Natural gas based hydrogen production with zero carbon dioxide emissions

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Natural gas based hydrogen production with zero carbon dioxide emissions Jorge A. Pena Lopez, Vasilios I. Manousiouthakis* Department of Chemical and Biomolecular Engineering, University of California, Los Angeles, 5531 Boelter Hall, Los Angeles, CA 90095-1592, USA

article info

abstract

Article history:

A novel process flowsheet is presented that co-produces hydrogen and formic acid from

Received 27 April 2011

natural gas, without emitting any carbon dioxide. The principal technologies employed in

Received in revised form

the process network include combustion, steam methane reforming (SMR), pressure swing

14 July 2011

adsorption, and formic acid production from CO2 and H2. Thermodynamic analysis

Accepted 15 July 2011

provides operating limits for the proposed process, and the use of reaction clusters leads to the synthesis of a feasible process flowsheet. Heat and power integration studies show this flowsheet to be energetically self-sufficient through the use of heat engine and heat pump

Keywords:

subnetworks. Operating cost/revenue studies, using current market prices for natural gas,

Hydrogen

hydrogen and formic acid, identify the proposed design’s operating revenue to cost ratio to

Natural gas

be 9.29.

Formic acid

Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

Energetically self-sufficient

1.

Introduction

Of all energy resources, natural gas is the resource most commonly used for hydrogen production. Typically, natural gas (which consists mainly of methane) is converted to hydrogen through steam methane reforming (SMR). In 1999, 99% of hydrogen’s world production was reported to be carried out using SMR [1,2]. This figure changed to 90% by 2001 [3], and to 50% by 2005 [4], with the remaining hydrogen production coming from oil (30%), primarily for hydroprocessing applications in oil refineries, from coal (19%) primarily for ammonia production, and the remaining (w1%) coming from electricity via water electrolysis. Aside from their potential operating cost advantages, SMR plants also require lower capital costs. An SMR plant with a capacity higher than one million Nm3 H2/day has the lowest investment cost in comparison with other hydrogen production processes [5,6].

High temperature methane and steam are fed to the SMR reactor, where three reversible reactions occur: ðgÞ

ðgÞ

CH4 þ H2 OðgÞ 4COðgÞ þ 3H2 ðgÞ

ðgÞ

COðgÞ þ H2 OðgÞ 4CO2 þ H2 ðgÞ

ðgÞ

ðR1 Þ ðR2 Þ

ðgÞ

CH4 þ 2H2 OðgÞ 4CO2 þ 4H2

ðR3 Þ

(1) (2)

(3)

The kinetics of these reactions on a Ni/MgAl2O4 catalyst have been studied by Xu and Froment [7]. The overall reaction is endothermic, thus combustion of a fuel (methane) is needed to provide the required heat. Hydrogen is first produced through the SMR reactor and then additional hydrogen is produced in a water-gas shift reactor which depletes the content of carbon monoxide in the stream exiting the SMR reactor by carrying out only reaction (R2) at a lower

* Corresponding author. Tel.: þ1 310 206 0300; fax: þ1 310 206 4107. E-mail addresses: [email protected] (J.A. Pena Lopez), [email protected] (V.I. Manousiouthakis). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.07.061

12854

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

temperature. Next, water is condensed and removed to be recycled back to the process. The hydrogen rich stream is then sent to a pressure swing adsorption (PSA) unit, to obtain hydrogen at 99.999þ% purity [8e10], through a multi-step cyclic process, in which the unit is first pressurized, then species other than hydrogen are selectively adsorbed (along with a fraction of hydrogen) on a solid bed (e.g. activated carbon, 5A zeolite) by contacting the gas with the solid, thus producing a highly purified hydrogen gas stream [9]. When the bed approaches saturation, it is depressurized, leading to desorption of species from the solid and thus to the bed’s regeneration, and to the generation of PSA waste gas. Continuous flow of product is maintained by using multiple, properly synchronized adsorption beds. The present work puts forward the position that “hydrogen can be produced from fossil fuels without generation of carbon dioxide emissions”. In 2006, Posada & Manousiouthakis [11] set a precedent with an integrated steam methane reforming process which generated hydrogen and carbon dioxide in dry ice form so it can be sequestered at the bottom of the ocean [12]. The next level proposes that carbon contained in the fossil fuel will not be emitted as carbon dioxide, but rather in the form of a valuable carbon econtaining chemical. This approach will not only eliminate carbon dioxide emissions from hydrogen production plants, but will also enhance the economics of these plants since they will no longer be generating a byproduct that needs to be dealt with at some cost, but rather will be producing a valuable co-product that can be sold, thus contributing to a reduced (or possibly even negative) cost for hydrogen. This co-production of hydrogen with valuable carbonecontaining chemicals from a fossil fuel energy source holds therefore the promise of producing “clean” hydrogen at reduced cost, thus accelerating the onset of a hydrogen carrier transportation era. The particular fossil fuel and co-product considered in this work are methane (natural gas) and formic acid respectively. The industrial uses of formic acid and its global and regional production capacity can be found in Refs. [13,14]. The remainder of this work is structured as follows: First, a novel hydrogen producing reaction scheme is proposed, that utilizes natural gas as both a raw material and an energy source and does not emit any carbon dioxide. Then the concept of energetic self-sufficiency, for an open thermodynamic system is introduced. Then, using the 1st and 2nd laws of thermodynamics, an exact thermodynamic analysis is carried out that identifies a range of values of the water/methane molar ratio (X ) over which the proposed hydrogen production scheme can be energetically selfsufficient. Subsequently, pinch analysis is carried out to estimate the energetic self-sufficiency of a process realization for the proposed reaction scheme, which utilizes steam methane reforming, natural gas combustion, reverse watergas shift, and hydrogen/carbon dioxide to formic acid technologies. Having established thermodynamic feasibility, a flowsheet is proposed, which co-produces hydrogen and formic acid, does not emit carbon dioxide, and consists only of available technologies. The flowsheet’s operating conditions have a significant dependence on the aforementioned ratio X. Detailed heat and power integration

studies are thus carried out to identify the exact values of X corresponding to an energetically self-sufficient flowsheet. Finally, conclusions are drawn. A detailed description of the notation used in this paper is included in a separate section.

2. Thermodynamic feasibility of proposed hydrogen production method First, the concept of energetic self-sufficiency is introduced. Let U be a steady-state open system with inlets in SI; outlets in SO; heat transfer rates Q_ j j˛SQ entering ðQ_ j > 0Þ or exiting ðQ_ j < 0Þ the system at temperatures Ts;j j˛SQ ; and shaft work _ s;j > 0 or exiting _ s;j j˛SW entering (consumed by) W rates W _ (produced by) Ws;j < 0 the system. This system has to satisfy the overall mass, element, energy, and entropy conservation laws. These latter two laws are often referred to as the 1st and 2nd laws of thermodynamics, which at steady-state require that the energy entering the system must equal the energy exiting the system, and that the system’s rate of entropy generation is non-negative. Under the assumption of steadystate and neglecting the effects of kinetic- and potentialenergy, the above conservation laws are represented by Eqs. (4) through (7) [15]: 0¼

X

_i m

i˛SI

0¼

X

XX

nj;k

i˛SI j˛SC

0¼

X

X i˛SI

(4)

X X xi;j xi;j _i _i m nj;k m Mj Mj i˛S j˛S

_i Hi m

i˛SI

0¼

_i m

i˛SO

O

X

_iþ Hi m

i˛SO

_i Si m

X

X j˛SQ

_iþ Si m

i˛SO

ck˛SE

(5)

C

Q_ j þ

X

_ s;j W

(6)

j˛SW

X Q_ j þ S_ G ; Ts;j j˛S

S_ G  0

(7)

Q

where, SC and SE are the sets of all chemical species and chemical elements respectively comprising the system, yj,k is the stoichiometric coefficient of the constituent element k in the formation reaction of the j-th chemical species, xi,j is the mass fraction of the j-th chemical species in the ith stream, Hi and Si are the specific mass enthalpy and entropy respectively of the ith stream at its temperature and pressure conditions Ti and Pi; and S_ G is the total rate of entropy generation due to irreversibilities both within the system’s control volume and in the heat transfer across finite temperature differences between the control volume and its surroundings. Definition. Let U be a steady-state open system with inlets in SI, outlets in SO, no heat transferred from the surroundings to the system, heat possibly transferred from the system to the surroundings Q_ j  0 j˛SQ at the uniform surroundings temperature ðTs;j ¼ T0 Þ, and the system’s net shaft work to be P _ Ws;j  0 (Fig. 1). Such a system is called non-positive j˛SW

energetically self-sufficient. The definition of energetic self-sufficiency can be mathematically stated as:

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

X

Q_ 0  0;

_ s;j  0 W

(8)

j˛SW

0¼

X

_i m

i˛SI

0¼

i˛SI j˛SC

X

(9)

O

_iþ Hi m

i˛SI

X i˛SO

X

(10)

X

_i Si m

 Q_ 0  0

_i Si m

i˛SO

X X xi;j xi;j _i _ i ¼ 0 ck˛SE m nj;k m M Mj j i˛SI j˛SC i˛SO j˛SC ! X X _i _ i  T0 $S_ G  0 Q_ 0 ¼ T0 $ Si m Si m nj;k

i˛SI

X X _i _i ðHi  T0 $Si Þm ðHi  T0 $Si Þm

_ s;j ¼ T0 $S_ G  W

j˛SW

i˛SI

! 0

i˛SO

(13) The above imply that necessary conditions, independent of S_ G , for a system to be energetically self-sufficient are: X X _i _i¼0 m m i˛SI

i˛SO

XX

i˛SI

(14)

Let Gi aHi  Ti $Si be the specific mass Gibbs free energy of the ith stream at its temperature and pressure conditions Ti and Pi. When all the system inlets and outlets are at the

n X

xi;k =Mk ¼

n X

k¼1

j¼1

_i0 Hi m

(17)

i˛SO

_ i Gi m

X

_ i¼ Gi m

i˛SO

X

_i ðHi  T0 $Si Þm

i˛SI

X _ i 0 ðHi  T0 $Si Þm

(18)

where DH_ s ðT0 Þ and DG_ s ðT0 Þ are respectively the system’s enthalpy and Gibbs free energy change rates at T0 ¼ 298.15 K. In this work, carrying out a thermodynamic analysis of energetic self-sufficiency requires evaluation of DH_ s ðT0 Þ and DG_ s ðT0 Þ, which in turn requires evaluation of the inlet and outlet streams’ specific mass enthalpies and Gibbs free energies. To this end, all streams are considered to consist of pure substances, except for the inlet air stream which is considered to be an ideal gas mixture. In addition, all streams are considered to be ideal gases, and to enter and exit the system at T0 ¼ 298.15 K and P0 ¼ 1 bar, except for hydrogen which is considered to be a real gas and is considered to exit the system at T0 ¼ 298.15 K and P0 ¼ 354.6 bar. Under these assumptions, the specific mass enthalpies and Gibbs free energies via ideal gas state, residual, and heat of formation properties are calculated as follows p. 140, p. 170 and p. 393, [15]: Case 1, i ¼ mixture; ideal gas; i ¼ CH4 ; H2 O; air; j ¼ CH4 ;

For mixtures with n components the following mass-molar fraction relations hold: yi;j Mj xi;j =Mj ; yi;j ¼ n j ¼ 1; n; i˛SI WSO n X X yi;k Mk xi;k =Mk k¼1

i˛SO

Hi aHi ðTÞ ¼ hi ðTÞ$



xi;j ¼

i˛SO

X X _i _i0 ðHi  T0 $Si Þm ðHi  T0 $Si Þm i˛SI

X

X

H2 O; N2 ; O2 ; Ar :

X X xi;j xi;j _i _ i ¼ 0 ck˛SE m nj;k m M Mj j i˛SI j˛SC i˛SO j˛SC X X X _ s;j ¼ _i _i0 W Q_ 0  Hi m Hi m nj;k

j˛SW

_i Hi m

(16)

C

i˛SO

S_ G  0 X X _i _i¼0 m m

X

X

i˛SI

i˛SI

i˛SO

O

i˛SI

(12)

This can be equivalently stated as:

i˛SI

i˛SI j˛SC

DG_ s ðT0 Þa

!

(15)

X X xi;j xi;j _i _ i ¼ 0 ck˛SE m nj;k m Mj Mj i˛S j˛S

(11)

j˛SW

X

nj;k

DH_ s ðT0 Þa

_ s;j W

_i¼0 m

i˛SO

XX ck˛SE

X

C

_i Hi m

i˛SO

XX

_i m

i˛SI

X X xi;j xi;j _i _i nj;k m nj;k m Mj Mj i˛S j˛S

T0 $S_ G ¼ T0 $

temperature condition of the surroundings (T0 ¼ 298.15 K), then the above necessary conditions can be written as: X

_i m

i˛SO

XX

Q_ 0 ¼ 

X

12855

k¼1

Additionally, the following relations for the calculation of mixture properties hold: Then, from p. 393, [15]

n n n X X X xi;j =Mj ig ig ig hj ðTÞ n $ xi;k =Mk ¼ hj ðTÞxi;j =Mj ¼ Hj ðTÞxi;j X k¼1 j¼1 j¼1 xi;k =Mk k¼1

X 2 3 Dhof;j ðT0 Þ þ nj;k $hok ðT0 Þ T ig Z n X6 R Cp;j k˛SE 7 dT þ ¼ 4 5xi;j M R M j j j¼1 T0

n X xi;j =Mj xi;j =Mj Pn lnPn $ xi;k =Mk x =M x =M k k k¼1 k¼1 i;k k¼1 i;k k¼1 j¼1 j¼1 1 0 ! n n n n X X X X     ig A @ ¼ Sj ðTÞxi;j  R xi;j =Mj ln xi;k =Mk xi;j =Mj ln xi;j =Mj þ R

Si aSi ðT; PÞ ¼ si $

n X

xi;k =Mk ¼

j¼1

n X

j¼1

ig

sj ðTÞxi;j =Mj  R

n X

j¼1

k¼1

2 3 P 1 0 ! nj;k $sok ðT0 Þ   Dsof;j ðT0 Þ þ ZT Cig n n n n X X X X     R P k˛SE p;j 6R 7 þ dT  ln xi;j =Mj Aln xi;k =Mk xi;j =Mj ln xi;j =Mj þ R@ ¼ 4 5xi;j  R Mj RT Mj Mj P0 k¼1 j¼1 j¼1 j¼1 T0

12856

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Gi aGi ðT; PÞ ¼ gi $

n X

xi;k =Mk ¼

k¼1

n X

ig

gj ðTÞxi;j =Mj þ RT

j¼1

n X j¼1

n X xi;j =Mj xi;j =Mj ln n $ xi;k =Mk n X X xi;k =Mk xi;k =Mk k¼1 k¼1

n X

¼

ig Gj ðTÞxi;j

þ RT

j¼1

n X

0

    xi;j =Mj ln xi;j =Mj  RT@

j¼1

k¼1 n X j¼1

1

xi;j =Mj Aln

n X

! xi;k =Mk

k¼1

X 3 Dhof;j ðT0 Þ þ nj;k $hok ðT0 Þ ZT Cig k˛SE p;j 7 6 R dT þ 7 6 1 0 7 Mj R M n 6 n n X X X 7 6 j T0     X 6 ¼ xi;j =Mj A xi;j =Mj ln xi;j =Mj  RT@ 3 7xi;j þ RT 2 6 Dsof;j ðT0 Þ þ nj;k $sok ðT0 Þ 7 7   ZT Cig j¼1 6 j¼1 j¼1 7 6 R R P k˛SE p;j 75 4 T$6 þ dT  ln 5 4 Mj RT Mj Mj P0 T0 ! n X xi;k =Mk ln 2

k¼1

Case 2, i ¼ pure; real gas; i ¼ H2 :   ig ig ig ig Hi aHi ðT; PÞ ¼ Hi ðT; PÞ  Hi ðTÞ þ Hi ðTÞ  Hi ðT0 Þ þ Hi ðT0 Þ  ig ZT Cig hi ðT; PÞ  hi ðTÞ R p;i dT þ ¼ R Mi Mi Dhof;i ðT0 Þ þ

X

Gi aGi ðT; PÞ ¼ Hi ðT; PÞ  T$Si ðT; PÞ  ig ZT Cig hi ðT; PÞ  hi ðTÞ R p;i dT ¼ þ Mi R Mi Dhof;i ðT0 Þ þ

Mi

  ig ig ig Si aSi ðT; PÞ ¼ Si ðT; PÞ  Si ðT; PÞ þ Si ðT; PÞ  Si ðT0 ; P0 Þ



ig

¼

þ Si ðT0 ; P0 Þ  ig si ðT; PÞ  si ðT; PÞ Mi Dsof;i ðT0 Þ þ þ

X

þ

R Mi

ZT Cig p;i RT T0

Mi

2 ig ZT Cig s i ðT; PÞ  si ðT; PÞ R p;i 6 dT  T$4 þ Mi RT Mi

ni;k $hok ðT0 Þ

k˛SE

þ

T0

ni;k $hok ðT0 Þ

k˛SE

þ

T0

X

dT 

  R P ln Mi P0

ni;k $sok ðT0 Þ

k˛SE

Mi

  R P þ ln Mi P0

Dsof ;i ðT0 Þ þ

X

T0

3 ni;k $sok ðT0 Þ k˛SE 7 5 Mi

where n is the number of species, SE is the set of constituent elements comprising the system, ni;k  0 k˛SE i ¼ 1; n is the stoichiometric coefficient of the constituent element k in the formation reaction of the j-th chemical species, and the hydrogen’s residual properties are calculated from the generic cubic equation of state as shown in Appendix A. The standardstate specific mass enthalpy and specific mass entropy of the ig

ig

ith species at T0 are denoted as Hi ðT0 Þ and Si ðT0 ; P0 Þ respecig

ig

tively. Hi ðT0 Þ ðSi ðT0 ; P0 ÞÞ is equal to the molar standard enthalpy (entropy) of formation of the ith species at T0, ! Dhof ;i ðT0 Þ  Dgof ;i ðT0 Þ o , over its molecular weight, Mi, Dhf ;i ðT0 Þ T0 plus the weighted sum of the standard-state specific mass enthalpies (entropies) of its constituent elements, hok ðT0 Þ ðsok ðT0 ÞÞ, with weights the ratios of the stoichiometric coefficients of the constituent elements ni;k  0 k˛SE i ¼ 1; n over the molecular weight, Mi [15]. Typically, hok ðT0 Þ, are chosen to be zero for all elements, while sok ðT0 Þ are evaluated based on the 3rd law of thermodynamics, p. 188, [15]. Fig. 1 e Energetically Self-Sufficient Open Steady-State System. U is a well delimited open system with streams from the set SI entering the system and streams leaving the system into the set SO, at temperature T0; heat (Q0, red arrow) and shaft work (Ws, blue arrow) are negative and therefore generated by the system. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

To simplify the preliminary thermodynamic calculations, the case in which oxygen is being fed to the system in lieu of air is considered. This simplification is not employed in the detailed pinch assessment calculations which consider that air containing nitrogen, oxygen and argon enters the flowsheet, while nitrogen and argon are among the flowsheet outlets. The standard-state molar enthalpy and Gibbs free energy of formation values at T0 were taken from the literature (Table 1) [15e18].

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

12857

a

Table 1 e Standard Enthalpies and Gibbs Energies of Formation (J/mol). Speciesa

Dhof,i(298)

Dgof,i(298)

Reference

CH4(g) CO2(g) H2O(g) H2O(l) H2(g)b CO(g) HCOOH(l) HCOOH(g) O2(g)

74,520 393,509 241,818 285,830 459.888 110,525 424,760 378,610 0

50,460 394,359 228,572 237,129 15,085 137,169 361,414 351,000 0

[14] [14] [14] [14] [15] [14] [16] [17] [14]

a (g) ¼ gas state, (l) ¼ liquid. b Standard-state value plus contribution calculated from departure function of RK equation of state at 354.6 bar.

b As discussed earlier, the most common industrial process for the production of hydrogen is steam reforming (SMR) of natural gas, which involves the incomplete endothermic transformation of methane and water to hydrogen, carbon dioxide and carbon monoxide [19e21]. Since the overall process does not release any carbon monoxide, it can be captured by the following overall chemical reaction:   Z O2 þ Z H2 O/CO2 þ ð2 þ ZÞH2 CH4 þ 1  2

(19)

An industrially implemented realization of this was discussed earlier in some detail [8,9]. An inputeoutput representation of the above overall chemical reaction is shown in Fig. 2. It considers that all process inlets and outlets are at T0 ¼ 298.15 K and P ¼ 1 bar, except for the purified hydrogen product which is delivered at T0 ¼ 298.15 K and P ¼ 354.6 bar, so as to meet the first-generation hydrogen-fueled automobile storage requirements. The single energy resource for this process is the incoming natural gas (which is used both as an energy source and as raw material for the hydrogen production) and it is thus desired to identify the water to methane molar ratio Z so that the overall process is an energetically self-sufficient system. Fig. 3a shows the enthalpy change rate ðDH_ s Þ and Gibbs free energy change rate ðDG_ S Þ for Eq. (19) per 1 mol/s of CH4. Based on the graph in Fig. 3a we can reach the conclusion that the

Fig. 2 e Energetically Self-Sufficient Hydrogen Production System. A well delimited hydrogen production system with methane, water and oxygen streams entering the system; hydrogen and carbon dioxide exiting the system. No heat/work entering the system and heat/work generated leaving the system.

Fig. 3 e (a) Enthalpy and Gibbs Free Energy Change in Mass Balance of Eq. (19) at 298 K and 1 bar (H2 at 354.6 bar); (b) Enthalpy and Gibbs Free Energy Changes in Hydrogen/ Formic Acid Co-production Process, Eq. (20). The horizontal line marks the level at which both of the values are zero and the vertical line indicates the energetically selfsufficient parameter value at this intersection.

self-sufficiency conditions of the system carrying out reaction (19) are met, for maximum hydrogen production, at Z þ 2 ¼ 3.325 moles of hydrogen per mol of methane which occurs at a water to methane molar ratio of Z ¼ 1.325. Various inefficiencies in bringing about a realization of reaction (19) may lead to a hydrogen/methane molar ratio smaller than Z þ 2 ¼ 3.325. As discussed in Ref. [22] the realization of reaction (19) has a methane/hydrogen molar ratio of Z þ 2 ¼ 2.797. To eliminate carbon dioxide emissions we explore in this work the feasibility of coproducing hydrogen with formic acid, while maintaining energetic self-sufficiency. The overall chemical reaction for this proposed hydrogen production process is shown below:   X O2 þ XH2 O/HCOOH þ ð1 þ XÞH2 CH4 þ 1  2

(20)

The requirement that this alternative hydrogen production process is energetically self-sufficient is first analyzed based on

12858

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

the 1st law of thermodynamics. All process inlet and outlet streams are considered to be at 298 K and 1 bar, except for the hydrogen product which is to be delivered at 298 K and 354.6 bar (meeting the first-generation hydrogen-fueled automobile storage requirements). The process enthalpy and Gibbs free energy changes, normalized on a per mole per second of methane basis, are then evaluated as a function of the water to methane molar ratio (X ). As Fig. 3b indicates the proposed system is energetically self-sufficient for values of X  1.195 thus placing a thermodynamic upper bound of X þ 1  2.195 on the number of moles of H2 that can be produced per mole of CH4 with simultaneous co-production of a mole of formic acid (HCOOH). This kind of 1st law of thermodynamics based analysis of energetic self-sufficiency can be carried out for a variety of other carbon containing chemicals that can be co-produced with hydrogen from natural gas. In Table 2, several such possible carbon containing chemicals are considered. For each of these chemicals, a similar 1st law based energetic selfsufficiency analysis is carried out, and the maximum hydrogen to methane ratio is quantified. As can be seen from Table 2 [23e26], co-production of hydrogeneformic acid, hydrogeneacetic acid, and hydrogenemethanol, all lead to maximum hydrogen to methane ratios that are smaller than the corresponding ratio of 3.325 achieved when CO2 is hydrogen’s co-product. Since the hydrogeneformic acid maximum hydrogen/methane ratio of X þ 1  2.195 is the closest to 3.325, it is this production process that is the focus of our investigation. The next step in formulating a viable hydrogeneformic acid co-production process is to identify technologies that have been industrially implemented for the production of these chemicals individually, and then to integrate these technologies. Since no process exists that can carry out the desired transformation in a single step, its implementation can be pursued through the use of a reaction cluster, that is, through a set of reactions whose sum is the overall reaction. Each step within the reaction cluster must be thermodynamically feasible [27e29]. A set of reactions satisfying these conditions consists of steam methane reforming, water-gas shift, combustion, and formic acid production from hydrogen and carbon dioxide (Fig. 4), as shown below. The reaction cluster shown in Fig. 4 consists of five reactions and possesses two degrees of freedom X  1; Y  0, due to the bifurcation of the methane input into hydrogen production (SMR) and energy production (combustion). To assess whether this system requires no external heat input, a 2nd law based thermodynamic analysis based on pinch

inequalities can be carried out [30e32]. Such an analysis can, and will, be rigorously carried out following the construction of an integrated co-production flowsheet. However an approximate pinch analysis can be carried out before the construction of any particular flowsheet based only on knowledge of the cluster reactions listed in Fig. 4, and of the temperatures at which these reactions are carried out industrially. This approximate analysis ignores the sensible heat effects in transporting reactants and products from one reaction temperature to another (considering that appropriate heat integration will approximately cancel out any such effects), considers that all reactions proceed to completion, and also ignores any heat/work input required to carry out any needed separations. The enthalpy and Gibbs free energy change of each reaction is calculated using Eqs. (17) and (18) at the corresponding temperature of the reaction. In this context, this approximate pinch analysis is carried out for the cluster using a minimum approach temperature DTmin ¼ 10 K, and considering that the first reaction (combustion) is carried out at 1150 K, the two endothermic reforming reactions take place in the SMR at 1090 K (considered as 1100 K in the cold stream temperature scale shifted upward by the minimum approach temperature DTmin ¼ 10 K), the WGS reaction (exothermic) takes place at 623 K, and finally the formic acid reaction (exothermic) takes place at 353 K. The resulting pinch inequalities represent the requirement that the cluster’s net available enthalpy at each of the above (hot stream scale) temperatures should be negative, since the heats of reaction are negative (positive) for exothermic (endothermic) reactions: ð0:5  0:25XÞDH1  0 .T1 ¼ 1150 K

(21)

ð0:5  0:25XÞDH1 þ ðYÞDH2 þ ð0:5 þ 0:25X  YÞDH3  0 .T2 ¼ 1090 K (22) ð0:5  0:25XÞDH1 þ ðYÞDH2 þ ð0:5 þ 0:25X  YÞDH3 þ ð0:5 þ 0:25X  YÞDH4  0 .T3 ¼ 623 K

(23)

ð0:5  0:25XÞDH1 þ ðYÞDH2 þ ð0:5 þ 0:25X  YÞDH3 þ ð0:5 þ 0:25X  YÞDH4 þ DH5  0 .T4 ¼ 353 K

(24)

ð0:5  0:25XÞDH1 þ ðYÞDH2 þ ð0:5 þ 0:25X  YÞDH3 þð0:5 þ 0:25X  YÞDH4 þ DH5 þQc ¼ 0 .T5 ¼ 298 K

(25)

Table 2 e Carbon Containing Chemical Candidates for Hydrogen Co-production. Max H2/CH4 molar ratio

Main Application

World Production Capacity [1000 tons/year]

Carbon Dioxidea Formic Acid

3.325 2.190

e 610

Acetic Acid Methanol

1.700 1.140

e Leather treatment, Food Preservative Vinyl acetate monomer synthesis Methyl tertiary-butyl ether synthesis

Carbon Containing Chemical

a Reference point.

4815 9230

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

12859

Fig. 4 e Reaction Cluster for Hydrogen and Formic Acid Production. Shown are 5 chemical reactions as a function of two parameters (X, Y ), whose addition results in an overall reaction with only one parameter (X ).

0:5 þ 0:25X  Y  0

(26)

Qc  0

(27)

DH1 ¼ 801943; DH2 ¼ 192607; DH3 ¼ 226256; DH4 ¼ 38505; DH5 ¼ 29406; all in kJ=mole The constraints in Eq. (26) ensure that the direction of the reactions in the cluster is to the right. The constraint in Eq. (27) ensures that the cold utility load Qc, that must be removed from the system at the bottom of the temperature cascade to keep both the 1st and 2nd laws of thermodynamics satisfied, be non-negative, i.e. Qc  0. The above set of inequality constraints, combined with X  1; Y  0, describes a two dimensional feasible region in XeY space (Fig. 5). The point of maximum production of hydrogen lies at X ¼ 1.225, Y ¼ 0.8, where constraints (22), (23) and (26) intersect. This suggests there is a possibility of increasing the production of hydrogen by carrying out the reforming reactions separately, but the technology employed in this case study requires reforming reactions occurring in a single reactor.

3. A realization of the proposed hydrogen production method The overall hydrogeneformic acid co-production process can be divided in four parts: 1. Steam Methane Reforming Plant. This consists of a pressurization system, the SMR reactor and a flash distillation process. 2. Hydrogen PSA Plant. The hydrogen PSA receives the stream coming from the top of the flash distillation process located downstream of the SMR reactor. Hydrogen is purified and pressurized for its final delivery. 3. Air separation and Combustion Plants. Air separation PSA system, a burner, an atmospheric pressure water condenser and a high pressure water condenser. 4. Formic Acid Plant. Feedstock pressurization system, the formic acid reactor and the dual pressure distillation column system.

3.1. Steam methane reforming and Hydrogen PSA subsystems The plant is fed with water and natural gas (methane) at atmospheric conditions, which are then brought to typical SMR conditions (10.1e30.4 bar, 800 Ke1200 K) [8,9,21,22,33e36]. The overall reaction occurring in the SMR ðgÞ

ðgÞ

ðgÞ

CH4 þ 2H2 OðgÞ /CO2 þ 4H2

Fig. 5 e Feasible Region Describing Cluster Constraints. ) Constraint (21), ( ) Constraint (22), ( ) Constraint ( ) Constraint (24), ( ) Constraint (26). Shown is the (23), ( region constructed through the set of constraints given by Eqs. (21e26); any point in the region satisfies constraint 21 since it has a degree of freedom in Qc.

ðR3 Þ

(28)

is an endothermic reaction and, by the Le-Chatelier principle, hydrogen production increases with temperature. Thus for high conversion and heat integration purposes an operating temperature of 1090 K is chosen, which is 60 K below the combustion reactor’s temperature of 1150 K. To avoid excessive pressurization, the SMR pressure of 21.3 bar is chosen, which is the same as the high pressure of the hydrogen PSA subsystem. The SMR is followed by a series of Water-Gas Shift (WGS) reactors where additional hydrogen is produced and carbon monoxide is significantly reduced. These reactors are considered part of the SMR subsystem in Fig. 6. The stream exiting the WGS reactor is then cooled to the operation temperature of the H2 e PSA and subsequently sent to a flash column where most of the water is removed. The water is recycled back into the SMR reactor, with a portion of it being sent to the combustor feed to be used as steam dilution to lower the speed of combustion and act as a heat sink that keeps the combustor from overheating. The light ends of the separator

12860

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Fig. 6 e Steam Methane Reformer and Hydrogen Pressure Swing Adsorber Diagram. A simple methane reforming diagram is shown; water and methane are compressed, heated and fed into the SMR reactor, along with other indicated streams; the exit is cooled and separated through a separation column and a PSA block, from which hydrogen is purified and delivered.

are sent to the H2 e PSA. The conditions at which the PSA is operated (21.3 bar maximum pressure and 311 K) are taken from Ref. [22]. Two streams come out of the PSA system, one containing H2, CO and CO2 at 1 bar, which is sent to the combustion plant and the other one consisting of pure H2 at 21.3 bar (99.999þ%). The latter is separated into two streams, one of which is sent into the formic acid plant and the other of which is pressurized to 354.6 bar to be delivered as a final product.

3.2. Two-stage PSA air separation and combustion subsystems Two fresh feeds come into this section of the plant; one consisting of methane and the other one of air. The air stream is fed into an air separation subsystem, which consists of a two-stage PSA. In the first stage, nitrogen is removed from the air by a zeolite packed PSA unit [33,34] and in the second stage argon contained in the first stage’s oxygen-rich product is removed through a PSA packed with carbon molecular sieve (CMS) adsorbent. Conventional air separation PSA systems use only a zeolite packed PSA, however this adsorbent does not remove the argon, leaving the oxygen product with a purity less than 95% [37,38]. The addition of the CMS PSA allows a purity up to 99.5% [33,39]. The energy needs of the two-stage PSA process are estimated from the literature [9] under the following assumptions: 1) Oxygen is produced at 100% recovery and purity. Incomplete recovery only affects the amount of air fed into the plant. In regard to energy consumption, the incomplete recovery of oxygen can be readily taken into account [40]. Ar is the leading impurity present in oxygen produced from air using zeolite PSA. If the Ar was not completely removed, it would exit the system in small quantities together with the formic acid stream, not affecting the functionality and energetic performance of the flowsheet. 2) The air enters the air separation PSA system at 298 K and 1 bar, and the oxygen product leaves the system at the same conditions. The oxygen produced from the PSA is then mixed with the fresh methane, the waste from the hydrogen PSA, water

stream from the SMR, and a recycle stream coming from the formic acid plant. The mixture is heated to 1130 K and fed into the combustion reactor which operates at 1150 K and 1 bar. The feed to the burner contains 20% steam to prevent the combustor from overheating. The rest of the feed is stoichiometric (to the methane, hydrogen, carbon monoxide, and formic acid entering the reactor) and the combustion is considered to be a complete conversion reaction, generating only water and carbon dioxide as its products. The combustor exit stream is cooled down to room temperature so most of its water content is condensed and subsequently separated from the carbon dioxide using two condensers, the first condenser operating at 1 bar and the second one at 21.3 bar (matching the pressure of the hydrogen stream with which it will be mixed). Pure CO2 is collected at the light end of the second separator, while water collected at the bottom of both separators is mixed and sent to the SMR plant. The water is sent to the SMR reactor and the carbon dioxide to the formic acid plant (Fig. 7).

3.3.

Formic acid plant subsystem

In this subsystem formic acid is produced from hydrogen and carbon dioxide. Two streams come into this part of the plant: the hydrogen stream which is derived from the hydrogen exiting the H2-PSA; and the carbon dioxide dehydrated exiting the distillation system following the combustion reactor. These streams (21.3 bar, 308 K) are mixed and then compressed and heated to 40.5 bar and 354 K, the pressure and temperature at which the reactor operates [41e43]. Water is used as reaction solvent and is recycled within the reaction and purification systems [41,44e46]. Most of the water is recovered from the bottom of the first distillation column and recycled to the entrance of the formic acid reactor. The rest of the water comes from the second distillation column along with formic acid, due to the azeotrope these two species form. The exit of the formic acid reactor is depressurized to 11 bar, heated close to the water’s boiling point at this pressure (470 K) and then fed into a distillation column operating at 11 bar (high pressure distillation column). The light end stream at the top of this high pressure distillation column contains mostly water with traces of carbon dioxide, while the heavy end stream at the column’s bottom consists of the high

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

12861

Fig. 7 e Dual Pressure Swing Adsorber (PSA) Air Separation and Combustion Plant Diagram. A simplified diagram of a dual PSA, burner and separation columns is shown. Oxygen is obtained from the dual PSA separation of air, then fed into the burner along with methane and recycled material; the exit is separated into carbon dioxide and water through two separators.

pressure azeotropic mixture of water and formic acid. This mixture is expanded to atmospheric pressure and fed into a second distillation column (low pressure distillation column) operating at 1 bar. The heavy ends of the low pressure distillation column consist of the atmospheric azeotropic mixture of water and formic acid, while the light ends consist of 99.9% pure formic acid, which is the final product (Fig. 8).

4. Simulation and heat/power integration of proposed process realization The proposed process flowsheet is first simulated and then heat and power integrated. The flowsheet simulation is carried out using the software UniSim Design R360 and R380. Heat and power integration is carried out using the UCLA in-house software which implements the theory developed in Ref. [30]. A process flow diagram is presented in Fig. 9. The system is initially simulated at the operational point for X ¼ 1.195, previously explained. The converged flowsheet’s temperatureeenthalpy data are then entered into the heat and power integration software, to identify the flowsheet’s hot/cold/electric utility needs. This process is repeated for various values of X. A value of X ¼ 0.654 is identified as the one leading to the

highest production of H2 for this process realization, while maintaining the flowsheet to be energetically self-sufficient. The PengeRobinsoneStryjekeVera equation of state is used to simulate most of the plant except for the Formic Acid plant in which the van Laar liquid activity coefficient model coupled with an ideal gas model for the vapor phase [47] was found to be the thermodynamic model that most accurately captured the watereformic acid azeotropic vaporeliquid equilibria. The material inputs to the flowsheet are the raw materials: natural gas (methane), water, and air at ambient temperature (298 K) and atmospheric pressure. The material outputs consist of pure nitrogen and pure argon at 1 bar and 298 K, hydrogen at 354.6 bar and 298 K and high purity formic acid (99.9%) at 1 bar and 298 K. The natural gas input is split into a SMR input and a Combustion plant input, with the split ratio being a direct function of the previously defined water to methane molar ratio (X ). A higher value of X yields a higher production of hydrogen and a larger flow of natural gas into the SMR, while a lower value of X yields higher energy production and a larger flow of methane to the combustion plant. The SMR and combustor are represented by Gibbs free energy reactors (without specified reactions) and the formic acid reactor is represented as a conversion reactor with a conversion fixed at 20%. The heat exchangers (heaters and

Fig. 8 e Formic Acid Production Plant Diagram. A simplified diagram of the formation and purification of formic acid is shown. Carbon dioxide and oxygen are fed into a reactor at the indicated conditions; the gaseous exit of the reactor is partially recycled, while the liquid phase is sent to a system of distillation columns to extract and purify formic acid.

12862

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Fig. 9 e UniSim Flowsheet of Hydrogen- Formic Acid co-production process. Shown is the simulation created in the process modeling software UniSim Design.

coolers) were assumed to have negligible pressure losses, and the compressors to work at a 75% adiabatic efficiency. The pressurization operations are represented by a single compressor and a cooler that may in reality correspond to a multi compressor intercooler train system. A mixture of recycled water, fresh water and natural gas is fed into the reformer at 1090 K and 21.3 bar, resulting in a molar ratio of steam/methane of 3.52 at the reactor entrance. This steam excess over the stoichiometric steam/ methane molar ratio of 2 leads to reduced byproduct carbon formation [21,48,49]. At the given conditions, the equilibrium conversion of natural gas in the reformer is 81.17%, generating a stream containing primarily water and hydrogen with significant amounts of carbon dioxide, monoxide and methane. The water in this stream is condensed and separated, through a flash separator at 21 bar and 311.15 K, to a liquid stream consisting mostly of water and to a vapor stream that is rich in hydrogen and is sent to the pressure swing adsorber (PSA) unit. The PSA process is considered to be adiabatic and almost isothermal, thus only requiring its power consumption to be incorporated in the heat and power integration analysis. A cyclic operation of a series of PSAs allows the overall system to operate in a steady-state mode. For purposes of heat and power integration, the PSA system is represented in the process diagram by a component splitter. The exits of the PSA are set to comply with the literature reported compositions [50,51]. The high pressure adsoption

step generates hydrogen while the low pressure (1.3 bar) desorption step outputs a waste stream. The waste stream generated is sent to the combustion plant to be oxidized to carbon dioxide and water along with other waste recycles and fresh natural gas. The hydrogen exiting the H2-PSA is split into the final product (1.654 mol of hydrogen per mole of methane) and the hydrogen feed to the formic acid plant. The final hydrogen product stream is compressed and cooled to be delivered at commercial conditions (354.6 bar and 298 K). The burner is operated at 1150 K and 1 bar. This high temperature allows the unit to generate a high quality hot stream that is used to satisfy the energy requirements of energy demanding units such as the SMR. An air stream undergoes oxygen purification through the air-PSA units previously described; these units are also represented in the process diagram as component splitters. The air-PSA process requires work of 48,000 kJ per kmol of oxygen produced [9]. This work load requirement is accounted for during heat and power integration analysis. The amount of oxygen required is already given by the overall balance equation which is the stoichiometric amount to oxidize all carbon containing products and unreacted hydrogen entering the burner into carbon dioxide and water. The remaining streams going into the burner come from waste streams from the hydrogen PSA, formic acid reactor, water from the SMR plant and fresh natural gas. The combustion is carried out at 1 bar and 1150 K and is modeled on a Gibbs free energy minimization basis, yielding a vapor stream

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

exit composed essentially of water and carbon dioxide. Water is separated by condensation in two stages, at 308.15 K and 1 bar in the first stage, and subsequently at 308.15 K and 21.3 bar in a second stage. The condensed water is then sent to the SMR, and the CO2-rich stream is then mixed with an equimolar amount of hydrogen and sent to the formic acid reactor. The hydrogen/CO2 mixture is pressurized from 21.3 bar to 40.5 bar and heated up to 354.15 K, conditions at which the reaction takes place [41e43]. According to the literature, this reaction can be carried out at various conditions with a great variety of solvents and catalysts [42,43]. For this work, a water-soluble Rhodium-based catalyst (RhCl(TPPTS)3) was chosen that allowed the use of mild temperature and pressure conditions in the formic acid reactor. The input to the formic acid reactor consists of an equimolar mixture of hydrogen and carbon dioxide with water and traces of formic acid, the water being used to dissolve the Rhodium-based catalyst (RhCl(TPPTS)3) and largely originating as recycle from the formic acid purification system. The formic acid reactor is considered to have a low conversion (20% conversion), thus generating an exit stream with significant amounts of unreacted hydrogen and carbon dioxide. The liquid output of the reactor has a 1.08 formic acid to water molar ratio, along with 0.8% and 2.3% mole of dissolved hydrogen and carbon dioxide respectively. This mixture is depressurized to 11 bar where it is separated via distillation. Water and formic acid form an azeotropic mixture whose purification through distillation is carried out in a two-step process. The first step is at 470 K and 11 bar, conditions at which the water is the lighter component. This first step is also necessary since hydrogen and carbon dioxide must be recovered to be recycled to the formic acid reactor, while the dissolved catalyst remains in the liquid phase stream. A second distillation step at 1 bar and 381.2 K is carried out, and the phase equilibrium simulations show water as the heavier component, meaning water along with catalyst will be recovered at the liquid exit of the column, thus producing a highly pure formic acid (99.5%) stream at the top of the column that will be delivered as final product at 1 bar and 298 K. Having completed the detailed simulation of the proposed hydrogeneformic acid co-production flowsheet, a rigorous heat and power integration analysis is now carried out on the converged flowsheet. To this end, the technique outlined in Ref. [30], for the calculation of the minimum hot/cold/electric utility cost problem for a heat exchange network, is employed. In this approach, individual heat exchange/heat engine/heat pump units are not considered, but rather a global thermodynamic approach is considered which holds for all possible networks of such units. The problem is stated as follows “given a set of streams with specific flowrates, inlet temperatures, and fixed outlet target temperatures; hot and cold utility streams with known inlet and outlet temperatures and unit cost; an electrical (work) utility with known unit cost; identify among all possible heat exchange/pump/engine networks, the minimum total (hot/cold/electricity) utility cost necessary to accomplish the desired changes”. To carry out this integration analysis, we first examine all enthalpy changing processes in the converged flowsheet that require external heating and cooling. These processes are further classified into isothermal and non-isothermal, and the

12863

streams involved are characterized as “hot streams” if their required enthalpy change is negative, and as “cold streams” if this change is positive. Since we wish the flowsheet to be energetically self-sufficient, hot utility use must be zero. Additionally, cold utilities are allowed, as well as the use of heat engines and heat pumps. Cold utility is available at 293 K, the work/cold utility cost ratio is 25 to 2, and the minimum approach temperature is 10 K. The objective function for the heat and power integration problem minimizes the total power and cold utilities cost. The resulting optimization problem is linear and it is solved using an optimization solver. The details of the formulation are discussed in Refs. [22,30].

5.

Results and discussion

A number of iterations were carried out to find the maximum hydrogen/methane molar ratio X þ 1 that would still allow the converged flowsheet to be energetically self-sufficient. The starting point is the thermodynamic limit of X þ 1 ¼ 2.195 identified earlier in Fig. 3b. For this value of X a converged flowsheet is first obtained, and subsequently analyzed from a heat and power integration viewpoint using the cold/power utility minimization procedure outlined earlier. It turns out that the obtained flowsheet is not energetically self-sufficient, i.e. it requires the use of hot utility to satisfy the 1st and 2nd laws of thermodynamics. Repeating this utility cost minimization procedure for various values of X yields a total work (work generated from heat and power plus work demand of the flowsheet) that varies in a nearly linear manner with X (Fig. 10). Close examination of Fig. 10 identifies the optimal water/ methane molar ratio to be X ¼ 0.654. At this ratio, close examination of the temperatureeentropy change diagram, Fig. 11a, associated with the minimum cold/power utility cost solution, reveals that optimal heat-power integration leads to three heat engine and two heat pump regions. Indeed, the solid line corresponding to the hot composite curve and the dotted line corresponding to the cold composite curve at the

Fig. 10 e Total Work vs. Parameter X in Process Flowsheet Simulation. Indicated are the regions corresponding to the work generated and consumed, and the equation correlating the total work to parameter X.

12864

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Fig. 11 e (a) TemperatureeEntropy Diagram: Heat Engine/Pump Network Featuring Minimum Cold Utility for the Overall Process; (b) TemperatureeEnthalpy change Diagram: Heat/Power Integration Featuring Minimum Cold/Power Utility Cost Shown in Fig. 11a is the temperatureeentropy diagram obtained from the heat and power integration of the identified optimal flowsheet; indicated are the heat engines (red) and heat pumps (blue) regions required for the integration of the flowsheet. Shown in Fig. 11b is the temperatureeenthalpy diagram from the heat and power integration of the optimal flowsheet; the width of the pictured column corresponds to the amount of power generated from the network of heat engines and heat pumps. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cold/power utility cost minimum, intersect each other at four locations. The two heat pump regions are the first between the first intersection point at 370 K and the second intersection point at 382 K, and the second between the third intersection point at 453 K and the fourth intersection point at 485 K. These points correspond to the condenser and the reboiler of the two

distillation columns separating the azeotropic mixture of formic acid and water. Within these temperature regions the hot composite curve is below the cold composite curve and heat pumping action is required. The overall enthalpy balance of the heat and power integration network is represented in the temperatureeenthalpy

12865

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

diagram (Fig. 11b). The imbalanced energy corresponds to the work generated by the heat exchange/engine/pump network, and assuming 100% of the work is converted to electricity, this energy is the input to all the electricity-powered (pump, compressors, PSA) units of the plant. In creating this diagram, the hot (cold) composite curve is lowered (raised) by 5 K so that the minimum approach temperature of DTmin ¼ 10 K, holds throughout the heat exchange/engine/ pump network. Having established the optimal hydrogen/methane molar ratio for maximum hydrogen production, an operating cost analysis of the converged flowsheet can be carried out. The system’s operating cost consists only of the cost of natural gas. The cost of water is negligible with respect to the cost of natural gas and all of the plant’s electricity needs are met by the plant itself, since it is energetically self-sufficient. According to the U.S. DOE, the national average industrial price of natural gas during April 2011 [52] was 5.23 $/(1000 ft)3. This figure can be translated into 4.52 $/kmol of CH4. With this information, a hydrogeneformic acid price diagram can be constructed which illustrates the price region in which these two chemicals could be sold at a profit. Marked in Fig. 12 are the commercial price for formic acid (0.70 $/kg) [23,26] and the energy-price equivalent (3.64 $/kg) [53]of hydrogen to gasoline (1 kg of hydrogen is energetically equivalent to 1 gallon of gasoline). An operating cost analysis of the plant suggests a return of 6.76 times the money invested in natural gas, when selling the formic acid at current commercial prices and dispensing hydrogen for free, or 9.29 times when selling formic acid at current commercial

prices and hydrogen at a price equivalent to gasoline. This suggests that the proposed process possesses a large profit margin (e.g. selling formic acid at 20% of its current commercial price and hydrogen at 0 $/kg is still profitable). Future research plans include the replacement of steam reforming with adsorption enhanced reforming, as discussed in Refs. [54e56]and references therein.

6.

Conclusions

An energetically self-sufficient process for the production of hydrogen and formic acid from natural gas was developed. Its maximum hydrogen production was 1.654 mol of H2 and 1 mol of formic acid per mole of CH4.The heat and power integration featuring minimum cold/power utility of the process determined that the heating and cooling requirements of the plant can be met through 3 heat engine and 2 heat pump subnetworks. This heat engine/pump network generates enough power to supply all the electricity needs of the flowsheet, including pumps, compressors and PSA systems. An operating cost analysis of the plant suggests a return of 676% when selling the formic acid at current commercial prices and dispensing hydrogen free of charge, or 929% when selling formic acid at current commercial prices and hydrogen at a price equivalent to gasoline.

Role of the funding source Funding for this project came from NSF through grant CBET 0829211 and CONACYT Mexico. NSF provided financial support to both authors and CONACYT Mexico provided financial support to the first author (Jorge A. Pena Lopez). The funding sources were not directly involved in the contents presented in this work.

Acknowledgments The authors gratefully acknowledge financial support from the National Science Foundation through grant CBET 0829211 and CONACYT Mexico.

Appendix A. Fig. 12 e HydrogeneFormic Acid Profitable Region. The diagonal line intersecting both axes is the lower bound operating cost corresponding to the optimal flowsheet. The region labeled as “profitable region” comprises of all the points at which one can price the hydrogen and formic acid and generate revenue, from an operating cost perspective. Marked are also the commercial price of formic acid and the gasoline-equivalent price of hydrogen. Since these two points are away from the lower bound line, there is a competitive opportunity to implement this technology.

Hydrogen thermodynamic properties using residual properties Constant pressure heat capacity of hydrogen in ideal gas state, p. 684 [15] ig

CP ¼ 3:249 þ 0:422$103 T þ 0:083$105 T2 ; R

TðKÞ˛½298; 3000

Generic Cubic Equation Thermodynamic Model for Real Gases, p. 93e98 [16].

12866

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

1 T P a½Tr R2 T2c ra ; Tr ðTÞa ; Pr ðPÞa ; aðTÞaJ ; Pc v Tc Pc baU P¼

Nomenclature

RTc bP Pr a½Tr  Pv ; ba ¼ U ; qðTr ÞaJ ; za Pc Tr RT UTr RT

RT aðTÞ RT aðTÞ=bðs  εÞ  ¼  v  b ðv þ εbÞðv þ sbÞ v  b ðv þ εbÞ

aðTÞ=bðs  εÞ 1 rb 5z ¼ q þ ðv þ sbÞ 1  rb ð1 þ εrbÞð1 þ srbÞ The generic cubic equation has as special cases the Van-der Waals, Redlich/Kwong, Soave/Redlich/Kwong, and Peng/Robinson models. Table A.2 summarizes the parameter values corresponding to each of these models. The thermodynamic properties of hydrogen are evaluated by combining ideal gas and residual properties. Let T0 ¼ 298.15 K, P0 ¼ 1 bar. Then, from p. 212e219, [15]. Specific mass enthalpy H

Cop;i , J/mol K standard state constant pressure heat capacity Cp of the ith species ig Cp;i , J/mol K ideal gas constant pressure heat capacity Cp of the ith species gi, J/mol molarGibbs free energy of material stream i Gi, J/kg specific mass Gibbs free energy of material stream i hi, J/mol molar enthalpy of material stream i Hi, J/kg specific mass enthalpy of material stream i ig hj , J/mol molar enthalpy of j-th species at ideal gas state ig Hj , J/kg specific mass enthalpy of j-th species at ideal gas state _ i , kg/s mass flowrate of material stream i m Mj, kg j/mol molecular weight of j-th species

h i h i h i HðT;PÞ  HðT0 ;P0 Þ ¼ HðT;PÞ  Hig ðT;PÞ þ Hig ðT;PÞ  Hig ðT0 ;P0 Þ þ Hig ðT0 ;P0 Þ  HðT0 ;P0 Þ 3 2

    

 ZT 16 d½lna½Tr  d½lna½T  7 r ig ¼ 4RT z  1 þ  1 qðTr ÞI þ CP dT  RT0 z  1 þ  1 qðTr ÞI 5 M dðlnTr Þ dðlnTr Þ T;P T0 ;P0 T0

Specific mass entropy S h i h i h i SðT; PÞ  SðT0 ; P0 Þ ¼ SðT; PÞ  Sig ðT; PÞ þ Sig ðT; PÞ  Sig ðT0 ; P0 Þ þ Sig ðT0 ; P0 Þ  SðT0 ; P0 Þ 3 2

 

 ZT 16 d½lna½Tr  dT P d½lna½T  r 7 ig  R lnðz  bÞ þ qðTr ÞI þ qðTr ÞI CP ¼ 4R lnðz  bÞ þ  Rln 5 M dðlnTr Þ T P0 dðlnTr Þ T;P T0 ;P0 T0

Q_ j , W SC

  9 8 1 z þ sb > ln if εss > = < s  ε z þ εb : where I ¼ > > ; : b if εss z þ εb

rate of heat entering the system at temperature Ts;j index set of components (species) present in at least one material stream si, J/mol K molar entropy of material stream i Si, J/kg K specific mass entropy of material stream i ig Sj , J/kg K specific mass entropy of j-th species at ideal gas state index set of constituent elements SE S_ G , J/K s rate of entropy generation index set of inlet material streams SI index set of outlet material streams SO index set of heat rates entering and exiting the SQ system

Table A.1 e Hydrogen Thermodynamic Properties. Species

Hydrogen (h2)

Critical Acentric Critical Molar mass M (kg mol1) Temperature Pressure factor u Pc (bar) Tc (K) 2.016e-3

33.19

13.13

0.216

Table A.2 e Parameter Values Corresponding to Van-der Waals, Redlich/Kwong, Soave/Redlich/Kwong, and Peng/Robinson models. Equation of State

a½Tr ðTÞ

vdW (1873) RK (1949)

1 1=2 Tr

SRK (1972) PR (1976)

aSRK ðTr ; uÞa aPR ðTr ; uÞb

s

ε

U

J

0 1

0 0

1/8 0.08664

27/64 0.42748

1 pffiffiffi 1þ 2

0 pffiffiffi 1 2

0.08664 0.07780

0.42748 0.45724

1=2

a aSRK ðTr ; uÞ ¼ ½1 þ ð0:480 þ 1:574u  0:176u2 Þ$ð1  Tr Þ2 ; uSRK ¼ 0:480 þ 1:574u  0:176u2 : 1=2 b aPR ðTr ; uÞ ¼ ½1 þ ð0:37464 þ 1:54226u  0:26992u2 Þ$ð1  Tr Þ2 ; uPR ¼ 0:37464 þ 1:54226u  0:26992u2 :

da ½Tr ðTÞ dTr 0

1 3=2  Tr 2 1=2 uSRK2  uSRKð1 þ uSRKÞTr 1=2 2 uPR  uPRð1 þ uPRÞTr

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

Sw

index set of work rates entering (consumed by) and exiting (produced by) the system reference temperature T0, K temperature at s; j surroundings region Ts;j , K _ s;j , W rate of shaft work consumed by j-th source in the W system _ s , W rate of shaft work entering (consumed by) the W system xi;j , kg j/kg mass fraction of j-th species in stream i X water to methane molar ratio with formic acid production yi;j , mol j/mol molar fraction of j-th species in stream i Z water to methane molar ratio without CO2 capture Greek letters Dhof;i , J/mol standard state enthalpy of formation of the ith species DH_ s , J/s rate of enthalpy change of system Dgof;i , J/mol standard state Gibbs free energy of formation of the ith species DG_ s , J/s rate of Gibbs free energy change of system Dsof;i , J/mol K standard state entropy of formation of the ith species stoichiometric coefficient of element k in j-th species nj;k U open, well delimited system or control volume

references

[1] Hart D, Freund P, Smith A. Hydrogen-today and tomorrow. Cheltenham: IEA Greenhouse Gas R&D Programme; 1999. [2] Midilli A, Ay M, Dincer I, Rosen MA. On hydrogen and hydrogen energy strategies: I: current status and needs. Renew Sustain Energy Rev 2005;9(3):255e71. [3] Das D, Veziroglu TN. Hydrogen production by biological processes: a survey of literature. Int J Hydrogen Energy 2001; 26(1):13e28. [4] Stiegel GJ, Ramezan M. Hydrogen from coal gasification: an economical pathway to a sustainable energy future. Int J Coal Geol 2006;65(3e4):173e90. [5] Barreto L, Makihira A, Riahi K. The hydrogen economy in the 21st century: a sustainable development scenario. Int J Hydrogen Energy 2003;28(3):267e84. [6] Barreto L, Makihira A, Riahi K. Medium and long-term demand and supply prospects for fuel cells: the hydrogen economy and perspectives for the 21st century; 2002. [7] Xu J, Froment GF. Methane steam reforming, methanation and water-gas shift: I. Intrinsic kinetics. AIChE J 1989;35(1): 88e96. [8] Sircar S. Pressure swing adsorption. Ind Eng Chem Res 2002 Mar;41(6):1389e92. [9] Ruthven DM, Farooq S, Knaebel KS. Pressure swing adsorption. New York, N.Y.: VCH Publishers; 1994. [10] Hulteberg PC, Burford H, Duraiswamy K, Porter B, Woods R. A cost effective steam reformer for a distributed hydrogen infrastructure. Int J Hydrogen Energy 2008;33(4):1266e74. [11] Posada A, Manousiouthakis V. Hydrogen and dry ice production through phase equilibrium separation and methane reforming. J Power Sources 2006;156(2):480e8. [12] Murray CN, Visintini L, Bidoglio G, Henry B. Permanent storage of carbon dioxide in the marine environment: the solid CO2 penetrator. Energy Convers Manage 1996;37(6e8): 1067e72.

12867

[13] BASF Corporation. Formic acid 85% US grade technical data sheet; October 2007. [14] Yali L. Formic acid market, entrepreneur.com; 2006. [15] Smith JM, VanNess HC, Abbott MM. Introduction to chemical engineering thermodynamics. Boston: McGrawHill; 2006. [16] Reid RC, Sherwood TK. The properties of gases and liquids: their estimation and correlation. New York, N.Y: McGrawHill; 1966. [17] Sinke GC. The heat of formation of formic acid. J Phys Chem 1959;63(12):2063. [18] Stull DR, Westrum EF, Sinke GC. The chemical thermodynamics of organic compounds. New York: J. Wiley; 1969. [19] Scholz WH. Processes for industrial production of hydrogen and associated environmental effects. Gas Sep Purif 1993; 7(3):131e9. [20] Spath PL, Mann MK. Life cycle assessment of hydrogen production via natural gas steam reforming. Golden, CO: National Renewable Energy Laboratory; 2001. [21] Armor JN. The multiple roles for catalysis in the production of H2. Appl Catal A 1999;176(2):159e76. [22] Posada A, Manousiouthakis V. Heat and power integration of methane reforming based hydrogen production. Ind Eng Chem Res 2005;44(24):9113e9. [23] ICIS Chemical Business. Product profile: formic acid; July 2003. [24] ICIS Chemical Business. Chemical profile: methanol; January 2007. [25] ICIS Chemical Business. Chemical profile: acetic acid and its use in VAM and PTA; July 2007. [26] ICIS Chemical Business. Chemical profile: formic acid; July 2006. [27] Andress RJ, Huang X, Bequette BW, Martin LL. A systematic methodology for the evaluation of alternative thermochemical cycles for hydrogen production. Int J Hydrogen Energy 2009;34(9):4146e54. [28] Holiastos K, Manousiouthakis V. Automatic synthesis of thermodynamically feasible reaction clusters. AIChE J 1998; 44(1):164e73. [29] Andress RJ, Martin LL. On the synthesis of hydrogen producing alternative thermochemical cycles with electrochemical steps. Int J Hydrogen Energy 2010;35(3): 958e65. [30] Holiastos K, Manousiouthakis V. Minimum hot/cold/electric utility cost for heat exchange networks. Comput Chem Eng 2002;26(1):3e16. [31] Linnhoff B, Hindmarsh E. The pinch design method for heat exchanger networks. Chem Eng Sci 1983;38(5):745e63. [32] Linnhoff B. Pinch analysis e a state-of-the-art overview. Chem Eng Res Des 1993 Sep;71(A5):503e22. [33] Armond JW, Webber DA, Smith KC, inventors; US Patent 4,190,424. Gas separation; 1980. [34] Richter E, Knoblauch K, Schlegel R, Korbacher W, inventors; U.S. Patent 4,566,881. Process and apparatus for producing oxygen with a low proportion of argon from air; 1986. [35] Brown LF. A comparative study of fuels for on-board hydrogen production for fuel-cell-powered automobiles. Int J Hydrogen Energy 2001;26(4):381e97. [36] Steinberg M, Cheng HC. Modern and prospective technologies for hydrogen production from fossil fuels. Int J Hydrogen Energy 1989;14(11):797e820. [37] Santos JC, Cruz P, Regala T, Magalhaes FD, Mendes A. Highpurity oxygen production by pressure swing adsorption. Ind Eng Chem Res 2007;46(2):591e9. [38] Hayashi S, Kawai M, Kaneko T. Dynamics of high purity oxygen PSA. Gas Sep Purif 1996;10(1):19e23. [39] Jee J-G, Kim M-B, Lee C-H. Pressure swing adsorption processes to purify oxygen using a carbon molecular sieve. Chem Eng Sci 2005;60(3):869e82.

12868

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 1 2 8 5 3 e1 2 8 6 8

[40] Vansant EF, Dewolfs R. Gas separation technology. In: Proceedings of the International Symposium on Gas Separation Technology, Antwerp, Belgium, September 10e15, 1989. Amsterdam, Netherlands; New York: Elsevier; 1990. [41] Gassner F, Leitner W. Hydrogenation of carbon dioxide to formic acid using water-soluble rhodium catalysts. J Chem Soc Chem Commun 1993;19:1465e6. [42] Jessop PG, Ikariya T, Noyori R. Homogeneous hydrogenation of carbon dioxide. Chem Rev 1995;95(2):259e72. [43] Jessop PG, Joo´ F, Tai C-C. Recent advances in the homogeneous hydrogenation of carbon dioxide. Coord Chem Rev 2004;248(21e24):2425e42. [44] Khan MMT, Halligudi SB, Shukla S. Reduction of CO2 by molecular hydrogen to formic acid and formaldehyde and their decomposition to CO and H2O. J Mol Catal 1989;57(1): 47e60. [45] Elek J, Na´dasdi L, Papp G, Laurenczy G, Joo´ F. Homogeneous hydrogenation of carbon dioxide and bicarbonate in aqueous solution catalyzed by water-soluble ruthenium(II) phosphine complexes. Appl Catal, A 2003;255(1):59e67. [46] Himeda Y, Onozawa-Komatsuzaki N, Sugihara H, Arakawa H, Kasuga K. Transfer hydrogenation of a variety of ketones catalyzed by rhodium complexes in aqueous solution and their application to asymmetric reduction using chiral Schiff base ligands. J Mol Catal A Chem 2003;195(1e2): 95e100. [47] Honeywell. UniSim thermo reference guide; 2008.

[48] Tindall BM, King DL. Designing steam reformers for hydrogen production. Hydrocarb Process Int Ed 1994;73(7): 69e75. [49] Rajesh JK, Gupta SK, Rangaiah GP, Ray AK. Multi-objective optimization of industrial hydrogen plants. Chem Eng Sci 2001;56(3):999e1010. [50] Hufton JR, Mayorga S, Sircar S. Sorption-enhanced reaction process for hydrogen production. AIChE J 1999;45(2):248e56. [51] Leiby SM. Options for refinery hydrogen. PEP Report No. 212. Menlo Park, CA: Process Economics Program. SRI International; 1994. [52] Energy Information Administration. Natural gas monthly; July 2011. [53] Energy Information Administration. Weekly U.S. retail gasoline prices, regular grade; July 2011. [54] Pena Lopez J, Manousiouthakis VI. A solid-gas ideal Gibbs reactor model for natural gas reforming in the presence of CaO. In: AIChE Spring National Meeting, paper 114d, San Antonio, TX; 2010. [55] Duraiswamy K, Chellappa A, Smith G, Liu Y, Li M. Development of a high-efficiency hydrogen generator for fuel cells for distributed power generation. Int J Hydrogen Energy 2010;35(17):8962e9. [56] Beaver MG, Caram HS, Sircar S. Sorption enhanced reaction process for direct production of fuel-cell grade hydrogen by low temperature catalytic steamemethane reforming. J Power Sources 2010;195(7):1998e2002.