EP2300147A1 - Method for the fabrication of a biosensor on a semiconductor substrate - Google Patents

Abstract

The invention relates to a method for the fabrication of a biosensor on a semiconductor wafer, comprising steps which comprise the forming of a central photosensitive zone (114) comprising at least one pixel device for biological analysis comprising a photosensitive layer, and a first peripheral zone (116) surrounding the central photosensitive zone, comprising electronic circuits. The first peripheral zone (116) is covered with a hydrophilic coating, and the central photosensitive zone is covered with a hydrophobic coating. A barrier of a biocompatible resin (119) is formed on the second peripheral zone (118).

Description

 METHOD FOR MANUFACTURING A BIOSENSOR ON A SEMICONDUCTOR SUBSTRATE

The present invention relates to a biological analysis device of the pixel type.

The present invention also relates to a CMOS biosensor comprising a plurality of pixel-type biological analysis devices arranged in a matrix of pixels.

The present invention also relates to a method of manufacturing a pixel-type biological analysis device and a method of manufacturing a CMOS biosensor comprising a plurality of pixel-type biological analysis devices arranged in a pixel array. The present invention generally relates to improvements to the biological analysis techniques described in US Pat. No. 6,325,977 or International Application WO 2006/082336.

These techniques envisage the use of pixel-type biological analysis devices comprising: a photosensitive layer and means for collecting photoelectrons in the photosensitive layer, and

means for reading and processing an electrical quantity supplied by the collection means for supplying a characteristic value of a light intensity detected by the photosensitive layer. In these documents, commercial CCD or CMOS imagers are used for the study of solid organic molecular structures, that is to say structurally stable and resistant to even extreme experimental conditions, such as for example oligonucleotides (RNA, DNA, ...). In these applications, the oligonucleotides can be directly grafted onto the outer surface of the biosensors. But biosensors made from such imagers are much less suitable for more fragile structures such as proteins, which, they easily lose their functionality, by changing their three-dimensional structure or denaturation, under the effect of temperature or dehydration. The intrinsic properties of the biomolecules of this family make their study standardized, multiplexed and miniaturized (in the format of a biosensor) remains at the moment, still delicate. However, they would make it possible to envisage applications in the field of the diagnostic test of infectious diseases, allergies, pollution, in the detection of drugs, or in the field of the defense, etc.

It may thus be desirable to provide a biological analysis device capable of using more fragile structures than the oligonucleotides envisaged in the documents of the state of the art.

One embodiment of the invention relates to a method of manufacturing a biosensor on a semiconductor wafer, comprising steps of providing a central photosensitive zone comprising at least one pixel-type biological analysis device comprising a layer photosensitive, realize a first peripheral zone surrounding the central photosensitive zone, comprising electronic circuits, covering the first peripheral zone with a hydrophilic coating, covering the central photosensitive zone with a hydrophobic coating whose roughness is chosen so that a single dose a solution that can subsequently be deposited on the central photosensitive zone does not spread beyond it until reaching the first peripheral zone, and form a barrier of a biocompatible resin on the second peripheral zone, the resin which binds naturally before polymerization onto the hydrophilic coating without endre on the hydrophobic coating.

According to one embodiment, the method also comprises the steps of making, in a second peripheral zone surrounding the first peripheral zone, electrical contact pads connected to the electronic circuits of the first peripheral zone, and covering the second peripheral zone by means of hydrophobic coating.

According to one embodiment, the hydrophilic coating is silicon oxide.

According to one embodiment, the hydrophobic coating is silicon nitride.

According to one embodiment, the method comprises a step of depositing, on the surface of the photosensitive layer of the pixel-type biological analysis device coated with the hydrophobic coating, a mixture of capture method comprising a probe protein grafted to a hydrogel, for the capture of target proteins.

According to one embodiment, the step of depositing the capture mixture comprises steps for preparing the outer surface of the photosensitive layer intended to receive the capture mixture, comprising a step of silanizing the outer surface; oxidation of the hydrogel and grafting of the oxidized hydrogel on the silanized outer surface; carboxylation and activation of the grafted hydrogel; and grafting the probe protein into the activated hydrogel. According to one embodiment, the step of preparation of the external surface by silanization comprises the following phases: cleaning of the external surface by successive rinsing with demineralised water, acetone and absolute ethanol in a hot water bath. ultrasound; alkaline oxidation of the outer surface by oxidation under the action of ozone plasma, followed by treatment with potassium hydroxide solution for several hours at room temperature; successive immersions in demineralized water and then in absolute ethanol under ultrasound and drying under an Argon flux; silanization of the outer surface by means of a solution of 3-aminopropyltriethoxysilane in ethanol for several hours; successive immersions in demineralized water and then in absolute ethanol under ultrasound and drying under an Argon flux; and heating the device for several hours and then maintained under an inert atmosphere.

According to one embodiment, the oxidation and grafting step of the hydrogel comprises the following phases: dissolution of the hydrogel in demineralised water, followed by treatment with a solution of sodium periodate whose volume is adjusted to obtain a mixture whose molar ratio between the sodium periodate and a predetermined and repeated monomer unit of the hydrogel is of the order of 50%, protection of the reaction mixture obtained against light and stirring for several hours at room temperature ambient, to obtain an oxidized hydrogel, grafting the oxidized hydrogel on the silanized outer surface and stirring at room temperature and protected from light for several hours, treatment of the assembly obtained with an aqueous solution of sodium cyanoborohydride for several hours in order to reduce formed Schiff bases, then rinsing with deionized water, with ethanol and drying under an Argon flux. According to one embodiment, the carboxylation and activation step of the grafted hydrogel comprises the following phases: treatment of the grafted hydrogel with a solution of bromoacetic acid in sodium periodate for several hours at room temperature and under Argon, purging the reaction mixture obtained and rinsing with demineralized water, ethanol and drying under an Argon flux, activation in the form of esters of the carboxylate residues formed, then rinsing with water.

According to one embodiment, the method comprises a step of dimensioning the outer surface of the photosensitive layer of the pixel-type biological analysis device to correspond to the size of a predetermined dose of a mixture. capture can be arranged on the outer surface in the form of a drop.

According to one embodiment, the outer surface of the photosensitive layer of the pixel device is sized to have an area of at least 25500 ^10-12 m ² .

According to one embodiment, the step of producing the photosensitive layer of the pixel-type biological analysis device comprises a step of forming, in the photosensitive layer, photoelectron collection means comprising a plurality of island-shaped collection zones and spaced from each other in the photosensitive layer.

According to one embodiment, the collection zones are spaced apart from each other by a distance less than a predetermined maximum distance and set according to a recombination distance of photoelectrons emitted in the photosensitive layer. According to one embodiment, the photosensitive layer is formed so as to have: a first portion capable of detecting an incident light intensity originating from a capture mixture and intended to provide a first electrical quantity and a second portion protected against the intensity incident light from the capture mixture for providing a second electrical magnitude.

According to one embodiment, the central photosensitive zone forming step comprises the steps of: forming in the central photosensitive zone a matrix of pixel-type biological analysis devices each comprising a photosensitive layer, and forming an insulating joint between adjacent pixels, the insulating joint being shaped to electrically isolating a photosensitive layer of a pixel from the photosensitive layer of the adjacent pixel.

According to one embodiment, the method comprises producing, in each pixel-type biological analysis device, an analog-digital converter for converting into digital data a characteristic value of a light intensity detected by the pixel.

Embodiments of the invention will be described in greater detail in the description which follows, given in a nonlimiting manner in relation to the appended figures in which: FIG. 1 schematically represents the structure of an embodiment of a biological analysis device according to the invention,

FIG. 2 is the electrical diagram of a biosensor comprising a plurality of devices such as that represented in FIG. 1,

FIGS. 3a, 3b, 3c are flowcharts showing successive steps of a manufacturing method of the device of FIG. 1,

FIG. 4a represents an embodiment of electronic means of a biological analysis device according to the invention,

FIG. 4b represents another embodiment of electronic means of a biological analysis device according to the invention, FIG. 5 schematically represents a differential pixel structure that can be used in a biological analysis device according to the invention,

FIGS. 6a and 6b show adaptations of a differential pixel structure including an analog / digital converter,

FIG. 6c is the electrical diagram of a biosensor comprising a plurality of differential pixels such as that represented in FIG. 6a,

FIGS. 7 and 8 show, in sectional views, embodiments of a photosensitive layer that can be used in a biological analysis device according to the invention,

FIGS. 9a, 9b, 9c, 9d show in more detail various embodiments of the differential pixel structure,

FIG. 10 represents a structure of adjacent pixels that can be used in a biosensor according to the invention, and

- Figures 11A, 11B show schematically, respectively from a top view and a sectional view, an embodiment of a biosensor according to the invention. In the following description, various specific details are given for a better understanding of embodiments of the invention. The disclosed embodiments may be implemented by dispensing with one or more of these details, or using other methods, equipment, materials, etc. In some cases, materials or operations that are well known in themselves are not described in detail to avoid obscuring certain aspects of the described embodiments. Referring to an "embodiment" in the description means that a particular feature or structure described in connection with this embodiment is included in this embodiment. Thus, the use of the terms "in one embodiment" or "in one embodiment" at various points of the description does not necessarily refer to the same embodiment. On the other hand, the particular features relating to each embodiment may be suitably combined to form one or more other embodiments. Finally, the analyzes, comments, deductions and conclusions set out in the following with respect to the prior art are deemed to be an integral part of the present invention.

FIG. 1 represents the general structure of a biological analysis device 10 and a biosensor according to one embodiment of the invention. The biological analysis device 10 is made on a semiconductor substrate 25. It comprises a layer of a capture mixture 12 comprising a certain amount of a probe protein 14 grafted to a hydrogel 16, 18 and intended to capture a certain type of target protein. The hydrogel comprises a hydrophilic medium 16 incorporating water molecules 18, such as for example Dextran (registered trademark), nitrocellulose, hyaluronic acid or a derivative of these saccharide polymers or polyamide. More generally, this hydrogel is a natural or synthetic polymer capable, according to its formulation, of retaining up to 99% of water.

Each molecule of the probe protein 14 is adapted to mate with a corresponding molecule of target protein likely to be in a substance to be analyzed. In case of pairing of the two molecules, when a labeled third molecule is added and associated with the target molecule, for example a secondary antibody, a quantity of light can be emitted, for example by chemiluminescence or fluorescence.

The biological analysis device 10 furthermore comprises a photosensitive layer 20. This photosensitive layer comprises an external surface 22 on which the capture mixture layer 12 is arranged. This photosensitive layer 20 is adapted to absorb chemiluminescence or fluorescence-emitted photons. from the capture mixture 12 and to collect photoelectrons using at least one collection area not shown in Figure 1. This is for example a photodiode. More specifically, the outer surface 22 is treated, for example by silanization, to allow the deposition of the capture mixture 12 and in particular of the hydrogel 16, 18 on the photosensitive layer 20.

Finally, the biological analysis device 10 comprises electronic reading means 24 for treating an electrical quantity supplied by the collection means of the photosensitive layer 20, for the supply of a characteristic value of a detected light intensity. by the photosensitive layer. These electronic means, represented here schematically, are implanted on the semiconductor substrate 25 and are configured to carry out a transfer of the charges generated in the photosensitive layer 20.

The biological analysis device 10 makes it possible to analyze a biological substance that may comprise a certain quantity of the target protein corresponding to the grafted probe protein 14, for example using a detection protocol for chemiluminescence. This chemiluminescence detection protocol comprises the following steps:

the biological substance may or may not be treated with a lysing agent in order to chemically decompose some of its components: this step renders certain molecules of the biological substance accessible to the molecules of the probe protein 14 and thus analyzable; the mixture obtained is then deposited on the capture mixture layer 12 of the biological analysis device 10 comprising the probe protein 14, and then incubated for a variable duration depending on the species treated,

a "secondary" antibody solution coupled to a chemiluminescence marker, for example the HorseRadish Peroxidase enzyme or HRP is added and the whole is incubated for a few minutes: this antibody is coupled specifically and with high affinity to the pairs formed by the probe protein and target protein molecules; the concentrations and incubation times can, again, vary and are optimized according to the intended application in a manner known per se,

several washes are then carried out in order to eliminate any trace of secondary antibody adsorbed nonspecifically on the surface of the biological analysis device 10 and capable of generating a false detection; the washing types (time, composition, concentration, number) are variable and can be adapted in a manner known per se according to the proteins manipulated during the detection protocol,

an activation of a chemiluminescence signal and a corresponding detection by the biological analysis device is carried out by adding a solution revealing the specific substrate of the HRP enzyme under the optimal conditions of concentration and pH for a chemiluminescence signal: this revealing solution can be obtained commercially or prepared, for example in the form of luminol; it is added over the entire surface of the biological analysis device 10 and the signal is immediately collected by the photosensitive layer 20, - a computer post-processing may possibly be necessary in order to amplify the signal obtained,

finally, the comparison of the luminous intensity value obtained by the biological analysis device with a predetermined calibration curve, or determined at the same time as the biological measurement, makes it possible to estimate the quantity of target protein present in the the biological sample.

The biological analysis device 10 can also be used to perform the analysis of a biological substance using a fluorescence detection protocol. Such a fluorescence detection protocol has the same steps as the chemiluminescence detection protocol except that the secondary antibody is coupled to a fluorophore instead of a chemiluminescence label and an excitatory light is provided to the instead of a revealing solution for the activation of a fluorescence signal. As shown in FIG. 2, several devices ₁₀ , such as the device 10 of FIG. 1, can be arranged in a matrix of pixels of a biosensor. Each device 10 ,, represents a line pixel i and column j of the matrix. In Figure 2, only the first two rows and the first three columns of a pixel array are shown, but it is easy to generalize this representation to m rows and n columns to form a complete biosensor.

The pixels 1O ₁ ,, 1O _{1 2} and 1O _{1 3} are connected to a first line bus I ₁ and the pixels 1O ₂₁ , 1O ₂₂ and 1O ₂₃ to a second line bus I ₂ for the transmission of the values of FIG. light intensity. The bus lines I ₁ and I ₂ are connected to a horizontal shift register 26. The pixels 1O _{1 1} and 1O ₂₁ are connected to a first column bus C ₁ , the pixels 1O _{1 2} and 1O ₂₂ to a second column bus C ₂ and the pixels 10 ₁₃ and 10 ₂₃ to a third column bus C ₃ for the transmission of the light intensity values. The bus columns C ₁ , C ₂ and C ₃ are in turn connected to a vertical shift register 28. The shift registers 26 and 28 are connected to a signal processing circuit 30 which may include in particular a digitizing circuit of the signals. signals.

To produce a biosensor arranged in a matrix of pixels, structures other than that which has just been described with reference to FIG. 2 are conceivable. In particular, each pixel can be provided with its own converter of analog signals into digital signals. Examples of such structures will be described later with reference to Figures 6a, 6b and 6c. In a conventional manner, the capture mixture 12 is arranged in two steps on the biosensor. In a first step, the hydrogel 16 is deposited on the entire surface of the biosensor. In a second step, a specific protein probe 14 solution is deposited locally in the form of drop. Several different probe proteins can thus be deposited on the same hydrogel substrate, drop by drop, each drop comprising a specific probe protein. It has been experimented that a drop of protein probe of a few nanoliters may be sufficient to allow a photodetection of target protein by chemiluminescence or fluorescence, which corresponds to a drop of about 150 microns (micrometers) in diameter. Below this size, ie evaporation during the deposition of the solution of protein probe becomes too important and no longer allows the maintenance of protein probes, or the amount of protein probe is not sufficient to then allow a photodetection by chemiluminescence (or fluorescence). In the prior art, given the size of the pixels of a conventional biosensor, which is well below 150 μm, a drop of specific probe protein covers a plurality of pixels of a biosensor because of their small surface area, generally ten, and the electrical signals obtained by the plurality of pixels of the biosensor must be combined by a posteriori treatment to obtain a value effectively revealing the amount of target protein in the biological samples. is complex and may be erroneous In addition, in order to prevent fluorescent or chemiluminescent light sources from interfering between two drops of neighboring specific probe protein, probe protein drops must be spaced a sufficient distance apart. In general, this distance can be up to three times the size of a drop of probe protein, ie approximately 450 μm. a significant photosensitive surface is thus lost in the biosensor.

According to a preferred embodiment of the invention, each pixel is advantageously dedicated to the detection of a particular target protein and has a photosensitive surface which has dimensions larger than those of a conventional imager, adapted to the size of a particular image. such a drop. Each pixel has, for example, a width of 150 μm or more (for a square-shaped pixel) corresponding to a surface of the order of 25500 μm ² or more (ie 25500.10 ^-12 m ² or more). As indicated above, this minimum surface corresponds to the size of a drop of probe protein below which either the evaporation during the deposition becomes too great and no longer allows the maintenance of protein probes, ie the amount of probe protein does not increase. is not sufficient to subsequently allow photodetection by chemiluminescence or fluorescence. Without these constraints, a pixel could reach much smaller dimensions, such as 20 microns x 20 microns, for example.

Thus, a drop of specific probe protein 14 is placed on the hydrogel substrate 16 above each biological analysis device, that is to say on each pixel 10 _g of the biosensor matrix so as to avoid any post-treatment as mentioned above.

The chemiluminescence or fluorescence detection protocol can be easily adapted to the biosensor shown in FIG. 2, considering each pixel of this biosensor as a biological analysis device such as the analysis device 10 previously described: in this case, each pixel comprises its own type of probe protein and the biological substance, optionally chemically decomposed, is deposited on the entire photosensitive outer surface of the biosensor. At the end of the protocol, the comparison of the light intensity values obtained for each pixel (and therefore for each type of probe protein of each pixel) with a predetermined calibration curve, or determined at the same time as the biological measurement, makes it possible to quantify a significant amount of proteins distinct from the biological sample. More precisely, for the calibration of the proteins to be quantified, a range of protein solutions of known increasing concentrations is systematically analyzed by the biosensor, under the same conditions as a real biological test. The value of the pixel for each of the analyzed concentrations is then plotted on a graph as a function of the protein concentration, or stored in memory on the biosensor. During the actual biological test, it is then sufficient to compare the obtained pixel value with the graph previously plotted, or with the value previously stored, to obtain an estimate of the concentration of the target protein in the biological sample tested. Deposit of the capture mixture

An embodiment of certain steps of a method for manufacturing a biological analysis device according to the invention will now be described with reference to FIGS. 3a, 3b and 3c. These steps are aimed at depositing a layer of the capture mixture 12 on the outer surface 22 of the photosensitive layer 20.

As shown in FIG. 3a, during a first step S100, a CMOS electronic assembly is provided comprising a photosensitive layer 20 and electronic means 24 for reading and processing an electrical quantity supplied by the photosensitive layer 20 (FIG. 1). The outer surface 22 of the photosensitive layer optionally comprises a hydrophobic coating, for example of the Si ₃ N _{4 type} which will be described later.

Then, during a step S102, the outer surface 22 of the photosensitive layer 20 of the CMOS electronic assembly is cleaned by successive rinsing with demineralised water, acetone and absolute ethanol in an ultrasonic bath .

During a step S104, the outer surface 22 is oxidized under the action of an ozone plasma, for example for 15 minutes, and then treated with a 2.2M solution of KOH in an aqueous mixture, for example an H ₂ mixture. O / EtOH (2: 3), for example for 3 hours at room temperature.

During a step S106, the CMOS electronic assembly is then immersed successively in demineralized water and then in absolute ethanol under ultrasound before being dried under an Argon flux. During a step S108, silanol residues 32 formed on the outer surface 22 are silanized with a solution of APTES in 0.4M ethanol, for example for 12 to 20 hours, at room temperature and under an inert atmosphere ( Argon).

Then the CMOS electronic assembly is again washed several times with demineralized water in an ultrasonic bath, rinsed with absolute ethanol and dried under a stream of Argon.

It is then heated, for example at 85 ° C., for several hours, for example 5 hours, and then maintained under an inert atmosphere during a step S110, to finally obtain a silanized external surface 34. As represented in FIG. 3b, a hydrogel 16, 18, for example Dextran (0.5 g, molecular weight of approximately 100 kiloDaltons), is dissolved in 50 ml of demineralised water, and then treated with a solution of 0.1M NaIO ₄ , the volume of which is adjusted in order to to obtain a mixture of molar ratio between the moles of NaIO ₄ and the moles of dextran monomer glucose of the order of 50%, during a step S112. Note in fact in FIG. 3b that Dextran is a polymer whose base unit is a derivative of glucose: it is therefore a question of adjusting the volume of NaIO ₄ 0.1 M solution so that there are two times more molecules of the base motif of Dextran than of molecules of NaIO ₄ . This results in the opening of every second base pattern in the Dextran, as illustrated at the output of step S112 in FIG. 3b. The reaction mixture is protected from light and stirred vigorously for several hours at room temperature, for example 20 hours, to obtain oxidized Dextran 36.

Then, in a step S114, the silanized outer surface 34 is treated with the freshly prepared oxidized Dextran solution 36. The reaction is stirred at room temperature and protected from light for 48 hours for example.

The reaction mixture is then purged in a step S116 and the CMOS electronic assembly treated with an aqueous solution of 0.1M NaBH ₃ CN for, for example, 3 hours in order to reduce formed Schiff bases. After aspiration of the suspension, the CMOS electronic assembly is rinsed intensively with deionized water and then with ethanol, and dried under an Argon flux, so as to obtain a grafted hydrogel 38.

As shown in Figure 3c, the grafted hydrogel 38 is then carboxylated and activated. To do this, during a step S118, it is treated with a solution of BrCH ₂ COOH 0.1 M in 2M NaOH for for example 16 hours at room temperature and under Argon. Once this oxidation is complete, the reaction mixture is purged and the sensor rinsed with deionized water and ethanol, and then dried under an Argon flow. The carboxylate residues formed are then activated in esters during a step S120, for example using an aqueous solution of EDC (0.2 M) / NHS (50 mM), for example for 15 minutes at room temperature. room. The CMOS electronic assembly is then thoroughly rinsed with water. Finally, in a last step S122, an appropriate solution of the protein of interest (protein probe) is grafted into the activated hydrogel.

To produce a complete biosensor as shown in FIG. 2, steps S100 to S120 must be performed on all the pixels of this biosensor. Finally, the last step S122 is applied differently to each pixel, for the grafting of a drop of specific probe protein on the activated hydrogel substrate above this pixel.

We will take note of the usual abbreviations used above: (CH3) 2O: Acetone C2H5OH or EtOH: Ethanol 03: Ozone

KOH: Potassium Hydroxide NaOH: Sodium Hydroxide H20: Ar Water: Argon

"APTES": 3-Aminopropyltriethoxysilane or (H2N (CH2) 3Si (OEt) 3)

NalO4: Sodium periodate

NaBH3CN: Sodium cyanoborohydride

BrCH2COOH: Bromoacetic acid EDC: 1- (3-Dimethylaminopropyl) -3-ethylcarbodiimide hydrochloride NHS: N-Hydroxysuccinimide.

We will now describe improvements of a biological analysis device and a biosensor according to the invention or a method of manufacturing such a biosensor. These improvements can be implemented separately or in combination, at the choice of the person skilled in the art depending on the desired result and economic or technical constraints that may need to be taken into account for the production of biosensors according to the invention.

Prediction of a Differential Pixel Structure In the biological analysis device 10 described above, the photosensitive layer 20 comprises for example a doped region within the semiconductor substrate 25 (FIG 1). The energy of an incident photon extracts electrons from atoms in the photosensitive layer 20, thereby generating a charge and therefore a current. The photosensitive layer 20 of the biological analysis device 10 uses an electric field at a PN junction to cause the separation of an ion and a photoelectron and to prevent recombination and loss of signal. However, these P-N junctions have a small loss current that the photosensitive layer 20 can not distinguish from a current that would actually be generated by incident light intensity, in this case chemiluminescence or fluorescence. This loss current is also present in the dark, so it is commonly called dark current. The term darkness is to be understood as a condition in which an incident light intensity is either absent or does not cause photo-generation of charges in the photosensitive layer 20 of the device This may be due to a protection of the photosensitive layer preventing the passage of incident light, or to the maintenance of the photosensitive layer 20 at a certain potential, for example an initialization potential, which prevents the accumulation of light. loads. This dark current is a limiting factor of the performance of the biological analysis device 10 or more generally of a CMOS biosensor. The dark current strongly depends on the temperature and is therefore difficult to compensate. It also varies considerably with any lack of uniformity in the doping gradients. Any improvement aimed at suppressing this dark current is particularly advantageous for the biological analysis device or the aforementioned biosensor, since the luminous intensity emitted by chemiluminescence or fluorescence is generally very low. In this type of application, it is important that the biological analysis device or the biosensor is very sensitive.

Fig. 4a is an electrical diagram showing an embodiment of the biological analysis device 10 having improved sensitivity.

The biological analysis device 10 comprises a photosensitive layer 20 comprising a first photosensitive portion 42 and a second photosensitive portion 44. The photosensitive portions 42, 44 emit signals applied to the electronic reading and processing means 24. In this exemplary embodiment, the signal supplied by the photosensitive portion 42 is applied to a negative input 46 of an operational amplifier 50 and the signal supplied by the photosensitive portion 44 is applied to a positive input 48 of the amplifier 50. Photosensitive portions are represented by photodiodes, but any other type of photosensitive sensor may be used.

The first photosensitive portion 42 is able to detect an incident light intensity from the capture mixture 12 (FIG 1) by chemiluminescence or fluorescence effect, while the second photosensitive portion 44 is protected against the incident light intensity from the mixture capture 12 or any other source of light intensity. Although the second photosensitive portion 44 is protected against any source of light intensity, it is still described as "photosensitive" in that it has the same material and electrical characteristics as the first photosensitive portion 42. was not protected from light, it would generate, as the first photosensitive portion 42, a charge in response to any incident light. For example, each of the first and second photosensitive portions may comprise photodiodes which, in one embodiment of the invention, may be of identical shape. An independent opaque protection is then formed above the second photosensitive portion 44. For the sake of clarity in the following description, the first photosensitive portion 42 will be qualified as "illuminated" photosensitive portion, while the second photosensitive portion 44 will be qualified as a photosensitive portion placed in the dark. The biological analysis device 10 thus produced can then be described as a "differential pixel" as opposed to pixels that do not have these two photosensitive portions, one of which is, by construction, protected from incident light.

The protection applied to the photosensitive portion placed in the dark 44 is obtained using mechanical means, for example by depositing a metal or other opaque substance above this photosensitive portion 44, or directly on its outer surface, either as a layer spaced from the outer surface of the photosensitive portion 44 with a suitable passivation layer. A metal layer or other spaced from the outer surface by a passivation layer is advantageous for reducing the coupling capacity between the metal shield and the photodiode of the photosensitive portion placed in the dark 44. On the other hand, this type of protection does not does not prevent the passage of a light intensity coming from the sides. Alternatively, the opaque layer can be removed and replaced by a mechanical barrier erected between the illuminated photosensitive portion 42 and the light-sensitive portion 44 in the dark, so that the capture mixture 12 is forced to remain above the illuminated photosensitive portion 42 without overflowing on the photosensitive portion placed in the dark 44. Therefore, the output of the photosensitive portion placed in the dark 44 represents only a dark current in the sense that no chemiluminescence or fluorescence reaction taking place in the capture mixture 12 generates a light intensity that can be captured by the photosensitive portion placed in the light. obscurity 44.

Alternatively, the mechanical protection may take the form of an opaque housing of the biological analysis device 10 which may have a plurality of housing portions arranged to cover a suitable surface to define the photosensitive portion placed in the dark. .

Each photosensitive portion 42, 44 of the photosensitive layer 20 is connected to a VRT initialization voltage via respective MOS switches 52 and 54. These MOS switches 52 and 54 are shown in FIG. 4 as NMOS transistors but could be made from PMOS transistors or any other type of suitable switch. An initialization switch 56 is also provided in parallel with a reference capacitance Cfb 58 also arranged in parallel between the output and the negative input 46 of the operational amplifier 50. The initialization switch 56 may be selectively controlled to discharge the operational amplifier 50.

It is important to ensure that the illuminated light-sensitive portion 42 and the dark-light-sensitive portion 44 are as similar as possible so as to generate the most similar dark currents possible. The constraints imposing this, as well as the different options for their arrangement will be detailed below.

For this, an adjustable capacitor Cs bearing the reference 59 is added between the ground and the positive input 48 of the operational amplifier 50.

In FIG. 4a, the photosensitive portions 42 and 44 are modeled using their respective intrinsic capacities Cpd1 bearing the reference 42a and Cpd2 bearing the reference 44a as well as their respective current sources bearing the references 42b and 44b. For the illuminated photosensitive portion 42, the current source is Iph + Id1, where Iph is the current generated by chemiluminescence or fluorescence effect and Id1 the dark current, through the photodiode. For the photosensitive portion placed in the dark 44, the current source comprises only Id2, that is to say the dark current that passes through it.

At the positive input 48 of the operational amplifier 50, the current and voltage values imply the following equation:

(Cpd2 + Cs) (δVINP / δt) + Id2 = 0 (Equation 1)

At the negative input 46 of the operational amplifier 50, the current and voltage values imply the following equation:

- Cpd1 δVINN / δt + Cfb (δVOUT / δt - δVINN / δt) - (Iph + IdI) = O

(equation 2)

By rearranging the previous equation, we obtain the following equation:

(Iph + Id1) + (Cpd1 + Cfb) δVINN / δt - Cfb δVOUT / δt = 0 (equation 3)

At the output of the operational amplifier 50, there is, in the general case, the following relation:

δVOUT / δt = (IPH (Cs + Cpd2) + Id1 (Cs + Cpd2) - Id2 (Cpd1 + Cfb)) / (Cfb (Cs + Cpd2)) (equation 4)

From equation 4 it follows that in the general case the effect of the dark current on the output voltage can be suppressed if the following relation is satisfied:

d1 (Cs + Cpd2) = Id2 (Cpd1 + Cfb) (Equation 5)

In particular, equation 5 can be satisfied, and the effect of the dark current canceled, if:

Id1 = Id2 and (Cs + Cpd2) = (Cpd1 + Cfb) (Equation 6) In other words, it is deduced from equations 5 and 6 that the effect of the dark current is canceled when the product enters the dark current in the illuminated light-sensitive portion 42 and the sum of the adjustable capacitance 59 and the intrinsic capacity of the dark-light-sensitive portion 44 is equal to the product between the dark current in the dark-light-sensitive portion 44 and the sum of the feedback capacitance 58 of the operational amplifier 50 and the intrinsic capacity of the illuminated photosensitive portion 42. In particular, the effect of the dark current is canceled in the output voltage, if the dark current is the same in both the photosensitive portions 42 and 44 and if the sum of the adjustable capacity 59 and the intrinsic capacitance of the dark-light-sensitive portion 44 is equal to the sum of the feedback capacitance 58 of the operational amplifier 50 and the In addition, the operational amplifier 50 may be designed to have a high input common mode rejection ratio so that the signals that appear on each output are ignored. This means that any noise on the mass is ignored and does not appear on the output.

On this basis, there are several options for designing a biological analysis device 10 corresponding to that of FIG. 4. The first is to impose identical photosensitive portions 42 and 44 and to adjust the adjustable capacity 59 so as to verify the equality (Cs + Cpd2) = (Cpd1 + Cf b). Assuming that Cpd1 = Cpd2, this amounts to setting the adjustable capacity 59 Cs to the feedback capability 58 Cfb. Several techniques for matching photosensitive portions will be discussed below.

It can be noted that the dark current is very sensitive to the profile and precise levels of doping of a substrate in which it flows, whereas in fact there is usually a difference in these parameters between adjacent pixels of a matrix. pixels or even between two photosensitive portions of the same pixel.

Another complication comes from the structural differences between the adjustable capacity Cs, on the one hand, and the feedback capability Cfb, on the other hand. The adjustable capacitor Cs has in particular a terminal connected to the ground whereas the operational amplifier 50 requires that the two terminals of the capacitor Cfd be floating. Since these structures are different, it is difficult to match them.

Equation 5 shows that, even if the currents Id1 and Id2 do not match, the effect of the dark current can always be canceled as long as the products Id1 (Cs + Cpd2) and Id2 (Cpd1 + Cfb) are equal.

FIG. 4b shows an embodiment of a biological analysis device 10 which comprises means for realizing the adjustable capacitance Cs and canceling the effect of the dark current even if the currents Id1 and Id2 are not equal. In this figure, the elements identical to those shown in Figure 4a have the same references.

The difference between the two figures lies in the configuration of the adjustable capacitor Cs, which is not represented here as a single capacitor 59, but as a plurality of capacitors 59a, 59b, 59c arranged in parallel between the mass and the input positive 48 of the operational amplifier 50. A basic positive input capacitance CsO (capacitor 59a) is connected between the ground and the positive input 48 of the operational amplifier 50. Optional positive input capacitors Cs1 (capacitor 59b) and Cs2 (capacitor 59c) are selectively connectable in parallel, using switches D1 and D2. The basic feedback capacity CsO has a lower value than Cfb. Of course, only three feedback capabilities have been shown in Figure 4b for purely illustrative and non-limiting, while a larger number may also be provided. Thus, in one embodiment, the values of Cs0, Cs1 and Cs2 are chosen so that Cs1 = Cs2 and CsO + Cs1 = Cfb.

Therefore, if the dark current Id1 is slightly higher than the dark current Id2, then to maintain the equality provided by Equation 5, Cs must be less than Cfb. This is achieved by leaving the switches D1 and D2 open, so that the adjustable capacitance Cs is provided by CsO only, which is less than Cfb.

If the two dark currents Id1 and Id2 are identical, then the switch D1 is opened and the switch D2 is closed. As a result, the resulting adjustable capacitor Cs across the operational amplifier 50 is Cs = Cs0 + Cs1 = Cfb. If the dark current Id 1 is slightly lower than the dark current Id 2, then to maintain the equality provided by Equation 5, Cs must be greater than Cfb. This is achieved by closing the switches D1 and D2, so that the adjustable capacitance Cs is provided by CsO + Cs1 + Cs2. This method can also be applied in the case where, even when the dark currents of the two photosensitive portions 42 and 44 are identical, it is desired to tolerate a small difference between the capacitance values Cfb and CsO.

In one embodiment, each capacitor of the plurality of capacitors arranged in parallel has the same common value, so that the adjustable capacitor Cs actually applied to the positive input of the operational amplifier 50 can be incremented by a constant value. . The number of capacitors in this plurality of capacitors arranged in parallel can be increased, so as to increase the resolution or amplitude of the variations that can be envisaged.

Alternatively, the capacitances of the plurality of capacitors arranged in parallel may have different values so as to increase the flexibility of the assembly. A preferred embodiment is to introduce a factor of two between each successive optional capability of the parallel arrangement, thereby limiting the number of switches required at constant resolution. By thus providing a basic positive input capacitance CsO and capacitances of 1, 2, 4, 8 and 16 times an increment Cs1, an effective adjustable capacitor Cs can be obtained whose value can vary between the basic capacitance CsO and CsO + 31 Cs1 with an increment of Cs1.

More generally, when N optional capacitors are used in parallel, to allow the electronic means 24 to adapt to manufacturing variations, generally of the order of +/- 25%, the CsO capacitor is chosen to correspond to 75% Cfb and the sum Cs1 + ... + CsN is chosen to correspond to 50% Cfb, so that the adjustable capacity Cs can effectively vary between 75% Cfb and 125% Cfb. The inequality of the dark currents Id1 and Id2 is often not known in advance and a calibration phase may be necessary to determine it. This calibration can be performed as a post-production test applied to the biological analysis device 10 and the configuration Switches may be stored in memory. On the contrary, the calibration can also be performed on the fly, just before the device is put into operation.

The output voltage VOUT can be measured by any appropriate technique, for example by means of an analog / digital converter or by using the device 10 in a lighting / frequency operation, in which the voltage VOUT is compared with a reference signal causing a reset of the device 10 when it is reached. The output voltage is then given by the frequency of the initialized pulses.

As has been seen previously, it is preferable for the suppression of the effect of the dark current that the first and second photosensitive portions 42 and 44 correspond as best as possible. But in reality, even if identical photosensitive portions are manufactured, the dark current may vary slightly from one to the other due to factors involved in their manufacture or operation. In practice, the respective dark currents of the two photosensitive portions will often be different. Indeed, strict equality can hardly be achieved. A realistic goal is therefore to try to minimize this difference.

The minimization of this difference is the object of the device shown in FIG. 4b. However, the resolution of this system is limited by the number and the value of the capacitances provided in the capacitive assembly disposed in parallel with the positive input of the operational amplifier 50. Another factor that can be played is the geometry of the capacitor. biological analysis device 10 and the geometry of a set of biological analysis devices as pixels in a biosensor. An example of pixel geometry is shown in Figure 5 in a top view.

In this figure, a photosensitive portion placed in the dark 60 is disposed in an upper left portion of the pixel. It is adjacent to an illuminated photosensitive portion 62 disposed in a top right portion of the pixel. The electronic means 64 for reading and processing this pixel are arranged in a lower portion of the pixel located under the photosensitive portions 60 and 62. For a good match of the photodiodes, the dimensions of the photosensitive portions 60 and 62 may be equal (DX1 = LX1 and DY1 = LY1) in which case RX1 = DX1 + LX1. In addition, these photosensitive portions may be square (DX1 = DY1). The size of RY1 depends on the precise composition of the electronic means 64, but it is preferable that it be equal to the dimensions DY1 and LY1, so that the pixel is generally square in shape.

On the basis of this pixel configuration, several pixels can then be arranged in different ways online and / or in a column to form a complete biosensor.

They can be arranged in line and in column, for example, by simple vertical or horizontal translation of the pixel shown in FIG.

Alternatively, they can also be arranged so that two successive rows of pixels are vertically inverted relative to each other and thus have their respective electronic means 64 or respective photosensitive portions 60, 62 in correspondence.

In addition, instead of being square, the pixels may also be rectangular, the electronic means 64 being located under the photosensitive portion placed in the darkness 60, itself located under the illuminated photosensitive portion 62.

Also, the photosensitive portions may be triangular to form a photosensitive layer 20 of generally square shape, or more complex shapes while having equal areas.

Several variants can thus be imagined by those skilled in the art and will not be detailed further.

To produce a biosensor arranged in a matrix of differential pixels, the structure that has been described with reference to FIG. 2 is conceivable. But others are also possible. In particular, each differential pixel may be provided with its own converter of analog signals into digital signals. This is also possible when the pixel is not differential.

A first example is illustrated in Figure 6a. In this figure, the differential pixel 10 comprises, in the upper part, a photosensitive portion placed in the dark 60 and an illuminated photosensitive portion 62. These two photosensitive portions are connected to a circuit amplifier 24 ', for example such as the electronic means 24 detailed in Figure 4a or 4b. This amplifier circuit outputs an analog voltage which is transmitted at the input of a converter 120 of analog signals to digital signals. The differential pixel 10 also comprises a digital decoding block 122 whose characteristics depend on the coordinates of the differential pixel 10 in the pixel array of the biosensor in question. This digital decoding block 122 monitors the state of an address bus 124 so that when a signal transiting on this address bus 124 corresponds to a predetermined value stored by the digital decoding block 122 and specific to the Differential pixel 10, block 122 transmits an activation signal Sa to converter 120 which in response provides the analog output voltage of the converted amplifier circuit to a digital value at a data transmission bus 126. A second example, illustrated in FIG. 6b and showing two adjacent pixels, represents a variant of the example of FIG. 6a. To simplify the illustration, the photosensitive portions 60 and 62 and the amplifier circuit 24 'of each differential pixel are not shown in this figure although present. In this variant, the digital decoding block 122 is replaced by a shift register 128 in each pixel, ie a shift register 128a in the pixel 10a and a shift register 128b in the pixel 10b.

Each shift register 128a, 128b receives an input signal SRI (Shift Register Input) and provides an output signal SRO (Shift Register Output). The output signal SRO of the shift register 128a forms the input signal SRI of the shift register 128b. Similarly, the output signal SRO of the shift register 128b forms the input signal of a next pixel shift register while the input signal SRI of the shift register 128a is formed by an output signal of a previous pixel shift register. The output signal of each shift register 128a, 128b is also provided to the corresponding converter 120a, 120b of the pixel 10a, 10b so as to enable this converter to transmit its digital output value to the data bus 126 at the same time. desired moment. In addition, a clock signal CLK (from English "Clock") is provided to the shift registers 128a, 128b to clock transmissions of digital pixel values to the data bus 126.

The operation of this variant comprises the following steps: firstly, each pixel of the pixel matrix is set to logic "0" (inactive) to block the output of the converters 120a, 120b,

a logic state is set to "1" (active) at the input of a first pixel 10a so as to activate the clock signal for this pixel and thereby set the signal SRO of this pixel 10a to the state "1", for the converter 120a to transmit its digital output value to the data bus 126,

the logic state of the first pixel 10a is then set to "0" (inactive) and the clock signal is again activated to activate the pixel 10b and deactivate the pixel 10a, - in the following process of reading the values In the digital matrix matrix converters, the output signal SRO of the pixel 10a remains in the state "0" and the active state on the signal SRO propagates in the matrix at each clock stroke.

The variant of FIG. 6b has the advantage of presenting identical electronic circuits on all the pixels of the biosensor, with fewer interconnections than in FIG. 6a, so that fewer connection pads are necessary to connect the biosensor to a external controller. On the other hand, the reading of the digital values of the converters offers less flexibility since it is predetermined by the wiring of the biosensor which can not, by definition, be modified during its operation. The random access made possible by the exemplary embodiment of FIG. 6b is therefore advantageous compared with the variant of FIG. 6b.

As shown in FIG. 6c, several devices 10 ,,, such as the device 10 of FIG. 6a can be arranged in a matrix of pixels of a biosensor. Each device 10 ,, represents a line pixel i and column j of the matrix. In Figure 6c, only four rows and four columns of a pixel array are shown, but it is easy to generalize this representation to m rows and n columns to form a complete biosensor. All pixels are connected to the same address bus 124 which determines the sequence in which the biosensor pixels can provide their digital output values. Likewise, all the pixels are connected to the same data transmission bus 126 which retrieves these digital output values.

Predicting a plurality of photoelectron collection areas

The differential pixel structure described with reference to FIGS. 4 to 6 is based on the assumption that the photosensitive portions are defined strictly by their boundaries, i.e., their external dimensions define the photoelectron collection zone. In this case, the requirement of correspondence between the first and second photosensitive portions means that the surfaces of these two photosensitive portions must be substantially equal.

However, a possible improvement in the structure of the photosensitive layer 20 of the biological analysis device 10, and in particular in the structure of the photoelectron collection means, makes it possible to design photosensitive portions of different surfaces while ensuring the correspondence of the photosensitive currents. generated darkness.

The principle to illustrate this structural improvement is described in the patent published under number US 6,998,659.

A first embodiment of a biological analysis device according to the principle presented in this patent is illustrated in FIG. 7. A P-epitaxial layer 66 is formed on a substrate P 68. This substrate 68 is the semiconductor substrate 25 described. higher or a doped region thereof. A photoelectron collection zone 70 is designed as an island-shaped N-well in the P-epitaxial layer 66. The collection zone 70 collects e1 e8 photoelectrons generated by incident radiation 72, including radiation. chemiluminescence or fluorescence. Electronic means 74 for reading and processing include a well 76 in which an NMOS transistor 78 is housed.

In a conventional pixel type biological analysis device, the N 70 well would extend over the entire available length between the electronic means 74 of successive pixels. But in the device of Figure 7, the well N 70 is designed as an island, that is to say, it is surrounded by P-epitaxial material not connected to the ground and very little doped in comparison with the well P 76.

The smaller size of the N 70 well means that the capacity of the photosensitive portion is relatively small. But the collection efficiency is not compromised. Indeed, the vast majority of photoelectrons, such as the photoelectrons e1, ..., e6 shown in FIG. 7, diffuse into the P-epitaxial layer 66 and are finally collected by the well N 70. The electron el can statistically to move indifferently either to the well N 70, or to the well P 76. As for the electron e8, it will certainly focus on the well P 76 and will be lost.

A second embodiment of a biological analysis device according to the principle presented in US Pat. No. 6,998,659 is illustrated in FIG. 8. In this embodiment, the electronic reading and processing means 80 comprise a thin layer 84 of P + material. extending over a large part of the device, so that the surface of the P-epitaxial layer 66 around the well N 70 is covered by this thin layer, with the exception of a narrow zone in the vicinity of the well N 70. The thin layer 84 extends from the well P 82 of the electronic means 80 and is therefore connected to the latter. The well P 82 is generally connected to the ground and therefore the thin layer 84 too. The thin layer 84 is also shallower and, at a lower potential than the N 70 well, the photoelectrons are more likely to go to the well N 70 and to be collected by it than by the thin layer 84 or well P 82. For example, the electron e7 in Figure 8 is more likely to go to the well N 70, while in Figure 7 it is as likely to go to the well N 70 only to the well P 76.

The size of a collection area such as the N 70 well is dependent on the manufacturing technology used, but can generally be as low as 1 μm x 1 μm. This minimum size is not recommended, however, because in these dimensions, the manufacturing hazards generate significant differences in size between the collection areas which generates significant differences in their characteristics. Since, in a differential pixel, the photosensitive portion placed in the dark must necessarily correspond precisely to the illuminated photosensitive portion, too small collection areas are not desirable. On the other hand, an increase in the size of the N-well increases its area and perimeter, leading to an increase in the dark current and the associated noise. Today, a good compromise size for a collection area such as the N well, given the aforementioned constraints, is a size between 5 microns and 15 microns.

An implementation of the improvement presented with reference to Figures 7 and 8 is illustrated in Figure 9a. In this figure, a differential pixel 86 comprises a first illuminated photosensitive portion 88 and a second photosensitive portion placed in the dark 90. As indicated above, it is not necessary that the dimensions of the first θt second photosensitive portions 88 and 90 be identical, it is sufficient that the dimensions of the collection areas 92 and 94 are. It is therefore possible to envisage great flexibility in the geometry of the differential pixels. On the other hand, it is necessary to take into account the distance that a photoelectron can travel in a substrate before recombining. This distance is called the recombination distance. It is determined firstly by the doping level of the substrate and then by the defects of the substrate. The higher the doping, the lower the recombination distance. This recombination distance therefore imposes a maximum distance between a collection zone and the edges of the photosensitive portion in which it is located so that all the photoelectrons can be collected before being recombined.

The limited size of the collection zones 92 and 94 therefore imposes a limited size of the photosensitive portions of the differential pixel 86. With a conventional substrate having a resistivity of 100 ohms per centimeter, the recombination distance is generally between 30 and 50 μm. However, it has been seen previously that it is advantageous to provide pixels whose photosensitive surface has dimensions adapted to the size of a probe protein drop of a few nanoliters for detection by chemiluminescence or fluorescence, that is to say that is, dimensions of the order of 150 μm. With a collection zone 92 whose size is between 5 and 15 microns, the photosensitive portion 88 can not reach these dimensions. To obtain a differential pixel 86 of larger dimensions, a solution shown in FIG. 9b consists of increasing the size of the collection zones 92 and 94. It is thus possible to reach pixel dimensions of the order of 150 μm. However, as already indicated above, increasing the size of the collection areas 92 and 94 increases the effects of the dark current and its associated noise.

To obtain a pixel of larger dimensions, it is also possible to provide a plurality of island-shaped N wells forming collection zones in the same pixel-type biological analysis device, spaced apart from one another by a distance less than the recombination distance.

In Figure 9c, a differential pixel 86 comprises two photosensitive portions 88 and 90 each having four collection areas 92 and 94 respectively. As illustrated in this figure, the geometry of the two photosensitive portions 88 and 90 is not necessarily the same. In addition, the collection zones 92 are not necessarily arranged in the same way in the illuminated photosensitive portion 88 as the collection zones 94 in the photosensitive portion placed in the dark 90. On the other hand, the four collection zones 92, on the one hand, and the four collection zones 94, on the other hand, have the same dimensions: they are for example 5 or 10 μm. Moreover, the four collection zones 92 are spaced from each other by 75 μm. It is thus possible to obtain a differential pixel 86 whose illuminated photosensitive portion 88 is a square of 150 μm on a side and whose light-sensitive portion placed in the dark 90 is a rectangle of 150 μm by 20 μm. In FIG. 9d, the differential pixel 86 comprises two photosensitive portions 88 and 90 each comprising nine collection zones 92 and 94 respectively. The dimensions of the collection zones 92 and 94 are for example 5 μm. Moreover, the nine collection zones 92 are spaced apart from each other by 50 μm, for example. It is thus possible to obtain a differential pixel 86 whose illuminated photosensitive portion 88 is a square of 150 μm on a side and whose light-sensitive portion placed in the dark 90 is a rectangle of 150 μm by 20 μm.

It is possible to generalize to pixels (not shown) comprising more collection areas (16 or more). It will be noted that since the geometries of the photosensitive portions

88 and 90 are not necessarily the same, the presence of a plurality of collection zones in FIGS. 9c and 9d makes it possible to envisage a geometry in which the photosensitive portion placed in darkness 90 surrounds the illuminated photosensitive portion 88 .

Prediction of an Insulating Joint Between Two Adjacent Pixels It has been described how the pixels of a biosensor according to the invention can be sized to match the size of a drop of probe protein. This advantageous dimensioning allows a simpler processing of the signals obtained. However, so that two drops of protein probe neighbors do not interfere by chemiluminescence or fluorescence, it is generally necessary to impose a minimum distance between two neighboring drops, generally several hundred micrometers. Another improvement that can be made to a biosensor according to the invention, shown in FIG. 10, makes it possible to considerably reduce the distance required between two neighboring drops. This figure represents two neighboring pixels 86a and 86b, sized to each receive a drop of specific probe protein. The first pixel 86a comprises an illuminated photosensitive portion 88a, here of square shape, for example 150 μm side. It further comprises a photosensitive portion placed in the dark 90a of rectangular shape, 150 μm in length and 20 μm in width. The second pixel 86b comprises an illuminated photosensitive portion 88b of 150 μm square shape. It further comprises a photosensitive portion placed in the dark 90b of rectangular shape, 150 μm in length and 20 μm in width.

The first pixel 86a is separated from the second pixel 86b by an insulating joint 92 shaped to electrically isolate the illuminated photosensitive portion 88a of this pixel from the illuminated photosensitive portion 88b of the second pixel 86b and to mechanically isolate the two illuminated light-sensitive portions, so that a source of chemiluminescence or fluorescence located in one of the two illuminated light-sensitive portions is not detectable by the other illuminated light-sensitive portion.

Thus, a photoelectron e1 emitted in the illuminated photosensitive portion 88a, on the edge thereof, and can in particular be directed towards the portion illuminated photosensitive 88b, would be absorbed by the insulating seal 92 in the event that it would actually go to the illuminated photosensitive portion 88b. Conversely, a photoelectron e2 emitted in the illuminated photosensitive portion 88b, bordering the latter, and in particular being able to point towards the illuminated photosensitive portion 88a, would be absorbed by the insulating joint 92 in the event that it actually moves towards the photosensitive portion. illuminated 88a.

Similarly, a molecule excited on the first pixel 86a diffusing a light intensity towards the second pixel 86b would see this light intensity absorbed by the insulating joint, and vice versa.

To ensure good mutual protection of the two neighboring pixels, the insulating joint may be a few micrometers thick, which is much smaller than the size of a pixel or a drop of protein probe. Thus, the presence of such an insulating seal 92 makes it possible to bring the pixels of a biosensor according to the invention substantially closer together and thus to considerably increase the number of specific probe protein drops which can be arranged on a predetermined surface of the biosensor.

The molecules or photoelectrons trapped by the insulating gasket 92 are lost for detection. However, considering the diffusion rate of the photon-emitting molecules by chemotuminescence (or fluorescence) and the thickness of the insulating joint 92, it is estimated that at least 2% in the least favorable configuration is the loss. of photoelectrons or photons generated on a pixel but not detected by it.

In addition, an additional insulating seal may also be interposed between the photosensitive portions (88a, 90a, and 88b, 90b) of each pixel 86a, 86b and their respective electronic reading and processing means. Indeed, the electronic means are not of the same nature as the photosensitive portions and may be a parasitic electron source that contribute to the effects of the dark current of the photosensitive portions. In addition, the electronic means have a higher potential than that of the photosensitive portions thus attracting the photoelectrons diffusing in the photosensitive portions. These phenomena are likely to deteriorate the detection of proteins by chemiluminescence or fluorescence. Thus, the presence of the additional insulating seal makes it possible to avoid this migration of undesirable electrons. This additional insulating gasket comprises, for example, a grounded P + / P- well and an N + / N- positive potential well. Surface treatment of the biosensor

The surface of a biosensor according to the invention is shown in FIG. 11A in plan view. The biosensor is arranged on a semiconductor wafer and has a central photosensitive zone 114 intended to be placed in contact with a biological substance, the zone 114 forming here a matrix of pixels of the type described above, and a peripheral zone 116 which surrounds the central photosensitive zone 114, the zone 116 comprising various electrical circuits (for example digital readout circuits, control logic circuit, etc.) and electrical connections between these elements (wiring). Finally, the biosensor comprises electrical contact pads 117 intended, for example, to allow the wiring of the biosensor ("bonding pads") to an interconnection support. During its manufacture or its use, the biosensor is subjected to a humid medium, in particular biological solutions or chemical reagents, or even acidic / basic or oxidizing / reducing solutions. The zone 116 is thus likely to be in contact with the humic medium, which could damage the electrical parts that it comprises.

An improvement aimed here is to protect the zone 116 of the wetlands intended to be in contact with the central photosensitive zone 114, and also to promote the deposition of the capture mixture as described above. This improvement includes the provision of the following steps or steps during the manufacture of the biosensor:

- Move the contact pads 117 away from the central photosensitive zone 114, for example at a distance of 850 μm micrometers thereof, by arranging them in a peripheral zone 118 which surrounds the peripheral zone 116. This distance avoids their contact with liquids during chemical treatments of the sensor or during the biological test;

- Cover the peripheral zone 116 with a hydrophilic coating with high roughness. This hydrophilic coating is, for example, silicon (SiO ₂ ) and form a kind of "ring", for example with a width of about 700 microns, around the central photosensitive zone 114;

- Cover the central photosensitive zone 114 with a hydrophobic coating of low roughness, for example silicon nitride (Si3N4). Preferably, the remainder of the surface of the biosensor is covered with the low-roughness hydrophobic coating, ie the peripheral zone 118, the zones 114 and 118 being able to be treated simultaneously. This coating allows a clear demarcation between the peripheral zone 116 to be protected and the central photosensitive zone 114. Once these steps are carried out, the peripheral zone 116 may be covered with a thick layer of biocompatible and non-conductive resin 119, forming a sort of barrier which completely encircles the central area, as shown in FIG. 11B by a sectional view. This resin barrier isolates from the wet environment the zone 118 receiving the contact pads 117 while also protecting the wiring zone 116, which it completely covers. Its height may be several millimeters and is thus much greater than the thickness of the biosensor, for example 30 to 100 μm.

Due to the hydrophilic variation between the zones 116 and 114, 118, the protective resin 119 tends, before curing by polymerization, to automatically distribute itself around the central photosensitive zone 114, without encroaching on it or Even in this way, the hydrophobic surface treatment has a repellent effect on the resin and limits its spreading until it is The hydrophobic coating on the central photosensitive zone 114 further makes it possible to ensure that solutions deposited on the zone 114 during the biological preparation steps do not extend beyond the pixel matrix by chemically attacking the particles. the elements on the inner edge of the zone 116. Such elements situated on the inner edge of the zone 116, for example aluminum conductors, could indeed be mopely covered by the resin 119 and thus be imperfectly protected.

In particular, the process for forming the hydrogel as described above comprises the step S104 of depositing a potassium hydroxide (KOH) solution that is capable of chemically attacking elements 009/000766

34

fragile at the edge of the zone 116. The solution is deposited on the pixel matrix in a determined dosage. The hydrophobic coating prevents the dose of potassium hydroxide solution from spreading beyond the central photosensitive zone 114 under the effect of surface tension forces. As a numerical example, a roughness of the order of 50 nm was measured on a coating of silicon nitride (Si3N4) using conventional measurement techniques of the roughness depth. The silicon nitride coating was deposited using a conventional manufacturing process carried out in the applicant's production units and achieved the desired result by limiting the spread of the potassium hydroxide solution. Those skilled in the art may, however, determine other coatings or even surface treatment processes making it possible to render the surface of the central photosensitive zone 11 hydrophobic. Thus, the combination of these characteristics allows optimal arrangement and adhesion of the surface. Protective resin 119 both by the chemical characteristics of the treated surfaces and by the arrangement of the biosensor components and their spacing.

It will be clear to those skilled in the art that a biological analysis device and a biosensor according to the invention, to which the above-mentioned improvements are optionally added, make it possible in particular to use the CMOS imager technology for chemiluminescence analysis. or by fluorescence of biological substances comprising fragile target proteins. It will also be apparent to those skilled in the art that the various embodiments as well as the various improvements described above can be combined to achieve other embodiments. Moreover, various modifications may be made by those skilled in the art to the embodiments described above, in the light of the teaching that has just been disclosed. In the following claims, the terms used are not to be construed as limiting the claims to the embodiments set forth in this description, but should be interpreted to include all the equivalents that the claims are intended to cover because of their formulation and whose forecasting is within the reach of those skilled in the art by applying his general knowledge to the implementation of the teaching that has just been disclosed to him.

Claims

1. A method of manufacturing a biosensor on a semiconductor wafer, characterized in that it comprises the steps of: - producing a central photosensitive zone (116) comprising at least one pixel-type biological analysis device (10) ,,) comprising a photosensitive layer (20),

- producing a first peripheral zone (116) surrounding the central photosensitive zone, comprising electronic circuits, - covering the first peripheral zone (116) with a hydrophilic coating,

covering the central photosensitive zone with a hydrophobic coating whose roughness is chosen so that a dose of a solution that can subsequently be deposited on the central photosensitive zone does not extend beyond it until reach the first peripheral zone, and

forming a barrier of a biocompatible resin (119) on the second peripheral zone (118), the resin being fixed naturally, before polymerization, on the hydrophilic coating without extending over the hydrophobic coating.

The method of claim 1, further comprising the steps of:

performing, in a second peripheral zone (118) surrounding the first peripheral zone (116), electric contact pads (117) connected to the electronic circuits of the first peripheral zone, and

- Cover the second peripheral zone (118) by means of the hydrophobic coating.

3. Method according to one of claims 1 and 2, wherein the hydrophilic coating is silicon oxide.

4. Method according to one of claims 1 to 3, wherein the hydrophobic coating is silicon nitride.

5. Method according to one of claims 1 to 4, comprising a step of depositing, on the surface (22) of the photosensitive layer (20) of the pixel-type biological analysis device coated with the hydrophobic coating, a mixture capture medium (12) comprising a probe protein (14) grafted to a hydrogel (16, 18) for the capture of target proteins.

The method of claim 5, wherein the step of depositing the capture mixture comprises steps of:

- preparing (S100, S102, S104, S106, S108, S110) the outer surface (22) of the photosensitive layer (20) for receiving the capture mixture, comprising a step of silanizing the outer surface,

- oxidation (S112) of the hydrogel and grafting (S114, S116) of the oxidized hydrogel on the silanized outer surface,

- carboxylation (S118) and activation (S120) of the grafted hydrogel, and - grafting (S122) of the probe protein in the activated hydrogel.

7. The method of claim 6, wherein the step of preparing the outer surface by silanization comprises the following phases: cleaning (S102) of the outer surface by successive rinsing with demineralized water, with acetone and with absolute ethanol in an ultrasonic bath,

- alkaline oxidation (S104) of the external surface by oxidation under the action of ozone plasma, then treatment with a potassium hydroxide solution for several hours at room temperature, - successive immersions (S106) in water demineralized then in absolute ethanol under ultrasound and drying under an Argon flux,

silanization (S108) of the external surface by means of a solution of 3-aminopropyltriethoxysilane in ethanol for several hours,

successive immersions (S108) in demineralised water and then in absolute ethanol under ultrasound and drying under argon flow,

heating the device (S110) for several hours and then maintaining an inert atmosphere.

8. Method according to one of claims 6 and 7, wherein the step of oxidation and grafting of the hydrogel comprises the following phases: dissolution (S112) of the hydrogel in demineralised water, followed by treatment with a solution of sodium periodate, the volume of which is adjusted in order to obtain a mixture whose molar ratio between the sodium periodate and a predetermined monomeric unit and repeated the hydrogel is of the order of 50%,

- protection (S112) of the reaction mixture obtained against the light and stirring for several hours at room temperature, to obtain an oxidized hydrogel,

grafting (S114) of the oxidized hydrogel onto the silanized outer surface and stirring at room temperature and away from light for several hours,

treatment (S116) of the assembly obtained with an aqueous solution of sodium cyanoborohydride for several hours in order to reduce formed Schiff bases, then rinsing with deionized water, with ethanol and drying under an Argon flux.

9. Method according to one of claims 6 to 8, wherein the step of carboxylation and activation of the grafted hydrogel comprises the following phases: - treatment (S118) of the hydrogel grafted with an acid solution bromoacetic acid in sodium periodate for several hours at room temperature and under Argon,

purge (S118) of the reaction mixture obtained, followed by rinsing with demineralized water, ethanol and drying under an Argon flux; - activation (S120) in the form of esters of the carboxylate residues formed, followed by rinsing with water .

The method according to one of claims 1 to 9, comprising a step of dimensioning the outer surface (22) of the photosensitive layer of the pixel-type biological analysis device to correspond to the size of the pixel. a predetermined dose of a capture mixture which can be disposed on the outer surface in the form of a drop.

The method of claim 10, wherein the outer surface (22) of the photosensitive layer of the pixel device is sized to have an area of at least 25500 ^10-12 m ² .

12. Method according to one of claims 1 to 11, wherein the step of producing the photosensitive layer (20) of the pixel-type biological analysis device comprises a step of forming, in the photosensitive layer, means for collecting photoelectrons having a plurality of island-shaped collecting regions (70; 92) spaced from each other in the photosensitive layer (20).

The method of claim 12, wherein the collection areas (70; 92) are spaced from each other by a distance less than a predetermined maximum distance and set according to a recombination distance of photoelectrons emitted in the region. photosensitive layer.

14. Method according to one of claims 1 to 13, wherein the photosensitive layer is made to have -. a first portion (42) capable of detecting an incident light intensity originating from a capture mixture (12) and intended to provide a first electrical quantity (Iph + ld1) and

- a second portion (44) protected against incident light intensity from the capture mixture (12) for providing a second electrical magnitude (Id2).

The method according to one of claims 1 to 14, wherein the central photosensitive zone forming step (116) comprises the steps of: forming in the central photosensitive zone a matrix of biological analysis devices of the type pixel (10 ,,) each comprising a photosensitive layer, and

forming an insulating joint (92) between adjacent pixels (86a, 86b), the insulating joint (92) being shaped to electrically isolate a layer photosensitive (88a, 90a, 88b, 90b) of a pixel of the photosensitive layer of the adjacent pixel.

The method according to claim 15, comprising performing, in each pixel-type biological analysis device (10 _M ), an analog-to-digital converter (120) for converting to digital data a characteristic value of a light intensity detected by the pixel.