US 20020008038 A1
An array of selectively addressible microelectrodes for combinatorial synthesis of complex polymers or alloys.
Field of the Invention The present invention is generally related to the combinatorial syntheses of compounds, particularly of biological compounds and metal alloys, on microelectrode arrays. The invention relates in particular to the combinatorial syntheses of oligonucleotides, peptides and biologically active compounds on microelectrode arrays under electrochemical control.
Background of the Invention: The simultaneous synthesis of massive numbers of different oligonucleotides, peptides, and biologically active compounds has applications in the identification of new drugs, in the mapping of genes, and in testing for interactions between biological molecules. Arrays of different genes are applied, for example, in diagnosing various diseases, such as cancer and hereditary diseases, in sequencing the human genome, and identification of drugs capable of blocking bioconjugation reactions.
 Conventional synthetic methods are time consuming and the step-by-step preparation of individual potentially bioactive agents consumes the largest fraction of the capital invested in the development of new drugs. To address these problems, a variety of methods for combinatorial syntheses has been developed. One commonly used method involves the definition of pixels where a nucleotide is added to an existing sequence through photochemical methods. Pixels for the occurrence or the avoidance of a chemically synthetic step can be defined by photolithographic methods. Alternatively, reactions in a particular pixel can be driven by light when photochemically active reactants are used. Non-photochemical methods of combinatorial syntheses involve valved grids of microfluidic channels, such that a reaction occurs in microreactors at the intersection points of the channels with open valves. In general, these methods have resulted in pixels or microreactors that were as small as about 35 µm x 35 µm in their cross-sectional area. The syntheses involved small but still significant amounts of expensive reactants, typically more than a billion molecules in each reaction in each pixel.
 Michael Heller et.al, (U.S. Patent No. 5, 605,662) discloses a procedure for producing arrays of selectively addressable microelectrodes. To transport and concentrate specific charged oligonucleotides, an electrophoretic field is used. For example, a positive potential is applied to a specific microelectrode in order to attract a negatively charged oligonucleotide and to concentrate the charged moiety at the site.
 The present invention provides a method for parallel syntheses where a specific reaction is induced and controlled at the microelectrode.
 This application is a continuation application of PCT Application No. PCT/US99/14459 A1, which claims priority to Provisional Patent Application Serial No. 60/090,520, which applications are hereby incorporated for all purposes.
 The present invention provides an electronic device and method for the combinatorial synthesis of biopolymers, alloys, and non-stochiometric inorganic compounds. The device includes an array of selectively addressable microelecrodes which preferably includes on its surface a conductive matrix, which may be a redox polymer or hydrogel. The matrix further comprises functional reactive groups, such as amines, aldehydes, carboxyllic acids, active esters, and the like, useful for the sequential addition of monomeric units to form polymeric compounds. For example, in one embodiment of the invention, the array includes amines in the matrix, to which nucleotides or oligonucleotides may be sequentially added by chemical reaction.
 In the method of the invention, a potential sufficient to induce a Faradaic reaction is selectively applied to each microelectrode to induce binding of a reactant to one or more of the functional reactive groups in the matrix. A complex compound is synthesized on the microelectrode by repeated Faradaic reactions, i.e., by repeated applications of a potential sufficient to induce the desired reactions.
 In an alternative embodiment, the method of the invention includes application of a potential to induce a Faradaic reaction and selectively deposit metals in layers onto a microelectrode. Optionally, a portion of one or more of the metal layers is etched or dissolved via subsequent Faradaic reaction at the microelectrode. Heating, oxadation, sulfadation, or consolidation of two or more layers causes formation of an alloy or non-stochiometric inorganic compound.
 Figure 1is a diagram showing the arrangement of the electrode or array surface bearing the reactive groups as a monolayer or in a polymeric matrix. The electrode surface is schematically drawn as a flat surface but the surface roughness (geometric area/actual area) may vary between 1 and 1000.
Figure 2 is a flow diagrams for the syntheses of oligopeptides (poly amino acids, peptide nucleic acids or other oligopeptides). The coupling is carried out with carbodiimide in conjunction with N-hydroxysuccinimide. The solid bar connecting the COOH function to the electrode is made of condensed amino acids. The amino acids may be natural amino acids or amino acids not found in nature.
Figure 3 is a flow diagram for the synthesis of oligonucleotides. The syntheses is modulated by the electrochemically generated protons. At potentials where protons are produced the local pH drops and the protecting dimethoxytrityl (DMT) on the 5" end of the oligonucleotide is cleaved. The cleavage enables the extension of the oligonucleotide.
Figure 4 is a flow diagram showing the synthesis of an oligopeptide with an electrochemically pH modulated enzyme catalyzed step.
Figure 5 is a schematic diagram showing the computerized automated potentiostatic and liquid delivery control to the microelectrode array.
Microelectrode ArrayThe electronic device of the invention includes an array of selectively addressable microelectrodes. A matrix comprising at least carbon (C) and hydrogen (H), and preferably capable of conducting electrons or holes, and also of conducting ions, is disposed on a surface of the electrodes. The conductive matrix is disposed on the microelectrode with a thickness greater than 3 nanometers, and preferably in the range of 3 nanometers to 20 microns, most preferably with a thickness of 5 or more nanometers. Most preferably, the matrix comprises a redox polymer or redox hydrogel, having a molecular weight of more than 104 daltons. Most preferably, the redox polymer or hydrogel comprises a fast redox couple, such as complexes of transition metals such as osmium, ruthenium, iron, copper, and cobalt.
 The matrix is modified to include functional reactive groups, such as amines, aldehydes, carboxyllic acids, active esters, and the like.
 The arrays of electrodes on which the combinatorial syntheses is carried out are usually large. Typically the number of electrodes in the array is greater than 4, greater than 100, and is preferably greater 1000 and it is most preferably greater than 10,000.
 The longest dimension of a microelectrode is preferably less than 100 µm, and is more preferably between 0.1 and 10 µm. The shape of a microelectrode may be, for example, oval, rectangular or preferably circular. Fine line electrodes, that are typically longer than 100 µm and have typical widths between 0.1 µm and 20 µm, can also be used. It is not necessary that the microelectrode surface be flat. A metal or carbon microelectrode surface may be, for example, concave or convex relative to the plane of the surrounding insulating layer. Preferably all the microelectrodes in the array are of the same dimensions and are spaced individually, or grouped in elements of the array, those elements forming organized patterns. It is preferred that the pattern have repeating elements and it is particularly preferred that the elements have rotational and/or translational symmetry. An example of a preferred pattern is one where all electrodes except those at the periphery of the array are equidistant from each other and have the same number of nearest neighbors.
 Typically it is desirable to space the electrodes at distances greater than 10 times the diameter or smallest dimension of the microelectrodes and it is most preferred to space the electrodes at distances greater than 20 times the diameter or smallest dimension of the microelectrodes. It is generally desirable to make the electrodes small, so as to minimize the amount of material consumed in the reaction driven at a subset of electrodes. Preferably fewer than one hundred million molecules are reacted at an electrode. It is more preferred that fewer than one million be reacted and it is most preferred that fewer than one hundred thousand be reacted. Reduction of electrode size also allows closer spacing of electrodes in an array, i.e. denser packing, meaning more electrodes per unit area. Typically the density of microelectrodes is greater than 100/cm2; and most preferably it is greater than 100,000/cm2.
 The electrodes are made of a conductor which is preferably non-corroding and may comprise gold, carbon, tantalum, ruthenium dioxide, a conductive metal carbide, a conductive metal nitride, or a conductive metal oxide, or a Group VIII metal such as platinum, iridium, rhodium, rhenium, palladium or ruthenium. The electrodes are insulated from each other by an insulator that may be organic or inorganic. Examples of inorganic insulators are silicon dioxide and silicon nitride and insulating doped silicon. Examples of organic insulators are polymers, such as poly(methyl methacrylate), polyimides, polyesters, or polyamides. The substrate of the structure can be any material having the desired mechanical properties such as glass, ceramic, silicon, plastic, or a metal coated with an insulating film. The electrical contacts to each element in the array may be in the plane of the electrodes, or in the plane of the electrodes and in a second plane, or in multiple planes including or excluding the plane of the electrodes. Contacts to the electrodes may be formed individually to each electrode, to groups of electrodes, or to rows and columns of electrodes in a grid.
 The elements of the array, i.e. its microelectrodes, can be connected to potential or current sources by either hard wiring or by light. Hard wiring means that the connection involves a continuous electrical path through electrical conductors, particularly metallic conductors or carbon. Connection by light is possible when a photoconductor is used, so that the illuminated areas conduct electrons or holes, while the non-illuminated areas do not conduct. Instead of applying an external potential or passing a current from an external current source, the potential at or the current passing through a microelectrode of the array can also be photogenerated. For generation of a photopotential or a photocurrent it is preferred that the microelectrode material be or comprise a semiconductor. The semiconductor may be an inorganic semiconductor, such as silicon or a III-V compound. It may also be an organic compound, such as a polymeric or non-polymeric organic compound used in light emitting diodes or photodiodes or photovoltaic cells.
 The electrochemical circuits require, in addition to the microelectrodes on which the syntheses are carried out, known as working electrodes, at least one reference electrode and may require at least one counter-electrode, that may or may not be the reference electrode. These added electrodes may be part of the array or may not be part of it. The reference electrode may be a conventional reference electrode such as a saturated calomel or silver/silver chloride electrode or it may be a pseudo-reference electrode consisting of a conducting material in contact with the solution. The counter electrode can be a conducting material with a surface area which is preferably to at least twice the sum of the combined surfaces area of the microelectrodes in the working electrode array. There may only be one reference electrode and one counter electrode for the entire array.
Reactive groups on the microelectrode surface: The electrode surface may be modified with reactive groups in a variety of ways depending in the electrode material of choice. Two examples are shown in Figure 1. On the left, a matrix with remote reactive functions R, such as a carboxylic acid functions, extending into the solution is shown. On the right, a polymeric gel/hydrogel support incorporating reactive functions R, such as carboxylic acid groups, attached to the back bone of the polymer is shown. Preferably the matrix should not be detached under the conditions of the reaction that is to be carried out. If the reaction is carried out under potentiostatic control, it is preferred that the potential which is applied should affect only the reactive groups or their reactions, not their attachment to the electrode.
 The conducting microelectrode, or group of microelectrodes, or the surface of the entire microelectrode array, including the surface of the insulating material between the electrodes, may be modified with a matrix. This matrix can, for example, be a matrix having affinity for a reactant, such as a polycation when the reactant is a nucleotide or a nucleotide derivative which is an anion. It can also be a matrix comprising a redox polymer or a redox hydrogel, the polymer or gel containing attached redox centers. When a redox hydrogel or conductive polymer is used, it is preferred that only the conducting region of the array be covered.
Method of the InventionIn the method of the invention, an array of desired chemical compounds is selectively synthesized on selectively addressable microelectrodes. A potential sufficient to cause a Faradaic reaction in the immediate vicinity of the microelectrode is selectively applied to each microelectrode to induce binding of a first reactant to one or more funcional reactive groups present in the matrix disposed on the microelectrode. After binding of a first reactant to a functional reactive group in the matrix, the selective application of potential is repeated to cause a second Faradaic reaction and induce binding of a second reactant to the first bound reactant. This procedure is repeated a sufficient number of times to synthesize a desired compound on the microelectrode, thus forming an array of individual compounds.
 The reactants may be, for example, nucleotides, amino acids, chemical moieties, or other molecular subunits such as oligonucleotides or peptides that are induced to bind amines, aldehydes, carboxylic acids, active esters, and the present in the matrix.
 The Faradaic reaction preferably causes a chemcial change in one or more of the fuctional reactive groups in the matrix, in a reactant supplied to the microelectrode, for example in a solution bathing the electrode, or in a further chemical species present in the solution. The chemical change can result in a change in pH, in the oxidation or reduction of a functional reactive group on the matrix, of the supplied reactant, or of a further chemical species in the reaction solution.
Control through a faradaic process: A Faradaic current is a current that passes when a species is electrooxidized or electroreduced at an electrode. This current is contrasted with a capacitive or polarization current where a current will flow for a period of time when a potential is applied without an electrooxidation or electroreduction reaction taking place.
 The capacitive or polarization current flows as a result of the redistribution of ions in the solution (or film) near the electrode. The electrophoretic process, where a macromolecular polyelectrolyte migrates to an electrode is such a non-Faradaic process. After some time this current stops because of the accumulation of cations near a negative electrode or of anions near a positive electrode.
 Upon flow of a Faradaic current, a substance is electrolyzed. At least one species in the immediate vicinity of the microelectrode (that is, the distance from the microelectrode where the microelectrode can cause redox reactions) is electrooxidized and at least one species is electroreduced. As a result, there is a net change in the chemical composition in the immediate vicinity of the microelectrode, not a mere re-distribution of ions in the cell.
 Faradaic reactions, unlike capacitive or polarization processes, take place only when sufficient potential is applied to an electrode. This is well known form any text in electrochemistry or even physical chemistry. The threshold potential is the formal half cell potential for the reaction driven. When the reaction is reversible, then the potential is the thermodynamic potential, - ΔG/(nF), where ΔG is the change in the Gibbs free energy, n is the number of electrons gained or lost by the reduced or oxidized species and F is Faraday"s constant. For example, the threshold potential for electrolysis of water is 1.23 V (measured relative the standard hydrogen electrode). A Faradaic current resulting in water electrolysis flows only when this potential is exceeded.
 Acid is produced by electrolysis of water at a half cell potential depending on the pH and temperature. The higher the pH, the more reducing (negative) this potential is, shifting at 25oC by 59 mV for each pH unit (when the partial pressure of hydrogen is fixed).Reversible potentials are listed in most handbooks of chemistry, such as the Handbook of Chemistry and Physics.
 The Faradaic reaction need not be reversible. For example, when ascorbic acid is electrooxidized the reaction is irreversible. The composition of the solution within the cell changes in the reaction. The compound is not only redistributed but converted.
 The local concentration of ions, particularly of protons, and thus the local pH can be controlled through controlling the current density passing through a particular electrode, which can be in turn, controlled by the applied potential. Upon electrooxidation, protons are usually released and upon electroreduction protons are usually consumed.
 The reactant in the faradaic process can be oxygen, the solvent itself, e.g. water; a readily electrooxidizable organic compound such as ascorbic acid, or a readily electroreducible organic compound, such as benzoquinone. The reactant can also be an inorganic or metal-organic ion, such as a complex of iron, cobalt, ruthenium, osmium or copper. For example, the reactant can be a bipyridine complex of ruthenium or osmium or cobalt; a cyanide complex of iron; or a metallocene derivative, such as a ferrocene derivative.
 The metal of the electrode or a metal ion may be reduced or oxidized to an oxidation state which is suitable for the modulation of the chemical reaction at the surface of the electrode. The reactant may also be an organic molecule which can be electroreduced or electrooxidized to a species which chelates with a dissolved metal ion. The electrochemically formed chelator reduces the local free metal ion concentration at the electrode surface.
Control through application of a potential without the occurrence of a substantial faradaic (electroreduction or electrooxidation) reaction:As taught in Heller et.al., U.S. Patent No. 5,605,662, the local concentration of ions at a particular microelectrode or group of microelectrodes may be controlled by a capacitive process, such as a process of attracting cations, repelling cations, attracting anions or repelling anions. Examples of the anions attracted or repelled are hydrated protons (H+), hydroxide anions (OH-) and Zn2+, Mg2+, Ca2+ or Cu2+ ions.
 The potential applied to the electrode for eletrophoretic transport and concentration of a charged species is mirrored by the ions in the solution. Thus a positive applied potential (relative to the potential of zero charge (PZC) of the particular metal in the solution) will draw anions from the bulk solution, causing a local build up of anions at the electrode interface relative to the bulk solution. The attracted anions will replace cations at the surface. The pH or ion concentration in the proximity of an electrode can be controlled through the potential applied, and/or the buffer added to the solution and/or the concentration of the relevant ion in the solution.
 The number of electrodes in the array is greater than ten thousand and the dimension of each microelectrode is 10 µm or smaller. The use of microelectrodes ensures that mass transport of species to and away from the electrode surface is efficient. Concentration polarization, which is the change in concentration of a reactant near an operating electrode, and is important in defining the necessary spacing of the electrodes, is also reduced when microelectrodes are used. Thus denser electrode arrays can be made.
 The extent of concentration polarization, meaning the distance to which a concentration change extends from a particular electrode depends on the diffusion coefficient of the reactant, the dimension of the microelectrode and the applied potential. The occurrence or prevention of a reaction at a particular electrode is controlled by applying a potential to, or passing a current through, the electrode at which the occurrence of the reaction is desired or in which the occurrence of the reaction is to be prevented. The application of a potential induces a faradaic or non-faradaic process, depending upon the specific potential applied and the species present in the vicinity of the electrode, which changes the concentration of at least one ion or molecule relative to the bulk solution.
 The concentration of an ion near a particular microelectrode can be controlled through a faradaic process, such as an electroreduction reaction or an electrooxidation reaction. It is well known that the potential where an electrooxidation or electroreduction reaction takes place depends on the presence of a particular electrooxidizable or electroreducible species in the solution. In general the local pH at and near the electrode surface is increased when an electroreduction reaction is taking place and is decreased when an electrooxidation reaction is taking place. Usually the higher the current density the greater the change in local pH. The magnitude of the change also depends on the concentration of buffering agents, decreasing when the buffer concentration is raised. For example, the local pH may be increased at a microelectrode by the electroreduction of oxygen, a reaction where protons are consumed.
 The pH can be decreased locally by the electrooxidation of water to hydrogen peroxide or to oxygen, or by the electrooxidation of a solute, such as ascorbate ion, that takes place at a less oxidizing potential than the potential for electrooxidation of water. In these reactions protons are released. The extent of pH change may be controlled by the potential applied or by the buffering capacity of the solution. In summary the local concentration of a particular species near a particular electrode is controlled through the potential applied or the current passed, or by the concentration and nature of an added buffer.
 There are different ways through which the local concentrations of ions at a microelectrode surface may control the occurrence or non-occurrence of a reaction. For example, a reactive species such as an O-acylisourea or an N-hydroxysuccinimide ester does not react with an amine in an acid environment, where the amine is protonated. However, at neutral or slightly basic pH the reaction with the amine does take place and an amide is formed. Through controlling the local concentration of protons, or of other ions, or of a molecule such as ascorbate which is required for a reaction to take place at a defined potential, it is also possible to modulate the activity of enzymes that catalyze either the formation or the breakage of chemical bonds, for example bonds formed in condensation or hydrolysis reactions. There are numerous well documented examples, found in textbooks of biochemistry, where an enzyme is active only in a well defined pH range. Also some enzyme-catalyzed reactions require the presence of a particular ion, such as Zn2+ or Mg2+, or Ca2+; others can be reversibly or irreversibly inhibited by the presence of other ions, e.g. Cu2+.
 Examples of non-faradaic processes whereby the local concentration of an ion is changed include local increase of the concentration of a cation when a negative potential is applied to an electrode; increase in the local concentration of an anion when a positive potential is applied; local decrease in concentration of a cation when a positive potential is applied; and local decrease in the concentration of an anion when a negative potential is applied. The above four processes do not require the occurrence of an electrochemical reaction.
 Examples of relevant reactions known to take place in a particular pH domain: Examples of such reactions include the syntheses of amides such as those of oligopeptides including peptides, proteins and in peptide nucleic acids via a carbodiimide involving reaction; syntheses involving N-hydroxysuccinimide esters; syntheses involving imidates; and synthesis involving polyphosphates. Figure 2 shows an example of a flow diagram of a synthesis utilizing a carbodiimide and Nfor the formation of a peptide bond on an electrode surface.
 When the objective is not to make amides but other compounds, other active reactants may be used. For example an epoxide may be used for a reaction with an amine; or an alkyl halide, particularly an alkyl iodide or an alkyl bromide, for a reaction of an amine. The reaction takes place in neutral or basic solutions, but not in acid ones. Control can be either by adjusting the local pH through a faradaic or a non-faradaic process such that the reaction proceeds, or through inhibiting the reaction through adjusting the local pH such that the reaction is prevented.
 For the electrically controlled synthesis of oligonucleotides, the well-known phosphoramidite method can be applied. Figure 3 shows the flow diagram of a current or a potential controllable synthetic route, based on the phosphoramidite method and involving a dimethoxytrityl (DMT) protecting group on the 5" terminus of the reacting base. When protons are generated at the microelectrode, the protecting group on the 5" is removed. This removal or de-protection allows the extension of the oligonucleotide once the cycle is repeated. Because the coupling step is also pH dependent, the coupling can also be modulated by an applied potential or a current passed through the electrode. Also the iodine required for the oxidation step may be electrochemically generated when the solution comprises iodide ions. Any unreacted oligonucleotide is capped by acylation with acetic anhydride to avoid extension of any undesired sequences. This capping is not shown in Figure 3.
Enzyme Catalyzed Reactions: There are families of enzymes known to catalyze the hydrolysis or formation of amides in peptides, proteins, and protein nucleic acids and of phosphate ester links in oligonucleotides or DNA and of glycosidic linkages in oligosaccharides. The families of these enzymes include, for example, kinases, peptidases, proteolytic enzymes and hydrolases, transferases, ligases. The enzymes" activity can be controlled by the local adjustment of the pH. Also the activity of some enzymes may be enhanced by the local electrochemical reaction or process (meaning application of potential or passage of current) increase in the concentration of ions such as Mg2+, Ca2+ , or Zn2+, or may be decreased by electrochemically increasing the local concentration of ions such as Cu2+. The enzyme may cause the addition, through formation of a covalent bond, of a dissolved species in the solution to which the array is exposed; or it may be such that it cleaves a terminal function on a molecule already on the electrode so as to enable a reaction at that particular electrode; or it may cause hydrolysis of a functional group of a molecule on an electrode.
 A schematic diagram of the steps involved in the enzyme catalyzed reaction is shown in figure 4. The method described is the step by step synthesis of a peptide nucleic acid modulated by a non-specific amidase, particularly acrylamide amidohydrolase from Pseudomonas aeruginosa. Enrichment of the zone near the microelectrode in protons prevents the enzymatic cleavage of the terminal amide and therefore the subsequent carbodiimide or N-hydroxysuccinimide ester (NHS) utilizing condensation reaction whereby an amino acid is added to the peptide, peptide nucleic acid or protein on the electrode.Automated system: The system may be automated by integrating a computer to control the flow of liquid containing reactants by a series of valves and also the potentiostat as shown in figure 5.
 For example, a series of different oligonucleotides can be formed on an array of electrodes using the apparatus illustrated in figure 5 and the process illustrated in figure 3. An array of carbon or metallic electrodes is formed. Each of the electrodes is coated with a coupling species. The coupling species includes a reactive functionality that is initially capped with a protective group. The array is placed within a flow device that is coupled to a valve, which is, in turn, coupled to four reservoirs of molecular subunits, corresponding to the four different bases of DNA; adenine, guanine, thymine, and cytosine. Each of the molecular subunits includes a first reactive functionality for coupling to a deprotected reactive functionality of the coupling species or a previously deposited molecular subunit. Each of the molecular subunits also includes a second reactive functionality that is initially capped by a protecting group.
 In operation, the valve is directed to open and allow one of the four solutions of molecular subunits to flow into contact with the array. Each electrode of the array is individually coupled to a potentiostat, typically, under computer-control. A potential is selectively applied to (or, alternatively, a current is passed through) those electrodes at which the particular molecular subunit is to be deposited. The electrical potential (or current) causes a change in the concentration of an anionic or cationic species (e.g., a change in pH) that leads to the removal of the protecting group on the coupling species and results in a reaction of the reactive subunit on the coupling species and the first reactive functionality of the molecular subunit.
 The solution containing the first molecular subunit is then removed and the valve is directed to open and allow a second solution with a different molecular subunit to flow into contact with the array. Again, a potential is selectively applied to those electrodes at which this particular molecular subunit is to be attached, including those electrode at which this second molecular subunit is to be attached to the first molecular subunit. The application of the potential results in the removal of the protecting group from the reactive group of the coupling species or the second reactive group of the previously coupled molecular subunit. This procedure is repeated until the desired oligonucleotide sequences are all formed on the array of electrodes.
 As an example, a four electrode array can be formed with the following oligonucleotides on the electrodes:
 Electrode 1 - AGTC
 Electrode 2 - ATGC
 Electrode 3 - GTGC
 Electrode 4 - TGCA
 One exemplary process includes the steps in the following table:
 Other sequences of steps could also be used to obtain the same array of electrodes. In addition, the same principle can be used to form other molecules, such as peptides or proteins, on electrodes that utilize a small set of subunits, such as amino acids.
Combinatorial Synthesis of Inorganic MaterialsIn the combinatorial syntheses of inorganic compounds or organic compounds having inorganic backbones a potential is applied to a microelectrode of the array or a current is passed through an electrode of the array, such that the local pH is changed or the local concentration of another ion is changed. When the reactive material on the microelectrode is a metal or metal oxide, then typically a reduction in pH, the application of a positive potential, or the occurrence of a local electrooxidation reaction can accelerate the dissolution of the oxide or the metal. For example, metals, such as zinc or aluminum, or an oxide of such metals are more rapidly dissolved by a local rise in pH. The rate of removal is adjusted through local control of the pH, the potential, or the current. Thus by varying in a gradual manner the potential or the current density the amount of residual material residing on an electrode after partial stripping of a layer can be increased or decreased. Similarly by driving an electroreduction reaction, whereupon the pH is increased, the amount of residual material may be increased when the film is in an etching solution, such as an acid solution.
 The removal and/or deposition of inorganic material can be useful in a variety of circumstances, including, for example, the combinatorial formation of non-stoichiometric materials. These materials are often tested for various properties, including, for example, fluorescence wavelength, fluorescence quantum yield, magnetic properties, and dielectric constant. It is often useful to test a range of different non-stoichiometric combinations. A material is "non-stoichiometric" if the composition can not be expressed as a chemical formula using numbers or integers of smaller than five.
 One method for forming a range of non-stoichiometric combinations includes forming an array of metal electrodes 102 on a substrate 114, as shown in Figure 1. By selectively applying different potentials or by varying the duration of the applied potential to the electrodes of the array, different amounts of the electrode can be removed.
 The rate of removal is determined, at least in part, by the local concentration of other anionic or cationic species around the electrode. This local concentration is modified by the potential applied to the electrode. For example, the pH can be altered by applying a potential, which can then cause the dissolution of the a portion of the metal electrode. The amount of the metal electrode that is removed depends, at least in part, on the potential, the current density through the electrode, and the period of time the potential is applied or the current is passed. By varying any of these parameters across the array, the amount of material removed varies.
 Next, a second metal can be deposited on the electrodes and the process of applying a potential is repeated. This can continue for any number of metal deposition steps. Alternatively, the metal may be selectively deposited by applying a potential that causes electroreduction of metal cations from a solution. The amount of metal that is deposited depends on the potential, the current density through the electrode, and the period of time the potential is applied. In yet another embodiment, porous films of metal compounds, such as, for example, metal oxides, can be formed on the electrode and a portion of the metal compounds can be selectively removed by applying a potential.
 After all of the metal and/or metal compound depositions are performed, and an array of electrode structures with different combinations of metals and metal compounds are formed, the metal and/or metal compounds can be converted into a desired alloy or compound by, for example, heating, oxidation, sulfidation, other chemical reactions, or consolidation of the various layers.
 This method can be useful for the formation of an array of different non-stoichiometric combinations of materials, each combination being determined, at least in part, by the particular potentials and durations applied to the electrode during each step.
 Yet another means of control involves increasing the local concentration of an anion with which the metal of the electrode may react to form a soluble complex. For example, gold is known to react at mildly oxidizing potentials with dissolved chloride ions. If the local concentration of chloride ions is increased by applying a positive potential, then the rate of the dissolution of the gold will also increase, even if the local pH is unchanged.
 In the hydrolytic reaction and precipitation of solution phase inorganic or mixed organic-inorganic compounds, such as halides or alkoxides of Si, Ti, Zr or Al, the nature and reactivity of the product may be controlled at an electrode of the array through the local change in pH. The formation of the solution phase inorganic or mixed organic-inorganic compounds, such as a polymer, can be controlled by application or non-application of a potential to each electrode in the array. In addition, the structure of the compound may, at least in some cases, be dependent on the potential applied at the electrode. For example, the polymer formed upon hydrolysis of a silicone precursor, such as methyl trimethoxysilane depends on the local pH. Ladder-type silsesquioxanes are often formed at higher pH.
Exemplary array and usePVP-Os-NH2: A redox polymer comprising a poly (4-vinyl pyridinie) backbone here about 0.10 of the pyridines are complexed with [Os(bpy)2Cl]+/2+; and about 20% are reacted with 2-bromoethylamine and are thereby quarternized ( bpy is 2,2'-bipyridine; PEGDGE is poly(ethylene glycol)diglycidyl ether , molecular weight 400-600; SCE is standard calomel electrode potential.
 An array of four gold electrodes is produced on a quartz plate, each electrode spaced a distance of 100 micrometers from the other, and insulated from eachother. Each gold electrode (2 micrometers in diameter) is connected to a contact pad. A solution of PVP-OS-NH2 (10 mg/ml) is incubated with PEGDGE (1 mg/ml) for two hours at 37oC, pH7.
 The electrodes are placed into contact with the PVP-OS-NH2 / PEGDGE solution, and a potential of -0.6 Volts (SCE) is applied for 10 minutes to electrophoretically deposit and crosslink the redox polymer. The solution is replaced with a solution of oligonucleotides (DNA or RNA) (10 mg/ml), and the potential is reversed to +0.6 Volts (SCE) for ten minutes. The array is then placed into a solution of ascorbic acid and sodium ascorbate (pH 7.5), total concentration 0.5mM. A potential of 0.4 volts (SCE) is then applied to two of the electrodes of the array.
 A drop of micrococcal nuclease (300,000 U/ml) from Staphlococcus aureus is added to the array. DNA or RNA is hydrolized only at those electrodes to which no potential is applied, and is not hydrolyzed at those electrodes at those electrodes to which a potential is applied.
 The foregoing description contains numerous references to publications and patents, each of which is hereby incorporated by reference, for all purposes, as if fully set forth.