a house of cards structure within the gel. Also impact- ing on the gel structure is the treatment of clay with orthophosphoric acid which can create edge-to-edge.
Mikrochim. Acta 128, 1-18 (1998)
Mikrochimica Acta 9 Springer-Verlag1998 Printed in Austria
Fundamental Review Clays as Architectural Units at Modified-Electrodes Susan M. Macha and Alanah Fitch* Department of Chemistry, Loyola University Chicago, 6525 North Sheridan Road, Chicago, IL 60626, USA
Abstract. Since Ghosh and Bard [80] first established the field of clay-modified electrodes some 15 years ago, great strides have been made in understanding the nature of the clay structural units and their impact on transport of a variety of electroactive probes (anions, neutrals, small cations, large cations, and compounds with distributed charge). Great strides have also been made in understanding the nature of the layered material in creating access of interlayer sites (size, charge, iron content, pillaring, and organic tailoring). In the last five years several successful applications of clay-modified electrodes have been achieved. Given the explosive growth in tailoring and construction of novel clay structures it seems reasonable to predict further significant advances in applications involving clay-modified electrodes. Key words: modifiedelectrodes,layeredhydrous silicates, tailored clays, catalysis, surfactants.
Contents I.
Whatare Clay-ModifiedElectrodes and Why are They Interesting. . . . . . . . . . . . . . . . . . . . . . 1 II. Claysand Layered Materials: The Structural Unit . . . . . 3 III. The Foundation: Physical Attachment to the Electrode.. 5 IV. The Corridors: Charge Transfer. . . . . . . . . . . . . . . . . . 7 v. Roomsand Room Design: Effect on Corridor Access .. 10 VI. Form Follows Function . . . . . . . . . . . . . . . . . . . . . . . 13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Clay-modified electrodes have reached a level of maturity and understanding in the last 15 years. This article takes the perspective that clays are used as a * To whom correspondenceshould be addressed
series of building blocks with the final structure containing functionally designed "rooms", "corridors", "elevators", a "foundation", and a " r o o f " . Access to the structure will occur both through the roof and the foundation. The first question to address is why one might use clays to build these architectures.
I. What are Clay-Modified Electrodes and Why are They Interesting? Clay-modified electrodes are a subclass of modified electrodes. Modified electrodes have been studied extensively during the last 20 years because of the possibility of intelligently designing the surface of an electrode to perform desired tasks that are coupled to an electrochemical reaction. Such tasks may consist of selective analysis (enzyme modified electrodes for glucose measurements via the glucose oxidase reactions) [1,2]; enhanced flux design (Nucleopore membranes); charge storage devices [3, 4]; catalysis [5, 6]; support matrices for electrocatalysis [7-12]. The vast majority of modified electrode surfaces consist of surface assembled monolayers (SAM) and polymeric coatings [13, 14]. A subclass of modified electrodes consists of modification with inorganic matrices. These inorganic matrices include Prussian blue films, ruthenium oxide coatings, aluminum oxide with etched pores, gamma-alumina [15], silica gels [16], zeolites [17-19], and clays. Clays in this context are defined as layered inorganic materials which include synthetic materials as well as native clays. This review focuses entirely on the layered inorganic materials as they are used in modified electrode studies.
2
S.M. Macha and A. Fitch
t
t 11
GroundWater Supply
ContaminatedSediment
Drilling~ f - ~
Drilling ~( fluid, s o i l ~ and clay OUt
slakin~
There are two major reasons why clay-modified electrodes (CMEs) are interesting. The first is that transport processes illuminated by clay-modified electrodes have applications to transport of pollutants in natural environments (e.g. under landfills, through capping layers in river or estuarine sediments, through nuclear waste repositories [20-22] and electroremediation of soils [23-26].) In addition, these processes shed light on other applications involving clays. An important example is the mechanism by which fluids from an oil well drilled through a shale (rock formed from compressed clay) interact with the clay. The solute diffuses into the shale and clay is swollen resulting in failure of the well [27] (Fig. 1). Finally, the mechanisms by which materials are transported within clay sheets is a fascinating subject which is related to molecular recognition, membrane sieving, and the effect of localized vs delocalized fields on enhanced transport. The second major reason for studying CMEs is, related to the possible use of the modified electrode as a device. Presumably modification of electrodes or optrodes with layered clay materials is of interest because one can control the architecture of the surface, creating a 'tinkertoy' assemblage of larger
ration
Fig. 1. Transportof solutes in clay films is important in use of clays as barriers (landfills, sediment capping) and in stability of oil well drilling through shales
ROOF
CORRIDOR
~
]
,OMS R
N-
m FOUNDATION
Fig. 2. Transport will play a role in the architecture of macrostructures based on clay units. The final structure may contain functionally active "rooms", "corridors", and "elevators" which may or may not provide direct access from the "foundation" (electrode) to the "roof" (solution)
sub-blocks, each one of which could be designed with a specific functionality or size characteristic (Fig. 2). Clays attract interest for this application because they come in a variety of starting shapes and sizes with different surface chemistries and with " r o o m " sizes of molecular dimensions (9-30,~).
Clays as Architectural Units
3
II. Clays and Layered Materials: The Structural Unit In order to properly use a clay-modified electrode it is important to understand the variety of structures which can be derived from clays. The types of structures that occur depend to a very large extent on a) the type of clay; b) the charge of the clay; c) the size of the clay; and d) the native and/or the ion exchanged major cation associated with the clay. Most clays used for clay-modified electrode (CME) studies are composed of sheets of tetrahedral silicates (Si-O) and/or octahedral aluminates (A1-O, A1-OH) (Fig. 3). Lateral sharing of oxygen atoms in tetrahedra or octahedra results in a sheet. Sharing of oxygen atoms between a tetrahedral and an octahedral sheet results in a layer. Minerals may consist of one silicate and one aluminate sheet, termed a 1:1 layer mineral. In a 2:1 layer mineral, two tetrahedral sheets sandwich an octahedral sheet. Most native clay platelets are negatively charged due to isomorphous
A.
C.
E.
F. !~ i \ i
i
.IP
c
!
(
Fig. 3. Clay layers are composed of A tetrahedral and B octahedral sheets. Sheets are linked into C l:1 and D 2:1 (tetrahedral/octahedral sheets) clay layers. E Octahedral sheets can be formed with all possible sites filled (trioctahedral) or with two sites filled (dioctahedral). F 0:1 structures are common among anion exchangers like brucite. Shown are two brucite layers oriented face to face
substitution within the clay structure. Clay particles do not have a net charge, therefore the surface charge of a platelet must be compensated by the accumulation of counter ions. An electrical double layer describes the association of counter ions at the surface of a charged platelet. Those counter ions native to each clay may be exchanged for one or more cations [28]. Clays are characterized by the total amount of exchangeable cations, the cation exchange capacity (CEC, meq/100 g). Classification of native clays depends upon the number and arrangement of tetrahedral and octahedral layers, the number of octahedral sites occupied, the extent of substitution (charge), and localization of charge in the octahedral or tetrahedral layer. Table 1 gives the classification of some commonly used clays in clay-modified electrodes according to these features. Several broad categories of clays can be listed. Those that have one silica/oxygen tetrahedraldayer and one octahedral layer bonded together are kaolinites. These clays contain very few substitution sites and therefore are non-swelling clays. A sub-type of kaolinite is halloysite, which differs by its internal water content. 2:1 clays have two tetrahedral layers and one octahedral layer bonded together. Among the 2:1 clays, when two of the three possible octahedral sites are occupied the clay is dioctahedral (Fig, 3). Montmorillonite is a 2:1 dioctahedral layer clay with substitution of cations in either the tetrahedral or octahedral sheets. Particle sizes range from 1 to 2 pm in diameter and approximately 9A in height. Three montmorillonites have been used extensively in claymodified electrodes. These are standard Wyoming, Arizona Cheto, and Texas montmorillonites. These clays differ in the extent of substitution within the tetrahedral layer and the native cation exchanger (Na + vs Ca2+). Progressive substitution within the tetrahedral layer of Fe 3+ for Si 4+ leads to nontronite. The high degree of iron in the surface tetrahedral layers contributes to extensive negative charge of the clay. The end member of the dioctahedral 2:1 clay series is the illite class of clays. These clays do not swell because the total charge on the clay (which can run as high as 150-200 meq/100 g) is compensated by dehydrated K +. The strong attraction of the clay layers across the cation bridge makes access to charged sites poor, resuiting in exchangeable charge only 10-40 meq/100 g. The starting member of the trioctahedral 2:1 clays (three octahedral sites occupied by Mg:§ is hectorite. Laponite is a synthetic hectorite of comparatively
S. M. Macha and A. Fitch
4
Table 1. Clays and clay types Classification
Type
1:1 2:1 dioctahedral
2.1 trioctahedral
Anion exchangers
Name
Formula
Charge meq/100 g
Kaolinite Halloysite
A18SisO2o(OH)16 A18Si802o(OH)16 wide (AI+Fe)/Si range
6-15 10-30
Prototype montmorillonite MontWyoming morillonite Cheto (SAz) Texas (STX) Nontronite Washington
(Al4_xMgx)(Sis)O2o(OH)4 (A12.94Fe0.35 Mgo.58)Si8 (A12.8Feo.34Mgo.84)(Si7.6Alo.24)
76.4 120 84.4 76
Illite
(All .06Fe2.73Mgo.26)(Si7.3Alo.7) (M+.nH20)Fe3+(Si4-xAlx)Ozo(OH)4 A14Si6A12020(OH)4
Prototype hectorite Hectorite SHCa Laponite Saponite Vermicullite
(Mg6-xLix) (Si8)O20(OH)4 (Mg5.34Lio.66)SisO2o(OH)4 a synthetic hectorite Mg6(Si7.34Alo.66)020(OH)4 (Mg6-xFex) (Si6A12)O2o(OH)4
Chlorite Brucite Hydrotalcite Layered double hydroxide
[(R2+, R3+)6(Si,A1)sO20(OH)4][(R2+R3+)6(OH)12] [Mg6A12(OH) 16]2+ [Mg6Alz(OH)I6][CO3'4H20] [Zn4Cr2(OH)12[C12.2H20] [Zn2A1(OH) 5] [(TA).xH2O] etc.
small particle size which can lead to imperfect stacking and disruptions of the corridors [29, 30]. Substitution within the octahedral layer of hectorite leads to saponite while substitution within the tetrahedral layer leads to vermiculite. Other layered clays are anion exchangers. These are relatively less common in the native form. Native anion exchanging clays are chlorites, 2:1 layer clays in which the octahedral sheet is a magnesium hydroxide or brucite sheet. Some of the Mg 2+ is substituted by A13+ resulting in a net positive charge in the sheet. The brucite sheet compensates almost entirely for the negatively charged tetrahedral sheets, giving the layer a very low negative charge. When brucite is compensated by carbonate the clay is known as a hydrotalcite [31]. The exciting proposition of using hydrotalcite-like clays is the ability to substitute any of a variety of metal ions into the brucite-like layer. This may have significant implications for electrochemistry [32-34]. Layered clays may be synthesized with structures similar to naturally occuring hydrotalcite. These synthetic clays are called layered double hydroxides
Particle Size, gm
10
1.29 0.8
10-40 available 2000 total
43 73
Area m2/g
12-158 97 83
43-195
19.81 50 nm
63 360
0.5-43 (120-200 total)
(LDHs). LDHs are attractive because they can be easily prepared by precipitation of Mg and A1 as nitrate and chloride salts in NaOH to form a single layer of the octahedral sheet (Mg(OH)2 with partial AI(OH)2) [35-37]. Anion guests may be intercalated easily [38]. However, ion exchange, intercalation, and pillaring proceed with difficulty because the anion is strongly bound by the high charge density of the layer [32, 34, 39]. As noted from the structural dimensions given in Table 1 the individual or microscopic structure of clay has dimensions both in the angstrom and micrometer regimes. In addition to these structural features, clays tend to stack into clusters or aggregates. An aggregate of platelets forms a particle. The particle size of a clay depends on the dimensions of the individual platelets, the type of exchange ions, the solvent, and the oxidation state of the clay [40]. A particle may consist of 5-30 platelets stacked in face-to-face association (stacked), edge-to-face (house of cards), "turbostratic" (rotated about the transverse axis), or overlapping (house of bricks) [41, 42]. A bridge may be formed between particles by slippage of one platelet
Clays as Architectural Units across platelets from smaller particles, thus extending the particle size [41]. In summary, clay-modified electrodes offer the hope of building blocks to larger macroscopic structures which are well defined in their nature. Some of the difficulties in achieving these larger structures lie in establishing and/or controlling the inter-particle structure. III. The Foundation: Physical Attachment to the Electrode In the last section it was implied that clay-modified electrodes are interesting because clays can be envisioned as the subunits necessary for the construction of larger structures. The final structure may have elements necessary for facile transport and/or catalysis. The first step in such construction is the foundation, or the attachment of the clay to the electrode surface. Because most of the clays (anion or cation exchanging) used are polar due to isomorphous substitution within the crystal lattice, and because most of the surfaces to which the clays are attached are also polar (metals and metal oxide surfaces), clays will self-adhere. Simple drying of the film in air, oven, or by spin coating should result in a film that adheres to the polar substrate, most likely through cation or anion bridges. LDH clays also attach by physical adsorption to the electrode surface in a manner similar to that described for the cation exchanging clays [33, 43]. This physical adsorption method of fixation has the advantage of simplicity and ease. One disadvantage is a tendency to oversimplify this procedure and assume that air drying, oven drying, and spin coating all result in similar "foundations". In order to understand the drying process which results in the clay film, it is necessary to understand the type of structure to be found in the gel solution and how that structure is driven by the type of clay, its particle size, and the type of counter ion associated with the clay. In a gel state clays are thought to be linked together by two processes controlled by the face-to-face bridging of the clay and the edge-to-face bridging of the clay. The face-to-face interaction is repulsive and is modeled by classical Derjaguin Landau Verwey Overbeek (DLVO) theory. When the diffuse double layers are compressed sufficiently, the negative charge of the clay is shielded and the individual clay sheets can rest at an equilibrium distance
5 determined by the overlap of the diffuse double layers. The counter-ion will affect this interlayer distance in a predictable manner [44, 45]. Edge-to-face structure depends predominately upon the pH of the gel. At low pH the edges of the platelets are protonated and edge/face interactions can result, creating a house of cards structure within the gel. Also impacting on the gel structure is the treatment of clay with orthophosphoric acid which can create edge-to-edge linkages expanding the lateral dimension of the layer [46]. Because of the salt effects on structure, well dispersed solutions are most easily obtained when the clay is exchanged with a mono-cation at dilute concentration. Flocculated solutions occur when more highly valent counter ions are used. Completely delaminated clays can be formed by exchange with quaternary ammonium surfactants and/or addition of polymers to the clay suspension. Air drying a pure gel results in loss of all but the surface attached water. If drying is slow, then time is available for the development of face-to-face orientation. If drying occurs at higher temperatures (200-240 ~ adsorbed water may be lost rapidly and the house of cards structure may be obtained. Drying at higher temperatures (550-650 ~ for montmorillonite) removes surface hydroxyls causing irreversible collapse, or calcination, with linkages formed between layers [47]. Spin coating at a moderate rate accentuates the trend toward face-to-face orientation [48-50]. Once a dry film is formed it may re-swell. This is particularly well known for cationic clays, but as yet is less well understood with the anionic clays. The extent of swelling again will depend upon the cations associated with the permeating solution. However, the extent of swelling is not solely governed by DLVO theory and diffuse double layer effects. Swelling will be driven to a large extent by the hydration energy of the salt. For example, bathing the dry film in KC1 results in an un-expanded clay film, since K + is easily dehydrated during the intercalation process [51-53]. Similar results are observed for Cs +. Sodium and lithium represent the alternative extreme. These cations have a higher energy of hydration and enter the interlayer in their hydrated form. This "forcing" apart of the dry face-to-face sheets requires substantial energy and is thought by some researchers to proceed non-uniformly. That is, intercalation reactions proceed layer by layer, not randomized over all layers. Once the first set of water is brought into the interlayer
6
subsequent expansion is more facile. X-ray diffraction studies suggest that expansion occurs in two more steps (3A in 1 M NaC1, 9A in 0.25 M NaC1, and >30A in I . 0 M ) of destabilizing cation is introduced, the swollen clay remains stable for extended time periods and it will not collapse in response to a more concentrated electrolyte solution. However, a collapsed clay will swell if the electrolyte solution is diluted. Third, for simple clay films the structure can change rapidly during the wetting process and less rapidly during a collapse. The kinetics of these events can be tracked electrochemically since the signal of a physically diffusing species tracks the pore space available. The kinetics of both the swell and shrink cycle of the clay-modified electrode are important as small scale systems which parallel environmental processes (solute/solvent fronts in clay barriers and slaking of stable clay structures from shales in oil drilling). They are also important for design applications where a large signal change is required in response to some triggering event, or where controlled diffusion release of a drug or reagent is envisioned. The kinetics of film swelling as tracked by probes traveling with the solvent front is rapid. For dry films 10 gm thick swelling will occur within 5 to 10 minutes or less. For electrolyte driven collapse of the film the time scale may be 6 to 10 hours [54, 55] (Fig. 4). Another factor governing the rate of film swelling and its final state is clay charge, cation exchange capacity (CEC). The initial phase of swelling is independent of cation type, controlled by hydration of the clay surface itself [56]. Swelling depends linearly on the CEC of a clay [56, 57]. Highly charged clays
S.M. Macha and A. Fitch 1
g
08!='!',,, IIIil,l~
|,
O6
In l
E 04.
10 M KC[ Je 9 aim m le, ~l== n 1.0 M CaCL~ "o l,
0.2. 0
';,,
[ 2
I 4
I 6
""..:i'".. .... I *'*''.I 8
10
'==='I 12
-02 Time, hours
Fig. 4. Shrinking kinetics. Well swollen films maintain their structure for prolonged periods of time even in the presence of cations which could prevent the initial swelling of the film. Plot of the current ratio (Ip.c~jIp.b~re) for the reduction of I mM Fe(CN)63- in a cyclic voltammetric experiment as a function of time at a SWy-1 clay film initially swollen in 0.01 M NaCI and subsequently exposed to either 1 M KC1 or CaC12. (S. M. Macha, A. Fitch, work in progress)
have a greater capacity for hydration and thus can support more layers of water. This gives rise to a greater extent of swelling assuming all other factors are held constant (that is, the counter ion is similar and the location of the charge (tetrahedral to octahedral) is similar). The effect of pH on the clay structure must also be emphasized here. A low pH reflects not only a high fraction of cation exchange sites occupied by H + and changes in the edge charge structure, but also can contribute to the dissolution of the clay mineral itself; such dissolution may enhance the catalytic activity of the surface [58]. The final structure of the acid-treated clay may differ significantly from the original clay structure which, in turn, will change the structure of the clay film. The foundation: chemical attachment. Some workers have directly attached clays chemically to the underlying electrode surface [8] (Fig. 5). A silane linkage was used to couple pillared clay to the electrode surface. The goal in these attachment strategies was the localization of the clay at some distance from the electrode surface in order to form a capacitive gap. Chemical modification of the clays depends upon the total surface hydroxyl groups on the clay. In general, many clays, while containing hydroxyl groups, have those groups buried in the inaccessible interior of the crystal lattice. Hydroxyl-containing clays which have promise for such siloxane chemistry are kaolinites, magadiites, and hydrotalcite-like (LDH) clays. Siloxane linkages can also be achieved at the edge surfaces of other (montmorillonitic) clays [59].
7
Clays as Architectural Units
~^
electrode
1
I
+l I
HU ~
bulk solution
.4 +I I+ I ,OH I OR ~...,'-Q~N(CH2)nN--R'--Sif-,
i+ I
OR/t
+1
I+
,
J
I
+1
e'ectr~
~
@
I+ I
particle
I+ / ~ ' - ( ~ -
t, film thickness
m
,"
zeolite or pillared clay particles
m q
m
I./OH
" ~-....- i~1(CH2)ni~l--R --S K,..OH
m
m
m
m
n
@
~'~___. @@(C~ )
Fig. 5. Attachment scheme of a pillared clay to an electrode surface to create a bilayer electrode. The cationic linker stabilizes an anionic, electroactive, porphyrin which serves as a charge shuttle to cationic groups localized in the negatively charged clay (D. Rong, T. E. Mallouk, Inorg. Chem. 1993, 32, 1454)
IV. The Corridors: Charge Transfer The ultimate goal of device-driven clay-modified electrodes is to create a series of rooms where chemistry can occur (charge storage, charge conversions, molecular recognition) and a series of corridors via which charge can be shuttled from the foundation (the clay/electrode interface) to the upper levels or from the roof (solution/clay interface) to the foundation (Fig. 2). Both the corridors and rooms are important, just as a room without access is of no interest and a corridor to nowhere is similarly of little use. Design of these structures must incorporate both features. In this section the construction of corridors is discussed. In the next section the design of specific rooms is examined. In a clay-modified electrode the conceptual corridors are often very analogous to physical corridors. Charge can be (predominately) shuttled by a mobile charge carrier. Attempts have been made to facilitate charge transfer using electroactive sites within the crystal lattice, or via in situ polymerization of electroactive wires in the interlayer. These latter "intercom" connections will be discussed at the end of this section. Electrochemical activity in any of these mechanisms is a measure of the efficiency of charge transport either via physical diffusion in corridors or via charge hopping. Typical experiments for the observation of
Fig. 6. Schematic of the diffusion layer 6 created by the electrochemical perturbation and the film thickness, g. 5 depends upon the rate of the electrochemical perturbation while g depends upon electrolyte driven swelling of the clay film
electrochemical activity involve cyclic voltammetry (CV), potential step experiments, and rotating disk electrode (RDE) experiments. In cyclic voltammetry the underlying electrode potential is rapidly perturbed in a linear fashion and the oxidation and/or reduction of the probe molecule residing within the diffusion layer of the electrode is monitored (Fig. 6). Current peak heights (Fig. 7), the potential at which the peak is observed, and the time for the signal to be developed [60, 61] (Fig. 8) contain useful information. The latter corresponds to the transit time of the charge carrying species (electroactive probe or charge compensating ion) to the diffusion layer at the clay/ electrode interface. Transient (film filling) charge transport can be assessed by the time required to develop the signal (a type of diffusion front propagation from the bulk solution/clay interface). Equilibrium charge transport (film filled) can be assessed by the steady state peak currents compared to those obtained at the bare electrode. This ratio contains information about the relative diffusion coefficients and relative concentration within the film as compared to the bulk solution (R = [cme/Ibare ~ (Dapp/Dbare)i/2(Ccme/Cbare)). The apparent diffusion coefficient (Dapp) contains terms for the accessible area, the tortuosity, and the electric field effects [56]. Alternatively, an absolute fit to the cyclic voltammetric peak as a function of the potential scan rate can determine the diffusion coefficient within the clay, assuming that an accurate measurement of the film concentration is known [62, 63]. Potential step experiments also are useful in determining the diffusion coefficient within the film, again assuming an accurate assessment of the film concentration can be made [63].
8
S.M. Macha and A. Fitch
A
0.05 M NaCI
I-
1.75 M NaCl , ' k
NN~
"73§
%" IRU,,,..x
l,OOnA
CME I
I
I
0.4
B
I
0
I
I
I
I
0.4
-0,4 EN
I
I
0
-0.4
SWa-1
8~-1
r
0.8
I
L
-
-
~
Red
NJ
o_~_o.4 II o
4
,
C
10 [NaCl]
8
,
,
,
,
4
8
10
"~ ,,
,
,
I
500 nA
,
120
.E 1 O0 e,
7~
Ox
J&
~
~
x = in NBCI solutions
C
80
4o xQ
*~ za ._-4" C -atz
Fig. 7. A Typical CV obtained for lmM Fe(CN) 3 at SWy-1 montmorillonite film (35 gg clay) and bare electrodes as a function of the concentration of NaC1 in the bathing solution (A. Fitch, C. Fausto, J. Electroanal. Chem. 1988, 257, 299). B A plot of the ratio of currents at clay (SWy-I or SWa-1) to bare electrode vs NaC1 concentration has a characteristic discontinuity related to expansion of the interlayer due to swelling. Ox and Red refer to native oxidized and chemically reduced clays (A. Fitch, J. Du, H. Gan, J. W. Stucki, Clays Clay Miner 1995, 43, 607). C X-ray diffraction data also shows the step function swelling of the clay as a function of electrolyte (K. Norrish, Trans. Farad. Soc. 1954, 18, 120)
Charge transport via physical diffusion of Ru(NH)~+ has been observed, as is also the case for several anions (Br-, Mo(CN) 4-, Fe(bphen) 4- [64], Fe(CN) 3-, IrCl~-[48], and Fe(CN)4phen 2-, Fe(CN)4bpy 2- [65], nitrobenzene [66], and anoinic water soluble porphyrins [41]). In many of these studies it may not be clear if physical diffusion is occuring in corridors, in "elevators", or in some combination of both. The transport between aggregates (elevators) in an anionic clay can be easily recognized in the
Fig. 8. Multisweep cyclic voltammograms (MSCV) for SWy-1 (CME) obtained at 50mV/s. A 0.01raM Ru(NH3)~ +, 0.01M Na2SO4, potential swept between 0.0 and 0.7 V vs SCE. Scale is 500nA. B 0.5 mM Co(sep)C13 in 0.1M NaC1 potential swept between 0.1 and -0.85 V vs SCE. Scale is 500 nA. C 0.5 mM Co(en)3C13, potential swept between 0.3 and -0.85 V vs SCE. Scale is 500 nA. The single sweep, time invariant, signal for the bare electrode is indicated. Potentials at the CME are shifted negative, and the magnitude of the signal is larger at the CME consistent with an electrostatic model of uptake (A. Fitch, J. Song, J. A. Stein, Clays Clay Miner. 1996, 44, 370)
electrochemical experiment by the behavior of a small, water soluble anionic probe molecule. The anionic probe molecule tends to be excluded from the interlayer domain, with exclusion increasing as the interlayer distance decreases. Considering the corridors alone, transport of the anion is controlled entirely by lateral motion through the interlayer domain, which is predominately unfavorable. Admission of the anion through corridors depends upon the width of the corridor, or interlayer swelling. Complete exclusion of anions is expected when the corridor is narrow and the negative electric field is high. To test if diffusion is corridor or elevator (pinhole) driven one monitors the current through a collapsed and then swollen film. A low current is expected for anions through a collapsed corridor, but a high, quickly developing current occurs
Clays as Architectural Units through a film controlled by pinholes. Subsequent expansion of the interlayer or corridors gives higher current through corridors and the same or lower current through pinholes. Two results, then, give a notion of the pinhole vs corridor driven transport. Immediacy of the signal for the anion in a collapsed film marks pinhole controlled transport. The time derivation of a more favorable signal upon swelling denotes corridor controlled transport. The case when the elevators directly access from foundation to roof can be modeled by a microhole array [67]. This model can be tested by showing the transition of the diffusion layer from outside the film (bulk solution, semi-infinite, linear diffusion) to the film interface (spherical diffusion) to inside the film (linear diffusion) (Fig. 6). The diffusion layer is tracked by monitoring the scan rate dependence of the maximum reduction current, It should be noted that there is an intermediate case between a film bisected by elevators or pinholes and a completely pinhole free film. This is the case of a structure in which pinholes are connected by transit down a corridor (Fig. 2). Access from solution to electrode occurs through elevators but is interrupted by a side trip down a corridor. The overall flux of material to the surface of the electrode will reflect this most constricted space. This model has been successfully applied to defect-dominated films and those in which defects were minimized [68]. Corridor width (interlayer spading) has been correlated with the transport of pore volume samplers [48, 54, 56] (Fig. 7). In summary, the clay film laid down has a unique pore architecture. For highly swelling clays, the size of the corridor can be controlled by the electrolyte solution. Expansion of the interlayer region can be effected by bathing in dilute sodium solution, however these propped-open corridors are in a metastable state and will collapse given enough time (10-12 hours). The transport of some weakly-adsorbing electroactive probes may be controlled by these corridor pores. Corridors designed for different populations. As implied, access to corridors will depend to a very large extent on the substrate which is chosen as the charge shuttle. Neutral and cationic charge shuttles will be expected to have greater access to the interlayer domain. These charge shuttles will have different effectiveness depending upon their ultimate attachment to the clay surface. What drives the attachment of these probe molecules to the corridor surface? Hydrophobic cations minimize their hydro-
9 phobicity by moving to the clay surface, particularly when large areas of the clay surface are hydrophobic (areas far away from the charge substitution sites). Hydrophobic molecules are unlikely to succeed as charge mediators required to deliver information to the active sites within the clay. Small cations are the most likely successful charge shuttle species. These compounds are highly water soluble and are given access to the interlayer due to their electrostatic interaction with the negatively charged corridors. As long as the species reside in the diffuse double layer where motion is controlled by random thermal processes they should remain viable charge shuttle candidates. The weak electrostatic nature of interaction of these compounds with the clay can be demonstrated by enhanced signals due to enhanced concentrations. In addition, the reduction potential of these Compounds in the clay often shifts negative. Potential shifts of the electroactive compound reflect the relative interaction of the oxidized and reduced forms of the complex with the clay. If the oxidized form is more strongly localized, the complex will be harder to reduce and the reduction potential shifts negative. If the reduced form is more strongly localized, the complex will be easier to reduce and the reduction potential will shift positive. For the strongly adsorbed (hydrophobic) compounds positive potential shifts are common implying that the charge minimization between near neighbors may become important. For many of the more weakly absorbed cations (Ru(NH3)~+[63, 69, 70], Co(en)~+[61]) potentials shift negative. Based on this, the oxidized form is favored, suggesting an electrostatic mechanism for the interaction with the clay (Fig. 8). Molecular modeling of the clay surface has confirmed the relative extent of electrostatic interaction with the clay surface [71]. The shift in potential is dependent not only on the relative interaction of the compound with the clay, but also upon the extent of loading, the bathing electrolyte solution, and temperature [61, 64, 69, 72]. Using these tools (CV, RDE, potential step chronoamperometry, potential shifts) the transport of Ru(NH3) 3+ has been extensively studied by several groups. Nearly all Ru(NH3)3+ in clay films appears to travel in the interlayer. Depending upon the extent of loading of the film, single file diffusion has been implicated [61, 63, 69, 70J. Like Ru(NH3)~ + other compounds (Co(en)~ + and Co(sep) 3+) move easily
10 within the film and have potential shifts suggestive of interlayer transport. These compounds contain nitrogen at their peripheries. This suggests limited attachment of the cation to the negatively charged clay surface. This hypothesis has been pursued with even larger complexes containing peripheral nitrogens such as the cobalt trisbipyrazine, trisbipyramidines, or edge charge such as chromium trisbipyridine [73]. Large size compounds can also serve as charge shuttles provided that there is a kinetic limitation to their adsorption to the clay surface. A kinetic limitation may be introduced by localized negative charge at the periphery of the compound. The hypothesis that the apparent diffusion coefficient is controlled by kinetics of attachment to the clay surface has been invoked for the unique behavior of Cr trisbipyridine. This compound has high electrochemical activity compared to Ru and Os trisbipyridine [57, 72, 74, 75]. Cr(bpy)~+ is unique in rapidly developing an enhanced current peak, shifted negative in potential. The peak decays to form a smaller positively shifted peak indicative of a one-to-one conversion of the species. This process was modeled as a rapid influx due to the electrostatic field of the clay followed by adsorption onto the clay. The difference between this Cr(bpy)~+ and other tris bipyridne complexes was attributed both to the trivalent state of the compound (more rapid flux into the clay) and the presence of larger negative charge at the periphery of the compound [76] which would decrease the rate of attachment to the clay. V. Rooms and Room Design: Effect on Corridor Access
Despite the control of a corridor in access through the film, the body of evidence suggests that for large compounds most corridors are blocked and that access to the majority of interlayer domains is restricted. Studies by several groups [64, 68, 70, 77, 78] reflect the difficulty of accessing these sites. Cations strongly bound to the exchange sites of a clay are electroinactive. Ion paired cations present in excess of the CEC are, however, electroactive. These ion pairs are absorbed but not strongly enough to prohibit mobility. Some ion pairs travel to the electrode through interparticle pores [78]. Ion pairs may also be responsible for charge shuttling from the electrode to immobile cations located away from the electrode surface [79]. Electron-hopping would also explain why some
s.M. Macha and A. Fitch cations are electroactive and others are more strongly electroinactive. Transport in anionic LDH apparently mirrors the transport in native cationic clays. Exchanged molecules may be strongly absorbed to the clay surface, or may be mobile [32]. Only a small fraction of the intercalated anions display redox activity. The electroactivity of the bound probe ion is strongly affected by anions in the supporting electrolyte [33, 35]. There are three ways in which this issue of electroinactivity has been addressed. The first is metaphorically related to modern architecture. The concept of the room is destroyed and the walls become irregular through delamination of the clay structure. The addition of polymers to the clay gel to form a composite can give a final electrode structure that is less well ordered. The disruption of the "rooms" and corridors may enhance transport by eliminating single file diffusion of single cations like Ru(NH3)~ + [64-] or by eliminating strong adsorption sites. Complete delamination of the clay structure can occur via polymerization reactions within the interlayer [80]. Delamination has also been approached by the use of smaller clays (laponite) where the face-to-face structure is energetically less favorable due to less face surface for interaction. Stacking in these clays is less favored [81, 82]. Workers with these clays have suggested that site accessibility has been achieved more easily in the less ordered film structure. What is lost is the unique size features of the layered clay, but what is gained is access to the highly charged surfaces of these clays. An alternative method of gaining access to the corridors is to tailor the inner surface of the clays in a more controllable fashion such that complete delamination of the clay is avoided, and yet the corridor dimension is expanded. Pillared clays fall into this category (Fig. 9). The purpose of a pillar is to control the interlayer dimension via a known pillar height. This can be applied to control of molecular sieve properties of a clay [8]. Rong et al. used aluminate (Al1304(OH)~ +) pillars of well known size. Addition of pillar molecules destabilizes the suspension of clay, so an additional monolayer of polymerized silane was added to adhere the clay layer to the electrode. Rudzinski and Bard [83] were able to produce electrode-bound stable pillared clays without the use of an adherent polymer. Some changes in the interlayer dimension, even when controlled by a pillaring agent, were found similar to swelling by dilute cation
11
Clays as Architectural Units
~3 nm
o ~ \ \ Sdlcate ~ \ \ Lay r
1. Amine-solvated Q+-clay
}1
~
2. TEOS-intercalated intermediate
it
3. Templated heterostructure
~ calcine,600C
4. Porous clay heterostructure
C.
9
9
t ~
~,.,x.~.~
~ .
~z.JL,NskX.,X/N.N.NI 9
a~.4 o
oo
solutions. Pillaring also can give to clay the ability to support electroactivity in nonaqueous solvent [83]. Other pillaring strategies introduce balls and rods into the interlayer region. An example of a "ball" is the pillaring of clay with SiO2/TiO2 sol gels [84]. The pillaring reagent is prepared [85] by mixing TIC14 with twice the volume of ethanol, stirring, then mixing with water and glycerol and, finally, mixing with clay. The formation of semiconductor particles can also be considered a method of pillaring of clays. Semiconductor particles of CdS and PbS have been formed within the clay interlayer [86]. Depending upon the clay type, the structure of the clay determines the size of the semiconductor particle or, conversely, the semiconductor particle determines (disrupts) the structure of the clay [87]. Organo tailoring of the clay can assist in the formation of inorganic tubes and rods within the interlayer region. The clay is first templated with an organic surfactant which is used to guide surfactant modified sol gel precursors. Heating results in the formation of rods, which become rigid on calcination [88-90] (Fig. 9). Pillared LDH have also been extensively investigated (Fig. 9). The high charge density of these layers makes pillaring difficult but not impossible [32, 33, 91]. The pillared electrodes are stable in room atmosphere and temperature on the order of a month and current density is not changed [34]. Intercalated redox active pillars (H2W12) are stable for up to 20 hours [33]. A particularly exciting method which has yet to be applied to clay-modified electrodes is to carry the delamination method to its logical conclusion and build structures inside out (Fig. 10). In this method an
9
oA
~?o .
.
F%xeN,xc'~x.~Kex,e%l P"z~2S?',.c%:N~2~"hl
Fig. 9. Various pillaring schemes for clays A SiO2/TiO2 (S. Yamanka, Y. Inoue, M. Hatori, E Okumara, M. Yoshikawa, Bull. Chem. Soc. Jpn. 1992, 65, 2494), B rod and pillared clays (A. Galaorneau, A. Barodawall, T. J. Pinnavaia Nature 1995, 374, 529), C terephthalic acid pillared LDH (W. Jones, E Kooli, J. Bovey, in: Sapporo Conference on the Chemistry of Clay and Clay Minerals, Sapporo, A. Yamagishi, ed. 1996, p. 15)
Fig. 10. Model of synthetic methacrylate/magnesium(nickel) phyllosilicate (Y. Fukushima, M. Tani, J. Chem. Soc. Comm. 1995, 241)
12
s.M. Macha and A. Fitch
"'\\\'\\\"\\'~'\N\\\\\\"~
~'~\\\\'~ s i lic a te lay e r ~ , \ \ \ \ \ N IN\"x\\"x\"~ .................. ,N\\\\\\"~l
Fig. 11. Structure of surfactant modifiedclays depends upon the clay charge and the total amountof surfactant intercalatedas well as the chain length of the surfactant(W. E Jaynes, S. A Boyd, Soil Sci. Soc. Am J. 1991, 55, 43)
organo-modified clay will adopt a smectite structure with surfactant tails pointing out. The synthetic route begins with fully exchanged DODAC (dioctadecyldemethylammonium) montmorillonite. Subsequent treatment with a non-anionic surfactant creates a bilayer [921. Surfactant-modified clays have also been used (Fig. 11). Surfactant-modified clays can run the range from expansion of the interlayer with retained access to compete delamination. Cationic surfactant-modifled clay films cause local concentration of anions and neutral molecules and exclude cations and multivalent ions [93]. The charged head group on the surfactant
causes an increased positive charge within the clay interlayer which is countered by confined anions. Neutral, non-polar organic molecules are drawn into the hydrophobic organoclay structure. Other neutral molecules that are not soluble in water can be solubilized by the surfactant-modified clay [93]. On the other hand, cations are ejected from the clay due to electrostatic repulsions. In addition, ions exhibiting too much hydrophilic character (multivalent ions) are also prohibited from diffusion through the clay [82, 94]. Counter anions can participate in ion exchange equilibria and are able to diffuse to the electrode. The configuration of the hydrocarbon chains may vary depending on the temperature of the system. A surfactant bilayer in which all of the hydrocarbon chains exist in trans c9nfigurations results in a rigid association of tilealkyl__ chains called a "gel". At room temperature most surfactant aggregates contain some cis-cofifi~fired segments, disturbing the rigidity of the gel system, This structure is more fluid and is called a "liqui d crystal" structure [93]. The fluid structure allows redox processes to diffuse through the clay through pores between particles [41] and in the disordered zones of the randomly oriented smaller sized particles (defect zones) [80]. A completely different strategy to linking corridors and active sites within the corridors is to do away with diffusion driven charge transport. In this strategy, most notably developed to date by Villemure et al. [95-97], electroactive sites are designed into the walls of the active room (the crystal structure) and charge is delivered through the walls (intercom) via these electroactive sites. This type of strategy requires that the clay material be electroactive, a proposition most easily achieved by synthetic LDH clays. Villemure has extended this concept to synthetic cobalt smectires. Other workers have also investigated the possibility of delivering charge directly to the active site using native redox centers: iron sites within the crystal lattice [98]. This method is intriguing particularly as it is obvious that such redox reactions do occur in nature. Grey soils observed under reducing environments (subsoils and rice patties) exhibit reduced iron giving a bluish or grey cast to high iron content clays which in an oxidizing environment are red in color. Several recent research groups hope to capitalize on the redox properties of these clays to inject charge into the front of advancing aquifers for electrochemical detoxification of contaminated ground waters.
Clays as Architectural Units Work by Stucki [99] indicates that such reactions are slow (h/2 > 1500 s) with respect to an electrochemical time scale. Reduction involves rearrangement of structural hydroxyl groups. The last method of delivering charge would be in hardwiring an electroactive polymer formed within the clay. This method has been investigated most notably by Bard's group [13, 100] and a few others [14] including Joo [87, 101]. The types of electroactive polymers investigated include polypyrole, polythiophene, and aniline.
13
~--~'Off"
I
~ u
iTPPii"
FeTPPS~44- Ru(bpy)33*12*
--I VI. Form Follows Function
The type of application will drive the form of the claymodified electrode. In this section several types of applications of clay-modified electrodes are reviewed. Some of these applications have met the criteria of success and others clearly represent attempts yet to be finalized. Applications are broken into a few broad categories and overlaps will be noted: molecular electronics, photocatalysis, electrocatalysis, enzymatic sensors, stripping sensors, and chiral sensors. Structuring the electrode surface with clays for the development of charge storage devices with unidirectional or controlled directional current flow has been attempted. Figure 5 shows the type of structure constructed in which a clay film concentrates and localizes Ru(NH3)~ + via electrostatic forces. The clay does not reside directly upon the electrode surface but is spaced by a cationic chain which stabilizes a water soluble anionic electroactive porphyrin. The redox potential of the porphyrin is controlled by the pH of the media. The pH determines the identity of the porphyrin axial ligand (water or hydroxide). The formal potential, the electrochemical reversibility, and the heterogeneous rate constant depend upon the nature of this ligand. When hydroxides are present, electron transfer is quenched. When water is present, electron transfer is possible and shuttling of charge through the porphyrin to the charge storage area (clay/ Ru(NH3)~ +) is possible. The gating scheme is shown in Fig. 12. A second type of unidirectional charge transfer was accomplished by mixing and matching of various redox probes within the clay film [65]. As discussed above the physical mobility of the charge carrier depends upon charge and hydrophobicity. Furthermore, the potential associated with reduction within the clay film also depends upon the strength of adsorption. As
potential of
~I.N~(CH2)nII...-VN J+ 7]+v~ "gating" layer
clay layer, charge storage
Fig. 12. Gated electrochemicalactivityfor the structure shownin
Fig. 5 is given schematically. When the pH is low, the anionic porphyrin retained by the cationic linker contains water as an axial ligand. The reduction potential of this compoundlies betweenthat of the electrodevoltageand the reductionpotential for Ru(NH3)~+ contained in the anionic clay.When the pH is raised, the change in the axial ligand of the porphyrin shifts the reduction potential and disrupts the electron transfer chain (adapted from D. Rong, T. E Mallouk, Inorg. Chem. 1993, 32, 1454)
a general rule the trisbipyridine class of metal complexes are strongly adsorbed, exhibit potential shifts positive, and are inaccessible electrochemically due to strong adsorption. By weakening the strength of adsorption slightly via substitution of one bipyridine ligand with two CN groups the compound can be distributed into two different populations, one which is adsorbed within hydrophobic domains (interlayer) of the clay and one which is aquated. The potential of the adsorbed species is shifted positive so that the aquated species can serve as a charge shuttle to it's adsorbed neighbors (Fig. 13). Charge storage is indicated by the asymmetry of the electrochemical peaks. An interesting feature of this system is the ability to recover the stored charge by the use of a second mediator, Fe(phen) 2+. The result is a rectifying structure at the clay surface. The structure is, unfortunately, unstable and the interesting behavior decays with time. A second application which requires directionally controlled charge transport is photocatalysis of water
14
S.M. Macha and A. Fitch Ru'(bpy)32+
' . . . . " -1.31V
fas MV2+/+
~(
s}ow
-0.69V ,,,,.,,,,,==,,, \
/c
iB ,11
k
V k
--
450
n m
\ \ \
H+/H2 -0,24V
\
SlOW k k \
fost
\ \ \
~i ie
,m,
~.28V
Ru(bpy)3 2+
D
Fig. 14, A scheme for photodriven catalysis. 450 nm light excites Ru(bpy)~ + which transfers the electron to a mediator like methyl viologen (MV) which delivers charge to the substrate. A scavenger (TEO or EDTA) is added to replenish the electron source. Fast forward eIectron transfer is required and slow electron transfer back is demanded. Some of these requirements may be met by spatially organizing the reagents
[ 1,0
&O
t
[ .~__L 5.0
__~.............. t 0,4
l
0.2 OJ V vs. SCE
Fig. 13, A Steady state CVof 0.14 mM Fe(CN)z(bpy)a at SWy-1 CME in 0.01 M NazSO4 recorded at 165 minutes: bar, 0.1 ~tA; scan rate 50 mV/s. B Steady state CV of 0.5 mM Fe(Phen)~ + at SWy-t C1V~ in 0,1 M Na2SO4 recorded at 140 minutes: bar, 0.1 gA; scan rate 50 mV/s. The CV shows the presence of two peaks attributed to void volume Fe(phen)~ + at 890 mV and to strongly adsorbed Fe(phen) 2+ adjacent to the electrode at 1015 inV. C CV of 0.14 mM Fe(CN)2(bpy)2 and 0.44 mM Fe(phen)~ + in 0.01M NazSO4 at bare Pt electrode: bar, 0.1 gA; scan rate, 50 mV/s. D Multisweep cyclic volteanmo~am obtained between i00 and 210 minutes soaking of.SWy-1 CME in 0.14 mM Fe(CN)2(bpy)2 and 0.5 mM Fe(phen)~' in 0.01M Na2SO4: scan rate 50roWs; bar, 0.1gA. Diode-like behavior of the CME is observed where charge is stored in the reduction sweep and released in the oxidation sweep 400 mV away (A. Fitch, E Subramanian, J. ElectroanaL Chem. 1993 36Z t77)
splitting. Research in the late 1970s and early 1980s indicated that clays can serve as hosts for sensitizers used for water splitting [102]. The generic reaction mechanism is given in Fig. 14. The photosensitizer accepts a photon which promotes an electron to an excited state where it is launched down an electron relay. The key is to make each of those steps rapid and the reverse reactions slow, a goal which can be accomplished by localization of the various reactants spatially. The second key is that the mediating species
should have good electron transfer kinetics even while localized so that the exchange reactions are fast. Several groups have dew)ted time to understanding the kinetics of charge transfer of various individual species along the proposed electron transfer relay. Specific dyes to be used, either as sensitizers or as mediators, are thione [103], methylene blue [101,104, t05], methyl viologen [106] and tetrathiafulvalene [1071. With thionine, photovoltages of up to 30 mV were detectable when the clay/thionine film was irradiated with 460 nm light. The responsible electrochemical. reaction was postulated to be electron transfer from triplet thionine to Fe(II) or EDTA. The reduced dye dismutates to leucothinoine, which in turn is reoxidized reversibly at the electrode surface. Light irradiated on the thionine-containing clay" catalytically generates the production of electrons in the presence of an elec~on acceptor Fe(II) or EDTA. A very similar mechanism was proposed for a reversible photogalvanic effect observed for methylene blue incorporated into montmorillonite [ 104]. One of the difficulties of opthnizing these reactions is the strength of adsorption of these dyes to the clay (10SM -t for methylene blue and 109M -l for thioflavin [t08] ), Because methylene blue is so strongly adsorbed to clays it can be used to displace exchange-
Clays as Architectural Units
able cations for a facile cation exchange capacity measurement [109]. In addition, these dyes are known to dimerize and/or aggregate processes which affect the efficiency of charge transfer within the clay. Of course the most studied photocatalytic process is one involving the use of ruthenium trisbipyridine as the sensitizer. This compound, too, is known to be adsorbed so strongly that most of the material within the film is electroinactive (see above). The species that are electroactive have been suggested to be an ion pair of the bipyridine complex with a counter ion. These ion pairs are those compounds adsorbed in excess of the cation exchange capacity of the clay. Other studies have focussed on the quenching of the compound by the clay itself, a process which would diminish the excited state lifetime and impede the electron relay. Despite these difficulties some success in photocatalysis has been reported. Photocatalytic reduction of oxygen by Ru(bpy)32+ supported on clays has been reported where Co(teta)(H2Oz)23+ serves as a charge mediator [7, 110]. Like the applications mentioned above, the use of clays as templates for enzyme based analytical sensors requires control of the directionality of charge transfer and the localization of various reagents in specific spatial organization. A novel application of clays relies on spatial control provided not by the individual aggregates of clay but upon entire films of clays [111]. A layer of nontronite at the surface of the electrode serves to localize methyl viologen near the electrode surface. This layer is followed by a layer of glucose oxidase enzyme, which is capped with a second layer of the nontronite. The methyl viologen serves as a charge shuttle to the trapped enzyme. The outer layer of clay protects the enzyme from denaturation and the electrode surface from fouling. The outer layer of clay also allows glucose to permeate to the enzyme, but prevents (via electrostatics) the loss of hydrogen peroxide to the bulk solution. The use of nontronite is driven by the high iron content of these clays and the nontronite is thought to promote the conversion of the hydrogen peroxide to water. Methyl viologen within the bottom or foundation layer of the clay then regenerates the catalytic iron site within the nontronite (Fig. 15). Interferences of anionic substates were minimal as might be expected even at the level of 0.5 to 0.1 mM. The detection limit for glucose was 5 btM. Spatial organization has also been achieved at claymodified electrodes where the goal is to protect the
15
Elea~ode~
k',~\\'\\\"~ x',,\\\",\\"~. ........
........
