Kinetics and Modelling

of the sulphur dioxide

oxidation to the sulphur
trioxide process

GROUP MEMBERS

Abdullah (19JE0014)
Aditi Nair (19JE0056)
Amlan Kumar Tripathy (19JE0119)
1 RESEARCH OBJECTIVE:

Through this research, we are making an attempt to select an appropriate kinetic equation, among

the numerous kinetic rate equations, to represent the experimental data for the SO2 oxidation to

SO3, as accurately as possible.

We are also interested in evaluating the range of validity for each kinetic model, i.e within what

range of pressure and temperature can a particular kinetic model give the most accurate result.

2.1 INTRODUCTION (IMPORTANCE OF MANUFACTURING H 2SO4):

Sulfuric acid (H2SO4) is an important chemical compound in the industry and is used in the

manufacture of a wide range of products. Sulfuric acid is a powerful mineral acid. It is a

colourless and viscous liquid that dissolves in water at all concentrations. Sulfuric acid is a very

important commodity and the production of national sulfuric acid is one of the indicators of its

industrial potential. To reduce the risk associated with its transport, sulfuric acid-producing

plants are often found near their surface, as they are corrosive and hazardous chemicals. Indeed,

it is more economical and safer to transport basic elemental sulfur than sulfuric acid. Sulfuric
acid is an important chemical in many process industries, such as fertilizer production, oil

refining, chemical compounding, lead-acid batteries.

The subject of this thesis is mainly focused on investigating the various kinetic equations

describing the conversion of SO2 to SO3 that are applicable for sulphuric acid plants, and

investigating the impact of various operating parameters on the oxidation reaction.

Then, an experimental design was used to select the most appropriate kinetic equation, which

best represents the experimental data. To select the most representative kinetic equation, the

residual sum of squares (RSS) of the differences between the predicted and experimental data in

the range of temperature of interest was minimized. The most appropriate kinetic model shows a

wide range of validity or the lowest average of RSS value in comparison with other kinetic

models.

Sulphuric acid is mainly produced by the contact process, which is today the leading technology

used in the manufacture of sulphuric acid because of its large production capacity of

concentrated sulphuric acid. This process consists of three main reactions carried in different

units:

1) elemental sulphur is burned with air to produce sulphur dioxide (SO2) ;

2) the sulphur dioxide is oxidized to sulphur trioxide (SO3) using oxygen (O2) from the air

(3) the sulphur trioxide is absorbed in a concentrated sulphuric acid column where SO3 reacts

with water to form additional H2SO4

As we can see, all three reactions are exothermic. The second step is the critical step in sulphuric

acid production mainly because SO2 oxidation to SO3 is an exothermic and equilibrium reaction

performed adiabatically such that a multi-bed catalytic reactor is required to achieve high

conversion.

The SO2 oxidation process is an equilibrium catalytic reaction, generally carried out in plug flow

reactors under adiabatic conditions. Industrially, a series of three to five catalytic beds are used to

perform this reaction with inter-stage gas-cooling. A schematic diagram of a four-bed catalytic

converter is presented in the figure below.

Figure 1:

The typical length of individual catalytic packed beds ranges between 0.2 and 2.0 m, with

diameters up to 12 m. Removing heat from the reacting gas mixture between catalyst beds needs

to be performed to increase the conversion. The conventional contact method is a very effective

process for making H2SO4 in view of the conversion of SO2 to SO3 (achieving 97.5%

conversion). It can be further improved by 1 to 2 % by using two stages of sulphur trioxide

absorption columns (double contact method). An intermediate absorption column is incorporated

between two catalytic beds to recover SO3 and enhance additional conversion.

2.1.1 Catalysts

The sulphur dioxide (SO2) oxidation to the sulphur trioxide (SO3) is a heterogeneous catalytic

reaction. Over time, with the development of the contact process, the use of alternative catalysts,

such as vanadium pentoxide (V2O5), having a longer lifetime and being significantly cheaper to

manufacture, V2O5 combined with alkali metal oxide promoters form the active phase of the

catalyst, which is typically added to porous silica as the support.

A composition of 1 mol V 2O5 with 2.5 moles of alkali metal sulphate is typically used in many

industrial vanadium catalysts.

2.1.2 Steps to find the appropriate kinetic model

In order to study the reaction kinetics of SO2 oxidation to SO3, we will make use of the kinetic

results obtained from the experimental design proposed by Doering et al.

While performing the experiment, the choice of catalysts was the same for both the kinetic rate

equations and the experimental data. As with other experiments, this experiment may also suffer

from outliers that are not in our control like experimental errors, incorrect data management or
model inadequacy. These outliers must be identified prior to the model evaluation; otherwise,

estimation errors may occur.

After performing the experiment and calculating the desired rate of reaction values, the residual

sum squares (RSS) of the differences between the predicted and experimental data were used to

assess the model validity over different ranges of operating conditions.

2.2 Mechanisms and rate laws

There are two main hypothetical reaction mechanisms that were proposed to explain the

observed experimental data on the alkali-V2O5 catalyst under certain operating conditions.

Calderbank and Eklund suggested a two-step mechanism:

(1) vanadium components oxidize the sulphur dioxide to the sulphur trioxide

(2) oxygen re-oxidizes V+4 to V+5. The degree of vanadium reduction in the first step was

measured based on the quantity value of [V+4 ] 2 ∙PSO3 / [V+5] 2 ∙ PSO2 over a large range of

reaction conditions and catalyst composition.

Calderbank has assumed that the vanadium reduction (step 2) was rate-determining and derived a

kinetic model.

Eklund worked on the same mechanism and proposed a different kinetic model that represented

experimental results more accurately. His model includes an equilibrium constant.

Villadsen and Livbjerg worked on the same mechanism which involves:

(1) vanadium components oxidize the SO2 and produce a [V+4.SO3] complex

(2) the complex dissociates into SO3 and V+4

(3) the V+4 reoxidizes to V+5 .

Villadsen et al. assumed that the oxidation of V4+ species (step 3) is a crucial step in the oxidation

of SO2 to SO3 and they derived a kinetic equation to explain their experimental results.

Collina et al.proposed a kinetic rate equation on the assumption that the reaction between

adsorbed SO2 and O2 from the gas phase is the rate-controlling step. In addition, the rate equation

properly explains their experimental data for a temperature range of 420-590°C.

The parameters of kinetic rate expressions are presented in Table 1.

 Table 1: Kinetic parameter and ranges of operating variables.

Parameters Temperature dependence T[C]

Model 1: Collina et al

k1 [kmol/kg cat atm² . s] k1 = exp (12.16 - 5473/T)

K2 [atm-1] K2 = exp (-9.953 + 8619/T)

420 - 590
K3 [atm-1] K3 = exp (-71.745 + 52596/T)

Kp [atm-0.5: Kp = exp (-10.68 + 11300/T)

Model 2: Eklund

k1 [kmol/kg cat atm s KP k1 = exp (848.14 - 97782.2/T - 110.1 InT) Kp = exp

[atm-0.5] (-11.24 + 11818.055/T) 420 - 554 -

Model 3: Calderbank

k1 [kmol/kg cat atm1.5. s] k1 = exp (12.07 - 15656.56/T)

370 - 450
k2 [kmol/kg cat k2 = exp (22.75 - 27070.7/T)

Model 4: Villadsen et al
k1 [kmol/kg cat atm . s] k1 = exp (-1.88 - 7466.08/T)

K2 [-] K2 = exp (2.10 - 286.74/T)

380 - 520
K3 [atm-1] K3 = exp (-1.51 + 2279.17/T)

Kp [atm-0.5] Kp = exp (-10.73 + 11318.3/T)

2.3 Maximum rate curve

Figure 2: Maximum reaction rate curve for the SO2 oxidation for a 10 mol% SO2, 11 mol% O2,

and 79 mol% N2 feed gas composition and 1.4 atm.

The figure above presents a series of constant reaction rate curves on a conversion temperature

plot for a 10 mol% SO2, 11 mol% O2, and 79 mol% N2 feed gas composition and 1.4 atm total

pressure.

The shape of the initial portion of a constant rate curve is due to the combination of the increase

in the reaction rate with temperature and the decrease of the reaction rate with conversion. As the

constant rate curve moves toward the equilibrium curve, it goes to a maximum point

corresponding to the maximum rate of reaction for a given combination of conversion and

temperature. Each constant rate curve is subjected to the same fate. For higher constant reaction

rates, the maximum occurs at lower conversion and higher temperatures.

If it would be possible to adjust the temperature of the packed bed reactor as a function of

conversion to constantly follow the maximum rate curve, it would require a length of bed or

catalyst weight that would be minimum for a given conversion. This would be an ideal packed

reactor. It is therefore interesting to determine this minimum bed length to offer a comparison

with different adiabatic process strategies. The maximum reaction rate curve is determined as a

function of conversion and temperature using the below expression.

For a given conversion, the temperature is changed until the maximum rate is obtained.
2.4 Experimental results

Given below is the experimental data obtained from the literature, using which we

analysed different kinetic rate models and observed their applicability.

T[°C] T[K] P[atm] %SO2 %O2 %X RSO2x105 [molso2/kg cat-s]

403.4 676.55 10 10.92 10.02 99.21 2.083

403.6 676.75 10 10.93 10.06 98.95 3.435

403.7 676.85 10 10.9 10.01 98.53 5.658

403.5 676.65 10 10.97 10.04 97.68 10.32

403.6 676.75 10 11.02 10.05 96.81 14.69

395 668.15 2.5 10.92 10.02 98.51 1.844

395 668.15 2.5 10.91 10.05 98.32 2.936

394.4 667.55 2.5 11.01 10.08 97.19 5.171

365.5 638.65 10 10.86 10.11 99.67 0.205

365.8 638.95 10 10.93 10.06 99.66 0.406

365.3 638.45 10 11.15 10.12 99.52 0.834

365.6 638.75 10 10.92 10.12 99.05 1.519

365.4 638.55 10 11.14 10.04 97.74 3.593

369 642.15 2.5 10.83 10.05 99.14 0.209

369.3 642.45 2.5 10.78 10.03 99.26 0.44

369.8 642.95 2.5 10.83 10.05 98.32 0.798

395 668.15 10 13.08 14.07 99.7 0.737

395.6 668.75 10 12.78 14.09 99.48 1.421

395 668.15 10 13.24 14.05 99.34 2.548

394.3 667.45 10 13.23 14.06 98.74 5.166

394.6 667.75 2.5 13.35 14.03 99.06 1.299

393.7 666.85 2.5 13.18 14.03 98.84 2.213

396 669.15 2.5 12.9 13.94 98.67 2.444

395 668.15 2.5 13.34 13.94 97.89 5.123

395.4 668.55 2.5 13.21 13.92 96.93 7.716

396.7 669.85 2.5 13.58 13.83 96.12 9.556

2.5 Results and discussion

The motive to do this experiment was to identify the kinetic model that best represents the

experimental data for the SO2 oxidation to SO3. The residual sum of squares (RSS) of the

differences between the predicted and experimental data was calculated for each kinetic model to

evaluate which kinetic model appears to be more reliable to represent the experimental data in

the range of temperature pertaining to the industrial SO2 reactor.

2.5.1 Experimental versus predicted reaction rate:

Kinetic data for the sulphur dioxide oxidation, obtained from an experimental design conducted

by Doering et al.13 was used in this investigation to assess the ability of each kinetic model to

represent the experimental data. In the experimental design of Doering et al.13, four to six

conversion data between 90% and equilibrium curve were used to measure the rate of reaction

for each gas feed composition, pressure and temperature.

The comparison of the data at 400˚C and for a feed composition of 11 mol% SO2, 10 mol% O2

and a pressure of 10 and 2.5 atm as well for a gas feed composition of 10 mol% SO2, 11 mol%

O2 and 1.12 atm are presented in Figure 2-6 Two important observations can be made from the

results of Figure 2-6. The Eklund kinetic model better represents the experimental data at a

pressure near 10 atm (Figure 2-6a). However, the Collina et al. kinetic model predicts with

relatively good accuracy the experimental data at the two lower pressures, including near

atmospheric pressure. Similar results were obtained at other experimental conditions. We can
conclude that there is no unique reaction kinetic model that represents the kinetic data over the

entire range of the experiment..

2.5.2 DATA TREATMENT AND ANALYSIS

● To better compare the different reaction kinetic models for the complete experimental

data set in the databank, the residual sum of squares of the differences between the

predicted and experimental data was calculated

● It is used to assess the predictive performance of each kinetic model.

Where,

➢ re,SO2i - The experimental rate of i-th experimental point

➢ rSO2i - The calculated rate of the i-th experimental point according to the reaction

mechanism model.

➢ N - Total experimental points considered.

 ● The residual sum squares of each kinetic model for different operating conditions are

presented in Figure

Figure - Residual sum squares (RSS) of each kinetic model for feed gas compositions of 10

mol% SO2, 11 mol% O2 for

(a) - for different temperatures at a constant pressure of 2.5 atm

(b) - different pressures at the constant temperature (400˚C)

Analysis & Insights: -

● When the temperature is increased, the RSS decreases significantly for all kinetic models,

which indicates that the models are more accurate if the reaction occurs above 400˚C.

● The models of Collina et al. and Eklund have the lowest RSS values compared to other

models.

● The model proposed by Collina et al. has the lowest RSS values at a low operating

pressures

● When the pressure increases, the RSS values increase. In contrast, the models of Eklund

and Calderbank show better predictions at higher pressures.

● No models exist that can be used for all experimental conditions.

● For instance, the model of Collina et al. better represents the experimental data in a

pressure range of 1 to 4 atm

● By considering the whole experimental pressure range (1 to 10 atm) and the temperature

range investigated, the Collina et al. kinetic model has a lower average RSS value and, in

general, shows a better fit considering all the experimental data in comparison with other

kinetic models.
Different Model Equations and RSS values: -

● A total of 135 experimental data points were used for different temperatures, pressures

and gas compositions, which can be included in the RSS function to find the best

equation that represents the entire experimental results.

Table gives the average RSS value of each kinetic model.

● The kinetic model proposed by Collina et al. has the lowest average RSS value (0.756).

● Using this criterion, this model better represents the entire experimental data than the

other models.

● To Perform a multi-objective optimization (MOO) for SO2 oxidation to SO3 in the plug

flow reactor, it is recommended to use the model of Collina et al., especially for pressure

near atmospheric and for the feed gas composition of 10 mol% SO2, 11 mol% O2.
CONCLUSION:

● This work presented the evaluation of four kinetic models for the SO2 oxidation to SO3

using vanadium pentoxide catalysts.

● The oxidation of SO2 to SO3 is an equilibrium reaction and the oxidation efficiency

increases with increasing the concentration of sulphur dioxide and oxygen in the feed gas.

● Moreover, in the higher pressures, the equilibrium conversion will be enhanced.

● By considering the residual sum of squares to compare different kinetic models using a

set of experimental data, it was found that the applicability of the models depends on the

operating conditions.

● A kinetic equation has been selected among the proposed ones to explain experimental

data in a narrow range of reaction conditions.

● Even though the four kinetic models propose different reaction mechanisms and were

obtained under different operating conditions, the information obtained in this

investigation is still very useful as a guide to decide an experimental plan to obtain the

kinetics of the proposed system.

● In this investigation, the Collina et al. kinetic model was chosen because it was more

comprehensive and detailed as it covers a wide range of reaction conditions.

● The determination of the residual sum of squares of all kinetic models allowed us to

evaluate the range of validity of the operating conditions for a specific kinetic model and

to guide in the selection of an appropriate kinetic model.

● Results showed that the equation was proposed by Collina et al. best represents the

experimental results in the range of operating conditions.

● Finally, this model could be selected to conduct a multi-objective optimization for the

catalytic oxidation reactor of the sulphur dioxide to the sulphur trioxide in the plug flow

reactor (PFR) for 10 mol% SO2, 11 mol% O2 and at the pressure near atmospheric due to

the lowest error under this operating conditions and to ensure the expected optimal region

within the real domain is enclosed.

 CONTRIBUTIONS BY TEAM MEMBERS

Amlan Kumar Tripathy

1. Introduction
2. Mechanism and Rate Laws
3. Maximum Rate Curve

Abdullah
1. Experimental Curve
2. Results and Discussions
3. Experimental versus Predicted reaction rate

Aditi Nair
1. Data Treatment and Analysis
2. Different Model Equations and RSS Values
3. Conclusion