of the elements that limit the efficiency, and the largest of these is the inability of a single junction to use all of the sun's spectrum. Thus by stacking multiple layers, each converting a different part of this spectrum, the Shockley–Queisser limit can be exceeded.

The figure below shows the response of a four-junction research cell produced by Fraunhofer Institute for Solar Energy in Germany, SOITEC S.A. and CEA-LETI, both in France. Note that the measured efficiency of this design is 44.7% for a research cell. Also, this design required concentrated solar light (297 suns) to achieve this efficiency.[5]

External Quantum Efficiency of a wafer bonded 4-junction GaInP/GaAs//GaInAsP/GaInAs solar cell measured on a concentrator solar cell device, the EQE is superimposed with the terrestrial concentrator normal spectrum

Note that multi-junction cells are different than the tandem cells mentioned in section 3.5. The former are tightly integrated into a single structure, whereas the latter are effectively stacked cells.

1.3.Concentrated Photovoltaics

Concentrated PVs (CPVs) have carved out niches that exploit their unique capabilities. They are the most efficient designs (see the prior subsection on multi-junction designs), but have specific site requirements, are more expensive (per watt) than silicon flat-panel designs and have other unique requirements. Nevertheless they have developed a significant market in large utility-scale projects and currently have >370 MW of installed projects (compare to 320 GW globally for all PV per Fraunhofer ICE).

Recent projects include several with capacity greater than 30 MW:[6]

  • Golmud, China, built by Suncore: 60 MW (2012) and 80 MW (2013) (the Golmud Project is the main image for this paper.)
  • Touwsrivier, South Africa, built by Soitec: 44 MW (2014)
  • Alamosa, Colorado, US, built by Amonix: 30 MW (2012)

    Note that the above reference (from Fraunhofer ICE and NREL) provided much of the information in this subsection.

1.3.1.Types of CPVs

There are basically two levels of solar concentration. High concentration PVs (HCPV) focus the solar light to 300 - 1,000 suns. Low concentration PVs (LCPV) focus the solar light to less than 100 suns. The differences in technology and requirements are described below. concentration PVs (HCPV)

HCPV is used in over 90% of existing projects. HCPV can only be used in areas where the direct normal irradiance (direct sunlight) is a very high percentage of the overall solar radiation since HCPV cannot use diffuse irradiance. This means sites are limited to arid regions with few cloudy times or days. In order to achieve the high concentrations, dual-axis trackers must be used to keep the plane of the Fresnel lens or mirror used to concentrate the sunlight normal (at right angles) to the direct sunlight. HCPV use three-to-five-junction cells similar to those described in subsection 3.2 (4-junction GaInP/GaAs//GaInAsP/GaInAs). Even though these are substantially more expensive than the silicon cells used in conventional solar panels, the level of solar concentration mostly offsets this difference.

A bottom-up analysis from NREL in 2014 based on a specific HCPV system with a Fresnel lens primary optic and refractive secondary lens estimated the total capex for cells and modules to be around $0.55/W. This is comparable to 2015 costs for a solar module (panel) in utility-scale projects as seen in the first figure in section 1. Of course very few conventional solar projects use dual-axis trackers, which are much more expensive than single axis trackers or fixed-tilt panel support structures.

One final strength of a HCPV system is that, due to its use of two-axis trackers and only using direct normal irradiance, it has a higher output in early morning and late afternoon than systems using conventional solar panels and either no trackers or single-axis trackers. concentration PVs (LCPV)

LCPV are currently used in less than 10% of installed CPV projects, and this share currently appears to be declining. LCPV has some capital cost advantages over HCPV, including:

  • LCPV modules typically use monocrystalline silicon. Although typical designs are specifically for LCPV, they use the same basic processes as monocrystalline silicon used in conventional PV panels. Typical module efficiencies are 20% to 25%.
  • LCPV can use single axes trackers, although dual-axis trackers provide higher efficiencies.
  • LCPV does use some diffuse irradiance, and thus can be used is areas that are marginal for HCPV.

1.4.Dye Sensitized Solar Cells

Dye Sensitized solar cells (DSSC) have been developed to the point that they are being sold by several firms. They have been under development for over 20 years and appear to have found a reasonable market niche: small cells for stand-alone consumer devices that are mainly used indoors. They have several characteristics that make them ideal for these types of applications, including:

  • Photon to electron conversion over a wide light spectrum
  • Conversion under low-light conditions
  • Lighter and more robust than silicon cells

    The operating principal is completely different than silicon solar cells. Instead of using P and N layers to produce holes and electrons, current DSSC designs uses a sandwich of electrolyte, titanium oxide, and dye to create electric current. The present best efficiency of production DSSC technology seems to be in the 10% to 15% range.

    Although DSSC technology continues to improve, it seems to be mainly focused on the market described at the beginning of this subsection. With sufficient performance improvement it may move into other markets.

1.5.Perovskite Photovoltaic Cells

According to the dictionary, perovskite is a mineral that is an oxide of calcium and titanium that generally contains other elements. Perovskite lends its name to the class of materials which have the same type of crystal structure as CaTiO3, known as the perovskite structure.[7]

Perovskite materials are used in the absorber layer of Photovoltaic (PV) cells (converts light to electric current). The most commonly used perovskites in these cells are methylammonium lead trihalide (CH3NH3PbX3, where X is a halogen ion such as I, Br, Cl).

The main reasons that perovskite cells have attracted much attention include:

  • In large production quantities, perovskite cells appear to be much simpler and less expensive to process than silicon cells. They can also be “printed” on a wide range of materials.
  • Perovskite cells are increasing their performance at an incredible rate. They were initially produced in 2009, and these initial cells had an efficiency of 3.8%. Current perovskite cells have an efficiency of well over 20%.
  • There are thousands of different variants of perovskite materials. This facilitates refining the PV cell design (with attendant performance increases) for many years.
  • Perovskites possess intrinsic properties like broad absorption spectrum, fast charge separation, long transport distance of electrons and holes and long carrier separation lifetime.
  • Perovskite are relatively easy to incorporate into conventional silicon cells as “tandem designs”. These use a silicon layer to harvest the red part of the light spectrum (which it prefers), and a perovskite layer to harvest the green and blue part of the spectrum. The tandem silicon/perovskite cells will probably be the initially produced products (see next section).

    The main problems with this technology are:

  • Withstanding high temperature, high humidity and other operational conditions
  • Lifetimes much shorter than silicon PV cells.

These problems are related, and rapid progress in being made in addressing them.

The following groups are the major researchers in, and early-stage developers of Perovskite technology.

1.5.1.Oxford Photovoltaics (a.k.a. Oxford PV)

Oxford PV appears to be the closest developer to shipping product.  There is a link to their website below:

They have already purchased a production facility.

The first product will probably be silicon/perovskite tandem cells. Co-founder of Oxford PV, Dr. Henry Snaith, indicated that they expected initial production (first production-prototypes) in 2018.

Oxford PV is also working on Building-Integrated PV (BIPV). Both BIPV and tandem-cells take advantage of the ability to print Perovskite cells on glass (BIPV) or silicon cells (tandem-cells) very inexpensively. BIPV will take the place of conventional building glazing, and generate a substantial amount of power.

1.5.2.Swiss Federal Institute of Technology (Ecoles Polytechniques fédérales de Lausanne (EPFL))

Although it appears that perovskite photovoltaics will be relatively easy to produce, especially as added layers to silicon cells (tandem cells), there are several remaining challenges in producing either a pure perovskite cell or a multi-layered perovskite cell (multiple absorber perovskite layers, each capturing a different light-spectrum) as described above.

Perovskite cell technology evolved from die sensitized solar cells (DSSC), as many of the processing methods and issues to overcome are very similar with perovskite cells and DSSC. A recognized leader in the development of DSSC was EPFL, and they have extended this capability to perovskite cells.

EPFL has demonstrated a production method that resulted in a 1 cm2 perovskite cell with an efficiency greater than 20%.

Another challenge is to identify a cost effective and stable material for the transparent P-type layer (hole-transport or cathode), directly above the perovskite-layer. EPFL has reported that they have developed a P-type layer material that is estimated to cost a fifth of existing materials and has the potential for much lower degradation by weather and biological agents

1.5.3.Greatcell Solar Limited (GSL)

Like EPFL (above subsection), Greatcell Solar, an Australian company, entered the perovskite cell development field via die sensitized solar cells. GSL has relationships with EPFL and other perovskite pioneers.

Greatcell’s primary function is to develop materials and manufacturing processes for production of cells. Currently they are partnering with Commonwealth Scientific and Industrial Research Organization (CSIRO), the University of New South Wales and other organizations to manufacture and test perovskite cells. They have recently signed a memorandum of understanding to transfer technology to JinkoSolar, a Chinese manufacturer.


[1] Research Cell Efficiency Records, U.S. Department of Energy, NREL, September 2015

[2] "Photovoltaic Module HIT®: Converting Sunlight into Electricity with the World’s Highest Conversion Efficiency",

[3] Kaneka News Release: World’s Highest Conversion Efficiency of 26.33%, Achieved in a Crystalline Silicon Solar Cell, Sep 2016,

[4]  Mark Osborne, "Tesla/SolarCity/Silevo/Panasonic 1GW Buffalo fab’s known unknowns", Jan 2017,

[5]AIP Conference Proceedings 1616, 45 (2014); doi: 10.1063/1.4897025

[6] Maike Wiesenfarth, Dr. Simon P. Philipps, Dr. Andreas W. Bett: Fraunhofer Institute for Solar Energy Systems (ISE), Kelsey Horowitz, Dr. Sarah Kurtz: National Renewable Energy Laboratory (NREL), "Current Status of Concentrator Photovoltaic (CPV) Technology", Version 1.3, April 2017,

[7] Wikipedia, “Perovskite”.


No discussions yet. Start a discussion below.