CA1249026A - Measurement of ligand/anti-ligand interactions using bulk conductance - Google Patents

Abstract

ABSTRACT OF THE INVENTION

Methods, apparatus and sensors are described for detection of specific ligands in a fluid sample by measuring ligand-specific changes in the bulk electrical conductance (or resistance) of a fixed test volume, with antiligand or ligand localized in or near that volume.

Description

MFASyREP~ENT_QE_LI bND~A~rI=LI~AN~
I NTERACT r ONS US I N~i EIULK CONDUCTANCE~:

BACKl;iROUND OF THE I NVENT I ON
___________________________ This i~vention relates to an~lyti~:al methods an~î apparat~s"
more parti~ularly to n~ethods, apparatus and sensors for dete~tion of a substance ~f interest in a fluid sample.
There are rhany types of standard tests or assays f~r detection of the prese~nce and/or conc~r~tration of 5p~Ci fic substanc:es in fluid~A Until recently many of these a~says ræqwired deYelopment of di f ~erent ræagents and protocol~; for each substanc~ lto be clet;ected. Examp}e~ of these approac:hes include various e~yme a55ay protocols in bioch~nistry l~bor~tories and thæ Technicon SMAC machine~i, duP~nt Automated Clinical An~ly2er, and ICodak Ektachem 700 machines in c:linical che~nistry l~borat~ries.
In the };~st two decacle~s a new typ~ of diagnostic teslt 4;r as~iay h~s gained increasing ~Ise ---- the antibocly--based assay7 or ~ m~r~oa~say. ln intmunc~as5ays~ al~ anl;ibocly may be IJs~d9 lF~r example,, to pr~be for th~ presence of a particular an~i~en, ha~ptq~n~ or other mol ecul e.

~ ~ *

Immunoassays have several potential advantagæs over previou~
a552ly5: .
~ a~ they are procedurally generalizable, that isr the same type of assay procedures and reagents can be used to dete~t ~ost antigens no matter what t5~e chernic21 properties of a partiular ant i gen ar e.
~ b) they are highly speci~ic; that is, thæy can potentially distinguish well between structurally related comp~unds.
~ .) they are potenti~lly v~ry sensitive.
Irn~unoassays are a speci ~ic type of a more general ~ssay strategy, the ligand/antiligand assay. All ligand/antiligand assays are based on two premises: (i) that c~rtain pairs of substances ~the ligand and the antiliyand) have a strong and specific affinity for ~ach other, that is9 they will tend to bind to each other~ whilæ binding littlæ or not at all to other sub~tances; and ~2~ tllat mæthods and apparatus can be devel~ped that allow detection of ligand/antiligand binding interactions once cornplexes have formed. As us~d herein~ }igand is dæfin~d as the substance to be detected, and antiligand thæ substance used to probe for the presence of the ligand. ~In some ligand/antili~and assays, an additional, perhaps modified~ ligand ~ay be used that competes with the substance to be detected for binding sites on the antiligand.) In many cases, detæction Gf a ligand/antiligand complex is mad~ pos~ible by labelliny o~e component o~ the complæx in som~
way, to make the entire co~plex "visible" to an appropriat~

detecting instrument~ For example~ radioi~mun~assay (RIA) uses a radioisotope as ~ labeln S~e U.S~ Pat~nt ~05~ 3,555,1~3 and ~,646,3~6~ Enzyme immunoassays (EIA) use an enzy~e that can produce a detectable color und~r appropriate conditions. See II.S.
Patent Nos. 3,654,0gO, 3,791 g~2 and 39~50J75~ Sirnilarly7 f 1 uor esc ence immunoassays ~FIA) use ~ fluorescent label~
Yet another ligand/antiligand a~say that i5 becoming increasingly importailt is the nuclæi~ acid hybridization assay, e.g., the DNA probæ assay, whictl uses ~ "probe" strand of nucleic acid as an antiligand to test for th~ prese-1c~ of a complem~ntary DN~ sequence. DNA probe ass~ys, like immunoassaysJ often use radioactive lab~ls, fluorescent labels or enzyrne labels.
~ ecently, other ligand/antiligand assay techniques have been dæveloped that us~ less standard tags. For example, particle c~unting im~unoassay uses small latex spheres which scatter light in kn.~wn ways when the sphæres agglutinat~ due to a~ltig~n-antibody i~teractionO ~oth i~muno~5say5 and DNA probe assays have used lumin~5c ænt labels ~s well.
Labæls can gYéatly increase the se~sitivity of ligand~antili~and assays by associating the presence of the labelled compl~x with a detectable signal. However, label-using assays also often llave associated disadvantages. First of all, the l~belled immunoreagents must be prepared, whi~h may require time and e~pænse. Al 50, several processing steps may be necessary during the ass~y, such as addition of label, incubation to allow reac~ion between components, washing awaly of ea~cæss l~bel, - ~æ~

transfer to a detectir19 instrllmetlt~ and detection o~ th~ la~elled complex. ~ecause of their relative cr)mplexity" many o~ these a~says may consumæ considerable time and recluire skillæd technic~l help~ ~uch assays also may be relatively ~lard to automate~ except by means of expensivæ equipment. Further, these a~says generally monitor the level of only one 1igand at a time. Further still the labelled reagents rn~y be unstable, inconvenient, or dangerous to handle, as with the isotopes I~5 or P3~ and carcinoyenic en~yrhe substrates. Also, usiny the above and other curretltly available label-depend~nt assays, it is not possible to mo~litor the level of antigen continuously; thus in order to follow the level of an antigen or antibody over an ~tendæd time, samples must be withdrawn at intervals, treated with label~ and tested.
A number of t ec hn i ques ar e i n c ommer c i al use or und er development that seek tv avoid some of the proble~s noted ~bove by avoiding the usæ of labels or by modifying how labels are used.
1~1any such techniques probe for the forb~ation olF ligand~antiligand complexes by opti~al means. For e~ampl~ rate n~phelGmetry, a type of immunoassay, measures chan~es in the angular distribution of scatteræd light as antigensi and an~ibodies form ag~regate under certain conditions~ This method is not very sænsitivæ, however~ and re~uires testing of two dilutions of the sample~
Pregnancy testing7 syphilis testing, and blood antigen testingp among others, are often don~ by vi~;ual inspection for for~atio~ of præcipitated or agglutinate~ antigen~antibody complex~s. Strictly speaki~y however, the cells or par~icles often ~sed in these tests to makæ a~gregati~n visible c~ld be call~d label 5. In any case they haYe s~me of the disadvantages associa~ed with labels noted bove~ SUC~l ~S the lack of potential for continuous measurement or ~or ~aking si~ultaneous deter~inations cf n~ultiple li~ands in the same sar~ple. A number of other immunoassay systems use some ~orm of ~ptic~l detection with~ut relying on precipitin f~r~,ation~ ag~lutination or standard l abel s. See U. 5. Pat .. Nos.. 3~ g75, 238, 4, 0S4, 646 anl:l 4, 3217 057.
There are al50 several types cf irn~unoassays based o~
electrical detection of ligand~antiligand ~o~ple~es. These assays focus on overc4ming one or rnore of the disadvantages noted above.
One such method is the resistive pulse te.hnique taught by U.S.
Pat. No. 4J1~1,73~ which makes a bulk condùctatlce me~surement.
This technique is a modification of thæ ~oulter counter approach, whereby a conducting fluid containing non~onducting particl~s is passed through a narrow constricted cha~nel, wh~se overall ~bul~;) resistance is measured. Th~ ov~rall resistance incre~ses whenev~r a nonconducti~g particle trav~rses the channel, and the size of the "pulse" pr~duced is proportional to thæ particlæ's volumæn Thus, if partïcles coated with antibody, f~r exa~p~e, a~e exposed to antigen under appropriate co~ditions, they will aggregate, and the increased ~ize and nu~ber of the aggregates ~an be related to ~hæ a~ount of anti~æn present. Howe~er, the sensiti~ity ~f this t~chnique is li~ited by the presenc~ of self-~ggregates that for~
during the manufacture of the antibody-coated parti~l~s them~lv~s~ U.S. P~t~ N~. ~191,73~ seeks to eli~inate this . ~, .%~

prob1em and hence increas~s the sensitivilty of this ~echnique by ro~ting two parltic1e preparations of di f fererlt sizes with antibody, and countin!3 only a~gregates cont~ining both large and ~i~nall particles as indicati~re of the presence of arltigen.
However~ the techniqu~, as disclosed, requires an expensive apparatus. Also, it requires a label ~the particle5 to which antibody or other antiligand is attached), and it cannot easily be adapted to llneasure multiple ligand~ at the 5ame tirhe in a single sample fluid" nor to measure a continuously varying sanp1e.
U.S. Pat. Nos. ~,2~6,89:3 and 4, ~2,0~6 relate to the use of a p i æ2 oel ec t r i c osc i 1 1 at or that has been c oat ed wi t h ant i 9 en t o detect the presence of antigen or antibody i~l a fluid sarnple.
More speci fically, there is a change in the frequency of the oscillator as its mass changes du~ to bindirlg of antibody to the antigen on it;5 ~iurface. Although this assay does not require labellæd re~ents or sophis~icatæd instrumetltation, as disclosed it requires a co~plex protocol including removing the sensor fYom 501utivn after æxposure to the fluid sampl~ ~nd dryin~ it before measure~ents can be mad e .
~" U.S P~t. No. 4,05~a3,646 relates to the use of capacitancæ as one o1' c:.everal ways to n~easure the relative thickness o1F an antigen/antibody bi~olecular layer. In particular, conducting ~ubstrate i~ exposed to an an~igen, antibo~y, or other antiligand, whereby a monomolecular l~yer of antiligand forms on the conducting s~r~ace that will in general be an ælectrical insulakor ~hen drya This surf~ce is then exposed to a soluti~ containing the ligand of interest~ a layer ~f whic~ is also an insul~tor when dry. Finally, a merc~ry drQp or other ~lectrode ls contacted with the dry ligand layer~ The condwcting substr~te and the mercury drop comprise t~le two conducting plates of a capacitor separated by an insulating layer whose thickness depends on the amount o~
ligand that has bound to ttle antiligand. The capacitance of this capacitor is then ~easured by ~eans of a suitable instrument.
This technique does not require labell~d irnmunoreagænts or expensive instrumentation. However, like the piezoelectris assay above9 it requires that the ligat~d-containing capacitor be re~oved fro~ solution and dried before measurements can be ~ade.
Other types of electrical assays have sought to simplify ligand/antiligand assay procedures by measurin~ changes in electrical properties associated with a surface or interfacæ in contact with an electrolyte~ For examplæ~ when a surface such as a metal electrodæ is exposed to an electrolyte solution, ~trong local gradients of electrical charge and potential arise in the regon of the electrode~electrolyte interface~ Because the gradients and associated electrical potentials are strongg liyand molec~les that interact with an i~unorea~e~t o~ otller antiligand i~obilized at the interface can have a considerable effect on the overall elæctrical prop*rties at the interface and thus ~an g~nerate strong ~lectrical signals of various kinds. Examples of assay~ based on the ~bove phenomena include voltamm~tri~ as5ay a~
t~ught in U~S~ Pat. Mo. ~,~33,144; field ef~ect transistor-~sed as~ays and oth~r semicc~nduct~r~based assays as taught in U.S, P~.

No~ 8,757 and ~7~34~B~Q, elec~rical reaCtance ~ss~ys as tau~ht in U.S0 P~t. ~o. 4~219~335; antibody ele~tr4d~ a55ay5 Analytical ChemO 567801 ~ig~4); Cile~ ~ Eng~ News, April 2, lg8~, p.~2) and other potentiometric assayst e.g~ as taught in U.S~
Pat~ Nos. 4,L51~049 and 4,081, 3 4.
Such interface-based ~55ay5 can be fast, simple to perforrn, and continuous~ since a sample c~n be ~onitored as long a~ it is in contact with t~le interface. Such assays can also b~
label-indepændent and ca~l be adapted t4 monitor multiple analytes simultaneously. ~owever, interface electrical properties ca~ be affected i~ many non-specific ways, 5uc~l as by variations itl pH or electrolyte cornposition or by nc,tl-spæcific adsorp$ion of speciæs in solution onto, or into, the surfac* or interface~ Thus, ~ethods based ott measurement of such properties are oftetl subject to in~erfere~lces tha~ are unpre~ictable or ~lard to control.
Further~ many ~f these methods require pr~paration of a r~er,-~rane or ~olecular layær at or on the interface that contains ligand or antiligand, and this can be hard to do reproducibly.
It would therefore ~e advantageous t~ have a ligand~antiligand assay syste~ which ret~ins the advantages while overcomi~g the disadvantages associated with the above - described a55ay5. M~re specifically, it would be advanta~eous to have an economical assay which is quick Ccan be completed in less than a ~inute)~ capa~le o~ being continuous, capable of d~tecting ~he presenc~ of m~ltiple liga~ds si~ultaneously in a fluid amplæ, simpl~ t~ perf~rm and label-independe~t~ It would also be advanta~eou5 to have an

2~

apparatus which enables su~h an as~ay to be performed and which i5 miniaturizable in the sense that it allows mea~rement of ligand~antiligand interaction in an ~xtremely small volume.
SUMMARY
_______ In ac~ordance with the present invention! there are provided methodsr apparatus and sensors for determ:inirlg thæ presence of a li~a~d in a fluid sample by measuring changes ln the bulk electrical c4ndu~tarl._e of a test volu~,e. The conductance chan~es are rnade ligand-specific by the presence :in or near the test volurne of a predeterrrlined re~ion containing locali7ed antiligand or ligand~ This predetermined region is exposed to the fluid sarnple, ligat-drantiligand interaction occurs, and the resultin~
conductance c~anges are monitored by any suitable conductance-rnæasurin~ instru~t.
This invention also r~latæs to methods and apparatus for elir~.inating no~-specific noise and dri~t by i~ co~,paring the bulk cond~ctance of a test volu~e with nearby negative or positive ,-ontrol volumes, and/or ii) minimi~ing effe~ts of pheno~ena ~ccurring at electrode~el~ctrolyte inteYfaces throuyh use of a zero-current~ four-electrodæ mæasuring technique and rece~sed electrodes.
More specifically, in one embodimeilt, the fluid sample ~low5 through a pred~t~rmitled regi~n consistins of a matri~ on which antiligand or ligand i5 i~mobili7ed. A suitable control mole~ule may be immobili~ed in a nearby predetermined re~io~ Otll test and contr~l predetermined regions are s~all ~typacally on the g_ o~der o~ 0.1 microliter~ and the ~onductances of the test volumes ~re monitored ~sin3 a zero-c~rrent fo~r electrode ~chnique, recessed ele~trodes, and an appropriate conduct~nce rne~sur~ng instrumænt.
Advantages of the present invention include its speæd and ease of use. The measure~bent can be continuous and i5 capable of monitoring seYeral ligands simultaneously~ If desired, it can be label-indep~ndent~ yet it can also use a vaYiety of convenient9 safe9 inexpensive labels to incYæase s~ns:itivity with little or no 1055 in speed and ease o~ use~ The present invæ~ltiotl also ~inir~izes effects of phet-ome~a tha~ occur at ~lectrode/electrolyte int~rfaces~ Further~,ore, the electYodes can be physically separate from the region containing localized antiligand or ligand; this allows thæ s~parate preparation 4~ el~ctrodes and localized antiligand or ligand ~o~positions~ allows their preparation to be potentially simple and inexpensivæ7 and allows flexibility in choosing methods, techniques and ~aterials.
Finally, a variety of instrun~ents exists which .an be used or modified to m~asure changes in bulk conductance due to ligand~antiligand i~1tæractions.
~RIEF DESCRIPTION OF ~HE D~AWINGS
Figure lA 5~10W5 the chang~ in thæ cond~ctance ratio betweærl a test sensor and a control sensor after additi4n of a fluid sample cont~inin~ a ._onstant concen~ra~ion of ligand; Figure lD is a c~libratio~ curve relating th~ initial rate of change of the ~ond~ctance ratio to a known ~onc~ntration of li~and in the fluid sa~ple.

Figure 2A shows the change in the conductance ratio in a Eluid sample in which the ligand concentration varies with time; Figure 2B shows the change in conduct-ance ratio after a fluid sample pulse.

Figure 3A is a calibration curve relating the per cent of binding sites occupied in a predetermined region at equilibrium to a known concentration of ligand;
Figure 3B shows the change in conductance ratio monitored via measurements made at equilibrium.

Figure 4A illustrates the binding of ligand/
particle complexes to a prede-termined region where the test volume includes at least the surface of the pre-determined region; Figure 4B illustrates the binding of ligand/particle complexes to the surface of a pre-determined region where the test volume does not include part of the predetermined region.

Figure 5A shows typical apparatus used in determining the presence of a ligand in a fluid sample;
Figure 5B shows the electrical equivalent. Figures 5C and 5D show other possible configurations of the apparatus.

Figure 6A shows a multiple sensor apparatus with several test cells and a negative and posi-tive control, each cell having its own e]ectric path and fluid path;
Figure 6B shows a multiple sensor apparatus with cells in series, all cells sharing a common current path and fluid path.

Figures 7A and 7B show two versions of a recessed planar sensor and apparatus.

Figure 8A shows a section of t:he measuring means in an IVEP
E;ensc~r (see definitions~ with the biolayer- inserted in the s~nsor; Figure 8B shows a c105e--Up of the sensiny region ~r~ th~
I VEP sen sor ~
FiguYe gA shows a sectic,n through the measuring means of an EVEP sensor ~see definitior\s); Fiyure 9B 5how5 a water bath for use wi th an IVEP c,r EVEP sensor .
Figure 10 shows speci fic dete.~tioll of an antibody directed against rabbit }9~ using an IVEP sensor~with subsequænt ampli fication of the sigr~al using a parti~le.
Figure 11 shows simultaneous det~ction of two di f ferænt ligands and use of a positive c.~ntr~l.
Figuræ 12 is similar to Figure 10 except that an EVEP sensor was use~d.
5:igure 13 show5 specific detectiorl of an antigen llsing a columnar s~ns~r.
FiguY~ 14 i5 a block diagram o~ ~ circuit for rneasurirlg conductivity f,~r use with test and control cells U5i~1 foLIr-~erminal measurements., FiguYe 15 i5 a schematic diagram showing furthær details of the comparator circuit of Figure 14.
Fi gur e l ~i i 5 a sc h emat i c d i ag r am show i ~9 f ur t her d et a i 1 s o f thæ transformær ~ircuit of Figure 14.
Figure 17 is a schematic diagrarn showing further details o~
the shield and ~u~rd driver amplifiers of Figur~ 14.

Figure 18 is ~ schematic diagram showing further details of an AC b;-quad amplifier used in the circuit vf F~gure 14.
Figure lg is a schematic diagran~ showi.ny fuYther ~etails of th~ phase shifter circuit of Figure 15.
Figures ~0, 21 and ~ are flow diagra~s showing ~ne r.~ethod by which a measurernent of conductivity may be perforrned by ~he circuitry described hærein; ~nd Figure 2~ shows an alternate embodiment for the embodiment shown in Figure 14.

DETA I LED DES CR I F''T I Ohl DF THE I NVENT I ON
____ _____ __________________________ The following dæfinitions are pr~vided in order to facilitate understanding of the present invention. To the extænt that thæ
definitions vary from meanings within the art~ the definitions below are tG c~ntrol.
Ligand rneans substance to be detectæd that specifically binds with another substance, the antilig~ld.
Antiligand mæans substan~e ~sed to pro~e for t~e pres~nce of a 1 i gand that spec i f i cal 1 y ~i nds to the li~and.
Test Volume means the volu~e whose conductance ls mæasured to determine the occurrence of ligand/antiligand interaction in a predetermined region.
Prædetermined Region mea~ region having lig~nd or antiligand localized in it.
Localizing ~eans is the mean~ used to localîze antiligand or ligand in the predetermined region4 !

~ ( - ~%~'~`2~
Contacting ~ean~ i5 the mæans used to insure ~nta~t bet~een th~ loc~liziny ~eans and fl~id sampl~ ~ontainiM~ the ~igand of interest.
~ easuritlg ~ean5 is the ~h~ans used to measure the bulk conductanc~ G f the test volu~e.
Sensor rnæans the localizing and measuYing means taken toget~7er.
Efficiency E of a sensor rneans the resistance of the prædet~rmined region divided by thæ r~sistancæ of th~ test volume.
~ onduc t anc e r atio ~ means t h e c c nduc t anc e o f t he t~st volume divid~d by the cor du.-tance of the negativ~ control ~olume.
IVEP Senscr means an Internal Voltage Equivalen.e Point sensor apparatusu E~EP Sensor n~eans an Extærnal Voltage Equivalæncæ Point sensor apparatus.
Bioregion means locali~ing n~eans that is mountæd on a bic~layer~
Biolayer mæans the combination of onæ or more bioregions and an appropriate supporting ~i~ture~ constructed for easy insertion or place~ent of bioregions into or onto a ~ensing appar~tus~
This invæntion relates to rnethoc3s, apparatus and sensors for det~rmining the presence of a liyand in a fluid samplæ by meas~ring liyand-induced changes in bulk ælectrical conductance.
I~ a preferred embodiment, a~tiligand to the lig~nd of intær~s~ is lccali~ed within a pr~determirled region.

': 1~

The predætermined region is expose~ to the fluid sample to be analyzed" and the~ bulk e:onductance of a voluns~e (the test volun~e) that at least par~ially contains the pred,i?t~rmined region is measuYed. Shanges ia~ ~hE? bulk conductanl:e of this tes~ volume that arise as a ræsult of ligand--antiligand irtter~ction in the predetermined region may be used to deter~ e the presence of 1 i gand i n tt7e f 1 ui d sar~pl e.
More ~enerally~ a variety of spatial relationships may exi5t b~tween the predetermined region and the~ ltest volume. F~or exarnpl~, thæ test volurne may e~tirely include the predeter7nin~d re~ion or alternativelyr thæ predetermined region may entirely include the te5t vc~lurne. The pr~determined region and the test volume rlay overlap, may be near each other but not overlap, or may be at a distance from each other.
Although the present m~thods and apparat~ls are de~;igned to detect the presence of a ligand in a Fluid sample, thæy are readily modified to detect the p~-esence ~ a ligand in a ga~v e.g~ by dissolving or bubbling the ga~ through an appropriate fluid, or in a solid, e.g. by di~solving the solid in an appropriate fluid.
While not wishing to be bound by th~ory, it i5 believed that a ligand~specific conductance change occurs in the test vclume because the presence or absence of ligand influences the distribution and/or con~entration of charyed sp~?cies which carry electric current thrcugh tbe test volume~ For exa~pl~ wl~en liga~d binds to antili~and immobilized in th~ predeter~in~d region 2~ , within the test volumey the b~un~ ligand occupie5 spacæ in the test ~olume w~ich is then no longer available ~o conduct surr~nt Cilanges in bulk conductance that arise from liyand/atltiligand c~mplex formation are in many cas~s s~all Con tfle order of a p~r~ent or le~55~ . Lar~e nonspeci fic chatlges ;n conductance can arîse frorn local environ~tltal variations, for e~ample, in tæmperatur~, composition of t5l~ fluid samp]Le~ visc4sity changes, or non-specific binding of proteins or other substances to th~
pr~deter !ni ned r~gi,n or test volume. Thæse changes catl t~nd to swamp out thæ ligand-specific changes c.ne i5 l~oking for. For ~xa~,ple, a changæ in temper~ture of 1 degree C~lsius can produce a 2~. to ~% ctlangæ in conductivity--perhaps lar~er than the entire ligand~specific effect on~ i5 se~king.
In order to be able to detect and quantify accurately the ligand-specific conductance change of int~rest in the presence of nonspecific changes, in a pre~err~d æ~bodirnent one cornpares th~
~ul~ conductarlc~ of a tæst volum~ with th~ ~ulk conductance of at least one control volume. ~oth positive and negativæ ccntrol volumes may be useJ.
ln general~ a n~gative control volumæ should be ~s ree as possible of non-specific interactions that cause changes in bulk c4nductance. At a minirnumr the nægativæ c4ntrol volume should not contain antili~and which reacts with t~e ligand of inter~st. lt may be suf~icient for the control volume simply not to contain any locali7ed substance therein. ~tow~ver, for the strqng~st application o~ this di~f~ren$ial, or comparat~ principle, th~

~16-test and negative control should be as similar as possiblæ. For example~ in a preferred ~l~bodime~t the negative control volume at least partially contains at least one pYedeteYmined re~ion that i5 exposed to the fluid samplQ, and this predetermined re~ion has locali~ed in it a rholæcule whose physical properties are similar to the physical properties cf the antiligand localized in the predetermined region of the test volume, but that does not bind 5peci fically to the ligand of interest~ For instance, if a test predeterrni~ed region contai?ls rnouse IgG directed against a hepatitis virus antigen~ a good negative control could ~e mous~
IgG from a naive animal. The use of this differe~tial tæchnique relaxes the need for præcise external temperature measurernent and contro~ This ~reatly simplifies conductance measure~ents and helps makæ the~ more præcise. ~pecific apparatus that mal~æs use o~ di~ferential techniques is disclosed later.
The conductance of a test volume can be compared with t~-at cf a negative co~trol volume in various ways. For ex~mple~ the value of each conductance n~dy ~e measured, then tt~at of the ~ægativæ
cot~trcl subtracted from that of the test volume man~ally or ælectricallyO Preferably~ the conductances are expressed ratiometrically as a function of time, using the conductance ratio C~ whære C = cQnductancæ_ f_test volume _______ conductance o~ negat~ve co~trol volume The test volume a~d30r the negative control volume can be ~urther compared with a po~itive control volume. In this ca5~ a positiYe control ligand is present in known concentratio~ in the fluid sa~ple~ and th~ positive ~ontrol volumæ at least partially co~tains a pred~terr,~inæd region which itself has locali7~d within it an antiligand which reacts with thæ positive control li~and~
The tæst volu~e and the positi~e control volu~,le m~y be ~ompar~d to tl7æ negati~e co~trol volume to correct for non-speciflc conduct~nce c~lange5, tllen to each other. Th~ positive control si~nal can be usæd as a ~alibration method, as a way to monit~r flow, or as a check on proper functioning of the methcJd and apparatus.
Locali~ation of antiligand can be accomplished in a vari~ty of ways. Preferably, localization comprises immobilizing antiligand on a matrix. A variety of commer~ially prepared matri~es aræ
available~ Such materials are easy to handle, and antiligatlds are easily bound to the~ in straightforward w~ys known to tho~e skill~d in the art. M~tricæs include gel beads, gel layers, glass beads, polymeric beads, micr~porous membranes, porous paper, filter m~$erial, or mi~tures thereo~. Preferably, the matrix i5 a porous ~embrane or paper that binds a variety of a~tilig~nds, such as nitr~cellulose filters (~illipore Corporation, ~edford, ~A~.
l~if~erent antiligal~ds rnay be bound to different predetermined regions of matrix material~ ~llowing multiple lig~nds to be detected in a single fluid s~ple.

( I

Localization of antiligand ~ay al50 be acc~mplished by confining the antiligand within a region defined by the bo~ndaries of ~ serni - permeable membrane that ig per~eable ~t least to the ligand but not to the ~ntiligand. For exah~ple, the predeterr,~ine~
Ye9i~n rbay consist of an atltibody solution confined by a dia~ysis membranæ.
Ti~e fluid sample and the predetærrnitled region may be brou~ht into contact in a vari~ty of ways as desf:ribæd in more detail herein below. For example~ a f l owi ng strear~ of sample f l ui d may he contactæd with the matrix of the predetermied region. The fluid may flow either through the n~atrix or past the rnatrix witi~
diffusio~ .-arrying the ligand of interest into the matrix.
Where the predetermined region corhprises at-~tiligand confined within a se~i-permeable m~mbrane, fluid is caus~d t~ flow past the memhrane. The ligand di ffu~;~s through the mæmbr~ne into the predeterr,litled reg i Otl.
T51e pre~ nt invention may b~ used in any situation where ~
li~and-antiliganli pair associate.. Examplæs of ligands wh}ch ~ay be detected using this syste~n are antigens bound to cell surfaces or other particles, free antig~ns, hapt~ns, antibodies, nucleic acids, enzy~es or ~utant enzymes, cofactors, enzyme substrates, receptor proteins, permease proteins~ transport proteins, speci fic binding proteins such as peripl~smic proteins~ molec~les boutld by a re,ceptor protein or ot~ler binding proteins; c~rb~lhydrates, lectit~s~ met~l iot~s~ etal ~inding proteins or other speci fic binding substances,. In ea~:h ~ the abs3ve cas4~s,, there i~i a substance that ~3peci fically binds to the ligan~ mentiotJedq For exampl~" if th~ ligand is an ar~tigen, the antilig~nd may be an antibody. If the ligand is an antibody~ the a~ntiligand rnay be an antigen. If the lig~nd i5 a nu~leic acid sequence, the antiligand may be a cornplerd~An~ary nucleic acid sequense. I~ the ligand is a horrnone" the antiligand may be a receptor protæin for that hormone, etc. This list of liga~lds and antiligands is not meant tc, be exhaustive but sirhply to illustrate the general principal that this invention may be use~ to det~ct any ligand f~r whi~h ~n ~ntiligand is 5 nown to exist9 can be ~ound, or ca~ be constru,-ted.
I~ a particularly useful application of this invention ligatds on cell surfaces may be used to detect the preserlce of whol~ Cell5 or particles~ fcr exarnple, red blood cells, white blood cells, bacteria, or rrærbbrane fragments derived from them. The rnethod of the present invention is particularly powerful in such cas~s bæcause the ligand is already associated with a nonconducting volurne that is rnuch larger than its own (molecul~-sized~ volurbe.
It is increasingly becomi~g possible to synthesize and~or modi fy rnolec~l~s so as to give the~ the propertiæs d~siræd for specific an~iligands such as a particular a~finity or specificity. For examplæ, mutant enzymes can be isolated wiSh altered a~finity, specificity or en~yrnatic a~tivity. Antiligands can be treated chemically to alter tt7eir propærties. N~cieic acid s~quences rnay be synthæsized, speci fiC peptide sequence5 may be synthesi~ed, or nucleic acid seque~cæs that direct the synthesis particul~r proteins rnay be specifically altered, ~.y. by using "

techniques of genetic engineering, thus allowing the synthesis oE altered proteins. These and other techniques known to those skilled in the art allow use of an in-creasing variety of antiligands with -the present invention.
References to some of these techniques may be found in the following: Advances in Enzyme Technology: "Artificial, Semisynthetic, and Designed Enzymes", (Technical Insights, Inc., Ft. Lee, NJ); Morinaga et al., Biotechnology, July 1984, p. 636.

Antiligands may be used that have either high affinity or low affinity for the ligand. Hiyh affinity means that the antiligand tends to bind strongly to the ligand, low affinity that it tends to bind weakly. A useful quantitive measure of affinity is the affinity constant K, which is equal to the concentration of free ligand needed to saturate half the antiligand binding sites at equi-librium (as long as the total number of ligand molecules is much larger than the number of binding sites). When the affinity is high, the affinity constant is low, and vice versa. More specifically,antibodies with affinity constants less than 10 10 M are generally considered high affinity, and antibodies with affinity constants greater than lO 8 M are considered low affinity. For the purposes of the present invention, high and low affinity are typically defined situationally: An antiligand is considered high affinity when -the concentration of the ligand of interest in the fluid sample is much higher than the affinity constant, and low affinfty otherwise. While not wishing to be bGund by theory, ~odels are described below that are helpful in und~rsta~ding how ligand~antiliga~d int~ractions may be quantit~ted.
In thæ high affinity si~uation, the rate of association r1 =
kl CL~AL3 of lig~rld and at~tiligand to form a corhple~ greatly exceeds the rate of dissociation r~ = k2 (L/AL) of ligand/antiligand complex as long as a signi ficant fraction o~ the bi~ldi~l~ sites are uno~cupied; furthermore, sub~tantially all binding sites are occupied at equilibrium. [In th~se expr~ssions ~1 i5 the forward rate constarlt, k~ is the ba~kwa~d ra~
constant, and ~L~, (AL~ and ~L~AL) are respectively th~
ccncentrations of freæ li~and, u~lbound antiligand, and ligand/antiligand complex in the predetermill~d regiorl.~ This follows from the definitiorl of the affini-ty constant K~ where ~ =

k2/k 1 .
Thus, to a first approximation, the rate of disso~iation r~ of ~igand/atltiligand complex rbay be ignored, and the association of ligand and antiligand is essenti~lly irreversible. As fluid san~ple passes through the matrix in 5uch a situation, binding is sub~tantially qua~titative as long as the ~luid passes through the ~atri~ slowly enough and a sig~ificant fra~tion of the binding sites ar~ aY~ilable; that i~, virtually lOOX. of the li~and binds to immobilized antiligand and is removed from the fluid sample.
In practice, the flow r~te needed for quantitative binding depends ~n the re~ction conditiGns 5uch as salt ~otcentratio~, t~mperature~ ~nd th~ ~hickn~s and por~ size of the ~atrix a3 well `~ -2~-~s on the nominal for~ard reaction rate~ Typically7 linæar flow rates on the order of millimeters or tenths of millin~ter~ per second allow quantitative or near-guantitative binding to a matrix that is 0~1 rnm thick) 0.45 urh in pore size and 10-4 ~ in immobilized atltiligand~
~ inding næed not be quantitativen In some cases it may be appropriat to pass fluid through the pred~ter~ined region at a rat~ that does not give time fGr quantitative binding. In such a ~as~ it ~ay be satisfactory tc,l.now th~ pr~nt of binding which o~curs7 ev~n if this is not 100'~. In ~n extrær~ case, fluid will be passed throuyh the predetermin~d region so quickly that only a small portiGn of the ligand entering the pr~deter~.inæd regio~
binds to it during its transit time.
~ uantitativæ binding cc,nditions have the advantage that th~
amount bound i5 not strongly dæp~ndent on the exact coslc~ntration o~ immobilized antiligand or on the exact thic~ness of the pred~termined re~io-l, as long as th~ m~trix remains unsaturated~
The extreme non-quantitive binding situatiGn~ which ~ay approximate that cf homogeneous kinetics, has th~ advantag~ that i~ is not strongly dependænt on the flow rate.
For qua~ltitative binding, thi? amount of ligand bound per ~nit time will be proportional both to the concentration of li~and in th~ fluid san~plæ and to thæ flow rat~. This yi~l~i5 a ~ligh af finityt or icineticy rneasure of ligand. For ~xample, i f th~?
conductancæ ratio C is plotted as a functi~n of time ~as not~cl above~, the~ slope o~ th~ resulting line is a function both of ligand con~ntrati~n atd flo~ raten Thus, if the flow rate i5 known, ligand concentration ~n be determined. For instance, re~erring now to Fig. lA9 i~ the concentration of th~ and is ~onsta~t~ th~ 514pe 10 will be constant until ~h~ pred~terrnined region becor~les saturat~d 1~, assur~ing a cc,nstant fl~w rate.

The observ~d value of the slop~ dC will d~pend not only ~n the dT
concentration of th~ particular ligand in the fluid sar~plæ atld th~
fl~w rate7 but also on oth~r particulars of the experiri-etlt s~t-up, su~h as C1) th~ pore si~ of the predæt~rrnined region; C~) thæ total volume of fluid in the bioregion available for ion flow; C~-) the efficiency E of th~ sensors, (4) the effe.~tive si~e of the ligand immobili~ed (~g. the partial n.olar v~lume if thæ li~and is a molæcul~ or ion, or the cell volu~ if th~ ligand is bound to a cell surface); and ~5~ chang~s in partial molar volume o~ ligand or antiligand or both resulting from changes in their structure due to bindin3n Hencel a standard curve such as in Figure i~ mu5t b~
determined for a particular set of experi~-ental con~itions used.
Ot~ce this i5 done, Sl~wever, ~he presence and concentration of li~and in an unknown ~a~ple ~luid may be detern~ined by refere~ce to ~he standard curve.
~ecause dC depends upon ~4) and ~5~ above9 the present dt invention may be used to stl~dy or deter~ine such characteristics~
This could be u~e~ul in r~search; e.g. studies of confor~ational chattyes in antibodies when they bind their antigens~ (See ~rotein_ C~nfor~tation 3~ a~ ~ olog~sal ~9~1, Celada e~ al., ~l~num, NY
~19~333.

Various ~eans ~ay be employed to ass~re a known flow rate9 a~
wi~l be ~pparent to ~ne skilled in the art~ For instance~ the apparatus may be designed to maintain ~ c~nst~nt, known flow rate throughout the co~rse of a measurement, e.g~ through use of appr~priate fluid resistors and~or a peristaltic pur,~p. Flow r~,ay be ~e~suYed by various com~ercially available devices (e.g., a G-1000 ~icroliter pær minute flow ~eter with reyulator valve~ from ~ilmon~
Instruments~ Great Neck~ NY). For particular sample fluids and li~ands the chang~ in corlductance of a matri~ due to both specific and non-speci~ic binding of rhaterial from the fluid sa~,ple may wel~
be correlated in a known way with the change in fl~w r~te through the filter at a gi~en applied pressure, since both changes may arise fro~ common causes such as the binding of n)at~rial from the fluid sampl~. ~he conductance ratio itself may someti~es thus be used as a measure of flclw rat~.
Referring tlOW to Fig. 2A, i~ the ligand concentration in the fluid passing through thæ predætermined region varies with time~ the slope dC will vary with ti~æ. For ~xa~ple, at point 17, the dt ligand concentration in the fluid passing through ~he apparatus is . si~nific~nt9 and as a result, the c~nductance ratio decreasæs at a significant rate as li~and binds to and occupies spaGe in the predetermined regi~n. Later~, at point 19~ the ligand concentration has dropped to ar~ undetect~ble level, ancl the conductance ratio is not changing d~tectably. 5till later~ at point 219 the ligand . , -2~-¢~
concentration has r~sen to a very high level, an~ lnally, at pclint :23, i t ha~s f al 1 en aç~ai n to a very 1 ow 1 ~vel . The 1 :i gand concentration at a par'cic-llar tim~ rnay be determan~d by m~asuring the instant~neous rate of change o~ the ~onductance ratio" i .e. " the slope of th~ graph at that time; again assur.~ir19 the flow rate i5 held constant, or i5 at least ~nown. This provides a cvntitluo~s rneasure~ of ligand c3ncentratiorl.
Saahple fluid rnay al50 be contact~d with the predetermined regi~n in a pulsed fashion resulting in a curve su~h as that shown in Fig. 2B~ For examplæy fluid sar.~plæ may be injected into a flowing buffer strear~-, of the sam~ conductivity U5itl9 standarti ccrirner~ial valve equipment ~e.g~ ~heodyne va~lves wi th s~Lmple injection loop, Rheodyne In~tru~nents, Co~ati, CA). Aft~r a known time, depænding on the 5ar!~plæ V41Uh~e and the flow rate, the sample pulse will pass entir~ly through the sensor~ Where ligand binding is ~uantitativ~ ei~her the rate o~ change 13 of the condu~tance ratio or thæ total changæ 1~ of the conductance ratio ~ay be used to quantitate the concentration of ligand as long as the predeter~ined r~gion has not becom~ s~turated. Thæ lattær method has t5~e advantag~ that it i5 not dependent ~pon knowi~g th~ exact flow rate. The form~r has the advat~tage that it i5 fastær~ Eikher method n~ay ~æ nllost appropriate under particular æxperiri~ental c i rcumstances.
As mentioned above, in some 5ituation5 it may Ibe desir~ble t~
use an antiligand with low affinity for its lis3and. Sllch a situal1ii4n can arise" ~or exa~ple wher~ th~ ligand of intereE;t i~

--~!6--present in high c~ncentration in the fluid sa~ple~ e~9D 10-3 ~, ~nd it i5 desirable to monitor th~ concentration o~ ligand c~ntinuously over ~n extended period of time. Since the predetermined region has a lirnited capacity to bind li~and, the time over which concentration co~ld be ~onitored, given quantitative binding~ would be lirnited. This potæntial probl*m is addressed by using an antiligand whose affinity constant ~ typially is within a~
order of rnagnitude of the concentration of ligand expected in thæ
fluid sarnple, i.e., a low affinity situation. In this situation, the rate of dissociation r2 of ligand~antiligand co~plex is significant compared to the rate of association rl Ci.e~, binding is reversible), and a significant fraction of thæ bindi~ig sites are unoccupied at equilibrium. The fraction of ~Inoccupied or occupieJ
binding sites will be a knowrl function of the concentration of li~and under gi~en experi~ental c~nditions; as shown in Fig. ~A~
CIn this figure, K, the affinity constant 15, repr~sents the concentratiotl of ligand needed to saturate hal~ of the binding sites in thæ predetermin~d region at æquilibrium ~s noted above.) A curvæ
sush as that of Fig. 3A ca~ thus be used as a standard curve to determine thæ præsencæ or concentration of a ligand in a fluid samplæ. The dyna~ic range of such low affinity~ or equilibrium~
methods can be increas~d if desire~ by using several predet~rmined ~egion~ containing antiligands for the s~me ligand but with different affinitiæs~ or by i~mobilizang several s~ch antiligands in a single pr~deker~in~d regisn.

-2~-....

( ~ eferring to Fig. 3B! ther~ is shown an æxample of the us~ of l~w affinity antiligand in the detection of varyin~ c~ncentYations of ligand in a fluid sample. Whæn thæ concentration of ligan~ in the fluid sa~,ple passing through the sensor rises ~16~, the conductan-e ratio fall 5 a5 ligand bind 5 t 0 and occupie5 spac e in the predetermined region, ~,aking that space unavailable to current flow. Whetl the ligand concentration falls (18), the conductatlc~
ratio rises. At ~5, thæ ligand cGncentration has f~llæn to a very low levæl, far below the affinity ~onstant, 50 that little or no ligand re~ins bound to the predeterrnined region, and the conductan e ratiC hen~e approachæs or equals its original valu~.
Later still, the ligand concentration rises again ~0)~ reaches a new high (~7), then f~lls again ~2~.
Thæ response time of this method depends on the dissociation rate r2=k~ (L~AL~ of the ligand/antiliyand co~plex. For antibodies, ~any o~ whi,-h have forward rate constants kl on the order ~f 2 X 107 per molar per s~cond, this me~ns that antibodies with affinity constants of 10-9 M or ~reater have response ti~es on the ordær of a minute or 1 es5.
It should be noted that antibodies or other antiligands with fast disso~iati~n times ~an servæ as the sensing æle~ents in convenient, fast reusable sensors. Such sensors require little or no treatment to remove bourd ligand.
Appropriate choice of low or high-affinity antili~and~ can be made by thcse skilled in the art. As no~ed a~ovæ, anti;igands them~lves m~y be altered to va~y th~ir affinityO ~150~ af~ y -2~-` :

~onstants can be altered in ~ v~riety of other ways, suc:h as variation in salt q~oncentration, temperaturey or amounts of detergent included in an experimer~t~ and c:hernical alterati~n oF the matrix to which antiligar7d is imfl~4bilized..
The sensitiviiy of the high affinity mæt5lod is it-~ersely proportional to the volurne o~ the predetermined region~ The larger the pr~dæterrhined region? the larger t5le total nurnber of binding sites (assuming ~ d~fined fraction o~ pr~deter~ined region by weight is a~tiligand~; thusy the lar~er the nurr.ber of liga~ld ri~olecules req~ired to result i~ a giYen fraction o~ occupied bin~ing sit~s.
Thærefore~ where sarhple si~e i5 limit~d or the co~centra~ion of ligand in the sarnple is low, it is d~sirable that the predetermined re~ion be srnall. Further, the sensitivity of the n~easureme~lt i5 also proporti4nal to thæ ef ficiency, E, of thæ particular sensirlg ~pparatus. Thus, both the test volume and the predetermined re~ion should be small for rnaximu~ sensitivity. Preferably" the Itest v~lume is ~tl l~he order of aboLIlt 0.1 mi~:r~liter to ot e mi~roliter or even less; the volu~ of the pred~termined region is also on this order or l~ss~ and the ef lFiciæncy E is 10~/. to n~rly 100%. !~uch alpparat~ls is described herein below~.
As an e:~;ample, suppose that or-e ml of a fluid sawcple containing 1 u9/h~ nd flows throush a sensor, the vol~n~e of wlho~ predetermined re~ion i~; ~bout one rnicroliter~ with quantit~tive bindi~g o~ ligand~ then, assuming th~ speci fic qra~.rity of tlle lir~nd is on the order cr one, the h~icroS~ram of bowld ligand wi~l ~ccupy a volum~ o~ about one n~noliter within the predet~rmined . ` '! ~ :

regior~ i..e., 0. l'X. of the volun~e of ~e predelterDflined regisn, yielding a conJuc~ance change of a~ou~ 0.1% Cor le~s i f the ef~iciency E of the ~;ensor is low). If the volurne ~f the predetermine;i region is ten times smaller ~:0.1 microliter~7 on t~le other harld, the sarre orle rhicrogram of ligarld will occupy 1~ rather than 0.1% of the predetermined region and thus will yi~ld a larger condu~tan~e change.
In some c~ses i~ rbay be desirablæ to amplify the chat79e in co~ldu.~tance tha~ occurs when a ligand and atltiligand ~ind to each other. As rnentioned aboYe, it i5 believed that the change in conductance o~ten is proportional to tt1e volurne occupied by bound ligand~ This signal can there~ore be amplified if the presence (or absence) of additional non-con~ucting volurhe can be associated Wittl the binding of ligand. Resulting conductance changes will t~len r e f 1 æc t t h e p r esenc e~ ( or ab 5etlC ~ ) O f t h i s add i t i ona 1 Yo 1 un e . Th i 5 m~y be accc:mplished in al variety o~ ways.
Mor e sp ~c i f i c al 1 y " a c omp 1 ex c an b e f or med b e t w e en a 1 i g a~d which interacts with the antiligand and a particle~ The particle rnay be COinpO5æ`d olF a variety of n~aterials ~uch as at~tili~and which interacts wi~h the ligand of interest, antili~and which interacts with ~he ligand of interest bound to an ampl i fying substance, macromoleculæs, molecular complexes, latex beads, lipid vesicle~;, ~on--conducting polymer beads, other non--conducting particIes, p~rtially conducting p~rticle~, magnetic particles or mixturf~s thereof. A variety of l-~etllod~ of attaching liyands to appropriate parlticles are k~own to thos~ ski}led in the art,.

~ he ~iimpleslt llse of ~ particle for enhancing the change in ~ulk conductance i~; whære ligand in the flu;d sa~ple is dire~tly boun~l to particles by incubation.. The li6~and/particle compl~x in the fl~lid sample is then exposæd to the pre~determinæd region, yiældin!3 a laYger sigt~al as ligand~particle con~plexes bind than if li~and alone had bo~nd. In a variation of the above; a known amcunt ~f lighnd that interacts with antiligand may be added to the ~luid sarQple~ This ligand can compete with ligar~d/pa~ticlæ comple~xe~s for antiligand7 th~s fo~mi~g the basis for a competitive par t i c 1 æ--enhat~c ed c onduc t anc e as say.
In another variation, a known am~unt of a known ligand that intæracts with antiligand is ~irst bound to particle5 and this co~plex is added to the fluid sample; ligand from the fluid samplæ
can then compæte with this li~and/particle complex~ forming another type of particle--enhanced competi tiYe conductance assay.
In ye~ another variation, a predetermin~d re~iGn containing immobilized antiligand i5 exposed t~ a preformed co~plæx comprising ligand and a particle, and the complex is allowed to bind to the region. This re~ion is then expos~d to sampl~ fluid containing ligand, which can compete off the ligasld/particle complex~ In a variation of this tectlnique, the ligand~particle complexes rel~ased from a prede~ermi~ed region ~ay themselvæs b~ detected, for example by monitoring the conductanc~ of a test volu~e associated with a filt~r matrix that tended to trap such released complexes, where the test volu~e i5 locate~ a~ a dis~ance ~ro~ thæ predætær~ined re~ion ~i ~ eO 7 downstream~ The detec~ion may ~lso b~ done ~y monitoring ., --3 1-f (~

color change or a pressure change across the filterp or by other ~eans.
In yet another vari~tion, refeYring now to Fig. ~A7 ~ known amount of li~and i~ complexed with a p~rticl~. This pr~form~d complex :24 is con~partrnentalized in at least one region 2~ that is adjacent to predetermined region 28. Comp~r~ment ~6 ~rlust itsel f be at least partially within teC.t volu~he 30. This for~ the basis of anoti7er type of particle-enhanced co~petitive co~lduct~nce ass~y. It ~ay be used to continuously monitor li~and in a fl~id s.arnplæ.
Compartn,ent 26 may be form~d~ for exampl*, by co~lfi~ling complex ~4 within a semipe~rneabl~ membrane 32 that is perrneable to the ligand cf interest. When the sample fluid contairls no free liga~ld~ .-omplex 2~ will tend to lie within thæ test volume~ since it will bind to the antiligand in the predetermined region. Wh~n the sampl~ fluid contains a hi~h levèl of the ligand, many complæxes 24 will lie outside test voluJI~e ~0, changing the conductance of the test volumæ, since they will be coripeted of f predetermined reyion 2a by the ligand in the ~luid ~ample and will be free to move through compartment ~6. The ligand~parti~le cor~plæx ~4 may bind only to the surface of the predetern~ined region, or it may bind internally as well~
It should bæ noted that the test volume and the predæt~rminæd regio~ do not have to overlap to produ~e a workabl~ method~ In Fig.
4E3, ~or æx~mple, test volume 30 and predeter7nirtæd region `8 do not overlap but nevertheles~ are proximate~ enoug~ to each other so that ligand-antiligand i~eraction occu~ring in the predetermi~ed re~ion in~l~en~es the condu~tance of the test volume.

More specifically ligand diffuses into cornp2rt~ent ~6 ~ndpredeterrnined region 28 and competes with particle ~o-hplex 24 for ~ntiligand binding sit~s. When the ligand cor7centration in thæ
fluid sample is loW7 ~no5t of cornplex 24 i5 iound to the predetermit7ed regior7 -8. Wh~n it i5 hiyhV rn~st o~ cornplex ~4 i~
free ~o move throug~-l compaYtrnænt 25. If test volu~e 30 i 5 at a distance frorn the surface of pre~eterrnined region 28, as showi1 in Fig. 4~, the test volutbe conductance will increase when the ligand conc~ntration i5 low, as ,-o~plæxes 2~ leave test v~lume ~0 and bind to predet~rr"ined regicn ~B. If test volur,e ~0 includes at least part of the region abcve the ~urface of th~ pr~deterrnined Yegi~n occupi~d by bound cornplexes, the test volurne conductance ~ay decrea~e when the ligand conc~ntration is low, as complæxes bind to the surface and cccupy part of thæ volurnæ of the test r~gion. In either case, measuring thæ conductanc~ of thæ tæ5t volu~ can be used to determine the presenc~ and con~entr~tio~ of ligand in the fluid san~ple.
Magnetic p~rticles ~o~plexed with ligand may b~ used in con junction with a magrletic field to rnodi fy the ability of p~rticles to stick to a predetærmi~ed region or to n~odi fy their distribution ~eaY the predetern~ined region. For example, a ~a~net placed under pred~terrnin~d region ~E~ in Fig. 4A or Fig.. ~B tends to attract ligand/particle complexes ~~ i f they contain magnetic material~ By varying the stYe~g~h or distanc~ of the magnet, one can vary the ability of partis:les to stick t~ prædetermitled regior ~ and a1BO the average ~oncerltraltion 4f par~icles at di f ferænt d i st ;~nc e~; ~bove t h i s r eg i on .

Similarly, a tangential flow field may be used to influence the ability of particles to stick to a ~urface~ For æxample~ one could remove particles fro~ a fluid sample whil~ ~asuring the amount of ligand present by passing the sample rapidly past a filter containing i~,mobilized antiligand; if t~e ~ilter's pc,re size is smaller than the particles? they will be excluded from the filter and also will be prevent~d from piling up on the filter by the "swe~ping" action of the fluid flowitlg tang~ntially over the surfa.e. Parti.-l*-fr~e fluid .-arrying li~and can ~leatlw~lile pass through the filt~rr allowing binding of ligand to the im~obilized antiliga~ld~ Or~ th~ oth~r ~)and, if one wishes to ~easure the presence of particles which themselves ha~e ligand boun~ to their surface, one might use a slower fiow~ strong enough to minimizæ
non-specific particlæ sticking but weak enough to allow ~ticking of ~hæ desired p~rticlæs via ligand~antiliga~d bindi~lg.
Conductance changes can also bæ amplified by forming a co~pl~x between a particle and antiligand sp~cific for the ligand of inte~est~ In one such æmbodimen~) a known amount o~ antili3and ~nd particles are incubatæd to for~ a preformed complæx. This comple~ i5 added to the fluid sa~ple. Ligand in the sa~ple binds to antiligand in the compl~x and also to antiligand lo~alizæd in the prede~ærrl~inæd region. The number of particles bound to the predeter~ined region depends on the amoutlt of ~ree ligand in the sa~ple. This gives a satldwich type of immunometric particle-enhatlc~d cond~tance assay. In a variatiot~ of this methody antiligand itsel~ i~ the particle~ or a p~ly~ ed c~mple~ o~ antili4and is the particle.

2~ .
Particle~; that Ihay be complexeci with anltili 7and are ~irnilar tothose that can be complOE~xE?d ~ith li~and, ~nd their methods of use with the present invention will be appa~ent to those ~killed in the ~rt~ For example~ a conduc~ance version of the sandwich Tandem assay oF ~Iybritech CSan ~iego~ CA, U.S. Pat. No. ~,~7~,110 is easily derivecl: A monoclc,nal antibody speci fic for an extended antigen i5 irnmobilized, æ.g. on a nitrocellulose filter. Fluid sample containing the antigen is expc,sed t~ the filter, all~wing anticen to bind. The~l a secor~d9 non--interferit-g m~noclotlal antibody specific for the antigen, complexed with ~ particl~ i f de~iired, i~i all~wed to bind to the antigen, and a condu~'car7ce c~lange resLIltitl9 from this binding is rnonitored.
In accordance with yet another variation, ligand that interacts with antili~and to the ligand of interest is lo~ali~ed within a predetermined region. The predetermined region is expos~d to the fluid sample and also to an antiligand which interacts b4th with the }iyand in the fluid sample and with ligand localized in the predetcrmined region. The bul~ co-lductance of a test volume that at le~7st partially contain~ the predætermined r eg i on i s m~asur ed ,, As with the previQusly described embodiment wherein antiligand is localized in the predetermined region, thæ bulk conductat7cæ of the test ~olllme may be cot~pared with the bul k conductance of one or mc~re cor7trol volun~es. The localized ligand may be immcL7ili7ed ot7 a rnatrix, and the matrix r~ay be contacted with a flowin~ stream of tl~lid sample. Alternatively, li~and rnay be lo~alized by being 35_ confined within the bound~rie~ ~f a membrane~ e.g~ a~ part of a particle ~omplex~ or in a polymerized formu This me~ttbrane i5 p~rltteable to the ligand to be detected and to the ~ntiligand that interacts with the ligand to be detected~ These ~bodiments are similar to the ærhb~dirnents taught above that use lo~alized antiligand~ Vari~tions will be apparent to one skilled in the art and are intended as part of this i~venti~n.
It should be understood that cGnventional labelling tec51niques may als,~ be us~d in con jun~tion with the present invention~ For exar~,ple~ in a sandwich assay where Iigand frGm th~ fluid sarnple has been bound to irnmobilized ~ntiligand and then been exp~sed to a preformed particl~ ~ornplæx containing a sæcond antiligand, the co~plex may in addition contain an en~yme label, in a rnanner si~lilar to ELISA assays. Enzyme substrate ls then passed over the sandwich fornted and thereby generates a desired conductance change n _itu in the predetermine~ region.
For æxa~ple, the enzyme may convert part of the suhstrate into gas bubblæs. Since t~)ese buhbles are nonconductive~ they lowæ~
the conductance of the tes~ volu~e, Alternatively~ a substr~te may be ~sed that is en7ymatically converted to an insoluble precipitate which binds to the predetermined regio~ and which again decreases the conductance o~ the test volume. These protocols are particularly advantageou~ embodiments because the label produced can be dete~ted in situ, in a higllly lo~alized re~ion. Of course~ other enzymes, substrates, and l~els may bæ
u~ed. For exa~ple~ an en~y~e c~uld be usæd th~t generates a char~ed 5pe~ie5~ th~s chan~in~ the conductivity.

Other techniques that can i~provæ performanc~ cfligand~ntiligand assays are known to those skilled in the art, and ~any of these ~an be e~ployed with the present inYention. For ' example~ U.S. Pat. No 4,092,408 teaches a method for ~chieving more reproducible, efficient binding of antibody to a rhatrix by using a precoat of second antibodyn There are also techniq~es wltich teach the use cf second ~ntibodi~s labællæd with biotin .~r en~yrnes, and ri,any of these rnay be errlployed alsc,, as long as thæ
label changes the conduc~atlce in a d~sir~d way.
The sensitivity of the above ligand/antiligand assays may be enh~r7ced by decreasitlg interfering effæcts such as rlorlsprci fic bindin~. The skilled practitioner ~ay use as a gui de i n e~hanciny the sensitivity of such conductance-based assays techniques used in si~ilar non conductance-based assays. For exa~ple~ in trying to detect a particular antigen bindin~ to i~mobili-æd antibody, it is possible to first imr~obilize to an appropriate ~atrix a protein that binds the antibody such as protæin A or second antibody.
The pri~ary antibody is then im~obili~ed by bindirg to the protein A or second antibody. This can result in less non-speci-ic binding when antigen binds than if the primary antibody had been directly bound to the ~atrix.
Because the pYesent invention depænds upon n~easuring changæs in the conduct~nce of a test ~olume7 appropriate care should be tak~n by the practitioner to avoid the introduction of btaterial into the test volum~ which ca~ interferæ ~i~l~ the r~)easure~ent, such a~ dirt particl~s~ cell~; and random bubbles. FoY e~a~ple, ~2~

pre-filtr~tion and deaeration of ~ampl~ or buffer may be appropriate in 50~ circun~stances. In othær cases, unwanted particles ~ay be excluded fYom part or all of the test volume ~hrough use of a ~embranæ, gel, or porous material which is in,pærrnæable to the particles. For exarnpl~ to rneasure the concentration of a rholecule in blood without interfærence from conductance changes duæ to variations in the nurnber or ~rientation of red blo~d cælls in the tæst volurne, th~ test volume .-~uld be occupied by a porous natrix that allows plasrna, but not cells, to entr7 e.~., by diffusior~.
The bulk conductance of the test volunæ can be measuræd in a variæty cf ways. In a preferræd embodimet-t, _ondu_tance i5 measured with ælectrodes that are in contact with an electrolyt~
using a 2æro-curr~nt, four-elæctrode rneasuritlg tæch-lique. In this method two electrodæs, called current electrodes or powær electrodes~ are usæd to providæ a current that passes through the test Yolume whose conductance is being measuræd9 and two othær elæctrodes, called voltage ~lectrodes or working ælæctrodes, are u~ed to measure the resulting voltage drop across the test ~lume. The voltage electrodes are part of a circuit with essentially infinite i~pedance, so that virtually zero net current flows thrcu~h the voltage ælectrode surfaces "~æncæ thæ name zero--curr~nt, four electrode technique. This te~nds to reduce prc~blems associated with phenomærla at ælectrode surfacæs, such as polari~at30n impædance5 volta~es associated with the presnce of a pol~rization irnpedance must be sllnall~ since ~o net current flows --3~1--~ $~

across the impedance. A further discussion of this method is given in Introduction to Bioinstrumentation, by C.D. Ferris, Humana Press, Clifton, NJ, pages 109-116 (1978).

Preferably, the voltage electrodes are recessed from the current flow and the current elec-trodes are recessed from thetest volume to reduce problems associated with polarization impedance. Specific apparatus incorporating these principles is disclosed herein below. Other recessed four-electrode conductivity cells are taught in U.S.
Pat. Nos. 3,963,979 and 3,939,408.

Two-electrode methods may also be used to measure conductance. (See Pederson and Gregg, IEEE Journal on Ocean Engineering, Vol. OE-4, No. 3 (1979)). These generally require higher frequency in order to obtain high accuracy.

Another method of measuring conductance involves the use of electrodes coated with an insulating material to avoid electrolyte-electrode interface effects. The coated electrodes are capacitatively coupled to the electro-lyte through a thin insulating coating. This method generally requires even higher frequencies. (See ~uebner, G.L., "Notes on radio frequency salinity measuring equipment at Texas A
and M College", Publ. 600, Nat. Acad. of Sci. National Research Council, 1958).

~ ~9432~ .o At eYen higher freq~encies, e.g~, ~igahertz frequencies~ i~ i5 possible to ~easure thæ dielectric properties of a fluid sample rather than the conductivæ properties~
Anothe~ tec~lnique which avoids the di~ect contact of electrodes with electrolyte is inductive coupling. An indu~tive salitlometær is described in an article by E~rowrl and Hanson in D~ep Sea ~æsearch, 8, 65-757 (1971). Also, comrnercial inductive ~alinometers are made by Beckman Instrurnents (Fullerton, CA3 and Aa~lder~a Ir1strurllents ~Berg~n~ Nl~rway~ arr,~ng ~thers. On~ ¦
Cinsulated) coil can be used t o pr oduc e a c ur r ent inductively in thæ test voluri,e; thæ magnitudæ cf this currænt, wi1ich i5 a ~easure of the conductance of the test volumæ, is dete{ted inductively by another, pickup, coil. This method generally requires lar~er test volumes~ Further, coil5 are difficult to miniaturize.
It should be understo~d that a variety of instrurnents ~an be used in conjunction with the various methods noted above to prvcess signals from a particular conductivity sensor or set of sens~rs. Instruments of particular use~ulness with the 2ero-curr~nt, recessed four-electrode ~næthods, apparagus and s~nsors described herein are illustYat~d in Examples 8 and 9 below. A simpler, analog circuit for makin~ ~erc-current, four-ælectrodæ measurements that can be modified for use with the present invention is i11ustrated on page 115 o~ Ferris~ 5Ue~a.
In accordanze with the presænt invention~ there is also pr~vided a~ apparatus for determining the præsencæ ~f a ligand in ~ fluid sample. I~ its simple~t configura~ion th~ app~ratus ha~

~ .

three basic c~ponents~ each of which perform~ a sp~cific function i~ d~termining th~ pr~s~nce o~ a li~nd in a fluid sample. They are: mean~ ~or localizin~ at~d or antiligand which i~teract~
with ïigand in a predetermined regiorl of the ~pparatus Cthæ
"locali~ing ~leans"S; ~eans for cotltacting the fluid sar,.pl~ Wittl th~ localîzing means ~the "contacting rreans"~; and means for ~easuring the bulk cond~ctance of a test volume which at least partially contains th~ predetermin~d region tthe "measurit7g m~ans")~ Thæ l~calizing means and the measuri~g means ~4g~ther cornprise the s~nsor.
~ ferring now to Fig. SA, th~re is shown a diagramnatic repr~s~ntation of a typical apparatus us~d in d~tern~ining th~
presence of a ligand in a fluid sample by measuring the bulk conductance of a test volume.. The t~st volume 109 at le~ast partially contains a predet~rmined region 108 that itsel f contains means for locali~ing antiligand which int~racts with the ligand of interest. The conductance o~ thi~ test volume is compared to the conductance of a co~trol volume 111 containitlg a control pr~determined r~gion 110 to increase the s~nsitivity of the meas~rementO Fig. ~A illustrates the physical arrang~ent o~ th~
appara~us, and Fig. SB shows the electrical equivalent. Figs. 5C
~nd 5D 5how variations of the meas~ri~lg mean~ and contacting means.
Pref~rably, the locali~ing ~eans co~prises a matrix ~aterial tha~ ~llows fl~id to fl~w throu~h it and that contains antilisand im~obili~d on it. The loc~ion of ~his ~atrix defines ,, ~4 ~ -predetermined region 108 of test volume 109 (Fig. 5~).
Likewise, the location of a matrix tha-t contains immobilized on it a molecule similar to antiligand but that is sub-stantially free of specific interactions which affect conductance typically defines predetermined region 110 of control volume 111. Antiligand in predetermined region 108 binds the ligand of interest as fluid sample flows through the matrix, changing the conductance of predetermined region 108 and test volume 109.

Typical matrix materials used as localizing means in several embodiments discussed below are microporous filters made of nitrocellulose or mixed esters of nitro-cellulose and cellulose acetate (Millipore Corp., Bedford, M~ or Schleicher & Schuell, Keen, NH). These filters have pore sizes ranging from 0.2 um or less up to 12 um.
Lower pore sizes generally bind more antiligand per unit volume but have slower flow rates. Particular requirements, e.g., regarding flow rate, binding capacity, and amount of non-specific binding that can be tolerated for an experi-ment must be taken into account in choosing an appropriate matrlx, Nitrocellulose filter has long been known to be an excellent material for absorbing proteins (Kuno and Kihara, Nature 215: 974-975; see also l'Millititer Methods", published by Millipore Corp.) Both proteins and nucleic acids bind to it virtually instantaneously, so immobilization is easy.
This and other immobilization procedures are well known to those skilled in the art (see, for example, Science 223 474-476 (1984)).

- ~2 -~,"., While r~ot wishing to be bound by theory the ~pproxim~e change in conduc~ance e~pe~ted from ~aximal bindiny of ligand to a m~trix ~nade ~f nitrocellulvs~ filter and loaded wilth antiligand as described in $~xample 1 will now be calculaterl..
Nitrocellulose filters such as HA Millipore (Millipore C~}rp.
bind about 1 uy of protein per mm~ of area9 given typical comrnercially available filters that are about Onl5 rnm thick tthis corresponds to a f i lt~r volu~læ of about 0..1~ ~.m~, or O.lS
rnicrQlitærs, per mm-- of area~. Also HA Millipore is 7'3% void voluri.e~ wit~ th~ remaining ~1~/. consisting of solid rnatrix material4 Thus, the volu~,e of the matrix available for fluid flow, el~ctri,- current flow, or antili~and bindiny~ is ~pproxirnately 0.1~ rnicrolitærs per mrn~ of matrix area. The partial rnol ar volume of rnost hydr at æd pr ot æi ns, i nc 1 ud i ng antibodies, is approximately 0.7 cc/g. ~Most organic molæculæs ha~e parti~l rnolar volurnes and densities around 1.~ Thus, 1 ug of protein occ~pies about 0.7 ~ 10 6cc, or about 0.7 nanoliters.
Thæ exp~ct~d decreasæ in co~lductanc~ causæd by immobili~ation of 1 ug of a protein antiligand per m~ of n~atrix area is th~s on the order of lZ., sinc~

0 7_nanolit~rs__ = Q OQ07 rnicr_liters ~ 0~6%
0.1~ rnicroliters 0.1~ microliters This concentration corresponds to a lo~al concentratio~ of antiligand in the predetermined region on thæ order cf 10--4~1 for a typicall~ sized protein antiliyand of about 100,00~ daltons, ~,.~. an antibody of lSO~OOO daltons.

~24~
If each immobilized a~tili~and molecule itsælf bis~ds about one ligand mol~cule of about the sa~e size~ such as another protein, the liga~d-specific conductance cha~ge associatæd with ~,-axi~nal ligand binding should then also be about 1~.
In 50~ cases, the antiligand will have multipl~ binding sites tfor examplæ~ see Exa~ple 1, in which the antiligand is an extended antigen with several antigenic d~ter~inatlts, and the liyand is a polyclonal antibody populatiGn fro~ an imb~unized ani~.al). Thæ expectæd ~na~i~al sigt~al str~ngth will then bæ
higher, since several volumes of li~and will bind per volume of immobi1i2æd antiligand, and the ri,axinlal concet~tration of bound ligand will also ~æ highær. Conversely, if the 1igand i5 a srnall rnolæcule9 the expected ~igna1 s~rength will be lower b~cause althou~h the n~aximal concentration of bound ligand ~oleculæ~ ~ay be, sayy 10-4 M, the vo1ume occupiæd by each bound ligand mol ecul e i s smal l ~
While not wishing to be bound by theory, in general it i5 expected that the observ~d change in conductance ratio ~C will be proportional both to the molar concentration b of bound liyand and to the partial mol~r vo1ume v of the li~and: ~C = 5 b v, w~lere 5 i~ a proportionality co~tant related to the signal strength. s itself will depænd on a variety of factors, such as the efficiæncy E o~ th~ sensor used; it m~y be dætermined fro~ a calibration cur~e such as that of Fig. lB.

,, . . ~ .

Other matrix materials ~nay be used as well ~as nitl-ocellulocje.
For example, the columnar ~iensor of Ex~n~plæ 4 uses cyano~en-bro~nide--activated Sepharose beads as a matrix mafterial ~Phar~nacia C:orp., Piscataway, N.J,, ) . ThiCU ma~rix material also binds abo~lt 1-/. by w~ight or volu~ne of n~any antili~ancls9 and binding i5 easy as described in detail by the Pharmacia Corp.
publication Affinity Chrornatography Principles and Mæt~ods, 197~.
The fact that for many matrix rfl,aterials, the maxirn~l conductance changæ is on th~ order of 1~/. if aMplifying rnethods are not used p~ints out the usefulness of r.~ethods and apparatus that allow accurate li.easuren)ent and that use negative controls as described elsewheret to help avoid obscuring of the signal by non-specific ~ffect 5.
The particular matrix rnaterial chosen will depend on many characteristics s~ch as bindin~ capaci ty, ease of binding, ease of handling, si2e and shape of desired predetermined regions and associated appara~us, flow rate desired, speed and sensitivity desired, etc.
While the lX. }ocalized conce~tration of ~any antiligands that is easily attainable using convenient localizing me~ns such ~s described above ~ay be ~ow in terms of the conductance chang~ it generates, it is actually quite high for ~nany biological molecule preparations æncountered~ being on the order of 10-4 M ~or an a~tiligand of 100~000 dalton 9ize7 as n~ted a~ove. Many ligands of interest are present at ~ch lower concentrations~ æ.~ R
M or 14wer~ in fluid s~mpl*s of int~rest~ such ~E; serur~p urin~

,_, f etc. Th~s all of the ligai~3d iM such a one ml s~mple can be concentr~ted and bound in a localizing means of 0.1 ul volume . hat i5 10--4 P1 in bindin~ site concentration. Usi~g the apparatusæs descrit:ed herein allows ~he fluidl sample to be intirt.ately cont~cted with a pred2terrnined r4yion containing such locali2ing means by ~llowing fluid to flow through the regionO This type of apparatus thus provides a pot~ntial ligand concentration ~actor of 10_~ M = 104 = 10, OOOX.
1--o--8~~,~
Such ~ concentrating of liga~d in the predetermined region is one reason that allows the methods and apparatus of the present invention to achieve considerable sensitivity without re~uiring a label, ~nd even greater sensitivity with a label.
P~efe~rrin~ again to Fig. 5A, there is shown a ræservoir 10:2 filled with an el~ctrolyte solution 100. Two channels 104 and 106 respectively provide paths by which the electrolyte 501ution can flow from the reservoir, through th~ test and control predeterminæd regions lOB and 110, ~hrough t~o exit channels 1 lZ
and 114, and eventually to a waste container tnot sho~n). Th~
~luid flow is illustrated by arrows 120. The fl~w thrvugh channels 104 and 106 may be driven by gravity, pnæu~atic or hydraulic pressure, or other means. If desir~d, the reservoir may contain a magnetic stirring bar or other means to help achieve h~mogeneous compo~ition and temperat~re of the ælectr~lyte. The m~teria~ fr~m which the reservoir and channe~s are fa~hioned is non-con~uctive~

-~6-Mea~uring ~eans ~re provided by rece~sed current and volta9e elec~rodes using the zero-current~ four-electrod~ method noted earlier~ A first current electrode 116 in channel 112 and a second ~urrent electrode 118 in channel 114 are respectively connected to terr~inals C1 and C~ A signal generat.~r i5 conn~cted to terrninals C1 and C~ which causes a current to flow ~rom current electrode 116 through channel 112, test p~edet~rmined r~gion 1089 channel 104, reservoir 10-, channel 106, control predeterrnined region 110~ and channel 114, to elæctrode 118. This cLIrrent flo~
causes a v41tage drop across test and control predeter~ined regions 108 and 110 whi._h can be r~asured to deter~.ine the presence of a particular ligand in th~ ~lectrolyte~
Two voltage elcctrodes l22 and 1~ are connected to terrninals Vl and V2 and contact thæ electrolytæ solution on either side of test predetermined region 108. 8y me~suring the volta~* betweæn termin~ls Vl and V2~ t~e conductance o~ predetern7ined region 108 may be measured. Similarly ~oltage *lectrodes 1~6 and 1~8 are c~nnected t~ tern~inals V3 and V4 t~ provide a r~eans for mea~uring the voltage drop across control prædetern~ined region 110.
Voltage electrodes 1 2 124~ 1~6 and 1--B are typic~tlly recessed from channels 10~ 106 112 and 114 throuyh which ttle electrolyte ~nd t~l~ electric current flow. Channels 130 itl Fig.
5A repre~ent this recession.
R~cession reduces errors which are caused by the effects o~
changing properties of the electrod~-elæctrolyte interfact9 5u~h as polarization impedan(:e Pola~izaltion irnp~danc~ may be ~2~
represented as a variable calpacitanf::e and resi~;ta~ce associated with the inte7~face. V~ltag*s developed acro~s this complicateclv and of ll;en ~changing~ impedance car~ result in si~ni ficant errors in the measured valuæ of the po~ent;ial ~t the electrode~ If the ~l~ctrod~ i5 ræmoved ~ine~, recessed~ frorl the electrolyte volume in which all or mGSt of the current flows, little or no current flows near or through this interface. There is ~hus little or no voltage drc,p across the i~tærfacæ7 and ~ence little or no eff~ct of chan~ing propertie~ of the inte~fa~ on the voltage rneasured by the elæ~tr~d~.
While not wishing to be bound by theory, the following model will b~ helpful in understanding how v.jltage electrodes 1~ and 12~ measure ~he conductance of test predetermined region 108 a~d voltagæ electrodes 126 and 1.8 measure the conductance of cot~trol p~edetermined region 110.
The v~ltage elect~odes 1~2, 124, 1~, 128 sen~e a potential approxi~atæly equal to that at the mouths of their corresponditlg recession channæls 130. Thæ potential sensed is an aværage of the potential across the mo~h of the recessio~ channel. Typic~lly, this p~te~tial may be thought o~ as that experienced by a point half way across the mouth called the voltage equival~ncæ point;
that i5 the electrode and the voltage equival~nce point lie on the ~am~ æquipotential. Points 131, 133~ 135 and 137 repres*nt the voltag~ equivalænce point~. Point 1~1 æxperiences the sam~
potenkial ~5 ~lectrode 1.47 pOillt 1~3 the sarne potential as electrode 1~2y point 135 the sa~e pote~tial a~ electrode 126, and : --~8--~29~
point 137 the s~m~ p~tential as elec~rode 128. The corresponding typical equipotentials ~re i~dicated by th~ dotted lines 125, 123 127 and 1~9, respectively~ Within the vol~me b2tween eguipote~tial~ 1~3 and 125, a local conduct~nc~ ~han~e typically af~ects the voltag~ drop measured by voltage ælectrod~s ~ and 1~ strongly~ while far outside this volurne9 such changes tend to have little or no effect. Thus, equipotentials 1~3 and 1~J help to loc~te test volurnæ 10~, tlle vol~lme whose conducta~ce i9 IlleaSUred tG deterrnine thæ o~curan~e of ligand/antiligand interaction in predetern~ined region 108. Sirnilarly the volur,~e betwæen ~quipotèntials 127 and 12g helps to locate control volume III. As liga~d binds to antiligarld immobili~ed in prede~rrfirled region 108 and ch~ng~s the conduct~nce of this region, the . .
con~uctanc~ of test volume 109 al50 changes si~cæ test volurf,~ 10 ~t least partly inclufies predetermined region 108.
~ urrent elæctrodes 116 and 118 are typically recessed dowtl channels 112 and 114 from reces~ion channels 130 that determi~e the positions of test a~d control volum~s 109 and 111. Such recession reduces errors which are caused by changing current distribution resulting from changes at the ~urface~i of c~lrrænt e~ectrodes 116 and 11~3, since it allows the current density to become unifoYm or nearly unif~rm by the time it reaches the test and control volun~es, irrespective of local variations in cllrrent density at the current electrode surYac~

--~4g--, ~2~
For effective recession of both curr~nt elec~rodes 2nd voltage electrodes, the electrode is preferably re~essed three ~r more time5 the width of the mouth of the rece~sion channel, The aboYe or ~quivalent steps potentially give the m~asuriny apparatus high accuracy. ~ith~t these steps, ~n accuracy of 1 part in IO,OOO i5 difficult or impossible to 3chieve.
Th~t part of the ~pparatus correspondirl~ to a sp~-ci fic test or contrcl vcIume i5 called a "cell". For exar,-ple, tile test cell i Fi~ 5A .-ornprises pr~deterrnined r~gic~r~ 1087 c~ntacti~g chantlels 104 ~nd 112, voltage electrodes 12~ and i~ with their respective recession channels 130, and the corresponding test volume 10~.
The contr~1 cell ~orfprises predetern~ d region 110, contacting chan~1eIs 106 and 11~, voltage electrodes 1~6 a~ld 1~8 with their respe~tive recession ch~nnels 1~0, and the corresponding control volu~e 111. Currænt electrodes 116 and lla and reseYvoir 10~ are common components needed for operation of both the te5t arld co~troI cells in this ~mbodi~ent.
Option l flow measurement appar~tus ~ay bæ provided t~ measure the flow of the elæctrc.lyte~ In Fi~ SA, two hot wire flow 5en50r5 132 a~d ~4 are shown in the exit channels 112 and 114.
Circuitry 1~6 passes a current flow through the sensors and measures the heat conducted away fro~ the sensors by the fluid flow to provide a ~easurement of the flow rates through test volu~e 109 and c~trol volume 111. Other means ~ay also be used t~ determit~e flow r~te. For e~ample~ flow mæters or fluid resist~rs t33 and 135 m~y be pla~d in e~it chan~el~ 11~ a~d 114 t~ re~lat~ fluid flow.
: -50-.... .

Fig. 5B is a simplified electrical model repre~entative cf so~7e ~f ~he the el~ctric~l characteristics of the ~rr~nge~ent of Fig~ 5A~ In Fig. 5~7 Rx is ~ resistance repr~senting the resistance of test volume 109. ~Either ræsistance or conductance may bæ rl~e~sured as con~enient, since each is t~le r~ciprocal of the other.~ Similarly, ~c represents the resistance of control volume 111. Test resistance Rx includes not only the resistance of test predetermined region 108 but al50 the resistance associat~d wit~7 those parts of inl~t ct1anr7el 104 and outlet channel 11~ that lie betweæn the test region ~08 and recession chat1nels 130. Si~ilarly control resistance Rc includæs the resistance ass~ciated with the æntir~ control voluniæ 111.
Clearly~ the most efficietlt sensor of this type will be one in which the resistanc~ of predetermined region 10~ is as lars~ a fraction as possible of the resistance of the entire test volume 10~) i.æ., where thæ efficiency E, d~fined as E = ~esistance_of_~r_determined re~ion 108 Resistance of test volume 109 approaches 100~..
Resi5tor5 R~V R3, R5, and R6 represent thæ resistance of the electrolyte solution in recession channels 130. ~esistors ~1 and ~7 represent the resistance o~ the electrolyte solution between current electrode~ 116 ~nd 11~ ~nd test and control volumes 109 and 111. Resistor ~4 represents the resistance between test and control volumes 10~ and 111~ which includes the resistance of the electrolyte in th~ reservoir.

`: --51-It ~hould be appreciated ~h~t ~he model shown in Fiy. 5B is only a 5ill~pli ~ied rnodel, CFc~r exarnple~ the polariz~tion irnpedance associa~ed with ~ach electrode has bee~l ~n~itted. ~f ~iesired, the polarization in~pedance~i may be indic~ted explicitly by pu~ing an impe~ance in seYies with each of resistarlces ~ 7.~ However~
this model can be helpful in underst~nding thæ electronic circuitry which rneasures the conduc~ance Cor resistance) ch~nges in test and control volumes 109 and 1119 a preferred embodirnent of whi~5l is des~ribed in Exarnple 8.
As noted above, the presence of a particular li~and in the electrolyte causes the conductancæ of the test volume to chan~æ.
To briefly describe the process by which this is done, test ~rerieter~lined region 108 contains a matrix which includes im~obilized rec~pt~rs or antiligands for detec t i ng th~ presence of selected ligand~ These antiligands bind the ligand rdolecules within predeter~ined region 108 as electrolyte flo~s through the region. While not wishing to be bound by theory it is believed, that as the amount of liga~d bound within test predeter~ined region ~08 increasæs, there is a corresponding decrease in the available v~lun~e through which elæctrical ~urrent may flow9 r~sulting in a decrease in the conductance of test predætermined region 108 and hence of the entire test volu~e 10~. Equivalently, there is an increase in the resistance Rx bætween terminal~ Vl ~nd V2.

-5~-., .

.. . ~.

Chan3es in the conducltance of test volun~e 109 c~n be used in det~rmininy the presence of a ligand in a fluid sample ~s describL?d in the rni?thods in rnore detail. Typically, a flow of pure elec~r41yte solution i5 est~blished through the ~st cell until a ~i~eady stat~ condition i5 reachQdr and the conductane is measured. Preferably, the conductivity ~f the electrolyte is n~atched to thalt of 'che fluid sarnple to be exab~ined for præsence of ligand. Ne~t, the fluid sample i5 added to the ~lectrolyte in the reservoir, and thæ chang~ in conciu.tatlcæ ~Jith tirne of test volun~æ
109 is monitored, as is the flo~ rate~. The observed change in conductance is cornpared with a standard curve 5u:::h as $hat of Fig.
lB ~o determine the conce~ntratibn of liyand in the fIuid sample.
The c ontac t i ng means may i nc 1 ud e a samp 1 e 1 Obp f or salr.pl æ
fluid injec~ion inste~ad of a reservoir. In this case a blank solution, or running buffær, of conductivity similar to that of the fluid s~mple i5 initially run through the sample loop, inIet channels and predetermined regions via a fluid inlet port. The using commercially available switching valv~s in coramon use ~in liquid chromatography, for example, available ~ro~ ~ainin I~struments7 Wob~rn, MA), the fluid sa~ple in~tead ~f the blank s~luti~n i~ injected into th~ fluid inlet port. Ligand-specific ch~nge~ in conductance can then be monitored either hy noting the rat~ 13 ~t which conductance change~i occur or the n~agnitude 14 of the total change in conduct~nce ~ter the ~luid sample has passed e~tirely through thæ sensor~ (S~e Fig. ~.) Use o~ ~lvæs to r~lat~ e~try Or fl~id sample has a~ addition~l potenti~l -5~-advantage: such ~alve t~chnolo~y is highly de~410ped and capable of b~ing ~utomated un~er co~puter controlO Thus the schedule of s~nsor expos~re to a fluid sample or set of fluid sampl~s can be controlled a~to~tically.
Many factors nlay affect the resistance between termin21s V1 ~nd V~v in addition to the change in condu~tance of the material in test predet~rmine{i reyion 108. The fluid sample itsel f may have a conductivity di ff~rent from that c,f the running buffer, though pref~rably its conductivity is adJusted tc be similar to that of tS1e running bu~fer. Small changes in ternp~rature, composition changes due to evaporation~ non-specific binding c~f fluid components to the matrix, etc., can also affec~ the c onduc t anc e ~ f t he m,at r i x or t he c onduc t ~nc e o f t h e t est vol urne as a whol e~
A second cell, the control cell shown in Fig. 5A is7 provided to help correct for these errors. Preferably, its control predet2r~ined region 110v voltage el~ctrodes 1~6 ~nd 1-~ and as~ocia~ed chann~ls 130, and inlet and outlet channels 106 and 11 are constructed identically to thos~ of the test cell. Control p~edetermined region ~10 includes matrix ~aterial similar to that in test region 108 ~xcept that the control cell matrix does n~t contain antiligand rn~terial anci is thus ~iubstantially fræ~ of li~and--specific interactions which effect ch~nges iM bulk conducltance., By m~asuring the conductance of test volume 1 relative to ~he! conductance of control vol~me 111~ it i~; possible ~o reJuce oSr eliminat~ variation in conduc~nce fr~n3 ef ~ects other "' , ~2~
. .
than specific bindin~ of lig~nd ~o antiligand site~3 in test predeterminetl re~ion 108 In Figs. SA and 5B~ the test and control ce11s are electrically in ~eri~s via a single current path7 thuugh they are hydraulically in a kind of partial parallel, ~ach cell having its own fluid path. However~ other el~ctrical configurations can be used to create and the~ rheasure a voltagæ drop across the test and control cells~ For exarnple, in Fig. 5C~ each cell has i*s own current path, wit~l the two cells sharirl~ a cornmon current electrode 144. In Fig. SD, each cell ~las its own current path, and additionally~ each cell has its own pair of current electrodes.
Referring tG Fi~. 5C, a sagn~l generator is connected to tær~inals C1 o~ electrode 140 and C~ of e1ectrode 144 that causes a curr~nt to flow through the t~st cell, ~rom current electrode 140 through channel 112~ test predetermined region 108 and tæs~
volu~e 10~, channel 104 and r~se~voir 102 t~ current electrode 144 in the reservboir~ Similarly a signal generator is connected to tern~inals C1 of electrodæ 142 and C~ of ælectrode 1~4 that c~uses a current to flow through the control cell, from electrode 142 through channel 11~, contro1 predeter~ined region 110 and contr41 volume 111, channel 106, and reservoir 102, then also to el~ctrode 144. Electrode 1~4 is thus a co~mon electrode tc the two cells.
Voltage electrodes 12X, 1~4, 1:;~6 and 128 and voltage electrode recession channels 1~0 function as in l~ig. cp~ to me~sure thæ
volta~ drop, a~d hem:e the resistancel, a~ro~s test volumæ

10~ and control v~lume 111 ~en~rated by the current ~Iow.
Cir~uitry whi~h rneasures th~ resistanc~r chatlg~s in th~ test and contYol vcllum~s of thæ apparatus in Fig~ 5C i5 describ~d in E.~ampl~
~ æferrirlg t,~ Fig. 5D, inc,tead of there being a currætlt elæ.tr~de in a reservoir cornm.,n to ti~e t~st and contr~l cellsv ~a--h c~ll has a .-urrænt eIæ.trod~ in its irl1et channæ1 ~4r in a r~,æssi"tl ,hannel that is in electri~al ,-onta,t with its inlæt .~hanne1~ and a ,-urrænt elæ~~trodæ in its outlet ,~lannel ~or irl a re,-rssi,-,n ci~annel that is in æle,-tri,al ._~,nta,~t with its ,-~utlet challnel~. Sp~cifi,-ally, in Fig. 5D the test cæll has a ._urrænt ~le.tr~~~dæ 1~8 in elæ~tri.al .,~nta,t with inl~t ,hannel :lO~ via currætlt ælectr,~de r~cæssi~n channel 131 and a .-urrænt æIæctr~,de 1~ in ælæ,-tri,al cnntact with outl*t .-hantlæl 11' via its .-,~rresp,~nding ~urrent ælæ,_trodæ recessi~n chann~l 131. Sirr,ilarly, thr ~~o~ltrc~l ,-æll has a curren~ elæctrodæ 1J-' in ele,_tri,_al c~ntact with inlrt ~hann~l lOG and a .urrænt ~lectr~~~de 150 in ælæ,_trical .-.~ntact with outl~t ~hannæl 11~, alsc, via ræcæssion ,ha~næls 131.
A signal gentrator connæ,_ted to tærrr-inals ~ f ælæ~trod~ 146 and C~ of elæctrcdæ 148 causæs a currænt to flow from ælæctr~dæ 146 thr.-,uyh .hanrlel 11 , t~st prædætærminæd region 108 and tæst volur,-læ
10~, channel 1~, to currænt elæctrodæ 148. Similarly, a signal g~nærator connæctæd tc. tærrr,inals 81 c~f ~læctrodæ 150 and C7 of æl~--tr.-.dæ 15~ caus~s a current to flow frorn æl~ctrod~ 150 through channæl 11~, ~~ontrol prædætærmirled region 110 and ,-ontrol vc,lumæ
111~ ch~nnæl 106~ to current electrode 15~. ~oltagæ æl~ctrodes .

, l26 and 1:28 and voltag~ ?ctrodæ rec~ssio~1 chantl~ls 1:30 functi~n as irl Fig~ 5A to rn~asure th* voltarg~ drop a, r~ss tæst volumæ 103 and contr-~l volun~ 3æn~2rat~d by the .~urrent fl-~w.
A ~omr"on fluid sourcæ inl~t 153 conn~cts t~ inlet chatlnels 10~' and lOG via c onne.-ting ~-hatlt7els 105 and 107 instead of via a cornmon r~servoir. Inlet 153 its*1 f typi. ally cQnne~ts tc~ a samplæ
lc,~p c on tai tl i ng a k n own ari":~urlt of f1uid sar"plæ.
Cir,-uitry whi.-h r"easures thë ræsistanc æ hang.-s in th~ t*st and .-..ntr..l v.~.lurnæs ..f the apparatus in Fi~. 5D is dæs. rib~d in E~,;ar,ipl e g.
Referring n,.w t.-. Fig. ~, th~rë is shl-wn a dia~1ramrllati represæntatiun ..f an.-.ther apparatus useli in det~rrrlir1ing th~
pr es~n. i~ o f a l i gand i n a f l ui d sarnp l ~, i n wh i C h t ~e pr *senc e l~ f several differ~nt ligands can b~ deterr"in~d at -.n.~æ. Fig. ~A
illustrat~s an arrarlger~ænt with multiple ~-ells in whi~_h each c~
ha~ its own ele~-tric cur rent and fluid path in th~ r"ann~r i 1 1 ust r at ed i ~l Fi 9 . 5D . Fi g . 6E~ i 1 1 ust r at *s an ar r angem~ nt wi t h multipl~ cælls in whi.-h ells are . ontlæ-~ted elæ.-tri.-ally and hydraulically i~l series, all ~ells shariny a .-.-,rr,rr"-.~l æle- trical and ~1 ui d pat h .
~ æferri~lg to Fig. 6A, each of ælls 1J~ 15~ 158 and l6~
pr~-færably in. ludæs a l...-ali~in~ means .-.-.n~prising a r"atri~; with antiliyarld immcbili~ed on it in th~ predetêrrninæd r~gi.Jn 16~, 154, 1~58 or 16~-J resp*,-tivæly. One ._~ æ.g. .-ell 15~, typically in~Iudes at l~ast part .~f a pr~d~termined re~i :.n 166 tllat dc.es not conta~n antiligand kn~wn tcl react sp~cifically with any component f ~ 9 f , .
of the fluid sarnple to serve as a negativ~ control. Measuring n~ns for ea~ ell compris~ tw~ rec~ss~d c~rr~t ~l~ctrodes ~onnected to t~rminals ~1 and C~ and tw~ r~cessæd v~ltage el~,-tr~d~s c~nn~ctæd to terrninals V1 and ~, in tt7~ manner illustrated in Fig. 5D. The v~ltage drc~p acr~ss the t~st vl~lurrle of any c~ll i5 ccmpared with the v,~ltage clr~p a~-r~ss that of control cell 1J~ ~r any other cell by c~nr\e~tit-g a sigrlal generat"r t,-, th~ urr~nt ele~trl~des ,~f thæ two desirecd c~lls t"
gætlerate arl ~l~:,-tri, c~rretlt that passLs through the cells~
Swit~hing m~an5 ar~ typi,ally pr"vid~d to all,-,w rapid cc,rhparis~r ~f diff~rerlt pairs ,~ ~lls. Fr~d~t~rmined regi,-,n 1~ f test .-ell 1~9 for e~;ar~pl~, r"ight typi,-ally ,_ontain antiligand that binds to a first ligand of int~r~st in the fluid sarnpl~, while pr~dct~rr"in~d r~ lns 164 and 168 ,~f t~st ~^lls 1J4 and 15a rr,ight c,-ntain se l~nd and third ir"rr,~bili~d antiliyatlds that bind to se,-ond and third ligands ,-f interc~st respæ,tively in t~ fluid sample. Arl-ther pred~termined rægi,"l, e.g. regi"n 17l~ f -ell 160, rr,ay ,_ontain ir~,rrlc~bili2ed arltiligand that bit-ds t,- a ligand -,f knc,wn cc,nc~ntration in th~ fluid sar,lple and rr,ay s~rvæ as a positive control. An arrangem~tlt such as that of Fig. 6A allcws simultan.?~us d~t~ctil~tl .~f multipl~ liga~lds in a fluid sar"pl~ an~
simultan~4us ._~rrlparison ~f t~st c~l ls with n~gative cc.ntrols, p~sitiv~ contr41s, ._lr eac~ oth~r.
In a related typ~ cf apparatL~s, ~ach c~ll has its clwn fluid path in th~ r,la~n~r l~f Fig. 6A, but it shar~ a comrrl-~n c~ ctrlcal pat~7 Witll th;~ c~ll with which it is being comparæd, in 1;he n,antlr~r ~f Figs. 5A and B. T~at is~ the two cells that arr being ~mparæd aræ placed in series wit~- ea~h ~th~r~ as in Fig~ 5B~ T~1is is ac~omplish~d by -~nne~ting a signal gen~rator to ~he ~1 current ~le~trode terr,~it~als of the tw,~ cells to g~n~rate a current that pass~s thr~ug~ the ~ells.
~ f~rrin~ to Fi~. 6~, ther* is shown an apparatus in whi-h a series of ~ell 5 spe~_ ifi~ for different ligat-ds of inter4st ar~
~orln~lted b~th el~,-tri~ally and ~ydrauli~~ally irl s~ri*5.
Typi~ally, ea_h ~-ell in~lud~-s a matrix with antiligatld imr"~bili-~d on it in a pr~deterr"ined re~i,_ln ~C~C1~ ?1~, ?c)&~ or ~(38, exc--pt again f,r a nr9ative ~ontrol .-11 w~ se pr~d~t~rr"ined r~gi~n _0~
dlles n~-~t ,-,~ntain antiligand kn~,wn t,, real~t spe~~ifi~ally witl~ arly ~or"ponent of the fluid. A ~nrnmon inl~t ~hantlel 175 and outl~t hannel 17~ tl~ all -elIs pr,-,vid~ pat~s by w~i,-h el~,trolyt~
s~luti~n 101~ -an flow frorr, or"rr",n reservoir 1~ ', thr~-~ug~ t~ lls and t~leir predetermin~d r~gions ~00, ~ &, ?C)~, and ~C~ and their resp~ tive test v~lur,les 1'3C)~ 13_', 13& ~r ,ntr~l v lur,es 1~, 198 to a waste ~~-,ntainer (not sllown). ~easuring mrans f,,r ea,-h ,ell are pr~,vided by two re.:essed vllta~e electr~des ~n either side of each predæt~rri,irled regiotl~ while all ~ells share tw,_ curre~lt ~l~~trcdes 17~ and 174. Cornmon first ,-urrent ælectr,-,de 17 in or outside ,hannel 17~ a~ld c"rnrm~tl s~cl~ltld ,~urrent *lectr"de 17~ in ~r outside ,hannel 17~ are respectively con~l~ct~d to tærminals C1 and C~~. A sigrlal generatcr is ,-ontlel-ted to terr"inals C1 atld C~ whici~ ,-auses ,urrent t~ fl w fr-"b ,-urrent electYode 172 itltO outlet c~la~lnel 173, through predet~rmitled _5~_ ,~

~ ~ a ~ ~f~.~
. ~ t~ ~ 7 re~gif~7tlS .00~ 20:2, 20~, ~0~ and ~08 intc~ in1et ~hannel 1757 lto current ~l~?ctrc,de 17~. This ~urretlt flcw ~-aus*s v,~ltaye drops ~cross t~st v~lurr,es lgO, 1~, lg6 and co~7tr~1 vc~lur~æs 194, 1~8 whic~ ca~ b~ n~æasured to detrrn7i~le th~ pres~tl,:* of ligands 4f intæræst in thr rl~{tr~ ytæ~ C:urrrnt æl~.~tr~~.dæ c hanlb~-r 171 n,ay br pr,~vidæd t.~ lowær tllæ p.~larizati4n in,pæd=7n.-e- .~f current æle.-tr.~dæ
17 ~.
T~læ ab.-,væ arrangæm~-lt pr.,vid~s f.~r the sarn~ elcctrical . urrænt a~d t~æ san,~ fluid . urr~77t t.~ f1 UW thr..ug7h all .-æ1ls. This . an h~lp sirf,pli fy th~ dæsigr1 .~f th~ ælæ, tri, al instrLIn~ tatibn and ,~f the fluid fl,~w n,~-asuri~1c,7 apparatus. (.It~ fa, t, sinl-r sic~t~al ~:læv~-l.-,pn,ænt will start in difftræt1t sætls,-,rs at diffrrrnt tir"rs, .~bs~rvati.~,n .~f su.-h tirnit7c~ di f f~ræ~1ces c an its~l f bæ used t..
n,~asur~ fl,~w ratæ. ~ H,-.w~vær, th t særiæs appr,.a. ~1 typi, a1ly itl. ræas-:-s th~ t.-,tal pr~ssur~ tl~edæd tc. . aus~:- fluid t" flc.w at a c3ivæn ratæ. Whi._~l cc.~lsideratil_"ls prædc~rnin2~tæ d~p~ds c~n th~-c..tlditi..~ls ræquir~d f.-.r a particular e~;p~riri~ætlt.
V,-.ltag~ .-tr.-.d~-s 17~, 178, 180, 18~, 184 at~d 18~ arr co~lnrcttd t,~ t~rrnitl~ls Vl-V6 rrspæl-tively a~d cc.~ltact t~l~
*Irctr.~llyt~ .-.n ~ither side .~f pr~cdet~rrfli7l~d r~c~ic~ns, C)0, ~O~y '04, ~ 6 and _C)8. 13y rn~dsuri~lg th~ v..ltagæ b~tw~-~n any tw,. ad.ja.-ænt voltag~ æle~tr.~d~ tærminals Vt~ and V~l+l thæ, ~lducta~lcæ ~f the pred~t~^rn,in*d regi,_.n and the tæst or cc~tltrc.l v..lur~le it~cluded b~tw~en thæ . .~rr~sp..nd1ng v..ltaq~ elæctrrd~s rnay b~ m~asl.lred~ As in pr~vic.LIs arratlg~^rfl~nts, v.~ltag~ ctr"d~s i76~ i78~ 180, 18'`, 184 and 1B6 ar~ typical ly r~cess~d fr.~rn c ~la~n~ls 173 anci 175 whær~r ('-J ~:
2~3 electrolyte and el~:tric curr~nt fl4w via vol tag~ el ~ctrode r~cæssi~n ~han~lels 188. The midpoints of th~ m~uths of adjacent re.~~ssion ~hatln*ls 1~8 typically d~finæ thæ p~siti4tls of tæst Y~lLlrfle5 l'~t3~ Ig2~ 1~6 and control volurf,~s l~g 1~8 includ~d b~tweetl adja.ent r~--æssi.tl .hantl~ls. Also7 currænt ele~-tr des 172 and 17~ ar~ typicaIly ræ~ssed dl~wn ~hannels 17~ atld ~75 fr-~rf the t~st volur~ s.
~ s with the apparatus of Fig. 6A, thæ predætf~rrbirlæd ræt3i_tls _'t~(3~ ~'0-~', ~'n~ 16~ -'t36 f the apparatus in Fib. ~B typi-ally c~tltain antiliyands to s~veral differtnt ligands of interest in a fluid san,pl~, as well as positive and negative .--~ntr~ls.
Gf.~ups .:f .~ells it) ~eri~s, of th~ typ~ sh.~wn itl Fig. 6B, can th~mselYes b~ pla, ~d 50 that ~a.:h.gr.~up ..f ceils has its own ~l~^ctri. .~urrent and fl~lid path in the ma~ er illustr~ted in Fig.
6A t.~ in reas~ th~ ~værall nur"ber 4-F test volurfles that can be mot~it_r~d sirhLIltane~usly.
An arranl3erf,ænt clf .ælls c4nnæ.~t~d in s~riæs can be r"ad~
thinnær tha~l that sh.-wtl in Fig. 6~ by c.~nne.ting in s~-riæs ~hydrauli.ally as w~ll as ælæctrically:) s~veral cælls rm difieri fr.~rn the type sh.~wn itl Fig. 5A, wher~ th~ reservoir is replaced by a .:hannel, and thP fluiri ..utlæt ~f orlæ c~ll contæcts tl a fluid inlet of thæ næ~t. This allows ~ach vclta~3~ ctrod~ tc. b~ -.n ~n~ of tw~ v~ls~ n~ matter hcw rr,any c~lls thæræ ar~ in thæ
s~ries, with fluid and ~læ-tric -urr~nt wæavi~lg back: and f._.rth b~twæen th~ tw~ lævels.

,................................................................... I

2~

Referring now to Figs. 7A and 7B, there is shown yet another appara-tus used in determining the presence of a ligand in a fluid sample by measuring the change in bulk conductance of a test volume as ligand interacts with antiligand localized in a predetermined region that is at least partly within the test volume~ In this embodiment, the measuring means lie in a plane, on the same side of the localizing means. In Fig. 7A, there is shown an arrange-ment where the localizing and measuring means lie in the bottom of a well that contains electrolyte. In Fig. 7B
there is shown a similar arrangement, except that the local-izing and measuring means are mounted on a support fixture that dips into a well, test tube, or other space that contains electrolyte. The arrangement of Fig. 7A allows easy injection of a fluid sample onto sensors. The arrange-ment of Fig. 7B allows easy insertion of sensors into a fluid sample.

Referring to Fig. 7A, there is shown (not to scale) a well 215 constructed from non-conducting material filled with electrolyte 210. Test predetermined region 214 and control predetermined region 21S lie on the bottom of the well. Preferably, test predetermined region 214 contains localizing means comprising a matrix that has immobilized on it antiligand which interacts with the ligand of interest or localizing means comprising a semipermeable membrane which encloses such antiligand. Control predetermined region 216 is constructed similarly to test predetermined region 214 but does not contain such antiligand and is substantially free of specific interactions which cause changes 2~ .

in bulk conductan,_e~. Ligand ent~rs predetermined regi~ns 214 anri 216 by di ffusi~n. Stirring pr~p~ r 21~ r ~ther m~a~s ~f causi~r~ bulk fluid flclw i5 uged if desired tc aid c~tlta, t bæ~ween thæ fluid and the pred~t~rmined r~gions.
Measuring rne2ns f,~r 1;~ test c~ll are pr~vid~:d by curret-t ~lectr~ s ~c18 and 2~0 c,~nnæ-t~d t~ t~rr~linals Cl atld ~- and v~ltage electr~d~s ~2~ and 228 c~nne}:ted t~, terrllit~als Vl and V~.
Measuri~g rr,ea~7s f,~r the ~ntr ~l -el l are pr ~vided by curretlt ele, tr~-~d*s ~_ and -~4 cclr~nected t,~ terr"itlals Cl and 1_2 and v,~ltagæ ~l~ctr-d*s ~3l~ and ~23~ nnel ted t,, terrninals vl and v!.
Th~ r"easuring rn~ans lie in the t~,~tt,~,rr ~ f well -1_J ir-n~tl-c~ndul:tirlg ri,at~-rial _3~, b*n~atrl pr~d~t~rnlitl~d regi},ns ~ 1~ and --clG ~
Tc rn~asur~ the, ~ ~t-ductanc~ ,- f th~ test c~l l, a si g~lal gen~rator is c~nn~ t~d t- terr"inals l_l and C~ c f ~ ctrod~s -18 and C-'07 ,_ausi~g a current t~ fl,~w frc,rb ~l~,~tr~de 21~1 thr,~ugh test predeterrr,ined regil_n -14, int~ ele~ tr~,lyt~ 210, ba k thr-,ug~
pr*determined region ~14, t,- electrocle ~ Such a ~urrent path i5 illustrated by the dc~tted line ansi arr,-,ws :37. Also, s-~me curr~nt r~r~ains entirely within predeterrr,ined regic~n ~1~. SU~h a current path is illLlstrated by the d ltted line and arrow '~3. The current flow i5 correlated with a voltage drcp alon9 the ,:urrent paths. Since at least part ~f æa~h :urrent path passes throu~3h test predeterrr,ined re~ic,n ~14, at least part ,-f the vnltage drop will be inflLl~n,-ed by the ~~o~ductan~~æ -,f this r*gi,~n. E~y rne~su~iny the vo1tage betwæa~n tr rminals V1 and V~ connected to ~ f ~IL2~9~26 electrod~s 226 and 28~ the conductance of test predet~Srr~it-ed region 214 rnay be measured.
Sirnilarly to measure the conductanc* c,f the contfol cell~ a signal generator is contlect~d to terminal; C1 and C~ of electrod~s 22~ and ~4~ causing a currænt t4 flow froril electrc,de ~ , through cor~trol predeterrnined r~gion 216, it-to elæctrolyte 210~ ba k through region ~16, to elæctrod~ , as illustrated by the dotted line and arrows 241. Alsc~, sorne curretlt rernains entirely within c~ntrol lo,-ali~ing r"eans 1~, as illustratæd by the dotted line a~ld arrc,w ~4~. The current flow is ,-orrelat~d with a voltag~ dr"p alor1g the curretlt paths. By r~,easuritlg the voltage drop betw~et terr"inals V1 and V -onne_t~d to ele,-tr"des ~30 and ~, the conductance of control predeternlined region ~i6 may be n,easured.
Preferably~ as in -~ther er"b~~,dirnents, the voltage ele~trodes at~d current electrc,des ar~ recessed. For exarnpl~, as illustrated in Fig. 7A, th~ ,-trodes are recess~d do~n chann~Sls ~5 a distancæ tilree or more tim~s the width of the ,~hannels.
~ urrent flow is densest in a volurne whose radius is ~f the order of t~le distal~c~ between the current ele~trodes. For a cell with a configuration ~uch as that in Fig. 7A, the test volume may be tllought of as extending out above the bottonl of the welI abo-lt this distance or a little more, i~e., the ~ffe,tive cell si-e r"ay be thl-,ught of ~5 being on t~le order of the distance bætween tlle current ele~trodes~ As a useful illustration, th~ test v~lume may be thougtlt of as bei~lg tt~e volume within the dott~d field line ~37~ and th~ control volun~ may bæ thought ~f as b~ing the volum~

within th~ d~tt~d fiæld lin~ 241; that is~ voltag~ ctr~des ~6, a, ~30 and 2~2 will typically sæns~ wæa~:ly or n-t at all c-ndu-tan-e c~lanses that ~c-_ur outside thi5 r~gi~t7.
Thæ ~ffisiænsy of a planar r~s~ss~d cell~ as F-~r c~lls prævisusly dis.-uss~d, .an bæ dæfin~d as E - ~æslstatlsæ slf_er~detæ__ltæd__ægiut F.~sistan e ~f test v~lumæ
Typisally, if th~ thi~_k:næ5s ~f the pr~dætærminæd r~gio?1 is .-.f thæ
ordæ~ of thæ ~~ell si_~, th~ .-æll will b~ æffisiæt1t. On thæ .-.th~r hand, if ttl~ pr~dætærn,in~d regi~n is thi kær tt1at1 ab..ut ~:i.1 r"m~
di f fusi~n b~---r"~s v~ry sl.~w. Thus, in rdær to ._snstru_t sensllrs that aræ ~ffi ient and that al50 ræsp nd qui.k:ly tsl thæ præsæn.~e ~f ligand, b.-.th thæ t~st v.-.lur"æ and prædet~rrtlitlæd r~gi~ln v.-lurnæ
shl-luld be small~ ~.9., ~ 0.5 mr" thick f~r thæ tæst vcllume and 0.1 rnr, thi-h: f~r thæ prædætærtnin~^d rægi-.n. Typisally thi5 requiræs th~ . ~nstru-ti~t- s f . I.~sæly spased curræt-t and v~~~ltagë
el~-tr..d~s, whi.h can bæ d_næ in a vari~ty .~f ways~ inl-ludi thisk film pr~.-æss~s, thitl film pr.~-~~ss~s and silic~n .hip t~.~hn.~ giæs.
~ hanges in thæ cslnductalll-e c.f t~st v~.lum~ 1~ ar~ .-srnpar~d, æ.g. rati-~rhetri-ally, t.-..-hangæs in th~ .- nducta~i.æ . f s~ntr..1 v~lum~ 216 and thæ ~hanging ~ndustatlc~ ratis. C is used to dætermine the pr~sen. ~ and -~n:æntrati~~~n .~f a lirland in a fluid sample~ Typi._ally, puræ ~læ.~trolyte is inje~ tæd int~. wæll ~15 and stirr~d until thæ ,:4ndu~tan.-æ rati.~ 1~ bæ~..mes steady. Præfærably th~ c;3rlductivity 4f ælectr.~lytæ 210 i5 matcll~d to that ~f the :
~ -&~-.", . .. . .

~ - ( - ~2~ .

fluid s~rnple t~ b~ ~xamined fcr th~ presænc~ ~f ligand. N~xt, th~
flLlid 5arl~ple i5 added t-~ ~lectrolyt~ 210 in well 215, and the chatlgæ in thæ cc,ndu~_tatl~_æ rati4 ~ wit~l tiril~ i5 m~lnit~r~d.
Th~ chang~ in the c4~lducta~l,e ratio ~b~erved during a particular experir"~tlt typi,-ally depends .-.~1 a variety 4f fa,_t,-.rs, su~h as thi~kness ~f t~le predet~rmi~l~d regi-~ns~ affirlity .~f antiligand f4r liga~ld, diffusi4n _4nstant .~f the ligand, s~n,-entrati4~ f ligatld, et.-. Ti-eræf-~re, a stanrJard .-urve f..r the ligand 4f int rest is ,-.-,nstru-ttd urlder k:nc~wtl experir"ental ...nditions. T~e 4hserved .~ha~lye in ~,~~ndu.-ta~lce rati,~, devel ped uniær the same c~lditi-~ls, is . .n,pared with thL standard .urve t.
deterrllin~ the J~l~entrati~l ,-f ligand in the fluid sanlple.
Thæ matrix }.f predet~rrrlin~d rægic,~ls -14 arld ~16 will e~;clude p~rti,les larger than the p,~re si-e 4f the matri~;. This ~ay aid i~l m~asurerr~ent c~f flL~id samples with a Iarg~ burd~n 4f unwanted particles even if su-h parti.}es .-a~ ter test v41un,e ~7 and contr~1 v.~lurhe ~-~1, e.g. if stirri~g ~.:eeps th~ particles h.~m.~gen~usly distributed. PreJeterrninL-d ræ~ -s ~1~ and ~16 .-a~
therrls~lves b~ rnade tc. exclude parti.-l~s fr.~rn v..lumes ~7 and ~4i if they are mu.h tnic~er than t~lese v41urlæs. Alt~rnatively, a layer :3r layers .~an b~ laid .-.ver pr~deterrnin~d regi..ns 14 and 21~;. If this layer i~ imperrleable t.~ parti. les pres~rlt in th~
~le. tr.~lyte, and i f th~ combined thick:ness .~f the predetern-ined regio~l and this layer is ~reater than that of the t~st and co~tr41 v.~lurn~s~ th~l t~is lay~r wiII e~.-lud~ parti~l-?s fron~ t~le test volume5 as noted earlier~ T~e above devices and variati~n5 on --6~--thærr~ can bæ us~ful to allow n~asurerr.ent of ligand presence it7 fluids c~ntaining particlæs that would oth-rwise cause noise in tll^ measur~n,ænt~ Planar recessed cells 5u.-h as those in t~le apparatus of Fig. 7A can be used to rr,easuræ the specific bindit7g of parti,l~ and c,~n,plæ.~,;es to thr surfacæ tor interi,-,r~ of a prædetern,inrd region ~ . In this 5i tuati~n, the te5t V41Urne i5 dælibæratæly kæpt larger t~lan th~ thicknæss of prc-dæt~-rmined ræ9il_l~ 50 as to inclLIdæ thæ surface of the rægi,-,~l. The tæst volurr,æ thæn e~tænds int,~ th- spac~ above th~ surfa.e of the rægi,-"l. As particlæfligand corl,plæ~,;es bind to antiliga~;d c"7 the surfac~ of predeterrnind region ~14, they oc,upy part of test v,~lurn~ ~37 ab,~,ve thr surfacæ, thus dæcræasing thæ ,-,Jndu,-tance of the test v,~lurr,-. A sir"ilar arrangærnætlt i5 usëd t.o rr,easure the spæ,ific binditlg of ,~æll5, væsiclæs, rrlerllbrane fragr"ænts .-,r .~.thær structuræs that carry particular ligands "n thæir surfa.-.-s and hen,* to dæterrr~itle thrir præsæ~l,_æ and con,~entratio~l in a fluid sarr,plæ.
~ In crder t,. increase the efficiæncy of a ræcæssæd planar sensc.r whos~ test v-~.lur"e is rr,uctl larger than its prædætern,inæd ræg7ic~l~ an c,pp-~sit7g wall of non-conducting rnaterial is brought near the surface of the pred~tærn,ined regic.n tc.-onfitlæ thæ
electric field, thus reducing thæ test v._.lurhæ. Alternativæly~ a rnc.dulating lay~r, perrneablæ to ligand but whosæ .-onductivity i5 lower thatl that ..f thæ prædætern,itlæd rægion, su~h as porous pclycarb~rlate filtær fr.-.m Nuclæpore (Plæasa~lt-~tl, i-A~, rnay bæ laid ove~ the præd~t~rrnin~d rer3ion. T~lis will tend to co~fit~e th~

~ ( ælæctric fi~ld 5~ t~lat thæ current densi~y in th~t part of thæ
predet~rmined r~9i4n n*-~;t tc the n~dulatin~ layer will be inrreas~d~ The vcltage drop o~rurritlg in thi~ part of the predetærrnitled regi--n will al50 increaset thus inrreasing the ~fficienry c.f the sens-~r.
Feferring to Fiy. 7B~ th~r~ is shown (not to scalæ~ a t~st tube or ~ther fluid-.ontaining spa-e 2~ that holds ~-le._tr~lyte ~lC). The sens-~rs in this arrangement ar~- forrr,ally ideMti-_al to th.-..se of Fig. 7A, a5 shc,wn by identi.-al nur"berins of .-orresp-.ndi~g el~r"~ts; .~nly th~ .~.~.nta--ting r"æans differ.
L.-calizitlg r"ea~7s .-.r,~prising ~7 rr,atri~; define the lo.-ati.~.n of test predæterrllirl~-d r~^gi.-.n ~ and ...ntr..l predeterrr,ined region ~1 as irl Fig. 7A. F.e~i..ns ~14 and ~16 are nl~urlted .-.n a non-. nducting layer 2~ that itself r"ay be ri.ounted cn a supp..rt fi~:ture ~t7. -rhis apparatus is designed t-. allcw easy ins~rti.~n of s~nsors intc the fluid ~10.
Measuring means und~rlie t~e lc.cali-ing rr,eat7s in pred~terr-ined reyi.ns ~14 and ~1~. Stirring bar ~1~ or 4t~7~r rnea~s .~f causing bul~ fluid fl~w i5 used if desired to aid .:onta~-t br~twe~n the fluid ~10 and the surfaces of predetern~ined r~gi.~ns 21~ and ~16.
A variety of strLtCtures, c.r biolayers9 can be used to fi:c a set . . f l o. al i ~ i n9 rneat7s f i r rr,l y and .- onven i ent 1 y i n posi t i ._.n over t~7e measurinrg means. This allows one to taka~ a rneasuræment, t~len rernove the used lo.-ali2itl9 means and fi~: a næw set in pla- e while using the sarne nleas~lrirl~ rr,eans. F..r e~an-ple, the locali~in~ r,eans rnay be mounted Otl a ~loldær that fits ~ver bac~ck ~ 4. Other bic,li3yer v~riations aræ discussed hærein b~!lclwa ..

~-- !
;2~ .
Multiplæ tæst cells and positiv~ as well as n~gative ~ontr~l cells n,ay be placed in w~115 such as those of Fig. 7A oY on dip-in ' sensor5 5uch as t~l,~.sæ of Fig. 7~. To aYoid interfe~enc~ betweæn cells, neighboring cælls sho~ld be far apart from ea.h ~ther; that~
is, tll~- distan.-æ betwætn thtrn should be c.-"lsiderably greater t~lan thæ distance betwææn thæ currænt electrodes of a single .ell.
Thus, recæssed plarlar sænsors of this typæ can bæ fashic.næd into rni._rotiter plates .-.r dip-in prob*s whæreitl sev~ral assays ar~
perforr"ed in ea,h wæll or .-.n æa.-h probæ. Probæs .an also bæ
fashiotled into sensors for indwellitlg .-atheters~ or the sænsors can liæ on thæ surfa._æ .-.r insidæ of a .~hannæl through whil_h fluid sampl* flows.
A practi,-al test apparatus and instrur"ænt should bæ capable of p*rf.-.rrning su,.æssivæ tæsts Otl various solutions qui.~k:ly and inæ~,;pænsivæly. Fref~rably the prædætærmitl~d rægions in thæ test and control .-ælls should be æasy t.. replacæ . Als," as dis,-uss~d earlier, each sæns.-.r prefærably has a small prædetærrnintd regi~n, e.g., le55 than 1 ul .-.r even 1 ss tha~ 0.1 u17 to allow increas*d sænsitivity. Further, the effi, ietl y E of the sænsor præfærably is high~ evæn when the predeterminæd regiot~ is small.
Two pr æf ærræd cc.tlfiguratic.ns that ~ w the abovæ featuræs aræ
illustratæd ~n~t to s.-ale~ in Figs. 8 and ~. Tilæ apparatus shown in Fig. 8, .allæd an "IVEP" ~Internal Voltagæ Equivaletlc~ Poi71t~
sensor apparatus, allows accurate measuref"etlt of the ~-ondu._tarl.-e of væry thin, layær~ e predætærrnitlæd r gions l:Fig 8~, ~56-~atld t~st volumes (Fig. 8~ 307)~ .~n the .~rdær o~ 0.1 mm thic~ and --6g--~1 -` ' f 2~.
0.1 ul ir~ volume9 or ev~n l~ss~ ~h~ ~pparatus shown in Fig. ~A, called an "EVEP' (External ~oltag~ Equivalenc~ P~int) s~ns~r apparat~ls~ allows ~onvrcni~tlt m*asurernent of events cccurritlg n~ar as well as in th~ pr~deterrl~in~d regions, and has a pot~ntially very 5i n,plæ bi 31 ayær.
~ *ferring tc Fiyure 8A~ there is 5$10Wn an IVEF sens~r apparatus cotr,prising a frcnt sæ~tion 5r) and a rear sæ ti~n ~5~
mad. 4f nc~ condu_ting rr,aterial su-h as a ryli- plasti~0 S.---tions 250 and ~ may be sæparated al-ng a plan~ dæsignated by r"id-lin*
~5~. A r"iddlæ lay~r -~61, callæd thæ bi-layær, c~nsists cf test prædætærmined regi~ns ~5~ and ~57 and - ntrol cr~ll predætærlflin~d ræyi-ns ~58 and ~5~ r"cu~lted on insulating layer ~G~:~. Biclayær ~1 is insærted betw~æn front and rear se~ti~ns ~ and ~5_ during c,p~raticn of th~ apparatus. Lo-ali~in~ rr,~ans pref~raby cc~rl,pris~
rf,atri~; rnat*rial c~nsisting f a t~lin p r us layær of nitroc~llulosæ cr ~ther p~r-us filter rnaterial, wi7ich defin~;
pr~det~rrr,irled regi-ns ~5~ . Th* prædæterr~,inæd r~gi~ns of the bi~layær ar~ callæd bicr~gicns. Pr~ferabiy, th* insulating layær itsælf is n,cunt~d ~n ~ rigid supp~rt ~n-t s~l~wn~ t~ aid in handIing a~7d positicning ~f the biclayer.
Fig. 8A shows the IVEP s~nsor in its ~pen p~siticrl", with frcnt and rear sænsor blocks 50 and ~5~ separated ænough (typica11y cabout 1/16 inc h ) to al1ow easy insertic.n o f bic,layer 261. Fig. 8~ sh~ws dætail of the sensing rægion of the test cell~
with sens~r blo-ks ~50 and ~5~ losed over bi~layer ~61, sq~lee~ing upper and l~wer bi~regions 56 and 57 in channels ~66 and ~68 -7~-C32~i formed b~tween surface 291 of frot~t s~nsor block ~501 insulating layær ~60, and surfacæ 29~ of rear setlsor blc.._k ~52.
The apparatus of Fig. 8A op*rates in a mann~r sirnilar to th~t of Fi~. 5A. (Howæv~r, it should be ~oted that an IVEP sensor apparatus could bæ rnad~- equally w~ll Wit~l the arrang~r"ætlts of Figs. 5'~ or SD.) T~st and control cel 15 ,are c~nrle.ted in seriæs and thæ test and control .-:æll paths are syr"mætri~al. ~efærring t,-, Fig. 8~, reservoir 10_ is filIæd with el~AItr~lyte s,~.luti"n 1l~8.
Two .~hantlels lC)4 and lOG pr"vide paths by which elæ~-tr,.lyte solutiot lO(:),~an fll.w fr,~rn resærv,,ir 1(:)~, thrl_.u~h tæst pr~^deterrnitled r~gi"ns ,5~ and ~57 and cotltr,-,l prædæt~rr~litled rægiotls ~58 and ' 'J'3, thr-J-Igh two e~,~it hatlnel a 11_ and 114, and eventually t,, a waste ~,ntainer ~n.-.t sh._~wn j. Thæ fluid fl..w is indicat~d by arr,-,ws 1~ . The flow may be driven by yravity~
pneu~nati. .~r hydraL~ pressure1 c,r other rrl~atls.
For thæ test cæll, fl~id travels fr.~,rn fluid inlæt p,_.rt 6_ and res~rv,-.ir 10--; through chatltlel 104, upper t~st bioregio openin~ provided in insulating layer ~61~, lower tæst bi.-,regi.
~7, atld thæn exits through ~hannel 11~-. For thæ c~ntrol ~~ell, fluid travæls fr.-~rn fluid inl~t port ~ and reserv~-.ir 102 thr.-.ugh chatltlæl 106, upper cc.ntrol bioregiln ~81 ,.pening _65 provided in insulating layer ~0~ lower .otltr,~l bioregiot- ~5g, and thetl æxits throu~h channel 114~
The rneasuring rneatls corr,prise recessed voltage and ,-urrent ~le-tr.des. A first ,-urr*nt ælæ.-trcldæ 116 in .hantlel 11~ and a se~ond current eIe~tr4de 118 in ~hantlel 11~ are respe tiv~1y . . .

49~?~ .
, .
c~nt ect~d t~ t~rn~inal~; C1 and (:~2. A signal g~n~rat4r i5 COtlneCt~d t~ trrrninals C1 and 1~ and ~auses a ~urrent to ~l~w fr~m current electr~d~ 116 thrc,-lgi- chann~ , lcwær tæst bi~r~r~ion :~57, ~p~ning ~&4, uppe~r te~st biore~gi,.n ~S6, chann~l 104~ r~s~rvoir 1 (~o ~ c ~annæl 106 ~ uppær c r~nt r l~l b i ~r æg i ,~n ~58, ~p~n i tl9 ~65, 1 .~wær bi,~regi4n :~5'~19 and channæl 114 t~. ælrctr.~dæ 118. This currænt fl.~w caus~s a v.~ltage dr.~p bætw~æn thæ t.~p and b..ttnm bior~gi,~ns 4f ~a._h .~æll w~i. h can b~ mæasuræd t., d-;t.:rr"inæd th~ pr~s~n. ~ of a parti,-ular 1 igatld in tn~ elæ.-tr.~lyt~^~
Tw.~ fluid .~nat-næl5 ~66 and 613 ab4væ and bæl"w th~:~ insulating layær ~60 ar~ fillæd Wittl ~l~.-tr41yt~ This may c.~nv~ni~ntly bæ
d.~næ by k:~rping b..th sætls..r bl.-.1 k:5 ;~50 and ~5_ subrnerg~d itl an elæ,-tr41ytæ bath (S~r Fig. '3E~. 8hannæls ~66 and -~68 aræ
equival~nt t,-. re.-~ssi.~n clla~ln~ls 130 shc,wn in fig. ~iA. As 5hl-.Wn cl~arly in fiy~ 8~, ræc~ssi-~n channæl :-~66 pr4vides an æ}æ. tri~al c4nn~rti.~n bætwæærl thæ fl.~wing el~-_tr-~lyt~ and v..ltagæ ~læ.-tr.-lde rc.nnæ.~tæd to tærn-,inaI V~. F.æcæssic~n .-hannæl ~68 providæs an ælæctri. al .~~nn~-_ti~n bætw~æn thæ ælæ,_tr,~lyte and v-31tage æl.-.-tr,~clæ 1~4 c4nne._tæd tc. terrf~inal vl. Thæ c4ntr~1 _æll .~pæerates sirf,ilarly, with ræcæssic"l .:hannels ~70 and ~7~ ræspectivæly pr.-.viding æl~._trical paths bætwæ~n thæ ælrctrcllytæ and voltagæ
æelrctr.~dæs 1'6 and l~S c4n~lect~d to tærrninals V8 a~ld V4. O-rings 27S thr4ugh Eil prc,vid, sealitlg t4 isolate thæ fluids in thæ test and c4ntr.~1 cells and prævætlt ~læ~~trical læakag~. S~lrfa.-æs ~0, S atld ~6 dætærn,inæ tlle dir~ænsi~ns .,f the ~æl1~3 by buttirlg against insl~latitlg layer :~60. Sur~aces :2C78~ ~:00, ~0:~: and ~04 rn~y - ~7~--al~o særv~ this functio~l. All r~cession c~lannels~ v~ltag~
electr-3dæs, and D-rings ar~ circular and symrnetrical ab~ut tile inlet and outl~t channels 104 and 11~ ~for th~ test ~ell) and 106 and 114 ~f~r tlle c,3ntr~1 ~ell~. V,~ltage el~ctrodæs 1~ 4~ 1~6 ' and 1~8 ar~ furth~r re--essed by b~ing pla.-~-d in re.-essi.-.n .-hambers ~78 c4nn*ct~d tc, recessi4n channæls ~ 68, ~71) and ~72, which alll~ws the electr4~s t4 bæ large and have 14w p-31ari~ati-3n imp~dan.-e.
In .~perati,.tl, the stru-tur~ .-.f Fig. 8A is c.pened al,.ng n,id-line ~54, and a n~w biolay~r ~1 is inserted with fr~sh test and .~ntr.31 cell bi.3rægi.-.ns. Fr,-.nt and rear se,ti.~ns ~50 and ~5 are th~tl r~asser"bl~d, with O-rings ~7~-_81 pr4viding sealing ,-.f ~ach cell~ This thre~-part structure all,~ws ~asy and inexpænsive r~pla,ement ,3f bi41ay~rs.
The sealing pr4vid~d by the O-rings may n4t be perfæ~t, h.3wev~r~ and seepage .3f el~-~tr~31yte past the O-rings r.~ay .-cur.
Re~erring t,3 Fi~. 8A, if there is secpage past O-rings ~74, ~75, ~76 and ~77, fc.r example, a .-~.ndu-tive path b~twe~t the test and C 4tlt r C1 cæl 1 5 may result in ærr.3rs in th~ measur~metlts 3f t~le c~ll conductances~ T~ reduce ~r elir~linate errors caused by such se~pagæ, guard ælectrc,des . 8~, _a~, ?86 and ~88 ,onnected t4 terri,itlals 1~1 thr4ur~h l,~ are pr.3vided if desired b*tween each .3f the voltage mea~uremetlt tærrninals V1 - v~ and its adj.3ining O-ring sæals. Th* .-.pærati-~n of thes* guard rings i5 discussæd in m~re detail in Exar,lpl*s ~ and C7.

~73-~:

Referring again to Fig. 8B, there is shown the sensing region of the test cell in closeup, wi~h the ~VEP sensor apparatus in the closed position. The bioregion matrix material of test predetermined regions 256 and 257 ~ills recession channels 266 and 268 respectively. Similarly/ though not shown here, the bioregion matrix material of control predetermined regions 258 and 259 fills recession channels 270 and 272. ~his is achieved by making sure that the surfaces 291, 293, 295 and 297 that form recession channels 266, 268, 270 and 272 lie a distance below coplanar surfaces 290, 292, 294, 296, 298, 300, 302 and 304 (Fig. 8A) that is less than the thickness of the particular biolayers used. With bioregions 0.125-0.150 mm thick, this distance is typically 0.1 mm. Filling recession channels 266, 268, 270, 272 with matrix material minimizes noise and drift due to bubble formation and convection of fluid from the recession channels into the test volume. This also helps ensure that the channels remain conducting and are not blocked, e.g., by bubbles, even if they are very shallow.
Referring again to Fig. 8s, the structure of khe IVEP
apparatus allows the voltages sensed by voltage electrodes 122 and 124 at the mouths o~ recession channels 266 and 268 to lie not only near, but actually inside test predetermined regions 256 and 257, as shown by voltage equivalence points 306 and 308; hence the term ~Internal Voltage Equivalence Point", or ~I~EP" sensor.
Voltage equivalence points 306 and 308 help locate test volume 307 whose conductance is measured to determine the occurrence of .~

li9andJ~ tili~a~ld irlteraCti4n itl pr~det~rmirled regio~s 256 ~nd ~57. Itl this sirnple rn.-,dæl, t~st vol-lm~ 307 may b~ thought of a5 typically conlprising 51al f the thi~knæss ~f t,~p rec~ssi._.~1 ,-hann~l ,~66 plus t~ thil_kt1ess ~f insulating layer ~60 plus hal f th~
thi.-l~ness of bc~ttnr~l re,-essi,-,~1 .~t1annæl :;~68. Sirnilarly, f,-,r the contr4l cell, thæ v~ltagæ s~nsed by the voltag~ ele~tr~des l~ and 1~8 at t~-e rr,ouths of re.~essi.~ han~1els ~70 and ,~7:2 ~Fig. 8A), allows the v,-.ltagæ eqLIivalence points f,~r th~ ntr,-.l sensor to lie insid~ ~ n~ltr~-ll predetærr"ined r~gil~lns ~J3 and '_J~ dæfinir19 a ,_orresp,~nding tæst v,-,lLIrile ~n"t shl~wn) that c,-,rr,prises hal f th.-thi.-k:n~ss "f t,~p re,_~ssi,_,n l har1n~ 70, plus the thic~;n--ss ,_,f it1sulating layer ~C), pl~ls hal-F thæ thil-k:~æss ,-,f b~tton, r~,-essi,-"
channel :~7~. .
The IVEP sens,-,r apparatLIs shQw5 that the lol-ati~1 and si~e ~,f th~ voltagæ ælæctr~dæs ~~an be ~ffæ,~tivæly de~ ouplæd fr.-~n~ th,-.se~ ,-,f t~l~ v,:~lur,l~ W~ lSæ` ~-o~ldu~ ta~-t^ t~l~y arr r~ a5uri~19 t~lr~ugh t~ LISe:~ ,-,f ræ,-æssi,-"l ,~hambers and ~-hannels. The flæxibility in særls._.r desig~
thLIs gai~1~d is ar1 irr,p,:,rtant advantagæ llf thc present inventi~n.
Again ræferring t" Fig. 8A, the abi l ity t,~ ~._"1stru,_t a tl1i~
test volume .~an be s~n to be limited only by th~ availability of thin rr,atrix rr,at~rials ~I.r th~ predeterrrlin~d regi~,~ls _~6-_~g, a thi~ insulating layer rrlat~rial ~6~ and th~ accuracy ~f te,-hniques availabl* to place the surfaces ~3i, 2~8~ ~g5 and ~7 that form r~ces6ion channæls ~66, ~8, _70 and ~7~ a small riistance beluw surfa,es _13t:), ?~?, ~4, '~6, ~8, ~ , 8t~' atld _04.

,, ~ ,ffoe As noted earli~r; a t~lin test volume allow5 t~st and control Y~lumæs to b~ very small~ which is ~n isnporta~lt advantar~e 4f thi5 ' ' IVEP ernbodirnent. Typically tæst volum~s are construct.?d using nitrocellulc!sæ n,atrix mat~rial from Millipor~ ~crp. ~ dfc~rd~ MA) that is l 5-150 micromet--rs thi.k, ræcession channels lQO
micron,---t-.-rs thick~ an insulating lay~r of l~xan from ~æneral El~.tri,- (S,henæctady N.Y~) 75 rnicromæters thick, asld an opening in the insulating lay~r 0.~ rnm in diarneter. This yields a distanle betw.?æn the vc,ltage equivalænc--- p,-,ints of one half th.--tni,_kness of th.- t,-,p re,~ssic,n channæl plus the thickn--ss of th.--insulating lay~r, plus onæ half the thickness of th-- bottor, re -ssion channæl, c,r 175 n,i,-ror,~ ters tcltal. Sin,æ the ar~a cf th~ c~p.-niny is about 0.5 mm~, an .-~stirnateri siz.~ fc.r tht- t..-st volumes is 0.0~ rni,-rolit--rs. ~-Evæn smaller test vc,lures can be cc~nstrLIct~d. As an ~xarnple, a matri~; that is appr.-.~ir"at~ly 25 or ~0 rnicrorneters thi.~ and r~cessic.n channels that are ~5 n,icromet~rs thi,-k: can bæ used alc.ng with an insulating layær that is appro~imat~ly 25 nli.~rorr~ters thi~. Als,-, if a half rr.illin,*ter dian~eter h~le in thæ insulating lay*r of thæ biolayer is usæd~ its ar~a will bæ 3~16 sq.
rnillirr,eter~ The distance bætwæen voltagæ equivalencæ p,-,ints i5 a total of .05 rnillirneters. Thus the total t~st volun~e would be about ~ nanol i t *r g or . OOg r -i c r ol i t er s .
Sensors with *ven smaller dime~nsions can be built up using thin layers ,.,f irlsulatitlg material o~ appropriate thic~;ne~ss and instrument grad~ machining techniques, injection molded parts"

,,~

2~i thick ~ilrn or thin filn~ techniqu~s~ silicon chip te.-~ln.~logy, or ot~er techniques knowtl to th~se skilled itl the art. Using su~~h te,-hniques it is p.-.ssiblæ t~ c~"lstru._t IVE~ setlsors whtr~ the widt~l of the oper7ing of the re,-e5si--,n cllan?lels ranges frorn 10 mi,-r.~rr,eters to ab~ut J r"i.-~-~.meters or even srmalltr and whært these 4penings are 10 mi._r~.r"eter~ apart wit~ recipe~t to each ~ther or less~ yieldirlg an even sr"all*r test v,~lume.
~ iolayer ~ an be varied in rf,any ways. For exan,ple either tht top .-.r b"ttor" predeterr"ined regi,.n '~'J~ ,.r ~!57 rnay be elirhinated, with substituti n of a ri,atri.~,; that dc,es n,-,t l._"~ali~e antili~ar7d~ ,~r ..f n,_.ri,atri.~; at all. Operlin~ ~&~ in the irlsulatir1g layer rnay be s~tbstarltially ~ ~mpletely filled with lo.ali_ing n,ear-s r~,atri~,; by .ompressiorl or by casti~lg nitrocellulos~ or .~th.er r"atri~
rnat~rial dire.tly in the openirlg. Filling .-,pening ~4 with lo,-ali_irlg meatls matri~; all..w~ .:otlstr~.-ti.~tl of a sensor Witl nearly lf~~/. ef~il-ietl::y, in Whil h s~b5tatltially all thæ test vc.lun,æ
~,~ntains l.~"~ali~ing n,eans. Als.-., in S~ l a s~nsor, s~bstarltially all ~f thæ test volume is in~lud~d within t~le l.-.cali_itlg m~eans.
~ eferritl~ t-. Fig. 3A, there is sh--.wn a typical EVEP serlsor apparatL~s cc,nsisting in part of fr~nt se~tion ~50 and rear section ~5~ whi.-~l are r,lade ..f n.~rl-corldu ting rflaterial SU~l as a.-ryli.
plasti._ and w~ h n,ay be separated al..ng a plane desiyrlatted by mid-line 3S~. A middle layer, bi--~layer 3~5, rompris~s test predet~r~f,ined regi~ln 356 and ,-orltrol predeterr,itled r~gion ~59 rr,.~utltæd if d~sired on an ins~lating ba k:ir-r~ 3~0 for .~onvænien.-e in tlandling~ ~iolayær 355 i5 insert~d between front and rear sections ~S0 and ~S~.

Irl this en,b4dimet7t, t~ locali~ing rr,~ans typi~~ally comprise a matrix mat.?rial consisting of a t~litl pc,rous lay~r ~f nitr~.ellulos.? or oth~r p~,r~us filtær material~ whi,-h defin~
predet.-rrr,in.?d regic"~s ~56 and 35B.
The apparatus ,-,f Fig. gA c~p..~r~t.?s in a manrl-?r sirrlilar t,~ th~
apparatus of Fig. S~ How--ver~ it sh._lulLI b-- nl~ted tilat tht- EVEP
s~nsor apparatus Can be mad-? equally w---ll with arrarl9~?met~ts such as thlls--l-lf Fir~s. 5A or 5D c~r in a s-~ri--~s flc~w configuYati,-,rl su,h as tilat ,-.f Fiy. G~). A tæst c--~ll and a ,-,-"ltrQl c-ll ea,-h has its oWr7 C ur r~nt path wit~l th~ twn l-~--lls sharirlg a cor"r,itl ,-urrent ~l.-.-tr"d.?, and tht? test and ,-otltr"l ,~?11 pat~ls are synlmetril~al.
~eferritly tc, Fi~. gA, r~s-?rvoir 1~ is fill--d with an el ~?CtrC~lytæ
s"luti"n 11~0 via fluid inlt-t port 3~3. Tw~ chanrl~?ls 104 and 106 r.?sp.?,tiv.-ly pr,~vid-? pat~ls by W~li,-~l t~le ~l~,-tr~lyt~ scll~ti~l fl~lws from r-?s.-rvoir l(:r~ tllrough tt?st pr-d~t-?rrr,irl~?d r-?giorl 3S6 and COrltrOl predt-t~rrl'lin~d re9iOtl ~53t through tw,,--~it ,hatlrlels 11-and 114, and ~ventually t.:, a wastt-,-ontairl-r ~rlot shl~wrl:). Th.-fluid flow is indicated by arrows 1~0. The flow may b~ driven hy gravity, pn~ulnatic ~r hydraulil pressur~, or ,-,th~r rr,--ans.
For thæ test cell, fluid trav~ls fronl r.?s.~rvoir 10~ thrc,ugh ~hann~l lC)4~ thr~:~ugh pr~d---termin~d region ~-56, through op~r~lrlg ~6 in insulatiny layer 860 l-lf biolayer 355, and tll---n ~its thr,~ugh chann~l 112. Sir~ilarly, f"r th* control c~ll, fluid travels frc~n res~rvoir 102 thrl.ugh chanr~æl 106~ through contr.~l pr~d~terrnin~d r~giorl 8~8, thro-lg~l clp~nirlg 3~4 in insulating lay~r 860 .,f biolay~r ~55~ a~d th~n e~its through ctlan~lel 114 -7~-~ ~f~Ct~ ~ "
Th~ measuring rrleans comprise r~cæssed voltagæ and curr~nt electrodes. A signal g~neratc,r is conne~ d to terrninals C1 ~f ~l~ctr~d~ 140 and C2 of rin~-shaped ~r arc:-shap~d elè-krode 144 that causes a ~urrent to flow frc~r~ el~ctrc~de 140 throu~h chann~l 11~, op~nirlg 36~, t~st pred~termin~d r~gi.-Jn 356, cha--n~l 104 and reserv~ir 100, to ~l~ctrod~ 14~. Similarly, a signal genærat.~r is ~onn~.-ted t~ t~rminals C1 4f ~l~ctrode 14~ and C~ ~f ~le--trode 14 that caus~s a current to flow fr.-rr, el~ctrbde 14~9 thrl-~U9 114, openirlg 3~4, c./1trol pred~terrrlin~d regi,_.n 3J8~ '-har\nel lC)~
and reservoir 10 ~ th~n als.~ t.-. ele-trode 1~4. Electrl-.d~ 1~4 thus is a c,~rbrr,cn ele,trcde to the tw~-cells. The -urrent flows .-ause voltage drl~ps acr.~ss th~ pr~det~rrnirled re~ ns and test volur"es (see belcw) c,f each cell whi.h can be measured to d~t~rr"ine the pr~sen.-e clf a parti.ular ligand in the ele--trolyt~.
Fluid charlnels 3~6 are ~ d with ~lectr41yt~ and prc.vide ~lectril-al paths t.~ recessed voltage eleltrl~des 368, 37~:l, 37~ and 374. Thes~ channels are er~uivalent tl~ voltage elæctr4d~ recessi-n ~hannels 130 5h'::~Wn in Fis~. ~ A and the simi lar ~hannels sh, ,wn in Fig. ~iC. P.~ints 374 and 37~ irl channels :3~6 are the v~~~ltag~
equivalen-~ p4ints of the test cell, ~nd poir~ts 378 and 380 are the voltagæ ~quivalerl,-e p4ints ,~f thæ ~ontr41 cæll. Thesæ help define thæ position of t~st volume 357 and c4ntr41 volume 85~
wh,~s~ .~onductanc~ will bæ measured. Voltage ~læctr~liæs 368 and 37~ ~ onrle~ted to tærr,linals V1 and V? mea~;uræ the~ v41tar~e dr~p b~tw~æn voltagæ eq~ival~ncæ points 37~ and ~76. It 5huuld b~ tl~t~d that in the EVEP config~ration, the test volume betwe~rl voltage ; ~ --79--. .

equivalenc~ pois~ts 374 and ~76, and hence typically the tæst volum~ 357, entirely ~n~ludes predætermined rægion 356.
Sirhilarly~ voltagæ ele~-trodes ~72 and 374 ornected tc, termi~als vl and V~ mæasur~ the voltage drop between voltage equivalærl,_e points 3~8 and 38~ which help define c,-,ntrol vc~lurr,e ~J'~.
Voltagæ ele~trode r~cessic,n charnbers 3E3~ recess the v"ltage ele~trodes even rn,-re, and allow them to bæ large, reducing their polari~atiotl impedan e. Curre~lt elæ~trode r~cessiotl char"bærs 383 a,t sir"ilarly for t~l~ ,-urrent ele trodes. Fluid outlæts ~-8;~ fror"
v-ltagæ electrod~ re cssiotl charbb~rs 38 alllw flushitlg -f the ~hamber5 with electrolyte if d~sired, while fluid ,~utlets 385 fr,-m the urretlt electr-,de ,-harf,bærs all"w rer"c,val of elæ,-tr,~,lyte t-waste containers.
Additionally or as an alternative to thæ voltag~ electrodæ
recession chamber outlæts 38ct, recæssic"7 ,-hannels 3&6 are filled Wit~l a cc,nducting matrix such as pc,rous papær or a gel to reduct nc,isæ due to bubbles or to entry of ,onter~ts of charr,bers 8a:~ intc~
test and ,_ontrol volumes 3~7 and ~S~, e.g., by convæl-tic)tl.
Likewise, fluid inlet and outl*t ~hantlels 1~l, 11~6, 11~ and 11c~
may co~tain a pc,rous matrix for similar reasons as lot79 as this dc,es n,~t intærfere unacceptably wit~l fluid flow, The volta~e equival~c~ poilts in the apparatus ~:onfi~uration of F~g~ ~A are outside pr~d~terrr,ined regions ~56 and 858i thus thæ
nar,~ External Volta~e Equivalencæ Point ~EVEP) sensor.
~ rin~s ~2G and J~38 provid~ sealing to isolate the fl~ids in the te~;t and control cel 15 an~:l prevent electrical leakagæ.

--~30--.

Th~ bi ol ~yer con f i ~urat i on ~Al l ows ~asy hat-dl i ng and positi~ning ,~f th~ r"atri~æs d~?fininy pr~d~t~rn in~d r ægi4ns 3~i6 ~And 35~. Bic,layær :3~5 . omprisæs pr~ t~rn,it7ed r~gions 35~ and 358 and spa~ ing lay~r 33C~ rro~Anttd on ins~AlaAtin~ ba~ i tl9 1 ay~r 3~0.
Spa~ ing layer 33C) i5 slightly thinnær (typi, ally Sln th~- c~rder .-,f 0. l mrf,~ than pr~d~termin~d r~gi~t1s 35~ and 358 (typi- al ly on th~
ord~r o f ().1~5 t~-l OA 1JO r~r":) 50 that bi"l ayær 355 wi l l bæ hæld f i rr.,l y wh~-n th~ S~t15. ,r i s , l . .s~d . Spa, i ny l ay~r 3~(:) and insLIldtirly lay~r 3~(:) aræ squ~2~d b~tw--~n surfa,-~s 3~-_ and '-'31 t,.
seal th~ O-fi~195 ar1d dætærr"ir1~ th~ dir,let1sic~r1s .-,f th~ t~st ar~d .-,ntrol c~lls.
Figure 3~ d~pil:t~; a fluid bath whi~-h ,_an b~ us~d f"r æith~r th~ IVEF ,~r thæ EVEP sæns"r t,-, mair1tain t~nlpæratuf~ æquilibri.um ,~f th~ t~st and ,-ootr~ ells~ t,o w~t th~ biolay~r as it i5 ins~rtæd, aAnd t,, k~æp t~1~, hanrl~-ls ,_,f th~A s~-ns~or fill~d Wit~1 fluid. tV~ r~
sp~.-ifi, ally, th~r~ is sh,~wn a t,~p vi~w .-.f fluid bath ~.ntaining ~l~ctr~-~lyt~ 8. Typi.-ally, thæ frl~r1t 35~l and ræar 35 bl~lrk:s ,~f an EvVEP set~sor ar~ rn,~unted in fl~;ibl~ wdlls ~0(:) and ~0~
of fluid bath ~ using s,-ræws ~(:k~ and rr,Gur1tir19 pi~. ~s ~ 6 and iC~8 or othær n-æans~ Th~ bl4~ks ar~ at a height that l~avæs the~
rr.~asuriny mæa~s lnGt ShOWn) in thæ bl..ck:s subr~ r9æd at all tir"~s so th~t whæn biolayærs ar~ insærtæd or r~rr,Gv~d, tl1~ ._hann~ls Gf th~ sænsor rænr,ain fillæd with flui(:l. As nnt~d ab~.væ~ th~ flL~id bath hælps mai~1tair- t~n,pæraturæ *quilibriurfl of t~1æ tæst and trGl ~ ~lls a~d we~ts t~ bi..-.,l~y~r ~5 it i5 i~.5~rt~ 'Ll~lti~l9 pi~æ5 ~06 and ~08 also h~lp maintain t~rr.p~ratur~ ~quilibrium by --81 '--z~
k;~eping s~nsor bl~s 3~0 a~d 3~2 away fror,~ walls 40~ and 404 of the bath~ In ~n~ral, n,~ additional t~rhperatur~ r~gulatio~7 is ne~d~d, ev~n for c4nductanc~ r"æasur~-rnænts sensitivf~ tcl lO0 parts p~r million or b~tt~r.
The arrany f rfl~'n t ,~f Fiy. ~ h,_,lds sens.-.r bl..cks 350 and 3 firmly at~d all.:,ws c~nv~nient ins~rti4n .~f biolayers. Aft~r a bic,lay~r ha~ be~n ins~rt~d, ~larnp 4lC) or c~thær n,æans squ~ es t~gæth~r th~ walls ,~f th~- bath7 and he-n,-æ squæ~-æs th~- bi~lay4r b~tw~n s~ns.-,r bl~~~cks 3~0 and ~,5-, is~lating th~ test and ,.ntr~l .-~lls frc~rl, ~a,-h clther ~l~ tri,-ally ar1d hydra~lli-ally via th~
acti-,n of th~ 0-rinys nl-lt~d su~ra.
Th~ E'~EP typ~ ,~f s*ns~r can b~ rl,c~difi*d in a vari~ty ,-,f ways.
A sarnpl~ ,-,p systæm rf,ay b~ used f4r fluid sampl~ inj~~ti~n, with c,r without a r~srrvc~ir. If d~sir~d, C~pf-nin9 3~- in ins-llating lay~r ~3l~ may, as sh,-~wn in Fiy. ~ b~ r"adæ stnall~r than tl1~
diarll~t~rs ,~f ,-hant7~ls lO~, lC)~ and ll~. This has th~ ~ff~t ,~f in,~ræasing th~ r~lativ~ ntributi~r7 t,~ th~ t~tal r~sistanc~
~b~^tw~en the v~,ltage ~quivalf~nc~ pc,ints~ c.f th~ r*r7ic~n i~1 th~
imrll~diate Vi,i~1ity of l~p~nin~ ~'--a partial J~.n~s c~}l-typ~
~f fect . Ml~st of the r~sistance asso~~iated with th~ nstrict~d op~ni~ly "c, urs withirl æ. v,~lurne assc~ciated with the~ c~p~ni7lg itself and t~le r~gi,,n orl either sid~ ..f th~ ,~pening within on~ diarnet~r~s distan. e~ This appr~a.~h, an improv~ th~ æf fi~ i~ncy ~f ~V~tl an inef fi~ient EVE~ s~ns,~r to 3b% tc, JO~/. i f the tllic~ness of pr~determin~d r~gic~n 3~& is, ~Imparable to .~r greater than b~th th~
thi~k:ness ~f insulating layær 360 a~ld thæ l:liamet~r cf .~pening ~:62, --Q~2--~ e~r~1 if the vc.ltagi equiYalæn.~ poi~ts 374~ ~76~ ~78 and 380 ar~ far awayO
Anothær way to incr~ase thr- r-ffi~iency of an EVEP sæns.~r configurati4n is t-~ brin~ voltag~ æquivalæn.-æ p~ints 374, ~76, ~78, 380 cl~,sr-r to th~ir prr-dætr-rr"in~d rægi,ns by ins~rtitlg hollllw tubular sheaths d.~wn thæ inl~-t and outlæt ports .~f thæ ~ontactitlg rnea~is~ partially blockitlg ræ,æssion chann~ls 3~ to thæ voltagr^
elæ,tr.-,dæs and pushing t~7æ voltagæ ~quival~n.æ points næar~r t..
thæ bi.,layær~ An æ~;pr-rimrtlt using t~is "s~at~æd .-oluriltl"
.-onfiguration ~f thæ EVEF sænsor is sh,-,wn in Exarr,pl~ 3.
Thr~ r-ffi,-iæn.-y of thr- EVEF, IVEF and ot~ær ~rnb;:,dimr~nts ,..an bæ
in~~r~as~d by pla,ing a ,_urrænt rr"_,dulating r,atr-fial nr-ar a lo._ali~ing r-æatls~ wh~rr-:in this matr-rial has a cr.-,ss-sæcti"nal ,_c.ndu,_tirl~ aræa lrss thatl thæ .ross-s~ti,~nal cllnducting aræa ass}l.iatæd wit~ th~ locali~ing rnæarls. Arl r-xamplæ of such a matr-rial is Nu.læp..ræ filt~-r~ a radiatilln-æxp-_.s~d, alk:ali dig~st~d pcr.-.us poly.-arb.,~atæ availabl~ cotrlm~r.-ially fr._.m Nucl~pc~r~
l~orporati4ll ~Plæasatltc~n, I-A~. This mat~rial ,-onsi~ts ~f ~
not~ du.ting sc.lid and 1~% .-onductin~ pores. It may b~ th._,ught of as ,r~ati~g a mi~roporous, partial Jo~es c~ ff~ct sir"ilar t~.
~ffect .~f ._onstricting ~p~ni~g ~6~ in ins-llating lay~r 3~C)~ 8ir el~ctric .~urr~nt rnust bæ .on.~ntrat~d in the n)odulatir-~g ratærial to pass through a conductitlg aræa 1~s5 than that as ociat~d with th~ Io.-ali.itlg m~ans a larg~r voltag~ drop will occur h~r~ du~ t4 tht? great~:r curr~nt d~nsity anr~ local r?sistanc~. To th~ t~nt that thæ l~caliziny neans i5 l-105~? ~?n~U9h to th~ curr*nt :: --67.~--f modulating n)at~rial to be at least partially in th~ ræ~i~n ~f ~ligher currerl~ density/ any volta~e change assc~ciated with changing condu.tancæ cf the lol-ali~ing r,-~æans that ~ urs in this region as a result of li~and-antilir~and a~ti4n will bæ ~reater Th~ l...ali_ing r"rans its~lf rr,ay have .hara-teristi-s that ri,a~;~
it a current rr,odulatirlg r,laterial. Nu--l~por~, for e~arrlplt~ can itself bind pr,.tein; it then servrs bc~th as l,~,alizing rr,eans and as a curret1t-m.-,dulating rr,eans. If Nu._l~pore is L~srd in an I~EP
typæ Jf sr~7s--r, an additi,-nal ---ndulting layer is laid Cltl t"p nf th~- Nu lep.-re layer sin-e the Nu.l~pore stru.-ture d.~l.s n~t cl-lt1dul-t ~le.tri.-ity in a dire.-ti.~n parallel t.-. its surfa.-~.
It sh.,uld b~ nl-ltrd that apparatus ,-.f the ab-~ve typeC .-an .~learly b~ constructed irl whi.-h the predeterrr,ined regi..ns contain lc..~ali 2 i ng rheans with irr,r"..bili~ed (-~.r ..therwise 1,-,.alized:) ligand irlst~ad of antiligand.
This inventi.-.n sp~._ifi,ally in,ludes th~ con ~pt c.f c~tlstru~ting a bi.~layer ._c.rnprising a ~onlbination of ~tle or rr.~re lo.-ali~ing rneans and appr,_,priate supp.-,rtiny fi~;ture that is d~sign~d for ~asy insertion .~r pla errl~nt .~f th~ -ali_ir-~ rr,~-ans intc. or Otlt~ the rest Qf the sensing apparatus~ The biolay~r may be dasp~sable. It rnay be s~parabl~ fron, the rr,~asuring rr,eans. Th~
ability t~ pr~pare su.h a t,odified l.-.cali~ing rr,eans separately frorn the rr,easuring rb~ans is an irflp~rtant advantage .~.f this inventiGn~ In scrne situatiQns, the biolayer may include rr,easuring rr~eans 5UC~ as ~l~ctr.,des. Th~- lo,-ali_ing r,~eans rr,ay cor,~prisr bound antilir~a~d rn~lecules. The bi~layer rnay cor,~prisæ at Ieast o~e insulating layer to whi~h th~ locali~ing rreans rr:ay be attac~1æd.
The insulatiny layer rhay rr.odulat~ t~1e electric curræt1t flow.
Appropriat~ biolayærs may be used with al~y of the apparatusæs dis.~losed abov~. It~ parti--ularr where the apparatus i5 designed fc.r diffære~tial or rr,ultiple r"easurer"e~1ts, the bi,.layer rnay .-..mprisæ two or r"oræ lo.-ali~ing r"eans.

E,YAMELE_i In this e~;perirr,ett, thæ prese~1-æ ..f a spe.-ifi.- antib._.dy, goat anti-(rabbit garf,rrla gl,-lbulin) antib.dy, .-.r bAF.', was dætected in a fluid sarnpl~- by r"easuring the .-onductan-_* change .f a bi..r*~ n .-,-.ntaining atl irr,r".~bili~æd antigen, rabbit gar"rrla gl.:.bulitl, ,-r F.AX
~st1or.t for "rabbit anti-X"~. An IVEP apparatus of the type sh.~w it~ Fig. 8 was used~ æ~;cept that t~e test and .-.-.ntr.~ lls æa.-h had its ..wn .urrent path and .-urre~1t ele.~tr.d*s as in Fig. 5D, and fluid san,ple was ~dded via a sar"plæ 1OIJP t.. a fluid inlet p.-.rt, with nc reserv ir, als~. as in Fig. 5D. Th* apparatus was ~ ne.ted t.~ a ..~ndu.:tan.e bridg-- of the type des,-ribed in E~;a~nplæ
g.
Test and .-o~1trol bi~rægions wære cnstructæd using 3.5 rrlm dlarlieter disks punched out of l5C) urn thi.~: nitro.-ællulose filter tsupplied by S~hlæicher and 5c~uell "~.eænæ, N.H., 5.~ urr, pc.re siz~ r"ount~d dire-:tly .~vær 0~8 r,1m ~ æ5 1 C.C ated in a ~:).003" s~eæt ~f Ie.~;an {p~ly.-arbonate, fr.~n1 Ge~1eral Electri~ 50 ul .Jf a ~.5 rr,g per rhl s,.luti.~ oat Igl~ ("GAX", short for ''I,~at-Anti-X'', Sigma Chemical ~., St. Louis, MO) wa~ applied to the control 8~ i ,.

bior~gio~ as a negative ~c,~ltrol. This Was all~wed to i~lcubate at room temperatur~ f~r approximat~ly 10 rnirlutes t,~ allow pr4t~in to bi~ld to th~ nitro~llulos~. Similarly, 5t) Lll of a ~.5 mg per ml s~luti,3n of rabbit Ig~ ( ~AX ~ Sigrna) was appli~d to th~ test bi,-,rAgion. I:Alt~ ,ugh FAX i5 it5~1 f an antibody fra~_tion, it h~re furlltions as an antigen to allow testing f~r th~ preserl,-e of ~AF.) Th~ biolayer ~ntaining th~ two prot~in~oated bior~giorls was t~en was~l-d wit~ rur~nir-g buff~r tO.A~ M Nat.1, O.Q M NaF04, pH
7.0~ ar1d pla.-~Ad int.-, the sens"r ~-har"b*r.
D~te~-ti-ln ,~f spe,-ifi,- ~AF antib,-.dy was i~itiated by fl,~wing dea~rated rur1r~ing buffær t~1r,-,ug~ ea,-tl bi,-llayer at the rate of appr,.xir"ately ~t3 ul per rr~ lte while r~"~r~itoring the l~~~rldu~-tanl~e ratio. T~lis pr,~"~edure established a stable base li~le.
,-ollapsible ba~ ,:c,~tait-ir~g runrlirly buffer l,~~,ated appr~ ;imately ~0 ab,-~ve the sensor rnaintained a pressure head that gave adequate flclw tiuri~lg th~ ,urse of t~ ;perim~t-t. 50 ug ~-~f bovi~ serurb album~n tE~SA, Fra,-ti,"l V from Sigma~ th~n 1~ ug of bi,_,tinylat~d bA~ tSiyma'9 then ~ ug of avidi~ ~Si~ma) were add~d s~que~ltially as i~ldicated in Fign 10 whil~ the ~-4tldu~t~ r~tirlu~d to be motlitor~d. Each sample was added in a 10~ ul volumæ via a 103 ul sar~.pl~ lo"p. Th~ results tFigure lt3) show the f~~~llowiny: 1~ A
r-or~-spe,-i fil- pr~t~ir-, BSA, ~-aused o~ly slight d~fle~~tior~ ~f t~e bas~ e sig~lal~ Howev~r, whe~l biotinylated CiAF~ was added, it ,_ausæd a ~egative defle, t~ of appr~ imately 5,0130 parts per r"illi,-,~ ~pprrl~ th~ ~:ondul_ta~,_e ratio, ~ r 13~ t~ a~ld did so in I e~is t h ~ a rn i nut e .

.. . . . .

A parti~l~ cor~sisti~g of th~ tetrarneric prot~in avidin w usæd t~ arnplify th~ ab.~v~ signal. Avidi~l hc~s a high affinity fnr bi~tin; t~us, avidin sh.~uld bind to t~e ir,~m-~biliz~d, bic,tinylated GA~. As is se~n in Fig. l~, additi.~n c.f avidin r~sult~d in ar ar~plifi atic ltl ,f 3,C1(~C) t.:. 4,t)t-)t~ pprn, ,-.r alrr,.-.st a fal-t,~r ,~f tw.~
Additi~rl ~f larg~r particl~s, e.g., p.-~lyavidit- p.~lyn,ers r"ade by cr~~,ss-lin~ir7g avidin m.-~lecules with gl-ltarald~hydæ, wl-.uld pr~du,-~an even larger signal. It sh~LIlri b~ t-t~d that this ~rnplifi~-ati,_lr e:,;ampl~ is a variaticn .-,n th~ sandwi,-h assay. The first, ir"rr,..bili~ed, at-tiligat-d was F.AX, the ligand was bi~~tinylat~d lAF, and th~ se-ct~d, r~ n-itlt~rf~rirlt3 a~ltiligand was avidin.

EXA~FLE_~
In this hypoth~ti~~al e~,:ar"ple, th- pr.-senl:e .-.f tw-. tæst ligands, I~AF and beta-gala,-t"sidasæ, atld a p,-.sitive ,-..rltr.-ll ligand~ DNP-~valburnirl, are det~-t~d simultanæ.-,LIsly in ~he sari,~
fluid sample~ The ~ rldu~tan ~s .~f test ar-d p~.sitive ,_,_,ntr,-.l bi._.regi.~ns are ~._mpared t,-. that ,~f a r\egc1tiv~ ntr,-.l. Th-.ls thL
apparatus ,-._lt~tains fllur c~lls in all. A multiple IVEP s~ns,-.r apparatus i5 used in whi.-h each ~ell has a s*parate fluid path in the rna~lr~r ,-.f Fig. ~A but shaYes a ..~r"n,.-.n el~.trical path with tile .ell with whi,-h it is b~ g .-.~r.,pared in the rf,a~ner .:,f figs. ~A and S~ That is9 each test ~-~r positiv~ c,~ntrc~l c~ll is plal-~-d ~l~--triL-ally in s~ries with the rlegative c.~1tr~l dLIring a m~asurement as d~s~ribed suera. The apparatus is ~:unn~cted t,~ a cc,nductanc~ bridgæ of ttle type described itl Exan~ple 8 to rnc~nitor ~o~lductarlc~ changes.
As in the pr*vious exan,ple ~AX and ~AX are applied to control and test bioregio~lsO 50 ul of a 2.5 mg/n~l solution of anti-DNF
antibody is applied to a third bioregio~-~ which serves as a p"sitive co~ltrolO S') ul of a ~5 n,g~ml soluti~t- cf antibody tc the et-2yn,~ beta-~ala,tosidase is applied to anc,th~r bi"rec,i-n, whi h serves as a second test bioregi~_,tl.
The bi~ iyer c-tltait~ing the f~-ur protein-cc,ated bi_r~gi~tls i5 washed as in Exar~pl~ 1 and pla,-ed intc, the s~nsor. ~un~ling buffer i5 flowæd through ea,-h biolayer as in E~,;arnple 1 at a rate ,,f appro~;irf,ately 50 ul per mitlute, and t~e ,~ctldu,-tatlcæ nf æa h test cr positive cotltrol ,_ell is monit,-,red in turn, about orl,-e every s~,-,-,nd~ erl ,:ondu,-tan,e ratios have stabili2ed~ lC)O ul of a fluid sarr,ple is injected into the sens,r apparatus via a 100 ul sarnple loop. Th~ sampl~ contaitls rutlni~g buff~r and, i~l additi ,~, 1~ ug of GAi~, 6 ug ~f beta galact sidase~ and ~ ur~, ~f ,valbur,lirl highly substituted with DNP~ Fir-,. 11 shows thæ respcnse e~pected for each sensc~r. The ~nsor ,_ontai~ling a~lti-beta-gala,t"sidase ræspotlds spæcifically to beta-galactosidase at the sam~ tirne as the sensor c ntaining ir"mc,bili2ed ~AX respotlds specifically to A~. The positive c~ntrol ,an serve both to estirhate t~e flow rate thrc,ug,tl thæ sensors and to calibrate the sensors. This hypothetical experirn~rlt shows the way in w~lich tl-,~ pres~tlt invention ,an simultaneously rneasilre rnultipl~ ligands in the same fluid, and also tæst for th~ proper wor,king and detailed fu~lction of the apparatus through use of a positivæ contrcil.
-~8-( 3 . -EX AYELE_3 In this e~peYirr,e~lt, th* presæ~ce of ~iAF antibody was agai~dete. ted in a fluid sarnpl~a The EVEF ~pparat~ls of Fig. ~ was used~ e~ ept that th~re was n.- insulati~g layer :~60 .-,r bi.-.layer 355 as su, h, th~ test arld .-.~ntrol bior~gions b~in~ pla~æd individuc711y in tt~e sæns.~r apparatus. Thr~ apparatus was c,nn~cted to a c~_lnductatl,-i bridge of th~ type des--rib--d i~l E~;ar,lple 3.
L,-,.-aliÆit~g m~at~s ~rrlprising dis s 5 rrrr, itl diarflet~r at-d 150 uri thi. I: weræ pu~l~hed out ..f H~ typ~ filters ~C~.~5 urr~ p,.re si-e!
.-ot~sisting .-.f r,li~;ed esters of nitro-~ellul._~s~, pur~ ~lased fr.-.m th.
Millipore C "rp,.-ation, ~edf"rd, Massa~-hus~tts. Sul_h dis 5 w~re thus 1~.6 mrr,- in area, and ~ mm~J or approximately 8 ul, in v.~.lurr,e. The a tive matri~; ar a throLIgh whi--h fluid sar,lple fl..ws was detern,ined by the .-ross se~ tiot-al area "f the it-let at~d l~utl~t p,~rts ~:1 n,m dia. :~ and was ab- .ut 1~. 7~ rnr,i--. Thus the pr~det~rn,itled re~gi.~.n v.~.lume was about b 11 rrli~r--lit~r7 ~.r 110 nan--.l iters.
Ea._h 5. r"rr, dia~ dis-~ -an bind appr.-.:,;ir"at~ly ~':' ug _-F pr.~teit~, to each disk, 100 ug of pr--t~ was applied so that the~ fiIter w~uld be .-omplætely saturated. I~is~:s w~re allc.w~d to sit f-~-r 5 mi~7uti-s at r.~or" terr,perature~ ~under .--:lver t--- avoid gross evapor at i otl :l, t hetl washed i rl i ml phosphat e-bu f f eY ed sal i ne solutio~7 iPE~S, 0.60 M Nai_1, 0.~:)1 M NaF0~, pH 8~:~, C). 1% sodium a;~:ide) atld st.~r~d in this salin~ s.~luti~l until ready for usæ.
Sp~?cifically~ to the t~st dis~ w.a9 appli~d 10 ul ~f a ~;olution of . --8 ~7--RAX (purchased from Sigma Chemical Company, St. Louis, Missouri), adjusted to be 10 mg/ml in RAX, 0.60 M in NaCl, 0.01 M in NaPO4, pH 7.3, and 0.1% in sodium azide. To the negative control disc was applied 10 ul of a solution of goat IgG (GAX), also purchased from Sigma and adjusted to be 10 mg/ml in GAX, 0.60 M in NaCl, 0.01 M in NaPO4, pH 7.3, and 0.1~ in sodium azide.

The discs were inserted in the apparatus of Figure 9. The apparatus was assembled, and 50 ul each of BSA, GAX and GAR in PBS were added sequentially to a reservoir chamber containing 5 ml of PBS, to yield final concentrations in the reservoir of 100, 100, and 36 ug/ml respectively. The results in Figure 12 show that these sensors detected a specific decrease in conductance of th~ RAX bioregion as a result of RAX/GAR complex formation.
Although not shown here, the conductance decrease continued for several more rninutes before reaching a plateau at about - 20,000ppm, or 2% total decrease in conductance of the bioregion.

In this experiment, the presence of an antigen, human gamma globulin ("HAX", short for "human anti-X"), was detected by measuring the change in conductance of a bioregion containing immobilized antibody, goat anti-(human gamma globulin) antibody, or "GAH." Apparatus of the type shown in Fig. 5C was used. The apparatus was connected to a differential conductivity bridge constructed by Neil Brown using two conductivity circuits of the type taught in PCT application US 83/01487 (Int~ No.: WO84/01218), published ~arch 29, 1985, entitled Automatic Temperature ~easuring Circuitry. These two clrcuits were compared using a manually switched precision ratio transformer (Dekatran decade transformer from Elec-tro-scientific Ins-truments, Portland, OR) in conjunction with a detection circuit consisting o~ the following: A high~grain pre-amplifier, a band pass filter with zero phase shift at the operating frequency, and a phase-sensitive detector whose output was connected to a strip chart recorder.

Test and control volumes comprised columns l mm in diameter and 10 mm long made by drilling holes in non-conducting acrylic plastic and filled with CNBr-activated bead preparations (Pharmacia Fine Chemicals, Piscataway, N.J.), which served as the localizing means. These beads covalently bind proteins or other molecules containing free NH2 groups. HAX purchased from Sigma Chemical Company, (St. Louis, MO) was covalently bound to gel beads of CNBr-activated Sepharose 4B according to standard techniques published by Pharmacia ~see Affinity Chromatography, Principles and Methods, 1979) to yield a bead suspension containing 9 mg HAX per ml of packed beads. These beads, 0.9% in HAX (w/v), were loaded into the control bioregion.
(In this example, the HAX itself serves as a negative control.) Similarly, fluorescent goat anti-(human gamma globulin) IgG (GAH) purchased from Antibodies, Inc., Davis, CA as a 56% pure preparation, was covalently bound to CNBr-activated Sepharose 4B beads to yeild a suspension containing 9 mg protein and 5 mg GAH per ml of packed beads = 0.5~ GA~ (w/v). The GAH beads were loaded into the test bioregion.
The following protein solutions were added in order to one ml of phosphate-buffered saline solution (PBS, 0.60 M NaCl, 0.01 M
NaP04, p~ 8.3, 0.1% or 0.05~ NaN3) in the reservoir:
(l) Bovine serum albumen (BSA), 100 ul, at a concentration of 15 mg/ul in PBS, (2) G~X, lO0 ul, at the same concentration in PBS;

(3) HAX, lO0 ul, at the same concentration in P~S.
The results illustrated diagrammatically in Figure 13 show that there was a change in the conductance ratio after addition to the sample fluid of the ligand HAX, which is bound specifically by the immobilized GAH antibody, but not after addition of GAX, which is a related IgG, nor after addition of BSA, a typical protein.
In this experiment it should be noted that:
(a) An antibody was immobilized this time, rather than an antigen as in Examples 1 and 3. This supports the aspect of the present invention which says that it is general with respect to binding of either component of a ligand/antiligand complex.
(b) Immobilized antigen was itself used to prepare the null sensor. As long as the negative control sensor is prepared in such a way that it is phvsically and chemically similar to the test sensor [except for the specific binding characteristics of the test sensor, of course], and as long the negative control immobilized substance is substantially free of specific interactions which effect changes in bulk conductance, it can serve as a control for nonspecific binding.

'~' ~ t ~

~ c) Th~ positiv~ ræsponse t,~ HAX was n.~t obscur~d by t~læpræs~nce of Othrr protrins (l~iAX and EISA~ in the fluid sarnplr d~lring thæ m~asur~rr~nt.

.' EXAM~LE_S
Sin.-~ tl~r mrasur~rrl~nt of ligand/antil:igatld bindirl~ according to the prrsent irlV~AntiOtl i5 bas~-d on alt~ratiorls itl tht ælr,-trical c . .ndul: t anc ~ ( or r ~si st atl~ ~ ) 4 f a setlsor ' s t est vol um~, any parti, ulat~ non-~ ond~lcting substat~.:r tllat a, . urr,ulat*s in thr tæst volur"e du,-: to spe. i Fic ligand~antiligand binditlg will alt~r th~
. ot~dLI,_ t anc e rr,~ asur ~r"~nt . F ar t i - l ~s su, h as ~r yt hr ,- yt es, 1 at æx b.:ads, plastic beads, or glass beads r"c-ty be d~t~ct~d .~r may be used as arnplifiratic"l agents t,i allow drte. tion of ~.tl~r sp~ciæs.
Hl-,w~ver, use ~-,f parti. les as ar"pli fi.-ati.-.tl ag~nts rnay aus~
non-sprl-ific filtef plugging undær s,~rr,~ ,-ir,-urr,statl,-tAs. A
strategy wherein insl-lluble particl~s ,-,r otl~er n,-,n-,~c"ldu,_titlg vc.lurr~es are forrr,ed frorr .-clmpletrly s.ilublr rtag~nts as a r~sult ..f ligarld~antiligand binding cc.uld s.-lv~ this pr..bler,.
As an t~;ample~ tll~ t~-yrl-e . atalase is .-ovd}ently linked to gOdt anti-(:rat)bit gan,r,~a slobulit~) atltibody ~l;AR~ by standard n.eth.-,ds ~uch as, rosslinki~lg witll glutaraldehydt. This . cmplr~; is us~rd, for rxample~ as a second antibody to tæst fc~r the binding c,f prir"ary rabbit antibody t o a ligand in a fluid sample wh i--h llas be~om~ boutld to a biorer~icin, tA.g. during a sandwicrl assay~ To drte~ct tllt prt;setlce clf bou~ld I~A~ tal~sr cor~plrx, ~nd llttl~:r th~
presenc~ of bound ligand, hydrogrn pe~roxid~ (O Ol~ is pass~d --~3--, throu!3h th~ sensc,r and is ~4nv~rted to O-, and H~O via the en~ymatic activity c~f the ell~yrne. Th.- evolved 0~ forrns bubbl~s in or næar th~ biorægi"n rnatri~;, a~d thes~ bubbles~ b~:-ing n~3n-conducting, CallSe dëCre.~aSed rl~^ctriral CC~tldUCta/lCe that i5 ri,onit,~red by the apparatus. E~le~ trod--s with r~,~ld bla, k -,r ~arbc~r surfa,-æs aræ used t~ rr,inimi~ non-speci fi~- ,-atalysis c,f H-l-l-b r ~a k d uwn t h at ^~"- ,- ur s wh ~tl p 1 d t i n um e 1 æ,_ t r "d es ar .- us ~d .
This m,eth,_,d p ,tentially~ all"ws fc,r an in, reasrd sensitivity uf d.-te ticltl ,~f ligand/anti-liga~ld bir1din~ ,-f s~væral "rders ,-,f rrlay~litud~ as -r"par~d t ut~ari,pli fif d r"~asUr~-ri,-lt. Th~ t,-,tal time re~-liræd t~ obtaitl t~is r"æasur~r"erlt sh,,LIld be sh~:rt ldppr,,~;ir"atæly minLItes). Alsc~ the me-th"d ~~an be desiyned 5" that it use5 ir~eXpetlsiV~ on-ha_ard,-,us ,-hærr,i, als. A varirty of c,ther *n~yr"es t51at prc,du,-~ gases, an bæ sirr,ilarly usëd~ as ,-an rn:yrrlæs ttlat produ,-r insolubl~ pr~,:ipitates.

EX_Iv~PLE_G
Hybridi_atiian ~f pr,,b~ nuclril- acids (eith~r F~IA i r I:NA) to ,rnpl~rr,etltary nu,-leic a ids irbr,",bili~rd Ot~ an insoluble rnat~ ; is a ,-or"rrlon lab,-,rat,-,ry prc" edur~. Althc~ugh this pr,-,cëdurr is ~ig~lly spe, i fic and s~nsitiYe, it is relatively di fficul t tc, prrfclrm~
~pënsive, and lrnr~tily~ and it may require lar~ arr"~~unts ,_f radica-tive reagents. Dete-tion of tlucl~i,- a, id ~lybridi-ati,_,n by n,~asur ing c onducta~~r changr~s thus - -uld b* us~ ful i f it ould be accornplish~?d rapidly and wit~lout the u~ie f radica~tive tracers.
Xn th~ hypoth~tical ~xampl~ bel~w, syrlth~tic hon~opolym~rs cf --g4--adenylic acid and uridylic ~cid are used in a n~odel ~;ystern to dærr"~lstrate this ~æneral prirlcipl~.
Control arld t~st bioregions ccmprising nitro~ellul4se filt~rs ~5 un, p~r~ size) are treated with polyuridy}i-: acid and p._.lyaderlyli.- a._id, resp-.-tively. UrlYea. ted birlding sites aræ
blo- l~ed by additi.-n of 50 ug of BSA or ~al f thyrr~us DNA to ea.~h bioregiotl. A bi.~layer -ontainislg thæ bioregi~ ls is pla ed into the IVEF særlsor~ and a fl.~w of runrling buffer of 61~ ul/rnitlutæ thru ea._~l s*ns..r is initiatæri to establish a bas~line, Usitlg appr.-.priat~ ~ybridi2atiorl .-onditi..rls .-.f salt .-on.-e-ltrati..n atld ~ernperaturæ f. r ti)æ parti. ular seqLIttl æs and pr-.. æssrs being studiëd. A sar"plæ ..f 1 u~ of p .lyuridylic a.-id in l(:)C) ul ru~l~litlg buffær is thtn addæd to the apparatus. ~ spæ.-ifi.: de-~reasæ irl ~orldu-tance of the test bi..r~gi.-.n as p.-.lyuridyli~ a. id birlds t..
th~ irrlrilobilized polyaderlylic acid indi~ates t~læ o.-._urrerl-æ of hybr i d i z at i ~n .
E:;ar"pl æ 7 Irl this hyp..theti~al e:v;perim~r1t, tht præsen.-æ an:l ~:oncæntrati~.n of a spæ._i fi~ anlirlo a~id, phenylalarlitle~ is detæ~tæd in a fluid samplæ usin~ an equilibriurn, rathær thatl a l~ineti~, ethod~ A dip-in typæ .~f s~nsor apparat-ls is usæd, SUC~l as that shown in Fig. 7B. The apparatus is . otlrlæcted to a ondu. tan.-t brid~ such as that describ~d in Example ~.
T~st and control biorægiorls ar~ . onstru. ted as follows. For the test bior~gi--,n~ atrnarl filter paper hl~ Wnatn,arl Lir"it~d, England) i~ activated to bind Grotein by treatn~erlt with ~yanogen --~S--~2~ 2~
bron-idæ. Thæ ~NBr-aLtivat.2d papær is nlount~d on a~ apprc~priate ~lL~ldi2r or Support~ th~1 æhposæ~d t- ~ s- luti~tl ~f a pr~t~in ti1at 1) binds phetlylala~line with a~ affinity constat t ~ wl~ic~l is similar ir ma~nit~de~ to th-2 C:?r. entration ::)f ph~tlylal~t iniS to b.2 td.2t.2~t.2d and that 2~ does not me~tabJlizi~ phetlylalatlitl~ t fl r ~t.~;amplæ ctle ~ LILllri us* a ph~nylalatlini~ hydrl~;ylas~ that; has b -tn mutat~d t~
inal_tivati2 t~l.2 ~n~ym~ and if desiræd tn ~hatlq~- its ~ffinity ~nstatlt) . Fl~r th~- .~ontrol bi Jregi ~n O~IBr-a~tivat~d pap.2r rr~ utlteSd in th~ hl~ld.2r is ~.~;p-ls*d t a s~JlUtiC~tl clf a pr~-~teirl that dl .~s rl~:lt bind i r r e^tab-li; e ph~nylalanitle ~:fLY ~ ;ar pl* ph*tlylalatlin~
hydr ;ylas.2 that has b.2.2n rrutat.2d t n/2it~l~r bind nl r rn*tab. l i~t phl2tlyldlanitl~ rray bL use;d:!. Ei-~th bi r-t3i-r-; are wd5h.2d thtn th*
hl ld~r is pl-siti.-l~ed s that t~l~ bil.r~-gi~~~n li.^ ~v.-r t~l-- s~nsitlq me2ans Qf thr- apparatus f-rr it-g t~le2 se2ns~r. T~le se~ns r is tdipp~d irltl~l l~tle r l ~ f d bufftreSd salitle~ s ~luti~l whllse; ~ ~-ltlductivity is mat hed t-l that c f t~- fluid sar pl* t-l be ttsted fLr t~l* prc S*tll~ë
cf ph~nylalatlitleS~ and t~lt c~-~tldLl~-tatl~-eS- rati S :.f tne teSSt and .~.~.ntr.-l bi. r*~i.-t-~s is m~asure~d. After th~ val le - f th* rati - has stabili~d t~. 1 ml -f th~2 fluid sample is add*d with stirrinq atld th.- ~llatlg~ itl t~l~ l-LlndLll-tarll-i~ ratiu i5 rl~-~t~d. On~:~ the2 rati.-. has stabiliz~d at3aitl t~l* hatlgt- is -ompar.2d to a tandard .~urv~ tv-allQw dë^termitlatil~n :f t~ on:~-rltratic~n ;:~f p~l*nylalanitle in thi2 fluid sarrlpltS-.

_g~;_ EXAMPL~ ~
A pr~ferred e~bodiment of an instru~æn~ for measuri~g the relative tonductance of a test volume ~nd ~ control volu~e i3 described below. Thi~ embodiment is u~ed whe~ the test ~nd control cells are electrically in seriesq Specifically, the instr~ent includes circuitry for ~easuring the relative change in conductivity between two cells which include ~ test bioregion and a control bioregio~ as the test cell bioregion conductivity c~langes as a result of ligand molecules binding to antiligand sites in its bioreyion. The instrument includes circuitry for driving a current through æach of t~ie cells and for me~suring the voltase drops across the test and control cells as a function of time. Th~ instrument i5 capable of measuring a change of 10 4 in t~l~ relative conductivity b~tween the two cells with an ~ccuracy of appr~xi~ately one p er c ent .
The ins~rument includes a two--stage A.C., ar~al~g--to--digital conversion circuit in which the initial conductivitiæs are measured via a successive approxi~ation process, while She relative cond~ctivities 2Ye monitored durin~ a test interval via a tracking technique~ The circuitry i5 controlled by a digital controller which can vaYy the parameters of the n~easuren~ent processes ~o as to ~ake the ci~cuit ad~ptable to ~easu~e~ent of a wid~ varie~y of ~iubstances using different cell conlFigurations and~or electrolytes~ E~tensive calibration and dia~nostic capabilities are provided t~ ensure the æccur~cy of the me~sur emerl t 5, Fig. 14 is ~ r;chematic dia~ram of circuitry whi~h can measure the relativ~ change in conductivity betwæt!n the test and control .
cells. As in Fig. 5B, the resistances of the test and csntrol cells are represented by resistors ~x and Rc; terminals V1 -- VJ, are the voltagæ me~sureme~-t termin~ls, terrninals C1 and C2 are thr current input terrninals~ and resistors 1~ 7 repre~sent the~ ilhpedanc~s of the fluid ~hannels~ which includæs contributi.~ns of the fluid ilr~pedances9 polari~atior. ir~.pedan,-~s, and other ef fectsO
In Fig. 1~, a signal generator 2 00 provides a very puræ sinæ
wavæ output sign~l at a frequency of 384 H~. Whil~ ~ higher frequæncy would mak~ the dæsign of the n)easurernent circuitry more simple, prcblems with maintaining the accuracy of four--terminal m~asurements which r~sult fron~ capacitatîve effec~s and other error sources are lesser~ed with the use of low~r frequencies.
The signal generator output signal is applied to terminal C1 through one of resistors 2~02, 2204, or 220~, ~a5 selected by a swi t c h --208 . I n t he desc r i bed æmbod i nnent, r esi 5t; c r s 2202, ~204, and ~-06 have values of S~C, 10~ and lSY ohms. Swi tch ~ 08 and resistors ~!202~ --06 are ~sed to m~tch the outp~Jt signal to the impedance of the test arld colltrol cells, dependin~ on the e1ectYolyte used and the cell configuration. Although switch 2208 is shown as a n~anually-operat~d switch in Fi~. IC" it could be controllelil by the digital proces50r in altern~a$e en~bodiments..
Th* si~n~al at ter~ninal Cl is also applied to a unity--~ain --9E~--`

buffer amplifier 2~10 which provides a buf~red output sign~l .
EC1 which is ~q~al to the vo}tag~ at t~rmin~l Cl.
Terminal C2 is driven by the output of an ~mplifier 2212. ' As discussed in more detail below, amplifi.er 2~12 dYives termin~l ~2 50 th~t the voltage at C2 i5 equal in r,lagnitude but opposite in phase ~o the voltage at t~rrninal Cl which c~us~s a current to flow fro~ terminal C1 throuyh t;he test ~nd control cell5 to C7, as described above in the description of Figs. 5A
a~d 5B. Noise pickup is minimized by providing a "balanced"
drive to the ~urrent terminals Cl and C2.
Th~ vol ta~ b~tween terrni nal s V1 and V~ i 5 applied to a high-input impedance, di fferential a~npli fier '7213 ~omprisæd of amplifier 2214 and transformers Z 16 and 2218. This circuit is discussed in mor~ detail below in connection with Fig. 16. The voltage across t~rminals V3 and V4 ss applied to a similar amplifier circuit 2215 including transformers 2236 and 7238 and a~plifier 2240. The operation of a~plifier circuit 2215 is similar to that of amplifier cir~uit 22i3. In practice~ the app~ra~us ~ontaining the te~t and contro1 cells may b~ some distance away f~om the ~lectronic measur~ment circuitry~ The signal~ from termina1s V1 - V~ are applied to the differential amplifiers ~213 and 2715 throuyh shielded cables Æ220~ 2222, 2232, and 2234.
Ter~inal ~ is connected to the signal or center conductor of c~bl~ 2220~ and the signal at terminal Vl i5 applied to on~
~de of the pri~ary windiny of transformer ~216. The ~i~nal on --9~

the cer~ter c~nductor of cable 2220 i5 applled to the input of a unity gain, high-impedance~, buffer ampli fier 2:;~24, which is physi c al 1 y 1 oc at ed c 1 ose t o t h ~ el ~c t r on i c mea~;ur em~n'c ~' circuitry. Th~ output o~ amplifiær ~ is connected to tSle shield olF cable 2220 and maintains the voltage on ~he shield at the same level as the signal on the si~nal c~rlductor of cable ~,00 As shown in Fig. 1~, thæ output of a~plifi~r 2~24 al50 drives a shield around the prih~ary windin~ ~f transforrf.er 2-'16 and a shield around the wire conne~cting the windings of transformers 2216 and 221~. The input to a second unity-gain buffer amplifiær 2~6 i5 connected to the signal conductor of cable 2222 ne~r its connection to transfor~er 221~. The output of amplifier ~26 drives the shield of cable ~. Providing an active drive for the shield conductor reduces noise pic~up and n~inimizes l~ss~s caused by capacitative leakage in cables 22~0 ~d ~
The shield cor~ductors of c~bles 2220 and 22~ may be connæcted to guard rings in the test cell to provide furthær r~d~ction of errors. Thi5 may be better understo~d by referring to the diagram of the test atld control cell configuration shown in Fig.. 8A,. As discussed above, fluid seepage around t;~e o-ring seal~ may result in recistiv~ leakage paths which will cause eYrors in the c~nductivity measure ments. The guard *lectrodes ~1 - G~ reduce or ~liminate s~ch errors. Th~ guard ælectrodæs ar~ driven by the 5i~nal5 on the shields of the cabl~s t~ each of the t~rmi~al6. The resist~i~sce of e1ectricaI paths pro~ ced by --~00--l~ak~e around the o-rings will be rel~tively high~ Since the guard electrodes ~re main~ained at a voltalye ~qual to the vsltage on the associated terminal ~1 ~ V4 by the low ~utput ~, impedance of ampli~iers ~224 ~230, any current flowing bet ween the two c~lls will be principally provided by the guard electrodes~ This reduces or eliminates errors whi~:h othærwise woul d be caused by el eCtY ical leakage alon~ paths resulting ~rom fluid seepage ~round the o-ring seals.
Signals El and E~ frorn the output~ of ,ampli fiers 222~ and 22:~; are buf fered signals equal or approxirnately equal to the v~ltages at terminals Vl and V2 . Si~nals GLl and GL2 are appli~d to amplifier 5 2~-4 and 2226 dur i ng d i agnost i ~
rc~utines to enable the leakage current fro~ the ~hiæld and,guard electrod~ circuit for each an~plifier to be measured, as discussed below~
Th~ voltage across terminals V~ and V4 is processed by circuitry similar to the circuitry dæscribed abovæ connected t~
terminals V1 ~nd V2 . The signals on terminals V3 and Y4 are conn~ct~d to the primary windings of transform~rs 2236 and 2~38 by shielded cable~ 2232 and 2234~ The signal~ on the center conductors of cablæs ~232 and 22~4 are appli~d to amplifiers 2228 and 2230 which maintain the shields of the cables at a potæntial equal to the sig~al on the associated termi~al and which pro~ide buf~ered signals E3 and E4 egual to the voltage~ at t~rmi~als Y3 and V4~ Amplifiers ~2.8~ 2230, and 2240, transformer 223 and the primary winding o~ transformer 2236 ar~ e~sentially id~tic~l to the c4rresponding circuitry ronnecked ~o termin~ls ~1 and V~ ex~ept ~hat terminals V3 and V4 are connected to thi~ circuitry with th~ ~pposit~ polarity. In ottler words, the output signals from trans~ormeys 2216 and 2236 produced by a current flowing throu~h th~ test and control cells will be Opp 05 i t e i n ph ase f r 4:~m eac h ot h er .
As ~escribæd aboveg it i5 desir~ble to apply a voltage to ~er~inal C2 which causes the signals at V1 and V~ to be equal in magtlitude b~lt opposite in polarity. This is dcne by me~ns of a negative feedback loop provided by amplifier 2212.
The El signal is appliæd via a resistor 22~8 to the input cf ampl i f i er 2212. The E4 si !3nal i s 5i Ihi larly appl i ed to the input of a~plifier 2212 via a resistor 2250. A~plifier 2212 i5 a relativæly high-g~inl invertins, AC amplifier. In the described e~bodiment, amplifier 2212 has a gain o~ apprcxi~ately 10~000 at the operating frequency of ~4 Hz~ The El and E~ input signals applied to amplifieY 2212 are a functio~ of the current driven through the test and control cells by amplifier - 21~7 and thæ circuit forms a closed lo~p feeclba~:k circuit which forces the signal level at the input of a~plifier 2~1~ to zero or ~rou~d potenti~l. In order f~r this to occur, the E4 signal must be eq~l ~nd opposite to th~ E~ signal~ which i5 the desired c~dition.
Each of ~he signal conductors of the cable~ may be sel*ctively conne~::ted by switctles ''~46 and resistors 2242 to a ~ec:~ node 2:244. Alth~u~3h not shown in Figla ~4~ the connectio~

--10~--of switch~ 2246 to the~ cables should be physit:ally cl~s~ to the inputs of the bulFfer amplifiers ~224 2~!30. In ~he descrihed embodi~nent~ resistors æ~2 each h~væ a v~lue of 1 kilohm.
~;wi'cc:~les 2246 ~re e~iecltronically controlle~ switches. Th~ four switches 2246 ~re respe~ctively contrvlled Iby signals SE, SF, SG~ ~nd SH from the digital controller. Switches 22~6 a~ay be selæctively closed durir~g diagnostic routir7es to provide a known resistance betwe~en th~ 5i~nal1 conductors of the cables. By comp~ring th~ volt~ges at terminals Vl - V~ with switches 22~6 open and closed, læakage paths between the termi~als ~.~y b~
determined, as discussed in more detail below.
The output winding 2:~70 of transfor~ner :i~216 has 12B turns ~nd proYides a signal proportion~l to the voltage drop across thæ
test cell. Transformer 2236 has multiple output windi~gsy each of which pro~ides a siynal proportional to the volta~e dlrop ~cross th~ control cell. ~y sel~ctively connectin~ the windings of transformer 2236 i~ s~ries the output of tr~nsfor~er 2~36 may b~ scal~d with respect to the output of transfoYmer ~Z~6~
Transformer 2236 has a 6~ - tuYn output winding ~272 one end of which is grounded. The second end of 64-turn winding 2272 is connected t~ or~e end of output winding 2270 of transformer 2:~16.
Tr~ns~ormær ~2~6 additionally has six binary-wei~hted winding~
2274 -- 22E14 having respectively 64, 32, lfi~ 8, 4, and 2 turns.
Switches 2275 through 2~E35 are sin~le--pole do~bl~--throw siwitches ~nd are typically low--P~ois~ FET switches which are e1ectrorlica11y contro11~d. Ir~ the præs~ærlt e~bodirn~nt, 5i~t sign~ls Sl - S~;

~Z~91,26 are applied to switches 2~75-2285 by the digital processor to control the states of the switches. These switch~s ~re connected to windings 227~-2284 as ~hown in Fig. 1~ 50 that any cornbinativn of the ~indings may be connected in series, depending on the settings of signals S1-S~. Thus, by appropri~tely settiny the swi tches Sl - S~, any even number of turns frorn 2 through 126 may be connected in series to provide an outp~t signal which is proportional to the voltage drop across the ~ontrol cell biolayer and which m~y be scaled over a rangæ of of 1:126 The si~nals from the vario~s o~tput windinys of thæ two transforn~ers 2216 ~nd 22~6 are s~med in the followirlg manner.
Olle end of 6~-turn winding ~272 on trarlsformer 2~3~ is grounded.
The second end of winding 2272 is connected to one end of the single ouSput winding 2~70 on transforn~er 2~16. The windings are connected 50 that the two windings are connected in seri~s and with the same phase. The second end of winding 2270 i5 connected to the ~ommon ter~ninal of switch Sl so that the selected windin~s of binary--weighted windings 2274--2:284 aYæ conne~ted in seriæs with windin~3s 2270 and ~27~. The output fror~ windings 22~4--22E3~ is also in phase with the outp~t from winding 2270.
Since, as ciescribed above, the input signals to transformers ~216 and Z36 are opposite in phase, the output signals from the two transformers aræ also opposite in phase.
The net e~fect of the above-described connection i5 that the signal fro~ output winding 2270 i5 connected in series, and thus subtracted tdue to the oppositely-phas~d outputs), fYom the . -104-~2~

OUtpLlt signal lF~oa~ 64 to lgO turns of transformer 223~y depending ~n the setiting~; of switches Sl ~ 56~ ~n other words9 by properly setting switch~s Sl ~ S6; the output signal from transformer 22:~:6 may be scaled so tllat it :i5 approxim~g;ely equal to the OLltpUt signal frorn l;ransforr,~er 2216 for voltage àrops across the test cell ranging from about 30';'. to 150'~ of the volt~ge drop across the control cell. The ~;ignal on line 2290 represents the exact difference of the output signals from transformer 5 ?2 1 6 and 2236.
The output frorn the 6~-l;urn winding 2r74 is applied via a unity - gain buffær amplifier 2294 to the analog input of a l~-bit~
multiplying digital-to-analog converter ~DAC~ ~_96. DAC ~2g6 ~ay be implement~d, for exa~pl~, by an ICL 713~ i~tegrated circuit.
Fourteen digital input signals D1 - D14 are applied to the DAC by the digital processor. In ræsponse to signals 1'rom the digital processor, the o~tput from windiny 2?72y which is proportional to the voltage drop across the control ~ell, may be scaled over ~ range of 214. The output from DAC ~296 i5 appliecl to a resistive~ 64--to--1 divider made of resistors --298 and 22g9. This divider scales the output of the D~C so that one MSB
of the DAC i5 eq~livilent to the output from one turn ot th~
windings on transformer Z~36. In this manner" the selectahle output wi~ldings of tr~nsformer ?2:36 ~r3d the DAC 22~6 allow the ~utput from the control cell to be scaled over 3 220--bit ran5e.

--10~--~%~ 6 The output from resistive divid~r 2~g~ is applied to o~e input of a ~ompar~tor 2292. The s~cond input to the cornparator i5 selected by switches 2 87 ~290~ Th~se switches are respectively controlled by ~ign~ls SA-SD fro~ ~he digit~l pro.-æssor. Closing ~witch SA connects the second input to comparator 22~2 to ground; switch S~ connects the inpu~ to the ~utput of 64-turr7 winding 2~7~; SC to the output of 2-turn winding 2~6; and swit,-h SD to the output of seriæs--connected windings -270, ~27~, and 2274-7?94, 2S dætærmined by the s~tting of swi tchæs Sl-S6. Swi tches SA-SC are u5el:1 du~i ny calibration, and the operatio/l of th~se switchæs is discussed in d~tail in connection with Figsn 20-~2.
During conductivity rn~asuremænt 5, 5W itch SD is close~ to COtln~Ct the se~c~ild itlpUt to the comparator to thæ
serie~-conn~cted windings 2~70~ 2~7-, and 2~.74-2784. In this mc,dæ, thæ test cell output signal from 128-turn winding 2~70 is e~fectively subtracted fro~ the control cell signal as scaled over a ran~e of approxin-ately 0.5 to 1.~9 depending ~n thæ
signals S1-~6 applied to the selectable windirlss ^~74~ 4.
The 51 - 56 signals ar~ set by the digital processor 5CI that the control cell output fro~ the windings of transformeY ~^36 approxi~ates as clos~ly as possible the t~st cell output si~nal from trans~ormer -216. The first input tc the compar~tor is provided by the DAC ~2~6 scal~d by 64:1 di~ider ~9~-~2g~. The ~xi~um output from the divider is eq~al to on~-hal~ o~ th~
s~allest in~reme~t provided by th~ settings o~ swit~hes ~;l s6, Thus, ~h~ ~ornbination of th~ six MS~'s provided by the trans~ormer winding swit~hes Sl-S6 and the 14 LS~5 pr~vided by the DAC provides a 2~0-bit c~nveY5ion ~ the test cell signal e~fectively using tlle control c~ ignal as a refer~rnce ~olt~ge for the conversion. The describæd circuit takes ~d~antage of the high accuracy of a precision-ratio transfor~er in deterrnining the ~ost significant bit~, which do ~ot change ofte~, while allowing fast tracking of the changing test cell si~nal by using the multiplying DAC 22~6 to provide the le~st signi ficant bits.
Fig. 15 shows the comparator circuitry in ~ore det~il~ The two i~puts to co~parator ~2~2 are applied to two, u~ity-gain, buffer amplifiers ~300 and 2302 whose outputs are applied to a pr~cision differential amplilFier 2304. The outpu~ 4~ ~mplifier 2~0~ is directly appli~d to one input of a multiplexer 2~-~4. The output of amplifier ~'304 is also applied to multiplexer 2324 via two amplifier stages 2306 ~nd 2308l e~ch of which have a gain of for a total gain of 64. A yain control sig~al G~l is applied to multiplexær 2~24 by the digital processo~ to sel~ct ~ gain of 1 o~ 6~ for the output signal from ~mplifiær ~304.
Amplifier st~ges 2306 and 2:308 are made up of op--amps 23?0 and 2322 connected as inverting amplifiers as shown ir- Fiy. 15.
Resistors 2~10-2316 are cho~en to give a gain of 8 for ~ach a~plifier stage. Opposed diodes 2~18 are connected in parallel across th~ feedback resi~t~r to the first amplifier stag~ 2~06 t~
limit th~ output signal amplit~d~. Although a gai~ vallle should be selected t~ keep the op~rati~n ~f atnpli t`ieY stages ~306--230B

and the following circuitry in the linear region, noise spike may be present. If such spikes were to drive any of the circuitry of comparator 2292 into saturation, large errors might result. Diodes 2318 serve to suppress any such noise spikes.

The output of multiplexer 2323 is applied via buffer amplifier 2326 to an AC bandpass filter circuit.
It is important that filter circuit 2328 has a high atten~
uation of frequencies removed from the signal frequency of 384 Hz while maintaining a very flat phase characteristic.
One circuit suitable for implementing filter 2328 is the bi-quad amplifer circuit described in U.S. patent No. 4,539,525 for "Bandpass Amplifier Filters" issued September 30, 1935. Other circuits known to those in the art may also be used to implement filter 2328.

The output of filter 2328 is applied to one input of a multiplexer 2330. Other signals applied to multi-plexer 2330 include the E1 - E4 signals, representative of the voltages at terminals V1 - V4, the EC1 and EC2 signals, representative of the voltages at the current terminals C1 and C2, the Fl and F2 signals, representative of the flow in each of the test and control cells, and ground. Multiplexer 2330 normally connects the output of filter 2328 to the input of amplifier stage 2334 when the circuit is measuring the conductance of the test or control cells. The digital processor provides the appropriate SEL input to multiplexer 2330 to measure the other signals applied to the multiplexer for calibration and ,A

dia5~nostic purpc)se~ir a5 discussed below~ It should b~ not~d t~at .-ll;he 5ignal5 applied to rnultiplexer 23~0 are A~ sign~ls which must be demodulated prior lto being me~sured~ ~' The output of multiplexer 2330 is appliæd to the n~n--invertit~g input ~f an op--amp 2:332 of ampli fier stage ~33~. A
feedba~k resistor 23.,6 i5 connected between the output and i nver t i ng i nput ~ f t he op--ar,)p . A 5eC ot~d r {~5 i st or ~338 i s connected to the inverting input of th~ op--amp a~d is al50 selectively conne~ctLd to ground through a 5witch 23~0. The ration of resistors ~33 and ~38 is 8 to 1. I~y slosing switch 23~0, a gain of 1 or of E) may b* se1ect~d for anlpli fi~r stage 2334. Switch ~340 i5 typically a low-noise FIET switsh whish is controlled by a signal GC-- providæd by the digital proc~ssor.
The output frorh ampli fier 2334 is applied to onæ it~put o~ a demodulator circuit ~34~ dire~tly, ancl to a seco~ld input of the den~odulator via a unity--gain ïnværting a~pli fiær stage ~:3~4, made~
up of op-amp 2:350 and ræsi stors ~ 46 atld 2348. Demodul ator 234 m~y be i~nplemented~ for ex~mple, by AD 534L bala~lcæd multiplier circuit. ~nverting an~plifier 2~44 provides a signal æqual to the signal from ampli fier 2~.,2 but 180 degrees out of pha~i?. Thi s signal i5 applied to the second input to the balan~ed dea~odulator~ It is desirabl~ tl~ use a balanced type of d~modulator to provide as stable a sigt~al as possibl* at the dem~dulator output. The output sigr)al frosn the demodulator represents the v~lue of the di f t~ærs~nc~ between the test and trol cell v~ltag~s.

. . ~ , ( ' _ The r*fer~nce i~put si~nal to de~odulator ~342 is selected by a multiple~er 23S2. The si~nal ~ro~ multiplexer Z35~ i5 a reference clock signal d~rived ~rom ~ignal genærator 2~00 w~ich i5 us~d to demodulate the AC signal applied ~o th~ balanced input5 to til~ demodulator. A 0-de~r~e ref~renc~ signal taken directly from the signal generator is applied to one input vf th~
multiplexer. This siynal is normally us~d to demodulat~ the input to the demod~llator, as disc~ssed below; when the conductivity of t~ corltrol and ~st cells i5 tne~sured.
Although, in theory, the signals applied to the comparator circuitry 2~2 should be e~actly in phase with thæ output si~nal fro~ signal g~nerator 2200~ s~all phase shift ~rrors may be introduced by parasitic irnpedancæs in the circuitry and othær ærror sourc~s. The~se phase errors can c~use errors in the amplitude of the demodulated signal, since th~ refer~nce signal will no longer be e%actly i~ phase with the signal to be demodulate~. Two oth~r reference signals ~ay be selected by ~ultiplexer 235~ ~hich allows such ~rrors to be measured so that p~oper compensation can be roade..
~ ~econd signal is applied to the multiplex~r by limiting ar.lplifie~ ~354a This signal is designated as the r~ndom reference signal ~ Si?~Ce its phase relationship to the clock sig~l is not exactly known. The output f~om ~ultiplexer 2330, which is the signal to ~e de~odulated, i5 appliæd via capacitor 2360 and resistor 2358 to one input of an op - amp ~354. Two oppo5~d diodes Z~S6 are conn~cted betwe~n t~ input of th~ op-a~p and ground to li~it the ~nput voltage to the op-amp~ The op-amp i~ operated in open loop mode and proYide~ at it~ ~utpu~ a sq~are w~ve sigt7al in phase with ~he ~utpu~ ~rom multiplexer 2330. ~y demodulating the output from multiplexer ~3~0 with this signal, the rnagnitude ~n be determined independerlt of any w~known phase shifts which the prior circuitry may ha~e introduced.
A third signal i~ ~pplied to multipl~xer 2352 from ~
gO-degree phase shi~ter circuit ~62. The 0-degree reference is applied to the input o~ 90-d~gree phase shiFting circuit ~362.
Phase shifter 2362 provides a precise 90 degree phasæ shi~t to the 384 Hz input signal and provides a refeYence clock signal in phase quadratur~ to the 0-degree signal. Dur~ng diagn~stics, the input to the demodulator may be demodulated using the random and gO-degræe refer~nc~ signals as well as the 0-degree ref~rence to deter~ine exactly the ~uadrature co~ponent introduced into the signal t~ be m~asuYed fro~ the ~st and cotltrol cell~. The digital prvcessor may use this i~formation to ensure that the quadratur~ error i~ not 50 large as to result in er~oneous measurements and optionally to correct the measured co~ductivity value~ i- such corrections are needed.
The output from the demodulat~r is applied to a pr~cision r~set-integrat~r circuit ~36~ Integrator 2364 is implemented ~sing a high gain precision amplifi~r 2372, input resistor 236~, and integrating capacitor 23~6. Capacitor 236~ should be a ~ polypropylenæ or other si~ilar type capacitor ha~in~ high stab~lity~ Typical value~ for re~i~tor 235~ a~d capa~itor ~366 , --1 1 1--ar~ lO0 kilohm~ and l~0 microfarads~ ~n elecgronic ~witch 2370 i5 c~nn~cted across the capacitor and i5 contr~lled by an integrator reset ~ig~al IRo A sec~d switch 2378 i5 connected in series with the input to the integrator and is controlled by an inværted IR sig~lal fro~h a~l i~verter 2~80. Thus~ when the IR Signal is inactive, switch ~370 is clos~ t~ reset t~le integrat~r whilæ switch 7378 disconnects the input si~nal to the integrator. When the IR signa:l goes active, th~ input sig~lal is applied to the integrator and the reset switch ~370 is opened, allowing the integrator 2~64 to integrate and filter th~ output signal from thæ de~o~ulator.
The ~utput signal from the de~odulator is applied to the input of an A~D conver~er ~376. Converter ~376 ~ay be imple~ented, for example, by means of an AD574AJ~ 12-bit bipolar A~D converte~r circuit manufactured 3:1y Burr Brown.. This corlverter p~rfor~s a conversion in approxi~ately 25 mi~roseconds~ whicll is essentially inst~ntane~usly with respect to the rate o~ chan~e of the output from integr~tor .,::~6~. A~D converte~r 2376 is controlled by sig~als to and from the di~ital proc~ssor, represented ~y the ~TL signals to and from converter ~37~ in Fig.
15. The digital output values ar~ read by the digital procæssor after each convers;ion i s per f~rmed.
As will b~?come more clear after reading the discussion c3f the rneasLsreDnent procedures explait ed below in con junction with Fig~i.
20-73~ the present invention allows a v~y 1~3w nois~ measuræn~nt to be made o~ th~ test and control cell conductivities using a 2~

~ethod which requires essentially ~o settling time during ~he period that the circuitry is ~racking the ~hang~ in conductivity of the test cell~ Further~ the finite ~i~e integration provided .' by integrator circuit ~6~ provide~ near opti~al filtering of noise which may b~ present in the signals ~rom the cell~.
Fig. 16 shows further dætzils of inp~t a~pli~ier cîrcuitry 2~13 connected to terminals V1 and V~ ~ ~he operation of the amplifier circuitr~ 2215 connected to ~er~inals V3 and ~4 i5 similar, except that transfor~er 22~6 ha~ multiple se~o~ldary windings while transfor~er 2216 h~s a single secondary winding, as discussed above~ In order to r"aintain accuracy~ i~ is important the the output signals from transformer 2216 have a precisæ rel~ti~nship to the signals from transfor~er 2~6. By using precision-ratio transfor~ers for transforn~ers 7~16 and 2236, the outp~t voltages frc~ the various windings can be n~ade accurate to better than one part in 10~.
In Fi~ 16, tra~sformer 2218 has an input winding with two taps to provide three winding segments 24Q~, ~404, and ~06 having 165~ i25~ and 235 turns respectively, and a single 1650-turn OUtp~.lt windi~g 2408. TransformeY Zi5 i5 a precision ratio transformer and has a 224-turn input wi~ding ~424, a 128-turn output winding 2410, discussed above in connection with Fig. 14~ and three single-turn output windin~s ~412, 2414, 3nd ~J,16.. Orle end ~f winding '`416 is conne~:ted t~ the lower tap of trans~ormer 2218 via a:~ switch 2422. Sin~ilarly the two taps to th~ conn~:tions~b~tw~en wiMding~i 241:2 and ~41~ are s*l~ctively connected to lthe tws:~ taps on winding 2~18 via switche ~418 and 2420. One of switches 2418--2~22 is ~lo~;ed during operaticin ~o s~l~ct the gain of the c~mpli fier ~tage. The imp~dance across th~
wir~dings of transformer ~218 is much larger tharl ~he impedance across the single t~rn windings ~12-241~, and transforr,ler 2~1E3 ensures th~t a high enough impedance is provided to the sensors.
Transformer 221~ al~;o serves to isolate the four el~ctrod~s from t h e anal og g r ound A
The secondary windin~ of 1;ransformer 2~18 i5 connected to the input of an AC ampli ~ier 221~ Ampli fi~r 2~14 has a v~ry high gain at the opæraticn frequency of 384 Hz, typically on th~ order of lOOyOOO~ while maintaitlitl~ a v~ry stabl~ phase shi-ft. Th~
output of an~pli fier 2214 is connected to input winding 2424 having ~2~ turn~; on transformer 2216. Th~ input of ar~pli fi~r Z224 is connected to the line going to t~rminal Vl. Its output às connected to the shields of the cables providing the input ~o tran~foYmer ~18., These shields include t37e shield of line ~220 and the shields c~f ~ach of th~ singl~-turtl wi~dings 2412-2416.
The shi~?ld of line :2~ to terminal V~ is similarly drivæn by arllpli fiær 22~ not shown in Fig. 16. The input windin~s ~0----406 of transformer 2218 are shiælded, and this shield is also connected to and driven by ampli fier 2Z2~i.
Amplifier 2213 serves to pr~vide a high imp~dance across terminals Vl and V~ and also provide~ a relatively high :impedance ts~ th~ input of ampli fier :Z:214. The gain of ampli fier 2213 i~; det~rmin*d by th~? r5~1;io o~ s;ingle--tur~ winding~ 24J2-24 1, I

-: ' - , , . -, - - ' ' ' to output windings 2410 and is thu~ very accurate and stable~ ..
This is done in the following manner.
The connection fro~ the CIUtput of amplifier Z214 through 1' transfor~7ers ~2:;!16 and 2218 back to the input of ampli fier ~214 provide~ negative feedback~ Due to the h:igh g~in of an7pli1'ier 2216y the negative feedback Imakæs the input to ampli fier 2214 a virtual ground~ The voltage across the output winding of t~ansformer 2218 must thereforæ be essentially zero, and h~nce the Yoltage acr~ss the selected input winding of transformer 2218 must also b~e zero. ~;ince the voltage drop acrc,ss the input winJing 2~02-:~406 to transforrner :ZZ18 and th~ single-turn windings ~!~lZ--2414 must equal the voltage across terminals V
and V~ essentially the entire voltage across Vl and V~
appears across the single tllYtl windings Z41~--Z416. The gain from lines 2~0 and 22Z ~connected to the inputs to a~plifier Z-13 to the o~tput windi~g 2410 is essentially det~r~inæd solely by the ratio betwe4n the number of windiny~i selected by switches :~41~ 4~ and th~ 1'`8--turn ~utput windin~ ~iO. By selectively closing switches Z418--2422, gains of 128~ 128~2, and l:i~8~3 may be selected. Switches ~!~18--~422 and t~e di f ferent ~aps on transformers 2~216 and :2218 enable a relatively constant input impedanc:~ to amplifier 2--14 to be mai~tained when thi? amplifi~
gain is changed.
FiS~o 17 shows one ~:ircuit su~table for ampli~i~rs ~ 6, Z228, ~nd 2230 which drive the shields o~ th* ~bles t~ ter~ninal~s Yl--V4 and the 0uard el~ trod~,, Th~ input Eaig~al is applied --1 ~5--to the non il~verting input of an op-amp 50~ through a cap~ci~or 502~ Op-amp 504 ~ay be implemeMte~ by LF356 amplifiers. The cutput of op-a~p 504 i5 connected to its inverting input to provide a unity-gain ampli~ier. Two l~r~e v~lue resistors 506 and 50~ are connected in series betw~en the non - inv~rting input to the op-amp and ground to provide a DC reference level at the input. A ~apacitor ~508 connects the juncltion of resistors 506 and 508 and the op-arnp output.
The output fror.l thc c:p-afnp is conn~cted through a capa~itor 51~ and a resistor ~10 in parallel with switcil 514 to provide the amplifier stage output sigrlal. Resistor 510 i5 typi~ally 1 kilohm. Switch 514 is an electronically controlled switch such as a FET switch at~d is normally closed during conductivity m~asure~ents. During diaynostic routines~ the processor provides a guard leaka~e measurement si~nal GL which opens switch ~14 putting resistor 510 in series ~ith the output signal t~ the shields and guard ~lectrodes. I~ there i5 no leakage current flowing frorn the~ guard and shield circui1;, the El through E~1.
voltages will re~ain the same when switch 51~ is ~losed. If there is any si~nificant leakage current, the leakage c~rrent will cause a volta~e drop across resistor 510. ~y measuring the E1 through E4 voltages with switches 514 cl~sed and open, the leaka~e current in each shield and guard electrode CiYCUit can be determined~

Fig. 18 is a schematic of an ~C amplifier circuit which can be used for the AC amplifier 2212. The circuit consists of four 5532 op-amps 641-644 connected as a bi-quad amplifier/filter shown in Fig. 18. This amplifier configuration is discussed fully in the aforementioned U.S. patent No. 4~539,525 entitled "Bandpass Amplifier Filters." The circuit provides an open loop gain of approxi-mately 30,000 while maintaining a very stable zero-phase shift characteristic of the operating frequency of 384 Hz.
Typical values for -the components shown in Fig. 18 are as follows:
602 100 ohms 618 4.99 K
604 10 K 630 0.0082 mfd.
606 150 K 632 0.002 mfd.
608 10 K 634 2.2 mfd.

616 49.9 K 644 5532 The circuitry of AC amplifiers 2214 and 2240 may be implemented by means of a two stage bi-quad amplifier circuit. One such circuit suitable for use with the described emobodiment and which includes compensation for DC offset errors is the above-referenced patent for "Bandpass ~mplifier Filters" in Figs. 8 and 9 thereof.

h ~ Fi~. 19 ~5 a ~che~a~ic dia~ra~ of ~ne circuit for implementing the ~0-des3ree phase shifter :~3620 The input ~3ignal to the phase shi~eer circllit is applied through series-connected ~apacit~r 708 ~nd resisltor 702 to the inverting lnput c3f an ~p an~p 714. The inverting input of the op--aDnp is grounded.
re~istor 704 and cap~citor 712 are ~onnec1;ed in paral lel bel;ween thæ ~utput and the i3~verting input to vp a~np 714. An op--amp 716 i5 co~nected as a ur~ilty-g~in buffer amplifier~ The o~ltpUt of ~p--~mp 714 is applied ~o the input of the second op--a~p 716 via a resistor ~06. A capa~itor 710 i5 connected b~tween the input to op--amp 716 and ground. Typical values for the comp~n~nts in Fig.
19 are ~s foll~ws.
702 40 K ` 710 0. 001 mfd.
704 41 megohms 712 Oc 01 rnfd.
706 8.~ K 714 I F356 708 1.... O ~nfd., 716 LF356 Figs. 20 throu~h 22 ~re flow di~3rams showing the steps and methods perforrr~ed by the digital prc~,~ess~r c~ntrol~r to carry out a c0~7ductivity n~æasurernent. It is iassualed that to start, the c~lls and re~ervoir ~f the apparatus are filled with e~ectrolyt~, that thæ solution i9 ~lowing through the system~ and that the material to be analy;!ed has not yæt been added to the electr~lyte.
Fig 20A shows the initial diagnostic routi~es performed by th~ instrur..ent prior to m;tkinsl c:GP~d-lctivity meaaure~7er~t~
F~rst~ th~ proc~r initia~ *1~ th~ ~chin~ includi~g ~?tting ~-.~i, ... . .
.
,, : .
.
~ r 1 1 8 l7 ' ~ .

all ~;wit~31es ~nd~ ~ f t;he ~low is under c~trol o~ t!he pr~cèsso~ 7`'-~
bl~ck 801. Swi~cche5 S~ ar~ all pUt into ~hei~
ps:~sition. 5wi~ch GC2 ls se1; 'CO the hi~h ~3~in p,t3sitionD and '~:he random reference ~;ignal i5 selected by multiplexer ~ 5:;Z 'Por the initial rneasuren~ents dLIring which the ~agnitude of various signal5 i5 measur~d~
Next D the instrument rneasures ;;he values of EC~ and EC27 ~nd El 'chrough F~, The digital1 proCessor cloes th3s by sending signal~ to r,~ulitiplexer 2330 which sequentially applies each of ttl~e signals to the inpu~ of amp1i ~ier ~832 and then rf~easuring the vo~ t~gæ, b}ock 803.. For each ~;ignal ~ thæ processor checks to ænsure that æach of thæse voltages are within predetermined lin~its, block l05. If not7 the processor goes 'cO
an error routine~ denot~d by branch point ~ where ~n appropri~t error me~sa~ge i5 output to the oper~tor and the rDeasur~rf~ænt process is aborted.
Fi~. 20B is a IbrielF flow di~grafn ot how voltage me~511rements are made during the diagnostic routine, such as ~ho~e perforn~
ir~ block 803. IFirst, multiplexeY ~330 is 1~ mmanded to select th~
proper siçlnal7 block 813. I~lext, the IQ si~ana1 is s~t high to r~?set the inte~rator, block ~315. The processor waits ~or at l~ast 1 millisecond to allow the itltgrator capacit~r to compl~ltely di~ch~rge.. The IR signal is therl set low ~o s~art th~
integratiorl of th~ de!#od~1ated sign~1, bloc~ . Th~ proc~ssor wait~ ~or a pr~detei~mined ~ " bls~ck 1319, and then commands th~
A~llt cz3rlvert~r to cs:llnv~rt and m~ 3uro th~ output volt~ ror~ th~
i nt *~ar ate~ , t3 1 oc k E3~!1, --1 ~g--~%~
Returning to Fig. 20A, if the voltages measured during block 805 are within the proper limits, the processor proceeds to determine the current flowing through the cells. This is done by measuring the E~o voltage at the output of signal generator 2200, block 821. The processor then computes the value of the cell current from the values of ECl, E~o, and the signal generator resistance selected by SR, block 823.
The processor next checks the impedances between terminals Vl and V2. This is begun by closing switches SE and SF, block 831, which connects the two l-kilohm resistors 2242 between terminals Vl and V2. Next El and E2 are measured, block 833. From the previously computed value of the current and the values of El and E2 with and without the resistors 2242 connected, the impedance between terminals Vl and V2 is computed, block 835. This value is then checked against its limit values, block 837.
If the impedance is within the limits, switches SE and SF
are opened, block 839, and the above procedure is repeated to measure the impedance between terminals V2 and V3, blocks 841-849, and between V3 and V4, blocks 851-859.
Next the guard electrode leakage currents are checked, blocks 871-877. First, switches GL1 through GL4 are closed, block 871, and the voltages El through E4 are measured with the resistors 510 shown in Fig. 17 in series with the guard electrode. From the change in each of the voltages El through E4 with and without resistors 510 in series with the guard ,~, 2~
electrodes, the leakage current for each of the four lines is computed, block 875, and compared against thP limit values, block 877.
If the flow rate is checked, this is done next, blocks 881-887. If necessary, the flow measurement apparatus is enabled by the FE signal, block 881. The flo~ signals F1 and F2 are then measured, blocks 883-885, and checked against their limit values, block 887~
To end the diagnostic routine, a check is made to ensure that the signals from the cells are properly phased, blocks 891-899.
This checX ensures that a short or other malfunction has not shifted the phase of one of the signals from terminals V1-V4. The previous measurements of E1-E4 were made with the random reference selected by multiplexer 2352 for the demodulation. First, the processor commands multiplexer 2352 to select the 0-degree reference signal, block 891. The values of E1-E4 are then measured, block 893. The multiplexer is then commanded to select the 90-degree reference, block 895. The values of E1-E4 are measured again, block 897. Finally, the measured values are checked to ensure that signals from terminals Vl-V4 are within allowable limits, block 899. If so, the processor proceeds to the calibration routines, block 900.
Fig. 21 shows the calibration routines. First, the zero offset error of DAC 22~6 is measured, blocks 901-915. To begin, multiplexer 2330 is commanded to select the EM signal for measurement, block 901, and multiplexer 2352 is commanded to select the 0-degree clock signal as the demodulating signal, block 903. Both gain control signals GCl and GC2 are set to the high gain state, block 905. Switch SA is closed ts provide a ground referencQ to one input of comparator 2292, block 907.
Next, the digital inputs D1-D14 to DAC 2296 a:re all set to zero, block 909. The differential input voltage to comparator 2292 at this point is the difference between ground and the zero output voltage from DAC 2296. This voltage is measured, block 911, and the zero offset error from DAC 2296 is computed, block 913. This value is checked against its limit values, block 915.
If the error is within acceptable limits, the offset error value is stored and used to correct later measurements.
Next, the A/D converter is calibrated to detQrmine the output from the A/D converter which is equivalent to one turn of the secondary windings on transformer 2236. With the gain still at high and the DAC output still at 0, switch SA is opened and switch Sc is closed to connect the output from the two turn winding 2286 to the measurement circuitry, block 9210 The voltage is measured, block 923, and the gain from the input of comparator 2292 through the A/D output is computed, block 925.
Next, the gain reduction factors controlled by GC1 and GC2 are computed. First, switch Sc is opened and Switch SB is closed, block 927. The input to the comparator is now precisely 32 times larger, due to the precision of the outputs from transformer 2236. The processor then sets the GCl and GC2 signals to select high gain, block 929. The voltage is measured, . : _ block 931, and the ratio between the two gain settings is computed and stored, block 935.
The gain of DAC 2296 and resistive divider 2298-299 is then calibrated. Switch SB is opened and switch Sc is closed to connect the 2-turn winding 2286 to the comparator, block 943.
The processor sets the Dl-D14 inputs to DAC 2296 all high, block 945. The voltage is then measured, block 947. By comparing this voltage with the voltage measured in block 923, the gain factor of the DAC and the 64:1 resistive divider can be computed, block 949. This completes the calibration routine.
The various correction factors determined during the calibration are used to determine the actual voltages from the measured voltages in the measurement routines described below.
Figs. 22A and 22B show the routine by which the conductivity change is measured. First, an initial routine, shown in Fig.
22B, is performed to quickly set switches Sl-S6 and the DAC
to their initial values. To do this, switch Sc is opened and switch SD is closed to measur~ the voltage from the series connection of windings 2270, 2272, and 2274-2284, block 1001.
The Dl-D14 inputs to the DAC are set to 0, block 1003, and switches S1-S~ are set to deselect all six of the selectable windings of transformer 2236, block 1007.
At this point the signal applied to the measurement circuitry is the di~ference between the test and control cell voltayes.
This voltage will typically be a fairly lar~e value due to the differences between the bioregions with and without the ., antiligands in the matrix, or due to other differences between the symmetry of the two fluid channels. Because of this the gain is set to the low value, block 1007, to prevent saturation of the measurement circuitry. The volta~e is then measured, block 1009. From this measurement, as corrected by the factors determined during calibratlon, the processor compu-tes the proper settings for switches S1-S6 and the two MSB's of DAC 2296, blocks 1011 and 1013.
S1-S6 and the MSB's D1 and D2 are set by the processor, blocks 1015 and 1017. The gain is set to high, block 1019, and the voltage is again measured, block 1021. The processor then computes the value of the 12 LSB's of the DAC, block 1023, and the processox sets DAC lines D3-D14 to these values, block 1025. At this point, the difference between the voltages to comparator 2292 should be less than one LS~ of the DAC output. The processor then goes to the trackiny routine shown in Fig. 22B.
At the beginning of the tracking routine, the fluid sample to be analyzed is added to the electrolyte, block 1041. This may be done under control of the processor or manually in response to a prompt from the processor. Next, the integrator is reset.
First, the processor sets the IR signal high to reset the integrator, hlock 1045. The processor waits for a short period of approximately 1 millisecond to allow the integrator capacitor to ~ully discharge, block 1045. The processor then sets the IR
signal low, block 1947, to begin the integration of the voltage being measured.

The integrator is allowed to integrate -for a selected period of time, block 1049, typically 50 -to 500 milliseconds. At the end of the integra-tion period, the processor commands the D/A
converter to do a conversion, block 1051. The conversion takes aproximately 225 microseconds, which is essentially instantaneously in terms of the rate of change of the integrator output signal. The A/D digital output is stored for later processing, block 1053.
Next, the processor checks to see whether the A/D converter is approaching the limits of its dynamic range, block 1055. If not, the processor returns to block 1043, and the above process is repeated. If the A/D is approaching the limits of its measurement range, the processor computes new settings for switches S1-S6 and/or DAC inputs D1-D14 which will return the voltage measured by the A/D converter to the center of its measurement range, block 1057, and these values are sent to the switches and DAC, block lQ61. The processor then returns to block ~0~3 and the above process is repeated.

Another embodiment of the circuitry of an instrument for measuring the relative conductance of a test volume and a control volume is shown in Fig. 23. This circuitry was used in making the measurements in several of the examples set forth in the present patent application, as noted. The circuitry of Fig. 23 is similar in many respects to the circuitry of Example 8, and the explanation below will point out the differences between the circuitry of Fig. 23 and the circuitry already discussed.

.~

``" ~2~ 6 The circuitry of Fig. 23 is designed to work with a test andcontrol cell structure in which each cell has its own current path, such as the apparatus of Fig. 5D, which provides individual current electrode pairs for the test cell fluid path and the control cell fluid path. These electrodes are shown schematically in Fig. 5D by electrodes 146 and 148 for the test cell and electrodes 150 and 152 for the control cell. The circuitry of Fig. 23 also works with an apparatus such as that of Fig. 5C, wherein the test and control cells share a common current electrode 144.
In Fig. 23, a signal generator 12Q0 applies a sine wave signal to an input winding 1201 of a transformer 1202 via a capacitor 1203. In this embodiment, the frequency of signal generator 1200 is 3kHz. Transformer 1202 has two, identical, output windings 1204 and 1205. Winding 1204 has four output taps, and the output is taken from one of these taps via a 4-pole switch 1206. The primary winding of transformer 1202 has 140 turns on a supermalloy toroidial core, and each of the taps on secondary windings 1204 and 1205 are single turn windings. The two secondary windings provide individual current sources for the test and control cells.
The output from one of the taps on winding 1204, as selected by switch 1206, is applied to one end of the input winding 1212 of a transfomer 1210. The othe~ end of winding 1212 is connected to the first voltage measurement terminal V1 of the test cell.
The test and control cell resistances are denoted by R~ and ~7 , J --, .

;92~ ~
Rc respectively. The second voltage terminal V2 of the test cell is connected to the common end of winding 1204.
Transformer 1210 has a 600-turn secondarv winding 1214 which drives a high-gain, tuned, AC amplifier 1216. Amplifier 1216 is similar to the AC amplifier shown in Fig. 18, but modi~ied to work at the 3kHz center frequency. The output of amplifier 1216 drives a 112-turn input winding 1218 to transformer 1220.
Transformer 1220 is a precision ratio transformer. A 64-turn feedback winding 1214 is connected in series with a variable resistor 1226 to the two current electrodes C1 and C2. A
64-turn output winding on transformer 1220 provides an output signal.
Transformers 1210 and 1220 and amplifier 1220 are connected to provide a negative feedback loop. Due to the high gain of amplifier 1216 and the negative feedback provided through the transformers, the input to amplifier 1216 can be considered to be a virtual short circuit, and the voltage drop across winding 1212 is essentiall~v zero. Thus the voltage across terminals Vl and V2 is nearly equal to the voltage from transformer winding 1204. The feedback winding 1224 drives a current through resistor 1226 and the test cell resistance Rx to maintain the voltage drop across the voltage terminals V1 and V2 equal to the drive voltage from winding 1204. The current through and the voltage across the test cell is determined by the setting of switch 1206 and value of variable resistor 1226. The value of resistor 1226 is much greater than the test cell impedance to ... .

: `

~2~2~i ~
maintain a relatively constant impedance level as the test cell impedance changes. Resistor 1226 is typically on the order of 20-50 kilohms. Resistors 1226 and 1246 are adjusted to obtain a desired signal level from the amplifiers so that the following circuitry does not saturate and to match the amplitudes of the outputs from the test and control cells.
The control cell voltage terminals V3 and V4 and current terminals C3 and C4 are driven by and connected to circuitry identical to the circuitry described above in connection with the test cell, including tapped transformer winding 1205, transformers 1230 and 1240, and amplifier 1236.
Transformers 1220 and 1240 are analagous to transformers 2216 and 2236 shown and described above in reference to Fig. 14.
Transformer 1220 has a single 64-turn output winding 1222 which provides a signal representative of the voltage drop across the test cPll impedance. Transformer 1240 has a 32~turn winding 1250 and six, binary-weighted windings 1252 having 32, 16, 8, 4, 2, and 1 turns respectively. Six switches 1254 selectively connect windings 1252 in series with a the 64-turn output winding from transformer 1220 and a fixed 32-turn winding 1250 on transformer 1240. These windings allow the control cell output to be scaled over a range of about 0.5 to 1.5 times the test cell output;
depending on the setting of switches 1252, with a precision of six bits.

~' , The series connection and phasing of these windings effectively subtracts the control cell output signal, as scaled by the settings of switched 1254, from the test cell output signal. This difference signal is across lines 1253 and 1257 and is applied to comparator circultry 1280 via an auto-transformer 1278, as explained below. The operation of the circuitry in comparator 1280 is similar to the comparator circuitry shown and described in Fig. 15.
Switches 1254 provide six bits of scaling range. Fourteen additional bits of scaling are provided by a 14-bit multiplying DAC 1260 in a manner similar to that of Fig. 14. DAC 1260 may be implemented by a ICL 7134U unipolar D/A integrated circuit. The output from an independent 32-turn winding 1256 is applied to the input of the DAC 1260. DAC 1260 is controlled by the digital processor, similarly to DAC 2296 in Fig. 14. The output from DAC
1260 i5 applied via a resistive network including resistors 1262-$268 and a capacitor 1276 to a unity-gain buffer amplifier 1270.
The output from buffer 1270 is applied to one end of the auto-transformer 1278. The other end of auto-transformer 1278 is grounded. The turns ratio of the two sections of the auto-transformer is 128 to 2. This scales the output oE the DAC
1260 by a factor of 64 so that one MSB from the DAC is equivalent to the output from a one-half turn winding on transformer 1240.
A 10 X potentiometer 1269 in conjunction with 2250 K resistor 1260 and a 10.2 K resistor 1262 all the output from the DAC to be ,,. .. ,~
f ~?

trimmed so that it is exactly equal to one-half turn. Thus switches 1254 and DAC 1260 provide a 20-bit range over which the control cell output can be scaled to track the test cell output.
A second DAC 1282 is also driven by winding 1256. DAC 1282 may be implemented by a ICL 7134B bi-polar D/A converter integrated circuit. The output from DAC 1256 is applied vla a 90~ phase shifter circuit 1284 and a 10 K resistor 1268 to buffer amplifier 1270. Thus, the total input to amplifier 1270 is composed of in-phase and quadrature components. During the diagnostic and calibration phases, the digital processor measures the quadrature component in the ouput from the test cell and sets the digital input to DAC 1282 so that the ~uadrature component is cancelled or reduced to prevent saturation of the following measuremant circuitry.
Further quadrature compensation may be optionally aclded by means of a potentiometer 1290 connected across the output terminals of signal generator 1200 and a capacitor 1292 connecting the wiper of pot 1290 to terminal V3. Typical values for these components are 5 kilohms and 100 pf. The setting of potentiometer 1290 may be adjusted to compensate for small amounts of quadrature error caused by capacitive loading.
The procedures followed in measuring conductivity using the alternate embodiment shown in Fig. 23 are similar to those described above and illustrated in Figs. 20-22. A source code listing of a program for implementing these procedures with the embodiment of Fig. 23 is attached hereto as Appendix A. This a ~ A ~
r.d~o"3~ ~4V
lis-ting is written in Basic and is for controlling a Hewlett-Packard Computer Model 9845 to perform the described measurements.

' ..

~ ( l z~ ~

APPENDI X A

:,.

~ *~*~ O0T 1~3~ S C~R~ ( GPi~ C~ PLOTTING~ ***********

~OM Ti~e,P$,T~Bits,Rnswer~Stabilise time~R~utine,Rns*lPl~t$,PrintS,Gain,Eit s$,I,R~$,5amples,Scale,Scans,Gr~$,Hardcopy$,Gain ,Interual,Cycle,B$
COM Date ~Run,TitleS~Rdd desc1S~Rdd desc~$,Biolsyer$,5en5c~r desc~,U~5]
CO~ Ratiol,Slope COM SHORT Out2(10000) ON KEY ~0 GOSUB Pr i nN KEY ~1 G0SUB Gra 100 ON KEY #15 GOTO Dump 110 ON KEY #3 GOTO 5900 140 Replot$="N"

170 IF Cycle=0 THEN GOTO 220 180 PRlNT PRGE
1~0 PRINT "DO YOU WISH TO REFLOT PRE~IOU5 ~hTR TO R DlFFERE~T SCRLE...~YxN~' 200 INPUT Fepl Ot *
2 1 0 IF Replot~="Y" THEN GOSUE Replot ~2a PfiINT PRGE
230 U~="...... PREVIOUSLY WRS..... "
~49 PRINT "ENTER SE~SOR DESORIPTION";U~;Sensor d~sc~
250 INPUT Sensor descS

270 PF;INT "ENTER ~IOLRYEF DESCRIPTION";U$;Biolayer*
2S0 INPUT Biol ayer~

300 PRINT "ENTER RNY RDDITIONRL DESCRIPTION";U~;Rdd descl~
310 INPUT Rdd desc1$
~20 PRINT PFIGE
330 PRINT "ENTER DRTE",Date 340 ~NPUT Date 360 PRINT "ENTER RUN NUM~ER...... PRE~lOUS RUN WRS ";Run 3~ Run=Run~1 3S0 PRIMT "NEW RUN WILL BE ";Run;" UNLESS hNOTHER IS ENTERED"
390 INPUT Run 410 PRINT "ENTER TITLE....... PREVIOUS TITLE ~hS..... ";Title~
420 INPUT Ti~le~

440 PRINT "DO YOU WISH TO CHRNGE RNY KEY~ORF~ INPUTS...... ~`f ~r N~"
450 INPUT Rnswer~

470 PRINT "DO YOU WRNT HRRD COPY....(YxN~"
4B0 INPUT H~rdc~py~

4~0 PRINT PRGE
500 IF Rnswer~="N" THEN G3TO Executi~e 510 PRINT "PRESENT GRIN LEYEL IS ";Gain ;" DO YOU WISH TO CHRNGE IT"

5~0 PRINT "IF NOT PRESS CONT KEY"
530 PRINT "OTHERWISE ENTER DESIRED GRIN LEYEL"

550 PRINT "ENTER D~SIRED GRIN LEVEL"
550 PRINT "Ncrmsl gain level is 3 for a stable sensor or where the ~ariation "
570 PRINT "~r6 small.For ~ situation where the output is ~arying r~r;i~ly"
5~0 PRINT "use lower gain levels (2 fc~r r"oderate situations I or 0 for"
590 PRINT "mcre extreme situations"
600 INPUT Gain 610 PRI~T PRGE
620 ! PRINT "ENTER SR~PLINb INTER~RL..~(Seconds~"
630 ! RINT "Recommended int~r~al is ~.5 Seconds....~DO NOT EXCEED 2 sec)"
640 ! RINT "If you do not want to ch~nge it press CONT key."
650 ! INPUT Int~rval A-l ~ ~ v ~ s ~
670 Time=200~*ln~
680 Stabilise tim~
fi90 ~its-20 710 PRINT "DO YOU WRNT ~NY DRTR RVERRGING....~Y~N~"
720 INPUT R~

74Q IF fl~5="N" THEN Samples=1 7S0 IF R~="M" THEN GOTO 800 760 PRINT "PRESENT NUM~ER OF SRMPLES IS ";Sa~,ple~;" DO YOU WISH TO CHRNGE IT "770 PRINT "E~TER NUMB~ OF SCRNS PER RVERRGE"
780 I~PUT Samples 750 PRINT P~E
800 PRINT "SELECT PRIR R ~ B OR P~IR C ~ D...(~ B C=C8D~"
810 INPUT ~$
320 IF ~$="R" THEN ~="R"
830 IF B ="C" THEN R~="B"
B40 GOSUB UnsCick 850 OUTPUT 720;"C";RS
860 PRINT P~GE
870 PRINT "DO YOU WRNT R GR~PHIC~L OUTPUT"
880 INPUT Gra~
890 IF Gra ="~" THEN GOSU~ Plot da~a 900 GOTO Executi~e g30 Plot dat~:!

950 PRINT "WHRT FULL SCRLE SENSITI~ITY ON THE GRRPH DO YOU WRNT..~ppm~"
960 INPUT Scale 980 PRINT "HOW M~NY SCR~S FULL SCRLE DO YOU WRNT ON THE TIME RXIS"
990 INPUT Scans 1010 !
1020 !
1030 Unstick:!
1040 SET TI~EOUT 7;T~3 1050 ON INT ~7 bOSU~ Bom~

1070 !
1080 !
1090 Init quad:!
1100 GOSU~ Unstick 1110 OUTPUT 720;"R0"
1120 OUTPUT 720;"L";32;",";0 1130 ~uadl=Quad2=Qbit=0 1150 !
1160 !
1170 Init re~:!
1180 GOSUB Unssick 1190 OUTPUT 720;"R9"
1200 OUTPUT 720;"L0,0"
1210 OUTPUT 720;"M0"
1220 Data1=Da-~2=Data3=Bit=0 1240 !
1250 !
1260 S~t hl re~
1270 GOSU~ Unstick 1280 ~=b-Bit 1290 Dat~1~Datal+2~M
1300 OUTPUT 720;"M";D~tal . ~3~

~ v 1340 Set_hi quad:! ~ ~ ~ ~-~,~
1350 bOSUB Un~tick r~ 1360 M=6-Qbit ; 1370 Qu~d1=Quad1~2AM
1380 OUTPUT 720;"L";Quad1~"l0"

1400 !
1410 !
1420 Set mid ref:?
1430 GOSU~ Un~tick 1440 M=12-Bit 1450 ~ta2=Data2~2~M
1460 OUTPUT 720;"L";~ata2;",0"
1470 RETUR~
1480 !
1490 !
1500 Set lo qu~d:!
1510 GOSUB Unstick 1~20 M=14-Qbi t 1530 auad2=~uad2+2~M
1~40 OUTPUT 720;"L";Quad1;",";Qu~d2 1560 !
1570 !
1580 Set lo ref:!
15~0 GOSU~ Unstick 1600 M-20-Bit 1610 ~ata3=~ta312^M
1620 OUTPUT 720;"L";D~t~2;",";~ta3 1c40 !
1650 !
1660 R to d:!
1670 SET TIMEOUT 7;T~3 1680 ON INT ~7 G05UB Bomb 1690 OUTPUT 720;"5";T
1700 SET TIMEOUT 7;T
1710 WRIT T~3 1720 ENTE~ 720;Y;Z

1740 OFF INT #7 1750 ~c=Y-128 17~0 RETURN
1770 !
1780 !
1790 BombO!

1820 OFF INT #7 1850 Gerr-G~rr~1 1870 !

1890 Gain1:1 1900 OUTPUT 720;"G1"
1910 Gain-1 19~0 RETURN
1930 !
~4~ l 1950 Gain2:1 1960 OUTPUT 720;"G2" A-3 1970 Gain=2 : ~3S' 1990 ! f 2010 Çain3:l ~ ~ ~ ~ 2 r 2020 OUTPUT 720;"G3"
2030 Gain=3 206~ 1 2070 Gain~:!
2080 OUTPUT 720,"G0"
2090 Gain=0 2120 !
2130 Executi~e:~
2140 I F GraS= " Y " THEN GOSU~ Gr~pl~

217~ Ratio2-Cycle=0 21g0 T=Time~25 2200 DISP "SET GRIN TO LEVEL ";G~in 2210 ON tain ~1 GOSU~ 5~in0,G~inl,G~in2,Gain3 2220 ! WR I T 1000 2230 ~ISP "INITIRLIZE QURDRRTURE"
22~0 GOSUB Init ~u~d 2250 ! ~RIT 10~0 2260 DISP "INITIRLIZE REFERENCE"
227C GOSU~ Init ref 2280 ! WRIT 10a0 22g0 ~ISP "DIGITIZE REFERENCE...... (PRELIMINRRY)"
2300 ÇOSUB Digitize ref 2310 DISP "DIGITIZE GURDRflTURE.o........... ..<PRELIMIN~kY) 2320 GOSUB Init quad 2330 GOSUB Quaarature 2360 Is quad ok:!
2370 GOSUB Unssi~kl 2380 OUTPUT 720;"R0"!
2390 GOSUB R to d ! THI5 RQUTINE CHECKS TO SEE THflT
2400 IF RBS~R~)<126 THEN ~OTO 2490! QURDRRTURE HhS BEEN CORRECTL~
2410 PRINT Pfl5E ! BRLflNCED.
2420 FOR Z=l TO 10!
2430 BEEP!
2440 NEXT Z !
2450 DISP "QURDRRTURE PROBLE~............... .QUR] JRRTuRE DOES NOT BRLRNCE"246C END
2470 !
24~0 !
2490 GO5UB Unstick 2500 OUTPUT 720;"R9"
2510 DISP "INITIRLIZE REFERENCE............. ..(PRECRLIBRRrE)"
25~ ! WRIT 250 253e GOSUB Init reS
2540 DISP "DIGITIZE REFERENCE...... ~PRECRLIBRRTE)"
2550 GOSUB Digitize ~e~
2560 GOSUB Cal ibra~e 2570 DISP "INITIRLIZE REFERENCE............. ...~PRE-TRRCK)"
2~80 GOSUB Ini~ ref 2590 DISP "DIGITIZE REFERENCE............... ..~PRE-TR~CK)"
2600 GOSUB D1gitize ref 261B IF Gr~S~"Y" THEN GRflPHICS
2620 Rati~1-(524288+D~tal*16384~Data2*256~Data3)~2^20 2630 D~lta~RBS~ Rat i ~I-Rat i o2) A-4 2650 G=16^(3-Gain~ f ~ ~ 4 2fi~0 IF Delta>.0001*~ .HEN GOTO 2589 2678 5err=0 2~80 T=Time~25 2690 K=30000*I~slope~T~l6^~Gain 2710 GOSU~ R to cd 2720 fldcl=Rdc 2730 Ratio=~524288+rlatal*163~4tData2*~56+~a-,a3-Rdc*~)~2^20 2740 IF RBS~Raticl-Ratio1)>.01 THEN GOSUB Init q~ad 2750 IF ~E~S~Ratio-Ratiol?>.01 THEN GOsur Quadrsture 2760 IF RBS(Ratio-R~tio1)>.01 THEN Ratio1=Ratio 2770 GOSUB Unstick 2780 OUTPUT 720;"R9'~
2790 IF Cyc1e=0 THEN ~er ratio=~atio 2300 lF Cycle=0 THEN Outl-Ratio 2810 Cycle=Cycle~1 2820 DISP Cycle 2830 Rver ratio=(Sarnples-1)/Samples*R~Jer ratio+F~atio~Samples 2840 out=INT~l0^7*~er-ratio+.5)~l0^7 2850 C=.1*Cycle-INT~.1*Cycle) 2860 lF C=0 THEN GOTO 2880 2880 IMRGE ''~l='',Drl,''¦D2='',rlrlD,''¦D3='',DDD,''¦Q1='',Drl,''¦Q2='',DDD,''¦R~ ='',5Ilrlrl,''lF~
ati 0= " 9 D.DDIIDrlD,''lR~e='',D.DDDDDrlD
2~90 PRINT USING 2S~0;Dat.al,Data2,~as,a3,Quacl1,t~uad2,Rdc,Ratio,Out, 2~00 GOSUB Check quad 2910 Out2(Cycle)=lNT((Out-Out1)~Out1*1E7)~10 2920 IF Cyclet10 THEN GOTO 2940 2930 IF Gra~="Y" THEN PLOT Cycle-10,0ut2(0ycle) 2940 IF RBS(Rdc)>126 THEN GOSUB Init ref 2950 IF RBS(~dc)>126 THEN GOSUB Di~iti~e ref 2~0 IF Rdc>100 THEN GOTO 2g90 2g70 IF Rdc<-100 THEN GOTO 3010 29g~ Incr=INT~960000,'Time) 30~0 GOTO 3020 3010 Incr=-INT(960000~Time) 3020 GOSUB Increment disi 305a 3060 Digitize ref: !
3070 Bit=Bitl1 3080 IF Bit<7 THEN GOTO 3120 30g0 IF Bit<13 THEN GOTO 3140 3100 IF Bit>l2 THEN Stage=3 3120 Stage=1 3140 Stage=2 3150 ON S-age GOSLIB Set hi ref~Cet mid ref,Set lo ref ~160 WRIT Stabilise_time 3170 GOSUB R to d 3180 IF ~dc~0 THEN GOSUB Bit off 3190 IF Bit=Bit 5 THEN RETURN
3200 GOTO 3~74 3210 !
3220 !
3Z30 Bit ~f~
3240 ON Stage GOTO R ngel,Rang~2,Ranc~æ3 3250 !
3260 !
327~ Range1~ 5 3280 M-b-~it :
o3~

33~ OUTPUT 720j"M"jD~--1 ( 331 RETURN ~ ~ ~ 49V 26 333~ l 3340 Range2:!
3350 ~=12-~it 3360 Dat~2=Dat~-2AM
3370 OUTPUT 720;"L";Data2;",~"

3390 !
3400 !
3410 Range3:!
3420 M=20-Bi~
3430 D~a3=Dat~3-2^M
3440 OUTPUT 720;"L";D~ta2;",";~ta3 3460 !
3470 !
3480 Quadra~ure:!
349~ Qbit=Qbi t+l 350~ lF Qbit<7 THEN GOSUB Set hi quad 3510 IF Qbit>6 THEN GOSUE Set lo quad 3520 WRIT St abilis~ time 3530 GOSU~ R t~ d 3540 IF Rdc>0 THEN GOSUB Qbit off 3550 IF Qbit>13 THEN RETURN
3560 G0TO 34g0 3570 !
3580 !
3590 Q~it off:!
3600 I F Qbit<7 THEN GOTO Qbit1 361~ IF Qbit~6 THEN ~OTO Qbit2 3620 !
3630 !
3640 Qbitl:!
3650 M=6-Qbit 3660 ~u~d1=Quad1-2~M
3670 QUTPUT 720;"L";Quad1;",0" --3690 !
3700 !
3710 Qbit2:!
3720 M=14-Qbit 3730 Quad2=Quad2~2AM
3740 OUTPUT 720;"L";Quad1;","jQuad2 3760 !
3770 !
3780 !
3790 !
3800 Graph:!
3810 PLOTTE~ IS 13,"GRRPHICS"

3840 MOVE 49~RRTIO,1 3860 LRBEL "SCRNS"
3870 MO~E 2.5~RRTIO,50 38~0 LDIR 90 3900 LRBEL "PPM"
3910 ~=R~TIO~9.S
3920 ~R~TI3*97 393~ LOC~TE R,B,10,~
3940 SC~LE 0,Scans,-Scale~Scale 3960 LINE TYPE 3 f f-3970 DRRW Sc~ns,0 3980 LINE TYPE 1 ~ Z49 J~g0 ~XES .2*5cans~.5*Scale~0,-Scale 4030 FOR Y=~Sc~le ro ~c ~1 e STEP .S*Scale 4040 110~JE B, Y
4050 IF Scale<100 THEN LR~EL USING "MDD.BX";Y
4060 IF Scale~=100 THEN LR~EL USI~G "MDDDBDDX";Y

4090 LOR~ 6 4100 FOR X=0 TO Scans STEP .2*5car~s 4110 MOVE X,-Scale*1.05 4120 LRBEL USING "~DDDD";X

4140 RETU~N
4150 !
41~0 !
4170 !
4180 !
4190 Encode:!
4~00 Da~a=~tal*16384+Data2*~56+~ata~+1ncr.

4220 !
42~0 !
4240 Decode:!
4250 ~atal=IMT(Datazl63~4~
42~0 Data2=lNT~(Data-Datal*16384)z256) 4270 Data3=Data-Datal~16384-Data2*256 4300 !
4310 Set d to_a:!
4320 OUTPUT 720;"M";~atal 4330 OUTPUT 720;"L";Data2;",";Data3 ~:

4350 !
4360 !
4370 Calibrate:!

~390 PRIMT PRGE
44~0 PRINT "CRLIBRRTION DRTfl"
4410 PRINT "*************~*~"

4430 Calibrate2:!
4440 GOSUB Gainl 4~ I=50 4460 T=l 4470 Rcc zæro=0 44S0 FOR N=l TO 10 4490 GOSUB R to d 4500 ~cc zero=Rcc zero+Rdc 4510 Zero(N)=Rdc 4Y~0 IF RBS~2ero(N)-Rcc zeroZN)~5 THEN Error flag=l 4S40 IF ~BS(Rcc zero)~20 THEN GOTO Try ~gain 4550 PRINT "Mean Zero=",Rcc z~rozl0;" I<~vsrage of 10 ~a~plee~"
4S60 Incrs30000 4S70 GOSUB Encod~
45~0 GOSUB Decode 4S90 GOSUB S~t d to a ~ 7 4600 ~c~0 ,~39 r u n~ I v 1 ~ ~
4620 GOSUB R to d ~ 9 ~2~ ( 4640 Rccum~N)=Rdc 465~ IF R~S(Rccum~N~-Rcc~N)>5 THEN Error ~l~g=1 4670 Slope~(Rcc-Rcc z~ro)~l0 4680 PRINT "~ean Rcc~m~ cc~10;" (R~er~ge of 10 cample 5 ) "
4~9a PRINT "~ean Slope F~ctor=";Slop~ R~erage of 10 ~ampl~s)"
470~ W2IT 5~00 4720 lF Error flag=0 THEN GOTO 4860 4740 PRINT "DURING THIS CRLI~RRTION RUN R~NORllRL ~RRIRTIONS IN YRLUES "
475~ PRINT "WERE DETECTED.THE DRT~ ~RE ~S FOLLOWS"

4770 PRINT "SRMPLE"I"ZERO VRLUE"~"FULLSCRLE ~RLUE","SLOPE ~RLUE"
4780 FOR N=l TO 10 47g0 PRINT N,Zero~N),flccum~N),Rccu~<N)-Z~ro(N) 4830 PRINT "OPER~TION SUSPENDED....pre~s CONT KEY if you wish to con~in~

485g PRINT PRGE
4860 ON Gain ~I GOSUB bain0,Gainl,G~in2,G~in3 4~70 IF Hardcc,py$='!N" THEN PRINTER IS l~
4880 Zncr=-3~00~
48~0 GOSU~ Encode 4900 GOSUB Decode 4~10 GOSUB Set d to a 4930 IF H~rdcopy5="N" THEN PRINTER IS 16 4950 !
~960 ~ !
4970 Increment digi:!
4980 GOSU~ Encode 4~90 GOSU~ Decode 50~0 GOSU~ Ses d to a ~020 ~ _ 5030 !
S04~ Pri:! ¦
5050 EXIT GR~PHICS
5~60 RETURN
5070 !
5~80 50~0 Gr~
s 1 a~ GRRPHICS
5 l l 0 RETURN

5140 Dump: I

5170 Rdd d~sc2$=" "
S180 PRINT "RNY RDDITIONRL DESCRIPTlON"
51~0 INPUT ~dd desc2 520~ PRINTER IS 0 5210 PRINT P~GE
5220 PRINT T~<25~,Title~
~23~ DU~P GRRPHICS
~240 P~INT "SEN$0R DESCRIPTION: ";Sensor desc*
5250 PRINT "BIOLRYER: ";Biolsyer* ~-8 5260 PRINT "DhTE: ";Da~e t ., O

_ ~ . ~J r r~ ~ U 1 7 I ~ U I ' I D C, I~ un ~
5280 PRINT "INITIRL REr-~TRNCE ~RTIO: ";INT(lE6*Ratif ~.5)~1E~
529Q PRINT "SR~PLES PE~ CRN: ";5ample~
5300 PRINT "GflIN LE~EL ";Gain ~ ~ 49 5310 PRINT "~EflN SLOPE F8CTOR: ";Slope 532~ PRINT "CO~MENTS:"
5830 PRINT Rdd descl~
5340 PRINT ~dd d~sc2 53~0 PRINT

5389 !
5390 Replot:!
5400 GOSUB Pl~t data 5410 Ba_e line=100 5420 Ou~3=0 5~80 PRINT " DO YOU ~NT BRSE LlNE ~R I FT OORF:ECT I ONS . ., . ~ yzyec N=No~"
5440 INPUT ~ns~
5450 IF ~ns~="Y" THEN GOTO 5480 5460 ~ri~t per sc~n=0 5480 PRINT "ENTER No OF SC~N'j FOR ST~RT OF Bh';E-LlNE ~RIFT CORRECTION"
5490 PR I NT " ~ Def 2u 1 t ual ue=0 sc ans > "
5500 INPUT Begin 5520,PRINT
5530 PRINT "ENTER N~ OF SCRNS FOR EN~ OF BRSE-LINE ~RIFT CORkECTlOM"
5540 PRINT " ( D~ault value=100 ~cans)"
5550 INPUT End 5560 FOR N=Begin~l TO .2*<End-Begin)~Begin 5570 Out3=Out3+0ut2(N~

55g0 Out4=Out3*5/(End-Begin) 560~ Out3=0 56~0 FOR N=Begin+1~.8*~End-~egin) TO End 5620 Out3=Out3~0ut2~N) 5640 Out5=Out3*5~<End-~egin) 5650 Drift per 5C an=(Out5-Out4)~.8~End-Begin) 5660 IF Scans~Cyclæ THEN Scans=Cycle 5670 GOSUB Graph 5680 FOR N=1 TO Scans 5690 PLOT N,Oue2(N)-Out2(Begin)-~N-Begin)*~rift per scan 5720 !
5730 !
5740 Check quad:!
5750 GOSUB Unstick 5760 OUTPUT 720;"R0"
5770 GOSUB ~ to d 5780 IF ~BS(hdc~<126 THEN GOTO 5810 5790 GOSUB Init quacl ~800 GOSUB ~uadrature 5810 OUTPUT 720~"R9"

584~ ~
5850 Try again:l 5860 D I SP " RETRY~NG TO B~L~NCE ~U~DR~TURE"
5870 FOR Z=1 TO 100 S880 ~EEP

~900 GOSUD Ini~ qu~d 5910 GOSUB Quacdr~ture ~ 9 5920 G~SUB Init r~f ' ~

J ~ " ~J 1~ U ;:~ U D J~ IL C ~ T
~ 9~0 G~O ~ brA~
~r~ ~ Sg7~ 2~2~
598~
5990 S~r:l 6000 DISP ~INIT-QU~D"
6010 GOSUB Inlt quad 6a20 DISP ~INIT-REF"
6a30 ~OSUB Init r~ -~4e Bit~20 Ç050 D~SP "DIGITIZE REF"
6060 GOSUB Dl~itlz~ ref 6070 DISP "INIT QURD ..... ~LINE 6025)"
60~0 GOSUB Init quad 6090 DlSP "QURDRflTURE"
6100 GOSUB Quadrature 6110 DISP "IMIT REF ~LINE S035)"
61~0 GOSUB Init r~f 6130 DISP "D~GITIZE REF ~LINE 60~8>"
6140 GOSUB Dl~i~ize ref 615~ R~tio~5242~8~D~t~1*16384+Data2*~56+Data3)~2A20 6160 PRINT "D1=";Dat~l,"D2=";Data2,"D3=";Data3,"Ratio=";INT(1E6*R~ti~+.5~1E~
617~ GOTO 6110 6lsa !
fil90 1 6200 Rdj msb lsb:!
6210 INPUT "ENTER DESIRED GRIN",G
6220 ON G GOSUB Gain1,Gain2,Gain3 6230 GOSUD Init ref 624~ GOSUB Dlgitize ref 6250 GOSUB Init quad 6260 GOSUB Quadr~ture 6270 GOSUB Ini t rer 6280 GOSU~ Low d~ta 6290 GOSUB Dig 6300 ~dc low=~dc ~310 GOSUB High data 6320 GOSU~ ~ig 63 0 ~d~ high=~dc 6340 PRINT ULO=";Rdc lowj"HI=";Rdc high 636~ ! .
6370 ~ .
6380 Low d~ta:l 63g~ D~tal-3~
6409 Dat~2~63 64~0 ~ S2S5

6~3~ l 6440 ~
6450 High d~t~:l 6460 D~ta1~32 6476 ~t~2=0 6480 D~t~3~0 649a RTURN
6S0~ l 651~ l 6S20 D1g:l 653a COS~ Sgt d 6S40 GOSUB ~_~o d ~-10 ,,, , _ , , . , .. , .. . _ . . . _ .. . _ _ .. , , , .. _ . .. .. . . . . . . . .. .. . . . ... _ _ ..... -- -- - / / ~

Claims (91)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY
OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method for determining the presence of a ligand in a fluid sample comprising:
(a) measuring the bulk conductance of a test volume, said test volume at least partially containing therein at least one predetermined region, said predetermined region being exposed to said fluid sample and also having localized therein antiligand which interacts with said ligand; and (b) determining the occurrence of ligand-antiligand interaction by detecting changes in the bulk conductance of said test volume.

2. The method of claim 1, wherein said changes in bulk conductance are determined by comparing the bulk conductance of said test volume with the bulk conductance of at least one control volume.

3. The method of claim 2, wherein said control volume does not contain antiligand which reacts with said ligand.

4. The method of claim 3, wherein said control volume is substantially free of specific interactions which effect changes in bulk conductance.

5. The method of claim 2, wherein said control volume does not contain any localized substance therein.

6. The method of claim 4, said control volume at least partially containing therein at least one predetermined region, said predetermined region being exposed to said fluid sample and also having localized therein molecules whose physical properties are similar to the physical properties of said antiligand.

7. The method of claim 2, wherein said fluid sample contains a known concentration of a control ligand, and control antiligand which interacts with said control ligand is localized in at least one predetermined region of said control volume.

8. The method of claim 4, wherein at least one of (a) the bulk conductance of said test volume, and (b) the bulk conductance of a second control volume is further compared with the bulk conductance of said control volume, and wherein said fluid sample contains a known concentration of a control ligand, and control antiligand which interacts with said control ligand is localized in at least one predetermined region of said second control volume.

9. The method of claim 1, wherein said localization of said antiligand in said predetermined region comprises immobilizing said antiligand on a matrix.

10. The method of claim 1, wherein said localization of said antiligand in said predetermined region comprises confining said antiligand within the boundaries of a membrane, said membrane being permeable at least to said ligand.

11. The method of claim 9, wherein said matrix is selected from the group of gel beads, gel layers, glass beads,polymeric beads, microporous membranes, porous paper, or mixtures thereof.

12. The method of claim 9, wherein said matrix is contacted with a flowing stream of said fluid sample.

13. The method of claim 12, wherein said contacting comprises causing said fluid sample to flow through said matrix.

14. The method of claim 12, wherein said contacting comprises causing said fluid sample to flow past said matrix.

15. The method of claim 1, wherein said antiligand is a substance which specifically binds to said ligand.

16. The method of claim 15, wherein said ligand is selected from the group of antigen, cell surface antigen, antigenic determinant, hapten, antibody, antibody fragment, monovalent antibody fragment, nucleic acid sequence, enzyme, cofactor, enzyme substrate, genetically or chemically altered protein, receptor protein, molecule bound by a receptor protein, permease, other transport protein, binding protein, molecule bound by a permease, a transport protein or a binding protein, carbohydrate, lectin, metal ion, metal-binding protein, or other metal-binding substances.

17. The method of claim 1, wherein said antiligand has high affinity for said ligand.

18. The method of claim 1, wherein said antiligand has low affinity for said ligand.

19. The method of claim 1, including the step of forming a complex between a ligand which interacts with said antiligand and a particle.

20. The method of claim 19, wherein said complex is formed in said fluid sample, said ligand component of said complex being derived from the ligand present in said fluid sample.

21. The method of claim 20, including the step of adding a known amount of ligand which interacts with said antiligand to said fluid sample containing said complex therein.

22. The method of claim 19, wherein said complex is preformed and added to said fluid sample, said ligand component of said complex being present in a known amount.

23. The method of claim 19, wherein said complex is preformed and prebound to said antiligand in said predetermined region.

24. The method of claim 19, wherein said complex is preformed and compartmentalized in at least one region which is adjacent said predetermined region containing said localized antiligand, said compartment being at least partially within said test volume.

25. The method of claim 19, wherein said particle is selected from the group of antiligand which interacts with said ligand, antiligand which interacts with said ligand bound to an amplifying substance, macromolecules, cells, small molecules, molecular complexes, latex beads, lipid vesicles, non-conducting polymer beads, other non-conducting particles, magnetic particles, or mixtures thereof.

26. The method of claim 19, wherein said particle is antiligand which interacts with said ligand bound to an amplifying substance.

27. The method of claim 26, wherein said amplifying substance can be further reacted to effect a change in the bulk conductance of said test volume.

28. The method of claim 27, wherein said further reaction comprises generation of a gas.

29. A method for determining the presence of a ligand in a fluid sample comprising:
(a) measuring the bulk conductance of a test volume, said test volume being proximate to at least one predetermined region, said predetermined region being exposed to said fluid sample and also having localized therein antiligand which interacts with said ligand; and (b) determining the occurrence of ligand-antiligand interaction by detecting changes in the bulk conductance of said test volume.

30. The method of claim 29, including the step of forming a complex between a ligand which interacts with said antiligand and a particle.

31. The method of claim 30, wherein said complex is preformed and compartmentalized in at least one region which is adjacent said predetermined region containing said localized antiligand, said compartment being at least partially within said test volume.

32. A method for determining the presence of a ligand in a fluid sample comprising:
(a) measuring the bulk conductance of a test volume, said test volume being at a distance from at least one predetermined region, said predetermined region being exposed to said fluid sample and also having localized therein antiligand which interacts with said ligand; and (b) determining the occurrence of ligand-antiligand interaction by detecting changes in the bulk electrical conductance of said test volume.

33. The method of claim 1, including the step of forming a complex between an antiligand specific for said ligand to be determined and a particle.

34. The method of claim 33, wherein said complex is preformed and said ligand in said fluid sample binds to said antiligand component of said complex and further binds to said antiligand in said predetermined region.

35. The method of claim 1, wherein the bulk conductance of said test volume is measured with electrodes which are in contact with an electrolyte.

36. The method of claim 35, wherein four electrodes are used to measure the bulk conductance of said test volume, two of said electrodes providing a current through said test volume whose bulk conductance is being meaured and two of said electrodes measuring the voltage drop across said test volume.

37. The method of claim 36, wherein at least said voltage electrodes are recessed from said current.

38. The method of claim 36, wherein at least said current electrodes are recessed from said test volume.

39. A method for determining the presence of a ligand in a fluid sample comprising:
(a) measuring the bulk conductance of a test volume, said test volume at least partially containing therein at least one predetermined region, said predetermined region being exposed to said fluid sample and also having localized therein a ligand, said predetermined region also being exposed to antiligand which interacts both with said ligand in said fluid sample and said localized ligand; and (b) determining the occurrence of ligand-antiligand interaction by determining changes in the bulk conductance of said test volume.

40. The method of claim 39, wherein said changes in bulk conductance are determined by comparing the bulk conductance of said test volume with the bulk conductance of at least one control volume.

41. The method of claim 39, wherein localization of said localized ligand comprises immobilization on a matrix.

42. The method of claim 41, wherein said matrix is contacted with a flowing stream of said fluid sample.

43. The method of claim 39, wherein said localization of said localized ligand comprises confining said ligand within the boundaries of a membrane, said membrane being permeable to at least said antiligand and said ligand present in said fluid sample.

44. The method of claim 39, including the step of forming a complex between antiligand which interacts with said ligand and a particle.

45. The method of claim 39, including the step of forming a complex between a ligand which interacts with said antiligand and a particle.

46. An apparatus for determining the presence of a ligand in a fluid sample, said apparatus comprising:
(a) in at least one predetermined region of said apparatus, means for localizing antiligand which interacts with said ligand;
(b) means for contacting said fluid sample with said localizing means; and (c) means for measuring the bulk conductance of a test volume which at least partially contains said predetermined region.

47. The apparatus of claim 46, wherein said region comprises at least one test region for determining the presence of said ligand and at least one control region to provide a comparative basis for making said measurement.

48. The apparatus of claim 46, wherein said localizing means comprises a matrix.

49. The apparatus of claim 48, wherein said matrix is selected from the group of gel beads, gel layers, glass beads, polymeric beads, microporous membranes, porous paper, or mixtures thereof.

50. The apparatus of claim 46, wherein said contacting means comprise a structure having:
(a) an inlet channel for bringing said fluid sample to said localizing means; and (b) an outlet channel for removing said fluid sample from said localizing means.

51. The apparatus of claim 50, wherein said structure comprises a non-conductive material.

52. The apparatus of claim 51, wherein said non-conductive material is selected from the group of plastic, ceramic, glass, quartz, or mixtures thereof.

53. The apparatus of claim 50, wherein said structure comprises at least three component parts, said first component part comprising said inlet channel, said second component part comprising said outlet channel, and said third component part comprising said localizing means, said first and second component parts being fastened in a sandwiched fashion about said third component part, said third component part being insertable between said first and second component parts.

54. The apparatus of claim 46, wherein said means for measuring the bulk conductance comprises a plurality of spaced-apart electrodes.

55. The apparatus of claim 54, wherein said plurality of electrodes are contactable with an electrolyte.

56. The apparatus of claim 55, wherein said plurality of electrodes comprises at least two electrodes.

57. The apparatus of claim 55, wherein said plurality of electrodes comprises at least four electrodes.

58. The apparatus of claim 57, wherein two of said electrodes provide a current through said test volume and two measure the drop in voltage across said test volume.

59. The apparatus of claim 58, wherein each of said current electrodes and each of said voltage electrodes is placed on opposite sides of said localizing means.

60. The apparatus of claim 58, wherein at least said voltage electrodes are recessed from current flow.

61. The apparatus of claim 60, wherein said voltage electrodes are physically recessed from current flow.

62. The apparatus of claim 61, wherein said physical recession comprises placing said voltage electrodes in chambers, said chambers being in electrical contact with said test volume via conducting channels.

63. The apparatus of claim 62, wherein said conducting channels additionally contain a matrix which retards bubble formation and convection.

64. The apparatus of claim 63, wherein said localizing means also comprises said matrix contained in said conducting channels.

65. The apparatus of claim 59, wherein said localizing means essentially entirely includes said test volume.

66. The apparatus of claim 62, wherein the width of the opening of said conducting channels ranges from about 10 um to about 1 mm, and said openings are from about 10 um to about 1 mm apart with respect to each other.

67. The apparatus of claim 62, wherein the width of the opening of said conducting channels adjacent to said test volume ranges from about 25 um to 100 um, and said openings are from about 40 um to 200 um apart with respect to each other.

68. The apparatus of claim 54, wherein said apparatus additionally includes a current-modulating material adjacent to said localizing means.

69. The apparatus of claim 54, wherein said localizing means additionally comprises a current modulating material.

70. The apparatus of claim 58, wherein at least said current electrodes are physically recessed from said test volume.

71. The apparatus of claim 70, wherein said physical recession comprises placing said current electrodes in chambers, said chambers being in electrical contact with said test volume via conducting channels.

72. The apparatus of claim 71, wherein said conducting channels additionally contain a matrix which retards bubble formation and convection.

73. The apparatus of claim 58, wherein said current electrodes and said voltage electrodes are on the same side of said localizing means.

74. The apparatus of claim 73, wherein said current electrodes and said voltage electrodes lie in a plane.

75. The apparatus of claim 74, wherein at least said voltage electrodes are physically recessed from said current flow.

76. The apparatus of claim 74, wherein at least said current electrodes are physicially recessed from said test volume.

77. The apparatus of claim 74, wherein said apparatus additionally includes a current-modulating material adjacent to said localizing means.

78. The apparatus of claim 58, wherein said plurality of electrodes also includes auxiliary electrodes.

79. The apparatus of claim 78, wherein said auxiliary electrodes comprise guard electrodes.

80. The apparatus of claim 66, wherein said test volume ranges up to about 1 ul.

81. The apparatus of claim 66, wherein said test volume is less than about 0.1 ul.

82. The apparatus of claim 66, wherein the thickness of said test volume ranges up to about 1 mm.

83. The apparatus of claim 66, wherein the thickness of said test volume is less than about 0.1 mm.

84. The apparatus of claim 66, wherein the volume of said predetermined region is less than about 0.1 ul.

85. The apparatus of claim 66, wherein the thickness of said predetermined region is less than about 0.1 mm.

86. The apparatus of claim 56, wherein said plurality of electrodes includes auxiliary electrodes.

87. The apparatus of claim 86, wherein said auxiliary electrodes comprise guard electrodes.

88. The apparatus of claim 46, wherein said test volume essentially entirely includes said localizing means.

89. An apparatus for determining the presence of a ligand in a fluid sample, said apparatus comprising:
(a) in at least one predetermined region of said apparatus, means for localizing a ligand which interacts with an antiligand, which antiligand also interacts with said ligand in said fluid sample;
(b) means for contacting said fluid sample with said localizing means; and (c) means for measuring the bulk conductance of a test volume which at least partially contains said predetermined region.

90. A sensor for determining the presence of a ligand in a fluid sample, said sensor comprising;
(a) means for localizing antiligand which interacts with said ligand; and (b) means for measuring the bulk conductance of a test volume which at least partially includes said localizing means.

91. A sensor for determining the presence of a ligand in a fluid sample, said sensor comprising:
(a) means for localizing a ligand which interacts with an antiligand, which antiligand also interacts with said ligand in said fluid sample; and (b) means for measuring the bulk conductance of a test volume which at least partially includes said localizing means.