15567265.2013.776153.pdf
TRANSCRIPT
-
7/28/2019 15567265.2013.776153.pdf
1/13
This article was downloaded by: [National Chemical Laboratory]On: 12 July 2013, At: 03:37Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Nanoscale and Microscale
Thermophysical EngineeringPublication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/umte20
Carbonation Characteristics of IsolatedCalcium Oxide Nanoparticles for Thermal
Energy Storage
Krishna V. Valavala a , Hongxiang Tian a & Sanjiv Sinha aa
Department of Mechanical Science and Engineering , University of
Illinois at Urbana Champaign , Urbana , Illinois
Published online: 27 Jun 2013.
To cite this article: Krishna V. Valavala , Hongxiang Tian & Sanjiv Sinha (2013) Carbonation
Characteristics of Isolated Calcium Oxide Nanoparticles for Thermal Energy Storage, Nanoscale and
Microscale Thermophysical Engineering, 17:3, 204-215, DOI: 10.1080/15567265.2013.776153
To link to this article: http://dx.doi.org/10.1080/15567265.2013.776153
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (theContent) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.
This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
http://www.tandfonline.com/page/terms-and-conditionshttp://www.tandfonline.com/page/terms-and-conditionshttp://dx.doi.org/10.1080/15567265.2013.776153http://www.tandfonline.com/action/showCitFormats?doi=10.1080/15567265.2013.776153http://www.tandfonline.com/loi/umte20 -
7/28/2019 15567265.2013.776153.pdf
2/13
Nanoscale and Microscale Thermophysical Engineering, 17: 204215, 2013
Copyright Taylor & Francis Group, LLC
ISSN: 1556-7265 print / 1556-7273 online
DOI: 10.1080/15567265.2013.776153
CARBONATION CHARACTERISTICS OF ISOLATEDCALCIUM OXIDE NANOPARTICLES FOR THERMALENERGY STORAGE
Krishna V. Valavala, Hongxiang Tian, and Sanjiv SinhaDepartment of Mechanical Science and Engineering, University of Illinois at
Urbana Champaign, Urbana, Illinois
Carbonation of lime is an attractive system for thermal energy storage. The typical use of
10-m particles creates well-known recycling problems. We propose the use of isolatednanoparticles dispersed over a high surface area to avoid sintering and model the conversion
time characteristics of such particles. Our calculations show that reactions on the surface
dominate in particles smaller than 70 nm, leading to fast conversion times. We compare
our predictions against experimental data on micrometer- and nanometer-size particles to
establish the validity of the model. This work shows that isolated nanoparticles arranged in
scaffolds are an ideal system for realizing fast and repeatable conversion for thermal storage
with high storage density.
KEY WORDS: porous CaO nano-particles, carbon sequestration, energy storage, random
pore model, shrinking core model, carbonation
INTRODUCTION
The recent focus on thermal energy storage [1] stems from the widespread availability
of waste heat and the inadequacies of current energy storage technologies. Thermochemical
energy storage has a potential cost advantage over electrochemical and mechanical storage
based on simple chemistries of abundant materials. Among thermochemical storage
systems, carbonation of lime provides one of the largest heats of reaction at 1.78 MJ/kgand is an attractive process for thermal energy storage in applications such as concentrated
solar power generation [2]. Though widely deployed in industrial processes and extensively
studied, carbonation suffers from poor recycling characteristics [3]. Unlike its reverse
calcination reaction, the carbonation reaction does not reach completion and the quality
of calcium oxide as a CO2 sorbent declines over repeated calcinationcarbonation cycles
[4]. In this work, we show that the incompleteness of the carbonation reaction is inherently
linked to the particle size: Nanoparticles exhibit fast conversion times and complete
reactions. We show that the loss of surface area caused by sintering results in reduced
Manuscript received 3 October 2012; accepted 10 February 2013.
The authors acknowledge support from the National Science Foundation under grant NSF-CBET-09-54696-
CAREER.
Address correspondence to Krishna V. Valavala, Department of Mechanical Science and Engineering,
University of Illinois at Urbana Champaign, 235 W. Mechanical Engineering Building, 1206 W. Green St., Urbana,
IL 61801. E-mail: [email protected]
204
-
7/28/2019 15567265.2013.776153.pdf
3/13
CARBONATION OF LIME 205
NOMENCLATURE
C concentration of CO2, mol/m3
De intraparticle diffusivity, m2/s
DP product layer diffusivityk reaction constant, m4/mol/s
km convective mass transfer coefficient,
m/s
L length of the pores per unit volume of
space, 1/m2
M molecular weight of CaO, kg/mol
R radius of the particle, m
S area of the unreacted CaO surface in the
pores per unit volume of space, 1/m
Sh Sherwood number
t time, s
V volume enclosed in the pores by the
unreacted CaO surface per unit volume
of space
X ratio of volume of CaO consumed to
initially available CaO at a given radial
position
X ratio of total volume of CaO consumed
to the total initial volume of CaO avail-
ableZ ratio of molar volumes of CaCO3 to
CaO
Greek Symbols
product layer thickness, m
nondimensional radial position, r/R0 density of CaO, kg/m3
Subscripts and Superscript
0 value at the beginning of the reaction
b bulk stream
e at equilibrium
i CaOCaCO3 interface
s outer surface nondimensionalized
reactivity. Sintering changes the surface texture over multiple cycles with the formation of
macroscopic pores as well as the shrinkage of microscopic pores. Such structural changes
largely dictate the decline by more than 60% over as few as 10 cycles [5, 6].Several methods such as sorbent doping, hydration, and preparation by wet precipita-
tion have been suggested [7, 8], with limited success to increase the reactivity and to sustain
it over repeated cycles. Based on our theoretical predictions, we suggest the use of isolated
nanoparticles dispersed on a scaffold to solve this issue in principle. Sub-100-nm isolated
particles offer the distinct advantage of short heat (4 ns) and mass diffusion times (8 nsfor CO2 in CaCO3). In 1974, Barker [9] experimentally demonstrated enhanced reactivity
of10 nm size particles in a powder to93% with no detectable decline over 30 cycles in a24-h period. More recent work [10] has demonstrated high conversion for nanometer-sized
CaO particles at even faster rates. Experiments on nanoparticles reported in the literature
[4, 5] did not yield repeatability in conversion over a large number of cycles mainly due tothe use of powdered samples that sinter easily. Advanced manufacturing techniques such as
electrohydrodynamic printing [11] can disperse isolated nanoparticles on textured surfaces
with high areas, avoiding sintering.
Figure 1 illustrates a conceptual scaffold that holds calcium oxide particles with-
out sintering. The scaffolding mesh structure has micrometer-scale pillars to increase the
surface area per volume of the structure. Calcium oxide particles are placed on the micropil-
lars such that the successive particles are separated by a distance greater than the expected
radial expansion of the particles in the carbonation reaction. This avoids sintering during
carbonation. We estimate an 30% increase in size in a 100-nm particle. Therefore, 100-nm particles should be placed with a minimum center-to-center gap of 130 nm to prevent
sintering. In this article, we assume that the calcium oxide particle held in a scaffoldingstructure could be treated as an isolated particle suspended in a gas mixture containing
-
7/28/2019 15567265.2013.776153.pdf
4/13
206 K.V. VALAVALA ET AL.
Figure 1 Concept design of a scaffolding structure to hold CaO particles without sintering. The supporting
structure has a high surface area per unit volume due to the micropillars shown in the figure. The distance
between any two successive calcium oxide particles should be greater than the radial expansion of the particle
upon carbonation to prevent sintering.
carbon dioxide. The model and the results presented in this article advance the use of iso-
lated sub-micrometer-sized lime particles in calcium looping as a means to achieve carbon
sequestration and thermochemical energy storage.
THEORY
An estimate of the carbonation time characteristics of CaO nanoparticles requires
consideration of the pore structure, CO2 diffusion rates, and reaction kinetics. The seminal
random pore model is a useful starting point in analyzing conversion behavior. In their
earliest version of the random pore model [12], Bhatia and Perlmutter assumed kinetic
control for reactions inside pores and correlated the local reaction extent X to the pore
structure parameter as X= 1 exp
1 +
4
, where is a dimensionless time
incorporating kinetics and porosity. Assuming uniform reaction extent within the particle
and a kinetic regime for reactions on the external surface, the overall conversion was given
as X= 1 1 3 exp1 + 4 , where the new parameter includes the effectof particle size. However, the pores inside CaO particles result in a CO2 concentration
gradient, leading to a nonuniform reaction extent within the particle, thereby invalidating
the assumption of uniform reaction extent. The kinetic control regime assumption is also
invalid for CaO carbonation reactions because CO2 diffusion through the product layer
controls the reaction rate. For these reasons, the expression above cannot be used to study
the particle size dependence of CaO carbonation reaction times. In a later version of the
random pore model [13], the effects of intraparticle diffusion resistance and the product
layer diffusion resistance were incorporated to account for the nonuniform reaction extent
X inside the pores. However, the reaction on the external surface of the particle, which is
important at nanometer scales, was not considered in calculating the overall reaction extent.
In our model, we calculate conversion in the pores as well as on the external sur-
face considering all possible resistances involved in the reaction. We calculate the reaction
-
7/28/2019 15567265.2013.776153.pdf
5/13
CARBONATION OF LIME 207
extent in the pores X(, t)using the random pore model and combine it with reaction on the
external surface calculated using the shrinking core model to obtain the overall conversion
X(t) as
X(t) = 1 3Ri (t)
R0
=0(1 X(, t))2d, (1)
where is the radial position nondimensionalized with respect to the particle radius RO,
and Ri(t) is the position of the external reacting surface at time t obtained by using the
shrinking core model.
Diffusion and Conversion Inside the Nanoparticle
In this section, we describe the random pore model used to calculate the reaction
extent in the core of the nanoparticle. We assume that the CaO nanoparticle is a porous
sphere exposed to a carbon dioxide and nitrogen gas mixture at a temperature of 650C and1 atm pressure with a CO2 mole fraction of 0.15.
As depicted in Figure 2, the concentration of CO2 decreases from Cb in the bulk
stream to Cs at the particle surface due to external mass transfer resistance, Rm,ER. It fur-
ther decreases to C(, t) due to intrapellet resistance, Rm,IPR, offered by the pores inside the
Figure 2 Schematic of a CaO nanoparticle suspended in a mixture of CO 2 and N2 environment at 650C. Carbon
dioxide gas diffuses into the particle through the pores (white) and the calcium carbonate product layer (grey) to
reach the reaction surface of calcium oxide (black) where the carbonation reaction takes place. Notations used for
concentration of CO2 at various positions with respect to the particle are shown along the horizontal line at thebottom of the figure.
-
7/28/2019 15567265.2013.776153.pdf
6/13
208 K.V. VALAVALA ET AL.
particle. Finally, the concentration drops to Ci(, t) at the reaction surface inside the pores
due to resistance from the CaCO3 product layer, Rm,PLR. The concentrations are nondimen-
sionalized using bulk concentration Cb as the normalizing quantity and are represented with
asterisks.
We assume pseudo-steady-state conditions during the carbonation reaction. Thisimplies that the change in CO2 concentration at any position over a short period of time
due to the reaction is negligible compared to the CO 2 flow rate. Under this assumption, the
diffusionconsumption equation for CO2 diffusion and the subsequent reaction is
1
2
2De
C
= R
2o
MDe0Cb(1 V0)
dX(, t)
dt. (2)
The associated boundary conditions reflecting the spherical symmetry and external mass
transfer resistance are
C
=0
= 0 (3)
C
=1
= Sh(1 C)
De. (4)
Solving Eqs. (2)(4) provides the local conversion profile X(, t), which can be used to
evaluate the overall conversion time characteristics.
We now describe the variables that are present in the above equations and provide
expressions to evaluate these variables. The initial intrapellet diffusivity De0 at the begin-ning of the reaction is defined as the harmonic mean of molecular diffusivity DM of CO2and Knudsen diffusivity DK through the pores
1
De0= 1
DM+ 1
DK. (5)
The intrapellet diffusivity De (, t) normalized with respect to its initial value De0 isexpressed in terms of the local conversion, X(, t) as
De (, t) =De
De0 = 1 (Z
1)(1
V0)X(, t)
V02
. (6)
The Sherwood number (Sh) in Eq. (4) defined as Sh = kmRO/De0 requires evaluation ofthe convective mass transfer coefficient km. The continuum flow regime expressions for Sh
cannot be used here because the mean free path of the gas is comparable to size of the
particle. The mean free path of CO2N2 gas mixture at 650C and 1 atm pressure is 0.3
m and the Knudsen number varies from, Kn 3 to 0.3 as the particle size varies from100 nm to 1 m. In our calculations, we use an interpolation formula for effective mass
transfer coefficient that is valid across various flow regimes [14, 15]:
km = DMRo
1 + Kn
43 Kn + 1.0161
43
Kn + 11
. (7)
-
7/28/2019 15567265.2013.776153.pdf
7/13
CARBONATION OF LIME 209
The rate of conversion dX/dt in Eq. (2) is expressed in terms of the concentration profile
C(, t) by correlating changes in the pore structure to the chemical kinetics of the reactionusing modifications to the random pore model [13]. The random pore model assumes that
pores are cylindrical with size distribution f(r), where f(r) is the length of pores with radius
in the range r and r+ dr. By assuming that no pores are formed or destroyed during thereaction, the following equation can be written to balance the number of pores
f
t+
t
f
dr
dt
= 0. (8)
Assuming that the reaction is of first order with respect to the concentration of CO 2 and is
proportional to the surface area of the solid reactant CaO available, the rate of change of
the pore radius is given by
dr
dt = k(Ci Ce), (9)
where k is the rate constant of the carbonation reaction, and Ce is the CO2 equilibrium
concentration in the carbonationcalcination reaction. Total pore length LE, pore surface
area SE, and pore volume VE per unit volume of space are obtained in terms of their initial
values L0, S0, and V0 by multiplying Eq. (8) by 1, 2r, and r2, respectively, and integrating
over the pore radius from zero to infinity. LE, SE, and VE are then corrected to account for
pore overlap [16] to obtain corresponding corrected expressions L, S, and V. The corrected
expression for the volume enclosed by the unreacted CaO surface V is given by
V= 1 (1 V0)expkCb t
0
(Ci
Ce
)dt
1 V0 . (S0 + kL0Cb t0(Ci Ce )dt)
. (10)
In the random pore model, the concentration of CO2 at the unreacted CaO surface, Ci, is
assumed to be independent of time and the integration is written as Ci t. Here we accountfor the changing concentration using t0(Ci Ce )dtinstead.
Using Eq. (10), the local conversion X(, t) inside the particle at the radial posi-
tion is
1 X(, t) = (1 V)(1 V0)
= expkCb t0(Ci Ce )dt
1 V0. (S0 + kL0Cb t0(Ci Ce )dt)
.
(11)
Differentiating Eq. (11) with respect to tprovides the local conversion rate dX(, t)/dtas
dX(, t)
dt= kCbSo
(1 V0)(Ci Ce )(1 X(, t))
1 + 2kL0Cb
t0(C
i Ce )dt
So
. (12)
Equation (12) expresses dX(, t)/dt as a function ofCi , and the left-hand side of Eq. (2) isa function ofC. Therefore, in order to be able to use dX(, t)/dt from Eq. (12) in Eq. (2),
Ci needs to be expressed in terms of C. This relation is obtained [13] by analyzing the
resistance offered by the CaCO3 product layer to the flow of CO2 from the CaCO3 surface
to the unreacted CaO surface. Assuming that the product layer thickness is small, growthof the product layer can be expressed as
-
7/28/2019 15567265.2013.776153.pdf
8/13
210 K.V. VALAVALA ET AL.
= ZkCbt
0
(Ci Ce )dt. (13)
Balancing the CO2 diffusion rate through the product layer of diffusivity Dp and the rate of
consumption of CO2 at the CaCO3CaO interface, we get
Dp(C Ci )
= k(C
i Ce )
M. (14)
Eliminating from Eqs. (13) and (14), the concentration of CO2 at the reaction surface
Ci can be related to the concentration of CO2 in the pores, C. Now, Eqs. (2)(4) can be
discretized and solved using the control volume method for concentration profiles C(, t),which, in turn, provides the local conversion profile X(, t) in the pores.
Conversion at the External Surface
As the reaction proceeds, the outer CaO reacting surface recedes from RO to Ri(t).
Because the molar volume of CaCO3 formed is greater than that of CaO, the particle radius
itself expands to Rs(t), given by
4
3 (Rs(t)
3 Ri(t)3) = Z4
3(R3O Ri(t)3). (15)
As discussed previously, the external surface reaction is incorporated in the overall calcula-
tion in the form of the upper limit Ri(t)/RO of the integral in Eq. (1). We calculate the radial
position of the external reaction surface Ri(t) by using the shrinking core model under a
pseudo-steady-state assumption [17]. The overall resistance to the flow of CO2 at time t,Hex, is [18]
Hex =1
4
1
kmR2O
+ 1DP
1
Ri(t) 1
Rs(t)
+ M
kRi(t)2
. (16)
The three terms in Eq. (15) are the external mass transfer resistance, product layer diffu-
sion resistance, and carbonation reaction resistance, respectively. The external resistance in
Eq. (16) was calculated using the convective mass transfer coefficient given in Eq. (7) and
was found to be 8 orders of magnitude smaller than other resistances even for particles as
small as 50 nm. The rate of mole consumption of CO2 during the time interval t to t+
dt
is now given by (Cb Ce)/Hex. Stoichiometry considerations provide the rate of change ofthe radial position of the CaCO3CaO interface as
Cb CeHex
= M
d
dt
4R3
3
Ri
= 4 M
Ri(t)2 dRi(t)
dt. (17)
RESULTS
The equations presented in the previous section were numerically solved by using
the control volume method [19]. The rate and diffusion constants were obtained from
experimental results [2022]. We validated the calculation methodology by matching ournumerical results to the analytical results obtained in the random pore model at particle
-
7/28/2019 15567265.2013.776153.pdf
9/13
CARBONATION OF LIME 211
sizes above 1 m. In this section, we compare our numerical results with experimental
data for both micrometer- and nanometer-sized particles. We follow this comparison with
a discussion on relative contribution of various flow resistances to the overall conversion
characteristics at various particle sizes.
For the purpose of comparing our theoretical model with experimental results, wechose two data sets [23, 24] in which the particle size is in the regime of interest and
information regarding pore distribution, carbonation conditions, and the conversion time
characteristics are available. The first data set was from the experiments of Martavaltzi
and Lemonidou [24] on pure nonisolated CaO obtained from calcium acetate with a par-
ticle radius of 550 nm. The second data set was from the work of Broda and Mller
[23], whose sorbent preparation method produces particles of radius 85 nm with mayen-
ite (Ca12Al14O33) as the inert supporting material. The weight percentage gain data in the
second publication were converted to conversion time data by assuming that all Al 2O3 was
converted to mayenite and, therefore, the remaining CaO weight percentage in the sorbent
was 69%.Figure 3 compares predictions from the model with experimental results. Because the
flow rate of CO2 increases from zero to its steady value in a thermogravimetric analyzer, the
initial reaction rates from experiments were smaller than those predicted at the beginning
stages of the reaction. The experiments did not provide measurements of product layer
diffusivity. Because the product layer diffusivity value is not universal but varies depending
on the source of CaO, precursor type, and preparation technique [25], we used the product
layer diffusivity as a fitting parameter in these calculations. The values used for the fit
were close to the experimentally determined values of product layer diffusivities [20, 21,
26], thereby establishing the validity of our model. The fitted activation energy and the
preexponential factor for the product layer diffusivity were 178.5 KJ/mol and 2.25
105
m2/s respectively for the first data set, yielding a diffusivity of 4.7 1015 m2/s at 690C,which is very close to the corresponding value of 4.85 1015 m2/s obtained at the same
Figure 3 Comparison of conversion time characteristics from our theoretical model and experimental results for
CaO particles with a radius of 85 nm (squares) and a radius of 550 nm (circles). The advantage of using isolatednanoparticles over nonisolated micrometer-sized particles can be seen from the results shown in the figure.
-
7/28/2019 15567265.2013.776153.pdf
10/13
212 K.V. VALAVALA ET AL.
temperature by Grasa et al. [21]. The fit to the second data set employed the same activation
energy and the preexponential factor was 8 105 m2/s. We suspect that the higher valueof product layer diffusivity in the alumina-supported CaO particles compared to the pristine
particles could arise from additional CO2 diffusion through unfused CaO grain boundaries.
The results in Figure 3 highlight the advantage of using nanometer-sized particles that aresupported by an inert framework over unsupported micrometer-sized particles.
We now investigate the impact of decreasing size on intraparticle diffusion. Diffusion
inside pores results in carbonation proceeding faster away from the center of the par-
ticle. For large particle sizes, the CO2 concentration gradient is significant and leads to
nonuniform reaction rates, decreasing the overall conversion. Figure 4 shows the effect of
intraparticle resistance on the conversion time characteristics for three particle sizes. The
solid lines in the figure were obtained by assuming the initial intraparticle diffusivity, Deo,
to be infinite. Under this assumption, the concentration of CO2 and therefore the reaction
extent are uniform throughout the particle. Thus, CaO particles reach complete conversion
regardless of size under this condition. However, smaller particles reach completion faster.This is due to reaction on the external surface, discussed later in this section.
The dotted lines in Figure 4 were calculated using the value of initial intraparticle
diffusivity obtained from Eq. (12). With finite intraparticle diffusivity, the concentration
of the CO2 decreased as we move from the external surface to the center of the particle.
Consequently, the local reaction rate was slower at the center of the particle and was faster
away from the center. In time, pores further away from the center of the particle closed
down and the reaction remained incomplete near the center. This was evident in particles
of size greater than 1.5 m. For example, in particles of size 5 m, conversion stopped
at around 75%. A finite intraparticle diffusivity did not lead to pore closure in particles
smaller than
1.5 m. Thus, pore closure is not an important issue with nanoparticles.
The large surface area available in nanoparticles significantly accelerates conversion.Based on measurements of specific surface area of pores [12], the ratio of external area to
Figure 4 Conversion time characteristics for isolated CaO particles with radii of 50 nm, 100 nm, and 5 m with
finite (dotted lines) and infinite (solid lines) intraparticle diffusivity. Particles larger than 1.5m experience poreclosure.
-
7/28/2019 15567265.2013.776153.pdf
11/13
CARBONATION OF LIME 213
Figure 5 Influence of external surface reaction on conversion time at various particle sizes. Reaction on the
external surface contributes significantly toward the overall conversion rate for particle size less than 500 nm.
pore surface increased twofold from 0.7 to 1.4 as the particle size decreased from 100 to
50 nm. Figure 5 shows the difference in times taken to reach half conversion with and with-
out inclusion of the reaction on the outer surface. The solid horizontal line corresponds to
the original random pore model and shows that the conversion time remained independent
of particle size, depending only on the pore structure. The dashed line represents the half
times calculated using the modified model including external surface reaction. It can beseen from Figure 5 that conversion time for a 100 nm particle is five times lower than that
for a 3 m particle due to availability of external surface area. This advantage is lost when
particles sinter, resulting in low conversion of nanoparticles [5].
Finally, we comment on thermal energy storage density for a system of dispersed
nanoparticles. The theoretical limit of storage density for CaCO3 is 4.8 GJ/m3. Recentadvances in the fabrication of high-surface-area alumina templates [27] make these an
ideal scaffold for dispersing nanoparticles. The hexagonal cells (50-nm periodicity) in thealumina template enable large area fraction coverage (>75%) of nanoparticles. Using two
particles per pore led to an effective storage density >0.2 GJ/m3, excluding CO2 storage.
CONCLUSION
In conclusion, we investigated the theoretical carbonation of CaO nanoparticles for
thermochemical storage using a modified random pore model. Comparisons with exper-
imental data validated the proposed model and clearly highlight the advantage of using
isolated nanoparticles. Our model further provides insight into three important advantages
of using isolated nanoparticles to achieve fast and repeatable conversions. First, pore clo-
sure is unimportant for particles smaller than1.5m. Second, reaction on external surfacedominates over reaction in pores in particles smaller than70 nm. Irrespective of the domi-nating reacting surface, sub-micrometer sized particles achieve complete conversion. Third,
subcontinuum transport of the reactant to the nanoparticle surface does not pose any limi-tation on the conversion time, even when the Knudsen number is 6 (for a 50-nm particle).
-
7/28/2019 15567265.2013.776153.pdf
12/13
214 K.V. VALAVALA ET AL.
All of the above advantages of using nanoparticles are lost if particles are allowed to sinter.
However, the use of templates or scaffolds potentially avoids this issue and represents an
important direction for future experimental work in this area.
REFERENCES
1. I. Gur, K. Sawyer, and R. Prasher, Searching for a Better Thermal Battery, Science, Vol. 335,
pp. 14541455, 2012.
2. M. Bosatra, F. Fazi, P. Lionetto, and L. Travagnin, Utility Scale PV and CSP Solar Power Plants,
presented at Power Gen Europe-2010, Amsterdam, The Netherlands, June 810, 2010.
3. J.C. Abanades and D. Alvarez, Conversion Limits in the Reaction of CO2 with Lime, Energy
and Fuels, Vol. 17, pp. 308315, 2003.
4. N.H. Florin and A.T. Harris, Preparation and Characterization of a Tailored Carbon Dioxide
Sorbent for Enhanced Hydrogen Synthesis in Biomass Gasifiers, Industrial & EngineeringChemistry Research, Vol. 47, pp. 21912202, 2008.
5. C. Luo, C. Luo, Q. Shen, N. Ding, Z. Feng, Y. Zheng, C. Zheng, Morphological Changes of Pure
Micro- and Nano-Sized CaCO3 during a Calcium Looping Cycle for CO2 Capture, Chemical
Engineering & Technology, Vol. 35, pp. 547554, 2012.
6. H. Lu, P.G. Smirniotis, F.O. Ernst, and S.E. Pratsinis, Nanostructured Ca-Based Sorbents with
High CO2 Uptake Efficiency, Chemical Engineering Science, Vol. 64, pp. 19361943, 2009.
7. J. Blamey, E.J. Anthony, J. Wang, and P.S. Fennell, The Calcium Looping Cycle for Large-Scale
CO2 Capture, Progress in Energy and Combustion Science, Vol. 36, pp. 260279, 2010.
8. U. Rockenfeller, L.D. Kirol, K. Khalili, J.W. Langeliers, and W.T.Dooley, Heat and Mass
Transfer Apparatus and Method for SolidVapor Sorption Systems, U.S. Patent 6224842, 2001.
9. R. Barker, The Reactivity of Calcium Oxide towards Carbon Dioxide and Its Use for Energy
Storage, Journal of Applied Chemistry and Biotechnology, Vol. 24, pp. 221227, 1974.10. N.H. Florin and A.T. Harris, Reactivity of CaO Derived from Nano-Sized CaCO3 Particles
through Multiple CO2 Capture-and-Release Cycles, Chemical Engineering Science, Vol. 64,
pp. 187191, 2009.
11. S. Mishra, High-Speed and Drop-on-Demand Printing with a Pulsed Electrohydrodynamic Jet,
Journal of Micromechanics and Microengineering, Vol. 20, pp. 095026, 2010.
12. S.K. Bhatia and D.D. Perlmutter, A Random Pore Model for FluidSolid Reactions: I.
Isothermal, Kinetic Control, AIChE Journal, Vol. 26, pp. 379386, 1980.
13. S.K. Bhatia and D.D. Perlmutter, A Random Pore Model for FluidSolid Reactions: II. Diffusion
and Transport Effects, AIChE Journal, Vol. 27, pp. 247254, 1981.
14. N.A. Fuks and A.G. Sutugin, Highly Dispersed Aerosols, Ann Arbor Science Publishers, Ann
Arbor, MI, 1970.15. S. Loyalka, Condensation on a Spherical Droplet, II, Journal of Colloid and Interface Science,
Vol. 87, pp. 216224, 1982.
16. M. Avrami, Kinetics of Phase Change. II Transformation-Time Relations for Random
Distribution of Nuclei, Journal of Chemical Physics, Vol. 8, pp. 212224, 1940.
17. C.Y. Wen, Noncatalytic Heterogeneous SolidFluid Reaction Models, Industrial and
Engineering Chemistry, Vol. 60, pp. 3454, 1968.
18. O. Levenspiel, Chemical Reaction Engineering, Wiley, New York, 1972.
19. S.V. Patankar, Numerical Heat Transfer and Fluid Flow, Hemisphere Publishing, Washington,
DC, 1980.
20. S.K. Bhatia and D.D. Perlmutter, Effect of the Product Layer on the Kinetics of the CO2Lime
Reaction, AIChE Journal, Vol. 29, pp. 7986, 1983.
21. G. Grasa, R. Murillo, M. Alonso, and J.C. Abanades, Application of the Random Pore Model tothe Carbonation Cyclic Reaction, AIChE Journal, Vol. 55, pp. 12461255, 2009.
-
7/28/2019 15567265.2013.776153.pdf
13/13
CARBONATION OF LIME 215
22. C. Ellis and J. Holsen, Diffusion Coefficients for HE-N2 and N2-CO2 at Elevated Temperatures,
Industrial & Engineering Chemistry Fundamentals, Vol. 8, pp. 787791, 1969.
23. M. Broda and C.R. Mller, Synthesis of Highly Efficient, Ca-Based, Al2O3-Stabilized, Carbon
Gel-Templated CO2 Sorbents, Advanced Materials, Vol. 24, pp. 30593064, 2012.
24. C.S. Martavaltzi and A.A. Lemonidou, Development of New CaO Based Sorbent Materialsfor CO2 Removal at High Temperature, Microporous and Mesoporous Materials, Vol. 110, pp.
119127, 2008.
25. R.H. Borgwardt, K.R. Bruce, and J. Blake,An Investigation of Product-Layer Diffusivity
for Calcium Oxide Sulfation, Industrial & Engineering Chemistry Research, Vol. 26, pp.
19931998, 1987.
26. D. Mess, A.F. Sarofim, and J.P. Longwell, Product Layer Diffusion during the Reaction of
Calcium Oxide with Carbon Dioxide, Energy and Fuels, Vol. 13, pp. 9991005, 1999.
27. H. Masuda, Highly Ordered Nanochannel-Array Architecture in Anodic Alumina, Applied
Physics Letters, Vol. 71, pp. 27702772, 1997.