PV: where are we now?

Feature article

54 Renewable Energy Focus July/August 2007

Cost: the big issue

One of the main challenges the

PV industry faces is to make solar

electricity cost eff ective across a

wider range of scenarios.

“One of the biggest hurdles that the

total solar industry has to overcome

is how we can bring down the cost,”

Kyocera general manager Hidehito

Hisa told us. “Right now solar

power is one of most expensive

technologies you can buy. Some day

the power generated by solar has to

be equal or below the conventional

grid energy source.”

In the last three decades,

significant research has been

carried out to reduce the cost,

and the price of solar modules

has gone down from US$21/watt

in 1984 to about US$3/watt in

2006.

Today, if you live away from the main grid, and the choice of obtaining electricity is from

diesel power or PV, solar cells

are already cost-effective.

Many countries already offer

incentives to encourage the use

of PV systems, and the market

has been growing annually at

more than 20%, with a total

revenue exceeding US$10 billion

in 2006 (see FIT for purpose, page

60).

PV: where are we now?RENEWABLE ENERGY FOCUS TAKES A SNAPSHOT OF CURRENT

DEVELOPMENTS IN PV, AND INVESTIGATES TO WHAT EXTENT CRYSTALLINE

TECHNOLOGY WILL CONTINUE TO BE THE MAINSTAY OF THE INDUSTRY. David Hopwood

Building integrated photovoltaic (BIPV) – combining function with design using fl exible UNI-SOLAR PV laminates (Image courtesy of UNITED SOLAR OVONIC, a Michigan, US-based company, which manufactures thin fi lm amorphous PV technology). Its proprietary technology has been developed to deposit solar cells simultaneously on to 6 rolls of stainless steel, each 1.5 mile long, using an automated line. According to the company, the fl exible modules off er “nearly complete freedom of design to architects as they can also conform to curved surfaces and hence are meeting the increasing demand for Building Integrated PV (BIPV). The modules also off er advantages in low and diff use light conditions, due to higher absorption of light in the blue wavelength range.”

ref_0804_pg54_59.indd 54ref_0804_pg54_59.indd 54 25/07/2007 15:18:0225/07/2007 15:18:02

Sub_Electricity (PV)/Cat_Solar

Renewable Energy Focus July/August 2007 55

Material challenges – is less more?

The fi rst solar cells used were

crystalline silicon wafers, and the

workhorse of the industry is still

single crystal or polycrystalline

silicon-based solar cells. The

manufacturing process is,

however, energy-intensive, and

there are many steps involved that

make the product expensive. The

modules have glass covers that

are fragile, and need expensive

support structures to install.

There are many ways in which

companies are trying to reduce

the cost of PV technology, such as

reducing the amount of material

per module.

Companies like CSG Solar have

designed technology that allows

the production of solar modules

using less than 2µm thick crystalline

silicon – placed on a textured sheet

of glass. This enables CSG modules

to capture the established strengths

of crystalline silicon, in a much

thinner and cost-eff ective form. “All

other thin fi lm contenders use more

exotic semiconductor materials and

so miss out on the proven durability

and/or environmental benignness

of silicon,” says CSG’s Joanna Ziegler.

“Just about everything in CSG’s

technology is innovative, from the

method of texturing the glass before

the silicon is deposited (enabling

light to eff ectively get trapped in

the thin silicon) to the use of inkjet

printing in the production of the

electrical contacts enabling power

to be extracted from the silicon.”

Thin Film (TFT)

Go to any PV conference though, and

the talk is of thin fi lm technology;

many companies are working on

diff erent thin fi lm technologies,

namely amorphous silicon, copper-

indium-di-selenide and cadmium-

telluride, because they can reduce

the cost of modules substantially.

But compared to crystalline solar

cells the challenge for thin fi lm

technologies lies in increasing the

effi ciency level achieved under

laboratory conditions to mass

production settings. While overall

system costs are already lower using

thin fi lm technologies, increasing

the effi ciency will reduce the larger

area currently required to generate

the same output as crystalline

modules, and assist in reducing the

costs for solar energy.

Whether, and to what extent, thin

fi lm and other novel technologies

will replace crystalline technology is

the subject of intense debate in the

industry. This has been highlighted

by the well-documented silicon

supply problems of the past few

years. According to Professor Eicke

Weber, director of the Fraunhofer

Institute for Solar Energy Systems

ISE in Germany, about 90% of the

market demand is for crystalline

silicon solar cells – with effi ciencies

ranging from 14% to 20% – and

this has created supply problems as

the PV industry competes with the

semiconductor industry for silicon

feedstocks: “there has not been

enough high-purity silicon available

to fulfi l capacities of the production

lines which have been built,” he

says, “and some production lines

are working at only 60% capacity

because of shortages.”

This shortage of material has been a

factor in the price of the PV modules

not coming down as the industry

would have hoped, and this has

been a further catalyst for the

development of newer technologies

such as thin fi lm that allow modules

to be produced at lower cost (it is

easier to deposit these materials on

large areas). “Most of the materials

– even Amorphous silicon – are very

independent from the supply of the

normal high purity silicon from the

crystalline silicon market, so they

are less expensive and they have

more access to the market,” Weber

explains.

But the downside is that these

technologies lack effi ciencies; “for

amorphous silicon you cannot

really buy anything above 8% of

long term stable effi ciency, and

other thin fi lm technologies have a

magic limit near 10%-11%”, he says,

well below what has been achieved

for crystalline technology.

And Weber, who admits that

Fraunhofer doesn’t concentrate

on developing thin fi lm

technologies, believes we need

to be cautious and realistic about

claims that the effi ciency of thin

fi lm will increase dramatically:

“I always read business plans

which predict optimistic jumps

in effi ciency. But just look at the

situation with amorphous silicon

– this is a material which has

been developed in the last 20

years with enormous amounts

of government support; more

government support has been

given to thin fi lm and amorphous

materials than for crystalline

silicon research, but this 20 years’

of research has not done anything

to substantially increase the

effi ciency – increasing effi ciency is

simply constrained by the physics

and the properties of thin solar

cells.”

Market niches for diff erent applications

Of course, he says, that’s not to

say that the thin fi lm technologies

don’t have their place in the

market, or even an important role

to play in keeping a lid on the price

of crystalline silicon PV systems. It’s

all to do with the diff erent needs of

the end user; a trade off between

whether high price/higher

effi ciency is your main concern or

lower price/lower effi ciency will do

the job:

“Right now thin fi lm solar cells

cost around �1-�2/watt, whereas

crystalline silicon solar cells cost

something like �3.50/watt. If you

are in the Sahara desert making a

power plant, for example, you don’t

care too much about the effi ciency

because you have enough area and

you can use the least US$/watt. On

the other hand if you have a house

and a roof, and you live in Northern

Europe, then you have a choice

whether you want to have 1.5Kw or

3Kw PV on your roof. Especially in

places like Germany with its Feed-in

Tariff – where you get paid per Kw/

hour per produced PV power – it

makes sense to pay for the higher

effi ciency.”

Do those involved in the production

of thin fi lm agree with this?

At this year’s Intersolar show, Schott

introduced a new framed version of

its ASI F-90 thin-fi lm module, with

an effi ciency of around 6% [10% at

an R&D scale] which the company

says produces nearly three times as

Image, courtesy of Isofotón, shows the company’s factory in Malaga, with PV concentration technology on display. Isofotón’s strategic focus on innovation has been directed towards concentration technology. The company’s systems reduce the use of semiconductor materials - through the use of optical systems - Isabel Cazorla told us.




much output as earlier models. The

company points to the use of thin

fi lm where, for example, roofs do

not face exactly the right direction

or off er the perfect angle, or when

temperatures rise signifi cantly

(independent studies reportedly

show that ASI modules can deliver

as much as 10% more power per

watt than crystalline modules),

claims Schott.

Lars Waldmann of Schott adds,

“thin fi lm does have applications

in areas like Building Integrated

PV, because the modules can be

integrated into the building quite

easily, and is more cost effi cient

in terms of price per m2 (and per

watt/peak in the future). Other

applications would be where you

need to keep the costs down in

rural electrifi cation, or in places

where solar radiation is good;

where you might have large

heat loads and the crystalline

PV is going down; and maybe

where you have diff used light

in the jungle or in Asian regions

– in these areas thin fi lm is a good

approach.”

Waldmann also points to larger-

scale ground mounted PV systems

as another application where thin

fi lm would fi t the bill.

Sharp Corp says that it has

successfully developed mass-

production technology for

stacked triple-junction thin-

film solar cells by turning a

conventional two-active-layer

structure (amorphous silicon

plus microcrystalline silicon) into

a triple-junction structure with

amorphous silicon (two active

layers) and microcrystalline

silicon (single active layer).

“This new architecture boosts

cell conversion efficiency

from 11%-13% and module

conversion efficiency from 8.6%-

10%. Creating two amorphous

silicon active layers significantly

increases voltage levels, and

structuring the cell to have three

active layers in combination with

microcrystalline silicon decreases

light-induced degradation (drop

in conversion efficiency)” a

source said.

On how much opportunity there

will be for thin film in coming

years, Weber of Fraunhofer

ISE adds, “I think there will

be enough space for the thin

film manufacturers and the

crystalline silicon manufacturers,

but my expectation is that the

market share of thin film will

never come above 10%-20%. It

will always stay a small segment,

and the biggest growth in the PV

market will be crystalline silicon.

In the next five years, I will be

very happy if thin film PV keeps

the lid on the price of silicon PV

– when the price differential to

silicon modules gets bigger it

will exert very healthy pressure

on the price of silicon modules.”

Waldmann agrees that there is no

way one technology will “replace”

the other. “For the retrofit market

– rooftop mounted PV – this is a

completely different market, and

you find the technology that

works for a certain situation, and

that’s why Schott is engaged in

both of these businesses.”

Purifi ed metallurgical silicon

Both Schott and Fraunhofer point to

other areas of development for PV,

away from the headline-grabbing

thin fi lm technologies.

Fraunhofer continues to push

the effi ciency of high effi ciency

crystalline silicon solar cells (“because

it still makes a diff erence if you can

get 20%, 22% or 24% effi ciency with

reasonably low cost production”).

Additionally, Weber has brought

research that he pioneered at the

University of California, Berkely,

into the Institute. Namely, what

he describes as “PV with unlimited

silicon resources,” or “crystalline

silicon PV based not on the high

purity silicon that the semiconductor

industry is using, but based on dirty

silicon, or purifi ed metallurgical

silicon.

He explains that what is actually in

short supply is the semiconductor

grade silicon, because of the market

shared between the IT and PV

industry.

Using intelligent defect engineering,

it is possible to produce solar cells of

essentially the same (or marginally

lower effi ciency), out of a material

that is practically in unlimited

supply. It also has other advantages;

low cost of production, and it needs

much less energy for production,

something that is important for an

industry that has thus far survived

on the back of its green credentials.

“Of course there is still the energy

required for making modules and

contacts etc, but the silicon part of

the total energy equation can be

substantially decreased by using

the purifi ed metallurgical silicon

concept,” Weber says.

Many would argue that we don’t

need this purifi ed metallurgical

silicon because the production

of semiconductor grade silicon

is currently being ramped up.

Weber’s response is that “if the

semiconductor grade silicon

comes down from the current price

(US$100-US$200/kg) to US$50/

kg (maybe a long-term price for

semiconductor silicon) and the

purifi ed metallurgical silicon can be

made for US$10-US$20 per kg there

is still a substantial diff erence.”

Organics

As part of Schott’s strategy, the

company is one of the few companies

looking into organic PV – together

with large companies like Bosch

and BASF. Organic PV is the term

used to describe solar cells based on

organic semi-conductive materials

that can generate electricity from

light. This means that, in the future,

they could replace the silicon that is

used today. The aim is to use new

materials, production processes and

installation technologies to make

the organic solar cells more effi cient

and cost-eff ective in the long term.

Furthermore, the new technology

will pave the way for sustainable

energy production and make solar

power more competitive.




Organic solar cells are fl exible and

as thin as a clear plastic folder.

They are both light and colour

tunable, which enables them, for

example, to be used in foldable

cell phone chargers, or on car

roofs.

From 2015 onwards, their main area

of application is expected to be in

the construction industry, where

the cells will be used in the form

of a thin layer of plastic on roofs,

windows or facades.

Tracking devices

Another important exploratory

focus lies in light concentrator

technology. Sharp Solar for

example, has developed a solution

that is not based on silicon. The

light is gathered by a fresnel-lense

and concentrated 700 times upon

a III-V compound solar-cell. The cell

itself is very small and reaches an

effi ciency of more than 37% under

concentration. A light-gathering

system already in operation

consists of 270 cells and a two-

axis tracking device. So especially

for the regions where large DNI

is available, the concentrator

represents an interesting

alternative to crystalline modules,

the company reports.

And Conergy’s Thorsten

Vespermann says that Conergy

has developed the SolarOptimus

tracking system, where, in sunny

regions, tracking systems can

“increase electricity output by up

to 35%.”

PV R&D roundup

Although the PV solar market

is dominated by mono/

multicrystalline silicon,

alternatives such as dye-

sensitised; polymer and

thin-fi lm amorphous; CdTe;

CIGS; CIS; and GaAs have not

abandoned the fi ght. Fuji

Keizai USA Inc predicts an

annual average growth of

43.1% for new technologies

over the next decade – to 13%

of the Mw market capacity;

Two newcomers on the block

are California’s Stion (US$15m

in series B funding) and UK’s

Imperial College, London

spin-off Quantasol (US$2.7m

seed funding). Formerly

NStructures, Stion uses TFT; it

is not using Si or CdTe, and is

thought to be taking a quantum

dot approach. Quantasol is

take a GaAs quantum well with

a tandem cell approach (fewer

junctions than the traditional

■

■

triple-junction cell) and

optimised spectral output for

its concentrator system;

Another newcomer is US

R&D fi rm Illuminex, which

has attracted more than

US$2m funding for nanowire

applications, among them PV

thread from Si nanowires and

conducting polymers;

In the US Prime Solar received

DOE funds of US$3m for 18

months work, to scale up

NREL’s record 16.5% CdTe cells

to “dramatically lower the

cost of PV energy generation.”

Simultaneously First Solar

scored fi ve agreements

for the manufacture and

supply of its CdTe modules,

totalling 485Mw sales of some

US$1.28bn. Agreements

include EDF Energies Nouvelles,

Sechilienne-Sidec, the Juwi

Group, SunEdison, and RIO

Energie GmbH. First Solar has

also approved an investment of

US$150m in the construction of

additional manufacturing plants

in Malaysia, rated for 120Mw

and scheduled to be on-line in

late 2009;

On the CIGS front, Arizona-

based Global Solar Energy

won MIL STD-810E ruggedised

■

■

■

certifi cation, claiming no other

fl exible solar product has yet

obtained this. Colorado-based

Ascent Solar, using TF CIGS on

plastic substrate, raised US$20m

from warrant conversions, and is

testing its R2R rapid prototyping

tools prior to operating its

1.5Mw pilot plant by year end,

and starting construction of the

fi rst of four 25Mw large volume

lines;

Solo Power, raising US$30m,

also uses CIGS with an

electrochemical process it

claims is more cost eff ective

than CIGS and Si, being built

on fl exible foil substrates;

Nano Solar with CIGS QDs

and nanoparticle ink and fast

roll printing has production

scheduled by year-end.

Plants in San Jose, California

and Berlin, Germany give it

647,000ft2 for cell and panel

manufacture, as well as R&D;

GaAs fares well, with Emcore’s

PV division awarded US$2m by

NASA’s Jet Propulsion Lab for

the Mars Cruise Stage spacecraft

solar panels, for delivery mid

2008; Aixtron AG has received

an order from Taiwanese WIN

Semiconductors for an ajAIX

2600G3 IC epitaxy reactor to

■

■

■

The magic numbers

What order of effi ciencies might be possible for PV technologies?

Sharp’s new thin fi lm cells consist of three cell layers: two amorphous and

one microcrystalline silicon layer, which convert greater portions of the light

spectrum and further boost effi ciency in comparison with microamorphous

tandem cells (with the cell thickness of just two microns being the same). “The

previous module effi ciency of 8.6% for tandem cells has now been increased

to up to 10% for triple cells”, reports a company source;

Lars Waldmann of Schott says that some companies have achieved

laboratory efficiencies of 21% for Monocrystalline technology [NB: Schott

itself doesn’t manufacture monocrystalline]. And he adds, “for polycrystalline

technology, we stand at 15%-17% on average – in the next five years we

will approach 16%-19%. I see a physical borderline at about 25%, and once

the industry has achieved this I think 36% may be possible, but only in

a 15-20 year R&D timescale. For thin film – cadmium telluride efficiency

is already at 10%. 13% could be achievable, but we don’t see that there

would be any possibility to increase efficiency any further. The next step in

thin film amorphous silicon we predict will be 10%-13%, and this will be a

combination of amorphous silicon, microcrystalline and a special dopine;

At the laboratory level, Kyocera recently reported that it had achieved

the world record in cell effi ciency, 18.5% (measured by Kyocera), with

150mmx155mm multicrystalline silicon.

■

■

■



Renewable Energy Focus July/August 2007 59

allow WIN to diversify into

solar cell manufacture of triple

junction solar cells;

Solar Millenium AG

(specialising in parabolic

troughs and solar ‘chimneys’)

has begun phase two of its

project near Granada – a solar

thermal facility covering 195ha;

A new test facility at Almeria

in Southern Spain worked by

Fraunhofer Institute for Solar

Energy System is to develop the

100 metre linear Fresnel refl ector

(low cost alternative to parabolic

mirrors) to focus sunlight onto

steel absorber tubes, in which

water is heated to 450oc driving

electricity producing turbines.

This not only improves on

parabolic trough weaknesses,

but is predicted as the route

to provide about 10% of world

electricity by 2050;

Not to be left out of new

developments, PV Powered

has produced a new,

commercial inverter PP30kw

that achieves a 94% CEC

effi ciency rating. The inverter

off ers remote performance

monitoring, night time

disconnect for reduced tare

losses, and a soft start circuit.

Using plastics for solar energy

has increased effi ciency by

6.5% thanks to a discovery at

the Center for Polymers and

Organic Solids (University

of California, Santa Barbara).

Nobel Laureate Alan Heeger

– professor of physics at UC

Santa Barbara, working with

Kwanghee Lee of Korea – has

created a new ‘tandem’ organic

solar cell of two multi-layered

parts working together to

gather a wider range of the

solar radiation spectrum at both

shorter and longer wavelengths.

Heeger is confi dent “we can

make improvements that will

yield effi ciencies suffi ciently

high for commercial products,”

and expects the technology

to be on the market in about

three years. Key element is TiOx,

transparent titanium oxide

which binds and separates the

cells, transporting electrons,

being a collecting layer for the

fi rst cell. As a stable foundation

■

■

■

■

it allows the fabrication of the

second cell to complete the

tandem architecture;

Conergy is investing around

�250 million in what it says is

the world’s fi rst and only fully-

integrated mass production of

wafers, cells and modules. The

fi rst modules, with total output

exceeding 50 megawatt, will

start rolling off the production

lines by the second half of 2007.

By 2008, a production capacity

of 300 Mw for wafers, 275 Mw

for cells and 250 Mw for solar

■

modules will have been reached.

The production process will

begin with a cell thickness of

200 µm but is designed for a

reduction to 160 µm;

Applied Materials focuses the

majority of its eff orts in crystalline

and thin fi lm silicon, although

it is conducting investigations

in other technologies. David

Miller explains, “in crystalline, we

see opportunities for improved

effi ciency driven by the

performance of process tools.

We expect a similar result in thin

■

fi lm silicon because our tools

will provide output that is more

uniform and consistent. This

combined benefi t will produce

more, higher value modules.

This will be true for both the

high throughput (3,300-5,500

uph) crystalline deposition

machine and on the large area

TF production line.”

Thanks to Andrea Bodenhagen

of United Solar Ovonic, and Gail

Purvis who both provided editorial

contributions for this article.


Documents

PV: where are we now?