, ,I'


. '


Depot Concept'





Presented at International Automotive Congress January 11-15. 1965 Published by:

, .





Ammonia as an Engine Fuel



L. William Huellmantel, and Harry R. Mitchell


Research Laboratories, General Motors Corp.


THIS EVA'LtJATION OF AMMONIA as an engine fuel was performed by the Gener~Motors Research Laboratories in support of the energy de t concept proposed by the Allison Div. of the General Moto s Corp. (1).- The objective of the

.energy depot concept is-to free the 'Armed Forces from reliance on hydrocarbon' fuels. One method is the production of a fuel from water and air .. Of the potential fuels which might be produced. anhydrous ammonia (NH3) was con-

sidered to offer the most advantages. The energy required to synthesrze this fuel would be provided by a mobile nuclear reactor. The General. Motors Research Laboratories undertook the task of evaluating anhydrous ammonia as a fuel for spark-Ignited reciprocating engines.

The chemical equation below describes the combustion of a stoichiometric mixture of ammonia and air:

2 NH3 + 1.5 (0 + 3.76 N ) ~3H20 + 6.64 N

2 2 11 . 2

Table 1 is a tabulation of those properties of ammonia and a typical commercial gasoline that have an effect on combustion,

As shown, the heating value of gasoline is 2.4 times that - . of ammonia) However. the stoichiometric air-fuel ratio for


- .. '

eNum bers in parentheses designate References at end of . paper.

Table 1 - Comparison of propertie~f Anhydrous Ammonia and Gasoline

Ammonia Gasoline'
Chemical formula NH3 CH
Density. Ib/gal 5.1 6.1
Boiling point (1 atrn), F -28 ".
Freezing point, F -108 -76
Vapor pressure ('70 F). J:sia 12S.8 t;,>
Heat of vaporization ('70 F),
Btu/lb 50S.6 116
Heat of combustion (Lower heat
value -- gaseous). Btu/lb 8000 18,900
Stoichiometric air- fuel 'ratio 6,06 14.5
Stoichiometric heat release
(weight), Btu/lb air 1320 1285
Stoichiometric heat release "
(Vol). '3
Btu/ft mixture 77.3 96.5
Octane rating - - Research
Method (Min) > III 91 ,/
""'Values not cfomparative with other data, ABSTRACT

" compression ratio, engine supercharging, and hydrogen ad-

Studies were conducted using spark-ignited reciprocating dition to the fuel. Dissociation of ammonia was invest i-

engines to evaluate ammonia as an alternate fuel for cer- gated as a practical means for supplying hydrogen to an en-

tain military applications Conventional engines were found gina,. The study indicates that satisfactory engine perform-

to perform poorly on ammonia. Several pracncat methods ance can be obtained while burning ammonia. AUXiliary

for improving engine performance while burn{tig amrnpnia equipment and controls necessary for-vehtcul ar lise will re-

are described which include increased spark energy. increased quire development.



, .



an ab ch

rn m til

wi ce , .$f' TI

,to fo m wi of cc oi T fc


ei Ii a:



t< h tl a


a "


a il

ammonia is 6.06:1 as compared to a much leaner ratio of about 14.5:1 for gasoline. Based on equal volumes of stoichiometric air- fuel mixtures. the heat content of the ammonia-air r1jixture_is about 800/0 of that of the gasoline-air mixture. T~erefore. since a redprocating engine is essentially a positive displacement device, the power produced with an ammonia-air mixture would not be expected to exceed- aboUt 80% of that produced with a mixture of g~soline and air. if the engine were normally aspirated in both cases, The high octane rating of ammonia is another important factor that must be taken into account when comparing performances of ammonia and gasoline in an engtne. This permits higher compression ratios and supercharging to be used. which improve performance. Also, the heat of vaporization of ammonia is 4.4 times that of gasoline, and the engine' consumes 2.4 times as much fuel by weight for equal power outputs because the heat of combustion of ammonia is 100"er. Therefore. ammonia fueLrequires 10.3 times as much heat. for vaporization as gasoline. This points out the need for a vaporizer when using gaseous ammonia as an engine fuel.

The use of ammonia as a fuel for internal combustion engines has been investigated in Europe. However. vefY limited information is available descriping engine performance. The first practical use of ammonia as a fuel on a limited scale is believed to have been performed by Ammonia Casale Limited in 1935 (2). A second and more extensive application. the Gazamo Process. was tried on vehicles in Belgium during 1942 (3). In the Gazamo Process, the engine was supplied WIth a mixture of ammonia vapor and coal gas. Hydrogen that was present in the coal gas was used to promote the ignition of ammonia. Flow regulation and proportioning of the ammonia vapor and the coal gas were accomplished manually by the operator of the vehicle. This particular program was undertaken because of a shortage of petroleum fuel created by World War II and was termina ted when this fuel shortage was relieved. "-

The experimental program undertaken at the General Motors Research Laboratories was to determine the feasibf l-

ity of burning ammonia in a spark-ignited reclprocattng en-

gine. The major objectives were: ".

1. To evaluate the effects of various engine 'design and operational parameters on engine perf~rmance while burning ammonia.

2. To determine what minimum modifications to a conventional engine are required to provide englfiepetf~rmance~ on ammonia equivalent to that developed whUe using com-

mercial gasoline. ~

Initial investigations were conducted on.a stngte-cyttnder ~ test eng ine.. These engine' studies were of a basic nature and fulfilled the first major objective of the fuel evaluatton program. In view of the encouraging results obtained on the . single-cylinder engine, a conventional multicylinder automoti ve engine was procured. and .perforrnance evaluations. of the engine were begun. Only. preliminary tests have ~. performed on the multtcyl inder engine. As a result. most of the experimental findings discussed in this paper are based ~~gle-CYlinder en,gine tests.


SiHgle-CY1ind~r Engine Installation +Pig. 1 shows the test cell installation of the single-cylinder engine. This overhead-valve engine has a displacement of 27 cu in .. a. bore of 3.375 in .. a stroke of 3.018 in .. and a nominal compression ratio of 9.4:1. The ignition system used initially I was similar to conventional production equipment used on

a 6.cyl automotive engine. with one exception: thestandard high resistance carbon ignition cable was replaced with a conventional high tension cable. A standard AC spark plug of heat range type 44 was used. A fuel-air mixing chamber was substituted for the carburet?r when ammonia was burned. It insured adequate mixing of the gaseous ammonia and air. A positive crankcase ventilation system was installed to safeguard against the possibi lity of a crankcase explosion. Engine airflow was measured by means of critical flow orifices; fuel flow was measured with a variable area' flow meter.

Fig. 1 - Test cell installation of singlecylinder engine l.nd ammonia fuel supply system·


302 \


Con butldin engine througl air ten charge'

Mul was a I charge and eq inurn "e

Th The su

The pressor mania' were e case 0' were n

302 \

Compressed air from the air supply system of the test building was used when supercharging of the single-cylinder engine was investigated. The compressed air was flowed' through electrical heating elements to simulate the rise in air temperature that would be incurred if an actual super-

charger had been used. ~

Multicylinder Engine Installation -The test engine used was a 215 cu in. V-8 engine equipped with a turbosupercharger. Fig 2 shows the engine installed on the test stand and equipped for operation on ammonia fuel. It is an aluminum engine with a nominal corrrpression ratio of 10.25:1-

The turbosupercharger is powered by engine exhaust gas.

The supercharge pressure was Iimtted to 18 in. Hg gage.

The carburetor. which is normally mounted on the compressor inlet flange. was removed and replaced by an ammonia-air mixing chamber. The ammonia and air flows were each manually controlled and proportioned. As in the case of the single-cylinder engine installation. the flows were measured separately and the air and fuel were then in-






traduced into the mixing chamber prior to admittance into the engine.

The standard engine ign'itipn system was employed dur~ ing preliminary engine tests. Ignition components similar to those used on the single-cylinder test engine were then substituted.

Fuel Systems - Gaseous ammonia was tnjected into the engine induction systems of both test engines .. The fuel systems provided to accomplish this were.similar in principle for the two engines but the system for themulticyllnder 'engine was more complex. The fuel systeirdor the singlecylinder engine was installed In the engine test cell and is shown in Fig. 1. Heathad to be,Rlovided to vaporize the ammonia in the mUlticY!JljdereI1~~~yste, whereas sufficient heat was transmitted-through the walls' of the storage vessels in the single- cylinder engine system to cause vaporization.

Fig. 3 is a schernattc of the ammonia fuel supply system • for the multicylinder engine. Fig. 4 shows the fuel storage

portio gine t conta: eratio 120 p! the he water am me chang the oJ betwe water ically flowe. and d centre admit four n

Til gated gatioi to the hydro SUpPI]

Ex ical c in the single ica l ( piing chern

Fig. 2 - Test cell installation of multicylinder engine p.repared for operation on ammonia fuel




Fig. 3 - Schematic of ammonia fuel system for mu-lt icy ltnder eng ine

Fig. ' ticyli


portion of the system that was located external to the engine test cell. Liquid ammonia was stored in six tanks each containing 150 lb of ammonia when full. During engine operation. the saturation pressure of ammonia (approximately 120 psigat room temperature) forced liquid ammonia into the heat exchanger where the ammonia was vaporized. Hot water was used to provide the' heat requ ired to vaporize the ammonia The level of the liquid ammonia: in the heat exchanger was controlled' With a_float switch·l'.'hich governed the operation of a solenoid valve located in the fuel line' between the-tanks and the heat exchanger. Th_e flow of hot water through the heat exchanger was controlled automatically. From the heat exchanger. the gaseous ammonia flowed through two pressure regulators. to reduce its pressure. and through two variable area flow meters and manually controlled throttling valves. The gaseous ammonia was then admitted into the ammonia-air mixing chamber through four nozzles.

The addition of hydrogen to the ammonia. was inve stigated only on the single-cylinder engine. For this investlga tion. gaseous hydrogen from high pressure bottles was added" to the ammonia in the engine fuel-air mtxing oharnber. The hydrogen supply system used was similar tothearnmonta supply system.

Exhaust Gas Sampling and Analysis Procedures - Chemical observations were employed. when feasible. to assist

in the interpretgtton of performance measurements on the single-cylinder and rnulticylinder engines. To obtain chernica~ data" it was necessary to develop spectalized gas satnpiing equipment and sampling techniques and to develop chemical and Chromatographic instrumentation and analy-

Fig. 4 - Ammonia fuel supply and control system for rnulticylinder engine

">' ), . ,'-f'

" 1'~

'-.' ... .r "t,

f':.f ~ ~. 303

sis procedures. In several insta~s;~alculating procedures

• had to be devised to reduce the e~~ntal data.

One important area of interest was the collection and. analysis of gas samples from the exhaust manifold of the stngle-cyttnder engine -. Gas samples were collected. in preevacuated bottles in such a manner. that the concentrations of exhaust gas constituents approxtrnatedthose 11)tlle actual

engine exhaust gas stream. . :

Each gas sample was a!1al>:z@d as r~9BY8~ for amm~nia. hydrogen. oxygen. and .oxi<!.es,"of nitr.,sfandard spectrophotometric procedures were employed to determine the concentration values of ammonia and oxidesof nitrogen. Oxygen concentrations were determined with an Orsat device. Gas chromatography was used to measure the concentration of hydrogen. The experimentally determined concentration values for the exhaust gas constituents. were then expressed in suitable weight units and were substituted int 0 appropriate reaction equations together with related engine fuel flow fnd airflow measurements.

Only three priI\ciple reactions were considered when characterizing the combustion process in mathematical terms. These reactions were: the simple oxidation of ammonia. the oxidation of hydrogen. and the dissociation of ammonia. The oxidation of ammonia to oxides of nitrogen was ignored since this reaction would have little effect on the calculated results,

An iterating procedure was employed to reconcile the reactant and.jiroduct values in the reaction equations. This procedure naturally became more involved as the number


of equations requiring simultaneous -solution increased. In

"the case of the ammonia-hydrogen fuel mixtures investigated. it was assumed that all of the hydrogen that was inducted into the engine was burned. The unreacted oxygen remaining after the hydrogen-air reaction was satisfied was then applied to the combustion of ammonia. In general. satisfactory solutions of. the combustion equations were realized after a relativel~.ew trial calculations were made. Some typical reactant data and exhaust product data ob-

tained from ba lancin these combustion equations are listed in Table 2.

Concentration values for the exha ust gas constituents ha ve been used to determine:

1. The per cent of the ammonia inducted into the engine that actually burned:

2. The effect of engine operation on air pollution. General Test Procedure - The single-cylinder engine was run on ammonia at bothspa rt-throttle and full-throttle settings over a wide engine speed range. Only full-throttle performance of the mu ltlcy linder engine was evaluated while buming ammonia. Normally-aspirated and supercharged modes of operation were i vestigated on both engines. At each engine operating con ition investigated .. performance data were obtained at t minimum spark advance for best

torque (MBT advance) and the leanest a-ir-fuel ratio

for best tor e (LBT air- fuel ratio).

When the single-cylinder engine was run ongaso1ine. performance data were obtained also at minimum spark ad-

.. A

.. A


vance and leanest air- fuel ratio settings for deve loprnent of maximum power. However. multicylinder engine perform ance with gasoline was obtained with a standard produ ction engine of a similar type.

All single-cylinder engine perf0jpance data were calculated on an indicated basis. When supercharged operation of the. engine was evaluated. the performance data were corrected to account for the power required to com press the engine air with a 750/0 efficient compressor.


, Initial Operation on Ammonia - At the outset of the fuel e.J1!lluation program. serious doubt was raised as to whether

an ammonia-air mixture could be ignited and IICImblfstion sustainedin a spark- ignited internal combustion engine. The. limited technical literature found on the subject of ammonia combustion was not encouraging. Therefore, it was heartening WhE)!l ignition of ammonia was achieved in the singlecylinder engine: using a conventional automotive-type' ignition system and a primary voltage of 12 'v. and the engine could be run oves a limited speed range and develop' sorn.e useful work.

. r Fig. 5 presents indicated horsepower and thermal eff i-



on g
the I
the i
de vr
~ que!
an cr
v wI
su ltr
gine .. 24

r-, /
" /-4
• GASOLINE _ - .-/
-1-<..!"'-1· /
.... -r: I I
l iA90UNJ/I. 1\, 2
N/71 ':'
~i ~ <16
I-- . 8
L.,. I I I h 20




W Vl'2


a W8



i5 ~4

o 000

1600 2400 3200 4000

~ .'


Fig .• 5 - Performanceof" unmodified single-cylinder engine on ammonia and gasoline -. - fu ll -thrott le normally aspirated 9D~rhtion. 9.4: 1 compression ratio . .


Table 2 - Reactant and Exhaust Product Data



Ammonia. lb/ hr Hydrogen. lbz'hr Air. Ib/hr

A / F. bJ.p. tical air mixture

Excess monia, Ib/hr

, ,

Ammonia burned. Ib/hr Ammonia burned. % of inducted

Ammonia in 1 pt sample bottle, mg Sample bottle pressure, in. Hg vac Hydrogen. % by vel

Barometer. 29,25 in. Hg abs

Ammonia Hydrogen Water Oxygen Nitrogen

Fuel Mixtures
2 3
13,6 13.0 11.6
0.0 0.0 0.24
7\.8 77,9 79,0
92.2 . 98.8 100.5
}',07 0.15 0,1)
Combustion Data
8.77 8.93 11.32
'64.5 68.6 97:5
Exhaust Products - Experimental Data
22.9 20.0 1.1
3.7 , 2.6 3.5
0.6 0.5 .0.'2 Exhaust Products - .Ca lcu lated Data
Ib/hr ft3/min lb/hr "3 lb/hr ft3/min
ft /min
4.63 1.63 3.92 . 1.37 .0,23 0.08
0.03 '0.'10 0.03 0,09 0.01 0.02 ~
13.95 4,64 14,20 4.72 20.11 6.69
5,29 0.99 5.55 1.04 0.56 0,10
65.51 14.00 67.17 14.38 69,93. 14.92 )



ciency curves that were obtatnedat wide-open throttle while running the engine normal ly aspirated both on ammonia and on gasoline An inspection of these plotted data shows that the engine performed poorly 6n ammonia. The maximum power developedat 2000 rpm was only .17.5~o of the maximum power obtained at 4000 rpm while burning gasoline.

A lso, the maximum indicated thermal efficiency of the engine was 21~o'as compared to 38~~ when gasoline was used. Development of useful horsepower ceased wh en the engine speed was raised above 2400 rpm.w~ile burning ammonia.

The inability to burn ammoniaj1'fectively in the engine was judged to be the primary reason, 'for the poor performance of the engine with this fuel. -Tlris observation was supported by chemical analyses of the engine exhaust gas that disclosed the presence of significantly large amounts of ammonia in the exhaust gas at full-throttle operating conditions. Therefore. considerable effort was devoted to obtaining representative engine exhaust gas sam ples and to deve loping a procedure for. calculating from the exhaust data the per cent of inducted ammonia burned in the engine.

The fol lowing practical corrective actions were considered for improving the ignitability and combustion of ammonia in a spark - ignited engine:

I. Increase in spark energy. L. Multiple ignition.

3. Increase in compression ratio.

4. Fuel additive for pr?inoting the ,com~ustion of ammonia

It was also realized that complete combustion of ammonia in a normally aspirated engine will not result in the developmentof as great a maximum engine power as that obta ine d while burning gasorine unger similar engine operating cond itions. The difference in energy content of equal volumes of stoichiometric gasoline-air and ammonia-air mixtures pTesr~'Y this posstbtuty. Supercharging the ammoniafueled engine'~~ logical means forovercomtng this power disparity. The relatively high octane rating of ammonia makes this a feasible approach.

A 11 of these suggested corrective measures were subsequently eva luated on the single- cylinder engine with considerable success.

Effect of Ignition System Modificati0E - The first method that was investigated to improve the combustion of ammonia in the stngle-cyllnder engine was the modification of

the engine ignition system. The standard coil and 1.5 ohm primary circuit resistor were replaced with a high performance coil and a 1.0 ohm resistor to increase the spark "p-. ergy. The primary voltage was increased from 12 to :1?3f-6

v which approximaresthe voltage used in most current automotive engines.

Full-throttle engine tests were conducted to determine the effect of spark plug gap size on power output. It was found that engine performance was affected noticeably by variation in gap size.and that a gap of about 0.085 in. resulted in maximum power output.

The effect of these ignition system modifications on engine indicated power is shown in Fig. 6. together with in-


dicated power data obtained while burning gasoline in the single-cylinder engine. Also shown in the figure are engine power data obtained while bu mingammonia and ustng a dual ignition system. The maximum power of the engine was increased about 80~0 and useful power could be developed up to a speed of about 3200 rpm by replacing the standard ignition system with the single modtftedIgnttton system. A further gain of about 2~0 in power was realized when the dual modified ignition sys~em was used. - s-

Tests in which each of the two Ignition SY~E:s werelled separately disclosed that greiter engine powe.r was dever-' oped with the spark plug' in the standard location 'than in

the alternate location. Therefore, it is believed that further improvement in engine performance could have-been realized by locating. the second spark plug in a more favorable position.

Fig. 7 presents plots of the MBT spark advances and I13T air-fuel ratios established while operating on ammonia and using the single modified ignition system. Also plotted is the MBT spark advance curve for gasoline. A's shown in the figure. the spark advances for ammonia are considerably gre'!ter than those for gasoline. indicating the relatively slow

burning rate of ammonia. ,. ,

The maximum power air-fuel ratio lor ammonia varied from about 6.1: 1 to 6.8:1. These a ir-fue l ratios are slightly leaner than the stoichiometric air-fuel ratio of 6.06: 1 of

an ammonia-air mixture.

A Ithough the ignition system modifications resulted in

a considerable improvement in engine performance: the power differential between gasoline and ammonia fuels was still greater than the theoretical difference. Chemical analyses of the engine exhaust gases indicated that an_ appreci-

a::: w ~ 161--+----I1-o~""\--+--


W If)

~12 I o W ~ 81--+---

u i5



1600 2400 3200 4000


Fig. 6 - Increased power of single-cylinder engine th rough ignition system modifications - - ammonia fue 1. full -throttle normally aspirated operation. 9.4: 1 compression ratio'

able a
gine w
head r
ious c(
sian II
were (
tern a:
It ,
to 15:
deve k
did re
sian P.
that 0
.. power
/ rpm. c
I In (
L highel
\ can bl
-, prodUi
/ (!;
Fig. ~
campi I




able amount of ammonia was still passing through the en-

gine without burning. "

Effect of Increased Compression Rat~ further improve the combustion of ammonia in the single-cylinder engine, compression ratios greater than 9.4: 1 were investigated. Increases in engine-compression ratio to 11.5: 1. 15: 1. and 18: 1 were accom plished by substttutin g steppedhead .ptstons for the original flat-head piston,

Fig, 8 presents indicated horsepower curves for the various com pression ratios tested and the power curve obtained while burning ga_soHne in the engine at the, 9.4:1 com pression ratio. These ammonia data and all subsequent fullthrottle single-cylinder engine' data discussed ~ this paper were obtained while using th~ single modified ignition system and a primary voltage of 13.6 v.

It will be noted in Fig, 8 that a sizable gain in eng[ne power was obtained when the compression ratio was increased from 9.4:1 to 11.5:1. The maximum power was increased 590/0 Further increases in compression ratio from 11.5: 1

to 15: 1 and to 18: 1 had negligible effects on indicated power development at engine speedsbelow about 2400 rpm. but did result in increased engine power at higher speeds, The maximum-power was increased 680/0 wt th the 15: 1 com pression.ratlo and 840/0 with the 18: 1 compression ratio above ihat obtained with the 9.4: 1 compression ratio. Maximum

._ power occurred at 3200 rpm in the case of both of t11ese higher compression ratios. However at speeds above 3200 rpm. engine power fell off rapidly.

In comparing the power curves for ammonia at these three higher compression ratios with the gasoline power curve. it can be seen that at speeds below about 2600 rpm the power produced with ammonia wasabout 800/0 of that obtained with





a:: 120
u 100
a:: 80
Vl 60
~ 40
~ 20
a:: 0
~ ~
>--0"_ v-...., b
.. ~

~~S~INj:_-f-- -

,...",..-- I
- ~-~
--"v, < o


7 a:: ..J W



6 a::

o~ 40c0 .

8JO 1600 2400 3200


Fig. 7- MBT spark advances and LBT air-fuel ratios for Single-cylinder engine -- ammonia' fuel. full throttle normally aspirated operation. modified ignition system.9.4: 1

compression ratio



gasoline, This is approximately the theoretical power ratio for the two fuels based ~n heating values and stolchlometrlc air-fuel ratios.

Fig. 9 illustrates the Influence of engine compression ra.....




GASOLINE ------.-y' 9.4:1 C.R.



~ 16 f_+----~f_--~--t--. -,--+------1 LU


til 121__~---___,1__-----


U 5 ~

1600 2400 3200

ENGINE SPEED- RPM Fig. 8 - Improved performance of single-cylinder engine due to increased compression ratio - - ammonia fuel. fullthrottle normally aspirated operation. modified ignition



« i
0 !
w 501
u I
~ 40
u_ i
;l- Fig. 9 - Improved combustion. of ammonia in Single-cylinder engine by increasing compression ratio - - full-throttle normally aspirated operation. modified ignition system

.. tlu
ga! Fig due syst

'" .



tio on the percentage of the inducted ammonia burned in investigated .• The increased charge density due to super-

the engine. The beneficial effect of increased compression charging should overcome the theoretical power::diffilrential

ratio became measurable at an engine speed of 2000 rpm for the two fuels.

and became progressively more pronounced as the engine : .. · The Indicatedhorsepower data ohta Ined while burning

speed was increased further. These data emphasize the prob- ammonia at supercharged engine conditlons-are shown fp

lem that is encountered when burning a fuel with a decid - Fig. 10. Also shown again for the YJ.lrpose of comparison is

edly slower flame speed than that of a hydrocarbon fuel. the indicated horsepower curve for theengtne tliit'twas ob-

In this case. the combustion of ammonia \ ... as promoted by t~ined while operating normally ssptrated ori:g£ioline at a

increased cylinder pressure. temperature. and rurbu fence compression ratio of 9.4: 1. A superchargepressure of; 18 in.

that accompanied an increase in compression ratio. Hg gage was used. Failure of the 18: 1 compression. ratio

Piston shape may be partly responsible for the improve- piston due to insufficient diametrlcalclearauce precluded

ment in the combustion process realized by increased corn- testing of this piston at supercharged' engine operating con-

pression ratio, The greatest gain was made~ when the flat- ditions. ' ' -

head piston (9.4: 1 compression ratio) was replaced with a The curves in Fig. 10 sh~w t~i'I! compression ra?;A, had a

stepped-head piston (11.5:J compression ratio). The pro- " negligi~ effect on supercharged engine perfo2n~t! at

trusion of the uPFer step of each stepped-head piston into speeds b~9W approximately 2400 rpm. However. at higher

the cylinder head plobabty caused an increase in gas tur- speeds'¥1d,eaSing the compression ratio resulted in a sig-

bulence in the,cOl.nbustion' chamber and thus improved the nificant tmpcovement in engine indica"ted power.' These

combustion of ammonia. Further tests would have to be power ga!ns af.high engine speeds were the result of increased

made to deterrntnj, which factor. compression ratro or pist'cm burning of-the \mmonia in the engine (Fig. 11).

head shape. was aide tesponsible for the 'improved engiJ!'e,. Fig. 10 Shows'tISO that the engine power developed with

performance. No attempt was made tli~ev~lop a cornbus- ammonia over th~' entire engine speed range tested can be

tion chamber shape that would contribute to a more rapid " .}l!ade to exceed th obtained while burning gasottne at a

burning of the amrnoata.. . . : J ;:\., ~,A::l compression ra 0 and normally aspirated engine op-

Effect of Supercharging - To obtain a maximum·pow~r'·. erating conditions, 1llis las accomplish~dbyusingthe 15:1

output with ammonia commensutit'e with that obtai~ed with compression ratio pist~nd a supercharge pressure of 18 in,

gasoline, supe!yharging of the single-cylinder engine was' Hg gage. The other compression ratios evaluated resulted

. . in power outputs greater than those for gasoline over most of the speed 'range. but fell below those for gasoline at high

.'.speeds, -

Part-Throttle Engine Considerations - Previously described engine tests indicate that a spar~ ignited ammonia -fu e le d engine can be supercharged to provide fu l l-thrott le performance commensurate with tha; ;eaiized in current automotive gasoline engines. However. adequate full-throttle performance is but one of many requirements of a vehicular

Fig. lO - Increased performance of.single-cyltnder engine due to supercharging - - ammonia fuel. modified ignition system



Z 0::

EO 80



o 60 2:

2: ... «

8 40


2 20

u, o

~-~-.x J
--11- ..... ""
SUP '.

~ I
I /
I .0 ,

800 1600,,- 2400 3200




Fig. 11 - Improved combustion of ammonia in single-cyl-

inder engine by supercI1Arging -.- modlf ied ignition system,

9.4: 1 compression r~tio' •



engine Satis{t,ctory part- throttle performance cannot be overemphasized because a. vehicular engine is operated most of the't ime at pan loads. In the case of vehicle operation. the range. the amount of fuel required for acceptable range. and-the operating cost of the vehicle a~each dependent

on the part load fuel economy of the elyne.

Of a ll of the engine modifications evaluated on the single-cylinder engine. only compression ratio and ignition . systems should have -signtftcant influences on pan-throttle performance, Fot this reason, the effects of these two en~ine variables on part load efficiency of the single-cylinder a mmonia - fueled eng Ine were investigated.

Fig. 12 shows the effect of compression ratio and dual igHit10ij on indicated' thermal efficiency of the single-cytilJ.der engine at various load settings. Typical data are presentedthai were obtained ,at an engine speed of 1600 rpm. For comparison purposes. p~rt-throttle data .for one cylinder of a mu lticyltnder engine are included that were obtained dUrlJ1g operation on gasoline .. The displacement per cylinder of this engine is equal to that of the single-cylinder test engi)1<\. The roa~ load requirement per cylinder. of the mult icy l tnder engine is indicated also on the figure.

It can be seen that for each compression ratio investigated. the indicated thermal efficiency of the single-cylinder engine dimtnished rapidly as engine' load was reduced from full load to road load. At full throttle, the thermal efftcrency approximated that of the multicylinder engine but was considerably less than that of t·he gasoline engine

in the vicinity 'Of road load. Only a slight im provement in

5 8
4 2 3 4 5' 6


Fig, 12 - Minor improvements ill part- throttle thermal ef-· ficiency of .single+cy lirrder engine due to increased compression ratio-and dual ignition ~- ammonia fuel. mod: ified ignition systems . .'






'!I • ~

single-cylinder ~gine efficiency. was realized with dual ignition. As'a result: ff was felt that satisfactory part-throttle "performance of a spark-ignited ammonia-fueled engine could "not be realized by engin"e modifications.

, Hydrogen Enrichment of Amrn9,nia ~ The possibility of enrichening the amfnonla \'litl1hydl-ogenVas considered as

a means of improvlrrg the part-throttle iJerformance of a spark- ignited ammonia -fuele d engine.,:.]T9e"use of hydrogen to promote. the combustion of ammoriltlsa logical choice . because it could 6e produced on the veliicle itself by decomposing som_e of the ammonia fuel in a catalytic dtssociator. Hydrogen can be ignited readily-and has a high flame speed.

The addition of a relatively sman amount of hydrogen

to the ammonia fuel was found to re~ an al'j]re"ciable Improvement in ·the part-throttle performance of the singlecylinder test engine, A 2.50/0 by \"eight addition of hydrogen 'was found to result in the. best performance' at the speeds inyestigated.· The effect of this amount of hydrogen .en-

. ricltment on indicated thermal efficiency of the engine at part load is shown in Fig. 13 for an engiue speed of 1600

< rpm. As iJ Fig. 12. comparable multicylinder gasoline engine data are plotted.

It will be seen in Fig. 13 that hydrogen enrichment resulte d in a sizable gain in indicated thermal efficiency of the stng levcyltnder engine over the entire load range investtgated. 'The engine efficiency values obtained with the ammonia-hydrogen mixture. were higher than those for gasoline over most of the load range. In view of these test results and supporting data at other engine speeds. it would

36 « ?


32I----+---IdY1 I· I' '" ._

-/ ."~5.,"YOROG'" )'

.. 28 I--ADDITION ", -_

~ ,~",,, I I 01 "I--r--'

w 24 I zrtGA.SOUNE_!-

u I 2151N3 V-S





f: r. '-! ~

,~ 8 / I I)

~ 4 ~___+----ji_-t--__tl _ __+-_+-il _--j l--ROtD L0Af (TYPiCAL)








Fig. 13 - Improved pan-throttle thermal efficiency of sillgle-cyltnder engine due.to hydrogen addition to ammonia

- - modified Ignirion system. 9,4: 1cornpre;sion ratio' ~



appe .sible form ther. evaf on fl supe

F richt malJ only to CI engi equa

, amn real: .thej crea mOT( that cent. abou

F gen The spee, mix! passe sure inch


ffi 12
~ e
~ 4
. ~- 0
dl!.e t
-: ... __ - - fu
. press:
~\\.: -

.. '-.


appearthat hydrogen enrichment of ammonia offers a feasible, ~berne for obtaining satisfactory pan -threttle per- . formance of a spark- ignited ammonia-fueled engine. Further ,tests were performed on the single -cylinder engine to eval\}'Ate the influence of hydrogen addition' to ammonia on full-throttle 'performance. Both normally aspirated .and supercharged eng U;e eperations were investigated. .y>j

Fig. 14 iIItfitrates the beneficial effect of hydrogen enrichment on engine indicated power at full-throttle normal ly' aspirated operating con9itions. It can be seen.that only a very small amount of hydrogen addition is suff'icient to cause a significant increase in engine power. ~xiqipm engine power was approximately 'doubled when hY2i~gen equa I to 20/0 by lVei~ht of the fuel mixture was adcf8:J to the

. ammonia. Further gainsof only a negligibieamount were realized 'Over an engine speed range of 800~3600 rpm when . theconcentratton qf hydrogen in the fuel mixture was in-. creased to. 30/0. At 4000 rpm. a 30/0 hydrogenad5lif}fp.

more effective than a 20/0 addittori. In geI].eral. . nd

that the engine pOlver began to decrease slightly as t centage of hydrogen in t?e fuel mixture w'as increased above

about 30/0. . . . -.

Fi1'>' 15 illustrates the v'a.riant beneficial effect of hydro" gen addition on indicated power of. the supercharged engine. The illustrated data were.obtained at~ch ~f three" engine speeds by increasing the hydrogen concentration in the ~el mixture in small incremental steps until the engine power passed through a peak value. An engine supercharge pressure. of 18 In. Hg gage-was maintained at all times, Also included on the figure are maximum rnginepow~r data ob-

1 ·0.5·M-I2

I '.~

:%-~-+---~-+I-I --~-I,~ ~0:H2 r


'f I~

.• ' J

';609 2400 3200



.~ .

. , "

Fig: 14 - Improved performance of single-cylinder engine


d~e to- varying· amounts of hydrogen addition to ammonia

- - full-throttle ncrmally -asptrated operation. 9.4: r com-

pression ratio



aOjl ta ined I'I~ runmng the engine on' arnrnpnia ~nly.-~t cll~' same ru~e pressure a~d while operating the engine , OD gasoline at normally aspirated conditions. ,,'

e. It. ,vi.P be s~en in-the figure that hydrog~n ac;jditlon had

~ " if' , - . . . ..

a slight detrimental effect on engine power dev~lop'mll~t ,

at the 2000 rpm 'investigated'speep, but enhanced markedly en!i!£1e power outputs at s~e"gs ~{3600 ~nd 4000 rp~'. Optimum concentrations o fiJlyHrogen 9f apprQximat~ly 1.35%. and 1.190/0 in the fuel. mtxture \ve!e esiagUsh.ed't6rengine:. . speeds of 3600 and 4060 rpm res~tiVillY. employing :'-.., a hydrogen concentration of about 1. 0 re-sulted' in Cn"Bine: , power development over the entire en . e sp~e.d ta'nge' in;.: " vestigated that was, equal to or exceeded the l;ingine~powElt';~.

\ developed while' burning g~oline in the ~ngine' under nor- '

. mally aspirated operating conditions. . . . -

, Chemicalhna lysesof engin~gas s~~lr-,s c0.uected ·dur-.

ing these tests-revealed the role. of hy~_g_en asa cornbus- ' .

tion prorhoter.for ammonia. At most of d'ill test conditions ..

the addition of hydrogen t(l;.~rl1inonia was found ~o increase theperc~ntage of inducted ammonia burned in the engine.' ... - The~egr~e to which hydrogen eurichiaent abetteq the corn-. bustion of ammonia tended \0 vary directly with-the degree

of ineffectual bu"rning of the ammonia itself In the engine, Where bombustion of ammonia was relatively poor (for examplfl,~ during- part-rhrortjer-ett normally aspirated roii-


, i I


o· W'

~ 12~4----+----+---·LO ~ o

• z



1600 2400 '. 3200

EKlGINE ,$f".EEO- R,pM

Fig. 15 . Varying improverfients in perform~ce of stngle-' cylinder engine due to 'hydrogen addition to ammonia - - " supercha rged operation. mo,ttHi.ed "ignition' system: 9.4:1

. s;. '_ " " . .

compressIOn ratio r " • • • ,';- .




f' :: ·0

. . .,


-' ;,.'


. '

. 'c


"',,: <

If) " ).
-~l 310

throttle. and high-speed supercharged conditions). hydrogen addition to the ammonia was beneficial, On the other hand. supercharging the engine at low engine speeclprovided optimum bumin&. conditions for the am11!0nia. aqd _hence. hy-

drogen add ition was not needed. . .

The anomalous fact that engtne power 8ecliIled ~lightly at anfull-thronle'fpe~ing conditions investigated when hydrogen enrichment was increased past the optimum amount may b~ attr ibuted to the widely different burning tates of the two fuels which necessitated. a compromise MBT spark advance.


. It Was shown' by the single-cylinder engine tests ~han~ith certain engine modifications and the addition of a small amount of hydrogen to the ammonia, satisfactory performance of a spark- ignited rectprocattng engine cou ld be ,expeered. Therefor.e,· it was decided to prove 'this more conclusively by o_perating a multicylinder engine onarnmonia. The p~Hminary teststhat were conducted verify the belief thafamulticY,tirider engine could be developed that would produce power outP\l.¥hile burning ammonia equivalent

to thai optaineci whil€'operating normally aspirated on gasoline ..

. Normally Aspirated .Engine Operation{" Fi~.- 16 presents two full-throttle brake horsepower curves'roal w~re obtained




~ I ''!-o . ~-;.~.'

! u,.

I ./ '/

'GOO 2400 3200


fig. 16 - Performance of multicy linder engine on am~enia and gasoline' - - full-throttle normally aspirated operation. standard and 'modified ignition systems





.< p

w)lne bumtngamrnonta in the multicytinil~ test engine at

normally aspirated engine operating conditions. Curve A •

refers to data /that were taken with the tstandard ignition sys-

tem Installed. Data for curve B were obtained after the tg- ~ "

ninon system had been modified. ' Also' shown is 'a brake liorsepower cufve (C) for a s\n1nar engine of lowef compres-

sion -ratio (8.75; 1,j's compared to 10.2~':i:r;tfiat 'was:6btained

while operating tl~e;1gine porrnally ascp~.iteirOl\ gasoline.

A cornpartson Q~ the ,power outputs sho~s~~.l.legrad.at~on in maximum power of approximately 87o/~w:1i.en swttchlng from gasoline to ammonia fuel.,' W1ieff the modified ignition sys-

tem was install~,d, the loss in engine power was reduced to

700/0. It is realized tliat this comparison favor .. slfghtly the arnmonia-jileled engine since it was rug at maximum power

. spark advances whereas with gasoline tlie spark timing was 'retarded from the best power -spark advances. . "

Two othertjpes of ignition systems~ere tested, in. an attempt to-Improve engine 'performance while bum lng amrnonja. These were a capacitor discharge system' and a con" tact operated transistorized :s,ystem. Also. severalcoil-re'sistor combinations were tested. Spark plug gap sizes'. were varied from 0.0..Q{j to 0.100 in. when testing the qifferent systems. However, none of tliese 'ignition systems resulted

in engine performance that exceeded that obtained while using the modified ignition system described prevtously.vtn

. I'

fact. mostof.them'resulted in inferior engine performance »

Superchars,ed Engine Ogeration - Fig. 17 prese,nt~ engine



___ --

100 1-t--------+--------+-------7,,/'..- ..;

I / i

G3ASOLINE 8 75":C R /

NORMALLY--AS'f"$ATED ').-:.4:


~ ~

202' T w a::


I{l ',2 Ii



--~-+-----/"L_-r----~--r-------~8 ~

I U a:: w c. ~ V>

o 800 • 1600 2400 - -' 3200 400~


Fig. 11 - Improved performance of rmrlticyltnder _~e -'due to supercharging - - ammonia fuel.:modified ignition system





brake h gine or operati tween! lislied 1

Aco ing was Beglrini gan to desired rpm. I to llmf

The over th ing the burning [mum I

·96 blip than th oline. engine difficu

It sl' are in I


in. Hg bene fie 2800 rl de ere a none xf gence I ably Ie ability



Fig. Ii specifi monia tern

.~ 1


brake horsepower curves obta ined during supercharged engine operationon ammonia and normally aspirated engine operation on gasoline. Also shown is the relationship between supercharge pressure and engine speed that was established while running the engine on ammonia fuel.

According to the supercharge pressure curve. supercharging was not evidentInftil the engine exceeded 1200 rpm. Beginning at about 12.00 rpm. the supercharge pressure be.gan to increase rapidly with engine speed and reached the desired maximum pressure of 18 in. Hg gage at about 2800 rpm. A hove 2800 rpm. the turbine bypass va lve was opened to limit tfl'Supercharge pressure to 18 in. Hg gage.

The plotted data show-that 'the engine power developed over the entire speed range when burning ammonia and using the turbosupercharger was less than that produced when burning gasoline in the normally aspirated engine. The maximum power realized with ammonia as the fuel was about

96 bhp at 3500 rpm. This maximum value is about 10% less than the maximum engine power obtained when usi~g gasoline. Engine performance data were not obtained at an engine speed of 4000 rpm while burning ammonia due to difficulties encountered wlth the ftiel supply system.

It shc~ld be noted tfiat the tW? brake ~orse po we r cur('es -are in closest agreement at an engine sp ed of about 2800 rpm At this speed. the desired superch rge pressure of 18

• in. Hg gage was achieved while burning ammonia. and the beneficial effect of supercharging was a maximum. Below 2800 rpm. useful eng me work derived fr m supercharging decreased progressiveJ.y with speed and ecame practically nonexistent at engine speeds below 1'20 rpm. The divergence of the two power curves above 28 0 rpm is due probably to the progressive' decrease with e gine speed in the ability of the engine to bum ammonia efficiently.

Fig. 18 presents brake specific fuel consumption. brake

O'W\~800----L_--~~--_L--.-2~400----~--3-200L---~--~~ ENGINE SPEED- RPM-

Fig. 18 .- MBT sparkadvances, I1lT air- fuel ratios and brake specific fue I consumptions for multicylinder engine - - am. monia fuel. supercharged operation. modified ignition system

. 311


thermal efficiency. and spark advance curves for the supercharged ammonia-fueledand normally aspirated gasolinefueled rnulttcyltnder engines. A lsoshown is the air-fuel ratio curve for the supercharged-ammonia engine operation. The curves obtained while burning ammonia in the engine • extend only through an engine speed of 3200 rpm. Although corresponding engine power data were obtained at3600 rpm, the fuel system failure that was mentlonedprevlously prec~uded the recording of a full set of engine- performance data.

As shown on Fig. 18. the two b~ke thermal efftciency curves agree closely over the engine ;;peed range of 800-

. 2800 rpm. However, at this higher speed the curve for ammenta began to fall off while that for gasoline remained about constant.

The brake specific fuel consumption of the engine while burning ammonia was about 2-1/2 times as great as that obtained with gasoline. The marked difference in heating values for the two fuels ~s primarily the cause of this large difference.

Spark advance requirements for the two fuels are also significantly different. The spark advance determined during operation on gasoline increased gradually from 5 deg btdc at 800 rpm to 24 deg btdc at 11000 rpm. A. much greater MBT spark advance was employed while burning ammonia. It varied somewhat randomly between 103 btdc and 115 deg btdc over the engine speed range investigated.

The I1lT air-fuel ratios established for operation on am rnonia varied between 1f.2: 1 and 6.7:l over the engine speed range. These air-fuel ratios are slightly leaner than the stoichiometric air-fuel ratio of 6.06:1 for ammoni a and are much richer than the stoichiometric air- fuel ratio of about 14.5: 1. for gasoline ..

Again. samples of-the exhauit gases were collected and. the' constituents analyzed. At one supercharged engine operating condition. it 'jas determined that the per cent of inducted ammonia burned in the eight cylinders varied between 93 and 96.5"/0. An overall ammonia-burned value

of 94.2"/0 was determined from analyses of engine tailpipe gas sam pies.

trace' engtn ·adval advar tive E Ing Co

PE aboui aboui the c large were thl~ \ curre sure, were rappl

T Si@11 press chaq cylin cleo


tion ular .gtne.

engh and I

E gine ture ichir arnrr requ



Engine Noise - No engine knock was detected during any of the tests described in this report. Although the octane number of ammonia. is not known. these tests indicate that. it is exceptionally high. However. during operation of the single-cylinder engine at a compression ratlo'of 18:1 at speeds of 2400 rpm and above. a rapping noise similar to that produced by heavilyIoaded diesel engines occurred when the spark advance was set for best power., .This noise could be eliminated by retarding the spark Slightly although

this resulted in a slight power loss. .l- .

During norma lly-asptrated operation at-the 18':1 compression ratio. several cylinder pressure-time' traces were obtained to determine the cause. of the rapping noise. fig . 19 is a photograph of an oscnlogram showing several such

Fig. Iatk





traces. These pressure-time traces \'1~btained with the engine operating at 3600 rpm -- fut l throttle . TIle spark ·advance was set at 112 deg btc which was the MBT spark advance. The six upper traces were taken during consecutive engine firing cycles and the lowest trace is the motoring compression curve.

Peak pressure varied from cycle to cycle ranging from about 720-1700 psi. and the rate ol pressure rise varied from about 25 to 100 psi! deg. This is considered to be htgh for the conventional spark ignition engine. It is felt that these large variations in peak pressure and rate of pressure rise were the ca~se of the rapping noise.· Further evidence of this was. obtained by observing that the rapping noise occurred in phase with the variations of peak cylinder pressure. that is. when several nearly equal peak pressure traces were followed by an extjernely high peak pressure trace a rapping sound was heard.

This rapping noise was not present at the lower com pression ratios. Oscilloscope traces obtained at the 15: 1 compression ratio during both normally aspirated and supercharged engine operations showed less variations in peak cylinder pressure and rate of pressure rise from cycle to cycle.


Contribution to Air Pollution - The effect on air pollution must be weighed seriously when considering a vehicular application of the spark-ignited ammonia-fueled en-

. gine. Even a relatively small amount of ammonia in the engine exhaust is to be a voided because of its irritating odor and toxic effect.

Emission of ammonia from the Single-cylinder rest engine was minimized by burning an ammonta-hydrogen mtxture required for optimum engine performance at the stoichiometric air-fuel ratio. A stoichiometric mixture of 980/0 ammonia and 20/0 hydrogen on a weight basis satisfied this



Amm concentrations in the Single-cylinder engine

. . .

Fig. 19 - Cylinder pressure-time traces showing large var-

iations of peak pressures and rates of pressure rise for six consecutive firing events in single-cylinder engine -- ammonia fuel. full-throttle normallyaspirated operation. 18: 1 compression ratio


. ~'/"'. .• . I

/~-. .




exhaust were calculated to be about 81.000 ppm on dry basis (water formed Qy. combustion excluded) at.3600 rpm and full-throttle. Adding 20/0 by weight of hydrogen to the ammonia and running ata similar engine operating condition reduced the concentration of ammonia in the engine exhaust approximately 940/0 to 5300 ppm. This significant decrease in ammonia emission was the result "Of a decided improvement in combustion efficiency. when hydrogen was addeq

to the ammonia. The per cent of inducted ammonia burned in the engine was Increased from apprcxlmately' 69 to 97.50/0.

These computed ammonia concentrlltldn values are ap: -preciably greater than the tolerable .litnitestablished for

tlie human body. However. when these quantities of ammonia are discharged from the englrre tailpipe and mix with the ambient air. tolerable ammonia concentrations should result if sufficient ventilation is provided. This engine exhaust gas condition is similar to tha t which is experienced when carbon monoxide is emitted from the tailpipe of a gasoline engine.

Detectable amounts of hydrogen were emitted from the engine while burning rich air-fuel mixtures of both ammonia and ammonia-hydrogen fuels. Approximately seven times as much hydrogen was discharged while burning a 90"/0 theoretical air mixture of ammonia and hydrogen as was discharged while burning a 920/0 theoretical air mixture of only ammonia. ,

While a considerable improvement in the combustion efficiency of the single-cylinder engine has been realized. still it is reasonable to expect that a measurable amount

of ammonia will be emitted from any spark-ignited ammonia-fueled engine. If a decided reduction in ammonia emission from an engine should be required. it may have to be accomplished in the engine exhaust system. Dissociation

of ammonia to hydrogen and nitrogen. absorption of ammonia. or chemical conversion of ammonia to innocuous constituents a,re various means that might be em ployed.

Exhaust gas samples were collected and. were analyzed for concentrations of oxides of nitrogen during maximum power operation of the single-cylinder engine at several different engine speeds. It was found that the concentration

of oxides of nitrogen increased with increasing indicated thermal efficiency of the engine and ranged from a concen'tration of about 200 ppm for a thermal efficiency of 150/0

. to 1200 ppm for a thermal efficiency of 300/0. If gasoltne

were to be burned in the engine at similar operating conditions. similar concentrations of oxides of. nitrogen in the

engine exhaust gas .would be expected, .

Engine Oil Analyses - During the course of this i~stigatton, chemical analyses of the engine lubricating oil were made to determine if the use of ammonia as an engine fuel would have an adverse effect on theorl. Approx irnate ly 5 gal of commercial 20 weight oil were used to fill the lubricating systems of the single-cylinder and multicylinder test engines. This large quantity of oil was needed because the oil was circulated through a heat exchanger.

New oil was provided at the start of engine testing involving the use of ammonia as fuel. Samples of the oil

wei lasi aft! ·pOI ell! pIe ana

ter: oft of I oil die pta con

occ whc cy l: be:

WhE it iv gill I

of r crai whi non Ilia mo:


feci in t Iorn was sam pow enri

. droj this trib cati systl


_ qu ir , "sula

wou tran able its u

gen of h


were extracted at various times during the engine tests. The last sample from tbe single -cytinder engine was collected after approximately 75 hr of engine operation at diverse .power and speed settings. while that from the multicyltnder engine had 46 hr of simi lar opera tion. A 11 of these oil sampIes together with a sample of the new oil. were chemically

analyzed. j..

The chemical analyses disclosed that no noticeable deterioration of the oil had occurred. The total acid number of the oil remained ap,Pl'0xlmately constant as did the amounts of barium. calcium. phosphorus. and zinc additives in the oil No resins were found in any of the samples which indicated that no significant oxidation of the oil had taken place. The viscosity of the oil remained approximately constant.

Crankcase Gas Analyses - The possibility of an explosion occurring in the engine crankcase was considered seriously when the fuel evaluation study was instituted on the singlecylinder engine. It was reasoned that piston blowby would be greater when burning gaseQus ammonia and hydrogen than when burning liquid hydrocarbon fuels. As a result. a positive crankcase ventilation system was installed on the engine.

Several engine rims were made with the specific purposes of measuring hydrogen and ammonia concentrations in the crankcase gas. Samples of the crankcase gas were collected while running the engine at a variety of supercharged and nonsupercharged conditions and while burning both ammoIlia and ammonia-hydrogen mixtures. No hydrogen Or am,..rnon ia was detected in any of these crankcase gas samples


Single-cylinder engine tests have demonstrated the effectiveness of hydrogen enrichment of the ammonia fuel

in the realization/,f accept able part-throttle engtne performance. It was found that 2 .5o/~ by weight of hydrogen was the minimum required amount. On the other hand, these same tests also dJ.sclosed that suitable maximum engine power might be (reveloped without recourse to hydrogen enrichment of the ammonia.

A t a first glance. it would seem reasonable to supply hy-

_ drogen and ammonia from separate fuel tanks. However. this would not only require the production, storage, and distribution of JWO fuels, but it would also introduce complications through the necessity for an add lt'ional fuel supply system on the vehicle. A most critical disadvantage involving the use of hydrogen is the cryogenic temperature re-

. quired to maintain this fuel in the liquid state. Heavily in"sulated equipment in addition to specialized techniques would be required for its storage and handling. Losses in transfer and storage would be severe. Due to these unavoidable losses and the wide flammability limits of hydrogen. its use would be very hazardous.

The serious shortcomings ofa separate vehicular hydrogen supply system would be eliminated if the desired amount of hydrogen could be either dissolved in the ammonia at


the Energy Depot or produced by dissociating some of the ammonia, Studies have shown that sufficient hydrogen cannot be dissolved in Itquid-ammouta to provide the desired concentration (4). Therefore. the dissociation of ammonta was investigated.

The following simple chemical equation describes the ammonia dissociation pr<?cess:



(Ammonia) - . (Hydrogen)


To promote this decomposition process. a catalyst may be used (for exam pIe. nickel or iron) and heat must be SIJ Pplied (5). In a vehicle. the engine exhaust gas provides a readily available source of free heat. l~t electrical energy provided by an engine-driven generator !'I0Uld also be a possible source of heat.

From. a theoretical standpoint. a' dissociator. heated by the engine exhaust gases. shou ld fulfil'l i111 of the needs of an automotive ammonia-fueled engine. The temperature of the exhaust gas emerging from the cylinders of a con.ventional automotive engine is sufficiently high at all normal operating conditions to insure practically complete dis-

sociation of the ammonia. Unfortunately from a practical standpoint. the dissociator cannot be located close to the cylinder exhaust ports and the reaction cannot be completed because the fuel would not remain in the hot zone long enough. Consequently. a catalyst must be employedt~compensate for the short reaction time and for the inability to make use of max irnum 'exhaust gas tem peratures.

One of the points in favor of an ammonia dissoc iator is the fact that only a part of the ammonia supplied to the engine need be decomposed to provide the aptlrnu m concentration of hydrogen in the fuel mixture. The dissociation

of about 13% by weight of the ammonia is requir!d to provide the desired 2.5% by weight of hydrogen in the fuel mixture.

It was the original intent; to undertake the development of a full-scale dissociator in ~veral stages. The first stage was to ~e a basic study-of the dissociation process -- to evaluate different techniques for promoting dissociation. This basic study was to be followed in succession by the construenon of dissociators of increasing size for the single-cylinder and multtcytinder test engines. Only the initial stage of

the development program has been completed.

The basic ammonia dissociation study was conducted with miniature models using available laboratory equipment and procedures. A gas chromatograph was used to determine

the degree of dissociation accomplished. TIle concentrations of both ammonta and nitrogen in the effluent from the dissociator were measured. The three different 'designs of ammonia dissociators evaluated in the laboratory are illustrated in Fig. 20. Cross- sectional views are presented, and the small size of these models can be appreciated by comparison with the scale. In each model. nickel in various forms was used as the catalyst. The desirable characteristics

. 31

of ·m(


.(fe tri of les tnt by suI'

tio all Te E1( of

the COl OIl( the les: fill d!a

act is ( spe hal



Fig. ciat


Z 15
U 10
__: <f 5 '314

of nickel. as well as its established use in comme/cial ammonia dissociating devices. prom (Xed the initial use of this

material in the mlnature models ,

Model A in the figure contained a bed of nickel catalyst

,(for example. powder or shot) that was heated by an electrical furnace to various selected temperatures. In the case of Model B. the tubes were constructed of lnconel or stdlnless steel. betl1 of which contain nickel. and the temperature ofthe tubes was maintained at the desired temperature by an electrical furnace. Controlled electrical energy was supplied to the nichrome filament in Model C.

Figs. 21A-21C present some typical ammonia dissociation values that were obtained w ith these three models. In all cases. the ammonia flow rate was kept below 0.25 CFH. Temperatures ranged as high as 1000 F in Models A and B. Electrical energy to Model C was held below a maximum of about 20 w.

From comparisons of experimental data obtained with these models. it was concluded that the particle bed reactor containing nickel shot was the most effective one and the one best suited for engine application. Results obtained with the Inconel tube reactor were superior to those of the stainless steel reactor. The electrical power requirement of the filament 'type reactor was judged to be too high for immediate application to an engine. Actually. the filament reactorhad been designed for quite' a different purpose and it is conceivable that if a reactor of this type had been buil t specifically for this study. a more efficient device would

ha ve resulted.

A detailed analysis of the data from the particle bed reacto! (nickel shot catalyst) yielded some important informa-

,______.. 100 In

Fig. 20 - Experimental microreactors for studying dissociation of ammonia

~--~--,--- --



tion of a practical nature .. Catalysis appears to be a significant factor in the dissociation process at tern peratures below ·900 or 1000 F. 'At higher temperatures. the thermal effect appears to be the predominant OJ1C. J:,. dlssociator of a size reasonable for engIne use was scaled up on paper from the

m icroreactor data. Such a hypothetical dissoclator should satisfy ammonia-fueled engine requirements in the high part-throttle andfull-throttle operatingtanges of a conventional automotive engine. Its effectualperformance at engin~ Idle and in the low part-throttle operating range is debatable due to the difficulty of providing exhaust gas at sufficiently high temperatures. A comblnatlon of the particle bed and electrically heated filament types of dissociator might satisfy these engine requirements if the currently high power requirement of the filament type reactor can be reduced. It must be emphasized that these observations are






~ ..


36 ml/min of NH3 l'>

024.66 9 NISKEL SHOT . !?

25 x . 9 NI'l-KEL SHOT_-+_~ __ lJ

66.6 9 NICKEL SHOT /1 r:

Z ' /

g 20 17'-

~ '5r_-~r_-~-~~~,~~ [g /1/1

<f D~'--~---~//7717

~ ;'

6-------- ..- ;'


05~---4--~_/_'----~~------ ~

Fig wit hea

bas and ope


soc for cat. era and for


900 'KXX)


spal perf -t ivr be, the suff:

Fig. 21A - Typical ammonia dissociation data obtained with experimental rn icroreactors - - Mode I A - particle bed

20 Mt/fv1IN OF NH3
- r
°/1 cess sibk met but in tl of h

amo able


400 800



Fig. 21B - Typical ammonia dissociation data obtained with e xpertmental rrncroreactors -- Model Br multi tube



80~----+-----~----~-----+--i I'


g 601----t---l-----+---k

~ I




Vl 40~--4_---+----r/



OL_~~~ __ ~ __ ~_L _L ~

5 6 7 8 9 10


Fig. 21C - Typical ammonia dissociation data obtained with experimental microreactors - - Model C -electrica lly heated filament


based solely on laboratory e~aluations of miniature m~dels and hence. their direct application to full-scale designs and operating conditions must be considered specula rive.

Complementary ammonia dissociation studies are being conducted at the Allison Div. on a much larger scale dissociator of t~ particle-bed type. This dissociator is sized for a single-cylinder engine' and uses a promoted iron catalyst. Preliminary Allison test results corroborate. ingene ra l, the findings of this paper and indicate that a simple and reasonably sized dissoc tator of this type "Can be evolved for use on a rnult icy l inder engine.


The-results of this study indic ate that an ammonia-fueled spark-ignited reciprocating engine can be developed with performance equivalent to that obtained in current automo-t ive gasoline engines. Desired rnaxtrnurn engine power may be developed by two different approaches. In one method.

the addition of a 'supercharger to a conventional engine would suffice. The engine would have to be supercharged in excess of current automotive practices. which should be possible due to the high octane raring of ammonia. The other method would involve also the addition of a supercharger

but only moderate supercharging of the engine. an increase in the compression ratio. and the addition of a small amount of hydrogen to the ammonia. The addition of a small amount of hydrogen to the ammonia is a requisite for suitable pan-load engine performance.


. Hydrogen. when added to ammonia in small quantities. was found to act as a combustion promoter.In accelerating the burning of ammonia. Dissociation of a part of the ammonia fuel in the vehicle appears to be the most logical method for supplying the required hydrogen. A' preliminary study indicates that a catalytic dlssoctator heated by the engine exhaust gas offers promise offulfilling tile needs of

an ammonia-fueled engine. . .

The engine modifications involved' when replacing gasoline with ammonia appear to be straightforward. Probably the greatest problem will concern the auxlltary equi prnent and controls. TIle development of an ammonia evaporator and fuel metering systems would possibly follow along the lines of similar LPG system de"velopments. The dtssoc iator would be a novel development but appears to be f'easible according to preliminary Allison test results.

To drive equal distances in a vehicle. at least 2.8 times by volume and 2.35 times by we ight as much ammonia as gasoline will be required. The fuel system in an ammoniafueled vehicle will be bulkier than that in a conventionally fueled vehicle. However. it is believed that this will not be a serious handicap for the majority of military applica-



The authors wish to express their appreciation to W. A.

Turunen for his direction and advice during this invest tgation and to R. L. Gatrel l, of the Chemistry Dept.. for developing special chromatographic equipment and techniques and con.ducting the ammonia dissociation experiments.


l. "Energy Depot Program - Technical Feasibility Study."

AEC Res. ana De~,'Rpt. NYO-98'76. Vols. 1 ana 2. Allison £OR 2737. May. IS. 1962.

2. Patents of Ammonia Casale (French). 799.610 and 802.905.

3. E. Koch ... Ammonia -- A Fuel for Motor Buses." Journal of the Institute of Petroleum. Vol. 31. (July 1945). 213- 223.

4. R. Wiebe and T: H Trernearne .• The Solubility of Hydrogen in Liquid Ammonia at 25. 50. 7-5 and 100 deg and at Pressures to 1000 Atmospheres.' Journal American Chemical Society. Vol. 56. (November 1934).2357-2360 ..

5. Katherine S. Love and P. H. Emmett. "The Catalytic'.

Decomposition of Ammonia over Iron Synthetic Ammonia Catalysts." Journal American Chemical Society. Vol. 63. (December 1941). 3297-3308.'

Discussion of this paper appears on page 316 .





Discussion of paper> 650050 (p. 274), 650051 (p . .1). and 650052 (p 300).

R. 1. FLANNERY American Oil Co.

THIS AUTHOR GENERALLY concurs with the remarks within the scope of the presentation, Several comments should be made.

The first relates to the choice between convened 'in,ter!JjJcombustion and fuel-cell power units for vehicles. This choice might- be influenced by the attractive possibility of a simple conversion kit allowing use of standard hydrocarbon -fueled engines. However, this convertability may not be unique to IC engines in view of other work directed toward fuel cell vehicles powered by steam reformed petroleum fuels. Such vehicles would be readily converted by changing to ammonia cracking.

The second relates to scope. \Vhile the application detailed in the paper, vehicles, is one of the most difficult. an army requires energy for various other purposes, ranging from cooking and lighting to communications power. Safe efficient burners for depot fuels will be needed. A variety of electric poi ... er suppltes of capacity or type not possible to be provided by batteries iq the isolated arena w ill also be neede.d. In many of these cases, the projected fuel cell, with suitable power processing auxiliaries, will probably be more efficient than corresponding converted engine generator sets.

The third r~lates to the direct ammonia fuel cell. In the paper, this approach was discarded for consideration a! this time because of lagging technical advance. Yet this system holds a potential advantage which warrants at least continued attention to its possible development. This advantage is high efficiency on idle. The rate of the ammonia cracker cannot be changed rapidly, whereas the fuel consumption of'r€e direct cell readily drops to.a low level on -Id le , Important fuel savings can result , In addition, an ambtcnt-temcsrature , difect ammonia cell would not re-

. ,

ica in

quire 2f1'!o of the ammonia '5 hydrogen to heat the ammonia cracker.

The fourth relates to tile projection of the fuel cell performances , The projections presented do not seem too optimistic. BJIt; for the benefit of thosemorecasUiiUy acquainted with fuel cells, it should be noted that the comparison with present state of art should be.made for power densities at t!he required operating efflclency, not at the maximum efficiency. Fig. A summarizes the power density data or' Fig. 9 and 11 of the paper and shows the operating .points selected in the design study. The VOltage efficiency fines which have. been superimposed on the curves reveaf that the design points are all in the 60-7ff'/o range. The corresponding point on the 1963 state-of-an curve given

is well below 100 wft 2.

Finally. in a truly isolated arena, some unexpected materials can become "fuels" in the sense that a supply of them can be used up. For example in the processes detailed, air purification and ion exchange water purification are required. Taking the KOH requirements for air purification given in the paper~ some 30. tons/year of KOH would be required for the utilization In fuel cells of the yearly reactor output L'Chernic al needs for regeneration of the ion exchange bed would depend on available water and on undisclosed process details, but could be substantial and would apply whether engines' or file') cells were used. Use of thermally regenerable treatment chernic a ls would help alleviate a problem in this area.

pOI vie Fro lin cul pre clu cal

hal d1f on in

am clu mi rna pOI

tha lin bUI be; the SOT otl ric me of

E. B. RIFKIN Ethyl Corp.

f1a in sib sto lat afT sto sur of thE pOI api 98

THIS PA PER IS an important contribution tothe literature _ on combustion in reciprocating engines as well as a stgnif-





me tha ch, bill fla of


Fig. A - Power density d~ta and operating points


ic ant coUee-tio o. data on the possible utility of ammonia in the energy de, t concept.

The authors nave clearly pointed out many of the important factors that must be considered if ammonia is to be viewed as a serious contender for use .. s a military fuel. From their data. it is obvious that, in contrast with gasoline, ammonia suffers from low heat of combustion. difficult ignitability. and low flame speed. However. these problems can be overcome to some extent by use of supercharging, high compression ratios, ignition system modifications, and partial dissociation of the fue l before induction.

As the authors' data show, ammonia-air mixtures are hard to ignite. Although minimum ignition energies are difficult to reproduce and to interpret precisely: some light on the problem is shown by the work of Buckley-and :5" in (1) who showed that the mtnimum ignition energy fo an ammonia -air system was 680 mi llijoulcs in a condenser ischarge. A corn parable figure for ~ -heptane in air was 0.3 mtllijoules , This probably explains why the ignition system modifications descri~d in tile paper rnarkeetly increased power output .

Another serious limitation on ammonia-air combustion invO'lves the flammable limits, or range of concentrations that will sustain, a flame . In the case of hydrocarbons, these limits are very broad. Thus, n-heptane, for example, will bum under ambient conditions in afi at any concentration between 1.1-6.7 vols "70 (2). This gtves wide latitude to

the engine designer in his quest for maximum power under some conditions (rich mixtures) or maximum economy under other conditions (lean mixtures). Further, thesrotchiornetric concentration occurs at 2.3"70 a-heptane . .!fllis allows more than a factor of !?'tn 'fuel concentration on either side of !,be stoichiom~trij:co:rrsition,

The ammonia -air system is quite different. Here, the

fl arnmab iljty limits are 15 and 28 Vol "70 ammonia (I). which in itself meant..lhattherl isIess than" twofold v artation possible in arrrmonia concentration for a burnable mixture. The sroichlometrtcmtxture occurs at' 21. fJI/D, with only' a narrow Ia titude allowed on either side of this point. Thus, under ambient conditions, all flammable mixtures mu"t be near stoichiometric. Undoubtedly, higher temperatures and pressures wi ll somewhat broaden this range, so that a wider range of mixtures can be burned in an engine .than indicated by these figures. An interpretation of some of the exhaust composition data in the authors' Table 2 can be based on this approach. Thus, fuel mixture 2, containing ammonia plus 98. fPIo of the theoretical air, stil1 burns .only 68. fPID of the ammonia inducted. Recognizing that some mixture in homogeneity probably exists in this system, we can theorize that the-part of the charge in the center of the combustion chamber is well insulated from the walls, and thus probably burns completely because its temperature is high'arn! its

flammability limits fairly broad. However , the portion

of the charge near the walls is cooler and has reduced ff am «

mability limits. so that quite lfke ly a substantial portion .

of it is outside the com bustible range and does not burn.

In addition to the, flammability problem encountered in

317 •

dry air, moist air would likely impose an additional penalty, since high humidity tends.to substantially }larrOW the flammability limits in the ammonia-air system (3).

To overcome this ftammabniry limit problem, the authors have used several techniques with substantial success. Supercharging and high compression ratios wo.uld both be expected to broaden flammability limits, .The 'cracking of the ammonia prior to induction is another route, Which is effective partially because of the very broadflammabiltty limits of hydrogen (4-74 Vol "/0). Another possible route is available, which involves the use of limit-broadening additives'in the fuel. ,Such an approach is not needed in the case of hydrocarbons, with their wide flammability limits, but could be verY useful in ammonia combustion. Some indication that 'flammability limits can be broadened in this way is found in the work of Egerton and POw ling (4), who showed that additive amounts of ethyl nitrate had a stgnlficant effect in raising the upper limit of flammability of several Itght hydrocarbons.

There is also a second reason for anticipating that additives may improve the combustion properties of ammoniaair mixtures. This is based on some evidence indicating that the ammonia molecule must drssoctate (at least par-

.tiatly) into nitrogen and hydrogen before it will burn. As the authors point out, a catalyst is required to make this reaction proceed.rne asurably at temperatures below 900

F. This opens the possibility for an additive to be introduced for the purpose of promoting such dissociation during compression. An ammonia-soluble metal compound. decomposing in the engine to. produce a fine dispersion of solid metal or metal oxide, might 'be one possible approach. If effective, it could facilitate ignitability, broaden flammability limits; and increase flame speed. If this approach were successful technically, il might greatly add to the practicabtllry of the military -use of ammonia, since it would substantially reduce the need for major engine modifications.

In ~their work \vitlffiydrttgen additton, . tl:}~.authors have shown how this approach is an 'attracti¥~6& for solution of some of the ammonia combustion problernsi iThelr results may be somewhat optimistic in relation to practical ap- ;;. pro aches, since the cracking of ammonia would introduce nitrogen as well as hydrogen, and the increased dilution would operate to reduce some of the gains shown when h,..drogen is added. From an energy standpoint, the cracking

of ammonia prior to induction into the engine has the'effett of increasing the , available energy by nearly 13"10 9n a weight basis. However, because the dissociation products-occupy twice the volume of the ammonia, the energy would decrease on a v()lumetric basis by about fJ'jo. ThUs, 'an addi-: tional problem may be encountered in practice where; it". would be desired to maximize the mass of -fue l and air,in" ducted into the cylinder. '"' ,', 01

Some time ago, Messrs , Kerley, 'l"e1t, and Adams i,'n Ethyl Corporation's Detroit Research taborarories brij<,flyex':' amined the engine performance of ammonia in a siir~eyb inder Waukesha engine with variable compression ratio' "

"' », '

'.t ..




~. l,


, ,

.J'_, 41;,;


and rei tem~1 pressio atmosp CR 4-J speed , Indicai compa mITE! plugs, crease, Ing the fidem from tl

Thr would vi lian

Eve to be ( concej usefulr needs engine to the flculry Iry lin econoi ergy d engtne effect terms ent m the fa a1fty, .




:'.' plpsro

-. ;~ , 3,_ 'don, >. , " 4 .. 190h

. ,

. '

"~I;','O· '4meri }:~



A condu

'th{lI~ the~


Co~p moni, much


and removable dome head. They matntatned intake air

temperature at 130 F, jacket temperature at 230 F. compression ratio at 8.5, and manifold pressure essentially' , atmospheric. SPark energy, supplied by the engine's Bendix CR 4-1 magneto, was about 80 mtl lijoules . E'ven at the low speed of 600 rpm, they encountered poor combustion, "as Indicated by a brake thermal efficiency of 13.75%. This compares with a ... a lue of 22,20/0 for Isooctane ponratnmg 3 ml TEL/gal under the same conditions. when two spark plugs were fired simultaneously, the'~ alue for ammonia increased to 18.so70. By the additional technique of increasing the compression ratio to 12.65, the brake thermal efficiency WqS raised to 21')10. In general, their conclusions from thfs work support those of the authors.

, Through their study, the authors have shown th'at there would be no incentive to use ammonia as a fuel in the civilian market as long as hydrocarbons are available.

Even in the military context,' much further work needs

to be done to clearly define the potenti a liry of the overafl ' concept, The present work is an excellent beginning on the usefulness of ammonia in gasoline engines, Additional work n~ds to be done ondteset-cycle'engtnes and on multifue l engines. These engines will encounter other problems, due to the probable low cetane number of ammonia and the difftculry of keeping a stratified charge within the flammability limits. Detailed 'consideration needs to be given to the economics of the entire concept, including ~ only the energy depot aspects, but also the complex task of modifying engines without rendering them so cumbersome' as to be ineffective. Finally, the overall evaluation.of the system in terms of cornpetitfve cost-effectiveness, in relation to present methods of wartime fuel supply. will probably be

the factor that determines whether it ever becomes a reality.



. .- '-,

,1..,BUl:kley and' HuSIIJ 'Chern, Eng. Prog . 58, 81 (1962).

, 2',' Lewis end von.Elbe, '-Combustion, Flames and Ex:-.' plpsrons, • Nt1~:Ymk" ~951. '

";' , 3., ,Perry, ·Che~nk~l Engineers" Handbook, • Tflird' Edt-

'tion, .P~.'l587'. .

. ... 4.~,Egerton and !>owlirlg,. Proc .. Roy, Soc, A 193, 172,.

~~ .. , .

, 190 (1949) , " " . '

i· '. I ~



. '

_", .

',~ T;,{). \1rAGN'ER , .~ rnerican, once r

;, '. . -c , ~ tt


•• '. \" ,< -', " ••

',<tllI~~~cted a ~ry~sen~ible IU'1'q logical program to explore

thl'!uItability, 'of ammonia as an engine fuel. We commend

therit for .tl:re~ !o'(6~ " .

.. DUring ~68,'€. f. Domke and I at the Americars-Ofl COIrypany t<iborator!es'i~vestigated the peiforman~e of aIJ1- J11Oni~.1n engin~, .Theapproach and sco~ of our work was much like ihat of the GeA'eral Motors work except that we

. - . ~




stor a
All (
to st
the I
to sl:
ily a
co a
all I
of a
crea did not investigate the effects of disso,!;ia ion of ammonia, and we extended our work to compression -tgi tion engines. we-agree completely with the results the authors present.

We would like to offer ,a few observations on the performance of compression-ignition engines operatetH)il ammonia.

The Armed Forcesin"entory includes~l1igh fraction of vehicles with compresston-Ignition engfues,and tlHiy. .aswetl as spark-tgnitlon, must be accommodat~a in tile energy depot concept. We were able to operate a CFR cetane method engine on pure ammonia and ammonia with several addi.ti~es, Ammonia (or ammonia plus additives) was injected in conventional fashion, although it was necessary to ad-

, vance injection timing greatly and to use a plunger' and bush- . ing assembly in the injection pump larger in displacement than the one normally used in the cetane method engine.

The engine was starred on keros~!le, and then switched to ammonia. The en~ine would not restart on pure ammonia

at 35:1 compression ratio with normal coolant and inlet air temperatures. However, we were able to attain ignition

and regular combustion by raising coolant temperature

to 370 F and air temperature to 270 F. Power output

at these conditions was slightly less 'than that obtained at normal conditions with kerosene. We did not analyze the exhaust, but it had a very strong odor of ammonia. Our investigation of additives was limited to only a few choices. Several were effective; they permitted ignition and regular combustion to be attained with somewhat lower compression ratio and coolant and air, temperatures.

From these results and ideas they generated, we believe that with some engine modifications and with proper additive treatment, it may be possible to make compressionignition engines operate satisfactorily on ammonia-based fuel. The alternate, of course, is to convert all compressiou-Ignltion engines to spark ignition. More work is nee - ' essary to determine which would be the best solution.

P. S. MYERS AND O. A, UYEHARA University. Of W!sconsin ,


ruE ENERGY DEPOT concept has been of considerable in - . terest to tlii' discussors , In 1962 the discussors and other coworkers presented a. paper entitled "Portable Power From . Nonp6rtq.ble,En~y Sourc,!?s",'" .,Thu pa~r was prese-riled": at the 'National SAE Powerplant meeting in Philadelphi~

.' and published- in the i96~>SAE Transactions .. the basic-purpose of, this paper wasro see :What. solutions might .be found ' 'to the .problcrn of providing .l\l energy supply suitable fclf'" portable power plants when our present petroleum 'supplies were exha,}1sted. Whne this'{sa slight1y different problem than the one under discussion- today, there are manystmi-

larittes and the conc lusions are remarkably strnilar . ' '

Inasmuch, as two different groups .constdermg two d ifferent, but rela1'ed,' problems reached similar cone lusions,

it would appear that these conclusions were fundamenial a:nd should be emphasized. Tbe 'first of these common cC:n<;lusions is that, barring unexpecteh br~ktrlf(iughs, chemical


. '

geni the,

.. ; 'vie~ , vall

Inte and only

t» mor Is it card mor: beer the



storage of energy is the only practical engine, m que .

A 11 other possible techniques are all too bulky. either because of high shielding requirements or becauseof the large bulk of the energy storage material and system,

The second common conclusion is that only matertals available on a large scale. that is, water and air. can be considered for production of fuel because of the large quantities of material involved, The implication here: is that hydrogen is going to be the basic means for storing the energy originally obtained, from nuclear or some other stationary power soiirce .

The third common conclusion is that it will be necessary to store the fuel in the Ii quid form or its equivalent, Hydrogen can. of course, be combined with other compounds such as nitrogen to forrn.easi ly Ii quifi able fue ls, I'f this procedure is fol lowed, one of the essential requirements is that the resulting 'fuel be stable both""chemically and with respect to shock,


The fourth conclusion is that it was undesirable to manufacture a fuel contatnii1g oxygen inasmuch as oxygen is readily available in the air free of charge,

It is interesting to note that the papers by Mr. Rosenthal and Mr. Grimes. as well as our earlier paper. reached these common conc luslons . It should be clear t~at these conclusions are a pplicable when considering portable power plants, Different conclusions might be reached when considering stationary power plants.

It is pointed out .in the papers that the number of potential fuels containirig hydrogen and other readily available compounds. are very limited. As the paper by Cornelius and

co authors points out, the fuel finally chosen, ammonia.

has certain combustion problems. These combustion problems seem to arise primarily because of the low energy content per cubic foot of mixture. In this respect the behavior of -the ammonia -air mixture reminds one of the lean mixtures of convent ional fuels which also have comparatively low heating values per cubic foot of mixture, Thus. it is very interesting to note. with the exception of turbulence.' all of the steps taken to improve combustion performance

of ammonia as an engine fuel increased the energy content p~r cubic foot of mixture at the timi.af ;,park. that is, increased compression ratio. turbocharging: and so forth.

The thought of improving. combustion by enriching the ammonia -air mixture with hydrogen is intriguing and ingepidus. vie do have some questions. however. regarding the wayIn which this was accomplished. particularly. in

: '-vieJ,'f"of the comments just made regarding the low heating

, ,viilue per cubic foot of the ammonia-air mixture. If we interpret properly the experiments conduc ted by Cornelius and his associates, the tests were run by adding hydrogen only to the ammonia -air mixture. The dissociation of am-

;.' monia, however. will produce both hydrogen and nitrogen.

Is it proposed to separate the hydrogen and nitrogen.discarding the rutrogen and adding the hydrogen to the ammonia -a ir mixture " If not; should not the experiments have been run with the addition of both hydrogen and nitrogen to the ammonia-air m i xt.ure ? Would the heating value of the



I n
a mixture have been significantly less if hygrogei1 and nftrogen rather than JUSt hydrogen had been added? Would it have not been just as easy. experimentally. to have added hydrogen and nitrogen in the proper proportions?

The discussors would a lso like to raise the question of whether or not the thermal inertia of the dissociator c which decomposes the ammonia to produce the hydrogen. would

be a complicating factor in the design of the engtne-dlssociator system? For examplevIf the engine were operating ~ at low load with consequent low drssoclator temperatures

and the throttle were suddenly opened.lequir~I1g increased quantities of ammonia to be d'issociated: )'Iould the disso-

c iator be able to supply these increased quantities?

It Isalso interesting to note in Fig. '14 of the Cornelius paper that ~,,/o hydrogen addition does npt produce a?}t significant improvement over '21'/0 except at the very highest speed-of 4000 rpm .• Can the authors give any explanation of this "leveling off" with Increased hydrogen addition?

'. ~ /'

MORTON S. SILBERSTEIN United Nuclear corp.

THE PAPERS PRESENTED in this session 00 the Ertergy Depot concept deal primarily with the p~oduction fuel in the field and the utilization of this fuel t? power army vehicles. In all of the concepts discussed. an inherent part of the energy depot is a mobile nuclear power plant which generate, the electric power required as input for the fuel production process. I would like to offer some brief remarks about the nuc lear-power generating portion of the energy depot.

This power source is being developed under the~Military Compact Reactor (MCR) program. for whi~h AlIis~n Division. General Motors Corporation is the prime c~ntractor, United Nuclear Corporation. whom I represent. has the responsibility as subcont~actor to Allison for design and development of the nuclear reactor portion of the MCR.

One of the accomplishments of this effort has been the design of an .MCR unit for an electric output in the same 3000 kwrange for which the Allison and Allis-Chalmers en ergy depot conceptual designs were evaluated.,

To meet the'size and weight requirements of rnobi l ity , the MCR is packaged in modules comparable to those described for the various energy depot fuel processing units:

These packages are trailer or truck mounted for operation' in the field. and are transportable overland on their trailers or 'alternatively by air or sea.

The nuclear reactor, surrounded by its biological shield. which must meet extreme low weight requirements compared with ordinary reactor shields. is 'the basic source of thermal energy. Its heat is transferred to a liquid metal coolant. which \jltimately he.ats air in a .heat exchanger. The heated air drt.es an open Brayton cycle gas turbine engine. which drives an alternator whose e.lectric output is the 'power source for the fuel manufacturing units. Auxiliary electrical equipment and controls for the nuclear plant are housed in separate units connected to the reactor and engine

by electric cabling." ,


The energy depot constitutes an extremely attr active-and logical application for nuclear power. T~,is is an example of where the demand for long time operation without additional fuel supplies is uniquely supplied by a nuclear energy source.

Mr . Rosenthal's paper pointed out the weight attractiveness of the energy depot in comparison with the continued supply of gasoline fuel. and the cost studies that have now been initiated. There are, of course. uses of the nuclear powered energy depot where; independence from external supplies is.an asset wh iclr cannot be measured in dollars. ' for. instance. the holding of a key spot which would otherwise

be lost. Nevertheless, ~t is of some interest to note, that although the CQSt of uranium fuel an7E:i ssoc iated plant are generally considered to be high un er military circumstances. the cost of delivered gasolin can also be surprisingly high.

A simpliflcd analysis of several special wartime situations shows that the predominant cost of supplying a major defense position is 6'ften that of the replacement value of aircraft and ships lost in bringing in the supplies: On.this basis, considering such instances as Tobruk and Malta in World W <U' Il (for which the necessary statistics ate described by w{nstcin Churchill in his books). delivered gasoline is found to cost $50-j80/ gal. In different circumstances. then, gasoline may range in value from the $,30 or so. per gal. available at the local filling station up to numbers hundreds of times as great under front line fighting conditions:

E. 1. GAY '\ Consulting Eng~er

•• THE MOBILE ENERGY depot modules present several logistical

problems: which I think might be illusrra.ted by Figs. B - N .

B. This is China on the old Burma -China Road between

Kunming and Chungking, This section of the road-Is called . the "Ladder," It has no guard rails and steep grades, with

a 1/4 in of slime. When raining the Goer just wouldn't go! Only a 6 x 6 or 8 x 8 would do the job,

C. This is another view of the old Burma -China Road,


D. The photograph shows typical terrain between China ..

and northeast India, We flew petroleum products CJ<Ie~ this

area unti 1 the pipeJtnes were bui lt .

E. This is up front on the Ledo Road near Myitkyina; you needed snowshoes here,

F The Ledo Road where a temporary wooden bridge with a' load limit of 4 to 5.tons and a.4 in. pipeline on 'each side

was but 1 t' , _.J

G. Two of the 4 in. pipelines -' ahead of the finished road. These 4 in. coupled pipelines handled 1500 gal/hr/

, line. We shipped aviation gas - motor gasol.ine and diesel fuel via these lines. There was a pumping station evety 8 miles. One of these lines could refuel seven of the old M-4 , tanks/hr.

H. A portion of the Ledo Road that was near' completion

and included' a Bail~y bridge and mulesl , v

, ' . w. EO.RN~.;TAi

I .. The picture snows a finished portion' of the Ledo Road: I • Grades in this area were as high as 14 to 16"/0. There-was

one short stretch at 250/0,'

. :~.

J. This is a Korean rivet with a bridge out.

K. Korea - just backofthe combat line, This dirt road )'Iitil up to 120/0 grades was the only line of cornmunicatton '. l'\ pipeline~ame wi1h1n 15 miles of thij. road.

L. Typical Korean terraln.. "

'}vl. This photograph wastaken InKorea and shows' fuel storage for the tent stoves. ~erose~~w~s used. What,do

we do when we use arnmoniaj, <.. ' .

N. Here are 6 in, pipelines and a pumping station in Ko~ea. these'6 in. lines, with a pll,rt!pfng station every. 16,5 miles would handle 3300 gal/hi, 'They would refuel 10 6ew M-pO tanks per hour: The newer we lded.B in. and ; 8 in. plpe ltnes operating at higher pressures will handle several timci as much fuel ..

, Th~~ is~ert:.r~ason to be1iev~ that South Vietnam (and . t~e,. teriitori:")1'~rtl1 and. ~e~ _0 SOuth Vietnam) has equant

difficult terrain. cle~t1l.1l.1 t 'cent visit to Thailand,

" .. ' of .',

would- confirm, this .' .'-~

The problems I visuali lowing, comments:

Apparently. these Mobile Energy Depot modules weigh about 30,000 lb and a complete unit about 100.000 lb. It takes a plane nearly the size '~f the 707 to carry 30. 000 pounds, There are not many military air strips in the area

Fig. B, - Burma -China road


-.~:-'!;. • ...,; ~~


':» ~
~ 2
. il
tr .




, .'/'

~ 1~' .

. , " ..

, "



• <


.. '


" ,

where we might engage if combat~t will take the 707.

, - A 16-tonGoer \"e1gh11g 39. OOQ"lb wttnour a paY~(id wiiL, be needed \0 carry. the 3D: 000 lb Jilodules: The 8-tdn doe~ suggested for-transporting, the hydrogen orarnrnorua weJghs 28, ZOO 10 without a payload .. These tw~ vehicles win also: need aircraft of large payload capacity, ~, .

In wartime, en,ergy is needed for many things besides inoving vehicles. Cooking and space hearing ,are two major items. Fifty per cent of the Ju~l supp1i for. Korea, durihg-

tile wintersmcuths.: .was for.space hcaung . " •

At no time during World War oIl or the Koreal'!\Var was there a~ acute shortage of fuei. There ;"ere some c lose calls however' There was fuel in Sicily, before there was food. General 'Paton ran away from his fuel supply. \Vhen tbld to col lee this 5 -gal gasoline cans,he replled.hc was not a garbage collector. He was a grea'r general, but-not good at' logistics,

W~ blew up the rail tank cats and fuel dumps.ar Gafsa·

in North Africa when we thought we were going to_lose the town" We had to refuel and regather our equipment after

. ~


• 0'

,. '


• 'tV


I ""l ,

'~ 4:".;~'"




':'~':r"""'C,~" .

". .. -.-_ . _" .

.. ..... '-}l-

Fig, E --Ledo road,




" .,

Fig, C - Burma -China road


,. ,


\ '



" '

, '

. 'r

" ,

."' .. -e


F(1), D :'Typical te~rain 'irlc.lina and Ilonhea;,

Indt"a ~ ." »: "

": (.'." /'


. '




.• Q








...... ~~

. .".~ ...

_' j


.,;:.~Fii; F -Ledo r6ad showing 4


-I '

in. p~pelines

Fig. G - 4 in. pipelines



Fig. H - Ledo road under construction

-e ,







Fig. I - Finished portion of the Ledo road



Fig. J - Korean river

. .

F.ig. L - Typical Korean terrain


, "

Fig, K - Korea' - just back. of the co~b.at~ine "


the kicking around we got at Kasserine Pass., We did the refueling atnight from conventional 5, 000 gal tractor-trailers (with no lights) directly into vehicles and into 5 -gal cans.

Fig ,M - Korea - fuel. storage for the tent stoves


The Army's performance specifications for vehicles require that they operate satisfactorily from + 115 F to -25 F without. the 'use of starting aids other than . .mantfold heaters or glow plugs. From -25 F, to -65 F, starting and heating a'ids may be used. Fuel, temperatures of' ... 145 F have been measured in the Msert.,There is no reason to believejha; the storage' vessels for hydrogen

, and/or ammonia would not be exposedto the same tern,perat'Ures.

Our M-60 tank is capable of a Z4htbatilefield day without refueling. Newer concepts talk Of a 36-hr battlefield day. To achieve this with ammonia, would, require 2.8 times thevolume of fuel, which would make all impossible sized tank; or .would reduce the battlefield day by the same amount, for example, from 24 hr to 8.6 hr .

For many years, ou~ top military command has fai led to recognize the need for maximum effort in studying the protection of our supply lines by sea in time of war. Preoccupation with the missile may have been the cause, I am glad to note from recent published data that antisubmarine warfare planning is now beiIlg emphasized'

Fig, N - Korea - 6 in pipelines and a pumping station

\ AM
to r
.J a p
a h
a n
a g
ab l
of I
leu /.



LT. COL. KERMIT O. LINDELL United States Army

, J


AS A RESULT of the discussi~S presented, I would like to amplify and perhaps clarify a few of the points presented during the session. First of al-l, the cost-effectiveness of the concept has not been evaluated. The Stanford Research Institute -has just recently been awarded a government contract to perform an operational analysis of the Nuclear Powered Energy Depot concept to include an evaluation of the costeffectivell.ess. As the study is just getting underway, even

a preliminary estimate is. not available at this time as to whether or not the concept is economically attractive,

The cost -effcctiveness of a concept such as this is extremely difficult to evaluate. With respect to Mr , Gay's remarks conceming the Ledo Road, some 300 planes and many lives were lost during World War II delivering supplies over "The Hump" into China. If a concept may save lives and equipment, how do you factor this element into a cost-effectiveness analysis? Is a life worth ten thousand dollars,

a hundred thousand dollars, or a million dol lars? As a matter of fap, I might suggest that if a Nuclear ,Powered Energy Depot had been available du~ing World War II, it may not have been nece,ssary to construct the Ledo Road to provide

a means of supplying troops in China; a substitute for gasoline, the major item of supply -involved, probably could have, been manufactured in China, This type of operation. I believe, i-s an exce lleru exampl~fan actt;al situation where an Energy Depot could hive been profitably employed.

'Also, it is extremely difficult to forecast the nature of future warfare, nuclear warfare. World War II and Korea can no longer be considered as valid historical examples. The exact nature of the conditions on a nuclear battlefield are still a matter of speculation. To quote, in substance,


about the basic problem, let ~e illustrate it In a different manner. You have, on one hand, the tremendous energy density availablein a nuclear reactor; and, on the other hand, a large, growing requirement for ~nergy to operate vehicles, aircraft and other equipment. How do you convert the energyof the reactor into a form which will be usable a's fuel by vehicles and aircraft" If, at the same time, you can eliminate the requirement for an exteI1SiVe~istdbutl?? system and relatively fixed and vulnerable fuel supply depots, a major military advantage may accrue.'l'!iis is the basic problem and the Nuclear Powered, Energy Depot concept offers one possible means of solving this problem . Direct transmission of electrical energy, without wires, offers,' perhaps, another means of solving the. basic problern .


THE DISCUSSION OF Dr. R. Flannery raise~ severa"l significant points. Some of these are covered in greater detail in Mr, Rosenthal's paper (p . 274i.'

The major use of fue l in the Army is in vehicles; this more difficult case was treated first . Energy depot fuels can, be used in modified space heaters and to power the equivalent of motor-generator sets.

DiIect consumption of ammonia in a fuel cell would be a more ideal solution if adequate performance could be obtained. Rand D should be dircofed tOi"ard this end. More recent studies have indicated that 25<t fuel consumption in the dissoc iator may be a conservative estimate for fuel utilization,

The state of the art of fuel cells has advanced rapidly, Fuel cell people today talk of the possibility of building fuel cell modules within a factor of 2 of the projected fuel ce'lI

a general officer I heard speak several years ago, "You prob- "modules.

ably won't know what type of tactics you will employ until The air purification chemical requirements were projected

you become actively engaged in a nuclear war," It is re-cognize d, however, that the Army 'must seek advanced concepts of supply to supplement or replace the present concepts of logistical operations,

Secondly. I would like to point out that the Energy Depot concept is not, at this stage in the development. fixed in concrete. If there is a means-of producing, in quantity, a more suitable fuel under the same constraints. we would be extremely interested; for an example. a process for pr od uc >

lI1g .b.l1e thanol. . " ,

Thirdly, let me reiterate that during the evolutionary phase of development, it is intended that ammonia be used as a supplement to normal petroleum supplies. Not all units would necessari ly be equipped with vehicles capable of burning ammonia. The capability to use ammonia as well as petroleum products in vehicles and Army' aircraft would be limited to designated units and/or special situations such as airhead operations or deep penetrations. in these particular instances. it might be extremely difficult to supply petro- 1eum fuels uti liz ing present distribution procedures,

In conclusion, to insure that-there is no misunderstanding


on the maximum theoretical requirements to remove all of the acidic gases from three times the air requirements of the cell. Contil11-!,fd experimental studies have indicated that marked reductions can be made and air purification chemical. would be a small requirement in the future .



THE AUTHORS WISH to express their appreciation for the compliment ary and constructive remarks regarding. the paper by the several d iscussors , The prepared discussions ha ve aided in clarifying certain important aspects of the energy depot concept.

With regard to Dr. Rifkin'scomments. we agree that it would be worthwhile to investigate the use of fuel additives other than hydrogen to promote the combustion of ammonia. In our investigation, we used only hydrogen to conform w: th:

. the initial ground rules that a ll fuel consti tuents be produced' from readily available materials. It might prove to Be more feasible to use an ammonia -soluble metal compound as sug-



a d a d fl

81 (I


t c d n f



gesred by Dr. Rifkin rather t an hydrogen if the addition of paper that satisfactory full throttle engine power may be de-

<?nly a small amount of the co pound to the fuel is required veloped solely by supercharging without any recourse ,to h~ - ', and, the provision of this co pound does not invalidate the drogen addition. In addition to this questionable thermal

Mobile Energy Depot conce to Interia of a dtssociator, .the possibility exists that the. engine,

The 60/0 decrease in energy content of the fuel charge on exhaust temperature may not be high enough at low engine a volumetric basis as mentioned by Dr, Rifkin caused by the part load conditions to insure adequate hydrogen emission

dissociation of the ammonia is incurred only if all of the from a thermally heated dissociator , Ii was for this reason

ammonia is dissociated " However, in order to provide the that an electrically heated type of dlssociatorwas 'Investt-

desired 2,5,,/0 by weight cencentration of hydrogen in the ,gated by the authors.

fuel mixture, only about 12, 5°l' of the ammonia must be dis - The question raised by Drs. Myers and pyehara regarding

soc i ated , and hence the decrease in energy content of the the general "leveling off" of-engine power with il1creasing

fuel is only about i. 7"/0. This ·dect.",ase;is?mall when com- hydrogen addition above ?!'fa can be explained bi comparing

pared to the large incre.ase in combustion efficiency that is' the r~latjve energy contents of stpichiotnetncarnmonia air ;!

realized when hydrogen is added to the ammonia, , and hydrogen air mixtures on an equal volume basis , The

• We agree with Dr, Rifkin that there is presently no incen- heating value of the hydrogen air mixture is only about 4,,/0 tive to use ammonia as a fuel for commercial vehicles. The greater than that of the ammonia air mixture, Therefore, current cost of producing ammonia would have to be reduced only a relatively small increase in power can be expected, . drastically t? make -it competitive in cost with gasoline. As \ as the concentration of hydrogen in an ammonia -air mix-

mentioned in the paper, - an engine would have to be modi - ture is increased above the amount required as a combustion

Tied and ~ertain auxiliary systems and controls would have 'promoter fer ammonia, We found also that the spark advance

to be incorporated, Any Significant emission of ammonia required for maximum engine power differed appreciably

from the engine tailpipe could not be tolerated because of for ammonia air mixtllres and the investigated ammonia hy-

its i;ritating odor. Regarding the military use of this fuel. drogen air mixtures. We conjectured that some loss of p~er

we agree with Dr. Rifkin that a much mere thorough evalua - occurred in trying to compromise these incompatible spark

lion of the entire concept will be required before its feasi - advance requirements,

bility can be ascertained.

Drs. Myers and Uyehara raised the question as to whether it would have been more appropriate during our engine testing program to have added a mixture of nitrogen and hydrogen to the ammonia rather than hydrogen alene to simulate the ammonia dissociation products. W~ jiRnsidered doing this when we initiated our hydrogen4adcrf'tion studies. However our calculations indicated that the decrease in energy content of the fuel mixture due to this additional nitrogen would amount to only' about 1,7,,/0. Also.wedid not wish

to further complicate the fuel supply system by adding a third gas supply.

Regarding the thermal inertia of the dissoc iator, we believe that a full scale dissociator will have to be constructed and tested on an engine in order to determine the magnitude of this problem, We appreciate that some lag in engine accelerat4</,: will result through the use of a dissociator. This may be compensated partially by the fact that less hydrogen (a's a percent of the total fuel supplied) is required when operating at full throttle with a supercharger than when running at part load, In fact. we have mentioned in our

@society of Automotive Engineers, Inc ,; 1966 Library of Congress Catalog Card Number: 12-14987


l\lr. Wagner's comments regarding his investigation of

ammonia as a fuel for compression ignition engines are appropriate since our work was confined to spark ignition engtnes , We concur with him that the combusron of ammonia in a comp~ssion ignition engine should also be investigated, since a larg~ number of mi !itary vehicles are powered presently by diesel engines ,

In summary, we wish to emphasize that the work described in our paper was performed during a relatively short period of time. Engine testing was started in March of 1963 and was concluded in January of 1964. Basic ammonia dissociation studies 'were conducted through the summer of 19G4, " Engine tests were performed only at steady state operating J;ondltiolls while using manual controls. No transient engine conditions were investigated. In view of the corrstructive criticism of the paper by the several discussors as well as our' own thoughts on the matter. we believe that additional studies are needed to provide a more exhaustive evaluation of ammonia as a fuel for military vehicles powered by spark-ignited engines,

Printed in U. S . A

Sign up to vote on this title
UsefulNot useful