EP1668654B1 - Independently addressable resistance matrixes and method for the production thereof - Google Patents

Abstract

The arrays of independently-addressable resistors are commonly used to control miniature elements. The invention proposes solving the problem caused by the loss of power dissipated in the addressed resistor by choosing, for this resistor, a material with a negative thermal coefficient resistance, which enables the addressing output of this resistor to be increased.

Description

TECHNICAL AREA

The invention relates to matrices of passive components, more particularly to resistors connected together by lines and columns, and to their manufacture. These matrices of resistors can be used in various fields, in particular to activate components by Joule effect.

STATE OF THE PRIOR ART

In order to gain space and control density, resistance matrices have been developed where a large number of resistive elements are condensed on a small surface, while being individually activatable.

EP-A-0 813 088 discloses an optical relay which is activated by thermal stimulation. The relay consists of a matrix of resistors. In EP-A-1 188 840 a matrix of selectively operable heating elements is described.

As shown in FIG. 1, a resistance matrix comprises N command lines (indexed N _i , with i strictly positive integer), M control columns (indexed M _j , with j strictly positive integer), and NM resistors (indexed R _ij , each resistor R _ij being controlled by the line N _i and the column M _j ). To control a resistor, the "switches" of its line and column are "closed": for example, the voltage "+ V" can be applied to the line N _i and "0" to the column M _j ; the resistance R _ij is then "addressed", that is to say subjected to a current, unlike the others.

Whatever the use of these matrices, one of the challenges is to precisely locate the control power on a determined resistance in order to achieve the effect expected by the control, while limiting the power dissipated in the other elements of the matrix. , in particular the resistors, because of induced or derived currents, both to increase the power in the resistor addressed and for the control to remain specific.

Indeed, the maximum power is dissipated in the resistor addressed. However, there are also other non-zero currents flowing in the rows and columns, as well as in the other resistors, which also induce power losses in and through these elements. This results in the control power being not completely dissipated in the addressed resistor (loss of efficiency) and that the unaddressed resistors also dissipate an undesirable power (loss of sensitivity). Simulations have shown that for a matrix of 150 points for example, about 15% of the power is dissipated at the addressed point, while the other points where the dissipated power is the highest give off about 5% of the power.

One of the known ways to remedy these effects is to couple each resistor to a diode or a switch to block the current in the unaddressed resistors. However, this solution is very cumbersome because it involves doubling each resistor, resulting in detrimental manufacturing costs and loss of compactness.

Another technique would be to segment the matrix into subunits such that the power loss is reduced, which reduces the number of diodes. This solution does not eliminate the complexity problems inherent to the diodes, nor the residual parasitic heating in each of the matrices.

Another alternative is to control each row and column with voltages that are adjusted and servocontrolled by a control system. By this means, it is possible to precisely control the residual power in the unaddressed resistors and modify the parameters. If this solution is efficient, it is clear that it requires a control system expensive and complex to implement.

STATEMENT OF THE INVENTION

The object of the invention is to propose a simple solution, which avoids the drawbacks inherent in existing solutions, for the production of a matrix of resistors making it possible to locate power on one of the resistors of the matrix by limiting the dissipated power. in the rest of the matrix. Thermally, this resistance activates an associated component.

More particularly, one of the aspects of the invention concerns the choice of the thermal properties of at least one resistor, in order to increase its addressing efficiency, that is to say the power dissipated by this resistance compared to the total power dissipated, power to thermally activate an associated component. This resistance (or these resistances) is thus chosen to negative temperature coefficient, that is to say that the value of the resistance decreases with its temperature. During its use, by the release of power, the temperature of the resistant element increases; according to the invention, the value of its resistance will then decrease, and therefore its power increase at constant voltage during heating. The accuracy of the activation of associated components is thus increased.

The invention thus relates to a resistance matrix of which one of the resistors has a negative temperature coefficient and is associated with a thermally activatable component. Advantageously, these negative temperature coefficient resistors consist of a single material having this property, which simplifies all the manufacturing process.

A preferred embodiment relates to a matrix whose all resistors are negative temperature coefficient, and in particular identical. Indeed, whatever the matrix, the power released in the unaddressed resistors is lower than the power dissipated at the addressed point. The temperature of the resistor addressed therefore increases faster than the temperature of the rest of the circuit: even if all the resistors are at a negative temperature coefficient, or even identical, the value of the unaddressed resistances will decrease less rapidly over time than that of the resistance addressed. A phenomenon of increase in power released by the unaddressed resistances occurs, but less than the increase in the power dissipated by the resistance addressed. There is therefore also in this case a gain in yield compared to that achieved in a conventional matrix.

Advantageously, the material used for some or all of the rows and columns has a positive temperature coefficient, which leads to an increase in the resistance of these elements and therefore a decrease in lost power.

Several resistors of the matrix according to the invention, see all, can be coupled to components to activate them. The invention also relates to a device using this matrix, such as a biochip or a reaction card.

Advantageously, to optimize its efficiency, it is possible to adjust, for example by a programmable pulse generator, the time of application of the control voltage on a resistor.

The invention also relates to the method of manufacturing a resistor matrix whose resistance, associated with a thermally activatable component, is formed of a material put in place, for example by depositing, on a substrate, the material having a resistance to negative temperature coefficient.

BRIEF DESCRIPTION OF THE DRAWINGS

• FIG. 1 : schéma d'une matrice de résistances, avec indication d'un courant induit.
• FIG.2: évolution par rapport au temps de différents paramètres en cours d'utilisation d'une matrice de résistances à coefficient de température positif (FIG.2a) et d'une matrice de résistances à coefficient de température négatif (FIG.2b).
• FIG.3: synopsis d'un exemple de fabrication d'une matrice préférée selon l'invention.

The invention will be better understood by means of the following figures, which serve only to illustrate the invention and are not restrictive:

• FIG. 1: diagram of a resistance matrix, with indication of an induced current.
• FIG. 2: evolution with respect to the time of various parameters in use of a matrix of resistances with positive temperature coefficient (FIG. 2a) and of a matrix of resistances with negative coefficient of temperature (FIG. 2b) .
• FIG. 3: a synopsis of an example of manufacturing a preferred matrix according to the invention.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

As described previously, FIG. 1 represents a conventional matrix of separately addressable resistors comprising N rows, M columns and NM resistors. These resistors can be controlled either simultaneously, or successively, or again according to a combination of these two modes.

avec U tension aux bornes. La puissance P_ij peut notamment être utilisée pour activer thermiquement un composant associé à la résistance R_ij.The resistance R _ij is addressed, and dissipates a power ${\mathrm{P}}_{\mathrm{ij}}\mathrm{:}{\mathrm{P}}_{\mathrm{ij}}\mathrm{=}\frac{{\mathrm{U}}^{\mathrm{2}}}{{\mathrm{R}}_{\mathrm{ij}}}\mathrm{,}$

with U terminal voltage. The power P _ij may in particular be used to thermally activate a component associated with the resistor R _ij .

The efficiency Q _ij of the resistor R _ij addressed is equal to the power P _ij referred to the total power released. However, the other elements of the matrix also react to the addressing voltage: an example of induced current is thus represented in dotted lines, which causes in this configuration a power release in particular by the resistors R _{i + 1 j} , R _{i + 1 j + 1} , R _{i j + 1} , R _{i j + 2} , as well as by the segments of lines and columns separating them. These parameters are to be taken into account for performance evaluation.

Moreover, any dissipation of power is accompanied by a heating of the resistance concerned and an increase in temperature. The temperature of the addressed resistance increases more and faster than that of the other elements.

However, conventional materials for making resistors have their resistance increase as the temperature increases: see the curve R _ij in FIG. 2a. The power dissipated (curve P _ij ) by the resistance R _ij will therefore decrease over time, and more rapidly than the power generated by the other resistors, whose temperature and resistance (curve R _na ) increase less rapidly. The efficiency of the resistor R _ij addressed therefore decreases as it is activated (curve Q _ij ), and the increase in temperature, which is the desired objective in the context of the control matrices for heating by Joule effect of elements, slows down.

In the context of the invention, is used to manufacture the resistance R _ij a material whose resistance decreases with temperature, that is to say a negative temperature coefficient resistance, or NTCR (" Negative Thermal Coefficient Resistance " ). This material may be one of the components of the resistor or the resistor may be made entirely of such material. Examples are the Nitride of Tantalum, Nickel-Chrome alloys, or nitrides of refractory materials. The temperature coefficient (TCR) can be adjusted either by the combination of materials or by the parameters chosen during the manufacture of the resistor. Depending on the needs, the NTCR can thus vary from -100 to -3000 ppm / ° C.

In this case of a matrix of NTCR illustrated in FIG. 2b, over time, the dissipation of energy by the resistor addressed R _ij increases as well as its temperature, its resistance (curve R _ij ) decreases, and therefore its power dissipated (curve P _ij ) increases all the more.

It can also be seen in FIG. 2b that, in this case where all the resistors are NTCRs, the other negative temperature coefficient resistors that are not addressed also have their resistance decrease (curve R _na ), but to a lesser extent because their temperature evolves less quickly, the power released by them remaining lower than the power dissipated P _ij . The output of the addressed resistor (curve Q _ij ) therefore increases.

A combination of the two examples is possible, where the resistance addressed R _ij is negative temperature coefficient, and the other R _na positive temperature coefficient: it would be found that the efficiency Q _ij of the addressed point increases all the more (no illustrated), and in particular in even greater proportions than in the case of a fully NTCR matrix. Other combinations are possible, for example with a line and / or a column NTCR only.

Moreover, the resistor R _ij is addressed by a control power which determines the voltage U at the terminals and the power P _ij dissipated by this resistor.

A modulation factor of P _ij other than the value of each resistor is therefore the power "really" addressed to R _ij . This power is lower than the initial control power, with partial losses in the other resistors as described above, but also losses related to the intrinsic resistance of the rows and columns.

It may therefore be advantageous to use a positive TCR material, such as aluminum or copper, for these lines and columns: by thermal conduction from the heated resistor, the material used in the rows and columns is capable of heating. Thanks to the use of a positive TCR material for these lines and columns, the resistance of the lines and columns will then increase, and the power lost in them will decrease, increasing the power addressed, and thereby even the efficiency of the resistance addressed.

The power addressed, and therefore the voltage across the resistor addressed, can also be modulated during use by adjusting the duration of application of this voltage. This last time parameter makes it possible to optimize the desired efficiency for each resistor R _ij addressed, and the desired temperature to activate the component concerned by this resistor. Indeed, the process allowing heating by Joule effect is a dynamic phenomenon. Thus, the application of a voltage for a short duration, for example 0.2 s, will make it possible to obtain moderate temperature increases, of the order of 100 ° C., and the application of the control for a period of time. longer, for example 10 s, will result in higher temperatures, of the order of 500 ° C (see Figure 2b). By way of example, there is shown in FIG. 1 a pulse generator (1) connected to the rows and columns, which makes it possible to apply voltages determined in amplitude and duration across said lines (N) and columns.

Example 1

Either a network of 144 resistors addressed by 12 lines and 12 columns, with heating resistors to be addressed of 1000 ohms and a resistance inter lines and inter columns of 1 ohm, that is to say an intrinsic resistance of 1 ohm of each line and / or interconnect column.

By simulation, it has been found that for resistances with a zero temperature coefficient, the power dissipated at the addressed point is 15% of the total power dissipated in the network, and that the maximum power generated by the other resistors is 4.5. %.

If the resistors have a TCR of -2500 ppm / ° C, when the temperature of the resistor addressed reaches 300 ° C, the other resistances have a maximum of 100 ° C, and the power dissipated by the resistor addressed reaches 40% of the resistance. total power instead of 15%, that is, more than doubling.

The matrix according to the invention therefore makes it possible to obtain very high temperatures, of 500 ° C. and more, at very localized points, for matrices which make it possible to address many points (50 to 1000 and more), and this fast way. An adjustment of the maximum power required is possible by checking the value of the resistance TCR. These effects are furthermore possible without a diode or switch device to weigh down the system, and the matrix can be made on different types of substrates, by means of manufacturing methods avoiding heavy technologies.

Indeed, to achieve a matrix according to the invention, standard technologies of microelectronics, including deposition and photolithography, are used in a preferred manner. However, any other technique that can be used for the manufacture of microsystems is conceivable: screen printing of glues, adhesives, conductive or non-conductive polymers, screen printing pastes, inkjet technology, etc.

FIG. 3 represents an example of a manufacturing method: a substrate (10) such as silicon is chosen. An aluminum layer (12) is deposited by sputtering (FIG. 3a). Photolithography and chemical etching make it possible to obtain line patterns (14) (FIG. 3b). A layer of NTCR resistive material (16) is deposited by sputtering (FIG.3c); the resistive patterns (18) are obtained by photolithography and etching (FIG. 3d). A dielectric layer (20) is then deposited to isolate rows (14) and columns (FIG. 3e), with photolithography of the contact resumption patterns (22) on the columns (FIG. 3f). Finally, an aluminum layer (12) is deposited by cathodic sputtering (FIG. 3g), the column patterns (24) being made by photolithography and etching (FIG. 3h). The thermally activatable compounds are combined according to known techniques.

Typically, the aluminum layer (12) has a thickness of 500 to 50000 Å (50 to 5000 nm), preferably 500 nm; the thickness of NTCR (16) is typically between 500 to 5000 Å (50 to 500 nm), preferably 100 nm. The NTCR can be adjusted preferably between -100 and -3000 ppm / ° C depending on the deposition conditions and the desired use parameters.

As the dielectric insulator (20), it is possible to use a polymer or a mineral such as SiO ₂ or Si ₃ N ₄ . The substrate (10) is insulating and comprises, for example, silicon, a polymer, a glass, a ceramic, etc., or a combination of these materials.

Application

The matrices according to the invention find their application in many fields, such as, for example, biology, imaging or flat screens, where the control systems must be miniaturized. More particularly, the matrices according to the invention can be used to manufacture biochips or " Lab On Chips ", also called reaction cards. Such a reaction map is known for example from WO 02/18823. In general, we will call later device for use biological any structure suitable for use in applications in biology, such as reaction maps or biochips.

However, to produce such devices for biological use, a microfluidic network is integrated on the support card of the device: the liquid to be analyzed must flow for example between the different reagents. In order to circulate a liquid in a network of micro-channels, micro-valves are actuated.

Micro valves have been developed for applications in microsystems, biochips and reaction maps. An example is given in document FR-A-2 828 244, which relates to micro-valves actuated by pyrotechnic effect. The startup of the micro-valves requires localized heating below the microsystem, for example by heating a resistance under each micro-valve which will then be actuated by Joule effect.

For this preferred application, the network of micro-valves must be consistent, with a high density of these components to activate: for example 50 to 1000 micro valves on a surface typically of the order of the size of a credit card must be addressed separately. The use of matrices of resistors therefore seems appropriate.

The matrices according to the invention add the advantage of optimizing the efficiency of each addressing, and therefore a better efficiency and specificity of the analyzes performed.

Claims (22)

1. Resistor array comprising N lines of commands N_i, with i being a strictly positive integer, M columns of commands M_j, with j being a strictly positive integer, and NM resistors R_ij, each resistor R_ij being commanded by the line N_i and the column M_j, characterised in that at least one of the resistors has a negative thermal coefficient resistance and is associated with a thermally activatable component.
2. Array according to claim 1, characterised in that each resistor R_ij is associated with a thermally activatable component.
3. Array according to one of claims 1 or 2, wherein at least one of the activatable components is a microvalve.
4. Array according to one of claims 1 to 3, wherein all of the resistors R_ij have negative thermal coefficient resistances.
5. Array according to one of claims 1 to 4, characterised in that at least one of the negative thermal coefficient resistors is made of a single material.
6. Array according to claim 4, characterised in that all of the negative thermal coefficient resistors are made of a single material.
7. Array according to one of claims 1 to 6, characterised in that all of the resistors are identical.
8. Array according to one of the previous claims, wherein the negative thermal coefficient resistor includes tantalum nitride, a nickel-chromium alloy, or a nitride from refractory material.
9. Array according to one of the previous claims, wherein the negative thermal coefficient resistor has a temperature coefficient of between -100 and -3000 ppm/°C.
10. Array according to any one of claims 1 to 9, characterised in that the material used for at least one line and/or at least one column has a positive thermal coefficient resistance.
11. Array according to claim 10, characterised in that all of the lines and/or all of the columns are made of a material with a positive thermal coefficient resistance.
12. Array according to one of claims 1 to 11, characterised in that all of the lines and all of the columns are made of the same material.
13. Array according to one of claims 1 to 12, which is associated with an insulating substrate.
14. Array according to one of the previous claims, also including means for adjusting the time for which a command voltage is applied to at least one of the resistors R_ij, in particular to each resistor R_ij, so as to obtain the desired output.
15. Method for producing a resistor array, wherein at least one of the resistors is obtained by placing a resistive material (16), of which the resistance has a negative thermal coefficient, on a substrate (10) and including the association of this resistor with a thermally-activatable component.
16. Production method according to claim 15, including the deposition of the resistive material by cathode sputtering.
17. Production method according to one of claims 15 or 16, including the deposition of a conductive material (12) on the substrate (10) so as to form lines (14) before the resistive material is deposited.
18. Production method according to one of claims 15 to 17, including the deposition of a conductive material (12) so as to form columns (24) after the resistive material has been deposited.
19. Method according to one of claims 15 to 18, including a step of depositing a material (20) insulating the lines from the columns on said substrate.
20. Method according to one of claims 17 to 19, including the choice of a material of which the resistance has a positive thermal coefficient for the lines and/or columns.
21. Method according to one of claims 15 to 20, including the association of the array with a microvalve array.
22. Device for biological use, including an array according to one of claims 1 to 14, associated with a microfluidic array.

Images