Publicaciones

Absorción de dióxido de carbono en disolucionesde dietanolamina en isopropanol

Ventura Perera, J.R.; Díaz García, C.; Limiñana de la Fé, G.; Álvarez Díaz, M.Dpto. de Ingeniería Química y T.F. Facultad de Química. Universidad de La Laguna.

Avda. Astrofísico Francisco Sánchez, s/n. 38200 La Laguna.

Carbon dioxide absorption into diethanolamine in isopropanol solutions.

Absorció de CO2 en dietanolamina en dissolucions d’isopropanol.

Recibido: 18-X-2001

AFINIDADREVISTA DE QUÍMICA TEÓRICA Y APLICADA

EDITADA POR LA ASOCIACIÓN DE QUÍMICOS E INGENIEROS

DEL INSTITUTO QUÍMICO DE SARRIÁ

Afinidad (2002), 59 (502), 641-647

Página 104

AFINIDAD LIX, 502, Noviembre-Diciembre 2002 641

Absorción de dióxido de carbono en disolucionesde dietanolamina en isopropanol

Ventura Perera, J.R.; Díaz García, C.; Limiñana de la Fé, G.; Álvarez Díaz, M.Dpto. de Ingeniería Química y T.F. Facultad de Química. Universidad de La Laguna.

Avda. Astrofísico Francisco Sánchez, s/n. 38200 La Laguna.

Carbon dioxide absorption into diethanolamine in isopropanol solutions.

Absorció de CO2 en dietanolamina en dissolucions d’isopropanol.

Recibido: 18-X-2001

RESUMEN

La cinética de la reacción entre el CO2 y la dietanolami-na (DEA) en isopropanol ha sido estudiada en una colum-na de laboratorio de paredes humedecidas a 298 K. Elorden de reacción respecto a la DEA, n, y el coeficientecinético aparente, k, han sido determinados bajo condi-ciones experimentales que aseguraban un régimen dereacción de pseudo 1er orden. Los valores obtenidos fue-ron

n = 1,7 y k = 76 (m3/kmol)1,7 s–1

La reacción puede ser descrita según el mecanismo dezwitterion, originalmente propuesto por Caplow(9) y desa-rrollado por Danckwerts(12), con valores de los coeficien-tes cinéticos de las etapas elementales,

k1 = 208,9 m3/(kmol.s) k2/k3 = 1,71 kmol/m3

Palabras clave: Absorción. Dióxido de carbono.Dietanolamina. Orden de reacción. Coeficiente cinéticoaparente.

SUMMARY

The kinetic of the reaction of carbon dioxide with diet-hanolamine (DEA) in isopropanol solutions was studiedat 298 K using a laboratory wetted-wall contactor. Thereaction order respect to the DEA, n, and the apparentkinetic coefficient, k, were determinate by experimentalconditions that assured pseudo-1er order regimen. Foundvalues were:

n = 1,7 and k = 76 (m3/kmol)1,7 s–1

The reaction was described with the zwitterion-mecha-nism originally proposed by Caplow(9) and developed byDanckwerts(12) with reaction rate constants of mechanismgiven by 208,9 m3/(kmol.s) for k1 and 1,71 kmol/m3 fork2/k3.

Key words: Absorption. Carbon dioxide. Diethanolamine.Order of reaction. Apparent kinetic constant.

RESUM

La cinètica de la reacció entre el CO2 i la dietanolamina(DEA) en dissolucions d'isopropanol ha estat estudiadaen una columna de laboratori de parets humides a 298K. L'ordre de reacció respecte a la DEA, n, i el coeficientcinètic aparent, k, se han determinat a partir d'un con-junt d'experiments en condicions d'operació que asse-guraven un régim de reacció de pseudo 1er ordre, i es vanobtenir els següents valors de:

n = 1,7 k = 76 (m3/kmol)1,7 s–1

La reacció pot ser descrita segons el mecanisme de zwit-terion, originalment proposat per Cplow(9) i desenvolu-pado per Danckwerts(12), amb els valors dels coeficientscinètics aparents de les etapes elementals,

k1 = 208,9 m3/(kmol.s) k2/k3 = 1,71 kmol/m3

Mots claus: Absorció. Diòxid de carboni. Dietanolamina.Ordre de reacció. Cooeficients cinètics.

INTRODUCCIÓN

La separación de gases ácidos, como el dióxido de car-bono y el sulfuro de hidrógeno, de corrientes gaseosasmediante absorción con reacción química, es un procesode interés industrial. Los disolventes generalmente emple-ados son las disoluciones tanto acuosas como orgánicasde alcanolaminas; entre estas cabe destacar la monoeta-nolamina (MEA) y dietanolamina (DEA). El estudio cinéti-co de las reacciones entre el CO2 y las alcanolaminas hasido llevado a cabo por diferentes autores, destacando eltrabajo de recopilación realizado por Benítez García(3, 4).Los investigadores que han trabajado en disoluciones acuo-sas, admiten que el orden de reacción respecto a la MEAes uno,(11, 16, 22) mientras que para la DEA existen discrepan-cias en los valores del orden de reacción.(6, 10, 23, 25) En mediosno acuosos, tanto para la MEA como para la DEA seencuentran valores comprendidos entre 1 y 2.(2, 24, 27)

En este trabajo se estudia la absorción de CO2 (A) por diso-luciones de dietanolamina (B) en isopropanol a 298 K. Si

Página 105

se considera la reacción irreversible y rápida (CAo = 0), conun coeficiente cinético aparente km,n.

la ecuación cinética vendrá dada por:

(1)

A su vez la densidad de flujo de transferencia podrá expre-sarse según:

(2)

donde E es el factor de aceleración de la transferencia demateria debido a la reacción química. Las expresionesanalíticas para E han sido obtenidas por diferentes auto-res(14, 7), en función del módulo de Hatta, Ha,

(3)

y del factor de aceleración instantánea, Ei, definido segúnla teoría de la penetración como:

(4)

La mayor parte de los investigadores aceptan para la absor-ción de CO2 por etanolaminas, tanto primarias como secun-darias y en medios acuosos u orgánicos, el mecanismopropuesto por Danckwerts(12) a partir de las ideas deCaplow(9), dado por:

(5)

(6)

que conduce a la ecuación cinética:

(7)

donde en aminas primarias R2 será H, como en el caso dela MEA; en aminas secundarias R1 y R2 pueden ser igualeso diferentes, en el caso de la DEA R1= R2 = -CH2-CH2OH.Esta ecuación permite explicar los órdenes de reacciónuno, dos o intermedios, en función de los valores de k1, k2

y k3, que dependen de la basicidad de la amina, el efectoestérico y la polaridad del disolvente.En el caso de que los experimentos se encuentren en laregión de reacción de pseudo m-n orden, definida por:

(8)

se cumple que:

E = Ha (9)

Por tanto, combinando las ecuaciones (9), (2) y (3), se obtie-ne:

(10)N

P

kH

mD k C C

AA

GA m n Ai

mBon

=+

+−

1

21

1,

22

< <<HaEi

rCO R R NH

kk

k k R R NH

=[ ][ ]

+ [ ]

2 1 2

1

2

1 3 1 2

1

CO R R NH R R N HCOOk2 1 2 1 2

1+ ← →⎯ + −

EDD

DD

CzCi

A

B

B

A

Bo

Ai= +

Hak m

D k C CL

A m n Aim

Bon=

+−1 2

11

,

N k p p Ek CA G Ao Ai L Ai= ( ) =–

– ,r k C CA m n Am

Bn=

A zB PK m n+ ⎯ →⎯⎯⎯,

Los autores que han trabajado con sistemas CO2-etano-laminas, tanto en medio acuoso como alcohólico, han obte-nido para el orden de reacción respecto al CO2 el valor deuno(1,2,6,8,15,26). En este caso, la ecuación (10) se simplifica a:

(11)

La interpretación de los datos de absorción con reacciónquímica requiere pues, del conocimiento de las propieda-des físicas así como de los coeficientes de transferenciade materia, los cuales se relacionan con variables hidro-dinámicas, propiedades de las fases y características geo-métricas de los contactores.Por tanto los experimentos que se realicen bajo estas con-diciones de pseudo primer orden, permitirán calcular elorden de reacción respecto a la amina, n, y el coeficientecinético aparente, k1,n. Además si se compara la ecuacióncinética así obtenida con la resultante de la aplicación delmecanismo propuesto por Danckwerts, ec. (7), se podráncalcular los coeficientes cinéticos de las etapas elemen-tales ó relaciones entre ellos.

MATERIAL Y MÉTODO

Los experimentos se realizaron en una columna de pare-des humedecidas de 1,995.10–2 m de diámetro y 0,10 m dealtura. La instalación utilizada está descrita en trabajosprevios(13, 17) y se muestra en la Figura 1.

– El análisis de la fase gaseosa se llevó a cabo mediantecromatografía de gases usando un detector de conduc-tividad térmica y una columna de Carbosiever-S mesh100/200, de 0,55 m de longitud y 1/8" de diámetro. Comogas portador se empleó helio N-55, y las condiciones detrabajo seleccionadas que permitieron una buena sepa-ración de los componentes de la fase gaseosa fueron:

Temperatura de inyección 423 KTemperatura de la columna: 398 KTemperatura de la interfase: 448 KTemperatura del detector: 473 KCorriente del puente: 0,15 ACaudal de gas transportador: 5,5.10–7 m3 . s–1.

– Propiedades físicas

Las medidas experimentales de las densidades y viscosi-dades se realizaron usando densímetros de inmersión yviscosímetros Cannon-Fenske sumergidos en un baño ter-mostático. Las ecuaciones para su cálculo a 298 K son:

ρ = 785,8 + 31,4 CB (12)

lnμ = –6,021 + 0,480 CB (13)

en las que la densidad viene dada en kg m–3 y la viscosi-dad en kg m–1 s–1, para concentraciones de amina inferio-res a 2,610 kmol m–3.Para la solubilidad del CO2 en disoluciones de DEA-iso-propanol se tomó la del CO2 en isopropanol, ya que la mayorparte de los investigadores que han trabajado con MEA yDEA en medios tanto acuosos como no acuosos(1, 2, 5, 10, 18, 23)

consideran que la solubilidad es independiente de la con-centración de la amina hasta valores de ésta de aproxi-madamente 2,0 kmol m–3, y que si hubiera variación, éstano excedería del 10%, y por lo tanto podría considerarseconstante.

NP

kH

D k C

AA

GA n Bo

n

=+1

1,

642 AFINIDAD LIX, 502, Noviembre-Diciembre 2002

R R N HCOO R R NH R R NHCOO R R N Hk1 2 1 2 1 2 1 2 2

3+ − − ++ ⎯ →⎯ +

k2

Página 106

Para la solubilidad del CO2 en isopropanol se utilizó la ecua-ción obtenida experimentalmente a presiones parcialesinferiores a 1 atm y en el intervalo de temperaturas entre298 y 313 K:

(14)

donde el coeficiente de la ley de Henry, H, se expresa enatm m3 kmol–1 si la temperatura viene en K.La difusividad del CO2 en disoluciones de DEA en isopro-panol, DA, a 298 K, se estimó a partir de la ecuación:

DA = 3,6510–12μ–1,104 (15)

Esta ecuación se obtuvo de un estudio bibliográfico reali-zado a partir de trabajos que han utilizado CO2 con MEAo DEA en disoluciones alcohólicas(1, 19, 24, 27).Para la estimación de la difusividad de la DEA en disolu-ciones de isopropanol, DB, se utilizó la ecuación:

DB = 1,02310–5μ–0,943 (16)

log,

,1 606 0

3 158H T

= −

obtenida de la aplicación de la ecuación de Wilke y Chang,donde el volumen molar normal de la DEA se determinóutilizando el método Le Bas(21) (VA = 0,1267 m3 kmol–1) y elparámetro de asociación para el isopropanol siguiendo lasideas de Partington(20) (φ = 1,3). En estas ecuaciones DA yDB se expresan en m2 s–1 y μ en kg m–1 s–1.El coeficiente de transferencia de materia referido a la faselíquida, kL, se determinó a partir de la ecuación

Sh = 1,392103 Re0,26 (17)

obtenida mediante experimentos de absorción física deCO2 en isopropanol, a 298 K y variando el caudal de la faselíquida entre 6,1 10–6 y 12,1 10–6 m3 s–1.

RESULTADOS Y DISCUSIÓN

Con el fin de determinar el orden de reacción respecto ala DEA, n, y el coeficiente cinético aparente, k1,n, se reali-zaron cuatro series de experimentos a 298 K con concen-


Figura 1. Instalación experimen-tal. (1,2) Medidores de caudal.(3,4) Medidores de temperatura.(5) Mezclador. (6) Manómetro deHg de rama abierta. (7) Baño determostatización. (8) Saturador.(9) Columna. (10) Controlador depresión. (11) Depósito de faselíquida. (12) Cambiador de calor.(13) Manómetro diferencial. (14)Controlador de temperatura. (15)Calderín.

Página 107

traciones de DEA en isopropanol de 0,40, 0,60 1,00 y 1,50kmol m–3. En cada una de ellas se varió el caudal de la faselíquida entre 0,80 10–6 y 3,10 10–6 m3 s–1. El rango de pre-siones parciales del CO2 a la entrada estuvo comprendidoentre 0,05 y 0,16 atm. El caudal de gas y la presión totalse mantuvieron constantes con valores de 1,5 10–5 m3 s–1

en c.n. y 0,947 atm respectivamente.La densidad de flujo experimental de transferencia, NA, secalculó mediante medidas cromatográficas de la concen-tración de CO2 a la entrada y salida del contactor, conoci-da el área de transferencia y realizando un sencillo balan-ce, tomando como base el caudal de gas inerte. Losresultados obtenidos se muestran en la Tabla I.

Si se supone que la ecuación que representa al sistema

k1,nCO2 + 2 DEA → Carbamato

es irreversible y de orden uno respecto al CO2 y "n" res-pecto a la DEA, la ecuación cinética correspondiente será:

–rA = kl,n CA CBn (18)

Admitiendo que la reacción es rápida, de pseudo 1-n ordeny que no hay resistencia en fase gaseosa, la ecuación (11)se simplifica:

(19)

lo que indica que la densidad de flujo es independiente delvalor del coeficiente de transferencia de materia referidoa la fase líquida. En la Tabla I se muestran asimismo losvalores de los cocientes NA/pA para cada serie, observán-dose que son prácticamente constantes para cada con-centración de amina, independientes del caudal de líqui-do, esto es de kL, lo que confirma, en primera aproximación,que el régimen de reacción puede considerarse como rápi-do, de pseudo 1,n orden y con resistencia de la fase gase-osa despreciable.Reordenando la ecuación (19) se obtiene:

(20)

de forma que si se representa el logaritmo de NA/(pA DA1/2 )

frente al de CBo, se debe obtener una recta de pendienten/2 y ordenada en el origen log (k1,n

1/2 /H).Los valores medios calculados de NA/(pA DA

1/2 ) se muestranen la Tabla II, cuyos logaritmos se representan, frente alos de la concentración de amina, en la Figura 2.Se observa que los puntos están bien alineados a una rec-ta, de cuya pendiente se puede obtener el orden de reac-ción respecto a la amina, n, y de la ordenada en el origen,el coeficiente cinético aparente de la reacción, k1,n. Losresultados son los siguientes:

n = 1,7

k = 76 (m3/kmol)1,7 s–1

Si se compara el valor del orden de reacción respecto a laamina se observa una disminución del valor respecto alobtenido por otros autores(2, 23), señalando que en algúncaso éstos han forzado el orden a números enteros.Con estos valores se calcularon, para todos los experi-mentos, los del módulo Hatta, que ahora toma la forma:

(21)Hak

D k CL

A Bo= 1117

17; ,

,

N

P D

k

HCA

A A

lnBon

1 2

1 22

/,/

/=

NPH

D k CAA

A n Bon= 1,


TABLA IResultados experimentales obtenidos de NA.

CBo L·106 pA·102 NA·106 (NA/pA)·105

kmol/m3 m3/s atm kmol/(s.m2) kmol/(atm·m2·s)

0,40 0,82 14,129 2,133 1,5082,00 14,099 2,208 1,5630,99 9,779 1,441 1,4721,97 9,728 1,492 1,5313,07 8,654 1,375 1,5861,00 7,801 1,205 1,5430,98 6,564 1,006 1,5312,03 6,551 1,036 1,5801,35 4,689 0,742 1,5811,98 4,688 0,730 1,555

0,60 2,50 12,499 2,561 2,0493,00 12,252 2,524 2,0602,00 10,260 2,135 2,0813,00 10,064 2,203 2,1892,50 7,938 1,649 2,0773,07 7,809 1,706 2,1841,00 5,533 1,131 2,045

1,00 2,00 12,162 3,543 2,9132,50 12,201 3,481 2,8543,00 12,041 3,636 3,0202,00 10,099 2,945 2,9172,50 10,043 2,874 2,8613,00 10,080 2,987 2,9631,00 8,460 2,431 2,8742,00 8,387 2,393 2,8541,00 5,526 1,594 2,8842,00 5,411 1,537 2,840

1,50 2,00 12,022 4,328 3,6002,50 11,870 4,411 3,7163,00 11,855 4,470 3,7712,00 9,721 3,531 3,6332,88 9,850 3,788 3,8461,00 8,617 3,063 3,5551,64 8,648 3,201 3,7023,05 8,887 3,275 3,6852,00 8,187 3,050 3,7252,00 5,426 1,931 3,5602,49 5,330 1,996 3,745

TABLA IIValores medios del cociente NA/pA para cada serie

experimental.

CBo (NA/pA) · 105 NA/(pA·DA1/2 )

Exps. kmol/m3 kmol/(atm·m2·s) kmol/(atm·m3·s1/2)

APA 0,40 1,545 0,294APB 0,60 2,098 0,420APC 1,00 2,898 0,646APD 1,50 3,685 0,937

Página 108


Figura 2. Efecto de la concen-tración de DEA sobre NA.

TABLA IIIValores calculados de E, Ha y Ei.

Exps. E Ha Ei

APA1 2,8 2,9 10,0APA2 2,3 2,3 10,1APA3 2,6 2,7 13,4APA4 2,3 2,3 13,4APA5 2,1 2,0 14,8APA6 2,7 2,7 16,1APA7 2,7 2,7 18,7APA8 2,3 2,3 18,7APA9 2,6 2,5 25,2APA10 2,3 2,3 25,2

APB1 3,4 3,5 15,3APB2 3,3 3,3 15,6APB3 3,7 3,7 18,2APB4 3,5 3,3 18,5APB5 3,5 3,5 22,7APB6 3,5 3,3 23,1APB7 4,4 4,4 31,5

APC1 7,6 7,7 24,8APC2 7,0 7,2 24,8APC3 7,1 6,9 25,0APC4 7,6 7,7 29,4APC5 7,1 7,2 29,5APC6 7,0 6,9 29,4APC7 9,0 9,2 34,6APC8 7,5 7,7 34,9APC9 9,0 9,2 51,7APC10 7,4 7,7 52,7

APD1 15,3 15,5 37,0APD2 14,9 14,7 37,5APD3 14,4 14,0 37,5APD4 15,4 15,5 45,2APD5 14,8 14,1 44,7APD6 14,6 14,5 50,5APD7 14,0 13,9 49,2APD8 18,1 18,6 50,7APD9 15,8 15,5 53,2APD10 15,1 15,5 79,1APD11 15,0 14,7 80,5

TABLA IIIValores calculados de E, Ha y Ei.

Página 109

Estos valores, junto con los del factor de aceleración, E, ylos de aceleración instantánea, Ei, calculados según (2) y(4) respectivamente, se muestran en la Tabla III. En ella se aprecia que, para todos los experimentos, losvalores de Ha son mayores de 2; además se verifica laigualdad E = Ha, con una desviación menor del 5%.Estos resultados permiten considerar que las hipótesissupuestas del régimen de reacción rápida y de pseudo-orden son válidas en el rango de operación de este traba-jo. Se puede observar además, que Ha < E i /2, otra de lascondiciones representativas de este régimen.Comparando la ecuación cinética

–rA = k1;1,7 [CO2] [DEA]1,7 (22)

con la ecuación (7) y reordenándola se obtiene:

(23)

por lo que, si se correlacionan los valores de 1/(k1;1,7 [DEA]0,7 )frente a los de 1/[DEA], se obtienen los valores de los coe-ficientes cinéticos de las etapas elementales del meca-nismo propuesto por Danckwerts, representado por lasecuaciones (5) y (6), obteniéndose:

k1= 208,9 m3 kmol–1 s–1 k2/k3 = 1,71 kmol m–3

En la ecuación (7) se aprecia que el cociente k2/k3 es el queinfluye de forma fundamental sobre el orden de reacciónde la DEA, e indica la importancia del disolvente en el pro-ceso de absorción. La influencia de la naturaleza del disol-vente se puede atribuir a la constante dieléctrica del disol-vente puro, esto es del isopropanol, y al carácter básicode la amina.

CONCLUSIONES

El estudio experimental realizado de la absorción del CO2

en disoluciones de dietanolamina, DEA, en isopropanol a298 K en una columna de paredes mojadas, permitió deter-minar los valores de concentración de DEA, presión par-cial y caudal de líquido que aseguraba las condiciones depseudo-1er orden, E = Ha.

– Esas condiciones permitieron calcular el orden de reac-ción respecto a la DEA, n, y el coeficiente cinético apa-rente obteniéndose los valores de:

n = 1,7 k = 76 (m3/kmol)1,7 s–1

– El resultado anterior puede explicarse conforme al meca-nismo propuesto por Caplow y desarrollado porDanckwerts,

k1CO2 + RRNH ↔ RRN+HCOO– (5)

k2

k3RRN+HCOO– + RRNH → RRNCOO–RRN+H2 (6)

cuya ecuación cinética resultante es

(7)r

CO DEA

kk

k k DEA

=[ ][ ]

+ [ ]

2

1

2

1 3

1

1 1 1

11707

1

2

1 3k DEA kk

k k DEA, ,

,[ ]= + [ ]

si los coeficientes cinéticos de las etapas elementalestoman los valores

k1 = 208,9 m3/(kmol.s) k2/k3 = 1,71 kmol/m3

Estos valores son de gran utilidad para discernir el meca-nismo de la absorción del CO2 en disoluciones acuosas uorgánicas de aminas en función de la basicidad de la mis-ma y de la influencia de la naturaleza del disolvente.

NOMENCLATURA

C Concentración molar.D Coeficiente de difusión.h Altura de la columna.H Constante de la ley de Henry.km,n Coeficiente cinético aparente.k1, k2, k3 Coeficientes cinéticos de las etapas elementales

definidas por las ecs. (5) y (6).kG Coeficiente individual de transferencia de mate-

ria referido a la fase gaseosa.kL Coeficiente individual de transferencia de mate-

ria referido a la fase líquida.Lm Caudal de mojado.Re Módulo adimensional de Reynolds, dado según

Re = 4 Lm ρ μ–1.Sh Módulo adimensional de Sherwood, dado según

Sh = kL h DA–1.

T Temperatura.z Coeficiente estequiométrico.ρ Densidad.μ Viscosidad.

Subíndices

i Condiciones en la interfase.o Condiciones en el seno.

BIBLIOGRAFÍA

(1). Álvarez F. C.; Midoux, N.; Laurent, A. y Charpentier, J.C.(1980b): «Chemical Kinetics of the Reaction of CarbonDioxide with Amines in Pseudo m-nth Order Conditions inAqueous and Organic Solutions». Chem. Engng. Sci. 35,1717.(2). Álvarez F. C.; Midoux, N.; Laurent, A. y Charpentier, J.C.(1981): «Chemical Kinetics of the Reaction of CO2 withAmines in Pseudo m-n th Order Conditions in Polar andViscous Organic Solutions». Chem. Engng. Sci. 36, 1513.(3). Benítez García, F.J. (1989a): «Reacción entre Dióxido deCarbono y Aminas. Parte I. Aminas Primarias ySecundarias». Ingeniería Química, Octubre, 319.(4). Benítez García, F.J. (1989b): «Reacción entre Dióxido deCarbono y Aminas. Parte II. Aminas Terciarias». IngenieríaQuímica, Noviembre, 215.(5). Belhaj, M.S. (1980): «Contribution a Etude des ReacteursGaz-Liquide a Cube Agitee Mecaniquement: Hydrodynamiqueet Transfert de Matiere en Milieu Aqueous et Organiqueavec les Liquides Moussant et ne Moussant pas». Thése,INPL, Nancy.(6). Blauwhoff, P. M. M.; Versteeg, G. F. y Van Swaaij, W. P.M. (1984): «A Study on the Reaction between CO2 andAlkanolamines in Aqueous Solutions». Chem. Engng. Sci.,39, 207.(7). Brian, P.L.T.; Vivian, J.E. y Matiatos, D.C. (1967):«Interfacial Turbulence During the Absorption of CarbonDioxide into MOnoethanolamine» AIChE J., 13(1), 28.


Página 110

(8). Camacho, F.; Sánchez, S.y Pacheco , R. (2000):«Absorción de Dióxido de Carbono Puro en DisolucionesAcuosas de Monoetanolamina». Afinidad, LVII, 485, 17.(9). Caplow, M. (1968): «Kinetics of Carbamate Formationand Breakdown». J. Am. Chem. Soc., 90, 6795.(10). Clarke, J.K.A. (1964): «Kinetics of Absorption of CarbonDioxide in Monoethanolamine Solutions at Short ContactTimes». I & EC Fundamentals, 3, 239.(11). Danckwerts, P.V. y Sharma, M. M. (1966): «TheAbsorption of Carbon Dioxide into Solutions of Alkalis andAmines». Chem. Engng., 10, CE244.(12). Danckwerts, P.V. (1979): «The Reaction of Carbon Dioxideinto with Ethanolamines». Chem. Engng. Sci., 34, 443.(13). Delgado, S.; Limiñana, G.; Díaz, F.; Pérez, J.F.. (1986):«Absorción de Dióxido de Carbono por Disoluciones deMonoetanolamina en Etanol». Anales de Química, 82, 573.(14). Hikita, H. Y Asai, S. (1964): «Gas absorption with (m-n)-th Order Irreversible Chemical Reaction». Int. Chem. Engn.,4, 332.(15). Hikita, H.; Asai, S.; Ishikawa, H. y Honda, M. (1977): «TheKinetics of Reactions of Carbon Dioxide with Monoethano-lamine, Diethanolamine and Triethanolamine by a RapidMixing Method". Chem. Engng. J., 13, 7.(16). Laddha, S.S.; Díaz, J.M. y Danckwerts, P.V. (1981): «TheN2O Analogy: the Solubilities of CO2 and N2O in AqueousSolutions of Organic Compounds». Chem. Engng. Sci., 36,228.(17). Limiñana, G.; Díaz, F.; Pérez, J.F. y Rodríguez, J.M.(1987): «Absorción de Dióxido de Carbono por Etanol enColumna de Paredes Mojadas». Anales de Química, 83,229.(18). Nunge, R.J. y Gill, W.N. (1963): «Gas-Liquid Kinetics: theAbsorption of Carbon Dioxide in Diethanolamine». AIChEJ., 9, 469.

(19). Oyevaar, M.H.; Morssinkhof, R.W.J. y Westerterp. K.R.(1990): «The Kinetics of the Reaction Between CO2 andDiethanolamine in Aqueous Ethylenglycol at 298 K: aViscous Gas-Liquid Reaction System for the Determinationof Interfacial Areas in Gas.Liquid Contactors". Chem. Engng.Sci., 45, 3283.(20). Partington, J.R. (1951): «An Advanced Treatise onPhysical Chemistry». Vol. II. «The Properties of Liquids».Longmans, Green and Co. London.(21). Reid, R.C.; Prausnitz, J.M. y Polling, B.E. (1987): «TheProperties of Gases and Liquids». Fourth Edition, Mc-GrawHill, London.(22). Sada, E.; Kumazawa, H., y Butt, M. A. (1976a): «ChemicalAbsorption Kinetics over Wide Range of Contact Time:Absorption of Carbon Dioxide into Aqueous Solutions ofMonoethanolamine». AIChE. J., 22, 196.(23). Sada, E.; Kumazawa, H. y Butt, M. A. (1977): «Solubilitiesof Gases in Aqueous Solutions of Amine». J. Chem. Eng.Data, 22, 277.(24). Sada, E.; Kumazawa, H.; Han, Z.Q. y Matsuyama, H.A.(1985): «Chemical Kinetics of the Reaction of Carbon Dioxidewith Ethanolamines in Non Aqueous Solvents». AIChE J.,31, 1297.(25). Savage, D.W. y Kim, C.J. (1985): «Chemical Kinetics ofCarbon Dioxide Reactions with Diethanolamine andDiisopropanolamine in aqueous solutions». AIChE J. 31,296.(26). Sridharan, K. y Sharma, M.M. (1976): «New Systems andMethods for the Measurement of Effective Interfacial Areaand Mass Transfer Coefficients in Gas-Liquuid Reactors».Chem. Engng. Sci., 31, 767.(27). Versteeg, G. F. y Van Swaaij, W. P. M. (1988b): «On theKinetics between CO2 and Alkanolamines Both in Aqueousand Non Aqueous Solutions. I. Primary and secondary ami-nes». Chem. Engng. Sci., 43, 573.


Página 111

Dilute SO2 Absorption Equilibria in Aqueous HCl and NaClSolutions at 298.15 K

Juan Rodrıguez-Sevilla,* Manuel AÄ lvarez, Gabriel Liminana, and Marıa C. DıazDpto. de Ingenierıa Quımica, Facultad de Quımica, Universidad de La Laguna, Avda, Astrofısico FranciscoSanchez, s/n, 38200, La Laguna, Spain

The solubility of dilute SO2 in aqueous HCl and NaCl solutions was determined at 298.15 K as a functionof ionic strength up to 3 mol‚L-1, in the partial pressure range between (0.075 and 1.8) kPa. Measurementswere carried out by a saturation method using a laboratory batch reactor. Equations to correlate theapparent Henry’s law constant, Hm, and the apparent first dissociation constant, Km1, were obtained asa function of ionic strength. Comparisons between experimental results and theoretical predictions werealso made. A model based on the classical Sechenov equation was tested, and a new value for the SO2

gas-specific parameter was obtained to calculate Hm. Two different models were considered for calculatingthe activity coefficients to determine Km1. An extended version of the Debye-Huckel theory describedthe experimental results for ionic strengths below 0.1 mol‚L-1. The Pitzer model was in good agreementwith the experimental data in the ionic strength range between (0 and 3) mol‚L-1. The close fit betweenmeasured and calculated data showed that the selected models can be successfully used for estimatingthe solubility of SO2 in salt solutions at low partial pressures.

Introduction

Sulfur dioxide is an important atmospheric contaminant.Its main source is stack gas due to burning of fuels withhigh sulfur content. SO2 can be removed either during orimmediately after combustion. In the latter case, the fluegas may be washed with an alkaline slurry or solution.Knowledge of the equilibrium data is an indispensablerequirement for the design of the absorption and desorptionprocesses. Several methods for flue gas desulfurizationhave been employed according to whether the reagent willbe recycled or not. Seawater washing has been identifiedas an option among the nonregenerative processes, sinceit provides a natural alkaline phase. Such a process offerspotential advantages for power stations located nearby thecoast, since it is of simple design, requires no bulkchemicals, and has low capital and operating costs.1Seawater is a complex liquid phase, with NaCl being itsmain component and with bicarbonate and sulfate speciesbeing responsible for seawater alkalinity. The electrolytecomposition of the liquid phase plays an important role inthe absorption equilibrium, since it determines the non-ideal behavior of the solution by means of the activitycoefficients. Frequently, experimental measurements onsingle-electrolyte solutions, for instance NaCl solutions, areused to test methods for the estimation of the activitycoefficients. The absorption equilibrium of SO2 in aqueousNaCl solutions has already been studied at a pressurearound 100 kPa,2 but few literature data are available inthe low partial pressure range of SO2. The low concentra-tion range of SO2 is of interest in the purification of largeamounts of waste gases and in the food industry where SO2

is used as a preservative.Thermodynamic constants for the SO2 absorption equi-

libria in NaCl solutions can be described as follows:

(1) The gas/liquid equilibrium:

where H°m is the Henry’s law constant, Hm is the apparentHenry’s law constant, φSO2 represents the fugacity coef-ficient, γSO2 represents the activity coefficient, PSO2 repre-sents the partial pressure of SO2, and m represents themolal concentration. The subscript m refers to molalconcentration.

(2) The first dissociation of SO2(aq):

where K1 is the first dissociation constant and Km1 is theapparent first dissociation constant. Qγ represents a termthat includes the activity coefficients of the different speciesinvolved, and aH2O is the activity of water.

The influence of the electrolyte concentration on thegas-liquid equilibrium has been studied for many gaseoussolutes and salt solutions.3-9 With increasing salt concen-tration, gas solubility is nearly always found to decrease(“salting out” effect). At moderately high salt concentration,the solubility of a sparingly soluble gas compared to thatone in pure water could be described by the classicalSechenov equation:

where Hc is the apparent Henry’s law constant, Cs is thesalt molar concentration, the superscript zero refers to pure

* To whom correspondence should be addressed. E-mail: [email protected]. Telephone number: +34 922 318058. Fax number: +34 922318004.

SO2(g) T SO2(aq)

H°m )φSO2

PSO2

γSO2mSO2

)φSO2

γSO2

Hm (1)

SO2(aq) + H2O T H++ HSO3-

K1 )γH+γHSO3

-

γSO2aH2O

mH+mHSO3-

mSO2

) QγKm1 (2)

logHc

H°c) kcCs (3)

1339J. Chem. Eng. Data 2002, 47, 1339-1345

10.1021/je015538e CCC: $22.00 © 2002 American Chemical SocietyPublished on Web 09/11/2002

Página 122

water, and the subscript c refers to molar concentration.The proportionality constant kc, the Sechenov molar con-stant, is specific to the gas and salt and depends partiallyon temperature. There are several empirical models for theestimation of kc. Recently, Schumpe,7 Hermann et al.,8 andWeisenberger and Schumpe9 have suggested a modelconsidering individual salting-out effects of the ions,

Here hi and hG are the ion-specific and gas-specific param-eters, respectively, and Ci represents molar concentrationof the ion i. For a single salt, the relationship between theSechenov constant and Schumpe parameters is

where ni is the index of the ion i in the salt formula.With regard to the activity coefficients, Robinson and

Stokes10 developed a simple thermodynamic model usingan extended version of the Debye-Huckel theory, valid fordilute electrolyte solutions. The model equation is givenby

where I is the ionic strength, A is the constant in theDebye-Huckel equation, B is the Debye-Huckel param-eter, z is the charge number, σ is the minimum approachdistance, and � is the specific interaction parameter.

More recently, Pitzer et al.11-17 developed methods toestimate the activity coefficients of solutes in complexconcentrated electrolyte solutions. The standard definitionsand thermodynamic transformations yield the followingequation for the activity coefficients:

The first term on the right of the equation is an extendedform of the Debye-Huckel parameter. The quantity λij

represents the short-range interaction between soluteparticles i and j in the presence of the solvent. This binaryinteraction parameter, or second virial coefficient, does notdepend itself on the composition of neutral species, but inthe case of ions it is dependent on the ionic strength. Itdoes depend on the particular solute species i and j and ontemperature and pressure. The equivalent quantity fortriple interactions is μijk; in principle, it might be ionicstrength dependent, but with a single possible exception16

there is no evidence of this. On the other hand, λ′ is thepartial derivative of λ with respect to the ionic strength.

The purpose of the present paper is to test several modelsused to predict the activity coefficients for the absorptionof SO2 into aqueous solutions at low partial pressures. Anexperimental study was carried out to evaluate the effectof the electrolyte concentration on the thermodynamicequilibrium involved in the absorption process. The solubil-ity of dilute SO2 in several electrolyte solutions wasdetermined along with the values of the apparent Henry’s

law constant and the apparent first dissociation constant.To study independently the effect of the electrolyte con-centration on the gas-liquid equilibrium, the dissociationof SO2(aq), eq 2, was inhibited by using different acidsolutions.

Experimental Section

(a) Materials. The sulfur dioxide and nitrogen used inthe present work had purities higher than 99.9 vol % (AirLiquide). Sodium chloride was an analytical grade reagent(Merck) with a purity higher than 99 mass %. Hydrogenchloride, iodine, and sodium thiosulfate solutions wereprepared from standard Tritisol ampules (Merck). Thewater used to prepare each solution was distilled anddeionized with a Milli-Q Plus device to 18 MΩ‚cm (Milli-pore).

(b) Experimental Procedure. Experiments were car-ried out at 298.15 K and 97.6 kPa, using SO2 + N2 mixturesin several electrolyte solutions in the SO2 partial pressurerange between (0.075 and 1.8) kPa. The electrolyte solu-tions used were NaCl(aq) (0 to 3) mol‚L-1, HCl(aq) (0.5 to1) mol‚L-1, and HCl + NaCl mixtures (1 to 4) mol‚L-1 totalionic strength.

The solubility of sulfur dioxide in aqueous solution wasdetermined by a saturation method where the gas wasbubbled through the liquid phase. The apparatus used isshown in Figure 1. The aqueous solutions were located ina jacketed vessel, and the temperature inside the solutioncould be regulated within (0.1 K by circulating water froma thermostatic bath (Haake). Thermometers with subdivi-sions of (0.1 K were used to monitor the bath and solutiontemperatures. A control valve adjusted the total pressureinside the vessel with (0.05 kPa accuracy.

SO2 + N2 mixtures were prepared from pure gases withcalibrated mass-flow controllers (Brooks Instruments). Theaccuracies of the mass-flow controllers were (10-5 and(10-3 mol‚min-1 for SO2 and N2, respectively. From thetotal pressure and temperature, the partial pressure of SO2

was calculated as follows:

logHc

H°c) ∑

i

(hi + hG)Ci (4)

kc ) ∑i

(hi + hG)ni (5)

log γi ) A[ -zi2�I

1 + Bσi�I+ �iI] (6)

ln γi ) f γzi2 + 2∑

j

λijmj +zi

2

2∑

j∑

k

λ′jkmjmk +

3∑j∑

k

μijkmjmk (7)

Figure 1. Experimental installation: (1) calibrated mass-flowcontrollers; (2) mixing unit; (3) thermostatic bath; (4) jacketedvessel; (5) temperature controller agitator; (6) IR SO2 analyzer;(7) dehumidification unit; (8) cooler box; (9) main gas fluid; (10)additional gas fluid; (11) digital pH meter.

1340 Journal of Chemical and Engineering Data, Vol. 47, No. 6, 2002

Página 123

where P is the total pressure, PW is the vapor pressure ofwater at the operation temperature, and GSO2 and GN2 arethe molar flow of SO2 and N2, respectively. The overalluncertainty in the SO2 partial pressure adjustment wasestimated to be (2%. For SO2 partial pressures below 500Pa, an infrared SO2 analyzer (ADC) directly measured theSO2 molar fraction. The accuracy in the SO2 partialpressure measurement was (3 Pa.

The sulfur(IV) concentration was determined, once equi-librium was reached, by adding a known volume of solutionfrom the vessel to a known volume of standard iodine. Theexcess iodine was back-titrated with standard sodiumthiosulfate. The overall uncertainty in the determinationof the sulfur(IV) concentration was estimated to be (1%.Measurements of pH were carried out using a combinedelectrode (Ingold), located in the vessel solution, and adigital pH meter (Radiometer). Buffer solutions of pH 1.68( 0.01 (Radiometer), 4.00 ( 0.02 (Merck), and 7.00 ( 0.02(Merck) were used for the instrument calibration. Todetermine the concentration of H+ from the pH measure-ments, the electrode response was calibrated in the NaClmedia using solutions with known concentrations of HCl.For each NaCl medium, an experimental correlation be-tween the electrode response and the H+ concentration wasobtained. The overall uncertainty in the determination ofthe H+ concentration was less than (2%. In the solutions,the SO2 concentration was calculated from the experimen-tal data of sulfur(IV) and H+ concentrations by means ofthe sulfur balance and the electroneutrality condition, asfollows:

When the dissociation of SO2(aq) was inhibited, the totalsulfur(IV) concentration, mS(IV), equals the SO2 concentra-tion, mSO2.

Discussion and Results

(a) SO2 Solubility in Aqueous NaCl Solutions.Measurements at 298.15 K allowed the effect of ionicstrength on the SO2 solubility to be evaluated. Figure 2shows the experimental solubility of SO2 and H+ concen-tration as a function of ionic strength for three SO2 partialpressures. Values of the apparent Henry’s law constant andthe apparent first dissociation constant, determined fromSO2 equilibration measurements, are given in Table 1.

Since this study concerns only low pressures, the fugacitycoefficient of SO2 can be set equal to unity. In accordancewith eq 1, Hm/H°m gives the SO2 activity coefficient, andthe superscript zero denotes the value in pure water. TheγSO2 values were correlated through a Sechenov typeequation as follows:

The value of H°m (82.9 kPa‚kg‚mol-1) has been discussedin a previous work.18 Literature data on the Henry’s lawconstant of sulfur dioxide in water are extensive. Thecorrelations of Edwards et al.19 and Goldberg and Parker,20

based on selected data, give values at 298.15 K of 80.8 and81.3 kPa‚kg‚mol-1, respectively, which are in good agree-ment with our experimental result and with the valuesreported by Rabe and Harris21 (84.6 kPa‚kg‚mol-1) andRumpf and Maurer22 (85.2 kPa‚kg‚mol-1). Some differencesare observed with other literature data, such as thosereported by Renon23 (88.7 kPa‚kg‚mol-1), as cited byHunger et al.,24 Camacho et al.25 (74.3 kPa‚kg‚mol-1), andMillero et al.2 (89.7 kPa‚kg‚mol-1). In relation to theSechenov constant of SO2 in NaCl(aq), few literature dataare available. Chang and Rochelle26 cited a molar constantof 0.076 L‚mol-1, taken from Harned and Owen.27 Milleroet al.2 measured the solubility of SO2 in NaCl solutions,giving a molal constant of 0.0246 kg‚mol-1 at 298.15 K,

PSO2) (P - PW)

GSO2

GSO2+ GN2

(8)

mS(IV) ) mSO2+ mHSO3

-

mHSO3-) mH+

mSO2) mS(IV) - mH+ (9)

log γSO2) log

Hm

H°m) 0.0188I/mol‚kg-1 (10)

Figure 2. Experimental results in NaCl solutions for three SO2

partial pressures: (2) 0.5 kPa; (9) 1.0 kPa; (b) 1.5 kPa; (a) totalsolubility; (b) H+ concentration.

Table 1. Experimental Values of the Apparent Henry’sLaw Constant and the Apparent First DissociationConstant in SO2 + NaCl Solutions at 298.15 K andP ) 97.6 kPa

I Hm

mol‚L-1 mol‚kg-1 kPa‚mol-1‚kg r2 a Km1 r2 a Nb

water 82.9 0.999 0.0114 0.986 160.1 0.10 82.7 0.997 0.0200 0.991 190.5 0.51 86.6 0.998 0.0255 0.996 141.0 1.02 85.4 0.994 0.0250 0.984 201.5 1.55 91.2 0.982 0.0299 0.980 152.0 2.09 94.4 0.987 0.0220 0.983 103.0 3.21 91.7 0.999 0.0182 0.994 10

a r2: correlation coefficient. b N: number of experimental data.

Journal of Chemical and Engineering Data, Vol. 47, No. 6, 2002 1341

Página 124

similar to that obtained in the present work (0.0188kg‚mol-1). The application of the Schumpe model,9 eq 4,with hNa+ ) 0.1143 L‚mol-1, hCl- ) 0.0318 L‚mol-1, andhSO2 )-0.0817 L‚mol-1, leads to a Sechenov molar constantof -0.0173 L‚mol-1, equivalent to a molal constant of-0.0249 kg‚mol-1. Figure 3 shows the experimental valuesof γSO2 as a function of the ionic strength, together withthe values predicted by the Schumpe model. Although theionic strength effect is small, it can be noted that theexperimental results are contrary to the predictions of themodel. Since the ion-specific parameters for Na+ and Cl-have been widely used in the literature with other gasesbut not with SO2, the gas-specific parameter given for SO2

may not be valid in NaCl solutions. For this reason, someadditional experiments were done with acid solutions, toevaluate the gas-liquid equilibrium separately. The resultsare discussed in section b.

The apparent first dissociation constant of SO2(aq) wasdetermined from experimental data of H+ and SO2 con-centrations. The values shown in Table 1 were correlatedas follows:2

where pKm1 is (-log Km1). In this equation, the value 1.94is the pK1, correspondent to the thermodynamic constantK1 at 298.15 K, and the sum of the other terms on the rightside corresponds to (log Qγ) defined by eq 2. Literature dataon thermodynamic constants for the dissociation of SO2(aq)in water are extensive, and some of them have beendiscussed in a previous work.18 The pK1 obtained at 298.15K is in reasonable agreement with the correlations ofEdwards et al.19 and Goldberg and Parker,20 which givevalues of 1.89 and 1.87, respectively, and with the data ofMillero et al.,2 who reported the value of 1.86. Somedifferences are observed with other data, such as thosereported by Renon23 (pK1 ) 1.76), as cited by Hunger etal.24 and Camacho et al.25 (pK1 ) 1.73).

To obtain semiempirical expressions that allow the totalsolubility of SO2 and the H+ concentration to be estimatedas a function of the SO2 partial pressure, eqs 1, 2, and 9were combined to give

By considering eqs 10 and 11, using the value of theHenry’s law constant in pure water, the following equationsare derived:

Figure 4 shows the correlation between the experimentaldata, for H+ and total sulfur(IV), and those calculated byeqs 14 and 15. As can be observed, these equations allowthe estimation of the SO2 + NaCl(aq) equilibrium with anoverall uncertainty of (10% in the ranges studied.

(b) SO2 Solubility in Acid Solutions. Solubility mea-surements were made at 298.15 K with different acidsolutions inhibiting SO2 dissociation in order to study theeffect of the electrolyte concentration on the gas-liquidequilibrium. All the liquid phases used showed that thedissociation of SO2 was less than 5%. The apparent Henry’slaw constant and the activity coefficient of SO2, given inTable 2, were calculated from experimental solubility data.In that table, Cs represents the molar concentration of theelectrolyte solution.

A value for the SO2 gas-specific parameter was obtainedby minimizing the following function:

Figure 3. Activity coefficient of SO2 in NaCl solutions as afunction of ionic strength: (2) experimental; (- - -) Schumpemodel.9

pKm1 ) 1.94 - 0.74�I/mol‚kg-1 + 0.35I/mol‚kg-1 (11)

mS(IV) )PSO2

Hm+ �Km1

HmPSO2

(12)

mH+ ) �Km1

HmPSO2

(13)

Figure 4. Comparison of experimental and calculated values inNaCl solutions: (2) experimental; (- - -) (diagonal (10%) deviation;(a) total solubility; (b) H+ concentration.

mS(IV)/mol‚kg-1 ) 10-(4.919+0.0188I/mol‚kg-1)PSO2/

Pa +10-(3.430-0.37�I/mol‚kg-1+0.185I/mol‚kg-1)�PSO2/Pa (14)

mH+/mol‚kg-1 )

10-(3.430-0.37�I/mol‚kg-1+0.185I/mol‚kg-1)�PSO2/Pa (15)


Página 125

In this function, log(γSO2

c )calc is obtained by means of theSchumpe model, eq 4, assuming that the fugacity coefficientfor SO2, φSO2, is equal to unity; then γSO2

c is given by Hc/H°c.The ion-specific parameters for Cl-, Na+, and H+ werefitted to those reported in the literature9 (hCl- ) 0.0318m3‚kmol-1, hNa+ ) 0.1143 m3‚kmol-1, hH+ ) 0 m3‚kmol-1),and then hSO2 was varied. The optimized parameter is

This value is different from that reported in the literature9

(hSO2 ) -0.0817 m3‚kmol-1). Figure 5 shows the correla-tion between the experimental and calculated values oflog(γSO2

c ), using the new value obtained for hSO2. As can beobserved, this new value allows us to represent theexperimental data of γSO2

c within a deviation of (10%. Theregression coefficient and the standard deviation were0.835 and 0.012, respectively.

(c) Thermodynamic Models for the SO2(aq) Dis-sociation Equilibrium in NaCl Solutions. For theevaluation of the activity coefficients of solutes and thewater activity, two different models were considered. Thefirst one used a water activity equal to unity and eq 6 forthe activity coefficients of solutes. The model parametersare shown in Table 3.

The second one was the Pitzer model,11-17 where theactivity coefficients for H+ and HSO3

- in NaCl solutionsare given by

For systems containing a single salt MX, for example NaCl,the model parameters are given by

where Aφ represents the Debye-Huckel parameter for theosmotic coefficient (0.392 kg1/2‚mol-1/2). The parameters λi,j,λ′i,j, and CMX

(φ) describe the thermodynamic properties of apure electrolyte; λ and λ′ represent measurable combina-tions of second virial coefficients, which arise from binaryinteractions given by �MX

(0) and �MX(1) , while CMX

(φ) representsthird virial coefficients and considers tertiary interactionsof the type M-M-X and M-X-X.

The additional parameter θMN, introduced for electrolytemixtures, represents the difference between the secondvirial coefficients for ions M and N of the same sign andinteractions between the average second virial coefficientsM-M and N-N. Because second virial coefficients aredetermined mainly by short-range interactions and ions ofequal sign are seldom close together, θMN is expected to besmall. Pitzer and Kim15 include further mixing parametersψ arising from differences in third virial coefficients, butthese terms are smaller than θ and are negligible for singleelectrolytes.

The parameters �MX(0) and �MX

(1) and CMX(φ) were evaluated

for a large number of electrolytes by Pitzer and Mayorga14

and Pitzer and Kim,15 but few activity coefficient data

Table 2. Experimental Values of the Apparent Henry’s Law Constant for SO2 in Acid Solutions at 298.15 K

Cs I Hc Hm

liquid phase mol‚L-1 mol‚L-1 kPa‚mol-1‚L kPa‚mol-1‚kg r2 log γSO2

c

HCl 0.5 0.5 75.9 75.1 0.999 -0.03941.0 1.0 74.8 73.2 0.999 -0.0457

HCl + NaCl 0.5 + 0.5 1.0 78.5 76.2 0.999 -0.02470.5 + 1.5 2.0 80.9 76.8 0.999 -0.01170.5 + 3.5 4.0 85.7 77.7 0.999 0.0134

Figure 5. Comparison of experimental and calculated values oflog γSO2

c : (2) HCl + NaCl; (9) HCl; (b) NaCl; (- - -) (γSO2

c (10%)deviation.

[log(γSO2

c )exp - log(γSO2

c )calc]2 (16)

hSO2) -0.0607 m3‚kmol-1

ln γH ) zH2 f γ + 2mCl(λHCl + ECHCl) + 2mHSO3

λHHSO3+

mNamCl(λ′NaCl + CNaCl) + mHmCl(λ′HCl + CHCl) +mNamHSO3

(λ′NaHSO3+ CNaHSO3

) + mHmHSO3λ′HHSO3

+

mNa(2θHNa + mClΨHNaCl) (17)

Table 3. Parameters in the Robinson and StokesEquation, Eq 6

A/kg1/2‚mol-1/2 0.508510-10 B/kg1/2‚mol-1/2‚m-1 0.32811010σH+/m 9.01010σHSO3

-/m 4.0�H+/kg‚mol-1 0.4�HSO3

-/kg‚mol-1 0.0

ln γHSO3) zHSO3

2 f γ + 2mNa(λNaHSO3+ ECNaHSO3

) +

2mHλHHSO3+ mNamCl(λ′NaCl + CNaCl) + mHmCl(λ′HCl +

CHCl) + mNamHSO3(λ′NaHSO3

+ CNaHSO3) +

mHmHSO3λ′HHSO3

(18)

f γ ) -Aφ[ �I

1 + 1.2�I+ ( 2

1.2) ln(1 + 1.2�I)] (19)

λMX ) �MX(0) + (2�MX

(1)

4I )[1 - (1 + 2�I) exp(-2�I)] (20)

λ )�MX

(1)

2I2[-1 + (1 + 2�I + 2I) exp(-2�I)] (21)

CMX )CMX

(φ)

2|zMzX|1/2(22)

E )1

2∑

i

mi|zi| (23)


Página 126

are available for SO2 and NaHSO3 solutions. Values of�NaHSO3

(0) , �NaHSO3

(1) , and CNaHSO3

(φ) were estimated by Rosen-blatt,28 from the effect of the ions on the structure of liquidwater. Hungers et al.24 modeled the thermodynamic equi-librium of dilute SO2 absorption into Na2SO4 or H2SO4

electrolyte solutions and estimated the values of �NaHSO3

(0)

and �NaHSO3

(1) , in accordance with Rosenblatt’s theories. Inaddition, Millero et al.2 determined the values of �MX

(0) , �MX(1) ,

and CMX(φ) for NaHSO3 and Na2SO3 from experimental

measurements of the SO2 solubility in NaCl solutions. Thewhole set of parameter values used in this work is listedin Table 4.

The water activity was calculated by the general Pitzerprocedure compiled by Rumpf and Maurer:29

Figure 6 shows a comparison between the experimentalpKm1 data obtained in this work and the correspondingvalues calculated using the above thermodynamic models.The Robinson-Stokes equation is in good agreement withexperimental data for ionic strengths below 0.1 mol‚L-1,but a deviation is observed at higher ionic strengths. Bycontrast, the Pitzer model is satisfactory in the studiedrange of ionic strength when Rosenblatt’s parameters forNa-HSO3 interactions are used.

Conclusions

The equilibrium distribution of SO2 between a gaseousphase and several electrolyte solutions was studied experi-mentally at 298.15 K and 97.6 kPa in a partial pressurerange between (0.075 and 1.8) kPa. Equations were derivedto correlate the apparent Henry’s law constant, Hm, andthe apparent first dissociation constant, Km1, as a functionof the ionic strength, which ranged between (0 and 3)mol‚L-1. Comparisons between experimental results andtheoretical predictions were also made. With regard to Hm,the Schumpe model,7-9 based on the classical Sechenovequation, was used and a new value for the SO2 gas-specificparameter, hSO2, was obtained. With regard to Km1, twodifferent models were tested to calculate the activitycoefficients. The simplest one, an extended version of theDebye-Huckel theory valid for dilute electrolyte solutions,described the experimental results for ionic strengths below0.1 mol‚L-1. The second one, a Pitzer model, was in goodagreement with experimental data in the ionic strengthrange between (0 and 3) mol‚L-1. In this case, Rosenblatt’sparameters28 were used to characterize Na-HSO3 interac-tions. The close agreement between the measured andcalculated data showed that the selected models couldsuccessfully be used to estimate the solubility of dilute SO2

in HCl and NaCl solutions.

Literature Cited

(1) Baty, R.; Coughland, J.; Reynolds, S. K. Technical and Environ-mental Implications of Desulfurization by Seawater Washing.Environ. Prot. Bull. 1991, 012, 21-31.

(2) Millero, F. J.; Hershey, P.; Johnson, G. The Solubility of SO2 andthe Dissociation of H2SO3 in NaCl Solutions. J. Atmos. Chem.1989, 8, 377-389.

(3) Van Krevelen, D. W.; Hofiijzer, P. J. Sur la Solubilite des Gasdans les Solutions Aqueuses. Chimie et Industrie (Numero Specialedu Xle Congres International de Chimie Industrielle); Bruxelles,1950; pp 168-173.

(4) Danckwerts, P. V. Gas-Liquid Reactions; McGraw-Hill: NewYork, 1970.

(5) Onda, K.; Sada, E.; Kobayashi, T.; Kito, S.; Ito, K. Salting OutParameters of Gas Solubility in Aqueous Salt Solutions. J. Chem.Eng. Jpn. 1970a, 3, 18-24.

(6) Onda, K.; Sada, E.; Kobayashi, T.; Kito, S.; Ito, K. Solubility inAqueous Solutions of Mixed Salts. J. Chem. Eng. Jpn. 1970b, 3,137-142.

(7) Schumpe, A. The Estimation of Gas Solubilities in Salt Solutions.Chem. Eng. Sci. 1993, 48 (1), 153-158.

(8) Hermann, C.; Dewes, L.; Schumpe, A. The Estimation of GasSolubilities in Salt Solutions. Chem. Eng. Sci. 1995, 50 (10),1673-1675.

(9) Weisenberger, S.; Schumpe, A. Estimation of Gas Solubilities inSalt Solutions at Temperatures from 273 K to 363 K. AIChE J.1996, 42 (1), 298-300.

(10) Robinson, R. A.; Stokes, R. H. Electrolyte Solutions, 2nd ed.;Butterworth: Boston, 1959.

(11) Pitzer, K. S. Thermodynamics of Electrolytes. 1. Theoretical basisand general equations. J. Phys. Chem. 1973, 77, 268-277.

(12) Pitzer, K. S. Theory: Ion Interaction approach. In ActivityCoefficients in Electrolyte Solutions; Pytkowicz, R., Ed.; CRCPress: Boca Raton, FL, 1979; pp 157-208.

(13) Pitzer, K. S.; Mayorga, G. Thermodynamics of Electrolytes. II.Activity and Osmotic Coefficients for Strong Electrolytes with oneor both Ions Univalent. J. Phys. Chem. 1973, 77, 2300-2308.

(14) Pitzer, K. S.; Mayorga, G. Thermodynamics of Electrolytes. III.Activity and osmotic coefficients for 2-2 electrolytes. J. SolutionChem. 1974, 3, 539-546.

(15) Pitzer, K. S.; Kim, J. J. Thermodynamics of Electrolytes. IV.Activity and Osmotic Coefficients for Mixed Electrolytes. J. Am.Chem. Soc. 1974, 96, 5701-5707.

(16) Puthela, R. C.; Pitzer, K. S. Thermodynamics of ElectrolyteMixtures. Enthalpy and the Effect of Temperature on the ActivityCoefficient. J. Solution Chem. 1986, 15, 649-662.

(17) Pitzer, K. S. A Thermodynamic Model for Aqueous Solutions ofLiquidlike Density. Rev. Mineral. 1987, 17, 97-142.

(18) Rodriguez Sevilla, J.; Alvarez Diaz, M.; Diaz Garcia, C.; Liminanade la Fe, G. Thermodynamic Equilibrium of SO2-H2O System atLow Partial Pressures. Afinidad 2001, 492, 141-146.

Table 4. Single-Electrolyte Solutions and MixingParameters at 298.15 K

electrolyte �MX(0) �MX

(1)MXC(φ) ref

HCl 0.1775 0.2945 0.0008 17NaCl 0.0765 0.2644 0.00127 17HHSO3 0.15 0.40 24NaHSO3 0.0249 0.2455 0.0004 28θHNa ) 0.036; ψHNaCl ) -0.004 17

Figure 6. Comparison of experimental and calculated values ofpKm1 for the dissociation of SO2(aq) in NaCl solutions: (2)experimental; (- - -) Pitzer model (with Rosenblatt’s parameters);(s) Robinson and Stokes equation.

ln aH2O ) MH2O{2Aφ I1.5

1 + 1.2�I-

2[mHmCl(�HCl(0) + �HCl

(1) exp(-2�I)) +

mHmHSO3(�HHSO3

(0) + �HHSO3

(1) exp(-2�I)) +

mNamCl(�NaCl(0) + �NaCl

(1) exp(-2�I)) +

mNamHSO3(�NaHSO3

(0) + �NaHSO3

(1) exp(-2�I))] -

(mH + mNa + mCl + mHSO3)} (24)


Página 127

(19) Edwards, T. J.; Maurer, G.; Newman, J.; Prausnitz, J. M. Vapor-Liquid Equilibria in Multicomponent Aqueous Solutions of Vola-tile Weak Electrolytes. AIChE J. 1978, 24, 966-976.

(20) Goldberg, R. N.; Parker, V. B. Thermodynamics of SO2(g) in Waterand Aqueous Sulfur Dioxide Solutions. J. Res. Natl. Bur. Stand.1985, 90, 341-358.

(21) Rabe, A. E.; Harris, J. F. Vapor Liquid Equilibrium Data for theBinary System, Sulfur Dioxide and Water. J. Chem. Eng. Data1963, 8, 333-336.

(22) Rumpf, B.; Maurer, G. Solubilities of Hydrogen Cyanide andSulfur Dioxide in Water at Temperatures from 293.15 to 413.5 Kand Pressures up to 2.5 MPa. Fluid Phase Equilib. 1992, 81, 241-260.

(23) Renon, H. In Thermodynamics of Aqueous Systems with IndustrialApplications; Newman, S. A., Ed.; ACS Symposium Series 133;American Chemical Society: Washington, DC, 1980; pp 173-186.

(24) Hunger, T.; Lapicque, F.; Storck, A. Thermodynamic Equilibriumof Diluted SO2 Absorption into Na2SO4 or H2SO4 ElectrolyteSolutions. J. Chem. Eng. Data 1990, 35 (4), 453-463.

(25) Camacho, F.; Bravo, J.; Tortosa, J.; Ruiz, J. M. AbsorptionEquilibrium of SO2 in Water. An. Quım. 1986, 82, 399-405.

(26) Chang, C. S.; Rochelle, G. T. SO2 Absorption into NaOH andNa2SO3 Aqueous Solutions. Ind. Eng. Chem. Fundam. 1985, 24,7-11.

(27) Harned, H. S.; Owen, B. B. The Physical Chemistry of ElectrolyticSolutions; Reinhold: New York, 1958.

(28) Rosenblatt, G. M. Estimation of Activity Coefficients in Concen-trated Sulfite-Sulfate Solutions. AIChE J. 1981, 4, 619-626.

(29) Rumpf, B.; Maurer, G. Solubility of Sulfur Dioxide in AqueousSolutions of Sodium and Ammonium Sulfate at Temperaturesfrom 313.15 K to 393.15 K and Pressures up to 3.5 MPa. FluidPhase Equilib. 1993, 91, 113-131.

Received for review November 22, 2001. Accepted August 6, 2002.

JE015538E


Página 128

Absorption Equilibria of Dilute SO2 in Seawater

Juan Rodrıguez-Sevilla,* Manuel AÄ lvarez, Marıa C. Dıaz, and Marıa C. Marrero

Departamento de Ingenierıa Quımica, Facultad de Quımica, Universidad de La Laguna, Avda. AstrofısicoFrancisco Sanchez, s/n. 38200 La Laguna, Spain

The solubility of dilute SO2 in seawater, from mixtures SO2 + N2, was determined in a temperaturerange between (278.15 and 318.15) K and a SO2 partial pressure range between (0.050 and 1.5) kPa.Measurements were carried out by a saturation method using a laboratory-stirred batch reactor. Equationsto correlate experimental data were obtained as a function of temperature and SO2 partial pressure.Comparisons between experimental results, literature data, and theoretical predictions were also madeat different temperatures. Two different models were considered to evaluate the activity coefficients ofionic species. Theoretical calculations were made using both an extended version of the Debye-Huckeltheory and the Pitzer ion-interaction model. Both models led to similar predictions and were in goodagreement with the experimental results. The reasonable fitting between measured and calculated datashowed that the selected models can be successfully used for predicting the absorption equilibria of dilutesulfur dioxide in seawater at different temperatures.

Introduction

Sulfur dioxide is an important atmospheric contaminant.Its main source is flue gases from the burning of fuels witha high sulfur content. SO2 can be removed either duringor immediately after combustion. In the latter case, theSO2 may be absorbed by an alkaline slurry or solution.Several methods for flue gas desulfurization (FGD) havebeen employed according to whether the reagent will beregenerated or not. Seawater scrubbing has been identifiedas an option among the FGD nonregenerative processes,since it provides a natural alkaline phase. Such a processoffers potential advantages for power stations locatednearby the coast, since it has a simple design, requires nobulk chemicals, and has low capital and operating costs.

Knowledge of the equilibrium data involving the sulfurdioxide-seawater system is an indispensable requirementfor the design and operation of the absorption process.Bromley and Read1 reported a few experimental andestimated data on solubility of sulfur dioxide in seawater.Abdulsattar et al.2 proposed a simplified chemical modelusing an extended version of the Debye-Huckel theory forelectrolyte solutions developed by Bromley.3 This modelpredicts the solubility of sulfur dioxide seawater in atemperature range between (283.15 and 298.15) K and aSO2 concentration range between (10-5 and 1) mol‚kg-1. Acomparison of the predicted SO2 solubility in seawater withthat one in freshwater indicated that sulfur dioxide is abouttwo or three times more soluble in seawater than infreshwater. The higher solubility in seawater can beexplained by the presence of alkaline components inseawater (mainly HCO3

- and SO42-). Douabul and Riley4

measured the solubility of sulfur dioxide in distilled waterand decarbonated seawater at a SO2 pressure of 101.3 kPain the temperature range between (278.95 and 303.15) Kand a salinity range between (0 and 40) g‚kg-1. Themeasurements showed that the SO2 solubility was lower

in decarbonated seawater than in distilled water, decreas-ing slightly with the increase of seawater salinity. Recently,Al-Enezi et al.5 measured the SO2 solubility in distilledwater and seawater, from mixtures of SO2 + N2, at a SO2

partial pressure of 22.4 Pa in the temperature rangebetween (283.15 and 313.15) K and a salinity rangebetween (5 and 65.1) g‚kg-1. They provided a second-orderequation to correlate the SO2 solubility as a function oftemperature and initial salinity. Their results indicated aSO2 solubility data lower than those predicted by Abdul-sattar et al.2 and an increase in the SO2 solubility at higherinitial salinity, which is the opposite effect to that foundby Douabul and Riley.4

As shown above, limited data on solubility of sulfurdioxide in seawater are found in the literature, and somediscrepancies are even observed among those data. Hence,the present work is focused on to provide new equilibriummeasurements at low partial pressures and differenttemperatures. Low partial pressure range is of interest inFGD by seawater scrubbing. The performance at differenttemperatures is necessary to simulate variations in theoperational conditions of SO2 absorbers. Also, the experi-mental results will allow testing several thermodynamicsmodels in order to predict equilibrium data for the absorp-tion of sulfur dioxide in seawater.

Experimental Section

(a) Materials. The purity of sulfur dioxide and nitrogenused in the present work was higher than 99.9 volume %(Air Liquide). Natural seawater was collected from thesupply stream of a Flakt-Hydro seawater FGD plantsituated in a coastal power station located at the CanaryIslands, Spain. The ionic composition of the seawater wasmeasured according to the Standard Methods;6 results areshown in Table 1. Artificial seawater samples were pre-pared according to the procedure of Lyman and Fleming,as cited by Riley and Skirrow.7 The ionic composition ofthose samples is also shown in Table 1. Chemicals usedfor preparing artificial seawater were analytical-grade

* To whom correspondence should be addressed. E-mail: [email protected]: +34 922 318058. Fax: +34 922 318004.

1710 J. Chem. Eng. Data 2004, 49, 1710-1716

10.1021/je049833l CCC: $27.50 © 2004 American Chemical SocietyPublished on Web 09/29/2004

Página 136

reagents (Merck) with purities higher than 99 mass %.Iodine and sodium thiosulfate solutions were prepared fromstandard Tritisol ampules (Merck). The water used toprepare all solutions was distilled and deionized with aMilli-Q Plus device, 18 MΩ‚cm (Millipore).

(b) Experimental Procedure. Experiments were car-ried out at a total pressure of 97.6 kPa, using SO2 + N2

mixtures in the SO2 partial pressure range between (0.050and 1.5) kPa. The solubility of sulfur dioxide in aqueoussolution was determined by a saturation method where thegaseous mixture was bubbled through the liquid phase. Theinstallation used has been described in previous works.8,9

SO2 + N2 mixtures were prepared from pure gases withcalibrated mass-flow controllers (Brooks Instruments). Theuncertainties of the molar flow rate measurements were(10-5 and (10-3 mol‚min-1, for SO2 and N2, respectively.The overall uncertainty in the SO2 partial pressure adjust-ment was estimated to be (2%. For SO2 partial pressuresbelow 0.5 kPa, an infrared SO2 analyzer (ADC) directlymeasured the SO2 gaseous molar fraction with an uncer-tainty in the SO2 partial pressure measurement of (0.003kPa. A control valve adjusted the total pressure inside thevessel with an uncertainty of (0.05 kPa.

Seawater samples were located in a stirred jacketedvessel, and the temperature inside the solution was regu-lated within (0.1 K by circulating water from a thermo-static bath (Haake). Thermometers with subdivisions of(0.1 K were used to monitor both the bath and solutiontemperatures. Measurements of pH were carried out usinga combined electrode (Ingold), located into the vesselsolution, and a digital pH-meter (Radiometer). Buffersolutions of pH 1.68 ( 0.01 (Radiometer), 4.00 ( 0.02(Merck), and 7.00 ( 0.02 (Merck) were used for theinstrument calibration.

Before starting the measurements, the vessel of about0.5 L of capacity was filled with deionized water and theinstallation was purged with nitrogen until steady-stateconditions were obtained for temperatures and gas flowrate. Once steady-state conditions were achieved, the vesselwas emptied, dried with the nitrogen flow, and filled againwith 0.350 L of a fresh seawater sample. An additional timewas required to stabilize temperatures again. During thistime and through each experimental run, the nitrogen flowwas saturated with water at the operation temperature,and pH and temperature of the seawater sample werecontinuously monitored. When a constant temperature wasreached, a SO2 flow at the required rate was injected intothe nitrogen main flow and the SO2 + N2 mixture wasbubbled through the seawater sample.

The sulfur dioxide was absorbed until saturation condi-tions were attained. At that point, the pH of the water

leveled off at a constant value and the SO2 concentrationin the gas stream was the same in the outlet and in theinlet. Once equilibrium was reached, the total sulfur (IV)concentration in the liquid was determined by adding aknown volume of seawater sample from the vessel to aknown volume of standard iodine. The excess iodine wasback titrated with the standard sodium thiosulfate. Theoverall uncertainty in the determination of the sulfur (IV)concentration was estimated to be (1%.

Discussion and Results

(a) SO2 Solubility in Seawater. Measurements at298.15 K with natural and artificial seawater allowedcomparing the total absorption capacity of both solutions.Figure 1 shows the experimental data of sulfur (IV)concentration in seawater as a function of the SO2 partialpressure, together with experimental correlations for theSO2 solubility in distilled water8 and NaCl solutions withsimilar ionic strength.9 Small differences (less than 5%) canbe observed between the total absorption capacity of thenatural and artificial seawater samples. However, the SO2

solubility in artificial seawater increases within (20-60)%in relation to distilled water and within (6-30)% in relationto NaCl solutions. The higher differences are observed atthe lower partial pressures of SO2.

To establish the effect of temperature, measurements at278.15 K and 318.15 K were also carried out with artificialseawater. Experimental data of pH and total solubility of

Table 1. Composition of the Seawater Samples

natural seawater: pH ) 7.79;S ) 36.87 g‚kg-1; I ) 0.775 mol‚kg-1

artificial seawater: pH ) 7.92;S ) 35.00 g‚kg-1; I ) 0.732 mol‚kg-1

ion Ci/mol‚L-1 mi/mol‚kg-1 mi/mol‚kg-1 Ci/mol‚L-1

Cl- 0.5628 0.5700 0.5600 0.5663SO4

2- 0.0306 0.0310 0.0291 0.0288HCO3

- 0.00326 0.00330 0.00282 0.00279Br- 0.00085 0.00084F- 0.00042 0.00043 0.00007 0.00007Na+ 0.5061 0.5126 0.4857 0.4803Mg2+ 0.0640 0.0648 0.0553 0.0555K+ 0.0129 0.0131 0.00927 0.00917Ca2+ 0.0116 0.0117 0.0154 0.0152Sr2+ 0.00016 0.00016B 0.00056 0.0057 0.00045 0.00045

Figure 1. Solubility of sulfur dioxide in aqueous solutions at298.15 K: dotted line, distilled water; dot-dashed line, 0.75mol‚kg-1 NaCl solution; 2, artificial seawater; b, natural seawater.


Página 137

SO2 were correlated as functions of temperature and partialpressure of SO2 as follows

The R2 values for the above correlations were 0.991 and0.998, respectively; the root-mean-square deviations fromthe fitted curves were 0.036 and 0.0010 mol‚kg-1, respec-tively (Figure 2). The validity ranges are a salinity of 35.00g‚kg-1, a temperature between (278.15 and 318.15) K, anda SO2 partial pressure between (0.050 and 1.5) kPa. Both

equations can be combined to derive the following correla-tion between total SO2 solubility, pH, and temperature

The validity ranges are a salinity of 35.00 g‚kg-1, a pHbetween (1.5 and 3.0), and a temperature between (278.15and 318.15) K.

As it was mentioned in the Introduction, experimentaldata for the SO2 + seawater system are scarce in theliterature.1,2,4,5 Data from Bromley and Read,1 as werereported by Abdulsattar et al.,2 provide equilibrium valuesof total solubility and pH at 12.8 °C and 33.9 g‚kg-1 insalinity but do not provide values of partial pressure of SO2.Data from Douabul and Riley4 and Al-Enezi et al.5 aregiven only at a single SO2 pressure of 101.3 kPa and0.022431 kPa, respectively. Al-Enezi et al. correlated theirexperimental results in the temperature range between(283.15 to 313.15) K and a salinity range between (0 to65.1) g‚kg1 by a function of the form

where S is salinity. The original data were given in molarconcentrations, and the fitting constants have been recal-culated using the seawater density, by means of theinternational equation of state of seawater.10 Values of theconstants for molal concentrations are shown in Table 2.Solubility data of Douabul and Riley were obtained in thetemperature range between (278.15 and 303.15) K and asalinity range between (0 and 40) g‚kg-1. They have beenalso correlated by a function of the form as eq 4, showingTable 2 the set of fitting constants.

Table 3 compares, at 285.95 K, solubility calculated byeq 3 and experimental data from Bromley and Read1,2 andAl-Enezi et al.5 in the pH range between (1.5 and 3.0). Datafrom Bromley and Read are close to those calculated by eq3, but a significant lower solubility is observed for thedatum at pH 2.55, from Al-Enezi et al., compared both withthe calculated value and the experimental value at pH 2.5.This low solubility was observed even by the authorsthemselves in their own solubility data in distilled water.

Figure 2. Comparison of experimental and calculated values oftotal solubility, mS(IV), and pH in artificial seawater: b, 278.15 K;2, 298.15 K; 9, 318.15 K; dotted line, diagonal (5% deviation.

pH ) A(T) + 9.6876ln(pSO2

/Pa)

A(T) ) 5.589 - 1526(T/K)

(1)

mS(IV)/mol‚kg-1 ) B(T)(pSO2/Pa) + C(T)(pSO2/Pa)1/2

1B(T)

) 1.10847 × 106 - 2.9917 × 108

(T/K)

C(T) ) 5.2302 × 10-3 - 1.51407 × 10-5(T/K) (2)

Table 2. Fitting Constants of Eq 4, a1, a2, a3, a4, and a5,to Correlate Literature Data of SO2 Solubility inSeawater

constants Douabul and Riley4 Al-Enezi et al.5

a1 148.23507 2.669204 × 10-2

a2 -0.941699 -1.607938 × 10-4

a3 3.314327 × 10-4 1.045986 × 10-5

a4 1.50622 × 10-3 2.445612 × 10-7

a5 4.525683 × 10-6 1.295351 × 10-7

validityrange:

pSO2 ) 101.3 kPa;pH ) 0.8;T ) (278.15-303.15) K;S ) (0 to 40) g‚kg-1

pSO2 ) 0.022431 kPa;pH ) (2.5-2.6);T ) (283.15-313.15) K;S ) (0 to 40) g‚kg-1

mS(IV)/mol‚kg-1 ) A(T) exp[ 9.6876pH - C(T)] +

B(T) exp[ 4.8438pH - C(T)] (3)

1A(T)

) 1.10847 × 106 - 2.9917 × 108

(T/K)

B(T) ) 5.2302 × 10-3 - 1.51407 × 10-5(T/K)

C(T) ) 5.589 - 1526(T/K)

mS(IV)/mol‚kg-1 ) a1 + a2(T/K) + a3(S/g‚kg-1) +

a4(T/K)2 + a5(S/g‚kg-1)2 (4)


Página 138

(b) Thermodynamic Models. A simplified chemicalmodel, proposed by Abdulsattar et al.2 and used by Al-Eneziet al.,5 has been applied to predict the equilibrium solubilityof sulfur dioxide in seawater in the temperature rangebetween (278.15 and 318.15) K. The model considers thepresence of eight species (apart from water): SO2 (g), SO2

(aq), HSO3-, SO3

2-, HSO4-, SO4

2-, H+, and OH-. Thefollowing equilibria are involved

Here, Hm° is the Henry’s law constant in molal unit, φSO2

is the fugacity coefficient, K is the equilibrium constant, γis the activity coefficient, m is the molal concentration, andaH2O is the water activity. Since this study concerns onlylow pressures, the fugacity coefficient can be set equal tounity.

Chemical equilibria involving inorganic carbon species(CO2, HCO3

-, and CO32-) have been omitted since the low

pH values are dominant at equilibrium and CO2 would bedesorbed from the liquid phase. As no CO2(g) is present inthe gas, which is composed only by N2 and SO2, thedesorbed CO2 will be removed from the installation andinorganic carbon species are assumed not to be present inthe equilibrated system.

Table 4 reports the selected expressions for the variationof Hm° and the equilibrium constants with temperature.The correlations for Hm° and K1 were established previ-ously8 under the same experimental conditions as the

present work and fits well with data from Goldberg andParker11 and Siddiqi et al.12 The correlation for K2

11 isfrequently used as a reference correlation for the H2SO3

second dissociation constant.13 The correlation for K314 fits

also very well with literature data.15

The model uses pSO2 as an independent variable. Then,7 concentrations and 7 activity coefficients are unknown,and 14 independent equations are needed to solve thesystem for each temperature. The equilibrium relation-ships, eqs 5-9, are completed with seven equations for theactivity coefficients and two ionic balances:

Total S(VI) balance

Hydrogen balance

Here, mS(VI)° and pH° are the initial concentration of SO42-

and the initial pH of the seawater, respectively.The activity coefficient for SO2 (aq) has been estimated

by the equation proposed by Schumpe,17 considering thefugacity coefficient of SO2 equal to unity

Here, hi is the specific parameter of the ion i, hSO2 is thegas-specific parameter of the SO2, and Ci is the molarconcentration of the ion i. Values for the different hi at298.15 K were taken from Weisenberger and Schumpe;18

the value for hSO2 at 298.15 K was proposed in a previouswork;9 those data are shown in Table 5. The temperatureinfluence on the specific parameters was considered as gasspecific, assuming a linear function in the range of (283.15to 363.15) K given by18

where hSO2° is the value at 298.15 K.Activity coefficients for the ionic species involved in eqs

6-9 have been evaluated by two different sets of equations:(1) Bromley’s equations.3 This model considers an ex-

tended version of the Debye-Huckel theory to estimate thesingle ion activity coefficients as a function of the ionicstrength, I, and the liquid-phase composition. Ionic strengthis defined as

Table 3. Comparison between the SO2 Solubility inSeawater Calculated by Eq 3, ms(IV) (Eq 3), andExperimental Data from Literature, mS(IV) (exp), at285.95 K

pH mS(IV) (eq 3)/mol‚kg-1 mS(IV) (exp)/mol‚kg-1 ref

1.5 0.082 0.07 1, 22.0 0.019 0.025 1, 22.5 0.0090 0.0096 1, 22.55 0.0085 0.0012 52.7 0.0074 0.0067 1, 23.0 0.0056 0.0046 1, 2

Table 4. Selected Expressions for the Henry’s LawConstant, Hm°, and the Equilibrium Constants K1, K2, K3,and K4, from T ) 278.15 K to 318.15 K

constants expression ref

Hm°/kPa‚mol-1‚kg ) exp [14.642 - 3058/(T/K)] 8K1/mol‚kg-1 ) exp [2335/(T/K) - 12.319] 8K2 /mol‚kg-1 ) exp [-358.57 + 5477.1/(T/K) +

65.31 ln(T/K) - 0.1624(T/K)]11

K3/mol‚kg-1 ) exp [2825.27(T/K) - 14.0321] 14K4/mol‚kg-1 ) exp [(-6723.67/(T/K) - 9.72048)] 16

SO2(g) T SO2(aq) Hm° )φSO2

pSO2

γSO2mSO2

(5)

SO2(aq) + H2O T H+ + HSO3-

K1 )γH+γHSO3

-

γSO2aH2O

mH+mHSO3-

mSO2(6)

HSO3- T H+ + SO3

2- K2 )γH+γSO3

2-

γHSO3-

mH+mSO32-

mHSO3-

(7)

HSO4- T H+ + SO4

2- K3 )γH+γSO4

2-

γHSO4-

mH+mSO42-

mHSO4-

(8)

H2O T H+ + OH- Kw )γH+γOH-

aH2OmH+mOH- (9)

Table 5. Schumpe (h) and Bromley (B) SpecificParameters at 298.15 K

ion or gas hi/L‚mol-1 Bi/kg‚mol-1

Cl- 0.0318 0.156SO4

2- 0.1117 -0.009HSO3

- 0.0549 -0.013SO3

2- 0.127 -0.087HSO4

- -0.013OH- 0.0839H+ 0 0.087Na+ 0.1143 -0.035K+ 0.0922 -0.087Ca2+ 0.1762 -0.035Mg2+ 0.1694 0.043SO2 -0.0607

mS(VI)° ) mHSO4- + mSO4

2- (10)

mH+ )10-pH°

γH++ mHSO3

- + 2mSO32- - mHSO4

- (11)

log γSO2) ∑

i

(hi + hSO2)Ci (12)

(hSO2- hSO2

°)/m3‚kmol-1 ) 0.275 × 10-3(T/K - 298.15)(13)

I )1

2∑

i

mizi2 (14)


Página 139

Bromley equations are for a generic anion X

and for a generic cation M

Subscripts X and M denote anions and cations, respectively,z is the ion charge, Aγ is the constant of the Debye-Huckelequation, and the terms Bi are individual ion parameters.

For water as solvent, temperature dependence of Aγ inthe range between (273.15 to 328.15) K is given by19

where Aφ is the Debye-Huckel constant for the osmoticcoefficient.

Bi values2 at 298.15 K are also given in Table 5. Valuesat other temperatures were estimated by assuming thatvariations in individual Bi values would follow the varia-tions of overall B for the sea salt as a whole.2 Bromley etal.20 reported the temperature dependence of Bseasalt. Fromthis information, the following temperature dependence forBi has been derived

(2) Pitzer’s equations.21-23 This model is used frequentlyfor the prediction of activity coefficients in multielectrolytesolutions with ionic strengths between (0-6) mol‚kg-1.Equations for single ion activity coefficients and for thewater activity are derived from the formulation of the totalexcess Gibbs energy of the solution.

For an anion X

For a cation M

For the water activity

where MH2O is the molecular mass of water (0.018 kg‚mol-1).In the above equations, the third virial terms for neutral

solutes are omitted. The subscripts a, a′ and c, c′ denoteanions and cations different to X and M, respectively; thesubscript n denotes neutral solutes; a < a’ and c < c′indicate that the sums are over the various anions a, a′and cations c,c′, respectively.

The quantity F includes the Debye-Huckel’s parameterfor the osmotic coefficient, Aφ, and other terms as follows

Here, b is a constant with the value 1.2 kg1/2‚mol-1/2 andB′ and Φ′ are the ionic strength derivatives of B and Φ,respectively.

The quantity E is given by

The terms CMX are related to the empirical parametersCMX

φ by the expression

The terms B and B′ include the specific parameters of theelectrolyte MX, �MX

(0), �MX(1), and �MX

(2), and depend on theionic strength as follows

The functions g and g′ are given by

where the independent variable x is RI1/2. For electrolyteswith one monovalent ion, at 298.15 K, R1 and R2 take thevalues 2.0 kg1/2‚mol-1/2 and 0, respectively; for 2-2 electro-lytes at 298.15 K, the optimum values are (1.4 and 12)kg1/2‚mol-1/2, respectively. For many applications, those

log γX ) -AγzX

2I1/2

1 + I1/2+ BX ∑

M

mM + ∑M

BMmM (15)

log γM ) -AγzM

2I1/2

1 + I1/2+ BM ∑

X

mX + ∑X

BXmX (16)

Aφ/kg1/2mol-1/2 )Aγ

3) 1.134 + 1.4052 × 10-3(T/K -

273.15) + 1.122 × 10-5(T/K - 273.15)2 (17)

Bi(T/K) ) Bseasalt(T/K) - [Bseasalt(298.15 K) -Bi(298.15 K)]

Bseasalt(T/K) ) 1.94357 + 0.031102 ln(1 - 243T/K) -

78.565T/K

- 0.27491 ln(T/K) (18)

ln γX ) zX2F + ∑

c

2mc(BcX + ECcX) + ∑a

ma(2ΦXa +

∑c

mcΨcXa) + ∑c

∑<c′

mcmc′Ψcc′X +

|zX| ∑c

∑a

mcmaCca + 2∑n

mnλnX + ... (19)

ln γM ) zM2F + ∑

a

2ma(BMa + ECMa) + ∑c

mc(2ΦMc +

∑a

maΨMca) + ∑a

∑<a′

mama′ΨMaa′ +

|zM| ∑c

∑a

mcmaCca + 2 ∑n

mnλnM + ... (20)

ln aH2O ) MH2O{ 2AφI3/2

1 + bI1/2- 2[∑

c∑

a

mcma(Bcaφ +

2ECca) + ∑c

∑<c′

mcmc′(Φcc′φ + ∑

a

maΨcc′a) +

∑a

∑<a′

mama′(Φaa′φ + ∑

c

mcΨcaa′) + ∑n

∑c

mnmcλnc +

∑n

∑a

mnmaλna + ....] - ∑i

mi} (21)

F ) -Aφ[ �I

1 + b�I+ (2b) ln(1 + b�I)] +

∑c

∑a

mcmaB′ca + ∑c

∑<c′

mcmc′Φ′cc′ + ∑a

∑<a′

mama′Φ′aa′

(22)

E )1

2∑

i

mi|zi| (23)

CMX )CMX

φ

2|z +z -|1/2(24)

BMXφ ) �MX

(0) + �MX(1) exp(-R1I

1/2) + �MX(2) exp(-R2I

1/2)

BMX ) �MX(0) + �MX

(1)g(R1I1/2) + �MX

(2)g(R2I1/2)

BMX′ )�MX

(1)g′(R1I1/2) + �MX

(2)g′(R2I1/2)

I(25)

g(x) )2[1 - (1 + x) exp( - x)]

x2

g′(x) )-2[1 - (1 + x + x2/2) exp( - x)]

x2(26)


Página 140

values may be assumed as independent of pressure andtemperature.

The terms Φ, Φ′, and Ψ take into account interactionsbetween ions of the same sign, which arise only formultielectrolyte solutions. For solubility calculations, thefollowing considerations are usually assumed

where θij, arising from short-range forces, is taken as aconstant for any particular anion a, a′ or cation c, c′ at agiven pressure and temperature.

Thus, the Pitzer’s ion-interaction model provides anexpression for the activity coefficients in multielectrolyte

solutions in terms of seven types of empirical parameters:�MX

(0), �MX(1), �MX

(2), CMXφ, θij, Ψijk, and λni. No data for all

those parameters are available in the literature, particu-larly the Cφ , θ, Ψ, and λ parameters. Tables 6 and 7 listdata found at 298.15 K19,21-24 for solutes involved inequilibria 6-9. Temperature dependencies are only avail-able for some of those parameters.

Previous calculations have found a relative insensitivityof the calculated activity coefficients to moderate changesin �(0) and �(1) by temperature. Also Rosenblatt19 estimatedthat the only temperature-dependent quantity that changessignificantly in the range (273.15 to 323.15) K is Aφ, givenby eq 17. Thus, in the studied temperature range, notemperature dependency of the interaction parameters hasbeen considered. However, temperature effects should betaken into account to achieve significantly greater accuracythan the present work estimates, particularly at a tem-perature around 323.15 K or higher.

Model calculations for each temperature were carried outassuming an initial value of ionic strength, I, and hydrogenconcentration, mH. Then, values of species concentrationsand activity coefficients were calculated from the equationsystem by an iterative procedure with two nested loops.In the inner loop, a new value of mH was generated fromeq 11 until convergence within (0.1%. In the outer loop, anew value of I was generated from eq 14 until convergencewithin (0.1%.

Calculations of SO2 solubility were made at differenttemperatures using both Bromley and Pitzer activitycoefficient equations. Water activity was calculated by eq21 when Pitzer’s equations were used, while it was setequal to unity when Bromley’s equations were used. Figure3 shows the experimental data together with the modelpredictions. As it can be observed, both Bromley and Pitzerequations lead to similar predictions. Experimental resultsat 298.15 K are fairly well represented by the models.Results at 318.15 K show more differences between ex-perimental data and theoretical values. Those differencescould be attributed to both the large decrease in the gassolubility and the assumptions used for taking into accountthe temperature effect on the specific interaction param-eters.

Conclusions

The solubility of SO2 in seawater, from mixtures of SO2

+ N2, has been studied experimentally in the temperaturerange between (278.15 and 318.15) K and a SO2 partial

Table 6. Available Single Electrolyte SolutionParameters �MX

(0), �MX(1), �MX

(2), and CMXO for the SO2 +

Seawater System at 298.15 K

cation M anion X �MX(0) �MX

(1) �MX(2) CMX

φ ref

H+ HSO3- 0.15 0.4 19

H+ HSO4- 0.2065 0.5556 22, 23

H+ SO42- 0.0298 0.0438 22, 23

H+ Cl- 0.1775 0.2945 0.0008 22, 23Na+ HSO3

- 0.0249 0.2455 0.0004 22, 23Na+ SO3

2- 0.021 1 19Na+ OH- 0.0864 0.253 0.0044 22, 23Na+ HSO4

- 0.0454 0.398 22, 23Na+ SO4

2- 0.0196 1.1130 0.00497 22, 23Na+ Cl- 0.0765 0.2644 0.00127 22, 23Mg2+ HSO3

- 0.49 1.804 19Mg2+ SO3

2- 0.2 3.00 -41 19Mg2+ HSO4

- 0.4746 1.729 22, 23Mg2+ SO4

2- 0.221 3.343 -37.23 0.0025 22, 23Mg2+ Cl- 0.35235 1.6815 0.00519 22, 23K+ HSO3

- -0.096 0.2481 19K+ SO3

2- 0.065 1 19K+ OH- 0.1298 0.32 0.0041 22, 23K+ HSO4

- -0.0003 0.1735 22, 23K+ SO4

2- 0.04995 0.7793 22, 23K+ Cl- 0.03 0.2122 0.0084 22, 23Ca2+ HSO3

- 0.438 1.76 19Ca2+ SO3

2- 0.18 2.38 -61.3 19Ca2+ OH- -0.1747 -0.2303 -5.72 22, 23Ca2+ HSO4

- 0.2145 2.53 22, 23Ca2+ SO4

2- 0.2 3.1973 -55.7 22, 23Ca2+ Cl- 0.3159 1.614 0.00034 22, 23

Table 7. Available Mixing Parameters, θij and ψijk,22,23

and Ion-Neutral Interaction Parameters, λni,,24 for theSO2 + Seawater System at 298.15 K

ψijk

i j θij k ) Na+ k ) K+ k ) Ca2+ k ) Mg2+ k ) H+

Cl- SO42- 0.030 0.000 -0.005 -0.002 -0.008 0.013

Cl- HSO4- -0.006 -0.006

Cl- OH- -0.05 -0.006 -0.006 -0.025SO4

2- HSO4- -0.0094 -0.0677 -0.425

SO42- OH- -0.013 -0.009 -0.05

ψijk

i j θij k ) Cl- k ) SO42- k ) HSO4

-

Na+ K+ -0.012 -0.0018 -0.010Na+ Ca2+ 0.07 -0.007 -0.055Na+ Mg2+ 0.07 -0.012 -0.015Na+ H+ 0.036 -0.004 -0.0129K+ Ca2+ 0.032 -0.025K+ Mg2+ 0 -0.022 -0.048K+ H+ 0.005 -0.011 0.0197 -0.0265Ca2+ Mg2+ 0.007 -0.012 0.024Ca2+ H+ 0.092 -0.015Mg2+ H+ 0.010 -0.011 -0.0178

n i λni

SO2 Na+ 0.0283SO2 Mg2+ 0.085SO2 Cl- 0

Φij ≈ Φijφ ≈ θij Φij′ ≈ 0 (27)

Figure 3. Comparison of experimental solubility against modelpredictions for the SO2 + seawater system: b, 278.15 K; 2, 298.15K; 9, 318.15 K; dotted line, Bromley’s model; solid line, Pitzer’smodel.


Página 141

pressure range between (0.050 and 1.5) kPa. Low partialpressure range is of interest in flue gas desulfurization byseawater scrubbing, and different temperatures are neces-sary to simulate variations in the operational conditionsof SO2 absorbers. Measurements at 298.15 K with naturaland artificial seawater allowed comparing the total absorp-tion capacity of both solutions. Small differences (less than5%) could be observed between the total absorption capacityof the natural and artificial seawater samples. However,the SO2 solubility in artificial seawater increased within(20-60)% in relation to distilled water, and within (6-30)%in relation to NaCl solutions of similar ionic strength. Toestablish the effect of temperature on solubility and pH atequilibrium, measurements at 278.15 K and 318.15 K werealso carried out with artificial seawater. Correlation equa-tions were derived as a function of the SO2 partial pressureand temperature, and comparisons with literature datawere also made. Experimental results allowed testingseveral thermodynamic models in order to predict thesolubility of SO2 in seawater at low partial pressures anddifferent temperatures. Two different models were consid-ered to evaluate activity coefficients of ionic species.Calculations were made at different temperatures usingan extended version of the Debye-Huckel theory and thePitzer ion-interaction model. Both models led to similarpredictions and were in good agreement with the experi-mental results. The reasonable fitting between measuredand calculated data show that the selected models can besuccessfully used to estimate equilibrium data for theabsorption of sulfur dioxide in seawater.

Literature Cited(1) Bromley, L. A.; Read, S. M. Removal of Sulfur Dioxide from Stack

Gases by Seawater. In Research Project S-15. University ofCalifornia Research Reports; University of California: Berkeley,CA, 1970.

(2) Abdulsattar, A. H.; Sridhar, S.; Bromley, L. A. Thermodynamicsof the Sulfur Dioxide-Seawater System. AIChE J. 1977, 23, 62-68.

(3) Bromley, L. A. Approximate Individual Values of � (or B) inExtended Debye-Huckel Theory for Uni-Univalent AqueousSolutions. J. Chem. Thermodyn. 1972, 4, 669.

(4) Douabul, A.; Riley, J. Solubility of Sulfur Dioxide in DistilledWater and Decarbonated Seawater. J. Chem. Eng. Data 1979,24, 274-276

(5) Al-Enezi. G.; Ettouney, H.; El-Dessouky, H.; Fawzi, N. Solubilityof Sulfur Dioxide in Seawater. Ind. Eng. Chem. Res. 2001, 40,1434-1441.

(6) APHA-AWWA-WPCF. Standard Methods for the Examinationof Water and Wastewater, 17th ed.; American Public HealthAssociation: Washington, DC, 1989.

(7) Riley, J. P.; Skirrow, G. Chemical Oceanography, 2nd ed.;Academic Press: London, 1975.

(8) Rodrıguez-Sevilla, J.; AÄ lvarez Dıaz, M.; Dıaz Garcıa, M. C.;Liminana de la Fe, G. Thermodynamic Equilibrium of the SO2-H2O System at Low Partial Pressures. Afinidad 2001, 492, 141-146.

(9) Rodrıguez-Sevilla, J.; AÄ lvarez, M.; Liminana, G.; Dıaz, M. C.Dilute SO2 Absorption Equilibria in Aqueous HCl and NaClSolutions at 298.15 K. J. Chem. Eng. Data 2002, 47, 1339-1345.

(10) Fofonoff, N. P.; Millard, R. C. Algorithms for Computation ofFundamental Properties of Seawater, UNESCO Technical Papersin Marine Science No. 44; UNESCO: Paris, 1983.

(11) Goldberg, R. N.; Parker, V. B. Thermodynamics of SO2 (g) inWater and Aqueous Sulfur Dioxide Solutions. J. Res. Natl. Bur.Stand. 1985, 90, 341-358.

(12) Siddiqi, M. A.; Krissmann, J.; Peters-Gerth, P.; Luckas, M.; Lucas,K. Spectrophotometric Measurementsof the Vapor-Liquid Equi-libria of (Sulfur Dioxide + Water). J. Chem Thermodyn. 1996,28, 685-700.

(13) Millero, F. J.; Hershey, P.; Johnson, G. The Solubility of SO2 andthe Dissociation of H2SO3 in NaCl Solutions. J. Atmos. Chem.1989, 8, 377-389.

(14) Pitzer, K. S.; Roy, R. N.; Silvester, L. F. Thermodynamics ofElectrolytes. 7. Sulfuric Acid. J. Am. Chem. Soc. 1977, 9, 4930-4936.

(15) Hunger, T.; Lapicque, F.; Storck, A. Thermodynamic Equilibriumof Diluted SO2 Absorption into Na2SO4 or H2SO4 ElectrolyteSolutions. J. Chem. Eng. Data 1990, 35, 453-463.

(16) CRC Handbook of Chemistry and Physics, 67th ed.; Weast, R. C.,Astle, M. J., Beyer, W. H., Eds.; CRC Press Inc.: Boca Raton,FL, 1987.

(17) Schumpe A. The Estimation of Gas Solubilities in Salt Solutions.Chem. Eng. Sci. 1993, 48, 153-158.

(18) Weisenberger, S.; Schumpe, A. Estimation of Gas Solubilities inSalt Solutions at Temperatures from 273 K to 363 K. AIChE J.1996, 42, 298-300.

(19) Rosenblatt, G. M. Estimation of Activity Coefficients in Concen-trated Sulfite-Sulfate Solutions. AIChE, J. 1981, 4, 619-626.

(20) Bromley, L. A.; Singh, D.; Ray, P.; Sridhar, S.; Read, S. M.Thermodynamic Properties of Sea Salt Solutions. AIChE J. 1974,20, 326-335.

(21) Pitzer, K. S. Thermodynamics of Electrolytes. 1. Theoretical Basisand General Equations. J. Phys. Chem. 1973, 77, 268-277.

(22) Pitzer, K. S. A Thermodynamic Model for Aqueous Solutions ofLiquid-Like Density. Rev. Mineral. 1987, 17, 97-142.

(23) Pitzer, K. S. Ion Interaction Approach: Theory and Data Cor-relation. In Activity Coefficients in Electrolyte Solutions, 2nd ed.;Pitzer, K. S., Ed.; CRC Press: Boca Raton, FL, 1991.

(24) Roy, R. N.; Zhang, J. Z.; Millero, F. J. The Ionization of SulfurousAcid in Na-Mg-Cl Solutions at 25 °C. J. Solution Chem. 1991,20, 361-373.

Received for review April 30, 2004. Accepted July 18, 2004.

JE049833L


Página 142

Environmental TechnologyVol. 30, No. 7, June 2009, 715–723

ISSN 0959-3330 print/ISSN 1479-487X online© 2009 Taylor & FrancisDOI: 10.1080/09593330902896187http://www.informaworld.com

Nitrogen transformation of reclaimed wastewater in a pipeline by oxygen injection

L.E. Rodríguez-Gómez*, M. Álvarez, J. Rodríguez-Sevilla, M.C. Marrero and A. Hernández

Department Chemical Engineering, University of La Laguna, Av. Astrof. Fco. Sánchez S/N., 38200 La Laguna, SpainTaylor and Francis (Received 29 January 2009; Accepted 15 March 2009)10.1080/09593330902896187

A study of oxygen injection was performed in a completely filled gravity pipe, which is part of the South Tenerifereclaimed wastewater reuse scheme (Spain), in order to inhibit the appearance of anaerobic conditions by anitrification-denitrification process. The pipe was 0.6 m in diameter and 62 km long and made of cast iron with aconcrete inner coating, A high-pressure oxygen injection system was installed at 16 km from the pipe inlet, wheresevere anaerobic conditions appear. Experiments on oxygen injection were carried out with three differentconcentrations (7, 15 and 30 mg l

−1 O2). In all experiments, oxygen dissolved properly after injection, and no gasescapes were detected during water transportation. Most oxygen was consumed in the nitrification process, due to thelow COD/NH4-N ratio, leading to a maximum production of oxidized nitrogen compounds of 7.5 mg l

−1 NOX-N withthe 30 mg l

−1 O2 dose. Nitrification occured with nitrite accumulation, attributed to the presence of free ammoniawithin the range 1.2–1.4 mg l

−1. Once the oxygen had been consumed, an apparent half-order denitrification tookplace, with limitation of biodegradable organic matter. The anoxic conditions led to a complete inhibition of sulphidegeneration.

Keywords: denitrification; nitrification; oxygen injection; pipeline; reclaimed wastewater

Introduction

Most of the Canary Islands (Spain) cannot meet theirwater demand with traditional water resources, due toover-exploitation of aquifers, forcing the establishmentof conservation measures, water reuse being one ofthem. The South Tenerife reclaimed wastewater reusescheme was put into operation in 1993. Water from thetwo biggest treatment plants of the island is reused inthe south of the island for the irrigation of crops (bananaplantations) and golf courses. Because of the largedistance separating Santa Cruz (site of the main waste-water treatment plant) from South Tenerife (site of themain agricultural and water demanding region of theisland), the construction of a large reclaimed wastewa-ter reuse system was necessary. One element of thissystem is a large, completely filled gravity pipe (62 km)which has been used in this study.

This pipe behaves as a biochemical reactor, where atransformation of the reclaimed wastewater compoundstakes place. Anaerobic conditions prevail during mostof the water flow, sulphide generation being one of themost important processes occurring under these circum-stances [1–4]. This process should be avoided due to thecorrosiveness related to it, malodorous problems andhigh toxicity. In the reclaimed wastewater pipe of theTenerife reuse scheme, under ordinary transport condi-

tions, sulphide generation starts after the depletion ofoxygen, leading to sulphide concentrations at the end ofthe pipe above 2 mg l–1, for an average chemical oxygendemand (COD) concentration of 65 mg l

−1, which iswithin the common COD range of the effluent of theSanta Cruz treatment plant. However, in previous stud-ies, in the same pipeline transporting water with CODabove 120 mg l

−1, concentrations of 15 mg l

−1 S2

− havebeen found [2].

Oxygen and air injection have been used for decadesto avoid the appearance of anaerobic conditions insewer systems [5–10], but few studies have been carriedout on reclaimed wastewater systems. In the studiesperformed in sewers, oxygen is consumed mainly inorganic matter transformation. In previous research onthe system under study, the feasibility of a nitrificationstep in the transportation pipeline was stated [11,12],with the beneficial effect of inhibiting anaerobic condi-tions. Hence, an injection of oxygen in the pipe wasplanned in order to provoke nitrification and maintainanoxic conditions during water transportation, so inhib-iting sulphide generation. The amount of NO3–N neces-sary to maintain anoxic conditions up to the end of thepipe should be within the range of 2.5–5 mg l

−1 [13],according to the typical reclaimed wastewater charac-teristics of the reuse system (60–90 mg l

−1 COD). In

*Corresponding author. Email: [email protected]

Downloaded By: [Rodríguez-Gómez, L. E.] At: 10:04 8 June 2009

Página 168

716 L.E. Rodríguez-Gómez et al.

addition to this, any reduction in the ammonia nitrogenof water would help diminish the chlorine requirementsfor the final reclaimed wastewater disinfection. Theinjection of oxygen in the reclaimed wastewater trans-port system reveals itself as an innovation towards theinhibition of sulphide generation, since it combines twomethods already applied: the use of oxygen injectionand the application of nitrate and/or nitrite to avoid theappearance of anaerobic conditions.

The objective of this work was to study the inhibi-tion of anaerobic conditions, during reclaimedwastewater transportation, by oxygen injection at highpressure, leading to a nitrification–denitrificationprocess.

Materials and methods

The pipe under study is part of the South Tenerifereclaimed wastewater reuse scheme. This completelyfilled gravity pipe, 0.6 m in diameter and 62 km long,made of cast iron with an inner coating of concrete, ispermanently operated with an average water flow rateof 750 m3 h

−1, corresponding to 0.74 m s

−1 water veloc-ity and 23.4 h mean retention time. Although the trans-portation is by gravity, the pipe behaves as a force main,with the pressure during transportation ranging between10 and 25 bar.

The ideal site for the installation of the oxygen injec-tion system was at 16 km from pipe inlet, for three mainreasons: a) it is the section where severe anaerobic condi-tions usually appear; b) at this site the pipe pressure isabout 23 bar, favouring the dissolution of oxygen inwater, c) the oxygen-supply company is located nearby,making it easier to monitor the system. The installationof the oxygen injection system was carried out byBALTEN (the organization in charge of management ofthe reuse scheme) and AL Air Liquide España S.A.

The injection system consists of a liquid oxygentank of 6 m3 capacity (35 bar), two atmospheric vapor-izers in parallel, pressure and flow metres, and adiffuser for the oxygen injection. All the elements arecontrolled by a Programmable Logic Controller (PLC),the oxygen flow injected being a function of the waterflow in the pipe, in order to keep a constant oxygendose. Because one of the usual problems related tooxygen injection in water mains is related to the gasdissolution, it was decided that the oxygen injectionshould be undertaken at the centre of the cross sectionof the pipe, through a porous diffuser of high efficiency,and, as a preventive measure, the first vent valve down-stream of oxygen injection was closed to avoid possibleoxygen losses.

Experiments were performed with three differentoxygen concentrations (7, 15 and 30 mg l

−1 O2).Oxygen was continuously injected for three weeks

before taking samples, in order to achieve the bestconditions for nitrification. Reclaimed wastewatersamples were collected at different points along thegravity pipeline, by using valves which are distrib-uted along it. Dissolved oxygen (DO), temperature,pH and oxidation–reduction potential (ORP) weremeasured in situ, and samples for analysis wererefrigerated and transported to the laboratory wheretotal suspended solids (TSS), total and soluble chemi-cal oxygen demand (COD, CODS), soluble organiccarbon (SOC), SO4

2

−, S2

−, NO3–N, NO2–N, NH4–N,CO3

2

− and HCO3

− were measured according to Stan-dard Methods [14]. Samples for sulphide analy-siswere pretreated after sampling, in situ, with zincacetate and NaOH, and refrigerated during transporta-tion, as well.

Results and discussion

When making this study, one matter of concern was toachieve a good oxygen dissolution and to avoid theoxygen moving along the top of the pipe due toinadequate mixing [5]. During each experiment, directobservation was carried out of the vent valves located inthe 5 km long stretch after oxygen injection, and no gasremoval was detected through the vent valves. Thismeans that the oxygen dissolved properly after injection,achieving complete oxygen dissolution. Three factorshelped oxygen to dissolve properly: 1) the high pressurein the injection site (23 bar), 2) the turbulent flow regime(Reynolds number of about 400,000), 3) the injectiondone at the centre of the pipe cross section.

Dissolved oxygen was consumed in a short sectionof the pipe (less than four hours of retention time), evenin the experiments carried out with the highest oxygendose (30 mg l

−1 O2). The oxygen consumption ratewas 5.1, 7.3 and 6.4 mg DO l h

−1 for the 7, 15 and 30mg l

−1 O2 doses, respectively. In Table 1 average char-acteristics of the reclaimed wastewater during theperiod of study are presented, measured in 12.4 km ofthe pipe. The characteristics of the reclaimed wastewa-ter during the study period fit in with the typical concen-trations of parameters of the effluent from the SantaCruz treatment plant.

Figures 1–3 show the variation in DO, NH4–N,NO2–N and NO3–N during transportation, for eachoxygen dose. With the three doses applied, a nitrifica-tion–denitrification process takes place, as can be seenin the figures. Nitrification occurs with nitrite accumu-lation, and denitrification generally is incomplete,except with the 7 mg l

−1 O2 dose, where a significantlylow concentration of oxidized nitrogen compounds (0.5mg l

−1 NOX–N) is achieved at the end of the pipeline(Figure 1). In all cases the appearance of anaerobicconditions and sulphide generation is completely inhib-


Página 169

Environmental Technology 717

ited, and two clearly defined zones can be distinguishedafter the oxygen injection: aerobic and anoxic. Thenitrification–denitrification process leads to anoverall nitrogen removal (NH4–N + NOX–N) of 7.6,11.7 and 18.0% with the oxygen doses of 7, 15 and 30mg l

−1O2, respectively. In relation to NH4–N, reduc-tions of 4.8%, 10.4% and 17.1% were achieved for theoxygen doses of 7, 15 and 30 mg l

−1, respectively,which would result in a significant reduction of the

chlorine requirement for reclaimed wastewater disin-fection before reuse.Figure 1. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection, 7 mg·l

−1 O2 dose.Figure 2. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection, 15 mg·l

−1 O2 dose.Figure 3. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection, 30 mg·l

−1 O2 dose.

Nitrification

The amount of oxidized nitrogen compounds (NOX–N)produced during nitrification is presented in Table 2,along with the relative amount of nitrite and nitrate.Nitrite accumulation occurred in all experiments, espe-cially with the 7 and 30 mg l

−1 O2 doses, with more than65% of NOX–N as nitrite.

The main factors affecting nitrification are tempera-ture, organic matter/NH4–N ratio and the presence ofinhibitors (free ammonia and nitrous acid) [15,16]. Asufficient amount of carbonate/bicarbonate is necessaryas well, which is assured by the high bicarbonate contentof the water under study (Table 1). In the experiments,the COD/NH4–N ratio was relatively low (1.8), whichresults from a low organic matter content (69 mg l

−1

COD) and a significant NH4–N content (38.4 mg l

−1).Under these circumstances the nitrification is greatlyfavoured instead of organic matter oxidation. The lowerthe organic matter/NH4–N ratio, the higher the amountof nitrifying bacteria [17].

The oxygen consumption for nitrification was esti-mated (Table 2), taking into account the production ofNOX–N, and a theoretical nitrogenous oxygen demandfor nitrification of 4.25 mg DO per mg NH4–N oxida-tion to NO3–N, and 3.22 mg DO per mg NH4–N oxida-tion to NO2–N [18,19]. The results show a considerablyhigh oxygen consumption in nitrification (92%, 76%and 89% for the 7, 15 and 30 mg l

−1 O2 doses,

�

�

�

�

�

��

�

��

��

��

��

��

��

� � ��

Figure 1. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection,7 mg·l−1 O2 dose.

Table 1. Reclaimed wastewater characteristics in 12.4 km ofpipe, prior to oxygen injection.

Parameter Reclaimed wastewater at 12.4 km

T (ºC) 23.2

± 1.0pH 7.76

± 0.06DO (mg l

−1) 0.0

± 0.0ORP (mV) –7

± 54COD (mg l

−1) 69

± 17CODS (mg l

−1) 49

± 15SOC (mg l

−1) 13.0

± 1.7TSS (mg l

−1) 11.8

± 4.6Turbidity (NTU) 10.2

± 3.4CO3

2– (mg l

−1) 0.5

± 1.4HCO3

− (mg l

−1) 775.2

± 68.1NH4-N (mg l

−1) 38.4

± 6.1NO2-N (mg l

−1) 0.6

± 0.5NO3-N (mg l

−1) 0.5

± 0.3SO4

2– (mg l−1) 65 ± 13S2– (mg l−1) 0.7 ± 0.7


Página 170


respectively), in comparison with the oxygen consump-tion in conventional sludge systems. This is in agree-ment with the low COD/NH4–N ratio which indicatedgood conditions for nitrification. The amount of nitrousacid was negligible in all cases.

The main factors affecting the nitrite accumulationare pH, temperature, presence of inhibitors (free ammo-nia and free hydroxylamine) and DO concentration

[16,20,21]. The water temperature during the experi-ments was constant and around 23.2 ºC, being withinthe optimal range for nitrification [15,16, 22] and closeto the optimal temperature for the activity of ammonia-oxidizing bacteria and the inhibition of nitrite-oxidizingbacteria, which is around 25 ºC [21].

When it comes to the presence of inhibitors, the freeammonia concentration was within the range of 1.2–1.4

�

�

�

��

��

� � ��

��

�

��

��

��

��

��

��

Figure 2. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection, 15mg·l−1 O2 dose.

� � ��

�

�

��

��

��

�

��

��

��

��

��

��

Figure 3. Variation of DO, NH4-N, NO2-N and NO3-N during reclaimed wastewater transportation after oxygen injection,30 mg·l−1 O2 dose.


Página 171


mg l−1. Free ammonia is one of the most important inhib-itors of nitrite-oxidizing bacteria activity. It has beenindicated [15,16] that free ammonia concentrationsbetween 1 and 5 mg l–1 could inhibit nitratation, but notnitritation. At pH 8.5 and 20 ºC temperature, the optimalfree ammonia concentration for maximum nitritationand minimum nitratation is about 5 mg l–1 NH3 [15].Hence, a concentration of free ammonia between 1.2and 1.4 mg l−1 has probably helped nitrite to accumulatein the oxygen injection experiments. Finally, the influ-ence of free hydroxylamine, an intermediate in theammonia-oxidizing process, on nitratation was notverified although its presence can inhibit it [23].

Dissolved oxygen concentration is an operationalvariable that is difficult to assess, since its concentrationvaries in the aerobic stretch from the initial values of 7,15 and 30 mg l–1 DO, depending on the dose applied,until absence of oxygen, at the end of the aerobicsection. It is known that by maintaining a low DO value(less than 1 mg l–1), oxygen limitation induces nitritationbut limits nitratation, which has been shown at DO/NH3

ratios of less than five [24]. This condition is achievedat DO concentrations below 6 mg l–1, for a free ammoniaconcentration between 1.2 and 1.4 mg l–1. During thisstudy, this condition could only be assured in the wholeaerobic stretch for the experiments done with the 7 mgl–1 O2 dose.

Denitrification

Once the oxygen has been consumed, a denitrificationprocess occurs (Figures 1–3). In Table 3 the initialNOX–N concentration at the anoxic start point ispresented for each oxygen dose utilized, along with theNOX–N consumption and other denitrification parame-ters. It must be noted that in all experiments, the initialNOX–N concentration was always higher than theNOX–N production in the nitrification step, presented inTable 2, due to the presence of NOX–N in reclaimedwastewater coming from the treatment plant, beforeoxygen injection, attributed to a slight nitrificationprocess at the beginning of the pipe [25].

Factors affecting denitrification are temperature,organic matter concentration, NOX–N concentration,COD/NOX–N ratio and the presence of inhibitors(HNO2) [15,26]. The concentration of HNO2 was

always negligible. The COD/NOX–N ratios and NOX–Nconsumption are presented in Table 3. In this researchthis ratio was 21, 14 and 12 for the applied oxygendoses of 7, 15 and 30 mg l–1 O2, respectively. The COD/NOX–N ratio, above which the denitrification iscomplete, is between 5 and 15 [15,24,27], which isconsistent with the almost complete denitrification forthe COD/NOX–N ratio of 21, corresponding to theoxygen dose of 7 mg l–1 (final amount of NOX–N of 0.5mg l–1). It is assumed that denitrification is completewhen the concentration of NOX–N is below 0.5 mg l–1,regarding that the reaction takes place inside thebiofilm, and that the concentration of NOX–N ismeasured in the bulk liquid. However, the applicabilityof the COD/NOX–N ratio can be hampered if the reac-tion is kinetically limited or if the organic matter quan-tified by COD has low biodegradability, which is thecase under study (reclaimed wastewater).

Anoxic organic matter reduction

The evolution of organic matter is a key factor in definingthe performance of denitrification. In order to distinguishbetween stoichiometry and kinetics limitation, an anal-ysis of the organic matter behaviour is necessary. InFigure 4 the variation in organic matter during transpor-tation is presented for the 30 mg l–1 oxygen dose, as repre-sentative of the behaviour observed with the three doses.In all experiments carried out, a significant reduction inorganic matter occurred, especially for the anoxic stretchof pipe. The pipe section where aerobic conditionsprevailed is very short (retention time less than fourhours), and the average reduction in organic matter is lowin this zone because of the small amount of oxygen avail-able for organic matter oxidation (nitrification highlyfavoured with respect to organic matter oxidation).Figure 4. Organic matter parameters performance during reclaimed wastewater transportation after oxygen injection, 30 mg·l −1 O2 dose.In Table 3 the variation in COD, CODS and TSS forthe anoxic section of the pipeline is presented for thethree oxygen doses. In general terms, the consumptionof organic matter is proportional to the oxygen doseapplied and, subsequently, proportional to the initialNOX–N concentration of the anoxic pipe stretch. MostCOD was soluble (71%), regarding the values of CODand CODS at the pipe inlet, and no variation in thepercentage of soluble organic matter was observedduring transportation. In order to estimate the amount of

Table 2. Nitrification data.

O2 doseNOX-N produced by O2 injection (mg l−1) % NO2-N % NO3-N

Free ammonia (mg l−1)

% oxygen consumed in nitrification

7 mg l–1 1.8 66.7 33.3 1.4 9215 mg l–1 3.0 43.3 56.7 1.2 7630 mg l–1 7.5 65.3 34.7 1.3 89


Página 172


COD consumed in denitrification an average stoichio-metric consumption ratio of 7 g COD/g NO3–N hasbeen used [28], which is within the range of data foundin literature (8.7 g COD/g NO3–N [18], 6 g solubleCOD/g NO3–N [29], 7.7 g COD/g NO3–N [30]). On theother hand, for the estimation of the COD consumed inthe nitrite reduction it has been taken into account thatthe denitrification via nitrite leads to a 40% reductionin the demand of carbon source [31].

The amount of COD consumed in denitrificationmore consistently corresponds to the soluble fraction ofCOD when readily biodegradable organic matter is avail-able [32]. In this pipe stretch, the experimental consump-tion of CODS is far below the theoretical COD needed fordenitrification in all cases. Since the removal of CODS

corresponds to only a part of the theoretical consumptionof COD based on the nitrate and nitrite removal, it may beconcluded that the other part is taken from particulate

compounds (the COD removal is higher than thetheoretical COD consumption, in all cases, except withthe 15 mg l−1 O2 dose, where the experimental CODreduction is slightly below the theoretical one). A netremoval of total COD higher than the net removal ofCODS indicates a high particulate organic matterconsumption [32]. The TSS depletion in the anoxic pipesection (Table 3, Figure 4) confirms this assumption.These results reveal that the biodegradable fraction oforganic matter is small (reclaimed wastewater), althoughmost organic matter is in soluble form (71%). Taking intoaccount that there was not enough biodegradable matter,denitrification was stoichiometrically limited.

Denitrification kinetics

The inside wall of the pipe of the study was coveredby a biofilm layer, which provides a far higher cell

Figure 4. Organic matter parameters performance during reclaimed wastewater transportation after oxygen injection,30 mg·l−1 O2 dose.

Table 3. Denitrification data.

Oxygen dose

7 mg l–1 15 mg l–1 30 mg l–1

Initial NOX-N conc. of the anoxic stretch (mg l–1) 3.0 3.7 7.9NOX-N consumption (mg l–1) 2.5 2.9 4.5COD/NOX-N ratio 21 14 12Theoretical COD consumption in denitrification (mg·l-1) 11 17 25Experimentalconsumption of organic matter

COD (mg l–1) 21 14 32CODS (mg l–1) 5 7 19TSS (mg l–1) 5.8 7.8 11.1


Página 173


concentration than the low concentration of suspendedsolids (5–15 mg l–1 volatile suspended solids, VSS).Hence, biological processes are assumed to take placein the biofilm rather than in the bulk [11]. Regardingthat NOX–N reduction occurred in the biofilm, the rateshould be first, half or zero order depending on thecombination of mass transfer, transport and reaction inthe biofilm [22,33,34]. In addition to this, otherresearch has indicated that denitrification rate can alsodepend on the availability of soluble organic matter[35]. An apparent half-order kinetics, in relation toNOX–N is obtained, when plotting (g NOX–N m−2)0.5

versus retention time (surface rates), for the threeoxygen doses utilized (Figure 5), at the same tempera-ture (23.2 ºC), which leads to three parallel straightlines, providing a half-order rate constant k1/2 = 0.066(g NOX–N)0.5 m–1 h–1. This is in agreement with otherstudies where it has been stated that at over 1 mg l–1

the denitrification order is 0.5, and for NOX–N concen-trations below 1 mg l–1 it is first order [24] Also, arecent study [35] indicated that an apparent half-orderkinetics for denitrification in sewers can be a result ofbiofilm biomass concentration decrease along thepipe. This half-order denitrification kinetics obtained inthis study contrasts with a first order found in a nitratedosage study carried out in the same pipeline [13]. Thisdifference may be related to the amount of NOX

–

species available in each case (a mixture of NO2– and

NO3– with the oxygen injection experiments and only

NO3– with the nitrate dosage) and to different biofilm

characteristics, since the duration of the experimenta-

tion with oxygen was larger than that with the nitratedosage.Figure 5. Apparent half order denitrification kinetics with the three O 2 doses applied ((NOx-N)i is the initial NOx-N concentration, at the denitrification start point).

Enhancement of organic matter quality of reclaimed water

Table 4 presents average values for reduction of organicmatter parameters during the ordinary transport condi-tions (without oxygen injection) and during the experi-ments of oxygen injection (in the anoxic stretch).Reduction of organic matter during the ordinary trans-port conditions correspond to the period between 1994and 1999 (26 values). It is evident that during the ordi-nary operation of the system the average consumptionof organic matter is lower than with the oxygen injec-tion operation (14.6% COD, 13.0% CODS, 33.7%TSS). With the oxygen injection the organic matterconsumption increases significantly achieving reduc-tions within the range of 26–38% for COD and 50–60%for TSS.

It must be noted that, during the ordinary operationof reclaimed wastewater transportation, the organicmatter reduction is produced in the aerobic pipesection, since no anoxic zone exists, and in the anaero-bic zone no significant organic matter consumptionoccurs in the pipeline [36]. In contrast, during theoxygen injection experiments, the organic matterconsumption is carried out mainly in the anoxicstretch, as previously explained. Significantly lowvalues of COD and TSS are achieved at the end of thepipe during the experimentation with oxygen (COD

��

��

��

��

� � ��

��

Figure 5. Apparent half order denitrification kinetics with the three O2 doses applied ((NOx-N)i is the initial NOx-N concen-tration, at the denitrification start point).


Página 174


and TSS final concentrations of 40–50 mg l–1 and 6–11mg l–1, respectively), compared with the valuesmeasured during the ordinary operation of the pipeline(Table 4). This result if of great interest for thereclaimed wastewater reuse managers, since it reducesthe cost of the advanced treatments to be applied tothe reclaimed wastewater before its final reuse in irri-gation.

Conclusions

(1) Oxygen dissolved completely after injection inthe reclaimed wastewater pipe, and no gasescapes were detected during water transporta-tion.

(2) Most oxygen (76–92%) was consumed in thenitrification process due to the low COD/NH4–N ratio.

(3) Nitrification occurred with nitrite accumulation,attributed to the presence of free ammonia in therange of 1.2–1.4 mg l−1.

(4) A complete inhibition of sulphide generationwas achieved by the presence of NOX–N inreclaimed wastewater.

(5) NOX–N was reduced in the pipe with a half-orderkinetics (k1/2 = 0.066 (g NOX–N)0.5·m−1 h−1).

(6) Denitrification was limited by the small biode-gradable fraction of organic matter.

AcknowledgementsThis work was funded by the PROFIT Program, SpanishMinistry for Science and Technology, project FIT-140100-

2003-178. We want to express our gratitude to BALTEN(Tenerife Water and Wastewater Reservoirs Agency) and ALAir Liquide España, S.A., for their participation and continu-ous cooperation throughout this study.

References[1] A.G. Boon, Septicity in sewers:Causes, consequences

and containment, Water Sci. Technol. 31 (1995),pp. 237–253.

[2] S. Delgado, M. Álvarez, E. Aguiar, and L.E.Rodríguez-Gómez, H2S generation in a reclaimedurban wastewater pipe, Case study: Tenerife (Spain),Water Res. 33 (1999), pp. 539–547.

[3] P.H. Nielsen and T. Hvitved-Jacobsen, Effect of sulfateand organic matter on the hydrogen sulfide formationin biofilms of filled sanitary sewers, J. Water Pollut.Control Fed. 60 (1988), pp. 627–634.

[4] R.D. Pomeroy and J.D. Parkhurst, The forecasting ofsulfide build-up rates in sewers, Prog. Water Technol. 9(1977), pp. 621–628.

[5] A.G. Boon, C.F. Skellet, S. Newcombe, J.G. Jones, andC.F. Forster, The use of oxygen to treat sewage in a risingmain, Water Pollut. Control 76 (1977), pp. 98–112.

[6] A.G. Boon and A.R. Lister, Formation of sulphide inrising main sewers and its prevention by injection ofoxygen, Prog. Water Technol. 7 (1975), pp. 289–300.

[7] G.H. Chen and D.H.-W. Leung, Utilization of oxygenin a sanitary gravity sewer, Water Res. 34 (2000),pp. 3813–3821.

[8] S. Newcombe, C.F. Skellett, and A.G. Boon, Anappraisal of the use of oxygen to treat sewage in a risingmain, Water Pollut. Control 78 (1979), pp. 474–504.

[9] T. Ochi, M. Kitakawa, and S. Tanaka, Controllingsulfide generation in force mains by air injection,Water Sci. Technol. 37 (1998), pp. 87–95.

[10] N. Tanaka and K. Takenaka, Control of hydrogensulfide and degradation of organic matter by air injec-tion into a wastewater force main, Water Sci. Technol.31 (1995), pp. 273–282.

Table 4. Comparison of the organic matter consumption with the ordinary transport conditions and with the experimentationwith oxygen injection.

Without oxygen With the oxygen injection (anoxic stretch*)

1995–99 7 mg l–1 O2 15 mg l–1 O2 30 mg l–1 O2

COD Pipe inlet conc. (mg l–1) 82 60 54 85Pipe outlet conc. (mg l–1) 70 39 40 53Reduction (mg l–1) 12 21 14 32% reduction 14.6 35.0 25.9 37.6

CODS Pipe inlet conc. (mg l–1) 69 43 37 58Pipe outlet conc. (mg l–1) 60 38 30 39Reduction (mg l–1) 9.0 5 7 19% reduction 13.0 11.6 18.9 32.8

TSS Pipe inlet conc. (mg l–1) 17.5 9.0 11.0 21.0Pipe outlet conc. (mg l–1) 11.6 4.2 4.2 9.9Reduction (mg l–1) 5.9 5.8 7.8 11.1% reduction 33.7 53.3 61.8 52.9

*With the oxygen injection, the beginning of the anoxic stretch was considered to be the pipe inlet.


Página 175


[11] S. Delgado, M. Alvarez, E. Aguiar, and L.E.Rodríguez-Gómez, Effect of dissolved oxygen inreclaimed wastewater transformation during transpor-tation. Case study: Tenerife, Spain, Water Sci. Technol.37 (1998), pp. 123–130.

[12] S. Delgado, M. Alvarez, L.E. Rodríguez-Gómez, S.Elmaleh, and E. Aguiar, How partial nitrification couldimprove reclaimed wastewater transport in long pipes,Water Sci. Technol. 43 (2001), pp. 133–138.

[13] L.E. Rodríguez-Gómez, S. Delgado, M. Alvarez, andS. Elmaleh, Inhibition of sulfide generation in areclaimed wastewater pipe by nitrate dosage and deni-trification kinetics, Water Environ. Res. 77 (2005), pp.193–198.

[14] APHA, AWWA, and WPCF, Standard Methods for theExamination of Water and Wastewater, 17th ed.,American Public Health Association, Washington, DC,1989.

[15] U. Abeling and C.F. Seyfried, Anaerobic-aerobic treat-ment of high strength ammonium wastewater – nitrogenremoval via nitrite, Water Sci. Technol. 26 (1992), pp.1007–1015.

[16] B. Balmelle, K.M. Nguyen, B. Capdeville, J.C. Corn-ier, and A. Deguin, Study of factors controlling nitritebuild-up in biological processes for water nitrification,Water Sci. Technol. 26 (1992), pp. 1017–1025.

[17] US Environmental Protection Agency, Process designmanual for nitrogen control, Technology TransferSeminar Publication, Office of Technology Transfer,Cincinnati, Ohio, USA, 1975.

[18] Metcalf and Eddy Inc., Wastewater Engineering:Treatment and Reuse, 4th ed., McGraw-Hill, NewYork, 2003.

[19] B. Sharma and R.C. Ahlert, Nitrification and nitrogenremoval, Water Res. 11 (1977), pp. 897–925.

[20] A.C. Anthonisen, R.C. Loehr, T.B.S. Prakasam, andE.G. Srinath, Inhibition of nitrification by ammoniaand nitrous acid, J. Water Pollut. Control Fed. 48(1976), pp. 835–852.

[21] H. Yoo, K.H. Ahn, H.J. Lee, K.H. Lee, Y.J. Kwak, andK.G. Song, Nitrogen removal from synthetic wastewa-ter by simultaneous nitrification and denitrification(SND) via nitrite in an intermittently aerated reactor,Water Res. 33 (1999), pp. 145–154.

[22] M. Henze, P. Harremoës, J. la Cour Jansen, and E.Arvin, Wastewater Treatment. Biological and Chemi-cal Processes, Springer-Verlag, Berlin, 1995.

[23] L. Yang and J.E. Alleman, Investigation of batch wisenitrite build up by an enriched nitrification culture,Water Sci. Technol. 26 (1992), pp. 997–1005.

[24] F. Çeçen and I.E. Gönenç, Nitrogen removal character-istics of nitrification and denitrification filters, WaterSci. Technol. 29 (1994), pp. 409–416.

[25] S. Delgado-Díaz, M. Álvarez, L.E. Rodríguez-Gómez,and E. Aguiar, Using oxidation-reduction potential assepticity control parameter during reclaimed wastewa-ter transportation, Water Environ. Res. 72 (2000),pp. 455–459.

[26] F. Lefevre, J.M. Audic, and B. Bujon, Automatic regu-lation of activated sludge aeration – single tank nitrifi-cation denitrification, Water Sci. Technol. 28 (1993),pp. 289–298.

[27] A. Lynga and P. Balmér, Denitrification in a non-nitri-fying activated sludge system, Water Sci. Technol. 26(1992), pp. 1097–1104.

[28] J. Patel and G. Nakhla, Interaction of denitrificationand P removal in anoxic P removal systems, Desalina-tion 201 (2006), pp. 82–99.

[29] A. Æsøy and H. Ødegaard, Denitrification in biofilmswith biologically hydrolyzed sludge as carbon source,Water Sci. Technol. 29 (1994), pp. 93–100.

[30] H. Siegrist and W. Gujer, Nitrogen removal in activatedsludge system including denitrification in secondary clar-ifiers, Water Sci. Technol. 30 (1994), pp. 101–111.

[31] J. Surmacz-Górska, A. Cichon, and K. Miksch, Nitrogenremoval from wastewater with high ammonia nitrogenconcentration via shorter nitrification and denitrifica-tion, Water Sci. Technol. 36 (1997), pp. 73–78.

[32] A. Æsøy, M. Storfjell, L. Mellgren, H. Helness, G.Thorvaldsen, H. Ødegaard, and G. Bentzen, A compar-ison of biofilm growth and water quality changes insewers with anoxic and anaerobic (septic) conditions,Water Sci. Technol. 36 (1997), pp. 303–310.

[33] S. Elmaleh and A. Grasmick, Mathematical modelsfor biological aerobic fluidized bed reactors, inMathematical Models in Biological Waste WaterTreatment, M.J. Gromiec and S.E. Jörgensen, eds.,Elsevier, Amsterdam, 1985, pp. 525–581.

[34] P. Harremoës and M. Riemer, Report on pilot scaleexperiments on down-flow filter denitrification, Report75-1, Department of Sanitary Engineering, TechnicalUniversity of Denmark, Lyngby, 1975.

[35] V.L. Mathioudakis and A. Aivasidis, Heterotrophic deni-trification kinetics in a pressurized sewer biofilm reactor,9th International Conference on Protection and Restora-tion of the Environment, Kefalonia, Greece, Universityof Patras and Stevens Institute of Technology, 2008.

[36] L.E. Rodríguez-Gómez, Reclaimed wastewaterevolution during pipeline transportation, Ph.D. diss.,University of La Laguna, 1998 (In Spanish).


Página 176

Publicaciones

Documents

Transcript of Publicaciones