File D 52.416 / 9 1'. t_ 1 McCOOK FIEl,D REPORT, SERIAL No .

1715

'AIR SERVICE IN ~ QRiVLION

~ CIRCULAR
(AVIATION)

PUBLISHED BY THE CHIEF OF AIR SERVICE, WASHINGTON, D. C.

Vol. IV March 15, 1922 No. 308

INVESTIGATION OF THE EFFECT OF DOPED FUELS

ON FUEL SYSTEM
( MATERIAL SECTION REPORT No. 152)

Prepared by Engineering Division, Air Service

McCook Field, Dayton, Ohio
September 14, 1921

&21~
~ W,.;y.
- RldburnUhiv6¥SityJ

WASHINGTON
GOVERNMENT PRINTING OFFICE
1922
 CERTIFICATE: By direction of the Secretary of War, Uie matter contained herein is published as a<lministrative
information and is required for the proper transaction of the public business.
(n)
 ,··
INVESTIGATION OF THE EFFECT OF DOPED ~FUELS
ON FUEL SYSTEM.
GENERAL. plate' was included for the purpose of finding out how it
stood up in comparison with the above. All of these
The use of antiknock compounds in aircraft operation
materials were obtained in sheet form from stock at Mc-
has practically become an absolute necessity for two
Cook Field, with the exception of the zinc plate, which
main reasons, the constantly increasing use of higher
was prepared in the chemical laboratory of the Material
compression ratios and the necessity for use of a cheaper Section. ·
fuel. This, the addition of antiknock compounds makes
Rubber hose of the following description was included
possible. in this investigation:
Accompanying the experimental use of a high test
Goodyear gasoline, line,!.- inch inside diameter.
aviation gasoline containing antiknock, occurred the,
Cloth wrapped gasoline line, % inch inside diameter.
obstruction of the respective gasoline systems by con-
Unarmored pump hose, %' inch diameter.
siderable amounts of extraneous matter. It was found
This material was likewise obtained from stock at Mc-
that the corrosion of the gasoline tanks and the subse-
Cook Field.
quent accumulation in tubes and strainers of the corroded
The action of the following fuels on the abpve materials
matter caused the obstructions of the lines. The belief was investigated:
arose that antiknock accelerated this action. ,It was
(a) High test gasoline.
considered advisable to investigate the effect of fuels and
· ( b) Low test gasoline.
doped fuels on various materials in order. to determine if
( c) Antiknock compound.
these compounds would produce any injurious results on
(d) 91 per cent high test gasoline containing 9 per cent
the fuel system. antiknock.
PURPOSE.
(e) 97 per cent high test gasoline containing 3 per cent
The purpose of this investigation 'was to determine the antiknock. -
effect of airplane fuels , dopes, and doped fuels on the (!) 50 per cent high test gasoline, 50 per cent benzol.
fuel system, with particular attention to the problem of (g) 84 per cent low test gasoline, 16 per cent antiknock.
corrosion prevention. The high test gasoline, low test gasoline, and benzol
were obtained from stock at McCook Field. The anti-
CONCLUSIONS. knock compound is Antiknock No. 1. It was obtained
It is found: from the General Motors Corporation, Research Division,
1. That the following metals are very slightly, if at all,' Moraine City. It is claimed that this antiknock com-
affected by doped fuels: pound is composed of 70 per cent aromatic amines and 3Q
(a) Aluminum, per cent benzol. The amines are probably orthotoluidine,
(b) Zinc. paratoluidine, xylidine, aniline, or mixture of these.
2. That the following materials are very badly affected A complete chemical analysis (page 4) was ma.de on all
by such fuels: metal stock used, and distillation curves (see fig. 8) were
(a) Copper. run on fuels. These data are included for purpose of
(b) Brass. reference.
(c) Terneplate. METHOD OF PROCEDURE.
(d) Iron. In the case of the metals, the sheet stock was cut into
3. That tin plate is moderately affected by these fuels.
small strips 4 inches long and 1 inch wide. These were
It is to be expected, therefore, that considerable diffi-
placed in the above fuels and doped fuels in such a manner
culty will be experienced with terneplate gasoline tanks that only half of the specimen was covered with the liquid.
and copper lines, brass jets, liners, etc ., and to a less
The other half was thereby exposed to atmosphere satu-
extent tin-plated tanks. Caution should be exercised in rated with the vapors of each fuel. Two series were run
. the use of each of these. in parallel; one series was placed in flasks which were
Reasoning by analogy, it is seen that alloys containing
securely stoppered so as to exclude all possibility of
magnesium, cobalt, nickel, tungsten, copper, iron, and ventilation and entrance of moisture and, therefore,
lead are apt to be very seriously affected by these fuels,
represented the conditions present in a gasoline tank
while alloys containing aluminum, copper, zinc, and which was partially filled with liquid and the re'st of the
silicon (and in rare cases silf er and platinum, as well as
tank filled with practically moisture-free air, 11aturated
osmium and iridium) are apt to be quite resistant to these
with the fumes of the fuel in the tank. The other series
fuels. was run under identical conditions, with the exception
MATERIALS.
that all flasks were allowed to remain unstoppered and
The materials used in this investigation are determined were placed in a large container, which was sealed to pre-
bv the nature of the materials used in a typical fuel sys- vent entrance or exit of fumes. This container was satu-
t;m (see fig. 1). These materials are largely aluminum, rated with moisture vapor, thereby reproducing the con-
copper, brass, steel, terneplate, and tin plate. Zinc ditiona that are present in a pa~tially filled gasoline tank
96724~ 22 (1)
I
 2
when the atmosphere is very 'high in moisture content.
It was believed that this latter condition would impose a
much more rigorous test on the materials than ·the above
practically moisture-free series. This was found to be
the case.
The rubber gasoline lines were tested from specimens
made by cutting the hose into sections 6 inches in length,
stoppering one end with a cork stopper and inserting in
the other end a 6-inch glass tube. The gasoline hose and
affected. The exposure extended over a period of 145
days.
DISCUSSION OF RESULTS.
It is observed that the metal which stood up better than
any other examined is aluminum . Second to this is zinc,
and third in order is tin, while steel, terneplate, and par-
ticularly copper and the copper alloy, brass, are very
badly affected by fuels containing antiknock compounds.

DIAGRAM OF TYPICAL FUEL SYSTEM.

~a¥.lf1C1Cflrol? . .
~f.tl't.: Y w.:,a.c 10Ptt.or a:'

"''~ LlfrC MU"T .XOPC

OOHf'f1'l1'iE'C f'IT RLt. POIITTa

lhPf"C -WfiY COCK

TYPCC-1

J
THECC tmY cx.t;' CX'ffOCXD
MAIN T A NK canrerx. ,p ru™rrr fJOfN?D

Tl!'Rl'ITMl~~f'l<Z»&C
fY/0 OR"i:z 1Rl«D fl80Vl"
ormnce roertor oTffl/11£~
CFJP mo xeeen

t:!Rff:2·
ML L/flCli M/¢T 8C
IJfK1{XQn17lfffQ A'lf('fT
IJT~Krt '1'1/J.

tr"'32 OfT fX..L nK'L flU2

q,Jm /!K,OTB(AIJITTWKO

---------
--~~:.-~----
, L80TTCi'rf Ot' ro.,(L,%C

--- ---- - - ------ --- - -

' F10 . !.-Materials used in fuel system.

glass tube were then filled with the fuel and securely
stoppered. The fuels in these tubes evaporated within a
few days. The tubes were refilled and allowed to evapo-
rate a second time. They were left undisturbed until
the time of examination at the conclusion of the test.
RESULTS.
The results obtained in this investigation are charted in
schematic form in figures 5 and 6. These data were ob-
tained by noting the effect of the various fuels and doped
fuels on the materialJ, and stating the results in the form
of slightly or somewhat affected, badly affected, and un-
The deleterious action of these compounds on the metals
present in the fuel system is considerably increased by
the presence of moisture. Figure 2 shows the results of
the action of fuels and doped fuel s on the various metals
in the absence of moisture. Figure 3 shows the action of
these fuels in the presence of moisture. Figure 4 presents
the metals in their original condition before t4e tests were
begun. It will be noted that the metals in figures 1 and 2,
which have been affected by the various compounds, are
considera];>ly darker than the originals shown. The actual
corrosion which took place is very difficult to show by
means of photograph. However, some idea of the relative
 3

Sneefflt,.,m/num Sheet Copper S heef-8n,ss

- Liqvid /;&par' M,pcr
Hi9i, Tesr
(,J'ISOJ,/ 1'/E C

e+:Zlo,desr
ljq;1.c/lrie
jG'Z_flnfi-Knoek
'1/?;'/li9!r 'le., f
t;c»lli,e
-Z/f l{lefi- knock
"fl'% Hi!lh 7iisf-
'Gc.solil't4
~ J%!{_11h'-'f!,o<-11 _

,J'o% llig/TisNiqsoline h
r-:-.-~- ---"~.,• ~-~-2-•
.so%1Je>,:r.ol___::-::':'~-*:::-::::;::::==~u;ts::;.__,f!l::1111

lfNT;- KNO(,f( _'111)2 -'--"--.,....'J

............,.....

Term!: Plo+e
L (ijvid
H19h Te.st<joso/ine

'/1%/1~ Je:,f(igc,cli11e
9% linf1-kr,.&1<
.. ,· ___ h'<:efe.'-
T/,,,,port- of each specimen
/al,e!/e<t 'L1qvicl" wos immersed
in tite/uel; fhepo,-f lobe/led
•lrJpor"wos exposed to flte corn-
l>~11ed n;pcrS of fl1e se fuels in
011''.

' f/;vrt- /(,voc}f

FIG. 2.- l;ondition of matena , alter 145 days· exposure.

Sfteel Copper Sheer B ross

H igh ~sr
GRSOLIIYE

Low Test
{j!JSOI-IIYE

'II% /f,9h Tesr

t:jos a/,,.,..ie
9% llnt-i- knock

97% lfi9t, 7esf'

. (jqsclhre
J%/}nt/ -kn"cA

Hi9hTest(jos,/1'ne ~
Sheets-tee/
'-''!"it/ Vapor
-- Ttn Pktfe
L i9t1/.:i /{,por

/Yo fe :-
Theporf of each spec/met?
/,be/led "t ,,;uid• wo.s itnmersec/
9ti /fig/, Tesf (josof,i,e /, ,;, thefvel; f/te parf hbe/leti
n 111111- 1<#de;, ·Vapor w...s exposed ro tbe com -
bined Vapors of +hese fvels in
air satvra'fed W'/fh m o i sfvre;
9T%11,'g'1 Tesfi,osoli11e
Jr% /lhti-/(1"'1&1<

FIG. 3.-Condition of material after 145 days' exposure (moisture series.)

 4
effects can be obtained(by comparing the varying degrees
of darkness of the samples as compared with their originals.
It will be noted that the darkest samples, that is those
that were most affected, are copper and brass. This
"effect" was in the form of a combined corrosion of the
metal and a deposition from the fuel, the total deposit
being in the form of a tacky residue.
It is, therefore, to be expected that these compounds
will affect, to a considerable extent; copper tubes, terne-
plate gasoline tanks, brass or bronze couplings, while
aluminum housings, aluminum pumps, etc., will prob-
ably suffer only slightly, if at all .
Tin plate (dairy stock) can be expected to hold up much
better than terneplate, and it is, therefore, a natural con-
clusion that dairy stock should be used for gasoline tanks
in preference to terneplate.
The ideal material for gasoline tanks appears to be
aluminum, for not only has aluminum withstood the tests
of this investigation better than any of the other metals,
but it h~s also been found that aluminum is the least
-
'!emePlate
lin 'Plafe .
FIG. 4.- 0riginal condition or material.
affected by sulphur and other compounds extracted by gaso-
line from rubber gasoline lines. (See Naval Aircraft Fac-
tory report on "Effect of corrosive material in rubber hose
--..upon gasoline tanks") and that aluminum gasoline tanks
suffer least from corrosion (see report "Tests on fuel tanks
for aircraft," Naval Aircraft Factory). The corrosive
action of doped fuels will be particularly marked if these
fuels are allowed to remain for some length of time in the
fuel system.
Reasoning by analogy, the results obtained from cer-
tain other metals and alloys of various metals can be an-
ticipated after the manner outlined in figure 5. Figure 5
give.s a graphical representation of the effect of antiknock
compound on various metals, arranged in descending
order of magnitude, that is, the metals at the top of the
curve are probably the least affected, while those at the
bottom of the curve are most affected, and those in be-
tween these two components are affected in the order of
magnitude relative to their position.
It is, therefore, to be expected· that alloys containing
rather high percentages of ahiminum, zinc, and in some
rare cases, silver and platinum, will not be affected by
antiknock compound, while alloys containing magne-
sium, copper, cobalt, nickel, and tungsten will be very
greatly affected by such compounds, and alloys contain-
ing chromium, manganese, tin, and leaq moderately
affecte~. ' That is to say, it is to be expected that such
materials, as for example the B . G. spark plug, containing
largely nickel with a trace of manganese, and the B. G.
Ferronica spark plug, containing chromuim, nickel, and
iron, will both suffer considerable fouling from these fuels.
It is possible, however, that the high heat of the cylinder
may prevent this. Other materials containing such ele-
ments are certain to cause trouble if placed in the gasoline
system.
In the case of rubber lines, it is believed that difficulty
may be expected only in those cases where high percent-
ages of antiknock compounds are used. It is to be ex-
pected that in such cases a thick gummy substance will
accumulate in the lines and possibly in the tanks. This
of course is prohibitive. The rubber lines subjected to
the action of fuels containing small percentages of anti-
knock compounds seem to have suffered insufficient
deterioration to warrant their consideration.
EFFECT OF FUELS AND DOPED FUELS ON IiUBBER.
Goodyear Gasoline Line, -h-inch inside diameter.-N o
apparent effect of the abo ve fuels on this line except that a
heavy tacky deposit is formed when high per<'entages of
antiknock are used.
Cloth wrapprd gasoline line, %-inch inside diameter. -
Slight discoloration by all fuels and also some deposit.
Antiknock produced a heavy tacky deposit.
Unarmored pu:mp ho.~e, :U:-·inch dimneter.-No appa.r~nt
e£fect produced by high-test gasoline, llJw-teBt gasoline,
and benzol. Each fuel, using antiknock, discolored the
inside of the hose and left more or less of a tacky deposit.
Straight antiknock produced th" same deposit to a
greater extent.
CHEMICAL ANALYSIS OF SHEET STOCK.
Tinned sheet steel, Specification No. 10207 A: 5 pol!nds per
base box of 112 sheets, 14 by 20 inches.
Terne sheet steel, Specification No. 10209.A: 8 pounds per
base box of 112 ~heet~, 28 by 20 inches:
Tin ... _____ _. __ . __ .. _.... _._ .per cent._ 20.0
Lead __ _. __ : . _. __ ... _. _...... . . _. _do .. __ 80.0
Sheet copper, 99.9 per cent copper.
Sheet aluminum:
Silicon .... _..... - . __ ..... _ ... _per cent._ 0.08
Covper. ·--· - · -... . ......... . ..... _do .. __ 4.00
Iron __ ._. __..... ___ . .. _. _. ___ .. .. do .. _. 0.62
Aluminum ._······-··- ..... _.... __ do .... 95.30
Sheet brass:
Lead_ ··-· -······· - ----- · ---------·-- ··- Trace.
Copper. ______ ______ .. ___ . ____ per rent._ 79. 77
J ron ____ . __ . ____ ____ . ____ ' _. __ . __. do .. __ (). 21
Zinc ___ - . _. - .. - _. _______ . __ ... ___ do __ . _ 20.02
Sheet st.eel, 0.023 ga., Specification No. 10201:
No. 1020 st.eel:
Carbon ___ - - . _. - _. ____ - - - .. - .. per cent. . 0. 15-0. 25
Do. -. - _.. - . - - . - ___ - - - . . - . . - .. rlo_ ___ 0. 20- 0. 30
Manganese ___ . - . _. _.. _- .. - - .. _.. _do____ 0. 30-0. 60
Do. --... _. . - - - . . __ . - - - - . - .... do_ ___ 0. 50-0. 80
Phosphorus ________ . ___ - - __ . __ .. __ do._._ 0. 045
Maximum ___ ___ ___ --- - . . ·-_._do. ,_ _ 0. 045
Sulfur __ __ __ _.- ____ .. _____ . __ . _.. _do .. __ 0. 050
Maximum·-··--- __ ___ --- _.. . _do __ ._ 0. 050
Galvanized sheet steel, 0.001 inch thick:
Zinc. --...... . - .. - _. - _- - - - . - __ per cent. _ 99. 9
 5
\.

i"'IE:TALS

L.ow4'f"ce.,r
G,.,,oli'\~

11 %)11fTeot-
q0.A.rlo-t<ud,
84-%1.io wTa,~
iG %11•1;. K•""-
q1')'a >Hilt Te ,r
3 0 fh;l.t- k'ftqd._
.5 o" . }Iii\. Tc~\~
so{- 'i)t.fl ol

FIG. 5.- The effect of fuels on fuel systems in the absence of moisture .

M ET}IL,~
I

I{,,~ Tul-
G,~• li-.,
.Li • ., ......
G•~oUt.~

FIG. 6.- The eliect of fuels on fuel systems in. t he presence of moisture.

F ro. 7.
NOTE-From the above chart it is possible to predict the effect of doped fuels on various metals. The points aluminum, copper, zinc,
tin, iron, and lead have been experimentally determined.
 ·--
oC"' P,'i' , ·"j ,..
6 _...
. - --- $ - .
-
I -

V
oc. •F. . ;..

160
I/ ...
/
/
320
'.
.,:,_

0
I
_/, ~ 4
140 284
./
/ ff
. .V /P
120 248
J
V D
/ /' I
100

L.
V
-
/
_, - .-- --- ..--i ~-- __,,, 'I"'

,--
./
/ '
_.,/
J 212
I/

80,
~ I ~
-
~- .c-, ,- \
'-

176

/; v- ,....... i..--- 140

. 60,
/ ~
/

F /

40 1 . 104
'-

20 68
FIRST 10 · 20 30 40 50 60 70 80 90 ORY
DI\OP POIMT
P~c:u1t 015tlll.d -

- - o - - 1-ri sh Tesr Go.so\tne

./ ~
--~~-- LowTe"3r G~soline l~1 r 392
--o-- E!enzol
- - A - - so% l1 i ih Tts, soi., Benrol
V J ~
356
- v - 91 %tlijh Test 9% fln1i-l\nock
V I Of.
iV ii
- x - 9'7% 11iyh Test 3%r'\rr1Hfoock
- o - S4%l,o\"t 1esr 16,:. Rnli-ijnock
.v I j
320

I I 284
J ·'
120 v· .I/ / 248
oc. /
V Vi /
'
? V
100
l./'
./
/p

~
-·
...... 212·
~
- -- ~

-~ ·- - i.-i- -
~7
/
:;;.,--~
80 176
/.,
v' ·-
./ Ir- \

I
6 ...... 140
/'
•
40 104

zo 68
FIRST
OROP
10 20 30 40 so 6() 70 80 90 ORT
l"OIN'f

FIG. 8.- Motor fuel distillation.

0