Analyst, Decwnhei- 1996, Vol. 121 ( I 751-I 758) 1751 Rigid Carbon-Polymer Biocomposites for Electrochemical Sensing* A Review Salvador Alegret Grup dc Sensors i Biosensors, Departament de Qui'mica, Uniwrsital Autdnoma de Barcelona, 08293 Bellatei-ra, Catalonia, Spain This paper reviews the use of biocomposite materials in the construction of amperometric biosensors. These rigid composites are formed by dispersing graphite particles in assorted polymers (especially epoxy resins). These composites are bulk-modified biologically (adding enzymes and cofactors) and chemically (blending mediators and catalysts). Keywords: Biocomposites; conducting composites; rigid c*arbon-polyniei- hioconiposites; amperometric biosensor; elrc~ti.ochemic.al seiisor; reijien) Introduction Most strategies in analytical chemistry today call for complex instrumentation and considerable support, including special laboratory facilities and highly skilled personnel.Chemical sensors are a key element of novel strategies applied to analytical instrumentation. Sensors and sensor-based devices provide original solutions without the need for complex instruments or a huge support infrastructure. Chemical sensors are devices that are small, robust, portable and easy to use. Additionally, they do not need reagents to operate and they can yield reliable information continuously. A chemical sensor has two distinctive parts: a selective recognition component (receptor) and an element (transducer) that converts the primary signal produced by the receptor during the recognition event into a more useful secondary signal.The nature of the primary signal can be thermal, mass, electro- chemical or optical and usually has to be transduced to an electrical signal. This secondary electrical signal contains the codified chemical information from the sample. Several disciplines have to converge in the design of these devices. The design of sensors with biological recognition components such as enzymes, immunological species, che- moreceptors and DNA strands is receiving great attention nowadays. The chemical selectivity shown by these biocom- ponentc is very high. Sensors of this kind are known as hiosc~nsors. Biosensor science and technology use physical and chemical immobilization procedures to couple biological recog- nition elements to appropriate transduction devices.Generally, the biological material is fitted on the surface of transducers using complex and wet immobilization procedures. However, these procedures are seldom suitable for mass production. Amperometric biosensors, usually formed by biologically surface-modified voltammetric electrodes, are gaining increas- ing importance owing to their high reliability, robustness and sencitivity. I Efforts continue to increase the quality of the electrochemical response. Additionally, new materials and ' Prcscnted 'it the 6th European Conference on Electroanalysis, Durham. March 25-29. 1996. immobilization techniques are being tried for the mass produc- tion of these devices. In this context, the present review covers recent work in the field of amperometric biosensors based on new types of materials known as hiocomposites.These materials are formed by rigid conductive composites based on carbon-polymer matrices where the biological material (enzymes) as well as other modifiers (cofactors, mediators, catalysts, additives, etc.) are jointly bulk-immobilized. Conducting Composites A composite is formed by the combination of two or more phases of different nature. Each phase maintains its individual traits, but the mixture may show new physical, chemical or biological properties. If one of the phases is an electrical conductor, the overall electrical properties of the conducting composite will be determined by the nature, the relative content and the distribution of each phase. Electrical resistance depends on the connectivity of the conductor particles in the matrix of the composite.Several conductimetric chemical sensors are based on the disruptive action of organic vapours on the conducting filaments of the material.2J Simple and inexpensive all-solid-state potentiometric sensors have been developed by replacing the metal substrate with a graphite+poxy or a metal- epoxy composite. These composites are mouldable before curing so sensors of different shapes and sizes can be constructed. Ion-selective membranes adhere better to these materials and the resulting devices are simple and inexpensive, show prolonged lifetimes4.5 and the quality of their response is acceptable for analytical applications. An extensive review of ion-selective electrodes based on conducting epoxy composites appeared recently.6 Conducting Composites for Amperometric Sensing The polymer gives the biocomposite a certain physical, chemical or biological stability.The biocomposite acquires particular electrochemical traits from the distribution of the conductive phase in the bulk and, consequently, on the surface of the biocomposite. Carbon materials (graphite, carbon black, etc-.) are ideal conductive phases for composites used in amperometric sensors. These materials have a high chemical inertia and show a wide range of working potentials. They also have a low electrical resistance (approximately 1 0-4 52 cm) and a crystal structure responsible for low residual currents. If the surface of a macroelectrode is reduced, the signal and the associated noise also diminish.In microelectrodes, accord- ing to Oldham,7>8 the perimeter of the sensing surface has a greater influence on the signal. By means of this edge effect, non-linear diffusion is established and the quality of the signal is enhanced. This enhancement is shown by a higher signal-to- noise ratio and lower detection limits. These features and their1752 Analyst, December 1996, Vol. 121 inherent small size have raised interest in microelectrodes. However, the low currents produced call for complex and expensive instrumentation. If small sensors are not required, an alternative is to build carbon fibre arrays separated by an insulating matrix and connected in parallel.9 The signal produced by this macroelectrode formed by a carbon fibre ensemble is the sum of the signals of the individual microelectrodes.The size of the resulting signal is equivalent to the signal produced by a carbon rod of the same active surface but showing the signal-to-noise ratio of a microelectrode. The construction of these ensembles is difficult. However, an equivalent device can be constructed when a composite is made of small conductive particles dispersed in a polymer matrix. Additionally, these devices are easier to build. The selectivity and sensitivity of an amperometric sensor are greatly enhanced if the surface is modified with certain chemical and biological species. One of the key advantages of composite-based sensors is the ease of bulk modification compared with the modification of the surface of a pure conductor, which is usually complex and costly.Conductive composites are modified easily because of their plasticity before curing. Modifying fillers can be blended into the matrix, conferring new abilities on the resulting composite. These new abilities include immobilization of reagents involved in the electrochemical reaction, electrocatalysis, preferential preconcentration and surface structuring. Soft versus Rigid Conductive Composites Adams'O proposed the use of soft carbon pastes to build amperometric transducers. These pastes are built by mixing an inert conductor (e.g., graphite powder) with a non-conducting liquid (e.g., paraffin oil, silicone, Nujol). This insulating liquid has a specific viscosity and the paste has a certain consistency. The resulting devices are easy to prepare and inexpensive and can be coupled to simple instruments.However, these pastes have limited mechanical and physical stability, especially in flow systems. Additionally, the pastes are dissolved by some non-polar electrolytic solvents, leading to a deterioration of the signal. The general degradation of these devices occurs quickly and has limited their use to the research laboratory. Reviews on chemically' 1 and biologically12 mod- ified carbon paste electrodes have appeared recently. On the other hand, amperometric sensors and biosensors based on rigid composites do not show the problems mentioned above. Further, the fabrication of these devices can be adapted for mass production at a low cost. Rigid Carbon-Polymer Biocomposites Rigid Carbon-Polymer Matrices Creasy and co-workers reported the copolymerization of styrene with divinylbenzene (a cross-linking agent) and vinyl- ferrocene (a modifier), using carbon black' 3 (semigraphitic carbon particles) or carbon fibre9314 as a conductor.This was the basis for the construction of chemically bulk-modified elec- trodes. These devices showed better physical properties (the sensing surface was renewable by polishing) and enhanced chemical traits (they were stable in organic solvents) compared with carbon paste electrodes. Wang and co-workers used a commercially available graphite epoxy resin (Grade RX, Dylon, Cleveland, OH, USA) to build chemically and biologicallyI6 bulk-modified electrodes. The use of this commercial compos- ite rendered the Fabrication of the sensors easier, quicker and more reproducible than the procedure proposed earlier by Creasy and Shaw.9 The approach followed by Wang and co- workers for the preparation of rigid biocomposites was the first report concerning this procedure and these materials.It has been adapted in our laboratories using a non-conducting epoxy (Epo- Tek H77, Epoxy Technology, Billerica, MA, USA), graphite powder (Merck, Darmstadt, Germany) (particle size below 50 pm), biological materials and additives. All these elements are mixed to build a particular biocomposite. l7,lx Our procedure has been expanded to include other polymer matrices such as silicone, polymethacrylate, polyester19 and polyurethane. All these polymers can be prepared it? situ, they readily admit the biological material and additives (catalysts, mediators, co- factors, etc.), they have a simple curing process and are commercially readily available. Graphite-Teflon electrodes were developed originally for vol tammetric and amperometric applications .2",2l These mate- rials with bulk-immobilized enzymes have served for the development of biosensors.22-z4 In this particular instance, the graphite and the powdered Teflon are mixed with the other ingredients and the mixture is pressed to form pellets.Biological Materials and Other Modifiers Immobilized in Rigid Carbon-Polymer Matrices The immobilization of some lyophilized en7ymes (oxidases, peroxidases, dehydrogenases and cholinesterases) in rigid carbon-polymer matrices has been reported (see Tables 1 4 ) .In some instances, the enzyme is covalently bonded to graphitG2 or silica25 particles before blending it to the polymer matrix. Different redox mediators and catalysts have been added to the biocomposites in order to enhance their selectivity and sensitivity. These modifiers may be substances related to ferrocene,1"22.26 tetrathiafulvalene27 and tetracyan- oquinodimethane25 or a metallic catalyst such as gold and palladium28-30 or platinum.31 Dehydrogenase biocomposites have been produced featuring the nicotinamide adenine dinu- cleotide (NAD+) cofactor.32,33 This has opened up the possibil- ity of reagentless biosensors for alcohols and lactate. An alcohol biosensor has been developed by confining dry yeast to a graphite-epoxy matrix. '6 Eisenia hicyclis, an alga, has also been immobilized in a graphite-epoxy matrix, forming a composite used in bioaccumulation assays and in the voltammetric measurement of metal ions.1 6 In our laboratories, biocomposites based on immunospecies immobilized in graphite-polymer matrices are being tried. These biocomposites have a surface that may be regenerated by polishing after each immunological assay.34 Preparation of the Biocomposites and Biosensor Construction The biocomposites are prepared very easily. The powdered graphite is dispersed homogeneously by hand with the appro- priate amount of polymer. According to Tallman and Petersen,x these materials can be classified as dispersed composites since the conductor particles have an equal opportunity to occupy any point throughout the matrix.The polymer material is activated when its components are blended. The activation happens when a volatile fraction evaporates or when a hardener, catalyst or initiator acts on the resin. The resin may be epoxy, silicone, methacrylate, polyester or polyurethane (see Table 1). The contents of the graphite, the modifier (enzyme, catalyst, mediator) and the additives are optimized for a particular polymer matrix. Graphite particles are smaller than 50 pm.19 The goal is to achieve the maximum electrical conductivity and the highest response quality with an appropriate biocomposite rigidity. Graphite content may vary from 20% (epoxy) to 60% m/m (silicone).I9 As mentioned earlier, there is a commercial epoxy that already contains the graphite. 16 The fraction of biological material may vary from 1 % (acetylcholinesterase)35 to 25% m/m (tyrosinase).l 6 The homogeneous mixture is introduced 2-3 mm into a tube made of PVC, glass, etc. A metal disk coupled to a wire is usedAnalyst, December 1996, Vol. 121 1753 to contact the composite inside the tube. The ensemble is left at room temperature or slightly higher (40 "C) for one or more days as needed by the curing of the polymer. When it is hardened, the biocomposite is polished with abrasive papers of decreasing grain size. If the matrix is Teflon,22~~~ the granular polymer is mixed with graphite in mass proportions of 7 + 3. The biological material is previously immobilized on particles of graphite powder22 or is homogenized with Teflon and graphite particles at -20 OC.23 Once mixed, the material is pressed at 7000 Table 1 Glucose biosensors based on rigid conducting biocomposites Biocomposite components (% m/m) c Linear response range (mmol I-') 0.1-5 0.1-5 0.2-5 0.4-20 0.1-5 0.05-5 2.5-30 0.2-1'- 0.0 1-2 1-10 gl-'-'- 1-6-1 0.1-2 0.1-5 '"ippl vel'sus Ag/AgCI/V +1.15 +].I5 +1.1 +1.15 +1.1 +1.15 +0.9 +0.8 +0.9 +0.5 +0.3 +O.15 +0.2 Mediator/ catalyst Enzyme* GOD (2) GOD (2) GOD (2) GOD (2) GOD (2) GOD (2) GOD (20) covalently bound to graphite GOD ( 1 5 ) GOD (20) GOD (1.5) GOD (2) Carbon Polymer Graphite (19) Epoxy (79) (Epo-Tek H77) Graphite (49) Epoxy (49) (Epo-Tek H302) Graphite (49) Methacrylate (49) (Sealer-Healer 1540) Graphite (62) Silicone (36) (Sellaceys) Graphite (36) Polyester (62) (Resipol 9 144) Graphite (60) Polyurethane (38) P" 7.0 7.0 7.0 7.0 7.0 7.0 7.4 7.4 7.0 7.4 6.5 7 .0 7.0 Ref.19 19 19 19 19 This 22 work 28 29 16 26 27 This work Graphite (10) Teflon (70) (7A Dupont) Graphite (15.8) Epoxy (63.0) Graphite-epoxy (Dylon) (54) (Epo-Tek H77) Gold (1 1.8) and 1 ,l'-Dimethyl TTFl (19.7) palladium (7.9) ferrocene (26) Graphite (15.8) Epoxy (63.0) (Epo-Tek H77) Silicone (28) (Sellaceys) TTF, TCNQI: (70) ' Glucose oxidase (GOD) (100-200 U mg-I). + Flow-injection. 1 TTF = tetrahiafulvalene; TCNQ = tetracyanoquinodimethane. Table 2 Rigid conducting biocomposite-based biosensors for phenol and phenolic substrates Linear response range (pmol 1-I) Ref. 16 Biocomposite components (5% m/m) Eappl Ag/AgCl/V -0.2 vei'sus -0.2 -0.1 -0.1 -0.1 -0.05 PI3 (working solution) 7.4 (methanol 50% v/v) Carbon/ polymer Catalyst Substrate Graphite-epoxy Catechol' (Dylon) (92.5) Enzyme Tyrosinase (7.5) Dopamine Phenolics Phenol 50-350 0.Y 1" 1 0.04*,' 1 .OW.' 42 43 30 6.7 Mushroom tyrosinase (5) (6300 U mg- I ) Mushroom tyrosinase (3 (12600 U) Tyrosinase (1) (2400 U mg Phenolics Catechol Phenol Catechol Phenol 6.0 Graphite-epox y ( D y W (99) 6.0 (acetonitrile 5-20% v/v) (methanol 5-20% v/v) 6.0 Tyrosinase (1) Tyrosinase (1.8) (2400 U nig-1) (3900 U mg-I) Grapi te-epoxy Gold (8) Catechol (Dylon) (79) Palladium ( 12) Phenol Graphite (1 8) Catechol Graphite-Teflon (10-30%) (80.2) 30 0.2-25' 23 7.0 (methanol 10% v/v) * Detection limit.1 Flow iiijection.I754 Analyst, Decwnher 1996, Vol. 121 kg cm-2, producing 2 mm thick disks. These pellets are coupled to a tube to form an electrode.According to Tallman and Peterseqx these materials can be classified as consolidated composites, since the conductor particles extend throughout the matrix in a random, reticulated fashion with regions of pure insulator and pure conductor. num electrodes.36337 It is known that carbon electrodes that have metal particles (Pt, Ru, Rh, Pd, etc-.) on their surface show great catalytic action.3X.39 The same happens when the metal is dispersed in carbon pastes.3() The addition of catalysts (gold, palladium) to a GOD graphite-epoxy biocomposite for the oxidation of hydrogen peroxide increases the stability of the signal and reduces the response time. Further, the oxidation potential of hydrogen peroxide is lowered by 250 mV.l* This decrease is also found in experiments with carbon rods where Au-Pd was sputtered to the surface of the electrode."' Therefore, metal bulk-modified composites represent more viable alternatives than those surface-modified electrodes produced by sophisticated technologies.However, the inclusion of metal catalysts in the biocomposite does not hinder the action of the usual interferents found in biological samples (ascorbic acid, uric acid, etc.1.28 On the other hand, it has been observed in our laboratory that this material retains the enzymic activity in dry storage for more that 1 year. The lower working potential and the higher quality of the signal observed in biocomposi tes containing Au-Pd has permitted the use of these materials in flow injection systems. Biosensors with these materials have been used to monitor glucose in fermentation processes.29 Artificial electron acceptors may be added to the biocompo- site.These substances act as electron mediators between GOD Amperometric Biosensors Based on Rigid Carbon-Polymer Biocomposites Glucose Biosensors Several glucose biosensors based on biocomposites have been reported (see Table 1). Glucose oxidase (GOD) has been used in our laboratory as an enzyme model to study the biocatalytic characteristics of rigid conducting biocomposites that feature immobilized enzymes. This oxidase is compatible with matrices of graphite and several polymeric materials such as epoxy resins, polymethacrylate, silicone, polyester, polyurethane and Teflon. These biocomposites have been applied to glucose measurement based on the direct oxidation of the hydrogen peroxide produced by the action of the enzyme [see Fig.1 (A)]. This happens at extreme potentials (0.9-1.15 V versus Ag/ AgCI) (see Table I). When a graphite-polymer composite is used, a shift towards more positive potentials is observed compared with measurements realized with graphite or plati- Table 3 Rigid conducting biocomposite-based biosensors for hydrogen peroxide and organic peroxide substrates PH mediator (working solution) 7.4 hexacyano- ferrate(1r) o-pheny lene diamine 7.4 7.4 Linear re spon \e range (nimol I-') Ref. 16 Biocomposite components (5% m/m) E.'PPl Carbon/ polymer Mediator/ l'f2YSI4.5 catalyst Substrate Ag/AgCl/V H202 -0.2 Enzyme Horseradish peroxidase (25) Graphite-epox y ( D y W (75) Horseradish peroxidase (1 5 ) (94 U mg-I) Horseradish peroxidase covalently bound to graphite (16) Organic -0.2 peroxides 47 Graphite ( 10) Teflon (70) Ferrocene (4) H202 0.0 Butan-2-one peroxide 2.5 pmol 1 I ' 22 20-200 pmol 1 * I 3.0 pmol I-'* H202 0.0 Butan-2-one peroxide Butan-2-one Ferrocene H202 -0.1 7.4 22 (acetonitrile 90% v/v) (reversed 1-100 pmol I-] micellar media) 7.4 1-60 pmol 1 ~ 24 7.4 ?-0.02 48 Horseradish peroxidase Graphite Teflon Horseradish peroxidase (1 5 ) (90 U mg- I ) mixed with human serum albumin ( 5 ) Hz02 -0.25 B utan-2-one peroxide '!-0.05 Cumene peroxide peroxy- benzoate hydro- peroxide terr-Butyl tel-r-Buty 1 H Z 0 2 -0.35 ' L O .I '?-(I. I ?- 1 0.005-0.5 19 Horseradish peroxidase (2) (318 U mg-I) peroxidase (2) peroxidase ( I .9) Horseradish Horseradish Graphite (1 9) EPOXY (79) Graphite (20) Graphite ( 19.6) Epoxy (76.6) Epoxy (78) H202 -0.3 7.0 0.03-7 31 Platinum (1.9) H202 -0.05 7.0 0.09-9 31 * Detection limit.-1 Flow injection.Analyst, December 1996, Vol. 121 1755 and the electrode [see Fig. l(B)] and include 1 ,l'-dimethylfewo- ceneI6.'6 and tetrathiafulvalene.'7 The addition of these media- tors permits the use of working potentials in the range 0.5-0.15 V. The action of interferents is greatly reduced at these working potentials. In the biocomposite modified with tetrathiafulva- lene, ascorbic acid interference is reduced by 90% and the detection of uric acid is negligible.27 Phenol Biosensors Biocomposites featuring tyrosinase have been used in biosen- sors for (see Table 2). In this enzyme system, the species produced electrochemically (catechol) is also the enzyme substrate (see Fig.2). This amplifies the electrochemical response.44 That is the reason for the low detection limits found in these biosensors (see Table 2). However, Onnerfjord et al.,43 using tyrosinase-based rigid biocomposites, found detection limits higher by one to two orders of magnitude than those produced by thyrosinase biosensors based on carbon pastes. If gold and palladium particles are introduced into the biocomposite, an increase in current is achieved. 30 On the other hand, the products of the enzyme reaction (quinones) are highly unstable in water. Furthermore, they polymerize quickly into polyphenols that block the enzyme, and may passivate the electrode.Wang et al.42 reported a 4% decrease in the response of a tyrosinase biosensor after 10 successive discontinuous measurements of 1 X 10-5 mol 1-1 phenol samples. This decrease was explained as being due to slow fouling of the measuring surface by the products of the reaction. This deleterious effect may be minimized by working in flow ~ y s t e m s ~ ~ 3 ~ 3 or renewing the surface of the biocomposite by polishing. Tyrosinase keeps its biocatalytic action when confined to graphite+poxy matrices for moderate periods of times (3% decrease in 10 d with the device in dry storage at 4 "C). However, the biocatalytic activity could be regained after polishing (7040% of the original activity for catechol).4' Wang et al.42 proposes that this stabilizing effect may be due in part to the protective action of the epoxy matrix, not unlike the reported effect in non-aqueous media.4"46 This tyrosinase-graphite-epoxy biocomposite has been used to measure phenols in a partially aqueous medium containing 50%16 or 5-20%30 v/v methanol and 5-20% v/v acetonitrile.") Amperometric biosensors incorporating tyrosinase-graphite- Table 4 Biosensors for bilirubin, alcohols, lactate and pesticides, based on rigid conducting biocomposites Biocomposite components (% m/m) Enzyme Bilirubin Horseradish Yeast (20) oxidase ( 5 ) peroxidase ( 5 ) (Sac i hul-omyes cel-r\~i.siuP) Yeast alcohol dehydrogenase (7.5) (350 U mg-I) Lactate dehydrogenase (6) (148.7 U mg-I) aterase ( 12) ( 1 120 U mg-1) Acetylcholine Acetylcholine esterdse covalently bound to silica (2) Butyrylcholine esterase covalcntly bound to silica (2) Carbon/ Cofac lor/ polymer mediator Graphite-epox y (Dylon) (90) Graphite-epoxy NAD+ ( 10) (Dylon) (82.5) Graphite-epoxy NAD+ (12) (Dylon) (82) Graphite ( 1 7.6) Epoxy (70.4) Substrate Bilirubin Ethanol Alcohols Ethanol Ally1 alcohol Propan- 1-01 Butan- 1-01 Propan-2-01 Lactate Acety lthiocholine P" E"Wl mediator Linear versus (working response Ag/AgCl/V solution) range Ref.-0.2 7.4 4-1 00 pmol I- I 49 hexacyanoferrate( 1 1 ) +0.6 7.4 16 hexac yanoferrate( 11 I ) NAD+ +0.7 7.4 32 04.4 rnmol 1- I (-5.8 mmol I-' 0-8.2 mmol 1-1 0-1 1.7 mrnol 1- 13-32 mmol 1- I 0.5-20 mmol 1-11 +0.7 7.4 0.08 mmol 1-1 33 +0.7 7.0 Graphite (18) TCNQ (9)t Acetylthiocholine +0.3 7.5 Epoxy (7 1 ) Graphite (18) TCNQ (9):i Epoxy (7 1 Butyrylthiocholine +0.3 7.0 5-120 pmol I-' 35 20 pg I-]* carbofuran 27 pg I-' paroxon carbaryl dichlorvos 20 pg 1-1 22 pg I-' 2.5-1 00 pmol 1- I 25 2.2 pg 1 - 1 ' carbofuran 27.5 pg 1-1 paroxon carbaryl 2.0 pg I-' 5-100 pmol I- I 25 22.1 pg 1 - 1 - 2.8 pg 1-1 3.6 ug I-' carbofuran paroxon chlorfenvinphos * Detection limit.+ Flow injection. * TCNQ = 7,7,8,8-tetracyanoquinodimethane.1756 Analyst, December 1996, Vol. 121 Teflon biocomposite have been developed for the detection of catechol.23 Studies of the operational stability of the bio- composite response in organic media (10% v/v methanol or acetonitrile) were carried out in the flow injection mode. Better stability was achieved in methanol.23 Tyrosinase shows poor selectivity with respect to phenol substrates.The reported selectivity sequences for different tyrosinase biocomposites show inconsistencies among them- selves42.43 and with conventional tyrosinase biosensors. The hydrophobic nature of the graphite-epoxy resin modifies the selectivity sequence. Peroxide Biosensors Horseradish peroxidase (HRP) has been immobilized in rigid carbon-polymer matrices. This is the basis for the development of sensors for hydrogen peroxide and small organic peroxides. The first reports in literature follow the approach shown in Fig. 3(A). Reducing agents, such as hexacyanoferrate(I1) ion16 or o-phenylenediamine,47 are added to the solution to regenerate the enzyme to its reduced form. In this way, the oxidized form of these mediators can be detected at lower voltages (-0.2 V versus Ag/AgCl) than those used for the direct detection of hydrogen peroxide (see Glucose Biosensors section). Perox- idase has been immobilized with the mediator ferrocene in graphite-Teflon mat rice^.^^,^^ This opens up the possibility of developing reagentless sensors, capable of working at potentials around 0.0 V versus Ag/AgCl.These devices simplify the measurement process as they function as direct sensors that do a-D-glucose It p-D-glucose ~~ ox , GOD 2H+ 2H' 6-gluconolactone \red 1 n H20 D-gluconate + H+ ox Fig. 1 Reaction sequences for the amperometric detection of glucose using biocomposite electrodes: (A) GOD-graphite-polymer biocomposite; and (B) GOD-mediator-graphite-polymer biocomposite. I Biocomposite 2H' + 2Hi I 2H+ Fig.2 phenolic substrates using a tyrosinase-graphite-polymer biocomposite. Reaction sequence for the amperometric detection of phenol and Fig. 3 Reaction sequences for the amperometric detection of hydrogen peroxide and organic peroxide substrates using biocomposite electrodes: (A) mediated HRP-graphite-polymer biocomposite (mediator in solution or in biocomposite); and (B) mediatorless HRP-graphite-polymer bio- composite. not require additional reagents. The rapid response shown by these biosensors makes them ideal for flow applications.zz The biocomposite HRP-ferrocene-graphite-Teflon is stable in a medium of acetonitrile water (9 + 1 v/v). Biosensors based on this composite have been applied to the determination of hydrophobic organic peroxides in this mediumz2 and in reversed micellar media.24 Biocomposites based on HRP-graphite-epoxy have been used to prepare mediatorless biosensors where direct electron transfer takes place between the active sites of the enzyme and the graphite particles when the substrate is present 19,31,48 [see Fig.3(B)]. We have observed31 that the addition of platinum particles in these biocomposites permits one to work with a lower potential than the optimum working potential of unmodi- fied HRP-graphite-epoxy biocomposite electrodes (see Table 3). The trend of sensitivity for mediatorless biocomposites4~ is in accordance with data obtained for o-phenylenediamine- mediated HRP-carbon paste electrodes47 : hydrogen peroxide > butan-2-one peroxide > tert-butylperoxy benzoate > cumene peroxide > tert-butyl hydroperoxide.The response and surfxe-to-surface reproducibility have been improved by mixing HRP with human serum albumin (HSA).48 This can be attributed to the stabilizing effect of HSA on HRP during the curing process of the graphite-epoxy resin (reaction of epoxy with amino groups), similarly to glutar- aldehyde inactivation of pure enzymes due to a cross-linking reaction with amino groups.48 Bilirubin Biosensor The co-immobilization of HRP and bilirubin oxidase in a graphite-epoxy composite and the addition in solution of ferrocene as a mediator complete the construction of a bilirubin biosensor,49 as seen in Fig. 4 and Table 4. The rapid passivation induced by the adsorption of bilirubin or biliverdin calls for frequent polishing of the surface of the biosensor.The renewable surfaces associated with graphite-polymer biocom- posites lend themselves well for this task. Alcohol Biosensors The co-immobilimtion of alcohol dehydrogenase (ADH) and NAD+ in a graphite-epoxy matrix has allowed the development of reagentless alcohol biosensors32 (see Table 4), following the scheme shown in Fig. 5(A). This type of biosensor shows a rapid decrease of the signal on continuous use owing to a fouling of the biocomposite, incomplete recycling of NAD+- NADH system or a loss of this cofactor. If the surface is polished, the initial activity is restored reproducibly. ADH from yeast, used in these biocomposites, oxidizes primary alcohols quickly (with the exception of methanol). Secondary alcohols are also oxidized but more slowly. The sensitivity sequence of these biosensors (i.e., ethanol > ally1 alcohol > butan-1-01 > propan- 1-01 > propan-2-01) is slightly different to the sequence observed for the same enzyme in solution.32 I biliverdin 0, 2H+ 2H' Fig.4 Reaction sequence for the amperometric detection of bilirubin using a biocomposite electrode: mediated bilirubin oxidase (BOX)- horseradish peroxidase (HRP)-graphite-polymer biocomposite (mediator in solution).Analyst, December 1996, Vol. 121 1757 In the first paper in which Wang and Varughese16 singled out polishable and robust biological electrode surfaces, they reported a graphite-epoxy biocomposite containing dry yeast. These materials had enzyme activity blended into the rigid conductive composite.The resulting alcohol biosensor worked in a buffered medium containing the cofactor and hex- acyanoferrate(II1) as a mediator [see Fig. 5(B)]. Lactate Biosensor Following the strategy mentioned earlier [see Fig. 5(A)], lactate dehydrogenase (LDH) and the cofactor NAD+ have been immobilized in rigid matrices consisting of graphite and epoxy33 (see Table 4). The resulting reagentless biosensors may experience a rapid decrease in sensitivity, as found in ADH- NAD-graphite-epoxy biocomposites. Owing to the high work- ing potential (0.70 V versus Ag/AgCl) needed for the regeneration of NAD+, these biosensors show significant interferences from several species such as acetaminophen, ascorbic acid and uric acid. These biosensors have a fast response and are suitable for continuous-flow measurements. In flow injection procedures, where the sample is briefly in contact with the sensor, passivation effects are much less noticeable than in discontinuous measurements.33 Pesticide Biosensors Organophosphorus and carbamate pesticides have been deter- mined with biosensors based on biocomposites containing acetyl~holinesterase.2~~~~ This enzyme hydrolyses both its natural substrate and thiocholine esters.The hydrolysis of acetylthiocholine produces thiocholine. This electroactive spe- cies is detectable at a potential of 0.7 V versus Ag/AgCl applied to the biocomposite. Fig. 6(A) shows the biosensor response to the substrate. This response is inhibited if organophosphorus and carbamate pesticides are present. This enzyme inhibition is irreversible, calling for the renewal or the reactivation of the enzyme content in the electrode surface either by replacing more enzyme or by regenerating it with special reagents.Biosensors based on rigid biocomposites are an attractive proposition here, since this enzyme ‘reloading’ is achieved by simple polishing of the biosensor surfxe. If the mediator TCNQ Biocomposite Fig. 5 Reaction sequences for the amperometric detection of substrates (S) as alcohols or lactate using biocomposite electrodes: (A) dehy- drogenase-NAD-graphite-polymer biocomposite; and (R) mediated dehy- drogenase-graphite-polymer biocomposite (cofactor and mediator in solution). @ H,O + thiocholine Fig. 6 Reaction sequences for the amperometric detection of thiocholine esters using biocomposite electrodes: (A) cholinesterase-graphite-polymer biocomposite; and (JS) cholinesterdse-mediator-graphite-polymer bio- composite.The enzymic hydrolysis of thiocholine esters is inhibited by the presence of some pesticides. (7,7,8,8-tetracyanoquinodimethane)25 is added to the bio- composite [see Fig. 6(B)], thiocholine can be detected at a potential of 0.3 V versus Ag/AgCl, curtailing the effect of interferents. Biocomposites made of butyrylcholinesterase- TCNQ-graphite-epoxy have been prepared for the measure- ment of butyrylthiocholine at this same potential25 (see Table 4). If the origin of the cholinesterase (electric eel, horse serum and bovine erythrocytes) is altered, serious inconsistencies are noted in the biocomposites prepared due to the leakage of the enzyme.A satisfactorily reproducible response is attained only with acetylcholinesterase from bovine erythrocytes. If the enzyme is immobilized on silica particles for stability, good reproducibility is attained regardless of the origin of the enzyme. This step does not alter the curing process of the biocomposite.25 Conclusions Biosensors based on rigid polymer-graphite composites are a recent development and examples of their design and applica- tion are still scarce in the literature (see Tables 1 4 ) . However, several advantageous qualities of biocomposites based on rigid graphite-polymer mixtures can be envisaged from the present review. The preparation procedure for these biocomposites is simple and involves dry chemistry techniques for the most part.In some cases, the enzyme has first to be immobilized on some sort of support particles. Before curing, these biocomposites are highly mouldable. This permits the easy construction of amperometric sensors of various shapes (cylindrical, planar, tubular, flow-through, etc.), and sizes. After curing, these materials are very stable from a mechanical point of view. The surface is stable, rigid and polishable and can be drilled or otherwise altered mechani- cally. The components of the sensing surface can be controlled by defining their content in the bulk. The presence of enzymes, cofactors, mediators, additives, etc., on the sensing surface can be tailored by adjusting their content in the bulk of the biocomposite. Biosensors prepared with the techniques described here have great biological stability.The biocomposite acts as an im- pervious reservoir for the biologically active components. The decrease in sensitivity on the surface is recovered by a simple polishing procedure. Each new surface yields reproducible results if all the individual components of the biocomposite are dispersed homogeneously in the bulk. Epoxy resins and Teflon are employed as polymer matrices because they are well known materials. They provide chemical stability and the resulting biosensors can be used in partially aqueous media (methanol-water, acetonitrile-water, etc.). The morphology, size and distribution of the conducting particles define the behaviour of the biosensor as a microelec- trode array. These microelectrode arrays or ensembles show efficient mass transport and a better electrochemical response (high signal-to-noise ratio, low detection limits, fast response times). The resulting biosensors are suitable for flow systems because of these electrochemical, chemical, mechanical and biological features.Finally, the preparation of the biocomposites and the construction of the biosensors are inexpensive. Final Remarks Surface characterization is a key point in understanding the function of modified electrodes. This knowledge is useful in the design of surface microstructures suitable for the construction1758 Atialyst, December 1996, Vol. 121 of more selective sensors. Surfaces of the type reviewed here have not been studied thoroughly. Scanning tunnelling micros- copy (STM) has been useful in the study of graphite distribution in the surface of an ADH-NAD-graphite-qoxy biocompo- site,32 but this technique cannot produce useful information about the other non-conductive components of the material.The biosensors reviewed here (see Tables 1-4) have been constructed manually in cylindrical shapes. Thick-film tech- nologys" may be the fastest, most reproducible and economical way of mass producing biosensors. Screen printing and ink-jet printing techniques have shown great potential in this respect. These procedures have been used for the sequential deposition of the layers on the device (conductor. receptor, mediator, permselector, insulator, etc.). Using the biocomposites de- scribed here, these methods can be transformed into one-step processes.s1-s4 In this fashion, the printing process becomes simpler and more reproducible. Biocomposites can be rendered more fluid and applied as inks for these methods of mass production.The coupling of the biocomposites mentioned here and printing processes for the production of biosensors have great promise. This work is in progress in our laboratories. All the biocomposites reviewed have been developed for using in amperometric devices. However, recently a novel potentiometric biosensor based on a peroxidase-graphite- epoxy biocomposite has been rep~rted.~' With this approach a challenging field is envisaged. Financial support from the Commission of the European Communities, Environment and Climate Programme (EVSV- CT94-0407) and the Interministerial Commission for Science and Technology (CICYT), Madrid, is gratefully acknow- ledged.References 1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 Scheller, F., and Schubert, F.. Biosen.soi.s, Elsevier, Amsterdam, 1992. Ruschau, G. R., Newnham, R. E., Runt, J., and Smith, B. E., Sons. Actuatoix, 1989, 20, 269. Harsanyi, G., Polymer Films in Sensor Applications, 'Technomic, Lancaster, 1995, pp. 1 13- 125. Lima, J. L. F. C., and Machado. A. A. S. C., Analyst, 1986, 111, 151. Alegret, S . , and Martinez-Fabregas, E., Biosensors, 1989, 4, 287. Machado, A. A. S. C., Analyst, 1994, 119, 2263. Oldham, K . B., J . Elrc~troanul. Chem., 198 I , 122, 1. Tallman, D. E., and Petersen. S. L., Electroanalysis, 1990, 2, 499. Creasy, K. E., and Shaw, B. R., Anal. Chem., 1989, 61, 1460.Adams, R. N., Anul. Chem., 1958.30, 1576. Kalcher, K., Kauffmann, J.-M., Wang, J., Svancara, I., Vytras, K., Neuhold, C., and Yang, Z., Electroanalysis, 1995, 7, 5. Gorton, L., Electroanalysis, 1995. 7, 23. Shaw, B. R.. and Creasy, K. E, Anal. Chem., 1988, 60, 1241. Wang, C., Creasy, K. E., and Shaw, B. R., J . Electroanal. Chenz., 1991,300. 365. Wang, J., Golden, T.. Varughese, K . , and El-Reyes, I.. Anal. Chem., 1989, 61, 508. Wang, J., and Varughese, K.. Anal. Chem., 1990, 62, 3 18. Alegret. S., Alonso. J., Bartroli, J.. Martinez-FBbregas, E.. and Valdes, F., in Uses of Inrmohili~ed Biological Compounds. ed. Guilbault, G. G., and Mascini, M., Kluwer, Dordrecht, 1993, pp. 67- 79. Alegret, S., Alonso, J., Bartroli. J., Ckspedes, F., MartineL- Fabregas, E., and del Valle, M., Sensors Mat., 1996.8, 147. Alegret, S., Cespedes, F., Martinez-Fibregas, E., Martorell, D., Morales, A., Centelles, E., and Mufioz, J., Biosens. Bioelectron., 1996. 11, 35. 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Klatt. L., Connell, R . , Adams. R . , I-lonigbel-g, I., and Price, J.. AnuI. Chem., 1975, 47, 2470. Shah, M., and Honigberg, I., A n d . Lett., 1983, 16, 1149. Wang, J., Reviejo, A. J., and Angnues, L., Elrc~ti-ocrnu1ysi.r. 1993, 5 . 575. Puig. D., Ruzgas, T., EmnCus, J.. Gorton, L., Marko-Varga, G., and Barcelo, D., EIec,tr.oaizcrlysi.s, in the press. Ortiz, G., Manuel del Valle. F. J., Reviejo, A. J., and Pingamon, J. M., submitted for publication. Martorell, D., Ckspedes, F., Martinez-Fiibregas.E.. and Alegret, S., Anal. Chinr. A(.tu, in the press. Wang, J . , in Floni Injection Analysis (FIA) Bused on Enzymes utrd Antihodies, ed. Schmidt, R. D., VCH, Weinheim, 1991, pp. 277- 286. CCspedes, F.. Martinez-Fiibregas, E., and Alegret, S., Electrou- nulysis, 1994, 6, 759. CCspedes, F., Martinez-Fhbregas, E., and Alegret, S.. And. Chinz. Acta, 1993, 284, 21. CCspedes, F., Valero, F., Martinei.-Fabregas. E., Bartroli, J., and Alegret, S., Anulyst, 1995, 120, 2255. Lutz, E. S. M., and Dominguez. E., E/cctroant~l~~si.s, 1996, 8, I 17. Morales, A,, CPspedes, F., Muiioz. J . , MartineL-Fibregas, E.. and Alegret. S., Anal. Chinz. Acta, in the press. Wang, J., Gonzilez-Romero, E., and Ozsoz, M., Elcr,troanulysi.s, 1992, 4, 539. Wang, J., and Chen, Q., Elec.ti.oc~rralysis, 1994, 6, 850. Santandreu, M., Cispedes. F., Alegret, S., and Martinez- Fibregas. E., in Pt.oceedi??gs of IVth Europeun Workshop on Biosensors jbi. Eni~ironnientul Monitoritig, Environment and Climate Programme of the Commission of European Communities. Barcelona, 1996, pp. 33- 37. Martorell, D., CCspedes, F., Martinez-Fabregas, E., and Alegret, S., Aiiul. Chim. Actu, 1994, 290, 343. Kauffinan, J. M., Linders, C. R., Patriarche. G. J., and Smyth, M. R., Talantu, 1988, 35, 179. Cespedes, F., Martinez-Fibregas. E., Bartroli, J., and Alegret, S., AnuI. Chin?. Actu, 1993, 273, 409. Gorton, L., Anul. Chin?. Acta, 1985, 178, 247. Cox, J . A,, Jaworski. R. K., and Kulesza. P. J.. Electr~oatzulysis, 1991, 3, 869. Wang, J., Naser, N., Angnes, L., Wu, H., and Chen, L., Anul. C'hrni., 1992,64, 1285. Yang, X., Johansson. G., and Gorton, L., Mikrochim. Acta, 1989, I . 9. Wang, J., Fang, L., and Lhpez, D., Analyst, 1994, 119, 455. Onnerfjord, P., Emneus, J., Marko-Varga. G., and Gorton. L.. Biosens. Bioelectron., 1995, 10, 607. Marko-Varga, G., EmnCus, J., Gorton. L.. and Ruzgas. T., Trends Anal. Chem., 1995, 14, 319. Zaks, A., and Klibanov, A. M., Science, 1984, 224, 1249. Wang, J., Lin, Y., and Chen, Q.. Electroanalysis, 1993, 5, 23. Wang, J., Freiha, B., Nasger, N., Gonzalez-Romero, E., Wollenber- ger, U., Ozsoz, M., and Evans, O., Anal. Chim. Actu, 1991, 254, 81. Wollenberger, U., Wang, J., Ozsoz, M., Gonzilez- Romero, E., and Scheller, F., Bioelectrochem. Bioeneq., 1991, 26, 287. Wang, J., and Ozsoz, M., Elec.troatiulysis, 1990, 2, 647. Galan-Vidal, C., Mufioz, J., Dominguez, C., and Alegret, S., Trends Anal. Chem., 1995, 14, 225. Wang, J., and Chen, Q., A n d . Lett., 1995, 28. I ! 3 I . Nagata, R., Yokoyania, K., Durliat, H., Comtat, M., Clark, S. A., and Karube, I., ~lci.ti.ounul~sis, 1995: 7, 1027. Nagata, R., Yokoyama, K., Clark, S . A,. and Karube. I., Biosens. Bioelrr.tron.. 1995, 10, 261. Nagata, K., Clark. S. A., Yokoyama, K., Tamiya, E.. and Karube, I., Anul. Chirn. Actu, 1995, 304, 157. Zulfikar, Hibbert, D. B., and Alexander. P. W., Elec.ti-ounulysis, 1995, 7, 722. Paper- 6103420K Receiid M q 17, 1996 Accepted Jiirze 13, I996