STUDIES OF RDX AND RELATED COMPOUNDS

VII. RELATION BETWEEN RDX AND N M X PRODUCTION

IN THE RACHMANN REACTIONL

-\BSTRACT
T h e reactions t o form R D X and H M X in Rachmann-type ~ n i s t i ~ r are
es co~np-
arable in respect of optimum nitric acid concentrations and in the fact t h a t
optimal amounts of acetic anhydride and ammonium nitrate are necessary for
masimum yields of either explosive. T h e activation energies for formation of R D X
and H M X were also found to be comparable, at 15 f 1 kcal. per mole. However,
withholdinganimo~~iuninitratefrom thereaction mixture \\.as found t o have a more
tleleterious effect on R D X production than on H M X production. A n ~ e c h a n i s nis~
proposed which a t t e m p t s in a general way t o represent the relation between
R D X a n d H h I X production in the type of reaction mixtures used.
INTRODUCTION
An accurate method for analysis of HNIX (cyclotetramethylenetetranitra-
mine) in RDX (cyclotrimethylenetrinitramine), produced by the Bachmailil
reaction (I),was described in a preceding paper (2). In the present investigation
the method was applied to a study of the relative rates of formation of R D X
and H M X , and of the influence of certain factors on the relative yields of R D S
and H M X in Bachmann type reaction mixtures.
EXPERIMENTAL
The materials employed were all of commercial grade. The hexamine, acetic
acid, acetic anhydride, and ammonium nitrate were 99.0, 99.3, 95.0, and 99%
pure respectively, while the nitric acid was 97y0 HNO3.
Ammoilium nitrate of accurately weighed amount was placed in 250 cc.
Erlenmeyer flasks and the desired molar ratio of nitric acid added to bring the
ammonium nitrate into solution. The ammonium nitrate was then reprecipitated
as fine crystals by addition of acetic acid to the dilution required. -Acetic an-
hydride of a designated amount was then added and the flask and its contents
shaken in a constant temperature bath, the temperature of which was controlled
within =t 0.2' C. After some minutes, the reaction was initiated by adding 5 cc.
of 1 M hexamine -acetic acid solution, which caused any solid anlmonium
nitrate present to dissolve almost immediately.
By this sequence of addition and by using large proportions of acetic acid,
any effect due to rate of solution of anlmonium nitrate was practically eliminated
and the total amounts of all reagents were brought together initially in a virtually
homogeneous reaction medium. In this manner, satisfactory reproducibility of
results was realized.
The reaction was stopped a t the desired time by addition of approximately
30 cc. of water. Nitric acid (5 cc.) was then added and the resulting
mixture simmered on a hot water bath for approximately four hours
il.lunuscript received M a y 16, 1952.
Cnntribz~tionfrorirthe Plcysical Clce?rcistryLaboratory, McGill Urziuersity, Jfontrecrl, Qzie.,
with firlar~cial assistarcce from the National Research Coz~ncilof Carzada.
 EPSTEIN A N D I1'1rYKLER: RDX. 171 733

t o destroy any B S S (1,7-diacetoxytetramet1~ylene-2,4,6-tr1nitra1n111e) pres-

ent (1). Blanli experiments with the product fro111 reaction mixtures con-
taining no amrnoniu~nnitrate showed that a three hour simmer was sufficient to
clestroy all the solids present. The Aask was then filled to the 11ecli with water
and permitted to stand a t room temperature for taro clays to ensure complete
precipitation of RDX. T h e precipitate was filtered on n sintered glass crucible,
dried a t 10SOC. for a t least four hours, and weighed. Appropriate empirical
COI-rections were applied for loss by solubility in the filtrate. The resulting R D X
was analyzed for HR4X by the method of differential homogeneous hyclrolysis ( 2 ) .
The percentage yield of R D X was calculated on the assumption that 100yo
yield represents the production of 2 moles R D X per mole hexamine. The H M X
yield was calculated on the basis that 1 mole of hexamine is capable of yielding
1 mole of H M X . The reported yields generally represent the mean value for two
or more identical experiments for which the results agreed within about 5y0.
T h e rates of prodilction of R D X and HMX were estimated as initial rates, fro111
tangents to the reaction-time curves a t zero time.
RESULTS
Efect of A?nmoniu?n Nitrate on Relative Yields and Rates of Production of RDX
and HllifX
Results given in Fig. 1 show that, under the conditions of high dilution used
MOLAR RATIOS
ACETIC NITRIC ACETIC HEXAMINE
---
ACID ACID ANHYDRIDE
FINAL YIELD 104 5.2 20 I
0 194 5.2 2D I
A " 194 9.5 20 I

50L- 1 2 3 4 5 6 7 8
MOLES A M M O N I U M N I T R A T E P E R MOLE H E X A M I N E
Fig. 1. Effect of arntllo~liumnitrate on relati\.e.yields of RDX and HMX.
in these experiments, the optimal mole ratio of ammonium nitrate (about
2.75 moles per molc hexamine) for R D X production a t 4S0 C. did not var!-
with clilution, but increased with increase of nitric acid concentration in the
reaction mixture. One experiment a t 55' C. showed no variation of the optimal
ammonium nitrate mole ratio with temperature. The production of IJMX
showed rather remarkable sensitivity to t h e ammonium nitrate mole ratio anti
to dilution as shown in the lower curves of the figure; the optimum yield of Hn/IX
occurred a t about 2 . 3 moles a r n m o n i ~ ~nitrate
m per mole hexa~iiine.
The initial rates of R D X procl~~ction increased ~ ~ n tthe
i l optimal ammoniunl
nitrate mole ratio was reached, but the initial rates of HhiIX procluction were
practically independent of ammonium nitrate mole ratio in the range studied.
E f e c t of N i t r i c A c i d o n Relalive Y i e l d s a,)td R a t e s of Productiolz of R D X ~ n Hdl I f X
Optimum nitric acid - hexa111ine ratios were determined for the follo\ving
conditions :
---
Temp.
45°C. I 55°C.
NIoles per mole hexamine
.Acetic acid 104 150 194 370 104
Acetic anhydride 20 40 20 20 20
.\mmonium nitrate 3 and 2 3 3 and 3 3
6.25
I

The effect of dilution, temperature, and acetic anhydride on the optimu~n

nitric acid mole ratio is shown in Fig. 2, while the effect of ammonium nitrate
is shown in Fig. 3. For given conditions, the opti~nunlnitric acid ~iloleratio for
RDX production was usualljr about the same as that for H M X production..
However, the optimum was much less pronounced for the latter, as a result of
which the percentage H M X in the explosive produced generally showed asteady
decline, which continued even after the optimum nitric acid mole ratio had been
considerably exceeded.
Dilution with acetic acid a t given ammonium nitrate mole ratio appeared to
decrease the yield of R D X and to increase the H M X content of the explosive
mixture (RDX + HMX). On the other hand, increase of ammonium nitrate
mole ratio a t given dilution increased the yield of R D X a t the optimum nitric
acid mole ratio, but clecreased the HR4X yielcl. Loss in R D X yield as a result
of a limited increase in nitric acid beyond the optimum value, or by dilution a t
given ammonium nitrate level, was apparently reduced or overcome by increase
in the a m m o ~ l i i ~ nitrate
m mole ratio.
An increase of temperature increased the optimum nitric acid mole ratio for
R D X production (from 5 . 2 to 6 moles per mole hexamine) without appreciable
alteration in optinlum yield. The optimum of about 5 . 2 moles of nitric acid per
mole hexamine for H M X production a t 45' C. gave way, however, to a steady
increase in yield of H M X a t 55' C. as the nitric acid level was i~lcreaseclfrom
4 to 8 nloles per mole hexamine.
Doubling the acetic anhydride mole ratio (with adjustment of acetic acid
 E P S T E I N ..IND I1.INRLER: RDX. 1.11

HEXAMINE AMMONIUM
NITRATE
3
.
0
ACETIC
ACID
104
194
0 370
ACETIC
ANHYORIDE
20
20
20 4
TEMP.

45-
45O

I # , # I I I I

4 5 6 7 8 9 1 0 1 1 1 2
MOLES N I T R I C ACID PER M O L E HEXAMINE
Fig. 2. Effect of dilution and acetic anhydride on opti~nurnnitric acid mole r a t i o lor
RDX a r i c l HMX production.
7
HEXAMINE ANHYDRIDE ACETIC AMMONIUM
----
20
ACID
0 1 0 4
I
NITRATE
3
MOLAR A 194 3
RATIOS A 104 2
6.25
70 -

50-

10 I I I I I I I I I

30

$20-
I
X
5 10-
0
I

, I

3 4
5 6 7 8 9 1 0 1 1
M O L E S N I T R I C A C I D PER M O L E HEXAMINE

Fig. 3. Effect of a m t n o n i ~ ~ nitrate

m on opti~nurnn i t r i c acid mole ratio for RDX and
H h l S production.
736 CAN.IDIAIV JOURIVAL OF CHEBIISTRY. T'OL. 30

content t o retail1 ~ l l esallle total volun~e)increased the optimu~n~ n o l el-atio of

nitric acid somewhat and decreased the optilnu~nHRiIX ~rieldwhile the I i D x
yield was increased slightly.
T h e maximum rates of R D X and H M X formation were generalll. observed
a t the nitric acid nlole ratios corresponding t o optimum yields.
Effect of Acetic Anhydride on the Relative Yields and Rates of Formation of R D X
and lll14X
,4t given a~nrnonil~m nitrate and nitric acid mole ratios (3 moles and 5 . 2 moles,
respectively, per mole hexamine), the acetic anhydride content of the reaction
mixture was changed over a range from 5 moles t o 70 moles.per mole hexamine,
with corresporlding acljustment of the acetic acid from 147 t o 17 moles per mole
hexamine to retain constant volume of the reaction mixture. T h e results, given
in Fig. 4, show that in these systems of high dilution, there was little change in
HMX
RATE OF FORMATION 0 A

0 5 10 15 20 25 30 35 40
MOLES A C E T I C A N H Y D R I D E PER M O L E H E X A M I N E
Fig. 4. Effect of acetic anhydride on relative yields and initial rates of formation of R D S
and HMX.
yield of R D X , while the H M X yield fell off, beyond about 15 moles anhydride
per mole hexamine. On the other hand, the initial rate of production of RDX
continued t o increase with anhydride mole ratio while the initial rate of HMX
production reached a steady value.
Effect of Dilution on Rate of R D X and I f M X Formation W i t h and Withoz.~tAdjzut-
ment to Optimum Nitric Acid Mole Ratio
T h e initial rates of formation of R D X and HR4X were found t o suffer some-
what less by dilution when the nitric acid mole ratio was adjusted to its optimal
value than when no such adjustment was made. Some typical data were as
follows:
I I
\~olumeof reaction medium (cc.) Nitric acid - hesamine Initial rate, .% yield
per 0.005 moles hexamine ratio per mln.

*Oplmrnlfor dilzllio?~i?zdicnled.
 E P S T E I N A N D IYINKLER: R D S . V I I 739

'I'here is some indication from the relative extents to which dilution affected the
rates that the reaction to form R D X is of higher order than that to produce
HMX.
Effect of Temperature
T h e rates of production of R D X and H I 4 X mere determined a t three te~npera-
tures (3S0, 4S0, and 55' C.), using reaction mixtures containing 0.075 ~nojes
(hexamine) per liter a t each temperature. Initial rates were estimated inde-
pendently by two individuals, using both large and small scale graphs, by drawing
tangents a t zero time. From the initial rates, an activation energy of 15 & 1 kcal.
per-mole was estimated for both R D X and H M X formation.,
Production of R D X and HA4X when Addition of A m m o n i u m Nitrate i s Delayed
Stiidies of a type previously reported (4) were made to determine the effect
oE delayed addition of ammonium nitrate on the relative production of R D X
and H M X under a variety of conditions. T h e only alteration in procedure was
to add the ammonium nitrate a t given intervals after reaction had been initiated
in its absence by addition of hexamine. Reaction time was calculated from the
time of addition of ammonium nitrate. The effect of increasing the length of time
between initiation of reaction and addition of ammonium nitrate is shown in
Fig. 5. As in the earlier study (4), serious loss (approximately 30%) of R D X was

Fig. 5. Effect of delayed addition of ammonium nitrate 011 relative yields of RDX and
HMX.

suffered when the addition of ainmonium nitrate was delayed for as little as
15 ~ n i n T
. h e corresponding loss in H b l X was not Inore than 10-12%.
Sufficiently prolonged delay (five hours or more) reduced the yield of both
R D S and HI\/IX to about one-third of the yield. available when ammoniun~
nitrate was added initially.
IYhen the addition of ammonium nitrate was delayed 20 min. (i.e. beyond tlre
first rapid decrease in R D X yield), the relation between yielcl and mole ratio of
a m ~ n o n i i ~nitrate
n~ e\~entuallyadded was practically identical with that vihen
the acldition was not delayed. An optimal mole ratio for R D X production of
2 . 7 5 moles per mole hexamine, and for HR'IX production of 2 . 5 moles per mole
hexamine, \\:as observed. The maximum yields of R D X and I-IMX were 48%
and 26y0 respectively. Excess ammonium nitrate decreased the H M X yielcl,
but not the R D X yield, below this maximum value.
T h e optiinum nitric acid mole ratio in systems where ammonium nitrate
acldition was delayed 20 min. was also identical with that when the addition
was not delayed. Likewise, the activation energies for the formation of R D X
and H M X , determined from initial rates, were essentially the same (15 and
13 Itcal. respectively) when addition of a~nmoniumwas delayed 20 min. as
when the nitrate was present froin the start of reaction.
Comparison ol the initial rates in reaction mixtures in which ammonii~nn
nitrate addition was delayed 20 rnin. with those obtained \vhen the nitrate was
not withheld revealed t h a t the initial rate of H M X forination was essentially
unaltered, while the initial rate of R D X production was reduced to about one-
quarter the normal value by delaying the addition.
DISCUSSION
From the similarity in optimum nitric acid mole ratios and in changes of
these mole ratios with varying conditions, together with the observation that
optimum amounts of acetic anhydride and a~nnloniumnitrate are necessary for
nlaximunl yields of both R D X and H M X , it seems reasonable to conclude that
the reactions to produce R D X and HRlX are not only similar, but closely
related. I t appears, however, that hexamine is converted into a t least two com-
pounds under the conditions of the experiments, from one of which both R D X
and H M X nlay be derived, while the other gives rise oilly to R D X . This is
indicated by the fact that when ammonium nitrate is withheld from the system,
one-third of the prospective R D X is lost rapidly without correspondingly
serious loss of HNIX production, while longer delay in adding aininonium nitrate
permits essentially parallel production of R D X and H M X . Also, with all other
conditions constant, increasing amounts of ammonium nitrate or acetic an-
hydride increase ~narlcedlythe rate of R D X , but not of HRlX for~nation,while
a decrease in H M X yield is roughly compensatecl by an increase in R D X yield.
The rate-controlling step for the formation of H M X from the colnmon precursor
is therefore apparently not influenced by excesses of these reagents, while t h a t
for R D X formation is favorably affected.
Since a t least one-half of the nitric acid necessary for maximum yield is
required before any RDX or H M X is produced, whereas some R D X and H M X
may be produced with as little as one-sixth of the amount of ammonium nitrate
necessary for maximum yield, the hexamine molecule appears to be attacked
first by nitric acid - acetic anhydride. Further indication of this sequellce of
reaction is obtained from the experiments in which the addition of ammonium
nitrate is delayed. T h e initial rates of H M X formation in these experiments are
of the same order of magnitude as those in the experiments in which a~nmonium
nitrate addition is not clelayecl. This might be expected i f nitric acid ancl acetic
anhydride determine the initial steps in the conversion of hesaminc to the
precursors forming H M X and R D X , ancl the subsequent addition of arnn~onium
nitrate permits conversion of the precursors, which have not been destroyed,
into R D X and HRIX. Comparison of the slo~vrate a t which nitric acid destroys
the precursors colnmon to I l M X and two-thirds of the R D X with the rn~lch
faster rate a t which these precursors are converted to R D X and H M X on
addition of ammonium nitrate indicates that ammonium nitrate does participate
later in the formation of H M X and RDX, a t least fro111their con~monprecursor,
ancl does not act merely as a buffer to prevent side reactions which might be
favored over R D X formation in the absence of ammonium nitrate.
T h e majority of the results obtained in the present stucly would appear to
be explained by the following scheme:

Stcp 1
HNOz (very fast)

HNOz

Six membered Ac.,O P r e c u r s o r (11) to H M S

ring precursor ( I ) and some R D ~(Possibly
.
to RDS 1 D P T (3,s)')

AcZO
(slow)

BSX (1) 1 RIIX HbIX Linear

+
fragments
+ +
fragrne~iis f r a g ~ n c ~ l t s
+
fragrnen ts
products
+
fragrnenls

followed b y :
Fragments + .%c?O + H N O j + N H , N O , +RDX + Hh4X
In this scheme, only the important general steps are included and it must be
remembered that any one of the steps above may involve a series of chemical
changes.
On the basis of the experimental evidence presented previously, steps 1, 2,
and 3 are assumed to involve hexamine, nitric acid, and possibly acetic anh>.dride
and acetic acid, but to be relatively little influenced by a~nmoniumnitrate.
If these steps are fast, the initial rates of formation of HhlX in the experiments
in which the addition of a m m o n i ~ ~ nitrate
m is delayed should be the same as
those forming HI\/IX when a~nmonii~rn nitrate is present initially, since the
concentratior~of prerursor (11) should be the same regardless of \\-hcther aln-
monium nitrate was present initially or not.
Although one-third of the potential R D X is destroyed within a few mini1tes
in the absence of ammonium nitrate, R D X and HhlIX may be produced from
their common precursor after several hours. I t would appear, therefore, that
the reaction proceeds even in the absence of an~n~oniilrn nitrate to the stage of
the sis-membered ring precursor and precursor (11). The six-membered ring
precursor is converted to R D X in the presence of ainn~oniumnitrate but is
destroyed rapidly in its absence. However, precursor (11), common to H M S
and part of R D X , is destroyed a t a slower rate, with parallel destruction of both
potential R D X and HRIX. I t may also be noted that in the absence of am-
monium nitrate the destruction of potential R D X and H M X a t 5S0 C. becoines
ver). illuch slower after four hours. This may be due to destruction of most of
precursor (11) after four hours, leaving only the relatively stable fragments to
be converted to R D X and H M X . These considerations indicate that the steps
5, 6, and 7 are rate controlling in production of R D X in the Bachmann reaction.
If step 5 is eliminated by delayed addition of ainmonium nitrate no decrease
in the consumption of nitric acid should occur, since BSX (1) contains the same
number of NOPgroups as R D X and is formed in place of R D X . Also, although
anlmonium nitrate is necessary for the production of R D X through step 5,
neither it nor its equivalent is iilcorporated in the R D X molecule, hence the
consumption of ammoniuin nitrate should be the same with or without the
occurrence of step 5. I t is interesting to note that the optimal nitric acid and
arnrnonium nitrate mole ratios are approximately the same whether step 5 is,
or is not, present. I t is also significant that the delayed addition of alnmonium
nitrate decreases the rate a t which R D X is formed to one-third of that in the
experiments where the addition of ammonium nitrate is not delayed. This is
not surprising since it would be a rerrlarkable coincidence if the rate of R D X
foi-mation through step 5 were the same as through 6. T h e eliminatioil of step 5
should result in slow formation of R D S by conversion of precursor I1 and
resynthesis of fragments.
T h e d a t a presented for the effects of temperature, dilution, and varying pro-
portions of reagents 011 the rates of formation and yields of R D X are in no way
opposed to the rnechanisnl suggested.
REFERENCES
K , E. arid SHEEHAN,
1. B A C H ~ I A NW. J . C . J. Am. Chem. Soc. 71: 1842. 1949.
2. EPSTEIN,S. and WINGLER,C . A. Can. J . Chem. 29: 731. 1951.
3. McI<al;. A. F., R I C H ~ ~ O N , and WRIGHT,
H.DH., G. F. Cali. J. Research, B,27: 4G2. 1948.
4. RALPH,A. O., ~ I A C H U T C HJ. , arid WINGLER,C . A. Can. J . Chem. 29: 725. 1951.
I NG.,
5 . RICHMOYD, H . H., MYERS,G. S., and WRIGHT,G. F. J. Am. Chem. Soc. 70: 3659. 1918.