~L
`.~~s;,Z~ERING
`
`':wASH..
`CHEMISTRY
`PRODUCT RESEARCH
`AND DEVELOPMENT
`
`78
`
`IEPRA6 17(1) 1-92 (1978)
`ISSN 0019-7890
`
`VOLUME 17 NUMBER 1
`
`MARCH 1978
`
`/,
`
`ODUCT
`ESEARCH
`AND
`DEVELOPMENT
`
`GE-1022.001
`
`

`
`
`
`
`
`
`
`GE-1022.002
`
`GE-1022.002
`
`

`
`GE-1022.003
`
`

`
`INDUSTRIAL &
`ENGINEERING CHEMISTRY
`PRODUCT RESEARCH
`AND DEVELOPMENT
`
`Editor: HOWARD L. GERHART
`Associate Editor: Jerome A. Seiner
`Managing Editor: Katherine I. Biggs
`Advisory Board: David Gushee, Vladimir
`James D.
`Idol,
`James E.
`Haensel,
`McEvoy, Norbert Platzer, and Robert G.
`Squires
`
`Published by th e
`AMER ICAN CHEM ICAL SOCIETY
`
`BOO KS AND JOURNALS DI VI S IO N
`
`D. H. Michael Bowen, Director; Marjorie
`Laflin, Assistant to tbe Director
`Editorial Department: Charles R. Bertsch,
`Head: Marianne C. Brogan, Associate
`Head; Edward A. Borger, Associate
`Editor
`Magazine and Production Department: Bacil
`Guiley, Head
`Researc h and Development Department:
`Seldon W. Terrant, Head
`© Copyright, 1978, by the American Chem(cid:173)
`ical Society. P e rmissio n of the American
`Chemical Society is granted for libraries and
`other users to make reprographic copies for
`use beyond that permitted by Sections 107 or
`108 of the U.S. Copyright Law, provided that,
`for all articles bearing an article code, the
`copying organization pay the stated per-copy
`fee through the Copyright Clearance Center,
`Inc. For further information write to Office
`of the Director, Books and Journals Division,
`at the ACS Washington address.
`Published quarterly in March, June, Sep(cid:173)
`tember, and December by the American
`Chemical Society at 20th and Northampton
`Sts., Easton, Pennsylvania 18042. Second
`
`class postage paid at Washington. D.C. and
`at additional mailing offices.
`
`Ed itor ia l Informa ti on
`G UIDE FOR AUTHORS, published in
`the first issu e of th e yea r, ~i v es co py re(cid:173)
`quirements for manuscripts.
`Manuscripts for publication should be
`submitted in quadruplicate to
`lanaging
`Editor, I&EC Product Research and Devel(cid:173)
`opment, at the ACS Washington address.
`Cor respondence rega rding acce pte d pa(cid:173)
`pers and proofs shou ld be di rected to the
`Ed itorial Depart ment at the add ress below.
`Page charges of $60.00 pe r page may be
`paid for papers published in this journal.
`Payment does not affect acceptance or
`scheduling of papers.
`Bulk reprints or photocopies of indi(cid:173)
`vidual articles are available. For information
`write to Business Operations, Books and
`Journals Division, at the ACS Washington
`address.
`The American Chemical Society and its
`Editors assume no responsibility for
`the
`statements and opinions advanced by con(cid:173)
`tributors. Views expres. ed in the editorials
`are those of the authors and do not neces(cid:173)
`sarily represent the official position of the
`society.
`
`S u bscr ipti on and Business Information
`1978 Subscription prices, printed or mi(cid:173)
`crofiche, including postage. Microfiche by air
`mail; printed by surface mail. Printed edition
`air mail or air freight rates available from
`Membership & Subscription Services at the
`address below.
`
`Members
`Nonmembers
`
`U.S.
`$8.00
`32.00
`
`Foreign
`$11.00
`35.00
`
`mailing label. Send all address changes to
`Membership & Subscription Services . Please
`allow six weeks for change to become effec(cid:173)
`tive. Cla ims for missing numbers will not
`be allowed if loss was due to failu re of notice
`of change of address to be received in the time
`specified; if claim is dated (a) North America:
`more than 90 days beyond issue date, (b) all
`other foreign: more than one year beyond
`issue date; or if the reason given is "missing
`from files." Hard copy claims are handled by
`Membership & Subscription Services.
`Mi crofilm editions of all ACS primary
`publ icatio ns, by single volume or entire back
`issue collection, are availahle. For additional
`information,
`microfilm (and microfiche)
`contact Microforms Program at
`the ACS
`Washington address. or call (202) 872-4554.
`
`To order single issu es or ba ck volu mes,
`printed or microfiche, contact Special Issues
`Sales at the ACS Washington address, or call
`(202) 872-4:~65. Cur rent year single issue
`$10.00. Prior yea r sing le issue $12.00. Back
`volume $46.00. Foreign postage additional.
`
`S u pple me n ta ry material mentioned in
`the journal appears in the microfilm edition.
`Papers containing supplementary material
`are noted in the Table of Contents with a • .
`See Supplementary Material notice at end of
`article for number of pages . Orders over 20
`pages are available only on 24X microfiche.
`Orders must state photocopy or microfiche.
`Full bibliographic citation including names
`of all authors and prepayment are required.
`Prices are su bject to change.
`
`Microfiche
`Photocopy
`1-8 pages
`9-20 pages
`
`U.S.
`$3.00
`
`5.50
`6.50
`
`Foreign
`4.00
`
`7.00
`8.00
`
`New and r en ewa l subscriptio ns should be
`sent with payment to the Office of the Con (cid:173)
`the ACS Washington addre. s.
`troller at
`Changes of a dd ress must include both old
`and new addresses with ZIP code and a recent
`
`Single microfiche or paper copies of Supple(cid:173)
`mentary Material may be ordered from
`Business Operations, Books and -lournals
`Division at the ACS Washington addre. s, or
`call (202) 872-4559.
`
`American Chemical Society
`.W.
`1155 16th Street,
`Washington. D.C. 20036
`(202) 872-4600
`
`Editorial Department
`American Chemical Society
`P.O. Box 3330
`Columbus, Ohio 43210
`(614) 421-6940, Ext. 3171
`
`Membership & Subscription Service.'
`American Chemical Society
`P.O. Box 3337
`Columbus, Ohio 43210
`(614) 421-72:1O
`
`, ....
`, .
`
`GE-1022.004
`
`

`
`86
`
`Ind. Eng. Chern. Prod. Res. Dev., Vol. 17. No.1, 1978
`
`Deposit Formation from Deoxygenated Hydrocarbons. 3.
`Effects of Trace Nitrogen and Oxygen Compounds
`
`William F. Taylor " and John W. Frankenfeld
`
`Exxon Research and Engineering Company , Government Research Laboratories. Linden . New Jersey 07036
`
`The effect on deposit formation rate of the presence of trac e amo unts of nitrogen and oxygen containing impuri(cid:173)
`ties in deoxygenated JP-5 was investigated. The change in deposit format ion rate , following the addition of repre(cid:173)
`sentative nitrogen and oxygen com pounds, was dete rmined over a temperature range of 150-450 DC in fuels
`with molecular oxygen contents reduced to less than 1 ppm. The addition of nitroge n com pounds as pure materi(cid:173)
`als did not increase deposit formation over the temperature range studied. However, certain nitrogen compounds
`led to sludge formation at temperatures in the range of 20-25 DC. Of the oxygen compounds studied, peroxides
`as a class were found to be highlydeleterious to fuel stability. Some acids, esters, and ketones were moderately
`deleterious while othe rs had no significant effect on de posit formation. In general. cyc loalkyl compounds were
`less harmful than their aliphatic or aromatic counterparts. Several interacti ons betwee n trace impurities were
`discovered which affect deposit formation rates.
`
`Introduction
`The deposit formation tendencies of jet fuel range hydro(cid:173)
`carbons have been the subject of considerable research (Nixon ,
`1962). Initial work was carried out with air -saturated hydro(cid:173)
`carbons in a narrow, near ambient temperature range in order
`to investigate storage stability characteristics. Subsequent
`studies were extended to higher temperatures in order to in(cid:173)
`vestigate the stability of such fuels when used in high-speed
`supersonic aircraft (Nixon, 1962; Churchill, 1966). Such
`studies were carried out mainly with fuels saturated with
`molecular oxygen via exposure to air although some limited
`work has been reported with reduced oxygen containing fuels
`(Nixon and Henderson, 1966;Taylor and Wallace, 1967). This
`laboratory studied the variables which control the kinetics of
`the deposit formation process in air-saturated jet fuels at
`temperatures up to 250°C at reduced pressures (Taylor,
`1968a,b, 1969a,b; Taylor and Wallace, 1967, 1968). Also,
`studies were carried out using tetralin as a model jet fuel range
`hydrocarbon (Taylor, 1970a,b, 1972). We are now extending
`our study of the kinetics of deposit formation to deoxygenated
`jet fuels (i.e., fuels in which the molecular oxygen content has
`been drastically reduced). The range of conditions investi(cid:173)
`gated was extended to include temperatures up to 649°C and
`pressures up to 69 atm.
`Previously (Taylor, 1974), the effects of deoxygenation on
`deposit formation in jet fuels were described. With most fuels,
`removal of molecular oxygen ma rkedly lowered the rate of
`deposit formation. However, certai n poor quality fuels showed
`less improvement. This surprising result led to a study of the
`influence of trace impurities on deposit formation in deoxy (cid:173)
`genated fuels, in order to determine whether such impurities,
`likely present in poor quality fuels , were negating the bene(cid:173)
`ficial effects of.deoxygenation. The effect of sulfur compounds
`was reported in part 2 of this series (Taylor, 1976). This paper
`describes the effects of trace impurities which contain one or
`more atoms of nitrogen or oxygen. Subsequent work will dis(cid:173)
`cuss the effects of hydrocarbon types on deposit formation in
`deoxygenated pure compound systems.
`
`Experimental Section
`Ap paratus. A schematic of the Advanced Kinetic Unit used
`to measure the rate of deposit formation was shown previously
`(Taylor, 1974). The molecular oxygen content of the fuel to
`be tested was adjusted in a fuel treatment vessel by sparging
`
`the fuel at atmospheric pressure using helium. The treated
`fuel was passed through an oxygen sensor cell and delivered
`to a double piston fuel delivery cylinder. The oxygen sens or
`cell contained a polarographic sensor and the oxygen content
`of the total fuel was monitored. Oxygen analyses were also
`made on selected samples using a thermal conductivity gas
`chromatographic analyzer. The fuel was delivered to the unit
`by means of high-pressure nitrogen. The treated fuel was
`separated from the nitrogen drive gas by use of two individual
`pistons, separated by a small water layer. The fuel then passed
`through a heated tubular reactor section consisting of a 1/4-in.
`o.d., 0.083-in. wall S.S. 304 tube which was contained inside
`of four individually controlled heaters. Each heater zone was
`approximately 12 in. long and was controlled by a proportional
`temperature controller. Unit pressure was controlled by means
`of a MITY-MITE controller.
`The rate of deposit formation was measured after a 4-h run .
`The reactor tube was cut into 16 sections, each 3 in. long, (4
`sections per reaction zone) and the tube sections were ana(cid:173)
`lyzed for carbonaceous deposits using a modified LECO low
`carbon analyzer system (Taylor, 1974). The analytical system
`was calibrated against known standards. The deposit forma (cid:173)
`tion rate was obtained by dividing the net carbon production
`per section by the corresponding inner surface area and ex(cid:173)
`pressed as micrograms of carbon per square centimeter per
`four -hour reaction time.
`Reagents. The jet fuel employed was JP-5 (MIL-T-5624H)
`whose properties were previously reported (Fuel A in Table
`I; Taylor, 1974).This fuel was a highly refined, stable material.
`It contained 234 ppm of total sulfur and less than 1 ppm of
`thiol sulfur. The nitrogen content was less than 1 ppm and
`only traces of peroxides were present. This fuel was obtained
`additive free from the Baton Rouge Refinery of Exxon Com(cid:173)
`pany, U.S.A.
`The tubing employed in the reactor section of the Advanced
`Kinetic Unit was If4-in. o.d. X 0.083-in. wall stainless steel type
`304. P rior to use, the tubing was cleaned inside and out with
`reagent grade acetone and chloroform and blown dry with
`nitrogen.
`Pure organic compounds containing nitrogen and oxygen
`of the highest quality available were obtained commercially
`and employed as received. The nitrogen and oxygen com(cid:173)
`pounds were added to the base fuel at the 100 ppm of Nand
`o level. This level was chosen as representat ive of probable
`maximums for jet fuels derived from petroleum.
`
`0019-7890178/1217-0086$01.00/0
`
`© 1978 American Chemical Society
`
`GE-1022.005
`
`

`
`Table I. The Effect of Individual Nitrogen Compounds on Total Deposit Formation in a Deoxygenated Jet Fuel
`
`Ind. Eng. Chern. Prod. Res. Dev., Vol. 17, No.1, 1978 87
`
`Pyrrole
`
`Pyridine
`
`Primary amine
`
`Miscellaneous
`
`Class of added
`nitrogen compound a
`
`Added compound
`
`ug
`of carbon
`
`Total deposits"
`As ppm
`based on fuel
`
`Oxygen
`content,
`ppm of O2
`0.68
`1310
`0.3
`2,5-Dimethylpyrrole
`0.68
`1316
`0.2
`Benzo(b)pyrrole (indole)
`0.54
`1028
`0.2
`Dibenzopyrrole (carbazole)
`1.02
`1977
`0.2
`2,4,6-Trimethylpyridine
`0.75
`1457
`0.2
`Benzo(b)pyridine (quinoline)
`0.69
`1330
`0.1
`2-Methylquinoline
`0.75
`1441
`0.2
`2,6-Dimethylaniline
`0.63
`1228
`0.3
`Hexylamine
`N -Methylcyclohexylamine
`0.73
`1411
`0.2
`0.54
`1049
`0.1
`2-Methylpiperidine
`0.71
`1380
`0.2
`Decahydroquinoline
`0.96
`1844
`1
`Hexanamide
`0.77
`1485
`0.4
`Base fuel
`a All compounds added to the 100 ppm N level. b Cumulative deposits produced in 4 h in the Advanced Kinetic Unit. Conditions:
`69 atm, zone 1,371 °C; zone 2, 427°C; zone 3, 482°C, zone 4, 538°C.
`
`1 0 0 0 r - - - , - , - - r . - - - - - . . - - - r - - , - - , - - ,
`
`1000 r - --,--,--
`I
`
`r.-- - , . - - - r - - , - - - . - - - ,
`I
`I
`I
`I
`
`~0
`
`~ ~ ~'
`
`i'
`N"-
`E5 100
`
`~0
`
`~0
`
`~ ~'
`
`i'
`N"-
`E5 100
`
`~0
`
`NO ADDED AU''''ES
`
`.~'..' .
`.......... ~. ...,
`:
`"I.
`, .-. ~ .
`
`.
`
`.~.
`
`10
`
`~ W~.
`
`,
`
`~ Z0e~ ~!
`
`5.0
`1.1 0
`
`I' 000"
`1 . 20
`
`19 00"
`1 .30
`
`I 600"
`•l OOO/ · K
`1." 0
`Figure 2. Deoxygenated fuelat 69atm (0.2 ppm of O2): e, withadded
`2,6-dimethylaniline; _ , with added N -methylcyclohexyl amine. All
`compounds added at 100ppm of N.
`
`700"
`
`I
`1 .50
`
`I
`1 .60
`
`1. 70
`
`NO ADDED AMINES
`
`! ...-~
`. ;:. ~..----.,-- ~. .
`- ...~,,~ --
`~ _#~. _.--.! •..!.-' •...I.--.
`'>:
`:
`
`1.70
`
`10
`
`5.0
`1.10
`
`~.,
`:.,
`
`~ Z0;
`
`~ ~V
`
`i
`~
`
`looor K
`Figure 1.Deoxygenated fuelat 69atm (0.2 ppm of O2): e,with added
`indole; A, with added 2,5-dimethylpyrrole; _ , with added carbazole.
`All compoundsadded at 100ppm of N.
`
`Results
`The source of petroleum is believed to be the remains of
`marine animal and vegetable life deposited with sediment in
`coastal waters (Hodgson, 1971). Bacterial action evolved
`sulfur, oxygen, and nitrogen as volatile compounds which were
`never completely eliminated despite the ever
`increasing
`pressure of sediment. Hence, crude oil is a mixture of hydro(cid:173)
`carbons containing varying quantities of sulfur, nitrogen, and
`oxygen compounds. The nitrogen content of crude oil ranges
`from practically zero to a few percent (Ball, 1962). Nitrogen
`compounds identified in jet fuel range petroleum cuts include
`pyrroles, indoles, carbazoles, pyridines, quinolines, tetrahy(cid:173)
`droquinolines, anilines, and amides (Sauer et al., 1952; Hen(cid:173)
`drickson, 1959; Nixon and Thorpe, 1962). The most prevalent
`compound types are pyrroles, indoles, carbazoles, and quin(cid:173)
`olines. Few reliable analyses of the oxygen content of petro(cid:173)
`leum products are available. However, oxygenated compounds
`are more abundant than nitrogenous species and somewhat
`less abundant than sulfur compounds. Oxygen compounds
`identified in jet fuel ra nge petroleum fraction are carboxylic
`acids, phenols, furans, ketones, alcohols, esters, amides, hy(cid:173)
`droperoxides, and pe roxides (Hendrickson, 1959; Nixon and
`Thorpe, 1962). Representatives of the above classes of nitro-
`
`gen and oxygen compounds were tested in this study at levels
`of 100 ppm of N or lOa ppm ofO. In addition, interactive ef(cid:173)
`fects, that is, the combined influences of different types of
`trace impurities, were examined to see if such influences were
`merely additive or were in some way synergistic.
`The rates of deposit formation were measured in the Ad(cid:173)
`vanced Kinetic Unit at 69 atm with the temperature zones at
`371-540 °e. The molecular oxygen content of fuel mixture was
`reduced to 1 ppm of O2 or lower. The JP-5 fuel alone, without
`added compounds, was tested similarly and for purposes of
`comparison. The results of the effects of various added ni(cid:173)
`trogen compounds on deposit formation in JP-5 are summa(cid:173)
`rized in Table I. Some representative Arrhenius plots are
`shown in Figures 1 and 2. None of these compounds studied
`promoted deposit formation to any appreciable extent in the
`deoxygenated systems. The total deposits observed were of
`the same order of those obtained from the base fuel (Table 1).
`This effect is in contrast to their deleterious nature in oxy(cid:173)
`gen-saturated fuels (Nixon, 1962; Taylor, 1968b). The Ar(cid:173)
`rhenius plots for fuels with pyrrolic compounds added have
`different slopes than that for t he base fue l (Figure 1). All
`showed slightly higher deposits in the lower temperature re(cid:173)
`gime when compared to the base fuel. Other heterocyclic
`
`GE-1022.006
`
`

`
`88
`
`Ind. Eng. Chern. Prod. Res. Dev., Vol. 17, No.1, 1978
`
`Table II. The Effect of Individual Oxygen Compounds on Total Deposit Formation in a Deoxygenated Jet Fuel
`
`Class of added
`oxygen compound a
`
`Peroxide
`
`Carboxylic acid
`
`Added compound
`
`Oxygen
`content,
`ppm of O2
`0.2
`Di-tert- butylperoxide
`0.1
`Cumene hydroperoxide
`0.2
`tert- Butylhydroperoxide
`0.1
`Cyclohexanecarboxylic acid
`0.1
`n-Decanoic acid
`0.2
`Cyclohexanebutyric acid
`0.2
`2-Ethylbutyric acid
`2,4-Dimethylbenzoic acid
`0.3
`0.2
`2-Methylphenol
`2,6-Dimethylphenol
`0.1
`0.1
`2,4,6-Trimethylphenol
`Benzo(b )furan
`0.2
`0.2
`Dibenzofuran
`0.9
`n -Dodecyl alcohol
`0.3
`4-Methylcyclohexanol
`0.7
`5-Nonanone
`4-Methylcyclohexanone
`0.3
`0.2
`Cyclohexyl formate
`0.7
`Methyl benzoate
`0.8
`Pentyl formate
`0.4
`Base fuel
`a All compounds added to the 100 ppm 0 level. b See Table I for run conditions.
`
`Phenol
`
`Furan
`
`Alcohol
`
`Ketone
`
`Ester
`
`Total deposits b
`As ppm
`based on fuel
`
`~g
`of carbon
`
`2879
`7219
`8934
`1563
`2997
`1730
`1291
`1801
`1561
`2048
`1451
`1505
`1410
`2046
`1356
`2422
`1244
`1318
`2488
`1894
`1485
`
`1.49
`3.73
`4.62
`0.82
`1.54
`0.89
`0.67
`0.93
`0.81
`1.06
`0.75
`0.78
`0.73
`1.06
`0.70
`1.26
`0.64
`0.68
`1.29
`0.98
`0.77
`
`1 , 0 00
`
`-o
`0
`
`, ,
`,
`,
`I,
`
`I
`
`II
`
`I
`
`100
`
`~
`0
`r
`N~
`
`~
`0
`
`5
`~..~
`:;:
`
`0
`
`~ ;
`
`;;
`~
`
`)000
`
`100
`
`10
`
`~ ~orN
`
`;5
`
`u,
`o
`
`l OOO/'"K
`
`Figure 3. Deoxygenated fuelat 69 atm (0.2ppm of O2) : e, with added
`cyclohexanecarboxylic acid; _ , with added mixed naphthenic acids;
`., with added n-decanoicacid. Allcompounds added at 100ppm of
`O.
`
`amines such as pyridines and quinolines showed similar ef(cid:173)
`fects. Primary amines, on the other hand, appeared to have
`no influence either in total deposits formed (Table 1) or on the
`shape of the Arrhenius plots (Figure 2).
`The effects of various oxygen compounds studied are
`summarized in Table II and representative Arrhenius plots
`are shown in Figures 3 and 4. These compounds varied con(cid:173)
`siderably in their influence on fuel instability. The furans,
`most carboxylic acids, and alcohols were generally not harmful
`to fuel stability. An exception was n-decanoic acid. It pro(cid:173)
`moted significantly higher deposit formation rates than either
`the aromatic or naphthenic acids studied (Table II, Figure 3).
`Carboxylic esters were only mildly harmful and their influence
`also varied with structure. The aromatic ester, methyl ben(cid:173)
`zoate, was mildly deleterious. The naphthenic ester, cyclo(cid:173)
`hexyl formate, exerted no apparent influence on deposit for-
`
`Figure 4. Deoxygenated fuelat 69 atm (0.1 ppm of O2) : e , with added
`tert- butylhydroperoxide; _ , with added cumene hydroperoxide; .,
`with added di-tert-butylperoxide. Allcompoundsadded at 100ppm
`ofO.
`
`mation. The purely aliphatic ester, pentyl formate, was in(cid:173)
`termediate in its effect. Of the two ketones studied, 5-nona(cid:173)
`none was mildly harmful while 4-methylcyclohexanone had
`no observable effect (Table II). Peroxidic compounds, re(cid:173)
`gardless of structure, cause significantly higher deposit for(cid:173)
`mation rates and may be classed as highly deleterious (Table
`II, Figure 4). The most active peroxide studied, tert- but(cid:173)
`ylhydroperoxide, afforded a 600% increase in total deposits
`over the base fuel even though it was added only at the 100
`ppm 0 level. This increase in deposit is of the same order of
`magnitude as that observed with the most harmful sulfur
`compounds wh ich were present at the 3000 ppm S level
`(Taylor, 1976) .
`An interesting result, obtained in all classes studied, is the
`consistently lower deposits observed with naphthenic (cy-
`
`GE-1022.007
`
`

`
`Table III. Interaction Study of the Presence of Pyrrole
`and an Acid in Deoxygenated JP-5
`
`Presence
`of an
`acid"
`
`Total JLg of carbon
`Presence of a pyrrole a
`No
`Yes
`
`1310
`1485
`No
`5071
`2997
`Yes
`a 100 ppm of N as 2,5-dimethylpyrrole added to base JP-5.
`b 100ppm of 0 as n-decanoic acid added to base JP-5; see Table
`I for run conditions.
`
`cloalkane) compounds. This result is especially noteworthy
`for the esters and ketones where the naphthenic compounds,
`cyclohexyl formate, and 4-methylcyclohexanone, afforded
`only half the deposits of their aromatic or open chain analo(cid:173)
`gues. However, the effect is also significant among the alco(cid:173)
`hols, phenols, and carboxylic acids (Table II).
`In order to test for interactive effects of synergisms, a 2 X
`2 factorial experiment was designed (Bennett and Franklin,
`1954) using the presence of the two added compounds as
`variables. An example is given in Table III. In the example
`shown, a strong interaction between 2,5-dimethylpyrrole and
`n-decanoic acid was uncovered. The total deposits for the
`interaction run, in excess of 5000 JLg of carbon, are consider(cid:173)
`ably greater than expected from the sum of the additive con(cid:173)
`tributions (2822 JL g from a contribution of 1485 J.Lg from the
`base fuel plus 1512 J.Lg for the effect ofthe addition of the acid,
`minus 175 J.Lg for the effect of the addition of the pyrrole).
`Several interesting interactions were observed when two
`different compounds were added simultaneously to the base
`fuel. For some, the total deposits formed were significantly
`higher than could be accounted for by additive effects alone.
`For others, stabilizing interactions were encountered. That
`is, lower deposit formation rates were observed than would
`be expected from either of the added compounds alone.
`A summary of the more important interactions is given in
`Table IV. The two compounds showing the greatest tendency
`to interact were 2,5-dimethylpyrrole and n-decanoic acid.
`These compounds also exhibited strong interactions with
`sulfur-containing impurities and olefins (Taylor and Frank(cid:173)
`enfeld, unpublished observations). Significantly, all of the
`interactions among purely oxygenated species were stabilizing;
`that is, they tended to reduce deposit formation. Other com(cid:173)
`binations of compounds from Tables I and II were studied but
`gave no significant interactive effects.
`
`Discussion
`As pointed out previously (Taylor, 1974), not all fuels ex(cid:173)
`hibit enhanced thermal stability when rigorously deoxygen(cid:173)
`ated. Since those fuels which fail to show the expected im(cid:173)
`proved stability are of generally poor quality, the presence of
`trace impurities is a likely cause. Taylor (1976) has shown that
`
`Ind. Eng. Chem. Prod. Res. Dev., Vol. 17, No.1, 1978
`
`89
`
`the presence of certain sulfur compounds can markedly in(cid:173)
`crease deposit formation in deoxygenated jet fuels. The
`present study indicates that certain other trace impurities,
`especially oxygen containing, can also contribute significantly
`to high-temperature instability.
`The formation of deposits in jet fuels is known to involve
`both chemical and physical changes (Taylor, 1974; Boss and
`Hazlett, 1969). In oxygen-saturated systems below the py(cid:173)
`rolysis temperatures, deposits form as a result of free radical
`chain reactions (autoxidation). However, in deoxygenated
`fuels, the process appears different and even more complex
`(Taylor, 1976).
`Previous studies have shown that nitrogen compounds can
`be highly deleterious to air-saturated fuels both under storage
`(Nixon, 1962) and "empty" wing tank conditions (Taylor,
`1968b). However, in deoxygenated fuel, nitrogen compounds
`have little effect on deposit formation under high-temperature
`conditions. This is apparent from data in Table I and the
`Arrhenius plots given in Figures 1 and 2. None of the nitrogen
`compounds caused a significant increase in total deposits over
`the entire temperature range. In the low-temperature regimes,
`however, the heterocyclic amines 2,5-dimethylpyrrole, indole,
`and carbazole promoted deposit formation to a small extent
`when compared to the base fuel (Figure 1). The primary
`amines studied had no observable influence on deposit for(cid:173)
`mation throughout the temperature range (Figure 2).
`The foregoing observations pertain to an effect on deposit
`formation on the tube surface at high temperatures (700-1000
`OF). It should be noted that during storage at ambient tem(cid:173)
`peratures, some heterocyclic amines, in particular 2,5-di(cid:173)
`methylpyrrole, promoted the formation of considerable
`flocculent sediment in both air-saturated and deoxygenated
`fuels. The appearance of dark colored sediment was evident
`within a few hours after the pyrrole had been added to the fuel.
`This was observed previously in air-saturated fuels and has
`been attributed to autoxidation of the pyrrole (Oswald and
`Noel, 1961; Dinneen and Bickel, 1951; Angeli, 1916). It is
`somewhat surprising, however, to find such sediment in rig(cid:173)
`orously deoxygenated systems where autoxidation would be
`very slow. A preliminary analysis of the sediment from
`deoxygenated JP-5 suggests that it is quite different from that
`reported by others in air-saturated fuel. At any rate, the
`sediment formed at low temperatures, suggesting that it was
`either broken down to fuel -soluble fragments at high tem(cid:173)
`peratures or that it did not adhere to the surface walls and
`form a deposit.
`The compounds containing oxygen were considerably more
`deleterious than those containing nitrogen (Table 11). Par(cid:173)
`ticularly noteworthy were peroxides although n-decanoic acid,
`methylbenzoate, and 5-nonanone also increased deposit for(cid:173)
`mation by at least 50% over the base fuel. Several aliphatic
`alcohols and phenols produced a moderate increase in deposit
`formation. This effect with phenols is in contrast to air-sat-
`
`Table IV. Interactions between Oxygen and Nitrogen Containing Impur-ities"
`Total deposits,
`JLg of carbon b
`
`Compounds
`2,5-Dimethylpyrrole + n -decanoic acid
`2,5-Dimethylpyrrole + 2,4,6-trimethylphenol
`2,5-Dimethylpyridine + n -decanoic acid
`2,4,6-Trimethylphenol + n -decanoic acid
`5-Nonanone + n-decanoic acid
`Dibenzofuran + n-decanoic acid
`2,4,6-Trimethylphenol + Methylbenzoate
`
`2822
`1276
`3489
`2963
`3934
`2922
`2454
`
`5071
`1784
`1993
`1484
`1741
`1566
`1808
`
`Interaction effect
`i.e., increase or
`reduce deposits
`
`Increase
`Increase
`Reduce
`Reduce
`Reduce
`Reduce
`Reduce
`
`a Compounds present at levelsof 100ppm of Nor 0 in deoxygenated JP-5; run conditions in Table I. b Predicted from data in Tables
`I and II assuming each component acting independently. c See Table I for run conditions.
`
`GE-1022.008
`
`

`
`90
`
`Ind. Eng. Chern. Prod. Res. Dev., Vol. 17, No.1, 1978
`
`urated fuels where they often act as antioxidants and, there(cid:173)
`fore, stabilizing agents. On the other hand, the cyclic ethers,
`benzofuran and dibenzofuran, showed no tendency to promote
`deposit formation. It appears that oxygenated impurities
`behave similarly to their sulfur analogues in deoxygenated
`JP-5. Taylor (1976) reported that di- and polysulfides, an (cid:173)
`alogues of peroxides, were highly deleterious to fuel stability
`while thiophenes, analogous to furans, were inert.
`The stabilizing effects of napthenic rings is illustrated by
`data in Table II and Arrhenius plots in Figure 3. As shown in
`Figure 3, the aliphatic acid, n-decanoic, afforded significantly
`greater deposits than either of the two napthenic acids shown.
`The napthenic alcohol, 4-methylcyclohexanol, gave lower
`deposits than any of the phenols or aliphatic alcohols (Table
`II). The same trend is noted for both ketones and esters.
`This difference cannot be due to increased stability to py(cid:173)
`rolysis since methyl benzoate is much more resistant to
`thermal decomposition than esters of formic acid (Hurd,
`1929). Similarly, alicyclic alcohols appear at least as heat
`sensitive as their aliphatic counterparts and phenols are much
`more stable than either (Hurd, 1929). It would appear that the
`lower deposit formation rates observed with the napthenic
`compounds are due to the enhanced solubility of their pyrol(cid:173)
`ysis products rather than to any enhanced stability to heat.
`One of the salient features of deposit formation with air(cid:173)
`saturated fuels is the complex Arrhenius plot which results
`from the sharp drop in rates in the 350-430 °C (650-806 OF)
`range. This was pointed out by Taylor (1974) and is illustrated
`by the typical curve for the air-saturated fuel given in Figure
`4. Deposit formation rates show a sharp drop in this temper(cid:173)
`ature range before continuing upward again at even higher
`temperatures. Previous studies at sub -atmospheric pressures
`have shown that as a hydrocarbon jet fuel is heated, the rate
`of autoxidative deposit formation increases until the liquid
`phase is lost, at which point the rate of deposit formation
`drops sharply (Taylor and Wallace, 1967). It would appear
`that much of this drop can be attributed to a reduction in the
`autoxidative reaction rate revealing the concentration of re(cid:173)
`active species as the system passes from the liquid phase to
`the vapor phase (Mayo, 1968). With increasing temperature,
`the autoxidative rate constants continue to increase, so that
`ultimately this concentration effect is overcome and the
`overall rate of reaction again increases. The very similar shape
`of the curves obtained when deoxygenated fuels are doped
`with various peroxides (Figure 4) suggests that peroxides
`formed by autoxidation of air-saturated fuels are, indeed, the
`reactive species, whose drop in concentration with the phase
`change causes this discontinuity. The curve for the rigorously
`deoxygenated fuel, with no added peroxides, does not exhibit
`this effect.
`Furthermore, the deoxygenated fuels doped with various
`peroxides show an immediate increase in deposit formation
`(low-temperature regimes in Figure 4) while the air -saturated
`base fuel, with no added peroxides, exhibits a gradual increase
`in deposits. This suggests a steady generation of per oxidic
`compounds due to autoxidation of the base fuel by dissolved
`oxygen . These results strongly suggest that peroxides are a
`major cause of deposit formation in air-saturated fuel sys(cid:173)
`tems.
`A number of interactions between trace impurities were
`observed in the present study. In some instances, these led to
`
`significantly increased deposits over those expected from
`additive effects alone while at other times the interactions
`actually tended to inhibit the formation of deposits. The most
`important interactions between 0- and N-containing im(cid:173)
`purities are summarized in Table IV. The most active material
`tested was 2,5-dimethylpyrrole, which interacted in a dele(cid:173)
`terious manner with both acids and phenols. A deleterious
`interaction with olefins was also observed which will be re(cid:173)
`ported later. Significantly, the interactions between oxygen
`compounds were nondeleterious. That is, they tended to re(cid:173)
`duce deposit formation rates below that expected from ad(cid:173)
`ditive effects.
`These findings suggest that trace impurities must be taken
`into account when assessing the thermal stability of fuel for
`high-speed aircraft. Deoxygenation procedures will be of op(cid:173)
`timal effectiveness only when trace impurity effects are con(cid:173)
`sidered and eliminated or controlled.
`
`Acknowledgments
`Helpful discussions with C. J. Nowack, L. Maggitti, Jr., and
`J. R. Pichtelberger are gratefully acknowledged.
`
`Literature Cited
`Angeli, A., Gazz. Chim. Ital., 46 (II), 279 (1916).
`Back, M. H., Sehon, A. H., Can. J. Chern ., 38, 1076 (1960).
`Ball, J. S., Proc . Am. Pet. Inst ., Sect. 8, 42, 27 (1962).
`Bennett, C. A., Franklin, N. L., "Statistical Analysis in Chemistry and the Chemical
`Industry", Wiley , New York, N.Y., 1954.
`Boss, B. D., Hazlett, R. N., Can. J. Cbem ., 47,4175 (1969).
`Braye, E. H., Sehon, A. H., Darwent, B. de B., J. Am. Chern. soc., 77, 5282
`(1955).
`Coleman, H. J ., Hopkins, R. L., Thompson, C. J., Am. Chern. Soc., Div. Pet. Chern.
`Prepr. , 15 (3), A17 (1970).
`Churchill, A. V., Hager, J. A., Zengel, A. E., SAE Trans., 74,641 (1966).
`Dinneen, G. U., Bickel, W. D., Ind. Eng. Cbem., 43, 1604 (1951).
`Fabuss, B. M., Duncan, D. A., Smith, F. 0 ., Satterfield, C. N., Ind. Eng. Chern.
`Process Des. Dev ., 4, 117 (1965).
`Hendrickson, Y. G., Am. Chern . Soc ., Div. Pet. Chem. ,