Mechanism, Kinetics, and Equilibrium of

Thermal Decomposition of Ammonium Sulfate

Raisaku Kiyoura and Kohei Urano

Tokyo Insitute of Technology, 0-okayama, Meguro-ku, Tokyo. Japan

The thermal decomposition mechanism of ammonium sulfate, the reaction rate and
equilibrium of the deammoniation from ammonium sulfate, a n d the reaction rate of
the dehydration of ammonium bisulfate were investigated b y DTA, TGA, a n d x-ray
diffraction. Ammonium sulfate on heating decomposed into triatnmonium hydrogen sulfate
or ammonium bisulfate, which subsequently decomposed into ammonium pyrosulfate
or sulfamic acid, and finally into several gases. The activation energies for the deammonia-
tion a n d dehydration were calculated, a n d the enthalpy change of the deammoniation
was evaluated. The results of this work should be helpful in many industrial processes
dealing with ammonium sulfate or its thermal decomposition products a t elevated tem-
peratures.

THE thermal decomposition of ammonium sulfate has

been studied since the early part of the 20th century.
chemical analysis. The amounts of NH,' in solutions of
the heated residues were analyzed by colorimetry, and
Dixon (1944) investigated the residue of heating in an the amounts of H - and Sod2- were evaluated by titration.
atmosphere of wet room air. Duval et al. (1959) and Kinetics and Equilibrium. Studies on the thermal decom-
Erdey et al. (1964) studied the thermal stabilities of the position mechanism concluded that the deammoniation
reagents by TGA. Rafal'skii and Ostrovskaya (1964) pro- reaction occurred first, producing ammonium bisulfate,
posed a theory for the decomposition mechanism from which decomposed mainly into ammonium pyrosulfate by
kinetic studies. However, many problems, such as the dehydration, and further into gases.
decomposition equilibrium or the boiling point of With sufficient water vapor in the system, the weight
ammonium bisulfate, are not yet solved. decrease on heating could be attributed only to the release
Many industrial processes dealing with ammonium sul- of ammonia. Therefore, the deammoniation rate may be
fate or its thermal decomposition products a t elevated obtained from the rate of weight decrease. Likewise, the
temperatures have been proposed. John (1962) and Datin rate of ammonium bisulfate dehydration may be obtained
(1965) proposed production of fertilizers from rock phos- by heating in dry air when other decompositions are
phate with ammonium sulfate or bisulfate. McCullough negligible. It is also necessary to select sample quantities
(1960), Montgomery (1962), Haeter and Reichau (1963), and temperatures such that the diffusion of the released
and Bonfield and Bohn (1966) proposed recovery of gases does not control the rates of decomposition. I n
ammonia and sulfur oxides by the thermal decomposition this case, the real chemical reaction rates corresponding
of ammonium sulfate. Kiyoura (1966a,b) developed a t o the maximum decomposition rates a t the given tem-
process for the production of ammonium sulfate from hot peratures can be measured. About 250 mg. of ammonium
flue gases by recovery of sulfur dioxide. sulfate was used in the temperature range from 140" to
The present work gives the mechanism, kinetics, and 240" C. for measurement of the deammoniation reaction
equilibrium of the thermal decomposition of ammonium rates, About 50 mg. of ammonium bisulfate was used
sulfate to enhance understanding and its application in in the range of 160" to 200°C. for measurement of the
industry. dehydration reaction rates. The rates of weight decrease
were observed with a thermobalance a t constant tem-
Experimental
peratures.
Mechanism. DTA and TGA curves were plotted by The total equilibrium of the thermal decomposition of
apparatus assembled in the laboratory with automatic ammonium sulfate cannot be measured by conventional
temperature programmers and recorders. X-ray powder methods, because it releases too many kinds of gases.
diffractions were carried out with CuK, radiation and The authors devised a method, where only the deammonia-
a nickel filter a t room temperature. The released gases tion equilibrium was determined with the thermobalance
were analyzed qualitatively by gas chromatography and (Figure 1). A quartz container was preferred to platinum,

Ind. Eng. Chem. Process Des. Develop., Vol. 9, No. 4, 1970 489
 the given atmosphere. At this temperature, the equilibrium
pressures of ammonia will be equal to its concentration
in the given atmosphere. The gas flow rates had no effect
from 25 to 100 ml. per minute. Accuracy of the tem-
perature measurements was within f 2 " C., and weight
measurements were within + O . l mg.
Results and Discussion

Mechanism. The DTA curve for ammonium sulfate

(Figure 2 ) shows two endothermic peaks, indicating that
decomposition occurs in at least two steps. The TGA
curves (Figure 3) at the lower heating rate show an
inflection point corresponding to a loss of 1 mole of
ammonia from 1 mole of ammonium sulfate. The release
of ammonia was observed by gas chromatography, and
found to occur above 100°C. The reaction expressed by
Equation 1 was inferred.

I I
However, the x-ray patterns of the heated residue of
Figure 1. Diagram of thermobalance ammonium sulfate (Figure 4) show no ammonium bisulfate
A. Gos storage G. Thermocouple at 160°C., but confirm the presence of triammonium
8. Pump H. Transformer hydrogen sulfate, (NH4)sH(SO4j2.
C. Water bath I. Temperature programmer The x-ray patterns of a mixture of 1 gram-mole of
D. Differential coil 1. Displacement meter ammonium sulfate and 1 gram-mole of bisulfate ground
E. Furnace K. Recorder
in a mortar showed that triammonium hydrogen sulfate
F. Sample container
was easily formed on grinding. This double salt released
ammonia on heating a t 180" C. and changed to ammonium
as the latter promoted decomposition, catalytically. bisulfate. However, when heated a t 160°C. for 4 days,
Ammonium sulfate was heated in an atmosphere of nitro- it showed no weight decrease. Figure 2 shows the DTA
gen, ammonia, and water vapor. Water vapor repressed curve of this salt.
dehydration and subsequent decomposition. The heating
temperature was elevated step by step a t intervals of
about 2 hours. The authors determined the highest tem-
perature at which no weight decrease was observed in

Temperatm 'C.
100 200 300 400
(NHhSQ

Temperature "C.
Figure 2. DTA curves of ammonium sulfate, bisulfate, and
the double salt heated a t 4°C. per minute in dry air Figure 3. TGA curves of ammonium sulfate

490 Ind. Eng. Chem. Process Des. Develop., Vol. 9, No. 4, 1970
 and decomposition. Qualitative analysis of the released
gases showed the presence of water, ammonia, sulfur diox-
ide, sulfur trioxide, sulfuric acid, hydrogen, and nitrogen.
The TGA curves (Figure 5 ) show no inflections even
a t the lower heating rate of 0.25'C. per minute, but
show that the weight decrease ceased a t a point of 8 %
decrease in weight, when it was heated a t 200'C. for
12 hours. The decomposition was repressed by water
vapor, confirming that dehydration is the first step in
the decomposition of ammonium bisulfate. The reaction
expressed by Equation 3 was inferred.
20 30 40
D i f f r a c t i o n Angle 28'
Figure 4. X-ray diffraction pattern of ammonium sulfate However, sulfamic acid (NH2S03H) formed a brown melt
heated at 160"C.for 2 weeks in dry air around its melting point (205"C.) and decomposed to
many gases, such as hydrogen, ammonia, and sulfur diox-
ide. The DTA curve of the mixture of sulfamic acid
and ammonium bisulfate (Figure 6) shows a large
I t is concluded that ammonium sulfate releases ammonia exothermic peak. The x-ray pattern of the heated sample
above 100°C. as in Equation 1, but from 100" to about of this mixture (Figure 7) shows the presence of ammonium
170" C., the ammonium bisulfate produced forms tri- pyrosulfate, confirming that sulfamic acid easily reacted
ammonium hydrogen sulfate directly with the ammonium with ammonium bisulfate to form ammonium pyrosulfate
sulfate. Thus the deammoniation reaction above 170" C. as in Equation 4.
can be written as in Equation 1, but the reaction up
to 170°C. may be expressed by Equation 2 .
NHZSO3H + NH,HSO, + (NH4)ZSzO; (4)
The x-ray patterns of the heated residues of ammonium
2(NH,)?S04 2 (NHI)3H(S0,)2 + NH, T (2) bisulfate did not show any sulfamic acid, but confirmed
the presence of ammonium pyrosulfate. However, the pres-
The DTA curve of ammonium bisulfate (Figure 2) shows ence of small amounts of sulfamic acid in the heated
two endothermic peaks corresponding to the melting point residue was ascertained, because the heated sample was

1CO
Temperature 'C.
90. 100 200 300 4co
A. 4'C. per min. in dry air
80 . B. 0.25'C. per rnin. in dry air
C.4'CC. per min. i n dryair
7o kept at 200'C.for 72 hrs.
.0
_ D. 4'C. per min. in 25 per cent HP
w
2 60 -
a,
%
$ 50:
8
+
5~40 ~

%
30 -
1I ,

Figure 6. DTA curves of sulfamic acid heated at 4°C.per

Figure 5 . TGA curves of ammonium bisulfate minute in dry air

Ind. Eng. Chem. Process Des. Develop., Vol. 9,No. 4, 1970 491
 According to these studies on the mechanism, the
decomposition temperature of ammonium sulfate is lower
than previously stated. The boiling point of ammonium
bisulfate in Lange's handbook (1961) is unreliable, because
ammonium bisulfate decomposes before boiling. Erdey's
decomposition mechanism, expressed by Equations 1 and
6, is rather simple and inadequate for the thermal decom-
position of ammonium sulfate.

20 30 40 Kinetics and Equilibrium. The decomposition reaction

Diffraction Angle 2 0 " rate is in proportion to the amount of undecomposed
ratio of sample. I t can be expressed by Equation 7,
Figure 7. X-ray diffraction pattern of mixture of sulfamic
acid and ammonium bisulfate heated a t 210°C. for 2 minutes dX
in dry air -= h(100 - X) :.ln(100 - X)= -ht (7)
dt
brown and more acidic than the mixture of ammonium where X = decomposed ratio in percentage, t = heating
bisulfate and pyrosulfate. A slight increase of N H 4 + in time in hours, k = reaction rate constant in hrf'.
its solution was observed on heating, and many gases The deammoniation reaction of ammonium sulfate is
such as hydrogen were released below the decomposition expressed by Equation 2 (up to 170°C.) and Equation
temperature of ammonium pyrosulfate. 1 (above 170°C.). Therefore, the decomposed ratio X
I t can be concluded that ammonium bisulfate releases can be written as in Equations 8 and 8' for the deammonia-
water above 150°C. as in Equation 3, but most of the tion. For the dehydration of ammonium bisulfate
sulfamic acid produced changes to ammonium pyrosulfate expressed by Equation 5, it can be written as in Equation
by reaction with ammonium bisulfate. The melt of 9.
ammonium bisulfate becomes stable solid again, and the
dehydration reaction is written as in Equation 5.
X=
W" - w
t -~wo x 100 = 1553 wo-wt*
,

MI '2M2 wo '
up t o 170°C. (8)
The reversibility of the reactions expressed by Equations
1, 2, 3, and 5 was ascertained by TGA, x-ray measure- - w,-wt -~
. wo x 100 = 777
wo - wt.
ments, and chemical analyses.
Erdey et al. (1964) reported that ammonium pyrosulfate
MI ' Me wo '
released sulfur trioxide and changed to ammonium sulfate above 170°C. ( 8 ' )
above 250" C. However, ammonium sulfate decomposes
quickly into ammonium bisulfate, and changes further X = W,I - wt -~
. Wo
x 100 = 1279
wo - wt (9)
a t this temperature. M.3 ' 2M4 W"
Ito and Uchida (1952) reported the presence of a double
salt of sulfamic acid and ammonium sulfate. The DTA where W o and W , = weight of salt after 0 and t hours,
curve of this salt prepared from the solution (Figure 6) and M1, M z , MB,and M , = molecular weight of ammonia,
shows that this salt is not stable above 150"C., and may ammonium sulfate, water, and ammonium bisulfate,
be irrelevant to the decomposition of ammonium sulfate. respectively.
Further studies on the complex decomposition of sul- From the rate of weight decrease measured with a
famic acid have to be carried out. thermobalance a t constant temperatures, the reaction rates
The mechanism of the thermal decomposition of were calculated by Equations 7, 8, 8', and 9.
ammonium sulfate is illustrated by Figure 8. Figures 9 and 10 show the logarithmic undecomposed
percentage (100 - X) as functions of the heating time,
t. These are projected as straight lines a t least in early
stages, when the other reactions were negligible. Figure
11 shows the relations between temperatures and the reac-
tion rate constants calculated from the gradients of Figures
9 and 10, and expressed by Equations 10 and 11. The
activation energies calculated by Equations 10 and 11
are, respectively, 16 and 13 kcal. per mole for deammonia-
tion and dehydration.
log hl = -3.46 x 103/T+ 5.03 (10)
log hz = -2.83 x 103/T+ 5.53 (11)
where h l and k l = rate constants of deammoniation and
Figure 8. Thermal decomposition mechanism of ammonium dehydration in hr.-', and T = heating temperature in
sulfate K.

492 Ind. Eng. Chem. Process Des. Develop., Vol. 9, No. 4, 1970
 100
200
c. 150
90 250
80
70
\
0 : De-ammoniation -1
.-0 "
%eo
01 2 io-' 10-I
In
CI
c
ru
cIn
-'
c
0
0

0 2 4 6 8 1 0 1 2 1 4
Time days

Figure 9. Reaction rate of deammoniation of ammonium

sulfate

10' 1 0-3
100 1.9 2.0 2.1 2.2 2.3 2.4
90
'
.-0
80
Temperature 103/"K.

Figure 11. Arrhenius plots of deammoniation of ammonium

sulfate and dehydration o f ammonium bisulfate
T 70
CL
'C.
200 150

\
4v ' 1

0 1 2 3 4 I

Time hrs.
b
Figure 10. Reaction rate of dehydration of ammonium bi- \
sulfate

Figure 12 shows the relation between the equilibrium

i
pressures and temperatures for deammoniation of
ammonium sulfate, as expressed by Equation 12. The b
enthalpy change of 34 kcal. per mole is evaluated by
Equation 12.
log K , = -7.34 x 103/T+ 11.52 (12)

where K , = equilibrium pressure of ammonia in

atmospheres.
By this equilibrium relation, the theoretical yield of
Kiyoura's process for the synthesis of ammonium sulfate
from hot flue gases can be evaluated. Flue gases will
contain sufficient water vapor, so that dehydrated mate-
rials will not be produced. Ammonium sulfate, bisulfate,
or the double salt of these two will be produced. Other 1.8 1.9 2.0 2.1 2.2 2.3 24
experiments have shown that the synthesis reaction rate
Temperature 10YK.
is very rapid. Therefore, ammonium sulfate will be pro-
duced until the concentration of ammonia equals its Figure 12. Equilibrium pressure of deammoniation of
equilibrium pressure at the temperature of the final state ammonium sulfate

Ind. Eng. Chem. Process Des. Develop., Vol. 9,No. 4, 1970 493
of the synthesis process. An insufficiency of ammonia Literature Cited
will produce ammonium bisulfate or triammonium hydro-
Bonfield, J. H., Bohn, F. L., U. S.Patent 3,282,646 (1966).
gen sulfate-for example, the theoretical yield of Datin, R. C., U.S. Patent 3,172,751 (1965).
ammonium sulfate can be calculated when 0 . 2 7 of the Dixon, P., Nature 154, 706 (1944).
sulfur dioxide in flue gas is completely oxidized to sulfur Duval, C., Wadier, C., Servigne, Y., Anal. Chim. Acta
trioxide, and 0.4% of the ammonia is added to the system 20, 263 (1959).
at 200°C. The equilibrium pressure of ammonia a t 200°C. Erdey, L., Gal, S., Liptay, G., Talanta (London) 11, 913
calculated from Equation 12 is atm., which is equal (1964).
to 0.014.Therefore, 0.39% of the ammonia will react, Haeter, L., Reichau, U., Ger. Patent 1,151,492 (1963);
leaving behind 0.01”C a t 200°C. Thus 0.19% of the sulfur C A 59, 8386h (1963).
dioxide will change to ammonium sulfate, and the residual Holowaty, M. O., U. S. Patent 2,899,277 (1959).
0.OlLCwill change to ammonium bisulfate. Ito, S., Uchida, S., J . Chem. SOC.Japan. Pure Chem. Sec.
73,348 (1952).
Miller’s (1955) process for pickling iron with ammonium John, C. M., U. S. Patent 3,047,369 (1962).
bisulfate, Holowaty’s (1959) process for recovering Kiyoura, R., Chem. Erg. News 44 (26), 23; (27), 36 (1966a).
ammonia from coke oven gas, and the many industrial Kiyoura, R., J . A i r Pollution Control Assoc. 16, 488 (1966b).
processes mentioned above can be investigated in detail Lange, N. A., Ed., “Handbook of Chemistry,” 10th ed.,
by utilizing Equations 10, 11, 12, and Figure 8. p. 219, McGraw-Hill, New York, 1961.
McCullough, R . F., U. S. Patent 2,927,001 (1960).
Miller, C. O., U. S. Patent 2,700,004 (1955).
Montgomery, J. C., U. S. Patent 3,047,369 (1962).
Acknowledgment Rafal’skii, N. G., Ostrovskaya, L. E., Geterogennye Reaktsii
i Reakts. Sposobnost 1964, 95 (1964); C A 64, 15057h
(1966).
The authors express sincere thanks to H. Kuronuma
and G. Uwanishi of the laboratory for their useful sugges- RECEIVED
for review December 18, 1968
tions. ACCEPTED
April 25, 1970

Model Predictive Time-Opti mal

Control of Second-Order Processes
D. A. Mellichamp
Department o f Chemical and Nuclear Engineering, University of California, Santa Barbara, Calif. 931 06

Set point changes for second-order processes are implemented using a hybrid computer
and a fast-time model of the controlled process to search out the time-optimal trajectory.
Process outputs are used as initial conditions to monitor the process transition in continuous,
closed loop fashion. Model scan rates up to 100 per second (with convergence to
the optimal solution in milliseconds) furnish a continuous updating of predicted switch
times for the controlled process and make possible the multiplexing of a single hybrid
controller to a large number of controlled processes. The controller has excellent charac-
teristics for set point changes and also as a regulator with both analog simulated
processes and an experimental two-tank liquid level system.

C o n t r o l methods that use predictive techniques are body of mathematical theory has been developed which
applied in many critical aerospace and military situations. can be used with a linear system model to generate a
The ability to predict and program system changes in control procedure that will be optimal in some sense.
advance is a well-recognized practice in the supervision Procedures differ, depending on the immediacy of the
of space probes and satellite orbit changes, and control situation: The rocket burn time for a deep space probe
of high performance aircraft and submarines. Most predic- obviously can be computed months in advance; a similar
tive techniques rely on a rather sophisticated mathematical calculation for satellite rendezvous may have to be made
model of the controlled system; a computer is used to as approach occurs. I n either case, although minor correc-
scan a large number of possible control schemes, from tions will have to be made during the transition to take
which the most useful is selected. Alternatively, a large into account discrepancies between the actual system and

494 Ind. Eng. Chern. Process Des. Develop., Vol. 9, No. 4,1970