selective oxidation into elemental sulfur in aqueous

8th International Conference on Environmental Science and Technology Lemnos island, Greece, 8 – 10 September 2003

HYDROGEN SULFIDE SEPARATION FROM GAS STREAMS: SELECTIVE OXIDATION INTO ELEMENTAL SULFUR IN AQUEOUS SOLUTIONS OF IRON CHELATES. O.G. TSAPEKIS, C.E. SALMAS and G.P.ANDROUTSOPOULOS School of Chemical Engineering, Chemical Process Engineering Laboratory, National Technical University of Athens, 9 Heroon Polytechniou Street, GR 15780, Athens Greece. EXTENDED ABSTRACT The present work concerns the investigation of the kinetics of H2S selective oxidation into sulfur, in a wetted wall batch recycle reactor, by gas absorption with chemical reaction in Fe3+(ΝΤΑ) aqueous solutions. The conversion of Fe3+ to Fe2+ was determined over the pH range ca. 3–6 and temperature range ca. 30–60οC. The penetration theory was employed for the evaluation of intrinsic specific reaction rates and enhancement factors. Activation energy values determined from the pertinent Arrhenius plots fall in the range, Εa=17.2– 22.8 kcal/mol. These values compare satisfactorily with that of Εa=24 kcal/mol obtained from a similar kinetic study performed in a bubble column reactor and indicates a chemical reaction control of the overall gas absorption phenomenon. Enhancement factors varied as follows: Ε =2.7–7.5 (pH=3), E=3.2–14.1 (pH=4), E=5.9–17.0 (pH=5), και Ε=6.7–20.1 (pH=6) indicating a substantial increase of the mass transfer coefficient due to chemical reaction.

888

1. INTRODUCTION Hydrogen sulfide is a regular component of gaseous streams of fuels conversion and upgrading processes i.e. liquid fuels hydrodesulphurization, coal gasification, etc. It is also met in biogas, natural gas and geothermal fluids. The separation of H2S from these streams is necessary prior to its conversion into sulfur or other useful products due to his odor, toxicity and corrosive properties. Removal of H2S from the said streams is also required before using them either as fuels or as reactants for the synthesis of chemicals. Historically, various methods for hydrogen sulfide separation from gaseous mixtures have been developed. Αn important family of methods encompasses those based οn the selective oxidation of H2S into elemental sulphur ίn aqueous solutions of iron chelates or other oxidative agents, (i.e. redox methods). Α brief review related to the development of new methods for sulphur recovery from gas streams is given in [1],[2]. The most important advantages of the redox methods are (i) the mild reaction temperatures (e.g. 20-60 °C), (ii) H2S conversion is practically complete (99.99%) and occurs ίn a single step, (iii) the use of air as oxidant for the regeneration of the oxidative solution. A possible drawback of these methods is the increased cost of chemicals, especially when high consumption of reactants is required due to low selectivity or decomposition of the chelating agent. In any case the choice of the desulfurization method, for a specific application, depends on the technical features and process economics. 2. LITERATURE SURVEY At least two industrially applied processes for the selective oxidation of H2S into elemental sulphur using solutions of Fe3+ are cited ίn the Iiterature. The first bears the commercial name Lo-cat (Αir Resources Inc. [3]), takes place ίn aqueous solutions of a Fe3+ complex with a proprietary agent i.e. ARI-30 and the regeneration is accomplished by blowing air through the solution. The operating conditions of the Lo-cat method are as follows: pΗ=8.0-8.5, temperature 4-50 oC, low pressure (ί.e atmospheric ~0.5 MPa) is recommended for the processing of streams containing a H2S percentage higher than 5%, iron concentration 500-2000 ppm, production capacity 1-5 Ith Id (1Ith=1016 kg). The second commercial process for the oxidation of H2S ίn aqueous solutions of iron chelates is named SulFerox [4], operates economically ίn stable solutions of an iron chelate of 4% concentration, pressures υp to 1.5-2.0 MPa and high temperatures, that are not noted by number. A commercial application of this method is the desulphurization of natural gas [3], [5]. The majority of the research results referred to the oxidation of H2S with Fe3+(NTA) are not published in the open literature. Because of the special significance, the chemistry of the complex compounds involved is patented. Regarding the use of the ΝΤΑ (NitriloTriacetic Acid) in the oxidation of H2S into elemental sulphur, Newmann and Lynn [6] studied absorption with chemical reaction of H2S in an aqueous solution of Fe3+(NTA) using a continuous flow, wetted wall column operated at steady state. Absorption runs were carried out ίn a narrow temperature (ί.e. 60-65 oC) and pΗ (i.e. 3.5-4.5) range. Ιn the latter research work it was concluded that liquid phase mass transfer controls the overall process which can be characterized as "absorption with instantaneous chemical reaction". Demmink et al. [7] carried out a similar investigation using a cocurrent down flow column packed reactor, equipped with static mixers (type SMV-4), which constituted the main element of the experimental setup. The latter was operated in semi-batch mode, i.e. continuous with respect to the gas phase and batch recycle with respect to the liquid phase. It was reported in [7] that the reaction of H2S with Fe3+(NTA) solution proceeds at

889

a finite rate and is first order with respect to each one of the reactants separately a result contrasting the findings reported in [6]. However, it should be noted that ίn the latter investigation oxidation was performed at 13oC and pΗ=6.7–8.2. These conditions are quite close with those of industrial methods. Hydrogen sulphide oxidation results ίn Fe3+(NTA) aqueous solutions using a bubble column laboratory reactor, was reported ίn [8]. Absorption runs were carried out in the temperature range Τ=36–63oC and pΗ=3.6– 6.5 and the application of the two films theory confirmed a chemical reaction control of the gas absorption phenomenon. Α first order kinetic with respect to H2S concentration and zero order with respect to Fe3+(NTA) were deduced and an activation energy Ea=24 kcalΙmol, was obtained. Ιn the present work the results are reported, of a kinetic study for the selective oxidation of H2S ίn aqueous Fe3+(NTA) using a wetted wall column reactor under temperature, pΗ of solution and iron chelate concentration comparable to those applied ίn the study reported in [8]. Particular care was given for pH stabilization using appropriate buffering of acetic acid-sodium acetate solutions during a specified H2S oxidation procedure. The scope of the present work is to confirm the presence of a finite reaction kinetic rate and to figure out the effect of temperature, concentration and pH upon the reaction rate constant. The wetted wall column reactor ensures a reliable value for the specific interfacial area. The penetration theory is being used for interpreting the experimental data. The investigation of the intrinsic phenomena controlling reaction kinetics is very important since such information allow for the development of an improved design of the pertinent industrial processes. It is of particularly significance for the choice of the process specifications i.e. composition and concentration of the chelating agent, operating conditions (temperature, pressure, pH, stirring, etc) and finally the development of the overall process flow sheet. 3. THEORY 3.1

Penetration Model for a Wetted Wall Reactor

The time of contact of a differential liquid volume that wets the reactor wall is equal to the reactor height divided by the surface velocity of the falling film because of the prevailing plug flow conditions, [4],[6],[7],[9],[10]. Thus,

τ =

where:

3 us = 2

h

(1)

u s

 gρ i   3µ

  

1

3

   

Q

i

πd c

   

2

3

(2)

τ= exposure time, h = column height, us = surface velocity, g = acceleration of gravity, ρl = density of liquid, µ = viscosity of liquid. Q, = volumetric velocity of liquid. During the entire of gas-liquid contact time, hydrogen sulphide penetration occurs into the liquid bulk and the following chemical reaction,

2 Fe3+⋅NTA + H2S

Æ

2 Fe2+⋅NTA + 2Η+ + S2-

progresses as a function of liquid depth and contact time. For a first order reaction with respect to H2S concentration and zero with respect to Fe3+ the instantaneous mass balance for hydrogen sulphide reads:

890

2

D

i

∂ CH H2S

∂y

2S 2

=

∂C H

2S

∂τ

+ k CH i

(3)

2S

where D: diffusion coefficient of H2S ίn liquid, C: H2S concentration, y: distance from the interface and ki: reaction rate constant. Βy taking account of the appropriate boundary conditions (where i indicates the position on the interface of gas–liquid) the instantaneous rate of H2S absorption reduced with respect to the contact surface following a contact time period τ, is expressed by, [4],[10].

i N(τ) = CH

2S

(

DH Sk 2

i

)

0.5

- kτ   0.5 e  erf (ki t) + 2πk τ   i 

(4)

where: Ν molar rate of mass transfer reduced with respect to interface area. 3.2

Application of the Penetration Model

Given the kinetic order of the oxidation reaction it remains to determine the reaction constant k as a function of temperature Τ and pΗ. The Penetration Model was applied to determine the value of k for a typical H2S oxidation reaction run using a wetted wall batch recycle reactor. The integrated mass balance over the entire column height Η and consequently over the whole contact time of phases, i.e. τ=H/us, [4], is, −k Iτ   Cbot Ybot τ 0.5 0.5 e i l erf k τ +  dτ ∫ 2Q dC 3+ = − ∫ GdYH S == ∫ 2πrCH Sus DH SkI Fe 2 2 2 i i Ctop Ytop 0 2(πk τ)  

(

)



( )

i



(5) The first two members of equation (5) constitute the mass balance in the reactor for a specified moment of time. The mass balance involves known or given values of parameters. The determination of k is possible through a numerical integration of the third member of Eq. (5) and the application of an error minimization numerical procedure (e.g. Marquard-Levenberg). It is noted that integration requires a correlation between Y and Ci for H2S and is provided by the mass transfer rate equation (ί.e. Eq. (6)). The direction of transport is from the gas bulk towards the interface combined with the Henry distribution law. Thus,

(

N = k g P yH i

P yH

2S

2S

i

− yH

= He

)

2S i CH S 2

(6)

(7)

where kg: H2S mass transfer coefficient ίn the gas phase, Ρ: total pressure, He: Henry distribution coefficient and Υ=Y/(1+Υ): molar fraction in the gas phase. 4.

EXPERIMENTAL

The flow sheet of the experimental set υρ employed for the separation of H2S from a gas mixture of H2S+N2 is shown ίn (fig. 1). Details of the mechanical parts, of the measuring and control instrumentation and the operating conditions are reported in [11], together with an outline of the construction details and rheological design. The experimental procedure included the continuous feeding of the gas mixture (i.e. H2S+N2) counter currently with respect to the falling film of the re-circulated liquid (ί.e. aqueous solution of Fe3+ΝΤΑ) through the vertically oriented column. The concentration of Fe3+ ίn the liquid collection tank was being determined at regular time intervals. Systematic recording of the primary experimental data namely temperature, physical properties of materials,

891

phase equilibrium, reaction stoichiometry, flow rate, rheological and mass transfer, are presented in [11], [12]. Isothermal absorption runs were performed at controlled temperature and pH levels i.e. Τ= 30, 40, 50, 60oC and pΗ=3, 4, 5, 6.

TMA

V-3

PR1 PR-2

PI-1

CV3

CV1 αισθητήρας H2S

PI-2 TIC-1

N2

TIC-0

H2S

TIC-2 PCLD

Η/Υ

Figure 1. Flow Sheet of the Wetted Wall Reactor Experimental Setup.

(R-1) Gas/Liquid Falling Film Reactor, (C-1, C-2) Ν2 and H2S Pressure Cylinders, (ΡI-1, ΡI-2) Reactor Inlet – Outlet Mercury Manometers, (FM-1,FM-3) Ν2 Rotameter, (FM-2) H2S Soap Film Flow meter, (FM-4, FM-5) Rotameters for the outlet stream of the reactor (CV1, CV-2, CV-3) Control Valves of Reactor Feed Streams, (CV-4) Control Valve of the gas Scrubber Feed Stream, (Ρ-1) Liquid Feed Peristaltic Pump, (V-1) Liquid Flow Pulse Dumping Vessel, (V-2) Thermoregulated Liquid Collection Vessel, (V-3) Liquid trap in the reactor effluent gas stream, (S-1) H2S gas Scrubber (i.e. 20% NaOH solution), (S-2) Auxiliary H2S Gas Scrubber (i.e. CdSO4 solution), (TIR) Reactor Inlet-Outlet Temperature Indicator/Recorder, (TIC) Temperature Indicator/Controller of Liquid Collection Vessel, Hydrogen Sulphide Indicator TOX-PEM, (ΤΜΑ) Magnetic stirrer–heater. 5. RESULTS AND DISCUSSION The primary results of Fe3+ conversion with respect to time ίn the liquid tank are depicted ίn (fig. 2). It is readily seen from this figure that the drop of Fe3+ concentration with time is linear. The penetration theory was applied for evaluating the specific reaction rate constant, k and the corresponding enhancement factors Ε (Table 1) for all runs (i.e. sixteen pΗ and temperature combinatorial cases). The enhancement factors of (tab.1) are indicative of the acceleration of the absorption process, particularly at higher

892

temperatures, due to the occurrence of the chemical reaction. Arrhenius plots for each particular pH value are depicted in fig. 3. Activation energy values were derived from a linear fit. From the unified Arrhenius diagram of (fig. 4a) an average activation energy Εa=19.083 kcal/mole was deduced. The latter value indicates that the overall phenomenon is chemical reaction controlled. As illustrated in fig. 4b, a linear dependence of the frequency factor of the specific kinetic constant upon pH, was confirmed. The kinetic equation (eq.8) for H2S selective oxidation into S, in aqueous solutions of Fe3+ (NTA) was validated. It should be noted that the specific interfacial area of the reaction system was about a=4m2/m3 (liquid).

R

H S 2

= [ − 10.62 + 4.15 ( pH ) ]10

14

  

exp  − 19083

 C R T H S g 2 

(8)

Table1. Enhancement factor for the oxidation of H2S with Fe3+(ΝΤΑ)

pH

3

4

5

6

2.71 3.24 4.99 6.67

3.62 4.75 5.87 7.99

7.34 9.70 10.52 13.47

7.55 14.13 17.07 20.95

ο

T C 30 40 50 60

110

100

100

90

80

80

3

Fe (mole/m )

70 60

3+

3+

3

Fe (mole/m )

90

50 40

70 60 50 40

30 0

500

1000

1500

2000

2500

30

3000

0

500

1000

t (s)

2000

2500

3000

3

80

3

80

60

3+

60

Fe (mole/m )

100

3+

t (s)

100

Fe (mole/m )

1500

40

40 0

500

1000

1500

2000

2500

3000

t (s)

0

500

1000

1500

2000

t (s)

3+

Figure 2. Fe Concentration versus Reaction Time Lines. Isothermal Experimental Runs at Constant pH as follows: (■) 30, (●) 40, (▲) 50, (♦) 600C.

893

ln(k i),

[k i] = (1/s)

7 6 5 4 3 2 1 2.95

3.00

3.05

3.10

3.15

1/T * 10

◆ ○ ▲ ■

3.25

3.30

3.35

(1/K)

pH=3 pH=4 pH=5 pH=6

ln(k)=-9018.5*(1/T) + 31.371 ln(k)=-11488 *(1/T) + 39.866 ln(k)=-9396.4*(1/T) + 33.770 ln(k)=-8661.3*(1/T) + 31.995

6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5

E=17920 E=22827 E=18670 E=17210

cal/mol cal/mol cal/mol cal/mol

16

pH = 3 pH = 4 pH = 5 pH = 6

14

kmno= [-10,62+4,15*pH]*10

14

-14

1/s

12

kmno*10

-1

3.20

Arrhenius diagrams for the oxidation of H2S in aqueous solutions of Fe3+ (NTA) at the specified pH.

Figure 3.

lnk i [k i]=s

3

10 8 6 4 2

3.00

a Figure 4:

3.05

3.10

3.15

3.20

3.25

3.30

3.0

b

3

1/T*10 (1/K)

3.5

4.0

4.5

5.0

5.5

6.0

pH

(a) Unified Arrhenius diagrams for Ea=19.083 Kcal/mol (mean value) for the Oxidation of H2S in aqueous solutions of Fe3+ (NTA) at the specified pH, (b) Dependence of the frequency factor of the specific kinetic constant upon pH.

6. CONCLUSIONS The oxidation reaction of hydrogen sulphide in aqueous solutions of Fe3+ occurs at a finite rate with kinetics of zero order with respect to Fe3+ and first order with respect to H2S concentrations. For the pΗ and temperature ranges investigated the overall absorption phenomenon is chemical reaction controlled with an activation energy ca. Εa=19 kcal/mole. This value approach the corresponding value Εa=24 kcal/mole obtained from a bubble column reactor, [8]. However, ίn the present case the enhancement factor is certainly higher. The results obtained ίn the present study are superior in comparison with relevant literature data since they apply to wide pΗ and temperature ranges.

894

7. LITERATURE 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11.

12.

Caruana, C. M., (1996) “Processes Aim to Improve Sulfur Recovery from Gases”, Chem. Eng. Progress, 92, 1– 17. Hardison, L. C. (1985) “Go from H2S to S in one Unit”, Hydrocarbon Processing, 4. 70 – 71. Bedell, S.A., Kirby, L.H., Buenger, C.W., McGauge, M.C. Hydrocarbon Processing (1988) 1, 63 – 66. Sherwood, T.K., Pigford, R.L., Wilke, C.R., (1975) "Mass Transfer", Mc Graw Hill, New York. Anonymous, (1996). “Steamflood Expansion Includes Liquid Redox Process”, ΟίΙ and Gas Journal, Jan. 29, 99. Neumann, D. W. and Scott Lynn, S., (1984) “Oxidative Absorption of H2S and O2 by Iron Chelate Solutions”, AIChE Journal, 30:1 62-69. Demmink, J. F., Wubs, H. J., Beenackers, A. A. C. M., (1994) “Oxidative Absorption of Hydrogen Sulfide by a Solution of Ferric Nitilotriacetic Acid Complex in a Cocurrent Down Flow Column Packed with SMV – 4 Static Mixers”, Ind. Eng. Chem. Res., 33, 2989 – 2995. Pitsinis, Ν., Androutsopoulos, G. (1997) 1st Hellenic Conference of Chemical Engineer, University of Patra, 383 – 388. Danckwerts, P. V. (1970) “Gas – Liquid Reactions” McGraw – Hill 34, 193, 113. Astarita, G., (1966) "Mass Transfer with Chemical Reaction", Elsevier, Amsterdam. Tsapekis, Ο.G. (2002)“ Gas Absorption from Liquid with Chemical Reaction: Kinetics of Selective Oxidation of Hydrogen Sulphide by a Solution of Ferric Nitrilotriacetic Acid in a Wetted Wall Column”. Graduation Dissertation, NTUA, School of Chemical Engineering, Chemical Process Engineering Laboratory (2002). Androutsopoulos, G., Hatzilyberis, Κ., Pitsinis, Ν., Salmas, Κ. Inve-stigation of Design Parameter for Greek Lignite Allothermal Gasification and H2S Separation Processes. Final Report, Project No 8875/28.8.96,PENED, Ministry of Development of Greece,

895