|
Cost of Ownership Issues for Hydrogen Gas Purification
Eric Bretschneider
UNIROYAL Optoelectronics
Abstract
In compound semiconductor growth, the choice of gas
purification methods has traditionally been based solely
on technology. Little attention has been paid to the
operating cost associated with different gas purifiers.
In order to compete in high volume markets, the operating
cost and specifically the cost of ownership should be
a crucial factor in selecting a gas purification method.
Accordingly, a detailed study has been completed to
compare the cost of ownership for a variety of gas purification
methods at varying gas flow rates. Depending on the
volume of gas being purified, the cost of ownership
for different purification methods can vary widely.
The impact of bulk compressed or liquid hydrogen and
point of use versus bulk purification on the cost of
ownership is shown to be particularly significant.
Background
From its very beginning, decisions relating to the
compound semiconductor industry (CSI) have been made
based on technology. The early support of the defense
industry that was vital to the development of today's
compound semiconductor industry placed a premium on
performance. There was little reward or incentive for
manufacturers to reduce costs. As the industry has matured,
remnants of this attitude remain. It is still not uncommon
to hear sales pitches that focus on technology and performance
with little attention paid to the reality of economics.
Continued incremental advances in silicon based technologies
have pushed the performance of silicon based devices
into realms that as recently as five years ago were
thought to be the exclusive domain of compound semiconductors.
Although compound semiconductors have managed to flourish
in many niche markets such as optoelectronics where
the indirect band gap of silicon places it at a significant
disadvantage, manufacturers need to pay close attention
to economic considerations in order to successfully
compete against existing and emerging elemental semiconductor
technologies.
Additionally, the technologies used for MOCVD growth
of compound semiconductors have advanced to the stage
where most of the commercial equipment is capable of
growing materials and structures of comparable quality
and performance. Under these circumstances, it behooves
the manufacturer to pay close attention to the relative
economic capabilities of different types of hardware.
Indeed, it is not uncommon to find that many hardware
manufacturers who supply the compound semiconductor
industry to quote Cost of Ownership (CoO) metrics for
their equipment. Most of these metrics are based on
the SEMI E-35 standard for CoO. While the growing prevalence
of CoO metrics helps to illustrate the relative maturity
of the CSI, it is important to realize that these metrics
are typically based on a generic version of a process
and therefore may not correlate well to the actual economics
of operating hardware in a manufacturing environment.
Cost of Ownership
One of the first steps in the design and specification
of a new manufacturing facility or the expansion of
an existing facility, should be an economic analysis
to determine the magnitude of the potential revenue
stream. At the concept stage, the uncertainty in this
estimate may exceed 50%! Despite this large potential
error, it is still critical to begin a cost model at
the earliest stages of design. Careful attention to
economics at all stages of design will help to maximize
the revenue stream from a facility and may also help
to avoid unnecessary design efforts and expenses. As
the details of the design are refined, the estimated
costs and revenues from the model will begin to converge.
It is not unreasonable to achieve economic estimates
within 5% of the actual costs and revenue streams before
construction on a facility has even begun.
As mentioned above, many of the details required to
calculate an accurate CoO metric are not available to
hardware manufacturers. These include, but are not limited
to: facilities design (including support equipment),
layout, installation costs and specific details of the
manufacturing process. The details of a manufacturing
process potentially have the greatest impact on the
economics of a facility. Key factors that may be affected
include: total cycle time (wafer throughput), labor
requirements, the maintenance schedule (uptime, spare
parts inventory) and also impact the choice of raw materials
(process yield, cost, safety stock, maintenance schedule).
In light of uncertainty of the CoO metrics obtained
from equipment manufacturers, it would be useful to
establish a new metric that encompasses all of the economic
factors in a process. Such a Global Economic Metric
(GEM) that incorporates all the factors that impact
the economic viability of a process would provide a
single figure of merit that would be perhaps the most
useful and easily understood metric for comparing hardware,
raw materials, facility and process configurations.
By necessity, any GEM model will require a tremendous
amount of information that may or may not be available
to equipment suppliers. Uncertainties in the input data
are best handled by performing sensitivity analyses
for input parameters and using statistical methods to
generate an uncertainty factor. A full discussion of
the preferred methods is beyond the scope of this article,
but can be found in standard references on derivatives
and statistics.
Hydrogen Delivery Systems
Before discussing the technologies available for hydrogen
purification, it is important to discuss the choices
for the storage/delivery of bulk hydrogen: compressed
(BHY) and liquid hydrogen (LHY). As will be seen later,
this decision may have a significant impact on the overall
economics of hydrogen purification.
For small facilities, BHY is often the preferred form
of delivery. Typical storage and delivery pressures
are 3000 psig and 200 psig respectively. Depending on
the supplier, delivery specifications may range from
99.95% to 99.995%. Facility configurations may require
transfilling high volume cylinders that are permanently
installed on site, or banks of high volume cylinders
on flat bed trailers may be exchanged. Both of these
options impose different requirements on the facilities
design team. While permanently mounted cylinders minimize
the total square footage required they result in the
highest unit cost of material due to the inefficient
delivery mode.
With a minimum purity specification of 99.9995% and
a unit cost that is 30% to 50% that of BHY, LHY would
seem to be the obvious choice for a new or existing
facility. Unfortunately, the higher purity of a LHY
system comes at a significant price with respect to
the design of facility. It is vitally important that
strict attention is paid to all OSHA , NFPA and local
codes related to LHY systems at every stage of design.
Adherence to these codes may require explosion proof
electronics and no permanent openings or flammable materials
within a significant radius of any component of the
system. LHY systems also have a low delivery pressure
of ~100 psig. Although this pressure may be increased
by use of a hydrogen compressor, this option may negatively
impact the delivered purity and may significantly increase
maintenance requirements for the system.
One additional issue with operation of a LHY system
is indirectly related to its extremely low temperature.
Unless there is a constant draw of hydrogen, the system
will vent an appreciable amount of hydrogen. This typically
limits the use of LHY systems to facilities that operate
on a 24/7 schedule. It should be noted that although
LHY has a significantly lower unit cost than BHY, at
low consumption rates, BHY provides the lowest operating
cost. This is due to a combination of several factors
including simpler facilities design, loss of hydrogen
due to venting during low draw periods of operation
on a LHY system, and a lower monthly service charge
for hardware leased from a gas supplier.
|
