Moulson A.J., Herbert J.M. Electroceramics: Materials, Properties, Applications

Подождите немного. Документ загружается.

However, the introduction in 1913 by Ford of the mass-produced car completely

ousted the battery for vehicle traction. The reason is clearly demonstrated in Fig.

4.25. More precise data well illustrate the comparison; for example, one litre of a

modern lead/acid battery weighs 2.4 kg and can store 0.07 kW h, suﬃcient to

drive a vehicle a few hundred metres. By comparison one litre of petrol weighs

0.85 kg and stores 11 kW h of energy, suﬃcient to drive the vehicle approximately

10 km.

Strong contenders for automotive power are the sodium/sulphur and sodium/

nickel chloride batteries, the latter known as the ZEBRA cell. ‘ZEBRA’ was

originally (c. 1979) an acronym devised for commercial security reasons but now

it stands for the very apt ‘Zero Emissions Batteries Research Activity’. Several

European car manufacturers including BMW and Mercedes have incorporated

the ZEBRA cell into prototype cars, vans and buses. The performance of the

battery far outstrips that of the lead/acid counterpart, as is evident from Fig.

4.25. Its technical performance coupled to features relating to safety and overall

economics give grounds for optimism concerning its future as an automotive

power source. It should prove especially attractive for applications such as in the

‘neighbourhood electric vehicle’ (NEV) where reasonable range and not

acceleration is the signiﬁcant attribute. Such vehicles would be used for

shopping, taking children to school, leisure pursuits, etc.

A further possible and important application for the Na/S and ZEBRA cells is

in ‘load levelling’. By this is meant the storage of electrical energy from power

supply generators when demand is relatively low and its release into the supply

network (the ‘grid’) at peak demand times. In this way the generators can be run

continuously at their optimal eﬃciency speeds.

The Na/S and ZEBRA batteries, which incorporate ceramic electrolytes, will

be discussed in detail below but ﬁrst the elementary basic science is summarized.

Elementary Na/S battery science The basic science of the Na/S cell is

identical to that of the fuel cell. In both cases the essential is an electrolyte. On

one side of the electrolyte membrane there is a source of sodium ions at a

particular chemical potential, and on the other side a ‘sink’ for the ions at a

relatively lower chemical potential. The elements of the process are illustrated in

Fig. 4.28.

The ideal electrolyte has a high conductivity with Na ions being the only

charge-carrier, that is, the transport number for Na is unity. (The ‘transport

number’ for a particular mobile charged species is deﬁned as the conductivity

contributed by the species expressed as a fraction of the total conductivity.)

Consider the discharge of the cell. At the anode two Na ions enter the electrolyte

having given up an electron each; that is the sodium is chemically oxidized. At

the cathode a polysulphide unit gains two electrons and is chemically

reduced.

182 CERAMIC CONDUCTORS

TEAMFLY

Team-Fly

This process leads to the development of a voltage diﬀerence between anode

and cathode and when the external circuit is ‘open’, this potential diﬀerence is the

‘electromotive force’ (the e.m.f.) of the cell. When there is an electrical load in the

external circuit, for example a motor, then a current will ﬂow and

electrochemical energy is converted into mechanical energy.

Just as in the case of the fuel cell, current also has to ﬂow through the

electrolyte doing work against its electrical resistance, generating heat and

leading to a terminal p.d. value lower than the e.m.f.

It follows from this brief outline that if maximum useful power is to be

extracted from the cell then the electrolyte resistance must be kept to a minimum.

This is dependent not only on the material but also on the geometry of the

membrane which must be as thin as is practicable. As in the case of the fuel cell,

the e.m.f can be calculated from thermodynamic principles. In fact the chemical

reactions occurring when the sodium ions react with the sulphur are rather

complex and the sodium:sulphur ratios change as the battery discharges. The

ﬁrst product as the cell discharges is the compound Na

At the anode the half cell reaction may be written:

2Na ! 2Na

þ 2e

(4:34Þ

and at the cathode

5S þ 2e

þ 2Na

! Na

(4:35Þ

The overall cell reaction is, therefore

2Na þ 5S ! Na

(4:36Þ

FUEL CELLS AND BATTERIES 183

Fig. 4.28 Elements of the Na/S battery.

When all the sulphur has been converted to Na

this compound is gradually

converted to Na

as the discharge proceeds. The sodium/sulphur cell operates

at a temperature of approximately 350 8C (the melting points of both Na

and

are close to 240 8C).

There has been a reduction in free energy of the chemical system which is

manifested as electrical energy. Under the assumption that the chemical reaction

occurs reversibly, that is no energy is ‘lost’ but only exchanged between chemical

and electrical forms, the cell e.m.f. can be calculated. The change in free energy

(DG) for reaction (4.36) at 350 8C is approximately 400 kJ. For each mole of the

sulphide formed two moles of Na

ions are transported across the cell. The

calculation (Eq. (4.33)) leads to a cell e.m.f. of 2.08 V.

As with the fuel cell this small voltage must be multiplied up by connecting

many cells in series, and in parallel to increase current availability and battery

reliability. This is a relatively straightforward procedure with the separately

packaged individual cells being connected with metallic conductors to form a

battery.

Elementary ‘ZEBRA’ battery science There is no new principle involved.

The basic reaction is:

2Na þ NiCl

! Ni þ 2NaCl (4:37Þ

The open circuit e.m.f., calculated from the free energy change for the reaction at

250 8C is approximately 2.59 V.

4.5.3 Electroceramics for fuel cells and batteries

Solid fast-ion conductors

Fast-ion conductors: general comments The essential element of ceramics-

based fuel cells and batteries is the electrolyte, a solid, fast-ion conductor.

Ionic transport under an applied ﬁeld takes place to some extent in all ionic

solids but generally involves the movement of ions into vacant sites over large

energy barriers resulting in very low mobilities, even at high temperatures.

Certain materials known as ‘fast-ion conductors’ or ‘superionic conductors’,

exhibit higher ionic mobilities and opened up the possibility of practical

applications, particularly in fuel-cell, battery and sensor technologies. In liquid

ionic conductors both positive (cations) and negative (anions) ions act as

current-carriers, although one or the other may dominate with a higher transport

number. In solids there is usually only one species of mobile ion, and it may be a

cation or anion.

The precise structural conditions for fast-ion conduction have yet to be

established (especially when the eﬀects of microstructure are involved), but there

184 CERAMIC CONDUCTORS

clearly must be a high concentration of sites into which the mobile ion can move

(see Section 2.6.3). These may be lattice vacancies or they may be features of the

crystal structure such as tunnels or under populated planes within which speciﬁc

ions can move by a mechanism with a low activation energy.

The size and charge of mobile ions and the size of the spaces within a lattice

through which they may move are important factors. Small singly charged

cations such as Li

and Na

are mobile in a number of environments, in contrast

to the more highly charged and smaller cations. The reason for this is the

polarization of the surrounding anions by the smaller more highly charged

cations. This interaction constitutes a high-energy barrier to the movement of the

cations. However, there are factors other than geometry and size that have

important eﬀects. For example, on a rigid sphere model, the interstices between

the anions (radius r

) in a face-centred cubic lattice would only allow penetration

of a cation of radius 0.15 r

. However, in the case of I



of radius 220 pm, where

the interstices oﬀer an aperture of only 34 pm radius, the transport number of

55 pm) is signiﬁcant, and the room temperature ionic conductivity of LiI

is 510

5

1

. This can be attributed to the following factors:

. ions are not rigid spheres,

. simple concepts relating to the interaction of massive bodies are not

applicable in detail to the interaction of ions; and

. thermal energy will lead to random enlargements of the interstices and to the

creation of anion and cation vacancies facilitating the movement of ions.

With respect to the last point, it is signiﬁcant that room temperature (300 K) is

approximately 0.4 the melting point of LiI (723 K) and so the concentration of

thermally induced lattice vacancies will be high. LiI, which is used as a solid

electrolyte in cardiac pacemakers, can be contrasted with RbAg

, a fast-ion

conductor that, in polycrystalline form, has an ionic conductivity due to Ag

ions of 21 S m

1

at 25 8C. Even though silver has a large radius (r

¼115 pm), the

RbAg

structure is such as to oﬀer apertures approaching this diameter for the

ions.

The fast-ion ceramic conductors of interest here are cubic stabilized zirconia

(CSZ), an oxygen ion conductor, and a sodium aluminate (b@-alumina), a

sodium ion conductor. Both are discussed in detail below. The overview by

T.A. Ramanarayanan et al. [8] covering CSZ and its applications is

recommended.

Cubic stabilized zirconia (CSZ) Pure zirconia (ZrO

) is either chemically

extracted and purified from the mineral zircon (ZrSiO

) or purified from

baddeleyite. It occurs as three crystalline polymorphs with monoclinic,

tetragonal and cubic structures. The monoclinic form is stable up to 1170 8C

FUEL CELLS AND BATTERIES 185

when it transforms to the tetragonal modification, which remains stable up to

2370 8C; from 2370 8C to the melting point (2680 8C) the cubic form is stable.

The monoclinic-tetragonal conversion is accompanied by a contraction in

volume of approximately 5% which can cause mechanical failure in ceramic

pieces. This difficulty is overcome in CSZ by stabilizing the cubic form down to

room temperature by substituting lower-valence cations for some of the

zirconium.

Cubic ZrO

has the ﬂuorite structure with the O

2

ions arranged in simple

cubic packing and half the interstices in this lattice occupied by Zr

4þ

ions (Fig.

4.29); the substitution of lower-valence cations leads to O

2

ion vacancies as

indicated. The vacancies which stabilize the structure also lead to high mobility

in the oxygen sub-lattice and to behaviour as a fast-ion conductor.

The elements that stabilize the cubic ﬂuorite structure in zirconia include all

the lanthanides, scandium, yttrium, magnesium, calcium, manganese and

indium. The main qualiﬁcation would appear to be an ionic radius close to

that of Zr

4þ

¼84 pm). Ce

3þ

with r

¼114 pm is one of the largest ions fulﬁlling

this function. Ca

2þ

¼112 pm) is the most commonly used substituent at about

the 15 mol.% level. Yttrium (r

¼101 pm) stabilizes the cubic phase when present

in the 13–68 mol.% range, but maximum conductivity is obtained with 7–

8 mol.% which results in some admixture of tetragonal and monoclinic phases

with the cubic. Calcium is also sometimes used at the 7–8 mol.% level, resulting

186 CERAMIC CONDUCTORS

Fig. 4.29 The ideal ﬂuorite structure showing half a unit cell including a dopant cation and a

charge-compensating oxygen vacancy.

in a partially stabilized zirconia. Scandium has been found to give a material with

a higher conductivity, which is particularly valuable at lower temperatures. From

the fact that the conductivity reaches its maximum at an intermediate

concentration of lower-valence ions, it appears that only a fraction of the

vacant O

2

sites contribute to O

2

mobility when the concentration of vacant

sites is high. Ions of variable valence need to be excluded if the electronic

conductivity is to be minimized. Thorium is added to yttria-containing bodies as

a grain growth inhibitor.

For refractory applications, when the ‘mixed oxide’ route is followed,

temperatures as high as 1900 8C are necessary to sinter to high density. The

high interest in SOFCs has stimulated eﬀorts to develop novel fabrication routes

to zirconia ceramics and yttria-stabilized sub-micron sized powders have been

prepared which sinter to 95% theoretical density at a temperature as low as

1150 8C [4].

Chemically prepared, highly reactive and uniform zirconia powders are

commercially available. One production route is the co-precipitation of a mixed

solution of ZrOCl

(derived from zircon) and YCl

(derived from yttria) solution

followed by calcinations, milling and spray-drying. The best commercially

available, chemically prepared, powders of 8% Y

/ZrO

(8YSZ) can be

sintered to full density at 1400 8C.

Careful attention has to be given to the purity of the precursors to avoid

detrimental eﬀects on conductivity. In a polycrystalline ceramic the conductiv-

ities of grain boundaries and bulk contribute to overall conductivity. In the case

of polycrystalline YSZ, because of its unusually high intrinsic (bulk) conductivity

the grain boundaries are far less conductive than the crystal, typically by a factor

of 100. The eﬀect the grain boundaries have on overall conductivity will

depend on grain size and, of course, on impurity content (e.g. silica), since

impurities tend to concentrate there. It is the eﬀort to understand more of the

various contributors to overall conductivity which has led to the application of

impedance spectroscopy (see Section 2.7.5).

b@-alumina The b-aluminas are a family of non-stoichiometric aluminates of

which the most important have the approximate formulae equivalent to

O11Al

(b-alumina), and Na

O5Al

(b@-alumina). They have a layer

structure (Fig. 4.30 (a)) with layers approximately 1 nm thick consisting of blocks

of close-packed O

2

ions in which the Al

3þ

ions occupy octahedral and

tetrahedral interstices in the same arrangement as the Mg

2þ

and the Al

3þ

in the

spinel structure. The spinel layers are separated by mirror planes containing Na

and O

2

ions. The Na

ions can move quite freely within this plane with the sort

of concerted motion indicated in Fig 4.30(b); as a result the conductivity is high

in these planes but negligible in the perpendicular direction.

It is instructive to write the compositions of the b- and b@-aluminas with the

oxygen lattice content the same in both cases, that is Na

and

FUEL CELLS AND BATTERIES 187

, respectively. This suggests that compared to b-alumina the b@

modiﬁcation contains an aluminium vacancy compensated by three extra

sodium ions in the conduction plane. The consequent higher conductivity of

the b@ modiﬁcation makes it favoured for battery electrolytes. The conductivity

of polycrystalline b-alumina at 350 8C (the temperature appropriate to

battery operation) is about 5 S m

1

and for polycrystalline b@-alumina about

50 S m

1

The preparation of ‘pure’ b@-alumina is not easy and dopants (e.g. Mg and Li)

are added to stabilize the modiﬁcation. The dopants have the added advantage of

reducing interstitial oxygen in the conduction plane and therefore facilitating the

movement of the sodium ions.

A route described by J.L. Sudworth et al. [9] for the commercial production of

an almost pure b@-alumina tube starts with the mineral boehmite {AlO(OH)}

which is calcined at 800 8C to produce a microcrystalline mixture of the g-, d-

and y-forms of alumina. The calcined powder is mixed with sodium carbonate

and lithium hydroxide and ﬁred at 1200 8C when complete conversion to b@-

alumina occurs having a composition close to (wt%) 90Al

, 9.2Na

O and

0.8Li

O. The b@-alumina is wet-ground to a particle size of approximately

1–2 mm after which the slurry is spray-dried to produce a free-ﬂowing powder

suited to forming the tubes by isostatic pressing.

Sintering is a critical step with schedules extending over typically 24 h and a

peak temperature of approximately 1600 8C. At this temperature there is an

appreciable sodium vapour pressure over beta aluminas and special precautions

are taken to maintain the required composition. The component may be enclosed

in a sealed spinel (MgAl

) saggar which can be re-used.

188 CERAMIC CONDUCTORS

Fig. 4.30 Structure of b@-alumina: (a) alternating spinel blocks and conduction planes;

(b) migration pathway of Na

ions indicating paths of concerted motion.

b@-Alumina tends to suﬀer from discontinuous grain-growth during sintering

with the result that an undesirable duplex (excessively large grains embedded in a

small grain-sized matrix) develops. Since mechanical strength is reduced by the

presence of large grains the sintering schedule must be adjusted so that the rate of

densiﬁcation is very much greater than the rate of grain-growth. In this way a

uniformly small grain-sized ceramic can be made and the necessary optimal

technical strength (200 MPa) achieved.

Other fabrication routes have been tried including forming by electrophoretic

deposition and densiﬁcation by ‘fast-ﬁring’. In ‘fast-ﬁring’ (or ‘zone-sintering’)

the tubes may be fed sequentially through an inductively-heated furnace with the

peak temperature approximately 1700 8C. The sintering takes only a few minutes

and so Na-loss is negligible, and the tube is rotated as it passes through the

furnace to prevent bending.

The solid oxide fuel cell (SOFC)

Details of the ceramics and approaches to fabrication involved in SOFC

technology are many and complex and well beyond the scope of the present text.

The following summary provides the framework to which the details can be

added. The articles by N.Q. Minh [10], B.C.H. Steele [11,12], B.C.H. Steele and

A. Heinzel [3, pp. 345–52], and S.C. Singhal [13] serve as good entries to the more

specialized literature.

Electroceramics exploited The four major functions electroceramics perform

in the present generation of SOFCs are as the electrolyte, the anode, the cathode

and the interconnect. The favoured ceramics are discussed below.

Electrolyte–cubic stabilized zironia Almost without exception cubic stabilized

zirconia is the chosen ceramic for the electrolyte in SOFCs. This is because of its

adequate conductivity and almost total absence of electronic conductivity, and

because it is stable against the wide range of oxygen partial pressures (1 atm. to

10

20

atm.) encountered in a fuel cell. Also, because of a combination of

availability and cost the favoured compound is yttria-stabilized zirconia,

ZrO

þ8–10 mol.% Y

(YSZ).

The ceramic is required in the form of dense, polycrystalline layers, as thin as

practicable to minimise cell resistance, consistent with the need to ensure

mechanical integrity. Adopted layer thickness is, of course, also determined by

the operating temperature of the cell. Self-supporting plates 150 mm thick can be

used at 900–1000 8C and supported thick ﬁlms (down to 15 mm) to as low as

700 8C.

Fig. 4.31 shows Arrhenius plots for YSZ together with other candidate

electrolyte ceramics illustrating how the combination of temperature and ceramic

FUEL CELLS AND BATTERIES 189

type inﬂuences the electrolyte thickness for a given areal resistance value,

arbitrarily but realistically set at 15 mO m

2

Anode The anode, or fuel electrode, must be stable against reduction, be

electronically conducting, have a high specific surface area and a connected

porosity allowing free passage of the fuel gases to the anode/electrolyte interface;

it must also facilitate the counter flow of the oxidation products away from the

interface. Although the temperature and reducing atmosphere conditions permit

the use of partially sintered metallic nickel as an anode, prolonged use would

lead to further sintering and undesirable microstructural changes; there is also

the need to match thermal expansivities. Coating the partially sintered nickel

with YSZ gives a better thermal expansion match between anode and electrolyte,

190 CERAMIC CONDUCTORS

Fig. 4.31 Ionic conductivities for candidate electrolyte ceramics. The arbitrary assumption

that for a planar cell format a resistance of 515 mO m

2

is required places an upper limit on

the permitted thickness of the electrolyte; lower values of conductivity demand thinner

membranes whilst higher values permit correspondingly thicker membranes. Electrolyte

thicknesses greater than approximately 150 mm are considered mechanically self-supporting.

After B.C.H. Steele [11].

reducing thermal fatigue during operation and improving adhesion between

anode and electrolyte. A further advantage is that the microstructure offers an

increase in reactive area at the three-phase boundary where oxidation of the fuel

occurs.

Cathode In contrast to the anode the cathode operates in an oxidizing

environment but, like the anode, it must have high electronic conductivity and a

pore structure enabling the gaseous oxidant to reach the cathode/electrolyte

interface.

The favoured material is modiﬁed lanthanum manganite (e.g. La

0.8

0.2

MnO

3þx

)

which has the perovskite structure. It is a p-type semiconductor the electron

transport occurring by electron-hopping (see Section 2.6.2) between the þ3 and

þ4 states of the Mn ion. The strontium-doping enhances the conductivity.

Interconnect The property requirements for the interconnect ceramic are the

most demanding. The conductivity should be as high as possible and, ideally,

100% electronic. This is necessary to reduce internal resistance and to avoid

severe problems which would arise were ionic space-charge polarization (see

Section 2.7.1) to develop.

Because the interconnect is exposed to high temperatures and oxidizing

environments it must be stable against both. It must also serve as an eﬀective

barrier to the oxidant and fuel gases so that direct reaction is avoided, and it is

required to be chemically inert with respect to the cathode, anode and the inter-

cell connector material, for example nickel (see Fig. 4.32(a)).

These requirements are satisﬁed by a doped lanthanum chromite (see Section

4.1.3) e.g. La (Ca, Mg, Sr, etc.) CrO

. However for SOFCs operating in the

temperature range 500–750 8C a stainless steel interconnect plate can be used.

Geometries and construction The two basic geometries for the SOFC are

planar and tubular.

Planar geometry Although there are variations in detail, the design of the

planar or ‘flat plate’ cell, and how cells are interconnected are essentially as

illustrated in Fig. 4.27. The construction may involve a self-supporting, coated

electrolyte or, if the operating temperature permits, the anode–electrolyte–

cathode structure may be supported on a stainless steel bipolar plate.

Brief summaries of the various processing technologies used in fabricating cell

components are given in Section 3.6.9.They include tape-casting, calendering,

screen-printing, electrophoretic deposition, electrochemical vapour deposition,

chemical vapour deposition, spray-pyrolysis and sol–gel.

As an example the electrolyte may be tape-cast or calendered. Flat and dense

plates of YSZ having thickness down to 150 mm are readily fabricated. The anode

and cathode materials can then be applied (by a slurry) to the faces of the

electrolyte and ﬁred on. Alternatively the anode, electrolyte and cathode layers

FUEL CELLS AND BATTERIES 191