CA1189348A - Method for the determination of species in solution with an optical wave-guide - Google Patents

Abstract

A B S T R A C T

An analyte in solution is made to react with a specific reac-tant coated on the wave-guide thus modifying the optical properties thereof. The index of refraction of the wave-guide material is high-er than that of the reaction medium which ensures that a light sig-nal injected into said guide be carried by multiple total reflection, the distance of penetration of the evanescent wave associated with the totally reflected signal being of the same order of magnitude or greater than the thickness of the analyte-reactant product layer.

Description

. -- 1 ME~ ~:)R T~E DE~I21~ION O~ SP~ 3S IN SOLUTION

W:F~I AN (~eTIQL ~V~OETIDE

The present invention relates to the use of waveguide for the determullation of the concentration of a species (or analy~ej in so~
lution in a liquid by measuring the rate (concentration dependent) of its combination with or dissociation frcm a specific reactant there--to, e.g. a oonjugate mDiety in a complexation reaction. Mbre speci-fically, the invention conoerns a method for determining an analy-te in solutio~ in which a layer of analyte-reactant product is form-ed at the surface of a lit waveyuide carrying a totally an~ multip-ly reflected elec~rcmagnetic w~ve signal and changes the optical pro-~erties thereof so as to modify said signal, said mcdification be-ing measured and used for said determination~ Thus, the invention applies to a variety of chemucal and biological systems and, parti-cularly well~ to the determination of bioactive molecules in low con-centration by im~uncassay type reactions, i.e. reactions based on the formation of complexes by the addition of antibcdy (AB1 to an-tigen (AG) molecules or vice-versa.
Man~ methods already exist in the field for achievin~ the above mentioned determination based on the classical technigues of bicche-mistry. For instance, chemic~l reactions can be used ~o detect a giv-en anal~te in a number of different ways~ Classical systems inclu-de titration or reaction with a specific reagent that gives a col ored product or precipitate. The requir~ment for this detection sys tem is that the reagent is in equivalence or in excess, so that the product can be measured b~ conven~ional photcmetry, turbidimetry, ~5 colorime~ry, etc ~ The measuring system is chosen according to the size of the signal to be measured. At very lcw analy~e concen~rations, detection beccme~ difficult and greater discrimlna~ion can be obtain~
ed, for example, by concen~ra~ing the reaction prcduct locally e.g.
by solven~ extraction, centri~ugation, etc.~ which may beccme tedious and ccstly. Hcwever, the above disadvantage was strcngly reduc~d when a practical system for the measurement o~ biochemical analytes in extremely low concen-trations was made available in 1960. This microanalytical system (radioimmunoassay) too]c advantage o:E -the cha.racteristics oE biological systems for molecular recognition :.
(an-tigen-antibody reactions) and -the extreme sensitivity of radioactive measurements (radioactive iso-tope labell.ing). An essential ;Eeature of this break-through was the concept oE
limited reagent assay with the tractor label used to measure the distri-bution of the analyte to be measured between the reagent-bound and the free moieties (see for e~ample: Revier Paper "The theore-tical aspects of satura-tion analysis" R.P. E~INS in "In vitro procedures wi-th radiosotopes in medicine", International Ato~ic Energy Agency, June 1970). Al-though i~munoassays were first described as limited reagent assays, equally practical sys-tems were la-ter described for reagent excess methods (see MILES et al~ Biochem. J. 108, 611 (1968)).
In addition to volumetric and gravimetric analysis, the present methods thus involve highly sensitive methods such as colorimetry, spectroscopy and radioactive measurements.
However, many of such techniques are now becoming obsolete as they are tedious, require a relatively large quantity of analyte to be accurate, are based on hard -to prepare and difficult to store reagents or require expensive and cumbersome equipment and highly skilled operators. Thus, there is a trend now to develope more subtle methods, which require less reagents and which can be performed safely, quickly and accurately by moder-ately skilled personnel. Among such methods whi.ch have been disclosed lately, some involve the use of optical wave guides including the reactant. For analysis, the wave guide is con-tacted with the analyte in solution whereby a reaction with the reactant on the wave guide occurs with the consequence that the optical properties of the latter are modified. The measurement of such modification then provides the required data for the analyte determination. According to the teachings oE some recent references, for instance U.S. pa-tent no. 4,050,895, the guides can consist of a porous light transmitting cora impregnated with the reactant into which the analyte will diffuse during the ,8 reac-tion. Or, the wave-guide can consis-t of a non porous light transmit-ting core (e.g. glass) coated with a porous or permeable sheath impregna-ted with the reactan-t and into which the analy-te will dif-, .

~ 3 ~ ~L~ 3~

fuse. Furthermore, in one specific case applied to immunoassay ff3~-r-E~a~e~ a rod-shaped wave-guide is coated with an antibcdy layer bcnded by diphenyldimethoxysilane and reacted with polystyrene latex spheres treated with an antigen. The antigen treated beads will then attach to the guide and mcdify the light signal output of the latter, which variation is used for the analytical de~ermination.
The above techniques have merit but they do not well apply to some typical analyses involving reaction kinetics. Indeed, it is well kncwn that rates may provide essential analytical data, particular-ly in the case of automated test systems and/ since reactions occur-ring within perm~able or porous bodies always involve a prelLmina-ry diffusion of ~he analyte into said body, and since diffusion pro-cesses are generally much slcwer than chemical reactions, the rate of the latter cannot be measured directly; in such a case, onl~ equi-librium data can be obtained. Also in the known prior-art, embodi-ments are avoided involving the use of a transparent core coated with a reactan'c sheath with a refraction index smaller than that of the core for ~he reason that, admittedly, lcw sensitivity would be ex-pected to result. Indeed in such case, most of the light signal in-jected at the input of the guide will travel within the core by atotal reflecticn process and, in such case, as is cc~monly accept-ed, the interaction of that signal with the reaction products locat-ed in the sheath, i.e. outside the core should only be minor. Con-sequently, ~are was taken in the prior art that the refractive in-dex of the sheath n2 (where the reaction takes place) be always larg-er than that (nl) of the core for allowing the ligh~ injected in the core at the input to be refracted into the sheath and~ from that point on, to ccntinue to travel in the sheath right to the output of the guideO Now, contrary to scme of the priox disclosures, the output signal (the result of the light having been mLdified by passing through the products of reaction: reactant ~ analyte within the sheath) will not readily reenter the core and reach the back end there-of (this behavior results fram elementary optical principles ~o be discussed later) and, for the measurem~nts, the ou~put light detect or must be located in the very near vicinity of the testing probe (i.e. the back end of the sheath). Such arrangement is not always practical constructionwise, namely when the guide (plus sheath) is ,g3~L~

dipped in a liquid for measuring an analyte in solution.
The present invention remedies these drawbacks as it involves no porous core or sheath and no diffusion through a matrix structure (sheath or core). Also, in the present invention, the light signal travelling inside the wave-ywide by tot~l reflection is neither trans-mitted by nor yuided within the reactant analyte product; rather, only the evanescent wave co~ponent of the signal input (i.e. that part of the wave that extends into the region outside the core in the case of total reflection) is involved. mus, since the range of action of the evanescent wave is only a fraction of wavelength (~)~
the quantity of product needed (reactant ~ analyte) is ex~remely small and utmost sensitivity (wi~h regard to the total amount of species to be analyzed) can be achieved.
m us, one object of the invention is to provide a method for the fast and accurate determination, with an optieal wave-guide, of the concentration of a chemical species or analyte in solution in a liquid. Another objection of the invention is to provide an ana-lytical meth~d for determining biological analytes with great spe-cificity and sensitivity. Another object is to provide an analytic-al method which can be easily implemented by moderately skilled work-ers and which requires only a minor amount of analytical solution.
Still another object of the invention is to provide versatile and automated measuring devioe s adapted for carrying out the above men tionned method.
These objects ~and still further objects which will appear in the course of this description) are appropriately fulfilled by ~he present method which ccmprises using a wave~guide core, the index of refraction (nl) of which is selected to be higher than that (n2) of said analyte solution and to provide a ratio nl/n2 such that the depth of penetration in the solution of the electromagnetic field associated with said light signal travelling in the guide practical-ly matches or exceeds ~he thickness of the said layer of analyte react-ant product. Thus, for instance, the method involves contacting a section of a lit (non porous) wave-guide core coated with a thin film of a reaetant specific to an analyte with a solution of said analy-te thereby enabling said analyte to react with said reactant o~ the film and form a reactant-analyte product layer, observing the cor-3~

responding optical changes with time occurring to the light signaltravelling through the core at the output of said core as the result of said product layer formation and co~paring the rate data obtainr ed wi-th standard reference data ob~ained in a similar manner from calibrating samples of said analyte, the refractive index (nl) of said core being greater than that (n2) of said solution and the thick-ness of said film being only a fraction of the signal wavelength.
It will be useful at this stage to provide so~e general inform~
ation on the transmission of light in a core by the ~o-called total reflecti~l pro oe ss. rrhis can be better done with the help of part of the accompany mg drawing in which:
Fig. l illustrates schematically the total reflection process of an incident beam N at the boundary between a dense (nl) and a ra-re (n2) medium at an angle 3 larger than ~ c~ the critical total re-flection ~ngle. In this figure, Eo is the initial magnitude of theelectric field component of the light at zero depth in the rarer me-dium, Z is the depth of penetration axis and dp is defined in the discussion below. R is the reflected beam.
Fig. 2 illustrates the fractional penetration depth of electro-magnetic field in rarer bulk medium for total internal reflectionversus angle of incid~nce for a number of interfaces. The penetrat-ion depth is infinitely large at the critical angle and is abou~ one ten~h the wavelength at grazing incidence for relatively high index media. ~ /nl is the wavelength in the denser medium. (Taken fr~m N.Y. HARRICK, Internal Reflection Spectrosccpy, Wiley 1967).
Fig. 3 illustrates the interaction of an evanescent wave and a layer of reaction product~
Physical insight into the interaction mechanisms at a reflect-ing surface can be obtained from a more fundam~ntal approach wi~h the aid of Maxwell's equations. In the present case, i.eO reflect-ion in a dense mediun at the bound æ y with a rare medium, the fol-lcwing question must be answered: what is the electromagnetic field distribution in the rarer medium beyond the reflecting interface for total internal reflection?
In this case, there exis~s a w~ve function in the rarer medium which propagates parallel to the interface. Its electric field am-plitude falls off exponentially with distance from the surface (see 3~

Fig. 1); therefore it is called an evanescent wave. In the ideal ca-se and if the rarer medi~m has no absorptive property of its own at the wavelength considered, there is no net flow of energy into the non absorbiny rarer medium, since the tLme average of the energy des-S cribed b~ Poynting's vector (see, for ins~ance M. BORN & E. WOLF,"Principles of Optics" Pergamon Press (1959)) is æero. Mathematical-ly, the electric field can be described by the expcnential function:

E = EO.e ~

m e depth of p~netration dpr defined as the distance required for the electric field amplitude to fall to e~l of its value at the surface, is given by ~1 P 2~r(sin2O-n~

where A l = ~/nl is the wavelength in the denser medium and n2_1 - n2/nl is the ratio of the refractive index of the rarer me-dium divided by that of the denser medium. The meaning of these re-lations is illustrated by Fig. 2 in ~hich the pene~ration depth ~i-vided by the incident and reflected wavelength ~ 1 is plotted versusthe angle of incidence- ~ for a number of interfa oe s ~i.e~ for dif-ferent ratios of n2/nl). It should be noted that the penetration depth is only about one-ten~h the wavelength in the cases when the differ-ence between the refraction indices is large, i.e. when n2/nl is small, this being near the grazing angle (~ = 90~. This penetrat-ion beccmes indefinitely large as A approaches Ac. At a fixed angley the penetration depth is larger in the case of small index differ-ences (i.e., as n2 1 approaches 1). The pene~ration depth is also proportional to wavelength and hence is greater at longer wavelengths.
As an example if the dense medium is slass (nl ~ 1.5) and the rare medium is an aqueous analyte (n2 ~ 1~3), n2/nl = 0.867 which corres-ponds approximately to curve ll of Fig~ 2~ In this case, the pene-_ 7 ~ 3~8 tration would be about 1/3 ~he ~avelength at the grazing angle, theo-retically infinite at the critical angle ( 60) but already below 1 at 65.
Since E decreases exponen~ially, the region beyond the bounda-ry interface in which the amount of energy for interacting with theproduct is stil]. significant corresponds to the depth Z where the electric field maynitude is still a reasonable fract.ion of Eo, say a value of at least 0.1 Eo/ better, the region in which E is between Eo and E d3. Thus, for optin~lm interaction efficiency, the thickness of the reaction produc~ film should approximately match the depth of that region4 This is illus~rated on FigO 3 which shcws an inci-dent (N) and a reflected beam (R), the zexo depth vector Eo of the evanescent wave and a film of product reactant plus analyte (A), the thickness of which approximately matches the penetration depth dp where E is about E~3. In Fig. 3, the influence of the refractive mdex of the thin film A is not considered significant because the thickness of this film does not exceed the depth of penetra~ion of the evanescent wave. Indeed, the change of refractive index of the rare medium due to the growing of the analyte-reactant film in the reaction area is so small that the corresponding change of the va-lue of the critical angle of reflection is practically negligible except for the reflec~ing n~des quite close to that reflecting an-gle~ Support to this view which constitutes an unexpected advanta-ge of the inven~ion over the prior art can be found for instance in the aforesaid N.Y. HARRICK reference, p.Sl.
Ano~her point which should be emphasi2ed for comparison with ~he prior art concerns the efficiency of the int~raction of the light si~nal with the reaction praductO In the classical spec~rsmRtri~ sys-tems, a light signal is passed through a ~ransparent holder contain-ing the analyte (beaker or cuvette) and part of the energy i9 absorb-ed by the sample which lea~s to scme degree of absorption that is measured. Yet, this method is not particularly efficient as the amount of analyte should be relatively large to provide signific2nt inter-action with the light sign~l under usual conditions. In contrast, in the present invention where the interac~ion of an evanescent war ve with a film the thicknes~ of whicn approximately matches with the penetration of that wave is involv~d, the efficiency is considerab-~ 3 ~

ly higher since there is a strong light amplification effect in theinteraction arèa. Indeed, as sh~n for instance in the previously menticned H~RRICK reference~ the field strength of the evanescent wave within its xange of interaction with the analyte-reactant lay-er is much stronger than that of the inccming signal. This is actuc~l-ly due to the simultcm~ous presence of both the incoming and outgo-in~ beam field amplitudes.
The above discussed cptical fundam~ntals useful for ~mderstand-ing the operating principles of the invention refer to the use of unpolarized light. In practice, it is important to note that the ini-tial magnitude of the electric field at zero depth (Eo) is dependent an the state of polarization of the incident light wa~e. ~lus, in some cases, polarized light in place of ordinary light can be advan~
ta~eously used in practising the invention ~it will ~e s~en herein-after that in case of measuring signal changes by ellipsometry, pc-larized light is essential) and, in such cases, the various optic-al paLameterS can be controlled and optimized for naximum response and sensitivity; for instance a selection of an apprapriate incident polarization angle (e.g. polarizatian parallel or perpendicular to the plane of incidence) can be made for maximizing Eo~
In view of the akoYe considerations, ~he following advantages of the mvention relative to the prior art can be fully appreciat-ed. Thus, in the case of a test involving one particular specific reactant for an analyte, the thickness oE the prodw t layer will usual-ly be determined by the respective size of the product molecules.Fbr instance, in a typical L~muncassay, the prodw t layer m~y be cons-tituted of a first film of an antibody and a second film of an an-tigen. m e thickness of this may range dep~n~ing on the molecule ty-p~s fr~m several Angstrcns to several hundreds of Angstroms or mo-reO ~w, in view of the thickness of the layer, the index of refract-ion of the core may be selected and also in some cases the wavelength so that the above discussed parameters will be matched as much as possible. To give an example by way of illustration, if the layer involved is relatively thin, core.s with high refractive indices will be selected (for instance, sapphire, n = 1.8; silicon, n = 3.4) and, if ocmpatible with the cptical processes involved (i.e. absorptian, scat~ering, fluorescence, etc..), shorter wavelengths will also be g ~ 3~

selected~ This will permit m m imizing interaction of the evanescent wave with regions of the analyte solution deeper than the thickness of the space where the analytical reaction is taking place, thus mi-nimizing the inEluen oe of undesirable extraneous factors (background noise, presence o~ impurities in the solution and the like). Obvious-ly, none of the methods of the prior art can achieve such possibi-lities. It should also be kept in mind in appreciating the differ-ences between the present invention and the prior art where the thick-ness of the sheath extends well beyond the penetratio~ of the eva-nescent wave (whereby the refractive index (n2) of the sheath beco-mes detenminant in contrast with what happens in the invention) and in whicb n2 is mAde greater than the index (nl) of the core, that the light signal initially injected in the core and refracted into the sheath will not readily r~enter into the core and be present at the output thereof as apparently believed by some (see for instan-ce WO 81 100 912). Indeed, when a light signal is travelling in a rarer medium surrounded b~ a denser medium, refraction of said light signal int:o the shea~h will occur. Then, this refracted wave will be total~y reflected by the outside bou~dary o the shea~h and will b¢unce back toward ~he core. Now, since the index nl of that core is snaller ~han that of the sheatb, the wave will do either of the two Eoll~wing things: if the incident angle is larger than the cri-tical angle, the wave will be again reflec~ed and will stay defini-tely in the sheath. If the incident angle is smaller than the cri-tical angle, the wave will go through the core and pene~l-ate on the other side into the sheath and so on. So, in no case will a wave ori-ginally injected into the core and havir.g been refracted in the sheath return solely in the core and be present therein at ~he output of the core unless it is still surrounded by the sheath. Thi5 is per--fectly illustrated in Fig. 3B of USP 4,050,895. No shortccming ofthat sort exists with the invention in which the key light signal only travels within the core and not m the outside layer contain-ing the reactant and analyte.
The optical changes involved in the method of the invention can relate to different kinds of phen~mena; for instance, the follcwing phen~na can be involvedO absorption of the light travelling in the core; scattering of the light signal by the reaction prcduct; fluc-rescence of the reaction product upon excitation by the lignt sign-al in the core. Further, the excitation signal in the core can be polarized and ellipticity polarization factors may be subject to mo-dification by the analytical reaction and be monitoredO Each of the-se possibilities are disclosed in more detail hereafter in this spe-cification.
m e types of analytical measurements which can be acccmplish-ed with the present methcd are so many that it is practically impos-sible to list them all. However, a few examples will be given here-inafter by way of illustration. Hcwever, before going any furtherin this direction, it is useful to develop sGmewhat the fundamentals pertaLning to ~he application of the inventi~n to "lLmited reagent"
and "excess reagen~l' assays desirably used in biological and diagnost-ic analysis~ Fbr the purpcse of such discussion, we shall convention-ally call th~ analyte the "antigen" ~ and the reagent the "antibo-dy" AB. N~edless to sa~ that the reverse ccndition is also validO
"Excess reagent" assay re~ers to cases in which an excess of a reactant in respect to the analyte is used. 'gLimited reagent" es-sentially involves the use of a system in which the test substance or analyte (containing the antigen to be measured) is treated with a limited amount of a specific reagent (the antibody) to give an ana-lyte-reac~ant product (eOg. an AG.AB complex) plus some residual ana-lyte. When ~he reaction is allcwed to go to ccmpletion, i.e. if the assay proceeds to 2quilibrium (l'saturation assay")~ that is, the li-miting conjugate reagent (AB) is saturated with the analyte, it isn~cessary to add, prior to the reaction (or, in sequen~ial assays, at some time before the final equilibrium i5 reached, i4e. prior to measurement), a fixed amount of a labelled form of the analyte (A~) to the reaction mixture being under testO For the exa~ple of an anr tigen to be assayed with an antibody reagent, the proportion of the labelled antigen (AG~) to the unlabelled one (unkncwn) shall stay the same in said residual analyte as it was at the starto Since the kncwn amount of AB used will bind a kncw~ amount of the AG + AG~ mixt-ure, it suffices to detenmine the residual ~G~ or the AG ~ bound to the AB (by means of its label) ~o calculate the amount of AG origi-nally present in the sample. To give a simplified example, suppose that the sample contains x equivalents of an enzyme (AG) to be meas-3~1~

ured by means of a kncwn amount ~ o an enzyme conjugate (AB~ thatforms an AG.AB ccmplex (with, for ins~ance, a 1:1 molec~lar ratio of both components). Then, prior to the reaction, a equivalents of the same enzyme to be measured but in labelled form (AG~) are added to the sample. Thus, in the course of the reactiont a portion of y equivalents of antigen (AG ~ AG~) is consumed by the g equivalents of antibcdy. Now, after removing the complex from the mixture, the residual AG ~ is ascertained by conventional means. If it is found, by substracting the value measured for the remaining AG~, that the amount actually used up was b equivalents, it becomes evident, sin-ce AG and AG~ are chemically indentical and consumed at the same rar te, that the ratio of consumed AG~ to consumed ~G, i.e. b/g-b should be equal to the original ratio a/x, from which x = a(g-b)/b can be calculated~
mis type of approach is quite attractive al.hough, in the prior art applications, it suffers from scme drawbacks, one of them being the general requirem~nt that the complex (mixture of AG~.AB ~ A~.A~) must be separated from the reaction medium which is scmetimes tedious and a possible source of errors. Ncw, when applied to the present invention, this drawback is non-existent because the complex that forms autcmatically removes the analyte from the solution as it de-posits onto the wave-guide~ To illustrate the application of the pre-sent mRthod to "saturation type assays" one shall again begin with a wave~guide, say an optical fi~er coated with a specific antibody AB which is immersed in~o a buffer solution and allowed to e~uili-brate with it. me unkncwn amDunt of complemen~ary antigen AG to be determined is then added as before but simultaneously with a known small amount of the s~me antigen labelled with a molecule having spe-cific optical properties (AG~) e.g. optical absorption, fluorescen-ce, etc..) which may be detected by the evanescent wave interactionat the surface of the coated fiber using suitable optical arrange-ments to be described in this specificationO ~ow, since both label-led and unlakelled AG are essentially identical in reactivity tcwards the AB-ccated fibre optic~ but only the l~belled species can be de-tected via its label, the apparent change in the op~ic 1 propertydetected through the fibre te.g. fall in absorbance if an absorbing label is used) will be inversely proporti~nal to the unknown concen-3~1 tration of AG and can be determined with re~erence to a suitable se ries of kncwn standards. m is kind d application is illustrated in one of the Examples hereinafter.
At this point it is usef~ to make a clear statement concern-ing the label used to measure the reaction that takes place. ~ormal-ly, in imm~noassays, it is not possible to measure the analyte di-rectly because the concentrations of analytes and reagents are ex-tremel~ lcw. Sin~e the equilibrium mixture in limited reagent assays essentially only contains excess analyte and a fixed am~unt of bound ccmplex, when the former cannot be measured directly and the latter is a fixed amount, no quantitative estimation of the origin~l ana-lyte con oe ntration can be obtained even after separation of the cam~
plex and the excess antigen. The added labelled tracer (a small quan-tity of Labelled analyte) is necessary to allGw measurement via the lab 1 according to its distribution between the bound moiety and the free moie~y.
If tke analyte, hcwever, has an intrinsic property that can be detected when it i5 concentrated locally (i.e. in situ separation and in situ ooncentratian), e.g. on the surface of a fiber optic pro-be, then the addition of a labelled analyte tracer is no longer ne-cessary. Thus, in the present invention, the in situ concentration of the analyte-reagent oomplex may allow for the detection of the analyte without res~rt to a labelled tracer. Hcwever, local concen-tration of the analyte reayent oomplex (since the amount of reagent is fixed) will not allow the quantitative determination of the ori-ginal analyte concentration unless the reaction between analyte and reagent is measured kinetically or unless the total amount of ana-lyte is less than ~he numter of reagent bonding sites. Thus, us m g the in situ seFaration and concentration of the ccmplex bound anar lyte in conjunction with a sensitive detection system in the kine-tic mcde as in the present invention allcws quantitative detecticn of the analyte in limited reagent system without resort to a label-led tracer.
As far as labelled systems æ e concerned, a distinction can be drawn between ~racer systems in which a labelled version of the ana-lyte is added in tra oe amounts to the reaction and the labelled rea-gent systems in which a lakel is attached ~o the specific reagent.

13 ~ 3~

The former tracer reagent is nor~ally used in limited reagent assays (e.g. s~andard radio immunoassays), while the latter is normally used in excess reagent assays (e.g. standard sandwich assays) and can be used in certain forms of kinetic reagent assays (e.g. KRCNICK et al, U5P 3,939,350); labelled analyte and labelled reagent systems are suitable for the present invention if the label used is capable of being detec~ed in an optical wave guide system (e.g. absorbing or fluorescent label~).
For the next part of the discussi~n, reference will be made to another part of the annexed drawing in which Fig. 4 is a diagran illustrating a first type of assay called 7'direct type of ~ssay".
Fig. 5 is a diagram illustrating a compekitive "limite~ reagent"
type of assay.
Fig. 6 is a diagram illustrating an indirect conpetitive ~91i-mited reagent" type assay.
Fig. 7 is a diagram illustrating a sequential l'saturation" ty-pe assay.
Fig. 8 is a diagran illustrating a "sandwich" type assay.
The m~st straightforward case of assay to which the invention is applicabie is schemati ed on Fig. 4. m is is called the "direct"
t~pe assay. In this assay, a wave-guide core 1 of which only a port icn (with reractive index nl~ is represented is provided with a film of antibody AB and this core is immersed in an analyte solution hav-ing refractive index n2 (nl ~7 n2) containing an antigen A~to be de-termined. The antigen will attach to the A~ molecules at a rate pro-portional to the antigen cQncen~ration ~AG?3 (since the amount of AB film on the core is a fixed en~ity) and when this rate is deter-mined, it can be correlated with standard rates Qbtained from cali-brat m g AG solutions and ~AG~ can be determined. So, the am~unt of AB available can be "limited" or it can be in excess and the reac~-ion can go to an equilibri~un. Fbr detecting the AB-AG complex form-ation by means of the optical changes occurrin~ in the core, the va-rious a~oresaid techniques can be used ti.e. extinction of the sig-nal by absorption, scattering and fluorescence phen~mena, etc..),provided the formation of AB.AG generates the required optical chan-ges. So, the test applies particularly well to large molecules able - 14 ~ 33~

to scatter light or having ob~ervable properties at determin~d war velengths or emitting fluorescence under excita~ion by certain wa-velengths. If such properties are missing, then the test is of lit-tle value and a "limited reagent" competitive type tes~ shall be pre-S ferable used with an amount of labelled AG (AG~) added to the ana~lyte. This is illustrated in Fig. 5 in which the signs (letters and numerals) used are like in Fig. 4. In this test, the amount of AG~
needs not to be kncwn provided it is standardized (i.e. always cons-tant for a set of calibration curves and tests), sin oe the rates of the analytical reactions with various guantities of AG? will always be in direct relation to the AG~/RG? ratio (beLng naturally keFt in mind that the observable optical change in the wave guide is due on-ly to the labelled reaction product AG~.AB).
~n another very useful variation of the "limited reagent" ty-pe assay, the wa~e-guide core is coated with a known or fixed am~unt of the same antigen (pre~erably in pure form) which should be deter-mined in the analytical sample, after which the core is contacted with the sample and, simultaneously, a fixed amount of free detect-able antibcdy is added. This is illustrated in Fig~ 6 in which the signs are the same as used before. It is seen on the figure that the reference amount of antibody AB will simultaneously react with the reference AG on the core and wi~h the AG? to be measured. mus, the observed rate (as depicted b~ the optical changes oc~urring in the oore) will be related to the AB/~? mole ratio and correlation with standard reference rate curves will provide the desired results. Ob-viously, in this case the AB mwst have observable properties in the optical me~hcd used for the test, i.e. AB can have absorbance at suit-able ~'s or be adapted to scatter light or the liked Alternatively if AB is not directly ob6ervable on the guide when reacting with AG, it can be labelled e.g. coupled with a fluorescent or any other opti-cally detectahle label.
Still another system involves the scrcalled sequential testing illustrated by Fig. 7. In this test, there is provided a core with a fixed amount of ABr this amount being in excess of the correspond-ing equivalen~y needed to bind the AG? to ~e determined. So, in afirst sbep, the coated core is ccntac~ed with the sample whereby the avail~ble AG? is bound and thus rem~ved from the sample~ Then the - 15 ~ 33~

core is contacted with labelled ~G~ which will fill the voids in the antibcdy layer on the core. Measuring thereafter or during reaction the o~tical changes due to the label of ~ will give the necessa-~y data for calculating the original amount of AG?.
Finally, the present invention also applies well to the assay case called "sandwich" a.ssay relating to the determination of anti-gens having ~ore than one bonding sites for antibodies, i~e. anti~
gens having sites numbered (1~, (2)o~ (n~ capable of bonding with 1, 2 or n different antibcdies. This is schematically represented on Fig. 8 where the first antibody is indicated as AEl, the seoond antibody as AB2 and a two-site antigen as ~2)AG(1). In this case, it is assumed that the antigen is not detectable per se optically whereas AB2 is detectable after reacting on the appropriate second site of AG. Thus, in this case in which (2)AG(l)? is to be determin-ed, a core is used with an initial reference coating of a first an-tibody (AEl) and is ccnbacted with the an-tigen solution. ffl e latter will thus bind to the core by its first binding site after which the second antibody ~E2 in reference amount is added and its rate of bind-ing on site 2 of the antigen is measured in view of the optical chan~
ges occurring in the core as the result of this binding. Of course in this procedure, the first reaction ABl ~ (l)AG(2) can be allow-ed to go to equillbrium before adding the reference quantity of AB2;
or, alternatively, a simultaneous type test can be und~rtaken, i.e.
the AB2 can be added b~ the solution simultaneously with the contact-ing thereof with th2 AEl coated guide core. This is particularly ap-plicable where ABl and AB2 are different mcncclonal antibcdies (see for example, UCTIL~ et al. Jc~rnal of ~mmunological Methods 42 (1981) 11 15). Also, any variations of the afore described assay system can be adapted by those skilled in the art without depaLting from the spirit of the inven~ionr the operating parameter of any of such variation being defined with reference to calibrating solution samr ples of the analy~e.
Exa~ples of devices for carrying the various embodiments of the me~hod of the invention in the fields of optical absorption, fluo-rescence and scattering as well as ellipscmetric measurements willnow be presented with reference to a subsequent part of the acccm-panying drawings in which, 16 ~ a 3 3 ~t 8 FigO 9 is a schematic represen~ation of an apparatus for measur-ing in situ changes in the light signal travelling through an optic-al fi~er caused by the deposition of a oomplex film on said optic-al fiber, Fig. 10 ~epresents schematically on an exaggerated sc~le, a fi~
ber optic probe usable in the apparatus of Fig. 9; Fig. lOa is a part-ial top view of said probe and Fig. lOb is a side sectional view the reof along the line B-B of Fig. lOa~
FigO 11 is a schematic representation, on an enlarged scale, of another probe embcdiment particularly adapted for light mode va riaticns measurements. Fig. lla is a front view and Fig. lIb is a side view, Fig. 12 is a diagrammatic representation of the light signals at the detector of an apparatus similar to that in Fig. 9 but used for fluoresc~nce me~surem~nts. Fig. 12a concerns the situation be~
fore the reaction is startedO Fig. 12b ccnoerns the situation sGme time during the test.
FigO 13a represents schematically (tcp view) an apparatus for m2asuring optical changes caused in a wave guide by light scatter-ing.
Fig. 13b is a partial schematic side view of the apparatus of Fig. 13a.
Fig. 14 represents schematically on an exaggerated scale part of a m~dification of the e~bodLment of Figs 13a and 13b~
Fig. 15 represents schematicall~ part of another m~dificat~on of said embodiment.
Fig. 16 repres~nts schematically ~n apparatus for measurin~ op~
tical changes caused in a wave-guide by fluorescence phenomena.
Fig. 17 represen~s schematically on a much exaggerated scale a portion of the appara~us of Fig. 16 to show the angles of reflec-tion of the incident and Eluorescent light involved.
Fig. 18a represents schematically (top view) an apEaratus for measuring ellipscmetrically optical changes in an optical fiber caus-ed by the formation of a complex film on said fIber.
Fig. 18b is a cross-secticnal view of a portion of such appa-ratus along line A~A in Fig. 18a.
m e apparatus shcwn schematically on Fig. 9 essentially compri-- 17 ~ 3 ses the follcwing co~ponents:
a) A fiber op~ic l, the central part of which passes thralgh a container or cuvette 2 for ho1ding a liquid analyte to be deter-mined; the cladding of the fiber section Lmmersed in the liquid has been removed so that ~his section can be coated, before operation, with a thin film of a specific ccmplexing reagent of the species dis-solved in the li~uid and which should be determined. m e assembly of the fiber l and the holder 2 constitutes the test probe of the aFparatus.
b) A chopper disk 3 rotated by a motor 4 and provided with two diametrically oppcsed windows with filters S and 6, plus a hole 7.
c) A main light scurce 8, a collimating lens 9, an a~nular aper-ture 10 and a focu6ing lens 11 for injecting into the probe fiber core an angularl~ ~elected light beam. In this particular embodiment, the fiber is multimcde and the aperture lO is arranged for passing specific modes acco~ding to some requirements to be discussed here-inafter. Hcwever, with modifications to be further discussed later the pres~lt aFparatus can also be used m conjunctiQn with lower mc~
de fibers, e.gO single mode. The focussing lens, represented sche-matically on the drawing is actually an optical system analogous toa microscope objective but which can be adjusted, in addition to its displac~ment along the Gptical path, sidewise and up and dcwn for accurately positioning the beam in front of the fiber front end.
d) Then, the aFparatus ccmprises a main light detector 12 for transfoDming the exit light signal rcm the core into an electric signal, a lock-in a~plifier 13 and a display device 14. A referen-ce signal is also provided by another ~ ce lS, the ligh~ of which is pulsed b~ passing through hole 7 of the chopper and directed to a detector 16, said detec~or ~hen applying corresponding electric pulses to the lock-in amplifier 13.
One embodiment of an analytic probe usahle with the present apr paratus is depicted on Fig. 10 (Figs lOa + lOb) which shcws a plast-ic holder tank 2 in two oppssite sides of which S shaped grooves or slots 21a and 21b have keen cu~. A portion of a fiber optics 1 with its cladding 22 has been inserted into and is being held by said groo-ves. The cladding of the middle portion of the fiber has been remov~
ed, for instance, by etching with a suitable solvent~ and thereaf-- 18 ~ 3 ter coated with a thin film 23 of a specific complex conjugate of the species to be determined, such species being represented by the small squares 24.
m e description of the cperation of the present device then fol-S lcws: Filter 5 is first selected for passing through the core a wa-velength ~l which is modified by absorpticn at each internal reflect-ion site because of absorption of part of the evanescent wave by the coating of the complex material that will form between the film 23 and the dissolved molecules to be analyzed 24. Fbr illustrative pur-pcses, it may be said that 23 represents a very thin layer of an an-tibody (AB~ and 24 represents molecules of a specific an~igen (AG) to be determined, the assay being of the "direct type" kind illus-trated on Fig. 4. Filter 6 is selected for passing a wavelength ~2 through the core that is essentially not absorbed by such oomplex ccating and henoe not affected by the immunological reac~ion involv-ed. Thus~ when ~he chopper 3 rotates two signals ~l and ~2 are al-ternately fed to the probe core, one signal ~ l that will be progres-sively abso~bed wi~h grcwth of the coating 23 + 24, and another sign-al ~2 which will serve as a reference signal i.e. for calibration and oompensation purposes (cell or fiber replacement, etc.~). m e probe is preFared by selecting a suitakle length of optical fiber (multimode in the case shcwn on Fig. lO), inserting it into the groo-ves 21a c~nd 2Ib so that the middle portion is held horizontal in the cuvette 2; water tightness is provided by then filling the grooves with a rubber grout or cement soluticn and allowing it to dry. Then the cladding of said middle portion is etched away. For a resin clad fiber, this is done with a suitable organic solvent for this clad-ding; for a gla~s clad fiber, the etching is performRd wi~h dilu~e hydrofluoric acid. In this last case, i~ is not necessary that tbe cladding be e~ched ~ompletely for reasons explained below. Tben, af-ter the dissolving or etcbing solution has been remcved and the cell cleaned thoroughly with distilled water, the fiber core i~ ooated with a film of antibody (AB3 by usual mean~, i.e. by fillina the cell with an AB solution so that AB depcsits anto tbe core as a uniform layer (if, b~ any chance, the surface of the fiber has not sufficient affinity for the molecule to be deposited thereon, it can priorly be made ~onding by the special treabments kncwn in the art, e.g. graft-ing bonding sites,applying an intermedia-te reac-tive layer, etc ln this respec-t, an abundant literature exists on the subjec-t (as disclosed in European Pa-ten-t no. 29~11 granted November 10, 1983 to sattelle Memorial Institute). After emp-tying the cell and rinsing, the la-tter is mounted on -the appara-tus at -the place inclicated between the lens ~ and the detec-tor 12 and a suitable bu:Efer solvent is introduced therein. The apparatus is then activated and the chopper is rotated at a convenient speed, e.g. 120 rmp. The signals ~1 and ~2 are fed to the core of the probe and the detector 12 conver-ts them into approximately squared electric pulses which are fed to the lock-in amplifier 13. A reference locking si.gnal i9 also provided once every turn, via the hole 7, by source 15 and detector 16 enabling the am-plifier to distinguish (by coincidence) between the pulses and ~ 2. In prac-tice, the controls are preferably set up so that the said pulses have about the same magnitude before star-ting the reaction. Then, at zero time, the sample to be analyzed (a solution of AG) is added to the cuvette of probe 2 and mixed rapidly with the buffer (for instance with a stirrer or a gas bubbler not shown). The signal due to ~\1 then starts to pro-gressively change at the exi-t of the core of the probe, as a consequence of the molecules of the analyte AG binding to the fiber core (for building the AB.AG complex) and said complex absorbing a part of the evanescent wave energy of ~1~ at a rate proportional to the concentration of AG, corresponding variations being provided to the amplifier in the form of the corresponding electric pulses from the detector 12. Thus, -the computing circuits associated with the lock-in amplifier 13 will compute the data resulting from the ratio of the ~1/ ~2 signals and pro -vide the results to the display system 14, for instance in the .
form of a rate curve recorded on a chart (of course, any other type of display can also be used if desired: digital or os- ;
cilloscope, etc.. ). The obtained rate data are thereafter .
compared to standard data obtained from solutions of AG of known concentrations and the unknown concentration is pro-vided by interpolation. Such calculations can be made manually .

:.

. .

or can be done automatically by means of a microcomputer, thereference data being stored in the memory thereof. By selection of appropriate periods during the time course of the reactlon, -the desired reaction ra-te data can be selec-ted and dis-tinguished from other rate data due to the ~5 i.

',' ~ l9 a -, .

-~ lB~34~

reacticn of inte~fering and undesired reactions proceeding more or less simultaneously ~ut at different rates. Additionally or alter-natively, the equilibrium conditicns can be derived by extrapolat-ing and by this the time taken to do each measurement in comparis-on with normal equilibrium measurements is reduced. Further consi-deration about this question can be found in "Kinetic Versus Equi-lihrium Mbthods of Analysis" by B.W. RENOE et al., in "Centrifugal Analysis in Clinical Chemistry, PRICE & SPENCER; PRAEGER (1980) and Analytical Chemistry S0/02, (1978), 1611 - 1618.
It is u~eful to note at this point that the sensitivity of the measurements can be varied depending on the light mcde orders apFlied to the fiber input, i.e. changing the aperture 10 annular setting and width. mis is easy to understand when it is remembered that the absorpticn phencmena pertaining to a totally reflected beam inside of a core and due to a cladding outside said core only affect the evanescent wave component of said beam, i.e. the electric compcnents penetrating into the cladding. Thus, the total reflec~ion angles of the modes selected must be 5hallow enough with regard t~ the fiber axis to ensure full reflection (even when the refractive index of the aqueous solutiGn i5 slightly modified by the grcwing of the co~at-ing surroundin~ the f~ber core) and steep enough to provide a suf-ficient density of reflecting sites along the fiber (indeed, it is only at such sites that the evanescent wave is interacting with the complex coating). Thus, a test sensitivity optimizaticn can be reach~
~5 ed by properly ajusting the aperture parameters, this being depend-ent on the respective core, test solution and oomplex refractive in-dices. Such adjustments can be determined by those skilled in the art for each type of measurements and can be then design incorporat-ed for the intention of field operators.
It is al~o in~eresting to note that by changing the settings of the aperture 10 parameters, the present apparatus can be made to operate in a different "class". For instance, by prcFerly adjusting s~ch aperture, the apparatus can be made to operate with light mo-des in the near vicinity of the critical angle of total reflectîon ~c for the starting system considered. Then, in the case when in the course of the test the difference ~ n = nl-n2 shall decrease with ocm-plex formation, the angle of total reflection at the core-coat mg 3~

boundary will change such that sc~e of the m~des (or all) will be suddenly refracted outside the fiber and sharp cut-off of the sign-al will occur. Such arrangement will there~ore provide ubmost sen-sitivity for very small amo~nts of analyte lecules. It is hcwever less weLl adapted to quanti~ative measurements.
It has been said above that, when using a glass d ad fiber opt-ic, such cladding need no~ be entirely etched away for the present use. Indeed, since during operation, the evanescent wave will a~tual-ly penetrate the d adding a few tens of nm, a residual glass clad-ding around the core with a thickness below that necessary for theevanescent wave to significantly interact with the reaction rnaterials is still possible which impcses less stringent control conditions on the etching operation. Also, ~hicker fibers are less fragile than thinner ones.
15A type of fiber holder that will multiply the sensitivity of the test by a larga factor is pictured schematically on Fig. lla and b. This holder consists of two helically grcoved flanges made of an inert plastic (plexiglass or polyester or instance) on which a piece of cptical fiber is wound and clamped by clamps 33 and 34. The front and back ends of the fiber, respectively 35 and 36, are bent side-ways so as to be usakle with the light injecting lens 11 and detect-- or 12. The middle straight sections of the windings 37 of the fiber are bare while the remaining curved sections 38 resting in the grooves keep their original protective cladding. The etching of the bare port-ions is done b~ fir3t covering the lcwer flange of the holder with a suitable siliccne rubber cement solution and letting it dry. The-reafter the holder i~ immersed in an etching solution dawn to the lower face of the upper flange whereby all the intermediate sections of the cladding will be removed, the lGwer rubber protec~ed flange remaini~g untouched. Fbr doing the tests and after coa~ing the ba-re portion of the fiber with ~G or AB as before, the holder is cor rectly positioned with ~he aid of a hook 39 in the optical path of the apparatus. Then a cuvette with the reacticn medium is brought in from belcw, the re~gent is added and the measurements are done as before.
It should be pointed out that all the above discLssions on the present invention concerned the use of fiber op~ics to be used in - 22 ~ 3~

an aqueous or organic medium the refractive index is near to that of water (i.e. around 1.3) whereas the index nl of the glass core of the fiber i5 rather near to 1.5 and the index n2 of the active cDating is in between i.e. ~ 1.5 but 7 1.3~ This condition is essent-ial here since i~ n~ were ~7 nl the light would be refracted into thecladding, ~hen back into the core and so on. As we have seen befo-re, such method is not i~possible per se, for instance, it has been disclosed in Nature 2 (1975) by HARD~ et al who worked with glass rod wave-guides and had to use rather thick polymer coatings contain-ing, embedded, some reagents to be determined, but it would not bepractical in the present case since the surrounding solutions always have refractive i~dices s~aller than the fiber core nl and only ve~
ry thin coatiny~ (of the order of a few ~ngstrom to a few tens of nm~ are involved. In ccntrastl H~RDY et al o~erated in air and did not measuxe rates of reactions.
Another important point that should be discussed is the quest-ion of mono~mode fibers. Indeed, with very minor ~odifications kncwn in the art and mainly pertaining to light injection into the fiber, single-mode fiber optics can be used as well in the prese~t method~
~0 Such fibers generally have a very thin oore (a few microns) surround-ed by a relatively thick cladding and their transmission is strong-ly wavelength dependent, e.g. a iber with only 5% attenuation per meter at 850 nn ~ay have attenuation of about 40% at 800 and 900 nm (see T.G. GIAILCRE~ZI in Proceedings of ~he I~ 66 (1978), 748).
~ow, for monormode conduction in a fibre at ~avelength ~ th~ follow ing expression V = 809a ~ n/~ must have a value below 2.4 ~n = nl-n2 and a is the ~iber radius in microns). Above V = 2~4, the fIber is no longer single-mode. When the value V is small (e.g. below 1), the guided single mode beccmes more lGosely bound, i.e. the field spreads considerably beycnd the pbysical core of the fiber (e.g. two or three times the core diameter). Thus, the ~ransmission characteristics of single-mcde fibers are very sensitive to tiny changes ~n changes) in the refractive index o the cladding, i.e. that of the organic coating which forms during the test of the present invention. Indeed, this refractive index is the key variable involved in the present assay performed with single-mode fibers as the thickness cf the ccm~
plex layer grcws duri~g the reaction. Whereas internal reflections 3~

in a multimcde fiber allcw the evanescent wave in the immediate vi~
cinity of the reflective bo~ndary to pass parallel to the core for a very short distance (which can be defined by Maxwell's equations) at each reflection, the mono-mode fiber allcws for the passage of this evanescent wave frac~ion parallel to the core along the whole effective length of the fiber. m us, the use of singl~-mode fibers in the present apparatus facilitates an increase in sensitivity as ccmpared with multi-mode fibers.
Single-mcde fibers can be used wit~out etching the original clad-ding completely sinoe the guided light fields extend significantlyinto the cladding and also into the oomplex layer formed during the reaction.
The mcdificaticns which must be made to the apparatus when using s m gle-mode fibers are essen~ially Qf an optical nature, i.e. inject-ion o~ the light i~to the fiber and detecting the signal. Such optic-al variations are not describ~d here as they are known from people skilled in the art and well described in the literature (see for ins-tanoe the above GIALLORENZI article and the references therein).
The present aFparatubs can also be used for performing the me-thod of the invention but using fluorescence effects instead of ab-sorption or refractive index changes. For doing this, the following modificatiGns should be made thereto:
a) A fluorescent labelled antigen (AG~) will be added to the analyte AG. m us, the me~hod will operate under the oonditions of 'llimit~d rea~ent" ~munuassays as discussed hereintofore (see Fig.
5).
b) -Filter 5 will be selected for a ~ 1 specific of the excitat-ion of the fluoresc~nce to be m~asured and filter 6 will be select-ed for blocking i~l but passing the fluorescence emission wavelength 3o ~2.
c) An additicnal filter of optical characteristics identical with filte~ 6 will be inserted between the fiber back-end and the detector 12.
Then, the probe with bufrer mediun and its fiber optics 1 ccat-ed (as usual) with an ~B film will be installed and the apparatuswill be adjusted for response as shown diagramatically in Fig. 12a.
In this figure~ there is shown a succession of pulses frcn the de-- 24 ~ 33~3 tector 12; the pulses marked R are the reerence pulses preduced when filter 6 is in the beam; the plllses E are the emitted fluorescence pulses injected into the fiber by the fluorescent complex A~.~G~ mo-lecules that ~orm onto the fiber core. Thus, at the start, the E le-vel (Fig. 12a) is about zero from no fluorescence or onl~ residualback-gro~d noise. Ihen, the analyte containing the antigen to be determined plus a known quantity of fluorescent labelled PG~ is ad-ded; a fluorescent coating graduall~ forms at a rate proportional to the ~G concentration and an emission fluorescence signal appears in the fiber and progressively increases with time as shcwn on Fig.
l~b. This variation is detected, applied to the amelifying and camr puting sections of the apparatus and recorded as the rate measure-ment desired. Then, afterwardsr visual or computerized comparisons are made of the reaorded data and standard referenoe data from which the desired analy~ical results are obtained using the techniques spe-cific to "limited reagent" assays discussed aboveO
In the present embodiment, the fluorescence genera~ed by the coating outside the fiber core is substantially reinjected into the fiber and provides the signal used for the test. A related phenome~-non has been recently described (see The JouLnal of Cell Biology 89,141-145 (1981) with regard to a glass block.
When using a fiber optic advantage can also be taken frcm a great sensitivit~ as the useful length of the fiber can be made quite si-gnificant although oonfined in a very small space and using very small volumes of solutlons. It should also be mentioned that the measure of the fluo~escence of coatings in immuncrassay has already been re-ported (see M.N. KRCNICR et al, Journal of I~munological MetbQds (1975), 235 - 240). However, in such case, the f'uorescen~e was mea-sured through the antibody solution a~d only one internal reflection site was used which gave poor sensitivity since only a very limit-ed portion of the total fluoresc nce emission could be processed In the case of using fiber optics like in the present invention, si-d~wise fluorescence pick up c m also be measured, for example, by using a fla~-coiled piece of fiber and placing the detector axial-ly to the coil thus collecting a larger part of the flu~rescent lightenitted by the fiber~
The present embodiment provides many advantages over the pre-- 25 ~

sent testing me~hods and devices kncwn in the art. For instance~
a) The sensitivity being proportional to the length of the imr mersed fiber 5~c~ion, very sensitive probes can be made altholgh hav-i~g sm~ll dimensions.
5h) Measurements can be made over a wide range of wavelengths, in the visible, W and IR bands.
c) A wide variety of substances (bioloyical or non-biological) can be tested among which one can cite drugs, haptens, enzymes, pep, tides, proteins, hormones, bacteria, viruses and cells. A more com-plete list of detectable analytes can be found, for example in US
pat~nts Nos. 3,817,837 and 4,299 t 916.
Qne interesting specific ca e is when testing blood samples for transfusion with regard to possible antibodies in the recipients blood. Thus, optical fiber probes can be prepared with a film con-taining the blood oonstituents of said patient and tes~ed against the blood cells of potential donors. In case of cross-reactivity, the cells will precipitate onto the fi~er (because of the reacticn of their ~wn ~G oenters with the AB of the fiber) and this reaction can be easily monitored by one of the typical absorption bands of hemc~lobin (e.g. at 555 nm). Another interesting specific case is the pcssibility of using a wave-guide sensor in vivo to make quali-tative or quantitative measurements of analytes within or secret~d from the body. For example, in diagnosis, it would be Fossible to assay the quantity of circulating insulin in response to a glucose loading test and, ~n treatment with injected honmones, the in vivo sensor w~uld detect the circulating hormcne concentration~ The p~s-sibility of using in vi~o sensors has ncw been described by EUCKLES, but the present inven~ion would have the added advantages of absen-ce of a label in the ~mmunoassay mode (where many labels are high-ly a~tive oampounds with possible toxic or carcincgenic e~fects) andimproved light transnission thus obviating the need for special trans-mission fibers to couple the signal in and out of the sensor as des-cribed by EUC~LE5.
The aFparatus shcwn schematically on Fig. 13a and 13b compri-ses the following key components:
An optical probe cc~pc~ed of a slanted edge wave~guide plate41 held by brackets (not shcwn) and a oounter pla~e 42 maintained - 26 ~ 3~3 in accuratel~ controlled parallel facing relationship in respect of the wave-guide 41. The plate 41 can be made of hiyh quality float glass accurately cut or quartz. The plate 42 can be a microscope sli-de. The space 41a between the plate~ 41 and 42 is of the order of a ~raction o~ ~n; this ensures that an analyte solution can be easi ly introduced into this space whexe it will be held by capillary for-ces~ The apparatus further oomprises a light source 43 (in this par-ticular embodiment, the light ~ouroe is a He-Ne laser providing pcr larized light), a chopper disk 44 act~ted by a motor 45 and provid-ed wi~h a hole 46 for providing a pulsed light signal to be sent torward the optical probe throuyh an optical system comprising focus-ing lenses 47 and 48, a mirror 49, a polarizer plate 50, a cylLndri-cal lens 51 and a diaphragm 52 for minimizing stray light. me reason why the light signaL is pulsed is to provide for ultimate ~hase sen-itive amplification of thP detected signal.
The apparatus further ccm~rises a photomultiplier detector 53with its high voltage pcwer supply 54, the outlet of which is con-nected t~ a lock~i~ a~plifier 55 through a two way switch 56. The photcmultiplier is arranged to collect the light scattered by the product that forms on the wave-guide plate in the space 41a as the re~ult of a chemical reaction occurring between a reactant and an analyte of the svluticn filling space 41a. The scattered li~ht is indicated by arrows 57.
The apparatu~ further co~prises a det~ctor 58 for collecting the light emerg mg at the output of the wave-guide core af~er mul-tiple reflections therein as shown on the drawing. The output of this detector can be alternatively fed to the lock-in am~liier 55 through switch 56. Th~ references signals for the apparatus are provided by a semi-transparent mirror 59 which derives a small portion of the light of the source on a detector 60, the latter providing an inten-sity reference signal fed to an analog divider 61, this being for ccmpensating possible variaticns of the source during the measure-ments. A pulsed locking reference 5ignal iS provided by a mirror 61 and a detector 62, the oorresponding electric cignal being fed to the ampliier 55. The apparatus inally comprises an elec~ronic ~nit 62 which will comprise a displa.y element (like in the previous emr b~diments) and recorder for recordin~ the outpu~ data and, optional - 27 ~

ly, a microprocessor for co~puting the measuxed data and effecting the required csmpari~ons with the stored reference data ~rcm cali-brating experiments.
In practice, the present apparatus is operated as follcws. The first step is to preFaLe the optical probe by depositing a film of a reactive species on the fa oe of t.he wave-guide 41 that defines the space 41a. Details on how to do this are given hereina~ter. Then the guide 41 and the plate 42 are mcunted on the apparatu~ and the opr tical and electronic systems are started upO After a few minutes of warming up, the oontrolq are adjusted for zero response fro~ photo-multiplier 53 (or alternatively, full transmission frcm detector 58).
Then the analyte solution ~a few ~ll) is plpetted into space 41a where-b~ the analyte in solution begins reacting with the reactant film on the guide 410 When the analyte-reactant product furnishes scat tering sites (large molecules, agglomerates, etc..) a scattered sig-nal reaches the photcn~tiplier tube 53 which is amplified, proces-~ed and displayed with time by the display included in the unit 62.
This re~ul~s in a rate curve; the slope data are recorded and com-puted against standard data obtained from calibrating samples and tored in the memory of the electronic unit 62. This ccmputation pro-vides the desired analytical results (e.g. the concentration of the unknown species in the anal~te solution) according to usual means.
If the output of detector 58 is not used, the edge 41b of the gui-de 41 can be masked to minimize the formation o back scattered light.
Masking c3n be done for instance with black paper or paint.
An improvement to the aforedescribed embodiment is illustrat-ed on Fig. 14. Instead of using a perfectly hcmogen~ous working sur~
face for the wave-guide 41, one can first treat that surfaoe to pro-vide discontinuities thereon. F~r instance, one can modify the in-herent adhesiveness of the surface tcward the reactant according toa certain pat~ern (achievable for instance by the known photolitho-graphic techniques). In the case represente~ on Fig. 14, the surfa-ce 41c of the guide 41 has been slightly roughened at areas indicat-ed by numeral 41d, such areas being as parallel stripes about one wavelength wide and separated by a distance of the same order of ma-gnitude. Such a grating-like Fattern can be obtained by first cover-mg the surface with a photo-resis~, exposing said resist layer ~ 28 through a photographic filn with the negative image of the grating, developping (i.e. dissolving the unexpcsed areas in a suitakle sol-vent~ and etching slightly the bare areas, after developping, for instance with HF. After final removal of the resist, the plate has a grating pattern of stripes with alterna~ing zones of higher and lcwer affinity for proteins (antigens or antikGdies). Thus, when the plate is contacted with a reactant (AB), the latter will mainly at-tach to the etched areas as indicated on Fig. 14 by the letters AB.
At this stage~ the thickness of the pattern i5 not sufficient to pror vide scattering; hcwever, in the presence of the antigen (which is of the proper nature to scatter light), the latter will bind to the stxipes having thereon the antigen i.e. as indicated by AG cn the dra~ing. This discontinuous type of layer will provide distinct or-ders of difracted sc2ttered light in contrast to the stray scatter-ing mode of the main e~bodiment thus improving the directivity andthe collecting efficiency (by the photcmultiplier tuke) of the scat-tered signal. In Fig. 14, the arrows 65 indicate the zero order of diffractian and the arrcws 66 indicate the first order o~ diffrac-tion. By collecting a specific order (e.g., the first) the ccllec-ticn of the signal by the photcmultiplier tube 54 is then stronglyamplified and the signal-to-noise ratio is increased in this modi-fication as comFared with the collection of the stray scattered light which is only aro~nd 10~ of the totaL scattered light.
In another modification, see Fig. 15, a perfectly flat and re-gular surface 41c of a wave-guide was sprayed wi~h microdroplets of antibody (A~) and, during analysis, the antigen AG attaches only to such preferential areas. Such a structure constitutes a statisti~al scatterer which scatters light in the shape of a cone the size and geometrical parameters of which depend on the size and the distri-bution of the dropletsO Thus, also here, there is a directicnal ef-fect that contributes to increase ~he efficiency of the signal col-lecticn. Another advantage of providing diffracted light signal is to minimize the importance of accidental scatterers such as dust part-icles or scratches in ~he glass with larger sizes (of the next high-er order of magnitude, i.e. about 10~ or more~; in such case the an-gle of diffraction is smaller and, the resulting diffracted light not reaching the photcmultiplier tube, its presence can be neglect-39~

ed~
The apparatus repres~nted on Fig~ 16 is generally similar to that already discussed with reference to Fig. 13 but it is adapted for fluores oe n~e measurements instead of ~cattered light measurements.
5 miS apparatus essentially con~rises an optical system represented by block 70 which is peactically identical with the system used in the apparatus o Fig. 13. m erefore, the details are not repeated for simplification. This system 70 contains a pulsed light source and means for providing a test signal of the correct excitation wa-velength properly directed to a wave-guide 71 and a reference sig-nal to a referen oe detector 80. Like in the other emkodiment, there is a plate 72 determining with guide 71 a thin space 71 for intrc~
ducing the analyte solution, the layer of reactant being coated on the upper face of the guide 71. The apparatus also comprises a si-de-detector (the photomultiplier t~be 73 and its supply 74), a co-re output det~ctor 78 and blocking filters 76 and 77 (these items are missi~g in the previous embodiment). The electronic oompcnents of the apparatus oomprise two integrator amplifiers 81 and 82, a mul-tiplexer 83, an analog to digital converter 84, a microprocessor 85 an~ a display recor~er 86. All these elements are con~entional and their operations are familiar to thcse skilled in the art.
The operativn of the aEparatus in the side-pickup mcde is qui-te simiLar to that described for the scattering embodiment. Thus af-ter preparing the wave-guide 71 with a layer of reactant on its UE-per surface and thP plate 72, the analy~e solution is introduc~d toprovide a fluorescent reactant-analyte product. The optical unit 70 sends a test signal of the proper wavelength ~ l for exciting the fluc-rescence of wavelength ~2. The emitted fluorescent light goes across screen 76 (which screens off all other wavel~ngths including ~cat-tered exciting light) and hits the photo~ul~iplier tube 73 wherebya signal proportional to the fluorescence (and to the extent of the reaction) is produced. mis signal ~ together with the reference sig-nal from detector 80 is fed to the multiplex am~lifier 83 which al-ternatively feeds them to the remainder of the electronic elements whereby a processing similar to tha~ descri~ed previously will oc~
cur.
Simultaneously or alternately with the aforementiQned cperation~

_ 30 ~ 33~

the signal occurring at the core output and falling on detector 78 can also be monitored. Fbr underst~nding how this is pvssible, re-ference is rnade to Fig~ 17. On this fi~ure, there is shown a portion of the guide 71, the plate 72 and, in between the space 71a and, sche matically, a layer of AB.AG product (the product which is the result of the reaction of the reactant AB plus the analyte ~G of which the rate is being measured in the test). The figuee also shcws the var rious ligh~ beams involved in the excitation to fluorescence process, i.e. N is the incident beam at Al, R is the reflect0d beam of wave-length ~l and ~ are the generated bac~ward and forward fl~orescentbeams (~2)~ ~ i i5 the angle of incidence and ~ic the critical angle for ~ l ~fc is the critical angle foL A 2. N~w, as represented, the excitation light N hits the internal surfac~ of the wall at an an gle ~i larg~r than the critical angle ~ ic ~nd it is thus reflected ~R~. However, part of the evanescent excitation wave is absorbed by the ABcAG layer and energy is reemitted at a shorter wavelength ~ 2.
Now~ by virtue of the reciprocity principle (oonir~ed quantitati-vely by C.R. CORNIGLI~ et al. in J.O.S.A. 62, (4), 1972, 479-486), excited molecules emit evanescent photons which behave exactly as ~he incident evanescent wave photons. Thus, the fluorescence emit-ted b~ m~lecules clo~e ~o the dense-to-rare interface (the AB-AG lay-er) propagates into the dense medium at angles larger than the cri-tical angle ~ fc for A~. hctually/ peak fluoresoenoe is observed when ~i is close to ~ic and at angles close to ~fc (see R.E. ~RENNER et al., Fiber Optics, Advances in R & D, Providence, RI, U5~, 19-23 Ju~
ne 1978, NY Plenum Press (1979l). Thus, the emitted fluorescence mar xLmum inte~sity concentrates wi~hin a relatively 3nall angular ran-ge ~nd, when guid~d in the wave-guide, the output a~ se~eral inter-action sites add up to provide a higher intensity signal at the out-put of the oore. A second point of importance is that, because therefractive indices are different for different wavelengths, the ex-citing signal and the fluorescent signal are following paths with different reflection angles in the guide and they will energe from the output also with angles different frcm one another. ~ence, the-re is an inherent optical sepæ ation of the emitted beam frcm theexci~ation beam at the output of the core and detector 78 can be suit-ably placed to be in the path of the fluorescent signal (~2) while - 31 ~

avoiding the residual excitation beam (Al) even in the absence of ilter 77.
The apparatus represented on Figs 18a and 18b is intended for measuring optical changes occurring in an optical fiber wave-guide provided by the formation of a product on the surface of said fiber, this measuring being performed by means of ellipsometry. Ellipscne-tric measurement applications to bioassays have been detailed in our related Canadian application No. 405,379 and the prësent reader is referred to this disclosure for general considerations in this field. The main difference of the present application with the prior disclosure is the use of an optical fiber instead of the generc~lly known flat reflecting surfaces. Ellipsometry is based on the measu-rement of changes in the degree of elliptical polarization of light caused by the presence of a filn of matter deposited on the reflect-ing surfa oe s. In view of the particular nature and geometry of op-tical fibers as compared to flat surface wave-guide, additional featu-res must be considered. Firstly, in order to maintain a beam of light linearly polarized in a fiber, one must use a special family of light modes. These mcdes are defined as the HElm modes, with m = 1, 2, 20 3..... n (n being limited by tbe size and refractive index of the fi-ber oonsidered, see E. SNITZER and H. 06IE~BERG J. Cpt. Soc. Am 51, 499 (1961~). Secondly, to be able to measure the full extent of any polarization modification caused by a reaction occurring on the sur-face of the fiber, it is important that signals with any direction 25 of polarization ~e transmitted equally well along the wave-guide.
Thus, geometrical perturbation such as that caused by mechanical stxess on the fiber should be avoided because, in such case, one par-ticu1~r direction of polarization could be favored. Consequently, in the present apparatus, the optical probe (of which the key com-30 ponent is the optical fiber) is made short and rigidly immobilized to minimize the undesirable effect of vibration or other possible perturbations~ J
In order to preferer.tially excito the H ~ m modes in an optical fiber, the radial light intensity should have a Gaussian distribu-35 tion, e.g. that obtained from a laser radia~ing in the fundamental mode and the beam should be directed axially and centrally on the fiber end. This coupling can be achieved by means of a microscope .. .. ... . ... . . . ~ . . .

3~

objective. Now, for minimizing background perturbations, it is pre-ferable to operate with only one mode rather than several modes 5i-multaneously. In the present apparatus, the HEll mode (lowest sin-gle mode~ was selected for simFlid ty. In principle, multi-mode fi-bers are usable with single mcde signals~ This can be made possible(iOe. the proper Gaussian distribution can be selected) by letting the incident beam diameter at the Eiber input face be O.65 times the diameter of the fiker core. ~owever, this is not really satisfactc-ry because other modes with various polarization states are also ex-cited to scme extent and a partially depolarized signal is o~tain-ed at the probe output. Hence, it is greatly preferakle to use a sin-gle-node fiber in this embodiment since this will essentially only transmit the ~ 1 mcde whereby no conversion to higher order nodes and ~hus no decay in the degree of polarization will occur. Conse-quently, a higher signal to~noise ratio is ob~ained as ccmpared withusing multi-mode fibers.
In the present enbodiment, the fiber le~gth is preferably 10 cm or less for the aforementioned reasons. It is preferably a high quality single-mcde fiber with respect to dapolarization and bire-~ringen~e characteristics~ Fibers obtained by CVD techniques suchas that disclosed in H. AULICH et al., Applied Optics 19, 22, 3735 (1980) are preferred.
The apparatu~ represented on Figs 1 & and 18b comprises an op-tical probe comprising a partially etched piece of optical fiber 101 with a core lOla and a symmetrically rencvsd cladding lOIb. The tech-nique for removing the cladding is the same as ~hat disclosed for oth~r optical fiber applications in this specification. ~t a very short dis~ance from the bare fiber co~e (a fraction of a mm) on b3th sides thereof are placed tw~ glass plates 10~ rigidly maintained on "U" shaped brackets 103. ~he clad end portions of the fiber wave-gui-de are cLamped by means of clamps not shcwn mounted on an adjusta~le rack not shcwn~ ~he latter enabling the position of the fiber to be accurately controlled sidewise and up and da"n. The sp2ce 104 be~-ween the fiber core and the plates 102 is the reaction site, the li-quids to be tested being retained b~ capillary forces between the~iber and the plates.
The present apparatus further ccmprises t~o microscope object-- 33 ~ 3~3 ives 105 and 106 for coupling the input incident beam with the fi-ber and the output elliptically polarized signal with the detecting system, respectively. The incident beam is generated by a light sour-ce 107~ (in this embodiment a He-Ne laser) the output light of which goes across a polarizer plate lOB and a rotating chopper disk 109 driven by a motor 110 and having a hole 111 for chopping the signal.
Then, the detecting systen of the apparatus comprises a quarter-wa-ve plate 112, a polarization anal~zer 113 and a photodetector 114, these elements being similar to corresponding elements described in related Canadian application No. 405,379 and operating similarly. Final-ly the apparatus comprises a lcck-in amplifier 115 and associated cornponents, i.e. micro-ccmputer 116 and recorder 117, the functions of these electronic ccmponents being essentially the same as that of the corresponding elements already described earlier in this spe-cification. It is to be further noted that the reference chopped sig-nal used for reference purpose is represented on the drawing as ori-ginating frcm the chopper tor 110 and being transmitted to the am- ;
plifier 115 by a llne 118.
The operation of the apparatus is very similar in essence to that described in the previous embcdimer.tscombirled with the opera-tion of the a~paratus described in related ~Çanadian application No. :~
405,379. Thus, for making a test/ the bare portion of the fi ber is ~irst coated with a reactant (e.g. the antibody) as shcwn by numeral 120 on the drawing by filling the space with the proper so-lution as shown on the drawing by dashed patterns. Then after the usual rinsing and drying stages, zeroing of the instrument in the presence of a blank buffer by properly alternatively rotating the polarizing elements 112 and 113 (see description in application No~
EP 81 810 255.0), and removing the hlank solution, the solution to be tested is introduced for starting the reactiRn between the anti-gen analyte and the antibody coating which results in the grcwing of the reactant-analyte layer on the wave-guide core and consequent modification of the elliptically polarized output. The correspond-ing change from the photodetector 114 is amplified and monitored in the electronic associated ccmponents 116 and 117 and provides the desired data exactly as in the previous embodiments.
For the immobilization of a film of antibody on a wave-guide - 34 ~

core (or, of course, of antigen if this is the antibcdy that must be determined) many methods known in the art can be used as alre~-dy mentioned hereinbefore. In carrying out t'ne present invention, with reference to immunoassay testiny, it is generally preferred to take advantage of the fact that most types of glass will strongly adsorb proteins via their hydrophobic or hydrophilic areas. Thus, in such case, the etched fiber is first cleaned in an aqueous deterg-ent (e.g. a 2~ aqueous detergent solution). Then, it is rinsed unr der running water and immersed overnight in conc. H2S04. Thent it is rinsed with distilled H2O and dried in warm air. It then readi ]y binds proteins from solutions, e.g. human IgG in buffers.
In anot'ner preferred method, the active portion of the guide after cleaning with H2~04 and rinsing as above, is immersed for an hour in a 15% (w/v~ of TiC14 in an anhydrous organic solvent like acetone or ethanol. T`nen, it is washed with distilled water and 0.1 M
phosphate buffer (pH 7) and, thereafter, contacted with a h~man IgG
solution (2 g/l in 0.9~ NaCl solution or 0.1 M p'nosphate buffer, p~I
7)~ The fiber (or rather the probe) thus made ready for testing against AG's is stored wet in a suitable buffer (or even dry for short-er periods).
The Exc~Tples that follow illustrate the practical aspects of the invention.

Exa~le 1 An optical probe including a wave-guide oE the type described with reference to Figs 13a and 13b and a glass counter plate was pre-pared for undertaking immunochemical assays as follo~s: the glass items made of the following type of glass (n = 1,523): "Farbloses Brillenrohglass B-260, DESAG", were first washed with a warm deter-gent solution, rinsed in distilled water and air dried. ~hen, they were immersed for 5 min in concentrated H2SO4 at 9SC, rinsed in dis-tilled water and dried with clean tissue~ me subsequent manipula-tions were then carried out with utmost care the clean glass surfa-ces not being touched and the probe handled by the edges. The pla-tes were mounted on the apparatus as illustrated on Figs 13a and 13b, leaving a space of 0 3 mm between them. ~br coating purposes an ap-- 35 ~
3~

proximately (but accurately weighed) 1 9/1 antigen solution was pre-pared by dissolving the required amount of solid human immunoglobu-lin (5ERU~ BIOCHEMICALS, Heidelberg, Germany) in 0.9% NaCl solution.
About 0.2 ml of such solution was placed between the glass elements 41 and 42 of the probe by means of a syringe and allowed to stay there for two hrs at room temperature. During this time, approximately 1 to 10~ of the AG available in the sample became attached to the sur-face of the wave~guide (about 2-20 JUg). Then the cell was emptied by absorbing the liquid with an absorbent material and the cell was rinsed with distilled water, the rinsiny waters being remsved as be-foxe. Then a 0.1 M phosphate buffer (pH 7) containing 2 q/l of bo-;r~ vine serum albumin and 0.5 ml/l of liquid detergent (T~EEN 2~ was placed in the cell and lef~ ~here for 1 hr. m e serum albumin would fi U the "holes" left empty on the glass after coating with the anr tigen (i.e. since the antigen d oe s normall~ not attach to all avail-able areas on the glass and, sin oe the antibody to be tested has al-so affinity for the bare glass, the presence of uncoated areas, the "holes'l on the wave-guide, might subsequently introduce errors in the measurem~nts). Since the antibody has no affinity for the bovi-ne serum albumin, this treatment would there simply disahle the unrused bare portions of the glass. The detergent actually supplements such actionO
~ fter again washing and rinsing, the cell was allowed to dry in a current of pure warm air. It was then ready for the experiments.
All analysis were dane at rocm temperature. The various cptical and electronic components, except the photomultiplier tube, were turn-ed on and allowed to equilibrate and a sample of antibody solution (2 ml) was introduced into the cell by the same means. The antibo-dy selected was rabbit anti-human imm~noglobulin frcm D~KO IMMU~O-CEEMIC~LS, Copenhagen, Denmark in 0.9% NaC1. The initial solutionas r~ceived was a 10 g,~l protein solution with a titre of 900/ml, i.e. this number (provided by the manufacturer) means tha~ each ml of the antibcdy solution ac~uall~ neutralizes 900 ~g of the ant~gen (n2utralization here means that when such reciprocal quantities of both ingredients have reacted, the solution contains no re appre-ciable AG or AB). Thus, for calibrating purposes~ known antibody so-luticns oonveniently diluted were used for perfo~ming the test (see - 36 ~ 3~

Table I). A~ scon as the cell was fi11ed with the calibrating solu-tion (n = approximately 1.33), the room was darkened to avoid satu-ratin~ the photomultiplier tube and the latter was switched on. The rate curve generated from the scattered light signal picked up by the photomultiplier started to develop after some seccnds of equi-libration (probably due to the time necessary to properly wet the proteins of the coating). The reaction was allowed to proceed for about 2-5 min and the slope of the rate curve (nearly linear) was averaged over that period.
The results are summarized in Table I below which gives the var lues of dilution of the samples of antiserum and the corresponding calculated titre values as well as the slopes of the corresponding rate curves as measured with the apEaratus used.

TABLE I
; ~ . . .
Dilution of antiserum Antikody titreSlope of rate (1 ml ~ ml of 0.9 NaCl) of antiserumcurve (mV/min~
___ 1 + 20 42.9 2~.0 1 + 100 8.9 12.6 1 + 200 4.5 6.1 1 ~ 1000 0~89 0.~

Fbr the analysis of an unknown sample, the same procedure was follow d, ~he rate data being recorded and, afterwards, compared to the standard data of Table I. The desired analytical data were ob~
tained by plo~ting the analytical rate value obtained against the corresponding titre as yiven by said data of Tahle I.
~ ' The reverse oE the experiment described in Exa~ple 1 was per-formed with the same apparatus and under the same conditions. In this case, the wave-guide in the optical proke was coated with a film of antibady (rabbit antihuman ~mmunoglobulin) b~ means of 0.2 ml of the initial solution of titre 900/ml (see Example 1). All other oFera-_ 37 ~ 33~3 tions we~e the same as described in Example 1. m e calibrating sam-ples were prepared frcm kncwn dilutions of the antigen solution as given in Ta~le II which also summarizes the slopes of the rate cur-ves recorded.
S

T~BLE II
__ Antigen solutionsSlope of rate curve ~ (mV/
0~2 4.9 1.63 11.3 ~.3~ 190 0 10.16 21.7 Fbr analyzing unkncwn samples of AG, the procedure already des-crihed in Example 1 was follcwed, i.e. the slope of the correspond-ing rate curva ~nder test was measured and the value obtained was correlated with the corresponding antig~n concentra~i~n by ccmpar-ing wi~h the grap~ prepqred frcm the values summarized in Table II.

In this Exa~ple, the signal obtained from the light emerging25 frcm the back-end o~ the wave-guide and f~lling on detector 58 was used for generating the rate curve. m e oFtic~l probe wa~ essential-ly the same as that used in Example 2 (with a ooating of ~B, rabbit anti-human immunoglobulin). Samples of antigen solutions (like in Example 2) were used the ooncentrations of which are shcwn in Table III belcw. m e remainder of the operations were carried out exact-ly as explained in E~amples 1 and 2 with the difference that the pho-tomultiplier 53 was not used and that the signal actually decreas~
ed as the reaction went on (this is because the proportion of light not reaching the output of the wave-guide increases as the reaction proceeds). m e slo~e of ~he rate curves recorded as a functlon of the amount of AG present in the samples is given in Tahle III.

~ 38 T~BGE III

Antigen solutions Slope of the rate curved ~ (mV/~lin) ` ~------_ __ __ __ __ loO -0~42 O ~ 32 ~0 ~ 35 O ~ 10 -O ~ 31 0.03 nok readily measurable me followin~ ~xample illustrates a competitive assa~ method of the type discussed with reference to Fig. 5. Fbr this, a fluores-cen~e measuremen~ type of apparatus of the kind disclosed with re-ference to Fig. 16 was used. The optical probe was prepared by coat~
ing the c~rresponding wave-guide 71 with a film of human immunoglc-bulin (AG) by the same pro~edure described in Ex2mple 1. me anti-body AB~ used was labelled with fluorescein isothiocyanate (~rTC) a~d specific to the antigen tested (it was obtained from DAX~ IM-MUN~GLCBULDN5, Copenhagen~. m e unlabelled corresponding antibody AB was also obtain~d from the same source. In the following experi-ment, the output face 71b of the wave-guide was c~ated with black paint to munLmize scattered incident light.
In each experiment, a fixed quanti~y of traoer antib~dy AB~ was added. This qu~n~ity was 100 ~ul of the Dako produc~ diluted l/200 in 0~1 M phosphate buf~er at pH 7 and it was muxed with 100 Jul of the test solution of unlabelled AB in the dilutions given in Table IV. The cperation was performed as in the previous Examples, the mix-ed 200 Jll por~ions being injected between plates 71 and 72. me re-sults are reported in terms of the slopes of the corresponding ra~
te c~rves in Table IV.

~ _ 39 - ~ 3~

q~BL~3 ~
__ Dilution of antibcdy ~ntibcdy Slope of rate (1 ml ~ ml of buffer) titre 5~
1 + 20 42.9 11.6 l ~ 100 8.9 57.7 l + 200 4.5 90.3 l ~ lO00 0.89 95.5 For ~he measurem~nt of unknown solut.ions of antikody, the sar me techniqye was used (addition of lO0 ~1 of the labelled antibody3 and ~he results were obtained with reference to the standard data given above.

The previous Exa~ple was repeated but, this time, the output from detector 78 at the back-end o~ the wave-guide was nitored.
Using the same test concentraticns given in Table IV, the correspond-i~g rate results (in increasing order) were obtained (m~/min): 2.9;
408; 6.1; 6.6.

~ .
A probe cell was prepared ~y taking a piece of plastic clad si-lica fiber from "Quartz h Silice SA", type PC5 Fibrosil ~SF' W se-ries (diameter 380 Jum, core 200 Jum, losses 0.14 db/m at 350 nm, 97%
transmission) and inserting it by its ends in~o the grooves of a PVC
cell as pictured on Figs lOa and lOb. For caulking, a solution of silioone rubber cement in acetone was used. me cladding was etch-ed away by filling the cell with concentrated ~2S04 and leaving it there for an hour, (unclad section about 10 cm lcng), then the fi-ber was made active by the first preferred method described at the last p~ragraphs before the Examples.
~ film of human Lmmunoglobulin G (IgG~ purchased frcm 'IServa Biochemicals", Heidelberg, Germany was then deposited on the fiber ~ 40 ~ 3~

as follows- the phosphate buffer in the cell was removed and repLac~
ed by a 2 g/l solution of the IgG (10 ml) in a 0.9% NaCl solution~
After 2 hrs, the antibody solution was removed and the fiber was rins-ed with the buffer solution after which a 2 9/1 solution of bovine serum albumin oontaining 0.5 ml of "TWEEN 20" (2 detergent) in the 0.1 M phosphate buffer was intrcduced and left there for one hour.
men, t~e cell was again rinsed with buffer and filled with fresh buffer ( 10 ml). The cell was installed on the optical bench of the apparatus of Fig. 9, filter 5 being selected for ~ 1 = 280 nm ~pro~
tein absorpticn) and filter 6 passing~ 2 = 340 nm (no protein absorpt-ion), then, adjustmen~s were made in order to center the beam, ha-ve proper response of the electronic units and allcw the system to equi7ibrate with the minLmum of noise. men, 50 Jul of various dilut-ions of anti-serum (anti-IsG raised in rabbits purchased from "Da-ko Inmunochemicals, Copenhagen) were added and thoroughly mixed bybubbling air through. The recorder was then started at zero time and the absorption changes at ~ 1 = 280 nm were monitored for about 10 min. me average rate of change in this absorption signal was plot-ted and ~he slope was recorded in arbitrary units. m e results are sh~l in Takle V as a function of the AG concentration.

` T~BLE V

AG (anti IgG) cancentration SloEe of the rate curve ~g/ml) (mV/~in) 4.5 2.23

2.3 1.02 1.5 0.78 0.76 not readily measurable lhe method was essentially as described in the foregoing exam-ple but replacing ~he rabbit IgG by a corresponding labelled anti-s~
r~n. me labelling c~po~d was fluorescein-isothiccyanate "is~ner-1"

-- 41 - ~L~ 3~

(FITC) and the labelled antiserum was obtained from "Dako". This mar terial strongly aksorbs at 492 nm (this wavelength is actually al-so that which excites fluorescein). In ~he experiment, filter 5 was selected for a ~ 1 = 492 nm filter 6 for a ~ 2 ~ 600 nm where no ab-sorption occurs. m e tests were performed as already described withvarious concentraticns of FITC-labelled of IgG and also using, in another set of ~easurements, active fiber sections of only 5 cm. The results are reported in Takle VI.

TABLE VI
__ Concentration of labelled Slope of rate curve ~mV/min) IgG in test cuvet.te ~g/ml) for fiber lengths 5 and 10 cm 0.827 10.2 22.6 0.568 6.8 14.3 0.350 4.1 8.9 0.178 2.1 4.3 20 O.OgO 0.98 2.1 0.045 0.43 0.94 It was seen in the previous Example that tagging with a fluo-rescein deriva~ive could be used to modify the absorption spectrum of a protein and, consequently, increase the sensitivity of the test of the present invention with regard to the ~mlabelled species. Of course, this system can also be used to measure unkn~wn concentrat-ion of unlabelled human Ig~ tcompetitive type measuremen~s) by using, in additicn, a fixed concentration of labelled AB , the ratio of which over that of the labelled ~pecies will then be rate determinant ac-cording to the sch~me shcwn on Fig. 6. In the e.Yperiment~, the re-sults of which are reported in Table VII, the fixed concentration of labelled IgG was 1 Jug/ml. The values reported in the first column are that of unlabelled IgG.

- 42 ~ 3~i~

TP~

Unlabelled IsG Slope of rate curves concentration (~g/n~)(arbitrary units) O (i.e. 1 ~g/ml of 24.3 labelled ~gC7) 0.09 1907 0.83 10.2 ~.14 2.4 9.2 0.34 m e apparatus of FigO 9 was operated under fluorescence measu-rement conditions, i.e. the filter 5 or the excitation light was for A 1 = 492 nm and the value for filter 6 and for the additional excitation light hlocking fil~er after the fiber back-end was for ~2 = 518 nm~
m e experimQnt was actually perfomed exactly as in Exa~ple 8 but measurin~ an increase of signal (emitted fluorescence~ instead of a decrease as in absorpticn phencmenaO As in Example 8~ the ana-lyte cDntained 1 Jlg/mL of fluorescent IsG ~ various concen~rations of unlabelled human IgG a~ shcwn in the left oolumn in the Tah~e.

l~BLE`VIII
__ Uhlabelled IgGSlope of rate curves concentration (~lg/ml)(arbitrary units) 0 17.9 0.09 13.0 0.33 5.3 4.14 1.~
9.20 0.42

3~

It should ~e d ear to the reader that the results disclosed in Examples 1 to 9 have been used as the standards for comparison pur-poses with similar samples of unknown concentrations. Co~parisons and oomputations can be done, as usual, visually or by electronic processing in the microcomputers attachable to the disclosed appar ratuses.

.

3~

Claims (21)

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

1. A method for determining an analyte in solution in which a layer of an analyte-reactant product is built up at the surface of a lit wave-guide core carrying a multiply totally reflected light wave signal, this layer formation having the effect of changing the optical properties thereof so as to modify said wave signal and said modification being measured and used for said determination, which comprises using as a waveguide a core with an index of refraction (n1) higher than that (n2) of said analyte solution and having a ra-tio n1/n2 of such value that the depth of penetration in the solut-ion of the light component associated with said wave signal in the core practically matches with or exceeds the thickness of said ana-lyte reactant product.

2. The method of claim 1, comprising contacting a section of a lit non-porous wave-guide core coated with a thin film of a react-ant specific to an analyte with a solution of said analyte thereby enabling said analyte to react with said reactant of said film and form a reactant-analyte product layer, observing the corresponding optical changes with time occurring to the light wave signal travel-ling through the core at the output of said core as the result of said product layer formation and correlating the rate data so obtain-ed with corresponding standard reference rate data obtained in a si-milar manner from calibrating samples of said analyte.

3. The method of claim 1, comprising coating a section of a non-porous wave-guide core with a reference quantity of the analyte to be determined, injecting a light wave signal into an input of said section, contacting said lit section with the analyte solution whi-le adding beforehand or simultaneously thereto a reference amount of reactant for causing the latter to competitively react with the analyte in the solution and said reference quantity on the core, ob-serving the corresponding optical changes with time occurring to said wave signal at an output of said core section as the result of said reaction between the reactant and said reference quantity and cam-paring the rate data so obtained with standard d reference rate data obtained in a similar manner from calibrating samples of solutions of said analyte.

4. The method of claim 2, in which there is added in the ana-lyte solution a tracer quantity of pure analyte in labelled form, the presence of said label in the reactant analyte product layer on the wave-guide core causing the said optical changes to the light wave signal travelling through that core, the magnitude of said chan-ge being in direct relation with the ratio of labelled analyte to analyte in the solution.

5. The method of claim 2, comprising using a core with a coat-ing of a reactant in excess of that stoichiometrically needed for binding the analyte in the solution to be determined, contacting said coated core with the analyte solution for totally binding said ana-lyte to be determined, adding to the analyte solution a reference amount of analyte in pure labelled form to react with the still un-used excess of reactant, the presence of said label in the reactant analyte product causing the said optical changes occurring to the wave signal, the magnitude of said change being in direct proport-ion to the ratio of the labelled analyte to the; analyte originally in the solution.

6. The method of claim 2 in which the analyte posesses more than one binding site for specifically binding more than one kind of reac-tants, comprising contacting said core coated with a first reactant with said analyte solution, adding a reference quantity of a second reactant for having said second reactant to bind to a second bind-ing site of said analyte, the said optical change being the result of the binding of said second reactant on the said second binding site of the analyte bonded by its first binding site on said first reactant of the core.

7. The method of claim 2 in which said rate data are interpret-ed in terms of the slope of the rate curves pertaining to the reac-tion under test or the rate data are extrapolated to determine equi-librium conditions.

8. The method of claims 1 to 3, in which said change in the optic-al properties of the guide core refers to the absorption of the wa-ve signal, said modification leading to a decrease with time of the light signal measured at an output of said core.

9. The method of claims 1 to 3, in which said change in the optic-al properties of the guide core concerns the generation of a fluo-rescent light signal, said signal increasing with time as measured at an output of said core or sidewise thereto.

10. The method of claims 1 to 3 , in which said change in the optical properties of the guide core concerns the scattering of said light wave signal, the extent of such scattering increasing with ti-me and being measurable partly sidewise to the scattering region and partly at an output of the core.

11. The method of claim 1, in which a polarized light sign-al is used and in which said change in the optical properties of the guide concerns the elliptical polarization parameters of said sign-al, said parameters being measured at an output of said core.

12. The method of claim 1, in which the core is select-ed from light transparent plate, rod or fiber-like items with refract-ive index at least 1.4.

13. The method of claim 11, in which the core is a single-mo-de or multi-mode optical fiber core.

14. The method of claim 12, in which a multi-mode fiber is used, and comprising using modes sufficiently shallow, with regard to the fiber axis, to ensure total reflection and sufficiently steep to en-sure a high linear density of light signal coating interaction si-tes.

15. The method of claim 12, in which the core is a multimode fiber and in which the modes are selected with an initial reflect-ing angle close to the critical angle, whereby in the course of the reaction a slight change of the refractive index n2 of the rare me-dium that occurs due to the growing of the analyte-reactant product will result in partial or total refraction of the light outside the core.

16. An apparatus for measuring kinetic parameters in a reaction of an analyte with a specific reactant thereto said reaction occur-ring on the surface of a wave-guide and causing detectable changes to the optical properties thereof, which comprises a light source, means to inject a signal from that source into the input of said wa-ve guide, detecting means to detect the light signal having under-gone changes when travelling therethrough and emerging therefrom and converting it to an electric signal, and means for processing said signal into useful data pertaining to said reaction.

17. The apparatus of claim 16, wherein the wave-guide is select-ed from materials of refractive index higher than that of the reac-tion medium.

18. The apparatus of claim 17, wherein the wave-guide is in the form of a plate or a fiber of refractive index in the range of about 1.4 to about 3.5 and shaped to ensure that the light signal is trans-mitted through said guide by multiple total reflection.

19. The apparatus of claim 16, wherein the modified light sig-nal emerges as a scattered, fluorescent or partly absorbed light sig-nal.

20. The apparatus of claim 16, wherein the wave-guide is in the form of a spirally coiled optical fiber.

21. The apparatus of claim 18, comprising a counter-plate ar-ranged in non-touching close parallel relationship with the wave-gui-de plate, the reaction taking place in the space provided between said plates, the reaction medium being held in place by capillary forces.