United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Research Triangle Park NC 2771
Research and Development
EPA-600/S7-84-014 Mar. 1984
SERA Project Summary
Environmental Assessment:
Source Test and Evaluation
Report - Rectisol Acid Gas
Removal
K. W. Crawford and R. A. Orsini
In June 1982, a Rectisol acid gas
removal unit at a Texaco refinery at
Wilmington, CA. was tested. The
primary goal of the test was to provide a
data base for evaluation of Rectisol
performance in entrained coal gasifica-
tion applications. This Rectisol unit
processes gases from the partial oxida-
tion (POX) of oil by the Texaco Synthesis
Gas Generation Process. A secondary
goal of the test was to validate samp-
ling/analytical methods for constituents
such as HCN and NH3 which were ex-
pected to be present at very low levels in
the Wilmington gas samples.
During the tests, the Rectisol unit
handled a feed gas containing about
0.2% H2S, a few ppmv COS, and >30%
CO2 and generated a product gas con-
taining <0.1 ppmv total sulfur and
usually <200 ppmv CO2. CO2-rich
offgases from the Rectisol unit
contained <10 ppmv total sulfur. With
regard to minor constituents, the CO2-
rich offgases contained about 400-900
ppmv of methanol (derived from the
process solvent), <1 ppmv of hydrocar-
bons other than methane and ethane,
and <1 ppmv each of HCN or NH3. Due
to their absence in the vacuum residual
oil feed to the POX unit, no detectable
amounts of volatile heavy elements
(e.g., Hg, As, Se) were found in Rectisol
gases.
Data from the test are evaluated in
the report from the standpoint of
differences in feed gas composition
which would be encountered in coal vs.
oil applications and/or where only part
of the gas is shifted.
This Project Summary was developed
by EPA's Industrial Environmental Re-
search Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title /see Project Report ordering infor-
mation at back).
Introduction
EPA's Industrial Environmental
Research Laboratory, Research Triangle
Park, NC, (IERL—RTP), is responsible for
the Agency's environmental assessment
activities and for control technology
development research relating to
substitute fuels from coal. Part of the
assessment activities involve generating
background data on the characteristics of
waste streams generated by coal gasifi-
cation and associated technologies and
evaluation of the performance of control
technologies for such waste streams.
IERL-RTP assessment programs involve
three general areas: (1) collecting and
evaluating existing process and
environmental data, (2) acquiring supple-
mentary data by sampling and analyzing
selected process/waste streams at
domestic or foreign facilities, and (3)
environmental assessment and
necessary process engineering support
studies.
The proposed widespread use of
Rectisol in U.S. coal gasification facilities
and the importance of removed acid gas
as potential atmospheric emissions from
such facilities prompted EPA to investi-
gate a test program at an existing unit.
Currently, the only Rectisol units
-------
operating in coal gasification applications
are outside the U.S. However, two units
exist in the U.S. which process gases
from the Texaco synthesis gas generation
process (partial oxidation, or POX) of
heavy oil. After exploring both overseas
and domestic units, the Rectisol unit at
Texaco's Wilmington refinery was
selected for testing. Although this unit
processes shifted gas from oil (rather
than coal gasification), the design is
similar to at least one planned for a U. S.
coal gasification facility; hence, the data
for this unit can be useful for assessing
coal-based units.
The goals of the Wilmington Rectisol
test were threefold: (1) to evaluate the
performance of the specific unit from an
environmental standpoint by character-
izing both process and waste streams for
a variety of species of environmental
interest; (2) to establish the applicability
of key sampling and analytical techniques
as applied to the matrices of Rectisol
streams; and (3) to provide test data as a
partial basis for estimating the environ-
mental effects of Rectisol units in coal
gasification applications.
The Facility
The Rectisol unit at the Texaco refinery
in Wilmington, CA, designed and built by
Linde A. G. of Munich, Federal Republic
of Germany, is part of a hydrogen produc-
tion train referred to as the hydrogen
generation unit (HGU). The HGU, based
on gasification of vacuum residual oil to
generate synthesis gas, provides part of
the hydrogen used in the refinery. The
Texaco POX gasifier generates a hot
crude gas containing about 43% H2, 50%
CO, 5% CO2, 0.5% H2S plus COS, and
entrained soot consisting primarily of
ungasified carbon. The soot is removed by
water washing in a carbon scrubber
which also cools the gas and recovers
heat. Dust-free cooled gas is fed to a sour
(sulfur tolerant) shift conversion unit.
Shifted gas is cooled in stages with steam
production and condensate recovery.
Cooled condensates are either recycled
to the shift reactor or to the carbon
scrubber or discharged to wastewater
treatment. Cooled shifted gas is fed to the
Rectisol unit for removal of H2S and CO2.
For trace CO removal, a copper liquor
wash step follows the Rectisol unit.
The Wilmington Rectisol unit utilizes
high-pressure acid gas removal with a
design capacity of 4000 kmoles/hr of
throughput. The process utilizes cold
methanol as a solvent to remove acid
gases by physical absorption under
pressure, with solvent regeneration by a
combination of pressure reduction,
stripping with inert gas (IN2), and heating.
The wash process was designed to
remove H2S + COS to <1 ppm and CO2 to
10 ppm and to supply an H2S stream with
at least 10% H2S and a CO2 tailgas with 5
ppm maximum H2S.
The Wilmington Rectisol unit is
designed for completely selective
removal of H2S + COS and of C02. While
results from this unit are relevant to
current designs from an environmental
point of view, the economics of the
Wilmington unit are no longer typical.
Modern plants have lower specific energy
requirements and the tendency is
towards higher operating pressures,
higher concentrations of H2S + COS in
the sulfur-rich off-gas, and larger
capacities.
Figure 1 is a simplified flow diagram of
the process indicating streams of interest
for environmental testing. The feed gas
(Stream 1) is cooled against cold
separation products. After separation of
the condensate, H2S and COS are
removed from the gas in the bottom
section of the wash tower with CO2-
loaded methanol. The sulfur-free gas is
washed in the top section of the tower to
remove CO2. The gas leaving the top of
the wash tower (Stream 2) is delivered to
the copper liquor plant where CO is
removed to ppm levels.
The methanol leaving the bottom of the
H2S absorption section of the tower
contains all of the H2S and also CO2 and
H2 in amounts corresponding to their
partial pressures and solubilities.
Hydrogen recovery is increased by
flashing the loaded methanol to an
intermediate pressure and recycling the
flash gas back to the feed gas via a small
recycle compressor.
The flashed methanol still contains too
much CO2 to achieve the necessary H2S
concentration and must be further
flashed and stripped with inert gas in the
bottom section of the H2S enrichment
column. The stripping gas (Stream 6) is
nitrogen obtained from the oxygen plant
which supplies oxygen to the POX units.
The CO2-rich offgas (Stream 3b) from
the top of the H2S-enrichment column is
discharged to the atmosphere. From the
loaded methanol from the bottom of the
H2S-enrichment column, the absorbed
H2S and C02 are stripped in the warm
stripper column with reboiled methanol
vapor. The stripped gases (Stream 4) are
sent to the Claus plant for sulfur recovery.
The main stream of CO2-loaded
methanol leaving the CO2-absorption
section of the wash column, after H2-
recovery by flashing, is let down and
stripped with nitrogen in the cold stripper
to remove much of the dissolved CO2. The
column overhead (Stream 3a) is
discharged to the atmosphere. The cold,
partly stripped methanol is returned to an
intermediate tray of the C02 absorber.
The methanol/water mixture
condensed from the feed gas is separated
in a still which produces pure
methanol overhead and wastewater
(Stream 5) in the sump. This keeps the
water content of the circulating methanol
at a low level.
Test Description
Gas sampling ports are located in pres-
surized gas recycle loopsthrough which a
slipstream of the process gas may be
passed. The recycle configurations and
valving are such that all samples could be
taken at ground level. Samples of the still
bottoms were also available from a
ground level tap. Sample loop
configurations did not permit the use of
probes for flow rate measurements, and
all flow rate, temperature, and pressure
data were provided by Texaco from
existing in-line equipment. The sample
loops were modified and the in-line flow
measurement equipment was calibrated
by Texaco during a scheduled plant
downtime for maintenance.
In most cases, sampling lines were
several meters long. This was of some
concern relative to obtaining representa-
tive samples, particularly of the Rectisol
feed gas which contains both moisture
and water-soluble species (e.g., NH3). The
possible biases due to the loss of water-
soluble species from the feed gas and of
methanol contamination are addressed.
The test period was from June 6
through 17, 1982. The shift and Rectisol
units had been operating continuously for
over a week so that steady operation
existed at the start of the test. Six streams
were sampledduringthetests: Stream 1 -
Rectisol feed gas, Stream 2 - product gas,
Stream 3a - cold stripper offgas. Stream
3b - H2S-enrichment column offgas,
Stream 4 - H2S-rich offgas, and Stream 5
- methanol still bottoms.
The gaseous process/waste stream
samples acquired during the test period
and analyses performed are identified in
the report. Where possible, standard
sampling/analytical techniques were
employed and details of the procedures
are given. The first 3 days were devoted to
equipment setup, preliminary sampling
and characterization to obtain order of
magnitude stream compositions, and
evaluation of the adequacy of propose
-------
H2 Rich
Off Gas*
Stripping^
Gas
Feedgas(T]
COj Rich
Off Gas
Product
Gas
1 t
1
f"^™^^™
Pk
J
©
' '
Methanol
Water
Separation
•Steam
Compressor
Figure 1. Simplified flow diagram of the Wilmington Rectisol unit.
Still Bottoms
sampling and analytical methodologies.
Material balance data were obtained
during the last 5 days of testing.
Test Data Summary and
Conclusions
The Wilmington program resulted in
the collection of both process and
environmental data. Results are briefly
and partially summarized here.
• With regard to the characterization
of process performance, the most
significant parameter to vary in the
Rectisol feed was the H2S concen-
tration, which decreased during the
tests. As a result, roughly
proportional reductions of H2S con-
centration were observed in the
H2S-rich offgas.
• Measured and estimated flow rate
data for the Rectisol process/waste
streams were substantiated by
material/species balance estimates.
• The principal potential pollutant
species in the CO2-rich offgases
were H2S, COS, CO, and methanol.
Other inorganic species detected
were NH3, HCN, and Cl. Some
portion of both NH3 and Cl are
believed to be attributable to
entrained process methanol. Other
than methanol, nonmethane
organic compounds detected were
individually present at concentra-
tions <1.6 ppmv.
• H2 was present at levels of >1% by
volume in the CO2-rich offgases for
the latter part of the test period. This
level is considerably higher than that
for the plant design and may be
related to off-specification operation
of the recycle compressor.
• The principal constituents of
environmental interest detected in
the H2S-rich offgas were H2S, COS,
and methanol. Organics present in
the H2S-rich offgas consisted almost
entirely of methanol; however, ppmv
concentrations of C2 through C6
hydrocarbons, benzene, and toluene
were also detected.
Adequacy of Sampling and
Analytical Techniques
The tests provided the opportunity to
validate certain methods for the levels of
target species and matrices involved and
to identify a number of limitations to
methods which have been used in other
types of tests. For sources involving fuel
combustion, incineration, or sulfur
recovery in refineries and coke plants, the
gas matrices generally contain only a few
percent of CO2. In coal and oil gasifica-
tion, most waste gases contain much
higher levels of C02. Also, Rectisol gas
streams constitute a much more reducing
environment than do combustion gases
so that collecting media designed for
combustion sources are not necessarily
applicable to Rectisol testing. The
following are some of the findings related
to test methods. The findings should be
generally relevant to testing of acid gas
removal and sulfur control units in coal
gasification facilities.
• H2S and COS were readily quantifi-
able by gas chromatography(GC), as
used at Wilmington, in the range of 1
ppmv to about 1%. The relatively
poor results obtained with thermal
conductivity GC for high level H2S
samples (H2S-rich offgas or Claus
feed gas) are thought to relate to
column saturation problems
indicated by tailing peaks.
• H2S can be readily quantified at
percent levels using a wet chemical
procedure adapted from EPA
methods 7 and 11. The method
proved to be quite reproducible at
Wilmington when compared to
thermal conductivity GC data for H2S
at percent levels. Good H2S material
closure around the Rectisol unit was
obtained when these grab sample
-------
data were used, suggesting that the
method has no obvious bias.
• High levels of CO2 in gases present
limitations with the use of alkaline
impmger solutions for collection of
HCN, since C02 tends to neutralize
the solutions with consequent loss
of HCN collection efficiency. An
alternative collection technique
using H2S04 was verified prior to the
Wilmington test and utilized in the
field.
• When levels of H2S exceed a few
percent in the sample gas, the H2SC>4
technique cannot be used unless
high levels of HCN are also
encountered. With high HCN levels,
only small sample volumes need be
taken for sufficient HCN collection
for subsequent analysis. In such
cases the sulfur interference can be
tolerated.
• Ammonia could be collected in acid
impinger solutions from the high
C02 gases at Wilmington. Blank
levels in dilute HCI tend to be much
lower than in dilute H2SO4, and the
detection limit with HCI is about 0.1
ppmv.
• Na2CO3 solutions were used in
impingers to collect HCOOH, HCI,
and HFfrom gas streams. The effici-
ency of collection of these species is
not known si nee their concentrations
were very low.
• The collection efficiency of the heavy
element oxidative impinger train for
high CO2 streams at Wilmington is
believed to be similar to that
observed in sampling of combustion
gases. In the case of the H2S-rich
offgas, the high H2S levels quickly
exhaust the H2O2 in the first
impinger. Thus, an alternative
technique is required for collection
of volatile heavy elements in
streams containing high levels of
reduced sulfur species.
• XAD-2 resin trap samples were
taken of the Rectisol feed gas and
CO2 offgases. Although the C6
through C8 aromatics were present
near or below detection limits, the
XAD data and on-site GC data are at
least reasonably consistent for C7
and C8 aromatics.
Wilmington Rectisol Differences
Related to Coal Gasification
Applications
Several basic features of the Wilming-
ton Rectisol unit would be found in
essentially all Rectisol units, regardless
of the type of service. However, feed
gases which differ greatly in composition
from those at Wilmington will involve
somewhat different designs. Also, the
requirements for product gas purity and
offgas pollutant loadings will impact
Rectisol design. Since the results of the
subject test cannot be utilized for other
units without a consideration of design
differences, some of the factors which
influence Rectisol design and how they
relate to test results are discussed in the
report. In the discussion of other Rectisol
applications, the emphasis is on how the
data from or features of the Wilmington
Rectisol relate to these applications.
The characteristics of Rectisol feed
gases are determined by (1) the specific
gasification process, (2) the gasifier feed
characteristics, (3) the position of shift
relative to sulfur removal, (4) the end use
of the gas which influences the extent of
shift required, and (5)the final washing or
moisture removal step(s) preceding
Rectisol. The report discusses the
impacts of each of these factors on
Rectisol feed gas characteristics.
Although the major focus is on entrained
coal gasification applications, some
comments regarding other applications
are also included. Two examples of
Rectisol designs are examined to
illustrate the influence of feed gas char-
acteristics, product specifications, and
environmental requirements on design.
K. W. Crawford and R. A. Orsini are with TRW. Inc.. Redondo Beach. CA 90278.
William J. Rhodes is the EPA Project Officer (see below).
The complete report, entitled "Environmental Assessment: Source Test and
Evaluation Report—Rectisol Acid Gas Removal," (Order No. PB 84-153 238;
Cost: $16.00. subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 70^-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
irUS GOVERNMENT PRINTING OFFICE 1984-759-015/7323
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
SSS eS5l«2PROTECTION AGENCY
! TDE.Ri
CHICAGO IL 60604
------- |