Amperometric enzymatic glucose electrode based on an epoxy-graphite composite

AnaIytzca Chtmtca Acta, 273 (1993) 409-417 Elsevler Saence Pubbshers B V , Amsterdam

Amperometric enzymatic glucose electrode based on an epoxy-graphite composite F CGspedes, E Martinez-Fiibregas, Lkpartament de Quhca,

J Bartroli and S Alegret

Unrversuat Autdnoma de Barcelona, E-008193 Be&terra &am)

(Recensed 1st June 1992, revised manuapt

received 9th October 1992)

Abstract An mexpenswe, robust, pobshable and easy to mechamze amperometrrc transducer showmg a long bfetnne, based on a composite matenal made of graphite and non-conductmg epoxy resm (Epo-Tek H77), was constructed Thus composite was charactensed electrochemlcally usmg cychc voltammetry and Imear-sweep voltammetly The appbcabdlty of this amperometnc transducer was demonstrated m the cons&u&on of a glucose bmsensor based on the amperometrlc detection of hydrogen perozode produced by the catalytic action of glucose oxldase covalently nnmobdlzed on a nylon mesh The sensor shows a lmear response range for glucose m the range 1O-5-1O-2 M when a potential of 1150 mV 1s appbed with respect to an Ag/AgCl electrode m a pH 7 00 buffered solution wrth 0 1 M phosphate and 0 1 M KC1 The resulting blosensor was compared unth a commercial glucose analyser Keywor& Amperometry, Blosensors, Enzymatic methods, Voltammetry, Enzyme electrodes, Epoxy-graphite wmposlte electrode, Glucose

Analyt~al methodologies for process momtormg and control are constantly evolvmg, seekmg to mcrease snnphclty, rehab&.y, preclslon, rapidity and economy m the acqulsltlon of analmcal data In this context, sensors are seen as a viable alternative to complex analytical mstruments Hence great efforts are being made to develop new transducers that are mcreasmgly selectwe, senaWe, robust, mexpensrve, easy to build and amenable to apphcatlon m the momtormg and control of clmlcal, environmental and mdustnal processes Polymer technology offers a wide choice of mater& smtable for use m a great vanety of mdustrlal products, conferrmg advantageous physical and chenucal propertles on those products and often makmg them cheaper to manufacture Correspondenceto S Alegret, Departament de Quhmca, Umversltat Autanoma de Barcelona, E-008193 Bellaterra (Spam) OW3-2670/93/W

The field of chemical sensors m general, and electrochermcal transducers m particular, has taken advantage of the avallablhty of new matenals, yleldmg a generation of compoate-based electrodes [ll Conventional voltammetrlc electrodes built with a conductor the surface of which has been chemically or physlcally treated (modtiled electrodes) can also be considered to belong to this group [2,3] In this work, a new electrochenucal sensmg matenal 1s reported This material encompasses the dlsperslon of electncally conductwe matenal m a non-conductwe, plastic phase This rigld transducing mate& yields an electrode mth features srmllar to those of carbon paste electrodes that have a mouldable and soft structure [4] Electrodes constructed usmg these new, conductwe composite materials show several electrochemical advantages over those built using a smgle conductwe phase (platinum, gold, sliver, mercury, graphite, etc 1 The matruc of the composite

00 Q 1993 - Elsewer Science Publishers B V All nghts reserved

410

material can accommodate other substances to Improve the electrochemical response of the transducer [5,6] Addltlonally, the slgnal-to-noise ratio 1s higher m composite material electrodes than m smgle-phase electrodes, allowmg for lower detection hnuts [7] From the mechanical pomt of ylew, an electrode based on a composite material can be renewed by pohshmg its surface, thus lengthenmg its hfetlme [S], and it can be mechantzed as a hard sohd Further, composite materials have technologically appealing features, such as their mltlal mouldmg before the polymerlzatlon of the plastic phase This mouldmg permits the construction of sensors of various sues and shapes, such as flow-through configurations, and also permits the construction of electrodes by the deposltlon of conductive tracks on a non-conductive strip, wluch 1s a recurrent design m disposable amperometrlc sensors [9] This hne of research 1s bemg pursued at present Conductive composite materials have been used to construct Ion-selectlve electrodes (ISEs), where the ion-sensitive membrane 1s fixed dlrectly to the composite, replacing the usual ISE structure involving an Internal reference solution The performance of these ISEs 1s encouragmg, espeaally with respect to the hfetnne of the sensor, a consequence of the good adhesion of PVC-based membranes to the composite matenal

[lOI In this work, the apphcatlon of [0,3]-composite materials [11,12] to the constructlon of disposable electrochermcal transducers has been broadened Thus matrix has a conductrve phase formed by graphite powder and the insulating phase 1s an epoxy resin Followmg their electrochenucal characterLzatlon (usmg hydrogen peroxide detection), the apphcab&y of these transducers was demonstrated m the design, construction and evaluation of a conventional first-generation type of amperometric glucose sensor EXPERIMENTAL

Apparatus Voltage-current curves were recorded usmg a Polarecord E506 recorder (Metrohm, Hensau,

F C&edes

et al /Anal

Chm

Acta 273 (1993) 409-417

SHrltzerland) Cychc voltammograms were generated employmg a Dacfamov 05 10 CNRS-Mlcrotee unit controlled by an Apple II Computer (CNRS, Toulouse, France) Sensor current was measured with a PRG-DEL amperometnc unit (Tacussel, Lyon, France) pH was controlled usmg a Dlgllab 517 pH meter (Cnson, Barcelona, Spain) Responses were recorded with a Labograph E586 recorder (Metrohm) Reagents and solutwns To construct the electrodes, graphite powder with a particle size of 1 km (Merck) and Epo-Tek H77 (Epoxy Technology, Bdlenca, MA) and Araldlt M (Cuba-Geigy) epoxy resins were used Lysme, glutaraldehyde and dunethyl sulphate were obtained from Fluka D-( + )-Glucose monohydrate and hydrogen peroxide were obtamed from Merck Glucose oxldase (G-2133 type VII from As~rgzU~ luger) was obtained from Sigma The supporting electrolyte was an aqueous solutlon buffered at pH 7 00 wrth 0 1 M phosphate and 0 1 M KC1 Hydrogen peroxide solutions were fresh prepared dally Glucose solutions were prepared from a benzolc acid saturated solution (3 g drnm3) [13] Auxdmy efectroa!es

To characterize and evaluate the workmg electrodes, a platinum aumllary electrode and a double-Junction Ag/AgCl reference electrode (Onon 92-02-00) were used The external reference solution was 0 1 M KC1 Constructwn of the amperometnc transducer The composite material employed was prepared by dlspersmg graphite powder m Epo-Tek H77 epoxy resin The mlxlng proportions were 1 4 parts by weight Before its mlxrng with the graphte, the resm was prepared by nuxmg Its two components, part A (the resm proper) and part B (HR hardener), m a 20 3 weight ratio Once homogemzed, the rmxture was applied to one of the openmgs of a PVC tube (6 mm o d ) This tube formed the body of the electrode The composite mixture filled the tube to a depth of ca 3 mm (Fig 1) The assembly was kept at 50°C for 48 h to harden the composite After the

F Gpedes

et al /Ad

411

Chm Acta 273 (1993) 409-417

fled by the selective blocatalytlc generation of an electrochemlcally active species (hydrogen peroxide) by an enzyme (GOD) D-(

+

)-glucose + 02 3

glucomc acid + H,O, (1)

The hydrogen peroxide produced 1s measured amperometrlcally by the direct oxldatlon on the surface of the composite-material electrode when a potential E IS applied H,O, 2

Fig 1 Construction of the sensors 1 = Flexible supportmg rmg, 2 =electrode body, 3 = copper contact, 4= wue. 5= conductwe and pohshable graphite-epoxy resm, 6 = 0-rmg, 7 = Parafdm, 8 = enzyme membrane, 9 = d&w membrane Amperometrx. transducer, l-5, glucose bmsensor, l-9

hardening process, the outer surface of the composrte was fmely pohshed usmg 3qm alumma paper (pohshmg strips 301044-001, Onon) molstened with doubly dlstllled water Construction of the bwsensor The glucose blosensor was constructed by applymg a nylon 6,6 membrane to the surface of the electrode Previously, glucose ox1Llase (GOD) had been lmmoblhzed covalently on the nylon surface The nylon membrane used has 108 threads -2 and a thickness of 120 pm (A Bozzone, ~~plano Gentile, Milan, Italy) The mnnoblhzatlon process employed [14,15] consisted of activating the nylon wth dnnethyl sulphate m order to form an nrndo ester group This group was made to react wth one of the ammo groups from the lysme molecule used as a spacer The lmkmg of the enzyme to the extender arm was reahzed usmg glutaraldehyde The nylon membrane containing the enzyme was fiied to the electrode by a dialysis membrane and an O-ring and using Parafihn as a sealant (Fig 1) Charactenstlcs of the sensmg system The resulting glucose brosensor based on composite matenals 1s of the kmd known as a fiistgeneration sensor These electrodes are identl-

0, + 2H++ 2e-

(2)

RESULTS AND DISCUSSION

Constructwn of transducer with epoxy-graphlte composrte mutenal Conducttvrty of the compostte matenal An amperometrlc transducer should have a very low resistance The conductlvlty of an epoxy-graphite composite is determined by its graphite content The Epo-Tek H77 resin admitted a maximum of 20% graphite content (sensors made m this manner will be called ET units), while Araldlte M resin admitted a 60% graphite content (sensors made m this manner will be called AM units) If the graphite content 1s reduced for both ET and AM units, a decrease m conductrvlty and sensltrvlty will result (Fig 2) H202 oxzdatwn potentuzl The exact oxldatlon potential of hydrogen peroxide, somewhere between 1000 and 1200 mV, 1s dficult to determine because of the proxumty of the oxldatlon potential of the solvent (water) In the lmear-sweep voltammogram obtained at a glassy carbon electrode, the typlcal plateau m the oxldatlon curve, correspondmg to the maxlmum current, 1s present (Fig 3) This plateau 1s not observed when the voltammogram 1s recorded usmg both ET and AM units (Fig 4) The system was optmuzed by calibrating at different potentials between the llrmts 1000 and 1200 mV The best compromise between sensltrvity and residual current of the system was found at a potential of 1150 mV Heterogenezty of the composite mater& When domg cychc voltammetry wth a snnple, reversible

412

c

b a

F% 2 ~~-~~~ ~~1~~ wth a pH 7 00 buffer (0 1 M phosphate-0 1 M KC0 Sweep rate, 10 mV s-l G~p~i~ amtentaf the eornposlte mate& (a) 10, (b) 20, (c) 30, (d) 40, (eI50, (0 60%

redox system, such as ~e~lI)-Fe~~~~, the ET and AM tmts behaved d~e~ent~y Thw ~ere~c~ were obsemed by dete~~lng the ~~~ cathodrc currents (t,> and ~~~ anodtc currents (zap) and the ~oteut~~s at which those cur-

04v FIB 4 ~ma~-~ee~ ~it~~~ using(aI AM uart (60% grapbl~-~% Araldlte M epoxy resmj and (b) ET unit (20% ~aph~~-80% Epo-Tek H77 epoxy restn) (1) Kn pW 700 buffer (0 1 M ph~~~a~-O 1 M KCD, (2) m pH 7 00 buffer (0 1 M phosphate-0 1 M KCI) and 5x 10e4 M Ei,G,

rent values were reached (EW and J&J These values should be related by

~&?p = 1

(3)

~0~9~~~~~-~~}~~

(41 where n IS the number of electrons ~a~icl~at~g HI the redox process In Ah4 untts the ratlo between anodlc and cathodic currents and the value of n are close to

v

04v

Fxg 3 Lmear-sweep ~01~~ usmg a ~~r~l~~ avadable glassy carbon electrode (2 mm d~amater~ (a) In pH 7 00 buffer (0 1 M p~h~e-~ 1 M KCII, (W m pH 700 buffer (0 I M ph~phate-~ 1 M KC@and 1 tnM N*O,

FLL 5 Qfchc yolt~o~ usmg (a) an AM umt (60% ~aphite-~ Araldite M epow iesmj and (b) an ET untt (20% ~aphit~-~ E&o-Tek H77 epoxy resm) pnth 5 mM K,FeO, and 1 M KC! Sweep rate, 50 mV S-I between -2@3and +6OUmV

F C&pedes et al /AnaL Chm Acta 273 (1993) 409-417

Fig 6 Surface of composite matenal electrodes (a) Composite contammg 60% graphite-40% Araldlte M epoxy resm, (b) composite contammg 20% graphite-80% Epo-Tek H77 epoxy resm The top mlcrographs show magmficatlons of the white boxed areas Indicated below

1 The peaks m the cychc voltammogram (Fig 5a) are well defined, mdlcatmg that the composite wth a high graphite content (60%) IS homogeneous ET units, on the other hand, yield a value of n = 0 7, and show ill-defined peaks m the cychc voltammogram (Fig 5b) This 1s due to decreased homogeneity m this type of composite material [16] wth a low graphite content (20%) Electrochemical actwrty 1s not uniform m the totality of the electrode surface Composite materials behave as if they were arrays of very small electrodes For this reason, non-linear dlffislon appears, and mass transport by migration IS the dommant phenomenon, generating a steady-state current [171 The photomlcrograph m Fig 6 shows the dlstmct surface morphology of these electrodes Electrode stub&y At first, a better response was expected from electrodes constructed using Araldlte M epoxy rather than from those based on Epo-Tek H77 resin This was expected on the basis of their respectwe electrochemrcal re-

sponses However, after repeated cahbratlon runs, AM electrodes showed a loss of hneanty, whereas ET electrodes mamtamed their hneanty throughout the concentration range of Interest (Fig 7) The non-hneanty of the AM electrodes IS proba-

:*,I:

-450

Log

-350

[Wal

Rg 7 Stabdlty (0, 0) Ah4 and (A, A) ET units o, A = Cahbration graphs for freshly bmlt mnts, 0, A = Cahbratlon graphs for umts after bemg immersed m pH 7 00 buffer (0 1 M phosphate-O 1 M KCI) for 1 week

414

F Gkpedes et al /Anal Chm Acta 273 (1993) 409-417

bly mduced by absorption of water m the AM matnx. This absorption was demonstrated when the lmeanty recovered after the electrodes had been dared at WC for 48 h Given the stabtity shown by ET umts, they were chosen as the electrodes to be used m the glucose bmsensor. E~~tw~ of the ET trmsducer m Hz& Electrolyte concentratm and pH effects A study of the system was made at Merent concentrations of the supportmg electrolyte It was observed that at low concentrations of hydrogen peroxide, the lmeanty d&rushed sightly, the background Norse mcreaaed and current mstabihty occurred All these delete~o~ effects happened when a buffered solution (pH 7 00,O 1 M phosphate) was used as the background When KCl at a concentration of 0.1 M or lngher was added, the system became very stable The farad= process d-bed by Eqn. 1 constatutes a proton source If a low pH IS employed, E&n 2 wdl shift to the left, dimImsh~ng the value of the current and the overall sensltMty of the system If a high pH IS used, the ensumg proton deflclt will provoke a sluft of reaction 2 towards proton production, thus oxldrzmg more hydrogen peroxide and mcreasmg the current considerably (Fig 8) However, using a high pH also facllltates the oxldatlon of the solvent, producmg an overall detnmental effect Experunental data confirmed

12000 00 ,

I

10000 00

6000 00

2

~6ooo00 4Ofwoo

200000

000 0

2

PH

Fa 8 Effect of pH on the signal of an ET transducer at different concentrationsof hydrogenpermde o = 1 X 10m4, CI=IxID-~, A =OM

TABLE 1

Day

a

b

r

ET1

1 1 4 6 6 11 25

782 791 795 786 781 783 776

103 104 102 101 100 100 098

099995 099992 099995 099995 099998 099997 099999

ET2

1 5

802 805

101 100

099996 099990

ET4

1 9

776 779

099 loo

099990 099992

ET6

1 14

790 788

101 102

099980 099990

ET8

1 21

7 77 780

101 102

099995 099992

that the optnnum workmg pH was one nelghbourmg acid-base neutraltty Cafzbratmn parameters If the loganthm of the current and the logarithm of hydrogen peroxtde concentration are treated by the least-squares method, the experimental data fit a linear equatlon expressed by log[ I(nA)] = a + b log[H,O,(M)]

(3

where a IS the value of logZl(nAj] at the orlgm and b 1s the slope of the straight lme In Table 1, a, b and the correlation coefficient r are grven for several electrodes using a workmg solution of 0 1 M p~~phate-0 1 M KC1 (pH 7 00) The average values of these parameters (n = 15) for a 95% confidence level are a = 7 9 f 0 1 and b = 100 f 0 01 The concentration range exammed was 10-6-10-3 M hydrogen peroxide Lmeanfy range The lower hmrt of the lmear response (LLLR) and the upper lnmt (ULLR) were found graphIcally usmg a callbratlon graph The lmear range of the system IS between 10m7 and 10e3 M hydrogen peroxtde If hydrogen peroxide concentrations of the order of lo-’ M are used, the current generated IS very close to the

F Ctspedes et al /Anal

varlatlons of the baselme The upper lnmt of the lmear response 1s determmed by the ability of the transducer to oxldlze the analyte as it comes m contact mth its surface The d&t of the residual current generated by the system 1s 5 nA mm-l Response dynamxs The 95% response tune of the transducer IS 13 s when the hydrogen peroxide concentration increases from 0 to 2 x 10m5 M Durmg ensumg cahbratlons, the signal 1s very stable between consecutnre addltlons Repetatwrty of the SJ~FUZ~ For discrete measurements of hydrogen peroxide m separate solutions, with a 4 x 10m6 M concentration, an average error of 1% was found for a 95% confidence level (n = 8) Evaluation of the glucose bwsensor pH eflect The workmg pH range for glucose oxldase hes between 4 00 and 7 00, with an actlvlty maxnnum at pH 5 5 However, when the enzyme has been chenucally treated for nnmobdrzatlon, the optmmm pH depends on the nnmoblhzmg system used As expected from the results obtamed for the hydrogen peroxide system, the current rises with mcrease m pH Above pH 9, the signal dunmlshes owing to enzyme denaturatlon For the system described, the i&orkmg pH range 1s between 6 and 9 (Fig 9) Cahbratton parameters When the logarithm of the current and the logarithm of glucose concentration are treated by the least-squares method, the data fit a straight lme where a, the value of

3 5

JOW

-

2000 -

1000-

0, 4 00

, 5 00

415

Chm. Acta 273 (1993) 409-417

1

1

600

.

I

700

I

800

1

I

000

tow

PH FIN 9 Effect of pH on the glucose blosensor signal Glucose concentration, 1 mM

TABLE

2

Cahbratton sors

parameters

Bmsensor

for composite-baaed

glucose blosen-

Log[ZhA)] - a + b log[glucose (Ml] Day

a

b

r

ETG4

2 3 6 14

673 689 689 7 13

099 106 102 110

0 0 0 0

ETGS

1 7 10 17 30

659 664 653 6 60 646

104 106 103 106 103

099990 099995 099996 099993 0 99995

ETG7

19 30 33 40

6 6 6 6

106 103 104 102

099990 0 99995 0 99993 0 99992

ETG8

20 30

609 611

103 104

0 99993 099992

17 12 21 10

99995 99990 99995 99992

lo$l(nAll at the ongm, is 6 13 f 0 05 and b, the slope, 1s 104 f 0 01 for a 95% confidence level and n = 15 As can be seen m Table 2, the results show low scatter using different electrodes over a period of more than 1 month The scatter of these results is sumlar to that shown by the internal epoxy-graphite transducer When the sensors were not m use, they were kept m 0 1 M phosphate-O 1 M KC1 buffered solution (pH 7 00) at 4°C Lznearzty range The glucose blosensor response 1s lmear between 10m5 and lo-* M glucose This range covers most requirements for glucose measurements m industrial and clmlcal applications Response dynamics The 95% response time of the glucose blosensor 1s 32 s when the concentration increases from 0 to 1 mM This increase m the response time when compared with that of the hydrogen peroxide transducer can be explamed by the mfluence of two factors enzyme kmetlcs and the barrier effect of the dlalysls membrane In any event, this lengthenmg of the response tune does not constitute a hmdrance to cahbratlon, as the response signal settles rapldly with successwe glucose additions

416

F C&.&es et al /Anal

TABLE 3 Average measurements for three glucose samples measured with the reported bmsensor and the YSI glucose analyser Standard Cg/l>

[Glucose1 (s/l) ETG bmsensor (n==3)

YSI glucose analyser (n=4)

45f02 91rtO3 13 &2

446i~Ol 898kOOl 134 f01

458 909 13 64

of the srgnatThe

~~~t1~1~

dmlys~s mem-

brane stablltzed the signal slgmfficantly and mcreased the reproduclb&y of the response For glucose determmatxons m separate solutions of 10-’ M ~n~ntratlon the mean error was 1% for a 95% cotidence level and n = 10

The Yellow Spnng Inst~ents CYSI) Model 2000 glucose analyser is a ~~-automatic system based on the same prmcxple as the btoseusor reported here Its workmg electrode ISa platmum disc with a membrane contammg nnmobdrzed

Chute Acta 273 fI99314@-41?

GOD and other membranes kelhdose acetate and poiycarbonate) that help to prevent mterferences Table 3 compares the results obtamed when three dtierent sampies (aqueous glucose solutlons) were anaiysed by the composite-based sensor and the YSI glucose analyser When a correlation diagram 1s constructed, the Ime passes through the ongm and has a slope very close to unity Therefore, it can be concluded that the results obtamed by the bmsensor descrkd are very smnlar to those @ven by the commercially avadable analyser ~o~~ca~~

of epoqyyapkte

compos8te

Although good results were obtamed using the composite mater& described above, the potential apphed for the detectjon of hydrogen peroxide 1s high (1150 mV vs Ag/AgC& posmg several problems such as mterferences arrsmg from elect~he~c~Iy actwe species and the ~~lblll~ of a clear d~~~na~on of the opts oxtdatlon potentzal of hydrogen peromde from the oxtdatlon potential of water All these aspects are also present m the glucose measurement

c

@--i 3 4Q mV

-

A

40 mV

/f

Fa 10 Linear-sweep vanes m a pH 7 00 buffered rate, 10 mV s-l (1) In supportmg electrolyte (SE), Q catalyst Sweep rate, 50 mV s-l (1) In SE, (2) m tet~~l~~alene s~lutxon m 0 25% acetone and enzyme 4 4 and 6 4 mM glucose,respecbvely

I

M

UA

solution (0 1 M p~phate-0 1 M KC@ dA) X&mod&d ET umt Sweep m SE -I-5 x 10-4 M H,O, (B) Modrfied ET umt w&h a 10% Au-Pd SE + 4 x 10m4 M H,O, (C) MadBed ET umt with 3 x 5 ~1 of a membrane Sweep rate, 10 mV s-t (1) In SE, (2), (3) and (4) 111SE + 2 4,

F Gspedes et al /Anal

411

Chmz Acta 273 (19931 409-417

The nature of the composite mater& allows for the addltlon of further mater& to Its matm and pennrts the absorption of vaflous substances on its surface owmg to &s ingh porosity Owing to the above, work IS bemg carrted out on the alter&Ion of the composite matrur w&h a catalyst to promote hydrogen peroxide oxtdation The catalyst employed 1s made of a mncture of gold and palladium m 3 2 weight proportrons [18] When the electrode was modtied with the addstlon of the catalyst, the hydrogen peroxide o~dat~on potential was lowered to 900 mV vs Ag/AgCl Addltlon~~, the characterrstlc plateau due to the maxunum oxldatlon mtensity of the species can be dls~~~lshed very clearly (see Fsg 10) The segregation of the oxtdatlon potentials of hydrogen peroxde and the solvent permits greater slgnal stab&y and lower background noise However, the problem posed by mterference has not been completely solved as the applied potential of 900 mV IS std relatively high Tberefore, current work 1sfocused on the development of a second generatIon of polishable and robust electrodes m which the ~m~slte matenai 1s modtied by a redox mediator (te~at~a~~alene) capable of establlshmg electron transfer w&h the active centres of the glucose omdase enzyme In this mstance the applied oxldatlon potential IS 150 mV vs Ag/AgCl (see Fig 10)

REFEREN1 J E Anderson and DE Tallman, Anal Chem ,48 (1976) 209 2 K Aokr and J Gsteryound, J Electroanal Chem, 125 (1981) 315 3 A J Bard, J A Crayston, G P Ktttlesen, TV Shea and MS Wnghton, Anal Chem,58 0986) 2321 4 R N Adams, Anal Chem ,30 (1958) 1576 5 V St& and M Kopamca, Electroanalysis, 1 (19891251 6 J Wang, Electroanalysrs, 3 (1991) 255 7 S G Weber, Anal Chem ,61(1989)295 8 J Wang and K. Varughese, Anal Chem ,62 (1990) 318 9 W J Aston, Btosensors B1oelectron, 7 (1992) 85 10 S Alegret and E M~~ez-F~b~gas, Btosensors Bloelectron ,4 (1989) 287 11 R E Newnham, Ferroelectmx, 68 (1986) 1 12 G R Ruschau, R E Ne~h~, J Runt and BE Srmth, Sensors Actuators, 20 (19891269 13 J A. Lott and K. Turner, Cbn Chem ,21(1975) 1754 14 WE Homby and D L Moms, m H W Weetall (Ed 1, Immobdrsed Enzymes, Ax&gem, Antibodies and Peptides, Vol 1, Dekker, New York, 1975, p 141 15 M Mascml, M Iannello and G PaflescL, Anal Chun Acta, 146 (19831 135 16 R C Engstrom, M Weber and J Werth, Anal Chem ,57 (1985) 933 17 N Sleszyns~, J uterus and M Carter, Anal Chem , 56 (1984) 130 18 X Yang, G Johansson and L Gorton, h&krochim Acta, 1 (198919