A Luminescent Solar Concentrator with thin-film amorphous silicon ...
A Luminescent Solar Concentrator with thin-film amorphous silicon solar cells St even T. Velt h u ijsen
October 2009 Utrecht University
A Luminescent Solar Concentrator with thinfilm amorphous silicon solar cells A proof-of-principle experiment
Steven T. Velthuijsen Utrecht, October 2009 Type of work:
Master thesis
Student number:
0231568
Study:
Energy Science
University:
Utrecht University, Copernicus Institute, The Netherlands
Supervision: Dr. W.G.J.H.M. van Sark
Department of Science, Technology and Society Copernicus Institute, University Utrecht, The Netherlands
Address author: Address: Soendastraat 6 Bis 3531 HP Utrecht The Netherlands E-mail: [email protected] [email protected] No:
N W S-E-2009-36
2
Table of Contents TABLE OF CON TEN TS ....................................................................................................................3 ABSTRACT ........................................................................................................................................... 4 1
2
3
IN TROD UCTION .....................................................................................................................5 1.1
S OLA R EN ERGY ............................................................................................................ 6
1.2
COST -REDUCT ION ........................................................................................................ 6
1.3
T HE LUM IN ESCENT SOLA R CON CENT RA TOR ............................................................ 7
1.4
H IST ORY ........................................................................................................................ 8
1.5
A M ORPHOUS S ILICON ................................................................................................. 8
THEORY ................................................................................................................................... 10 2.1
OPT ICS ..........................................................................................................................10
2.2
ELECT RONICS ..............................................................................................................12
2.3
I N T EN SIT Y EFFECT ..................................................................................................... 15
EXPERIMEN T .......................................................................................................................... 17 3.1
M A T ERIA LS ................................................................................................................. 17
3.2
CON ST RUCT ION ..........................................................................................................21
3.3
R ESULT S ...................................................................................................................... 22
3.4
OBSERV A T ION S ON T HE RESULT S .............................................................................23
4
MOD ELLIN G ............................................................................................................................ 25
5
OUTLOOK ................................................................................................................................ 26
6
5.1
T UN IN G OF BA N D GA P A N D EM ISSION PEA K ......................................................... 26
5.2
QUA N T UM D OT S .........................................................................................................27
5.3
M ODELLIN G OUT LOOK ............................................................................................. 28
5.4
L IGHT SOURCE............................................................................................................ 29
5.5
S UGGEST ION S FOR FOLLOW -UP EX PERIM EN T S ..................................................... 29
CON CLUSION S ...................................................................................................................... 30
APPEN D IX A – LIGHT SOURCES .............................................................................................. 31
3
APPEN D IX B – D EPOSITION ...................................................................................................... 33 APPEN D IX C – CELL MEASUREMEN TS ................................................................................. 35 CURREN T -V OLT A GE BEHA V IOUR ......................................................................................................... 35 S PECT RA L R ESPON SE (I N T ERN A L QUA N T UM EFFICIENCY ) ............................................................. 37 APPEN D IX D – CELL D EGRAD ATION ...................................................................................38 APPEN D IX E – MOD EL IN PUT FILE ......................................................................................40 REFEREN CES ................................................................................................................................... 43
Abstract Luminescent Solar Concentrators are a low-cost alternative photovoltaic application. They consist of a transparent plate doped with a luminescent species, to the sides of which are attached mirrors and solar cells. In this thesis a proof-of-principle configuration of two thin-film amorphous silicon solar cells and two mirrors was built and tested. It is the first time that amorphous silicon was used. Its higher-than-crystalline-silicon band gap which reduces thermalisation losses and its low cost can be defining factors in developing a viable low-cost solar cell with applications in the built environment. A system efficiency of 1% was measured, while higher results may be obtained under improved measurement conditions.
4
1 Introduction The search for renewable energy has been going on for some decades now. H aving alternatives to fossil fuels lessens countries’ dependence on oil-producers and hence provides increased reliability. The investment in renewable energy has increased by a factor 4 in the last 4 years, with wind and solar PV capacity showing a 250 and 600% increase, respectively. Small hydropower and biomass also have a large share in the world renewable energy use, but these also have serious drawbacks: hydropower is only possible in very specific locations, while the production of biomass for energy use is limited because it competes with food production. Figure 1 shows the share of different types of renewable energy. I t can be seen that solar PV still has a very small share, but as stated above its growth is rapid. [1]
5
Figure 1 - Share of different renewable energy sources in the world and its regions. [1]
1.1
Solar Energy
There are two important reasons why solar energy has not yet attained a large market share: intermittency and cost of energy. The intermittent nature of solar energy, meaning that it is only available at irregular and unpredictable intervals, can be worked around by using solar energy in combination with non-intermittent energy sources and by storage of energy. The cost of solar energy is still high compared to conventional energy sources, mainly because much of the technology is still very new. I t can be said that cost reduction is the main driver for research into solar energy technology.
1.2
Cost-reduction
I n the very broad field of solar energy two main approaches can be defined in the quest for cost reduction: 1) Reduction of cost of components of existing solar energy systems and 2) development of cheaper new solar energy systems. The first approach is an example of technological learning and has led to PV systems now being available on the consumer market. Even though the cost of solar energy
6
may still be higher than that of fossil-based electricity, consumers will still profit because the price of grid electricity is higher still (ofttimes helped partially by investment subsidies or feed-in tariffs). The second approach in itself is a wide field of research that tries to reduce permegawatt cost by increasing the intensity of the cell (heliostats, power towers etc.), by decreasing the cost of constituent materials, as in the case of for instance organic PV, or by enhancing the coupling between the absorption characteristic of the cell and the incident spectrum (tandem cells). Luminescent solar concentrators try to combine these cost -reducing paths.
Figure 2 - Schematic Representation of a LSC without mirrors. Incident light that is absorbed by a luminescent particle will upon re-emission escape from the plate if it stays within the escape cone (rays 1). Otherwise total internal reflection will guide the light to the cell (rays 2). Source: [4]
1.3
The luminescent solar concentrator
A luminescent solar concentrator (LSC) consists of a sheet or plate of a transparent material which is doped with a luminescent moiety and to the sides of which are attached solar cells. The bottom is covered with a mirror. The luminescent moieties absorb incident light and re-emit it at a redshifted wavelength. D ue to total internal reflection most of the light is trapped inside the plate and guided to the edges, where it is absorbed by the solar cells and converted into electricity. This is shown schematically in Figure 2.
7
Light is concentrated onto the solar cells because the plate’s thickness is small compared to its length and width, the materials cost is low, because the active surface (cheap plastic) is larger than the solar cell surface (expensive semiconductor) and the red-shifted incident light is more efficiently transformed into electricity in the solar cells . I n this way the three cost -reducing paths defined above are combined in one device. Additionally, an LSC has the advantage of being able to utilize diffuse as well as direct (sun)light.
1.4
History
I n their 1977 paper, Goetzberger and Greubel first proposed ‘a new principle for solar energy conversion’[5], which was, as they called it, the ‘fluorescent collector’. They identified three practical difficulties to be solved: ‘synthesis of dyes with stringent requirements, identification of plastic materials with high transparency and development of solar cells with higher band gaps.’ Mainly due to the first of these difficulties, interest for their idea waned because it became clear that the luminescent species used (dyes) had stability issues under illumination. I n recent years the subject has again gained some attention, partly due to the promise of quantum dots as luminescent species, and partly in the framework of the Fullspectrum project, which is ‘a new PV wave making more efficient use of the solar spectrum’[6]. The three difficulties foreseen by Goetzberger and Greubel have been addressed extensively in the recent past. For instance, the quality of dyes on the market today has been shown to be much higher than in was thirty years ago[7]. Secondly, transparent polymer plates have been produced with absorption loss of the order 0.5 -1
m [4]. Thirdly, the development of solar cells with higher band gaps has been looked into which has already led to interesting results, such as a 7.1% efficiency achieved with an LSC with four GaAs cells attached to the sides[8].
1.5
Amorphous Silicon
A new consideration is the possible use of amorphous silicon solar cells on the sides of an LSC. Amorphous silicon (a-Si) has the advantage of lower cost and higher band
8
gap, but a-Si cells typically have a significantly lower efficiency. This thesis tries to answer the following question: How does an LSC system consisting of a doped PM M A plate with a-Si:H solar cells attached to it behave under illumination, how does it compare to similar systems with crystalline cells and how can it be modelled.
9
2 Theory I n describing the theory underlying the luminescent solar concentrator principle it is useful to follow a beam of light as it traverses the system, from its first contact with the plate to the ‘conversion’ of photons into electrons. I n the following discussion light will be described both consisting of photons and as a beam, because both ways of looking at light are useful at times.
2.1
Optics
Because in an LSC light has to travel through different media, undergoing absorption and (more or less isotropic) emission, the optical properties of the system are of prime importance. Table 1 lists the loss mechanisms that limit optical efficiency with typical values, as taken from [9]. T able 1 - Loss mechanisms (written as efficiencies); for an explanation see text ηtr ηabs ηqe ηst ηtrap ηre ηpl
transmission at the surface absorption efficiency internal quantum efficiency Stokes efficiency Trapping efficiency Fraction of light not escaping after absorption/re-emission Transparency efficiency of the plate
0.96 0.15-0.20 0.95-1.00 0.85-0.95 0.74 0.5-0.8 0.85-0.95
A beam of light, upon reaching the outer surface of the plate, is either transmitted into the plate, or reflected away from it. This external reflection is the first loss mechanism, denoted ηtr, and it depends on the difference in refractive index of the plate and the air (or whatever medium the LSC is in contact with). I t typically accounts for about 4%. W hen light ‘hits’ a luminescent particle, there’s roughly a 15-20% chance of absorption taking place. This is the absorption efficiency ηabs. Absorption mostly (95-100% of the time) results in shifted re-emission. O therwise, the energy is lost in the form of vibrations of the dye (heat). This is called th e internal quantum efficiency ηqe.
10
The Stokes efficiency ηst is loss because of the Stokes shift: light is emitted at a longer wavelength than it was absorbed at (and hence has lower energy). O n the one hand this energy loss is detrimental to the performance of the LSC, but on the other hand the Stokes shift is essential in coupling the re-emitted light to the solar cell spectral response and reducing re-absorption losses. Light in the plate can only escape if it hits the top surface at an angle steeper than the critical angle (so if it comes from within the escape cone). For a flat slab with refractive index of around 1.5, this means than roughly 25% of the light will escape through the top surface [10]. This describes the trapping efficiency ηtrap. Because re-emission is isotropical, more light will escape through the top surface if more re-absorption takes place. ηre is the fraction of light that manages to stay in the slab until absorbed. Light travelling through the plate can be absorbed or scattered by the plate material. This effect will be small for highly transparent materials like glass and PMMA but the efficiency is not 100% and denoted by ηpl.
Figure 3 - Possible paths the light can take upon encountering an LSC plate. T he luminescent particles are represented by circles. A number of loss mechanisms is shown. A : reflection at the top surface, B: escape through the escape cone, C: no absorption, D: re-absorption and QE
October 2009 Utrecht University
A Luminescent Solar Concentrator with thinfilm amorphous silicon solar cells A proof-of-principle experiment
Steven T. Velthuijsen Utrecht, October 2009 Type of work:
Master thesis
Student number:
0231568
Study:
Energy Science
University:
Utrecht University, Copernicus Institute, The Netherlands
Supervision: Dr. W.G.J.H.M. van Sark
Department of Science, Technology and Society Copernicus Institute, University Utrecht, The Netherlands
Address author: Address: Soendastraat 6 Bis 3531 HP Utrecht The Netherlands E-mail: [email protected] [email protected] No:
N W S-E-2009-36
2
Table of Contents TABLE OF CON TEN TS ....................................................................................................................3 ABSTRACT ........................................................................................................................................... 4 1
2
3
IN TROD UCTION .....................................................................................................................5 1.1
S OLA R EN ERGY ............................................................................................................ 6
1.2
COST -REDUCT ION ........................................................................................................ 6
1.3
T HE LUM IN ESCENT SOLA R CON CENT RA TOR ............................................................ 7
1.4
H IST ORY ........................................................................................................................ 8
1.5
A M ORPHOUS S ILICON ................................................................................................. 8
THEORY ................................................................................................................................... 10 2.1
OPT ICS ..........................................................................................................................10
2.2
ELECT RONICS ..............................................................................................................12
2.3
I N T EN SIT Y EFFECT ..................................................................................................... 15
EXPERIMEN T .......................................................................................................................... 17 3.1
M A T ERIA LS ................................................................................................................. 17
3.2
CON ST RUCT ION ..........................................................................................................21
3.3
R ESULT S ...................................................................................................................... 22
3.4
OBSERV A T ION S ON T HE RESULT S .............................................................................23
4
MOD ELLIN G ............................................................................................................................ 25
5
OUTLOOK ................................................................................................................................ 26
6
5.1
T UN IN G OF BA N D GA P A N D EM ISSION PEA K ......................................................... 26
5.2
QUA N T UM D OT S .........................................................................................................27
5.3
M ODELLIN G OUT LOOK ............................................................................................. 28
5.4
L IGHT SOURCE............................................................................................................ 29
5.5
S UGGEST ION S FOR FOLLOW -UP EX PERIM EN T S ..................................................... 29
CON CLUSION S ...................................................................................................................... 30
APPEN D IX A – LIGHT SOURCES .............................................................................................. 31
3
APPEN D IX B – D EPOSITION ...................................................................................................... 33 APPEN D IX C – CELL MEASUREMEN TS ................................................................................. 35 CURREN T -V OLT A GE BEHA V IOUR ......................................................................................................... 35 S PECT RA L R ESPON SE (I N T ERN A L QUA N T UM EFFICIENCY ) ............................................................. 37 APPEN D IX D – CELL D EGRAD ATION ...................................................................................38 APPEN D IX E – MOD EL IN PUT FILE ......................................................................................40 REFEREN CES ................................................................................................................................... 43
Abstract Luminescent Solar Concentrators are a low-cost alternative photovoltaic application. They consist of a transparent plate doped with a luminescent species, to the sides of which are attached mirrors and solar cells. In this thesis a proof-of-principle configuration of two thin-film amorphous silicon solar cells and two mirrors was built and tested. It is the first time that amorphous silicon was used. Its higher-than-crystalline-silicon band gap which reduces thermalisation losses and its low cost can be defining factors in developing a viable low-cost solar cell with applications in the built environment. A system efficiency of 1% was measured, while higher results may be obtained under improved measurement conditions.
4
1 Introduction The search for renewable energy has been going on for some decades now. H aving alternatives to fossil fuels lessens countries’ dependence on oil-producers and hence provides increased reliability. The investment in renewable energy has increased by a factor 4 in the last 4 years, with wind and solar PV capacity showing a 250 and 600% increase, respectively. Small hydropower and biomass also have a large share in the world renewable energy use, but these also have serious drawbacks: hydropower is only possible in very specific locations, while the production of biomass for energy use is limited because it competes with food production. Figure 1 shows the share of different types of renewable energy. I t can be seen that solar PV still has a very small share, but as stated above its growth is rapid. [1]
5
Figure 1 - Share of different renewable energy sources in the world and its regions. [1]
1.1
Solar Energy
There are two important reasons why solar energy has not yet attained a large market share: intermittency and cost of energy. The intermittent nature of solar energy, meaning that it is only available at irregular and unpredictable intervals, can be worked around by using solar energy in combination with non-intermittent energy sources and by storage of energy. The cost of solar energy is still high compared to conventional energy sources, mainly because much of the technology is still very new. I t can be said that cost reduction is the main driver for research into solar energy technology.
1.2
Cost-reduction
I n the very broad field of solar energy two main approaches can be defined in the quest for cost reduction: 1) Reduction of cost of components of existing solar energy systems and 2) development of cheaper new solar energy systems. The first approach is an example of technological learning and has led to PV systems now being available on the consumer market. Even though the cost of solar energy
6
may still be higher than that of fossil-based electricity, consumers will still profit because the price of grid electricity is higher still (ofttimes helped partially by investment subsidies or feed-in tariffs). The second approach in itself is a wide field of research that tries to reduce permegawatt cost by increasing the intensity of the cell (heliostats, power towers etc.), by decreasing the cost of constituent materials, as in the case of for instance organic PV, or by enhancing the coupling between the absorption characteristic of the cell and the incident spectrum (tandem cells). Luminescent solar concentrators try to combine these cost -reducing paths.
Figure 2 - Schematic Representation of a LSC without mirrors. Incident light that is absorbed by a luminescent particle will upon re-emission escape from the plate if it stays within the escape cone (rays 1). Otherwise total internal reflection will guide the light to the cell (rays 2). Source: [4]
1.3
The luminescent solar concentrator
A luminescent solar concentrator (LSC) consists of a sheet or plate of a transparent material which is doped with a luminescent moiety and to the sides of which are attached solar cells. The bottom is covered with a mirror. The luminescent moieties absorb incident light and re-emit it at a redshifted wavelength. D ue to total internal reflection most of the light is trapped inside the plate and guided to the edges, where it is absorbed by the solar cells and converted into electricity. This is shown schematically in Figure 2.
7
Light is concentrated onto the solar cells because the plate’s thickness is small compared to its length and width, the materials cost is low, because the active surface (cheap plastic) is larger than the solar cell surface (expensive semiconductor) and the red-shifted incident light is more efficiently transformed into electricity in the solar cells . I n this way the three cost -reducing paths defined above are combined in one device. Additionally, an LSC has the advantage of being able to utilize diffuse as well as direct (sun)light.
1.4
History
I n their 1977 paper, Goetzberger and Greubel first proposed ‘a new principle for solar energy conversion’[5], which was, as they called it, the ‘fluorescent collector’. They identified three practical difficulties to be solved: ‘synthesis of dyes with stringent requirements, identification of plastic materials with high transparency and development of solar cells with higher band gaps.’ Mainly due to the first of these difficulties, interest for their idea waned because it became clear that the luminescent species used (dyes) had stability issues under illumination. I n recent years the subject has again gained some attention, partly due to the promise of quantum dots as luminescent species, and partly in the framework of the Fullspectrum project, which is ‘a new PV wave making more efficient use of the solar spectrum’[6]. The three difficulties foreseen by Goetzberger and Greubel have been addressed extensively in the recent past. For instance, the quality of dyes on the market today has been shown to be much higher than in was thirty years ago[7]. Secondly, transparent polymer plates have been produced with absorption loss of the order 0.5 -1
m [4]. Thirdly, the development of solar cells with higher band gaps has been looked into which has already led to interesting results, such as a 7.1% efficiency achieved with an LSC with four GaAs cells attached to the sides[8].
1.5
Amorphous Silicon
A new consideration is the possible use of amorphous silicon solar cells on the sides of an LSC. Amorphous silicon (a-Si) has the advantage of lower cost and higher band
8
gap, but a-Si cells typically have a significantly lower efficiency. This thesis tries to answer the following question: How does an LSC system consisting of a doped PM M A plate with a-Si:H solar cells attached to it behave under illumination, how does it compare to similar systems with crystalline cells and how can it be modelled.
9
2 Theory I n describing the theory underlying the luminescent solar concentrator principle it is useful to follow a beam of light as it traverses the system, from its first contact with the plate to the ‘conversion’ of photons into electrons. I n the following discussion light will be described both consisting of photons and as a beam, because both ways of looking at light are useful at times.
2.1
Optics
Because in an LSC light has to travel through different media, undergoing absorption and (more or less isotropic) emission, the optical properties of the system are of prime importance. Table 1 lists the loss mechanisms that limit optical efficiency with typical values, as taken from [9]. T able 1 - Loss mechanisms (written as efficiencies); for an explanation see text ηtr ηabs ηqe ηst ηtrap ηre ηpl
transmission at the surface absorption efficiency internal quantum efficiency Stokes efficiency Trapping efficiency Fraction of light not escaping after absorption/re-emission Transparency efficiency of the plate
0.96 0.15-0.20 0.95-1.00 0.85-0.95 0.74 0.5-0.8 0.85-0.95
A beam of light, upon reaching the outer surface of the plate, is either transmitted into the plate, or reflected away from it. This external reflection is the first loss mechanism, denoted ηtr, and it depends on the difference in refractive index of the plate and the air (or whatever medium the LSC is in contact with). I t typically accounts for about 4%. W hen light ‘hits’ a luminescent particle, there’s roughly a 15-20% chance of absorption taking place. This is the absorption efficiency ηabs. Absorption mostly (95-100% of the time) results in shifted re-emission. O therwise, the energy is lost in the form of vibrations of the dye (heat). This is called th e internal quantum efficiency ηqe.
10
The Stokes efficiency ηst is loss because of the Stokes shift: light is emitted at a longer wavelength than it was absorbed at (and hence has lower energy). O n the one hand this energy loss is detrimental to the performance of the LSC, but on the other hand the Stokes shift is essential in coupling the re-emitted light to the solar cell spectral response and reducing re-absorption losses. Light in the plate can only escape if it hits the top surface at an angle steeper than the critical angle (so if it comes from within the escape cone). For a flat slab with refractive index of around 1.5, this means than roughly 25% of the light will escape through the top surface [10]. This describes the trapping efficiency ηtrap. Because re-emission is isotropical, more light will escape through the top surface if more re-absorption takes place. ηre is the fraction of light that manages to stay in the slab until absorbed. Light travelling through the plate can be absorbed or scattered by the plate material. This effect will be small for highly transparent materials like glass and PMMA but the efficiency is not 100% and denoted by ηpl.
Figure 3 - Possible paths the light can take upon encountering an LSC plate. T he luminescent particles are represented by circles. A number of loss mechanisms is shown. A : reflection at the top surface, B: escape through the escape cone, C: no absorption, D: re-absorption and QE
