Energy 30 (2005) 2580–2595

www.elsevier.com/locate/energy

Effective utilization of by-product oxygen from electrolysis

hydrogen production
Takeyoshi Kato*, Mitsuhiro Kubota, Noriyuki Kobayashi, Yasuo Suzuoki
Nagoya University, Furo-cho Chikusa-ku, Nagoya 464-8603, Japan

Abstract
To avoid fossil-fuel consumption and greenhouse-gas emissions, hydrogen should be produced by renewable
energy resources. Water electrolysis using proton exchange membrane (PEM) is considered a promising
hydrogen-production method, although the cost of the hydrogen from PEM would be very high compared with that
from other mature technologies, such as steam methane reforming (SMR). In this study, we focus on the effective
utilization of by-product oxygen from electrolysis hydrogen production and discuss the potential demand for it, as
well as evaluating its contribution to improving process efficiency. Taking as an example the utilization of by-
product oxygen for medical use, we compare the relative costs of hydrogen production by means of PEM
electrolysis and SMR.
q 2004 Elsevier Ltd. All rights reserved.

1. Introduction

Hydrogen is one of the most promising energy carriers for future energy systems; it can be used (in
gas or liquid form) to store and transmit energy, and can be supplied to fuel-cell vehicles (FCVs) as well
as to fuel-cell power-generation systems. Hydrogen, used as the main energy carrier, could offer an
answer to the threat of global climate change and help avoid the undesirable effects of the use of fossil
fuels [1]. Although it is estimated that hydrogen is more expensive than fossil fuels, hydrogen from
renewable energy resources is a virtually inexhaustible, environmentally benign, final energy carrier that
could meet most of our future energy needs while avoiding the environmental costs and health problems
associated with fossil fuels.
Hydrogen is a long-term option. Its production, storage, and distribution facilities must be improved
and developed. In the short term, hydrogen will be produced from fossil fuels, such as natural gas, by

* Corresponding author. Tel.: C81-52-789-5373; fax: C81-52-788-6724.

E-mail address: tkato@nuee.nagoya-u.ac.jp (T. Kato).
0360-5442/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2004.07.004
 T. Kato et al. / Energy 30 (2005) 2580–2595 2581

steam methane reforming (SMR), which is a very mature existing technology. In the long run, however,
hydrogen must be produced by renewable energy resources to avoid fossil-fuel consumption and
greenhouse-gas emissions. In other words, hydrogen has a considerable potential for overcoming the
limitations of intermittently renewable energy resources and can therefore benefit their development.
For example, in the long run, water electrolysis by proton exchange membrane (PEM) is considered to
be a promising method of producing hydrogen through renewable energy resources such as wind and
photovoltaic power, because of its high efficiency. Currently, even in the case of well-established
alkaline water electrolysis, the hydrogen production cost is high in comparison with fossil-fueled
hydrogen-production technologies such as SMR because of the high investment cost of the former and
the high electricity cost involved. Technological improvements, both in electrolysis technologies and
electricity-production technologies from renewable-energy resources, however, could make the
production of hydrogen by electrolysis very attractive for the future. Thus, as a long-term option,
hydrogen production by PEM water electrolysis would also contribute to the introduction of future
renewable electricity. Furthermore, the introduction of distributed hydrogen production, based on PEM
water electrolysis utilizing renewable electricity, would be enhanced.
When hydrogen is produced by the water electrolysis process, half the number of moles of oxygen is
produced simultaneously as a by-product of hydrogen. If large quantities of hydrogen need to be produced
from renewable resources via the electrolysis process, by-product oxygen will also be produced on a large
scale. In this situation, the by-product oxygen should be fully utilized, as oxygen is an important industrial
gas used in many processes such as combustion, semiconductor production, and wastewater treatment. The
effective utilization of oxygen would improve the energy efficiency of some industrial processes [2]. While
by-product oxygen from electrolysis hydrogen production can be harmlessly vented, it seems more
prudent to explore its possible large-scale utilization.
Fig. 1 shows the conceptual diagram of the simultaneous utilization of hydrogen and by-product
oxygen. The use of oxygen-enriched combustion air in a number of energy-intensive industrial
applications has the potential to reduce the amount of heat lost to the atmosphere by about two-thirds. As
concern for the global environment rises, demand for oxygen is expanding in such areas as electric

Fig. 1. Conceptual diagram of simultaneous utilization of hydrogen and by-product oxygen.

2582 T. Kato et al. / Energy 30 (2005) 2580–2595

furnaces and glass melting, as well as in the treatment of municipal solid waste (MSW) and wastewater.
Moreover, the utilization of by-product oxygen could contribute to reducing the large amount of
electricity consumed in oxygen production by air-separation technologies, such as cryogenic air
separation and pressure swing absorption (PSA). If by-product oxygen is to be fully utilized, the balance
between by-product oxygen and oxygen demand is very important. If oxygen demand is not very large
relative to the potential supply of by-product oxygen from water electrolysis hydrogen production, then
large quantities of by-product oxygen will be wasted.
In this paper, we discuss the possible demand for by-product oxygen and its potential contribution to
energy saving. First, we present a brief introduction to oxygen- and hydrogen-production technologies.
Second, we show current and future oxygen demand and the potential for improving energy efficiency by
utilizing by-product oxygen in industry. We then compare oxygen demand with the potential supply of
by-product oxygen from hydrogen production by electrolysis. Finally, we evaluate how economical it is
to utilize by-product oxygen for medical purposes.

2. Oxygen-production technologies

An air-separation unit using a conventional, multicolumn cryogenic distillation process produces

oxygen from compressed air at high recovery and purity. Cryogenic air separation is currently the most
efficient and cost-effective technology for producing large quantities of oxygen, nitrogen, and argon as
gaseous or liquid products. The energy requirement of the latest technology is about 0.5 kW h/Nm3 -O2.
Oxygen purity of cryogenic process can be higher than 99 vol%. Because of the high purity requirement,
oxygen for medical use is normally produced by the cryogenic process. No technology, except
electrolysis, is expected to challenge cryogenic air separation for the production of large quantities of
oxygen, especially at high purity. In Japan, oxygen demand was 9615!106 Nm3 in 2001 [3,4], and
oxygen was produced mainly by the cryogenic process. Because cryogenic air separation consists of five
major processes: air compression, air pretreatment, heat exchange, cryogenic separation and oxygen
compression, it is utilized for large-scale production in excess of 8000 Nm3/h and is not suitable for
small-scale, on-site oxygen production.
The adsorption process is based on the ability of some natural and synthetic materials, such as
zeolites, to preferentially adsorb nitrogen. The regeneration of the adsorbent is necessary and can be
accomplished by heating the bed of zeolitic material (temperature-swing adsorption (TSA)) or reducing
pressure in the bed (pressure-swing adsorption (PSA)). Because of the faster cycle time and simplified
operation, PSA is usually used. The oxygen purity of adsorption process is typically 93–95 vol%. The
required energy in the adsorption process is also about 0.5 kW h/Nm3 -O2. In contrast to the cryogenic
process, the adsorption process is used for smaller applications.
Some noncryogenic processes, for example, chemical air-separation process, membrane process and
ion-transport membrane, are also available.

3. Hydrogen production by electrolysis

Hydrogen can be produced from a variety of fossil and nonfossil resources by using energy such as
heat or electricity. The major processes for hydrogen production include SMR, catalytic decomposition
 T. Kato et al. / Energy 30 (2005) 2580–2595 2583

of natural gas, partial oxidation of heavy oil, coal gasification, water electrolysis, thermochemical water
decomposition, and photochemical, photoelectrochemical and photobiological processes. SMR, coal
gasification, and water electrolysis are the most important industrial processes for hydrogen production
today.
SMR is a well-established, commercialized process, and the most common method of producing large
quantities of hydrogen. Currently, about 99% of world hydrogen is produced from fossil fuels, primarily
natural gas [5]. Hydrogen production efficiencies of SMR with a capacity of over one million Nm3 -H2/
day are in the range of 63–85% (based on the higher heating value (HHV) of hydrogen), and the
investment costs range from 270 to 500 US$/kW [6–9]. At a capacity of over one million Nm3 -H2/day,
SMR technology has the least-expensive investment cost in comparison with other fossil-fueled
hydrogen-production methods. This process also produces carbon dioxide, one of the main greenhouse
gases, which is unwanted. To obtain pure hydrogen, purification steps in the downstream of the SMR
plant are necessary to remove undesired compounds of this nature.
Water electrolysis is also a well-established technology and the most widely used method of
producing high-purity hydrogen. Several processes are available for water electrolysis, ranging from
established alkaline systems to developing, advanced methods, such as PEM. The conventional
electrolytic methods are known as alkaline water electrolysis, which has been a mature technology for
decades, with efficiencies of around 70–80% (HHV). Efforts are being made to enhance the efficiency of
alkaline water electrolysis by increasing the operating temperature or electrolyzing under pressurization.
The current investment cost of alkaline electrolysis is approximately 500 US$/kW [6,8,10,11]. The
investment cost of hydrogen production is higher for alkaline electrolysis than for SMR, but their
efficiencies are comparable. PEM, rather than an alkaline aqueous solution, is used as the electrolyte and
is considered a promising method because of the extreme volume reduction it involves. Currently, the
investment cost for PEM electrolysis is over 1000 US$/kW [11,12]; the high cost of the components is
the main drawback of this technology. In addition to the high investment cost, the major cost factor of
electrolysis is the electricity, making water electrolysis the most expensive method among the current
commercial processes. Thus, electrolysis is used mainly in small plants.
Renewable energy resources, for example, wind, photovoltaic (PV), solar thermal and hydropower,
are all efficient sources of the electricity required for water electrolysis. Although the hydrogen produced
by renewable-resource-based electricity is very expensive in most cases, it is attractive because it is a
very pure and clean energy carrier. In the long run, hydrogen should be produced by renewable energy
resources to avoid fossil-fuel consumption and greenhouse-gas emissions.
Water electrolysis powered by renewable energy resources would produce only hydrogen and
oxygen, avoiding the emission of CO2. When large quantities of hydrogen are produced from renewable
resources by the water-electrolysis process in the future, by-product oxygen will also be produced on a
large scale. When DC electricity is passed between two electrodes (anode and cathode) immersed in
water, hydrogen collects at the negatively charged cathode and oxygen collects at the positively charged
anode. The main chemical reactions occurring at the two electrodes are:
þ
* Anode : 2H2 O/ O2 þ 4H þ 4e
K

Cathode : 4HC C 4eK/ 2H2

Global reaction : 2H2 O/ 2H2 C O2

2584 T. Kato et al. / Energy 30 (2005) 2580–2595

For example, when electrolysis efficiency is 71%, 5000 kW h of electricity would produce 1000 Nm3
of hydrogen and 500 Nm3 of oxygen. As 250 kW h of electricity is required for oxygen production of
500 Nm3 by cryogenic air separation, the full utilization of by-product oxygen corresponds to the
reduction of electricity consumption to 4750 kW h with electrolysis, raising the electrolysis efficiency to
76%. Electrolysis itself is not attractive for producing oxygen and cannot compete with the other
technology. The by-product oxygen might, however, be useful for making the PEM electrolysis
hydrogen production attractive.
In this case, the balance between by-product oxygen and oxygen demand is very important. In
hydrogen production by water electrolysis, if the oxygen demand is not too large relative to the
possible supply of by-product oxygen, large quantities of by-product oxygen must be wasted.
Oxygen itself is an important industrial gas used in many industries such as blast furnaces, electric
furnaces and glass melting. The by-product oxygen can thus be sold to these industries, reducing
the nominal cost for producing hydrogen by PEM electrolysis. On the other hand, if a large oxygen
consumer produces oxygen by PEM electrolysis, it can put the hydrogen on the market. In this
study, therefore, we discuss the potential demand for oxygen on the basis of a survey conducted
both on current oxygen demand and on new technologies utilizing oxygen for improving energy
efficiency.

4. Oxygen demand

Fig. 2 shows the oxygen demand in Japan. Total oxygen demand was about 9615!106 Nm3 in 2001.
There are a number of new applications utilizing oxygen to improve the energy efficiency of industrial
processes. Following are details of some types of industrial oxygen demand.

Fig. 2. Oxygen demand in Japan.

 T. Kato et al. / Energy 30 (2005) 2580–2595 2585

4.1. Blast furnaces

The blast furnace process involved in steel making is industry’s largest oxygen consumer. In the very
latest blast furnace process, oxygen is highly utilized to improve productivity. In Japan, blast furnaces
consumed oxygen of 8037!106 Nm3 in 2001, corresponding to 84% of the total oxygen demand.
Oxygen for blast furnaces is normally produced on site because of the large demand for it and, normally,
the cryogenic air-separation process is utilized. In Japan, almost 70% of cryogenic air separation is used
for supplying oxygen to blast furnaces, but as 75% of these systems were built before 1980, the existing
systems are going to be replaced with new oxygen-production systems in the near future. At that time, if
the hydrogen demand for FCVs is large enough and some of the hydrogen is produced by renewable
electricity, an electrolysis hydrogen production system for supplying hydrogen and by-product oxygen
to FCVs and blast furnaces, respectively, would be a good option.

4.2. Electric arc furnaces

With large quantities of the oxygen used in blast furnaces being produced and consumed on site, the
largest commercial consumer of oxygen is the electric arc furnace. In 2001, the oxygen demand for
electric arc furnaces was about 888!106 Nm3, corresponding to 9.2% of the total oxygen demand in
Japan with the adsorption process for producing oxygen being the main one utilized. In the electric arc
furnace process, production efficiency has been improved by increasing the amount of oxygen. In the
conventional electric arc furnace process, the electricity consumption rate/t of steel production was
about 380 kW h/t, with 33 Nm3/t of oxygen being consumed. Recently, the innovative electric furnace
developed by NKK of Japan, has been shown to improve heat efficiency remarkably [13]. This newly
developed electric arc furnace process is a fully closed melting-shaft furnace, into which scrap is
continuously loaded from the top of the shaft. As the scrap falls into the furnace, it mixes with molten
steel and melts rapidly. Because of the furnace structure, which enables scrap to be fed nonstop into
molten steel, heat loss on the water-cooled panels and fingers is effectively prevented thereby
maintaining the scrap-preheating temperature at around 1000 8C and resulting in lower electricity
consumption. In the new electric arc furnace, the consumption of electricity is only 150 kW h/t with
oxygen consumption of 45 Nm3/t. To produce 1 t of steel, therefore, the electricity input is 230 kW h
lower in the new electric arc furnace, requiring an additional 12 Nm3 of oxygen. As a result, the
reduction of electricity consumption per unit of oxygen use is 19 kW h/Nm3 -O2, which corresponds to a
reduction in primary energy of 182 MJ/Nm3 -O2 where the efficiency of electric power generation (he) is
38%.
In Japan annually, 29!106 t of steel are produced by electric arc furnace according to Fig. 2. If all
existing electric arc furnaces are replaced with newly developed electric arc furnaces, there would be a
reduction of 6634 GW h/yr of electricity (or 62,849 TeJ/yr with heZ38%), plus an additional oxygen
demand of 346!106 Nm3/yr, increasing total oxygen demand for electric arc furnaces to some 1300!
106 Nm3/yr.

4.3. Glass melting

Glass manufacture is a high-temperature, energy-intensive process. The majority of large glass tank
furnaces incorporate regenerative heat-recovery systems. As structural heat losses from such furnaces
2586 T. Kato et al. / Energy 30 (2005) 2580–2595

Fig. 3. Comparison between air-blown combustion and oxygen-blown combustion in glass melting.

typically account for some 40% of total heat input, improving efficiency can lead to significant
reductions in energy consumption. In a conventional air-blown combustion melting process, the energy
requirement is approximately 11 MJ/kg -glass [14]. Overall efficiency can be improved by rebuilding the
furnace to incorporate a number of energy-saving features, namely electric boost and supplementary
oxygen-blown combustion. Fig. 3 shows the comparison of energy consumption, NOx emission and CO2
emission between a conventional air-blown combustion furnace and a new oxygen-blown combustion
furnace. By utilizing oxygen, energy efficiency can be improved by 40%. NOx and CO2 emissions can
also be reduced outstandingly. In the oxygen combustion furnace, the ideal oxygen requirement is about
0.3 Nm3/kg -glass. Taking into account energy consumption per unit of production shown in Fig. 3, the
reduction of primary consumption per unit of oxygen use is 16 MJ/Nm3 -O2.
For cost reasons, the oxygen-blown combustion furnace is not used very much except in electric glass
production where the price of the glass product is relatively high compared with other glass products in
general. Therefore, if fossil fuel prices increase, the shift to the oxygen combustion furnace would be
accelerated. In Japan, the annual production of glass products including sheet glass, glass fiber wool
products, glass fiber continuous textiles, glass foundation products and glass containers was
approximately 4.4!106 t glass in the last 3 years. If these glass products were produced using the
oxygen-blown combustion furnace, the estimated oxygen requirement would be approximately 1313!
106 Nm3/yr. This would reduce the annual energy consumption of the glass-melting process by
20,951 TJ/yr.

4.4. Electric power plant

An electric power plant has the potential to be one of the largest consumer of by-product oxygen. We
are working on the development of a pure oxygen combustion burner for power-generation systems [15].
Based on the development of this, we have proposed a new concept for a power generation system with a
pure-oxygen/blown-natural-gas combined cycle (NGCC). Fig. 4 shows the schematic flow diagram of
our proposed system in which 85% of the exhaust is recycled as a working medium for the gas turbine.
One of the most important features of our proposed system is the exhaust component, consisting of only
CO2, H2O, and O2. After the condensation of H2O, CO2 can easily be captured without the CO2
separation unit. This feature could be an advantage over a conventional NGCC with air-blown
combustion, which is currently the cheapest option for electricity production. At present, electric power
plants with a CO2 -capture unit attract widespread attention as being environmentally benign
 T. Kato et al. / Energy 30 (2005) 2580–2595 2587

Fig. 4. Oxygen combustion natural gas combined cycle power plant.

installations. Although the capital cost of a conventional NGCC is lower than that of power plants using
other fossil fuels, the capital cost, including CO2 -capture unit, is estimated at about 1000 US$/kW, two
to three times higher than without CO2 capture [16–19]. The total thermal efficiency is around 45%
instead of the 53% with conventional NGCC without CO2 capture, as CO2 separation from flue gas
requires a large amount of power. As our proposed system does not need to have a CO2 separation unit, it
could be as cost-effective an option as a CO2 -capture power plant fueled with natural gas. The economic
assessment of our proposed system is not performed in this paper, but will be carried out as a future
study.
As shown in Fig. 4, the oxygen requirement of our proposed system is 72 kg/s (50 Nm3/s) for
producing 400 MW of electricity at the sending end. If this system is operated with a load factor (LF) of
80%, the annual oxygen requirement is 1273!106 Nm3/yr, corresponding to 13% of the current total
oxygen demand in Japan.

4.5. Gasification

A gasification process converts any carbon-containing material into a synthesis gas or syngas
composed primarily of carbon monoxide and hydrogen, which can be used as a fuel to generate
electricity or steam, or as a basic chemical building block for a large number of uses in the petrochemical
and refining industries. Typical raw materials used in gasification are coal, petroleum-based materials
(crude oil, high-sulfur fuel oil, petroleum coke, and other refinery residuals), gases, or materials that
would otherwise be disposed of as waste, for instance, MSW. If the syngas is to be utilized to produce
electricity, it is typically used as a fuel in an integrated gasification combined cycle (IGCC) power-
generation configuration. The syngas can also be processed using commercially available technologies
to produce a wide range of products, fuels, chemicals, fertilizer, or industrial gases.
Gasification is an endothermic chemical reaction, so heat has to be supplied externally or by partial
combustion. In the case of partial combustion, the oxidant for the gasification process can be either
atmospheric air or pure oxygen. In the simplest gasification system with a partial combustion process, air
2588 T. Kato et al. / Energy 30 (2005) 2580–2595

Fig. 5. Cumulative capacity of commercial gasification projects in the world [19].

is used. Although air gasification itself is relatively cheap, downstream gas cleaning is expensive
because of the large volumes that need to be handled. On the other hand, oxygen-blown gasifiers offer a
higher-heating-value gas and faster reaction rates than air-blown systems, but have the disadvantage of
additional capital costs associated with the oxygen plant. The use of by-product oxygen for gasification
might solve the economic problem inherent in using oxygen.
Based on the gasification plant survey, there are 163 commercial gasification projects consisting of a
total of 468 gasifiers [20]. Fig. 5 shows the cumulative capacity of gasification projects. The capacity is
large in coal and petroleum gasification and increases rapidly. Currently, gasification is not commonly
used for household waste. Nevertheless, the most significant growth in the market for waste-
management systems in the near future is likely to be for the treatment of MSW [21]. The requirement of
oxygen depends on the feedstock to be gasified. For example, the Api Energia IGCC plant being built at
Falconara Marittima on Italy’s Adriatic coast requires 62 t/h of oxygen to gasify 59.2 t/h of high-sulfur
heavy oil produced by the Falconara refinery, convert it to syngas, and use the gas to generate 280 MW
of electricity, plus steam and other gases for use in the refinery [22]. When this plant is operated with the
annual LF of 80%, the annual requirement of oxygen is 304!106 Nm3/yr.

4.6. Medical care

Because oxygen gas used for medical care is categorized as a medical supply, it is produced and
delivered within a highly controlled environment. Pure and clean oxygen produced by electrolysis is thus
suitable for medical use. The oxygen demand for medical use was about 105!106 Nm3 in 2001 in
Japan, which is only 1.1% of the total oxygen demand, including the oxygen produced on site for blast
furnaces, however, oxygen for medical use is the third largest sector for oxygen demand in Japan.
Utilizing oxygen for medical purposes is not related to the improvement of the energy system, but
because of the very high price of medical oxygen, the effective utilization of by-product oxygen might
decrease the high cost of electrolysis hydrogen production, especially electrolysis using PEM. We
surveyed the retail price of oxygen for medical use at 877 hospitals in Japan, as shown in Fig. 6.
Depending on total oxygen demand, the oxygen price is basically very high, ranging from 60 yen/Nm3
(0.82 US$/Nm3) to 10,000 yen/Nm3 (35 US$/Nm3). These prices almost correspond to the fuel price on
an Nm3 basis.
 T. Kato et al. / Energy 30 (2005) 2580–2595 2589

Fig. 6. Retail price of oxygen for medical use in Japan.

In Section 6, we discuss how economical it is to use by-product oxygen from electrolysis hydrogen
production for medical purposes based on a survey of oxygen demand and pricing in hospitals.

5. Potential for consumption of by-product oxygen

Hydrogen demand in Japan is estimated to be approximately 15,000–20,000!106 Nm3/yr. As most

hydrogen for chemical and steel plants, for instance, is produced and consumed on site, the hydrogen
demand in the market place is only about 150!106 Nm3, in other words less than 1% of actual
consumption. Fig. 7 shows hydrogen and oxygen demand in the last 5 years in the market place in Japan,

Fig. 7. Hydrogen and oxygen demand on Japanese market.

2590 T. Kato et al. / Energy 30 (2005) 2580–2595

with oxygen demand being about nine times greater than hydrogen demand, although large quantities of
oxygen are also produced and consumed on site.
In the future, there will be new demands for hydrogen. Replacement of fossil fuels by hydrogen in
vehicles throughout the world is predicted to occur over the next 50 years as mass production of FCVs
accelerates. For example, Japan’s target for installing FCVs is 50,000 vehicles in 2010 and 5,000,000
vehicles in 2020, with an estimated hydrogen demand of 160!106 Nm3 in 2010 and 4250!106 Nm3 in
2020, respectively [23]. In the case of southern California, the projected hydrogen demand is 630!
106 Nm3 in 2020, assuming cumulative numbers of FCVs of 350,000 for passenger cars, 150,000 for
light trucks and 330 buses [24]. Fuel-cell technology for stationary use will also consume large
quantities of hydrogen.
The new hydrogen demand must be met by a hydrogen production system in addition to existing
systems. If part of the incremental hydrogen requirement is met by the water electrolysis process, half
the number of moles of oxygen will be produced simultaneously as a by-product of hydrogen production.
In this study, assuming that the estimated hydrogen demand for vehicle use in 2020 in Japan is met by
water electrolysis, we compared the amount of by-product oxygen supply and potential demand for
oxygen mentioned above. When all hydrogen for vehicle use in 2020 is produced by water electrolysis,
the available by-product is about 2125!106 Nm3. On the other hand, the current oxygen demand is four
times more than the available by-product oxygen as shown in Fig. 2. In addition, as mentioned above, the
potential oxygen demand is 346!106 Nm3/yr in electric arc furnaces and 1313!106 Nm3/yr in glass
melting, where energy efficiency would be improved by the utilization of oxygen. When oxygen-blown
NGCC with a 400 MW e capacity is installed, the annual requirement for oxygen is 1273!106 Nm3/yr
(LFZ80%). Fig. 8 shows the new oxygen demand in these three processes (electric arc furnace, glass
melting and oxygen-blown NGCC on the horizontal axis). In Fig. 8, the potential supply of by-product
oxygen is shown by the dotted line. The total oxygen demand in only three processes is larger than the
potential supply of by-product oxygen from water electrolysis hydrogen. In addition to the oxygen
demand shown in Fig. 8, there would be large quantities of oxygen demand in other processes in the
future. Consequently, although the assumption that the future hydrogen demand would be met by
water electrolysis looks unrealistic, even in a situation such as this, the opportunity for consuming
the by-product oxygen would exist.

Fig. 8. Comparison between potential supply of by-product oxygen and oxygen demand.
 T. Kato et al. / Energy 30 (2005) 2580–2595 2591

In Fig. 8, the perpendicular axis shows the reduction of primary energy consumption per unit of
oxygen use. The colored square in Fig. 8 means ‘potential oxygen demand’ times ‘primary energy
reduction per unit of oxygen use’. Therefore, each colored square shows the potential of primary energy
reduction by the large-scale utilization of by-product oxygen of each technology. The energy-reduction
potential of utilizing oxygen in the electric-furnace sector is large, though the oxygen demand is not
large. The reduction of primary-energy consumption in oxygen-blown NGCC, which is the reduction
from oxygen-blown NGCC with a cryogenic air-separation system, is small compared with the other
industrial processes, though there would be much potential demand. The total reduction in primary
energy consumption by utilizing by-product oxygen in these three processes reaches about 89,000 TJ/yr.
In addition to the oxygen demand shown in Fig. 8, there would be large quantities of oxygen demand in
other processes. As a result, if by-product oxygen from water electrolysis hydrogen production is fully
utilized, it could contribute to improving the energy efficiency of many industrial processes.

Fig. 9. Electricity and heat-supply systems to hospital.

2592 T. Kato et al. / Energy 30 (2005) 2580–2595

6. Economics of utilizing by-product oxygen for medical purposes

6.1. Evaluation model

In this study, we performed an economic assessment of utilizing by-product oxygen for medical use.
Assuming two energy systems, as shown in Fig. 9, we calculated the difference in operational cost for
meeting electricity and heat demand in hospitals. In both systems, electricity and heat demand are
mainly met by proton exchange membrane fuel cell-co-generation system (PEFC-CGS) with utility
electricity and city gas compensating for the shortfall. The only difference between the two energy
systems is the hydrogen-production process, (i.e. PEM water electrolysis in system-A and SMR in
system-B). The difference in operational cost between the two systems therefore corresponds to the
difference in the cost of hydrogen production, including the cost of investment in PEM or SMR. Table 1
shows the assumed hydrogen cost produced by PEM and SMR [12,22] together with the retail price of
utility electricity and city gas used in this study. The hydrogen production cost is assumed to be about
63 yen/Nm3 (0.58 US$/Nm3) for PEM electrolysis, and 25 yen/Nm3 (0.23 US$/Nm3) for SMR. The
hydrogen cost of PEM is as much as 255% that of SMR. By using these hydrogen production costs, the
breakeven cost of by-product oxygen is calculated to be:
ð0:58 K 0:23ÞðUS=Nm3 of H2 Þ=0:5ðm3 of O2 =Nm3 of H2 Þ Z 0:7ðUS=m3 of O2 Þ (1)
This cost is much cheaper than the actual retail price of oxygen in most hospitals in Fig. 6, although
the cost of filling cylinders/tanks and shipping should be added to this cost to make it comparable with
the actual cost of oxygen used in the hospitals. Therefore, if the by-product oxygen is produced
somewhere else, and then transported and fully utilized in hospitals, it might offset the high cost of PEM
electrolysis hydrogen production. The by-product oxygen is still wasted, however, if there is not enough
demand for it. In this paper, therefore, we conducted the study on the utilization of the by-product
oxygen assuming on-site hydrogen and oxygen production by PEM electrolysis in hospitals.
In system-A, we assumed that by-product oxygen is available without any additional cost, while oxygen
cost is required in system-B. The oxygen cost in system-B is based on the actual price from a survey taken
Table 1
Assumption of electricity price and city gas retail price
Electricity Variable 14.47 yen/kW h (0.13 US$/ July–September
kW h)
13.15 yen/kW h (0.12 US$/ Others
kW h)
Fixed 1625 yen/kW/month
(15.08 US$/kW/month)
City gas Variable 50 yen/Nm3 (0.46 US$/
Nm3)
Hydrogen PEM 63 yen/Nm3 (0.58 US$/ Include investment cost
Nm3)
SMR 25 yen/Nm3 (0.23 US$/ Include investment cost
Nm3)
Lifetime of hydrogen production system: 10 years for PEM and 20 years for SMR annual interest rate; and 5% for PEM and
SMR.
 T. Kato et al. / Energy 30 (2005) 2580–2595 2593

Table 2
Assumption of electricity and heat demand in hospital
Annual demand Maximum demand
2
Electricity 170 kW h/m /yr 50 W/m2
Hot water 80 Mcal/m2/yr 40 kcal/m2/h
Heating 74 Mcal/m2/yr 82 kcal/m2/h
Cooking 80 Mcal/m2/yr 90 kcal/m2/h

at 877 hospitals shown in Fig. 6, ranging from 60 yen/Nm3 (0.82 US$/Nm3) to 10,000 yen/Nm3 (35 US$/
Nm3). In the following calculation, we focus on 160 hospitals where data on total floor space is available.
In most of the 160 hospitals, CGS is currently not installed. Therefore, to estimate the hydrogen
demand for PEFC-CGS in hospitals, we surveyed the size of existing conventional-type CGSs already
installed in hospitals, and examined the relationship between the size of CGS and the total floor space of
the hospital. As a result, we found the size of CGS per 1 m2 total floor space is 0.0194 kW on average.
We then formulated the relationship between the capacity of CGS and the total floor space of the hospital
as shown in Eq. (2)

capacity of CGS in hospital ðkWÞ Z 0:0194 !total floor space ðm2 Þ (2)

For example, when the total floor space is 10,000 m2, a suitable electrical capacity for the CGS would
be 194 kW e. As actual demand data is not available for each of the 160 hospitals, we used the typical
demand data set on electricity and heat as shown in Table 2. The maximum electricity demand in
hospital is assumed to be 0.05 kW/m2. Considering the capacity of PEFC-CGS assumed in Eq. (2),
almost 40% of maximum electricity demand can be met by PEFC-CGS.

6.2. Economic assessment for utilizing by-product oxygen

Assuming the constant output operation of PEFC-CGS throughout a year, we calculated the annual
hydrogen consumption and by-product oxygen supply for 160 hospitals. Fig. 10 shows the results.

Fig. 10. Comparison between annual oxygen demand and available by-product oxygen.
2594 T. Kato et al. / Energy 30 (2005) 2580–2595

Fig. 11. Comparison of hydrogen cost between system-A and -B including oxygen cost.

The by-product oxygen supply is at least 10 times larger than the actual oxygen demand. In terms of the
balance of demand and supply, the use of by-product oxygen in hospitals is not a sound application. As
mentioned above, however, the oxygen price for medical use is very high. Therefore, even if large
quantities of by-product oxygen are wasted, its use could have some economic merit.
By comparing the actual oxygen cost and the estimated energy cost required for electricity and heat
supply to hospitals, we made an economic assessment for utilizing by-product oxygen in hospitals. Fig. 11
shows the hydrogen cost in system-A relative to the total cost of hydrogen and oxygen in system-B. In this
study, the hydrogen cost of PEM is assumed to be as much as 255% that of SMR. When the oxygen cost in
system-B is taken into account, the cost in system-A decreases. In some hospitals, the hydrogen cost in
system-A is almost the same as that in system-B, including oxygen cost. Consequently, the effective
utilization of by-product oxygen in hospitals where the oxygen price is very high for safety reasons would
contribute to reducing the relative cost of hydrogen produced by PEM electrolysis. It looks meaningless if
only 10% of by-product oxygen is effectively used because of the assumption that all the hydrogen for
PEFC is produced by electrolysis. Therefore, some hydrogen should be produced by SMR to eliminate the
wasted by-product oxygen. We will examine an optimal condition regarding the capacity of PEFC-CGS
and the ratios of hydrogen production by electrolysis and SMR as a next step in this work.

7. Conclusion

In this study, we discussed the potential demand for by-product oxygen and its contribution to energy
saving. As there is a potential demand for large quantities of oxygen, there might be substantial potential
for utilizing the by-product oxygen of electrolysis hydrogen production. If the by-product oxygen can be
utilized effectively, it would contribute to the improvement of the energy efficiency of various industrial
processes and electric power production, as well as the reduction of CO2 emissions. Even if large
quantities of by-product oxygen are wasted, its use in hospitals would have economic merit.

References

[1] Barreto L, Makihira A, Riahi K. The hydrogen economy in the 21st century: a sustainable development scenario. Int
J Hydrogen Energy 2003;28(3):267–84.
 T. Kato et al. / Energy 30 (2005) 2580–2595 2595

[2] Jianwei Y, Guolong S, Cunjian K, Tianjun Y. Oxygen blast furnace and combined cycle (OBF-CC)—an efficient iron-
making and power generation process. Energy 2003;28(8):825–35.
[3] Oie I. Gas georama in Japan. Gas Rev 2002 [in Japanese].
[4] Japan Industrial Gas Association. Monthly statistics. Japan Industrial Gas Association; 2003. See also: http://www.jiga.gr.
jp/english/statistics/Monthly_Statistics/index.html.
[5] IEA (International Energy Agency). Hydrogen—today and tomorrow. IEA Greenhouse Gas R&D Programme; 1999. See
http://www.ieagreen.org.uk/h2rep.htm.
[6] Wurster R, Zittel W. Hydrogen energy. Published at the workshop on energy technologies to reduce CO2 emissions in
Europe: prospects, competition, synergy. Energieonderzoek Centrum Nederland ECN, Petten, The Netherlands, 11–12
April; 1994.
[7] Berry GD. Hydrogen as a transportation fuel: costs and benefits. Report UCRL-ID-123465. Livermore, CA: Lawrence
Livermore National Laboratory, US Department of Energy; 1996.
[8] Basye L, Swaminathan S. Hydrogen production costs—a survey. SENTECH Inc., report DOE/GO/10170-778.
Washington, DC: US Department of Energy; 1997.
[9] Block DL. Comparative costs of hydrogen produced from photovoltaic electrolysis and from photoelectrochemical
processes. Proceedings of the 12th World Hydrogen Energy Conference, Argentina; 1998.
[10] Mann MK, Dipietro JP, Iannucci JJ, Eyer JM. Exploring the technical and economic feasibility of producing hydrogen
from sunlight and wind. Proceedings of the 12th World Hydrogen Energy Conference, Argentina; 1998.
[11] Rogner HH, Scott DC. Building sustainable energy systems: the role of nuclear-derived hydrogen. Paper presented at
Nuclear Production of Hydrogen, First Information Exchange Meeting in France; 2000.
[12] WE-NET (World Energy Network). International cooperation in research and development of clean energy system with
particular emphasis on hydrogen. Phase II, Task 8: development of hydrogen production technology. Annual report on
FY1999; 1999.
[13] JFE Steel Corporation. Ecologically friendly and economical arc furnace "ECOARC". Japan: JFE Steel Corporation;
2000. See also: http://www.jfe-holdings.co.jp/archives/nkk_360/No.45/45n1.html [in Japanese].
[14] Yamane M. Handbook of glass engineering; 1999. p. 322–324 [in Japanese].
[15] Pak H, Nikzat H, Fuse T, Kobayashi N, Hasatani M. Bench scale experiment of pure oxygen pulverized coal combustion
in high temperature. Proceedings of the ninth Asia Pacific Confederation of Chemical Engineering (APCChE) Congress
and CHEMECA; 2002 p. 332–342.
[16] Herzog H, Vukmirovic N. CO2 sequestration: opportunities and challenges. Seventh Clean Coal Technology Conference,
Reno, Nevada; 1999.
[17] Delallo M, Buchanan T, White J, Holt N, Wolk R. Evaluation of innovative fossil cycles incorporating CO2 removal.
Proceedings of Gasification Technologies Conference, Queensland, Australia; 2000.
[18] Audus H. Leading options for the capture of CO2 at power station. Paper presented at Fifth International Conference on
Greenhouse Gas Control Technologies (GHGT-5); 2000; p.91–6.
[19] David J, Herzog H. The cost of carbon capture. Proceedings of the Fifth International Conference on Greenhouse Gas
Control Technologies (GHGT-5); 2000 p. 985–990.
[20] NETL Gasification Technologies. Worldwide Gasification Database. Washington, DC: National Energy Technology
Laboratory, US Department of Energy; 2000. See also: http://www.netl.doe.gov/coalpower/gasification/models/
dtbs(excel).PDF.
[21] Juniper Consultancy Services. The market for pyrolysis and gasification of waste in Europe: a technology and business
review; 1998.
[22] Spence JL. Api ENERGIAIGCC plant status. Proceedings of Gasification Technologies Conference, Queensland,
Australia; 2000.
[23] WE-NET (World Energy Network). International cooperation in research and development of clean energy system with
particular emphasis on hydrogen, Phase II, Task 1 investigation and study for system evaluation. Annual report on
FY2002, Japan; 2003.
[24] Ogden JM. Developing an infrastructure for hydrogen vehicles: a southern California case study. Hydrogen Energy 1999;
24(8):709–30.