&EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 2771 1
EPA-600 7-79-224
September 1979
Adipic Acid Degradation
Mechanism in Aqueous
FGD Systems
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
tfie views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-224
September 1979
Adipic Acid Degradation Mechanism in
Aqueous FGD Systems
by
F.B. Meserole, D.L Lewis, A.W. Nichols, and G. Rochelle
Radian Corporation
8500 Shoal Creek Blvd.
Austin, Texas 78766
Contract No. 68-02-2608
Task No. 58
Program Element No. EHE624
EPA Project Officer: Robert H. Borgwardt
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
CONTENTS
SECTION Page
1 Introduction 1
2 Summary of Results 3
3 Conclusions and Recommendations 6
4 Literature Review 10
4.1 Summary 10
4.2 Conjugated Oxidation of Carboxylic Acids 10
4.3 Other Reactions of Carboxylic Acids 11
4.4 Oxidative Decarboxylation by Metal Ions 18
4.5 Health Effects 19
5 Analytical Methods 23
5.1 Ion Chromatography 23
5.2 Gas Chromatography 23
5.3 Total Organic Carbon 25
5.4 Solid Phase Characterization 26
5.5 Carbon Dioxide 27
5. 6 Gas Chromatography-Mass Spectrometry 28
6 Results of Field Sampling Efforts 29
6.1 Gas Phase Sampling -. 29
6.2 Liquid Sampling 30
6.3 Results of IERL-RTP Material Balance Test (10/25/78). 32
6.4 Results of IERL-RTP Material Balance Test (11/15-
16/78) 36
6.5 Results of IERL-RTP Material Balance Test (12/18-
22/78) 37
6.6 Results of Shawnee Material Balance Test 45
6. 7 Solids Characterization 46
6.8 Field Data Correlation 54
7 Laboratory Tests 58
7.1 Screening Tests 58
7.2 Mechanisms Tests 64
Bibliography 74
Appendix A - Analytical Methods 77
Appendix B - Sample Chromatograms 83
111
-------
LIST OF FIGURES
Page
FIGURE
4-1
6-1
6-2
6-3
6-4
6-5
6-6
6-7
6-8
6-9
6-10
6-11
6-12
7-1
7-2
7-3
7-4
B-l
B-2
B-3
IERL-RTP scrubber configuration (natural oxidation) . . .
IERL-RTP scrubber configuration (forced oxidation)
IERL-RTP scrubber configuration (forced oxidation)
Calcium sulfite hemihydrate crystals taken from EPA's
Shawnee Test Facility no additives (x500)
Calcium sulfite hemihydrate crystals from natural
oxidation tests at RTP, 2500 ppm adipic acid (x590) ...
Calcium sulfate dihydrate crystals from forced
oxidation tests at RTP, 600 ppm adipic acid (x200) .
Calcium sulfate dihydrate crystals from forced
oxidation tests at RTP, 4000 ppm adipic acid (x200) . . .
Calcium sulfate hemihydrate crystals from forced
oxidation tests at RTP, 600 ppm adipic acid (x460) ....
Calcium sulfate hemihydrate crystals from forced
oxidation tests at RTP, 4000 ppm adipic acid (x460)
Adipate decomposition as a function of sulfite
Laboratory apparatus for screening tests
Laboratory apparatus for mechanism test with Porapak Q
Laboratory apparatus for mechanism test with
Laboratory apparatus for mechanism test with silica
gel and carbon traps
Sample 1C chromatogram with high sulfite
Sample 1C chromatogram with no sulfite
Sample GC chromatogram
13
31
34
39
43
48
51
51
52
52
53
53
56
59
66
69
72
84
85
86
IV
-------
LIST OF TABLES
TABLE page
2-1 Summary of Field Sampling Results 5
2-2 Material Balance Results of Dynamic Laboratory Degradation
Studies 5
3-1 Comparison of Potential Hydrocarbon Emissions from Degrada-
tion of Adipic Acid with Reported Values for Coal-Fired .
Industrial and Utility Boilers 8
4-1 Decarboxylation in the Air Oxidation of Cyclohexane in Acetic
Acid Catalyzed by Co'2 15
4-2 Rate Constants for Decarboxylation in Cumene Undergoing
Oxidation for the Reaction, RiCOOH + R-0-0-* Complex 16
4-3 Rate Constants for the Reaction of Cumyl Peroxide Radicals
with. Carboxylic Acids in Chlorobenzene 17
4-4 Toxicology of Adipic Acid and Related Compounds 20
6-1 Samples Obtained at IERL-RTP 32
6-2 Results of Analysis of IERL-RTP (10/25/78) Samples 33
6-3 Material Balance Calculations: IERL-RTP 10/25/78 35
6-4 Samples Obtained at IERL-RTP (11/15-16/78) 36
6-5 Results of Analysis of IERL-RTP (11/15/78-11/16/78) Samples.. 38
6-6 Material Balance Calculations: IERL-RTP 11/15/78-11/16/78... 40
6-7 Samples Obtained at IERL-RTP 41
6-8 Results of Analysis of IERL-RTP (12/18/78-12/22/78) Samples.. 42
6-9 Material Balance Calculations: IERL-RTP 12/21/78 44
6-10 Samples Obtained at Shawnee 45
6-11 Results of Analysis of Shawnee (2/2/79) Samples 47
6-12 Material Balance Calculations - TVA's Shawnee Facility 49
6-13 Compositional Analysis of Scrubber Solids 50
-------
LIST OF TABLES (continued)
TABLE Page
6-14 Adipate Decomposition Data from RTF and Shawnee 55
7-1 Results of Phase One Lab-Screening Tests 60
7-2 Summary of Phase Two Lab-Screening Tests 62
7-3 Results of Sealed Flask Experiments 63
7-4 Results of Sealed Flask Experiment with Material Balance 64
7-5 Results of Mechanism Test with Porapak Q Sorbent Trap 67
7-6 Carbon Balance for Mechanism Tests Using Porapak Q Trap 65
7-7 Carbon Balance for Mechanism Test Using Porapak Q Trap;
Hydrocarbon Values Neglected 68
7-8 Results of Mechanism Test with Recirculated Gases 70
7-9 Carbon Balance for Mechanism Test Using Recirculated Gases 69
7-10 Results of Mechanism Test with Silica Gel and Activated
Carbon Traps 73
7-11 Carbon Balance for Mechanism Test Using Silica Gel and
Activated Carbon Traps 71
A-l Response Characteristics of Species Analyzed by Dionex 78
A-2 1C Eluents 79
VI
-------
SECTION 1
INTRODUCTION
The addition of weak organic acids such as adipic acid to flue gas
desulfurization (FGD) wet scrubbers using limestone has been shown to bene-
fit both S02 removal and limestone utilization. The value of adipic acid
as a scrubber additive has been suggested in a theoretical analysis by
Rochelle (1). Adipic acid has the effect of buffering scrubber solutions,
thereby enhancing liquid phase mass transfer. Improved SOz removal can be
achieved economically due to the relatively low price of adipic acid. How-
ever, long-term scrubber tests at Industrial Environmental Research Labora-
tory-Research Triangle Park (IERL-RTP) and EPA's test facility at the TVA
Shawnee Power Plant have experienced substantial losses of adipic acid.
These losses were in excess of that expected from the scrubber solution
removed with the filter cake solids.
Tests have also shown that the adipic acid loss rate is considerably
higher during forced oxidation runs than during natural oxidation conditions.
Unaccounted losses of as much as 80% of the adipic acid make-up rate were
observed in the Shawnee venturi/spray tower scrubber system.
Several possibilities could explain this disappearance including:
inaccurate measurement of adipic acid in the scrubber
solution, i
precipitation of insoluble adipate salts,
chemical complexation and/or decomposition of adipic acid,
biodegradation of adipic acid,
physical adsorption on, or coprecipitation with the
scrubber solids, and
volatilization of adipic acid into the flue gas.
Radian was contracted by the Environmental Protection Agency (EPA) to
select methods to accurately measure levels of adipic acid in FGD scrubber
solutions and conduct material balance measurements at the IERL-RTP pilot
plant and at the Shawnee Test Facility. The primary objective was to verify
adipic acid loss and to determine its cause. Other objectives were to identi-
fy by-products if degradation occurs, determine mechanism, and suggest possible
courses for limiting the loss.
-------
During the period between October, 1978 and February, 1979 Radian person-
nel conducted three sampling trips to RTF and one sampling trip to the
Shawnee Test Facility. At both facilities samples were collected at the
various slurry and gas streams around the system. Analytical methods were
developed for the accurate measurement of adipic acid and valeric acid in
the samples and the data obtained were used to perform material balance
calculations.
A literature survey of known decomposition mechanisms of adipic acid,
and carboxylic acids in general, was conducted in order to ascertain any
properties or reactions of adipic acid relevant to its stability under
scrubber conditions.
Finally, a series of laboratory bench scale studies aimed at simulating
scrubber conditions was conducted. These tests were designed in such a way
that operating conditions were variable and degradation could be monitored
by the collection and analysis of decomposition products. The results of
these tasks are presented in this report.
REFERENCES
1. Rochelle, G. "The Effect of Additives on Mass Transfer in CaC03 and CaO
Slurry Scrubbing of SC>2 from Waste Gases, " Ind. Eng. Chem. Fundam. , 16:
67-75, 1977.
-------
SECTION 2
SUMMARY OF RESULTS
The samples collected during this program from the RTF and Shawnee
S02 scrubbers during the adipic acid addition tests were analyzed to deter-
mine the concentrations of adipic acid and any degradation products. Gas-
eous, liquid and solid samples were taken using a variety of sampling
techniques and were then characterized using several analytical procedures
including;- ion chromatography, gas chromatography, gas chromatography-
mass spectrometry, infrared spectroscopy and total organic carbon deter-
mination.
The major degradation products of adipic acid identified in the samples
collected from the RTP and Shawnee test units were:
valeric acid - CH3(CH2)3COOH and
glutaric acid - HOOC(CH2)3COOH.
Trace quantities of the following were also found:
butyric acid - CH3(CH2)2COOH and
succinic acid - HOOC(CH2)2COOH.
All four of the above acids were found in the liquid phase samples, but
only valeric acid was found in the scrubbed flue gases.! No measurable
amounts of organics could be detected in the washed solids using infrared
analysis. The detection limit using this technique is about 0.1% on a weight
basis.
The products measured in the samples collected during the laboratory
degradation studies were the following:
valeric acid,
carbon dioxide,
methane, and
butane.
In some cases, these degradation products were found to account for
most of the adipic acid lost whether in the field or laboratory tests. How-
-------
ever, in general, only 25-50% of the adipic acid added could be accounted
for. The results of the field sampling efforts associated with this program
are summarized in Table 2-1. The material balance calculations were per-
formed using analytical data measured by Radian personnel and flow rates that
were supplied by on-site personnel.
In addition to the field measurements, laboratory programs were carried
out to establish, the conditions necessary-for adipic acid degradation and to
identify the degradation products. Conclusions drawn from these tests
include:
Adipic acid degradation can occur by a chemical process
as opposed to a microbial process.
The similarity in the degradation rates between the labora-
tory and field tests indicate the catalytic and micro-
bial effects are small, if present.
The oxidation of sulfite in the presence of adipic acid
results in degradation, supposedly through a free-
radical mechanism.
During the laboratory phase, a series of material balance tests was
made using various techniques to trap low molecular weight hydrocarbon pro-
ducts. The results of these tests are presented in Table 2-2.
The degradation rate of adipic acid was found to depend upon the de-
gree of sulfite oxidation and adipic acid concentration in solution. Field
and laboratory data were compared using this correlation. Although there is
significant scatter a general trend appears to prevail. Much of this varia-
bility may be a result of the different scrubber systems used to collect
the data and different analytical techniques used to measure adipic acid.
Settling and dewatering measurements conducted at RTF on solids pro-
duced with and without the addition of adipic acid showed only slight dif-
ferences due to adipic acid. The settling rates were reduced by approxi-
mately 25% and some increase in the moisture content of the filtered sludge
was measured for the solids precipitated in the presence of adipic acid.
Electron micrographs of a limited number of solid samples suggest that
the average particle size decreases as the adipic acid increases. However,
the effect was not large enough to yield a detectable difference at Shawnee
when filter-cake solids were compared with and without the use of adipic
acid in the scrubber.
-------
TABLE 2-1. SUMMARY OF FIELD SAMPLING RESULTS
Facility
RTF
RTF
RTF
Shawnee
Hydrocarbon
Collection
Techniques
Poropak Q Trap
Reclrculatlon of
Product
Casea
Silica Gel
Trap
Date
10/25/78
11/15/78
12/18/78
2/2/79
Initial
Adlplc Acid
(mmolea C)
18.7
19.0
9.5
19.2
Scrubber
Type
TCA
TCA
TCA
Venturi/
Spray: Tower
TABLE 2-2. MATERIAL BALANCE
Final
Adlpic Acid Valeric Acid
(nmolea C) (mmole C)
11.4 4.4
11.5 2.4
4.1
12.6 1.5
Oxidation
Mode
Natural
Forced
Forced
Forced
Adipic Acid
Decomposition (%)
28%
69%
54%
91%
Fraction of Products
Measured (%)
84-97%
42-45%
53%
17%
RESULTS OF DYNAMIC LABORATORY DEGRADATION STUDIES
Methane Ethane
(mmole C) (mraole C)
1.6 .4
Propane Butane
(mmole C) (mmole C)
.2 .9
.02
C02 Total
(mraole C) (mraole C)
1.4 20.3
3.2 17.1
1.1 5.2
1.6 15.7
-------
SECTION 3
CONCLUSIONS AND RECOMMENDATIONS
The findings of this study support the field test results at RTF and
Shawnee that adipic acid added to an SQz wet scrubber is consumed in the pro-
cess. Furthermore, the degradation rate depends, in part, on the degree of
sulfite oxidation occurring in the scrubber system.
The degradation products identified in the laboratory and field tests
include the following:
valeric acid,
glutaric acid,
butyric acid,
succinic acid,
butane,
methane, and
carbon dioxide.
No cyclopentanone, a high temperature reaction product of adipic acid,
was found in any of the samples analyzed.
Adipic acid and the acidic decomposition products listed above were
found in the liquid 'portion of slurry samples but not in the solids. No
adipic acpLd was detected in the flue gas; however, measurable quantities
of valeric acid were collected in the flue gas samples from both the RTF
and Shawnee facilities. Laboratory tests generated COa, methane, butane
and valeric acid. Due to the naturally high concentration of C02 in flue
gases this could not be verified in the field. Attempts to measure the low
molecular weight hydrocarbons such as methane and butane in the field were
unsuccessful.
Additional conclusions derived from this program are summarized below:
the decomposition of adipic acid is primarily a chemical
not a microbial process,
the oxidation of sulfite in the presence of adipic acid was
a necessary and sufficient condition for degradation to
occur in a laboratory situation,
-------
the rate of adipic acid decomposition, -3A/3t, appears to be
proportional to the rate of sulfite oxidation, -9SOs /3t, and
the adipic acid concentration in solution which can be
stated mathematically as:
3A/3t = k (3S03=/3t) (A) (3-1)
according to Equation 3-1, reduction of the decompo-
sition rate can be achieved by lowering the sulfite
oxidation rate or the adipic acid concentration in solution,
the literature review indicates that the decomposition
products may vary depending on the relative concentrations
of sulfite and oxygen during the reaction and that the de-
carboxylation may occur through a free radical mechan-
ism,
the degradation products identified in both field and
laboratory test samples don't appear to pose any significant
health related problems, however, valeric and butyric acids
both can cause objectionable odors around the process, and
the failure to close the material balances during forced
oxidation runs at both RTP and Shawnee may result f roa
the inability to measure some of gas phase species.
The presence of adipic acid does appear to cause some reduction in
particle size of gypsum with the effect becoming greater at higher concentra-
tions. In addition to direct evidence obtained from electron micrographs of
scrubber solids, settling and dewatering tests also indicated some decrease
in particle size. The effect on sludge properties does not, however, appear
to be significant in practice.
Evidence of the growth of sulfur-reducing bacteria was observed in
several of the slurry samples collected during the natural oxidation tests
at RTP. The samples were sealed in polyethylene bottles after collection
and stored at room temperature. Black deposits characteristic of metal
sulfides formed in the samples and the odor of HaS was strong when the
bottles were opened. However, no similar observations have been reported
at the sites used for the disposal of the.RTP scrubber solid wastes.
As mentioned previously, not all of the degradation products of adipic
acid were accounted for during the field sampling efforts. Probable products
are methane and butane which naturally would be stripped from the scrubber
into the scrubber flue gases and oxidizer vents. Since the oxidizer vents
can be combined with the exit flue gases, potential hydrocarbon emission
levels, assuming that all of the missing adipic acid was converted either
to methane or butane, were calculated for the exhausted flue gases. The
emission values calculated for both methane and butane assuming that all the
degradation was either one or the other are 12,000 and 14,000 yg/Nm3. The
concentrations calculated are compared to hydrocarbon values reported for
-------
coal-fired industrial (1) and utility boilers (2) in Table 3-1. Laboratory
tests, however, indicate to as much as 50% of the degradation may be COz
which would not contribute significantly to the COz concentration in the
flue gas.
TABLE 3-1. COMPARISON OF POTENTIAL HYDROCARBON EMISSIONS FROM DEGRADATION
OF ADIPIC ACID WITH REPORTED VALUES FOR COAL-FIRED INDUSTRIAL
AND UTILITY BOILERS
Source
Industrial
Utility - Lignite
- Bituminous
Adipic Acid Degradation2
Methane
(Ug/Nm3)
15.0001
3,954
1,758
12,000
Butane Total C2 - C6
(y.g/Nm3) (mg/Nm3)
2,070
861
14,000
1Total hydrocarbon as methane.
2Based on Shawnee forced oxidation tests and assuming that all the unaccounted
adipic acid was converted to methane or butane.
Since material balance attempts failed to account for all of the adipic
acid degradation products, additional studies are recommended to further
characterize the degradation mechanism(s). Emphasis should be placed
on the quantitative collection and analysis of low molecular weight hydro-
carbons. Concurrent with tests to further elucidate the degradation products,
experiments should be performed to 'study the effects of such operating
parameters as:
i
temperature,
pH,
adipic acid concentration,
sulfite concentration,
dissolved heavy metals, and
free-radical scavengers.
Another area where additional work may be necessary is the development
of analytical techniques that can be used in the field for the measurement
of adipic acid and important degradation products.
-------
REFERENCES
1. Leavitt, C., et al. Environmental Assessment of Coal- and Oil-Firing
in a Controlled Industrial Boiler; Volume III. Comprehensive Assess-
ment and Appendices. EPA-600/7-78-164c. Environmental Protection Agency,
Research Triangle Park, North Carolina, 1978.
2. Personal communication from Mr. J. Warren Hammersma, TRW, Inc. to
Mr. Ronald A. Venezia, EPA, 20 October 1978.
-------
SECTION 4
LITERATURE REVIEW
4 . 1 SUMMARY
The most probable cause of adipic acid decomposition in limestone scrub-
bing systems is oxidation or decarboxylation by free radicals generated
during oxidation of SOa species. Such conjugated oxidation of carboxylic
acids has been observed, at temperatures as low as 80°C, in the air oxidation
of cumene, cyclohexane, cyclohexanol and hexane. With this mechanism
expected products include glutaric acid, valeric acid, and butane as well
as other shorter acids and hydrocarbons. The rate of adipic acid degrada-
tion should be proportional to the rate of oxidation and to the adipic
acid concentration and inversely proportional to the total concentration
of dissolved S02 .
Adipic acid degradation may be catalyzed by transition metals such
as cobalt (Co) and manganese (Mn) . The presence of chloride may result in
chlorinated degradation products.
4.2 CONJUGATED OXIDATION OF CARBOXYLIC ACIDS
At 25-100°C aliphatic, carboxylic acids (such as adipic acid) react
too slowly with free radicals to sustain continuous reaction with C^ in
aqueous or organic solvents. However, carboxylic acids will react with
free radicals generated by the sustained oxidation o'f other species such
as cyclohexane, paraffins, cumene, and cyclohexanol. An excellent re-
view of these phenomenomena is presented by Denisov et al. (2) .
The oxidation of sulfite or bisulfite undoubtedly proceeds by a free
radical reaction. One example of such a mechanism is given by Chen and
Barren (3) for cobalt catalyzed oxidation of sulfite:
4. kl +2
S03~ + Co 3 -> Co + 'S03~ (4-la)
S03 + 02 -» 'SO 5 (
= k3
SO 5 + S03 -> SO 5 + *S03 (4-lc)
= ktf
SO 5 + S03 -" 230^ (4-ld)
- k5-
SOs + 'SO 5 -> inert products (4-le)
10
-------
The rate of sulfite oxidation is determined by Reaction 4-lc and is
proportional to sulfite concentration and to the concentration of the free
radical, -SO 5". Hence, with a reasonable rate of sulfite oxidation, there
will be present a steady-state concentration of free radicals which usually
react with sulfite but may also react with carboxylic acids or other or-
ganics which may be present.
4.3 OTHER REACTIONS OF CARBOXYLIC ACIDS
In 1M Ha SOi* at 50-90 C, adipic acid will decompose to give CO and
hydroxylvaleric acids. Decarboxylation was catalyzed by H2 SOi*. At 50°C
the first order rate constant in 1M H2SOit was 3.0 x 10 2 min"1 (4).
The Hunsdiecker reaction is used to synthesize organic bromides from
carboxylic acid salts of silver or mercury (5). The proposed mechanism
involves an oxidative decarboxylation:
Br20 + RC02H -> RC02Br + HOBr (4-2a)
RC02Br -» RC02- + Br« (4-2b)
RC02- ->- R- + C02 (4-2c)
R* + RC02Br -» RBr + RC02- (4-2d)
Kolbe electrolysis of acetate solution produces ethane via a decarboxy-
lation:
RC02~ -> RC02- + e~ (4-3a)
RC02- -» R- + C02 (4-3b)
2R- * R2 ' (4-3c)
Similar reactions can occur by photochemical processes on semiconductor
materials (6). Irradiation of FeCls solutions at 38 C containing succinic,
glutaric, or adipic acid gave B-chloropropionic acid, y-butyrolactone and
5-chlorovaleric acid respectively (7).
Microbial decomposition of aromatic and aliphatic acids has been noted
(8, 9, 10). Aliphatic acids are readily digested by sewage microorganisms
(11).
At elevated temperatures (225 C) adipic acid will decompose to give
cyclopentanone (12, 13). This reaction is catalyzed by Ba(OH)2.
11
-------
The reaction rate constants at 15°C for acid plus Co 3 in_l.l_M
were 0.69 x 10~2 x 10 M"1 sec l for propionic acid and 1.2 x 10~2 M 1 sec"1
for adipic acid._ The estimated rate constant for propionic acid at 55 C
was 2.04 M"1 sec 1.
Anderson and Kochi (14) investigated the reaction of Mn with carboxy-
lic acid in the carboxylic acid solvent at 90-125°C. The reaction with
n-butyric acid proceeded slowly at 125°C and produced C02 and propane with
minor amounts of propylene and propyl esters of butyric acid.
Oxidative decarboxylation of acids by lead tetraacetate was reviewed
by Sheldon and Kocki (15). Secondary and tertiary carboxylic acids are
routinely decarboxylated by Pb "* in acetic acid solvent at 25-100 C.
Reactions with primary acids are much slower. Reaction products include
hydrocarbons and lower acids. A potential product of reaction with adipic
or glutaric acids is the corresponding Cs or Cit lactone.
Conjugated oxidation in the presence of a transition metal may result
in a synergistic effect on the rate of deearboxylation. Hence, Tinker (16)
found that oxidative decarboxylation of adipic acid conjugated with cyclo-
hexane oxidation was four times faster with 0.07 M Co 2 than without any
catalyst.
Free radicals may attack carboxylic acids at the 0-H bond of the car-
boxylic group or at C-H bonds of methylene groups. Attack at the 0-H bond
results in oxidative decarboxylation by a mechanism of the form (17):
R-0-0- + Ri - COOH -* R-O-OH + RiCOO* (4-4a)
RiCOO- -» Ri- + C02 (4-4b)
The residue of the acid, Ri, can react with Oa to give a peroxide radical
which reacts further to end up as an acid group:
i
Ri- + 02-*- Ri-0-0- (4-5a)
Ri-0-0- + RH * Ri-O-OH + R« (4-5b)
Ri-O-OH -* Ri- 0 + H20 (4-5c)
Ra=0 + 02 + RiOOH (4-5d)
The acid residue, RI, may also react with other species to extract H and
end up as a hydrocarbon RiH. Hence, the oxidative decarboxylation of adipic
acid can give valeric acid (CHa-(CH2)3-COOH) or glutaric acid (HOOC-(CH2)3-
COOH) as well as producing C02. The products are also subject to oxidative
decarboxylation and can give the full series of lower acids and hydrocarbons
as shown in Figure 4-1. The 3 dicarboxylic acid (malonic) is unstable and
decomposes to C02 in an oxidizing environment.
12
-------
C02
Formic
Methane
Figure 4-1. Oxidative decarboxylation of adipic acid.
-------
Oxidation of the C-H bonds in dicarb.oxylic acids produces no hydrocarbons
or monocarboxylic acids, but results in lower dicarboxylic acids (2).
Hence, attack of adipic acid can result in glutaric and succinic acids as
well as C02. Malonic (Ca) and oxalic (2) acids are unstable and decompose
quickly to COa.
Tinker (16) studied decarboxylation of adipic acid conjugated with
the air oxidation of cyclohexane at 90-150°C catalyzed by cobalt in acetic
acid solvent. The carboxylic groups of acetic, succinic, glutaric, and
adipic acid were tagged with C14. The evolution of tagged COz was followed
as a function of time. In the absence of cyclohexane, adipic acid did not
decarboxylate. The amount of decarboxylation was proportional to the acid
concentration and to the amount of 02 consumed. The relative rate of
decarboxylation, % decomposition/02 consumed per liter, was approximately
independent of acid concentration and Oa consumption. As shown in Table
4-1, the relative rate of decarboxylation was a function of temperature and
increased about a factor of six from 900C to 150°C. The relative decar-
boxylation rates of succinic and glutaric acids were about half that of
adipic acid. Acetic acid was very stable. Cobalt was not necessary for
decarboxylation, but did catalyze it significantly.
Emanuel Berezin, et al. (2) have studied decarboxylation of monocarboxy-
lic acids in heptane or octadecane undergoing oxidation at 130 C. The de-
carboxylation rate is the same for propionic, n-butyric, and valeric acids,
but about 50% less for acetic acid.
Russian workers have studied the decarboxylation of aliphatic acids
in cumene undergoing oxidation from 80°C to 140°C (2). The decarboxylation
of acetic, n-butyric, isobutyric, and stearic acids takes place as low as
80 C. The relative rates for acetic, n-butyric, n-valeric, and isobutyric
acids are 0.25, 1.0, 1.0, and 5.0, respectively. At 135 C, nonane was de-
tected in the decarboxylation of decanoic acid. The quantity of nonane
increased with a decrease in 02 pressure.
In the oxidation of cumene, it was determined that the primary reaction
resulting in decarboxylation was attack of a peroxy-radical at the ct-methyr
lene group. Rate constants for this reaction are given in Table 4-2. The
rate is uniform for unbranched monocarboxylic acids and is about four times
less with dicarboxylic acids.
Denisov et al. (2) have measured the overall rate of reaction of
cumyl peroxide radicals with acids in chlorobenzene solvent. The rate
constants and activation energies are given in Table 4-3. Longer chain
acids react faster because there are more available C-H bonds. Dicarboxy-
lic acids react about six times faster than monocarboxylic acids with an
equal number of carbon atoms.
14
-------
TABLE 4-1. DECAKBOXYLATION IN THE AIR OXIDATION OF^YCLOHEXANE_ ( 3.3 M)
IN ACETIC ACID (9". 75 M) CATALYZED BY Co (3.3 x 10 2 M)
Acid
Adipic
Glutaric
Succinic
Acetic
T(°C)
90
90
90
125
150
150
150
90
150
90
150
90
115
150
% Decarboxylationa/02 Consumed
b
1.8°
1.3
1.1°
1.8
8.3
d
2.8
10. 06
0.9
4.3
0.9
4.1
0.03
1
0.1
0.6
3.0 moles Oi consumed/liter unless noted otherwise
0.9 moles Oa/liter
C4.2 moles 02,/liter
no cobalt
6Co+2 = 6.6 x 10~2 M
15
-------
TABLE 4-2. RATE CONSTANTS FOR DECARBOXYLATION IN CUMENE UNDERGOING
OXIDATION, FOR THE REACTION, -RiCOOH + R-0-0"*- COMPLEX
(DENISOV et al. 1977)
Acid
Isobutyric
n-Butyric
n-Valeric
Decanoic
Stearic
Sebacic
Azelaic
Pimelic
Rate-constant
(£/mole-sec)
120°C 125°C
0.58
0.44
0.25
3.2
1.5
1.6
1.7
1.9
__
16
-------
TABLE 4-3. RATE CONSTANTS FOR THE REACTION OF CUMYL
PEROXIDE RADICALS WITH CARBOXYLIC ACIDS
IN CHLOROBENZENE (DENISOV,- et al., 1977)
Acid
k( liters/mole-sec)
110°C
130UC
55°CC
Activation
energy
(kcal/mole)
Acetic
Propionic
n-Butyric
Isobutyric
Valeric
Glutaric
Pimelic
Suberic
0.13
0.17
0.25
0.46
0.31
1.8
2.4
3.2
0.4-10~2
0.6-10"2
1.0-10"2
2.4-10~2
1.1-10"2
6.9-10
-2
-2
10-10
12-10 ~2
16.2
15.1
14.8
13.4
15.1
11.4
11.0
11.4
Estimated
17
-------
It appears that free radical attack at both the 0-H and C-H bonds
is possible. Therefore, we can expect the products of decomposition to
include glutaric and succinic acids, monocarboxylic acids from valeric
to formic, and hydrocarbons from butane to methane. The distribution of
products should depend upon the ratio of 02 to S02 in the solution.
If the rate of decomposition of adipic acid is proportional to the
concentrations of adipic acid and free radical then:
decomposition rate = k, [R*] [Adipic] (4-6)
Assuming that the rate of oxidation is proportional to the concentration of
free radical and S02 species gives:
oxidation rate = kQ [R-][S02]T (4-7)
Combining Equations 4-6 and 4-7, the ratio of decomposition rate to oxida-
tion rate can be expressed as follows:
k
moles adipic decomposed _ d [adipic]
moles S02 oxidized -r, , (4-8)
o L 2 JT
The ratio k,/k may be a function of pH, since the actual adipic and S02
species will depend upon pH and may react at different rates with the free
radical.
4.4 OXIDATIVE DECARBOXYLATIONBY METAL IONS
Transition metal ions such as Co 3, Mn 3, Pb ^, and Ce react with
carboxylic acids at 25-100°C and result in decarboxylation. Typical condi-
tions for such reactions require organic solvents such as acetic acid or
aqueous solutions at pH 0-2. The rate of reaction with aliphatic acids
is greatest when the a-carbon is tertiary (isobutyric acid) and least when
it is primary (acetic acid).
Clifford and Waters (18) measured the rate of reaction of Co+3 with
several carboxylic acids at 10-35°C in 1.1 M perchloric acid. They postu-
lated a reaction mechanism of the form:
RCOaH + Co(H20)63+ + R- + C02 + Co2+ + H+ + 6H20 (4-9a)
R. + Co3+ + R+ + Co2+ (4_%)
R+ + H20 -s- ROH + H+ (4-9c)
18
-------
4.5 HEALTH EFFECTS
Toxicology data for adipic acid and several identified and potential
degradation products were obtained from the Dangerous Properties of Indus-
trial Materials (19), Toxic and Hazardous Industrial Chemicals Safety
Manual for Handling and Disposal with Toxicity and Hazard Data (20),
Registry of Toxic Effects of Chemical Substances (21) and Federal Regis-
ter (22). These findings are summarized in Table 4-4. Both adipic and
butyric acid are listed in EPA's hazardous substances list in Category
D with a reportable quantity (RQ) consisting of 5000 Ibs (2270 Kg). Thus,
permits will be required to discharge these two compounds.
19
-------
TABLE 4-4. TOXICOLOGY OF ADIPIC ACID AND RELATED COMPOUNDS
Toxicity
Compound
Adipic Acid
Synonym
Hexanedloic Acid
Toxic Dose
orl-rat LDLO:3600 mg/Kg
Comments
Food additive
Feportable Quantities of Hazardous Substances*
Category
D
RQ in pounds (kilo)
5000(2270)
Valeric Acid
Pentanoic Acid
ipr-rat LD50:275 mg/Kg
orl-mus I,D50:1900 mg/Kg
ipr-mus LD50:275 mg/Kg
ivn-mus LD50:680 mg/Kg
ivn-mus LD50:1290 mg/Kg
orl-mus LD50:500 mg/Kg
scu-mus LD50:3590 mg/Kg
Corrosive, unpleasant
odor
co
O
Glutaric Acid Pentanediolc Acid
Cyclopentanone Ketocyclopentane
Butane Methylethylmethane
Butyl Hydride
Butyric Acid Butanoic Acid
No data found
ipr-mus LD50:1950 mg/Kg
ihl-rat LC50:658 mg/m3
TVL-air:500 ppm
orl-rat LD50:2940 mg/Kg
ivn-mus LD50:800 mg/Kg
orl-rbt LDLo:3600 mg/Kg
Moderately flammable
Flammable
Flammable, unpleasant
odor
5000(2270)
Abbreviations:
orl - oral
rat - rat
ipr - intraperitoneal
mus - mouse
ivn - intraveneous
scu - subcutaneous
ihl - inhalation
TLV - threshold limit value
rbt - rabbit
LDLO - lowest published lethal dose
LD50 - lethal dose 50 percent kill
LC50 - lethal concentration 50 percent kill
LDLo - lowest published lethal concentration
m3 - cubic meter
Taken from the EPA hazardous substances list contained in the Federal Register, Vol. 44 No. 34 - 16 February, 1979 (p 10280-10282)
-------
REFERENCES
1. Rochelle, G. "The Effect of Additives on Mass Transfer in CaC03 and CaO
Slurry Scrubbing of SO2 from Waste Gases," Ind. Eng. Chem. Fundam., 16:
67-75, 1977.
2. Denisov E. T., N. I. Mitskevich, and V. E. Agabekov. Liquid-phase
Oxidation of Oxygen-containing Compounds. Consultants Bureau, New
York, 1977.
3. Chen, T., and C. H. Barren. Ind. Eng. Chem. Fundam., 11:466, 1972.
4. Kiseleva, R. A. and M. S. Dudkin. Zh. Prik, Khim., 40:2513, 1967.
5. Bunce, N. J. J. Org. Chem., 37:664, 1972.
6. Kraeutler, B.; Bard, A. J. J. Am. Chem. Soc., 100(19):598-2-92, 1978.
7. Kuhnle, J. A., R. E. Lundin, and A. C. Waiss. J.C.S. Chem Com., 287,
1972.
8. Keith, C. L., et al. Arch. Microbiol., 118 (2):173-176, 1978.
9. Grula, M. M., Grula, E. A. BERC-RI-76-6, 1976. 61 pp.
10. Hammond, M. W., Alexander, M. Environ. Sci. Technol., 6(8):732-735, 1972.
11. Dias, F. F. and M. Alexander. Appl. Microbiol., 22:1114, 1971.
12. Messier, F.; DeJongh, D. C. Can. J. Chem., 55:2732-2740, 1977.
13. Potai, S. The Chemistry of Carboxylic Acids and Esters. Interscience,
New York, New York, 1969.
14. Anderson, J. M. and J. K. Kochi. J. Am. Chem. Soc., 92:2450, 1970.
15. Sheldon, R. A. and J. K. Kochi. In: Organic Reactions, W. G. Danben,
ed. 1972.
16. Tinker, H. B. J. Catalysis, 19:237, 1970.
17. Berezin, I. V., E. T. Denisov, and N. M. Emanuel. The Oxidation of Cyclo-
hexane. Pergamon, New York, 1964.
21
-------
18. Clifford, A. A. and W. A. Waters. J. Chem. Soc., 2796, 1976.
19. Sax, N. Irving. Dangerous Properties of Industrial Materials, 4th ed.
Van Nostrand-Reinhold, New York, New York, 1975.
20.. International Technical Information Institute. Toxic and Hazardous
Industrial Chemicals Safety Manual for Handling and Disposal with Toxicity
and Hazard Data. Tokyo, 1976.
21. Christensen, Herbert E. and Edward J. Fairchild, eds. Registry of Toxic
Effects of Chemical Substances. NIOSH, Rockville, Maryland, 1976.
22. Federal Register, "Hazardous Substances", EPA, Vol. 44 No. 34 - Friday,
February 16, 1979.
22
-------
SECTION 5.
ANALYTICAL METHODS
A variety of analytical methods were employed in the analysis of samples
collected during field testing and those generated in the laboratory experi-
mentation. The following sections address each of the methods which were
used, with emphasis on application, problems and accuracy.
5.1 ION CHROMATOGKAPHY
Ion chromatography (1C) is a relatively new technique for identi-
fying and quantifying ions in solution, which offers the advantage of
direct measurement of the concentration of the ions of interest. Ion
exchange columns are used to separate the ions which are then detected by
a conductivity meter. At low concentrations conductivity is directly pro-
portional to concentration and is virtually linear. Concentrations of
ions in the sample are determined by comparison with standard solutions.
Identification of ionic species is based upon retention time of the ion
in the system, which ^±s a function of the affinity of that ion for the
separator column under a given set of operating conditions. The reten-
tion time of specific ions is determined by analysis of standard solu-
tions containing the species of interest. A Dionex Model 14 ion chroma-
tograph was used for all 1C work done on this project. Operating condi-
tions are detailed in Appendix A with other pertinent information.
Ion chromatography was used as the principal method for quantitative
determination of adipic acid concentration in all phases of this project.
All material balance calculations related to testing at the IERL-RTP
facility (Section 6) are based on adipic acid concentrations as measured
by 1C. Also, 1C analyses were used for measurement of adipic acid degrada-
tion in all laboratory tests conducted as part of this project. In addition,
1C analyses of liquid samples taken from the bench scale test apparatus
during test runs were used to semi-quantitatively assess sulfite oxidation
during the course of £he experiment. This was done by comparing relative
concentrations of SOi* and S03 in the liquid phase of the test mixture.
Although 1C analyses were found to be the most reliable means for
quantifying adipate concentrations in FGD slurry samples and synthetic
slurry samples, the method does present certain problems and disadvantages.
The most significant problem encountered in the 1C analyses was that of
variations in peak resolution, with the sulfite and adipate peaks being
the most difficult to resolve. The evaluation of 1C data requires esta-
23
-------
blishing a standard curve relating peak height to concentration. Poor
resolution of the peaks in a standard or a sample gives rise to difficulties
in measuring peak heights accurately, since the 'true' baseline is difficult
to establish. Several factors tend to affect the resolution which can be
achieved with 1C. In several instances, observed decreases in the ability
of the instrument to give good resolution of the adipate peak under a
given set of conditions were traced to deterioration of the ion separator
column. In some instances, regeneration procedures restored the column to
previous performance levels. In other instances, it was necessary to modi-
fy parameters such as eluent strength to achieve ..acceptable resolution. In
the most extreme cases, the separator column had to be replaced to bring
instrument performance to an acceptable level. In order to achieve good
resolution of the adipate peak, especially in the presence of sulfite, condi-
tions had to be maintained such that each sample injection required some
18-24 minutes to elute all species completely.
The large number of samples collected'and generated during the course
of this project precluded replicate 1C analysis of every sample. A large
number of samples were analyzed two or more times, however, especially
during the laboratory testing phase of the program. When replicate analy-
ses were performed, the mean value is reported. The analytical data from
twenty two samples which were analyzed two or more times were statistically
evaluated to determine the accuracy of the method. The number of repeat
analyses of a given sample ranged from two to seven, and were separated
in time by up to several weeks. For each sample, the mean and root mean square
(RMS) error were calculated. The RMS error was calculated to be 7% over the
range of adipic acid concentrations measured (see data tables inSection 6).
Thus, based upon the above error value, calculated material imbalances of
greater than 10% are statistically valid.
5.2 GAS CHROMATOGRAPHY
Slurry and gas samples obtained from IERL-RTP and Shawnee facilities
were analyzed for adipic, valeric, glutaric, and butyric acids by gas chroma-
tography (GC). Gas samples obtained during the laboratory test studies
were analyzed for low molecular weight hydrocarbons by GC. Operating condi-
tions for GC analyses are detailed in Appendix A.
Slurry samples, NaOH impinger solutions and XAD-2 sorbent samples
were prepared for analysis by extraction, concentration and derivitization.
Liquid samples were acidified to pH 1 with concentrated HC1 and then ex-
tracted with diethyl ether by either continuous extraction or separatory
funnel extraction. Solid samples were extracted with diethyl ether in a
Soxhlet extractor. The extracts were then concentrated and treated with
diazomethane in order to form the methyl esters of any carboxylic acids
present.
24
-------
Individual samples were analyzed on the GC by direct injection. A
flame ionization detector (FID) was employed in the analysis due to its
sensitivity to all organic compounds. Identification of compounds in the
samples was made by comparison of retention times with those of known
standards prepared in the laboratory. A compound having the same retention
time as a standard compound was identified as that compound. While this
method of analysis is useful for the identification of a wide variety of
compounds, it is at best semi-quantitative. Several prepared solutions
of adipic acid ranging from 1500 ppm to 2500 ppm were extracted and analyzed.
The apparent recovery ranged from 84% to 118%. Duplicate injections of
standards indicate a variation between injections of up to 11%.
Multiple extractions and analyses of valeric acid standard solutions
indicate an extraction and concentration efficiency of about 60%. Dupli-
cate injections of samples indicate a variation between injections of up to
15%.
Gas samples from the laboratory degradation tests were of two types:
headspace gases and gases purged from sorbent traps. The gases were sampled
with a gas-tight syringe and loaded into a sample loop for injection onto
the column. An FID detector was again chosen for the analysis. Peaks
were identified by comparison of retention times with those of a standard
gas mixture.
As in the case of the field sample analysis the specificity of this
method is limited only by the reliability of assigning a positive identi-
fication to a peak with the same retention time as a standard. Accuracy
is dependent on a number of factors including the degree of mixing of the
gas and the reproducibility of the injection technique. A sample was
analyzed four times with an average concentration of 41.1 ppm butane. The
average deviation was found to be +_ 3.4 ppm or J^ 8.3%.
5.3 TOTAL ORGANIC CARBON
i_
Ion chromatography was chosen as the method for quantitative adipic acid
analyses of samples collected and generated during the course of this pro-
ject. While the method has been demonstrated to give reliable values for
adipate in solution, its ability to detect other organic species which might
be present as a result of adipic acid degradation is not well documented.
For this reason, all field samples which were analyzed by 1C were also ana-
lyzed for total organic carbon (TOC). A Dohrmann DC-52D total organic carbon
analyzer was used for these analyses. This instrument couples oxidative-
reductive pyrolysis with a flame ionization detector (FID). A- carbonate
bypass system eliminates instrument response to inorganic carbonates (HCOa
and CO3 ). Methane and ethane are vented during the carbonate bypass se-
quence; all other organic species are stoichiometrically converted to
methane and detected by the FID. Data obtained by TOC analyses were pri-
marily used for comparison to data obtained by GC and 1C methods.
25
-------
The TOG method is not specific for a particular organic species, such
as adipic acid. Quantitative determination of a given species is possible
only if it is the only one present. If, however, the concentration of one
compound has been determined by some other method, additional carbon indi-
cates the presence of other organics in the sample. Contamination can give
rise to indications of extraneous carbon if care is not taken in sample
handling. Filtration of samples through cellulose filters was found to give
greatly inflated TOG values. For this reason, all field samples were fil-
tered using polycarbonate membrane filters. Analytical data on twenty one
samples collected at RTF during the first and second sampling trips were
statistically evaulated to estimate the accuracy of the TOG method. The
number of replicate analyses performed on any given sample ranged from two
to fifteen. These analyses were performed over a period of several days and
all analyses of a given sample were not necessarily done on the same day.
The standard deviations expressed as percents of the means ranged from 0.4%
to 10.5% with an average of 3.8%.
5.4 SOLID PHASE CHARACTERIZATION
5.4.1 Sulfite Analysis
Solids obtained from bench scale degradation studies were analyzed
for their sulfite concentration by an iodometric titration with sodium
arsenite.
In this procedure the sample is added to an excess of buffered iodine
solution. The iodine remaining after the stoichiometric oxidation of sul-
fite is titrated with standard sodium arsenite solution. An amperometric
dead-stop method was used for the endpoint detection. The analytical pro-
cedure is detailed in Appendix A.
5.4.2 Infrared Absorption Spectrophotometry
Calcium sulfite solids, generated to simulate scrubber solids in the
bench scale tests, were analyzed by infrared spectrophotometry using a Perkin
Elmer Model 283 spectrophotometer. This analysis provided a preliminary
qualitative characterization of the CaSOa solids. A verification of the
solids' sulfite concentration was later made using an iodometric titration
method.
Infrared analysis was also performed on field samples collected from
the IERL-RTP pilot facility. These samples included adipic acid feed and
scrubber filtrate solids from the reaction tank.
The major limitation of infrared analysis is that the technique is pri-
marily qualitative. However, with care the relative amounts of CaC03,
CaS0^2HzO and CaS03'%H20 in scrubber solid samples can be determined pro-
26
-------
viding limestone utilization and oxidation ratios. The detection limit for
organic compounds in the solids by direct infrared analysis is approximately
0.1% or 1000 ppm.
5.4.3 Organic Analysis
In addition to the infrared analysis a direct extraction procedure
was used to determine the organic content of the solids. The dried solids
were placed in a sohxlet extractor and diethylether was used as the extrac-
tant. The ether was subsequently removed and analyzed for extracted organics
using the methylation-GC technique used in the liquid phase analyses. No
organics were detected above the 10 ppm (in the solids) detection limit of
this procedure. However, since the solids were not dissolved in this pro-
cedure organics incorporated in the solids would not be detected. Metal
salts of organic acids present in the solid phase would not be measured
using this technique. Future studies are needed to investigate the organic
content of the scrubber solids.
5.5 CARBON DIOXIDE
The flow-through laboratory tests described in Section 7 employed
sodium hydroxide impingers for trapping carbon dioxide generated in the de-
carboxylation of adipic acid. These NaOH solutions were analyzed for COz
using a non-dispersive infrared COz analyzer. In this system, COz is libera-
ted after injection of the sample into an acid solution. The COz is de-
tected by the infrared analyzer and monitored by a recorder equipped with
a disc chart integrator. The integrated peak area is then used to deter-
mine the amount of COa present in the sample.
Samples of the NaOH solution used to trap COz were preserved prior to
analysis by addition of an ammonia-EDTA buffer solution to maintain a pH of
'^10. Samples of this buffer solution and of the stock NaOH solution were
analyzed to determine background levels of COa present in these solutions.
These blank values were taken into account in the calculation of the amount
of COa generated in the degradation reaction.
i
This method of COz determination is quite specific, especially when the
sample matrix is well defined as in this application. The analyzing unit
allows samples to be analyzed without introducing atmospheric carbon dioxide.
Water vapor is removed from the gas stream by a series of drying tubes up-
stream of the detector. By using peak area instead of peak height, inaccura-
cies due to varying pH in the acid pool are virtually eliminated.
27
-------
5.6 GAS CHROMATOGRAPHY-MASS SPECTROMETRY
The tenax tubes, VGA's and some extracts were analyzed utilizing a Hew-
let-Packard 2982A combined gas chromatograph-mass spectrometer (GC-MS).
A Hewlett-Packard 5934A Data System was used for the collection, storage, and
retrieval of data.
The GC-MS instrument consists of a Hewlett-Packard 5710A gas chroma-
tograph and a Hewlett-Packard 5982A dodecapole mass spectrometer and GC-MS
interfaces. The instrument is equipped with a dual ion source for operation
in the electron impact or chemical ionization mode. The major features of
the system include: 3-1,000 amu mass range covered in a single scale; adjust-
able scan rate of 325 amu/sec.; sensitivity to picogram levels, even with large
samples; provision for membrane and jet separators; AID measurements at every
0.1 amu; and resolution permitting full separation of half masses.
28
-------
SECTION 6
RESULTS OF FIELD SAMPLING EFFORTS
During the course of this program Radian personnel conducted four field
sampling trips. Three of these trips were to the RTP pilot unit and one
to EPA's Shawnee test facility. Samples were collected during adipic acid
addition tests to measure adipic acid and degradation products around the
systems. The objective was to identify the degradation products and
determine the relative amounts leaving the systems in the various exit
streams such as the filter cake and scrubbed flue gases.
The sampling procedures and analytical results of the field studies
are presented in this section.
6.1 GAS PHASE SAMPLING
The inlet and outlet flue gases and the oxidizer vent, in the case of
forced oxidation, were sampled to measure the vapor concentration of
any volatile degradation products. Prior to and during gas sampling,
several parameters were measured in order to characterize the gas streams:
gas flow rate, ,
absolute pressure, ,
temperature,
percent COz and 02, and
percent
The gas flow rate was determined through the use of an S-type pitot tube
used in conjunction with either a water manometer or a 0 to 1 inch magnehelic
29
-------
gauge. The stack pressure was determined by disconnecting one leg of the
pitot tube and measuring the pressure with a water manometer. The barometric
pressure was determined by a barometer. The stack temperature was determined
by insertion of a thermometer into the gas stream. Carbon dioxide and oxygen
concentrations were determined by Orsat tests. Percent moisture was deter-
mined by drawing a known amount of stack gas through a pre-weighed,. ice-
cooled water impinger followed by a silica-gel impinger. The impingers
were then weighed to determine the amount of moisture collected.
Flue gas samples were collected in a number of different ways. They
include:
XAD-2 resin traps (RTF only)
Tenax sorbent traps (RTF only)
NaOH impingers (RTF and Shawnee)
Porapak Q sorbent traps (Shawnee only)
The XAD-2 resin traps were designed to collect semi-volatile and
volatile organic compounds. The traps consisted of a 10 inch long by one
inch diameter stainless steel tube packed with approximately 100 ml of the
resin. The sampling train consisted of a six inch stainless steel probe
followed by a Gelman filter. The resin canister was then attached and followed
by a cooled dry impinger to eliminate moisture. The impinger was followed
by a diaphragm pump, a needle valve to control flow and a dry gas meter. A
diagram of the sampling train can be seen in Figure 6-1.
The Tenax sorbent traps were designed to trap medium molecular weight
volatile compounds. They consisted of a 12 inch long by 1/8 inch diameter
stainless steel, glass lined tube packed with Tenax GC 60/80 followed by
silica gel 35/60 mesh. The sampling train was identical to that used for
the XAD-2 canister.
The NaOH impirigers were designed to trap volatile carboxylic acids.
Two Smith-Greenburg impingers charged with IN NaOH were used in each run.
The sampling train was identical to that used previously.
The Porapak sorbent traps were designed to trap low molecular weight
hydrocarbons. They consisted of a 24 inch long by 3/8 inch diameter stain-
less -steel tube packed with approximately 30 ml of Porapak Q 100/120.
During the run, the trap was preceded by one NaOH impinger. The sampling
train was identical to that used previously.
6.2 LIQUID SAMPLING
Slurry samples were filtered on site through polycarbonate membrane
filters into sample containers. A stainless steel filter holder was used
in conjunction with a peristaltic pump. Solids were removed from the filter
and placed in a sample container.
30
-------
LO
H
NaOH
IMPINGER PORAPAK TRAP
ICE BATH
NaOH IMPINGERS
XAD-2 CANISTER/TENAX TRAP
INSET
DRY IMPINGER
DRY GAS METER
PUMP
Figure 6-1. Gas phase sampling train.
-------
6.3 RESULTS OF IERL-RTP MATERIAL BALANCE TEST (10/25/78)
During the week of October 22, Radian personnel conducted a field
trip to IERL-RTP to sample the process streams around the pilot S02 scrubber
during natural oxidation conditions. The samples taken are summarized in
Table 6-1.
TABLE 6-1 SAMPLES OBTAINED AT IERL-RTP
Sampling Point
Type of Sample
1. Scrubber Flue Gas Outlet
2. Scrubber Outlet Slurry
3. Scrubber Feed Slurry
4. Adipic Acid Feed
5. Filter Cake
6. Filtrate from Filter
XAD-2 resin (2)
NaOH impinger
Tenax adsorbent (2)
Filtered liquid (4)
Filtered solids (4)
Volatile organics, unfiltered (4)
Same as #2
plus Tenax adsorbent (2)
of head space
Liquid composite
Solid Sample
Filtered liquid
Volatile organics, unfiltered
Slurry samples were analyzed by gas chromatography (GC), ion chromato-
graphy (1C), and total organic carbon (TOC) analysis. Flue gas samples were
analyzed by GC. The only degradation products identified were valeric
and glutaric acids. Results of these analyses are shown in Table 6-2.
Analysis of volatile organics samples by gas chromatography/mass
spectrometry (GC-MS) showed only trace amounts of organic materials at
the ppb level.
The configuration of the IERL-RTP scrubber under natural oxidation
conditions can be seen in Figure 6-2. Adipic and valeric acid concentrations
and mass flows are indicated.
32
-------
TABLE 6- 2 RESULTS OF ANALYSIS OF IERL-RTP (10/25/78) SAMPLES (NATURAL OXIDATION TESTS)
Stream/Set
Scrubber Outlel
Slurry
I
II
III
IV
Scrubber Feed
Slurry
I
II
III
IV
Filtrate from
Filter
Adipic Acid
Feed Composite
Flue Caa1
XAD-2
XAD-2
NaOU impinger
GC
Adipic Acid Valeric Acid
Time (mg/ID (mmalea CM) (mg/£) (mmolea C
1130-1300 2,060 85 70 3
1345-1430
1530-1650
1730-1745
1130-1300 2,260 93 70 3
1345-1430
1530-1650
1730-1745
1530-1650 2,070 85 80 4
1130-1745 10,610 436 ND2-
1200-1400 ND 1.0 <1
1600-1800 ND 2.3 <1
1430-1455 ND 0.7 <1
Glutaric Acid Adipic Acic
I/I) (mg/H) (mmolee C/H) (mg/fl)
KTP1
10 <1 2,560
2.4JO
2,560
2.560
Avg. 2,520
10 <1 2,410 2,150
2,560
2,630
2.480 2,130
Avg. 2,520
20 1 2,410
ND 9,500
1
(mmoles
105
99
105
105
104
99
105
108
102
104
99
390
1C
Chloride
C/Z) (mg/JO
4,790
4,540
4,720
4,610
4,670
4,610
4,720
4,790
4,820
4,740
4,720
ND
Sulfate
(ppm)
10,760
11,140
11,430
10,760
11,020
10,470
12.010
12,300
15,370
12,540
10,760
49,470
Organic Carbon
(mmule C/ 2}
114
118
112
113
113
112
113
115
111
113
107
406
Sum of Species
(,-naiole C/ SJ
108
102
108
108
107
102
108
111
105
107
104
390
z"Not detected"
'Concentrations are expressed as ppmv
-------
Valeric Acid 10-34 mmole/hr
Mist Eliminator
Adipic Acid
2520 ppra
Valeric Acid
70 ppm
S02 , HC1
Adipic Acid
2410 ppm; 105 mmole/Ir
Valeric Acid
80 ppm; 5 mmole/hr
Glutaric Acid
20 ppm; 1 mmole/hr
Glutaric Acid
10 ppra
Figure 6-2. IERL-RTP scrubber configuratlon (natural oxidation).
-------
TABLE 6-3. MATERIAL BALANCE CALCULATIONS: IERL-RTP 10/25/781
(NATURAL OXIDATION TESTS)
Process Stream
Concentrations
Adlplc Acid
(mmole/Z)
Valeric Acid
(mmole/t)
Glutaric Acid
(mmole/H)
Stream
Flow Rate
U/hr)
Species Flow Rates
Adipic Acid
(uunole/hr)
Valeric Acid
(mmole/hr)
Glutaric Acid
(ncnole/hr)
Total
(nmole/hr)
Inlet
Adipic acid feed
72.7
2.2
160
Total Inlet
160
160
Outlet
Outlet flue gas 3-10xlO~5 3.4xl05 10-34
Filter cake liquor 16.5 0.8 0.2 6.0 99 5
Scrubber liquor losses 17.3 0.7 0.8 1.0 17 1
Adipic
1
1
Total Outlet
Recovery Fraction
Acid Degradation
10- 3A
105
19
134-158
84-99%
28%
'Bloudown data based on IERL-RTP calculations for this period
-------
As can be seen from Figure 6-2, the rate of valeric acid loss in the
flue gas is in the range of 10 mmole/hr to 34 mmoLe/hr. Since the adipic
acid feed rate to the scrubber is 160 mmole/hr, the loss of valeric
acid represented from 6% to 21% of the amount of adipic acid fed into
the system.
To assess whether the measured gas phase concentrations of valeric
acid are reasonable, the stripping rate or change in the liquid phase
valeric acid concentration across the scrubber was calculated. Using the
extreme values of valeric acid measured in the gas phase, the change in the
liquid concentration would range from 0.2 to 0.8 mg/Ji. This represents a
change in the liquid phase concentration, 70 mg/£, or less than 2%, which
considering the boiling point of valeric acid, seems reasonable.
However, a direct measure of the change in the liquid concentration of
valeric acid across the scrubber would be difficult.
The material balance for adipic acid and its degradation products
shown in Table 6-5 closes within 15% or less depending on which gas phase
valeric acid value is correct. Approximately 28% of the adipic acid added
during the natural oxidation tests is degraded. Valeric acid appears to
be the major decomposition product with most of the valeric discharged in
the outlet flue gas. Some glutaric acid was also found in the aqueous
solutions within the scrubber system but none in the flue gases.
6.4 RESULTS OF IERL-RTP MATERIAL BALANCE TEST (11/15-16/78)
During the week of November 13, Radian personnel conducted a field
trip to IERL-RTP to sample the process streams around the pilot SOa scrubber
during forced oxidation testing. The samples taken are summarized in
Table 6-4-
TABLE 6-4. SAMPLES OBTAINED AT IERL-RTP (11/15-16/78)
Sampling Point
Type of Sample,
1. Scrubber Flue Gas Outlet
2. Scrubber Outlet Slurry
3. Scrubber Feed Slurry
4. Adipic Acid Feed
XAD-2 resin (2)
NaOH impinger (2)
Tenax adsorbent (2)
Filtered liquid (4)
Filtered solids (4)
Volatile organics, unfiltered
Same as #2
Liquid composite
(continued)
36
-------
TABLE 6-4. Continued
Sampling Point
Type of Sample
5. Filter Cake
6. Filtrate from Filter
7. Flue Inlet to Scrubber
8. Oxidizer Air Vent
Solid sample
Filtered liquid
XAD-2 resin
NaOH impinger
Tenax adsorbent
NaOH impinger
Slurry samples collected at IERL-RTP were analyzed by gas chromatography
(GC), ion chromatography(1C), and total organic carbon .(TOC) analysis.
Flue gas samples were analyzed by GC. Results of these analyses are shown
in Table 6-5. In addition, tenax adsorbent samples of the flue gas outlet
and flue inlet to scrubber were analyzed by GC-MS and compared. Differences
due to the scrubber were too insignificant and poorly resolved to warrant
meaningful identification.
The configuration of the IERL-RTP pilot scrubber under forced oxidation
conditions can be seen in Figure 6-3. Adipic and valeric acid concentrations
and mass flows are indicated in the appropriate places.
The rate of valeric acid loss indicated in Figure 6-3 is in the range
of 1.4 mmole/hr to 6.5 mmole/hr. The adipic acid feed rate was 127 mmole/hr.
Therefore, the amount of adipic acid lost as valeric acid in the flue
gas is in the range of 1% to 5% of that added.
The material balance for adipic acid and its degradation products
shown in Table 6-6 closes only within 58% or less depending on which gas
phase valeric acid value is correct. Approximately 69% of the adipic
acid added during this forced oxidation test is degraded. Again valeric
acid appears to be the major decomposition product, but in this case
glutaric acid makes a significant contribution to the material balance.
Valeric acid is discharged in the flue gas and liquor while glutaric acid
is discharged only in the liquor.
6.5 RESULTS OF IERL-RTP MATERIAL BALANCE TEST (12/18-22/78)
During the week of December 18, Radian personnel conducted a field trip
to IERL-RTP to sample the process streams around the pilot SOz scrubber
during forced oxidation testing. The samples taken are summarized in Table 6-7
37
-------
TABLE 6-3 RESULTS OF ANALYSTS OF IERL-RTP Ul/J 5/78-1 L/16/78) SAMPLES
(FORCED OXIDATION TESTS)
U)
00
Stream/Set Time
Scrubber Outlet
Slurry
I 1345-1415
III 06o£o610 COIBPO
IV 1340-1350
Scrubber Feed
Slurry
I 1345-1415
m llltllll Coi?pD
IV 1340-1350
Flue Gas"
XAD-2 t 1310-1510
SparEe Outlet 2158-2223
NaOH Imp Inge r 11
j
z"Not Detected"
GC
(rng/4) (inoioles C/Jt) (ng/X) (mmoleg C/£) (ng/i) (mules C/JL) (ing/O (
RTF1
730
,slte810 33 80 4 70 3 ^
880
Avg. 760
590
site 860 35 100 5 70 3 76°
730
Avg. 710
ND 0.07 <1
ND 0.4 <1
mmoles
30
27
32
36
31
24
30
33
30
29
1C
C/Z) (mg/O
RTF1
9080
8960 1040
9700
9870
Avg. 9400
8960
9080 1030
9820
9650
Avg. 9380
-------
Valeric Acid 1.4-6.4 mraole/hr
Mist Eliminator
Co
VD
Adipic Acid
127 ramole/hr
Adipic Acid
710 ppm
Valeric Acid
I00 ppm
Adipic Acid
440 ppm; 6.9 nimole/lir
Valeric Acid
50 ppm; 1.2 mmole/lir
Glutarlc Acid
50 ppm; 1-0 mmole/hr
Adipic Acid/8040 ppm
Figure 6-3. IERL-RTP scrubber configuration (forced oxidation).
-------
TABLE 6-6 MATERIAL BALANCE CALCULATIONS; IERL-RTP 11/15/78-U/16/781
(FORCED OXIDATION TESTS)
Concentrations Stream Species Flow Rate
Adipic Acid Valeric Acid Glutaric Acid Flow Rate Adipic Acid Valeric Acid Glutaric Acid
Process Stream (mmole/fc) (mmole/Jl) (mmole/£) (Jl/hr) (mmole/hr) (mmole/hr) (mmole/hr)
Inlet | Adipic acid feed 55.0 - 2.3 127
" Total Inlet
Outlet flue gas 3-13 x 10"6 4.8 x 105 1.4-6.4
Outlet Filter cake liquor 3.0 0.5 0.4 2.3 6.9 1.2 1.0
Scrubber liquor losses2 5.0 0.9 0.5 6.6 33 5.9 3.3
Total Outlet
Recovery Fraction
Adipic Acid Degradation 3
Total
(mmole/hr
127
127
1.4-6.4
9.1
42.2
52.7-57.7
42 - 45%
69%
'slowdown data based on IERL-RTP calculations for this period.
2Incidental scrubber solution losses other than those associated with the filter cake.
Degradation is defined as the difference between the adlplc acid feed rate and the measured discharge rate.
-------
TABLE 6-7 SAMPLES OBTAINED AT IERL-RTP
Sampling Point
Type of Sample
1. Scrubber Flue Gas Outlet
2. Scrubber Outlet Slurry
3. Scrubber Slurry Hold Tank
4. Adipic Acid Reaction Tank
5. Adipic Acid Feed
6. Filter Cake
7. Filter Liquor
8. Inlet to Flue
9. Oxidizer Air Vent
XAD-2 resin (composite)
NaOH impinger (composite)
..Tenax adsorbent (4)
Filtered liquid (8)
Filtered solids (8)
Volatile organics (8)
Filtered liquid (4)
Filtered solids (4)
Volatile organics (4)
Filtered liquid (7)
Filtered solids (5)
Volatile organics (5)
Liquid composite
Solid Sample (2)
Volatile organics (2)
Filtered liquid (2)
XAD-2 resin
NaOH impinger
Tenax adsorbent
NaOH impinger
Tenax adsorbent
Slurry samples collected at IERL-RTP were analyzed by GC, 1C and TOG.
Flue gas samples were analyzed by GC. Results of these analyses are shown
in Table 6-8.
The configuration of the scrubber under forced oxidation conditions
is shown with adipic and valeric acid concentrations and mass flows in
Figure 6-4.
The average rate of valeric acid loss indicated in Figure 6-4 is
8.3 mmole/hr. The adipic acid feed rate measured by Radian personnel for
the period during which samples were analyzed was 520 mmole/hr. The loss
as valeric acid accounts for 1.6% of the adipic acid added.
41
-------
TABLE 6- 8. RESULTS OF ANALYSIS OF IERL-RTP (12/18/78-12/22/78) SAMPLES
(FORCED OXIDATION TESTS)
Stream/Set
Scrubber Outlet
Slgrry (SC)
IX
X
Scrubber Feed
Slurry (HT)
IX
X
Filter Liquor XI
Adiptc
Feed
Flue Gas'
XAD-2
NaOH Impinger
Sparge outlet
GC
1C TOC
Adlplc Acid Valeric Acid Glutaric Acid Adipic Acid Organic Carbon Sum of Species
Time (mg/i) (mraoles C/Jt) (mg/i) (mmoles C/Jt) (mg/Jl) (mmoles C/Z) (rog/Jl) (mg/4) (mmoles Cli.) (mmoles C/Z) (mmoles C/S.)
12/21/78
1315 3.070 126
2150 3.370 138
1350 3,620 149
2210 2.2602 932
2/22/78
070 1.9802 812
composite 11.1802 4592
composite
composite
1320 _
280 14 320 12
120 6 160 °6
120 6 190 7
180 9 180 7
120 6 190 7
ND3 ND
*
0.43
0.41
0.23
RTF1
3,360 138 210 164
3,870 159 224 171
3,480 3,875 143 203 156
4,400 2,950 181 229 197
* **'"' l/o 152 18'9
26,520 1,090 1,334 1,090
'Concentration ae analyzed by Accurex personnel on site
Inefficient extraction la indicated
'"Not detected" X
'Concentrations are expressed as ppmv
-------
Valeric Acid 8.3 mmole/hr
Mist Eliminator
Adiplc Acid
520 imnole/hr
150 ppra
Adipic Acid 26,520 ppra Glutaric Acid
180 ppm
Figure 6-4. IERL-RTP scrubber configuration (forced oxidation).
Adipic Acid
4290 ppm; 67.6 mmole/hr
Valeric Acid
120 ppm; 2.8 mmole/hr
Glutaric Acid
190 ppm; 5.2 mmole/hr
-------
TABLE 6-9
MATERIAL BALANCE CALCULATIONS; IERL-RTP 12/21/7B1
(FORCED OXIDATION TESTS)
Process Stream
Concentrations
Adipic Acid
(mmole/i)
Valeric Acid
(mmole/Jl)
Glutaric Acid
(mmole/Jl)
Stream
Flow Rate
(Jl/hr)
Species Flow Rate
Adipic Acid
(mmole/hr)y
Valeric Acid
(mmole/hr)
Glutaric Acid
(mmole/hr)
Total
(mmole/hr)
Inlet } Adipic acid feed
181
2.9
Total Inlet
520
-P-
-P-
Outlet
Outlet flue gas
Filter cake liquor
Scrubber liquor loss
29.4
25.9
1.7 x 10
1.2
1.7
1.4
1.6
4.8 x 10
2.3
6.6
67.6
171.0
8.3
2.8
11.2
3.2
10.6
8.3
73.6
192.8
Total Outlet 275
Recovery Fraction 53%
Adipic Acid Degradation 3 54%
'Slowdown data based on IERL-RTP calculations for November 13-17.
2Incidental scrubber solution losses other than those associated with the filter cake.
Degradation is defined as the difference between the adipic acid feed rate and the measured discharge rate.
-------
The material balance for adipic acid and its degradation products
shown in Table 6-9 closes only within 47%. Approximately 54% Of the adipic
acid added during this forced oxidation test is degraded. Both valeric
acid, which is discharged in the flue gas and liquor, and glutaric acid,
which is discharged in the liquor, appear to be major degradation products.
In addition, an extract of the filter liquor w'as analyzed by GC-MS and found
to contain small amounts (<10 ppm) of succinic acid. The amount of succinic
acid was not great enough, however, to significantly affect the material
balance.
6.6 RESULTS OF SHAWNEE MATERIAL BALANCE TEST
Radian personnel conducted a field trip to EPA's Shawnee Test Facility on
February 2, 1979 to sample the process streams around the scrubber. The
samples taken are summarized in Table 6-10.
TABLE 6-10 SAMPLES OBTAINED AT SHAWNEE
Sampling Point
Type of Sample
1. Venturi Inlet
2. Venturi Outlet
3. Oxidizer Air Vent
4. Venturi Feed
5. Spray Tower Feed
6. Clarifierr Underflow
7. Filter Effluent
8. Spray Tower Effluent
NaOH impinger
Porapak sorbent
NaOH impinger
Porapak sorbent
NaOH impinger
Porapak sorbent
Filtered Slurry
Filtered Solids
Volatile Organics
Filtered Slurry
Filteres Solids
Volatile Organics
Filtered Slurry
Filtered Solids
Volatile Organics
Filtered Slurry
Filtered Solids
Volatile Organics
Filtered Slurry
Filtered Solids
Volatile Organics
45
-------
TABLE 6-10 Continued
Sampling Point Type of Sample
9. Clarifier Overflow Filtered Slurry
Filtered Solids
Volatile Organics
10. Venturi Effluent Filtered Slurry
Filtered Solids
Volatile Organics
Slurry samples collected at Shawnee were analyzed by GC, 1C and TOC.
Flue gas samples were analyzed by GC. Results of these analyses are given
in Table 6-11.
The configuration of the Shawnee scrubber is shown along with adipic
and valeric acid mass flows in Figure 6-5.
The rate of valeric acid loss in the flue gas is 1300 mmole/hr.
The adipic acid feed rate is 17,400 mmole/hr. The loss as valeric acid
accounts for 7.5% of the adipic acid added.
The material balance for adipic acid and its degradation products
shown in Table 6-12 closes only within 17%. Approximately 91% of the
adipic acid added during this forced oxidation test is degraded. The
major degradation product appears to be valeric acid which is lost from.
the. flue. Valeric and-glutaric acid were both found in the liquid phase
portion of the recycle slurry samples, but not in the liquid associated
with the filter cake. This lack of valeric and glutaric acid in the filter
liquor is significant in that it may indicate these species are consumed in
the scrubber slurry.
In an attempt to detect low molecular weight hydrocarbons in the flue
gas at Shawnee, a sample of the flue gas was drawn through a sorbent trap
containing Porapak Q. Earlier work had indicated that trapped hydrocarbons
could be thermally de-sorbed from Porapak Q. Subsequent testing however in-
dicated that upon heating Porapak Q bleeds methane, ethane, propane and bu-
tane, making identification of components of the flue gas difficult if
not impossible.
6.7 SOLIDS CHARACTERIZATION
The presence of many organic compounds during the precipitation of
inorganic (ionic) crystals is known to cause changes in the particle shape
and/or size. This change in crystal morphology may or may not be associated
with an incorporation of the organic species into the crystalline phase.
In order to evaluate the effects of adipic acid on the precipitation of
calcium sulfite and calcium sulfate solid samples collected during the
46
-------
TABLE 6-11. RESULTS OF ANALYSIS OF SHAWNEE (2/2/79) SAMPLES
(FORCED OXIDATION TESTS)
Adlpic Acid
(mg/t) (mmoles C/l)
Valeric Acid
(mg/Jl) (mmoles C/J.)
Glutaric Acid
(mS/£) (mmoles C/J.)
(mg/1)
Adlpic Acid
(morales C/i)
(mg/l)
Organic Carbon
(mmoles C/£)
Venturl Feed Slurry
Spray Touer Feed-Slurry
' 416
960 39
25 1
820
470
34
99
72
Clarlflar Underflow
419
Spray Tower Effluent
424
Clarifler Overflow
428
Venturl Effluent
431
Filtrate
425
1180
48
250 12
20 1
450
690
620
28
1544
80
Gas
Venuurl Outlet
Venturi Oxidizer Vent
0.8s
ND
'Concentration as analyzed by TVA personnel on site
2"Not Detected"
Concentration expressed as ppmv
-------
FLUE GAS
00
REHEAT
FLUE GASJ
ALKALI
ADIPIC
ACID
17,400 ramole/hr
{ MAKEUP WATER
Valeric Acid
1300 mmole/hr
SPRAY TOWER EFFLUENT
HOLD TANK
(CLARIFIED LIQUOR
^ OVERFLOW
0
o
f
*
ON TANK
_L-»
C
c
1
3
BLE
DEW/
.r
DESUPERSATURATION ^=
TANK
Adlpic Acid
1620 mmole/hr
BLEED TO SOLIDS
DEWATERING SYSTEM
Figure 6-5. Shawnee scrubber configuration.
-------
TABLE 6-12
Inlet >
Adiplc acid feed
MATERIAL BALANCE CALCULATIONS
(FORCED OXIDATION TESTS)
SHAWNEE FACILITY1
Process Stream
Adipic Acid
(mmole/S.)
Concentrations
Valeric Acid
(mmole/i)
Glutaric
(mmole
Acid
/*)
Stream
Flow Rate
U/hr)
Adipic Acid
(mmole/hr)
Species Flow
Valeric Acid
(mmole/hr)
Rate
Glutaric Acid
(mmole/hr)
Total
(ramole/htf
Outlet flue gas
Filter cake liquor
Scrubber liquor losses2
7.9
3.63 x 10 5
nd
3.58 x 107
204
17,400
1,620
Total Inlet
1,300
Total Outlet
Recovery Fraction
Adipic Acid Degradation !
17,400
17,400
1,300
1,620
2,920
17%
91%
Slowdown data, adipic acid feed rate and gas flows based on Bechtel calculations for October 10 through November 1.
2Incidental scrubber solution losses other than those associated with the filter cake.
'Degradation is defined as the difference between the adipic acid feed rate and the measured discharge rate.
-------
adipic acid addition tests were examined using scanning electron microscopy
(SEM)- Infrared analyses were also performed to semiquantitatively deter-
mine the ratios of calcium carbonate (CaC03), calcium sulfite hemihydrate
(CaS03'%H20), coprecipitate (xCaS03«%H20 (x-l)CaS(K -%H20) , and calcium
sulfate dihydrate
The electron micrographs of the solids collected from the IERL-RTP
pilot system during natural oxidation tests are shown in Figure 6-7.
These are compared to calcium sulfite crystals previously taken from
Shawnee during low oxidation conditions, (Figure 6-6) but with no adipic
acid present. The crystals grown in the presence of adipic acid appear to
be much more layered than usual. This appears to correlate with a
small change in filterability of the solids as measured during RTF tests.
The weight percent solids dropped to 49% at adipic acid concentrations
near 2000 ppm from a normal value of 55% (1) .
The calcium sulfate dihydrate crystals produced during forced oxida-
tion runs using adipic acid seem to show a decrease in average size as the
concentration of adipic acid increases. This effect can be seen in the
electron micrographs shown in Figure 6-8 to 6-11. .
Infrared spectral analyses were performed to characterize the general
composition of the scrubber solids. Samples from both natural and forced
oxidation tests were analyzed. No organic compounds could be detected
in these samples (detection limit is approximately 1000 ppm) . The inorganic
compositions as determined by infrared spectroscopy are shown in Table
6-13 along with average compositions of the solids analyzed by on-site
personnel during the appropriate test periods.
TABLE 6-13 COMPOSITIONAL ANALYSIS OF SCRUBBER SOLIDS
Technique
Sample Period
Limestone Utilization
m
Oxidation Fraction
Chemical Analysis (RTF)
Infrared Spectroscopy
Infrared Spectroscopy
Chemical Analysis (RTF)
Infrared Spectroscopy
Infrared Spectroscopy
10/23-27/78
10/25/78 (am)
10/25/78 (pm)
11/13-17/78
11/15/78
11/16/78
92-103
>95
>95
79-82
80
75%
30
20
69
97
75
>95
50
-------
Figure 6-6. Calcium sulfite hemihydrate crystals taken from EPA's Shawnee
Test Facility, no additives (x500).
Figure 6-7.
Calcium sulfite hemihydrate crystals from natural oxidation
tests at RTF, 2500 ppm adipic acid (x590).
51
-------
Figure 6-8. Calcium sulfate dihydrate crystals from forced oxidation tests
at RTF, 600 ppm adipic acid (x200).
Figure 6-9. Calcium sulfate dihydrate crystals from forced oxidation tests
at RTF, 4000 ppm adipic acid (x200).
52
-------
Figure 6-10. Calcium sulfate hemihydrate crystals from forced oxidation tests
at RTF, 600 ppm adipic acid (x460).
Figure 6-11.
Calcium sulfate hemihydrate crystals from forced oxidation tests
at RTF, 4000 ppm adipic acid (x460).
53
-------
6.8 FIELD DATA CORRELATION
The adipic acid degradation observed at RTF and Shawnee was correlated
with the sulfite oxidation rate and the adipic acid concentration in solu-
tion to determine the influence of these parameters. The test data used
in this correlation are presented in Table 6-14. The graphical relationship
is shown in Figure 6-12. Although there is substantial scatter in the data,
it does appear that increases in both the sulfite oxidation rate and the
adipic acid concentration will produce an increase in the degradation rate.
Therefore, for control purposes, the adipic acid concentration should be
maintained at a value just needed to accomplish the desired increase in
removal efficiency. Since the use of organic acids for increased mass
transfer purposes will be most desirable in forced oxidation solutions,
the oxidation rate will be set by the amount of SOa absorbed across the
scrubber. The stoichiometry of excess air to sulfite may have some
effect on the degradation mechanism.
54
-------
TABLE 6-14 ADIPATE DECOMPOSITION DATA FROM RTF AND SHAWNEE
Ul
Ul
Date
9/14-18/78
10/16-20/78
10/23-27/78
10/30-11/3/78
11/6-10/78
11/13-17/78
11/27-12/1/78
10/8-30/78
10/10-30/78
Location
RTF
RTF
RTF
RTF
RTF
RTF
RTF
Shawnee
Shawnee
Run Type
For.
Nat.
Nat.
Nat.
For.
For.
For.
For.
Nat.
Ox.
Ox.
Ox.
Ox.
Ox
Ox.
Ox.
Ox.
Ox.
Adipate Cone.
(ppra)
1650
2050
2090
1630
709
760
633
2350/1550
1620
Adipate Degradation Sulfite Oxidation
Rate (Ib/hr) Rate (Ib S02/hr)
.0300
.0142
.0143
.0144
.0255
.0238
.0243
4.9
2.3
5
0
0
1
4
5
5
490
110
.1
.90
.94
.2
.9
.4
.5
-------
o 0.025 .
u
rt
T)
0.020
01
0.015
o.oio
3
o
o
ri
U
0.005
X
X
x
X
X
500
1000 1500 2000
Adipate Concentration (ppm)
LEGEND:
Q - RTF - Nat. Ox.
Qf - RTF - For. Ox.
Q]H - Shawnee - TCA
Q]F - Shawnee - V/S.T.
^^ - Radian Laboratory Test
-i 1
25000
3000
Figure 6-12. Adipate decomposition as a function of sulflte oxidation and adlpate concentration.
-------
REFERENCES
1. Limestone Scrubbing of SOa at EPA/RTP Pilot Plant - Progress Report
37 - November 1978, Mr. Robert H. Borgwardt.
57
-------
SECTION 7
LABORATORY TESTS
A series of laboratory tests was conducted to study adipic acid degra-
dation under conditions simulating those found in FGD systems. The lab tests
were designed to maintain scrubber-like conditions while allowing for a
broad range of experimental flexibility. The tests were primarily aimed at
providing the necessary data for characterizing the conditions under which
degradation occurs, and the mechanism responsible for the degradation pro-
cess.
7.1 SCREENING TESTS
Initial bench scale testing consisted of a series of screening tests
designed to narrow the scope of subsequent testing. The first priority
in the screening tests was to establish a set of conditions sufficient for
degradation to occur, and then to demonstrate the degradation process in the
laboratory. Once degradation was induced under laboratory conditions, sub-
sequent tests were designed to define the essential elements in the degrada-
tion process. Manipulation of these elements was then used to yield data
pertaining to the nature of this mechanism.
Slurry samples collected by Radian personnel at IERL-RTP during a
natural oxidation material balance run on October 25, 1978 were composited
for use in the initial screening tests. The composite was divided into five
portions for the test, 01-05. An additional slurry sample, collected
November 15, 1978 during a forced oxidation run, was designated as 06. All
samples had been refrigerated in sealed plastic bottles since they were
collected.
Samples 02 and 04 were filtered through 1 ym membrane filters, and
the filtrate was used for the test. Samples 01, 03, 05 and 06 were left
unfiltered. Sample 05 was sterilized by boiling for '\>2 hours. Samples
03, 04, 05, and 06 were placed in fritted glass impingers and 01 and 02 were
placed in Erlenmeyer flasks. All six vessels were suspended in water baths
maintained at ^50 C. Samples 03, 04, 05 and 06 were supplied with air at
0-120 ml/min. The air for the purged samples was saturated with water by
first passing through an inipinger of DI water maintained at the same tempera-
ture as the samples. This was done to minimize the concentrating effect
due to evaporation. Samples 01 and 02 were not purged with air, but were
stirred in an open flask. The appratus used for samples 03-06 is diagrammed ,
in Figure 7-1.
58
-------
PURGE GAS IN
Ul
DI II20
FLOW METER
NaOII
SOLUTION
REACTION
VESSEL
Figure 7-1. Laboratory apparatus for screening
tests (construction consisted of
glass vessels with teflon tubing
and connections)
-------
Under these conditions, the test was expected to reveal the following:
the effect, if any, of an air purge on the adipic acid
concentration in the raw slurry as compared to slurry
which was not purged with air (03 vs_ 01) ,
the effect, if any, of an air purge on the adipic acid
concentration in slurry filtrate as compared to filtrate
which was not purged with air (04 vs 02),
- the difference, if any, in effect of an air purge on the
adipic acid concentration in natural oxidation slurry
as compared with forced oxidation slurry (03 vs 06), and
the ability of microbial activity to account for changes
in adipic acid concentration in scrubber slurry (03 vs 05).
The initial adipic acid concentration of samples 01-05 was determined
by analysis of the composite slurry sample by ion chromatography(IC).
The initial concentration of_sample 06 was determined by the same method.
After 36 hours each of the six samples was again analyzed to determine the
percent loss of adipic acid. The results are shown in Table 7-1.
TABLE 7-1. RESULTS OF PHASE ONE LAB SCREENING TESTS
Sample
01
02
03
04
05
06
Description
Slurry, no air purge
Filtrate, no air purge
Slurry, air purge
Filtrate, air purge
Boiled slurry, air purge
Forced oxy. slurry, air
purge
Initial
(mg/£)
2780
2780
2780
2780
2780
730
Final
(mg/£)
2190
2700
2050
3070
2120
730
Loss
(%)
21
3
26
-10
24
<1
Several implications can be drawn from the data in Table 7-1. First, de-
gradation of adipic acid occurs to a significant extent only in solutions
which contain solids. If degradation occurs at all in the filtrate, it is at
a much slower rate. The apparent 3% loss indicated in Sample 02 is more
likely due to analytical error. The increase in adipic acid concentration
in 04 is most probably due to evaporation, since none of the values reported
were corrected for evaporation. Samples 01, 03 and 05 all showed substantial
losses of adipic acid. The differences between the three are not considered
significant when analytical error is taken into account. This would seem
to indicate that an air purge is not necessary for degradation to occur
in the raw slurry. If air is necessary, the amount available to Sample 01
by stirring in an open flask was sufficient, and Samples 03 and 05 were sim-
ply provided with a large excess which did not significantly affect the net
60
-------
result. The analytical data also indicate that biological acitvity does not
play a significant role in the degradation process since Sample 05 showed a
loss of adipic acid comparable to that seen in 01 and 03.
Of the four samples in this test which contained solids, only 06 showed
no degradation of adipic acid. This sample differed from the other three
slurry samples in two major ways. First, its initial adipic acid concentra-
tion was approximately one fourth that of the other three. Second, it was
obtained during a forced oxidation run in which sulfite oxidation was about
98 mole percent while the other three were collected during a natural oxidation
run in which oxidation was about 30 mole percent. Therefore, Sample 06
contained virtually no sulfite while Samples 01, 03 and 05 had relatively
high sulfite concentrations in the solids. Since degradation has been observed
at RTF with adipic acid concentrations as low as those in Sample 06, this
is probably not a significant factor. That would imply a link between adipic
acid decomposition and the oxidation of sulfite.
The second phase of the screening tests was designed to confirm the
apparent link between sulfite oxidation and adipic acid degradation, and
to yield qualitative data pertaining to pH dependence. The following
comparisons were made in this series of tests:
unpurged. solution vs nitrogen purged solution vs air purged
solution,
initial pH 4 vs pH 5 vs pH 6 with an air purge and
solids (no attempts to control pH durin g test were made),
and
CaS03«%HaO added (50 grams/Jl) vs_ no CaS03-%H20 added.
This phase of the screening tests differed from the first in several
important ways. The primary difference was in the nature of the solutions
tested. Synthetic "slurry" solutions were used in these tests rather than
actual samples of real scrubber slurry. A solution of approximately 1700
mg/£ of adipic acid was used to simulate slurry filtrate. Unfiltered
slurry was simulated by adding calcium sulfite solids to the adipic acid
solution. Another difference in these tests was the use of bottled purge
gases. The initial set of tests relied on a purge of room air via a peris-
taltic pump. The third major difference was in the samples which were not
purged. In the initial tests, the two samples which were not purged (01
and 02) were stirred in an open flask. In these tests the unpurged samples
(07 and 08) were stirred in sealed flasks so that the only oxygen available
was that which was in the relatively small headspace of the flask at the
start of the run. All tests were run with the reaction vessels suspended
in a water bath maintained at 50 C. Tests 09-19 were conducted using
a series of fritted glass impingers, as before (Figure 7-2). In each
61
-------
test, the initial and final adipic acid concentrations were determined by
ion chromatography. The results of this series of tests are presented in
Table 7-2.
TABLE 7-2. SUMMARY OF PHASE TWO - LAB SCREENING TESTS
07
08
09
10
11
12
13
14
Test Designation
Adlpate Solution
Adipate/CaS03-JsH20
Adipate/CaS03'»sH20
Adlpate Solution
Adipate/CaS03.%H20
Adipate Solution
Adipate/CaS03'%H20
Adipate/CaS03-%H20
Initial Adipic Initial
Acid Cone. pH
.1640
1640
1640
1640
1640
1640
1770
1770
5
5
5
5
5
5
4
6
Purge
(sealed)
(sealed)
Nz
N2
Air
Air
Air
Air
Loss
of Adipic Acid
2%
1%
<1%
8%
38%
7%
24%
36%
Test
Duration
24 hr
24 hr
24 hr
24 hr
24 hr
24 hr
24 hr
24 hr
As indicated in the table above, a significant amount of degradation
occurred only when sulfite was available and oxidizing conditions pre-
vailed (Tests 11, 13 and 14). The degradation values for the other tests
are low enough to be considered insignificant and are most likely due to
analytical error. Although the data do not conclusively prove that degra-
dation does not occur in the absence of sulfite and/or oxygen, they do indi-
cate that sulfite oxidation is sufficient to cause the degradation of
adipic acid. The adipic acid losses in Tests 11, 13 and 14 are of the same
order as that measured in the initial tests. Comparing Tests 13 and 14, the
data show that the net loss was 50% greater in Test 14 than in Test 13 over
the same period of time. This would seem to indicate that the rate of de-
gradation is slightly higher at pH 6 than at pH 4. However, the glass
frit on the air sparge tube in the reaction vessel used in Test 13 became
clogged with, gypsum twice during the test causing the air flow to cease. The
total amount of time during which there was no flow is not known, but could
have been as much as eight hours. Therefore, this test cannot be regarded
as conclusive evidence of pH effect, but may simply reflect the dependence
of degradation rate on sulfite oxidation.
The conclusions that can be drawn from the screening studies are out-
lined in the following list:
adipic acid degradation occurs in the presence of sulfite
oxidation,
sterilization of slurry didn't prevent adipic acid degradation
indicating that bacterial degradation is not important,
62
-------
contact of adipic acid solutions with oxygen
doesn't result in degradation,
adipic acid is not stripped (volatilized from solutions
at 50°C),
adipic acid is apparently not adsorbed by either
CaS03*%H20 or CaSOi^HaO, and
dependence of degradation on pH was not established.
The third phase of the screening tests was specifically aimed at detect-
ing any low molecular weight hydrocarbons which might be generated in the
degradation process. Such compounds would not have been detected in the
field sampling efforts conducted prior to these tests due to the nature of
the methods employed. However, the oxidative decarboxylation of adipic acid
would be expected to yield such products to some extent.
Tests conducted during this phase of the laboratory studies used
sealed 500 ml flasks as reaction vessels to facilitate trapping and concentra-
tion of gaseous products. Since an air purge was not possible with this
type of apparatus, vigorous stirring provided the necessary gas-liquid
contact. A 10:1 gas-liquid ratio (volume) was used to provide sufficient
oxygen for the reaction to proceed to a measurable extent. Three reaction
vessels were set up for the first part of this test. One contained a 1500
ppm solution of adipic acid, another contained 1880 ppm solution of valeric
acid, and the third contained an equivalent volume of deionized water.
Calcium sulfite solids were added to each vessel in the ratio of 0.05
grams of solids per milliliter of solution. All three were sealed, then
stirred at 50 C for 36 hours. At the conclusion of the run, the headspace
of each flask was analyzed by GC. The results are presented in Table 7-3.
TABLE 7-3. RESULTS OF SEALED FLASK EXPERIMENTS
Sample
Description
Initial
Adipic Acid
(ppm)
Concentration
Valeric Acid
(ppm)
Headspace
Butane
(PPm)
Gas
Deionized HaO
+CaS03'%H20
Valeric Acid
Solution +
CaS03'%H20
Adipic Acid
Solution +
1880
1500
1280
20
Air
Air
Air
63
-------
The data from these three tests confirm that butane can be generated in
the degradation process. The amount of oxygen available in these tests would
be expected to be the limiting factor in the degradation reaction. Since the
solutions were not analyzed at the conclusion of the run, the amount of
butane generated in these tests cannot be correlated with the amount of adipic
acid or valeric acid which was decomposed.
A fourth test of this type was run to measure actual degradation as
well as quantify the products. This test was essentially the same as the
adipic acid version of the first three, with a few modifications. The main
difference was the use of pure oxygen as the headspace gas rather than air.
Thus, five times as much oxygen was available for reaction with the sul-
fite. In addition, air was allowed to enter the reaction vessel through a
water bubbler to replace the oxygen consumed in the reaction and maintain the
pressure within the vessel as near as possible to atmospheric pressure. At
the conclusion of the test run, again 35 hours at 50°C, the solution was
analyzed by 1C for adipic acid, then extracted and analyzed by GC. Ana-
lysis of the headspace gas was done in the same manner as in the previous
tests. A material balance for this test is presented in Table 7-4. Butane
is apparently not a favored product under the conditions encountered in this
test and represents only a small fraction of the products of the decomposi-
tion reaction. In addition to detection by GC analysis, however, its
presence was confirmed by GC-MS analysis of the head space gas.
TABLE 7-4. RESULTS OF SEALED FLASK EXPERIMENT WITH MATERIAL BALANCE
Valeric
Adipic Acid Adipic' Acid Acid Butane Final
Sample (Initial) (Final) (Final) (Final) Total % Initial1 Gas
Adipic Acid (ppm)
Solution +
CaS03'%H20 (nunole)
1400
0.4S
1140
0.39
120
0.06
30
5.6 x 10~" 0.45 94%
02
02
1 Percent of original material accounted for by analyses at conclusion of test.
7.2 MECHANISMS TESTS
The screening tests described in the previous section were designed to
yield information pertaining to specific aspects of the degradation process.
Based on this information, a series of mechanism tests was conducted, aimed
at providing data for evaluation of the interaction of the various parameters
and their relationship to one another. The tests were designed to reflect
actual scrubber conditions and to allow a material balance to be done on
a flow through system.
The apparatus used in the first flow-through laboratory test incorpora-
ted a sorbent trap for collection of butane and other low molecular weight
hydrocarbons, preceded by a sodium hydroxide impinger for collection of
CC>2 generated in the decarboxylation reaction and any valeric acid which
might be in the gas phase. The reaction vessel was preceded by an impinger
of deionized water for saturation of the oxygen stream with moisture to mini-
64
-------
mize evaporation of the adipic solution, and both were immersed in a water
bath maintained at 50°C. The reaction vessel was magnetically stirred,
and contained 15 grams of CaS03'%H20 in 300 ml of adipic-acid solution.
After addition of the calcium sulfite, the solution pH was adjusted to 6.0.
The apparatus is diagrammed in Figure 7-2.
The sorbent trap consisted of a 2' x 3/8" i.d. stainless steel tube
packed with Porapak Q 100/120. The trap was cooled to 7°C during the run and
was de-sorbed in the following manner: A two liter flask was evacuated and
attached to the inlet of the trap. A bag of nitrogen was attached to the out-
let of the trap. The flask was opened and the trap was back-flushed with
about two volumes of nitrogen to remove entrained oxygen. The trap was then
heated to 180°C and desorbed into the flask at a rate of about 50 ml/min until
the flask was full. The gas in the flask was then analyzed on the GC. At
least 28 liters of purge gas was necessary in order to reduce the concentra-
tions of the compounds being desorbed from the column to 10% of their con-
centration in the first two liters of purge gas. Calculations had indicated
that two liters of purge gas would be sufficient to desorb the sample that
had been trapped. Subsequent testing indicated that this sorbant material
bleeds low molecula'r weight hydrocarbons when heated. This problem is pos-
sibly due to reaction of the column material with residual oxygen that
cannot be efficiently flushed from the trap.
Samples of the solids and liquid in the reaction vessel were obtained at
intervals during the test run, as well as at the start and conclusion of the
run. Initial, intermediate and final adipate concentrations were determined
by 1C analysis. The adipic acid solution was extracted and analyzed by GC
at the conclusion of the test, as was the NaOH solution preceding the sor-
b.e.nt trap. Samples of the NaOH solution were also analyzed for COa . The
analytical results for this test are presented in Table 7-5.
Carbon balance calculations carried out using the data in Table 7-5
are shown below:
TABLE 7-6. CABBON BALANCE FOR MECHANISM TESTS USING PORAPAK Q TRAP
INITIAL FINAL
Adipic Acid Adipic Acid Carbon Dioxide Valeric Acid Hydrocarbons Total
(mmoles C) Cmmoles C) (maoles C) fomoles C) (mmoles C) (mmoles C) % Initial
18.7 11.4 1.4 3.7 3.2 19.7 105%
If the hydrocarbon values are rejected and the actual amount of
carbon as butane, propane, ethane and methane is assumed to be negligible
the carbon balance looks like this:
65
-------
02 IN
ON
C?
KJ,
'LI
50"C
DI
- THERMOMETER
SAMPLE PORT
r
J
z*c v_
REACTION FLASK,
STIRRED
NaOH SOLUTION
O O
5°C
FLOW METER
V
SORBANT TRAP
COOLING COIL
Figure 7-2. Laboratory apparatus for mechanism test with Porapak Q sorbent trap.
-------
TABLE 7-5. RESULTS OF MECHANISM TEST WITH PORAPAK Q SORBENT TRAP
ppm Adlplc mmoles Antmolcs Solids mmoles Aim»oJ.t*s SO3=
Sample Acid Adlplc Acid from start A% % SO 3= S03v from start
Initial 1520 3.12 79 92
6 hrs 1200 2.46 .66 21.0 34.8 39 53
24 hrs 980 2.00 1.12 35.5 17.3 19 73
40 hrs 930 1.90 1.13 38.8 4.7 5 87
ramol.es nunoles mmolcs dAd /9t
Valeric Acid Hydrocarbons COz 3s03~/3t
0.012
0.023
0.74 'Ci.0.23 1.4 0.001
C30.07
C20.20
Cjl.65
0.013
'Subsequent work Indicates, these values are in error due to hydrocarbon bleed from sorbent material.
-------
TABLE 7-7. CARBON BALANCE FOR MECHANISM TEST USING PORAPAK Q TRAP;
HYDROCARBON VALUES NEGLECTED
INITIAL
Adipic Acid
18.7
Adipic Acid
(mmoles C)
11.4
Carbon Dioxide
1.4
FINAL
Valeric Acid
/ « t+\
3.7
Hydrocarbons
(mnoles C)
Total
(mnw>les C)
16.5
Z Initial
88%
Since the use of the Porapak sorbant trap failed to provide a reliable
means of trapping and concentrating hydrocarbon byproducts of the degrada-
tion process, a second flow through test was done without the sorbant trap.
In order to concentrate the gas phase reaction products, a closed loop
system was used whereby the sparge gas was recirculated by a peristaltic
pump. The system was flushed with oxygen initially and additional oxygen
was bled into the system periodically. Two NaOH bubblers were used in series
to assure complete trapping of all C02. A gas mixing flask was placed in line
prior to the reaction vessel to provide a point for obtaining a homogeneous
gas sample. Two identical systems were set up and run simultaneously with the
only difference being the initial adipic acid concentrations. The reaction
vessel in system 01 contained 300 ml of 1550 ppm adipic acid solution and
17.25 grains of calcium sulfite while that in system 02 contained 300 ml of
770 ppm solution with 17.25 grams of calcium sulfite. The initial pH was
adjusted to 6.0 as in the previous test. The apparatus is diagrammed
in Figure 7-3 . A TOC analysis was done on the adipic acid solutions at the
conclusion of the test, in addition to the analyses described for the previous
test. The test results are presented in Table 7-8.
Carbon balance calculations carried out using the data in Table 7-8
are presented below-:
TABLE 7-9. CARBON BALANCE FOR MECHANISM TEST USING RECIRCULATED GASES
Initial Final :
Adipic Acid Adipic Acid Carbon Dioxide Valeric Acid Hydrocarbons Total
(mmoles C) (mmoles C) (mmoles C) (mmoles C) (mmoles C) (mmoles C) % Initial
19.0
9.5
11.5
4.1
3.2
1.1
0.5
nd
<0.1 15.2
nd 5.2
80%
552
nd - not detected
The small amount of butane detected in 01 followed by its disappear-
ance, coupled with the poor recovery and lack of degradation products in 02
suggest several possibilities:
68
-------
SAMPLE
PORT (GAS)
SAMPLE PORT
9> (LIQUID)
V
o
GAS MIXING FLASK
2°C
I
J
f'
\ f
1
1
REACTION FLASK, STIRRED
NaOII SOLUTION
Figure 7-3. Laboratory apparatus for mechanism test with recirculated gases.
-------
TABLE 7-8. RESULTS OF MECHANISM TEST WITH RECIRCULATED GASES
Sample
ppm Adipic
Acid
nunoles
Adipic Acid
A mmoles
from start
A%
Solids
% S03
mmoles
Ammolea SO 3=
from start
mmoles
Valeric Acid
pinoles nunoles
hydrocarbons CO2
mmo]es
Carbon
By TOC
3S03
01
Initial
5 hvs
17 lira
24 lirs
90 hrs
02
Initial
5 hrs
17 hcs
24 hrs
90 hrs
1550
1300
1190
1080
930
770
490
400
580
340
3.17
2.66
2.45
2.21
1.91
0.51
0.72
0.96
1.26
16.1
23.2
30.3
40.0
94.9
84.8
81.0
72.0
59.0
129
111
103
90
73
18
26
39
56
0.49
0
0.17
0.17
0.22
0.00
3.16
1.58
1.01
0.83
1.19
0.69
0.57
0.75
0.39
0.89
35.7
47.6
24.7
56.2
94.7
86.9
76.3
67.0
60.0
127
113
97
84
74
14
30
43
53
nd
nd
19.58
TOTAL
1.10 9.00
TOTAL
0.028
0.026
0.018
0.018
0.022
0.041
O.OU
-0.028
0.05
0.012
-------
the hydrocarbon products were further degraded by the
reaction,
the test set-up leaked, or
the hydrocarbons were adsorbed by the tubing and pump.
Since no significant amounts of low molecular weight hydrocarbons were
detected in the closed loop system, a return to the flow through system was
indicated. A diagram of this system can be seen in Figure 7-4.
The conditions of this test were identical to those in the first flow
through mechanism test with the following modifications:
two NaOH impingers were used in series in order to more
efficiently trap C02, and
two sorbent traps were used in series, one containing
silia gel and the other activated carbon.
The sorbent traps in this test consisted of 2' by 3/8" stainless steel
tubes, one packed with 35-60 mesh silica gel and the other packed with acti-
vated carbon. Laboratory tests indicated that the silica gel trap was effi-
cient at trapping n-butane but would not retain methane or ethane well. The
trap was cooled to 7°C during the run and then desorbed in the same manner as
the Porapak Q trap. A single two liter purge was found to be sufficient to
desorb all of the n-butane trapped on the silica gel trap. Desorption of
the carbon trap failed to yield any hydrocarbons.
The samples were obtained and analyzed in a manner identical to those in
the first flow-through mechanism test. The results of this test are tabula-
ted in Table 7-10. A carbon balance is shown below.
TABLE 7-11. CARBON BALANCE FOR MECHANISM TEST USING SILICA GEL AND ACTIVATED
CARBON TRAPS
Initial Final
Adipic Acid Adipic Acid Carbon Dioxide Valeric Acid Hydrocarbons Total
(mmoles C) (mrnolea C) (mrnoles C) (mmoles C) (mmoles C) (mmoles C) % Initial
19.2 12.6 1.0 0.3 <0.1 13.9 72%
This large discrepancy in the carbon balance indicates the possibility
that a degradation product that is not detectable by the sampling and analy-
tical techniques employed is being formed.
71
-------
02 IN
T-
THERMOMETER
q O
M
a O
50*C
SAMPLE
PORT
DI H20-
REACTION FLASK,
STIRRED
r
N,
£) O Q
w
J
I
1
\2°C /
SOLUTION
O O cb a
0
x»
c^\
SILICA GEL
TRAP
FLOW METER
- ACTIVATED
CARBON TRAP
Figure 7-4. Laboratory apparatus for mechanism test with silica gel and
carbon traps.
-------
TABLE 7-10. RESULTS OF MECHANISM TEST WITH SILICA GEL AND ACTIVATED CARBON TRAPS
ppm Adiplc mmoles A mmoles Solids mmoles Aramoles S0a=
Sample Acid Adiplc Acid from start A% % S03= S03= from start
Initial 1650 3.2 92.3 120
6 hrs 1320 2.7 0.5 15.5 78.8 100 20
24 hrs 1020 2.1 1.1 34.2 61.6 77 43
mmoles
mmoles p moles mmoles Carbon 3Ad /9t
Valeric Acid hydrocarbons C02 By TOC 3S03=/Dt
_
0.83 0.61 0.025
n-butane
0.4
methane
0.3 4.17 0.95 16.25 0.026
n-butane
0.6
methane 1.56
total total
n-butane
= 5.0
total
methane
- 1.0
-------
BIBLIOGRAPHY
Anderson, J. M. and J. K. Kochi. 1970. J. Am. Chem. Soc., 92:2450.
Berezin, I. V., E. T. Denisov, and N. M. Emanuel. 1964. The Oxidation of
Cyclohexane. Pergamon, New York.
Borgwardt, Robert H. November 1978 Progress Report. Limestone Scrubbing
of SOa at EPA/RTP Pilot Plant - Progress Report 37.
Bunce, N. J. 1972. J. Org. Chem.. 37:664, 1972.
Chen, T., and C. H. Barren. 1972. Ind. Eng. Chem. Fundam., 11:466.
Christensen, Herbert E. and Edward J. Fairchild, eds. 1976. Registry of
Toxic Effects of Chemical Substances. NIOSH, Public Health Service,
Rockville, Maryland.
Clifford, A. A. and W. A. Waters. 1965. J. Chem. Soc., 2796.
Denisov, E. T., N. I. Mitskevich, and V. E. Agabekov. 1977. Liquid-phase
Oxidation of Oxygen-containing Compounds. Consultants Bureau, New York.
Dias, F. F. and M. Alexander. 1971. Appl. Microbiol., 22:1114
Grula, M. M.; Grula, E. A. 1976. BERC-RI-76-6, 61 pp.
Hammersma, Warren J. 20 October 1978. Personal communication to Mr.
Ronald A. Venezia, EPA.
Hammond, M. W.; Alexander, M. 1972. Environ. Sci. Technol., 6(8):732-735.
International Technical Information Institute. 1976. Toxic and Hazardous
Industrial Chemicals Safety Manual for Handling and Disposal with
Toxicity and Hazard Data. Tokyo.
Keith, C. L., et al. 1978. Arch Microbiol., 118 (2):173-176.
Kiseleva, R. A. and M. S. Dudkin. 1967. Zh. Prik, Khim., 40:2513.
Kraeutler, B.; Bard, A. J. 1978. J. Am. Chem. Soc., 100(19):598-2-92.
Kuhnle, J. A., R. E. Lundin, and A. C. Waiss. 1972. J.C.S. Che. Com., 287.
74
-------
BIBLIOGRAPHY (continued)
Leavitt, C., et al. 1978. Environmental Assessment of Coal- and Oil-Firing
in a Controlled Industrial Boiler; Volume III. Comprehensive Assess-
ment and Appendices. EPA-600/7-78-164c. Environmental Protection
Agency, Research Triangle Park, North Carolina.
Messier, F.; DeJongh, D. C. 1977. Can. J. Chem., 55:2732-2740.
Pptai, S. 1969. The Chemistry of Carboxylic Acids and Esters. Interscience
New York.
Rochelle, G. 1977. The Effect of Additives on Mass Transfer in CaC03 and
CaO Slurry Scrubbing of S02 from Waste Gases, Ind. Eng., 16:67-75.
Sax, N. Irving. 1975. Dangerous Properties of Industrial Materials, 4th ed.
Van Nostrand-Reinhold, New York, New York.
Sheldon, R. A., and J. K. Kochi. 1972. In: Organic Reactions, W. G. Danben,
ed.
Tinker, H. B. 1970. J. Catalysis, 19:237.
75
-------
APPENDIX A
ANALYTICAL METHODS
77
-------
GAS CHROMATOGRAPHY
Samples were analyzed for adipic, valeric and glutaric acids on a Tracor
Model 560 gas chromatograph equipped with a flame ionization detector. Peak
areas were integrated with either a Hewlett-Packard Model 3380 A Integrator or
a Spectra Physics Model 4000.
Conditions for the analysis are listed below:
Column: 6' x 2mm i.d. glass, packed with Tenax GC, 60/80 mesh
Carrier Gas, Flow: Nz, 25 ml/min
Oven Program:
Initial: 100°C, hold 3 minutes
Rate: 10°C/min
Final: 300°C, hold 5 minutes
Injector Temperature: 250 C
Detector Temperature: 300 C
Samples were analyzed for low molecular weight hydrocarbons on an
Analytical Instrument Development Inc. Model 511 .Gas Chromatograph. Peaks
were recorded on a strip chart recorder and concentrations were determined
by peak height. Conditions for the analyses are listed below.
Column: 6' x 1/8" i.d. stainless steel, packed with Chromosorb 102,
80/100 mesh
Carrier Gas, Flow = NZ, 20 ml/min
Oven Temperature: 125 C isothermal
ANALYSIS FOR ADIPIC ACID BY ION CHROMATOGRAPHY
A Dionex Model 14 Ion Chromatograph was used for adipic acid analyses.
INSTRUMENT CONDITIONS
Standard Dionex anion separator (3 x 500 mm) and suppressor columns
were used. The eluent (see below) was 2.0 mm NaHCOs and 1.6 mM Na2COa
for most analyses. It-was made up in 3 liter batches from stock solutions
as needed. The flow rate was set at 30%. Chloride, adipic acid, sulfite,
and sulfate may all be resolved and quantified under these conditions.
Table A-l details the response characteristics of these four species with
the most commonly used eluent.
TABLE A-l. RESPONSE CHARACTERISTICS OF SPECIES ANALYZED BY DIONEX
Anion Retention Time Relative Response Working Range
Cl~
Adipate
SO 3=
30!+=
3.2 min.
6.8 min.
8.4 min.
14.3 min.
15 ymho/mMole
1 ymho/mM
4.5 ymho/mM
3.8 umho/mM
0.005-5mM
0.01-2mM
0.03-6mM
0 . 03-6:mM
78
-------
TABLE A-2. 1C ELUENTS
Eluent # mM NaHCO 3 ' mM Na2CO 3
1
2
3
4
5
6
3.60
3.00
2.40
1.90
2.40
2.00
3.00
2.40
1.80
1.50
1.90
1.60
PREPARATION OF REAGENTS AND STANDARDS
A. Eluent Stock Solution
Weigh Out
25.4g of Na2C03 and 25.2g of NaHCO3 and dissolve in 500 ml of deionized
water in a volumetric flask.
B. Eluent Solution (2mM NaHC03, 1.6 mM Na2Ca3)
Pipette 10 mis. of eluent stock solution into a standard Dionex defla-
table eluent container and dilute to 3 liters.
C. Adipic Acid Stock Solution
Weigh 1.46 g of desicated adipic acid to the nearest .01 mg. Place in
a 1 liter volumetric flask and add 700 mis of deionized water. Allow
the solid to dissolve. Mild heating and swirling may be necessary.
Allow the flask to cool to room temperature and dilute to the mark.
Working standards may be prepared by suitably diluting this stock solution.
PROCEDURE
1. The ion chromatograph is initially allowed to stabilize under the analysis
conditions (30% flow, with analytical eluent, etc.) until a stable base
line is obtained.
2. Several standards over the range of interest are analyzed and a calibra-
tion curve obtained, (plotting peak height vs. concentration.) In
practice, 0.1 mM to 0.6 mM adipic acid is a convenient range.
3. Samples are analyzed for 2% to 3% hours.
79
-------
4. One or two "check" standards- and a quality control check are analyzed
to see that the machine response is not drifting, and is still functioning
properly.
5. Samples are again analyzed for several hours.
6. Standards are again run and a calibration curve obtained at the end of
the day.
LABORATORY GENERATION OF CaS03 %H20
CaS03 %H20 solids, used to simulate scrubber solids, were synthesized
by titrating a 0.77 M CaCl2 solution with a 0.77 M solution of Na2S03.
Na2S03 was used as the titrant to minimize oxidation of sulfite solids to a
fulfate phase. Following the titration, the solution was vacuum filtered and
dried at 50 C. The resulting solids were then pulverized with a mortar and
pestle and stored in a sealed container until used in the bench scale tests.
Using this technique, approximately 90% ±5% pure CaS03 %H20 solids could be
generated.
SULFITE ANALYSES
Equipment
1. 50 ml buret
2. Magnetic stirrer
3. Pipets (10ml); bulb pipet fillers; and 100 ml graduated cylinder
4. 150 ml beakers
5. 4-ounce small mouth glass bottles
6. microammeter apparatus:
a. 2 1-cm square platinum-sheet electrodes mounted about 1 cm apart
b. 1.5 volt dry cell battery
c. electrometer or microammeter 0-25 yamp
d. voltage divider
7. Laboratory analytical balance accurate to 0.05 mg
REAGENTS & SOLUTIONS
1. Sodium Arsenite Stock Solution (M3.010 mole/£)
Dissolve 1.3 grams of NaAsOa into one liter of D.I. water. Standardize
this solution using reagent grade 0.1 N standard iodine.
, T .. , . .. , (ml la) (normality of I2)
(molarxty of NaAs02) = 2 (ml NaAsQ2) - L
2. Preparation of a 0.1 Normal Iodine Solution
Dissolve 25 grams of potassium iodide (KI) in as little D.I. water as
possible. Accurately weigh 12.69 grams of -resublimed iodine and dissolve it
completely in the saturated potassium iodide solution. When the dissolution of
80
-------
of the iodine is completed, dilute with D.I. water accurately to one liter.
A larger solution can be made up at one time, if desired. Iodine should be
kept in a well stoppered, dark colored bottle, in a cool place. It need not
be standardized if only used for the sulfite analysis since a blank is also
run.
3. pH 6 Buffer
This solution contains 1 mole/5, NaAc and 0.05 mole/JlHAc. Pipet
2.9 ml of glacial acetic acid into 500 ml of D.I. water. Stir in 82 grams
of NaAc and when completely dissolved make up to one liter. It is convenient
to prepare several liters of this solution at a time since 50 ml is used for
each analysis.
PROCEDURE
1. Weigh 50 mg,of dry solid .sample to the nearest 0.1 mg.
2. Pour 100 ml of pH 6 buffer solution into a 4 oz glass bottle with cap, and
and stirring bar.
3. Pipet 10 ml of 0.1 EL standard iodine solution, into the bottle, cap,, and
begin stirring.
4. Add the weighed solid sample and recap.
5. Stir for 15 minutes.
6. Empty bottle quantitatively into a 150 ml beaker.
7. Place the platinum electrodes in the solution, stir and connect the
microammeter. The current should be 30-50 microamps.
8. Begin the titration with the 0.11 mole/liter sodium arsenite solution
using the 50 ml buret. The color of the solution serves as a rough
indicator of the state of the titration. The iodine color (red) changes
to yellow 5 ml before the endpoint is reached. When the solution turns
light yellow, the titration should be carried out dropwise. The solution
will become colorless one to two drops before the endpoint. The titration
is completed when the current reaches 3-4 microamp.
9. After the titration is completed, record the volume of sodium arsenite used
and rinse the electrodes with with D.I. water after removing them from the
solution.
10. Blanks are to be run each day using th'e above procedure without the
addition of sample.
CALCULATIONS
The concentration (mmole/gram) of sulfite present in the solid sample
can be determined by use of the following equation:
_ (B-S)M
C ~ W
where,
C = concentration of sulfite(in mmoles/g)
B = volume of arsenite used to titrate the blank, in ml
S = Volume of arsenite used to titrate the sample in ml
M = concentration of the sodium arsenite (moles/A)
W = weight of sample used, in grams
81
-------
APPENDIX B
SAMPLE CHROMATOGRAMS
83
-------
Figure B-l. Sample 1C chromatogram with high sulfite.
84
-------
rhd±i±zii±t±t±i
! I ! , ' ' I : i i i r i
Figure B-2. Sample 1C chromatogram with no sulfite.
85
-------
61 Contaminant
19. 78 Contaminant
12. 82 Valeric Acid
17 30 Succinic Acid
9 59 Glutaric Acid
i. 2
. ___ Ad Ip In Af
'STOP
F:T TVPE
£0
76 V
9 8 !"l
± 2. 0 * M
5 *?1 T
1 0 7 0 T
12. 82 Tfl
17 30
13. 53
21. 01 Pi
HP 33S0R
DLY OFF
{1 V ,-' it 3S
flREfi ;
HREH
1311
1553
195395
99999993
57773
2 6 9 i 5 1
9432 4
15341
73395
827002
STOP 45
rtTTM 32
081 291
001 535
193
38. 43
0 56 9
2S5 3
092 9
. 015 7
. k!72 Zf
814 5
ft E J t L: f i 0 0 0
Figure B-3. Sample GC chromatogram.
86
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-224
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Adipic Acid Degradation Mechanism in Aqueous FGD Systems
5. REPORT DATE
September 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
F.B. Meserole, D.L. Lewis, A.W. Nichols, and
G. Rochelle
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
8500 Shoal Creek Blvd.
Austin, TX 78766
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
-- 68-02-2608, "Task 58
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 10/78 - 4/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES
919/541-2336.
IERL-RTP project officer is Robert H. Borgwardt, Mail Drop 65,
16. ABSTRACT _, . .. ^JT.-IJJ-II j JT i j
The report gives results of a field and laboratory study or the adipic
acid degradation mechanism in aqueous flue gas desulfurization (FGD) systems.
(Adding adipic acid to limestone-based, S02 wet scrubbers increases S02 removal
and limestone utilization. However, as much as 80% of the adipic acid added to
some systems is lost, supposedly through degradation.) The degradation is
associated with the oxidation of sulfite, possibly through a free radical mechanism.
At least one mechanism is an oxidative decarboxylation yielding valeric acid,
butyric acid, glutaric acid, and C02. The quantities of products measured during
laboratory testing only account for approximately 30% of the adipic acid degraded.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Air Pollution
Adipic Acid
Degradation
Limestone
Flue Gases
Scrubbers
Desulfurization
Sulfur Dioxide
Organic Acids
Buffers (Chemistry)
Additives
Air Pollution Control
Stationary Sources
Buffer Additives
13B
07 C
14B
08G
21B
07A
07D
07B
11G
IS^DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
93
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
87
------- |