International Conference on Solar Concentrators for the Generation of Electricity or Hydrogen, 1-5 May 2005, Scottsdale, Arizona (NREL/CD-520-38172)

COST ANALYSIS OF A CONCENTRATOR PHOTOVOLTAIC HYDROGEN

PRODUCTION SYSTEM

Jamal R. Thompson
Howard University, 2300 6th St., NW, Washington, DC 20059, USA
e-mail: [email protected]

Robert D. McConnell
NREL, 1617 Cole Boulevard, Golden, CO 80401-3393, USA
e-mail: [email protected]
Mohsen Mosleh
Howard University, 2300 6th St. NW, Washington, DC 20059, USA
e-mail: [email protected]

1. Abstract
The development of efficient, renewable methods of producing hydrogen are essential for the success of the
hydrogen economy. Since the feedstock for electrolysis is water, there are no harmful pollutants emitted
during the use of the fuel. Furthermore, it has become evident that concentrator photovoltaic (CPV)
systems have a number of unique attributes that could shortcut the development process, and increase the
efficiency of hydrogen production to a point where economics will then drive the commercial development
to mass scale.
Concentrating solar energy to produce electricity can occur at quite high solar conversion efficiencies.
The highest efficiency for solar concentrator cells, as measured at NREL, is now above 37%. Solar
Systems P/L of Australia has exhibited a 40% boost in hydrogen production by separating the solar infrared
radiation incident on concentrator solar cells and using it as the heat source for a solid oxide electrolyzer
cell operating above 1000 Celsius [9]. With today’s solar cell technologies, it is therefore possible to
achieve a 50% conversion efficiency of the solar energy to hydrogen through high temperature electrolysis.
With gasoline prices constantly increasing, the cost associated with producing hydrogen is becoming
more and more favorable. At approximately $3.10/kg, the cost of producing hydrogen through wind
electrolysis is becoming competitive with that of gasoline [10]. It is expected that hydrogen production
through thermal-CPV electrolysis has the potential to be equally as attractive, if not more so. Details of a
cost analysis for such a hydrogen generation system will be presented.

2. Introduction Another issue is the cost of producing hydrogen. For

The development of a hydrogen economy can have many hydrogen to be considered as an alternative to fossil fuels,
benefits for the environment. It could play a role in production costs from renewable resources need to be
reducing global warming and air quality problems in and substantially reduced. This paper examines the costs
around major cities. A large percentage of the pollution that associated with a system that uses solar energy to produce
contributes to these issues is easily traced to the power hydrogen from water.
demands of buildings and the emissions of vehicles.
Provided hydrogen can be produced from renewable 3. System Description
resources at reasonable costs, the use of hydrogen fuel cell To address the cost issues associated with hydrogen
technology in buildings and vehicles would effectively production from renewable resources, methods of
eliminate a major contribution to air pollution problems and performance enhancement should be explored. The use of
global warming. Additionally, the United States spectrum splitting of solar energy to harness the thermal
dependence on foreign oil would be curtailed, and providing portion of the solar spectrum offers a potential for higher
buildings with their own power generation capability would efficiency and increased hydrogen production capability. In
reduce the demand on the electric grid. These are only a this section, we will outline a method of producing
few of the many beneficial possibilities of a hydrogen hydrogen that combines solar concentration and spectrum
economy. splitting to increase system efficiency and production
For the hydrogen economy to come to fruition, there are capability.
many obstacles that need to be overcome. The most The CPV hydrogen production system (developed and
prevalent roadblock is the current lack of an infrastructure to patented by Solar Systems P/L of Australia) consists of four
support a hydrogen economy. Hydrogen production major components: a concentrator dish (including PV
facilities need to be constructed and methods for panel), optical filter, fiber optic light guide, and solid oxide
transporting hydrogen need to be developed. electrolyzer cell (SOEC) stack. The following is a
 are the optical filter and the fiber optic light guide. These
components will be discussed further in Section 4.

4. Cost Analysis
The cost of solar hydrogen production needs to be
competitive with similar hydrogen production processes for
it to be successful in the hydrogen market. As a prototype,
constructing a system today is relatively expensive.
However, efficient and cost effective design and mass
production can significantly reduce future production costs.
In this section, the projected cost for such a system will be
evaluated and compared to similar hydrogen production
systems.

Figure 1: System Description 4.1 Projected System Cost

Table 1 provides component cost information for a one
description of the hydrogen production process (See Fig. 1 kilowatt system at current cost as well as a projected per
for visual reference): kilowatt cost based on the future purchase of 1GW of CPV
1. The dish of the CPV focuses sunlight on the PV Electrolysis systems. All numbers provided in the projected
panel providing it with the energy equivalent of costs column are based on the purchase of 1GW worth of
250 – 300 Suns. This increase in energy boosts electrolysis systems (consists of fifty thousand 20kW CPV
the electrical output of the PV panel by the same Electrolysis dish systems). As seen in the table, CPV
factor. currently costs $10,000 per kilowatt. This cost includes the
2. The optical filter is placed in front of the PV panel dish, sun tracking capability, and the PV panel. The PV
and reflects the infrared energy. The remaining panel is only 10% of the system cost, allowing for room to
visible light is transmitted through to the PV panel reduce cost through package and structure design. The
where it is transformed into electric energy and projected cost for CPV is $800 per kilowatt [11].
supplied to the SOEC. The rugate and dichroic filters are the types of spectral
3. Meanwhile, the reflected infrared energy is focused splitters being considered for implementation into this
into the light guide and transmitted to the SOEC. system. As shown in Figure 2b, the rugate filter uses a
4. The SOEC combines this heat energy with the graded index film to bend the energy rays where as the
electricity to separate water into hydrogen and dichroic filter utilizes layers of thin film with different
oxygen. An electrical charge is supplied to the indices to achieve the same results (Figure 2a) [1-2]. When
electrolyzer creating a potential that draws the compared to the step-indexed filter, the rugate filter presents
hydrogen and oxygen molecules to the cathode and a slightly more efficient design, as it requires fewer internal
anode, respectively. Supplying heat to this process reflections to achieve the same result. These two
reduces the amount of electricity needed to technologies will need to be tested and compared to
separate the molecules and increases the efficiency determine what benefits the rugate filter has over the
of the process. dichroic filter. For the cost of the spectral splitter,
5. From the SOEC, the hydrogen and oxygen are information is provided for the dichroic filter. Costs for
released at atmospheric pressure and can be stored rugate filters were not available for this study. Costs for
in tanks for usage by fuel cells or other hydrogen dichroic filters ranged anywhere from $70 to $500, with the
powered devices. more efficient filters generally being the most expensive.
For this system to be effective, these components need to be Determining which filter is best designed for this system
implemented into a system that achieves high efficiency. will require some testing. Assuming that the solar hydrogen
Two components that can be the cause of significant losses system will demand the most efficient filter, cost for the
most expensive filter is factored into the system cost. For a
1GW system the cost of the dichroic filter is estimated to be
Table 1: System Cost Analysis $5.50 per kilowatt.
Component Current Cost 1GW Purchase Like the optical filter, there are two types of fiber optics
Cost for a 1kW Est. Cost per kW that will be considered for use in this system. The first is
(per kW) System (Yr. 2020++)
the step-indexed light guide, which transmits energy
CPV $10,000 $800.00 through reflections as the dichroic filter does. The second is
Dichroic Filter $500 $5.50 the graded-index fiber, which bends energy in the same
Light Pipe* $990 $14.75 manner as the rugate filter. The response is best described
SOEC** $2,000 $400.00 as sinusoidal in nature. The graded index fiber reduces the
Unknown Costs $5,000 $5.00 effects of many of the losses associated with attenuation and
Total $18,490 $1,221
dispersion in the fiber [3]. The performance of graded
indexed fibers needs to be evaluated and compared to that of
*Prices quoted by CeramOptec
**Manufacturing costs quoted by Ceramatec, Inc. step-indexed fibers to determine what benefits the
 Incident Ray Reflected Rays Incident Ray Reflected Ray

η0
Air
Air η0
Thin Film Layer 1 η1
d
Thin Film Layer 2 η2 η1(d)
Graded-index Film
Substrate ηs
Substrate ηs

Transmitted Rays Transmitted Ray

Figure 2a: Dichroic filter response Figure 2b: Rugate filter response

technology may have. Estimates for this study are provided assistance from Ceramatec, Inc., we have determined that an
for the step-indexed fiber. The fiber optic light-guide is electrolyzer, consisting of three 50 cell stacks, operating at
estimated by CeramOptec to cost $990/part for 1 to 2 parts 1000°C and consuming 20kW of electricity, would be
and $24.75/part for 500 parts. These estimates are used to capable of producing 5.719 kg of hydrogen over a time
forecast the future cost of 50 thousand parts (See Table 1). period of 7.7 hours. Using this calculation as the daily
This information is based on a fiber optic cable that is one average production for one unit, we expect a 10MW system
meter in length and 25mm in diameter, constructed of boro- (consisting of 500 units) to be capable of producing
silica fibers, and encased in a stainless steel interlock approximately 1.04 million kg-H2 during one year of
jacketing. production (See Table 3).
The final piece of the system is the SOEC stack. As a Now that the capital cost associated with the
custom made product, SOECs are fairly expensive. Their construction of a plant and the production capability of said
cost currently ranges from $1000 to $2000 per kilowatt plant are known, the price at which H2 will be provided to
depending upon the ceramic material used as the electrolyte. the consumer can be determined. To do so, a payback
However, Ceramatec has projected costs for this component period on the capital investment is set at 20 years.
to be reduced to as low as $400 per kilowatt by year 2020. Assuming the plant produces hydrogen at the rates provided
In addition to the costs associated with the components in Table 3, hydrogen can be sold to the consumer at
mentioned above, there are miscellaneous costs that cannot $3.18/kg ±30%. Furthermore, investors can expect to see an
be determined this early in the development of the system. 11% return on their investment with hydrogen sold at this
Other areas of concern have to deal with the components price. The error associated with this calculation accounts
required to connect the light pipe to the SOEC and to the for the uncertainties in investment analysis. This cost
CPV. Another system component that will affect system estimate shows that hydrogen production from CPV
cost is an automatic control system to manage and maintain electrolysis can be cost competitive with other methods of
hydrogen production. Estimates of what these additional production.
costs would add to the price of the system are factored into
the miscellaneous costs provided in Table 1. Table 2: Investment Analysis
Major Equipment Cost (per kW) $1,221
4.2 Plant Investment and H2 Production Analysis Total Major Equipment Cost (TMEC) $12,210,000.00
Provided that the projected cost for a CPV Electrolysis Installation (47% of TMEC) $5,738,700.00
system is $1221/kW, a capital cost analysis for the Piping (50% of TMEC) $6,105,000.00
Instrumentation (25% of TMEC) $3,052,500.00
construction of a 10MW CPV electrolysis farm is provided
Buildings and Structures (12% of TMEC) $1,465,200.00
in Table 2. The information in this table is based on the Electrical Systems Installed (11% of TMEC) $1,343,100.00
cost analysis of a chemical plant provided in [8]. Some Yard Improvements (10% of TMEC) $1,221,000.00
percentages, such as piping, service facilities, and buildings Service Facilities (35% TMEC) $4,273,500.00
and structures, have been adjusted to better reflect the costs Total Direct Costs (TDC) $35,409,000.00
expected to be associated with a CPV hydrogen production Engineering & Construction (E&C) (33% of TMEC) $4,029,300.00
Contingencies (35% of TMEC) $4,273,500.00
plant. The Total Fixed Investment (TFI) is a summation of Construction Expenses (41% of TMEC) $5,006,100.00
the Total Direct and Indirect Costs, and the Total Capital Legal Expenses (4% of TMEC) $488,400.00
Investment (TCI) is a summation of the TFI and the Total Contractor's Fee (5% of TDC) $1,770,450.00
Working Capital (TWC). The TCI calculation carries an Total Indirect Cost (TIC) $15,567,750.00
error factor of ±30%. Total Fixed Investment (TFI) $50,976,750.00
Total Working Capital (TWC)
For cost per kilogram of H2 calculations, the production (15% Total Capital Investment) $8,995,897.06
capability of the plant needs to be determined. With Total Capital Investment (TCI) $59,972,647.06
Table 3: Hydrogen Production Cost Analysis include the testing of optical filters and fiber optic light
Plant Size 10MW guides and the development of methods of exchanging the
heat to the water for the electrolysis process.
System Cost $1,221/kW
With oil prices constantly rising, the development of a
Hydrogen Produced in 1 day* 2,865 kg
hydrogen economy presents a reasonable method of slowly
Hydrogen Produced in 1 year 1,043,765 kg
reducing the dependency of the United States on foreign oil.
TCI $60.0 million With further development, CPV electrolysis systems could
Hydrogen Cost (Break Even) $2.87/kg ±30% be a major contributor to the support structure of such an
Hydrogen Cost (11% ROI) $3.18/kg ±30% economy.
*Assumes each electrolyzer is capable of producing 5.719 kg-H2 in a 7.7-
hour day.
6. Acknowledgements
The CPV Electrolysis System is designed and patented by
Table 4: Cost Comparison Solar Systems P/L (Australia). The authors thank John
Hydrogen Production Lasich and Dave Holland for their efforts in the
Process development of the project.
Cost (per kg)
The authors thank Joseph Hartvigsen and S Eangovan
Gas Reformation [4] $1.15
of Ceramatec, Inc. (Salt Lake City, Utah) for their assistance
Wind Electrolysis [10] $3.10
in determining the production capabilities and costs for the
Nuclear Electrolysis [5] $1.48 solid oxide electrolyzer.
PV Plate Electrolysis [6] $7.40 The authors thank Johanna Levene of NREL’s
CPV Electrolysis $3.18 Hydrogen group for her assistance in the development of the
plant cost analysis.
4.3 Cost Comparison 7. References
Other methods of producing hydrogen include gas 1. F.T.S. Yu & X. Yang, “Introduction to Optical
reformation and wind, nuclear, and PV plate electrolysis. Engineering,” Cambridge University Press, 1997.
As shown in Table 4, consumer costs for these technologies 2. A. Thelen, “Design of Optical Interference Coatings,”
range from as little as $1.15/kg for gas reformation to McGraw-Hill Companies, Inc., 1989.
$7.40/kg for PV plate electrolysis. While gas reformation is 3. J.A. Buck, “Fundamentals of Optical Fibers,” John
inexpensive and nuclear electrolysis is projected to be Wiley & Sons, Inc., 1995.
cheap, neither of these technologies utilizes renewable 4. D. Mears, M. Mann, J. Ivy, M. Rutkowski, “Overview
resources. Furthermore, a byproduct of gas reformation is of Central H2A Results,” 2004 US Hydrogen
CO2, which is the very pollutant a hydrogen economy seeks Conference Proceedings, April 26 - 30 2004.
to eliminate. While production from renewable resources is 5. W.A. Summers, Hydrogen Production Using Nuclear
more expensive, the environmental benefits out weigh the Energy, 15th Annual U.S. Hydrogen Conference.
cost. 6. L. Kazmerski, “PV Electrolysis,” ASES Renewable
While hydrogen production from CPV electrolysis Hydrogen Forum, Oct. 1, 2003.
would greatly benefit the hydrogen economy, it is important 7. EIC Country Analysis Briefs,
to note that production from other renewable methods is www.eia.doe.gov/emeu/cabs/usa.html.
necessary to supply the nation’s energy demands. In 2004 8. M.S. Peters, K.D. Timmerhaus, R.E. West, “Plant
automobiles in the US consumed 9 million bbl/d (378 Design and Economics for Chemical Engineers,”
million gal/day) of gasoline [7]. Considering that 1kg of McGraw-Hill Companies, Inc., 2003.
hydrogen is about the energy equivalent of one gallon of 9. J.B. Lasich, “Production of Hydrogen from Solar
gasoline, it is easy to conclude that no one renewable Radiation at High Efficiency,” US Patent No.
technology is capable of supplying this demand on its own. 5,973,825, October 26, 1999.
However, solar radiation is one of the most abundant 10. M. Mann, P. Spath, A. Watt, “The Economic Feasibility
renewable resources available. of Producing Hydrogen from Sunlight and Wind,”
Hydrogen Power Now: Proceedings of the 9th
5. Conclusion Canadian Hydrogen Conference, February 1999.
From a cost standpoint, we have projected CPV electrolysis 11. R.D. McConnell, J.B. Lasich, C. Elam, “A Hybrid Solar
to be a feasible method of producing hydrogen. Component Concentrator PV System for the Electrolytic
cost projections suggest that in the future, CPV hydrogen Production of Hydrogen,” To be published in the
production will be competitive with other hydrogen Proceedings for the 20th European Photovoltaic
production methods. Development of a system needs to be Solar Energy Conference, June 2005.
completed to verify production capabilities. Initial steps
toward the development of a CPV electrolysis system will