0% found this document useful (0 votes)
115 views7 pages

The ASTM Copper Strip Corrosion Test: Application To Propane With Carbonyl Sulfide and Hydrogen Sulfide

About method

Uploaded by

AJ Manurung
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
115 views7 pages

The ASTM Copper Strip Corrosion Test: Application To Propane With Carbonyl Sulfide and Hydrogen Sulfide

About method

Uploaded by

AJ Manurung
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

120 Energy & Fuels 2003, 17, 120-126

The ASTM Copper Strip Corrosion Test: Application to


Propane with Carbonyl Sulfide and Hydrogen Sulfide
Wendy C. Andersen, Aziz I. Abdulagatov, and Thomas J. Bruno*
Physical and Chemical Properties Division, National Institute of Standards and Technology,
325 Broadway, Boulder, Colorado 80305-3328

Received July 3, 2002. Revised Manuscript Received September 30, 2002

Carbonyl sulfide (COS), which occurs as an impurity in commercial sources of propane, can
hydrolyze in the presence of water to form hydrogen sulfide (H2S) and carbon dioxide (CO2). All
commercial and HD-5 grade propane is required to pass the copper strip corrosion test, ASTM
Test Method D 1838. While it is known that hydrogen sulfide will cause the failure of the copper
strip test, it is widely believed that the hydrolysis of carbonyl sulfide can also cause the failure
of the corrosion test. We have tested gravimetrically prepared mixtures of COS and H2S in pure
and commercial grade propane with a variant of the ASTM copper strip corrosion test. Minor
changes to the published ASTM corrosion test were implemented for diagnostic or monitoring
purposes in making the measurements. Surprisingly, mixtures containing as much as 1000 ppm
(mass/mass) COS did not cause a failure of the test, even when the copper strip was in contact
with the COS mixture for 2 h (double the normal contact time). Mixtures containing H2S caused
the failure of the test; however, concentrations as high as 3.5 ppm (mass/mass) H2S passed the
test. Moreover, we were not able to produce the colors and patterns shown on the ASTM copper
strip corrosion standards lithograph.

Introduction Test Method for Copper Strip Corrosion by Liquefied


Petroleum (LP) Gases, ASTM D 1838.4,5 In this test, a
Carbonyl sulfide, (COS or SdCdO, CAS: 463-58-1), strip of cleaned, polished copper is suspended in a vessel
is a gas that is present in minor amounts in petroleum that has been rinsed with water, and pressurized with
and coal refinery gases and, to a greater extent, in coal an appropriate quantity of LPG. The filled vessel is then
gasification streams.1,2 This material occurs as an maintained at 38 °C (100 °F) for 1 h, after which the
impurity in commercial sources of propane (referred to strip is removed and immediately “read”. Reading a
as liquefied petroleum gas, LPG, available as com- copper strip is done by comparison with lithographed
mercial or HD-5 grades), and can hydrolyze in the standard strips provided by ASTM. The lithographs are
presence of water to form hydrogen sulfide (H2S) and divided into five classifications (Table 1). First, a
carbon dioxide (CO2). While COS is not itself corrosive, pristine, freshly polished strip is displayed on the ASTM
the hydrolysis product H2S is corrosive, especially in the lithograph. This pristine strip does not have a rating
presence of water. Because propane and carbonyl sulfide or classification beyond “freshly polished”. Next, four
have relatively similar normal boiling temperatures subdivided levels of progressive sulfur-related corrosion
(-42.1 °C for propane, -50.2 °C for COS), approximately are presented: level 1 (with 1a and 1b slight tarnish),
90% of petrochemical COS will be found in the propane level 2 (with 2a-2e moderate tarnish), level 3 (with 3a
fraction of refinery gases. Only 10% will be found in the and 3b dark tarnish), and level 4 (with 4a-4c severe
ethane fraction, while virtually none will be found in corrosion). For commercial LPG or HD-5 quality pro-
the butane fraction. It is not unusual to find up to 200 pane samples, 1a and 1b are considered passing, while
ppm (mass/mass, g COS/106 total g) COS in the propane anything higher is considered failing. LPG batches that
refinery gas fraction. Carbonyl sulfide also occurs in fail the copper strip corrosion test must be reprocessed,
natural gas, and reports of COS in downstream natural or down-graded to a lower value product, thus resulting
gas are increasing.3 Natural gas is usually saturated in a loss for the producer and additional costs for the
with water at the well head, and much of the COS that consumer.
is present is hydrolyzed to H2S before processing. Engineering literature indicates that a COS concen-
The corrosivity of certain sulfur species in fuel gases tration of 58 ppm (mass/mass) will cause failure of the
such as LPG is determined by failures in the Standard copper strip corrosion test, with the presumed mecha-
(4) Standard Test Method for Copper Strip Corrosion by Liquefied
* Author to whom correspondence should be addressed. Fax: (303) Petroleum (LP) Gases, ASTM Standard D 1838-91; American Society
497-5927. E-mail: bruno@boulder.nist.gov. for Testing and Materials: West Conshohocken, PA, 1991 (Reapproved
(1) Ferm, R. J. Chem. Rev. 1957, 57, 621-640. 2001), p 1.
(2) Svoronos, P. D. N.; Bruno, T. J. Ind. Eng. Chem. Res. 2002, 41, (5) Standard Test Method for Detection of Copper Corrosion from
5321-5336. Petroleum Products by the Copper Strip Tarnish Test, ASTM Standard
(3) Panel Discussion. Symposium on Natural Gas and LPG Odor- D 130-94; American Society for Testing and Materials: West Consho-
ization, Gas Technology Institute, July 23-25, 2001, Chicago, IL. hocken, PA, 1994 (Reapproved 2000), p 1.

10.1021/ef020145m CCC: $25.00 © 2003 American Chemical Society


Published on Web 11/19/2002
ASTM Copper Strip Corrosion Test Energy & Fuels, Vol. 17, No. 1, 2003 121

Table 1. ASTM Copper Strip Classificationsa


classification designation description
freshly polished strip ----
1 slight tarnish a. light orange, almost the same as a freshly polished strip
b. dark orange
2 moderate tarnish a. claret red
b. lavender
c. multicolored with lavender blue or silver, or both, overlaid on claret red
d. silvery
e. brassy or gold
3 dark tarnish a. magenta overcast on brassy strip
b. multicolored with red and green showing (peacock), but no gray
4 corrosion a. transparent black, dark gray or brown with peacock green barely showing
b. graphite black or lusterless black
c. glossy or jet black
a Ref 5. Copyright ASTM INTERNATIONAL. Reprinted with permission.

nism being hydrolysis of COS to H2S.6,7 It has also been tested pentane solutions containing elemental sulfur,
observed that a tank car of LPG will pass the copper H2S, COS, and other sulfur compounds as a model for
strip test in the morning, but fail the test in the LPG corrosion, yet their results in the heavier hydro-
afternoon after sitting in the sun for several hours. One carbon did not duplicate those of Pyburn.19
author has reported a passing sample of LPG failing
the copper strip test less than 30 min later.8 Here again, Experimental Section
the presumed mechanism is COS hydrolysis. It should
Sample Preparation. To assess the potential failure
be noted that many basic compounds will also tarnish mechanisms that might play a role in the copper strip corrosion
a copper strip, even in the absence of sulfur.9 test, we prepared a series of mixtures under controlled
Since the development of the corrosion standards,10 conditions. These mixtures contained only the pure compounds
several researchers have conducted studies to correlate of interest, to eliminate possible side reactions or competing
the corrosion classifications with the actual quantity of effects. The only departure from this protocol involved experi-
sulfur and sulfur compound contamination in the pe- ments that were performed with commercial propane.
troleum product. Matthews and Parsons prepared cor- Mixtures of COS or H2S in propane were prepared gravi-
roded copper strips of various classifications by exposure metrically. All samples contained 50 g of propane plus the
to naphtha and kerosene containing elemental sulfur quantity of COS or H2S necessary to make the desired
concentration. The mixture concentrations ranged from 500
and mercaptans, and then analyzed the corroded surface
to 1000 ppm (mass/mass) for the COS mixtures and 3.5 to 500
of the copper using chemical analysis and cathodic ppm (mass/mass) for the H2S mixtures. Mixture preparation
reduction.11 Kashima and Nose related test strips involved the cleaning and evacuation of a 475 mL sample
corroded by sulfurous paraffin with the electron diffrac- cylinder, and mass additions of the components measured to
tion patterns of the copper oxide and copper sulfide films (1 mg. Mixture cylinders were allowed to reach thermal
responsible for the corrosion classifications.12 More equilibrium before the mass was recorded. The uncertainty
recently, the sulfur content of corroded surfaces was in the mass (measured with an electronic load cell) resulted
correlated with both the ASTM classification and the in an uncertainty in the COS mixture concentrations of (2%.
sulfur content of naphtha mixtures containing sulfur, It was not possible to directly prepare the lower concentration
mercaptans, disulfides, and organosulfides.13-16 Only H2S mixtures with the same uncertainty, because the mass of
H2S required was below or near the uncertainty of the balance.
two studies have focused on interpreting the copper strip
The 3.5 to 100 ppm (mass/mass) H2S mixtures were thus
test for lighter hydrocarbons. In the late 1970s, Pyburn prepared by serial dilution from a 2.5 mass percent mixture
et al. exposed copper strips to known concentrations of of H2S and propane (all components were gas phase). The
hydrogen sulfide and elemental sulfur in LPG and resulting uncertainty in concentration for the lowest concen-
correlated ASTM corrosion classifications directly with tration (3.5 ppm (mass/mass)) H2S mixture prepared from this
sulfur concentration.17,18 In 1991, Clark and Lesage stock mixture was (13%. Sample cylinders were constructed
of carbon steel, and all valves, fittings, and tubing were 316
(6) Fails, J. C.; Rumbaugh, G. H. Hydrocarbon Process. Pet. Refin. stainless steel (AISI designation). Ultra high purity propane
1962, 41, 211-217. (99.7%), COS (99%), and H2S (99.5+%) were obtained from
(7) Derdall, G.; Hyne, J. B. Quart. Bull.; Alberta Sulfur Research,
Ltd. 1978, 15 (1), 24-37. commercial vendors, and their purity was confirmed by gas
(8) Mick, M. B. Hydrocarbon Process. 1976, 137-142. chromatography. Commercial grade, odorized propane was
(9) Andersen, W. C.; Bruno, T. J. Ind. Eng. Chem. Res., submitted purchased from a consumer filling station and was used as
for publication. received. To assess the reproducibility of the test method,
(10) Mallatt, R. C.; Demkovich, P. A.; Cropper, W. V. ASTM Bull.
1956, 218, 49-51. duplicate mixtures were prepared and corrosion-tested for the
(11) Matthews, F. W. H.; Parsons, D. F. Proc. Mid-Year Meet., Am. ultra high purity samples of 500 and 1000 ppm (mass/mass)
Pet. Inst. 1950, 30M, 24-37. COS and 3.5 ppm (mass/mass) H2S.
(12) Kashima, M.; Nose, Y. Proc. Int. Congr. Metal. Corros. 1963, Corrosion Testing Apparatus. The apparatus used for
2, 381-390.
(13) Garcia-Anton, J.; Monzo, J.; Guinon, J. L.; Gomez, D.; Costa, the copper strip corrosion test was made from a sight glass
J. Fresenius’ J. Anal. Chem. 1990, 337, 382-388.
(14) Monzo, J.; Garcia-Anton, J.; Guinon, J. L. Fresenius’ J. Anal. (17) Pyburn, C. M.; Cahill, F. P.; Lennox, R. K. Energy Process./
Chem. 1991, 341, 606-610. Can. 1978, 70, 40-44.
(15) Monzo, J.; Garcia-Anton, J.; Guinon, J. L. Fresenius’ J. Anal. (18) Pyburn, C. M.; Cahill, F. P.; Lennox, R. K. Proc. 57th Annu.
Chem. 1992, 343, 593-596. Conv. Gas Proc. Assoc. 1978, 57, 46-51.
(16) Garcia-Anton, J.; Monzo, J.; Guinon, J. L. Corrosion (Houston) (19) Clark, P. D.; Lesage, K. L. Proc. Laurance Reid Gas Cond. Conf.
1995, 51, 558-566. 1991, 41, 181-195.
122 Energy & Fuels, Vol. 17, No. 1, 2003 Andersen et al.

Figure 1. Apparatus used for copper strip corrosion testing.

gauge equipped with a type 316 stainless steel body, and As an independent blind review20 of our copper strip
capable of operation to 8.3 MPa (1200 psig) at 38 °C (100 °F) “readings”, three engineers who work in the LPG industry and
(Figure 1). The sight glass gauge was used instead of the routinely conduct the copper strip test were shown images of
commercial vessel proposed by ASTM guidelines in order to our copper strips after exposure to various COS or H2S propane
allow observation of the liquid level in the filled container. A mixtures. The individuals were asked to interpret the copper
type 316 stainless steel pipe plug at the bottom of the sight strips without knowledge of our test conditions, or knowledge
glass gauge was fitted with a length of type 316 stainless steel of our conclusions. In all cases, their interpretation of the
rod terminated in a hook, upon which the prepared copper strip results exactly matched our own.
was suspended. The apparatus was furnished with a manifold In addition to the experiments that adhered strictly to the
to allow filling, evacuation, and venting. A well-stirred water ASTM procedure, we conducted several experiments in which
bath (sized with a favorable filling factor to accommodate the the procedure was intentionally varied. The first variation was
to keep the test vessel in the water bath for double the ASTM
sight glass gauge) maintained at 38 ( 1 °C (100 °F) provided
mandated time, or 2 h, to determine how longer test residence
temperature control for the test. Between each test, the sight
times would affect the corrosion results. In the second varia-
glass gauge was thoroughly cleaned and the final rinse was
tion, we allowed for a separate water phase to be in contact
done with distilled, deionized water.
with the copper strip during the test, to assess how a wet LPG
Testing Method. Copper strip corrosion testing of the sample would influence the test results. For these experiments,
standard mixtures adhered closely to ASTM guidelines. The the vessel containing the pristine copper strip was evacuated,
tests began by polishing and cleaning a copper strip (70 mm and either 5 or 20 mL of distilled water was added to the
× 12.5 mm × 3 mm) according to ASTM guidelines, by use of vessel, then the propane sample was added to the vessel to
silicon carbide grit paper of varying degrees of fineness and form a condensed propane layer on top of the water layer (in
acetone as the wash solvent. Electrolytic tough pitch copper vapor-liquid-liquid equilibrium). The vessel was immediately
of high purity was used for all tests. A milliliter of distilled, placed into the water bath. No attempt was made to mix the
deionized water was added to the sight glass gauge, swirled phases.
to coat the walls, and then drained from the vessel. The freshly
prepared copper strip was then suspended from the hook on Results
the vessel plug, and inserted into the bottom of the vessel. The
vessel was evacuated to 0.13 Pa (10-3 mmHg), and the sample The copper strip corrosion test was performed on
was introduced through the manifold. The sample was verified mixtures of COS in both ultra high purity propane and
to completely submerge the copper strip by observation commercial grade propane. Commercial grade propane
through the sight glass. No external light source was used to was examined to assess how the lower purity product
illuminate the sight glass during a test; the test was conducted would influence corrosion. Finally, mixtures of H2S and
with only minimal exposure to ambient light. The loaded gauge propane were tested to quantify how much H2S was
was then submerged in the 38 °C bath for 1 h. At the end of necessary to produce a corrosion test failure.
the time period, the vessel contents were vented, and the Copper Strip Corrosion Tests on COS + Pro-
copper strip was removed and interpreted by comparison pane. To determine the amount of corrosion caused by
against the ASTM copper strip corrosion standards. Digital COS in propane, test mixtures were prepared with 500,
photographs were recorded of the pristine copper strip, the 750, and 1000 ppm (mass/mass) COS and ultra high
“before” image, and the exposed strip, the “after” image.
Moreover, the spent copper strips were coated with a layer of (20) Brunner, R. G. Gas Processors Association. Personal com-
polyurethane and stored in a dark location as an archive. munication, June 6, 2001.
ASTM Copper Strip Corrosion Test Energy & Fuels, Vol. 17, No. 1, 2003 123

Figure 2. Post-corrosion test images of copper strips (side 1 and 2) are shown after exposure to mixtures with COS concentrations
(ppm, mass/mass) from top: 0, 500, 750, 1000, 500 (2 h test), 1000 (2 h test), 1000 (commercial grade propane).

Table 2. Corrosion Classification and Test Results for discernible from the pristine strip, and were given an
COS in Ultra High Purity Propane (UHP) and Odorized ASTM 1a classification. Duplicate mixtures of 500 and
Commercial Grade Propane (LPG)
1000 ppm (mass/mass) COS were prepared and corro-
ASTM sion tested. There was no difference in test results for
sample (ppm (mass/mass)) rating result
duplicate measurements. Additional mixtures of 500
UHP, no COS added 0 pass and 1000 ppm (mass/mass) COS were tested for 2 h
500 ppm COS in UHP 0 pass rather than the ASTM-specified 1 h. In both cases, the
750 ppm COS in UHP 1a pass resulting copper strips were somewhat more tarnished
1000 ppm COS in UHP 1a pass
than the strips from the 1 h test, with an ASTM
500 ppm COS in UHP, 2 h test 1a pass
1000 ppm COS in UHP, 2 h test 1b pass
classification of 1a for the 500 ppm (mass/mass) mixture
and 1b for the 1000 ppm (mass/mass) mixture. In no
500 ppm COS in UHP, 20 mL H2O 1b pass
1000 ppm COS in UHP, 20 mL H2O 1b pass
case was a “failed” result obtained for a COS + propane
mixture.
LPG, no COS added 1a pass
600 ppm COS in LPG 1a pass Tests were also performed on 500 and 1000 ppm
1000 ppm COS in LPG 1a pass (mass/mass) mixtures where 20 mL of water was added
600 ppm COS in LPG, 5 mL H2O 1a-1b pass to the test vessel. For these experiments, there was
600 ppm COS in LPG, 20 mL H2O 1b pass more corrosion present on the copper strip; however,
purity propane. Samples of the ultra high purity pro- neither of the concentrations produced more than an
pane (with no added COS) were also subjected to the ASTM 1b rating (Figure 3). Our study of the COS
corrosion test. Figure 2 represents the before and after hydrolysis reaction rate supports this observation, in
images of the COS + propane mixtures, and Table 2 which we found that the rate of COS hydrolysis is
summarizes our interpretation of corrosion levels. The proportional to the quantity of water present.9 It is
pure propane produced a copper strip indistinguishable therefore not surprising that more corrosion is observed
from the freshly polished strip. The copper strip exposed under test conditions in which more water is present.
to the 500 ppm (mass/mass) mixture was also recovered In these experiments, the water covered the lower 2/3
unchanged from the pristine strip, and therefore was of the copper strip. It is interesting that the majority of
rated “0” (actually not ratable on the ASTM scale). For the corrosion occurred above the water phase, in the
the 750 and 1000 ppm (mass/mass) mixtures, the copper propane phase. This observation was unexpected,19 since
strips were recovered from the test in a condition barely any appreciable H2S that formed from the hydrolysis
124 Energy & Fuels, Vol. 17, No. 1, 2003 Andersen et al.

Figure 3. Post-corrosion test images of copper strips (sides 1 and 2) are shown after exposure to mixtures with COS concentrations
of 500 (top) and 1000 (bottom) ppm (mass/mass). Lower (right) 2/3 of the strip was submerged in water.
of COS during the test would be expected to have had graphs. The 13, 25, and 50 ppm (mass/mass) H2S
greater affinity for the aqueous phase. mixtures appeared to produce similar results, despite
For the studies with commercial (odorized) propane, the progressive increase in H2S concentration. More
the corrosion test was initially performed on the propane remarkable was the fact that 100 and 500 ppm (mass/
“out of the bottle” with no added COS. Unlike the ultra mass) mixtures did not completely corrode the copper.
high purity propane, commercial propane produced We could not even cause an ASTM 4c level classification
some slight tarnish, giving a rating of 1a on the ASTM by filling the moistened vessel with pure H2S at 1.55
scale. Mixtures containing 600 and 1000 ppm (mass/ MPa (225 psig)! While these results clearly demonstrate
mass) COS in commercial grade propane were tested a limitation of the copper strip corrosion test, the single
and both mixtures produced strips with class 1a tarnish, most disturbing result is that our 3.5 ppm (mass/mass)
which were indistinguishable from the commercial mixture of H2S in propane passed the copper strip test
propane-only sample (1000 ppm (mass/mass) result is with a 1a classification. LPG with this level of H2S
shown in Figure 2). Two additional 600 ppm (mass/ would not generally be considered merchantable.
mass) mixtures were tested with varying quantities of
water. In the first experiment, 5 mL of water was added Discussion
to the vessel to generate a water level just below the
base of the strip. The resulting strip was slightly darker The results of all of our copper strip corrosion testing
than in the trace water test, and was classified between are clearly contrary to previously reported results, and
1a and1b. In the second experiment, 2/3 of the strip was contrary to conventional wisdom, which holds that much
submerged in 20 mL of water. The submerged portion lower concentrations of COS should cause failure of the
of the strip was significantly darker orange than the copper strip corrosion test.
unsubmerged portion, indicative of class 1b tarnish. The copper strip corrosion tests that were performed
There seemed to be a silvery tone to the unsubmerged in this research indicate that the hydrolysis of COS is
portion of the copper strip, which may have been too slow to account for test failure with propane. This
indicative of a class 2d tarnish; however, the subjective result, at first glance, is somewhat surprising given the
nature of this test makes absolute interpretation impos- magnitude of the industrial difficulty that this reaction
sible. Despite the somewhat more dramatic results is thought to cause. Recall that plant operators and
obtained when preparing mixtures in commercial pro- processors expect corrosivity problems when the COS
pane, in no case was an unambiguous failure classifica- concentration is approximately 50 ppm (mass/mass).6
tion obtained. In our studies, mixtures of COS in propane were never
Copper Strip Corrosion Tests on H2S + Propane. observed to fail the copper strip corrosion test, even at
On the basis of the reports in the literature, one would 1000 ppm (mass/mass), a COS level which, in LPG,
have expected moderate concentrations (50 to 100 ppm, would be considered massive. In fact, our copper strips
mass/mass) of COS in propane to cause copper strip never extended beyond a 1a classification unless we
corrosion test failure. When we could not produce a doubled the test time or added copious quantities of
copper strip test failure under any conditions with COS water to the test vessel. Even in these cases, a “pass”
mixtures, we tested standard mixtures of H2S in ultra was recorded.
high purity propane to confirm the validity of the The inability of COS to cause LPG corrosion failure
corrosion test. By measuring H2S propane mixtures over is supported by our own measurements of the kinetics
a wide concentration range, we attempted to reproduce of the COS hydrolysis reaction in propane.9 In 1 h, the
the various levels of tarnish shown on the ASTM copper hydrolysis reaction is simply too slow at 38 °C to produce
strip corrosion standards to quantify how much H2S is the quantity of H2S necessary to fail the copper strip
necessary to cause the various classifications of corro- test. Regardless, it has been suggested that the very low
sion. Before and after images are shown in Figure 4. levels of H2S required for corrosion test failure, 0.35 to
Duplicate 3.5 ppm (mass/mass) H2S mixtures were 1 ppm (mass/mass), are produced even by slow hydroly-
prepared and tested. Test results were identical for the sis.6,8,17 By contrast, our results indicate that concentra-
two mixtures and the result for one of the 3.5 ppm tions of H2S up to an order of magnitude higher will
(mass/mass) mixtures is shown in Figure 4. While the pass the corrosion test with only a 1a rating. It is indeed
severity of corrosion does indeed increase with increas- noteworthy that we could not produce the standard
ing H2S concentration, it was impossible to interpret the ASTM corrosion classifications with high purity mix-
majority of our results within the accepted classifica- tures of H2S and propane.
tions of the corrosion standards. In fact, many of the It is apparent from this work that the presence of COS
results we obtained did not resemble any of the litho- alone, even at relatively high concentrations, and even
ASTM Copper Strip Corrosion Test Energy & Fuels, Vol. 17, No. 1, 2003 125

Figure 4. Post-corrosion test images of copper strips (sides 1 and 2) are shown after exposure to mixtures with H2S concentrations
(ppm, mass/mass) from top: 0, 3.5, 6, 8, 13, 25, 50, 100, 500, pure H2S.

in wet propane, is not enough to cause the repeated in LPG transport and storage vessels, may account for
failure of the copper strip corrosion test as practiced in corrosion test failure as these materials can generate
the LPG industry. While we by no means advocate the elemental sulfur in the presence of H2S.17,18
sale of high COS level LPG, it is clear that COS is not
the sole cause of failure. It is probable that the copper
strip corrosion test failures that are experienced in the Conclusions
field are caused either by a combination of sulfur species
being present, or by an as-yet-unidentified agent. There We have examined the ASTM D 1838 copper strip
are measurements and anecdotal evidence to support corrosion test for LP gases and have found that contrary
the former explanation.17 Situations are reported in to popular opinion, the hydrolysis of COS is not fast
industry in which two separate batches of LPG pass the enough to cause failure of the copper strip test. More-
copper strip corrosion test individually, but when the over, COS levels as high as 1000 ppm (mass/mass) in
two batches are combined, the mixture fails.18 This can propane will not tarnish a copper strip beyond a 1a
happen if one batch has a trace of H2S, and the other ASTM classification. H2S levels as high as 3.5 ppm
has a trace of elemental sulfur. The components can (mass/mass) in propane will not cause copper strip test
then act synergistically when mixed to produce the failure. Finally, using standard mixtures of H2S in
unacceptable corrosion test result. Catalytic surfaces propane ranging in concentration from 6 to 500 ppm
such as iron oxides and iron sulfides, which can be found (mass/mass), we were not able to produce the various
126 Energy & Fuels, Vol. 17, No. 1, 2003 Andersen et al.

classifications of the ASTM copper strip corrosion Acknowledgment. This work was performed while
standards lithograph. We believe that additional work W.C.A. held a National Academy of Sciences/National
is necessary to better define the conditions under which Research Council Research Associateship Award at
sulfur compound synergy can affect the copper strip NIST.
corrosion test. We further believe that additional work
should be done to improve the concept and practice of
the test itself. EF020145M

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy